Hígado Páncreas y Transplante Ahpba Capitulo Pancreas

BEAUX BOOKS PUBLISHING
P.O. BOX 53364

SHREVEPORT, LA 71135
Hepato-Pancreato-Biliary and Transplant Surgery: Practical Management of Dilemmas

Copyright © 2018 by Beaux Books Publishing ISBN: 978-0-9977721-9-7
All rights reserved. No part of this publication may be reproduced or transmitted in any form or by any
means, electronic or mechanical, including photocopying, recording, or any information storage and
retrieval system, without permission in writing from the publisher. Requests can be sent to Beaux Books
Publishing, c/o Trina Chu, P.O. Box 53364, Shreveport, Louisiana 71135. This book and the individual
contributions contained in it are protected under copyright by the Publisher (other than as may be noted
herein).
NOTICES
Knowledge and best practice in the field are constantly changing. As new research and
experience broaden our understanding, changes in research methods, professional practices, or
medical treatment may become necessary.
This work is provided “as is” and contains information obtained from authentic and highly
reliable sources. While all reasonable efforts have been made to publish reliable and up to date data
and information, neither the Publisher nor the editors, authors, contributors, the Americas-Hepato-
Pancreato- Biliary Association (AHPBA) can accept any legal responsibility or liability for any
errors or omissions that may be made. The Publisher wishes to make clear that any views or
opinions expressed in this book by individual editors, authors, contributors, or AHPBA are personal
to them and do not necessarily reflect the views/opinions of the Publisher. The editors, authors,
contributors, AHPBA, and the Publisher disclaim any and all warranties, express or implied,
including any warranties as to accuracy, comprehensiveness, or currency of the content of this
work.
Practitioners and researchers must always rely on their own experience and knowledge in
evaluating and using any information, methods, compounds, or experiments described herein. In
using such information or methods, they should be mindful of their own safety and the safety of
others, including parties for whom they have a professional responsibility. This work is no substitute
for individual patient assessment based upon healthcare professionals’ examination of each patient
and consideration of, among other things, age, weight, gender, current or prior medical conditions,
medication history, laboratory data and other factors unique to the patient. The Publisher does not
provide medical advice or guidance and this work is merely a reference tool.
To the fullest extent of the law, neither the Publisher nor the editors, authors, contributors, or
AHPBA assume any liability for any injury and/or damage to persons or property as a matter of
products liability, negligence or otherwise, or from any use or operation of any methods, products,
instructions, or ideas contained in the material herein. Healthcare professionals, and not the
Publisher, are solely responsible for the use of this work including all medical judgments and for any
resulting diagnosis and treatments.
Given continuous, rapid advances in medical science and health information, independent
profession verification of medical diagnoses, indications, appropriate pharmaceutical selections and
dosages, and treatment options should be made and healthcare professionals should consult a
variety of sources. When prescribing medication, healthcare professionals are advised to consult the
product information sheet (the manufacturer’s package insert) accompanying each drug to verify,
among other things, conditions of use, warnings and side effects and identify any changes in dosage
schedule or contradictions, particularly if the medication to be administered is new, infrequently used
or has a narrow therapeutic range. The reader is strongly urged to consult the drug companies’
printed instructions, and their websites before administering any of the drugs or selecting any
procedures recommended in this book. This book does not indicate whether a particular evaluation
and/or treatment is appropriate or suitable for a particular individual. Ultimately, it is the sole
responsibility of the medical professional to make his/her own professional judgements, so as to
advise and treat patients appropriately.
The Publisher, editors, authors, contributors, and AHPBA have also attempted to trace the
copyright holders of all material reproduced in this publication and apologize to copyright holders if
permission to publish in this form has not been obtained. If any copyright material has not been
acknowledged, please write and let us know so that we may rectify in any future reprint.
Except as permitted under U.S. Copyright Law, no part of this book may be reprinted,
reproduced, transmitted, or utilized in any form by any electronic, mechanical, or other means, now
known or hereafter invented, including photocopying, microfilming, and recording, or in any
information storage or retrieval system, without written permission from the Publisher.
Trademark Notice: Product or corporate names may be trademarks or registered trademarks,
and are used only for identification and explanation without intent to infringe.
Administrative Assistants: Marjorie Malia, Beverly Wright, Ashley Abrams, and Yvette Sanchez
Cover Design: Jeremy Ricci
Format Designers: Jeremy Ricci, Thuy-Tien Chu
Proofreaders: Barbara Pratt, Dr. Phillip Estes, Nina Tindall
Legal Assistant: Trina Chu, Esq
Artists: Yen Chu, Iwan Yasin Agustin, Kaylee Messina
Printed in The United States of America
Quyen D. Chu, MD, MBA
To my parents, Trinh V. Chu and Nhan T. Chu, who risked everything,
crossing a vast and perilous ocean so that we children could have a future in
America. To my wife, Trina Chu, and my two children, Thuy-Tien and Yen,
who endured countless days and nights without having their husband and
father at home. To my mentors and patients who have taught me not only the
art and science of surgery, but also the humility of knowing that there is much
more that I do not know than what I do know. To my alma mater, Dartmouth
College, who took a chance on me and gave me this wonderful life. Finally,
to this great nation of ours that has given me the opportunity to rise from a
world of poverty to one of limitless possibilities.
Charles M. Vollmer, MD
To my wife, Beth, and three children, Andrew, Julia, and Eric, for their
steadfast patience, understanding, and support. Also, to my parents, Charles
and Judith Vollmer, for providing me all the opportunities to become the
surgeon I always wanted to be. It is my sincere hope that this endeavor will
improve decision-making and outcomes for the next generation of HPB
surgeons.
Gazi B. Zibari, MD
To my parents, the late Baderkhan S. Zibari and Khanzad K. Zibari, who are
up in heaven. To my caring and passionate wife, Maysoon M. Zibari, who
single-handedly raised our children while bravely beating breast cancer. To
our four children, Susan, Renas, Rona, and Lehat. To my siblings, especially
my oldest brother, General Zibari, who was the father figure for all of us,
helping guide us through difficult life-and-death decisions. To my mentors,
colleagues, students, and organ donors and their families. To the AHPBA
Foundation Board and the leadership and management team for entrusting us
with this worthy project. To all authors for their outstanding contributions. To
my co-editors for their steadfast support, especially to Dr. Chu for
spearheading this noble work. To the victims of the Kurdistan of Iraq who
had to endure the brutal Anfal Campaign and the Halabja chemical attack. To
the Kurdish Peshmerga who, on the behalf of the world, fought bravely
against ISIS. Finally, to the United States of America, the best country in the
world, for allowing me to take refuge and empowering me to achieve the
American Dream.
Susan L. Orloff, MD
To all of my patients, who have given me the privilege and honor of caring
for them and who have provided me with true meaning in my life as a
physician and surgeon. To my parents, Ann and Marshall, both physicians,
who guided my five siblings and me on a path of embracing education and
humanitarianism, and gave all of us opportunities towards this endeavor. To
my husband, Bob, my son, Jackson, my daughter, Annie, and our various and
sundry animals that create our menagerie. To Marjie Malia, Director of
Association Management for the AHPBA, and Katherine Rose Franklin, my
Administrative Coordinator, who were incredibly helpful in navigating the
difficult task of ”herding cats” towards the goal of making this textbook
happen. To Jeremy Ricci, for his amazing and beautiful formatting, and to
Iwan Agustin, for his creative artistic talents. Finally, to all of the awesome
authors who participated in the content of this textbook that has led to the
creation of a novel and, hopefully, very impactful textbook. To the
readership, please read, learn, enjoy, and become invigorated by our field.
Mallory Williams, MD, MPH

To Patrice, who calls me Husband; to Shelby, Maya, and Zoe, who call me
Father; to my students, who judge me a worthy guide in their journey to
excellence; and to my patients, who entrust me with their lives and call me
Surgeon: I say thank you. I call you all an inspiration and devote this work to
you.
Mariano E. Gimenez, MD, PhD
To my dear wife, Mariana, and three children, Felipe, Ignacio, and Azul, for
their constant support and affection. To my parents, Nere and Antonio, for
teaching me that, with passion and work, everything is possible. To my
colleagues in Buenos Aires, Strasbourg, and here, from whom I learn every
day.
Forty years ago, well within the lifetime
of most of us, the practice of
hepatopancreatobiliary surgery barely
existed. While there was great interest
and research in HPB pathophysiology,
therapeutic options were limited, and
operative interventions, in particular,
remained rare. Attempted operations
were typically high-risk, resource-
intensive procedures associated with high morbidity and mortality,
undertaken only by the boldest of surgeons. Over the ensuing several years,
however, the work of many pioneering investigators set the stage for the
emergence of HPB as a specialty area of surgery in its own right, bringing it
out of the dark ages and into the mainstream of practice. Indeed, in the current
state of the art, HPB operations are now commonplace for a wide array of
disease processes, with the expectation of successful outcomes for most
patients.
Many factors have played a role in the remarkable transformation of the
field, too many to list here, but the point is that our surgical forebears laid a
solid foundation upon which the technical advances that have shaped our
field have been built. As the beneficiaries of these advances and hard-won
lessons from the past, we now have the ability to offer effective and safe
surgical therapies to our patients. But for those to whom much has been
given, much is expected, and it is our charge now to continue refining and
improving the clinical, technical, and investigative aspects of HPB surgery,
to work toward uniform standards of training and practice, and to continually
assess and reassess the long-term results of our interventions. It is imperative
that we do this, first and foremost, to ensure the best possible outcomes for
our patients, and also to honor the legacy of those great surgeons and others
who have blazed the trail that we all now have the privelege to hike.
It is also fitting and proper for the AHPBA to be a leader in this process.
Formally established in 1994 as the American Chapter of the IHPBA, the
society has grown from a small group of surgeon thought leaders to a
powerful, well-funded, and dynamic society of over 1500 members. Over the
years, this chapter has evolved to become renamed the Americas, proudly
claiming members from the northern reaches of Canada to the southernmost
regions of South America, and every country in between. The AHPBA is
now one of the largest and most active societies dedicated to advancing the
field of HPB surgery by disseminating knowledge, fostering research and
innovation, advancing education and training, and encouraging
multidisciplinary collaboration.
The current textbook was undertaken by the AHPBA leadership in an
ongoing effort to fulfill its mission. Written in a novel, dilemma-focused
format, with each chapter addressing a specific question, this is a unique
textbook addressing all aspects of HPB surgery by society members who are
leading experts in the field. The book is well illustrated, with several salient
points highlighted at the end of each chapter. The scope of the book is very
broad, and its content is thus appropriate for a wide range of readers, from
residents and fellows to established general surgeons and expert HPB
surgeons. We expect that this text will remain an important resource for many
years to come.
It has been a privilege for us to have played a small part in this project
and in the ongoing advancement of the field through the AHPBA. We
congratulate the book’s editors, the contributing authors, the AHPBA
leadership, and Marjorie Malia and the staff at Association Management for
the great effort in bringing this book to completion.
Surgical management of diseases affecting the liver, pancreas, and biliary
tract (HPB) is a mature but rapidly evolving discipline. Clinicians with
special interest in these 3 organ systems have diligently endeavored through
the decades to innovate their specialties to first and foremost optimize patient
outcomes. With a steadfast commitment to this principle, these clinicians
have created thriving societies such as the Americas Hepato-Pancreato-
Biliary Association (AHPBA) and the International Hepato-Pancreato-
Biliary Association (IHPBA). The former has over 1,200 active members
from over 40 countries throughout North, Central, and South America, while
the latter includes over 1,800 members from over 60 different countries.
In 2017, the AHPBA leadership set out to publish a practical textbook,
with chapters written by the experts in the field that would encompass a
broad range of HPB topics, with the intention of dispersing critical
knowledge towards the mission of educating practicing surgeons, physicians,
and the wider health care provider community. The goal was to encapsulate
the essence of this vibrant and exciting discipline in a more cogent, concise,
and practical manner than traditional texts. We editors hope that this
textbook, Hepato- Pancreato-Biliary and Transplant Surgery: Practical
Management of Dilemmas , has achieved its goals.
The editors aimed to develop a format of the HPB textbook that
encompasses a broad clinical scope of 5 disciplines: surgical oncology, HPB
surgery, transplantation, interventional radiology, and trauma. The content of
Hepato-Pancreato-Biliary and Transplant Surgery: Practical
Management of Dilemmas is further divided into 6 major sections: (1)
Hepatic, (2) Biliary, (3) Pancreatic (4) Transplantation, (5) Trauma, and (6)
Innovation and Technology.
We believe that one of the unique aspects of Hepato-Pancreato-Biliary
and Transplant Surgery: Practical Management of Dilemmas is its
format. Rather than assigning a chapter a traditional title such as “Gallstone
Ileus,” the title is presented in the form of a question, such as “Should a
Cholecystectomy Be Done for Gallstone Ileus?” This approach reflects the
mindset of a clinician who is facing an actual clinical dilemma and wants to
know how to best care for his/her patient. Another distinct feature of this
textbook is that each section is structured into 3 major components
(preoperative, intraoperative, and postoperative) that highlight dilemmas
across the spectrum of care. Furthermore, embedded within each of these
components are debates that we believe provide a wonderful opportunity for
the readers to have a more in-depth understanding of simmering
controversies in HPB diseases.
Hepato-Pancreato-Biliary and Transplant Surgery: Practical
Management of Dilemmas has been a global undertaking. Over 350 authors
from different corners of the globe have contributed to its final product. Each
chapter is written by recognized experts and begins with a clinical scenario
followed by a concise and in-depth text that is filled with compelling images,
detailed illustrations, comprehensive tables, thorough algorithms, and other
adjunctive tools that enhance the learning process. Key references are
provided for further review. The ‘”Salient Points” at the end of each chapter
reinforce key concepts for the readers.
We have strived to address a broad spectrum of clinically relevant
scenarios in this large- volume textbook, with the understanding that
additional content may have been omitted due to text length and content
evolution constraints. Hence, we welcome suggestions from you, the readers,
with the prospect that we will incorporate your valuable suggestions into
future editions.
We hope that you will benefit from and enjoy this textbook, and believe
that you will find Hepato- Pancreato-Biliary and Transplant Surgery:
Practical Management of Dilemmas to be not only an invaluable resource,
but also a must-have textbook for all interested in this exciting specialty.
Quyen D. Chu, MD, MBA

Charles M. Vollmer, MD
Gazi B. Zibari, MD
Susan L. Orloff, MD
Mallory Williams, MD, MPH
Mariano E. Gimenez, MD, PhD
We would like to thank the AHPB Foundation for their support, a legion of
authors for their valuable contributions, Marjorie Malia, Jill Willhite,
Beverly Wright, Ashley Abrams, and Yvette Sanchez for their administrative
support, Jeremy Ricci, Thuy-Tien Chu, and Lory Tubbs, for their design and
formatting services, Barbara Pratt, Phillip Estes, and Nina Tindall for their
proofreading services, Trina Chu, Esq for her legal assistance, and Yen Chu,
Iwan Yasin Agustin, and Kaylee Messina for their artistic talents.
Eddie K. Abdalla, MD, FACS
Hepatobiliary and Pancreatic Surgery & Surgical Oncology
Atlanta Liver & Pancreas Surgical Specialists
Medical Director & Liver Pancreas Program
Northside Hospital Cancer Institute
Atlanta, Georgia
Jad Abou-Khalil, MD, MSc (Epi) FRCSC

Division of Hepato-Biliary and Pancreatic Surgery
The Ottawa Hospital
Ottawa, Ontario
Marwan S. Abouljoud, MD, FACS, CPE, MMM

Benson Ford Endowed Chair
Transplant and Hepatobiliary Surgery
Director, Henry Ford Transplant Institute
Henry Ford Hospital
Detroit, Michigan
Peter Abrams, MD, FACS

Director of Pancreas and Islet Transplantation
MedStar Georgetown Transplant Institute
MedStar Georgetown University Hospital
Washington, DC
Pablo Acquafresca, MD
General Surgeon
DAICIM Foundation Staff
Buenos Aires, Argentina
Anya Adair, MBBS, PhD, FRCS (Eng)

Consultant Transplant/HPB Surgeon
Royal Infirmary Edinburgh
Honorary Senior Lecturer
University of Edinburgh
René Adam, MD
Head Department of Hepatobiliary Surgery, Cancer and Transplantation
Director Research Team
Hôpital Universitaire Paul Brousse
Centre Hépato-Biliaire
Villejuif, France
David B. Adams, MD, FACS

Professor of Surgery
Medical University of South Carolina
Charleston, South Carolina
Syed A. Ahmad, MD, FACS

Chief Division Surgical Oncology
Associate Director
University of Cincinnati Cancer Institute
Director Pancreas Disease Center
University of Cincinnati Medical Center
Cincinnati, Ohio
Joseph Ahn, MD, MS, FAASLD

Associate Professor of Medicine
Director of Clinical Hepatology
Oregon Health & Science University
Portland, Oregon
Okan Akhan, MD, FCIRSE, FESGAR

Professor of Radiology
Interventional Radiology and Abdominal Imaging
Chief of Non-Vascular Interventional Radiology
Hacettepe University
Co-Chair of World Health Organization-Informal
Working Group on Echinococcosis
Ankara, Turkey
Devrim Akinci, MD, FCIRSE

Department of Radiology
Nonvascular Interventional Radiology
Ankara, Turkey
Farzad Alemi, MD FACS

Director, HPB Surgery
Transplant & HPB Institute
St Vincent Medical Center
Los Angeles, California
Marc-Antoine Allard, MD
Department of HPB Surgery & Transplantation
Centre Hépatobiliaire
Université Paris Sud
Paul Brousse Hospital
Villejuif, France
Peter J. Allen, MD, FACS

Associate Director,
Rubenstein Center for Pancreatic Cancer Research
Murray F. Brennan Professor of Surgery
Memorial Sloan Kettering Cancer Center
New York, New York
John D. Allendorf, MD, FACS

Associate Professor of Clinical Surgery
Stony Brook School of Medicine
Vice-Chairman
NYU Winthrop Hospital
Mineola, New York
Mario Almada, MD
Assistant Surgeon
University of Montevideo
Montevideo, Uruguay
Thomas A. Aloia, MD, FACS

Associate Professor
Department of Surgical Oncology
University of Texas MD Anderson Cancer Center
Houston, Texas
Adnan A. Alseidi MD, EdM, FACS
HPB & Endocrine Surgery
HPB Fellowship Director
Virginia Mason Medical Center
Seattle, Washington
Eisar Al-Sukhni, MD, MSc, FRCSC

Clinical Associate
Hepatobiliary and Acute Care Surgery
St Joseph’s Health Centre
Toronto, Canada
Fernando Andrés Alvarez, MD

Staff Surgeon, General Surgery Service
Clínica Universitaria Reina Fabiola
Córdoba, Argentina
Christopher D. Anderson, MD, FACS

James D Hardy Chair
Professor and Chair
Department of Surgery
University of Mississippi Medical Center
Jackson, Mississippi
Yumi Ando, MD
Fellow, Gastroenterology
Oregon Health & Sciences University
Portland, Oregon
Gustavo Andreoli, MD
Associate Professor of Surgery
University of the Republic (UDELAR)
Uruguay
Oscar C. Andriani, MD
Assistant Professor
HPB Surgery
Swiss Medical Group
Universidad Austral
Wellington Andrus, MD, PhD

Liver Transplantation Program Coordinator
Division of Abdominal Organ Transplantation
Department of Gastroenterology
Clinics Hospital of Sao Paulo
University Medical School (HC-FMUSP)
São Paulo, Brazil
Anand Annamalai, MD, FACS

Chairman
Transplant and Hepato-Pancreato- Biliary Surgery
Director of Transplant and Hepato- Pancreato-Biliary Institute
Chan-Soon-Shiong Institute For Medicine
St. Vincent Medical Center
Avo Artinyan, MD, MS, FACS

GI Surgical Oncology
Chief of Colorectal Surgical Oncology
Baylor College of Medicine
Michael E. Debakey Department of Surgery
Houston, Texas
Horacio J. Asbun, MD, FACS

Chairman
Mayo Clinic College of Medicine and Science
Jacksonville, Florida
Reza Askari, MD, FACS

Co-Director, Surgical Intensive Care Unit
Director, SICU and ACS Fellowship
Associate Trauma Director
Director, Surgery Clerkship
Brigham and Women’s Hospital
Harvard Medical School
Boston, Massachusetts
Jenna Aziz, MS
Georgetown University
Washington, DC
Salim Aziz, MD
Clinical Professor of Surgery
George Washington University &
Howard University School of Medicine
Washington, DC
Michelle Babicky, MD
Senior Clinical Fellow in Surgical Oncology
New York, New York
Elizabeth A. Bailey, MD, MEd

Fellow, Center for Healthcare Improvement & Patient Safety
Fellow, Center for Surgery and Health Economics
Hospital of the University of Pennsylvania
Philadelphia, Pennsylvania
Emily J. Baird, MD, PhD

Assistant Professor of Anesthesiology
& Perioperative Medicine
Program Director of Resident Education & Scholarship
Portland, Oregon
Marshall S. Baker, MD, MBA, FACS

Clinical Associate Professor of Surgery
University of Chicago
Pritzker School of Medicine
Vice Chair of Quality and Patient Safety
North Shore University Health System
Chicago, Illinois
Chad G. Ball, MD, MSc, FRCSC, FACS

Associate Professor of Surgery and Oncology
Hepatobiliary and Pancreatic Surgery
Trauma and Acute Care Surgery
University of Calgary
Foothills Medical Center
Calgary, Alberta
Todd H. Baron, MD, FASGE

University of North Carolina
Chapel Hill, North Carolina
Savio George Barreto, MD

Senior Lecturer & Clinical Fellow
Hepato-Pancreato-Biliary Surgery
Flinders University & Medical Centre
Adelaide, Australia
Wesley Barry, MD
Research Fellow
Children’s Hospital Los Angeles
University of Southern California
Alfred E. Baylor, III, MD, FACS

Assistant Professor of Surgery
Wayne State University SOM / Detroit Medical Center
Michael and Marian Ilitch Department of Surgery
Detroit Receiving Hospital
Detroit, Michigan
Eliza W. Beal, MS, MD, FACS

The Ohio State University
Wexner Medical Center
Columbus, Ohio
Joal D. Beane, MD
Fellow
Complex Surgical Oncology
University of Pittsburgh Medical Center
Pittsburgh, Pennsylvania
Rachel E. Beard, MD
Division of Surgical Oncology
Hepatobiliary and Pancreatic Surgery
Alpert Medical School of Brown University
Rhode Island Hospital
Providence, Rhode Island
Stephen W. Behrman, MD, FACS

University of Tennessee Health Science Center
Memphis, Tennessee
Jacques Belghiti, MD, HACS

HPB Unit of Liver Surgery and Transplantation
Hopital Beaujon
Paris, France
Matthew Benns, MD, FACS

University of Louisville
Louisville, Kentucky
Marina Berenguer, MD
Professor of Medicine
La Fe University Hospital, IISLaFe & Ciberehd
University of Medicine
Valencia, Spain
Jennifer Berumen, MD, FACS

Division of Transplantation & Hepatobiliary Surgery
UC San Diego Health
Transplant Center & HPB Surgery
La Jolla, California
Marc G.H. Besselink, MD, PhD, MSc

Academic Medical Center
Amsterdam, the Netherlands
Sylvester M. Black, MD, PhD

The Ohio State University Wexner Medical Center
Columbus, Ohio
Adam S. Bodzin, MD, FACS

Assistant Professor
Section of Transplant Surgery
University of Chicago-Biological Science Division
Chicago, Illinois
Hillary J. Braun, MD
Resident Physician, General Surgery
University of California, San Francisco
San Francisco, California
L.D. Britt, MD, MPH, D.Sc. (Hon), FACS, FCCM, FRCSEng (Hon),
FRCSEd (Hon), FWACS (Hon), FRCSI (Hon), FCS(SA) (Hon)
Henry Ford Professor and Edward J. Brickhouse Chairman
Eastern Virginia Medical School
Surgery Department
Norfolk, Virginia
Christoph E. Broelsch, MD, FACS

Professor of Surgery (Ret)
Section of Hepato-Pancreato-Biliary Surgery & Liver Transplantation
Chicago, Illinois
Kimberly M. Brown, MD, FACS

Associate Chair of Education
Dell Medical School at UT Austin
Austin, Texas
Zachary J. Brown, DO
Surgical Oncology Research Fellow
Thoracic and GI Oncology Branch
Center for Cancer Research
National Cancer Institute | National Institutes of Health
Bethesda, Maryland
David S. Bruce, MD, FACS

Transplant Surgeon
Ochsner Multi-Organ Transplant Institute
Ochsner Clinic
New Orleans, Louisiana
Joseph F. Buell, MD, MBA, FACS

Professor of Surgery and Pediatrics
Tulane University
Mark P. Callery, MD, FACS

Chief, Division of General Surgery
Beth Israel Deaconess Medical Center
Andrew M. Cameron, MD, PhD, FACS

Chief, Division of Abdominal Transplantation
Johns Hopkins Hospital
Baltimore, Maryland
Andre Campbell, MD, FACS, FACP, FCCM

Attending Trauma Surgeon
Zuckerberg San Francisco General Hospital
Medical Director Surgical ICU
UCSF, School of Medicine
Lee Cartagena, MD
Senior Trauma/Critical Care Fellow
Department of Trauma and Burns
John H. Stroger Hospital of Cook County
Chicago, Illinois
Jamie Case, PhD

Scientific Program Director
Bio-Repository & Transplantation Research
Scripps Center for Organ & Cell Transplantation
Scripps Clinic and Scripps Green Hospital
François Cauchy, MD
HPB Unit of Liver Surgery and Transplantation
Hopital Beaujon
Paris, France
Eugene P. Ceppa, MD, FACS

Indiana University Department of Surgery
Indianapolis, Indiana
Guillermo Cervio, MD
Chief of the Clinical Area Hospital Garrahan
John A. Chabot, MD, FACS

Columbia University Medical Center
New York, New York
Carlos Chan, MD
Chief HPB Surgery
National Institute of Nutrition and Medical Sciences
Mexico City, Mexico
William C. Chapman, MD, FACS

Section Chief of Transplant Surgery
Washington University in St Louis, MO
St Louis, Missouri
Kenneth D. Chavin, MD, PhD, FACS

Director, Transplant Institute
Division Chief, Hepatobiliary and Transplant Surgery
University Hospitals Cleveland Medical Center
Vice Chair of Research, Department of Surgery
Case Western Reserve University School of Medicine
Cleveland, Ohio
Edward E. Cho, MD, ScM

Hepatopancreatobiliary and Foregut Surgery
Associate Director – Advanced GI Surgery Fellowship
Trinity Surgical Consultants
Methodist Richardson Medical Center
Richardson, Texas
Michael A. Choti, MD, FACS

Joan and George Bayoud Professor of Surgery
University of Texas Southwestern Medical Center
Dallas, Texas
John D. Christein, MD, FACS

University of Alabama-Birmingham
Birmingham, Alabama
Quyen D. Chu, MD, MBA, FACS

Charles D. Knight, Sr. Endowed Professor
Chief, Surgical Oncology
Director, Hepato-Pancreato-Biliary Services
LSU-Health Sciences Center- Shreveport
Shreveport, Louisiana
Turkmen T. Ciftci, MD
Associate Professor of Radiology
Nonvascular Interventional Radiology
Ankara, Turkey
Francisco G. Cigarroa, MD, FACS

Division Head of Liver & Pediatric Transplantation Surgery
Carlos and Malú Alvarez Distinguished University Chair
Ashbel Smith Professorship in Surgery
University of Texas Health San Antonio
San Antonio, Texas
Sean Cleary MD, MSc, MPH, FRCSC, FACS

Associate Professor of Surgery,
Division of Hepatobiliary and Pancreatic Surgery
Mayo Clinic
Rochester, Minnesota
Jordan M. Cloyd, MD
Columbus, Ohio
Robert J.S. Coelen, MD, PhD

Surgical Resident
Medical Center Slotervaart
Amsterdam, the Netherlands
Felipe José Fernández Coimbra, MD, FACS

Surgical Oncologist
President of the Brazilian Society of Surgical Oncology
Secretary AHPBA
Chief of the Department of Abdominal Surgery at
AC Camargo Cancer Center
São Paulo, Brazil
Waldo Concepcion, MD, FACS

Chief of Clinical Transplantation
Chief of Pediatric Kidney Transplantation
Division of Transplantation
Stanford University School of Medicine
Stanford, California
Claudius Conrad, MD, PhD, FACS

Hepato-Pancreato-Biliary Surgery
Houston, Texas
Kendra D. Conzen, MD, FACS

Division of Transplant Surgery
University of Colorado School of Medicine
Aurora, Colorado
Matthew Cooper, MD, FACS

Director of Kidney Transplantation
Washington, DC
Edward E. Cornwell, III, MD, FACS, FCCM, FWACS (Hon)

LaSalle D. Leffall, Jr. Professor and Chairman
Howard University College of Medicine
Surgeon-in-Chief
Howard University Hospital
Washington, DC
Thiago Nogueira Costa, MD

Staff Surgeon at Servico de Cirurgia das Vias Biliares
e Pancreas / Hospital das Clinicas
Faculdade de Medicina da Universidade de Sao Paulo
San Paulo, Brazil
Jorge Cardoso Cúneo, MD, PhD

Authorized Teacher of Surgery
Institute of Oncology “Angel H. Roffo”
University of Buenos Aires
Steven A. Curley, MD, FACS

Professor and Chief of Surgical Oncology
Houston, Texas
Horacio R.V. D’Agostino, MD, FACR, FSIR

Professor of Radiology, Surgery and Anesthesiology
Chairman, Department of Radiology
LSU Health Sciences Center- Shreveport
Luiz Augusto Carneiro D’Albuquerque, MD, PhD

Professor
Clinics Hospital of Sao Paulo University Medical School
(HC-FMUSP)
São Paulo, Brazil
Michael D’Angelica, MD, FACS

Enid A. Haupt Chair in Surgery
Attending Surgeon, Hepatopancreatobiliary Service
Program Director, Complex Surgical Oncology & HPB Fellowship
Cornell University, Weill Medical College
New York, New York
Leigh Anne Dageforde, MD, MPH

Abdominal Organ Transplant and HPB Fellow
Washington University in Saint Louis
Saint Louis, Missouri
Jashodeep Datta, MD
Fellow, Complex General Surgical Oncology
New York, New York
Allison David
Sidney Kimmel Medical College
Thomas Jefferson University
Catherine H. Davis, MD, MPH

The University of Texas MD Anderson Cancer Center
Houston, Texas
Xabier de Aretxabala, MD, FACS

Professor of Surgery Universidad de Chile
Staff Surgeon Clinica Alemana
Staff Surgeon Air Force Hospital
Santiago, Chile
Maria Luiza Leite de Medeiros, MD

Surgical Oncologist
AC Camargo Cancer Center
São Paulo, Brazil
Vincent E. de Meijer, MD, PhD

Division of Hepato-Pancreato-Biliary Surgery & Liver Transplantation
University of Groningen and University Medical Center
Groningen, Netherlands
Mauricio R. Liberato de Moura, MD

Staff, Department of Interventional Radiology
University of São Paulo
Staff, Division of Interventional Radiology
Sírio Libanês Hospital
São Paulo, Brazil
Eduardo de Santibañes, MD, PhD

Professor and Chairman, Liver Transplant Unit
Hospital Italiano de Buenos Aires
Gabriela del Ángel, MD

General Surgery Resident
National Institute of Nutrition and Medical Sciences
Mexico City, Mexico
Mary Ann Decoteau, MD

General Surgery Resident / Transplant Research
Scripps Center for Organ Transplantation
Aram N. Demirjian, MD, MA, FACS

Gunjan Desai, MS, MRCS, FMAS

Department of Gastrointestinal Surgery
Lilavati Hospital and Research Centre
Mumbai, India
Siddharth Desai, MD
Chief Resident in Surgery
Vikrom K. Dhar, MD.

Cincinnati, Ohio
Mashaal Dhir, MD
Upstate Medical University
Syracuse, New York
Michele Diana, MD, PhD

Director of the Research Unit on
Endo-Laparoscopic Procedures
IHU-Strasbourg, Institute of Image- Guided Surgery
Senior Researcher at IRCAD
Research Institute against Cancer of the Digestive System
Strasbourg, France
Geraldine C. Diaz, DO
Associate Professor of Anesthesia & Critical Care
University of Illinois at Chicago
Chicago, Illinois
Ornella Agnes Ditulio, MD

Assistant Professor
Elijah Dixon, MD, BSc, MSc (Epi), FRCSC, FACS

Professor of Surgery, Oncology, and
Community Health Sciences Chief
City Wide Section of General Surgery University of Calgary
Calgary, Canada
Maren Donato, MD
Clínica de Imágenes
Neuquén
Neuquén, Argentina
Jon D. Dorfman, MD, FACS

University of Massachusetts Medical School
UMass Memorial Medical Center
Worcester, Massachusetts
Cataldo Doria, MD, PhD, MBA, FACS

Nicoletti Family Professor of Transplant Surgery
Director, Jefferson Transplant Institute
Director, Division of Transplantation
Surgical Director Kimmel Cancer Center –
Jefferson Liver Tumor Center
Thomas Jefferson University Hospital
Austin Dosch, MD
Resident in Surgery
University of California, Irvine
Irvine, California
Maria B. Majella Doyle, MD, MBA, FACS

Director, Liver Transplant
Director, Transplant HPB Fellowship Program
Section of Abdominal Transplant
Washington University School of Medicine
St. Louis, Missouri
Jeffrey A. Drebin, MD, PhD, FACS

Chair of Surgery
Weill Cornell School of Medicine
New York, New York
Fernando Duek, MD
Staff Surgeon
Liver & Transplant Unit
Hospital Dr. Cosme Argerich
Elizabeth W. Duggan, MD
Chief of Service for Division of General
& Transplant Anesthesiology
Emory University School of Medicine
Atlanta, Georgia
Truman M. Earl, MD, FACS

Brett L. Ecker, MD
Resident, Department of Surgery
University of Pennsylvania
Devin E. Eckhoff, MD, FACS

Professor and Division Chair
Liver Transplant and Hepatobiliary Surgery
University of Alabama at Birmingham
Birmingham, Alabama
Aras Emdadi, DO
C. Kristian Enestvedt, MD, FACS

Assistant Professor
& Hepatobiliary Surgery
Portland, Oregon
Laura M. Enomoto, MSc, MD

Fellow in Surgical Oncology
Wake Forest Baptist Health
Winston Salem, North Carolina
Alberto Espino, MD
Assistant Professor
Director, Endoscopy Unit
Pontificia Universidad Católica de Chile
Santiago, Chile
Jeffrey H. Fair, MD, FACS

Professor of Surgery & Chief
University of Texas Medical Branch
Galveston, Texas
Alik Farber, MD, FACS

Chief, Division of Vascular & Endovascular Surgery
Associate Chair for Clinical Operations
Boston Medical Center
Professor of Surgery and Radiology
Boston University School of Medicine
Marcelo Fasano, MD, FACS

Staff Surgeon
“Dr Carlos A Bocalandro” Hospital
Carlos Fernández-del Castillo, MD, FACS

Jorge and Darlene Pérez Endowed Chair in Surgery
Director, Pancreas and Biliary Surgery Program
Massachusetts General Hospital
Alberto R. Ferreres, MD, PhD, FACS (Hon)

Professor of Surgery and Chair
“Dr. Carlos A Bocalandro” Hospital
Caetano Finger, MD
Percutaneous Surgeon,
Staff DAICIM Foundation,
Robert A. Fisher, MD, FACS

Chief, Division of Transplantation
Director, the Transplant Institute
The Anthony P. Monaco Transplant Surgical Service
William E. Fisher, MD, FACS

Professor of Surgery Clinical Vice Chair and Chief
Division of General Surgery
George L. Jordan, MD Chair of General Surgery
Michael E. DeBakey Department of Surgery
Director, Elkins Pancreas Center
Chief, Surgery Service Line
Baylor St. Luke’s Medical Center
Houston, Texas
David P. Foley, MD, FACS, FAST, FAASLD

Surgical Director, Liver Transplant Program
William S. Middleton VA Hospital
Program Director, University of Wisconsin Transplant Surgery Fellowship
Department of Surgery, Division of Transplantation
University of Wisconsin School of Medicine and Public Health
Madison, Wisconsin
Yuman Fong, MD, FACS

Sangiacomo Chair and Chairman
City of Hope Medical Center
Duarte, California
Zhi Ven Fong, MD, MPH

Henri R. Ford, MD, MHA, FACS, FRCS, FAAP

Dean
University of Miami
Leonard M. Miller School of Medicine
Miami, Florida
Kathryn J. Fowler, MD
Assistant Professor
Mallinckrodt Institute of Radiology
Washington University in St Louis
St Louis, Missouri
Wesley P. Francis, MD, FACS
Associate Lecturer
University of the West Indies
School of Clinical Medicine and Research
Nassau, Bahamas
Wayne A. I. Frederick, MD, MBA, FACS

President
Howard University
Washington, DC
Catherine H. Frenkel, MD
Fellow
University of Pennsylvania
Peter J. Friend, MD, FRCS

Professor of Transplantation
Nuffield Department of Surgical Sciences
University of Oxford
Oxford Transplant Centre
Churchill Hospital
Oxford, United Kingdom
Danielle M. Fritze, MD
University Transplant Center
University of Texas Health San Antonio
San Antonio, Texas
James Fung, MBChB, MD, FRACP, FHKCP, FHKAM

Consultant and Honorary Clinical Associate Professor
Queen Mary Hospital
The University of Hong Kong
Hong Kong, China
Amy Gallo, MD, FACS

Palo Alto, California
Dragos M. Galusca, MD, FASA

Chief, Division of Critical Care Anesthesiology
Department of Anesthesiology
Pain Management and Perioperative Medicine
Henry Ford Health System
Wayne State University School of Medicine
Detroit, Michigan
Armando Ganoza, MD
Thomas E. Starzl Transplantation Institute
Children’s Hospital of Pittsburgh of UPMC
Andrés Garelli, MD, FACS

Chief of Hepato Bilio Pancreatic Sector
Central Military Hospital
Director of the General Surgery Especialization of
Central Military Hospital
Buenos Aires University
David A. Geller, MD, FACS

Richard L. Simmons Professor of Surgery
Chief, Division of Hepatobiliary and Pancreatic Surgery
Director, UPMC Liver Cancer Center
University of Pittsburgh
Pittsburg, Pennsylvania
Georgios Gemenetzis, MD
Postdoctoral Fellow
Johns Hopkins University School of Medicine
Baltimore, Maryland
John F. Gibbs, MD, MHCM, FACS

Chairman, Department of Surgery JSUMC
Chief of Oncology
Professor of Surgery Seton Hall School of Medicine
Neptune, New Jersey
Richard Gilroy, MBBS

Medical Director of Hepatology & Liver Transplantation
Intermountain Medical Center
Murray, Utah
Mariano Giménez, MD, PhD

Director - Chair of General Surgery &
Minimal Invasive Surgery “Taquini”
University of Buenos Aires, Argentina
Scientific Director of Percutaneous Surgery
IHU-IRCAD
Strasbourg, France
Nicolas Goldaracena, MD
Clinical Associate in Transplantation
University of Toronto
Toronto, Canada
Jamie Golden, MD
Pediatric Surgery Research Fellow
Hiram Raul Muñoz González, MD

National Institute of Medical Science
and Nutrition “Salvador Zubirán”
Delegación, Tlalpan, Mexico
Claire Goumard, MD
Research Fellow
Houston, Texas
Paul D. Greig, MD, FRCSC, FACS

Toronto, Canada
Elliot I. Grodstein, MD
Donald and Barbara Zucker School of Medicine at Hofstra/Northwell
Manhasset, New York
Juan Francisco Guerra, MD

Assistant Professor
Department of Digestive Surgery
Director, Transplant Program
Pontificia Universidad Católica de Chile
Santiago, Chile
Amar Gupta, MD, FRCSC

HPB Surgery Fellow
University of Calgary
Calgary, Canada
Niraj J. Gusani, MD, MS, FACS

Associate Professor of Surgery,
Medicine, & Public Health Sciences
Penn State College of Medicine
Hershey, Pennsylvania
Esteban Halac, MD
Staff Surgeon
Liver Transplant Division
Hospital Juan P. Garrahan
Andrew D. Hardie, MD
Assistant Professor of Radiology
Julie K. Heimbach, MD, FACS

Professor of Surgery and Chair
William J. von Liebig Transplant Center
Mayo Clinic College of Medicine
William Scott Helton, MD, FACS

Director of Liver, Biliary and Pancreatic Surgery
Center of Excellence,
Digestive Disease Institute
Seattle, Washington
Alan W. Hemming, MD, FACS, FRCSC

Professor and Chief,
Division of Transplantation & Hepatobiliary Surgery
UC San Diego Health
Transplant Center & HPB Surgery
Paulo Herman, MD, PhD

Associate Professor
Head of the Liver Surgery Unit
Hospital das Clinicas
University of São Paulo School of Medicine
São Paulo, Brazil
Jonathan M. Hernandez, MD
Surgical Oncology and HepatoPancreatoBiliary Surgery
Thoracic and GI Oncology Branch | Center for Cancer Research
National Cancer Institute | National Institutes of Health
Bethesda, Maryland
Roberto Hernandez-Alejandro, MD, FACS

Chief, Division of Transplantation and HPB Surgery
University of Rochester Medical Center
Rochester, New York
Sara Houck, MSW, LICSW

Liver Transplant Social Work
University of Washington Medical Center
Seattle, Washington
Eduardo J. Houghton, MD
Staff DAICIM Foundation.
Associate Professor
Staff, Mini Invasive Surgery
Bernardino Rivadavia Hospital
Abhinav Humar, MD
Chief, Division of Abdominal Transplantation Surgery
Clinical Director, Thomas E. Starzl Transplantation Institute
Thomas E. Starzl Professor in Transplantation Surgery
David Iannitti, MD, FACS

Chief of HPB Surgery
Carolinas Health Care System
Charlotte, North Carolina
Oscar Imventarza, MD
Chief of Liver Transplantation Hospital Garrahan & Hospital Argerich
President International Hepato Pancreatic Biliary Association
President Elect STALYC (Latin American Transplant Society)
Kyle R. Jackson, MD
Johns Hopkins Hospital
Baltimore, Maryland
Lenworth Jacobs MD, MPH, FACS

University of Connecticut
Vice President of Academic Affairs and
Director of the Trauma Institute
Hartford Hospital
Hartford, Connecticut
Palepu Jagannath MS, FACS, FRCS (Eng)
Chairman, Department of Surgical Oncology
Lilavati Hospital and Research Centre
Mumbai, India
Ajay Jain, MD, FACS

Department of Hepatobiliary & Pancreatic Surgery
SUNY Upstate Medical University
Syracuse, New York
William R. Jarnagin, MD, FACS

Chief, HPB Surgical Service
Benno C. Schmidt Chair in Surgical Oncology
Weill Medical College of Cornell University
New York, New York
Nicolas Jarufe, MD, FACS

Chief Digestive Surgery Department
P. Universidad Catolica de Chile
Santiago, Chile
Shiva Jayaraman MD, MESc, FRCSC, FACS

HPB and General Surgery
St. Joseph’s Health Centre
Toronto, Canada
Vagner B. Jeismann, MD, PhD
Assistant Professor
Liver Surgery Unit
Hospital das Clinicas
University of São Paulo School of Medicine
São Paulo, Brazil
D. Rohan Jeyarajah, MD, FACS

Director, HPB/AGI Fellowship Program
Director, GI Surgical Services
Methodist Health System
Richardson, Texas
Sandeep Jhajj, MD
Resident
University of Arizona
Tucson, Arizona
Lynt B. Johnson, MD, MBA, FACS

Robert J. Coffey Professor and Chairman
Medstar Georgetown University Hospital
Washington, DC
Kimberly Joseph, MD, FACS, FCCM

Division Chair, Trauma Critical Care and Prevention
Department of Trauma and Burns
John H. Stroger Hospital of Cook County
Chicago, Illinois
Dong-Hwan Jung, MD, PhD
Professor
Division of Hepatobiliary Surgery and Liver Transplantation
Department of Surgery, Asan Medical Center
University of Ulsan College of Medicine
Seoul, Republic of Korea
Zeljka Jutric, MD
Assistant Professor, Department of Surgery
Division of Hepatobiliary & Pancreas Surgery
University of California, Irvine Medical Center
Orange, California
Piyush Kalakoti, MD
Postdoctoral Research Fellow
Zaheer S. Kanji, MD, MSc, FRCSC

Clinical HPB Fellow
Seattle, Washington
Matthew H.G. Katz, MD, FACS

Chief, Pancreatic Surgery Service
Vice Chair for Research
Houston, Texas
Dixon B. Kaufman, MD, PhD, FACS

Ray D. Owen Professor and Chair
University of Wisconsin School of Medicine and Public Health
Madison, Wisconsin
Tara S. Kent, MD, FACS

Adeel S. Khan, MD, MPH, FACS

Section of Transplant Surgery
Washington University in St Louis, MO
St Louis, Missouri
Khalid Khwaja, MD, FACS

Surgical Director of Solid Organ Transplantation
The Transplant Institute
The Anthony P. Monaco Transplant Surgical Service
Bradford J. Kim, MD, MHS

Houston, Texas
Na Eun Kim, MD
Boston Medical Center
Peter T.W. Kim, MD, MSc, FRCSC, FACS

Hepatopancreatobiliary Surgery/Liver Transplantation
Vancouver General Hospital
Clinical Associate Professor University of British Columbia
Vancouver, British Columbia
Dilip S. Kittur, MD, ScD, FACS

Professor of Surgery, Microbiology and Immunology
Syracuse, New York
Andrew Klein, MD, MBA, FACS

Director, Comprehensive Transplant Center
Esther and Mark Schulman Chair of Surgery and Transplant Medicine
Professor and Vice Chair, Department of Surgery
Cedars-Sinai Medical Center
Göran B.G. Klintmalm, MD, PhD, FACS

Chief and Chairman
Annette C. and Harold C. Simmons Transplant Institute
W.W. Caruth Chair in Organ Transplant Immunology
Baylor University Medical Center
Dallas, Texas
Michael D. Kluger, MD, MPH, FACS

Division of GI & Endocrine Surgery
Columbia University College of Physicians and Surgeons
New York-Presbyterian Hospital
New York, New York
Lisa Marie Knowlton, MD, MPH

Postdoctoral Fellow, Trauma and Surgical Critical Care
Stanford Hospital and Clinics
Bas Groot Koerkamp, MD, PhD

Division of Hepato Pancreato Biliary Surgery & Abdominal Transplantation
Erasmus MC, University Medical Center Rotterdam
Rotterdam, Netherlands
Gustavo Kohan, MD, PhD

Assistant Professor
David A. Kooby, MD, FACS

Chief of Emory Surgery, Northern Arc
Winship Cancer Institute
Atlanta, Georgia
Nisa Kubiliun, MD
Assistant Professor of Internal Medicine
Director of Endoscopic Services
Clinical Chief, Division of Digestive and Liver Diseases
UT Southwestern Medical Center
Dallas, Texas
Aleksandra Kukla, MD
Associate Professor of Internal Medicine and Nephrology
Mayo Clinic
Rupak D. Kulkarni, MD
Galveston, Texas
John W. Kunstman, MD, MHS

Fellow, Complex General Surgical Oncology
New York, New York
William P. Lancaster, MD, FACS

Eric Langewisch, MD
University of Nebraska Medical Center
Omaha, Nebraska
Julie N. Leal MD, FRCSC

HPB Surgeon
Dalhousie University
Moncton, Canada
Anna M. Ledgerwood, MD, FACS

Director of Trauma
Detroit, Michigan
Jeffrey E. Lee, MD, FACS

Professor & Chair, Department of Surgical Oncology
Irving and Nadine Mansfield &
Robert David Levitt Cancer Research Chair
Houston, Texas
Major Kenneth Lee, IV, MD, PhD, FACS

University of Pennsylvania Health System
Sung-Gyu Lee, MD, PhD

Professor
Division of Hepatobiliary Surgery and Liver Transplantation
Department of Surgery, Asan Medical Center
University of Ulsan College of Medicine
Javier C. Lendoire, MD, PhD

Vice-Chairman
Liver & Transplant Unit
Hospital Dr. Cosme Argerich
Chairman, Liver Transplantation
Sanatorio Trinidad Mitre
Sui Kwong Li, MD

Infectious Diseases Fellow in Training
Division of Infectious Diseases
Portland, Oregon
Michael E. Lidsky, MD
Fellow, Hepatopancreatobiliary Surgery & Complex Surgical Oncology
New York, New York
Keith D. Lillemoe, MD, FACS

Chief of Surgery
W. Gerald Austen MD Professor of Surgery
Perparim Limani, MD, PhD

Fellow, Surgical Oncology
Houston, Texas
Jason B. Liu, MD, MS, FACS

University of Chicago Medicine
Chicago, Illinois
George E. Loss, Jr., MD, PhD, FACS

Chief of Surgical Services
Chairman Department of Surgery
Chief, Ochsner Multi-Organ Transplant Institute
Ochsner Clinic
Michael Loudin, MD
Gastroenterology Fellow
Division of Gastroenterology and Hepatology
Portland, Oregon
Andrew M. Lowy, MD, FACS

Chief, Division of Surgical Oncology
Leader, Gastrointestinal Oncology Program
Moores Cancer Center
UC San Diego Health
Charles E. Lucas, MD, FACS
Michael & Marian Ilitch Department of Surgery
Detroit, Michigan
Stewart Jay Lustik, MD, MBA

Professor of Anesthesiology
Vice Chair for Clinical Affairs
Director, Anesthesia Technical Services
University of Rochester School of Medicine
Rochester, New York
Shishir K. Maithel, MD, FACS

Scientific Director, Emory Liver and Pancreas Center
Winship Cancer Institute, Division of Surgical Oncology
Atlanta, Georgia
Wuttiporn Manatsathit, MD
Omaha, Nebraska
Jacques Marescaux, MD, FACS, FRCS (Hon), FJSES (Hon), FASA

(Hon)
President and Founder of IRCAD
Chairman of the Institute of Image- Guided Surgery
Strasbourg, France
Christopher Lee Marsh, MD, FACS

Division Chief, Scripps Center for Organ Transplantation
Guilherme Lopes P. Martins, MD

Staff, Department of Interventional Radiology
Staff, Division of Interventional Radiology
Sírio Libanês Hospital
São Paulo, Brazil
Ronald Matteotti, MD, FACS, FMH, ABIHM

Medical Director Hepatobiliary &
Pancreatic Program Meridian Cancer Care
Chief Hepatobiliary & Pancreatic Surgery JSUMC
Neptune, New Jersey
Erin Maynard, MD, FACS

Abdominal Organ Transplant and Hepatobiliary Surgery
Oregon Health and Science University
Portland, Oregon
Stacey McCandlish, MSW, LICSW

Liver Transplant Social Work
University of Washington Medical Center
Seattle, Washington
James A. McClintic, BS
Galveston, Texas
Domenica McKenna, PharmD

Transplant Clinical Pharmacy Specialist
VA Portland Health Care System
Portland, Oregon
Marcos R. Menezes MD, PhD

Director of Radiology and Image Guided Intervention of
Instituto do Cancer do Estado de São Paulo
Director of Image Guided Intervention Center of Hospital Sirio Libanes
São Paulo, Brazil
Christine (Cooky) O. Menias, MD

Chair, Division of Radiology
Mayo Clinic
Phoenix, Arizona
Miguel-Angel Mercado, MD, ASA (Hon), ESA (Hon)

Professor and Chairman
National Institute of Medical Science &
Nutrition “Salvador Zubirán”
Delegación, Tlalpan, Mexico
J. Michael Millis, MD, FACS

Vice Chair for Global Surgery
Director, Hepatobiliary Surgery
Chicago, Illinois
Rebecca M. Minter, MD, FACS

A.R. Curreri Professor
Chair, Department of Surgery
University of Wisconsin School of Medicine & Public Health
Madison, Wisconsin
Takashi Mizuno, MD, PhD

Physician-Scientist
MD Anderson Cancer Center
Houston, Texas
Somala Mohammed, MD, MPH

Houston, Texas
Andre Luis Montagnini, MD, PhD

Staff surgeon at Servico de Cirurgia das Vias Biliares
e Pancreas / Hospital das Clinicas
Full Professor at Disciplina de Cirurgia
do Aparelho Digestivo e Coloproctologia
Faculdade de Medicina da Universidade de Sao Paulo
Sao Paulo, Brazil
Martin I. Montenovo, MD, FACS

Assistant Professor
Surgical Director Living Donor Liver Transplant Program
University of Washington
Seattle, Washington
Katherine A. Morgan, MD, FACS

Professor and Chief
Division of GI and Laparoscopic Surgery
Ronald S. Mowad, MD
Arvind R. Murali, MD
Assistant Professor of Medicine
Gastroenterology & Hepatology
University of Iowa
Iowa City, Iowa
Masato Nagino, MD, PhD

Professor and Chairman
Nagoya University Graduate School of Medicine
Showa-ku, Nagoya, Japan
Attila Nakeeb, MD, FACS
Indiana University School of Medicine
David Nasralla, MD
Clinical Research Fellow in Transplant Surgery
Nuffield Department of Surgical Sciences
University of Oxford
Oxford Transplant Centre
Churchill Hospital
Oxford, United Kingdom
Ibrahim Nassour, MD, MSCS

University of Texas Southwestern Medical Center
Dallas, Texas
Willscott Naugler, MD
Medical Director of Liver Transplantation
Portland, Oregon
Steven J. Nurkin MD, MS, FACS

Associate Professor of Oncology
Roswell Park Comprehensive Cancer Center
University at Buffalo
Buffalo, New York
Jon S. Odorico, MD, FACS
Department of Surgery,
University of Wisconsin School of Medicine & Public Health
Madison, Wisconsin
Ali J. Olyaei, PharmD

Professor of Medicine and Pharmacotherapy
Nephrology & Hypertension
Oregon State University and Oregon Health & Sciences University
Portland, Oregon
Mark S. Orloff, MD, FACS

Vice Chairman Department of Surgery
Division of Abdominal Transplantation & Hepatobiliary Surgery
Rochester, New York
Susan L. Orloff, MD, FAASLD, FACS

Chief, Division of Abdominal Organ Transplantation/Hepatobiliary Surgery
Adjunct Professor, Molecular Microbiology & Immunology
Chief, Portland VA Medical Center Transplant Program
Portland, Oregon
W. Shannon Orr, III, MD, FACS

Surgical Oncology
Houssam Osman, MD, FACS

Hepatopancreatobiliary Surgery
Associate Director, HPB fellowship
MRMC - Methodist Heath System
Dallas, Texas
Sergio Pacheco MD
Digestive Surgeon
Pontificia Universidad Catolica de Chile
Santiago, Chile
Esha Pai, MBBS, MS

Fellowship - GI & Thoracic Surgical Oncology
Specialist Registrar, Surgical Oncology
Tata Memorial Centre
Mumbai, India
Arun P. Palanisamy, PhD

Assistant Professor
Division of Transplant & Hepatobiliary Surgery
School of Medicine
Case Western Reserve University
Cleveland, Ohio
Mariano Palermo, MD, PhD, FACS

Staff at DAICIM Foundation
President of International Club of Young Laparoscopic Surgery
Alessandro Paniccia, MD
University of Colorado Anschutz Medical Campus
Aurora, Colorado
Janak A. Parikh, MD, MSHS, FACS

Michigan State University College of Medicine
Southfield, Michigan
Flavio Paterno, MD, FACS

University of Cincinnati College of Medicine
Cincinnati, Ohio
Timothy M. Pawlik, MD, MPH, PhD,
FACS, FRACS (Hon.)

Professor and Chair, Department of Surgery
The Urban Meyer III and Shelley Meyer Chair for Cancer Research
Professor of Surgery, Oncology, and Health Services Management and Policy
The Ohio State University, Wexner Medical Center
Columbus, Ohio
Lucio Lucas Pereira MD
Research Fellow
Mayo Clinic Florida
General Surgery - HBDF/Brasilia
Brazil
Henry A. Pitt, MD, FACS

Chief Quality Officer
Temple University Health System
Associate Vice Dean for Clinical Affairs
Lewis Katz School of Medicine at Temple University
Rebecca Plevin, MD
Clinical Instructor of Surgery
Trauma Fellow
Zuckerberg San Francisco General Hospital
UCSF, School of Medicine
David T. Pointer Jr., MD

Resident, General Surgery
Tulane University School of Medicine
Elizabeth A. Pomfret, MD, PhD, FACS

Chief of Transplant Surgery
Igal Kam, MD Chair in Transplantation
University of Colorado School of Medicine
Aurora, Colorado
Alessandra Pulvirenti, MD
Research Fellow
Hepatopancreatobiliary Service
Memorial Sloan-Kettering Cancer Center
New York, New York
Motaz Qadan, MD, PhD

Claudemiro Quireze Jr, MD, PhD

Federal University of Goiás
President of the Brazilian Society of
Hepato Pancreato Biliary Surgery (2018-19)
Goiânia, Goiás, Brazil
Majed A. Refaai, MD
Associate Professor, Pathology and Laboratory Medicine
Director, Hemostasis and Thrombosis Laboratory
Associate Director, Clinical Laboratory Medicine
Rochester, New York
Jeffrey Reha, MD, FACS
Roger Williams Medical Center
Providence, Rhode Island
John F. Renz, MD, PhD, FACS

Section of Transplantation
Chicago, Illinois
Jorge D. Reyes, MD, FACS

Roger K. Giesecke Distinguished Chair
Professor and Chief
Seattle, Washington
Taylor S. Riall, MD, PhD, FACS

Professor, Department of Surgery
University of Arizona
Tucson, Arizona
J. David Richardson, MD, FACS

University of Louisville
Louisville, Kentucky
Robyn Richmond, MD, FACS

Trauma, Surgical Critical Care
Acute Care Surgery
Texas Tech University Health Sciences Center
Lubbock, Texas
John Paul Roberts, MD, FACS

University of California San Francisco
Sean Rogers, MD
University of Missouri-Kansas City School of Medicine
Kansas City, Missouri
Selwyn O. Rogers, Jr., MD, MPH, FACS

Professor and Chief of Trauma and Acute Care Surgery
Director, Trauma Center
Executive Vice President, Community Health Engagement
The University of Chicago Medicine & Biological Sciences
Chicago, Illinois
Vinayak S. Rohan, MD
Charles B. Rosen, MD, FACS

Director, William J. von Liebig Center for Transplantation
& Clinical Regeneration Medical Director
Department of Contracting and Payer Relations
Mayo Clinic
Robert E. Roses, MD
Assistant Professor
Hospital of the University of Pennsylvania
University of Pennsylvania Perelman School of Medicine
Thomas M. Runge, MD
Division of Gastroenterology and Hepatology
University of North Carolina, Chapel Hill
Chapel Hill, North Carolina
Teviah E. Sachs, MD, MPH, FACS

Boston University School of Medicine
Ronald R. Salem, MBChB, FACS

Lampman Professor of Surgery
Chief, Section of Surgical Oncology
Yale University School of Medicine
New Haven, Connecticut
Hrishikesh Samant MD
Assistant Professor of Clinical Medicine
Division of Gastroenterology/ Hepatology
Transplant Hepatology
JCM Transplant Center, WK System
Navdeep S. Samra, MD, FICS, FACS

Associate Program Director-General Surgery Residency Program
Trauma, Acute Care Surgery and Surgical Critical Care
LSU Health Sciences-Shreveport
Dominic E. Sanford, MD, MPHS

Fellow in Hepatobiliary, Pancreatic, & Liver Transplant Surgery
St. Louis, Missouri
Guillermo P. Sangster, MD
Professor of Clinical Radiology and Anesthesiology
LSU Health Sciences Center - Shreveport
Director, Body Imaging Division
Heena P. Santry, MD, MS, FACS

The Ohio State University College of Medicine
Wexner Medical Center
Columbus, Ohio
Michael G. Sarr, MD, FACS

Emeritus J C Masson Professor of Surgery
Mayo Clinic
Amit Sastry, MD
HPB Surgery Fellow
Carolinas HealthCare System
Charlotte, North Carolina
Blayne Amir Sayed, MD, PhD

Administrative Chief Resident
Emory University Hospital
Atlanta, Georgia
Thomas M. Scalea, MD, FACS

Physician in Chief, R Adams Cowley Shock Trauma Center
Francis X. Kelly Professor of Trauma
University of Maryland School of Medicine
Baltimore, Maryland
Bruce David Schirmer, MD, FACS

Stephen H. Watts Professor of Surgery
University of Virginia Health System
Charlottesville, Virginia
Richard D. Schulick, MD, MBA, FACS

The Aragón/Gonzalez-Gíustí Chair
Professor & Chair
University of Colorado Anschutz Medical Campus
Aurora, California
David L. Scott, MD
Director of Kidney and Pancreas Transplantation
Oregon Health and Science University
Portland, Oregon
Aarti Sekhar, MD
Associate Professor of Radiology
Department of Radiology and Imaging Sciences
Atlanta, Georgia
Edgardo Serra, MD
Director of Centro CIEN
Bariatric Surgery, Argentina
Staff of the Percutaneous Image Guided Surgery TEAM of
DIACIM Fundation/IHU-IRCAD
Strasbourg, France
Adil Shah, MD
Surgery Resident
Howard University
Washington, DC
Shimul A. Shah, MD, MHCM, FACS

The James and Catherine Orr Endowed Chair in Liver Transplantation
Section Chief, Transplantation
University of Cincinnati College of Medicine
Cincinnati, Ohio
Junichi Shindoh, MD, PhD

Surgical Section Chief, Liver Disease Center
Toranomon Hospital
Attending Surgeon, Hepatobiliary Surgery Division
Department of Gastroenterological Surgery
Tokyo, Japan
Hosein Shokouh-Amiri, MD, FACS, FICS

Director, Liver Transplant at
WK/John C. McDonald Regional Transplant Center
Co-Director, WK/John C. McDonald Regional Transplant Center
Shailesh V. Shrikhande, MBBS, MS, MD, FRCS (Hon)

Chief, Gastrointestinal and Hepato- Pancreato-Biliary Service
Professor, Surgical Oncology
Tata Memorial Centre
Mumbai, India
Neeraj Singh, MD
Associate Professor of Internal Medicine and Nephrology
Medical Director, Kidney and Pancreas Transplant
John C. McDonald Regional Transplant Center,
Melanie K. Sion, MD
Advanced GI Surgery Fellow
Douglas P. Slakey, MD, MPH, FACS

Chief of Surgery
Christ Medical Center, Advocate Health Care
University of Illinois at Chicago
Chicago, Illinois
Rebecca A. Snyder, MD, MPH

Surgical Oncology Fellow
Houston, Texas
Thomas Soeprono, MD
Director of Transplant Psychiatry
Associate Residency Director
Attending Physician
Psychiatric Consultation and Telepsychiatry Program
Department of Psychiatry
Seattle, Washington
Ponnandai S. Somasundar MD, MPH, FACS

Vice Chairman of Surgery,
Director of Geriatric Oncology Program
Associate Professor of Surgery, Department of Surgery
Roger Williams Medical Center, Boston University
Ho-Young Song, MD, PhD

Divya Sood, MD
Cancer Therapeutics Research Fellow
Moores Cancer Center
University of California, San Diego
David A. Spain, MD, FACS

The David L. Gregg, MD Professor and
Chief of Acute Care Surgery
Stanford University
Zachary E. Stiles, DO
Research Fellow and Resident in Surgery
University of Tennessee Health Science Center
Memphis, Tennessee
Peter G. Stock, MD, PhD

UCSF Department of Surgery
Steven M. Strasberg, MD, FRCS(C), FACS, FRCS(Ed)

Pruett Professor of Surgery
Hepatobiliary-Pancreatic and Gastrointestinal Surgery
St. Louis, Missouri
Lynne Strasfeld, MD
Division of Infectious Diseases
Portland, Oregon
Ashley N. Suah, MD
Chicago, Illinois
Vijay Subramanian, MD
St. Louis, Missouri
Choichi Sugawa, MD, FACS

Michael & Marian Ilitch Department of Surgery
Detroit, Michigan
Avinash N. Supe, MS, FICS

Director (ME & MH) & Dean
Professor, GI Surgery
Professor of Medical Education
Seth GS Medical College KEM Hosp
Director, GSMC FAIMER Regional Institute
Past President Academy of Health Professions Educators
Past President, IHPBA- India
Ex Dean - Sion Hospital
Mumbai, India
Sharven Taghavi, MD, MPH

Fellow, Surgical Critical Care
Brigham and Women’s Hospital
Tze-Woei Tan, MD, FACS

Vascular Surgery
University of Arizona College of Medicine
Tucson, Arizona
Helen S. Te, MD, FAASLD, FAST, AGAF

Professor of Medicine
Medical Director, Adult Liver Transplantation
Center for Liver Diseases
Chicago, Illinois
Ronald Tesoriero, MD, FACS

University of Maryland School of Medicine
Chief of Critical Care
Associate Program Director
Surgical Critical Care & Acute Care Surgery Fellowship
R. Adams Cowley Shock Trauma Center
University of Maryland Medical Center
Baltimore, Maryland
Vinaykumar B. Thapar, MS, DNB

Consultant General & Laparoscopic Surgeon
Dr L H Hiranandani Hospital
Mumbai, India
Miguel Tobon, MD
Wayne State University/Detroit Medical Center
Detroit, Michigan
Jiaywei Tsauo, MD, PhD

Department of Interventional Therapy
National Cancer Center/Cancer Hospital
Chinese Academy of Medical Sciences and Peking Union Medical College
Beijing, China
Jeffrey S. Upperman, MD, FACS, FAAP

Division of Pediatric Surgery
Keck School of Medicine
University of Southern California
Professor Roberto Valiñas MD

Director Surgical Clinic “F”
Director Hepato-Pancreato-Biliary Services Hospital de Clínicas
Montevideo, Uruguay
George Van Buren, II, MD, FACS

Michael E. DeBakey Department of Surgery
Houston, Texas
Sven M. van Dijk, MD

Amsterdam, Netherlands
Thomas M. van Gulik, MD, PhD

Hepato-Pancreato-Biliary Surgeon
Chief of Experimental Surgical Laboratory
Jean-Nicolas Vauthey, MD, FACS

Chief, Hepato-Pancreato-Biliary Section
Houston, Texas
Adarsh Vijay, MBBS
Transplant Surgical Fellow
Washington, DC
Anthony M. Villano, MD
Resident Physician
Medstar Georgetown University Hospital
Washington, DC
Marcelo Vivanco, MD, FACS

Staff Surgeon Clinica Alemana
Santiago, Chile
Brendan C. Visser, MD, FACS

Hepatobiliary & Pancreatic Surgery
HPB Fellowship Program Director
Medical Director, Cancer Center GI Clinical Care Program
Rogier P. Voermans, MD, PhD

Department of Gastroenterology and Hepatology
Michael D. Voigt MB, ChB, M. Med. FCP

Medical Director Liver Failure and Transplantation
University of Iowa Hospitals and Clinics
Iowa City, Iowa
Angela S. Volk, MD
Tulane University School of Medicine
Charles M. Vollmer, Jr, MD, FACS

Director of Pancreatic Surgery
University of Pennsylvania Perelman School or Medicine
Mariano M. Volpacchio, MD
Head of Computed Tomography
Hospital de Clínicas, Universidad de Buenos Aires
Head of Body Imaging
Tyler S. Wahl, MD, MSPH

Birmingham, Alabama
Daniel Reis Waisberg, MD

Clinics Hospital of Sao Paulo University Medical School
(HC-FMUSP)
São Paulo, Brazil
Kerollos Wanis, MD
Resident
Division of General Surgery
Western University
Ontario, Canada
Kenneth Washburn, MD, FACS

Executive Director, Comprehensive Transplant Center
Chief, Division of Transplant Surgery, Department of Surgery
Columbus, Ohio
Jason Wellen, MD, MBA, FACS

Section of Abdominal Transplant Surgery
Surgical Director of Kidney and Pancreas transplantation
St. Louis, Missouri
Scott Westphal, MD
Assistant Professor of Medicine
Omaha, Nebraska
Jared A. White, MD, FACS

Assistant Professor
Liver Transplant and Hepatobiliary Surgery
Birmingham, Alabama
Joshua J. Wiegel, PharmD, BCPS
University of Wisconsin Hospital and Clinics
Madison, Wisconsin
Stephen J. Wigmore, MD, FRCSEd, FRCPEd

Professor of Transplantation Surgery University of Edinburgh
National Clinical Lead MCN HPB Cancers Scotland
Program Director Edinburgh Surgical Sciences Qualification
Vice President (President Elect) British Transplantation Society
Clinical Surgery
University of Edinburgh
Royal Infirmary
Edinburgh, England
Brian H. Williams, MD, FACS

UT Southwestern Medical Center
Dallas, Texas
Mallory Williams, MD, MPH, FACS, FICS, FCCP

Colonel, USAR, MEDCOM
Professor of Surgery & Chief
Division of Trauma, Critical Care, & Surgical Nutrition
Trauma Medical Director
Director of the Surgical ICU
Howard University Hospital
Washington, DC
Robert F. Wilson, MD, FACS, MCCM
Professor (Emeritus) of Surgery
Wayne State University SOM / Detroit Medical Center
Michael and Marian Ilitch Department of Surgery
Detroit, Michigan
John A. Windsor, MD, FRACS, FACS

University of Auckland
Auckland, New Zealand
Steven A. Wisel, MD
UCSF Department of Surgery
Christopher L. Wolfgang, MD, PhD, FACS

Chief, Hepatobiliary and Pancreas Surgery
Professor of Surgery, Pathology and Oncology
Paul K. Neumann Professor of Pancreatic Cancer Research
Member, Miller-Coulson Academy of Clinical Excellence
Johns Hopkins
Baltimore, Maryland
Brent T. Xia, MD
Cincinnati, Ohio
Charles J. Yeo, MD, FACS

8th Samuel D. Gross Professor and Chairman
Senior Vice President and Enterprise Chair
Jefferson Health
Natesh Yepuri, MBBS

Syracuse, New York
Mohammad Y. Zaidi MD, MS

Katz Foundation Research Fellow
Atlanta, Georgia
Roham T. Zamanian, MD, FCCP

Director, Adult Pulmonary Hypertension Service
Palo Alto, California
Manuel E. Zeledón, MD, PhD

Assistant Professor General Surgery
University of Costa Rica
San Jose, Costa Rica
Gazi B. Zibari, MD, FACS

Past President-AHPBA
Director, John C. McDonald Regional Transplant Center
Director, WK Advanced Surgery Center
Willis-Knighton Health System
Nathan R. Zilbert, MD, MEd, FRCS(C)

Hepato-Pancreato-Biliary Surgery Fellow
University Health Network/University of Toronto
Toronto, Ontario
Amer H. Zureikat MD, FACS

Co-Director, UPMC Pancreatic Cancer Center
Nicholas Zyromski, MD, FACS

Associate Professor
Indiana University School of Medicine
P D
1 What Imaging Modality Do I Need to Get to Assess Liver Lesions
and How Do I Interpret Them?
Guillermo P. Sangster, Maren Donato, and Horacio R.V.
D’Agostino
2 How Do I Prepare A Patient For Hepatic/Biliary Surgery, How Much

Future Liver Remnant (FLR) Is Needed In A Healthy Versus Cirrhotic
Liver, and What Steps Should I Take If the FLR Is Inadequate?
Cataldo Doria and Allison David
3 What Is the Ideal Margin Width After Hepatectomy for Colorectal

Metastases and Hepatocellular Carcinoma?
Ibrahim Nassour and Michael A. Choti
4 How Should I Manage Disappearing Colorectal Liver Metastases?

Eliza W. Beal and Timothy M. Pawlik
5 Should I Treat Viral Hepatitis Before Resecting HCC?

Junichi Shindoh
6 How Should I Manage A Patient with Recurrent Colorectal Liver

Metastases and One with Synchronous Colorectal Liver and Lung
Metastases?
Anya Adair and Stephen J. Wigmore
7 Is There a Role for Liver Resection in Patients with Non- Colorectal,
Non-Neuroendocrine Metastases?
Anthony M. Villano and Lynt B. Johnson
8 What is the Utility of Liver Scoring Systems in the Management of

Hepatocellular Carcinoma (HCC)?
Ronald Matteotti, Steven J. Nurkin, and John F. Gibbs
9 Liver Hemangioma, Focal Nodular Hyperplasia, and Hepatic

Adenoma: How Do I Tell Them Apart and When Should They Be
Resected?
Aras Emdadi, Gazi B. Zibari, and Quyen D. Chu
10 NASH, CASH, and Steatohepatitis: What Are the Concerns and How
to Diagnosis Them?
Hrishikesh Samant and Wuttiporn Manatsathit
11 How Should the HPB Surgeon Manage the Patient with Polycystic
Liver Disease?
Zaheer S. Kanji and Adnan A. Alseidi
12 How Do I Diagnose and Manage Amoebic and Pyogenic Liver

Abscess?
Avinash N. Supe and Vinaykumar B. Thapar
13 Should Hepatectomy be Performed in Octogenarians?
Catherine H. Davis, Bradford J. Kim, and Thomas A. Aloia
14 How to Prepare for Major Liver Resection/ Pancreatectomy in a

Jehovah’s Witness Patient?
Edward E. Cho, Houssam Osman, and D. Rohan Jeyarajah
15 Should We Resect Primary Colon Adenocarcinoma for Patients with

Unresectable Liver Metastases?
Jeffrey Reha and Pannandai S. Somasundar
16 Is There a Role for Living Donor Liver Transplantation for

Hepatocellular Carcinoma?
Khalid Khwaja and Robert A. Fisher
17 How Should I Handle Hepatocellular Carcinoma? Resection or

Transplantation?
Alagappan Annamalai and Andrew Klein
P /C D
18 A Patient with Colorectal Cancer and Synchronous Liver Metastases:
A) Simultaneous Resection of Primary and Metastases:
Roberto Hernandez-Alejandro and Kerollos Nashat Wanis
B) Liver Resection First (Reverse Strategy):
Marc-Antoine Allard and René Adam
C) Resect the Primary First:
Somala Mohammed, Steven A. Curley, and Avo Artinyan
19 Neoadjuvant Chemotherapy for Colorectal Liver Metastasis:

A) Pro:
Eddie K. Abdalla
B) Con:
Michael E. Lidsky and Michael D’Angelica
20 Should Patients Undergo Associating Liver Partition and Portal Vein

Occlusion for Staged Hepatectomy (ALPPS) or Two-Stage
Hepatectomy (TSH)
A) ALPPS:
Fernando A. Alvarez and Eduardo de Santibañes
B) Two-Stage Liver Resection:
Takashi Mizuno, Rebecca A. Snyder, and Jean-Nicolas Vauthey
21 A Patient with Liver Metastases from Neuroendocrine Tumor

A) Resection:
Amar Gupta and Elijah Dixon
B) Transplantation:
Wellington Andraus, Daniel Reis Waisberg, and Luiz Augusto
Carneiro D’Albuquerque
22 What Approach is Best for Hepatic Hydatid Cyst?

A) Percutaneous:
Okan Akhan, Turkmen Ciftci, and Devrim Akinci
B) Surgical:
Fernando Duek and Javier C. Lendoire
I D
23 How to Manage Tumors Involving the Retrohepatic Inferior Vena
Cava
Alan W. Hemming, and Jennifer Berumen
24 What Are the Different Techniques in Liver Parenchymal Dissection

and Transection?
Vincent E. de Meijer and Bas Groot Koerkamp
25 What Are the Different Vascular Control Techniques for Liver

Resection and On Whom Should I Use Them?
François Cauchy and Jacques Belghiti
P D
26 What Are the Post-Hepatectomy Complications and How Should I
Manage Them?
Miguel-Angel Mercado and Hiram Raul Muñoz Gonzalez
P D
27 How to Avoid Injuring the Biliary Tree During Laparoscopic
Cholecystectomy and How to Manage it When it Does Happen?
Dominic Sanford and Steven M. Strasberg
28 How Should I Manage Retained Common Bile Duct Stone(s) in

Patients With and Without A Previous Gastric Bypass?
Gazi B. Zibari and Quyen D. Chu
29 How Should I Manage Intrahepatic Biliary Stones?
Henry A. Pitt, Janak Parikh
30 How Should I Approach a Cirrhotic Patient with Acute Cholecystitis?

Jad Abou-Khalil and Scott Helton
31 How to Manage a Gangrenous Gallbladder. What To Do When the

Biliary Anatomy Is Unclear?
Felipe José Fernández Coimbra and Maria Luiza Leite de
Medeiros
32 How Should I Manage Gallstone-Related Acute Cholangitis?

Ronald S. Mowad and Quyen D. Chu
33 How Do I Diagnose and Manage Patients with Acute Acalculous

Cholecystitis (AAC)?
Ajay Jain, Natesh Yepuri, and Dilip S. Kittur
34 How to Diagnose and Manage Biliary Dyskinesia?

James McClintic and Kimberly Brown
35 How to Manage Cystic Duct Remnant and Post-Cholecystectomy

Syndrome?
Lucio Lucas Pereira and Horacio J. Asbun
36 How Should I Manage a Patient with a Choledochal Cyst?

Danielle M. Fritze and Francisco G. Cigarroa
37 Should a Cholecystectomy Be Done For Gallstone Ileus?

Marcelo Fasano and Alberto R. Ferreres
38 How to Diagnose and Manage Caroli’s Disease and Mirizzi’s

Syndrome?
Nathan Zilbert, Eisar Al-Sukhni, Shiva Jayaraman, and Sean
Cleary
39 What Should I Do with An Incidentally Detected Gallbladder Polyp?

Mohammad Y. Zaidi, Aarti Sekhar, and Shishir K. Maithel
40 Do All Porcelain Gallbladders Need Surgery and, If So, Do I Need to

Perform a Radical Cholecystectomy at the Onset?
Rachel E. Beard and David A. Geller
41 How to Differentiate Biliary Cystadenomas from

Cystadenocarcinomas and How Should They be Managed?
Doug P. Slakey and David T. Pointer, Jr.
42 Who to Resect for Gallbladder Cancer and How Extensive Should

the Resection Be?
Julie N. Leal and Brendan C. Visser
43 How Do I Diagnose and Treat Intrahepatic Cholangiocarcinoma?

Zeljka Jutric and Yuman Fong
44 Is Staging Laparoscopy Necessary for Potentially Resectable

Perihilar Cholangiocarcinoma?
Robert J.S. Coelen and Thomas M. van Gulik
45 Should the Caudate Lobe Be Routinely Resected for Hilar

Cholangiocarcinoma?
Masato Nagino
46 Should a Feeding Tube Routinely Be Placed For Patients Undergoing

Hepato-Pancreato-Biliary Operations?
Austin Dosch and Aram N. Demirjian
47 Is Biliary Drainage of the Future Liver Remnant Important After Liver

Resection for Perihilar Cholangiocarcinoma?
Laura M. Enomoto and Niraj J. Gusani
48 A) Portal Vein Is Involved In a Patient with Hilar

Cholangiocarcinoma: Proceed with Surgery and Resect Portal Vein:
Palepu Jagannath and Gunjan Desai
B) What Is The Best Approach For Hilar Cholangiocarcinoma With
Portal Vein Involvement?
Michelle Babicky and William R. Jarnagin
P D
49 How Should I Manage Postoperative Bile Leak?
Blayne A. Sayed and David A. Kooby
P D
50 A Patient Needs a Pancreatectomy But Has A Cardiac History
Demanding Anticoagulation Therapy. How Do You Manage This?
Zachary Brown and Jonathan Hernandez
51 Is There a Role for the Surgeon in the Management of Locally-

Advanced Pancreatic Cancer?
Michael D. Kluger and John A. Chabot
52 A Patient Has a Resectable Pancreatic Head Cancer But is 83 Years

Old. Should I Perform a Whipple?
Sandeep Jhajj and Taylor S. Riall
53 The Patient Presents with a Pancreatic Head Mass with Jaundice and
a Bilirubin of 20. Should You Operate Soon or Stent the Bile Duct?
Tyler S. Wahl and John D. Christein
54 A Resectable Pancreatic Head Tumor Invades the Transverse

Mesentery and Would Require A Hemicolectomy. Is It Worth the
Added Risk?
Joal D. Beane and Eugene P. Ceppa
55 A 25-Year-Old Female Presents with a Story of “Chronic

Pancreatitis”. The Pancreas is Radiographically Normal. Should You
Perform a Total Pancreatectomy with Islet Autoransplantation?
Katherine Morgan
56 A 55-Year-Old Patient Has Chronic Pancreatitis With a Head-

Dominant Disease. What Operation is Best – Beger, Frey, Whipple,
or Puestow?
Brent T. Xia, Vikrom K. Dhar, and Syed Ahmad
57 A Patient Has a 2.0 cm Side-Branch IPMN and is 45 Years of Age.

Should I Operate or Follow?
Georgios Gemenetzis and Christopher L. Wolfgang
58 What Are Grove Pancreatitis and Pancreatic Divisum and How Do I

Deal With Them?
David B. Adams and Andrew Hardie
P /C D
59 A 1.5 cm Neuroendocrine Tumor is Identified in the Pancreatic Head.
Should I Operate or Not?
A) The Case for Resection:
Catherine H. Frenkel and John D. Allendorf
B) The Case for Observation:
Alessandra Pulvirenti and Peter J. Allen
60 A Patient Presents with a Pancreatic Adenocarcinoma Which Is In

Apposition with the Porto-splenic Confluence. Should I Operate Now
or Later?
A) The Case for Up-Front Surgery:
Melanie K. Sion and Charles J. Yeo
B) The Case for Preoperative Therapy:
Jordan M. Cloyd and Matthew H.G. Katz
I D
61 A Pancreatic Cancer at the Portal Confluence Requires a Vein
Resection. How Do You Reconstruct the Vein – Primary Repair,
Venous Conduit, Synthetic Graft?
Jordan M. Cloyd and Jeffrey E. Lee
62 How to Manage the High-Risk Pancreatic Anastomosis?

Brett L. Ecker and Charles M. Vollmer, Jr.
63 Should I Place an Intraoperative Drain for Distal Pancreatectomy?
Zachary E. Stiles and Stephen W. Behrman
64 Should You Get Frozen Sections During a Whipple for Pancreatic

Cancer or IPMN?
Jashodeep Datta and Jeffrey A. Drebin
65 Should the Spleen Be Preserved in a Distal Pancreatectomy?

William P. Lancaster and Attila Nakeeb
66 A Single Metastatic Focus is Identified in the Liver During a Planned

Resection for Pancreatic Malignancy? Should I Proceed?
Alessandro Paniccia and Richard D. Schulick
67 You Find Locally Unresectable Disease During an Exploration for a

Periampullary Cancer. How Do You Best Palliate?
Motaz Qadan and Keith D. Lillemoe
68 On Whom Should You Perform Local Excision Versus

Pancreatoduodenectomy for An Ampullary Mass?
Andre Luis Montagnini and Thiago Nogueira Costa
P /C D
69 Should I Place an Intraoperative Drain for Whipple’s Resection?
A) The Case for Drainage:
George van Buren 2nd and William Fisher
B) The Case Against Drain Use:
John W. Kunstman and Ronald R. Salem
P D
70 A Septic Patient with a Completely Dehisced Pancreatico-
jejunostomy (ISGPF Grade C) Requires a Reoperation on POD#7.
How Do You Manage This?
Rachel E. Beard and Mark P. Callery
71 How Should I Approach Post-Pancreatectomy Hemorrhage?

Major K. Lee, IV
72 A Patient Presents with An Intraluminal Hemorrhage After Whipple.

What Do I Do Now?
Nisa Kubiliun and Rebecca M. Minter
73 A Patient Has a Chyle Leak Following a Pancreatic Resection. How

Do You Manage This?
Carlos Chan and Gabriela Del Angel
74 There Is Bile In the Intraoperatively Placed Drain Early After A

Whipple Resection. What Should You Do?
Jason B. Liu and Marshall S. Baker
75 When Should a Patient Who Requires a Splenectomy for a Distal

Pancreatectomy Be Vaccinated?
Na Eun Kim and Teviah E. Sachs
76 A Patient Undergoes a Pancreatectomy for Pancreatic Cancer. Should

They Receive Extended Anticoagulation Therapy Beyond Discharge?
Elizabeth A. Bailey and Robert E. Roses
77 A Stable Patient POD#7 From a Whipple is Ready for Discharge But

Has a High Likelihood for Readmission. What Should You Do?
Zhi Ven Fong and Carlos Fernández-del Castillo
78 How Do I Manage a Patient Who Has “Failure to Thrive” Following

a Pancreatic Resection?
Rachel E. Beard and Tara S. Kent
P /C D
79 Should I Use a Somatostatin Analogue After a Whipple Resection?
A) Pro:
Shailesh V. Shrikhande and Esha Pai
B) Con:
Divya Sood and Andrew Lowy
LIVER TRANSPLANTATION
P D
80 How Should I Optimize a Patient With Acute Cholecystitis and End-
Stage Liver Disease?
Jared A. White and Devin Eckhoff
81 Management of Hepatorenal Syndrome– When to Start Hemodialysis

and When is a Patient a Candidate for Simultaneous Liver Kidney
Transplant?
Eric Langewisch and Scott Westphal
82 How Should One Optimally Treat a Patient with Hepatocellular

Cancer Who is Waiting for a Liver Transplant?
Angela S. Volk and Joseph F. Buell
83 How Should Simultaneous Liver Kidney (SLK) Waitlisted Patient be

Managed? What Is The Optimal Timing of Transplanting the Kidney?
Dave L. Scott
84 What Are the Protocols for Downstaging Hepatocellular Carcinoma

(HCC) to Within Milan Criteria, Transplantation Outside of Milan
Criteria? – Pros and Cons of Center Specific Approaches
Adeel S. Khan, Kathryn J. Fowler, and William C. Chapman
85 Management of Hepatic Hydrothorax on the Waitlist – Pleural Tube

or Not? TIPS or Not? Indications and Strategies
Michael D. Voigt and Arvind Murali
86 How Should I Manage Patients with Hepatopulmonary Syndrome and

Portopulmonary Hypertension Who Are On The Waitlist?
Amy Gallo, Roham T. Zamanian, and Waldo Concepcion
87 How Should I Manage Spontaneous Bacterial Peritonitis While the

Patient Is On The Waitlist? When to Activate for Liver
Transplantation?
Helen S. Te
88 How Should I Manage a Leaking Umbilical Hernia in High MELD

Wait Listed Patients?
Devin E. Eckhoff and Jared A. White
89 Obesity In The Liver Transplant Candidate – What Is The Ideal
Strategy?
Julie K. Heimbach
90 How Much Steatosis To Tolerate In The Donor Allograft?

Arun P. Palanisamy, Vinayak S. Rohan, and Kenneth D. Chavin
91 To Anticoagulate or Not in the Setting of a Waitlisted Patient with

Portal Vein Thrombosis and Varices. What Agent to Use?
Michael Loudin and Joseph Ahn
92 Viral Hepatitis: How Should I Manage HBV in the Waitlisted Patient?

James Fung and Marina Berenguer
93 The Management of Hepatitis C Before and After Liver

Transplantation: When and When Not to Treat, and How to Treat
When Treatment is Indicated?
Richard Gilroy
94 How Should One Evaluate and Manage a Liver Transplant Candidate

Who Also Has A Substance Use Disorder?
Thomas Soeprono, Sara Houck, and Stacey McCandlish
95 When Is A Patient Too Sick to Transplant? Keeping the High MELD

Patient Alive While Awaiting for Transplant?
Yumi Ando and Willscott Naugler
96 How Can Donor-Recipient Matching Optimize Donation After

Circulatory Death (DCD) Liver Utilization?
C. Kristian Enestvedt
97 What Are The Indications for Living Donor Liver Transplant

(LDLT)?
Kendra D. Conzen and Elizabeth A. Pomfret
98 How Do You Decide Whether to Use the Right Lobe or Left Lobe in a
Live Donor Liver?
Dong-Hwan Jung and Sung-Gyu Lee
99 What Are The Indications for Split Liver Transplantation, When to

Use Both Livers; Right Lobe/Left Lobe, Left Lateral Section/Right
Trisegment, etc?
Esteban Halac and Javier Lendoire
100 How Should We Match Donor Organ Quality with Recipient Severity
of Illness?
Kenneth Washburn and Sylvester M. Black
101 Machine Perfusion Techniques to Enhance Quality of Marginal Liver

Organs and to Prolong Ischemia Times Without Deleterious Effects –
Which Approach is Best: Normothermic vs Hypothermic?
David Nasralla and Peter J. Friend
102 Pediatric Liver Transplantation – What Are The Indications, Timing,

Technical Aspects, and Complications?
Oscar Imventarza and Guillermo Cervio
103 Should Transplantation Be Considered for Patients with

Unresectable Cholangiocarcinoma?
Charles B. Rosen
104 Is There a Role for Liver Transplantation for Metastatic

Neuroendocrine Tumors?
John F. Renz, Geraldine C. Diaz, and Christoph E. Broelsch
I D
105 Management of Hepatorenal Syndrome, Renal Insufficiency
Intraoperatively-Intraoperative CRRT?
Dragos M. Galusca and Marwan S. Abouljoud
106 How Should I Address Anastomotic Size Mismatches Between

Donor and Recipient: (A) Bile Ducts, (B) Portal Veins, (C) Hepatic
Arteries, (D) Hepatic Vein/Caval Anastomoses?
Hillary J. Braun and John P. Roberts
107 What Are The Technical Aspects of Dealing with Anastomotic

Mismatches in Liver Transplantation?
Martin I. Montenovo and Jorge D. Reyes
108 Options For Aortohepatic Conduits in Liver Transplantation: When

to Use What?
Peter T.W. Kim and Goran B. Klintmalm
109 Biliary Reconstruction for Liver Transplantation in Patients with

Primary Sclerosing Cholangitis: Which Approach is Best -
Choledochocholedocostomy or Roux-en-Y Hepaticojejunostomy?
Kyle R. Jackson and Andrew Cameron
110 What Are The Technical Strategies in Retransplantation? How to

Explant the Liver? Anastomotic Variants and Approaches?
Flavio Paterno, T. Mark Earl, and Shimul A. Shah
111 When to Use Venoveno Bypass? (Rarely Needed- But There Are
Some Indications)
Shannon Orr and Christopher D. Anderson
112 How to Use Thromboelastography to Guide Blood Product Usage in

Liver Transplantation?
Mark S. Orloff, Stewart Jay Lustik, and Majed A. Refaai
113 What Are the Intraoperative Coagulation Issues During Liver

Transplantation: An Anesthesiologist’s Perspective
Emily Baird and Elizabeth Duggan
114 What Are the Issues in Intraoperative Management of Patients

Undergoing Liver Transplantation: An Anesthesiologist’s Perspective
Emily Baird and Elizabeth Duggan
115 How Should I Manage Intraoperative Pleural Effusions During Liver

Transplantation: Transdiaphragmatic vs Chest Tube Drainage?
T. Mark Earl and Siddharth Desai
116 Is tPA Protocol with Infusion into the Recipient Hepatic Artery in
Donation After Circulatory Death (DCD) Livers to Prevent Ischemic
Cholangiopathy Useful?
Nicolas Goldaracena and Paul D. Greig
P D
117 How Should I Manage Hepatic Artery Thrombosis After Liver
Transplantation?
Mary Decoteau, Jamie Case, and Christopher Lee Marsh
118 How to Manage Portal Vein Thrombosis in Liver Transplantation?

George E. Loss, Jr. and David S. Bruce
119 How Should I Manage Early Venous Outflow Obstruction in Liver

Transplantation?
C. Kristian Enestvedt and Erin Maynard
120 How Should I Manage Biliary Strictures and Leaks After Liver
Transplantation?
Juan F. Guerra and Alberto Espino
121 What Are The Indications and Timing for Liver Retransplantation?
Erin C. Maynard
122 What Should I Look for in a Liver Transplant Recipient Who Has a
Fever?
Sui Kwong Li and Lynne Strasfeld
123 What Are the Common Perioperative Cardiac Events During and
After Liver Transplantation?
Elliot I. Grodstein and David P. Foley
124 When Should You Repair an Umbilical or an Inguinal Hernia in a

Cirrhotic or a Post-Liver Transplant Patient?
Vijay Subramanian and Majella Doyle
125 What Are the Updates on Immunosuppression in Liver

Transplantation?
Joshua J. Wiegel
126 What Are the Complications and Management of Immunosuppressive

Therapy in Liver Transplant Recipients?
Domenica McKenna, and Ali J. Olyaei
127 What Are the Indications for Early Reoperation Following Liver
Transplantation?
Ashley Suah, Adam S. Bodzin, and J. Michael Millis
PANCREAS TRANSPLANTATION
128 Is There a Role for Islet Cell Transplantation in Patients with Type I
Insulin-Dependent Diabetes Mellitus? Should Islet Cell Transplant Be
Considered for Type II Diabetes Mellitus Patients?
Steven A. Wisel and Peter G. Stock
129 When Should Pre-emptive Solitary Pancreas Transplant Be

Considered for a Patient with Type I Diabetes Mellitus?
Adarsh Vijay, Matthew Cooper, and Peter Abrams
130 What is the Utility of Simultaneous Pancreas Kidney Transplantation

in Patients with Advanced Diabetic and Kidney Disease?
Dixon B. Kaufman
131 Who Should Get a Pancreas Transplant, Which Technique and How
to Manage Its Rare Complications?
Hosein Shokouh-Amiri and Gazi B. Zibari
132 How Should One Manage Pancreatic/Duodenal Leak after

Transplantation: Observation, Stenting, Percutaneous Drainage or
Surgical Correction?
Leigh Anne Dageforde and Jason Wellen
133 A Patient is Two Years Post Pancreas Transplant with Bleeding from
the Donor Arterial Conduit Aneurysm. What is the Role of Surgery
and Interventional Radiology?”
Armando Ganoza and Abhinav Humar
134 A Pancreas Transplant Recipient Has Acute Allograft Dysfunction:

What is the Differential Diagnosis and How Should It Be Managed?
Neeraj Singh, Aleksandra Kukla, and Jon S. Odorico
HEPATIC TRAUMA
135 Is There a Role for Non-Operative Management of Patients with

Right Upper Quadrant Penetrating Trauma?
Adil Shah and Edward E. Cornwell, III
136 Is There a Role for Operative Management of Blunt Hepatic Trauma

in the “Nonoperative Management Era”?
Matthew Benns and J. David Richardson
137 How Do We Best Manage Traumatic Retrohepatic Inferior Vena

Cava Injury?
Alfred E. Baylor, III and Robert F. Wilson
138 A Patient with Non-Expanding Hematoma from Vena Caval Injury

Due to Penetrating Injury. Is There a Role for Percutaneous Stenting?
Sean Rogers and Farzad Alemi
139 What is the Role of Vascular Shunting for Major HPB Trauma?
Chad G. Ball and Amar Gupta
140 What is the Role of Hepatic Transplantation for Hepatobiliary
Trauma?
Rupak D. Kulkarni and Jeffrey H. Fair
141 How Should We Manage Gallbladder Injury Concomitant with

Hepatic Trauma?
Jon D. Dorfman and Heena P. Santry
142 What Are the Different Hemostatic Techniques in Hepatic Trauma?

Lenworth Jacobs
143 How Do We Manage Hepatobiliary Injuries & Their Complications?

LD Britt
PANCREATICODUODENAL TRAUMA
144 How Do We Manage Trauma to the Distal Pancreas: Is Splenectomy

Always Required?
Wayne A. I. Frederick
145 Operative Management of Penetrating Trauma to the Pancreatic Head

and Duodenum: Is There Still a Role for Pancreaticoduodenectomy?
Charles E. Lucas, and Anna M. Ledgerwood
146 How Do We Manage the Challenging Duodenal Injury Between the

Ampulla and the Vessels?
Anna M. Ledgerwood, Miguel Tobon, and Charles E. Lucas
147 How Should We Manage Penetrating Pancreatic and SMA Trauma?

Rebecca Plevin and Andre Campbell
148 What is the Role of Non-operative Management in Pancreatic

Trauma?
Navdeep S. Samra, Piyush Kalakoti, and Mallory Williams
149 Is There a Role for Endoscopic Retrograde

Cholangiopancreatography (ERCP) in Pancreatic Trauma?
Choichi Sugawa and Charles E. Lucas
150 How Do We Diagnose and Manage the Complications of Pancreatic

Trauma?
Robyn Richmond, Mallory Williams, and Brian H. Williams
INJURY TO THE PORTA HEPATIS
151 How Do We Manage Penetrating Trauma to the Extrahepatic Biliary

Tree?
Kimberly Joseph and Lee Cartagena
152 Operative Management of Portal Vein/Superior Mesenteric Vein

Trauma: Can We Just Ligate?
Tze-Woei Tan and Alik Farber
153 Hepatic Artery Trauma: Repair or Ligate?

Ronald Tesoriero and Thomas Scalea
154 How Do We Diagnose and Manage Complications of Biliary

Trauma?
Lisa M. Knowlton and David A. Spain
SPECIAL TOPICS
155 How Do We Manage Hepatobiliary and Pancreatic Trauma in the

Damage Control Era?
Mallory Williams
156 How Do We Manage HPB Injuries in the Perioperative Period?

Sharven Taghavi, Reza Askari, Mallory Williams, and Selwyn O.
Rogers, Jr
157 Improving Quality Through Integration of the Hepatobiliary and

Transplant Surgeons Into The Trauma Team: How To Achieve The
Goal?
Mallory Williams and Wesley Francis
158 What Are The Special Considerations in the Management of

Pediatric HPB Trauma?
Jamie Golden, Wesley Barry, Henri Ford, and Jeffrey S.
Upperman
159 What is the Role for Extracorporeal Membrane Oxygenation

(ECMO) in Improving Survival After Hepatic Trauma?
Mallory Williams, Jenna Aziz, and Salim Aziz
P D
160 Will Technology, Along with Proper Training and Teaching, Improve
Patient Outcome?
Jorge Cardoso Cuneo and Andrés Garelli
161 What is the Future of Surgery?

Mariano E. Giménez
162 How Many Classifications Are There for Acute Pancreatitis &
Which One Should We Be Using?
Michael G. Sarr
163 Is Antibiotic Therapy Currently the First Step in the Treatment of

Infected Acute Necrotic Collections of the Pancreas?
Savio G. Barreto and John A. Windsor
164 A Patient With Severe Acute Pancreatitis and Necrosis Shows

Systemic Inflammatory Response Syndrome (SIRS) 9 Days After the
Onset of Pancreatitis - What Should We Do?
Sven M. van Dijk, Rogier P. Voermans, and Marc G.H. Besselink
165 When is Open Debridement Indicated for Necrotizing Pancreatitis?

Nicholas J. Zyromski
166 There is a Well-Formed Large Pseudocyst With a Thickened Wall.

What Should be Done?
Gustavo Andreoli, Roberto Valinas, and Mario Almada
167 A Patient with Hemorrhagic Pancreatitis Who Has Failed

Embolization. What is the Next Step?
Gustavo Kohan and Ornella Agnes Ditulio
168 Should Percutaneous and Endoscopic Therapies be Performed by

HPB Surgeons?
Mariano Palermo and Mariano E. Giménez
169 A Patient Requires a Whipple Resection-Should I Do This Using a

Traditional Open Approach or Robotically?
Mashaal Dhir and Amer H. Zureikat
P /C D
170 Preoperative HPB Biopsy:
A) Rarely Needed:
Perparim Limani, Claire Goumard, and Claudius Conrad
b) Always:
Claudemiro Quireze, Jr
171 Preoperative Percutaneous or Endoscopic Biliary Drainage for

Klatskin and Pancreatic Cancer?
A) Yes:
Vagner B. Jeismann and Paulo Herman
B) No:
Sergio Pacheco and Nicolas Jarufe
I /T D
172 Image Fusion, Virtual Reality and Hybrid Imaging in HPB Surgery -
Landmark Innovations or New-Fangled Technology?
Michele Diana and Jacques Marescaux
173 Single-Port Hepato-Pancreato-Biliary Surgery: Myth or Reality?

Xabier de Aretxabala and Marcelo Vivanco
174 Do Duodenal Stents Replace Palliative Surgery In Advanced

Pancreatic Cancer?
Jiaywei Tsauo and Ho-Young Song
175 Which Type of Local Ablation Should We Use for Treating Liver
Cancer: Ethanol, RFA, Microwave Ablation, Cryoablation or
Irreversible Electroporation?
Mariano E. Giménez and Manuel E. Zeledón
P /C D
176 Palliative Treatment of Malignant Jaundice - Percutaneous vs.
Endoscopic & Echoendoscopic Approaches?
A) Percutaneous:
Eduardo Houghton
B) Endoscopic:
Thomas M. Runge and Todd H. Baron
177 Should Patients Undergo Radiofrequency Ablation (RFA) or

Surgical Resection for Small Hepatocellular Carcinoma (HCC)?
A) RFA:
Marcos R. Meneze, Guilherme L. Martins, and Maurício R.
Liberato de Moura
B) Resection:
Oscar C. Andriani
178 Ablative Therapy: Percutaneous or Laparoscopic?

A) Case for Percutaneous:
Pablo Acquafresca
B) Case for Laparoscopic:
Amit Sastry and David Iannitti
P D
179 Bioethics - Implications of Use of Technology - Where Are We
Going?
Bruce Schirmer
180 Follow-up images - When and What? The Case of Gallbladder

Polyps
Mariano Volpacchio and Christine Menias
181 Laparoscopic, Endoscopic, or Percutaneous Techniques for

Postoperative Complications from HPB Surgery: Pros and Cons
Edgardo Serra and Caetano Finger
BACKGROUND
Each year approximately 10% of the 6 million people in the United States on
chronic antithrombotic therapy will require an operation or invasive
procedure(s), with the most common indications for anticoagulation or
antiplatelet medication being chronic atrial fibrillation, the presence of a
mechanical heart valve, or prior venous thromboembolism. When evaluating
a patient on chronic antithrombotic therapy for operative candidacy, the risks
of hemorrhage and thrombosis, including all attendant consequences, must be
taken into account. Although pancreaticoduodenectomy is considered a high-
risk procedure, it is rarely if ever an urgent undertaking, thereby permitting
appropriate planning. The goal of minimizing thromboembolic events and
major hemorrhage is best achieved by a careful evaluation of each individual
patient in concert with the appropriate consultants, particularly considering
the expanding armamentarium of available antithrombotic agents that are
neither evaluable in real-time nor readily reversed. We have therefore
chosen to present our approach to the management of patients on
antithrombotic therapy requiring a pancreaticoduodenectomy for cancer in
case scenarios likely to be encountered by the practicing surgeon.
Before we begin, a cautionary note regarding postpancreatectomy
hemorrhage and the profound implications it carries. Although there is no
available evidence to suggest that anticoagulation status will influence the
development of pancreas-specific complications, it is, however, conceivable
that full anticoagulation may precipitate a hemorrhagic event in a patient that
otherwise develops a pancreatic fistula or an intra-abdominal infection. It
should be noted that postoperative hemorrhage will occur in 5-8% of patients
undergoing pancreaticoduodenectomy, and its occurrence is associated with
significantly increased rates of both morbidity and mortality.
Postpancreatectomy hemorrhage can be classified based on the time of
occurrence in relation to the resection. While early hemorrhage occurs within
24 hours and is most commonly caused by technical failure, late bleeding
occurs greater than 24 hours and is typically secondary to a complication of
the operation such as pancreatic fistula or abscess. Moreover, about 1/5 of
patients who develop postpancreatectomy bleeding will do so after
discharge. Consideration therefore must be given to the individual’s risk for
procedure-related adverse events in general, underscoring the importance of
the surgeon’s role in decision-making with the consultants, who will have
little, if any, understanding of the complications associated with pancreatic
surgery.
CASE SCENARIO
A 68-year-old man is referred by his medical oncologist after a recent
pathologic diagnosis of pancreatic adenocarcinoma obtained during ERCP
and stenting for progressive jaundice. The patient has a history of
hypertension and rheumatic heart disease that required a mitral valve
replacement 8 years prior and is currently being treated with warfarin 5mg
daily. His most recent INR was 3.0. On review of the CT scan, the tumor
appears to be localized to the head of the pancreas with clear fat planes
surrounding the major vascular structures.
Management
The patient described will need to be seen in consultation with his
cardiologist, as well as by a staff hematologist who will continue to follow
throughout the perioperative period. The patient and consulting physicians
will need to be made aware of the potential complications and, specifically,
any complication that may require procedural intervention that could
necessitate delay or full reversal of anticoagulation. Given this patient’s
mechanical mitral valve, he is considered high-risk for cessation of
anticoagulation (see Table 1 ) and should therefore receive bridging therapy
with subcutaneous low molecular weight heparin. Warfarin should be
stopped 5 days prior to the operation. INR should be checked during the
ensuing days and corrected accordingly the day prior to resection with
vitamin K to a goal around 1.5.
TABLE 1: Overview of Risk Evaluation and Need for
Bridging Therapy. TIA: Transient Ischemic Attack, VTE:
Venous Thromboembolism, a-fib: atrial fibrillation, HTN:
hypertension, CHF: Congestive Heart Failure, CHADS2:
CHF, HTN, Age >75, Diabetes Mellitus, Stroke
Following completion of the procedure, prophylactic subcutaneous

heparin (we prefer heparin over other low molecular weight formulations in
the immediate postoperative period) should begin at 8 hours and continue for
48-72 hours. Given the high risk of thrombotic events, full anticoagulation
should be reinstituted after this two-to-three-day window. Delayed bleeding
events are usually associated with pancreatic anastomotic leaks and/or intra-
abdominal infections, which, when accompanied by hemorrhage, carry
significantly increased rates of morbidity and mortality. We therefore suggest
hypervigilance for early detection of post-operative complications, with
adjustment made accordingly and in conjunction with the consultants. Upon
discharge, our preference is to continue therapeutic low molecular weight
heparin and avoid early resumption of warfarin for primarily two reasons.
First, nearly all patients experience some form of altered eating habits and
diet modifications that can make consistency with attaining a goal INR
challenging. Second, discharge from the hospital does not render the patient
immune from postoperative complications, although the likelihood certainly
decreases with increasing time from the date of resection. It should be noted,
however, that these decisions should not be made unilaterally and should be
made in conjunction with the consulting hematologist and cardiologist.
DISCUSSION
Warfarin, a vitamin K antagonist, inhibits the synthesis of clotting factors II,
VII, IX, and X, as well as protein C and S. It was originally utilized as a
rodenticide in the 1950s and remains one of the most common
anticoagulation medications encountered in clinical practice today. The name
Warfarin was derived from its founding organization, the Wisconsin Alumni
Research Foundation, and “arin” from a similar molecule, coumarin.
Coumarin is present in high quantities in rotten sweet clover and was
responsible for the hemorrhagic deaths of large numbers of grazing cattle in
the 1920s. For patients requiring a Whipple procedure, or any major
operation, warfarin should be stopped approximately 5 days before the
operation, with a goal INR of about 1.5 at the start of the procedure. Vitamin
K should be administered as appropriate, with fresh-frozen plasma reserved
for immediate perioperative use.
The need for bridging anticoagulation after warfarin cessation is
dependent upon the risk of developing adverse cardiovascular and/or
thrombotic events. High, moderate, and low risk refers to the rate of
occurrence of acute thromboembolism per year in absence of anticoagulation
and is approximated at >10%/year (high), 5-10%/year (moderate, and
<5%/year (low). High- risk patients should undergo bridging therapy after
stopping antithrombotic medication, whereas low-risk patients typically do
not require anything beyond cessation of the medication. In patients
determined to be of moderate risk for perioperative acute thromboembolism,
there are no clear guidelines, and the decision should be made on an
individual basis. This information is summarized in Table 1 .
The incidence of prosthetic valve thrombosis in patients not appropriately
anticoagulated is approximately 2% per patient-year, whereas the incidence
of any thromboembolic event resulting in death, stroke, or peripheral
ischemia is 4% per patient-year. Of note, thromboembolic events following
aortic valve replacement are significantly lower than that of mitral valve
replacement. Patients with mitral valve prostheses require bridging
anticoagulation, which typically consists of subcutaneous low molecular
weight heparin at a dose of 1 mg/kg twice a day of actual (not ideal) body
weight. The last dose of low molecular weight heparin should be given 24
hours before the operation, or if the patient is admitted to the hospital and
receiving intravenous heparin as a bridge, the continuous infusion should be
stopped 4-6 hours before surgery.
CASE SCENARIO 2
A 62-year-old woman presents with a pathologic diagnosis of
adenocarcinoma of the head of the pancreas obtained during endoscopic
ultrasound. The patient’s CT scan reveals clear fat planes surrounding all
major vascular structures and a small hypodense mass in the uncinate
process. She has a history of atrial fibrillation and is being treated with
rivaroxaban (Xeralto). In addition, the patient has hypertension and type II
diabetes mellitus but no congestive heart failure, and has never suffered
symptoms of a transient ischemic attack or stroke.
Management
The patient described will need to be seen in consultation with her
cardiologist. As she has normal renal function, her anticoagulation
medication should be held ≥1 day prior to surgery. The half-life of
rivaroxaban is about 7 hours and there are no active metabolites. Her
CHADS2 score is two, with points assigned for hypertension and diabetes
mellitus. This CHADS2 score places her at low risk for a perioperative
cardiovascular event. No bridging therapy is required (Table 1 ).
Following completion of the procedure, prophylactic subcutaneous
heparin (we prefer heparin over other low molecular weight formulations)
should begin at 8 hours and continue throughout the hospitalization. This
patient is at low risk for perioperative events, and we believe patients with a
similar risk profile should generally not be restarted on their antithrombotic
therapy until the time of the first postoperative clinic visit. Moreover, this
patient may be at higher risk for a postoperative pancreatic fistula given the
size and location of her tumor, meaning she may require additional
postoperative interventions. In general, patients with low-risk profiles, such
as the patient presented in this scenario, will not require prophylactic
heparin therapy at home. This decision should of course be made in concert
with the consulting physicians.
DISCUSSION
Until recently the mainstay of anticoagulation therapy has been warfarin, but
over the past several years, new oral anticoagulant medications have
emerged. Both the intrinsic and extrinsic clotting cascades converge to
generate activated factor Xa, making it a potent target. Factor Xa, along with
factor Va, form the active prothrombinase complex. The prothrombinase
complex generates thrombin (from prothrombin), and ultimately the platelet-
fibrin thrombus. The new oral anticoagulation medications are summarized in
Table 2 .
TABLE 2: Overview of Anticoagulation Medications. FFP:

Fresh Frozen Plasma, PCC: Prothrombin Complex
Concentrate, INR: international normalized ratio, PT:
Prothrombin Time, PTT: Partial Thromboplastin Time
The new oral anticoagulant medications have a much broader therapeutic

window as compared to warfarin, allowing for fixed dosing in adults without
monitoring. These oral factor Xa inhibitors have been found to be non-
inferior to warfarin for the prevention of stroke and systemic embolism in
patients with atrial fibrillation in multiple trials (RE-LY, ROCKET-AF, and
ARIS-TOTLE). Additionally, rivaroxaban is approved in the prevention of
venous thromboembolism, based on results of the RECORD trials in patients
following orthopedic procedures, and has been shown to be non-inferior to
warfarin in treatment of symptomatic venous thromboembolism and
pulmonary embolism (EINSTEIN trials). In general, no specific reversal
agents exist for the new oral anticoagulants but prothrombin complex
concentrate or recombinant factor VIIa can be employed if required.
However, the half-life of these agents is relatively short, which prevents the
need for reversal in most situations.
CASE SCENARIO 3
A 58-year-old man is referred after presenting with weight loss and
symptoms of pancreatic exocrine insufficiency. He has a resectable
hypodense mass in the head of his pancreas and a serum Ca 19-9 level of
150. The patient has coronary artery disease and suffered a myocardial
infarction 10 months ago, for which he underwent percutaneous coronary
intervention with placement of two drug-eluting stents. The patient is
currently on anti-hypertensive medication, as well as aspirin and
clopidogrel.
Management
The patient described will need to be seen in consultation with his
cardiologist, as well as by a staff internist/cardiologist who will follow
throughout the perioperative period. Following completion of all necessary
preoperative cardiac assessment, the patient should be instructed to stop
clopidogrel 5 days before the operation. His drug-eluting stents were placed
greater than 6 months ago, meaning the likelihood of stent thrombosis is
significantly decreased when compared to the risk at earlier time points. The
patient should continue to take aspirin, as the available data indicates a
benefit from continuation of therapy for patients with coronary artery or other
cardiovascular disease and moderate- to high-risk for perioperative
cardiovascular events.
Following completion of the procedure, prophylactic heparin (we prefer
heparin over other low molecular weight formulations) should begin at 8
hours and continue throughout the hospitalization. Aspirin should be resumed
approximately 24 hours after surgery, and may be given per rectum if
necessary. Given that our patient’s coronary vessels were stented more than 6
months ago, we suggest delaying resumption of clopidogrel until the time of
discharge or at the time of the first follow-up visit. Clopidogrel is an
irreversible platelet ADP receptor antagonist and will necessitate platelet
transfusion if hemorrhage ensues or an additional procedure is required. This
decision regarding restarting of clopidogrel should of course be made in
concert with the cardiologist.
DISCUSSION
The incidence of cardiac stent thrombosis is 2-5% based on several
retrospective studies of patients undergoing operations within two years of
stent placement, which may seem small but the consequences can be
catastrophic. Patients with coronary artery stents usually require antiplatelet
therapy such as aspirin and clopidogrel to prevent stent thrombosis. Aspirin
is an irreversible cyclooxygenase inhibitor, while clopidogrel is an ADP
receptor antagonist. In general, no monitoring is required for antiplatelet
medications. See Table 3 for more information regarding antiplatelet agents.
If bleeding occurs and urgent reversal is required, the treatment of choice is
platelet transfusion.
TABLE 3: Comparison of postoperative outcomes by distal
pancreatectomy technique
Patients with coronary artery stents may present a dilemma in weighing

risk of stent thrombosis and perioperative bleeding, depending upon the
timing of presentation. Data suggests discontinuation of dual antiplatelet
therapy within 6 weeks of bare-metal stent placement or within 3-6 months of
drug-eluting stent placement significantly increases the risk of stent
thrombosis. Current recommendations suggest deferring surgery for at least 6
weeks after placement of a bare-metal stent and at least 6 months after
placement of a drug eluting stent, which may not be a reasonable option for a
patient with a new diagnosis of pancreatic adenocarcinoma. For a patient
requiring a pancreaticoduodenectomy within the above delineated time
interval from stent placement, we would favor consideration for neoadjuvant
chemotherapy. This approach affords both a delay in the requisite time off
antithrombotics in the perioperative period (to reduce the risk of stent
thrombosis) while providing exposure to potentially therapeutic agents. The
merits of a neoadjuvant approach to pancreatic cancer are beyond the scope
of this chapter.
CASE SCENARIO 4
An 86-year-old man presents for a second opinion regarding the management
of a recent diagnosis of pancreatic adenocarcinoma in the head of the gland.
Review of the CT scan demonstrates a large mass associated with short
segment occlusion of the portal vein. The patient has a past medical history
of atrial fibrillation, stroke one year ago with residual weakness of the left
upper extremity, hypertension, coronary artery disease with prior bare-metal
stent placement 2 years ago, diabetes mellitus, pulmonary embolus 4 months
before diagnosis, and mitral valve regurgitation. The patient is an active
smoker. He insists that the tumor be removed within the next week.
Management
Despite the insistence, this patient has a borderline resectable cancer and
should not undergo an operation but rather referral to a medical oncologist
for systemic chemotherapy. Moreover, he is at high risk for any intervention
given his advanced age and multiple co-morbidities. This patient is unlikely
to complete the prescribed 6 months of chemotherapy without incident. He
should, however, undergo repeat imaging following completion of systemic
chemotherapy to evaluate the extent of his disease, assuming his candidacy
for resection remains viable. If resection is entertained, realistic expectations
of life expectancy with and without resection should be thoroughly explained,
preferably using a comorbidity risk estimation tool, along with the potential
complications associated with pancreatic resections. He should be seen in
consultation with his cardiologist for the usual battery of work up, such that
perioperative risk can be accurately assessed.
DISCUSSION
In our opinion, patients with potentially resectable disease should not be
denied consideration for operative intervention. However, in some patients
the risks simply outweigh the potential benefits. We would be remiss without
a reminder that the median survival after resection is about 2 years, and that
80% of patients undergoing resection will die of disease recurrence within 5
years. For the patient in question, using a life expectancy risk tool based on
comorbidities (and not accounting for the cancer diagnosis), his life
expectancy is approximately 2 years. In other words, accepting all the risks
associated with a pancreaticoduodenectomy in this patient, including
management of his antithrombotic medication, makes little sense if he is
unlikely to survive beyond 2 years because of his other medical
comorbidities.
SALIENT POINTS
Pancreaticoduodenectomy is a high-risk procedure necessitating
cessation of anticoagulation and antiplatelet therapy prior to resection.
As there is rarely, if ever, an urgent need for a
pancreaticoduodenectomy, appropriate planning in concert with
consulting physicians is strongly advised.
The use of bridging therapy should be stratified depending on the
individual patient’s risk for perioperative adverse events.
Patients on antiplatelet medication for coronary artery or other
cardiovascular disease who are considered moderate- to high-risk for
perioperative cardiovascular events likely benefit from continuation of
aspirin therapy.
Given the increased morbidity and mortality associated with
postpancreatectomy hemorrhage, and recognizing that this complication
can occur after discharge, we prefer to delay reinstitution of
antithrombotic therapy until the time of the first postoperative visit for
most patients, except those at the highest risk for thrombotic
complications.
SELECTED REFERENCES
1. Baron TH, Kamath PS, McBane RD. Management of antithrombotic
therapy in patients undergoing invasive procedures. N Engl J Med.
2013. 368(22): 2113-24.
2. Bauer KA. Pros and cons of new oral anticoagulants. Hematology Am
Soc Hematol Educ Program, 2013. 2013: 464-70.
3. Douketis JD, Spyropoulos AC, Spencer FA, Mayr M, Jaffer AK,
Eckman MH, et al. Perioperative management of antithrombotic therapy:
Antithrombotic Therapy and Prevention of Thrombosis, 9th ed:
American College of Chest Physicians Evidence-Based Clinical
Practice Guidelines. Chest, 2012. 141(2 Suppl): e326S-50S.
4. Panduranga P, Al-Mukhaini M, Al-Muslahi M, Hague MA, Shehab A.
Management dilemmas in patients with mechanical heart valves and
warfarin-induced major bleeding. World J Cardiol, 2012. 4(3): 54-9.
5. Wente MN, Veit JA, Bassi C, Dervenis C, Fingerhut A, Gouma DJ, et al.
Postpancreatectomy hemorrhage (PPH): an International Study Group of
Pancreatic Surgery (ISGPS) definition. Surgery, 2007. 142(1): 20-5.
Chu Q, Vollmer C, Zibari G, Orloff S, Williams M, Gimenez M (eds).

CASE SCENARIO
A 67-year-old male presented with painless jaundice. Imaging revealed a
“heterogeneous head of the pancreas mass involving the hepatic artery.” An
uncovered biliary stent was placed and adenocarcinoma confirmed. The only
option offered was palliative chemotherapy. Despite excellent performance
status, dose-reduced gemcitabine was started. At the urging of his children
after rising Ca 19-9 levels, he sought a second opinion. A pancreas protocol
CT was performed (Figure 1 ). Nutrition expertise was able to reverse
weight loss with enzyme replacement therapy. Abraxane was added,
followed by radiotherapy. The patient underwent a pancreaticoduodenectomy
with en bloc resection of the common and proper hepatic arteries after
confirming the replaced right hepatic was uninvolved. The patient had an
uneventful recovery from a T4N1M0, R0 resection.
FIGURE 1: A) The arrow demonstrates the uninvolved origins of
the celiac axis and superior mesenteric artery. The tumor
(arrowhead) circumferentially wraps around the common and
proper hepatic arteries which supply the left liver. B) This oblique
view shows (arrow) the uninvolved superior mesenteric and
replaced right hepatic arteries, distant from the tumor.
█ BACKGROUND
Operative management with an R0 resection margin offers patients with
pancreatic adenocarcinoma the best odds at extended survival. Numerous
studies have demonstrated better short and long-term outcomes for patients
with pancreatic neoplasms operated on at high volume centers, and when
performed by high-volume surgeons. In addition, studies from MD Anderson
Cancer Center and Technische Universität München – two highly specialized
centers – have shown successful outcomes in patients undergoing resection
after being deemed inoperable at other centers.
These patients need to be treated by a team of specialists focused on the
diagnosis and management of pancreatic disease. The patient must be treated
by consensus, not ego or the tunnel vision of a single discipline. The
surgeon’s role is first to determine resectability. If initially unresectable, the
surgeon determines specific endpoints for which achievement may allow for
resectability. Collaboration amongst the multidisciplinary group then moves
toward achieving the endpoints while improving or maintaining quality of
life. The patient is regularly presented at a multidisciplinary conference
where relevant disciplines review and provide expert input, assess
achievement of objectives, and document next steps.
Approximately 50-70% of pancreatic tumors are considered unresectable
at diagnosis because of metastases, or arterial involvement. R0 resections
are possible in a proportion of the patients without metastases. The
remainder will progress during neoadjuvant therapy, or occult metastasis
will be identified at exploration. We therefore argue that locally advanced
pancreatic cancer (LAPC) must be treated by a multidisciplinary group with
high volume commitment to treating LAPC, not just experience treating
resectable disease. This will limit unnecessary explorations, and provide
safer outcomes with neoadjuvant and operative therapy.
DEFINITIONS
There is little variation in the definitions of potentially resectable, borderline
resectable, and LAPC among the commonly referenced criteria: MD
Anderson Cancer Center, National Comprehensive Cancer Network, and the
International Study Group of Pancreatic Surgery. No one set of criteria is
necessarily better than the other. Confusion mostly arises around borderline
resectable and LAPC, where there is great interpretative variability by
different institutions and the consequent treatments offered.
Potentially resectable disease can be characterized as tumor without
superior mesenteric artery, superior mesenteric vein, portal vein, celiac axis,
or hepatic artery involvement. These patients should go straight to surgery, or
enter a clinical trial evaluating neoadjuvant therapy. Although we have seen
patients advised by non-surgeons they were unresectable because of splenic
artery or vein involvement, this alone does not stand in the way of surgery. At
the opposite end of the continuum, metastatic disease is defined by the
presence of biopsy-confirmed adenocarcinoma in any distant organ, the
peritoneum, or lymph nodes that do not directly drain the pancreas. Under
such circumstances, surgery is contraindicated even if there is a response to
chemotherapy.
Borderline resectable cases are categorized based on venous or arterial
involvement. When the vein involvement can be resected and repaired by
simple suture or patch, we bring the patient directly to surgery. When there is
venous involvement requiring circumferential transection with or without
interposition grafting, we advocate neoadjuvant chemotherapy prior to
surgery.
When there is common hepatic artery or ≤180° superior mesenteric artery
involvement, we advocate neoadjuvant chemotherapy and radiotherapy prior
to surgery. This is because of the effectiveness of modern-era regimens in
triggering tumor regression away from essential vasculature and improving
R0 rates. Operative morbidity can interfere with the delivery of adjuvant
therapy, so upfront chemo- and radiotherapy in an optimally vascularized
and oxygenated field is practical. Further, the neoadjuvant course allows
insight into tumor biology, and better patient selection.
LAPC describes non-metastatic disease with >180° involvement of the
superior mesenteric artery, celiac axis, long segments of the hepatic artery, or
superior mesenteric/portal vein involvement without targets for
reconstruction. No tumor categorized as LAPC on a radiology report should
be considered unresectable until the appropriate imaging is reviewed by an
experienced surgeon, radiologist and multidisciplinary panel.
Misclassification based on non-specific imaging reports is common in our
experience and can be the difference between palliative and curative-intent
therapies. Even when reporting is done according to a standard template,
there is variability in interpretation, confusion over contact (solid versus
stranding), and templates typically will not comment on the length of
involvement. This is the first reason why a specialized surgeon should be
involved in the care of LAPC.
PANCREAS SPECIFIC IMAGING

After overt metastasis is ruled-out, clinical staging of pancreatic tumors must
be performed by cross-sectional arterial and venous phase pancreas
protocoled imaging. Either CT or MRI is sufficient, both are unnecessary,
and PET is rarely indicated. We prefer CT imaging because it is more
rapidly acquired and less subject to respiratory variation. Decisions can be
guided but not made based on EUS, even when performed by a very
experienced endoscopist, because of an inability to review the imaging in
three dimensions. Imaging should be standard ized and follow a structured
reporting template, such as those published by the Society of Abdominal
Radiology and the American Pancreatic Association. A radiologist
committed to pancreas imaging, definitive in judgement, and versed in
pancreatic surgery possibilities is an essential team member. The radiologist
develops this background by staffing multidisciplinary conference, being
respected as a crucial decision maker, and receiving feedback based on the
intra-operative findings and pathology.
A thorough review for vascular anomalies may provide opportunities in
LAPC. For example, an uninvolved replaced right hepatic artery (or proper
hepatic off the superior mesenteric) may allow for pancreaticoduodenectomy
en bloc with any length of artery from the celiac axis through the proper. To
the contrary, not recognizing these anomalies may result in inadvertent
transection, bleeding and disorientation while managing the bile duct or
uncinate. Such anomalies also provide vascular insurance when performing a
modified Appleby. When reviewing arterial imaging, it is critical that the not
only the circumferential involvement of tumor be measured, but also the
length of involvement. It is common to see a caudad uncinate lesion abutting
the superior mesenteric artery, while circumferentially encasing early
branches in the mesentery; this makes resection impossible.
It is a rare patient that cannot get some form of contrast enhanced imaging.
In a patient that cannot get iodine-based contrast, the risk of nephrogenic
systemic fibrosis from gadolinium (≈2%) should not stand in the way of
treating a cancer that is 100% fatal in the absence of curative-intent surgery.
TISSUE DIAGNOSIS, LABORATORY

EVALUATION & BILIARY
DECOMPRESSION
A core tissue biopsy, preferably obtained by endoscopic ultrasound, is
essential. If abdominal lesions cannot be reached by endoscopic- or imaging-
guidance, laparoscopy should be performed. Rarely pancreatitis will
demonstrate similar clinical signs and symptoms, and appear as LAPC on
imaging. When tissue demonstrates pancreatitis, we closely monitor these
patients until resolution. Biopsies are repeated as necessary to confirm the
benign diagnosis and that the pancreatitis is not a manifestation of an
underlying cancer.
Aside from routine laboratory studies, evaluation should include Ca19-9
and CEA levels, which can be followed sequentially through treatment.
Although inflammatory markers such as CRP have shown an association with
prognosis, the role of monitoring these levels in LAPC is not defined.
In the absence of elevated bilirubin, a stent should not be prophylactically
placed. If any degree of biliary obstruction is present, a covered metal stent
should be placed. Such a stent can serve as a fiducial for stereotactic body
radiotherapy, and be removed pre-operatively if irreversible electroporation
is to be considered. Plastic stents are prone to infection and further
manipulation, while an uncovered metal stent makes the future operation
more difficult. Patients with duodenal obstruction should be considered for a
duodenal stent. A double bypass procedure should be avoided unless
endoscopic measures are exhausted, because of operative morbidity, delay to
starting systemic therapy, and this may complicate the operative field at the
anticipated future operation.
Patient Selection
With a median age of 71 years at presentation, decisions about surgery for
pancreatic cancer cannot be age-based. Careful consideration of
comorbidities and performance status are necessary. While most individuals
can tolerate a complex operation performed with the assistance of
exceptional peri-operative anesthesia and critical care, consideration as to
whether the patient can tolerate an unexpected complication or prolonged
recovery is central to decision making.
Patients being considered for curative-intent surgery for LAPC should be
ECOG 0 or 1, that is, no worse than restricted in physically strenuous
activity but ambulatory and able to carry out work of a light or sedentary
nature. Secondly, the patient should have the cardiopulmonary physiology to
tolerate both a protracted procedure, and shifts in hemodynamics that may be
abrupt and severe.
Patients with a rapid decline in performance status following diagnosis
may not be good candidates for aggressive treatment of LAPC, because this
may represent a low threshold for decompensation. At the same time, it
should be considered whether appropriate ancillary services are necessary
to maintain or improve performance status through the neoadjuvant period.
These include physical therapy, frequent meetings with nutritional services,
consultation with psychiatry for depression and assistance coping with the
diagnosis, and palliative care to assist with symptom management and
individual prioritization of treatment objectives against risks. A
comprehensive pancreas program should have all of these services integrated
and committed to treating patients with LAPC.
Patient selection is also based on compliance with examinations and
appointments and a realistic understanding of the risks and realities of
pancreatic cancer. This understanding is reinforced by an open, comfortable
relationship with practitioners who provide the same comprehensive plan.
This calls for a team at the same center in constant contact. The patient must
also have adequate family support willing to make appropriate
accommodations at home, and in their personal and professional life to
support the patient through the process.
Only about 25-33% of patients with LAPC will be eligible for curative
intent resection. Patients with LAPC will need to understand these realities
from the beginning of their treatment. It will allow for self-selection of
motivated patients, and may mitigate severe disappointment if the operation
becomes impossible.
NEOADJUVANT CHEMOTHERAPY
In LAPC, we strongly encourage patients to receive their neoadjuvant therapy
at our center. This allows frequent presentation at multidisciplinary meetings,
a rigorous concentrated approach to the patient, and avoidance of
preventable derailing from the treatment course. If the patient is seen at
another comprehensive cancer center, the surgeon must follow the imaging
closely. Reports of local progression may be meaningless to the final
operative plan, but may be interpreted as progression to unresectability by a
provider not intimately involved with the process, or experienced with
LAPC.
Consistent with NCCN guidelines, we recommend use of either
FOLFIRINOX or gemcitabine-based therapy. No patient being considered
for surgery should receive mono-therapy. If this is all they can
physiologically tolerate, it is highly unlikely they can handle the stress of
operative management. Therapy is continued for approximately two months
(4 cycles of FOLFIRINOX or 2 cycles of gemcitabine/abraxane) before
patients are reimaged, markers are drawn and presentation at
multidisciplinary conference occurs. This is where center experience is
important. Local progression does not warrant regimen change unless there is
a major issue with tolerance. Rarely will radiotherapy be initiated after this
short course, but if there has been rapid regression, radiotherapy and sooner
to surgery may be recommended. The majority of patients go on to complete
another two months of neoadjuvant chemotherapy (4 cycles of FOLFRINOX
or 2 cycles of gemcitabine/abraxane). If they remain operable, then radiation
is pursued. If inoperable, chemotherapy is continued with palliative intent.
NEOADJUVANT RADIATION
In all patients with locally advanced disease, radiation therapy is delivered
after post-chemotherapy imaging rules out metastatic disease. Stereotactic
body radiation therapy or intensity-modulated radiation therapy can be used
at the discretion of the multidisciplinary team. After the delivery of radiation,
we maintain the patient off systemic chemotherapy and perform pancreas
protocolled imaging 4-6 weeks later with immediate plans for surgery in the
absence of metastatic disease.
When treating LAPC, a working assumption is that even the best imaging
will over stage some patients after neoadjuvant therapy. According to the
International Study Group of Pancreatic Surgery, imaging specificity ranges
from 67-91% due to the inadequate differentiation between soft tissue scar
and true cancerous invasion. We achieved R0 resections in 80.4% of 56
patients with locally advanced NCCN-defined unresectable pancreatic
cancer based on cross-sectional imaging after FOLFIRINOX or gemcitabine-
based neoadjuvant therapy. Although tumor shrinkage is desired, this seldom
occurs because of the desmoplastic nature and relative chemo resistance of
pancreatic adenocarcinoma. A combination of failure of progression, decline
in tumor markers, decrease in surrounding inflammation, and changes in
attenuation on imaging are seen as treatment success. Then, the only way to
confirm the presence of malignant tissue around the vasculature is by
exploration, palpation and biopsy as the dissection proceeds.
SURGERY
LAPC by definition involves the neck of the pancreas. Whether the burden of
tumor extends to the right or left dictates whether a pancreaticoduodenectomy
or distal pancreatectomy is indicated, respectively. It is expected that a
pancreaticoduodenectomy will extend just to the left of the aorta, whereas a
distal pancreatectomy will extend to the gastroduodenal artery. The areas of
high risk for extirpative failure, venous inflow and outflow, vascular control,
and potential conduits, need to be planned before entering the operating room
based on detailed study of the imaging. There needs to be clear limits in the
mind of the surgeon before the case begins as to what circumstances the
operation will be terminated before full commitment. This will prevent going
too far and harming the patient if the operation cannot be completed or result
in an R2 resection. Operations are started early in the day when all hospital
resources are available. Laparoscopy is performed to prevent futile
laparotomy. Any suspicious lesions are biopsied for frozen section. If
metastatic lesions are confirmed, additional tissue is obtained for complete
pathological evaluation.
The operation should move immediately to dissection beyond the
extremes of arterial disease, as this will determine whether the resection
proceeds. For example, if a modified Appleby is anticipated, dissection
begins along the cephalad portion of the pancreas widely around the celiac
axis down to the aorta to determine involvement (Figure 2 & 3). The
combination of neoadjuvant therapies and tumor-desmoplasia make tissue
differentiation difficult, so suspicious tissues should be sent for frozen
section to guide operative progression. The objective is to dissect in normal
soft tissue planes, while minimizing violation of tumor planes. Mobilization
of viscera may be necessary during these steps, but transection should be
avoided unless there is a need for bypass. Attention should next be turned to
identifying sites of vascular control. At this point we will perform test
clamps to make sure there is sufficient arterial flow to terminal organs if
arterial resection without reconstruction is expected.
FIGURE 2: Modified Appleby Operation. The original operation
was described for gastric cancer that involves the celiac axis. The
modified Appleby operation is for resection of locally advanced
tumors of the pancreatic body and tail that involve the celiac axis.
In this procedure, the celiac axis is ligated and flow to the hepatic
artery is established via retrograde flow through the intact
pancreaticoduodenal arcade and the gastroduodenal artery (GDA)
to maintain liver perfusion. CA: Celiac Axis; LGA: Left Gastric
Artery; SA: Splenic Artery; CHA: Common Hepatic Artery; GDA:
Gastroduodenal Artery; SMA: Superior Mesenteric Artery; IPDA:
Inferior Pancreaticoduodenal Artery; AIPDA: Anterior Inferior
Pancreaticoduodenal Artery, PSPDA: Posterior Superior
Pancreaticoduodenal Artery; SMV: Superior Mesenteric Vein
FIGURE 3: Completed Modified Appleby Operation
Irreversible Electroporation (IRE) is a technology that can be used to

non-thermally ablate tumor tissue around blood vessels and bile or
pancreatic ducts. For a tumor less than 4 cm in longest axis encasing the
superior mesenteric artery and celiac trunk and/or meso-portal venous
confluence that cannot be reconstructed, in situ IRE can be considered to
treat LAPC by ablating the tumor in the absence of resection. IRE is also
used for LAPC with curative intent resection, when there is concern for
microscopic involvement of essential vasculature that cannot be
reconstructed or resected. IRE should not be performed if an R2 resection is
anticipated; under these circumstances the approach should be in situ
ablation. IRE is increasingly popular, mostly because of a lack of options.
The technology is still being studied with regard to safety and mid- and long-
term efficacy. A multi-site registry has been organized by the Americas
Hepato-Pancreatico-Biliary Association with financial support from
Angiodynamics.
180-360° ENCASEMENT OF THE

CELIAC AXIS AND BRANCHES
PROXIMAL TO GASTRODUODENAL
ARTERY
The operation of choice is a modified Appleby procedure. After entry in to
the lesser sac and confirmation of the absence of metastatic disease, the
spleen should be mobilized of its attachments. The right gastric artery and
epiploic vessels are preserved. The tail of the pancreas is then dissected out
of the retroperitoneum from left to right with retro pancreatic tissue as
necessary until the left side of the aorta is identified. Resection of the celiac
axis, pancreas and lympho-neuro-vascular tissue off the base of the aorta and
left superior mesenteric artery is performed. We have not performed
preoperative coil embolization or reconstruction of the common hepatic
artery during modified Appleby procedures, and have not had any hepatic or
gastric ischemia.
180-360° ENCASEMENT OF PROPER

HEPATIC ARTERY DISTAL TO
GASTRODUODENAL ARTERY
In the absence of a replaced right hepatic artery that would allow sacrifice of
the proper hepatic artery, we advocate for IRE in this situation. An
alternative approach is a long segment bypass to the uninvolved proper
hepatic artery (usually concomitant portal venous involvement will need
reconstruction). However, these are timely reconstruction procedures and
have elevated morbidity and mortality. We thus advocate IRE to the vessel
bed performed prior to widely mobilizing tissues. When the
pancreaticoduodenectomy specimen is dissected, all lymphovascular tissue
along the involved vessels are skeletonized to the adventitia. The underlying
concept being the IRE will result in apoptosis of remaining R1 diseased
tissue.
180-360° ENCASEMENT OF BOTH THE

SUPERIOR MESENTERIC ARTERY AND
CELIAC AXIS
If a modified Appleby will remove the macroscopic tumor, this is performed
after IRE of the superior mesenteric artery with skeletonization of the
remaining tissues (patient must be informed about the resultant diarrhea that
will ensue). If the tumor burden cannot be resected by
pancreaticoduodenectomy or distal pancreatectomy, then in situ IRE alone is
utilized. If pancreaticoduodenectomy will remove the macroscopic tumor,
this should be performed following IRE to the superior mesenteric and celiac
arteries; with en bloc skeletonization of the peri-vascular tissues.
180-360° ISOLATED ENCASEMENT OF
THE SUPERIOR MESENTERIC ARTERY
A pancreaticoduodenectomy or distal pancreatectomy is performed based on
burden of disease following IRE, with en bloc skeletonization of the peri-
vascular tissues to the level of the adventitia. The “artery first” approach is
very useful in these complex dissections when IRE is to be used along the
superior mesenteric artery.
VENOUS RECONSTRUCTION
CONSIDERATIONS
Venous involvement is common and surgeons treating pancreatic cancer must
always be prepared for resections and reconstructions. The distal portal vein
target is almost always available; the portal vein is long and can be extended
by separating it from lymphatics and other structures in the hepatoduodenal
ligament, or by mobilizing the right liver. Although not essential, preserving
the cardiac vein can prevent a problematic gastropathy if the splenic vein
needs to be sacrificed.
The superior mesenteric vein is typically the critical target, because
proximally it is the confluence of multiple tributaries that may need to be
individually anastomosed to a graft. Under these circumstances, an internal
jugular graft with short distances of the middle and superior thyroid veins
can be used for multiple anastomoses. Alternatively, mesenteric veins
leading to the superior mesenteric vein can be joined side-to-side and then
anastomosed around the perimeter to the larger jugular graft. It is important to
examine the proximal jejunum prior to anastomosis. Performing the pancreas,
choledochal and gastric anastomoses to congested or poorly vascularized
intestine is sure to result in leaks and complications, so a further length of
jejunum should be resected. We do not routinely heparinize these
anastomoses during the recovery period, however, if we are concerned with
flow along these tributaries, we will start a heparin gtt in the operating room
with a goal of about 75% therapeutic ptt and discontinue this in the first 48-
72 hours if there is no clinical evidence of thrombosis.
POST-OPERATIVE CARE
After these complex procedures, most patients are recovered in the SICU for
24 hours to prevent hypo- or hypertension, both which risk vascular
anastomoses. Routine post-operative vascular imaging is not performed;
ultrasonography is employed in response to clinical symptoms, abruptly
rising AST/ALT, or excessive ascites drainage. Otherwise, no special steps
are taken with regard to feeding, mobilization, drain management, or
anticoagulation outside of our standard post-operative protocols.
ADJUVANT THERAPY
The decision for adjuvant chemotherapy is based on the success of the
neoadjuvant regimen, and final pathologic interpretation. Thus, this decision
should be made on an individual basis considering nodal status, failure to
achieve an R0 resection, and evidence of lymphovascular or perineural
involvement. Radiation therapy is not repeated in the adjuvant setting.
Adjuvant therapy is initiated within 6-12 weeks. Beyond this point, a plan
based on surveillance imaging and tumor markers should be considered.
CONCLUSION
Only an experienced surgeon who understands the technical details of the
operations for LAPC and inherent risks can identify the potential resectable
patients. As such, they should be the advocate for the patient, monitoring
progress through neoadjuvant therapy, reviewing all imaging, and being
attentive at multidisciplinary conference. The surgeon cannot be a passive
observer, waiting for the patient to be prepared and delivered.
When properly planned, these procedures can be performed safely at high
volume centers. The work of Dr. Joseph G. Fortner should not be used as a
deterrent to more radical pancreatic resections. Since Fortner’s pioneering
work, operative and peri-operative safety have improved. More importantly,
we are in a new era in chemotherapy and targeted radiotherapy, allowing for
better treatment and selection of patients that may benefit from these
operations. Neoadjuvant therapy allows the multidisciplinary committee
approximately 6 months to assess tumor biology, identify distant metastasis,
and properly select patients. Further study is unquestionably necessary.
Randomized operative trials are not feasible, but the comparative group of
inoperable patients is readily available.
SALIENT POINTS
Locally advanced pancreatic cancer (LAPC) must be treated by a
multidisciplinary group with high volume commitment to treating
LAPC, not just experience treating resectable disease.
A radiologist committed to pancreas imaging, definitive in judgement,
and versed in pancreatic surgery possibilities is an essential team
member. Imaging should be standardized and follow a structured
reporting template.
Decisions about surgery for pancreatic cancer cannot be age-based.
Careful consideration of comorbidities and performance status are
necessary.
FOLFIRINOX or gemcitabine-based neoadjuvant therapy followed by
radiation therapy is necessary before considering curative-intent
surgery for LAPC.
Patients must be presented at multidisciplinary conference every time
imaging is performed.
The specificity of contrast-enhanced cross sectional imaging to
differentiate between soft tissue scar and cancerous vascular invasion
is limited.
Final operative decisions are based on multidisciplinary review and an
identified plan to achieve an R0 margin
There needs to be clear limits in the mind of the surgeon before the
case begins as to what circumstances the operation will be terminated
before full commitment. This will prevent going too far and harming
the patient if the operation cannot be completed or result in an R2
resection.
Irreversible electroporation is an increasingly popular therapy for
LAPC. The technology is still being studied with regard to safety and
mid- and long-term efficacy.
SELECTED REFERENCES
1. Michalski CW, Kleeff J, Bachmann J, Alkhatib J, Erkan M, Esposito I,
et al. Second-look operation for unresectable pancreatic ductal
adenocarcinoma at a high-volume center. Ann Surg Oncol.
2008;15(1):186-92.
2. Al-Hawary MM, Francis IR, Chari ST, Fishman EK, Hough DM, Lu
DS, et al. Pancreatic Ductal Adenocarcinoma Radiology Reporting
Template: Consensus Statement of the Society of Abdominal Radiology
and the American Pancreatic Association. Radiology 2014; 270(1);
248-60.
3. Bockhorn M, Uzunoglu FG, Adham M, Imrie C, Milicevic M, Sandberg
AA, et al. Borderline resectable pancreatic cancer: a consensus
statement by the International Study Group of Pancreatic Surgery
(ISGPS). Surgery. 2014;155(6):977-88.
4. Kluger MD, Rashid MF, Rosario VL, Schrope BA, Steinman J, Hecht
EM, Chabot JA. Resection of locally advanced pancreatic cancer
without regression of arterial encasement after modern-era neoadjuvant
therapy. Journal of Gastrointestinal Surgery. Accepted 8/13/17. PMID:
28895032

CASE SCENARIO
Case 1 : The patient is an 83-year-old woman who presented with
obstructive jaundice. Workup revealed a mass in the head of the pancreas
(Figure 1 ) with biopsy-proven adenocarcinoma. She had a BMI of 32 and
well-controlled hypertension. She lived alone and had excellent performance
status. She was seen by a surgeon who recommended against surgical
management. On second opinion, she was felt to be resectable and a
reasonable surgical candidate despite her age. She underwent a
pancreaticoduodenectomy. Final pathology showed a node-negative, margin-
negative pancreatic cancer arising in an intraductal papillary mucinous
neoplasm (IPMN). Despite recovering from surgery, the patient opted against
adjuvant therapy. She ultimately passed away from metastatic pancreatic
disease 6 years after surgery. Up until her final months, her family recalls her
having an active life, highlighted by dancing at her grandson’s wedding.
FIGURE 1: Mass in the head of the pancreas in an 83-year-old
female; she was told that surgery was not an option elsewhere. A.
Imaging shows dilated CBD and PD in the pancreatic head. B. No
vascular involvement. This image shows a clear fat plane around
the SMA and SMV.
Case 2 : An 84-year-old man with no comorbidities who walks 4 miles per

day presented with obstructive jaundice. Workup revealed an ampullary mass
with biopsy-proven adenocarcinoma. He was felt to be a good operative risk
with a good prognosis tumor. He underwent a whipple and was discharged to
his assisted living facility on postoperative day (POD) 6. He failed to thrive
at home and was subsequently readmitted for dehydration and sent to a higher
level of postoperative nursing care. He never returned to living
independently and on POD 25, he was found dead in his assisted living
facility.
█ BACKGROUND
Pancreatic cancer remains the fourth leading cause of cancer deaths in both
men and women in the United States. In 2017, there will be an estimated
53,670 new cases of pancreatic cancer, with 43,090 estimated deaths. As life
expectancy in the United States increases and the population ages, an
increasing numbers of older patients are diagnosed with pancreatic cancer.
Currently, the median age of diagnosis is 70 years old, with 80% of patients
over the age of 60 and approximately 13% over the age of 84.
The treatment for pancreatic cancer primarily depends on the resectability
at presentation (Table 1 , Figure 2 ), determined by the presence or absence
of vascular invasion or distant metastatic disease. Approximately one third
of patients present with locoregional disease (resectable, borderline
resectable, locally unresectable) and even fewer have truly resectable
disease without significant involvement of the superior mesenteric vein
(SMV), portal vein (PV), superior mesenteric artery (SMA), or celiac axis at
the time of presentation. In resectable patients, multimodality therapy with
surgery and adjuvant or neoadjuvant chemotherapy offers the best survival.
While the overall 5-year survival is only 7%, multimodality therapy in
patients with resectable disease has a 2-year survival of 41%.
FIGURE 2: Algorithm showing the management options for
octogenarian patients with pancreatic cancer by resectability status
and functional status. Note that age is not a factor in the algorithm.
TABLE 1: Treatment Options Based on Resectability

Status
Despite the benefit of resection and chemotherapy in early stage
pancreatic cancer, surgery and multimodality therapy are underused.
Moreover, utilization of resection is decreasing as patients age, despite the
benefits.
Many factors complicate treatment decisions in older patients with
pancreatic cancer, including comorbidities, decreased baseline life
expectancy, higher complication rates, and higher failure-to-rescue rates once
complications occur. These factors may differentially impact not only
quantity, but also long-term quality of life, as many older patients are unable
to return home after surgery.
SURGICAL RESECTION IN THE ELDERLY
PATIENT
Pancreatectomy offers the only hope for cure in patients with pancreatic
cancer. In the 1970s, mortality following pancreatectomy exceeded 25%.
Currently, mortality is less than 2-3% at most major centers, but morbidity
remains as high as 30-50%. Despite the low mortality, significant
controversy remains regarding the role of pancreatectomy in older patients,
especially those over 80 years old. With the high morbidity following
resection and decreased physiologic reserve, failure to rescue is more
common following pancreatic resection in older patients.
Regardless of improvements in mortality and morbidity, recent studies
have shown that resection is underutilized in the United States. Furthermore,
this underutilization is even higher with older patients as they are less likely
to undergo surgical evaluation or resection. In an analysis of data linked to
Surveillance, Epidemiology, and End Results (SEER) and Medicare, older
patients were less likely to undergo evaluation by a surgeon; 81% of patients
<70 years were evaluated by a surgeon, decreasing to only 45% of patients
85 years and older (P<0.001). Even after surgical evaluation, 39% of
participants younger than 70 years of age underwent resection versus 5% in
patients older than 85 years of age. The decrease in surgical evaluation and
resection with age persisted even in a subgroup analysis of patients with no
comorbidities.
Many single-institution and population-based studies have evaluated
morbidity and mortality in older patients undergoing surgical resection, with
inconsistent findings. While results are mixed, large population-based
studies clearly document increased mortality with increasing age. In a Texas
population-based study (1999-2005), 3,736 patients underwent pancreatic
resection. Of those, 47.7% were less than 60 years old, 23.7% were 60–69
years of age, 22.9% were 70–79 years of age, and 5.7% were 80 or older.
Mortality increased with the increasing age group (2.4% to 5.8% to 7.4% to
11.4%, P<0.001). Likewise, in a Medicare analysis, perioperative mortality
increased from 6.7% in patients aged 65–69 to 15.5% in patients aged 80
and older (P<0.0001).
In the Texas data, length of stay also increased in the increasing age
group, with a median length of stay of 11 days for patients <60 years (mean =
16.0+15.6 days), 13 days for patients 60–69 years (mean = 16.5 +12.4 days),
14 days for patients 70–79 years (mean = 17.6 +13.2 days), and 15 days for
patients over 80 (mean = 17.9 + 12.4 days).
The data on short-term complications following pancreatic resection in
older patients are also mixed. A French study by Scurtu and colleagues
demonstrated a statistically higher incidence of delayed gastric emptying in
the older group (12.5% vs 0%, P=0.04), and a large study from Johns
Hopkins showed a 33% incidence of delayed gastric emptying in the patients
≥80 compared to an 18.6% rate in patients <80 (P = 0.03). The incidence of
cardiac and pulmonary complications was also higher in older patients,
while operation-specific complications such as pancreatic fistula, bile leak,
intra-abdomen abscess, sepsis, and cholangitis did not differ.
In an analysis of the NSQIP database, complication rates were actually
similar across age groups while in-hospital mortality increased in patients 80
years or older compared to patients younger than 80 years (3.0% vs 1.1%, P
= 0.02). However, older patients were more likely to die (failure to rescue)
following a complication. In patients 80 years or older, the failure-to-,rescue
rates were higher (7.7% vs 2.7%, P = 0.01). The increased failure-to-rescue
rates reflect the decreased physiologic reserve and decreased ability to
recover from complications in this vulnerable population.
Despite the increased morbidity and mortality with surgical resection in
older patients, these patients still derive a survival benefit from surgical
resection. In a population of Medicare patients with pancreatic cancer,
resection was associated with lower risk of death, regardless of age, with
hazard ratios of 0.46 (younger than 70), 0.51 (70-74 years), 0.47 (75-79
years), 0.43 (80-84), and 0.35 (85 and older) for resected participants
compared with unresected participants younger than 70 (P<.001). The lower
risk of death with increasing age suggests appropriate patient selection.
Finalyson et al. demonstrated similar 5-year survival rates for patients aged
65-69 (16.4%), 70-79 (15.6%), and 80 and older (11.3%, P = 0.28).
It has been consistently shown that mortality after pancreatic resection is
higher at low-volume hospitals compared to high-volume hospitals
regardless of the volume cutoff chosen. More importantly, the difference in
mortality between high- and low-volume hospitals is accentuated with age.
Low-volume facilities have higher rates of complications and baseline
morbidity, with likely higher failure-to-rescue rates in older patients once
complications occur. With the increasing age group (<60 years, 60-69 years,
70-79 years, and 80+ years), mortality increased from 2.0% to 3.5% to 4.5%
to 8.7% at high- volume hospitals (P<0.0001). However, at low-volume
hospitals, mortality increased from 3.0% to 9.5% to 11.4% to 14.7%.
While quantity of life is important, quality of life may be equally
important for many patients. Population-based data show that older patients
are less likely to be discharged home following a pancreatectomy, with 20%
of patients 70-79 years and 62% of patients 80 years and older requiring a
skilled nursing facility or nursing home care upon discharge.
MULTIMODALITY THERAPY
Although surgery is essential for curative treatment, adjuvant chemotherapy
has been shown to improve survival. Given the high incidence of recurrence
and poor survival in pancreatic cancer, it is thought to be a systemic disease
at the time of presentation. Therefore, in resectable patients, multimodality
therapy with surgery and adjuvant or neoadjuvant chemotherapy offers the
best survival and is the current standard of care. There is significant debate
over whether chemotherapy should be delivered in the adjuvant or
neoadjuvant setting, with pros and cons to each approach. The approaches
are difficult to compare, as most studies are not analyzed on an intent-to-treat
basis and randomized trials have been difficult especially in resectable
disease. However, recent population-based data suggests that survival is
similar in patients who receive both modalities, regardless of timing.
Multimodality therapy for locoregional disease in all patients has increased
over the last decade, from 31.3% in 2004 to 37.9% in 2011 (P<0.0001).
However, rates are lower in older in patients. Only 11% of Medicare
beneficiaries with locoregional pancreatic cancer patients received
multimodality therapy, with chemotherapy delivered in an adjuvant setting in
93.1% and neoadjuvant setting in 6.9%.
ADJUVANT THERAPY
Adjuvant treatment has been shown to improve survival in multiple
controlled trials including ESPAC-1, CONKO-001, ESPAC-3, RTOG 9704,
and GITSG. However, older patients are less likely to receive adjuvant
therapy following surgery. In Medicare patients undergoing surgery first,
more than half (51.2%) did not receive adjuvant chemotherapy. The underuse
of adjuvant therapy in older patients is likely multifactorial. The higher
complication rates following surgery may delay or preclude adjuvant
therapy. Additionally, it has been our experience that many older patients
decline adjuvant therapy when offered, as in our case scenario.
NEOADJUVANT THERAPY
While there is significant debate over the timing of chemotherapy, of those
receiving multimodality therapy, >90% of patients nationally receive it in the
adjuvant setting. While adjuvant therapy improves survival, postoperative
complications can delay or prevent receipt of chemotherapy, especially in the
older population. Over the last decade, the use of neoadjuvant therapy in
truly resectable disease has increased. Moreover, neoadjuvant therapy is
favored in the setting of borderline resectable disease; neoadjuvant therapy
can convert patients to resectable disease and has been associated with
higher R0 resection rates. This approach may be beneficial in older patients,
especially in those who present a high surgical risk.
In a retrospective study evaluating the receipt of multimodality therapy
over a nine-year period (2000-2008), 179 patient >70 years of age were
treated for resectable or borderline resectable pancreatic cancer; 153 (85%)
were treated with neoadjuvant chemoradiation with intent to resect, and 26
(15%) underwent surgery first. In the neoadjuvant group, 48% completed
multimodality therapy, while 52% progressed during chemotherapy or had
poor performance status and would not tolerate surgery. In the initial surgical
resection group, 11 of 26 (42%) received adjuvant therapy. Survival was
similar between the two groups. Lack of progression during chemotherapy
helps select the patients who will most likely benefit from surgery. If disease
progresses, surgical resection can be avoided. A neoadjuvant approach
allows the disease biology to become evident. Finally, the older patient’s
ability to tolerate chemotherapy may be a marker of their physiologic reserve
and predict surgical outcomes.
Newer neoadjuvant regimens such as FOLFIRINOX are more effective
but not well tolerated, especially in older patients. FOLFIRINOX is
associated with a higher incidence of adverse effects including neutropenia,
febrile neutropenia, thrombocytopenia, diarrhea, and sensory neuropathy.
FOLFIRINOX is only recommended in young patients with good cardiac
function and normal bilirubin levels. Additionally, older patients are
underrepresented in studies of neoadjuvant and adjuvant therapy.
CHEMORADIATION IN NONSURGICAL
CANDIDATES
In older patients who are not surgical candidates, either for locally
unresectable disease or poor functional/performance status, chemoradiation
is also possible. Chemoradiation without surgery is not curative, but may
slow tumor growth or cause tumor regression. It may also palliate symptoms
associated with unresected pancreatic cancer including gastrointestinal
bleeding due to tumor invasion of the duodenum. Chemoradiation prolongs
survival for months. Patients that undergo chemoradiation for unresectable
disease experience multiple side-effects of treatment, such as with
FOLFIRINOX. Overall, chemoradiation alone prolongs survival less than
surgery alone.
PALLIATION
In some cases, despite resectable disease, patients may not be candidates for
surgical resection. In older patients with a poor performance status, frailty,
poor functional status, and severe medical comorbidities, surgery may not be
an option. In addi tion, one must consider the patient’s preferences and goals
of care. Given the poor prognosis of pancreatic cancer and the toxicity
associated with treatment, the decision often balances the quality and quantity
of life. Many older patients may choose to forego aggressive treatment given
these factors.
Once the decision is made to not pursue aggressive therapy in the case of
resectable (or unresectable disease), palliation becomes tantamount. The
goal of palliative care is to improve quality of life. Most common
complications associated with pancreatic cancer are related to involvement
of adjacent structures and include obstructive jaundice, gastric outlet
obstruction, and intractable pain from celiac plexus tumor infiltration.
One half to three quarters of patients with pancreatic cancer develop
obstructive jaundice. With advances in endoscopy, most patients with
obstructive jaundice who are not candidates for surgical resection can be
managed endoscopically. Endoscopic stent placement has a success rate of
>90%. The median patency for plastic stents is 2-4 months; they are prone to
obstruction and bacterial overgrowth and should only be placed in the setting
of planned resection or short expected survival. In cases of palliation without
planned surgical resection in patients with a projected survival of 4 to 6
months, self-expanding metal stents are recommended due to longer patency
(7-10 months).
In the rare setting of failed endoscopic management, percutaneous
transhepatic biliary drainage is an alternative, nonoperative procedure. With
percutaneous transhepatic drainage, the bile ducts are accessed through the
liver parenchyma and a wire is passed through the obstruction. Once the
obstruction is crossed, metal stents can be place and the internal/external
drains removed. If the site of obstruction is not passed, obstruction can only
be relieved by proximal drainage and will be associated with significant
daily fluid and bile salt loss.
Gastric outlet obstruction is another common complication of unresected
pancreatic cancer. Historically, this was managed surgically with a
gastrojejunostomy; however, in older patients who are poor surgical
candidates, this is ideally avoided. Endoscopic stent placement for outlet
obstruction is now an option and is often successful in the short-term with
approximately 90% of patients resuming diet within 24 hours. Stent migration
can occur in approximately 27% of patients, therefore it is recommended to
stent only symptomatic patients. Stent occlusion is also common and most
patients can tolerate only clear or full liquid diets, but they can at least
manage their own secretions. Newer endoscopic technology allows for
creation of an endoscopic gastrojejunostomy, but this technology is in its
infancy, without good outcomes data. In our preliminary experience,
complication rates are high, but when successful, patients experience better,
more durable return to diet with endoscopic gastrojejunostomy compared to
stenting. Furthermore, surgery is avoided.
Finally, invasion of celiac plexus will produce severe abdominal and
back pain. In addition to oral analgesics, pain can be managed with
neurolysis of celiac access. Overall, patients have decreased pain at four and
eight weeks post procedure and consume decreased amount of opioids
overall.
Hospice involves a team-oriented approach to medical care, pain
management, and emotional and spiritual support expressly tailored to the
individual patient during the final stage of serious illness. Palliative care and
hospice teams help clarify patient and family goals of care, and the benefits
of hospice care have been well established. In a population-based study of
older patients who died of pancreatic cancer, hospice was underused. Only
59.6% of patients who died of pancreatic cancer were enrolled in hospice
care, and only 35.9% were enrolled for four weeks or more. In addition,
aggressive care at the end of life increased over time, and admission to the
ICU and receipt of chemotherapy during the last month increased from 15.5
to 19.6 % (P<.0001) and from 8.1% to 16.4% (P<0.001), respectively.
Hospice care should be considered early in the course once patients decide
against aggressive therapy.
CONCLUSION
We present two cases to highlight the challenge. In both, the patient was over
80 years of age, but in good health and living independently. In the latter
case, the patient has a better prognosis tumor. In our first clinical case, the
patient was initially told she was not a surgical candidate, and she could
have missed an opportunity to live an additional six years. She understood
the risks of surgery and chose to proceed. However, she declined adjuvant
therapy despite healing from surgery. She ultimately succumbed to pancreatic
cancer but enjoyed good quality of life in the interim. In the second case, the
patient failed to thrive and died shortly after surgery, despite readmission
and no obvious postoperative complications. He was never able to return
home and back to his prediagnosis quality of life. In fact, surgery shortened
his life significantly. These cases highlight the difficulty in making these
decisions. Despite all we know, we cannot reliably predict who will benefit
and who will not.
The management of resectable pancreatic cancer in older patients
requires careful consideration of the staging of disease; the patient’s frailty,
functional status, and comorbidities; the disease prognosis; as well as the
patient’s treatment goals and preferences. All patients with locoregional
pancreatic cancer should undergo surgical evaluation by a pancreatic
surgeon, as patients cannot make an informed decision if they are not fully
evaluated, not offered all the options, and do not understand the risks and
benefits of each option. In discussing surgery, age alone should not dictate the
patient’s management, as surgical benefit does not diminish with age. Older
patients with good functional status (living at home, no dementia, tending to
their own activities of daily living) should be considered for surgical
resection. It is critical to communicate the risks and disclose the potentially
higher morbidity, mortality, and failure-to-rescue rates. In addition, we
should discuss the long-term survival/prognosis as well as the potential
impact on quality of life, including inability to return to home or independent
living after surgery. The impact of complications is difficult to communicate
unless a person has experienced them. Older patients should also understand
the benefits of surgical resection and multimodality therapy.
Multidisciplinary care teams should be involved in the care of these complex
and vulnerable patients and the surgical and multimodality therapy options,
risks, and benefits should be discussed with the multidisciplinary team, the
patient, and the family. Based on the data presented, surgical resection should
be directed to high-volume centers and/or surgeons to optimize outcomes in
this vulnerable group when aggressive therapy is reasonable and chosen by
the patient and treatment team. Only after a complete discussion of the
benefits and risks of procedure as well as alternatives, can a patient can
make a decision that is truly right for them. This decision will be different for
each patient.
SALIENT POINTS
In discussing surgery, age alone should not dictate a patient’s
management.
The morbidity, mortality, and failure-to-rescue rates are higher in older
patients undergoing surgical resection for pancreatic cancer. Despite
these facts, older patients can still derive a survival benefit from
surgical resection.
The decision to resect an older patient with pancreatic cancer must
consider the complete picture, including tumor characteristics and
prognosis, patient comorbidity/ability to tolerate surgery, the opinion
of the multidisciplinary care team, and patient preferences and goals of
treatment.
Multimodality therapy is the standard of care, but many factors
challenge our ability to fully deliver this in older patients.
Consider neoadjuvant in older patients; this will allow the tumor
biology to become clear and avoid surgery in patients who will likely
not benefit.
While the current data suggest underuse of surgical resection this, in
part, reflects good patient selection.
When surgery is performed, it should be done at high-volume centers.
SELECTED REFERENCES
1. Parmar AD, Vargas GM, Tamirisa NP, Sheffield KM, Riall TS.
Trajectory of care and use of multimodality therapy in older patients
with pancreatic adenocarcinoma. Surgery. 2014; 156(2):280-9.
2. Bilimoria KY, Bentrem DJ, Ko CY, Stewart AK, Winchester DP,
Talamonti MS. National failure to operate on early stage pancreatic
cancer. Ann Surg. 2007; 246(2):173-80.
3. Riall TS, Sheffield KM, Kuo YF, Townsend CM, Goodwin JS.
Resection Benefits Older Adults with Locoregional Pancreatic Cancer
Despite Greater Short‐Term Morbidity and Mortality. J Am Geriatr
Soci. 2011;59(4):647-54.
4. Finlayson E, Fan Z, Birkmeyer JD. Outcomes in octogenarians
undergoing high-risk cancer operation: a national study. J Am Coll of
Surg. 2007;205(6):729-34.
5. Riall TS, Reddy DM, Nealon WH, Goodwin JS. The Effect of Age on
Short-Term Outcomes After Pancreatic Resection: A Population-based
Study. Ann Surg. 2008; 248(3):459-67.
6. Scurtu R, Bachellier P, Oussoultzoglou E, Rosso E, Maroni R, Jaeck D.
Outcome after pancreaticoduodenectomy for cancer in elderly patients.
J of Gastrointest Surg. 2006;10(6):813-22.
7. Tamirisa NP, Parmar AD, Vargas GM, Mehta HB, Kilbane EM, Hall
BL, et al. Relative contributions of complications and failure to rescue
on mortality in older patients undergoing pancreatectomy. Ann of Surg.
2016; 263(2):385-91.
8. Dimou F, Sineshaw H, Parmar AD, Tamirisa NP, Jemal A, Riall TS.
Trends in Receipt and Timing of Multimodality Therapy in Early-Stage
Pancreatic Cancer. J Gastrointest Surg. 2016;20(1):93-103.
9. Sheffield KM, Boyd CA, Benarroch‐Gampel J, Kuo YF, Cooksley CD,
Riall TS. End‐of‐life care in Medicare beneficiaries dying with
pancreatic cancer. Cancer. 2011;117(21):5003-12.

CASE SCENARIO
A 70-year-old, 200-pound, highly functional white male with no medical
comorbidities presents to your office with jaundice. He reports progressive
jaundice, icterus, and a 5% weight loss over the past few weeks. Prior
workup includes normal laboratory chemistries except for a total bilirubin of
20, mostly direct in nature. Patient underwent a pancreas protocol computed
tomography (CT) scan showing a pancreatic head mass and a dilated
common bile duct (CBD). Endoscopic ultrasound biopsy confirms
adenocarcinoma. You determine the patient has resectable disease without
evidence of metastasis.
█ BACKGROUND
PREOPERATIVE CONSIDERATIONS
Approximately 65-75% of patients suffering from pancreatic cancer develop
symptomatic biliary obstruction. Debate exists about whether patients
experience higher mortality following pancreatoduodenectomy in the setting
of hyperbilirubinemia. Further, controversy remains over whether to
manipulate the biliary system preoperatively with stent placement given
associations with morbidity and mortality postoperative, including higher
surgical site infection rates and wound-related complications.
At our institution, we follow current National Comprehensive Cancer
Network (NCCN) guidelines for preoperative biliary drainage and treat
symptomatic, obstructive hyperbilirubinemia preoperatively to improve
symptoms and liver function. We recommend preoperative biliary
decompression whenever feasible for symptomatic obstructive jaundice. If a
biliary stent cannot be successfully placed via endoscopic technique, we
recommend percutaneous biliary drainage. The NCCN also recommends
preoperative biliary decompression among septic, coagulopathic patients, or
those with impaired renal function prior to surgery. Additionally,
symptomatic patients suspected to have delays in surgery >1 week are also
stented for decompression. We are a high-volume center offering neoadjuvant
therapies for pancreatic adenocarcinoma and routinely place a stent for
preoperative biliary drainage in this sort of patient prior to undergoing
neoadjuvant therapy.
Neoadjuvant chemoradiotherapy (resectable/borderline resectable

scenarios)
Enrollment in neoadjuvant therapy trials is encouraged by the NCCN. If
your institution offers neoadjuvant chemoradiotherapy clinical trials, or the
patient wishes to be referred to a high-volume center to enroll in a clinical
trial, the patient requires preoperative biliary stenting for decompression
during treatment prior to surgical resection. If neoadjuvant therapy is not
available, we still recommend biliary decompression with stenting prior to
surgical resection in this patient.
Stent type
While stent technology evolves, your choice of stent type varies from
plastic or metal; fully covered, partially covered, or uncovered; rigid or self-
expanding. Overall, short, self-expanding metal stents are favored over
plastic regarding ease of placement and patency rates; however, depending
on the timing of surgery, plastic stents are appropriate for shorter operative
wait times. An ongoing trial is underway to compare metal and plastic stent
outcomes for preoperative biliary decompression (ClinicalTrials.gov
NCT01191814).
Optimization
If your patient is optimized for surgery, biliary stenting for decompression is
indicated for symptomatic obstructive jaundice in this patient with a bilirubin
of 20 or greater. Should other comorbidities or risk assessment require
further testing or intervention prior to surgery, we recommend biliary decom
pression with stenting while further tests or interventions are performed
preoperatively.
INTRAOPERATIVE AND POSTOPERATIVE

CONSIDERATIONS
Having a biliary stent does not compromise exposure or dissection during
pancreatic resection. We offer adjuvant chemoradiotherapy according to
current NCCN guidelines for pancreatic adenocarcinoma. In our experience,
we observe low morbidity and mortality rates, including low surgical site
infections among patients with preoperative biliary stent placement for
obstructive jaundice having a pancreatic head mass.
SALIENT POINTS
Preoperative biliary stent decompression in patients with symptomatic
obstructive jaundice is indicated and endorsed by the NCCN.
Patients should receive preoperative biliary stenting for
decompression with either a plastic or covered metal stent if not
planning to operate soon or if planning neoadjuvant
chemoradiotherapy.
Nonsurgical candidates or patients with metastatic disease should
receive palliative biliary decompression with a self-expanding,
uncovered metal stent if other palliative surgical decompression
procedures are not performed.
SELECTED REFERENCES
1. House MG, Choti MA. Palliative therapy for pancreatic/biliary cancer.
Surg. Clin. North Am. 2005;85(2):359-71.
2. National Comprehensive Cancer Network (NCCN). Pancreatic Cancer
(Version1.2017).2017;https://www.nccn.org/professionals/physician_gl
s/pdf/pancreatic.pdf. Accessed March 21, 2017.

CASE SCENARIO
A 65-year-old man presented with abdominal pain and, on CT scan, was
found to have a locally advanced mass in the head of the pancreas with
extension into the transverse colon and mesocolon with abutment of the
superior mesenteric vein (Figure 1 ). He underwent an ERCP with placement
of an endoscopic biliary prosthesis and brushing of the malignant-appearing
stricture. Pathology revealed a pancreatic ductal adenocarcinoma. He was
referred to medical oncology for neoadjuvant therapy, but was unable to
complete treatment due to repeated episodes of bacteremia. At time of
exploration, the tumor in the head of the pancreas was found to have
significant mesocolic involvement secondary to an adjacent abscess. A
classic pancreatoduodenectomy (PD) was performed in addition to an
extended right colectomy and regional lymphadenectomy. Surgical pathology
revealed a 5-cm pancreatic adenocarcinoma that had invaded into the
transverse colon with microscopically negative margins, lymphovascular
invasion, and 5/22 lymph nodes positive. His postoperative course was
notable for a Grade B pancreatic fistula treated with antibiotics and
percutaneous drainage. He was discharged to home on post-operative day 14
and ultimately received adjuvant chemotherapy.
FIGURE 1: Locally advanced pancreatic head mass with
transverse colonic involvement. A) Involvement of the transverse
colonic mesentery and proximal transverse colon. B) The classic
‘double duct’ sign with adjacent pancreatic hyperemia consistent
with pancreatitis. SMA: Superior Mesenteric Artery; SMV: Superior
Mesenteric Vein.
█ BACKGROUND
Surgical resection provides the only chance for prolonged survival and cure
in patients with pancreatic adenocarcinoma. Pancreatectomy has become
safer over the last three decades due to advancements in surgical technique
and critical care. As experience has grown, some surgeons have taken a more
aggressive approach to locally advanced tumors of the pancreas. While
significant strides have been made, post-operative morbidity remains high
and performing multi-visceral resection during pancreatectomy remains
controversial. Proponents argue that the dismal outcomes provided by other
local therapies necessitates an aggressive surgical approach in order to
afford the best chance for prolonged survival for the patient. Others would
suggest that tumor invasion into major vascular structures or adjacent organs
is a sign of poor tumor biology and portends a dismal outcome irrespective
of an aggressive surgical approach. Herein, we discuss our experience in
managing this complex clinical scenario and how this compares to the
limited data in the literature.
An analysis of patients at Indiana University between November 2006
and May 2011 who underwent pancreatectomy (PD, distal pancreatectomy,
and total pancreatectomy) and simultaneous colectomy (P+C) were analyzed
and outcomes compared to patients who underwent pancreatectomy alone
(P). Of 840 patients who underwent pancreatectomy, twenty underwent
pancreatectomy with concomitant colectomy (2.4%).
Pancreatoduodenectomy was performed in 9 patients (9/519, 1.7%) and
distal pancreatectomy in 11 (11/301, 3.7%). Outcomes are shown in Table 1 .
Mortality appeared to be increased following P+C compared to P alone
(10% vs 3% NS), and there was a statistically significant increase in
morbidity (70% vs 33%, p < 0.01). The increased morbidity was largely due
to increases in organ space infection (Figure 2 ). Given the substantial
increases in both morbidity and mortality, we concluded that careful patient
selection and strategies to prevent and/or control anastomotic leaks are
necessary to improve outcomes in patients undergoing simultaneous
pancreatectomy and colectomy.
TABLE 1: Comparison of Pancreatectomy Alone versus

Pancreatectomy with Colectomy
FIGURE 2: Incidence of deep organ space surgical site infections
in patients following pancreatectomy (P) and combined
pancreatectomy and colectomy (P+C).
Several retrospective series have addressed whether concomitant PD and

right colectomy is associated with increased morbidity and/or mortality
compared to PD alone. Most single institution series suggest comparable
post-operative outcomes to PD alone, but, due to sample size, are prone to
Type II error and selection bias. One of the largest series compared
outcomes of 28 simultaneous resections with 607 patients having undergone
PD alone and found increased operative time (530 vs 410 min, p < 0.001)
when right colectomy was performed at the time of PD, but no outcome
reported significant difference, although likely significant in a larger series,
in mortality (7% vs 1%, p = 0.068), major morbidity (25% vs 17%; p =
0.304), or pancreatic fistula (7% vs 13%; p = 0.472). The morbidity and
mortality in our series is increased compared to this and several other
smaller single institution reports, but is consistent with a recent propensity
score-matched analysis of patients identified in the American College of
Surgeons National Surgical Quality Improvement Program (ACS-NSQIP).
Major morbidity and mortality in 156 patients who underwent P+C was
50.0% and 9.0% compared to 28.8% and 2.9% (p < 0.001) in a propensity
score-matched cohort of 624 patients who underwent PD alone. Similar to
our experience, the incidence of deep organ space infection in the NSQIP
analysis was double for patients undergoing P+C (22.4% vs 11.1%, p
<0.001).
While the impact of colonic resection during PD on post-operative
morbidity and mortality is unclear, the oncologic benefit of such an
aggressive approach is even less understood. However, the median survival
of patients with locally advanced pancreatic cancer is less than one year in
unresectable patients, and palliative chemo-radiation therapy alone offers
minimal survival benefit. Most agree that local invasion into the mesocolon
or transverse colon is not necessarily a sign of aggressive tumor biology, but
a mere function of tumor location and/or size, and when a R0 resection is
performed, survival is comparable to patients without mesocolic
involvement. Several studies have shown that when carefully selected,
similar rates of R0 resection following P+C compared to pancreatectomy
alone can be achieved. When Suzuki et al. compared eight patients with
pancreatic adenocarcinoma who underwent P+C to patients who underwent
PD alone, they found the same median progression-free survival (6 vs. 6
months) and a similar median overall survival (14 vs. 12 months). In another
series of eight patients, the median survival was 20 months. Temple et al.
described a trend towards decreased survival in 28 patients with
periampullary tumors who underwent P+C compared to 607 patients who
underwent pancreatectomy alone at both 1-year (54% vs. 79%) and 3-years
(42% vs. 61%), but these differences were not significant. Again, due to the
retrospective nature of these series and reduced statistical power to detect
differences, the ability to obtain an R0 resection and effect on subsequent
survival controversy persists. The oncologic outcomes of P+C for locally
advanced colon cancer are better established. Multiple centers have reported
similar oncologic outcomes following P+C with acceptable increases in
post-operative morbidity compared to colectomy alone when an R0 resection
is obtained and patients are matched for stage. Yet, given the characteristic
aggressive tumor biology of pancreatic cancer, this data should not be
extrapolated to patients with locally advanced periampullary tumors.
While significant controversy exists, authors can agree on the importance
of patient selection in approaching this complex oncologic dilemma. Patients
should be referred to a tertiary referral center with experience in pancreatic
surgery where care can be orchestrated by a multidisciplinary team, with
consideration of both patient and tumor factors. The use of neoadjuvant
therapy in these cases specifically should be strongly considered as this will
assist in selecting patients who would maximally benefit from an aggressive
local approach and avoid operating on patients at risk for early local or
distant recurrence. Moreover, neoadjuvant therapy has been shown to
increase the likelihood of completion of multimodal therapy. In preparation
for surgery, patient nutrition, performance status, smoking cessation, and
other known modifiable risk factors should be optimized in order to reduce
the risk of post-operative complications; patients should be counseled on the
advanced risk of major morbidity and mortality following P+C.
The technical challenges of combined pancreatectomy and segmental
colectomy cannot be underscored. The median operative time is typically
100-130 minutes longer and intraoperative blood loss is often 500-1000ml
greater. The greatest technical challenge associated with these cases is the
involvement of the tumor with the central mesentery and the SMV. Typically,
the middle and right colic veins are sacrificed at their insertion into the trunk
of Henle prior to insertion onto the SMV. In addition, the mesentery is often
for-shortened and folded on top of the SMV due to inflammation,
desmoplasia, pancreatitis, or abscess--making the dissection that much more
tedious. Technical errors at this location can compromise the circulation to
either the small bowel or colon. In these situations, complete medial visceral
mobilization to expose the root of the small bowel mesentery is often
necessary via a Cattell-Brasch maneuver. Once resected, careful evaluation
of the mesenteric blood supply should be performed, and consideration
should be given to the creation of a diverting loop ileostomy. Wound
infections are a significant cause of morbidity following P+C and one must
consider wound protection, clean-closure protocol, primary closure with or
without negative pressure wound therapy, or delayed primary closure with or
without negative pressure wound therapy.
In summary, pancreatoduodenectomy with colectomy is considered an
aggressive surgical therapy given the potential for increased morbidity and
paucity of data showing a clear survival benefit. However, selected patients
may benefit when an R0 resection is achievable. Patients should be treated at
a high-volume pancreatic center and managed by a multidisciplinary team.
Our practice is to refer patients for consideration of neoadjuvant
chemotherapy prior to consideration of surgery. In patients who respond, and
those with no progression of disease, we perform extensive counseling on the
added risks of surgery while attempting to modify known risk factors and
optimize nutritional status preoperatively. The technical challenges of
patients with locally advanced pancreatic cancer who require segmental
colon resection cannot be overstated. Both operative time and blood loss are
augmented in these combined procedures. Post-operatively, patients must be
observed vigilantly in order to quickly rescue any patients who develop a
morbidity and fall off the expected post-operative course. Most of all, patient
selection is of utmost importance.
SALIENT POINTS
Pancreatoduodenectomy with colectomy is considered an aggressive
surgical therapy given the potential for increased morbidity and paucity
of data showing a clear survival benefit
While select patients may benefit, our practice is to refer patients for
consideration of neoadjuvant therapy prior to consideration of surgery
Patients should be treated at a high-volume pancreatic center and
managed by a multidisciplinary team
SELECTED REFERENCES
1. House MG, Kilbane M, Ceppa EP, Nakeeb A, Howard TJ, Schmidt
CM, et al. Simultaneous Pancreatectomy and Colectomy: A Safe
Combination? Paper presented at: Pancreas Club Annual Meeting 2012;
San Diego, CA.
2. Temple SJ, Kim PT, Serrano PE, Kagedan D, Cleary SP, Moulton CA,
M, et al. Combined pancreaticoduodenectomy and colon resection for
locally advanced peri-ampullary tumours: analysis of peri-operative
morbidity and mortality. HPB (Oxford). 2014;16(9):797-800.
3. Harris JW, Martin JT, Maynard EC, McGrath PC, Tzeng CW. Increased
morbidity and mortality of a concomitant colectomy during a
pancreaticoduodenectomy: an NSQIP propensity-score matched
analysis. HPB (Oxford). 2015;17(9):846-54.
4. Kimchi ET, Nikfarjam M, Gusani NJ, Avella DM, Staveley-O’Carroll
KF. Combined pancreaticoduodenectomy and extended right
hemicolectomy: outcomes and indications. HPB (Oxford).
2009;11(7):559-64.

CASE SCENARIO
A 25-year-old woman presents with debilitating epigastric abdominal pain.
She initially presented 5 years ago with an episode of severe acute
pancreatitis. She underwent cholecystectomy at that time. She subsequently
has had multiple episodes of pancreatitis, marked by pain and oral
intolerance requiring hospital admission. Her pain is now constant, requiring
daily narcotic medications, and she is unable to work. She has lost 20 lbs
despite pancreatic enzyme replacement therapy. She has undergone ERCP
with stenting, without relief. She does not drink alcohol or smoke and has no
family history of pancreatitis. Her genetic testing is notable for SPINK-1 and
CFTR gene mutations. A recent MRCP (Figure 1 ) shows small pancreatic
ducts, without clear evidence of chronic pancreatitis. She underwent EUS,
showing 6 of 9 conventional criteria for chronic pancreatitis.
FIGURE 1: MRCP showing a normal pancreas and main
pancreatic duct in a patient with clinical manifestations of chronic
pancreatitis.
█ BACKGROUND
Chronic pancreatitis is marked by severe, intractable, and debilitating
abdominal pain with associated devastating impairments in quality of life.
Patients may develop attendant diabetes and malnutrition due to endocrine
and exocrine pancreatic failure, respectively. Frontline therapies include
addressing inciting factors for inflammation and fibrosis, pain control, and
medical support. Intervention is undertaken for pain relief. Endoscopic
procedures may be employed in an attempt to optimize organ drainage.
Surgical therapies are considered once medical and endoscopic interventions
fail. Operative algorithms are based primarily on ductal anatomy and organ
morphology. Patients with a dilated main pancreatic duct (>6 mm) can
benefit from operative drainage (lateral pancreaticojejunostomy or localized
pancreatic head resection combined with lateral pancreaticojejunostomy).
Patients with small pancreatic ducts may experience pain relief with focused
resection of a dominantly affected pancreatic head or tail. Some patients,
however, have diffuse, small duct pancreatitis. These patients may benefit
from total pancreatectomy with islet autotransplantation (TPIAT) (Figure 2,
3 ). TPIAT can also be effective for pain relief and improvements in quality
of life in patients with genetic pancreatitis and as salvage therapy in patients
who have failed lesser surgeries.
FIGURE 2: Total pancreatectomy and reconstruction with a Roux-

en-Y Hepaticojejunostomy and Roux-en-Y duodenojejunostomy
FIGURE 3: Techniques of Islet Autotransplantation: Following a
total pancreatectomy, native beta islet cells are isolated and
reinfused back to the patient via the portal vein. The islet cells take
residence in the liver and produce insulin.
While TPIAT was initially described as a treatment option for

pancreatitis as early as 1978 at the University of Minnesota, enthusiasm for
this intervention has only recently emerged over the past decade or so, likely
due to increasing acceptance of pancreatic surgery for benign disease and the
evolving science of islet harvest. The clinical science of this rather radical
procedure therefore is still in its formative period.
SURGICAL EVALUATION
In particular, patient selection is recognized as being paramount to a
successful outcome after TPIAT, but is still being elucidated. Patients
presenting for evaluation for TPIAT are a heterogeneous and complex group
of patients. They are marked by debilitating abdominal pain with associated
poor quality of life and attendant social dysfunction; they have encountered
(and failed) multiple medical and endoscopic interventions; and they often
have complicated medical histories. Current best practice includes following
strict criteria for patient selection to better sort through these complex
patients and to optimally apply this invasive treatment option.
CONFIRMING THE DIAGNOSIS

Chronic pancreatitis
The first criterion for surgery (Table 1 ) is that the patient has objective
evidence of chronic pancreatitis or recurrent acute pancreatitis. Chronic
pancreatitis is well diagnosed with radiographic or endoscopic means.
Contrast-enhanced CT scan can show parenchymal calcifications, ductal
abnormalities, or peripancreatic inflammatory changes. MR pancreatography
can show ductal pathology including dilation, strictures, or stones, as well as
loss of parenchymal enhancement on T1 weighted images. ERCP criteria for
the diagnosis of chronic pancreatitis have been well established (Cambridge
classification), and these criteria have been applied in interpretation of
MRCP images as well. Endoscopic ultrasound is a useful modality to
evaluate for pancreatitis, and expert consensus has identified pertinent
features for diagnosis. It is operator-dependent, however, and can be
nonspecific such that the presence of multiple characteristics (greater than 5
of 9 criteria) is most reliable in determining diagnosis. In patients where
imaging characteristics are indeterminate, some centers employ endoscopic
pancreatic exocrine function testing for bicarbonate levels with the intent of
identifying patients with pancreatic dysfunction.
TABLE 1: Selection Criteria for Total Pancreatectomy
with Islet Autotransplantation
Minimal change pancreatitis

The challenge in evaluating patients for TPIAT is that these are a distinct
group of patients from the typical chronic pancreatitis patients surgeons have
historically encountered, with heavy calcific burden, dilated ducts, and an
alcoholic etiology. TPIAT candidates often have a different phenotype with
small ducts, minimal calcific burden, and an idiopathic (or increasingly
recognized genetic) etiology. Minimal change pancreatitis was described as
early as 1992 by Walsh and colleagues at the Middlesex Hospital in London.
The authors reported on their experience with 16 patients who had
pancreatic-type abdominal pain but minimal or equivocal findings on
imaging with ERCP, abdominal ultrasound, or CT scan. The patients were
treated with total (12) or partial (4) pancreatectomy. Pathologic examination
of the pancreas showed chronic inflammatory changes in all 16 patients and 9
were pain free on long-term follow-up. Wilson and colleagues have
described their experience with TPIAT in “minimal change” chronic
pancreatitis patients in the modern era, with 58% achieving narcotic
independence and all with improvements in quality of life postoperatively.
Important, however, is the definition of “minimal change.” Wilson and
colleagues did require radiographic confirmation of pancreatitis (CT, ERCP,
or EUS) to offer TPIAT, although they utilized less stringent criteria for a
pancreatitis diagnosis than most, considering only 4 (rather than 6) of the 9
conventional criteria on EUS as diagnostic. Given the invasive nature and
potential morbidity of TPIAT, it is most prudent to adhere to strict criteria for
diagnosis of chronic pancreatitis, particularly as optimal patient selection is
still being elucidated.
Recurrent acute pancreatitis

Another indication for TPIAT is recurrent acute pancreatitis (RAP). The
primary approach to RAP is endoscopic, both for evaluation and
management. Some patients, however, fail endoscopic therapies, and in
selected patients, TPIAT may be a reasonable therapeutic option. For clarity,
patients with RAP considered for TPIAT should have at least 2 documented
episodes of pain with associated elevations in amylase or lipase greater than
3 times normal that are not temporally associated with endoscopic
intervention.
Genetic pancreatitis
Genetic causes for pancreatitis are increasingly being recognized. In 1996,
Whitcomb and colleagues described the mutated cationic trypsinogen gene
(PRSS1) as the cause for hereditary pancreatitis, an autosomal dominant
disorder with 80% penetrance conveying the inappropriate activation of
trypsin. Patients affected develop recurrent acute pancreatitis and chronic
pancreatitis often beginning in early childhood and carry a significantly (up
to 70 times) increased risk for pancreatic cancer, particularly with smoking.
Since then, several other genes have been identified that predispose patients
to pancreatitis, predominantly through the inappropriate activation or
regulation of trypsin. Mutations of the pancreas secretory trypsin inhibitor
gene (serine protease inhibitor Kazal type 1, SPINK-1) and of the
chymotrypsin C gene (CTRC) cause impaired regulatory inhibition of trypsin
activity. Mutations in the calcium-sensing receptor gene (CASR) result in
altered intracellular calcium regulation and mutations of the gamma glutamyl
transferase 1 gene (GGT1) cause dysregulation of cellular glutathione levels,
both of which lead to activated trypsin and a predisposition towards
pancreatitis. Some mutations in the cystic fibrosis transmembrane
conductance regulator gene (CFTR) can compromise pancreatic ductal
flushing leading to pancreatitis. Finally, the claudin gene locus (CLDN2) on
the X-chromosome has been associated with alcoholic pancreatitis, which
suggests that even environmentally related etiologies for pancreatitis may
have an underlying genetic component. Currently PRSS1, SPINK1, CFTR,
and CASR mutations have readily commercially available tests and should
be a routine part of the evaluation for TPIAT. The presence of a pancreatitis-
associated gene mutation may support a diagnosis of chronic pancreatitis in a
patient with equivocal imaging, and may influence the decision for a TPIAT
as an initial surgical intervention rather than a partial resection or drainage
procedure.
PAIN ASSESSMENT
The clinical hallmark of pancreatitis is severe intractable abdominal pain.
This pain is the primary factor for determining intervention. Patients
considered for TPIAT should have pain that is significantly disruptive to
their work, school, family, and societal roles, as well as impairments in
quality of life. While patients with chronic pancreatitis, and in particular
hereditary pancreatitis, have an increased risk for pancreatic cancer, current
evidence does not support TPIAT in patients without pain for cancer
prevention.
During the preoperative optimization period, patients should have their
pain management optimized, both with pharmacological and behavioral
medicine support. Identifying a pain management provider during the
perioperative and postoperative period is ideal.
CONSIDERATION OF LESSER
INTERVENTIONS
Surgery is not the frontline therapy for pancreatitis, and thus medical
optimization and endoscopic options should be thoroughly explored as an
initial step. In patients who fail the nonoperative approach, surgery is
considered based on organ morphology and ductal anatomy. In patients with
dilated ducts, a drainage procedure is appropriate. In patients with
nondilated ducts, a targeted resection is indicated. Patients offered total
pancreatectomy are those that have small ducts and diffuse disease. The
potential exception to this principle is the patient with hereditary pancreatitis
(PRSS1 mutation). Once these patients demonstrate the pancreatitis
phenotype, they typically have a progressive disease course, and
consideration of TPIAT as the initial intervention appears warranted. Less is
currently known about the natural history of the other genetic etiologies of
pancreatitis, and thus the classic algorithm should not be altered for these
patients.
DETERMINATION OF PHYSIOLOGICAL
FITNESS
As in all elective major abdominal surgeries, patients need to be risk
stratified and medically optimized prior to considering surgical intervention,
as comorbidities are often the ultimate determinant of outcome.
Cardiopulmonary comorbidities should be evaluated. It is notable that
comorbid hepatic disease is a relative contraindication to islet transplant, as
the islet infusion is an embolic event to the terminal branches of the portal
vein. This normally small physiologic event can be significant in the setting
of underlying hepatic compromise.
Chronic pancreatitis patients are often malnourished on presentation due
to exocrine pancreatic insufficiency and pain, causing poor oral intake. A
formal nutritional evaluation and support is important prior to TPIAT.
Pancreatic enzyme replacement therapy should be instituted preoperatively to
assess patient tolerance of medications as well as to prehabilitate them for
surgery.
DIABETIC PATIENTS
Patients presenting for TPIAT may have comorbid diabetes, either from
pancreatitis-related endocrine failure, or due to other etiologies. Oral
glucose or mixed-meal tolerance testing can assess preoperative beta cell
function. In patients requiring insulin on initial evaluation, a stimulated C-
peptide study can assess the patient’s ability to synthesize endogenous
insulin. Patients with a stimulated C-peptide less than 0.7 nmol/L likely do
not have sufficient islet function to warrant the potential morbidity and cost
of an islet autotransplant.
DETERMINATION OF PSYCHOLOGICAL
FITNESS
Patients with chronic pancreatitis and chronic pain being evaluated for
TPIAT have significant attendant psychological comorbidities. There is a
high prevalence of depression (68%), opioid misuse (48%), and impaired
psychological quality of life. Patients often have poor coping mechanisms.
Behavior health psychology evaluation and support are essential prior to
surgery and postoperatively to optimize patient selection and outcomes. In
addition, behavior health specialists can teach patients coping skills and
adjunctive pain relief measures including biofeedback and guided imagery
that can improve opportunities for a successful perioperative period.
CASE SCENARIO WRAP-UP

In the case scenario presented, this 25-year-old woman needs objective
confirmation of her pancreatitis diagnosis, specifically by radiographic
determination, in order to be certain TPIAT has appropriate potential to
benefit her and to warrant surgical risk. If her axial imaging is indeterminant,
identification of 6 of 9 conventional criteria by EUS can be helpful. A history
more indicative of recurrent acute pancreatitis can be confirmed by obtaining
documentation of enzyme elevation during her pain episodes. Genetic testing
can be supportive of her diagnosis.
Less morbid interventions, including endoscopic therapies or a surgical
drainage procedure or partial resection should be considered based on her
ductal anatomy and gland morphology.
She needs physiologic assessment, including nutritional evaluation and
counseling, and pancreatic enzyme replacement therapy. If she has diabetes, a
stimulated C-peptide study should be undertaken to be sure islet
autotransplantation is warranted.
A behavioral health evaluation should be performed to address any
psychological comorbidities and optimize her coping mechanisms for the
perioperative period.
SALIENT POINTS
TPIAT is an effective means of pain relief and improves quality of life
in selected patients with chronic pancreatitis.
Patient selection is fundamental, but can be challenging in this complex
and heterogeneous patient group.
Following specific guidelines can facilitate optimal outcomes and is
essential as this emerging clinical modality evolves and is elucidated.
SELECTED REFERENCES
1. Sutherland DE, Radosevich DM, Bellin MD, Hering BJ, Beilman GJ,
Dunn TB, et al. Total Pancreatectomy and Islet Autotransplantation for
Chronic Pancreatitis. J Am Coll Surg. 2012; 214(4): 409-24.
2. Morgan K, Owczarski SM, Borckardt J, Madan A, Nishimura M,
Adams DB. Pain Control and Quality of Life After Pancreatectomy with
Islet Autotransplantation for Chronic Pancreatitis. J Gastrointest Surg.
2012; 16(1): 129-34.
3. Sutherland DE, Matas AJ, Najarian JS. Pancreatic islet cell
transplantation. Surg Clin North Am.1978; 58(2): 365-82.
4. Walsh TN, Rode J, Theis BA, Russell RCG. Minimal change chronic
pancreatitis. Gut. 1992; 33(11): 1566-71.
5. Wilson GC, Sutton JM, Smith MT, Schmulewitz N, Salehi M, Choe KA,
et al. Total pancreatectomy with islet cell autotransplantation as the
initial treatment for minimal-change chronic pancreatitis. HPB. 2015;
17(3): 232-8.
6. Whitcomb DC, Gorry MC, Preston RA, Furey W, Sossenheimer MJ,
Ulrich CD, et al. Hereditary pancreatitis is caused by a mutation in the
cationic trypsinogen gene. Nat Genet. 1996; 14(2):141-5.
7. Morgan KA, Borckardt J, Balliet W, Owczarski SM, Adams DB. How
are chronic pancreatitis patients selected for total pancreatectomy with
islet autotransplantation? Are there psychometric predictors? J Am Coll
Surg. 2015; 220(4): 693-8.

CASE SCENARIO
A 55-year-old patient has a history significant for recurrent acute pancreatitis
secondary to hypertriglyceridemia. His imaging studies demonstrate chronic
pancreatitis with a head- dominant- disease (Figure 1-2 ). In addition, there
is presence of calcifications in the head. ERCP revealed a proximal high-
grade stricture in the pancreatic duct which was unable to be traversed by a
guidewire (Figure 2 ).
FIGURE 1: CT imaging revealing ductal dilation and calcification
of the pancreatic head.
FIGURE 2: A) CT imaging demonstrating diffuse calcifications

along the pancreatic duct, and B) ERCP revealing a proximal
stricture in the pancreatic duct.
TABLE 1. Randomized Controlled Trials of Surgical
Intervention for Head-Dominant Chronic Pancreatitis
█ BACKGROUND
Chronic pancreatitis is a progressive inflammatory disorder characterized by
the hallmark symptom of persistent, disabling abdominal pain. In addition,
due to fibrosis and irreversible destruction of the pancreatic parenchyma,
patients can exhibit metabolic sequelae of endocrine and exocrine
insufficiency. Regardless of the disease etiology, environmental versus
genetic, more than half of all patients will require surgical intervention.
The most common indication for surgical intervention is intractable pain
in spite of optimal medical management. Multiple hospitalizations for acute
or chronic pancreatitis episodes interfere with a patient’s quality of life, and
secondary complications of pancreatic fibrosis (obstruction of the pancreatic
or common bile duct, and intestinal obstruction) and duct rupture (formation
of pseudocyst or pancreatic ascites) may also warrant surgical exploration.
In addition, the concomitant suspicion of malignancy reinforces the decision
to proceed with surgery.
Given the multifactorial etiology and pathophysiology of chronic
pancreatitis, various surgical options exist. Although randomized controlled
trials (RCT) have established the superiority of surgery over endoscopic
treatment, much debate exists with regard to the optimal operative approach.
Surgical techniques for chronic pancreatitis can be divided into two general
categories: drainage/decompressive procedures and resection procedures. A
cross-sectional study by van der Gaag et al. of 223 patients who underwent
surgical intervention for chronic pancreatitis demonstrated long-term pain
relief in two-thirds of patients, when the procedure was tailored to the
anatomical abnormalities. Thus, a surgeon dedicated to treating patients with
chronic pancreatitis should be facile with tailoring the operative approach to
the patient’s pancreatic parenchymal and ductal anatomy and should be
skilled in the available operative techniques. This chapter will highlight the
steps critical to working up a patient with head-dominant chronic pancreatitis
for surgery (Figure 2 ) and define the current operative procedures as well
as their indications, advantages, and disadvantages.
PREOPERATIVE WORKUP
The majority of patients will already have had numerous imaging studies
performed for acute exacerbations of pancreatitis. Prior to any surgical
intervention, the surgeon must be familiar with the patient’s pancreatic
parenchymal and ductal anatomy. Various imaging modalities are utilized to
gain this information. Computed tomography (CT) plays a key role in the
global assessment of chronic pancreatitis. It allows for delineation of
pancreatic parenchymal changes through the use of intravenous iodinated
contrast. The pancreas is enhanced homogenously after administration of
contrast, with optimal enhancement occurring 35-40 seconds after contrast
injection.
In addition to detecting parenchymal atrophy and calcifications, CT is
also useful in detecting pseudocysts. Multiplanar reconstruction and thin
slices allow for optimal visualization of the pancreatic duct and any
associated changes in caliber. The portal venous phase of enhancement
allows for detection of vascular complications, such as pseudoaneurysms or
venous thrombi.
Similar to CT, magnetic resonance imaging (MRI) allows for a global
assessment of the pancreas parenchyma and ductal anatomy. Normal pancreas
appears bright on T1-weighted noncontrast images, with loss of signal
alluding to fibrosis. Magnetic resonance cholangiopancreatography (MRCP)
uses T2-weighted images to evaluate ductal anatomy. Its sensitivity is
improved with secretin administration, which promotes duct distension (at
least 1 mm in the main pancreatic duct) by increasing pancreatic secretions
and sphincter of Oddi contraction.
Although MRCP and secretin administration improves assessment of the
ductal anatomy, it may be insufficient for small branch pancreatic ducts.
Ultimately, endoscopic retrograde cholangiopancreatography remains the
gold standard for assessing the pancreatic duct and may also be therapeutic
when complications are detected, such as obstruction of the right-sided main
pancreatic duct.
Finally, endoscopic ultrasound (EUS) is another useful modality in
assessing pancreatic parenchyma and ductal abnormalities. Due to the
increased risk of developing pancreatic adenocarcinoma in patients with
chronic pancreatitis compared to the general population, EUS with biopsy
may be useful in obtaining tissue diagnosis when there is a change in
appearance of the pancreas on serial imaging that is concerning for
development of cancer.
Resection Procedures
PANCREATICODUODENECTOMY
Traditionally, pancreaticoduodenectomy (PD) and pylorus-preserving PD
have been the main surgical approaches for chronic pancreatitis with head
enlargement and no upstream ductal dilatation, as the head and uncinate
process are viewed as the driver of chronic pancreatitis. At the University of
Cincinnati, we perform pancreatic head resection if the main pancreatic duct
is < 7 mm in caliber. The main argument in support of resection is that an
inadequate drainage procedure does not address the primary goal of pain
relief when leaving involved parenchyma behind.
Although the step-to-step details of PD vary, the procedure consists of
three anastomoses: pancreaticojejunostomy, choledochojejunostomy, and
duodeno- or gastrojejunostomy. When performed at high-volume centers, the
procedure is associated with mortality rates < 2%, though morbidity rates
remain up to 50%. In comparison to patients with periampullary
malignancies, those with chronic pancreatitis develop firm, fibrotic
parenchyma which facilitates anastomosis with the small bowel, resulting in
reduced rates of pancreatic fistulas and potentially less morbidity.
Pancreaticoduodenectomy Technique
A bilateral subcostal incision or midline incision is made to enter the
abdomen. After an initial abdominal exploration, the lesser sac is entered,
and the hepatic flexure is mobilized off the duodenum. Adhesions from the
pancreas head to the stomach, if present, are taken down and the
infrapancreatic SMV is identified. Next, an extended Kocher maneuver is
performed to mobilize the duodenum and pancreatic head to the level of the
superior mesenteric vein anteriorly and left renal vein posteriorly.
The portal dissection begins with a cholecystectomy. The common bile
duct is isolated, and the common hepatic duct is divided just above the
junction with the cystic duct. Next, the distal common bile duct is dissected
to the pancreatic head, and the gastroduodenal artery is identified and
ligated. The portal vein is identified, and the avascular plane between the
anterior portal vein and posterior pancreas is developed. Once this is
completed, the stomach is divided (or duodenum if pylorus-preservation is
desired), as well as the jejunum approximately 8-10 cm distal to the ligament
of Treitz. This is followed by division of the pancreas along the created
dissection plane from the superior mesenteric vein to its confluence with the
portal vein.
The reconstruction is performed by first performing a retrocolic end-to-
side pancreaticojejunostomy followed by a retrocolic end-to-side
hepaticojejunostomy. Finally an antecolic gastrojejunostomy (Figure 3 ) or
duodenojejunostomy (Figure 4 ) is performed. We selectively place feeding
jejunostomy or gastric tubes, taking into account the patient’s preoperative
nutrition status and prior use of parental nutrition, and weigh this with the
risk of developing delayed gastric emptying (DGE) or likelihood of needing
to supplement poor oral intake.
FIGURE 3: In the “classic” Whipple, the pylorus and distal part of
the stomach are removed along with the pancreatic head, and
reconstruction consists of a retrocolic end-to-side
pancreaticojejunostomy, followed by a retrocolic end-to-side
hepaticojejunostomy, and an antecolic gastrojejunostomy.
FIGURE 4: In the pylorus-sparing Whipple, the stomach and
pylorus remain intact, and reconstruction consists of a
duodenojejunostomy instead of a gastrojejunostomy.
BEGER PROCEDURE
Duodenum-preserving pancreatic head resection (DPPHR), in which the
duodenum is left intact and resection of the pancreatic parenchyma is limited,
has gained traction as an alternative to PD (Figure 5 ). This procedure was
first introduced by Dr. Hans G. Beger in 1972, under the notion that
conservative resection of inflamed pancreatic tissue would provide pain
relief and spare the patient from the high morbidity associated with an
extensive PD. The contemporary modification of the Beger procedure
consists of two pancreaticojejunostomies and a jejunojejunostomy. Similar to
PD, the indications are head-predominant, small duct chronic pancreatitis.
However, the Beger procedure is not appropriate when malignancy is
suspected.
FIGURE 5: Subtotal resection of the head and neck of the

pancreas above the portomesenteric axis is performed. A roux limb
brought through the transverse mesocolon is used to perform a side-
to-side anastomosis to the resection cavity in the pancreatic head
and an end-to-end anastomosis to the distal pancreas.
Beger Technique
A bilateral subcostal incision or midline incision is made. The lesser sac is
opened, and adhesions from the pancreas are taken down. A Kocher
maneuver is performed, and the infrapancreatic superior mesenteric vein is
identified. After ligation of the right gastroepiploic arteries, the pancreas
neck is transected. Next, sutures are placed in the pancreatic head along the
C-loop of the duodenum, to aid in both hemostasis and traction (our
preference is to place two rows of 3-0 polypropylene).
Subtotal resection of the head and neck of the pancreas above the
portomesenteric axis is performed with electrocautery or the ultrasonic
dissector. Next, a cholecystectomy is completed, if a gallbladder is present.
Using a roux limb brought through the transverse mesocolon, two
pancreaticojejunostomies (side-to-side anastomosis to the resection cavity in
the pancreatic head, and end-to-end anastomosis to the distal pancreas) are
created, followed by a downstream jejunojejunostomy to reestablish
intestinal continuity.
FREY PROCEDURE
The Beger procedure was revised in 1987 by Dr. Charles F. Frey to consist
of duodenum-sparing head resection combined with a lateral
pancreaticojejunostomy, a modification which now bears his name (Figure 6
). Similar to PD and Beger procedures, the Frey procedure is indicated for
head-predominant, small duct chronic pancreatitis. In addition, it is also well
suited for patients with the presence of “chain-of-lakes,” or multiple
strictures in the proximal pancreatic duct.
FIGURE 6: Approach is similar to Beger procedure. The
pancreatic duct is opened from the pancreatic head to 1-2 cm of the
pancreatic tail. After a limited resection of the pancreatic head, a
roux jejunal limb is used to construct a lateral
pancreaticojejunostomy to the parenchyma and exposed duct.
Given that the Frey procedure avoids dissection of the pancreas above the
portal and mesenteric vasculature, the operative time and morbidity is
significantly shorter than the Beger procedure and PD. A single institution
RCT by Bachmann et al. found the Beger and Frey procedures to be
comparable in postoperative mortality, quality of life, and pain control.
There were also no differences in survival, endocrine insufficiency, or
exocrine insufficiency at 16-year follow-up.
Frey Technique
The initial operative approach is similar to the Beger technique–enter the
lesser sac, conduct a Kocher maneuver, and ligate the right gastroepiploic
arteries. Next, the pancreatic duct is assessed by palpation or with the use of
intraoperative ultrasound and is opened using electrocautery. The duct is
opened from the pancreatic head to 1-2 cm of the pancreatic tail. Prior to
coring out the pancreatic head, we prefer to place stay sutures to gain
hemostasis and traction. After a limited resection of the pancreatic head, a
roux jejunum limb is used to construct a lateral pancreaticojejunostomy to the
parenchyma and exposed duct. Finally, intestinal continuity is re-established
with a jejunojejunostomy.
BERNE PROCEDURE
The Berne procedure is a combination of the Beger and Frey procedures–a
subtotal pancreas head resection is performed but avoids division of the
pancreas neck and drainage of the pancreatic duct via creation of a lateral
pancreaticojejunostomy (Figure 7 ). Proponents of the Berne procedure
claim that the modification reduces morbidity while adopting the highlights
of both Beger and Frey procedures. A single institution RCT from the
University of Heidelberg demonstrated superior perioperative outcomes and
equivalent long-term patient quality of life, endocrine and exocrine
pancreatic function, and need for additional interventions. Larger, multi-
institutional trials are needed to evaluate the Berne procedure with Beger
and Frey procedures.
FIGURE 7: This is a combination of the Beger and Frey
procedures-a subtotal pancreas head resection is performed by
avoids division of the pancreas neck and drainage of the
pancreatic duct via creation of a lateral pancreaticojejunostomy.
RESULTS OF PANCREAS HEAD

RESECTION
A review of our institutional experience by McClaine et al. demonstrated that
DPPHR procedures were equally as effective as PD in terms of narcotic
independence and improved quality of life, and with similar rates of
postoperative insulin requirement and incidence of endocrine insufficiency.
In addition, DPPHR outperformed PD in postoperative metrics of estimated
blood loss, operative time, and length of stay.
A meta-analysis of four randomized controlled trials comparing DPPHR
to PD demonstrated that both types of procedures were equally durable with
regard to postoperative pain relief and overall mortality. Although the
morbidity of PD has been previously reported to be 53% compared to 18%
following DPPHR, there was no overall difference in morbidity rates,
including the development of pancreatic fistula, on meta-analysis. Despite
this, DPPHR outperformed PD in specific perioperative clinical parameters,
and were associated with less blood transfusion, and decreased operative
time, rates of DGE, and length of stay. Analysis of long-term factors
demonstrated equivalence in incidence of postoperative endocrine
insufficiency between the two types of procedures, although patients who
received DPPHR had reduced exocrine impairment.
Our practice is to utilize PD if there is any concern of malignant
transformation. Otherwise, in the absence of strictures and a normal sized
pancreatic duct, we favor the Beger procedure. The Frey procedure is
utilized in the presence of strictures and focal obstructions with a “chain-of-
lakes” appearance.
DRAINAGE PROCEDURE: PUESTOW

Although retrograde drainage of the pancreas by pancreaticojejunostomy has
been extensively described since the early 20th century, the procedure did
not adequately address patients with chronic pancreatitis complicated by
numerous strictures and obstructions along the length of the pancreatic duct.
In 1958, Dr. Charles B. Puestow and Dr. William J. Gillesby described the
addition of longitudinal pancreatotomy to decompress the pancreatic duct,
after which they slipped the “tail of the pancreas into the opened, divided
jejunum and brought the jejunum to the right, using the jejunum as a sheath
over the opened pancreas.” Dr. Puestow’s procedure was subsequently
modified by Dr. Philip F. Partington and Dr. Robert E. Rochelle into the
operative technique commonly practiced today.
Modified Puestow Technique (Figure 8)
FIGURE 8: Similar to the Frey procedure, palpation or
intraoperative ultrasound is used to identify the pancreatic duct.
Using electrocautery, the duct is exposed. At lease 7cm of
pancreatotomy is created to ensure adequate drainage. After
clearing the duct of calculi, a side-to-side roux-en-y retrocolic
pancreaticojejunostomy is created, incorporating full thickness
bowel wall to the pancreatic capsule using 3-0 monofilament suture.
After a bilateral subcostal or midline abdominal incision is created, the

lesser sac is entered. A Kocher maneuver is performed to expose the
pancreas. We prefer to divide the short gastric vessels to allow greater
retraction of the stomach and exposure of the pancreas. Similar to the Frey
procedure, we use palpation or intraoperative ultrasound to identify the
pancreatic duct. Using electrocautery, the duct is exposed (within 1 cm of the
duodenum to include both duct of Wirsung and Santorini, and to within 1 cm
of the tail). To ensure adequate drainage, at least 7 cm of pancreatotomy is
necessary. Following adequate exposure of the duct and clearance of calculi
present, a side-to-side, roux-en-y retrocolic pancreaticojejunostomy is
created, incorporating full thickness bowel wall to the pancreatic capsule
using 3-0 monofilament suture.
RESULTS OF DRAINAGE PROCEDURES

The modified Puestow procedure has been demonstrated to be safe, with a
mortality rate < 1%. As pancreatic head resection has been associated with
new-onset diabetes, the avoidance of parenchymal resection with the
modified Puestow procedure preserves pancreatic endocrine and exocrine
function. In a case series by Nealon et al. of 124 patients who underwent
modified Puestow procedure, 106 (86%) patients demonstrated adequate
pain relief and narcotic independence. Further series have demonstrated
procedural success when the pancreatic duct size was at least 7 mm in
caliber. Among patients who failed following this procedure, leaving behind
inflammatory tissue, the “pacemaker of pain,” may be responsible. Thus, we
employ the modified Puestow procedure in patients with a minimal duct size
of 7 mm, and in the absence of head-predominant disease (<4 cm diameter).
CONCLUSION
Operative management of patients with chronic pancreatitis who have failed
conservative measures is dictated by anatomical involvement of the
parenchyma and ductal system. The primary goal of intervention is durable
pain relief, followed by preservation of functional pancreatic tissue.
Surgeons dedicated to caring for this unique patient population should be
familiar with both resection and drainage procedures.
For the majority of treated patients, a tailored surgical approach is
successful with regard to pain relief and narcotic independence. Up to 25%
of patients, however, may fail to achieve symptomatic relief following index
operation. Our practice is to reevaluate these patients and tailor the
revisional operation based on the anatomical constraints. Among patients
with diffuse involvement of the (remnant) pancreas gland, we utilize
completion pancreatectomy with islet cell transplant as salvage therapy. Our
experience with revisional surgery has been effective in relieving pain, with
a quarter of these patients demonstrating narcotic independence and more
than half maintaining decreased narcotic requirements. We have previously
demonstrated total pancreatectomy with islet cell transplant (TP/ICT) as an
effective option in patients with genetically linked pancreatitis and minimal-
change chronic pancreatitis.
Future work is needed on determining preoperatively which patients are
at risk for failure after the index operation and may benefit from TP/ICT as
the initial surgical intervention. Figure 9 summarizes our surgical algorithm
for patients with head-predominant chronic pancreatitis.
FIGURE 9: Surgical Algorithm for Patients with Head-Dominant
Chronic Pancreatitis
SALIENT POINTS
Operative approach for chronic pancreatitis is dictated by anatomical
abnormalities.
Pancreaticoduodenectomy is mandated for clinical suspicion and
diagnostic workup suggestive of malignancy.
Duodenal-preserving pancreatic head resection and
pancreaticoduodenectomy are equally effective with regard to
achieving durable postoperative pain relief.
Frey procedure is warranted for small duct chronic pancreatitis with
multiple duct strictures.
The modified Puestow drainage procedure is warranted for large duct
(> 7 mm) chronic pancreatitis ± multiple duct strictures.
SELECTED REFERENCES
1. Ahmad SA, Wray C, Rilo HL, Choe KA, Gelrud A, Howington JA.
Chronic pancreatitis: recent advances and ongoing challenges. Curr
Probl Surg. 2006; 43(3): 127-238.
2. van der Gaag NA, van Gulik TM, Busch OR, Sprangers MA, Bruno MJ,
Zevenbergen C. Functional and medical outcomes after tailored surgery
for pain due to chronic pancreatitis. Ann Surg. 2012; 255(4): 763–70.
3. Bachmann K, Tomkoetter L, Erbes J, Hofmann B, Reeh M, Perez D.
Beger and Frey procedures for treatment of chronic pancreatitis:
comparison of outcomes at 16-year follow-up. J Am Coll Surg. 2014;
219(2): 208–16.
4. Klaiber U, Alldinger I, Probst P, Bruckner T, Contin P, Köninger J.
Duodenum-preserving pancreatic head resection: 10-year follow-up of
a randomized controlled trial comparing the Beger procedure with the
Berne modification. Surgery. 2016; 160(1): 127-35.
5. McClaine RJ, Lowy AM, Matthews JB, Schmulewitz N, Sussman JJ,
Ingraham AM. A comparison of pancreaticoduodenectomy and
duodenum-preserving head resection for the treatment of chronic
pancreatitis. HPB (Oxford). 2009; 11(8): 677-83.
6. Diener MK, Rahbari NN, Fischer L, Antes G, Büchler MW, Seiler CM.
Duodenum-preserving pancreatic head resection versus
pancreatoduodenectomy for surgical treatment of chronic pancreatitis: a
systematic review and meta-analysis. Ann Surg. 2008; 247(6): 950–61.
7. Nealon WH, Matin S. Analysis of surgical success in preventing
recurrent acute exacerbations in chronic pancreatitis. Ann Surg. 2001;
233(6): 793-800.
CASE SCENARIO
A 45-year-old male undergoes a CT for assessment of abdominal pain after a
vehicle accident. Imaging is unremarkable for trauma-related findings, and
the patient is set for discharge. However, an incidental finding of a 2 cm
mass in the head of the pancreas is noted (Figure 1 ). The lesion has cystic
characteristics and is consistent with a branch-duct Intraductal Papillary
Mucinous Neoplasm (BD-IPMN). The patient is a smoker with no significant
past medical history. His father died of metastatic pancreatic
adenocarcinoma at the age of 61. Should I operate or follow?
FIGURE 1: Computed tomography in a 45-year-old patient with a
2 cm branch-duct IPMN in the pancreatic head (arrow).
█ BACKGROUND
The complexity of IPMN management proves to be one of the biggest puzzles
of modern surgical oncology; the million-dollar question is how to determine
who should undergo surgical resection, who needs surveillance, and who
needs no further attention. The clinical decision is difficult, since the risk of
malignancy in not operating must be balanced against perioperative
morbidity and mortality, and potential long-term complications, such as
endocrine insufficiency, in patients undergoing surgical resection.
IPMNs are mucin-producing cystic neoplasms arising within the
pancreatic ducts. These lesions are precursors to invasive pancreatic ductal
adenocarcinoma (PDAC). The progression of IPMN to invasive cancer
occurs through an adenoma-to-carcinoma sequence analogous to colorectal
polyps. Despite the precursor nature of IPMN, evidence suggests that the
overwhelming majority of these lesions will not progress to cancer. Thus, the
management of IPMN consists of thoughtful selection of patients for resection
versus observation. Patients with features associated with an increased risk
of either high-grade dysplasia (carcinoma in situ ) or invasive cancer will
benefit from resection. The evidence we use to stratify that risk is limited,
and we have little data on the long-term biological behavior of most
individuals with IPMN. Our current understanding of features associated
with high-grade dysplasia (HGD) or invasive cancer is based on
retrospective series of resected patients. These studies lack a common
denominator of all patients with IPMN lesions, since they are already
selected for surgery. Due to the lack of definitive evidence on numerous
aspects of IPMN biology, expert consensus statements have been developed.
These guidelines are based on judgment of experts in the field using extended
clinical experience and limited evidence.
The examined scenario is typical of what is frequently seen in the
Multidisciplinary Pancreatic Cyst Clinic of Johns Hopkins Hospital. Our
approach to a patient with a pancreatic cystic neoplasm is to start with the
premise that the clear majority of these lesions are benign at the time of
diagnosis and only a very small percentage will progress to malignancy.
Therefore, the decision to offer surgical resection should only be made to the
patients who have a high probability of benefitting–a small minority of the
total population. For each patient, we need to assess a variety of different
clinical, radiological, and pathological parameters, assign a risk of
malignancy, and conduct an individualized treatment plan. We generally
follow either the International Association of Pancreatology or the European
Consensus guidelines (ICG/ECG) in the management of our patients. The
general workup includes a careful assessment of age, gender, and medical
history. In addition, high-quality pancreas-protocol cross-sectional imaging
is obtained, if not already available; this will include either a CT scan or an
MRI. Based on this information alone, surgical resection can sometimes be
recommended and no further work-up is needed. If the initial analysis
identifies no indications for surgical resection (per ICG/ECG), we often
perform an endoscopic ultrasound (EUS) with fine needle aspiration (FNA)
and cyst fluid analysis for a cyst larger than 1 cm. The main indication for
this procedure is to evaluate for severe atypia and the possible presence of a
solid component. If these elements are absent, the patient can safely undergo
surveillance. With these things in mind, the following is how we would think
of the management of the individual presented in this scenario.
PRESENTATION
In a 45-year-old male with a pancreatic cyst of this appearance the most
likely diagnosis is a BD-IPMN and less likely an oligocystic serous
cystadenoma or a cystic pancreatic neuroendocrine tumor (PNET). Based on
appearance and demographics, this lesion is highly unlikely to be a solid
pseudopapillary neoplasm or a mucinous cystic neoplasm. Thus, we have
given this gentleman the presumed diagnosis of BD-IPMN. The workup will
be to confirm the diagnosis of BD-IPMN as best as possible with current
methodology. Once we make the presumptive diagnosis of BD-IPMN, we
then focus on ruling out any “high-risk stigmata” that would warrant
resection. As mentioned above, the vast majority of BD-IPMN can be
observed. In the assessment of this patient we will address the following
questions:
DOES THE PATIENT HAVE ANY “HIGH-
RISK STIGMATA ” THAT WARRANT
RESECTION OR “ WORRISOME FEATURES”
THAT WARRANT FURTHER ASSESSMENT
WITH EUS-FNA?
Imaging
The radiological assessment of BD-IPMNs is the big weapon in our arsenal
to stratify, as correctly as possible, patients who are at high-risk of invasive
cancer and may need surgery. In our practice, we have adopted a
combination of the ICG/ECG criteria to assess the imaging characteristics in
BD-IPMN patients (Table 1 ). This patient already has a high-quality CT
scan that does not show a mass lesion, ductal dilatation, or other concern for
malignancy. With this in mind, he will undergo a Magnetic Resonance
Imaging/Cholangiopancreatography (MRI-MRCP, Figure 2A ) to assess
further and more accurately the incidental BD-IPMN. Utilization of MRI-
MRCP as the imaging modality of choice increases the chances of a
definitive radiologic diagnosis and provides substantial information on a
possible communication of the BD-IPMN with the main pancreatic duct
(MPD). If MRI is not available, a multidetector computed tomography (CT)
with <2 mm sections over the area of interest is sufficient.
FIGURE 2: MRI (A) and Endoscopic Ultrasound (B) of the Same
Patient.
TABLE 1: Imaging Characteristics Utilized for Branch

Duct-IPMN Stratification
Identification of enhancing solid component in the cystic wall is

suggestive of invasion and an indication for surgical resection. Furthermore,
thickened cyst wall and presence of nonenhancing mural nodules
(“worrisome features”) are risk factors, and additional workup is required.
MPD dilatation or abrupt caliber change is rare in BD-IPMNs, and it usually
suggests a mixed-type lesion. If this patient has no suspicious BD-IPMN
imaging characteristics, he is considered a candidate for surveillance. An
EUS is utilized for further characterization of a BD-IPMN with “worrisome
features” or to verify the absence of enhancing solid component, if the lesion
is >1cm (Figure 2B ); an FNA is also performed simultaneously. With the
EUS we can accurately assess the thickening of the cyst wall and the status of
mural nodules. In summary, our patient has imaging findings consistent with a
BD-IPMN with no “worrisome features” or “high-risk stigmata.”
Key points:
Utilize MRI-MRCP or a multidetector CT for accurate assessment of the
lesion
Enhancing solid component in the cystic wall or a dilated MPD is an
indication for surgical resection, if clinically appropriate
Perform an EUS-guided FNA if the lesion has “worrisome features” in
order to assess for the presence of severe atypia or solid component not
seen on cross-sectional imaging.
Assessment of cyst-related symptoms

In this scenario, the pancreatic cystic lesion was discovered incidentally, and
the patient was asymptomatic. Patients with BD-IPMN who are
asymptomatic have a low risk of malignancy. Occasionally, they may present
with non-specific symptomatology, such as nausea, anorexia, and abdominal
discomfort. However, the symptoms that are strongly associated with
malignancy are jaundice, severe abdominal or back pain, unintentional
weight loss, and pancreatitis. Jaundice in BD-IPMN patients is included in
the ICG “high-risk stigmata” of malignancy and is considered an indication
for surgical resection. Additionally, chronic low-grade pancreatitis
secondary to obstruction may induce parenchymal atrophy and subsequent
endocrine dysfunction. Recent-onset diabetes mellitus is not a strong factor
for IPMN-associated malignancy, but one must be suspicious.
Key points:
Absence of clinical symptoms is favorable towards surveillance
Patients with jaundice, severe pain, and unintentional weight loss have
an indication for surgical resection, if clinically appropriate
Pancreatitis, pancreatic insufficiency, and recent-onset diabetes mellitus
are risk factors of malignant progression and resection may be
considered when accounting for other factors.
Are there any biomarkers that could be helpful?

Utilization of cancer-specific biomarkers in early detection of invasive
malignancy in IPMN is a field of intense investigation. Several studies have
correlated higher CA19-9 values with increased risk of malignancy in
IPMNs. Measurement of serum CA19-9 values should be performed in a
newly diagnosed BD-IPMN patient with “worrisome features”; nonetheless,
our experience indicates that the sensitivity and specificity of this marker is
low. One may consider suspicious a CA19-9 value >35 U/mL and investigate
further. Recent studies have shown that biomarkers of immunologic response,
such as the neutrophil-to-lymphocyte ratio may also prove useful in early
prediction of invasive development in BD-IPMNs.
The introduction of cytological and molecular assays for cyst fluid
analysis in IPMN has shifted the interest away from the traditional blood-
associated biomarkers. If the BD-IPMN in this scenario has imaging
“worrisome features” that increase the suspicion for malignancy, an EUS-
FNA and cytological analysis of the cystic fluid are advised to characterize
the pathological background of the lesion. However, the neoplastic epithelial
cells of BD-IPMN are rarely described in EUS-FNA specimens; in very few
cases, cells with high-grade atypia or malignancy may be identified in
cytology, and then a diagnosis of in situ or invasive carcinoma is likely.
When cytology is negative, its utility in BD-IPMN risk stratification and
surgical decision making is limited, and additional analysis of cyst fluid
biomarkers of malignancy may assist further.
The presence of mucin and carcinoembryonic antigen (CEA) levels in the
cystic fluid should be assessed. CEA is a molecular surrogate marker for
mucin and is utilized in differentiating mucinous from serous cysts, but is not
useful in assessing the risk of malignancy. That is, there is no correlation
between higher risk of malignancy and increased CEA. CEA levels in cyst
fluid have the highest sensitivity and specificity in differentiating a mucinous
pancreatic cystic lesion (73% and 84%, respectively). A cut-off CEA value
of 200 ng/mL is used for diagnostic purposes, in accordance with the ICG
criteria. Increased CA19-9 and amylase concentration in cyst fluid may also
be indicative of a pancreatic lesion with high malignant potential.
Key points:
Measure blood CA19-9 and consider high values as suspicious in
lesions with “worrisome features”
Proceed to cyst fluid analysis to evaluate atypical cells and fluid
biomarkers in a patient with difficult diagnosis.
ARE THERE ANY ISSUES TO CONSIDER

OUTSIDE OF THE STANDARD GUIDELINE
RECOMMENDATIONS? SHOULD WE BE
CONCERNED ABOUT HIS FAMILY HISTORY
AND/OR SMOKING HISTORY?
The patient is 45 years old, a current smoker, and has a family history of
pancreatic adenocarcinoma in a first-degree relative. We have previously
demonstrated that smoking history does not appear to be a risk factor for
harboring high-grade dysplasia or invasive carcinoma in BD-IPMNs.
However, smoking may be associated with the development of concomitant
pancreatic adenocarcinoma and cessation must be advised during patient
surveillance. Additionally, this patient’s father died at a relatively young age
from metastatic pancreatic adenocarcinoma. IPMNs are precursor lesions
and share genetic defects with pancreatic cancer. Several studies have
proven the presence of GNAS, KRAS, and TP53 mutations, even in IPMNs
with low-grade dysplasia. Considering the “two-hit” hypothesis for tumor
suppressor genes, germline mutations inherited by the patient’s father may
lead to early development of IPMN-associated or concomitant pancreatic
carcinoma. Familial clustering has been reported for about 10% of
pancreatic cancer cases. The risk of developing pancreatic cancer increases
substantially with two or more affected first-degree relatives. This alone
warrants close surveillance, preferably through an established protocol, but
does not change the indications for resection of IPMN. The relative risk of
developing pancreatic cancer in a patient with an IPMN and a positive
family history is unknown.
Key points:
Smoking is not considered a risk-factor for BD-IPMN progression;
however, always advise cessation, since there is a proven correlation
between smoking and pancreatic cancer
A positive family history of pancreatic cancer in first degree relatives
increases the risk for pancreatic cancer; that risk is substantially higher
with 2 first degree relatives.
Recommendation
The patient in this scenario has a presumed BD-IPMN with no features
associated with increased risk of high-grade dysplasia or invasive
malignancy. He will not benefit from resection at this time and should
undergo close observation.
NOW THAT WE DECIDED TO NOT

OPERATE ON THIS PATIENT, HOW SHOULD
WE FOLLOW HIM?
If the patient has no concerning features in imaging, and EUS-FNA is
negative for cells with high-grade atypia, he is a candidate for surveillance.
The protocol followed in our clinic for a 2-cm BD-IPMN is MRI/MRCP
every 6 months, combined with physical examination and serological marker
surveil lance. The follow-up period is continued indefinitely in cases of
positive family history for pancreatic adenocarcinoma, and assessment
intervals may increase to one year, if the BD-IPMN remains unchanged for
more than two years. We are now aware that the risk for disease progression
and for potential surgery increases over time. Patient management adjusts to
any changes in symptomatology or imaging characteristics of the lesion,
accordingly; surgical resection is indicated if the lesion develops suspicious
attributes (eg, “high-risk stigmata”).
IF THE IDENTIFIED LESION PROVED TO BE

A MAIN-DUCT IPMN (MD-IMPN), WOULD
OUR APPROACH DIFFER?
Main-duct IPMNs (MD-IPMNs) involve the main pancreatic duct (MPD) and
account for almost 25-30% of the resected IPMN lesions; in asymptomatic
patients, they are less frequent. In almost all cases, a MD-IPMN is
characterized by distinct dilatation (over 5mm) and/or abrupt change in the
caliber of the MPD, as a result of mechanical obstruction of pancreatic juice
outflow. Jaundice is also a common symptom at presentation when the lesion
is located in the pancreatic head, whereas new-onset diabetes and pancreatic
gland atrophy may be present in chronic cases. Preoperative association of
an IPMN lesion with the main pancreatic duct through imaging may prove
difficult; existing classification results primarily from pathological data of
resected cases. Large retrospective studies suggest that MD-IPMNs show a
significantly different and more aggressive biological behavior than BD-
IPMNs. A higher risk for presence of invasive component has been strongly
correlated with MD-IPMNs. This has a practical impact on the management
of patients with MD-IPMNs: surgical resection is recommended in all cystic
lesions that involve the MPD in imaging.
Currently, the definitive way to classify an IPMN lesion and its
relationship with the duct is with surgical resection and pathological
assessment of the specimen. However, imaging characteristics of the lesion
and patient’s signs and symptoms can provide information about the origin
and associations of the cyst. One must always keep in mind that a direct
connection of the cyst with the MPD in CT accompanied by a dilatation over
5mm is a strong indication that the lesion is a MD-IPMN or a mixed-type
IPMN, that involves both the main and branch pancreatic ducts. When in
doubt regarding the classification of an IPMN, assess the radiological
characteristics of the lesion and the patient’s clinical presentation, and
follow the flowchart for BD-IPMNs.
CONCLUSIONS
Most 2 cm BD-IPMN cases fall into the category where surgical resection is
not clearly indicated. The optimal therapeutic approach is to assess the
patient’s performance status, clinical signs, and radiology findings, and
consult with them over the advantages and disadvantages of surveillance and
surgery. An assessment of the case under a multidisciplinary setting can
provide an individual comprehensive approach. The main parameters that
dictate the management of a young patient with a 2-cm BD-IPMN are
available in Table 2 . The surgeon must keep in mind that each patient is
different, and all available clinical and imaging information are valuable in
determining the therapeutic approach. The patient must be actively involved
in the treatment plan, and if available, a multidisciplinary committee should
be consulted.
TABLE 2: Different pathways in Branch Duct-IPMN

management
SALIENT POINTS
Utilize MRI-MRCP or a multidetector CT for accurate assessment of
the lesion.
Enhancing solid component in the cystic wall or a dilated MPD is an
indication for surgical resection, if clinically appropriate.
Perform an EUS-guided FNA if the lesion has “worrisome features” in
order to assess for the presence of severe atypia or solid component
not seen on cross-sectional imaging.
Absence of clinical symptoms is favorable towards surveillance.
Patients with jaundice, severe pain, and unintentional weight loss have
an indication for surgical resection, if clinically appropriate.
Pancreatitis, pancreatic insufficiency, and recent-onset diabetes
mellitus are risk factors of malignant progression, and resection may be
considered when accounting for other factors.
Measure blood CA19-9 and consider high values as suspicious in
lesions with “worrisome features.”
Proceed to cyst fluid analysis to evaluate atypical cells and fluid
biomarkers in a patient with difficult diagnosis.
Smoking is not considered a risk factor for BD-IPMN progression;
however, always advise cessation, since there is a proven correlation
between smoking and pancreatic cancer
A positive family history of pancreatic cancer in first-degree relatives
increases the risk for pancreatic cancer; that risk is substantially higher
with 2 first-degree relatives.
SELECTED REFERENCES
1. Tanaka M, Fernandez-Del Castillo C, Kamisawa T, et al. Revisions of
International Consensus Fukuoka Guidelines for the management of
IPMN of the Pancreas. Pancreatology. 2017 Jul 13. [Epub ahead of
print]
2. Del Chiaro M, Verbeke C, Salvia R, Kloppel G, Werner J, McKay C, et
al. European experts’ consensus statement on cystic tumours of the
pancreas. Dig Liver Dis. 2013; 45(9):703-11.
3. Gemenetzis G, Bagante F, Griffin JF, et al. Neutrophil-to-lymphocyte
Ratio is a Predictive Marker for Invasive Malignancy in Intraductal
Papillary Mucinous Neoplasms of the Pancreas. Ann Surg. 2017
Aug;266(2):339-345.
4. Rezaee N, Khalifian S, Cameron JL, Pawlik TM, Hruban RH, Fishman
EK, et al. Smoking is not associated with severe dysplasia or invasive
carcinoma in resected intraductal papillary mucinous neoplasms. J
Gastrointest Surg. 2015;19(4):656-65.
5. He J, Cameron JL, Ahuja N, Makary MA, Hirose K, Choti MA, et al. Is
it necessary to follow patients after resection of a benign pancreatic
intraductal papillary mucinous neoplasm? J Am Coll Surg. 2013;
216(4):657-65; discussion 665-7.

CASE SCENARIO
A 58-year-old man with chronic abdominal pain was seen for management of
chronic calcific pancreatitis associated with biliary and pancreatic ductal
dilation. The patient has a history of inveterate alcohol and tobacco use. CT
scan showed a mixed density mass along the medial aspect of the second
portion of the duodenum that was unchanged from prior scans. On ERCP
there was narrowing of the distal pancreatic duct in the head with dilation in
the body and tail. CEA was 19.8 ng/ml and CA19-9 was 23 ng/ml. MRI of
the pancreas showed a cystic mass in the pancreatoduodenal groove with
associated fibrotic tissue, consistent with groove pancreatitis and associated
obstruction of the distal common bile and main pancreatic ducts (Figure 1a-c
). There were multiple large intraductal calculi in the dilated distal main
pancreatic duct. The patient underwent a pancreaticoduodenectomy that was
uneventful; he was discharged from the hospital on the 6th postoperative day.
FIGURE 1A: Groove pancreatitis showing dilated pancreatic duct
(thick arrow) and dilated common bile duct (thin arrow).
FIGURE 1B: Groove pancreatitis with peripancreatic fluid (thin

arrow) and edematous duodenum (thick arrow).
FIGURE 1C: Groove pancreatitis showing duodenum (thick arrow)
and largest of several cysts in the duodenal wall (thin arrow).
█ BACKGROUND
Groove pancreatitis is an uncommon presentation of chronic pancreatitis and
is frequently unrecognized by radiologists who are unpracticed with cross-
body imaging in the diagnosis of chronic pancreatitis. Groove pancreatitis is
a form of focal chronic pancreatitis within the pancreas tissue that lies
between the duodenal wall and the intrapancreatic portion of the common
bile duct. It has both a cystic and solid presentation and may be described as
cystic dystrophy of the duodenal wall, duodenal dystrophy, and paraduodenal
pancreatitis. The variations of nomenclature relate to the pathogenesis of this
disorder. An embryologic variation in the differentiation and fusion of the
pancreas and duodenum leads to ectopic or heterotopic location of
pancreatic tissue. This variation in pancreatic anatomy does not usually
present until the adult period. It is likely that the development of chronic
pancreatitis in ectopic pancreas tissue requires a genetic predisposition as
well as environmental factors such as tobacco and alcohol abuse as seen in
the index case.
Groove pancreatitis involves progressive cystic degeneration of ectopic
pancreas rests that lie within the duodenal wall. Pancreatic rests can occur in
other locations, but when located within the pancreaticoduodenal groove, the
development of fibrotic and inflammatory tissue leads to a typical radiologic
and clinical presentation. Groove pancreatitis may not involve the rest of the
pancreas with the inflammatory process, though the pancreatic duct and
biliary system are frequently secondarily involved in the process, as was the
case with this patient. Identification of this entity and its varied appearances
as a distinct pathology is important in order to direct the patient to a surgeon
who is willing and able to perform pancreaticododenectomy for chronic
pancreatitis. It is not uncommon for the diagnosis to be overlooked on initial
cross-sectional imaging, though the experienced radiologist readily identifies
the characteristic cystic and inflammatory periduodenal changes.
The imaging appearances and clinical features of groove pancreatitis may
overlap with pancreatic adenocarcinoma and EUS may have a role in
diagnosis when the question of a pancreatic mass is present. CA 19-9 may be
helpful in ruling out a cancer diagnosis. The typical patient is similar to the
patient in this case, a 50-year-old male with a history of chronic relapsing
abdominal pain and weight loss associated with alcohol and tobacco abuse.
Serum laboratory values usually show only mildly elevated amylase and
lipase that help to differentiate from typical acute interstitial or acute
necrotic pancreatitis. Serum liver transaminases may be mildly elevated.
Jaundice and hyperbilirubinemia are less common but can occur mimicking
ductal carcinoma. Patients may also present with symptoms of gastric outlet
obstruction when the inflammatory process leads to duodenal stenosis in the
descending duodenum.
Groove pancreatitis is evident histologically as chronic inflammation and
scarring in the anatomical space between the medial wall of the duodenum
and the head of the pancreas, which is termed the “pancreaticoduodenal
groove.” Initial presentations can be subtle both clinically and on imaging.
The earliest imaging finding is focal inflammation and fibrosis in the
pancreaticoduodenal groove without involvement of the pancreatic or biliary
ductal systems. Ductal dilatation may present at a later stage of the disease
process and can be related to increased fibrosis. Ductal obstruction may be
precipitated by the development of the cystic changes that exert a mass effect
on both the pancreatic and bile ducts near the ampullary region. The
accessory duct of Santorini at the minor papilla is the duct most frequently
involved in the obstructive process. When an associated obstructive
pancreatopathy develops due to focal inflammation and fibrosis, ductal
dilation, proteinaceous plugs, and intraductal calcific stones may develop. At
this stage the appearance of groove pancreatitis may overlap with typical
appearances of advanced chronic calcific pancreatitis. The features of
chronic calcific pancreatitis correspond to the severity of ductal
abnormalities, with chronic stenosis leading to upstream stasis and eventual
intraductal stone formation as demonstrated in this patient.
There are two established typical gross morphologic forms of this
disease: the pure form and the segmental form. In the “pure” form, the
abnormalities are confined to the pancreaticoduodenal groove, and the head
of the pancreas is spared. In the “segmental” form, the inflammation and
scarring extend to directly involve the head of the pancreas, which frequently
leads to intrapancreatic abnormalities including cysts and fibrotic masses,
which was the case with this patient.
Prior to undertaking pancreatic resection, a trial of smoking and alcohol
cessation is warranted. Endoscopic management of terminal biliary stenosis
and pseudocyst drainage may have a role while the benefits of smoking and
alcohol cessation are given some time to work. When chronic inflammatory
changes of an obstructive pancreatopathy develop in the body and tail of the
pancreas, resection of the head of the pancreas has a lower complication rate
because the anastomosis is constructed to a fibrotic pancreas with a dilated
duct. When that is the case, the postoperative course is similar to that of this
patient in which no pancreatic fistula developed, and the patient had an
expeditious discharge to home. Although operative complications related to
anastomotic leak and fistula formation are higher in the patient without
secondary inflammatory changes in the remnant pancreas, long-term patient
outcomes are better because of the absence of chronic inflammatory changes
in the remnant pancreas.
CASE SCENARIO
A 33-year-old woman was referred for surgical management of recurrent
acute pancreatitis associated with pancreas divisum. She had a 5-year history
of recurrent abdominal pain and nausea associated with multiple emergency
room visits and several hospitalizations with associated serum lipase
elevation and mild changes of acute pancreatitis on CT scan. Past surgical
history was significant for morbid obesity managed with a laparoscopic
gastric bypass. One year previously the patient had undergone on open minor
duct sphincteroplasty that decreased emergency department visits and
hospitalizations but failed after 6 months and led to a subsequent
pancreaticoduodenctomy, which also failed after 12 months. A completion
pancreatectomy with islet autotransplantation was performed. The resected
pancreas was atrophic and fibrotic with consequent poor islet cell yield.
█ BACKGROUND
The strategies utilized in the surgical management pancreas divisum are those
used in the surgical management of chronic pancreatitis: improve drainage,
resect damaged tissue, or combine the first two with a simultaneous resection
and drainage procedure. This patient demonstrates how all three can be
utilized in an ineffective way that delays optimal therapy. The prudent
surgeon always remembers that pancreas divisum is as common as left-
handedness: ten percent of the population has pancreas divisum. The concept
was promulgated in the 1970s that in pancreas divisum there is an anatomic
impediment to the normal drainage of pancreatic exocrine secretions that
results in an obstructive pancreatopathy. When divisum anatomy is present,
the dorsal duct, formerly known as the duct of Santorini, is called the
dominant duct. Variations of the dominant dorsal duct are many and
infrequent enough so that radiologic and endoscopic delineation may be
confusing. The classic radiographic image has the dominant pancreatic duct
crossing the terminal bile duct and entering the descending duodenum
(Figure 2 ). Usually a small ductal system drains the ventral head of
pancreas and enters the duodenum through the major papilla with the terminal
bile duct (Figure 3 ). Beware of the entity called acquired pancreas divisum
that is associated with chronic pancreatitis and malignancy. Either process
may result in total occlusion of the ventral duct, causing the duct of Santorini
to assume responsibility for pancreatic exocrine outflow via the minor
papilla.
FIGURE 2: MRCP demonstrating dorsal duct dominant pancreas
divisum with duct of Santorini coursing horizontally across distal
common bile duct to its downstream union with the duodenum
proximal to the ampulla of Vater.
FIGURE 3: The normal adult pancreas is formed from the fusion
of the dorsal duct and ventral duct. The ventral duct drains 70% of
the pancreas via the major papilla. Failure of the ventral duct to
fuse with the dorsal duct results in pancreas divisum, a name given
because the pancreas is now drained by two ducts. As a result 70%
of the pancreatic drainage is via the minor papilla. Although not
entirely cleared, it is thought that pancreatitis stems from the the
minor papilla being too narrow to adequately drain the pancreas.
With the development of diagnostic ERCP in the ’70 s, idiopathic

recurrent pancreatitis was attributed to pancreas divisum and became a target
for surgical management with duodenotomy and minor duct sphincteroplasty.
As therapeutic endoscopy developed, endoscopic minor duct sphincterotomy
or papillotomy supplanted the open surgical approach. Although
sphincteroplasty was successful in the majority of patients in early studies,
long-term failures were not uncommon. Environmental and genetic factors
are likely to interact with the divisum anatomy in contributing to recurrent
acute and chronic pancreatitis. Pancreas divisum is associated with a higher
prevalence of genetic mutations that are known risk factors for chronic
pancreatitis. Current thinking is that pancreas divisum is not a cause of
pancreatitis by itself but acts as a partner with genetic mutations and
environmental stressors such as alcohol and tobacco abuse.
The clinical presentation of pancreatitis associated with dominant dorsal
duct anatomy has two variations: acute recurrent pancreatitis and chronic
pancreatitis. The former leads to the latter frequently, as demonstrated by this
patient. Rarely does chronic pancreatitis associated with pancreas divisum
result in severe fibrosing pancreatitis associated with biliary, duodenal, or
splanchnic venous obstruction. Severe pancreatitis with pancreatic necrosis
is also uncommonly associated with pancreas divisum. Imaging of this
patient did not show radiologic evidence of acute or chronic pancreatititis.
Attempts at surgical drainage did not prevent the rapid onset of severe
fibrosing pancreatitis.
A hopeful hypothesis recommends that endoscopic minor duct
sphincterotomy be undertaken prior to the development of chronic
inflammatory and fibrotic changes. The chief indication for surgical therapy
of pancreas divisum is pain; endocrine and exocrine insufficiency is rare, as
are biliary, duodenal, and splanchnic venous occlusion. The difficulty of
waiting for chronic morphologic changes to develop is that once chronic
fibrosis has developed in the head of the pancreas, endoscopic and surgical
minor duct sphincterotomy have a limited success rate. Division of the sparse
smooth muscle fibers of the minor sphincter is unlikely to alter the hardened
fibrosis of the surrounding pancreatic parenchyma. In general, endoscopic
therapy for pancreas divisum is typically recommended for pancreatic-type
pain in association with obstructive morphology or documented recurrent
acute pancreatitis without another clear and reversible etiology. Minor
sphincterotomy for abdominal pain alone is unproven and a slippery slope;
once the orifice is compromised, there is always the possibility of orifice
restenosis and duct obstruction.
Operative sphincteroplasty of the sphincter of Henle was the mainstay of
therapy prior to the development of endoscopic sphincterotomy. It is hard to
imagine how an operative sphincteroplasty can be better than an endoscopic
sphincterotomy, given the sparseness of smooth muscle fibers that surround
the pancreatic duct as it courses into the duodenum. Occasional patients with
bypassed foregut anatomy may elect open surgical sphincterotomy and forego
a retrograde endoscopic approach or a hybrid laparoscopic and endoscopic
prograde approach. Surgical outcomes with open sphincteroplasty are best
when endoscopic sphincterotomy has not been done before. The first cut is
the best one. It is tempting to undertake open sphincteroplasty in patients with
altered foregut anatomy, which is how this patient was managed. Another
approach is a hybrid procedure with concomitant laparoscopic gastrostomy
of the bypassed stomach and endoscopic prograde sphincterotomy. A
gastrostomy tube is left in place to secure access for removal of the
anastomotic stent later. Patients with prior Billroth II gastrectomy are not
candidates for retrograde luminal endotherapy. A retrograde approach to the
minor papilla does not provide angulation that allows for endoscopic
sphincterotomy on the minor duct. Patients who have previously undergone
successful endoscopic sphincterotomy with recurrent fibrosis may be
candidates for operative sphincteroplasty. The concept is that open minor
duct sphincteroplasty with loop magnification and fine absorbable suturing of
the duct will have better outcomes than endoscopic sphincterotomy. The
difficulty in undertaking this course of action is that patients with chronic
fibrosis of the head of the pancreas do poorly with sphincteroplasty, and
preoperative identification of pancreatic fibrosis is imperfect.
Pancreatoduodenectomy may be indicated in patients with pancreas
divisum who have failed minor duct sphincteroplasty when changes of
chronic pancreatitis develop in the head of the pancreas. When a resection of
the head is indicated in pancreas divisum, the divided neck of the pancreas is
usually soft with a nondilated pancreatic duct, one at risk for anastomotic
leak. If head resection in pancreas divisum fails and patients become
candidates for Total Pancreatectomy with Islet AutoTransplantation (TPIAT),
islet yields may be uncompromised. Islet volume is greater in the tail and
body than in the head of the pancreas, and islet loss is limited in a head
resection for pancreas divisum. It was remarkable that this patient developed
such severe pancreatic atrophy and fibrosis one year after the Whipple
procedure. This is evidence for the view that when anastomotic obstruction
occurs after a Whipple procedure, acinar cells atrophy before destruction of
the islets. The resultant islet-rich, acinar-poor atrophied pancreas expected
in the case was not found, and the remnant pancreas was not favorable
isolation for islet autotransplantation. It was surprising that the patient
described above had so much islet atrophy and fibrosis and a resultant poor
islet yield. The issue of whether pancreas divisum patients with chronic
pancreatitis should undergo a speculative Whipple procedure or a TPIAT is
unanswered.
The Puestow procedure is an option in the management of pancreas
divisum when pancreatic ductal dilation is present. Variations in the Puestow
procedure and the Whipple procedure have been utilized in chronic
pancreatitis associated with pancreas divisum in a fashion similar to their
utilization in chronic pancreatitis not associated with pancreas divisum.
Though typically the Beger and Frey procedures are selected for patients
with an inflammatory mass in the head of the pancreas, they have been
utilized in pancreas divisum not associated with enlargement of the head of
the pancreas. The Frey procedure was developed due to the ascribed failure
of the LPJ to drain the head and uncinate process of the pancreas. In pancreas
divisum the head should be effectively drained with a longitudinal
ductotomy. The utilization of multiple resection and drainage techniques is an
indication of the difficulty in patient selection and management in chronic
pancreatitis associated with pancreas divisum.
TPIAT is indicated in patients with chronic pancreatitis with intractable
pain who have failed medical, endoscopic, and surgical management.
Pancreas divisum is a risk factor for chronic pancreatitis in about 15% of the
cohort in the reports of TPIAT in the management of chronic pancreatitis.
Typical pancreas divisum patients who are selected to TPIAT have
undergone endoscopic sphincterotomy, operative sphincteroplasty, and
pancreatic head resection in succession over many years. The unanswered
question is whether TPIAT should be undertaken sooner rather than later if
endoscopic treatment fails.
SALIENT POINTS
Groove pancreatitis is a relatively rare disease, which is best
described as chronic inflammation of ectopic pancreatic tissue in the
duodenal wall.
Groove pancreatitis is also known as paraduodenal pancreatitis and
duodenal cystic dystrophy.
Groove pancreatitis is associated with alcohol and tobacco abuse.
Groove pancreatitis may be complicated by duodenal stenosis and
biliary and pancreatic ductal obstruction.
Groove pancreatitis may mimic pancreatic cancer and vice versa.
Pancreas divisum occurs in about ten percent of the population.
In combination with environmental and genetic factors, patients with
pancreas divisum may develop recurrent acute pancreatitis and
subsequent chronic pancreatitis.
Endoscopic minor duct sphincterotomy has a role in the management of
pancreas divisum.
Operative sphincteroplasty may also have a role in patient
management, as may other resection and drainage procedures.
The exact timing and selection of resection and drainage procedures
including total pancreatectomy with islet autotransplantation has not
been determined.
SELECTED REFERENCES
1. Hungerford JP, Neill Magarik MA, Hardie AD. The breadth of imaging
findings of groove pancreatitis. Clin Imaging. 201;39(3):363-6.
2. Egorov VI, Vankovich AN, Petrov RV, Starostina NS, Butkevich AT,
Sazhin AV. Pancreas-preserving approach to “paraduodenal
pancreatitis” treatment: why, when, and how? Experience of treatment
of 62 patients with duodenal dystrophy. Biomed Res Int. 2014; Epub
2014 Jun 5.
3. Morgan KA, Romagnuolo J, Adams DB. Transduodenal
sphincteroplasty in the management of sphincter of Oddi dysfunction and
pancreas divisum in the modern era. J Am Coll Surg. 2008; 206(5):
908-14.
4. Schnelldorfer T, Adams DB. Outcome after lateral
pancreaticojejunostomy in patients with chronic pancreatitis associated
with pancreas divisum. Am Surg. 2003; 69 (12):1041-4.
5. Adams DB, Cote G. Pancreas Divisum and Other Variants of Dominant
Dorsal Duct Anatomy. Current Surgical Therapy. 12th Ed. Elsevier,
New York, 2017, Chapter 95.

CASE SCENARIO
A 56-year-old otherwise healthy female is referred to your office with a 1.5
cm tumor of the head of the pancreas, which was found incidentally during an
emergency room evaluation for abdominal pain. She has no current
complaints or symptoms, and her physical exam is unremarkable. Endoscopic
ultrasound-guided fine needle aspiration of the lesion is consistent with a
neuroendocrine tumor. Complete workup reveals no evidence of metastatic
disease or hormone production. The patient asks for you to, “Take the tumor
out,” because her best friend recently passed away from pancreatic cancer.
█ BACKGROUND
Pancreatic neuroendocrine tumors (PNETs) represent a minority, less than
4%, of pancreatic neoplasms. Compared to other gastrointestinal endocrine
tumors, PNETs are of greater concern because they have significantly poorer
survival. These neoplasms are not homogeneous. Rather, different PNET
diagnoses portend variable long-term outcomes dependent upon presentation
and histology. Management strategies are not uniform, can often be
controversial, and are typically individualized to the patient. Given the
increasing frequency with which high-resolution, cross-sectional imaging is
utilized, there is an increasing abundance of incidental, small, asymptomatic
PNETs. Management of such tumors with surgery versus observation is
controversial. However, a surgical approach is favored in the context of
several clinical scenarios, which are described by this chapter in detail.
TUMOR CHARACTERISTICS
There are many tumor-dependent characteristics of a PNET which favor
surgical resection. These characteristics include functionality, classification,
anatomy, and diagnostic ambiguity.
FUNCTIONAL TUMORS
Sporadic hormone-producing PNETs warrant resection in the absence of
metastatic disease. Resection offers patients symptomatic relief from
hormone overproduction and may improve overall and disease-free survival.
Functional PNETs associated with genetic syndromes, such as Multiple
Endocrine Neoplasia Type I (MEN-1), are discussed later in the chapter.
Hormone-associated tumor syndromes include gastrinomas in Zollinger-
Ellison syndrome, insulinomas leading to the Whipple triad, VIPoma in
Vemer-Morrison Syndrome, carcinoid syndrome, adrenocorticotropic
hormone-omas producing Cushing’s syndrome, glucagonomas,
somatostatinomas, growth hormone releasing factor tumors, parathyroid
hormone-related peptide-omas, and cholecystokininoma. Given the wide
variety of potential functional syndromes, a thorough history and physical, in
addition to metabolic workup is essential after PNET diagnosis. Failure to
identify a functional lesion coupled with the decision for observation over
surgical resection could lead to significant adverse health effects.
NONFUNCTIONAL TUMORS
Multiple classification systems exist for PNETs, and consideration of more
than one system is advisable when embarking upon a surgical pathway for
nonfunctional sporadic tumors. The American Joint Committee on Cancer
(AJCC) staging system utilizes the features of the primary tumor (T) and the
presence of regional lymph node involvement (N) or distant metastases (M)
to inform management and prognosis. T-status is substratified based on size
greater than 2 cm (T2) and whether the tumor extends beyond the pancreas
(T3) and involves the celiac axis or superior mesenteric artery (T4).
It is accurate, albeit controversial, to state that surgical resection is a
viable option for all T1-T3 nonmetastatic, non functioning, nonsyndromic
PNETs. Some experts, however, would dismiss this strategy as an
oversimplification and argue that observation is preferable to surgery for
select T1 lesions. This selection process, unfortunately, is not well defined
and is often surgeon dependent. Importantly, mere classification as an
asymptomatic T1 lesion is not, on its own, sufficient enough information to
safely defer surgical intervention.
AJCC staging characteristics are too limited to dictate management
decisions alone. For example, they do not account for tumor histology, and
they are not specific to PNETs; they also apply to pancreatic
adenocarcinomas, which herald an overall worse prognosis. Additionally,
they fail to consider tumor growth rate. For example, a 0.75 cm tumor that
doubles in size over a 3- to 6-month interval is concerning and warrants
resection. Furthermore, there is a well defined risk of lymph node metastasis
for tumors less than 2 cm (T1). Failure to identify regional lymph node status
withholds valuable prognostic information from both the patient and the
treating physician. Therefore, several reasons to resect a T1 lesion exist and
are further described.
Intermediate and high-grade T1 PNETs, based on the World Health
Organization (WHO) classification of neuroendocrine tumors, should be
strongly considered for resection. The WHO determines histologic grade by
the number of mitoses per high-powered field (HPF), and Ki-67 proliferation
index, which is a marker of tumor cell multiplication and is expressed during
chromosomal replication. Well-differentiated/low-grade (G1) tumors have
less than 2 mitoses/10 HPF, and a low (less than 3%) Ki-67 index. Grade 3,
or poorly differentiated tumors, should be resected irrespective of size
criteria, although there is some debate regarding the exact Ki-67 cutoff, in
favor of a higher threshold. Nevertheless, the literature has clearly shown
that tumors with Ki-67 less than 2% have a distinctly less aggressive initial
disease course when compared to lesions with Ki-67 between 2 and 20%,
and greater than 20%.3 Intermediate grade tumors may also have strong
consideration in favor of surgery, with greater emphasis given to Ki-67
proliferation index than to number of mitoses.
Grading may be difficult to assess in the preoperative setting. Ki-67
expression on FNA is frequently concordant with postoperative histologic
analysis, but agreement is less than 90%. Tumor heterogeneity is also a well-
recognized phenomenon that may lead to misclassification and
undertreatment. Other tumor characteristics, such as lack of somatostatin
receptor positivity in nonfunctioning tumors, may correlate with more
aggressive tumors and warrant consideration as well. When there is
diagnostic ambiguity about the lesion in question, surgical resection is often
the preferred alternative.
Perhaps the most convincing data favoring surgical resection for T1
lesions is that nodal metastases occur in PNETs despite their small size, and
nodal status directly correlates with patient outcome. Failure to address
nodal metastases contradicts the primary tenant of PNET management, which
is to resect all disease, thereby preventing distant metastasis and extending
patient longevity. When Hashim et al. applied an aggressive surgical
strategy–resection and lymphadenectomy were uniformly applied to 98% of
studied patients irrespective of tumor size– it was demonstrated that 12% of
patients with subcentimeter lesions and 13% of patients with tumors less than
1.5 cm had N1 disease. Patients with tumor diameter greater than 1.5 cm
were 4.7 times more likely to have regional lymph node metastases than
smaller tumors (95% CI, 1.81 to 12.35, P = 0.002). When one considers this
statistic, it is clear that size criteria cannot be considered alone in deciding
to resect or observe a lesion. A T1 lesion with imaging suggestive of
positive lymph node status is not appropriate for observation.
Numerous published reports have demonstrated that nodal metastases,
disease recurrence, and death from PNET disease, occur infrequently albeit
undeniably in T1 lesions, even those less than 1 cm. Although PNETs tend to
be indolent tumors, many experts agree that the data is not good enough to
reliably distinguish those small, aggressive tumors in order to support more
conservative stances that favor observation over surgery. This is particularly
true of well-conditioned patients with a reasonable life expectancy and the
absence of clear medical contraindications to surgery.
Particular attention should be given to tumor location when considering
surgery for a PNET. Pancreatic head lesions have been shown to be 2.8 times
more likely to have regional lymph node metastases (95% CI, 1.3 to 5.8, P =
0.007).3 In extrapolated data, the probability of N1 disease for tumors of the
pancreatic head was at least 1 in 5 patients, even with small tumors. Tumor
location may also impact the likelihood of duct obstruction, jaundice,
pancreatic insufficiency, or pain, all of which will influence the decision to
resect the lesion.
When considering a strategy of observation one must also critically
appraise published data on nonoperative management of small tumors. These
studies are not randomized and therefore have built in biases, particularly
selection bias. There may be a significant conversion rate to surgery after
imaging surveillance. For example, Sadot et al. converted 26 of 104 patients
selected for observation to the surgical arm after a median of 30 months.
Cited reasons for these conversions appear arbitrary rather than being based
on objective data–38% patient and 27% physician preference, making it
difficult to create a cohesive guideline that challenges surgery as the standard
of care. Most importantly, advocates of observation report on patients with
N0 disease, which can be presumed but not guaranteed prior to surgery.
Also, long-term follow-up on patients undergoing observation instead of
surgery is limited, especially in the context of the slow-growing but not
always benign evolution that typifies a PNET.
PATIENT CHARACTERISTICS
The majority of patients with sporadic functioning and nonfunctioning PNETs
should undergo surgical resection, unless patient characteristics result in
operative contraindications. Examples of contraindications include patients
with a high-risk profile, or patients for whom resection would unreasonably
interfere with their quality of life. Lifestyle factors include pre-diabetics or
patients with good oral hyperglycemic control, for whom blood glucose
control could become challenging. Advanced age has been shown to be an
independent factor associated with more aggressive disease course, thus it
should not be used in isolation to select out patients otherwise well-suited to
surgery.3 Patients who are unreliable and unlikely or unable to present for
sequential imaging may be better served with initial surgery. Some patients,
like the woman in the clinical scenario above, experience significant anxiety
after being given their diagnosis and prefer surgery over the fear of the
unknown.
SURGICAL MANAGEMENT
Once the choice in favor of surgery is established, one must decide upon the
extent of surgery. Options to address pancreatic parenchyma include a
limited resection (enucleation or parenchymal-sparing procedures, e.g.
central or middle pancreatectomy) versus formal resection
(pancreaticoduodenectomy or distal pancreatectomy). Lymph node resection
may be included, and if it is performed, it may be completed in a limited or
extended fashion. Insulinomas and gastrinomas may be enucleated if their
anatomy is favorable. Enucleation instead of formal resection for other
functioning PNETs is rarely recommended. This is, in part, due to the fact
that these other functional subtypes tend to present with more advanced
disease and behave in a more aggressive fashion, warranting regional
lymphadenectomy. In patients with functional tumors who are anticipated to
require long-term octreotide, cholecystectomy should be performed at time of
initial surgery.
Nonfunctional resectable lesions greater than 2 cm in size (T2 or T3)
typically warrant formal surgical resection. For lesions less than 2 cm and
limited to the pancreas (T1), limited resection can be an appropriate, less
invasive, alternative to traditional resection techniques. Tumors considered
appropriate for enucleation are typically less than 2 cm, superficial, away
from the pancreatic duct, and encapsulated. Review of the Surveillance
Epidemiology and End Results (SEER) database from 1998-2012 highlights
that, while surgeons are significantly more likely to consider enucleation for
smaller tumors, a minority of patients (57 patients in 981) undergo
enucleation.
Lymph node sampling with frozen section and possible node dissection may
be performed at the time of limited resection, but the utility of this is widely
debated. NCCN guidelines favor lymphadenectomy based on staging
criteria–lymphadenectomy is recommended for T2 lesions and suggested for
T1 lesions greater than or equal to 1 cm. However, Kaplan-Meier survival
statistics using SEER data show no difference in overall or disease-free
survival for T1 or T2 N0 versus NX patients, and extended
lymphadenectomy (greater than 10 lymph nodes) for all T groups and any N
status has not been associated with improved survival. Therefore, Conrad et
al. recommend that lymphadenectomy is best utilized to confirm
prognostication in patients with any size tumor and radiographic suspicion of
nodal involvement and argue that the risks associated with routine
aggressive, extended lymphadenectomy may not be justified.
INHERITED SYNDROMES
The most essential patient factor impacting surgical management is the
presence of an inherited syndrome. Genetic predisposition to PNETs occurs
in MEN-1, von Hippel-Lindau disease (VHL), von Recklinghausen
disease/neurofibromatosis type 1 (NF-1), and tuberous sclerosis (TS).
PNETs are relatively uncommon manifestations of NF-1 and tuberous
sclerosis, and resection of these lesions is recommended.
MEN-1 is characterized by hormone production, larger and multifocal
lesions, and metastatic disease. Malignant islet tumors are the major cause of
premature death in these patients. Surgery is rarely curative. Therefore, a
more conservative approach to surgery is preferred. Enucleation is often
reasonable when feasible because this strategy favors a curative outcome and
spares the majority of pancreas parenchyma. Formal pancreatic resection for
patients with MEN-1 is intended to minimize metastatic risk and maximize
survival. A Whipple may be indicated in a young patient with a large mass
(greater than 1 to 2 cm), a functional tumor refractory to medical
management, or rapid tumor growth. However, given the genetic basis for
tumor development in these patients, additional lesions are more likely to
develop in the unresected pancreatic tissue than in patients with sporadic
disease, and completion or total pancreatectomy is to be avoided in
syndromic patients.
VHL PNETs, as in MEN-1, are typically multiple. However, unlike in MEN-
1, they are usually asymptomatic and nonfunctioning. Metastatic potential is
unclear due to the limited number of patients with this disease, and primary
causes of morbidity tend to be renal or cerebral tumors, rather than those
arising in the pancreas. Surgical management for PNETs in VHL is heavily
dependent on size criteria. Tumors less than 1 cm are typically monitored
with serial imaging, tumors greater than 3 cm are resected, and intermediate
tumors are managed individually. Enucleation is generally favored, if
feasible.
CONCLUSION
In summary, most patients with resectable PNETs are best suited to surgical
resection, with consideration given to limited resection based on anatomic
and size criteria. Observation could be considered for nonfamilial T1 lesions
with grade 1 his tology, no evidence of lymph node involvement, no evidence
of hormone production, without interval growth, if surgery is of prohibitive
risk, or the patient prefers to forgo resection. However, data supporting
observation is limited. For the majority of patients with this diagnosis,
surgical management remains the standard of care.
The patient presented in this scenario will almost certainly undergo
resection. The notion of observing a solid pancreatic neoplasm is an
anathema to most patients even with detailed counseling, especially when
they have first-hand experience watching someone close to them die of
pancreatic cancer. Furthermore, at the age of 56, this patient will need
surveillance at least annually for the next 40 years. Given the nature of cross-
sectional imaging, there will undoubtedly be some change on imaging within
this time period (either real or created by technique and interpretation) that
elevates concern enough to proceed with resection.
From our perspective, the question is not whether this lesion should be
resected, but rather, when it should be resected and how extensive the
resection should be. Understanding the limitations of the available data, in an
otherwise healthy, nonsyndromic patient with a 1.5 cm PNET, our practice is
to offer a resection that achieves a microscopically negative margin and
provides a sampling of the regional lymph nodes, while preserving as much
pancreatic parenchyma as possible, using the most minimally invasive
technique feasible based on the tumor location and its proximity to the main
pancreatic ductal system.
Finally, we would like to reinforce that the NCCN guidelines recommend
resection for this patient. Supporting this recommendation are decades of
experience by multiple experts in the field of PNET management at many of
the leading institutions across the country. Deviation from these guidelines
should require compelling evidence, which in our estimation does not exist.
SALIENT POINTS
Functioning and/or symptomatic PNETs should be resected and
consideration should be given to whether cholecystectomy is also
appropriate.
Formal surgical resection is appropriate for T2 and T3 PNETs.
Sporadic T1 lesions should be strongly considered for surgical
resection, especially for young patients without contraindications and
poorly compliant patients. Greater weight should be given to resect
lesions in the pancreatic head or adjacent to ductal structures, those
with grade 2 or 3 histology, size greater than 1 cm, or exhibiting
diagnostic ambiguity or interval growth.
Regional nodal status should not be presumed to be negative based on
small tumor size.
PNETs associated with inherited genetic syndromes should be
considered separately from sporadic lesions, with management based
on existing guidelines.
SELECTED REFERENCES
1. Chiruvella A, Kooby DA. Surgical Management of Pancreatic
Neuroendocrine Tumors. Surg Oncol Clin N Am. 2016;25(2):401-21.
2. NCCN Clinical Practice Guidelines in Oncology (NCCN Guidelines®).
Neuroendocrine Tumors. (Version 2.2016).
https://www.nccn.org/professionals/physician_gls/pdf/neuroendocrine.
pdf. Accessed February 14, 2017.
3. Hashim YM, Trinkaus KM, Linehan DC, Strasberg SS, Fields RC, Cao
D, et al. Regional lymphadenectomy is indicated in the surgical
treatment of pancreatic neuroendocrine tumors (PNETs). Ann Surg.
2014; 259(2):197-203.
4. Sadot E, Reidy-Lagunes DL, Tang LH, Do RK, Gonen M, D’Angelica
MI, et al. Observation versus Resection for Small Asymptomatic
Pancreatic Neuroendocrine Tumors: A Matched Case-Control Study.
Ann Surg Oncol. 2016; 23(4):1361-70.
5. Conrad C, Kutlu OC, Dasari A, Chan JA, Vauthey JN, Adams DB, et al.
Prognostic Value of Lymph Node Status and Extent of Lymphadenectomy
in Pancreatic Neuroendocrine Tumors Confined To and Extending
Beyond the Pancreas. J Gastrointest Surg. 2016;20(12):1966-1974.
6. Alexakis N, Connor S, Ghaneh P, Lombard M, Smart HL, Evans J, et al.
Hereditary pancreatic endocrine tumours. Pancreatology. 2004;
4(5):417-33; discussion 434-5.

CASE SCENARIO
A 63-year-old man was incidentally discovered to have a mass in his
pancreas. His workup included a CT scan that showed a 1.5 hypervascular
mass in the head of the pancreas (Figure 1 ). The patient denied any
symptoms related to the pancreatic mass. Specifically, he denied abdominal
pain, weight loss, nausea, vomiting, or diarrhea. His past medical history is
remarkable only for hypertension and hyperlipidemia.
FIGURE 1: Uncinate process PanNet at contrast-enhanced CT.
Arterial phase at CT scan shows the hypervascular primary tumor
(yellow arrow) on the transverse (A) and coronal plane (B).
FIGURE 2: Workup for incidentally discovered PanNet =2cm
TABLE 1:
█ BACKGROUND
Surgical Management of Small Pancreatic
Neuroendocrine Tumors
The finding of a solid, well-circumscribed hypervascular lesion in the
pancreas on CT imaging is strongly suggestive of a pancreatic
neuroendocrine tumor (PanNet). PanNets are rare neoplasms that originate
from the endocrine cells of the pancreas and exhibit a wide spectrum of
behavior ranging from an indolent low-grade neoplasm with slow growth, to
highly aggressive neoplasms with a propensity for early metastases and poor
outcomes.
Over the last decade very small (≤2 cm) PanNets have been increasingly
diagnosed in asymptomatic patients due to the widespread use of high-quality
imaging techniques. Resection is the only curative treatment for localized
PanNets, and represents the traditional first-line therapy. However, the
increasing identification of small nonfunctional resectable neoplasms, often
in elderly patients, with an uncertain malignant potential has questioned the
role of resection, as many patients will not experience progression for many
years. In order to avoid overtreatment, operative recommendations must take
into account, and balance, the risk of malignant progression and the risks
associated with resection--including the development of perioperative
complications and the development of pancreatic exocrine and endocrine
insufficiency.
Several surgical series have demonstrated that incidentally discovered
nonfunctional-PanNets (NF-PanNets) that are <3cm in diameter are in the
majority represented by a low-grade pathology (84-95%) and no nodal
involvement (86-99%). A recent study by our group investigated the natural
history of small NF-PanNet and reported a 99% 5-year overall survival, and
none of the >100 patients in the nonoperative group experience metastatic
progression. In this case-matched evaluation, there was no difference in
survival between patients who were selected for radiographic surveillance
and size-matched controls who had undergone resection. The median follow-
up was almost four years and suggests that selected patients may be safely
monitored.
The patient presented above presents with a small lesion in the head of
the pancreas that is amenable to pancreaticoduodenectomy (PD) or possibly
enucleation (EN). Both of these procedures have a high rate of postoperative
morbidity (46-50%) and a measurable mortality rate (≤2%), even in
specialized centers. Based on the data presented above, treatment
recommendations must consider the competing risks of surveillance and
resection. If pancreaticoduodenectomy is recommended for 100 patients with
a similar presentation as above, 2 will die from complications of the
operation. If radiographic surveillance is recommended for 100 similar
patients, then we would expect all to be alive in four years. In addition, the
related long-term complications from pancreatic resection should be
considered in patients treated for indolent diseases. The rate of exocrine and
endocrine insufficiency for patients undergoing PD has been reported to be
approximately 55% and 5%, respectively.
Although some would argue that PD is not necessary, it should be
emphasized that parenchyma- sparing resections are not always feasible.
Enucleation is reserved for small neoplasms located superficially that are not
immediately adjacent to the main pancreatic duct. Enucleation has also been
reported to have significant rates of morbidity. In one study by Jilesen et al.,
the leak rate was 40%, and the rate of exocrine and endocrine insufficiency
for patients undergoing EN has been reported to be approximately 19% and
7%, respectively.
Before a surveillance strategy should be implemented, an accurate
diagnosis and thorough staging evaluation should be obtained. A proper
clinical evaluation should investigate for the presence of symptoms referable
to a hormone-related syndrome such as diarrhea, flushing, hypo- or
hyperglycemic events, and rarely a rash. When these symptoms exist, the
specific syndrome should be supported by an increase in the appropriate
serum hormonal marker and, when present, represents an indication for
resection. Staging evaluation should include anatomic (CT/MRI) assessments
and, in selected patients, functional techniques (111 In-pentreotide
scintigraphy/68 Gallium DATATOC-PET scan). Imaging and functional
studies may be complementary and should be used to determine the location
and size of the primary tumor and to assess for the presence of regional and
distant metastases. Usually, small NF-PanNets are incidentally discovered
when abdominal imaging is performed for unrelated indications. PanNets
appear typically on CT and MRI imaging as well-circumscribed
hypervascular lesions. Rarely, PanNets may present as a cystic lesion, have
heterogeneous enhancement, and very rarely may present as a hypovascular
mass with pancreatic duct dilation. These latter findings may resemble other
pancreatic cystic lesions, or pancreatic adenocarcinoma, and can make the
differential diagnosis challenging.
Functional studies exploit the expression of somatostatin receptors
(SSTR) on PanNets and include 111 In-pentreotide scintigraphy (111 In-
SPECT) and 68 Gallium DATATOC-PET scan (68 Ga-PET). These tests are
able to diagnose well-differentiated tumors that highly express SSTR and
may be more sensitive in the identification of metastatic disease. In
particular, 68 Ga-PET has been found to detect distant metastases with higher
sensitivity and specificity than conventional techniques, and when functional
studies are combined with anatomical studies, a change in management has
been reported in 33% of the cases. Given its high specificity, 68 Ga-PET
may be used in selected patients to confirm the diagnosis of well-
differentiated NF-PanNet. For both functional techniques, false positives are
rare and include intra-pancreatic spleen and pancreatic renal clear cell
metastases.
When the anatomic and functional imaging evaluation suggest a stage I
(American Joint Committee Cancer 7th ed.) NF-PanNet ≤2 cm with a typical
radiographic pattern (well-circumscribed hypervascular lesion), no further
evaluations are necessary and patients may be considered candidates for
radiographic surveillance. In the setting of atypical radiographic findings, or
when there is a lack of radiotracer uptake, an endoscopic ultrasound with
fine needle aspiration biopsy (EUS/FNAB) should be performed to confirm
the diagnosis of PanNet and to evaluate for an intermediate or higher grade
lesion. The concordance between the cytological and histological Ki67 range
between 83-89%, and it is almost completely consistent for tumors smaller
than 18 mm.
Small PanNet with a high-grade component (Ki-67>20%) are rare, and
surgical resection is recommended regardless of the tumor size due to the
more aggressive tumor behavior. PanNet with intermediate grade component
(Ki-67 2-20%) are characterized by uncertain behavior and have a higher
chance to develop distant metastases. These patients should generally
undergo surgical resection or, in selected cases (patient not fit for surgery),
for close radiographic surveillance (3-6 months).
SALIENT POINTS
Small NF-PanNets are neoplasms characterized by low-grade
pathology with an indolent behavior.
Pancreatic surgery is curative but has a high morbidity rate, a
measurable mortality rate, and long-term sequela such as exocrine and
endocrine pancreatic insufficiency.
An initial observation approach is reasonable and safe in patients with
presumed small (≤2cm), low-grade, incidentally discovered PanNet.
A thorough diagnostic and staging workup must be performed to
confirm diagnosis before a surveillance strategy can be recommended.
SELECTED REFERENCES
1. Sadot E, Reidy-Lagunes DL, Tang LH, Do RK, Gonen M, D’Angelica
MI, et al. Observation versus Resection for Small Asymptomatic
Pancreatic Neuroendocrine Tumors: A Matched Case-Control Study.
Ann Surg Oncol. 2016;23(4):1361–70.
2. Bettini R, Partelli S, Boninsegna L, Capelli P, Crippa S, Pederzoli P, et
al. Tumor size correlates with malignancy in nonfunctioning pancreatic
endocrine tumor. Surgery. 2011;150(1):75–82.
3. Jilesen APJ, Van Eijck CH, Busch OR, Van Gulik TM, Gouma DJ, Van
Dijkum EJ. Postoperative Outcomes of Enucleation and Standard
Resections in Patients with a Pancreatic Neuroendocrine Tumor. World
J Surg. 2016;40(3):715–28.
4. Sadowski SM, Neychev V, Millo C, Shih J, Nilubol N, Herscovitch P, et
al. Prospective study of 68Ga-DOTATATE Positron Emission
Tomography/Computed Tomography for Detecting Gastro-Entero-
Pancreatic Neuroendocrine Tumors and Unknown Primary Sites. J Clin
Oncol. 2016;34(6):588–96.
5. Díaz del Arco C, Díaz Pérez JÁ, Ortega Medina L, Sastre Valera J,
Fernández Aceñero MJ. Reliability of Ki-67 Determination in FNA
Samples for Grading Pancreatic Neuroendocrine Tumors. Endocr
Pathol. 2016;27(4):276–83.

CASE SCENARIO
A 68-year-old woman presented to an outside hospital with painless
jaundice, a fifteen pound weight loss, dark urine, and light stool. A CT scan
performed was read as “biliary obstruction due to a 3 cm mass in the
pancreatic head, abutting the porto-splenic confluence causing possible
distortion of the superior mesenteric vein” (Figure 1 ). The patient was taken
to an endoscopy suite by a gastroenterologist where an endoscopic
ultrasound and biopsy were performed, confirming pancreatic
adenocarcinoma;a biliary endoprosthesis was placed to relieve the patient’s
jaundice.
FIGURE 1: CT Scan image (L=Liver, GB=Gallbladder, T=Tumor,
IVC=Inferior vena cava, Ao= Aorta, B= Common bile duct with
endoprosthesis, *=Superior mesenteric vein, +=Superior mesenteric
artery.
█ BACKGROUND
Pancreatic ductal adenocarcinoma (PDA) is notorious for its challenging
biology and potentially perilous location among the vascular structures of the
retroperitoneum. Despite substantial ongoing research, only modest gains
have been made in the treatment and life expectancy following the diagnosis
of this disease. According to National Cancer Institute SEER database
cancer statistics from the year 2012, the 5 year survival after diagnosis of
pancreas cancer was 7.7%. New data from the Pancreatic Cancer Action
Network suggests the 5 year survival may now be up to 9%. Although PDA
is twelfth on the list of new cancer diagnoses annually (53,070 new cases in
2016), it is projected to become the second leading cause of cancer-related
deaths in the United States (41,780 deaths due to pancreas cancer in 2016).
Sadly, only one-half of patients with Stage I disease come to surgery in
academic centers, and fewer in community hospital settings (typically 30%).
Thus, it is clear that the identification, rapid referral, and employment of
aggressive treatment strategies need improvement. With the goal of
increasing the number of long term survivors following diagnosis with PDA,
we in the pancreas cancer treatment community have yet to universally agree
upon the best treatment algorithm for patients such as the one in the above
clinical scenario.
HIGH-VOLUME CENTERS
It has been incontrovertibly shown that pancreas cancer is best treated at
high-volume centers. We strongly support this and receive many referrals,
such as the one described in the above clinical scenario, to our institution.
Regardless of tumor extent or clinical stage, the challenges of treating PDA
patients are many, and all cases should be reviewed by expert,
multidisciplinary teams.
OPTIMAL IMAGING
Our approach to these patients is expeditious and begins (if the patient has
had a less than optimal CT or MR scan) with a requisite “pancreas protocol”
CT scan, which is specifically designed to visualize pancreas tumors, liver
parenchyma and relevant visceral vasculature. Using a multidetector CT
scanner, the optimal protocol obtains cross-sectional images at four different
time points which are then reconstructed in 3-dimensions as 3 mm slices.
First, the patient will orally take in 500 mL of water to distend the
duodenum. Oral contrast is not used. The initial images obtained are without
intravenous contrast. Next, images are captured at three different phases
following administration of a 100 mL intravenous iodinated contrast bolus.
When a bolus-tracker device is placed over the liver and contrast is
detected, the early arterial images are obtained. Approximately ten seconds
elapse and the late arterial images are obtained. Finally, approximately 35
seconds later, the portal-venous images are then obtained. As with many
images performed outside of high-volume centers, the timing of contrast in
the portal venous phase was sub-optimal in the case scenario above. We
proceeded to acquire CT images at our own facility following these
standard, recommended guidelines.
In this case scenario, we reviewed the patient’s new multidetector CT
images carefully in all planes and, in fact, found that the superior mesenteric
vein (SMV) and the portal vein (PV) appeared to be free and clear of tumor.
We routinely evaluate the venous vasculature in the axial, coronal, and
sagittal views, as the composite information can be quite reassuring
regarding the degree of difficulty in the dissection. Additionally, upon review
of our CT scan, we noted a previously unmentioned, replaced right hepatic
artery arising off of the SMA (Figure 2 ), which requires attention at the time
of surgery. This among other nuances, such as hepatic parenchyma
enhancement, are key features to treating pancreas cancer patients
safely,confirming our desire to see these patients in referral.
FIGURE 2: CT scan with RHA (L=Liver, GB=Gallbladder, r-
RHA=replaced right hepatic artery, SMA=Superior mesenteric
artery, IVC=Inferior vena cava, Ao= Aorta, B= Common bile duct
with endoprosthesis, D=Duodenum, P=Pancreas, S=Spleen).
DEFINING RESECTABLE, BORDERLINE &

LOCALLY-ADVANCED TUMORS
At initial diagnosis, it is of the utmost importance to categorize a
presumptive PDA by surgically relevant anatomic criteria found on CT scan
imaging. In the absence of spread to the liver, peritoneum, or distant
metastasis, it is common practice to first define these tumors as resectable,
borderline resectable, or locally advanced, which is based on the
relationship of the tumor to the surrounding mesenteric and retroperitoneal
vasculature (Table 1 ). We strongly recommend review of the CT scan
images by the attending surgeon with the radiologist. (As an aside, we also
routinely review the images with the patient and family for their educational
purposes.) We have found this can often lead to a modification of the original
outside radiology reading and categorization of the tumor. Differentiating
PDAs into these three above categories can drastically alter the treatment
approach for each case. This is why accurate categorization is critical, and
high-quality images are required to proceed.
TABLE 1: Criteria Defining Resectability Status 6

(Distant metastasis is considered unresectable; CA: Celiac
Artery; CHA: Common Hepatic Artery; SMA: Superior
Mesenteric Artery; SMV: Superior Mesenteric Vein; PV:
Portal Vein; IVC: Inferior Vena Cava).
Despite numerous efforts, national experts have yet to fully agree upon the
precise definitions of resectability. The definitions vary slightly between
academic centers and surgical societies with respect to tumor relationship
with the surrounding vasculature. This makes communicating clinical
decisions more difficult and makes standardizing research efforts more
challenging. It is our preference to follow the guidelines outlined by the
National Comprehensive Cancer Network (NCCN) in terms of definitions
and treatment protocols.
In brief review: two pancreas tumor locations within the gland are
broadly defined by the NCCN as head/uncinate process or body/tail (Table 1
). The arterial and venous systems are each independently evaluated in
relation to the tumor location. A resectable tumor located anywhere in the
pancreas is one that displays no arterial contact, with a clear fat plane around
the celiac axis (CA), the superior mesenteric artery (SMA), and the common
hepatic artery. Additionally, a resectable tumor does not contact greater than
180 degrees of the SMV or PV and causes no vein contour irregularity. An
unresectable tumor in the head/uncinate process contacts greater than 180
degrees of the SMA, CA, or contacts the first jejunal SMA branch. In the
body/tail, an unresectable tumor contacts greater than 180 degrees of the
SMA or CA or has evidence of contact with the CA and aortic involvement.
With regard to the venous system, a PDA anywhere in the gland is deemed
unresectable if the SMV/PV are believed to be unreconstructable either due
to tumor involvement or occlusion, or if there is tumor involvement with the
first draining venous branch of the jejunum. (Of note, in rare instances we
have resected PDAs that meet this last venous unresectability criterion, and
we have not done a reconstruction, leaving the patient’s own varicosities
intact, to allow venous drainage from the small and large bowel.)
Identifying and acting expeditiously once a resectable tumor has been
discovered, we believe, is crucial in the treatment of early stage PDA.
Surgery is the cornerstone of curative therapy, as no other treatment modality
has the capacity to cure this disease. Achieving a margin negative, termed an
R0, surgical resection is the best chance for a cure that we can provide to a
patient. Therefore, we believe the best initial approach is to remove these
tumors, provided we are able to do so safely. It has been shown that when an
R0 resection is achieved, overall survival is improved and cure, even
without adjuvant therapy, is possible, thus making this an important goal in
therapy. While attributing improved prognostic implications of an R0
resection entirely to the resection itself is controversial; nonetheless, there is
no evidence to the contrary.
It is our current practice to treat all patients with resectable PDA with
upfront or initial surgery and, once knowledge of the pathologic stage is
gained, treat with post-operative adjuvant therapies as indicated. We obtain
standard templated surgical consents (for pancreatectomy) and educate
patients well in advance of their actual operation. All patients are offered
participation in various prospective randomized clinical trials (RCTs) that
we may offer. Alternative treatment strategies (such as neoadjuvant
chemotherapy options) are also explained and offered to the patient in a
multidisciplinary fashion.
When a borderline resectable tumor is discovered, the treatment
algorithm is variable between clinicians. Individual surgeon interpretation of
the radiographic studies also affects how to approach these cases. As
previously mentioned, the terminology used in the definition of borderline
resectable disease differs between national experts and, furthermore, results
in variable approaches. The area of greatest variation involves tumors that
demonstrate venous involvement. The key factors to note regarding venous
involvement are whether there is:
Greater or less than 180 degrees of tumor contact or abutment with the
SMV-PV confluence;
Presence or absence of venous occlusion or thrombosis, or wall contour
irregularity amenable to resection or reconstruction.
Again, our stance is aligned with the NCCN guidelines (Table 1 ). We
believe many borderline cases, in particular those due to minor venous
abutment of the porto-splenic confluence, warrant and benefit from an
exploration. In our practice, these patients are taken to the operating room
without delay, as long as the venous structures do not appear deformed or
show evidence of contour irregularity. In doing so, direct assessment and
visualization of the anatomy can be performed at surgery. A consensus
statement was published on this point, authored by Bockhorn et al in 2014
from the International Study Group of Pancreatic Surgery (ISGPS). It was
recommended that if venous reconstruction were possible, a patient should
be offered an exploration and resection. Furthermore, whether or not venous
involvement should even be classified as borderline has been called into
question. Although resectability in the face of vascular involvement remains
controversial in terms of improvement in overall survival, there is a general
acceptance nationally of the efficacy and appropriateness of venous
resections.
In cases of locally advanced PDA and particularly in cases of arterial
involvement, we are in agreement that neoadjuvant therapy should precede
surgical treatment strategies. Unfortunately, as a surgical community, we have
not been able to complete a phase III randomized-controlled trial to evaluate
the efficacy of neoadjuvant therapy in the context of resectable or borderline
resectable PDA. One study, first authored by Katz et al in 2012, performed
with gemcitabine-based chemotherapy showed that neoadjuvant therapy
made little difference in tumor burden size, resectability status, or ability to
achieve an R0 resection. Another 2015 study from the group at the
Massachusetts General Hospital, first authored by Ferrone et al, reported on
the results of a FOLFIRINOX-based neoadjuvant regimen resulting in no
correlation between post-treatment imaging and resectability status.
OUR TECHNIQUE
In the case of the patient highlighted in this chapter with pancreas
protocol CT images showing only venous abutment, we went through the
steps of preparing the patient for possible venous resection (see V. Venous
resections below). Indeed, we were pleased when the dissection for this
patient went smoothly; no vein resection was necessary, and ultimately a
standard pylorus-preserving pancreaticoduodenectomy was performed. To
describe this operation, we begin through a vertical midline incision from the
xiphoid process to umbilicus. Upon entering the peritoneal cavity we
perform a thorough exploration, ‘run’ the small intestine and look for
metastatic disease to the liver, ligament of Treitz, root of the mesentery, and
peritoneum. We begin the extirpative phase of this operation by first
performing a “top down” cholecystectomy. We dissect out the extrahepatic
biliary tree from the hepatoduodenal ligament, elevate the common hepatic
duct from the portal vein, and divide it. From there, we identify and dissect
out the gastroduodenal artery (GDA). In this particular case, the replaced
right hepatic artery was identified and carefully preserved. We test clamp the
GDA to ensure the absence of celiac stenosis and the presence of good flow
in the proper hepatic artery, and then divide it. We then elevate the first part
of the duodenum from the pancreatic neck and divide it with a linear dividing
stapler, approximately 3 cm below the pylorus, preserving the pylorus
proximally. We then take down filmy adhesions between the posterior distal
stomach and the anterior aspect of the pancreas. We identify and divide the
downstream GDA caudally, inferior to the stomach, which allows further
mobilization of the stomach to the left side. Next, we “Kocherize” the
duodenum, elevate the head of the pancreas up out of the retroperitoneum and
off of the IVC, and continue to dissect the uncinate process from the
transverse mesocolon. We then dissect out the SMV inferior to the pancreatic
neck. When able, in the absence of vascular involvement, we elevate the
pancreatic neck ventrally from the SMV-PV confluence and place a Penrose
drain through the tunnel we have created. We use electrocautery to divide the
neck and carefully dissect the neck and head of the pancreas away from the
SMV, PV, and SMA. Again, in this case, care was taken to preserve the
replaced right hepatic artery (Figure 2 ).
We then turn our attention to the ligament of Treitz, divide the proximal
jejunum 20 cm distally with a linear dividing stapler, and separate it from its
mesentery. The duodenojejunal junction is mobilized behind the mesenteric
vessels, and we deliver the proximal jejunal stump under the root of the
mesentery to the patient’s right side. We then divide the remaining
attachments between the specimen and the retroperitoneum (the so called
mesopancreas) and remove the specimen.
The reconstruction phase begins with an end-to-side invagination
pancreaticojejunostomy (PJ) using a 3-0 silk outer layer and 3-0 vicryl
continuous inner layer, performed over a pediatric feeding tube which is later
removed during the hepaticojejunostomy (HJ). Downstream from the PJ we
perform an end-to-side HJ using a single layer of interrupted 5-0 PDS suture,
beginning with the posterior layer and followed by the anterior layer,
performed over a T-tube modified to be an I-tube, which is also subsequently
removed intraoperatively. (Hence, no transanastomotic stents are left
behind.) Finally, for our duodenojejunostomy (DJ), we deliver the jejunum
through a rent in the transverse mesocolon and perform the anastamosis in
two layers, approximately 30 cm downstream from the HJ, using 3-0 silk and
3-0 vicryl suture. The DJ is then pulled caudally, to rest below the
mesocolon. We conclude by placing two intraperitoneal drains and closing
the mesenteric defects.
VENOUS RESECTION
It has been shown that patients who undergo a partial vein resection with
pancreaticoduodenectomy, when compared side by side with patients who
undergo a pancreaticoduodenectomy without vein resection, have the same
oncologic outcome, provided an R0 resection is achieved. Additionally,
there was no difference in survival in patients undergoing resection who
were found to have vein involvement. We have retrospectively collected an
unpublished series of patients considered to have borderline or locally
advanced disease who underwent neoadjuvant therapy, a quarter of which
ultimately came to surgery for resection. We have found in this small series
that no survival benefit was found for patients who received neoadjuvant
chemotherapy when compared to published survival rates of borderline
cases that undergo surgery first without neoadjuvant chemotherapy. From this
and other studies we conclude that in cases of isolated vein involvement,
proceeding with surgery followed by adjuvant therapy as indicated is a
reasonable approach and prevents the opportunity for a surgical delay and
disease progression to unresectable/locally advanced status. It has also been
shown that the addition of a venous resection is safe in experience hands and
does not cause any additional morbidity or mortality.
Despite optimal pre-operative imaging, we have found the degree of
venous involvement and associated challenges in dissection difficult to
interpret preoperatively. While some studies show a correlation between
venous invasion on preoperative imaging and postoperative histologic
venous involvement, our experience has found the degree of tumor adherence
unpredictable. This is due to misinterpretation of peritumoral inflammation
and other factors. If there is any question of vein involvement, we make the
appropriate preoperative preparations for a more aggressive operation, such
as a vein resection. We believe a patient with a tumor with venous
involvement that is less than or equal to 180 degrees of the PV-SMV
confluence without vein contour irregularity is one that can be safely taken to
the operating room with a surgery first approach. Often the dissection of the
vasculature and removal of the tumor will be more facile than predicted.
If we feel a venous resection may be necessary, we begin by evaluating
the type of reconstruction that may be required and evaluate available
vascular conduits for harvest. On the day of surgery, we communicate with
the anesthesia team to leave one internal jugular vein exposed and available
to us for harvest. We will sterilely prep out the patient accordingly, with one
side of the neck for potential internal jugular vein harvest, as well as
bilateral groins for saphenous vein harvest. We ask and ensure that the
anesthesia team has the necessary vascular access and blood products
readily available should we encounter blood loss requiring transfusion.
Intraoperatively, when we discover and decide that we will be
performing a venous reconstruction, we adhere to a number of principals.
First, we will complete all other aspects of the pancreatectomy before
attempting separation of the tumor or any part of the pancreas from the PV
and/or SMV. It is wise to have as much of the resection accomplished before
working on the areas of potential venous attachment or encasement. When
appropriate for the size of an anticipated small defect, we will harvest a
segment of the left renal vein to use for a patch venoplasty. Using the left
renal vein prevents the morbidity of a second operative site, while venous
drainage of the left kidney flows through the adrenal and gonadal veins. If we
believe a larger segment of vein is required, our options are to use saphenous
vein or internal jugular vein. Saphenous vein is useful in the event that we
need a longer segment interposition graft for segments of the SMV; internal
jugular vein we have found to be an excellent size-match for portal vein
reconstructions. Again, before attempting separation of the PV or SMV, we
will complete all other aspects of the resection. We obtain proximal control
of the SMV and side branches, splenic vein, and distal control of the portal
vein.
Four types of venous resections listed below and outlined in Figure 3 are
agreed upon by the ISGPS and range from simple closure to interposed
conduits. These are:
Type 1: Partial venous excision with direct closure (venorraphy) by
suture closure
Type 2: Partial venous excision using a patch venoplasty closure
Type 3: Segmental venous resection with primary veno-venous
anastomosis
Type 4: Segmental venous resection with interposed venous conduit and
at least two anastomoses
FIGURE 3: Four types of venous reconstruction PV: Portal Vein;

SV: Splenic Vein; SMV: Superior Mesenteric Vein; IJV: Internal
Jugular Vein (Adapted with permission from Elsevier in Maley WR &
Yeo CJ (2016). Vascular Reconstruction During the Whipple
Procedure. Cameron JL & Cameron AM (Eds.), Current Surgical
Therapy, 12th ed. (pp. 549-553) Philadelphia.
Optimally, and with the help of preoperative imaging, the most commonly
encountered situation is the focal adherence of the tumor to the side wall of
the SMV or PV and a simply venorraphy can be performed.
POST-OPERATIVE CARE
Approximately 80% of patients who undergo a pancreas resection at our
institution adhere to our standard clinical pathway protocol for postoperative
care. This pathway both intensively monitors patients early in the
perioperative period and ensures a timely discharge, generally on post-
operative day six or seven following pancreaticoduodenectomy and day 5
following distal pancreatectomy. Initially, these patients spend one night in
the ICU with clinical monitoring devices inclusive of a central venous
catheter, an arterial line, a Foley catheter and a nasogastric tube (NGT). On
post-operative day one, if the patient is clinically well, we begin de-
escalation of monitoring devices with removal of the NGT and arterial line.
We will transfer the patient to the general surgery floor and ensure that the
patient is out of bed and ambulating. We remove the Foley catheter and begin
early diuresis on post-operative day two and monitor the patient’s intake and
output as well as daily weight. One drain is typically removed on post-
operative day three. By post-operative day four, the patient is started on a
regular diet, all medications are taken orally, and if the remaining Jackson-
Pratt drain output is non-sinister, it is removed.
For the patient in the case scenario, her post-operative course followed
the standard pathway without deviation. Her pathology showed a 2.5 cm
moderately differentiated adenocarcinoma of the pancreatic head with all
surgical margins negative for tumor and zero out of fifteen lymph nodes with
evidence of metastatic tumor. As for adjuvant therapy, our preference is for
our patients to establish a relationship with the medical oncologists while
inpatient and schedule their port placement for four to five weeks after
discharge. In this case, the patient was enrolled in the APACT trial, a phase 3
randomized, multicenter trial evaluating the use of adjuvant nab-paclitaxel
plus gemcitabine versus gemcitabine alone. Generally we advise adjuvant
therapy to commence six weeks from the date of the pancreatectomy.
CLINICAL TRIALS
We have a number of investigator-initiated trials, which we explain and
enroll patients in routinely. In anticipation of a pancreaticoduodenectomy, we
will enroll appropriate patients in a trial entitled the WARP trial (Whipple
Accelerated Recovery Pathway, NCT #02517268), which will aid in
defining an optimal, streamlined post-operative clinical care pathway.
Patients on the experimental WARP trial arm have increased daily physical
activity, slower advancement of their diet, and discharge on post-operative
day five, with close telephone follow up and intense outpatient follow-up
through telehealth appointments. We also will enroll patients in the WASH
trial (abdominal lavage of distilled WAter or Saline at High volumes, NCT#
02757859) which evaluates intraoperative tumor cell shedding and survival,
using serial lavage and dilutions. As a result of our HYSLAR trial (the use of
a restrictive fluid regimen with 3% Hypertonic Saline versus Lactated
Ringers, NCT# 01428050), the patient discussed herein was intraoperatively
and postoperatively maintained on 3% hypertonic saline and did quite well
without any perioperative complications.
SUMMARY
Venous abutment does not preclude patients from a surgery- first approach to
treating PDA. Optimal CT imaging and interpretation in conjunction with the
attending surgeon is imperative for proper resectability status categorization
of PDA. The term “Borderline Resectable” and the treatment implications
from this categorization are not nationally agreed upon. A randomized
controlled trial is needed to elucidate the best initial approach to PDA
patients with resectable and borderline resectable disease. Venous resections
for PDA are well tolerated and rarely necessary; if anticipated, appropriate
preparations should take place.
SALIENT POINTS
All patients with PDA should be referred to high-volume centers and
treated by experienced individuals in a multidisciplinary fashion.
Achieving an R0 resection has been shown to improve overall
survival.
Avoid over-classifying tumors as locally advanced, unresectable or
borderline that on suboptimal imaging appear to have isolated venous
involvement, as many of these patients could benefit from a resection.
Carefully review all available radiographic studies yourself; avoid
misinterpreting sub-standard studies and prepare for all pre, peri, and
post-operative aspects of patient care.
All patients with PDA should be enrolled in clinical trials as
appropriate.
SELECTED REFERENCES
1. Bockhorn M, Uzunoglu FG, Adham M, Imrie C, Milicevic M, Sandberg
AA, et al. Borderline resectable pancreatic cancer: a consensus
statement by the International Study Group of Pancreatic Surgery
(ISGPS). Surgery. 2014; 155(6): 977-88.
2. Kelly KJ, Winslow E, Kooby D, Lad NL, Parikh AA, Scroggins CR, et
al. Vein involvement during pancreaticoduodenectomy: Is there a need
for redefinition of a “Borderline Resectable Disease?” J Gastrointest
Surg. 2013;17(7):1209-17.
3. Katz MH, Fleming JB, Bhosale P, Varadachary G, Lee JE, Wolff R, et
al. Response of Borderline Resectable Pancreatic Cancer to
Neoadjuvant Therapy Is Not Reflected by Radiographic Indicators.
Cancer. 2012; 118(23):5749-56.
4. Murakami Y, Uemura K, Sudo T, Hashimoto Y, Nakashima A, Kondo N,
et al. Benefit of Portal or Superior Mesenteric Vein Resection with
Adjuvant Chemotherapy for Patients with Pancreatic Head Carcinoma.
Journal of Surgical Oncology. 2013;107(4):414-21.
5. Maley WR & Yeo CJ. (2016). Vascular Reconstruction During the
Whipple Procedure. Cameron JL & Cameron AM (Eds.), Current
Surgical Therapy, 12th ed. (pp. 549-553). Philadelphia: Elsevier.
6. National Comprehensive Cancer Network, NCCN Guidelines, version
2.2016, Pancreatic Adenocarcinoma.

CASE SCENARIO
An 80-year-old man with diabetes mellitus, hypertension, obstructive sleep
apnea, and an Eastern Cooperative Oncology Group performance status of 1
presented with a biopsy-confirmed cancer of the pancreatic head in the
setting of abdominal pain and weight loss. A computed tomography scan
demonstrated a 2.4 cm hypoechoic mass in the head of the pancreas that had a
circumferential interface with the porto-splenic venous confluence measuring
180° (Figure 1 ) and his serum CA 19-9 level was 1400 U/L. He was treated
with 5 cycles of gemcitabine and nab-paclitaxel followed by chemoradiation
with capecitabine, during which time his comorbidities were optimized and
he was enrolled in an exercise program. Following treatment, his CA 19-9
decreased to 175 U/L, and imaging revealed stable disease. He therefore
underwent pancreatoduodenectomy with segmental venous resection.
Pathology demonstrated negative lymph nodes and a negative margin of
resection. Recovery was uneventful.
FIGURE 1: Case example of a patient with a technically resectable
adenocarcinoma of the head of the pancreas (thick arrow) and a
circumferential interface between the tumor and the porto-splenic
confluence (thin arrow) measuring approximately 180°
█ BACKGROUND
Pancreatic ductal adenocarcinoma (PDAC) is one of the leading causes of
cancer-related mortality in the United States. Only about 7% of patients
newly diagnosed with PDAC live 5 years or longer. Even patients who
present with a radiographically localized cancer and who undergo de novo
resection of their primary tumor and regional lymph nodes are likely to
develop recurrent disease after this “potentially curative” treatment. Eighty
percent of them will do so within 2 years following surgery, and their
recurrent cancer will invariably lead to their death. Although overall
survival following pancreatectomy for PDAC is definitively prolonged by
the postoperative administration of systemic chemotherapy, that improvement
is incremental (on the order of 3 months!), and as many as 50% of patients
who undergo surgery never receive intended chemotherapy thereafter.
Major advances in the care of patients with localized PDAC have come
with time, but they have come slowly. Indeed, a series of patients who were
treated with pancreatectomy de novo over four decades at a high-volume
pancreatic treatment center justifiably considered to be one of the world’s
finest showed important but only relatively marginal improvements in
survival over recent years. Today, a highly functional patient with localized
PDAC who undergoes pancreatectomy performed by an expert surgeon and
who receives postoperative chemotherapy as prescribed can expect to live
only slightly longer than 2 years. Additionally, although about 20% of
patients who undergo resection de novo are effectively cured of their cancer,
an equivalent or even higher percentage will die within a year of surgery.
Clearly, the strategy of treating patients with PDAC with surgery upon
diagnosis—though it remains the standard of care—is unacceptable.
The delivery of nonoperative therapies prior to surgical resection, instead
of following it, represents an alternative treatment strategy. Although high-
level studies have not demonstrated survival benefit to this approach relative
to the standard of care, significant empirical data and a sound theoretical
basis exist to support it. Purported benefits of preoperative therapy include
facilitation of a margin-negative resection, the early treatment of
micrometastatic disease that is present in all patients, the selection of a
physiologically robust population of patients with “biologically favorable”
disease, and the assurance that all patients who undergo surgery receive
combined modality therapy. Long discounted by surgeons concerned about
losing the window for a “potentially curative” operation, this strategy is
becoming increasingly accepted nationwide, and it was recently recognized
in guidelines published by the American Society for Clinical Oncology as
entirely appropriate for at least some patients with localized PDAC.
Here we argue that the administration of preoperative therapy is rational
for many—if not most—patients with localized PDAC because, as we will
show, the primary use of surgery for PDAC is contrary to everything we have
learned about the biology of this cancer and the physiology of the patients
who are afflicted with it.
Preoperative Therapy Makes Good Surgery Better
The safe conduct of a margin-negative resection represents the primary
obligation of the pancreatic surgeon. Unfortunately, both the anatomy and
biologic behavior of PDAC tumors are such that positive surgical margins
and locoregional cancer recurrence represent significant problems for even
the most technically gifted surgeon.
Because most pancreatic cancers arise in the head or uncinate process
within millimeters of the superior mesenteric artery, and because cancer
cells infiltrate into the perineural and lymphatic tissues of the
retroperitoneum, proper conduct of pancreatoduodenectomy requires
skeletonization of the right lateral aspect of the superior mesenteric artery
from the level of the first jejunal branch of the superior mesenteric vein to the
takeoff of the artery from the aorta. Even when this dissection technique is
used at pancreatoduodenectomy, however, microscopic disease is often left
behind in the local tumor bed (Figure 2 ). Although the likelihood of positive
margins at surgery can be minimized by the application of meticulous
surgical technique in patients with well-selected tumor anatomy, it cannot be
entirely eliminated. Indeed, series from large, high-volume pancreatic
treatment centers report microscopically positive resection margin (R1) rates
following pancreatoduodenectomy as high as 50%. Moreover, recent studies
using rigorous histopathologic protocols have found cancer cells in at least
one surgical margin in as many as 90% of all pancreatoduodenectomy
specimens—an observation congruent with the reality that up to 80% of
patients who die from PDAC do so with locally recurrent cancer.
FIGURE 1: Superior mesenteric artery margin distance estimated
radiographically and measured histopathologically in two patients.
Red line depicts path of surgical resection. Small arrow highlights
tissue between SMA and tumor; large arrow demonstrates cancer
cells in proximity of inked margin. A,B: A patient who underwent
pancreatoduodenectomy with tangential vein resection and who had
tumor cells within 1 mm of the inked retroperitoneal margin. C,D: A
patient who underwent pancreatoduodenectomy with a positive
retroperitoneal margin. From Katz et al. Effect of neoadjuvant
chemoradiation and surgical technique on recurrence of localized
pancreatic cancer. J Gastrointest Surg. 2012;16(1):68-78-79
Preoperative therapy appears to improve the likelihood of a margin-

negative resection dramatically. Retrospective studies have demonstrated
that the delivery of radiation therapy to the primary tumor and regional lymph
nodes prior to pancreatectomy concurrent with radiosensitizing doses of
chemotherapy is associated with wider tumor clearance (measured
histopathologically by the microscopic distance between cancer cells and the
inked margin) and improved locoregional control relative to
pancreatoduodenectomy alone. In fact, the rate of locoregional recurrence
may be reduced by as much as 50% following preoperative chemoradiation
and pancreatoduodenectomy, a result remarkably similar to the degree of
locoregional control conferred by preoperative radiotherapy for both
esophageal and rectal cancer.
Despite this observation, the utilization of preoperative radiation therapy
remains low nationwide, possibly due to concerns that it may cause
peripancreatic fibrosis that results in technical difficulties during surgery, or
that it may prolong wound healing and lead to postoperative adverse events.
However, the preoperative setting is an ideal one within which to radiate the
abdomen, since it allows for treatment within a well-oxygenated
environment, in which the efficacy of radiation is enhanced and toxicities to
adjacent normal tissue are minimized. We and others have clearly shown that
radiotherapy may be delivered safely in the preoperative period, and that it
actually reduces the rate of postoperative pancreatic fistula—the most
serious complication following pancreatectomy.
PREOPERATIVE THERAPY TREATS A

SYSTEMIC CANCER
The provision of individualized, stage-specific treatments to patients with
cancer represents a fundamental principle of modern oncology practice.
Unfortunately, clinical staging algorithms for localized PDAC—which focus
entirely on the radiographic diameter of the primary cancer or its regional
anatomic extent—are almost useless in their ability to prospectively
discriminate between patients who will and will not be cured by
pancreatectomy. This is because postoperative cancer recurrence and
survival are dictated not only by the anatomic relationships of the primary
tumor, but also by the extent of coexisting metastatic disease. Also, current
reality is that imaging technology is not sufficient to detect or quantify the
micrometastases that exist in most, if not all, patients with localized PDAC.
Indeed, although modern scanners and imaging protocols allow the
preoperative evaluation of critical tumor-vessel relationships with great
precision, grossly macroscopic metastases to the liver or peritoneal surfaces
are missed in as many as 20% of patients who undergo exploration. Even
more alarming, the vast majority of patients who undergo a “potentially
curative” resection of an ostensibly localized cancer, even those in whom
lymph nodes are negative and in whom resection was microscopically
complete, develop distant metastases within 2 years—clear evidence for the
growth of clinically occult metastases that existed at the time of surgery.
These clinical observations have been corroborated in the laboratory by
data that clearly suggests that pancreatic cancer metastases develop earlier
than we once believed. In one particularly jarring study, the pancreata of
mice that were genetically engineered to develop pancreatic cancer were
fluorescently labeled, and cells were tracked over time. Investigators
identified labeled cells in the bloodstream and the liver prior to the
development of macroscopic tumors in the pancreas. In another study that
used computational methods and clinical data to model the likelihood of
coexisting micrometastases based on the size of the primary tumor,
investigators calculated that the probability that metastasis-enabled cells
exist in patients with a 4 cm cancer is 100%, and it is not much lower than
that except in patients with the smallest of primaries. We surgeons are clearly
well behind the eight ball.
To the extent that micro- (if not macro-) metastatic disease exists in most
if not all patients who present with PDAC, resection delays the initiation of
requisite systemic chemotherapy by 2 months or longer when applied as
primary therapy. Within that time, this untreated cancer is allowed to grow
and progress—a process almost certainly accelerated by the growth factors
and inflammatory mediators spilled into the circulation as a result of surgery.
Reduction of the burden and the growth rate of clinically occult metastases
with chemotherapy, in contrast, allows for the possibility of earlier systemic
disease control and the more appropriate use of pancreatectomy in patients
with a locally predominant cancer phenotype.
In practice, the serum CA 19-9 level can be effectively used to estimate
the metastatic burden of patients with otherwise clinically localized PDAC,
to help assess the response to preoperative therapy, and to determine the
subsequent role of surgery. For example, we would consider primary
resection for a patient with a tiny, resectable cancer and a serum CA 19-9 of
1000 U/ml as inappropriate as it would be in a patient with a locally
advanced cancer and synchronous macroscopic liver metastases. On the
other hand, normalization of that patient’s serum CA 19-9 level following
treatment with systemic chemotherapy would suggest therapeutic response,
the ability to maintain disease control, and a patient who might benefit from a
potentially morbid operation.
PREOPERATIVE THERAPY ALLOWS TIME

FOR PHYSIOLOGIC O PTIMIZATION
Patients who present with PDAC typically do so in the context of physiologic
and functional depression due to multiple causes including advanced age,
biliary and/or enteric obstruction, malnutrition, cancer-related cachexia, and
preexisting comorbidities. Performance of major surgery in the setting of
adverse clinical parameters such as age, serum albumin, a history of weight
loss, advanced comorbidities obtained on medical history, and others is
associated with a high rate of postoperative morbidity and mortality, and a
low rate of long-term survival. Using data from over 8,000 patients reported
in the National Surgical Quality Improvement Program database, for
example, we found that about one-third of patients who underwent
pancreatectomy between 2005 and 2010 could be considered to have
“borderline resectable pancreatic cancer type C” physiology—a
classification we have used to describe patients medically unfit for surgical
resection—and those patients suffered 50% higher mortality than patients
who did not meet these criteria. The use of frailty scores and anthropometric
indices may allow detection of even subtle physiologic derangements with
greater precision.
In many patients, such deficits are at least partially reversible. The period
within which preoperative therapy is administered represents an ideal time
to assess and to address physiologic concerns and to optimize nutritional and
physiological indices with targeted interventions. This occurs most
efficiently through a structured multidisciplinary program that emphasizes the
judicious use of internal medicine expertise, nutritional counseling, and
“prehabilitation.” In our center, patients who are scheduled to undergo
pancreatectomy for PDAC are treated with preoperative therapy concurrent
with a structured program of exercise and nutrition modeled after American
College of Sports Medicine guidelines for exercise in cancer survivors. This
process effectively enriches the population who undergoes surgery with
patients who have both a robust physiologic profile and a biologically
favorable cancer—the very patients most likely to see the benefits of
pancreatectomy.
PREOPERATIVE THERAPY ENSURES

REQUISITE MULTIMODALITY TREATMENT
Finally, although surgery is routinely touted as the only modality with the
potential to cure a patient with PDAC, it is rarely curative on its own.
Indeed, in the seminal trial that showed that postoperative systemic
chemotherapy improves survival over surgery alone, patients randomized to
observation following pancreatectomy had a median disease-free survival of
only 7 months! Further, the patients enrolled in that trial did not represent the
“typical” patient with PDAC—they were “best of the best”: they were, as a
group, relatively young, they had a low postoperative serum CA 19-9, and
they had recovered rapidly enough to be enrolled in the trial within 3 weeks
of surgery. The typical patient undergoing pancreatectomy alone would be
expected to fare even worse.
Unfortunately, many patients who undergo pancreatectomy nationwide in
anticipation of subsequent systemic chemotherapy do undergo surgery alone;
they never receive the systemic chemotherapy that has been clearly shown to
improve their therapeutic outcome. In a review of NSQIP and the National
Cancer Database data, patients in the United States who underwent
pancreatectomy between 2006 and 2009 received any therapy following
pancreatectomy only about 57% of the time. Failure to regain functional
status following surgery (a function of both the magnitude of the operation
and the physiologic depression of patients who often undergo it) and the
early development of metastatic disease are the most common factors that
prevent the timely initiation of therapy in the postoperative setting; dose
reductions and the inability to complete all intended cycles of therapy are
common even among those who initiate it. To the extent that the likelihood of
an individual patient’s rapid functional recovery and the biologic behavior of
their cancer can be predicted, the administration of preoperative therapy to
those patients at highest risk for omission of postoperative therapy can
effectively guarantee the delivery of multimodality therapy to essentially all
patients with localized PDAC.
OUTCOMES OF PREOPERATIVE
THERAPY
We recently reported our experience with the administration of preoperative
therapy at the University of Texas MD Anderson Cancer Center: a total of
622 patients who received chemotherapy and/or chemoradiation prior to
pancreatectomy over a 25-year period. Despite an increase in the relative
percentage of patients who were operated for “borderline resectable” or
locally advanced tumors as well as an increase in the use of vascular
resection and reconstruction concurrent with pancreatectomy over time, we
observed a steady increase in survival. The median overall survival of
patients treated between 2010 and 2014, the most recent time period we
examined, was 43 months, and no patient operated on within this period died
within 90 days of surgery. Although it is easy to overgeneralize these single-
institution results, they are clearly remarkable when compared to the median
survival duration of 28 months among patients treated with pancreatectomy
and the most modern postoperative therapy regimen and a nationwide
perioperative mortality rate of 7.4%.
A PLANNED REBUTTAL
Critics of preoperative therapy for PDAC will object to the arguments for
preoperative we have presented. They will lament the possibility that the
window for surgery—and therefore the potential for cure—will close during
the administration of therapies that have never been shown to be curative on
their own. Or, they will voice concerns about the adverse events associated
with preoperative chemotherapy or radiotherapy that may prohibit surgery. In
reality, however, isolated locoregional progression that precludes subsequent
resection is exceedingly rare, and treatment-related complications sufficient
to prevent surgery are also distinctly uncommon in the absence of systemic
disease progression, when care is delivered at experienced centers.
Critics will also note the total absence of high-level data that
unequivocally demonstrates the superiority of this alternative approach over
standard care. They will reference two randomized controlled trials (RCT)
from Europe, both of which randomized patients with resectable PDAC to
either surgery or preoperative chemoradiation, and neither of which detected
a difference in median overall survival between the treatment arms.
However, neither of these studies met their primary endpoint because they
failed to accrue patients: of a combined total of 318 patients targeted for
accrual, only 111 patients were enrolled. A larger European study that
randomizes patients to surgery or preoperative chemotherapy may be more
successful, but irrespective of its findings, the tide has already turned. There
is not a major academic treatment center in the United States that does not
have an active preoperative therapy trial open for accrual. ASCO guidelines
support the use of preoperative therapy at least for some patients. And, the
only two trials for localized pancreatic cancer offered through the National
Clinical Trials Network have been designed to evaluate preoperative
strategies, and neither of these randomized studies offers a surgery-first
control arm.
CONCLUSIONS
Significant advancements in the care of patients with localized PDAC
continue to lag behind the strides made elsewhere in oncology. It is clear
from decades of experience that the current dictum of surgery followed by
adjuvant therapy leads to unacceptable long-term results even in the best of
patients. In contrast, a sound theoretical argument can be made to justify the
delivery of nonoperative therapies in the preoperative setting, and increasing
empirical data support it. As systemic regimens improve in the near future,
decisions regarding treatment sequencing will only become more important.
Ultimately, the care of these patients is optimized when a community of
medical, radiation, and surgical oncologists alongside dedicated
radiologists, pathologists, gastroenterologists, and other team members are
engaged in their multidisciplinary management. In this chapter, we have
shown that the administration of preoperative therapy is appropriate for at
least some patients with PDAC—irrespective of the technical resectability of
the primary tumor. Therefore, the debate over this issue, we believe, should
now evolve to more mature questions regarding which regimens to use, the
optimal duration of therapy, the best sequencing of components, identifying
measures of response, personalizing therapy, introducing newer agents, and
more.
SALIENT POINTS
Potential advantages of a therapeutic approach for patients with
localized PDAC in which chemotherapy, radiotherapy, or novel agents
are administered prior to pancreatectomy include facilitation of a
margin-negative resection and improved local control, early systemic
treatment for a presumably systemic disease, optimization of
nutritional, physiologic and comorbid impairments prior to major
surgery, and ensuring delivery of all nonoperative therapies.
Although data from randomized controlled trials does not exist,
retrospective data demonstrates favorable long-term outcomes which
have improved over time
Complications from treatment, declines in performance status, or
locoregional progression occurring during neoadjuvant therapy that
prevent subsequent resection are rare.
SELECTED REFERENCES
1. Khorana AA, Mangu PB, Berlin J, Engebretson A, Hong TS, Maitra A,
et al. Potentially Curable Pancreatic Cancer: American Society of
Clinical Oncology Clinical Practice Guideline. J Am Soc Clin Oncol.
2016;34(21):2541-56.
2. Cloyd JM, Crane CH, Koay EJ, Das P, Krishnan S, Prakash L, et al.
Impact of hypofractionated and standard fractionated chemoradiation
before pancreatoduodenectomy for pancreatic ductal adenocarcinoma.
Cancer. 2016;122(17):2671-9.
3. Denbo JW, Bruno ML, Cloyd JM, Prakash L, Lee JE, Kim M, et al.
Preoperative Chemoradiation for Pancreatic Adenocarcinoma Does Not
Increase 90-day Postoperative Morbidity or Mortality. J Gastrointest
Surg. 2016; 20(12):1975-1985.
4. Rhim AD, Mirek ET, Aiello NM, Maitra A, Bailey JM, McAllister F, et
al. EMT and dissemination precede pancreatic tumor formation. Cell.
2012;148(1-2):349-61.
5. Cloyd JM, Katz MH, Prakash L, Varadhachary GR, Wolff RA, et al.
Preoperative Therapy and Pancreatoduodenectomy for Pancreatic
Ductal Adenocarcinoma: A 25-Year Single-Institution Experience. J
Gastrointest Surg. 2017; 21(1):164-174.

CASE SCENARIO
A 65-year-old woman was found to have a borderline resectable pancreatic
ductal adenocarcinoma (PDAC) with >180° involvement of the superior
mesenteric vein (SMV) (Figure 1 ). She received 4 cycles of FOLFIRINOX
followed by 30gy of external beam chemoradiation. After completing
preoperative therapy, the CA 19-9 had decreased from 670 to 29 U/L and
repeat imaging demonstrated stable disease. At the time of
pancreatoduodenectomy (PD), the mass was encasing the portal vein (PV)
and SMV at the portosplenic confluence. After an en-bloc resection of the
tumor, together with the portosplenic confluence, the PV-SMV was repaired
primarily in end-to-end fashion. In addition, given that the inferior mesenteric
vein (IMV) entered the SMV rather than the splenic vein (SV), and therefore
could not decompress the divided SV, a splenorenal shunt was constructed.
FIGURE 1: Preoperative computed tomography of borderline
resectable pancreatic ductal adenocarcinoma with >180°
involvement of the superior mesenteric vein (SMV). In order to
perform a meticulous dissection of the superior mesenteric artery
(SMA), given its proximity to the tumor (T), the splenic vein (SV) is
divided. The SMV-PV is reconstructed primarily and a splenorenal
shunt is created to prevent sinistral hypertension.
█ BACKGROUND
PDAC is one of the leading causes of cancer-related mortality in the United
States, largely due to the fact that the majority of patients have metastatic
disease at the time of presentation. Survival rates are better in patients with
localized disease amenable to surgical resection, especially when combined
with multimodality therapy. Historically, vascular involvement, whether
diagnosed radiographically prior to surgery or at the time of laparotomy, was
considered a contraindication to surgical resection. Therefore, patients with
localized cancers and major vessel involvement were treated solely with
chemotherapy and radiation. Since surgical resection is considered
necessary, though not necessarily sufficient, for curative intent, attempts to
extend the criteria of resectability to patients with vascular involvement have
long been of interest.
The first large series of vascular resection at the time of pancreatectomy
was described by Fortner as he introduced the concept of “regional
pancreatectomy” in the 1970s. Although Fortner’s original description
demonstrated the technical feasibility of venous and arterial resections with
pancreatectomy, the procedure was not associated with improved
oncological outcomes and therefore was largely abandoned. Nevertheless,
over the past several decades, developments in cross sectional imaging,
preoperative staging, vascular techniques, and multimodality therapy have
led investigators to again consider the merits of vascular resection at the time
of pancreatectomy. In fact, pancreatectomy with venous resection comprises
a substantial proportion of operations for PDAC at high volume centers and
large series suggest no worse survival in patients that are able to undergo a
margin negative resection.
PATIENT SELECTION
Since microscopic and macroscopic resections are associated with both
locoregional recurrence and reduced overall survival, the likelihood of
obtaining a margin negative resection must be critically assessed when
evaluating whether to proceed with surgical resection. As the ability to
safely perform complex vascular resections at the time of pancreatectomy has
evolved, a new radiographic classification system was needed to more
accurately characterize the anatomic relationship of the tumor to its nearby
vascular structures. Broadly defined as potentially resectable, borderline
resectable (BR) and locally advanced (LA) (Table), these designations
provide important prognostic information regarding the ability to achieve a
margin negative resection, the likely need for vascular resection, and the
selection of preoperative therapies prior to surgery.
TABLE 1: Criteria Defining Resectability Status (Data
derived from the National Comprehensive Cancer Network.
NCCN Guidelines, version 2.2016, Pancreatic
Adenocarcinoma).
In addition to anatomic staging, patients must be appropriately selected in

terms of their potential for harboring occult metastatic disease as well as
their physiologic capacity to safely undergo major surgery. Patients with
significantly elevated CA 19-9, indeterminate liver or lung lesions, or subtle
signs of peritoneal disease should be considered for preoperative
chemotherapy and/or chemoradiation therapy as an opportunity to select for
patients with biologically favorable disease. Those who demonstrate
normalization of their CA 19-9, as well as stable radiographic findings,
could then be considered for surgical resection. Similarly, those patients
whose comorbidities or performance status make PD prohibitive should also
be considered for preoperative therapy, simultaneously focusing on
improving their suitability for surgery through medical optimization and
prehabilitation.
SURGICAL ANATOMY
Venous Anatomy
The pancreas is anatomically intimately associated with critical vascular
structures including the PV, SMV, celiac artery (CA), hepatic artery (HA) and
superior mesenteric artery (SMA). The SMV drains the midgut and courses
posterior to the neck of the pancreas, joining the splenic vein (SV) in order to
form the PV. The IMV typically drains into the SV but occasionally enters in
to the SMV or forms a common confluence with SMV and SV. The
gastrocolic trunk is an early anterior tributary of the SMV and often divides
into the right gastroepiploic and middle colic vein. The first jejunal vein
typically runs posterior to the SMA to join the ileal branch in order to form
the SMV. The first jejunal vein carries particular significance as it can be
directly invaded by tumors of the uncinate process, preventing separation of
the tumor from the SMV. In addition, injuries of the first jejunal vein can lead
to excessive hemorrhage since vascular control at the root of the mesentery
can be extremely challenging.
Knowledge of the venous anatomy is critical to understanding and
applying sound technical principles for venous reconstruction. In general, the
limits of venous resection extend proximally from the bifurcation of the left
and right PV to the first jejunal/ileal branches in the root of the mesentery.
The two most important factors to consider are the location of the tumor
relative to the portosplenic confluence and the extent of involvement.
Classification systems exist that describe the extent of tumor involvement as
a means of guiding the reconstruction (Figure 2 ).
FIGURE 2: Schematic of the five forms of venous resection and

reconstruction. PV: Portal Vein; SV: Splenic Vein; SMV: Superior
Mesenteric Vein; IJ Graft: Internal Jugular Vein Graft (Adapted
with permission from Springer in Tseng JF, Raut CP, Lee JE, et al.
Pancreaticoduodenectomy with vascular resection: margin status
and survival duration. J Gastrointest Surg 2004; 8(8):935–49).
Arterial Anatomy
The SMA originates from the aorta and courses posterior to the pancreas as
it enters the root of the mesentery. The SMA maintains a fairly constant
posteromedial relationship to the SMV and is close in proximity to the SMV
throughout its length. Therefore, it is uncommon for pancreatic tumors to
directly invade the SMA in the absence of venous involvement. On the other
hand, the SMA is surrounded by a rich network of nerve and lymphatic
tissues that frequently harbor microscopic tumor cells directly invading from
adenocarcinomas of the head or uncinate process. For this reason, a
meticulous dissection during PD along the periadventitial plane of the SMA
is encouraged to maximize the retroperitoneal margin. Pancreatectomy with
concomitant SMA resections have historically been associated with
prohibitively high morbidity and mortality and are therefore not routinely
performed nor discussed further in this chapter.
TECHNICAL ASPECTS
Lateral Venorrhaphy
While the need for venous resection should be anticipated preoperatively
based on a thorough examination of cross-sectional imaging, a formal
assessment of the tumor-vein interface may be conducted after transection of
the pancreatic neck and obtaining vascular isolation and control. Venous
resection is indicated when, in the estimation of the surgeon, tumor adherence
prevents separation of the PV-SMV from the pancreatic head and uncinate
process. That is, venous resection is not performed routinely, for example, in
attempts to achieve a wider margin or enhance the lymphadenectomy.
When tumor involves only the right lateral wall of the SMV, a segment of
vein may be excised en bloc with the tumor. If the subsequent defect
comprises <25% of the vein diameter, lateral venorraphy with primary suture
may occasionally be performed. Transverse closure, as is done during
pyloroplasty for example, is the least likely method to result in stenosis. If
the venous defect is larger, or is located at the confluence, then patch
venoplasty with either autologous vein graft or bovine pericardial patch is
preferred (Figure 3 ).
FIGURE 3: Intraoperative photograph of patch venoplasty using
internal jugular autograft for repair of a V1 venous resection. This
reconstruction method is preferred when segmental resection is not
required but the venous defect is >25% of the diameter of the SMV-
PV or located at the portosplenic confluence. PV, portal vein; SV,
splenic vein; SMV, superior mesenteric vein
SEGMENTAL RESECTION AND
RECONSTRUCTION
For more extensive venous involvement, segmental resection should be
performed. Primary anastomosis is preferred when possible. In fact, defects
as long as 4cm can be repaired primarily after removal of the specimen and
complete mesenteric mobilization. When a tension free primary anastomosis
is not possible, an interposition graft should be used. Autologous grafts are
preferred over synthetic grafts due to their reduced risk of infection and
thrombosis. Our preference is the internal jugular (IJ) vein because of its
favorable size match to the PV-SMV and ease of harvest. Other autologous
options include the left renal vein (LRV) or saphenous vein. Alternatively, a
customized bovine pericardial tube graft (xenograft) can be fashioned to the
appropriate size using an endovascular stapler.
The possible need for interposition vein graft should be anticipated prior to
the operation and the patient prepped and positioned appropriately. If IJ vein
is the preferred conduit, a shoulder bump is placed and the head positioned
to the right in order to expose the left neck, which is then prepped and draped
at the start of the case. Once the tumor is deemed resectable and the surgeon
commits to proceed with PD (i.e., generally following division of the
pancreas and dissection of the specimen from the SMA), a longitudinal
incision is made along the left sternocleiodomastoid muscle and the internal
jugular vein is isolated and harvested from the clavicle to the facial vein; it
is then kept in heparinized saline until needed.
Once the SMA dissection is completed and the retroperitoneal tissues
divided, the tumor remains only attached to the PV-SMV. The right colon is
returned to its normal anatomic position in order to de-rotate the mesentery.
Low-dose systemic heparinization may be administered but is not required.
Vascular clamps are placed 2-3cm proximal and distal to the area of venous
involvement. SMA clamping is not routinely performed but may be used as a
means of preventing intestinal wall congestion that can interfere with
subsequent anastomoses, particularly if the reconstruction is anticipated to be
especially complex, or take a longer-than-average amount of time to
complete (e.g., oblique segmental interposition; or placement of a graft prior
to a difficult SMA dissection due to the presence of a large uncinate process
tumor, a close SMA margin or the presence of a replaced right hepatic
artery). Either primary anastomosis or insertion of an interposition graft is
then performed using 6-0 non-absorbable monofilament suture (Figure 4 ). In
the case of interposition grafting, we recommend performing the more
challenging side first. This is generally the distal anastomosis, as inadvertent
displacement of the distal vascular clamp can lead to significant hemorrhage,
which is difficult to control at the root of the mesentery. Occasionally,
however, the proximal anastomosis is more challenging and should be
performed first. This could be the case, for example, when the tumor is
located cephalad in the pancreatic head and extends to the undersurface of
the liver, forcing a portal venous anastomosis high in the hilum.
FIGURE 4: Intraoperative photograft of SMV-PV reconstruction
using internal jugular (IJ) vein as an interposition autograft. The
graft is fashioned using interrupted 6-0 Prolene and care is taken
to ensure that is not rotated. PV, portal vein; SMV, superior
mesenteric vein; LHA, left hepatic artery; RHA, right hepatic artery;
SMA, superior mesenteric artery
Postoperatively, patients are routinely administered antiplatelet therapy as

well as standard chemical and mechanical prophylaxis for deep vein
thrombosis; therapeutic anticoagulation is not necessary. Routine
radiographic monitoring (e.g. ultrasound) of the venous reconstruction is not
required, but attention to venous patency should be given on any follow-up
imaging studies.
MANAGEMENT OF THE SPLENIC VEIN

Venous resection and reconstruction is traditionally deferred as the last step
of PD after the retroperitoneal dissection has been completed; this helps
ensure an optimal configuration to the final reconstruction. In the absence of
venous involvement, the PV-SMV may be separated from the pancreas and
reflected to the left in order to maximally expose the SMA. However, when
separation of the venous structures is not possible, the tumor and vein must
be reflected together to the right. In this scenario, however, exposure of the
SMA is limited by the portosplenic confluence. While the SMA dissection
may proceed in many cases by intermittently retracting the SV cephalad and
caudal, division of the splenic vein at the splenoportal confluence is often the
most oncologically appropriate strategy in order to maximize exposure of the
SMA. It will also shorten the distance between PV and SMV after removal of
the specimen, since the SV is no longer tethering either structure, thereby
enabling a primary anastomosis.
However, when the SV has been divided during the course of the
operation, the surgeon must ensure the patient has adequate gastrosplenic
outflow. In most cases, outflow continues retrograde through the IMV as it
typically enters the SV. However, when the IMV enters the SMV,
gastrosplenic outflow may be compromised, especially if the coronary vein
has been divided previously during the dissection. In this scenario, in order
to prevent postoperative sinistral hypertension, creation of a distal
splenorenal shunt is recommended. The surgeon exposes the LRV to the left
of the SMA. The left adrenal vein and/or gonadal vein may be divided in
order to maximize exposure if necessary. The SV is mobilized and
anastomosed to the LRV in an end-to-side fashion using 6-0 non-absorbable
monofilament suture (Figure 5 ).
FIGURE 5: Intraoperative photograft of a distal splenorenal shunt

(DSRS), which is constructed to prevent sinstral hypertension when
adequate splenic outflow is suspected. This is most commonly the
case when the splenic vein (SV) has been divided during the course
of pancreatoduodenectomy and the inferior mesenteric vein drains
into the superior mesenteric vein (SMV). PV, portal vein; IVC,
inferior vena cava; LRV, left renal vein; HA, hepatic artery; SMA,
OUTCOMES
Short-Term Outcomes
PD with venous resection and reconstruction is a technically challenging
endeavor and one might expect higher rates of adverse postoperative events.
In fact, a study utilizing the National Surgical Quality Improvement Project
found higher postoperative morbidity (39.9% vs 33.3%) and mortality (5.7%
vs 2.9%) rates in patients undergoing PD with vascular resection compared
to those who did not undergo vascular resection. However, this finding, and
those of similar studies, are limited by the inability to differentiate planned
vascular resection and reconstruction from repair of inadvertent vascular
injuries at the time of surgery. Other single-institutional series, multi-
institutional studies, and meta-analyses have suggested that elective, planned
venous resection can be performed safely without significantly increased
morbidity rates compared to standard PD.
Long-Term Outcomes
Compared to non-operative therapies, PD with venous resection and
reconstruction leads to improved long term outcomes. In fact, it is now
universally accepted that venous involvement should not be an absolute
contraindication to surgical resection. Lygidakis et al randomized patients
with PDAC and limited vascular involvement to either PD with vascular
resection or double bypass and found 2 year survival rates of 81.8% vs 0%,
respectively. Other comparative studies have suggested similar oncologic
outcomes between patients with PDAC who underwent standard PD and
those who underwent PD with venous resection, especially when
preoperative therapy was utilized.
IMPACT OF PREOPERATIVE THERAPY

Increasingly, patients with PDAC are receiving chemotherapy and/or
chemoradiation therapy prior to surgical resection, especially those with BR
or LA tumors. Preoperative therapy offers multiple advantages, including the
guaranteed delivery of multimodality therapy to all patients, the early
treatment of micrometastatic disease presumed present in nearly all patients
with PDAC, facilitation of a margin negative resection, and, most
importantly, the selection of patients with favorable tumor biology who are
most likely to benefit from major surgery. Furthermore, evidence suggests
that the administration of preoperative therapy is not associated with
increased adverse events. On the other hand, preoperative radiation has the
advantage of being associated with reduced rates of postoperative pancreatic
fistula.
PATENCY
In general PV-SMV resection and reconstruction is associated with high
patency rates. Venous thrombosis can occur at any time after surgery, but
acute thrombotic events are associated with the highest morbidity. Late
thrombotic events are often related to cancer recurrence. Little literature
exists on risk factors for venous thrombosis after reconstruction, though some
studies have suggested that size mismatch, interposition grafting, duration of
the operation, and postoperative complications are associated with patency
failures. Primary suture repair, either lateral venorraphy or end-to-end
anastomosis, is associated with the lowest incidence of thrombosis.
SALIENT POINTS
With meticulous attention to preoperative cross-sectional imaging,
careful patient selection, the use of multimodality therapy, and sound
technical principles, the benefits of surgical resection can be safely
extended to those patients with PDAC and venous involvement.
The need for vascular resection should be anticipated based on the
preoperative imaging but critically assessed intraoperatively after
division of the pancreas.
Attempts at splenic vein preservation should be undertaken but not at
the expense of an oncologically inadequate SMA dissection.
The venous resection and reconstruction is deferred as the last step of
the operation and attempts should be made to perform primary repair
when possible.
When the venous defect cannot be repair primarily, an interposition
graft using autologous vein, should be fashioned.
When performed at experienced high volume centers in appropriately
selected patients, PD with concomitant venous resection is associated
with excellent short and long term outcomes.
SELECTED REFERENCES
1. Chua TC. & Saxena A. Extended pancreaticoduodenectomy with
vascular resection for pancreatic cancer: a systematic review. J.
Gastrointest. Surg. 2010; 14(9), 1442–52.
2. Katz MH, Pisters PW, Evans DB, Sun CC, Lee JE, Fleming JB, et al.
Borderline resectable pancreatic cancer: the importance of this
emerging stage of disease. J. Am. Coll. Surg. 2008; 206(5), 833-46;
discussion 846-8.
3. Katz MH, Fleming JB, Pisters PW, Lee, JE, Evans DB. Anatomy of the
superior mesenteric vein with special reference to the surgical
management of first-order branch involvement at
pancreaticoduodenectomy. Ann. Surg. 2008; 248(6), 1098–102.
4. Tseng JF, Raut CP, Lee JE, Pisters PW, Vauthey JN, Abdalla EK, et al.
Pancreaticoduodenectomy with vascular resection: margin status and
survival duration. J. Gastrointest. Surg. 2004; 8(8), 935-49; discussion
949-50.
5. Cloyd JM, Katz MH, Prakash L, Varadhachary GR, Wolff RA, et al.
Preoperative Therapy and Pancreatoduodenectomy for Pancreatic
Ductal Adenocarcinoma: a 25-Year Single-Institution Experience. J.
Gastrointest. Surg. 2017; 21(1), 164–174.

CASE SCENARIO
During a pancreaticoduodenectomy (PD) for cholangiocarcinoma, the
resection proceeds without much difficulty, and now it is time for the
reconstruction aspects of the operation. There has been 500 mL of blood loss
to this point. The pancreaticobiliary drainage limb is introduced into the right
upper quadrant, and the remnant pancreatic duct is searched for at the cut
surface of a soft, fatty gland. After considerable inspection, a diminutive duct
measuring 1 mm is visualized (Figure 1 ).
FIGURE 1: Clinical scenario of a pancreatic gland with soft
texture and a diminutive main pancreatic duct during a
pancreatoduodenectomy for cholangiocarcinoma with 500 mL of
estimated blood loss.
█ BACKGROUND
FRAMING THE ISSUE

Even the most experienced pancreatic surgeon may be humbled by such a
high-stakes pancreatico-enteric anastomosis and may ask, “What do I do with
this?” Pancreatic anastomotic failure, with the subsequent development of a
clinically relevant postoperative pancreatic fistula (CR-POPF), can be
expected in nearly one-third of all such high-risk cases and is the greatest
contributor to morbidity and mortality in these scenarios. Fortunately, or
unfortunately, pancreatic surgeons have a full tool kit at their disposal, with
multiple options and permutations of anastomotic reconstruction techniques
and technologies available to mitigate such risk. One may be under the
impression that “more is better” in this high-risk scenario, and feel the need
to bootstrap the anastomosis with as many of these tactics as possible. Yet,
through the precise definition of outcome and application of risk
stratification, it is now possible to sort through the available data and strip
down the process of anastomotic creation to a few simple tenets.
DEFINING FISTULA RISK

First of all, just what do we mean by “risky”? One surgeon’s impression of
hazard may be vastly different than another’s. For decades, this concept of
risk for the development of pancreatic fistula was mired in subjectivity—
basically, risk is in the eye of the beholder.
To begin, it is useful to understand just how we perceive risk. A recent
survey of approximately 900 pancreatic surgeons from around the globe
demonstrated that most surgeons believe the most common contributors to
CR-POPF following PD are: 1) Gland characteristics (e.g. parenchymal
quality, pancreatic duct size; 91% of those surveyed), 2) Patient features (e.g.
comorbidities, medications; 71%), and 3) Intraoperative factors (e.g. blood
loss, hemodynamic instability; 69%). On the other hand, institutional features
(e.g. procedural volume), or the contribution of available perioperative
fistula mitigation strategies (e.g. stents, drains, somatostatin use), were NOT
perceived to be influential. Most surgeons consider the development of
fistula to be multifactorial, citing an average of three factors, with
presumably a higher frequency in the risky anastomosis.
More precise identification of discrete risk factors, and their relative
contributions to CR-POPF, first required standardized definition of the
outcome of interest. Prior to 2005, there were myriad definitions of what
constitutes a pancreatic fistula—often defined by local, institutional customs.
The International Study Group on Pancreatic Fistula established a consensus
definition of CR-POPF (recently refined in 2016), providing the first step in
scientifically reproducible assessment of fistula risk. In the context of this
consensus definition, patient-specific risk factors for CR-POPF following
PD have subsequently been identified as (1) soft gland texture, (2) small
pancreatic duct size, (3) pathology exclusive of pancreatic cancer or
pancreatitis, and (4) elevated blood loss. These factors contribute, in a
composite fashion, to the Fistula Risk Score (FRS), an easy-to-use 10-point
scale that can be applied in the operating room, which accounts for the risk
profile at “time-zero”—the point of creating the pancreatic anastomosis.
Importantly, it segregates nicely in to four discreet “Risk Zones” (Negligible
– FRS 0; Low – FRS 1-2; Moderate – FRS 3-6; High – FRS 7-10), which
correspond to escalating CR-POPF occurrence (Figure 2 ).
FIGURE 2: The Fistula Risk Score and its relationship to
clinically-relevant pancreatic fistula following
pancreatoduodenectomy.
It is important to note that this tool readily demonstrates how risk is

multifactorial and more complex than just considering one or two commonly
ascribed risk factors. As proof, Table 1 displays the results from the multi-
national, multi-institutional Pancreas Fistula Study Group, culled from over
5300 PD patients cared for by over 60 surgeons at 17 pancreatic surgical
specialty institutions. The most commonly recognized and cited risk factors
in isolation are soft gland parenchyma and small duct size (generally
considered anything ≤3 mm). Many consider that the combination of these
two findings portends additive, if not synergistic, risk. Real-world results
would suggest that this is actually only a modest escalation of risk. Instead,
the FRS relies on the interplay of four factors, with smaller duct size and
elevated blood loss being the most strongly weighted. In fact, the presence of
soft parenchyma alone provides only 20% of the possible total contribution
to the score. Notice from the table how the quantitative risk category of High
FRS (7-10), using all four variables, provides the best calibration of hazard.
TABLE 1: Association of clinically-relevant pancreatic
fistula (CR-POPF) with purported risk factors for its
development after pancreatoduodenectomy, taken from
5,323 procedures performed by the Pancreas Fistula Study
Group.
Hence, we believe that the current body of literature regarding

management of the high-risk pancreatic anastomosis (or any, for that matter)
should be evaluated in the following manner: (1) did the tested mitigation
strategy reduce the risk of fistula using the consensus definition of CR-POPF,
and (2) did the study use risk-stratification and, in particular, identify a focal,
risky cohort for strategy evaluation? The advent of this externally validated
prediction tool (FRS) has allowed for deeper scrutiny of this topic. Table 2
demonstrates that there are 20 specific combinations of risk factors that can
equate to a High-Risk scenario.
TABLE 2: The various combinations of risk factors for
clinically-relevant pancreatic fistula that constitute the
High Risk zone of the Fistula Risk Score (FRS).
WHAT DO WE MEAN BY “MITIGATION
STRATEGIES”?
Broadly speaking, mitigation strategies are any technique or approach that the
surgeon might take to decrease either the occurrence or the severity of
pancreatic fistula formation. These can be sub-categorized in to either: 1)
Anastomotic techniques, or 2) Technological adjuncts. Anastomotic
approaches include: pancreaticogastrostomy (PG), pancreaticojejunostomy
(PJ), duct-to-mucosa technique, invagination, dunking, Roux (double) limb
creation, or no anastomotic creation (duct occlusion, stapled, or total
pancreatectomy). Adjunct technologies include the use of external drains,
somatostatin analogues, the use of trans-anastomotic stents (either internal or
external), tissue sealants, and autologous tissue patches. Obviously, the
various permutations of all of these factors is overwhelming, and it is
therefore near-impossible to account for all these variations in randomized,
controlled studies. What follows is a critical assessment of each of these
options, particularly in the setting of high Fistula Risk Score situations
(FRS7-10), using data from the aforementioned multi-center, international
Pancreas Fistula Study Group.
ANASTOMOTIC APPROACHES
Pancreaticogastrostomy
Of the various mitigation strategies that have been proposed to reduce the
occurrence of CR-POPF, much academic attention has focused on the
technical variation between gastric versus jejunal drainage of the pancreatic
remnant. There have been five randomized controlled trials (RCTs) to assess
the value of PG using contemporary definitions of outcome, with conflicting
findings, and only two included any form of risk stratification. In two of these
RCTs, PG did not improve CR-POPF rates in cases of a soft gland, but did
provide benefit when the duct size was small (<3 mm).
There are certain hesitations prior to endorsement of PG reconstruction
for a high-risk anastomosis. A major limitation includes the simplistic
stratification of risk in these RCTs (i.e., without analysis beyond one risk
factor). In fact, in a propensity-matched, risk-stratified analysis of the
Pancreas Fistula Study Group data, PG was not associated with improved
fistula rates at any FRS risk zone. Surgeons demonstrated preferential use for
PG reconstruction with increasing fistula risk, suggesting selective use for
the high-stakes anastomosis. But even in the High FRS cases, there was no
observed benefit of PG in mitigating fistula risk or severity. Moreover, when
surgeons facing a High FRS scenario used an external stent—which was
strictly used with PJ reconstruction (although hypothetically could be used
with PG reconstruction)—the fistula rate was markedly reduced as compared
to PG. In this regard, current RCTs evaluating the utility of PG are limited in
their testing of individual mitigation strategies in isolation, whereas an
optimal analysis requires assessment of multiple strategies in conjunction.
Additionally, in the United States (and other areas of the world), there are
technical concerns about this less commonly utilized (and taught)
reconstruction. Globally, only 10% of pancreatic surgeons claim that PG is
their preferred form of pancreatic reconstruction. That is, if High-Risk cases
identified by the FRS generally constitute just a tenth of all PDs, the possible
benefits of such a rarely-used technique may be offset by its infrequent
application. In support of this notion, a recent position statement by the
International Study Group for Pancreatic Surgery suggested that surgeons
faced with high-stakes scenarios adhere to the reconstruction techniques that
they regularly practice. When surveyed, two-thirds of surgeons indicate that
they apply the same technique all the time, while one-third modulate their
approach given the particular circumstances they face.
Pancreaticojejunostomy
On the other hand, PJ reconstruction remains the overwhelmingly applied
anastomotic technique around the world (89% of surgeons). From a technical
standpoint, there are myriad ways a duct-to-mucosa PJ can be created,
making head-to-head testing of efficacy nearly impossible. The author’s
preferred technique is described in great detail in the book chapter reference
provided below. Certainly, other approaches to PJ reconstruction with
different nuances are available, but this described duct-to-mucosa technique,
refined over hundreds of cases now, has resulted in a more than satisfactory
outcomes for us. This includes a 2.7% CR-POPF rate in the last 150 PJs,
including no fistulas in all (0/11) high-FRS cases.
Alternative Forms of Pancreatojejunostomy

Occasionally, a duct will not be visible or will be too small (or fragile) to
allow for a duct-to-mucosa anastomosis. There have been two ISGPF-era
RCTs to address the value of dunking or invaginating reconstruction, although
neither provided risk-stratified analysis for the high-stakes anastomosis.
Still, these technical variations can be a helpful ally when needed, and we
have employed them successfully in a handful of cases. Similar to the duct-
to-mucosa PJ, the invagination anastomosis is usually created in an end-to-
side fashion in two layers, although and end-to-end version is also
described. Here, the jejunum is opened for a length slightly shorter than that
of the cut pancreatic surface long-axis. An inner layer is crafted, which
approximates the cut-edge of the pancreatic capsule at the transection margin
to the open, sero-mucosal edge of the bowel. The second, outer layer
effectively rolls over the jejunal wall to attach the serosa of the jejunum to
the pancreatic capsule about 1-2 cm from the inner anastomosis (Figure 3A-F
). Another variation of an invagination is the Blumgart technique where
through-and-through mattress sutures are placed across the entire pancreatic
gland (Figure 4A-B ).
FIGURE 3: Depiction of an invaginated anastomosis
FIGURE 4: The Blumgart technique for an invaginated
anastomosis.
Technically, the terminology “dunking” refers to the projection of the

pancreatic tissue in to the enterostomy site by a distance of a few
centimeters. Many PGs are created in that fashion. The seromucosal edge of
the bowel enterostomy is attached in a single layer directly to pancreatic
capsule about 2 cm from the transection plane. Thus, the free transection
edge, and the orifice of the main duct within it, are essentially floating within
the lumen (Figure 5A-C ).
FIGURE 5: A “dunking” anastomosis
Another variant of the end-to-side PJ is a single layer attachment of
bowel directly to the edge of the transection plane (Figure 6 ). This is how
Whipple performed his pancreatoenterostomies (once he began to employ
that component of the procedure). Some authorities feel that fragile
pancreatic parenchyma should have as few stiches applied as is necessary to
create a seal. Therefore, the outer layer of the invaginating approach (with its
distortion of the bowel and tensions on the pancreatic parenchyma) is
avoided. Instead, to create this, an enterotomy is created that is slightly
smaller than the dimensions of the pancreatic cut face (similar to the
invaginating technique). We prefer to perform a single-layer apposition of the
cut edge of parenchyma to a sero-mucosal bite of bowel with interrupted 2-0
silk sutures. Of course, care is taken to avoid any central aspects of the
parenchyma where a diminutive pancreatic duct might lurk. Extreme care
must be taken to apply just enough tension to lock the knots, without ripping
the fragile pancreatic parenchyma. We believe this particular approach has
the advantage over the others in that more substantial bites of pancreatic
tissue can be achieved tension-free, while using fewer overall sutures on the
particularly fragile gland.
FIGURE 6: Single layered anastomosis performed originally by
Whipple
The utility of these variations has been studied within the Pancreas Fistula
Study Group database. In a series of 459 High FRS PJs, 88
dunking/invaginating anastomoses were employed - most commonly when the
pancreatic duct was ≤1 mm. In this subset, it was not associated with a
decreased occurrence of CR-POPF (22.9% vs. duct-to-mucosa: 17.5%,
p=0.454). In summary, given this data, we use the dunking technique only in
the very rare circumstance when there is a 1mm or imperceptible duct (<3%
of all cases), where a duct-to-mucosa approach is practically impossible.
Roux Limb Creation

It has been postulated that dissociating the mixture of bile from pancreatic
juice may decrease the formation of pancreatic fistulas following PD. In
order to achieve this, the separate hepatico- and pancreatico-jejunal
anastomoses must be divorced on different bowel conduits. This is achieved
by creating a separate Roux-en-Y drainage limb. This brings with another
anastomo sis (entero-enterostomy) and, therefore, added operative time, as
well as the possibility of leaks or bleeding. Accordingly, it is actually
seldom employed (frequently or always used roughly 11% worldwide).
Historical data from empirically weak studies conducted in the pre-ISGPF
era has never demonstrated value to this approach when applied to all PDs.
Furthermore, its efficacy in high-risk scenarios has never been specifically
tested.
Duct Occlusion
Occlusion of the pancreatic remnant duct without anastomosis, as originally
implemented by Whipple during his initial cases, remains a possibility in
certain, usually extreme, operative circumstances. When a duct cannot be
identified and the operation need be hastened for physiologic instability or
other reasons, the distal pancreatic remnant can be closed in a number of
ways. Through-and-through 2-0 silk sutures may be placed in an interlocking
U fashion at the cut pancreatic end. If a duct can be visualized, it can
separately be suture ligated or sealed with a polymer prior to closing the
pancreatic capsule. Alternatively, a stapled closure, using an appropriate
sized staple for the parenchymal thickness encountered, is a swift alternative
in cases of physiologic instability.
Regardless of the method, the adjacent area should be externally drained
as these occlusions tend to leak pancreatic juice over the short term, while
rarely becoming a significant, externalized fistula. While exocrine
insufficiency may develop, islet cells are often preserved and overt diabetes
is infrequent. Due to its infrequent application, and lack of studies in the
modern era, there are no significant data to support its regular use in all PDs
– particularly given obligate exocrine insufficiency. We offer the Whipple
procedure as a “bail-out” maneuver in a case of extreme risk or physiologic
decompensation. This approach is quicker and less impactful than the
alternative of a “completion” total pancreatectomy, or external drainage of an
uncontrolled, open pancreatic duct.
TECHNOLOGICAL AND PHARMACOLOGIC
ADJUNCTS
Following creation of the anastomosis, there are several strategies available
to potentially mitigate the risk of pancreatic fistula. These include the
application of intraperitoneal drains, anastomotic stents, prophylactic
somatostatin analogues, and tissue sealants or autologous patches. As stated
earlier, the temptation may be that employing more of these is better – in
other words, “Throw everything and the kitchen sink at it”. However, we
believe that is not necessarily the best course.
Intraperitoneal Drainage
Contemporary randomized studies suggest that intraperitoneal drains are
valuable in cases of high fistula risk, and may decrease both the rate and
severity of CR-POPF. Following PD, the author routinely places one fluted
19-Fr Blake (silicon, closed suction) drain anterior to the anastomosis in
cases of moderate and high anastomotic risk (FRS 3-10). We avoid placing
rigid drains directly behind the anastomosis, as they can cause unnecessary
trauma to the posteriorly located portal and superior mesenteric veins. We
then adhere to a protocol of early drain removal, which is driven by the drain
amylase concentration on POD1. If the POD1 drain amylase is < 5000, and
there is no sinister effluent, then the drain is removed early (on POD3). This
has been successful and safe even in cases of a High FRS. If it is >5000, the
drain is removed later, at the surgeon’s discretion. It is important to note that
early drain removal for cases of POD1 amylase greater than 5000 has never
been tested in a randomized, controlled fashion, so we do not yet have
evidence that early drain removal in this scenario is either safe or valuable.
Transanastomotic Stents
Anastomotic stents can facilitate the creation of the duct-to-mucosa
anastomosis and may prevent its dehiscence by shunting digestive pancreatic
enzymes. Externalized transanastomotic stents diminished the rate of
clinically relevant fistula in two ISGPF-era RCTs of “risky” anastomosis
(i.e., non-dilated ducts +/- soft pancreatic texture). In a more rigorously risk-
adjusted analysis of the Pancreas Fistula Study Group, externalized stents
were not associated with improved fistula outcomes for Negligible, Low and
Moderate risk PDs (FRS 0-6). However, externalized stents in patients with
High FRS risk demonstrated markedly lower rates of CR-POPF. Moreover,
when stratifying comparisons by individual FRS risk factors in isolation,
there were no significant differences in rates of CR-POPF in the setting of
either a soft gland, a small duct, high-risk pathology, or elevated blood loss –
suggesting that the benefit of external stents may solely be in the cohort of
composite risk identified by multiple FRS factors (rather than individual risk
factors in isolation).
In contrast, according to a FRS risk-adjusted analysis, internalized stents
led to increased complications and should generally be avoided since they
do not diminish CR-POPF rates. It is unclear why this difference between
externalized vs. internal stents exists; what is even more intriguing is that
external stents appear to only be beneficial in the highest risk (i.e., High
FRS) anastomosis. One hypothesis involves competing factors between the
benefits of a stent, as mentioned above, and the detrimental influence of a
foreign body at the anastomosis. In cases of low and moderate risk, these
potential benefits may be outweighed; however, an externalized stent appears
to have a uniquely beneficial role in High FRS cases. Internal stents, by
definition, do not drain digestive enzymes to the external environment, and
thus may be a less effective shunt.
Tissue Patches and Sealants
In certain instances, the pancreas will have a fibro-fatty layer surrounding its
anterior capsule. If so, preserve this layer when diligently dissecting the
pancreas, and keep it attached to the pancreatic body if possible. It can be
tacked with silk sutures to the pancreaticobiliary limb to create a “hood” for
the anterior aspect of the anastomosis. The omentum, or alternatively the
umbilical ligament with its vascular pedicle, may also be used in a similar
fashion to “wrap” the anastomosis. There are no reasonable data to support
any value to either strategy in any degree of risk, but such tissue patches are
readily available, devoid of cost, and possibly beneficial. These adjuncts are
used always or frequently by just 13% of surgeons who perform PD.
In contrast, biologic sealants add increased cost to the operation without
proven efficacy. Interestingly, while infrequently used with regularity, 18%
of surgeons profess they use sealants “occasionally” (1-25% of their cases)
—probably reflective of cases with perceived higher risk for fistula
development. There is only one ISGPF-era RCT on the use of fibrin glue,
which distinguished between biochemical leaks and CR-POPFs; it did not
find a clinical value to this strategy. No inclusion criteria related to fistula
risk were employed; hence, the utility of sealants specifically in high-risk
cases is so far understudied. Likewise, an older RCT which included only
PDs where there was a soft gland and non-dilated duct did not find that
topical administration of a fibrin glue sealant altered the rate of fistula or
other intra-abdominal complications, with the caveat that this trial preceded
the consensus definition of CR-POPF. Neither sealants, nor patches, have
demonstrated any protective value in FRS-based risk adjusted analyses using
the Pancreas Fistula Study Group data.
Somatostatin Analogues
Prophylactic use of somatostatin, and its analogues—octreotide and
pasireotide—are some of the most studied, and controversial, adjuncts to
possibly mitigate the risk of pancreatic fistula. Octreotide prophylaxis has
been prospectively evaluated in over 20 trials; while large meta-analyses
have reported a decreased rate of pancreatic fistula, only three RCTs
employed established definitions for CR-POPF. Octreotide did not reduce
fistula rates in any of these contemporary studies, and the rates were instead
higher in the octreotide cohort in two of them. Moreover, none specifically
studied the high-risk anastomosis, identified by either the FRS or other
criteria.
On the other hand, retrospective, risk-adjusted analysis of the Pancreas
Fistula Study Group data has shown intriguing results. In fact, worse CR-
POPF and overall outcomes were seen with the application of prophylactic
octreotide; furthermore, the negative effect appears to be potentiated with
higher risk. In other words, when applied to High FRS scenarios, CR-POPFs
developed at an alarming rate of 65%, while that rate was just 18% when it
was not used in that setting. Based on these findings, the author has omitted
prophylactic octreotide from his practice and observed markedly improved
overall fistula outcomes in his last 100 PDs.
In contrast, the preliminary data for pasireotide is promising, with
reduced rates of clinically relevant fistulas in a single-center trial. However,
validation in a multi-institutional setting, increased access to the medication,
and data specific for the high-risk anastomosis are still lacking.
PUTTING IT ALL TOGETHER

Developing an Optimal Approach to the High Risk
Scenario
A recent analysis from the Pancreas Fistula Study Group has concentrated on
the characteristics, outcomes, and optimal management of 522 patients
considered High Risk (FRS 7-10) for CR-POPF development. Thirty-seven
different permutations of available mitigation strategies were employed,
from the most simplistic anastomosis (PJ without any stent, drain, tissue
patch or octreotide) to the most elaborate (PG with an internal stent, drain
and prophylactic octreotide), suggesting a lack of consensus on how surgeons
deal with this high stakes scenario. Using multivariable logistic regression,
externalized stents and octreotide omission were identified as two strategies
that were each independently associated with improved fistula outcomes.
Moreover, the combination of externalized stents and octreotide omission
was associated with fistula outcomes that were superior to the use of either
strategy in isolation (decreasing the CR-POPF rate to just 13% in this high-
risk cohort). This “optimal” mitigation bundle was always used in the context
of PJ reconstruction and with an intraperitoneal drain; hence, these four items
can be considered the essential components to mitigating CR-POPF
occurrence in the High FRS scenario.
Tips and Perils

While performing a PD in this specific setting, there are a few points to
consider. Firstly, while innate risk factors have been elucidated and have
been codified in the FRS, we contend that nothing is more important to
prevent a CR-POPF than a technically precise anastomotic creation. While
this certainly applies to all cases, it is even more tantamount in situations of
high risk where there is no margin for technical error on usually fragile
tissue. In our opinion, this is a situation that requires the skill of the attending
surgeon with experience, rather than a trainee.
Second, in homage to basic surgical tenets, assure that the
pancreaticobiliary drainage limb is healthy, with good blood flow and sits in
the right upper quadrant without undue tension or evidence of obstruction.
This reconstruction is fragile enough at the anastomotic site, and therefore the
drainage aspect distal to it must also be ideal. While we usually place this
limb in a retro-mesenteric position to create a so-called “neo-duodenum”,
this sometimes creates undue tension and compression on the limb,
especially in circumstances of significant central adiposity. Accordingly, if
this layout does not appear suitable, the limb may be brought forward through
a retrocolic window created in the transverse meso-colon.
Lastly are some perils and pitfalls to consider. Prior to beginning the
anastomosis, gently dilate the duct using a slow, prolonged opening of a
pediatric right angle instrument within the lumen, which can be valuable in
providing more caliber to sew to. Care must be taken, however, not to
“crack” the duct. When applying an external stent, ensure that the Silastic
pediatric feeding tube is sized appropriately for the pancreatic duct, and
allow it to traverse the anastomosis without fixation. It will be externalized
and secured with an absorbable, purse-string suture at the enterotomy, either
several centimeters prograde from the hepaticojejunostomy, or beyond the PJ
site on the redundant end of the bowel limb. The stent may also afford greater
precision when suturing the inner anastomotic layer. For this, consider the
use of a 7-0 caliber double armed suture in extremely small diameter ducts.
Double-armed sutures are useful in providing the most flexibility for needle
placement. The number of sutures placed is variable given the size and
integrity of the duct; extremely small ducts may only accommodate 4
“corner” sutures, but this will generally be adequate. The tissue is usually
very fragile, so multiple “attempts” at placing the sutures in the appropriate
orientation are not advised. Needle angle placement needs to be correct on
the first attempt. Of course, this is the sort of situation where magnified
vision with Loupes, and the use of a Castro needle driver, provide added
value in affording exact precision with suture placement.
SALIENT POINTS
Risk adjustment methodologies, such as the Fistula Risk Score, are
useful for identifying high risk anastomoses as well as for comparing
outcomes of mitigation strategies across studies.
This situation demands the best skill and expertise available. It is not a
time to practice.
Stay true to your tenets: if you regularly use a duct-to-mucosa
pancreaticojejunostomy for reconstruction, this may be your best
choice. Alternative approaches (Dunking, Invagination, modified
Blumgart, single-layer) are options when there is a diminutive duct that
cannot be either identified or sewn.
Intraperitoneal drains and externalized transanastomotic stents may
improve patient outcomes in high-risk scenarios. There is insufficient
evidence to recommend the use of octreotide prophylaxis, biologic
sealants, Roux limbs, or tissue patches.
An optimal approach to the high-risk scenario has been identified,
which includes a pancreaticojejunostomy with an externalized
transanastomotic stent, an intraperitoneal drain, and the omission of
prophylactic octreotide.
SELECTED REFERENCES
1. McMillan MT, Malleo G, Bassi C, Sprys MH, Ecker BL, Drebin JA, et
al. Pancreatic Fistula Risk for Pancreatoduodenectomy: An International
Survey of Surgeon Perception. HPB(Oxford) 2017; 19(6):515-524.
2. Bassi C, Marchegiani G, Dervenis C, Sarr M, Abu Hilal M, Adham M,
et al. The 2016 Update of the International Study Group (ISGPs)
Definition and Grading of Postoperative Pancreatic Fistula: 11 Years
After. Surgery. 2017; 161 (3):584-591.
3. McMillan MT, Sprys M, Malleo G, Bassi C, and Vollmer CM. Defining
the Practice of Pancreatoduodenectomy Around the World.
HPB(Oxford). 2015; 17(12);1145-54.
4. Callery MP, Pratt WB, Kent TS, Chaikof E, and Vollmer CM. A
Prospectively Validated Risk Score Model for Pancreatic Fistula after
Pancreaticoduodenectomy. J Am Coll Surg. 2013; 216(1):1-14.
5. Vollmer CM. Pancreaticoduodenectomy: Pancreaticojejunostomy. In
Mulholland M, Hawn M, Hughes S, Albo D, Sabel M, and Dalman R,
eds. Operative Techniques in Surgery. First Edition. Lippincott
Williams & Wilkins, 2015.
6. McMillan MT, Soi S, Asbun HJ, Ball CG, Bassi C, Beane JD, et al.
Risk-Adjusted Outcomes of Clinically Relevant Postoperative Fistula
Following Pancreatoduodenectomy: A Model for Performance
Evaluation. Ann Surg. 2016; 264(2):344-52.
7. Ecker BL, McMillan MT, Asbun HJ, Ball CG, Bassi C, Vollmer CM, et
al. Characterization and Optimal Management of High-Risk Pancreatic
Anastomoses During Pancreatoduodenectomy. Ann Surg. 2017. In press.

CASE SCENARIO
A 57-year-old woman was incidentally found to have a 3 cm lesion within
the distal body of the pancreas on CT scan performed for an unrelated
reason. Endoscopic ultrasound demonstrated en-casement of the splenic
vessels without any additional vascular involvement and biopsy returned as
adenocarcinoma. On operative exploration, the lesion was noted to possibly
involve the posterior wall of the stomach, and a partial gastrectomy was
done as part of an en-bloc resection of the spec-imen in a plane anterior to
the left adrenal gland. The pancreatic neck margin was free of involve-ment
and was noted to be soft in consistency. Despite considerable effort, the main
duct of Wir-sung was unable to be identified. The remnant was oversewn
with 2-0 pledgeted Prolene vertical mattress sutures and covered with the
falciform ligament. A 15 French round Blake™ drain was left in place near
the pancreatic remnant and brought out through a separate stab incision. On
post-operative day 3, the drain amylase was 78 U/L and the volume was 50
mL/24 hours. The drain was removed and the patient discharged on
postoperative day number 4. The patient returned five weeks postoperatively
with complaints of early satiety and nausea. Computed tomography re-vealed
an 8 cm fluid collection in the operative bed with compression of the
stomach. The patient underwent successful endoscopic drainage via
cystgastrostomy. The endoscopic stent was re-moved four weeks later with
no further sequelae.
█ BACKGROUND
“Throughout the surgical literature, in the past as well as currently, the
issue of drainage has been the subject of heated debate and remains the
source of abundant confusion…. The presence of a drain should not be
relied on instead of proper surgical management… drains cannot com-
pensate for poor surgical judgement and technique.” Margaret Levy, The
American Journal of Surgery, 1984.
The utilization of an intraperitoneal drain following pancreatectomy may best

be described as an anecdotal reaction founded on little more than long-
standing tradition. The placement of a drain following distal resection has
been based, in no small part, on the known serious morbidity associated with
these pancreatic resections including bleeding, fistula, and abscess
formation; all with their often inherent need for postoperative therapeutic
intervention. In theory, routine drain placement may allow for early
recognition and/or initial control of these complications thereby mit-igating
the potential need for other postoperative invasive procedures that may result
in an unto-ward delay in recovery.
In other gastrointestinal procedures including those involving the
gallbladder, biliary tree, colon, and liver, studies have demonstrated there is
no advantage to prophylactic, operatively placed drains. As such, the routine
use of drains in these procedures has been largely abandoned. In
contradistinction, there is a paucity of literature investigating the need for
operative site drainage following distal pancreatectomy. In recent years, the
routine use of drains as an adjunct to this pro-cedure, as with
pancreatoduodenectomy, has been challenged.
POTENTIAL BENEFITS OF DRAIN

PLACEMENT
Pancreatectomy is associated with a morbidity rate of 10-30%, which has
remained unchanged in the recent literature and clinically can manifest as a
spectrum ranging from innocuous to life threat-ening. These procedure-
related complications often require therapeutic intervention including the
need for percutaneous or endoscopic drainage and, in some cases,
reoperation. In theory, the prophylactic utilization of a drain may identify
early postoperative hemorrhage and control the drainage of blood, chyle, or
pancreatic fluid if a leak should occur from the remnant gland, thus
potentially allowing early identification of patients that may require
reoperation (bleeding) and theoretically minimizing the development of
postoperative pancreatic fluid collections and infectious sequelae.
Moreover, the presence of a pancreatic fistula, when it occurs, may be
recognized sooner based on the measurement of amylase in the drainage.
Accordingly, the need for further postoper-ative intervention may be
diminished.
POTENTIAL DISADVANTAGES OF DRAIN

PLACEMENT
Several potential complications associated with the use of intraperitoneal
drains may outweigh their proposed benefit. As a foreign body within the
peritoneal cavity, drains may contribute to the postoperative inflammatory
response, provoking the formation of adhesions. While drains are generally
considered to be ”soft,” there have been published reports of their possible
association with erosion into structures within the operative bed, including
the surrounding vasculature and hollow viscera. The negative pressure
gradient generated by bulb suction may potentially promote pancreatic fistula
formation from the distal remnant. If left in place for prolonged periods,
drains may serve as a nidus for infection by allowing retrograde bacterial
seeding to an otherwise sterile operative bed. Drains may become isolated
from the pancreatic remnant by surrounding structures soon after closure of
the abdomen, thereby defeating the purpose for which they were placed. Fi-
nally, modern interventional radiology or endoscopic techniques now often
allow for safe and ef-fective percutaneous or endoscopic drainage, should a
fluid collection develop in the postoperative period, obviating the need for a
drain in most cases.
EVIDENCE
The literature regarding routine drainage following distal pancreatectomy is
sparse, as most studies have focused on pancreatoduodenectomy rather than
distal resections. Table 1 summarizes key studies to date.
TABLE 1: Studies examining the use of routine surgical
bed drainage following distal pancreatectomy
In a retrospective review, Paulus and colleagues identified 69 patients

that underwent elec-tive distal pancreatectomy over a 14-year period, 39 of
whom received a drain. Only one surgeon in this study was responsible for
all patients not receiving a drain. Fifty percent of patients that un-derwent
operative bed drainage experienced an intra-abdominal complication,
similar to the 43% among those not receiving a drain. There was no
significant difference between the groups with respect to the need for
therapeutic intervention postoperatively. The presence of a drain was useful
in identifying or treating a complication in only 3 patients, contradicting the
belief that a drain leads to early recognition and management of morbidity
following pancreatic resection.
Correa-Gallego et al. analyzed their institutional experience of 1122
pancreatic resections over a 5-year period. Drain placement was at the
discretion of the operating surgeon and a variety of remnant closure
techniques were utilized in those undergoing distal pancreatectomy. Of 350
distal resections in this study, over half (56%) did not receive a drain at the
time of operation. In the subset analysis of left-sided resections, those with a
drain experienced significantly more grade III/IV morbidity (32% vs 22%)
and had a significantly greater length of stay (7 vs 5 days). How-ever, those
receiving a drain also had significantly higher operative times and blood
loss. When comparing those with and without a drain, there were no
significant differences in the rate of pan-creatic fistula or need for
postoperative intervention, including interventional radiology drainage or
reoperation.
The impact of drain placement following distal pancreatectomy was
further assessed utiliz-ing the American College of Surgeons-National
Surgical Quality Improvement Program (ACS-NSQIP) Pancreatectomy
Demonstration Project database that was accumulated over a 14- month
period. Of 761 patients undergoing elective distal pancreatectomy, 606 had
prophylactic drainage. Propensity score matching was utilized to compare
those receiving a drain to those that did not (116 in each arm). Overall
morbidity and pancreatic fistula rates, as defined by ACS-NSQIP, were
significantly greater among those receiving a drain (43% vs 30% and 21.7%
vs 7%). In this study, drains did not reduce the incidence of clinically
relevant pancreatic fistulas, intra-abdominal septic morbidity, or the need for
postoperative therapeutic intervention. It should be noted that of the 8 deaths
in this database, none were captured in the propensity scored analysis.
Most recently, Van Buren et al. published the results of a multi-
institutional prospective tri-al of patients undergoing distal pancreatectomy
at 14 high-volume centers in North America ran-domized to either undergo or
forego an intraperitoneal drain (174 and 170, respectively). Drain
management was predefined and directed by postoperative amylase values
and volume of output. In these two well-matched groups, there was no
significant difference in the rate of grade 2 or higher complications (42%(no
drain) vs 44%). While there was a higher rate of intra-abdominal fluid
collections among patients without a drain (22% vs 9%), there was no
difference in clinically relevant postoperative pancreatic fistula (12% vs
18%) or intra-abdominal abscess (8% vs 9%) be-tween those without and
those with a drain. There was no difference between groups relative to the
need for postoperative imaging or therapeutic intervention, hospital
readmission, or quality of life scores. Overall 90-day mortality was not
significantly different between treatment groups; however, the only 2 deaths
in this series occurred on those that did not receive a drain.
WHEN SHOULD A DRAIN BE USED

FOLLOWING DISTAL PANCREATECTOMY
AND HOW SHOULD IT BE MANAGED?
During the last 10-15 years, the need for drainage of the operative bed
following pancreatoduode-nectomy, along with the management of an
operatively placed drain, has been rigorously studied in the surgical
literature. This refinement of drain management (if one is placed) has been
more scien-tifically examined based on modern stratification systems,
including the International Study Group of Pancreatic Fistulas and the Fistula
Risk Score. Similarly, indications for drain removal are now better defined
based on prospective trials. However, the role of intraperitoneal drainage
following distal pancreatectomy fundamentally differs from that of
pancreatoduodenectomy. Left-sided re-sections commonly require less time
and are associated with less blood loss. They also typically involve duct
ligation and parenchymal closure and do not include enteric anastomoses. As
op-posed to Whipple procedures, the remnant gland is nearly always soft
following distal resection. Thus, extrapolating the surgical literature
regarding the need for drainage following pancreatoduo-denectomy to those
undergoing distal pancreatectomy is problematic at best.
The limited studies to date regarding the need for operative bed drainage
following distal pancreatectomy, including Level 1 multi-institutional data
recently reported by Van Buren et al., have failed to demonstrate outcomes
inferior to routine drainage if a drain is withheld at the time of operation. The
counterargument would suggest that while placement of a drain may not
necessarily re-duce morbidity or the need for postoperative therapeutic
intervention, its presence certainly does not lead to negative consequences.
Since no method of remnant management has proven superior following
distal pancreatectomy on a multi-institutional level, the technical aspects of
gland transec-tion and closure are not helpful in guiding the decision to place
a drain or not.
Given the relatively low rate of procedure-related abdominal morbidity
and the ready avail-ability of interventional radiology and endoscopic
management of postoperative pancreatic fluid collections, we feel drains
have no meaningful impact on improving surgical outcomes following distal
pancreatectomy and can be safely withheld. While not scientifically studied,
our impression is that extended resections to the right of the superior
mesenteric/portal confluence rarely develop fistulae or postoperative fluid
collections. Perhaps, in cases when sharp transection of the gland is
performed and the main duct cannot specifically be identified and ligated, a
prophylactic drain might be warranted. Surgeon discretion for drain
placement is vital however, and should never be discouraged or the
placement of a drain disparaged.
If a drain is utilized, we recommend following guidelines for drain
removal as proposed for pancreatoduodenectomy. While various schemas for
drain management exist based on prospective analyses, they typically favor
amylase assessment by postoperative day 3 and drain removal if val-ues are
5000 U/L or less and/or drain effluent is less than 50 mL/24 hours. Whether a
drain is uti-lized or not, the clinical status of the patient should guide any
need for postoperative imaging or therapeutic intervention. As noted in the
preceding case scenario, delayed morbidity is not un-common following
distal pancreatectomy and must be anticipated in the extended postoperative
course.
SPECIAL SITUATION: SUBPHRENIC

FLUID COLLECTIONS
Postoperative fluid collections are not uncommon following distal pancreatic
resection, but many are grade A fistulas or benign postoperative pancreatic
fluid collections. Typically, these are of lit-tle consequence clinically and
may be observed with serial imaging and patient evaluation. How-ever, the
advent of signs and symptoms of early satiety, nausea/vomiting, or sepsis
necessitate in-tervention whether or not an intraperitoneal drain was placed
at the index operation. This may be troublesome when the fluid has collected
in the subphrenic space. Options for treatment include endoscopic (often not
possible due to location) or percutaneous drainage if a window exists,
though occasionally reoperation is required via either an anterior approach
or through the bed of a resected 12th rib. We have occasionally proceeded
with transpleural percutaneous drainage and have experienced few instances
of contamination of the pleural space. Reoperation is commonly needed if a
sequestrum of infected, necrotic debris is present in order to manually
debride the area. A posterior retroperitoneal approach may be particularly
advantageous, especially if remote from the index operation, to allow access
to the infected space while minimizing contamination of the abdominal cavity
as well as mitigating the difficulty of accessing the subphrenic space
anteriorly due to adhesions and inflammation.
CONCLUSION
Current evidence does not support the routine utilization of intraperitoneal
drains following elective distal pancreatectomy. Trials, including Level 1
evidence, suggest similar outcomes whether patients receive a drain or do
not. Ultimately, drain utilization should be at the discretion of the operating
surgeon.
SALIENT POINTS
Routine use of an intraperitoneal draiin after distal pancreatic resection
should not be considered mandatory.
The current literature suggests there is no advantage to intraperitoneal
drainage following elective distal pancreatectomy.
Inability to identify or ligate the main pancreatic duct or a soft remnant
gland may prompt consideration for drain placement.
Postoperative drain management should mimic that following
pancreatoduodenectomy.
SELECTED REFERENCES
1. Levy M. Intraperitoneal drainage. Am J Surg. 1984; 147(3):309-14.
2. Paulus EM, Zarzaur BL, Behrman SW. Routine peritoneal drainage of
the surgical bed after elective distal pancreatectomy: Is it necessary?
Am J Surg. 2012; 204(4):422–7.
3. Behrman SW, Zarzaur BL, Parmar A, Riall TS, Hall BL, Pitt HA.
Routine Drainage of the Operative Bed Following Elective Distal
Pancreatectomy Does Not Reduce the Occurrence of Complications. J
Gastrointest Surg. 2014; 19(1):72–9.
4. Correa-Gallego C, Brennan MF, D’Angelica M, Fong Y, DeMatteo RP,
Kingham TP, et al. Operative Drainage Following Pancreatic Resection.
Ann Surg. 2013;258(6):1051–8.
5. Van Buren G 2nd, Bloomston M, Schmidt CR, Behrman SW, Zyromski
NJ, Ball CG et al. A prospective randomized multicenter trial of distal
pancreatectomy with and without routine intraperitoneal drainage. Ann
Surg. 2017 Sep;266(3):421-31.

PRELUDE
The goal of this chapter is to serve as a decision aid for the HPB surgeon
who, not infrequently, is faced with the challenging clinical scenario of
acting upon intraoperative frozen section(s) (FS) of the transected pancreatic
neck margin during a Whipple’s resection (i.e., pancreaticoduodenectomy).
Questions to be addressed include:
Should a FS from the transected pancreatic neck margin even be sent
routinely? What does the data inform us about its utility, its ability to
change surgical management, and impact disease prognosis/outcome?
How does one interpret positive neck margins in patients demonstrating
response to neoadjuvant therapy?
How important is the role of the pathologist assessing these specimens?
Are there any novel methods of neck margin analysis in the pipeline that
might improve our biological selection of patient subsets that warrant a
more aggressive surgical approach?
This particular decision-making paradigm is encountered during surgery for
two predominant pathologies—pancreatic ductal adenocarcinoma (PDAC)
and intraductal papillary mucinous neoplasm involving the main (pancreatic)
duct (MD-IPMN) with or without a priori knowledge of incident
dysplasia/carcinoma. Since the intraoperative considerations regarding
pancreatic neck margin FSs vary considerably for these two related, but
distinct, entities, they will be discussed separately in this chapter.
CASE SCENARIO
A 77-year-old male with a past medical history of hypertension, stage III
chronic kidney disease, hyperlipidemia, and a history of coronary artery
bypass grafting in 2002, presented with 2 weeks of vague right upper
quadrant fullness and unintentional 10-lb weight loss. Laboratory and
physical examinations were unremarkable. Computed tomography of the
abdomen demonstrated an irregular mass in the head/neck of the pancreas
with abutment, but not encasement, of the superior mesenteric vein, and
sparing of the superior mesenteric artery (SMA). He underwent 6 cycles of
neoadjuvant gemcitabine and nab-paclitaxel with excellent response, as
demonstrated by post-treatment imaging. He underwent exploration and
planned Whipple resection. FS analysis from the intraoperative neck margin
(IONM) demonstrated a focus of residual invasive adenocarcinoma. An
additional margin of approximately one centimeter was resected, which
returned negative for carcinoma, and pancreaticojejunostomy (PJ)
reconstruction was performed to this margin. The patient had an uneventful
recovery and remains disease-free 9 months later.
█ BACKGROUND
Does A Positive Pancreatic Neck Margin Reflect
Disease Biology Or Extent Of Initial Surgical
Resection?
Of the pancreatic margins in a Whipple’s resection (i.e., neck, radial,
SMA/uncinate, retroperitoneal), the IONM is most amenable to surgical
remediation; consequently, investigators have questioned whether the
practice of resecting additional parenchymal neck margin after initially
positive IONM FS is worthwhile, particularly if it converts the resection
from FS:R1 to R0 on permanent section (FS:R1→PS:R0). Although earlier
studies on this subject espoused an aggressive approach to neck margin
clearance in PDAC either by re-resection of the neck margin or via
conversion to total pancreatectomy (TP) in IONM-positive cases, the
preponderance of contemporary data suggest that this practice does not
impact survival.
The largest and most recent analysis of this question is a retrospective
study by Kooby et al. of 1399 PDAC patients undergoing
pancreaticoduodenectomy at 8 US academic medical centers between
January 2000 and August 2012. On FS, 152 patients (10.9%) were R1, and
51 patients (3.6%) had false-negative (FS-R0, PS:R1) margins. PS:R0
margins were achieved in 1196 patients (85.5%), 131 patients (9.3%)
remained PS:R1, and 72 patients (5.1%) were rendered FS:R1→PS:R0 by
additional resection. Median overall survival (OS) for PS:R0 patients was
21.1 months; however, OS for PS:R1 and FS:R1→PS:R0 patients were
similarly dismal (13.7 vs. 11.9 months; P<0.001; Figure 1 ). FS:R1→PS:R0
conversion was associated with a 55% increase in risk-adjusted mortality.
FIGURE 1: Data from Kooby et al. reporting on 1399 patients
demonstrating that patients converted from frozen section R1 to
permanent section R0 (FS:R1?PS:R0) margin have similarly dismal
outcomes compared with patients with R1 on PS (Reprinted with
permission from Wolters Kluwer in Kooby DA, Lad N, Squires M,
Maithel S, et al. Value of intraoperative neck margin analysis
during Whipple for pancreatic adenocarcinoma. A multicenter
analysis of 1399 patients. Ann Surg 2014;260:494-503).
The forerunner to this study was an analysis of 382 patients from Emory
University, some of whom were included in the previously discussed study
by Kooby et al., with similar conclusions—patients rendered FS:R1→PS:R0
and PS:R1 patients had similarly poor OS outcomes (11.3 vs. 11.0 months,
respectively). On multivariate analysis, poor histologic grade, larger tumors,
and positive retroperitoneal margin —but not FS of the IONM—were
independently associated with decreased OS. In a separate report of 202
Whipple’s resections for PDAC from the University of South Florida,
FS:R→PS:R0 (median OS 11 months) and PS:R1 (13 months) patients had
significantly worse OS compared with PS/FS:R0 patients (21 months). In the
discussion, the authors refer to the practice of re-resecting pancreatic neck
margins based on IONM-positivity as “doomed to failure because of
aggressive tumor biology.”
A closer dissection of these analyses underscores the bio logic
underpinnings of this phenomenon. In the multi-institutional study by Kooby
et al. discussed earlier, FS:R1→PS:R0 and PS:R1 patients had similarly
higher rates of larger tumors, nodal positivity, and perineural invasion
compared with PS/FS:R0 patients; moreover, when controlling for
SMA/retroperitoneal margin-positivity (which has confounded interpretation
of results from earlier studies), as well as after excluding patients who had
received neoadjuvant therapy, outcomes in the FS:R1→PS:R0 and PS:R1
cohorts were similarly dismal. These data speak to biologic tumor
aggressiveness in these patient subsets that is unlikely to be mitigated by
extent of surgical resection.
It is worthwhile discussing the role of conversion to TP—if necessary—
to achieve R0 resection in IONM-positive cases, as advocated by Schmidt et
al.; in 61 eligible patients for analysis (PS:R1 n=28 and FS:R1→PS:R0 via
TP n=33), median OS was significantly higher in patients undergoing TP
compared with PS:R1 patients (18 vs. 10 months, P=0.04). This study has
been criticized for failing to disregard the bias inherent in selecting patients
who are more likely to tolerate the morbidity of TP, as well as for its lack of
generalizability to practice patterns beyond high-volume tertiary care
centers. As such, long-term survival data of TP for PDAC derived from
national cancer registries have not reproduced these outcomes, even for R0
resection—in a recent review from the National Cancer Data Base (NCDB),
5-year actuarial survival was 13% and median OS 15.5 months in nearly
2600 patients (1998—2006). Of course, given its retrospective study design
and inability to decipher if TPs were being performed as a priori decisions,
or as a deliberate reaction to serially positive IONMs, these data are
difficult to interpret in the context of the question posed herein. Although
conversion to TP in IONM-positive cases is clearly not standard practice—
or our practice in most cases, for that matter—consideration may be given to
this approach with a critical view on patient fitness for TP, tolerability of the
ensuing exocrine/endocrine insufficiency, and assessment/interpretation of
disease biology, including response to neoadjuvant therapy (if applicable).
Value of IONM Assessment In The Era Of

Neoadjuvant Therapy For PDAC
Neoadjuvant therapy (chemotherapy ± radiotherapy) has been well
entrenched in the management of borderline resectable PDAC for several
years now; an emerging paradigm advocates for its application in resectable
disease as well. The momentum for this approach is, in part, due to the
advent of multi-agent regimens (i.e., mFOLFIRINOX and gemcitabine/nab-
paclitaxel) with improved activity against PDAC; indeed, the Southwest
Oncology Group (SWOG) Trial 1505 will test these two regimens in the
perioperative setting for patients with resectable tumors.
Proponents of the neoadjuvant approach cite early treatment of
micrometastatic disease, avoidance of laparotomy in those with rapidly
progressive disease, the ability to assess in vivo responsiveness and
tolerability of therapy, and possible augmentation of margin-negative
resection rates as purported advantages. Select phase II studies of
neoadjuvant gemcitabine-based therapy for resectable PDAC from Memorial
Sloan-Kettering and M.D. Anderson Cancer Centers have suggested R0
resection rates in excess of 70%; conversely, the Australian phase II AGITG
GAP trial investigating neoadjuvant gemcitabine/nab-paclitaxel in resectable
PDAC demonstrated an R0 rate of 52%, failing to meet the intended R0
endpoint of 85%.
While it remains to be seen if newer cytotoxic regimens will improve R0
resection rates in general, it would not be unreasonable to imagine being
more aggressive in pursuing margin clearance when faced with FS IONM-
positivity in patients thoroughly vetted with neoadjuvant therapy—i.e.,
demonstrating tumor regression, conversion from borderline resectable to
outright resectable disease, or lack of progression to metastatic disease—
suggesting chemoresponsiveness and/or indolent biology. The introductory
case to this discussion represents one such scenario. There are no data to
guide us in this matter; only 10% of patients in the multi-institutional Kooby
et al. study received neoadjuvant therapy, and all underwent R0 resection.
We anticipate this will be an active area of investigation as neoadjuvant
therapy in PDAC becomes more ubiquitously applied.
“Closeness” Of Microscopic Tumor Clearance In

Pancreatic Resections-Can We Do Better?
A provocative, albeit dated, study by the Heidelberg group—titled “most
pancreatic resections are R1 resections”—was conducted on the premise that
recurrence rates after curative-intent PDAC resections are significantly
higher than published R1/R2 rates, and that this discrepancy could perhaps
be explained by underreporting of margin-positivity due to lack of
standardized pathologic protocols. A composite (neck; circumferential soft
tissue including medial, anterior surface, superior, and posterior; bile duct;
and intestinal) margin analysis from 111 consecutive Whipple specimens
(2005-2006) was performed according to an exhaustive standardized
pathologic protocol and compared with non-standardized pathologic
assessments of 188 Whipple specimens (2002-2004); R1 was defined as a
distance of the tumor from the resection margin of ≤1 mm. Incredibly, R1 rate
(not restricted to neck margin alone) in the standardized protocol era was
76% compared with 14% in the non-standardized protocol era.
Given current standards of tumor clearance in other solid tumors (e.g., 5
cm for proximal margin in gastric cancer; at least 2 cm in most non-rectal
gastrointestinal [GI] cancers, etc.), could we be doing better than 24% >1
mm R0 clearance in PDAC? These data strongly argue for improvements in
surgical as well as pathologic quality, particularly since the survival
implication of such imperfect practices is not trivial—an analysis from the
Massachusetts General Hospital (MGH) group revealed that survival in ≤1
mm “R0” resections was similarly bleak compared with R1 resections (16
vs. 14 months, respectively), and significantly lower than ≥1 mm R0
resections (35 months). These data engender further nihilism for the practice
of additional parenchymal resection simply for the sake of “microscopic
clearance”—biologically, anything other than a wide clearance is no
clearance at all—and argues for the liberal use of neoadjuvant strategies to
assess in vivo tumor responsiveness.
On a related note to standardization (or lack thereof) of pathologic
reporting, and its substantial impact on resectional outcomes in PDAC, is the
widely held belief that IONM assessment by subspecialized GI, versus
general, pathologists may improve FS interpretation accuracy. To this end, a
single-institution analysis of IONMs from 77 consecutive Whipple’s
resections compared overall accuracy and interobserver variation of FS
diagnoses between GI and general pathologists. Overall accuracy (93.5% GI
vs. 85.1% general), positive predictive value (78.6% vs. 41.4%) and
specificity (97.8% vs. 87.5%) was higher in FS interpretation by
subspecialized GI, compared with general, pathologists; moreover, GI
pathologists rendered a lower rate of false positive IONM diagnoses
compared with general pathologists (2.2% vs. 12.5%). These results, while
not surprising, are likely a surrogate for the well-documented superiority of
cancer-specific outcomes in PDAC seen at high-volume tertiary care centers
—such institutions are more likely to recruit subspecialty clinicians
(surgeons, pathologists, oncologist, etc.) and provide comprehensive
multidisciplinary care that serve to reinforce these volume-outcome
relationships.
Novel Methods of Detecting IONM

Novel techniques of detecting IONM-positivity, although in their infancy,
may obviate much of the ambiguity and variability encountered in
conventional histopathologic reporting of IONMs. One such method, reported
by the Stanford group recently, utilized molecular assessment by ambient
mass spectrometry to evaluate Whipple neck resection margins. After cross-
validation in training and independent sets, 32 neck margins were examined
prospectively. Interestingly, although in agreement with histopathologic
analysis in 24 (75%) of 32 margin-negative resections, 8 margins (25%)
reportedly negative by histopathology were found to be positive by mass
spectrometry; this discordance translated into an OS discrepancy—median
OS in the spectrometry-positive/histopathology-negative (“R1”) cohort was
10 months, compared with 26 months in “true” R0 (i.e., both spectrometry-
and histopathology-negative) resections. Such techniques, although
promising, require larger scale multi-institutional validation before adoption
into mainstream practice can be recommended.
Suggested Approach
Therefore, in PDAC patients, our practice leans towards recognizing that
disease biology often trumps surgery, acknowledging that IONM is a
surrogate for biologic aggressiveness that is unlikely to be mitigated by
extent of surgical resection. That said, it is still our practice to routinely send
FS of the IONM during Whipple for PDAC; if positive, we perform another
transection of 1-2 cm thickness, and perform our PJ to this margin regardless
of FS result. Other experts in the field, however, have questioned the routine
practice of sending IONM FSs since it has little bearing on long-term
outcome—particularly in an era where cost containment and resource
limitations are a growing concern. In general, the HPB surgical community
agrees that the focus of our efforts should shift towards thoughtful and
evidence-driven patient selection, either by vetting disease biology with
constantly improving neoadjuvant systemic therapies or utilizing
molecular/immune prognosticators of outcome, in order to minimize the
incidence of IONM positivity in the first place.
CASE SCENARIO #2: MAIN DUCT-

IPMN
A 75-year-old BRCA-2 positive female, with previous history of ovarian
and breast carcinoma, presents with abdominal imaging, revealing a dilated
main pancreatic duct (MPD) to 1.1 cm with 10 unilocular cystic lesions in
the pancreatic head (Figure 2 ). She denied prior history of pancreatitis, and
is not currently jaundiced. Endoscopic ultrasound-guided FNA demonstrated
hyperplastic mucinous epithelium with focal papillary change and atypia;
fluid CEA was 746 ng/mL and amylase 1256 U/L. A diagnosis of MD-IPMN
was made. Patients underwent Whipple’s resection; FS of the IONM
revealed high-grade dysplasia (HGD)—a subsequent 1 cm parenchymal
margin was resected and sent, revealing low-grade (LGD) at the revised
margin. PJ was performed at this margin. Pathology revealed MD-IPMN
with HGD in the HOP, and confirmed LGD at the neck margin. Her post-
operative course was uncomplicated, and she remains disease-free 8 months
post-resection.
FIGURE 2: A. Main-duct intraductal papillary mucinous neoplasm
(MD-IPMN) with maximal dilation to 1.1 cm in the pancreatic head
(asterisk), with progressively tapering main duct dilation throughout
neck and body (arrow); B. endoscopic retrograde pancreatography
showing dilated main duct without mural nodules (asterisk).
Should IONM During Whipple For MD-IPMN

Inform Extent Of Resection?
The most recent international consensus guidelines for management of MD-
IPMN—so-called Fukuoka criteria, published in 2012—recommend routine
FS analysis of IONM during Whipple’s resection for MD-IPMN. These
recommendations, derived largely from retrospective single-institution
experiences, suggest that FS of the IONM: a) informs the extent of resection;
b) reliably allow adequate resections in a vast majority of patients; and c)
can safely dissuade prophylactic total pancreatectomy in IONM-“negative”
cases. An analysis of 127 pancreatic resections for IPMN by Couvelard et al.
demonstrated that results of FS of transection margins were confirmed by
definitive examination in 94% of cases; more importantly, FS analysis
changed the extent of resection in 30%, and allowed “adequate” margin-
negative resection in 97% of patients.
The guidelines recommend that additional resection of the pancreas be
attempted if FS of the IONM is positive for HGD or invasive carcinoma, in
order to obtain a “negative” margin. Furthermore, if “exuberant papillary
nodules” are present at the margin, there may be abundant residual tumor in
the remnant pancreas; further resection is indicated in these cases too. As
such, the MGH group reviewed their experience of 223 resected MD-IPMNs
(1990-2013); 5-year OS and disease-free survival (DFS) was 69% and 83%
respectively, and estimated recurrence rate at 10 years was 25%. Excluding
18 TPs, final parenchymal transection margin (n=155) was classified as
negative (i.e., absent for IPMN, or showing only LGD or moderate-grade
dysplasia [MGD]) in 138, or positive (i.e., HGD or invasive carcinoma) in
17. The two most prominent factors independently associated with OS were
presence of invasive component and positive (HGD/invasive) final margin
(both P<0.001); 15-year DFS was significantly higher in margin-negative
compared with margin-positive cohorts (78% vs. 31%).
The Fukuoka guidelines are more ambivalent in situations where IONM
demonstrates LGD or MGD—“further resection is controversial,” and
“presence of lesser grades of dysplasia (moderate or low-grade) may not
require any further therapy.” This issue continues to be debated at national
and international meetings. Proponents of surgical aggressiveness maintain
that total pancreatectomy be applied liberally in such cases, particularly in
younger patients who can not only tolerate the postoperative sequelae of
brittle diabetes and exocrine insufficiency, but also have more years at risk
for progression of disease in the remnant pancreas. The more conservative
alternative—i.e., withholding further resection if IONM reveals LGD/MGD
—may in fact be supported by recent evidence. Shi and Hruban reported a
low recurrence rate in the remnant pancreas, ranging from 0-8%, in patients
with a resection margin positive for LGD/MGD. More recently, Tamura and
colleagues dichotomized transected neck margins in resected MD-IPMNs as
free from (n=31) and positive for (n=16) LGD/MGD IPMN; cumulative DFS
and remnant pancreas recurrences rates were similar between patients with
and without LGD/MGD at the resected margin at initial operation. An
additional analysis from this retrospective review is worth discussing: all
patients recurring in the remnant pancreas (n=7; all with HGD or invasive
carcinoma lesions at initial resection) successfully underwent completion
TP; not only was the cumulative DFS in this cohort equivalent to patients
without remnant recurrence, but it was also significantly higher than patients
with extra-pancreatic recurrence. Based on these data, the authors concluded
that prophylactic TP can be safely avoided in patients with MD-IPMN if
surgical margins are negative or reveal LGD/MGD.
What Are The Adjuncts To Intraoperative Frozen

Section That May Aid Determination Of Pancreatic
Transection Margins?
While FS is the dominant method of IONM assessment in the US, several
adjuncts to FS have been described—mostly from Asian centers—to aid
determination of the final parenchymal transection margin in MD-IPMN.
These include peroral pancreatoscopy using the SpyGlass™ system,
intraductal ultrasonography during endoscopic pancreatography, and
combining intraoperative cytology with FS analysis. Utilization of these
modalities, each with its unique advantages and associated learning curves,
are dependent on regional/institutional expertise, and generalizability across
distinct practice settings is currently unclear.
Suggested Approach
When performing Whipple’s resection for MD-IPMN, in general, we adhere
to Fukuoka guidelines—we: (a) send FS of the IONM routinely; (b) take
additional margin if FS is positive for HGD or invasive carcinoma, but
desist further resection in the setting of negative margin or LGD; (c) consider
TP very selectively in preoperatively informed and medically fit patients if
margins are serially positive for HGD/invasion. Moreover, when PS reveals
HGD or invasive carcinoma at the final neck margin, undetected on FS, we
offer interval completion TP selectively to surgically fit and motivated
patients.
SALIENT POINTS
Although the practice of requesting intraoperative frozen section
analysis of neck margins from Whipple’s specimens in PDAC is fairly
ubiquitous, the data do not clearly support its routine application—
intraoperative conversion of R1 to R0 resection based on neck margin
analysis does not appear to impact long-term outcome; these patients
do similarly poorly compared with patients with definitive R1 margins
on permanent section.
It may be reasonable to consider re-resection of the pancreatic neck to
negative margins in PDAC patients demonstrating response to
neoadjuvant therapy; data guiding decision-making in this scenario are
scant.
In the era of improving anesthetic and perioperative surgical/medical
care, completion total pancreatectomy in serially neck margin-positive
PDAC cases may be considered very selectively in surgically fit
patients, particularly in cases demonstrating response to neoadjuvant
therapy.
Based on the current evidence, frozen section analysis of the neck
margin should be considered standard of care in MD-IPMN
Serial resection (and even completion total pancreatectomy in
appropriate candidates) should be offered to if HGD or invasive
carcinoma is detected at the margin; LGD does not mandate further
resection.
SELECTED REFERENCES
1. Kooby DA, Lad NL, Squires MH 3rd, Maithel SK, Sarmiento JM,
Staley, et al. Value of intraoperative neck margin analysis during
Whipple for pancreatic adenocarcinoma: a multicenter analysis of 1399
patients. Ann Surg. 2014;260 (3):494-501.
2. Schmidt CM, Glant J, Winter JM, Kennard J, Dixon J, Zhao Q, et al.
Total pancreatectomy (R0 resection) improves survival over subtotal
pancreatectomy in isolated neck margin positive pancreatic
adenocarcinoma. Surgery. 2007; 142(4):572-8.
3. Esposito I, Kleeff J, Bergmann F, Reiser C, Herpel E, Friess, et al.
Most pancreatic cancer resections are R1 resections. Ann Surg Oncol.
2008; 15(6):1651-60.
4. Tanaka M, Fernández-del Castillo C, Adsay V, Chari S, Falconi M, et
al. International consensus guidelines 2012 for the management of IPMN
and MCN of the pancreas. Pancreatology. 2012; 12(3):183-97.
5. Marchegiani G, Mino-Kenudson M, Sahora K, Morales-Oyarvide V,
Thayer S, Ferrone C, et al. IPMN involving the main pancreatic duct:
biology, epidemiology, and long-term outcomes following resection.
Ann Surg. 2015; 261(5):976-83.

CASE SCENARIO
A 45-year-old woman presents with left upper quadrant pain. Abdominal CT
scan reveals a 3 cm cystic lesion in the tail of the pancreas (Figure 1 ).
Endoscopic ultrasound with fine needle aspiration demonstrates a unilocular
cystic mass without worrisome features. Cyst fluid analysis reveals elevated
carcinoembryonic antigen (CEA) and low amylase levels. A laparoscopic
distal pancreatectomy with possible splenectomy is recommended for a
symptomatic mucinous neoplasm in the tail of the pancreas. The patient is
concerned about losing her spleen unnecessarily and wants to know why it
would need to be removed.
FIGURE 1: CT scan showing mucinous cystadenoma of the body
of the pancreas
█ BACKGROUND
Distal pancreatectomy is the procedure of choice for benign, premalignant,
and malignant lesions of the body and tail of the pancreas. In recent years
distal pancreatectomy has become a safe procedure that is being performed
more frequently because imaging studies are identifying more benign and
premalignant lesions. For patients with malignant lesions of the pancreas,
splenectomy is recommended to ensure an adequate lymphadenectomy. In
patients with benign lesions, distal pancreatectomy with preservation of the
spleen may be considered.
Spleen preserving distal pancreatectomy (SPDP) was first described by
Mallet-Guy and Vachon in 1943. This classic technique preserves the splenic
artery and vein by identification and ligation of the multiple small, short
vascular connections to the body and tail of the pancreas. In 1988, Warshaw
published an alternative technique in which the splenic artery and vein(s) are
ligated proximally and distally, with retention of the short gastric and left
gastroepiploic vessels to preserve blood flow to and from the spleen). In
subsequent years, the relative risks and benefits of distal pancreatectomy
with and without splenectomy have been studied, as well as method of spleen
preservation (splenic vessel ligation versus splenic vessel preservation).
There are several variables to take into account when considering distal
pancreatectomy with splenic preservation. The objective of this chapter is to
summarize the advantages and disadvantages to preservation of the spleen in
the setting of distal pancreatectomy and to compare methods of splenic
preservation.
SURGICAL TECHNIQUE
Minimally invasive pancreatic surgery has become increasingly common
over the last two decades. With improved laparoscopic technology and
increasing surgeon experience, a minimally invasive approach has been
shown to be comparable to open surgery with respect to morbidity and
mortality, and it is generally agreed upon that minimally invasive surgery is
not inferior to open surgery in terms of operative objective. A discussion of
the merits of surgical technique with respect to laparoscopic, robotic, or
open surgery is beyond the scope of this chapter and will be discussed
elsewhere. Basic considerations include surgeon expertise with each
approach, as well as patient-specific factors including body habitus,
pathology, location of the lesion, etc.
The technique of vessel preserving spleen preserving distal
pancreatectomy (VP-SPDP) is described below (Figure 2 ). The body and
tail of the pancreas are exposed by dividing the gastrocolic omentum, taking
care to stay outside the gastroepiploic vessels. The short gastric vessels
usually do not need to be divided, and every effort should be made to
preserve them if a splenic preserving procedure is being attempted. The
peritoneum is then incised along the inferior pancreatic border, and the
pancreatic body is separated from the retroperitoneum by means of sharp and
blunt dissection along its inferior border. The splenic vein can then be
identified from underneath the pancreas, and careful circumferential
dissection around the splenic vein is performed and a vessel loop placed
around the vein. The splenic artery can be identified from the under surface
of the pancreas by retracting on the vessel loop around the splenic vein, or it
can be identified along the superior border of the pancreas anteriorly. Once it
is dissected circumferentially, it is also controlled with a vessel loop. These
precautionary measures allow quick control of bleeding should a vascular
tear occur later in the procedure. Once the pancreatic body has been
adequately mobilized from the splenic vessels, the pancreatic parenchyma is
divided proximal to the lesion. Once the proximal pancreatic tissue is
divided, the specimen is grasped and gently retracted anteriorly to allow
further dissection of the vessels. The dissection proceeds toward the splenic
hilum in a medial-to-lateral direction. The pancreatic branches of the splenic
vein are sequentially identified, dissected free, and divided with ties, clips,
or an energy device. The branches of the splenic artery, which runs just
superior to the vein, are treated similarly. Special care must be taken as the
dissection approaches the hilum of the spleen.
FIGURE 2: Vessel preserving spleen preserving distal
pancreatectomy
The alternative approach to SPDP is VL-SPDP, which involves dividing

the splenic vessels proximally and distally while preserving the short gastric
and left gastroepiploic vessels to maintain splenic perfusion (Figure 3 ). The
initial steps of this technique are essentially the same as those already
described (see above), up to the division of the pancreas. In the VL-SPDP,
after pancreatic transection, the splenic artery and vein are divided with
ligatures or an endovascular stapler. The left portion of the pancreas is lifted
up and mobilized posteriorly along with the splenic artery and vein, and the
vessels are again divided as they emerge from the pancreatic tail to enter the
hilum of the spleen. The spleen is then supplied solely by the short gastric
and left gastroepiploic vessels.
FIGURE 3: Vessel ligating spleen preserving distal pancreatectomy

(Warshaw Technique)
If an en bloc distal pancreatectomy with splenectomy (DPS) is performed,
the splenic artery and vein are divided after the pancreas is transected. The
distal pancreas is dissected free in a medial-to-lateral direction. The short
gastric vessels are divided with care, taken not to injure the stomach wall.
The retroperitoneal attachments of the spleen and the tail of the pancreas are
then divided freeing the specimen.
LESION PATHOLOGY
Distal pancreatectomy with splenectomy is an operation that adheres to
oncologic surgical principles in that it is an en bloc resection and provides
adequate regional lymphadenectomy. Preserving the spleen necessarily limits
the regional lymphadenectomy, particularly with respect to the lymph nodes
within the splenic hilum, and in cases where the splenic vessels are spared,
lymph nodes along the vessels are potentially left in situ. Thus in cases of
adenocarcinoma or high grade neuroendocrine tumors of the pancreatic body
or tail, distal pancreatectomy with splenectomy is the appropriate operation.
In the setting of benign or premalignant lesions, e.g. intraductal papillary
mucinous neoplasm, mucinous cystic neoplasm, etc., distal pancreatectomy
without splenectomy can be considered.
In cases of chronic pancreatitis, it is our practice to remove the spleen as
our experience shows that patients have inferior postoperative outcomes with
respect to chronic pain when the spleen is spared.
ANATOMIC CONSIDERATIONS AND

LOCATION OF THE LESION
The body and tail of the pancreas are intimately associated with the splenic
vessels. The splenic artery is frequently deeply embedded in the pancreatic
parenchyma at multiple locations along the superior aspect of the gland,
making separation of the gland from the artery technically challenging and
potentially at the expense of significant hemorrhage. Additionally, the splenic
vein is intimately associated with the dorsal surface of the pancreatic body
and it can be difficult or impossible to separate the gland from the vein. This
is especially true in cases with pancreatitis or with lesions abutting the vein.
The branching pattern of the splenic vessels is variable. Vessel preservation
may not be possible in circumstances where the vessels branch relatively
proximally and the pancreatic tail extends into the splenic hilum amongst
theses branches. Likewise, splenic preservation may not be possible if the
lesion is located at the distal extent of the pancreatic tail and abutting the
spleen Ultimately, the decision to perform splenectomy is frequently made
intraoperatively, often with the initial plan to preserve the spleen, but should
that prove to be impossible, or significantly compromise the operation, the
spleen is removed.
SPLENECTOMY VS. SPLENIC

PRESERVATION
Splenectomy has been associated with several adverse consequences in the
setting of pancreatic resections. Shoup and colleagues demonstrated a
significantly higher serious complication rate and infectious complication
rate in patients undergoing distal pancreatectomy, plus splenectomy,
compared with patients undergoing spleen preservation (Table 1 ) in a series
of patients with benign pancreatic lesions. There is evidence to suggest that
splenectomy in any setting is associated with increased risk of early
infectious complications due to suppressed immunity. Splenectomy has also
been associated with thrombocytosis, thrombosis, increased myocardial
infarction risk, and an increased risk in malignancy. The most dreaded
consequence of splenectomy is the development of overwhelming post-
splenectomy infection. The spleen is responsible for immunity against
encapsulated organisms namely Neisseria meningitidis, Haemophilus
influenzae, and Streptococcus pneumoniae. Mortality rates from this
complication range from as low as 10% to as high as 70%. It is somewhat
difficult to truly estimate the risk of overwhelming post-splenectomy
infection because the exact incidence is unknown and, in addition, the
incidence varies according to age at time of splenectomy and indication for
splenectomy (e.g. trauma, hematologic disease). The estimated yearly
incidence following splenectomy is 0.25% per year, and the estimated
lifetime incidence is 5%. Sparing the spleen obviates this risk.
TABLE 1: Distal Pancretectomy With and Without Splenic

Preservation
THE CHOICE OF SPLENIC

PRESERVATION TECHNIQUE
The spleen can be preserved either by sparing the splenic vessels or ligating
the splenic vessels with splenic viability maintained via the short gastric and
left gastroepiploic vessels. The purported advantages of the VL-SPDP
operation are relative technical ease, less blood loss compared to
splenectomy, decreased operative time, and preservation of the spleen. There
is increased risk of splenic infarction (Figure 4 ) with the potential for
development of splenic abscesses and need for subsequent splenectomy. The
vessel ligating technique should be avoided in patients with a large spleen,
as the spleen may not adequately perfuse all of the parenchyma via the short
gastric vessels. There is also the theoretical risk of isolated gastric varices,
owing to increased venous drainage via the short gastric vessels. The short
gastric veins have been noted to be dilated radiographically after the
Warshaw operation consistent with sinistral hypertension. However, reports
of gastrointestinal hemorrhage secondary to this are scarce.
FIGURE 4: CT scan showing splenic infarct and postoperative

pancreatic fluid collection following splenic vessel ligating spleen
preserving distal pancreatectomy
Sparing the splenic vessels is associated with minimal risk of splenic

infarction and need for subsequent splenectomy. There is also decreased risk
of postoperative abscess requiring a drainage procedure or reoperation. The
procedure is technically demanding and is often associated with increased
operative time and bleeding risk; though, if successfully accomplished,
operative blood loss is comparable or less than the VL-SPDP technique or
splenectomy. Postoperative length of stay is often shorter in patients in whom
the spleen is preserved.
It is our experience that postoperative outcomes are similar between
splenectomy and the vessel ligating spleen preserving operation and that
splenic vessel preservation is associated with less postoperative morbidity
compared to either of the former (Table 2 ). Thus it is our preference to spare
the splenic vessels when preserving the spleen; otherwise, a splenectomy is
performed.
TABLE 2: Comparison of postoperative outcomes by distal

pancreatectomy technique
SALIENT POINTS
Spleen preservation should be considered for benign and premalignant
lesions of the body and tail of the pancreas.
Splenic vessel preserving technique is superior to the vessel ligating
technique and splenectomy in terms of postoperative morbidity.
The primary constraints to splenic vessel preservation are vessel
anatomy and lesion location.
All patients should be vaccinated prior to distal pancreatectomy in
anticipation of potential splenectomy.
SELECTED REFERENCES
1. Mallet-Guy P, Vachon, A. (1943). Pancreatites Chroniques Gauches.
Paris: Paris, Masson et cie, 1943.
2. Warshaw AL. Conservation of the spleen with distal pancreatectomy.
Arch Surg. 1988; 123(5): 550-3.
3. Shoup M, Brennan MF, McWhite K, Leung DH, Klimstra D, Conlon
KC. The value of splenic preservation with distal pancreatectomy. Arch
Surg. 2002; 137:164-8.
4. Davidson RN, Wall RA. Prevention and management of infections in
patients without a spleen. Clin Microbiol Infect 2001;7(12):657–60.
5. Beane JD, Pitt HA, Nakeeb A, Schmidt CM, House MG, Zyromski NJ,
et al.
6. Splenic preserving distal pancreatectomy: does vessel preservation
matter? J Am Coll Surg. 2011;212(4):651-7.
7. Jean-Philippe A, Alexandre J, Laurent C, Collet D, Masson B,
Fernández-Cruz L, et al. Laparoscopic spleen-preserving distal
pancreatectomy: splenic vessel preservation compared with the
Warshaw technique. JAMA Surg. 2013;148(3):246-52.

CASE SCENARIO
A 63-year-old female is diagnosed with a pancreatic solid lesion during
evaluation for abdominal pain. She is otherwise healthy, without any prior
surgical or medical history. A pancreatic protocol CT reveals a cT3N0M0
located in the head of the pancreas without evidence of distant metastatic
disease (Figure 1 ). Her overall performance status is excellent, ECOG 0
and ASA class II. The neoplastic pancreatic lesion appears resectable, by
imaging criteria, and she elects to undergo a pancreaticoduodenectomy.
However, at the time of the operation, a 0.8 cm solid lesion is identified in
the left hemiliver (segment 3). You obtain an intraoperative biopsy of the
liver lesion, and the frozen section histopathological analysis reveals an
adenocarcinoma of pancreatic origin.
FIGURE 1: Abdominal computed tomography, venous phase,
demonstrating (A – axial view; B – coronal view) an
hypoattenuating pancreatic head mass measuring 1.9 x 1.7 cm
(solid arrow) with minimal focal stranding that extends to the
superior mesenteric vein, with no other major vessels involved by
the mass; (C– axial view) severe dilation of the pancreatic duct up
to 1 cm – of the body and tail of the pancreas, with parenchymal
atrophy (solid arrow); (D – coronal view) the dilated pancreatic
duct abruptly tapers (solid arrow) at the level of the
hypoattenuating pancreatic head mass. SMV, superior mesenteric
vein; SMA, superior mesenteric artery.
█ BACKGROUND
This is a familiar scenario to many pancreatic surgeons. In fact, despite the
current advancements in imaging modalities, distant metastatic disease is still
diagnosed at the time of surgical exploration in approximately 12% of cases.
If we consider the patient described in the case vignette, and assume no
evidence of metastatic disease at exploration, we are faced with a potentially
treatable disease – perhaps with curative intent. Indeed, the survival rate for
surgically treated PDAC, with negative microscopic resection margins (R0),
is estimated at about 25% at five years (observed median overall survival
[OS] of 22–24 months) with approximately 4% considered to be long-term
survivors at ten years.
The presence of metastatic disease (stage IV) dramatically reduces the
expected OS of patients with PDAC, which is estimated between 2 and 6
months, without treatment.
Furthermore, despite significant improvements in modern
chemotherapeutic regimens, such as FOLFIRINOX or gemcitabine plus nab-
paclitaxel, the expected OS offered by these regimens, in patients with stage
IV PDAC, remains an unsatisfactory median of 11 and 8.5 months,
respectively.
Once again, let us consider the patient in our case vignette. She is brought
to the operating room with the expectation of having a complete resection of
her PDAC and a prospect of a 25% chance of being alive at five years;
however, with the discovery of the metastatic lesion, she suddenly becomes a
member of a whole different group with a far shorter life expectancy.
It has been established that complete surgical removal of PDAC is the
only chance of a cure, but its role in the setting of stage IV PDAC has not
been well investigated, especially in the era of modern chemotherapy. The
surgeon’s struggle becomes even more apparent when dealing with a locally
resectable PDAC that presents concomitantly with a completely resectable
metastatic lesion, as is illustrated in our case vignette. Therefore,
considering the recent success of resecting metastatic disease from other
solid organ malignancies such as colorectal cancer, sarcoma, and gastric
cancer, it is only natural to query if there is a role for metastatectomy in the
setting of PDAC.
One of the proposed advantages of resecting metastatic disease – if both
the primary lesion and the metastases can be completely resected – relies on
the concept that reducing the tumor burden would allow for a more
efficacious action of systemic chemotherapy; de facto permitting the
chemotherapy and the immune system to concentrate solely on undetected
microscopic disease. However, it remains unknown if the above stated
concept applies to PDAC once the tumor has already metastasized; thus,
extrapolating data from other cancers might not be possible nor advisable.
A downside of performing a major pancreatic resection and a
concomitant liver resection is represented by the potential mortality and
morbidity that can arise by combining both procedures. Although pancreatic
resection mortality has decreased significantly during the last decades – now
estimated around 5% for most centers and < 1% in high-volume pancreatic
centers – the morbidity remains a high 40-60% for all comers. Morbidity
following pancreatic surgery for PDAC often leads to delays in initiation of
systemic chemotherapy, changes in regimen or dosages, or, at its worst,
omission of systemic treatment altogether. In the setting of stage IV PDAC,
the inability to tolerate or the omission of systemic chemotherapy can have
severe and far-reaching consequences commonly resulting in a significant
decrease of life expectancy.
The current national and international guidelines for the treatment of
PDAC do not recommend resection of the primary tumor and synchronous
liver metastases (Tempero et al. 2014). Nonetheless, in the literature there
are isolated reports or case series where such an approach has been
undertaken. However, the available results are inconsistent and often in
conflict with each other, thus leaving an open question as to whether
complete resection of the PDAC with combined resection of liver metastases
will lead to a survival benefit.
Hackert et al. published the largest series on the subject: 128 patients
undergoing PDAC and metastases resection, between 2001 and 2014, of
whom 85 patients had liver oligometastatic disease and the remaining 43
patients had distant aortocaval lymph node metastases. In their study, 86% of
patients with liver metastases underwent a synchronous atypical resection,
while formal hepatic resection was reserved for only 14% of the patients,
and in most cases, it was carried on in a metachronous fashion. Most of the
resected lesions (72%) had a diameter measuring less than 2 cm and the
overall number of metastases in the resected specimen was up-to-three in
96% of the cases. The authors reported a 30-day surgical morbidity and
mortality of 45% and 3%, respectively, after resection of synchronous M1
tumors. These figures changed to 21.7% and 4.3%, respectively, after
resection of metachronous lesions. The median survival for patients
undergoing liver resection (synchronous or metachronous metastatectomy)
was 12.3 months with an estimated 5-year survival of 8%.
Data regarding the use of postoperative chemotherapy were available in
only 74% of patients, of which the majority (79.5%) received gemcitabine-
based regimen. It is worth noting that the 12-month median survival,
described in this study, is certainly superior to the 6- to 7-month median
survival reported with single-agent gemcitabine, yet this result is comparable
to the reported 11-month survival with modern FOLFIRINOX regimen,
therefore adding little benefit to the palliative systemic chemotherapy
approach in terms of overall survival.
Tachezy et al. reported on the experience of six large European pancreas
centers with resection of the primary PDAC and synchronous liver
metastases, between June 1994 and July 2014, focusing on the operative and
oncologic outcomes of such an approach. The authors identified 69 patients
who underwent synchronous pancreatic and liver resection, for PDAC, and
compared their outcomes to a matched cohort of patients, with liver
metastatic PDAC, undergoing surgical exploration without tumor resection.
The results of this study are particularly interesting as the authors reported a
definitive survival advantage for patients undergoing synchronous surgical
resection–when PDAC was localized in the pancreatic head–compared to
patients undergoing surgical exploration without resection (median OS 13.6
vs 7 months, P < .001). Though, in their study, the survival advantage offered
by surgical resection was no longer present when the primary PDAC was in
the body/tail of the pancreas (median OS 14 vs 15 months, P = .312).
However, no definitive explanation for this behavior was evident. In contrast
to the study conducted by Hackert et al. (previously described in this
chapter), the patients described by Tachezy et al. underwent solely non-
anatomic liver resections (atypical wedge resections). Once again, the data
regarding post-operative adjuvant therapy was available only in 83% of the
patients, and of these, 80% received adjuvant therapy mainly in the form of
single-agent gemcitabine (70% and 65% for the resected and nonresected
group, respectively), with a minority being treated with FOLFIRINOX (7%
and 6% for the resected and nonresected group, respectively). Of note, the
authors reported a significant increase in perioperative morbidity in the
resected group compared to the nonresected group (68% vs 48%, P = .025);
however, the reoperation rate and the 30-day mortality were similar between
the groups. Furthermore, 5-year survival was 0% in the nonresected group,
while 4 of 69 evaluable patients (5.8%), after combined resection, were still
alive after 5 years. The results of this study, although encouraging, must be
considered with caution due to the retrospective nature, the limited number of
patients described, the long-time interval spanning over two decades, and the
lack of definitive selection criteria for the allocation of patients in the
resection vs. nonresection group, among others. Nevertheless, the study
shows that synchronous resection, in the setting of metastatic PDAC to the
liver, is safe and feasible and is associated with acceptable morbidity and
mortality.
On the other end of the spectrum, several studies have argued against
surgical resection in patients with stage IV PDAC, suggesting that such an
aggressive approach is associated with minimal to no survival advantage –
compared to palliative systemic treatment alone – at a cost of increased
morbidity and potential mortality.
Takada et al. presented the first study assessing the survival benefit of
surgical resection of stage IV PDAC (with isolated liver metastases)
compared to palliative by-pass surgery. The authors reported on 33 patients,
between the years 1981 and 1995, who were diagnosed with PDAC and
synchronous liver metastases at the time of laparotomy. A total of 11 patients
in this group underwent a pancreaticoduodenectomy combined with
partial/wedge hepatic resection and the remaining 22 patients underwent a
palliative by-pass procedure (n=22). The size of the resected liver metastatic
lesions ranged from 0.5 to 5 cm, and the number of resected lesions per
patient ranged from 1 to 7. In the resected group, the median survival was 6
months and the longest survival was 10 months; in the palliative by-pass
group, the median survival was 4 months and the longest survival was 6
months. Moreover, the author reported a perioperative mortality of 9% in
both groups. Takada et al. observed no improvement in OS between the two
groups, but noted higher surgical morbidity and mortality rates in the group of
patients undergoing pancreatoduodenectomy with synchronous liver
resection. Furthermore, all 11 patients in the resected group died from
multiple recurrent liver metastases within 10 months of the resection. The
high perioperative mortality rate reported in this study appears excessive
compared to modern series, but it is representative of the study period
(1985–1991). Furthermore, the authors do not report on the use of systemic
chemotherapy in either of the two groups, thus adding additional limitations
to the interpretation of this study’s results.
Another example arguing against the benefit of surgical resection in the
setting of stage IV PDAC is offered by Gleisner et al. The authors described
a cohort of 17 PDAC patients on whom synchronous pancreatic resection and
hepatic metastatectomy were performed. The majority of patients, in their
cohort, underwent a nonanatomic liver resection (wedge resection), and most
patients selected for resection had a solitary hepatic metastasis, with a
median size of 0.6 cm. Of the 17 patients with PDAC who underwent
synchronous resection, only 6 received adjuvant chemotherapy. Ultimately,
the authors concluded that the OS in patients undergoing hepatic resection of
synchronous metastasis was not different from the OS of matched patients
who underwent palliative bypass (5.9 vs. 5.6 months, P=0.46).
Understanding the complex biology of PDAC is of the utmost importance,
especially when dealing with systemic disease. There is a concrete risk in
PDAC that visible resectable disease represents only the tip of the iceberg,
and undetected microscopic disease might already be present, in the liver
parenchyma or elsewhere, at the time of surgical resection. Furthermore, the
aggressive biology of PDAC, characterized by short tumor-doubling time
interval and by its well-known metastatic propensity, might make resection of
detectable metastatic disease futile without an effective mechanism to control
the occult microscopic disease and prevent its rapid resurgence.
The adoption of surgical resection, in stage IV PDAC, will require the
establishment of effective patient-selection strategy and the development of
effective tumor markers able to pre dict the tumor response to existent
systemic chemotherapeutic strategy.
Indeed, some patients might benefit from complete resection in stage IV
PDAC; however, the characteristics of this group remain for the most part
unknown and recommendations for routine resection in stage IV PDAC are,
at this stage, unsupported by reliable data.
Nonetheless, a true opportunity exists to evaluate the effect of complete
surgical resection, especially in the setting of more efficacious systemic
treatment as represented by FOLFIRINOX and nab-paclitaxel, a topic that
has not been fully investigated. The results of a clinical trial could
potentially provide further insight to guide patient-selection strategies and
provide valuable data on the efficacy of modern chemotherapeutic agents in
the setting of resected stage IV PDAC.
To conclude our case-vignette: based on the available data in the
literature and in accordance with the current guidelines, surgical resection
should not be performed. Nevertheless, the surgeon should assess the utility
of a palliative by-pass procedure (gastrojejunostomy with or without an
hepaticojejunostomy) to prevent or resolve a gastro/duodenal or choledochal
obstruction caused by a potentially enlarging pancreatic mass. In this
situation, every effort should be made to minimize morbidity and recovery
time, inevitably caused by the laparotomy, to allow for prompt initiation of
systemic therapy within the shortest interval from surgical exploration.
SALIENT POINTS
Only 10% of newly diagnosed PDAC presents with resectable
localized disease, 40-50% presents with locally advanced or
borderline resectable disease, and the remaining 50-60% of cases
presents with distant metastatic disease (ranging from an isolated liver
lesion to widely diffused disease).
Despite the current advancements in imaging modalities, distant
metastatic disease is still diagnosed at the time of surgical exploration
in approximately 12% of cases.
The current national and international guidelines for the treatment of
PDAC do not recommend surgical resection for stage IV disease,
without differentiating single isolated metastasis vs. diffuse disease.
Retrospective studies, which evaluate the role of surgical resection in
the setting of liver oligometastatic PDAC, consist of isolated case
reports or small case series. Their results are often discordant, patient
selection criteria are inconsistent, and the use of systemic
chemotherapy varies and mostly entails single agent gemcitabine.
The role of surgical resection, in the setting of stage IV PDAC with
oligometastatic liver disease, should be further investigated, and a
reliable answer to this controversial topic, in the era of modern
systemic chemotherapy, can only be generated through a randomized
control trial.
SELECTED REFERENCES
1. Gleisner AL, Assumpcao L, Cameron JL, Wolfgang CL, Choti MA,
Herman JM, et al. Is resection of periampullary or pancreatic
adenocarcinoma with synchronous hepatic metastasis justified? Cancer.
2017; 110(11): 2484–92.
2. Hackert T, Niesen W, Hinz U, Tjaden C, Strobel O, Ulrich A, et al.
Radical surgery of oligometastatic pancreatic cancer. EJSO 2017;
43(2): 358-63.
3. Tachezy M, Gebauer F, Janot M, Uhl W, Zerbi A, Montorsi M, et al.
Synchronous resections of hepatic oligometastatic pancreatic cancer:
Disputing a principle in a time of safe pancreatic operations in a
retrospective multicenter analysis. Surgery. 2016; 160(1): 136–44.
4. Takada T. Yasuda H, Amano H, Yoshida M, Uchida T. Simultaneous
hepatic resection with pancreato-duodenectomy for metastatic
pancreatic head carcinoma: does it improve survival?
Hepatogastroenterology. 1997; 44(14): 567–73.
5. Tempero MA, Malafa MP, Berhman SW, Benson AB 3rd, Casper ES,
Chiorean EG, et al. Pancreatic Adenocarcinoma, Version 2.2014:
featured updates to the NCCN guidelines. JNCCN. 2014; 12(8): 1083–
93.
CASE SCENARIO
A 65-year-old man presents with painless jaundice and is found to have a
locally advanced pancreatic ductal adenocarcinoma (PDAC) in the head of
the pancreas (Figure 1 ). After placement of a metallic biliary stent, the
patient is enrolled in a neoadjuvant protocol, which includes 8 cycles of
FOLFIRINOX (5-FU, oxaliplatin, and irinotecan) followed by 28 treatments
of chemoradiation for a total of 50.4Gy, boosted to 58.8Gy around involved
blood vessels. Following a good radiologic response, the patient is taken to
the OR for a pancreaticoduodenectomy. Intraoperatively, the patient is noted
to have residual tumor involving the superior mesenteric artery.
Intraoperative biopsy confirms adenocarcinoma deeming the tumor
unresectable.
FIGURE 1: Computed axial tomography scan of a patient with
locally advanced pancreatic ductal adenocarcinoma (PDAC) in the
head of the pancreas with near circumferential involvement of the
█ BACKGROUND
INTRODUCTION
Pancreatic cancer is a challenging disease to treat and cure. Approximately
80% of patients present with distant, or locally advanced unresectable,
disease. Of those who undergo surgical resection with curative intent, 5-year
overall survival remains low, in the order of 20%. Without treatment,
expected survival is short, ranging between 4 and 6 months. With
combinations of modern systemic agents (such as Gemcitabine/Abraxane or
FOLFIRINOX), with or without radiation therapy, and surgical resection,
median survival rates of up to 44 months have now been reported.
Improvements in multidisciplinary care with delivery of systemic therapies,
better treatment of associated toxicities, and optimization of surgical
technique and postoperative care have accounted for the considerable
improvements in survival following treatment of pancreatic cancer, including
in the palliative setting.
In this chapter, we focus on palliative treatment of locally advanced
unresectable disease. It is important to note that pancreatic cancer may be
considered unresectable because of either distant disease (lungs, liver, or
peritoneum), or locally advanced unresectable disease. The distinction is
potentially important, as implications for palliative therapy may differ based
on local extent of the tumor and anticipated progression of local disease.
In general, practice has shifted away from performing palliative
interventions at the time of discovering unresectable disease during surgical
exploration. Improvements in systemic regimens and interventional
techniques are thought to account for this paradigm shift. However, this is a
generalized observation, since locally advanced disease differs significantly
in its potential to cause biliary and/or gastrointestinal obstruction, compared
with small tumors that metastasize early, where the need for any palliative
intervention may be diminished.
We focus our discussion on a case of locally advanced unresectable
disease of the pancreatic head, where impending biliary and gastric outlet
obstruction may be present or anticipated, as an example of a periampullary
malignancy commonly encountered in the authors’ practice. With continued
improvements in the longevity of patients treated with palliative intent (e.g.
use of FOLFIRINOX in the palliative ACCORD-11 trial), the potential for
obstruction may be further increased. While substantial shrinkage of tumors
may be achieved, radiographic changes are difficult to assess. Therefore,
operative exploration may still be commonly performed. Given the potential
obstructive nature of locally advanced periampullary malignancies, practical
knowledge of surgical palliative techniques remains an important tool in the
armamentarium of the versatile pancreatic surgeon, who should be involved
early in the work up and management of all patients with localized pancreatic
cancer. This chapter addresses biliary bypass, gastric bypass, treatment of
pain, locally ablative procedures, and palliative pancreatectomy in the
management of locally advanced periampullary malignancies.
BILIARY BYPASS
In patients with periampullary cancer presenting with obstructive jaundice,
endobiliary stent placement offers non-operative placement of durable stents
with rapid recovery and low complication rates. Between endoscopic
techniques and percutaneous transhepatic access, cases in which access to
the bile duct is completely precluded are rare. However, endobiliary stent
placement and biliary instrumentation is still associated with complications,
particularly infectious morbidity, and often associated with need for repeated
interventions and hospital readmissions. Metallic biliary stents are now used
routinely in preoperative patients, including for prolonged periods in patients
undergoing neoadjuvant therapy.
Jaundice is the most common presenting symptom of patients who present
with periampullary cancer. Jaundice can lead to debilitating pruritus,
malnutrition, liver dysfunction, and death if not addressed. While endoscopic
management of malignant biliary obstruction is considered the gold-standard
modality in relieving jaundice, a role for surgical biliary bypass still exists
in patients who present with biliary obstruction who undergo surgical
exploration. The exact indications are difficult to define. For example, when
endobiliary stents have been placed preoperatively for jaundice, either with
or without neoadjuvant therapy, the surgeon may still choose to retain the
indwelling endobiliary stent and forego surgical bypass, particularly if a
self-expanding metallic stent is in place. However, the superiority and
durability of surgical bypass compared with endobiliary stenting is
indisputable, as even metallic stents can obstruct. There are no consensus
data by pancreatic surgeons to favor one modality over the other. On the
other hand, when exploration in a jaundiced patient is performed without a
biliary stent or even with a plastic stent, most experienced pancreatic
surgeons would favor a biliary bypass unless precluded for technical
reasons.
In cases where endobiliary stents are technically impossible to place (e.g.
following prior gastric-bypass or gastrectomy, or with duodenal obstruction),
the decision to perform surgical palliative bypass is an easier one, as
percutaneous access, while still reliable, frequently involves an external
drainage component initially and is associated with a higher complication
rate than endoscopic placement. Finally, in some cases, final determination
of resectability cannot be accomplished without dividing the bile duct,
thereby requiring biliary reconstruction.
The authors’ recommendation, based on the extensive experience of the
senior author with this scenario, is that if a jaundiced patient with a stent in
place undergoes exploration via laparotomy and is found to have a locally
unresectable cancer, then surgical palliative bypass is indicated, with stent
removal. If, however, the patient is found to be unresectable from a smaller
tumor that has metastasized within the abdomen, or when disease is detected
during staging laparoscopy, reliance on postoperative endobiliary stenting
may be justified. If in either of these scenarios a plastic stent is in place,
conversion to a metal stent can be performed at a later date. Finally,
regardless of tumor size, performance of a biliary bypass in a patient without
existing jaundice would seem unwarranted.
Multiple options exist for bypass of the biliary tree, which can be
performed via open or minimally-invasive techniques. Although Roux-en-Y
hepaticojejunostomy or choledochojejunostomy are most commonly
performed, other options include gallbladder bypass (cholecystogastrostomy,
cholecystoduodenostomy, or cholecystojejunostomy), as well as alternative
common bile duct or hepatic duct bypass options, such as
choledochoduodenostomy or hepaticoduodenostomy. These options are
discussed below.
CHOLECYSTOENTERIC BYPASS
The gallbladder provides the simplest conduit for enteric bypass of the
biliary tree, particularly given its enlargement in cases of obstructive
jaundice (Courvoisier’s sign). With simple mobilization of the gallbladder
fundus from the cystic plate, routine anastomoses, either handsewn or
stapled, to the stomach or duodenum may be carried out. However, such
procedures are largely of historic interest, due to either proximal extension
of the tumor into the bile duct or involvement of the proximal gastroduodenal
tract with tumor. Cholecystojejunostomy, however, is an equally simple
procedure that avoids these complications, requires no gallbladder
mobilization, and can also be safely performed using hand-sewn or stapled
techniques (Figure 2A-B ). The jejunal loop is brought up either in an
antecolic or retrocolic fashion, and a Roux limb or a Braun
enteroenterostomy is created to divert enteric contents from refluxing into the
gallbladder (Figure 2C-D ).
FIGURE 2: Cholecystojejunostomy performed using hand-sewn or
stapled techniques (Reprinted with permission from Lippincott
Williams & Wilkins and Wolters Kluwer Health. In Lillemoe K,
Jarnagin W (eds). Master Techniques in Surgery-Hepatobiliary and
Pancreatic Surgery. Philadelphia 2013).
While gallbladder bypass is simple, additional elements are critical for

safe and durable bypass to occur, including lack of involvement of the cystic
duct juncture by an obstructing cancer, the presence of a relatively distended
gallbladder, and the presence of free-flowing bile into the gallbladder, all of
which may be confirmed intraoperatively with a cholecystogram, as needed.
Thus, as stated, procedures employing the gallbladder as a conduit for biliary
drainage are now rarely employed by experienced pancreaticobiliary
surgeons.
CHOLEDOCHOENTERIC AND
HEPATOENTERIC BYPASS
Use of the common bile duct or common hepatic duct as biliary conduits are
usually favored by most pancreatic surgeons, due to both long term reliability
and comfort with the procedure. Even large tumors are unlikely to involve
the proximal biliary tree, rendering these procedures safe and effective for
biliary drainage. A choledochoduodenostomy or hepaticoduodenostomy are
simple bypasses that can be performed following Kocherization of the
duodenum, which permits duodenal mobilization towards the duct. A side-to-
side or end-to-side anastomosis is subsequently performed, often with
concomitant cholecystectomy (Figure 3 ). The procedure is simple and safe,
although complications including bile reflux, cholangitis, and sump syndrome
may occur. In sump syndrome, the accumulation of bile debris in the
defunctionalized distal common bile duct can lead to significant symptoms,
including pain and cholangitis, and revision with a Roux-en-Y
hepaticojejunostomy may eventually be required. Once again, encroachment
of tumor on the proximal gastroduodenal tract precludes this type of bypass.
FIGURE 3: A choledochoduodenostomy or hepaticoduodenostomy

performed as an end-to-side anastomosis (Reprinted with
permission from Lippincott Williams & Wilkins and Wolters Kluwer
Health. In Lillemoe K, Jarnagin W (eds). Master Techniques in
Surgery-Hepatobiliary and Pancreatic Surgery. Philadelphia 2013).
The authors favor Roux-en-Y hepaticojejunostomy as their preferable
biliary enteric bypass (Figure 4 ). The anastomosis is similar to that
performed routinely during reconstruction after pancreaticoduodenectomy, is
distant from the proximal gastroduodenal tract, and is usually distant from the
point of biliary obstruction. While a jejunal loop bypass is possible, the risk
of associated cholangitis due to enteric reflux may be problematic, and
formation of a Roux-en-Y limb for the anastomosis is preferred. Variations
on formation of the hepaticojejunostomy exist; the authors perform a single
layer of interrupted anterior and posterior row 4-0 polydioxanone (PDS)
sutures placed radially from the jejunum to the hepatic duct, following
transection of the duct and oversewing of the distal portion.
FIGURE 4: Roux-en-Y hepaticojejunostomy is the authors’

preferable biliary enteric bypass and is performed using a single
layer of interrupted anterior and posterior row 4-0 polydioxanone
(PDS) sutures placed radially from the jejunum to the hepatic duct
following transection of the duct and oversewing of the distal
portion (Reprinted with permission from Lippincott Williams &
Wilkins and Wolters Kluwer Health. In Lillemoe K, Jarnagin W
(eds). Master Techniques in Surgery-Hepatobiliary and Pancreatic
Surgery. Philadelphia 2013)..
Alternative surgical biliary bypasses, such as segment 3 biliary bypass,

are rarely necessary for periampullary tumors.
GASTRIC BYPASS
Unlike biliary bypass, locally advanced tumors that result in gastric outlet
obstruction in the author’s opinion are best addressed with surgical bypass
procedures for long-term efficacy, given the overall lack of reliability of
endoscopically-placed stents that frequently migrate, perforate, or cannot be
technically placed. Furthermore, due to the more frequent need for
reintervention, plus the inability tolerate a full range of oral intake, the
authors believe that duodenal stenting is unlike endobiliary stenting in its
reliability and durability to provide long-term palliation. Even when
unresectability is detected during staging laparoscopy, including for
metastatic disease without gastric obstruction, gastrojejunostomy can be
reliably addressed using minimally-invasive techniques by most pancreatic
surgeons, a scenario which is clearly different from biliary bypass. Finally,
in older data that examined the role of surgical gastric bypass in patients
without established gastric obstruction, prophylactic gastrojejunostomy in
randomized-controlled trials and subsequent meta-analysis demonstrated a
reduced incidence of subsequent gastric outlet obstruction, without adversely
affecting short-term outcomes. Although these earlier data did not provide
direct comparisons with endoscopically placed stents, and while the
incidence of gastric outlet obstruction continues to be debated
(approximately 20%), the authors’ threshold to perform gastric bypass is
lower than for biliary bypass, particularly as patients treated with palliative
intent exhibit increased longevity with modern systemic treatment regimens.
Surgical gastric bypass is usually simple, and there is little variation in
outcomes seen between the different anastomotic techniques, including
retrocolic vs. antecolic, and isoperistaltic vs. antiperistaltic loop
gastrojejunostomy (Figure 5 ). Our preferred approach is a retrocolic
isoperistaltic configuration, which is thought to best avoid encroaching
malignant obstruction. Both authors favor a two-layer hand-sewn
anastomosis, although single-layer and stapled anastomoses may also be
safely performed.
FIGURE 5: A stapled surgical gastric bypass anastomosis is shown

using an antecolic or retrocolic loop gastrojejunostomy (Reprinted
with permission from Lippincott Williams & Wilkins and Wolters
Kluwer Health. In Lillemoe K, Jarnagin W (eds). Master Techniques
in Surgery-Hepatobiliary and Pancreatic Surgery. Philadelphia
2013).
When considering combined biliary and gastric “double” surgical bypass,

procedure complexity increases slightly, owing to the numerous options
available for reconstruction. While double loops can be utilized to perform
bypass, the authors’ preference is to utilize a Roux-en-Y biliary anastomosis
as described above, and a simple loop gastrojejunostomy using the
pancreaticobiliary limb, which should be slightly longer at the point of the
jejunal transection (20-40cm instead of 10-15cm), in order to facilitate
formation of the loop gastrojejunostomy anastomosis prior to the
jejunojejunostomy.
PAIN MANAGEMENT
Pain control for locally advanced unresectable pancreatic cancer is an
important cornerstone in the palliative treatment algorithm. Pain is
classically central epigastric with radiation to the back, and gnawing pain
frequently affects patients during sleep at night, when it may be initially most
noticeable. Quality of life may eventually be significantly hindered due to
unremitting severe pain that disrupts activities of daily living. The disruption
to life can be significant, and the efficacy and side effects of narcotic
medications is often problematic. While endoscopic and percutaneous
techniques may be utilized to provide celiac plexus blockade, these
procedures are highly operator-dependent and are limited by access and
visualization.
Although the pendulum has swung away from doing more intraoperatively
for patients found to have locally advanced unresectable disease, Level 1
evidence from a prospective randomized-controlled trial from the Hopkins
group showed that prophylactic intraoperative celiac axis blockade is
indicated, given the benefit attained in preventing the development of pain,
with minimal additional risk or morbidity. Using 20ml of 50% alcohol
injections, the plexus, which lies behind the hepatic and splenic arteries, is
chemically ablated (Figure 6 ). The vessels can be palpated with relative
ease during surgical exploration and without the need for extensive
dissection. With tactile feedback, the ability to perform repeated injections
effectively illustrates the superior efficacy of this simple, quick, and cheap
modality over any endoscopic or percutaneous approach.
FIGURE 6:
While celiac plexus blockade is frequently sufficient to control, delay, or

prevent the development of severe disease-related pain, this procedure may
not always work, and additional consideration may be given to alternative
procedures, such as thoracic splanchnicectomy. Although this technique has
historically been reported with mixed results for patients with chronic
pancreatitis, thoracoscopic intervention to alleviate pain should be
considered in cases of unremitting pain despite celiac axis blockade.
LOCALLY ABLATIVE THERAPIES

Several experimental, locally ablative, modalities have been evaluated in the
treatment of locally advanced disease, including alcohol ablation,
cryoablation, radiofrequency ablation, irreversible electroporation, and
intraoperative radiation therapy (IORT). While these therapies do not
represent the standard of care in the treatment of patients with locally
advanced unresectable disease, the authors have experience with use of
IORT in this setting, which is therefore included briefly herein. In a study at
the Massachusetts General Hospital, investigators evaluated the use of IORT
in the era of intensive neoadjuvant chemotherapy and chemoradiotherapy for
PDAC. In a retrospective analysis of 68 patients with locally advanced
unresectable, or borderline-resectable, disease between 2010 and 2015, 41
patients underwent successful resection, 18 had locally advanced
unresectable disease, and 9 were found to have distant metastases. In patients
who underwent resection, median survival was 35.1 months for patients who
underwent concomitant IORT for close margins on frozen section analysis vs.
24.5 months for those who underwent resection alone, although this
difference was not statistically significant. Importantly, however, IORT was
extended to 17 of 18 patients with locally advanced unresectable disease,
with a resultant median survival of 24.8 months, and with reports of long-
term survivors who are still alive from that study period. IORT was
associated with increased hospital stay (4 vs. 3.5 days), but without
adversely affecting operative times or morbidity. While the authors believe
use of IORT in the treatment of locally advanced unresectable PDAC remains
experimental in nature, its use may be recommended in facilities with
appropriate expertise to deliver this highly specialized strategy in the event
locally advanced unresectable disease is encountered intraoperatively, given
these favorable data from the authors’ institution.
PALLIATIVE PANCREATECTOMY
At this point in time, there is no role for prophylactic palliative
pancreatectomy, neither in the locally advanced unresectable setting, nor in
the presence of metastatic disease. In the former, prior reports exist on the
role of palliative (R2) pancreaticoduodenectomy and distal pancreatectomy
for locally advanced unresectable disease. However, in many of these older
reports, it was hard to distinguish whether authors truly embarked on
resection with curative intent, which were subsequently resected with
grossly positive margins. Given that there are no data to suggest benefit from
tumor debulking in pancreas cancer, pancreatectomy is not indicated.
More recently, the role of pancreatectomy has resurfaced in the metastatic
setting, where patients with low volume disease that has remained stable on
modern systemic agents are selectively considered to have exceptionally
favorable tumor biology. In this setting, reasons such as providing
‘chemotherapy holidays’ or improving survival have been used to justify
resection. It is important to note, however, that such reports are anecdotal
and do not represent the current standard of care. Given that the majority of
surgical morbidity is related to the pancreatic anastomosis and/or distal
pancreatic stump, the authors strongly recommend against palliative
pancreatectomy outside of experimental study protocols at this time.
SALIENT POINTS
If a jaundiced patient has undergone exploration via laparotomy and
found to have a locally advanced unresectable cancer, then palliative
surgical biliary bypass in the form of a Roux-en-Y
hepaticojejunostomy should be considered.
Given the high incidence of gastric outlet obstruction and the reduced
reliability of endoscopically-placed duodenal stents for long-term
palliation of locally advanced disease, loop gastrojejunostomy (open
or laparoscopic) is indicated at the time of exploration for most
patients with unresectable pancreatic cancers.
Prophylactic surgical celiac axis blockade is beneficial in preventing
the development of pain associated with locally advanced disease with
minimal risk or morbidity.
IORT for locally advanced unresectable disease appears to be
associated with improved survival in centers that possess technical
expertise in use of this locally ablative modality.
SELECTED REFERENCES
1. Lillemoe KD, Cameron JL, Hardacre JM, Sohn TA, Sauter PK,
Coleman J, et al. Is prophylactic gastrojejunostomy indicated for
unresectable periampullary cancer? A prospective randomized trial.
Ann Surg. 1999; 230(3): 322-8.
2. Lillemoe KD, Cameron JL, Kaufman HS, Yeo CJ, Pitt HA, Sauter PK.
Chemical splanchnicectomy in patients with unresectable pancreatic
cancer. A prospective randomized trial. Ann Surg 1993; 217(5): 447-
55.
3. Reid-Lombardo KM, Barton J, Sarr MG. (2013). Operative palliation
of pancreatic cancer. Master Techniques in Surgery: Hepatobiliary and
Pancreatic Surgery. By Lillemoe KD and Jarnagin WR. 1st Edition.
Philadelphia, PA. Lippincott Williams & Wilkins.
4. Bhutani N, Cheadle GA, Bahr MH, Vitale GC. Comparative efficacy of
bilateral thoracoscopic splanchnicectomy for intractable pain secondary
to pancreatic cancer vs. chronic pancreatitis. J Am Coll Surg 2017;
224(4): 566-571.
5. Conrad C, Lillemoe KD(2013). Palliative therapy for pancreatic cancer.
Current Surgical Therapy. By Cameron JL and Cameron AM. 11th
Edition. Philadelphia, PA. Elsevier Saunders.

CASE SCENARIO
A 62-year-old female presented with abdominal pain and jaundice. Upper
digestive endoscopy plus endoscopic ultrasound showed a 1.9 cm nodular
lesion in the duodenal ampulla (Figure 1 ) confined to the mucosa and no
apparent lymph node involvement. An endoscopic ampullectomy was
performed (Figure 2 ). Final pathologic assessment revealed a well-
differentiated tubulovillous adenoma with foci of noninvasive high-grade
dysplasia. All margins were free of adenomatous involvement. Twenty
months later patient presented with jaundice and a new endoscopy showed
local recurrence with 1.5 cm extension into the distal bile duct. A
pancreatoduodenectomy was performed. Final pathology revealed a 1.8 cm
ampullary adenocarcinoma T1N1M0.
FIGURE 1: Endoscopic view of a 1.9 cm ampullary mass.
(Courtesy of Dr. Ernesto Quaresma Mendonça. Department of
Endoscopy. Hospital das Clinicas, University of São Paulo).
FIGURE 2: Final aspect of the endoscopic ampullectomy. A plastic
stent was temporarily placed inside the bile duct to prevent
obstruction due to local edema. The biopsy revealed tubulovillous
adenoma with foci of noninvasive HGD. (Courtesy of Dr. Ernesto
Quaresma Mendonça. Department of Endoscopy. Hospital das
Clinicas, University of São Paulo).
█ BACKGROUND
The ampullary region is the anatomical area that includes the papilla of Vater,
the confluence of the main pancreatic duct, distal common bile duct, and the
sphincter of Oddi. Although uncommon, the ampullary region can harbor
benign and malignant neoplasms. They represent roughly 5% of all digestive
tract tumors.
The diagnosis of ampullary tumors may be incidental or due to the
investigation of digestive symptoms, such as jaundice, abdominal pain,
bleeding and/or acute pancreatitis episodes. The decision of whether a
patient should be treated or observed must be based on the type of tumor and
the presence of symptoms. Therapeutic options are serial endoscopic
surveillance, local resection (LR), or pancreatoduodenectomy (PD).
Endoscopic surveillance may be indicated for patients with asymptomatic
benign tumors without risk of malignant transformation.
Local resection (LR) can be performed either by a surgical approach
(transduodenal ampullectomy, TDA) or an endoscopic approach (endoscopic
ampullectomy, EA). LR may be considered for patients with benign tumors,
for small tumors that do not harbor carcinoma, and for poor surgical
candidates. LR presents lower complication and mortality rates but can be
associated with a high risk of local recurrence and inadequate resection
margins. EA has the disadvantage of not performing lymphadenectomy, and
its application to tumors at risk of metastases should be carefully considered.
TDA allows lymphadenectomy and node sampling in the peri-pancreatic
region.
PD is the treatment of choice for larger lesions, invasive lesions, and
malignant tumors with risk of lymph node involvement. PD follows
oncological principles, makes possible a radical resection with adequate
lymphadenectomy, and allows for correct staging of the disease. Morbidity of
the PD is around 40% and mortality can reach 6%. Often, these scenarios are
at a significant risk for postoperative pancreatic fistula (POPF), given that
they are often performed in the setting of a soft pancreatic parenchyma and a
normal to small, unobstructed main pancreatic duct. All PD candidates
should be clinically evaluated and fit for major surgery.
LOCAL EXCISION OR PD FOR

AMPULLARY MASSES?
The decision whether to apply LR or PD to treat an ampullary mass should
be based on the histological type of the tumor, extension of the tumor into the
local papillary area and ducts, risk of local recurrence, and risk of lymph
node involvement. A correct pre-operative diagnosis is important in order to
avoid unnecessary radical resection.
DIAGNOSTIC PROCEDURES
Occasionally the diagnosis of ampullary lesions is incidental in
asymptomatic patients, but in most cases it is identified during investigation
of upper gastrointestinal symptoms.
Endoscopy: Allows the ability to diagnose if the lesion is in the mucosa,

submucosa, or intra- mural layer, as well as if ulceration is present. Forceps
biopsies have sensitivity and specificities varying from 45% to 80%, and a
sphincterotomy may be necessary to increase sensitivity.
Endoscopic ultrasound (EUS): Plays an important role in the diagnosis of

ampullary lesions. It allows more precise characterization of the lesion,
whether it is nodular or infiltrative, degree of mucosa and submucosa
invasion, extension within the distal bile duct or pancreatic duct, lymph node
and vascular involvement. EUS-guided fine needle aspiration biopsy
increases the accuracy of pathological diagnosis and should be used
whenever possible.
ERCP: May help in demonstrating extension to bile duct, need for

sphincterotomy, duct cytology, and stenting, if necessary. However, this
should be applied with caution given the significant chance of inciting acute
pancreatitis, which may negatively affect the ability to treat the tumor
adequately.
Computed Tomography Scanning (CT) or Magnetic Resonance Imaging
(MRI): Can identify local or regional lymph node or distant metastases, as
well as vascular or adjacent organ involvement as well.
TYPE OF TUMOR
Nonadenomatous, Benign Tumors
Rarely seen in the ampullary area, lipomas, hamartomas, fibromas,
leiomyomas, and adenomyomas are benign tumors whose malignant potential
has not yet been established in the literature. For individuals with clinical
suspicion or pathological diagnosis of benign tumors, indication for surgical
resection will depend on the presence of symptoms.
In asymptomatic patients, close endoscopic observation is recommended.
For patients with symptoms of biliary obstruction, mechanical duodenal
obstruction, or pancreatitis, limited surgery is recommended. Endoscopic
sphincterotomy, EA, or TDA can be indicated in this situation. Whenever
possible the less-invasive procedure should be chosen. Pathological
examination of the surgical specimen should confirm the preoperative
diagnosis. Figure 3 suggests an algorithm for such benign ampullary lesions.
FIGURE 3: Management of Benign Ampullary Lesions EA:
Endoscopic Ampullectomy; TDA: Transduodenal Ampullectomy; PD:
Pancreatoduodenectomy
Adenomas
Similar to colon cancer, ampullary adenocarcinoma originates from an
adenoma-to-carcinoma sequence in a rate of about 30%. Complete resection
of ampullary adenomas is the goal for curative intent.
EA has been indicated for small adenomas (<3.0 cm) confined to
ampullary region, those not invading the muscularis propria, and those with
intraductal invasion < 1.0 cm, but without malignancy or evidence of lymph
node involvement. Recurrence rates range from 0% to 30%. However, a
correct preoperative diagnosis is not accurate. Endoscopy biopsies of
ampullary adenoma show false negative results for malignancy in between
11% and 60%. In contrast, the presence of high-grade dysplasia (HGD) and
ductal dilatation are associated with malignancy in ampullary adenoma. EUS
and MRI are valuable tools to define T and N stages.
Surgical resection (TDA or PD) is indicated when ampullary adenomas
do not fit the criteria for EA. Even in early-stage ampullary adenocarcinoma,
metastatic lymph nodes are present in 40% of cases and disease-free
survival is significantly longer after PD when compared to TDA. PD must be
the first-line treatment of choice for every patient in good clinical condition.
Recently a systematic review comparing surgical resection (TDA or PD)
and EA for ampullary adenoma showed surgical resection to be superior to
EA for complete primary resection, primary success, and recurrence
outcomes. There was no difference regarding complication rates between the
two groups. Figure 4 provides a framework of thought for the management of
ampullary adenomas.
FIGURE 4: Management of Ampullary Adenomas EA: Endoscopic

Ampullectomy; TDA: Transduodenal Ampullectomy; PD:
Pancreatoduodenectomy
Neuroendocrine Tumors
Ampullary neuroendocrine tumors (NET) are rare. They can be aggressive
and present high rates (35-45%) of lymph node metastasis at diagnosis.
Unfortunately, no clinical or pathological factors can accurately predict the
absence of nodal involvement.
Local excision (EA or TDA) for ampullary NETs is an inadequate
oncological resection because it may understage the disease and is
associated with high recurrence rates. For localized ampullary NETs, PD
should be considered the treatment of choice for patients with good
performance status. Figure 5 depicts an algorithm for managing this entity.
FIGURE 5: Management of Ampullary Neuroendocrine Tumors:

EA: Endoscopic Ampullectomy; TDA: Transduodenal Ampullectomy;
PD: Pancreatoduodenectomy
* According to institutional cancer protocol
Malignant Tumors / Carcinoma

PD is the treatment of choice for localized ampullary carcinomas regardless
of ampullary, biliary, or pancreatic origin. Radical resection is the only
chance for cure and should be chosen as the first option for all patients with
satisfactory clinical indications for surgery.
Neoadjuvant therapy (chemotherapy or chemoradiotherapy) should be
considered in cases of borderline or locally advanced tumors, but this is
more infrequently the case than it is with pancreatic adenocarcinoma.
Jaundice control can be performed either by endoscopic biliary stenting or
bilioenteric anastomosis. PD can be performed in the circumstance of
clinical or radiological response.
When metastatic disease is present, the treatment aims are palliative and
should include the biliary and/or gastric bypass (endoscopic or surgical) and
chemotherapy. In rare exceptions, PD may be indicated in cases of bleeding
tumors that do not respond to non-invasive treatments. Figure 6> indicates
how one might approach management of malignant tumors of the ampullary
area.
FIGURE 6: Management of Malignant Ampullary Tumors

PD: Pancreatoduodenectomy
SALIENT POINTS
A correct preoperative diagnosis is important to avoid unnecessary
radical resection.
For benign ampullary tumors, local resection can be recommended as
the first option treatment for symptomatic patients.
For ampullary adenomas, the decision on LR or PD should be based on
tumor size, presence of malignance, and length of extension to bile or
pancreatic duct. Endoscopic approach is associated with significantly
higher recurrence rates.
For localized ampullary NETs, PD should be considered the treatment
of choice in well-conditioned patients.
PD is the treatment of choice for localized ampullary carcinomas
regardless of ampullary, biliary, or pancreatic origin.
SELECTED REFERENCES
1. Baptiste GG, Postlewait LM, Ethun CG, Le N, Russell MC, Kooby DA,
et al. Is There an Optimal Surgical Approach to Neuroendocrine
Tumors of the Ampulla? A Single Institution Experience Over 15 Years.
Am Surg.2016;82(7):637-43.
2. Dixon E, Vollmer CM Jr, Sahajpal A, Cattral MS, Grant DR, Taylor BR,
et al. Transduodenal resection of peri-ampullary lesions. World J Surg.
2005;29(5):649-52.
3. Mendonça EQ, Bernardo WM, Moura EG, Chaves DM, Kondo A, Pu
LZ, et al. Endoscopic versus surgical treatment of ampullary adenomas:
a systematic review and meta-analysis. Clinics (Sao Paulo).
2016;71(1):28-35.
4. Panzeri F, Crippa S, Castelli P, Aleotti F, Pucci A, Partelli S, et al.
Management of ampullary neoplasms: A tailored approach between
endoscopy and surgery. World J Gastroenterol. 2015; 21(26):7970-87.
5. Song J, Liu H, Li Z, Yang C, Sun Y, Wang C. Long-term prognosis of
surgicaltreatment for early ampullary cancers and implications for local
ampullectomy.BMC Surg. 201;15:32.

CASE SCENARIO
An 80-year-old man with a history of diabetes, hypertension, obesity and
CAD presented with abdominal pain and jaundice. CT revealed a 1cm mass
in the head of the pancreas. There were no contraindications to resection.
Biopsy confirmed adenocarcinoma. Cardiology deemed the patient
“moderate risk” for a perioperative cardiac event and recommended holding
Plavix. At surgery, operative time was 6 hours, EBL <400 mL, the pancreatic
duct was 5 mm, the pancreas was firm, and you were confident about your
pancreatic anastomosis. Should you leave a drain?
█ BACKGROUND
Drains have traditionally been placed at the time of pancreatic resection to
evacuate pancreatic juice, bile, or lymphatic fluid that may accumulate in the
operative bed. They may also serve to herald post-pancreatectomy
hemorrhage. However, drains have been shown to be unnecessary or even
detrimental in other operations, such as splenectomy, gastrectomy, and
colectomy. In a recent global survey of experienced pancreatic surgeons,
14% reported they never leave drains and 27% use drains selectively. This
is mainly due to concern that drains, particularly in patients without
clinically relevant post-operative pancreatic fistula (CR-POPF), may serve
as a portal of entry for bacteria and increase complications. Also, surgeons
reason that an image-guided drain can be placed postoperatively if needed.
Until recently, the literature on this topic was limited and contradictory.
The need for routine placement of drains at the time of pancreatic
resection was first questioned by Jeekel in 1992. Since then, multiple
retrospective cohort studies and three randomized clinical trials have
addressed this question (Table 1 ). All of the cohort studies have shown
either no difference or a decreased overall complication rate with
elimination of routine drainage. However, these studies are retrospective,
small, single-institution cohort studies, which are susceptible to multiple
sources of bias. Changing of institutional practices and evolution of surgical
technique over time are not controlled. Most importantly, these studies lack
randomization, introducing a great deal of bias and potential confounding
differences between the two treatment groups.
TABLE 1: Studies assessing outcome of pancreatectomy
with and without drains
A randomized controlled trial by Conlon et al. evaluated the use of drains

in 179 subjects (drain, n=88; no drain, n=91). There was no significant
difference in the number or type of complications between the two groups
(drain 63% vs no drain 57%). The authors noted no difference in 30-day
mortality. The authors did report an increased incidence of intraabdominal
abscess and fistula formation in the group randomized to drain placement (p
≤ 0.02). The interpretation of the study is confounded by combining patients
undergoing both pancreaticoduodenectomy and distal pancreatectomy. Also,
the study was performed at only one center, and the follow up was short—
only 30 days.
A dual-center randomized controlled trial focused on
pancreaticoduodenectomy randomized 395 patients to drain or no drain
following pancreaticoduodenectomy (PD). The conclusion was that
elimination of routine drainage was not inferior to routine drainage. In-
hospital mortality, re-intervention (re-laparotomy or radiologic intervention),
and overall morbidity was similar in each group. Clinically relevant
postoperative pancreatic fistula (CR-POPF) was reduced (11.9% drain;
5.7% no drain; P=0.030). The screening and randomization process in this
study has been criticized because 3200 patients were eligible for enrollment
at the 2 centers, suggesting a screening bias. Also drains were placed in 21%
of patients who were allocated to the no drain group, suggesting surgeons
deviated from the protocol in higher risk situations, resulting in additional
selection bias.
In contrast, a multicenter randomized controlled trial by Van Buren et al.
found that the absence of intraoperatively placed drains was associated with
an increase in the number of patients that had at least one ≥ grade 2
complication (drain 52% vs no drain 68%; p= 0.047) and a higher
complication severity (p=0.027). In addition, mortality increased from 3 to
12%, indicating that the routine elimination of drains in all patients
undergoing PD is associated with unacceptably inferior outcomes. Patients
without intra-operatively placed drains who developed a pancreatic fistula
were promptly treated with percutaneous drainage but their outcomes were
still inferior to patients who had a drain placed at the time of resection. This
study included only pancreaticoduodenectomy patients, was conducted at 9
high-volume pancreas specialty sites, and patients were followed more
rigorously; 60 days for morbidity and 90 days for mortality. There is a linear
association between fistula risk factors and mortality following
pancreaticoduodenectomy. However, when drains are omitted, this
relationship becomes exponential. A drain may not prevent the occurrence of
a CS-POPF, but we believe it mitigates the risk of morbidity and mortality
when one occurs.
Most experienced pancreas surgeons would agree that leaving a drain at
the time of pancreaticoduodenectomy is wise in the presence of risk factors
for CR-POPF. We think it is very clear that failing to place a drain at the time
of resection in a patient that ultimately develops a CR-POPF will result in
very inferior outcomes, including more deaths. This ultimate and
unacceptable bad outcome must be balanced against the risk of leaving a
drain. What really is the down side of leaving a drain in a setting where CR-
POPF is less likely, such as cases with a firm pancreas and dilated
pancreatic duct? Leaving a drain is very unlikely to result in a mortality. Yes,
data suggests that a drain left >5 days in patients who are low risk may
increase the risk of a CS-POPF. However, if the drain is removed early, this
potential harm caused by the drain itself is eliminated. Selective use of
drains after PD should not be justified by an effort to eliminate risk caused
by a drain, because any perceived increased risk for a drain is likely
minimized or eliminated by early removal when drain amylase
concentrations and the patient’s initial postoperative course support an
uncomplicated recovery.
SUGGESTED APPROACH
Returning to our case, the risk of a fistula in this patient is negligible, but the
patient is an octogenarian with multiple co-morbidities, making the risk of
mortality higher in the presence of a CR-POPF. The wise course, in our
opinion, is to leave a drain at the time of resection, monitor the patient
closely in the early postoperative period, and, if the patient clinically
appears well and the drain amylase concentration is low, remove the drain
early. Otherwise, leave the drain in place, as it will mitigate the bad
outcomes associated with a CR-POPF.
SALIENT POINTS
If a patient develops a CR-POPF, absence of a drain placed at the time
of pancreaticoduodenectomy is associated with increased morbidity
and a 4-fold increase in mortality.
Placing a drain at the time of pancreaticoduodenectomy is very unlikely
to ever cause an operative mortality.
In patients who have few risk factors, and have demonstrated in the
early postoperative period by their clinical course and drain amylase
concentration to be at lower risk for CR-POPF, early drain removal
will decrease potential harm caused by routine drainage.
SELECTED REFERENCES
1. McMillan MT, Malleo G, Bassi C, Sprys MH, Vollmer CM Jr. Defining
the practice of pancreatoduodenectomy around the world. HPB
(Oxford). 2015; 17(12): 1145–54.
2. Jeekel J. No abdominal drainage after Whipple’s procedure. Br J Surg.
1992 ;79(2):182.
3. Heslin MJ, Harrison LE, Brooks AD, Hochwald SN, Coit DG, Brennan
MF. Is intra-abdominal drainage necessary after
pancreaticoduodenectomy? J Gastrointest Surg. 1998, 2 (4): 373-8.
4. Conlon KC, Labow D, Leung D, Smith A, Jarnagin W, Coit DG, et al.
Prospective randomized clinical trial of the value of intraperitoneal
drainage after pancreatic resection. Ann Surg. 2001, 234 (4): 487-93.
discussion 493–4.
5. Fisher WE, Hodges SE, Silberfein EJ, Artinyan A, Ahern CH, Jo E, et
al. Pancreatic resection without routine intraperitoneal drainage. HPB
(Oxford). 2011, 13 (7): 503-10.
6. Adham M, Chopin-Laly X, Lepilliez V, Gincul R, Valette PJ, Ponchon T.
Pancreatic resection: drain or no drain? Surgery. 2013; 154 (5): 1069-
77.
7. Correa-Gallego C, Brennan MF, D’Angelica M, Fong Y, Dematteo RP,
Kingham TP, et al. Operative drainage following pancreatic resection:
analysis of 1122 patients resected over 5 years at a single institution.
Ann Surg. 2013, 258 (6): 1051-8.
8. Van Buren G, Bloomston M, Hughes SJ, Winter J, Behrman SW,
Zyromski NJ, et al. A Randomized Prospective Multicenter Trial of
Pancreaticoduodenectomy With and Without Routine Intraperitoneal
Drainage. Ann Surg. 2014, 259 (4): 605-12.
9. Witzigmann H, Diener MK, Kienkötter S, Rossion I, Werner B, Pridöhl
O, et. Al. No Need for Routine Drainage After Pancreatic Head
Resection: The Dual-Center, Randomized, Controlled PANDRA Trial.
Annals of Surgery; 2016 (264)3:528-37.
10. McMillan MT, Vollmer CM Jr, Asbun HJ, Ball CG, Bassi C, Beane
JD, et al. The Characterization and Prediction of ISGPF Grade C
Fistulas Following Pancreatoduodenectomy. J Gastrointest Surg.
2016;20(2):262-76.
11. Bassi C, Molinari E, Malleo G, Crippa S, Butturini G, Salvia R, et al.
Early versus late drain removal after standard pancreatic resections:
results of a prospective randomized trial. Ann Surg. 2010;252(2):207-
14.

CASE SCENARIO
A 58-year-old male is referred for pancreatic ductal adenocarcinoma
(PDAC). Appropriate staging evaluation reveals a resectable 2.5cm mass in
the pancreatic head (clinical stage IB). Prior to referral, a plastic biliary
stent was placed for hyperbilirubinemia. At operation,
pancreaticoduodenectomy was performed. A soft pancreas with a 2mm duct
was encountered and reconstructed. During transection of the bile duct, 5mL
of purulent bile trickled into the operative field. Biliary continuity was
restored via choledochojejunostomy, performed by your trainee under close
supervision and appeared sound, but a leak test necessitated redoing the
anastomosis. The trainee queries you regarding necessity of operative
drainage.
█ BACKGROUND
The surgical objective of pancreaticoduodenectomy (PD) has remained
constant since its description by Whipple, Kausch, and others in the first half
of the twentieth century – extirpation of the pancreatic head, distal common
bile duct, and duodenum followed by reestablishment of gastrointestinal
continuity. However, few procedures in the practice of general surgery have
evolved to the same degree as PD – nearly every aspect of the procedure and
its perioperative adjuncts have been scrutinized. It is vital to recall that in the
1970’s and 1980’s PD was nearly abandoned at many centers worldwide.
Then, as now, PD was the only curative option for malignancy of the
pancreatic head. The dismal overall cure rate, even with successful resection
coupled with a persistently high surgical mortality in excess of 20% at that
time, greatly limited the utility and indications for PD. Moreover, major
morbidity and prolonged inpatient hospitalization were considered the norm.
It is one of the great success stories of general surgery over the past thirty
years that PD and pancreatic surgery at large are now performed with
mortality rates of less than 3% and for an ever-growing list of indications,
including removal of pre-cancerous lesions. The transformation has been
multi-factorial. The introduction and relative standardization of more
effective surgical techniques by luminaries such as Cameron, Warshaw,
Blumgart, and Brennan was crucial, as was the establishment of surgical
societies dedicated to the field and the proliferation of high-volume
pancreatic surgeons and centers. Perhaps equally important, advances in both
diagnostic and interventional radiology greatly enhanced the pancreatic
surgeon’s ability to manage diverse causes of postoperative morbidity.
Improvements in anesthesia, nursing, and critical care have been important
contributors as well.
Operative drainage has been a constant component of PD since its outset.
Historically, drainage modalities have been multiple and varied. Biliary T-
tubes, pancreatic duct drains, and feeding gastrostomy or jejunostomy have
all been described as routine components of PD in the past; all are now
rarely employed. Currently, operative drains following PD are generally
understood to be one or two intraperitoneal drains placed at the time of
surgery in the region of the pancreatic and/or biliary anastomoses. Such
drains remain in regular employ, and descriptions of their placement,
management, and criteria for removal are included in most modern surgical
texts and atlases. The utility of such drains and the necessity of their usage
following PD is a topic of great controversy amongst pancreatic surgeons.
Morbidity following PD can mirror standard complications observed
following any major abdominal surgery, such as surgical site infections or
ileus. In addition, complications that are specific to pancreatic surgery that
occur after PD include delayed gastric emptying, postoperative pancreatic
fistula (POPF), and post-pancreatectomy hemorrhage. Of these, delayed
gastric emptying is troublesome but rarely results in prolonged incapacity.
Hemorrhage, particularly late extra luminal hemorrhage, although quite rare,
can be deadly to the patient if not recognized and managed expeditiously.
Conversely, POPF is a frequent occurrence and is defined by international
consensus as amylase concentration from any measurable volume of
intraabdominal fluid greater than three times serum amylase concentration on
or after postoperative day three. POPF, as shown in Table 1 , are classified
by severity as grade A (no clinical impact, biochemical diagnosis only),
grade B (requiring readmission or specific treatment such as percutaneous
drainage or antibiotics), and grade C (requiring reoperation or associated
with sepsis, organ dysfunction, or death). The profound improvements in
operative mortality and the widening indications for pancreatic resection
have resulted in an explosion of studies exploring means to improve surgical
morbidity following PD. Accordingly, a vast array of interventions have been
examined, attempting to mitigate the occurrence and severity of POPF,
including changes in surgical technique, pharmaceuticals such as
somatostatin analogues, adhesives, stents, and others. Remarkably, very few
of these myriad efforts have resulted in widespread practice changes due to
cost concerns, ineffectiveness, or feasibility. Notwithstanding this lack of
consensus, compared to a few short decades ago when the merits of
pancreatic resection were openly questioned, it is encouraging to see modern
clinical research questions focus on such granular details as the utility of
operative drains and their impact on outcomes.
TABLE 1: Diagnosis and classification of postoperative

pancreatic fistulas following pancreatic surgery by the
International Study Group on Pancreatic Fistulas (Adapted
with permission from Elsevier in Bassi C, et al. The 2016
update of the International Study Group (ISGPS) definition
and grading of postoperative pancreatic fistula: 11 years
after. Surgery 2017;161(3):584-591).
RATIONALE FOR ROUTINE DRAINAGE

Operative drains allow evacuation of fluids that collect in the surgical bed.
Any blood, bile, pancreatic secretions, or other fluids are removed and
readily assessed. From a practical standpoint, surgeons utilizing operative
drains after PD usually employ two drains placed at the conclusion of the
case. One drain is placed adjacent to the pancreatic anastomosis and the
other adjacent to the biliary anastomosis. Surgeons who routinely place only
a single drain usually place it adjacent to the pancreatic anastomosis.
Removal of the drain(s) is considered in uncomplicated cases once the
patient has resumed a regular diet and lack of leakage has been confirmed.
The amylase concentration of the pancreatic drain is usually assessed prior
to its discontinuation.
By allowing the surgeon to monitor the status of the peritoneal cavity after
the conclusion of surgery, several apparent advantages are conveyed. The
benefit most frequently cited by surgeons who regularly employ drains after
PD is early diagnosis. Dehiscence of the biliary, or gastrojejunal
(duodenojejunal, in the case of pylorus-preserving PD) anastomoses, could
be detected by the presence of bile or succus within the drain. Both potential
complications are concerning due to the sepsis that usually follows
uncontrolled intraabdominal leakage of bile or gastrointestinal contents. In
the case of biliary leakage, a well-placed drain may serve as a controlled
fistula and accelerate healing if the leak can be managed non-operatively.
Similarly, presence of a drain may facilitate early diagnosis of extra luminal
post-pancreatectomy hemorrhage.
Of the three anastomoses constructed during PD, the pancreatic
anastomosis is the most common source of leakage and early diagnosis of
POPF formation is a compelling reason for drain placement. As mentioned
above, POPF frequently complicates PD, with high-volume surgeons still
reporting POPF in 7-18% of cases. While many pancreatic surgeons use the
terms POPF and pancreatic leak interchangeably; POPF is recognized as a
defined, distinct complication (Table 1 ). This international consensus
definition of POPF, first described in 2005 and recently reinforced in 2017,
relies on recognition of amylase-rich peritoneal fluid after postoperative day
three, and routine drain placement allows measurement of amylase
concentration in the drain effluent. Importantly, simple biochemical leaks of
amylase-rich fluid (originally termed grade A POPF), which are clinically
silent and diagnosed only via biochemical assay, can only be identified in
patients with operative drains. Patients with an evolving POPF frequently
demonstrate an increase in volume of drain output following initiation of an
oral diet, which may further increase diagnostic suspicion. In patients that
develop a POPF, surgeons that routinely place drains infer that the drain may
serve to create a controlled fistula and prevent the abdominal sepsis and
patient decompensation that associated with high-grade POPF. Certainly,
these are laudable goals as there are ample data demonstrating that
development of a high-grade POPF increases cost, prolongs length of stay,
and delays the administration of adjuvant systemic therapy.
RATIONALE FOR OMITTING ROUTINE

DRAINAGE
The first series of cases in which PD was performed without operative
drainage was reported in 1992. Given the accepted motivations for routine
drainage described above, the notion of forgoing drain placement was quite
novel. However, several developments in the 1990’s and 2000’s coalesced
that motivated some pancreatic surgeons to consider omitting drains in
selected patients. Better surgical techniques increased confidence in both
pancreatic and biliary anastomotic construction. Closer monitoring in
specialized, high-volume centers and ICU care facilitated earlier recognition
of the subtle signs of POPF formation or other complications. Perhaps most
importantly, improvements in interventional radiology greatly decreased the
number of ‘failure to rescue’ scenarios and unplanned returns to the operating
room. With these developments, the number of critical management decisions
that relied upon data solely derived from the surgical drain declined.
Certainly, drains provide data, but circumstances in which that data proved
decisive were becoming uncommon. Importantly, surgeons experimenting
with omission of operative drains recognized that conditions existed in
which some patients undergoing PD would still benefit from drain placement.
Hence, selective rather than total omission of drains was the norm.
In considering the consequences of omitting routine drainage, the
supposed benefits from drainage enumerated above must be examined,
especially with regard to anastomotic leakage. Leakage from the
gastrojejunostomy (or duodenojejunostomy, in the case of pylorus-preserving
PD) is a serious complication but occurs rarely, owing to the durability and
excellent blood supply of the stomach and intestinal walls. Similarly,
disruptions of the biliary reconstruction are concerning, but also uncommon.
Amongst experienced pancreatic surgeons, both are rare events (combined
incidence of less than 2% of cases in our experience), particularly leakage of
the gastro-/duodenojejunostomy. Moreover, they usually occur in when
extenuating surgical factors are known to exist, such as poor quality tissue for
the biliary anastomosis, and therefore are somewhat predictable.
Pancreatic anastomoses are unique due to the nature of the pancreatic
parenchyma; in patients with operative drains placed following PD, the
majority will be found to have an increased amylase concentration in their
drain effluent on postoperative day one, even though most will not develop a
POPF. As described above, if such amylase-rich drain effluent persists
beyond postoperative day three without negative clinical effects, it will be
termed a biochemical leak only, or class A POPF in the original international
consensus definitions (see Table 1 ). These will only be apparent in patients
with drains; omission of drainage in this group of patients is often cited as
being harmful by advocates of routine drainage. However, this notion pre-
supposes that undrained class A POPF will evolve into clinically-relevant
class B or C POPF. There is no direct evidence of this phenomenon. In a
study seeking to validate the utility of a risk metric for POPF development,
we examined 20 patients in a cohort of 259 that underwent PD without
operative drains (POPF incidence of 7.7%) who later developed a
clinically-relevant POPF. Fever, abdominal pain, and leukocytosis were the
most common presenting signs or symptoms and 90% were diagnosed
following percutaneous drainage of an amylase-rich fluid collection
observed on an imaging study after presentation. Therapy was augmented
with parenteral nutrition and antibiotics as needed; two patients required
laparotomy (classified as grade C POPF). These data demonstrate that
clinically-relevant POPF present, and are managed, in an almost-identical
fashion irrespective of drain placement. In other words, even in patients with
operative drains, clinically relevant POPF frequently require secondary
drainage and occasional return to the operating room. Consequently,
supporters of selective drain placement believe that class A POPF truly has
minimal clinical relevance and that class B or C POPF can be recognized
and treated just as successfully in patients without operative drains.
A unique consideration involves late extra luminal post pancreatectomy
hemorrhage. A possible etiology of these bleeding events after PD, which are
rare but can lead to large-volume blood loss, is dissolution of the vessel
wall and ligatures of the transected gastroduodenal artery from pancreatic
enzyme-rich secretions collecting in the operative bed. Theoretically, this
accumulation of secretions could be mitigated by a surgical drain, which
could also facilitate efficient diagnosis by the presence of blood in the drain
effluent. In practice, these events are very uncommon. Operative drains may
not be positioned to collect the extravasating blood or may be clogged by
thrombus, creating a false sense of confidence by improperly implying lack
of bleeding. Finally, these hemorrhage events have been clearly demonstrated
in patients with operative drains as well. The key to post pancreatectomy
hemorrhage management after PD remains maintenance of high clinical
suspicion with early angiography and embolization or covered stent
placement as needed.
A final consideration is the possibility of potential harm from drain
placement. As with all surgical drains, those placed after PD create a conduit
for microorganisms to access the intraperitoneal cavity. This risk seems to be
reduced by the usage of closed suction drains in several trials comparing
them to open silicone (Penrose-type) drains, although not decisively so. The
duration of drainage seems to increase the risk of complications. Asian and
European studies have examined randomized patients who underwent PD
with mandatory drain placement and were randomized to early
(postoperative day three or four) or late drain removal. Those with delayed
drain removal had a higher incidence of POPF and infectious complications.
The relationship between drains and other complications remains somewhat
unclear. It has been reported that closed suction drains exert a continuous
negative pressure of up to 125mmHg, and ‘stripping’ the drain can increase
that to over 200mmHg. Several authors have opined that this negative
pressure may actually induce POPF by creating suction on the fragile
pancreatic parenchyma. We have also demonstrated in a comprehensive
analysis of patient and operative factors that operative drain placement is
associated with an increased rate of delayed gastric emptying, which has
been corroborated by other authors. Finally, routine placement of drains
invariably results in another surgical wound with associated patient
discomfort and rare technical complication associated with drain removal.
EVIDENCE FOR DRAIN EMPLOYMENT

STRATEGIES
Few clinical topics in pancreatic surgery have received as much attention
over the past decade in the literature as role of operative drainage. A
randomized trial from Memorial Sloan Kettering Cancer Center (MSKCC)
by Conlon and colleagues in 2001 was the first widely discussed paper on
the subject since the initial report of forgoing operative drainage in 1992. In
this paper, the MSKCC group randomized 179 patients undergoing pancreatic
surgery (PD or distal pancreatectomy) with no difference observed in
operative mortality or overall morbidity (overall, 2% and 59%,
respectively). In the drained group, 11 patients (12.5%) developed a POPF
while no patient in the undrained group (0%) developed a POPF. Patients
with a drain were more likely to develop intraabdominal collections or
enterocutaneous fistulae as well. The authors noted that their results seemed
to mirror those of other practice-changing trials that led to abandonment of
routine drain usage following other forms of intraabdominal surgery such as
cholecystectomy, colon resection, or gastric surgery. The high-level evidence
provided by this study served to ignite widespread interest in forgoing
routine drain placement following pancreatic surgery in general and PD in
specific.
Multiple retrospective reports and cohort studies from the 2000s to the
present day (including our own) have followed this seminal report and are
presented in Table 2 . These studies are larger and have reported more
granular data. Also, more modern reports exist in the era of standardized
definitions of PD-specific complications, such as POPF by international
consensus groups (the MSKCC report preceded these definitions), making
comparison more straightforward. Three such studies are listed in Table 2
that are largely congruent in their conclusion that routine drainage could be
abandoned in the majority of patients. The study by Mehta, et al analyzed 709
patients undergoing PD and found that drain placement had a higher
incidence of POPF, overall morbidity, and longer hospital stay. However,
patients that had drain placement had higher blood loss and longer operative
time, possibly suggesting a more difficult operative case. Conversely, duct
size was smaller in the undrained group. Perhaps the most crucial
contribution of this study is that the need for secondary (percutaneous) drain
placement during the postoperative course did not differ between the groups.
Correa-Gallego, et al provided follow-up from the MSKCC group 12 years
after Conlon’s original trial. In this large retrospective report of 739 patients
undergoing PD (over 1,100 were included in the entire study, which included
distal pancreatectomy as well), no overall difference in morbidity or
clinically-relevant POPF was observed (the authors employed a proprietary
institutional grading system for adverse events, but it correlates well with
internationally-accepted definitions). Undrained patients had a shorter length
of stay, but did have a higher mortality rate of 3% versus 1% in the drained
group. While this did not reach statistical significance and was within
accepted norms, subsequent analysis of these 11 patients in an editorial reply
(Ann Surg. 261(6):e160, June 2015) found that only 3 experienced a POPF
during their postoperative course prior to their demise (2 of the 3 patients
that died in the drained group had a POPF). Their manuscript also clearly
illustrates the current controversy in operative drainage – of the six high-
volume surgeons at MSKCC participating in the study, two routinely place
surgical drains following PD.
TABLE 2: Recent reports examining patient outcomes
following pancreaticoduodenectomy with and without usage
of operative drains.
Fisher and colleagues in Texas authored a cohort study in 2011 that

suggested decreased length of stay, readmission rates, and POPF formation in
patients undergoing PD managed without operative drains with no change in
mortality or overall morbidity. This study prompted the same group to initiate
a multi-center randomized trial involving nine centers that was reported in
April 2014 that examined 137 patients – 68 received operative drainage (the
type and number of drains was left to the discretion of the surgeon) while 69
did not. The study was designed with a target accrual of 376 patients but was
terminated by its Data Safety Monitoring Board. To much surprise, at 90 days
postoperatively, 8 patients (12%) in the undrained group had died versus 2
(3%) in the group that received a drain, a trend that nearly reached statistical
significance (p-value 0.097). In addition, an increased need for
postoperative drainage and a trend towards increased clinically-relevant
POPF was observed in those patients without operative drains. Six of the
eight patients who died did so in the setting of multisystem organ failure
associated with POPF.
These data were completely at odds with previous publications and are
difficult to reconcile. In general, prospective randomized trials are
considered superior due to minimization of bias. However, in surgical trials,
randomized controlled trials are often hampered by slow accrual, restrictive
inclusion criteria, and problematic surgical quality control. The study groups
appeared equitably randomized in this trial, although it is notable that 22% of
patients in the ‘No Drain’ group required a vascular resection during their
PD versus 13% in the ‘Drain’ group. While surgical quality in this trial was
undeniably high, utilizing 15 surgeons at 9 high-volume institutions, the
practices of those surgeons were clearly variable, as evidenced by the
variety of PD techniques reported in the trial. Postoperative management was
similarly not standardized. The inclusive nature of the trial design precluded
such standardization, and the experience of the study participants with
managing post-PD patients without drains is not detailed. However, several
findings are notable. The ‘No Drain’ group experienced a significant
increase in length of stay for their index admission, again at odds with
reported retrospective data. Despite this, percutaneous drain placement
occurred in the ‘No Drain’ group anywhere from 13 to 39 days
postoperatively (median, 23 days). In a study described above, where we
examined 20 patients in a cohort of 259 that underwent PD without operative
drains (POPF incidence of 7.7%) who later developed a clinically-relevant
POPF, the day of POPF diagnosis ranged from 7 to 42 (median, 18).
Furthermore, 6 undrained patients in the Van Buren trial required re-
laparotomy (8.7%), compared to 4.8% of undrained patients in our series.
The study authors note that POPF were “managed aggressively” when
diagnosed, but these data, combined with the late nature of the deaths (half
occurred after postoperative day 30), suggest that for at least some of the
patients with POPF an opportunity may have existed for earlier intervention.
It is unclear whether this would have resulted in rescue, but the overall
mortality for the trial of 7.3% is higher than the usual accepted figure of less
than 3% for PD at high-volume centers. Finally, there was no opportunity for
crossover in this trial. Advocates for selective drain placement generally
agree that there are some cases where drain utilization is not only prudent,
but strongly preferred (such as a contaminated field). For patients
randomized to the ‘No Drain’ group, there was no opportunity for surgical
judgement to allow placement of a drain. Even if such patients remained in
the ‘No Drain’ cohort and the trial were analyzed in an intent-to-treat
fashion, the results may have been quite different. For most surgeons that are
routine non-drainers, a selective drainage approach is the case in actual
practice; as such, examining drain utilization in high quality, large
retrospective trials may be more representative.
Since the trial by Van Buren et al several studies have been reported that
seek a way of identifying those cohorts that would benefit from routine drain
placement and those in whom drains may be obviated. For instance,
employing the widely-used and validated Fistula Risk Score metric to
stratify patients based on risk for POPF is one modality that has been
examined. These efforts are complemented by additional studies from both
American and Italian groups focusing on very early (postoperative day one)
drain removal based on amylase measurement in the drain effluent. These
laudatory efforts will hopefully coalesce in the future to provide an objective
method of selecting patients for drain usage. Furthermore, these data have
been gradually entering into generalized clinical practice. Enhanced
Recovery After Surgery (ERAS), or ‘fast-track’ protocols, which employ a
standardized set of pre-, intra-, and post-operative parameters in an effort to
minimize time to discharge and operative morbidity, have been examined in
patients undergoing pancreatic surgery. A fundamental component of early
ERAS protocols, which were first proposed in patients undergoing colon and
rectal surgery, was forgoing routine drain placement. Existing reports on
employing ERAS protocols in pancreatic surgery are encouraging with
regard to decreasing lengths of stay without increased readmission rates but
have a high degree of institutional heterogeneity and are either retrospective
in nature or utilize historical controls. Joint recommendations from the ERAS
Society, the European Society for Clinical Nutrition and Metabolism, and the
International Association for Surgical Metabolism and Nutrition published in
early 2013 note that evidence is currently insufficient to recommend avoiding
drains following pancreatic surgery entirely, but that evidence for ongoing
routine usage is weak at best. When drains are employed, ERAS guidelines
encourage their early removal postoperatively.
SUGGESTED STRATEGY
It is our practice to employ operative drains in a highly selected fashion
following PD. While we currently place drains in less than 5% of cases, we
do not advocate total abandonment of surgical drainage in appropriate
patients as determined by sound surgical judgement. In general,
contamination of the surgical field or concern for anastomotic fidelity are the
most common indications for drainage in our practice. Specific instances
could include:
1. Presence of or concern for persistent leakage from the biliary
reconstruction.
2. Gross contamination of the surgical field, either from pre-existing
abscess or uncontrolled spillage of a large volume of infected bile or
succus from hollow viscera (i.e. during concomitant bowel resection).
3. Abnormally difficult or tenuous pancreatic anastomosis.
The third indication above is rarely applied in our practice and is not
universally utilized in the case of a pancreas thought to be high risk of POPF
formation due to a soft gland texture or small pancreatic duct if the
anastomosis is otherwise sound. In the case scenario above, in our practice
we would consider placement of a drain in the region of biliary anastomosis
if doubt persisted regarding its integrity. Should the operative field no longer
be soiled from spillage of stent-associated sludge, this would not necessarily
mandate drain placement, nor would the soft pancreas with a small duct.
The operative technique for reconstruction during PD we employ is as
follows: the pancreatic duct is transected sharply and
pancreaticojejunostomy is carried out in a two-layer end-to-side fashion with
duct-to-mucosa reconstruction. Single-layer hepaticojejunostomy in the
Blumgart fashion is similarly performed in a running fashion unless the duct
is small or tissue quality is poor. Following the re-establishment of biliary
continuity the area is copiously irrigated. Once the field is clean, an unsoiled
laparotomy sponge is firmly applied circumferentially to the biliary
anastomosis. Any evidence of biliary leakage, easily visible as greenish
discharge on the white laparotomy sponge, results in reinforcement or
revision of the anastomosis until its integrity is ensured. If drains are
employed, they are placed in a directed fashion (i.e. – if being placed for
biliary concerns, a second drain need not be placed adjacent to the
pancreatic reconstruction). During recovery, diet is re-instituted as soon as
patient tolerance allows (usually, liquids on day two and solids on day three
or four). Once oral intake has been established, the drains are examined (and
amylase concentration assessed, if the drain is peri-pancreatic) and removed
if no leakage is evident. For patients without drains, a high index of
suspicion is mandatory for POPF development, especially if a high-risk
gland exists (soft texture/small duct).
Slow return of dietary tolerance, unexplained fevers (even if intermittent),
leukocytosis, and vague abdominal discomfort should be closely monitored
with a low threshold for further investigation.
It is our belief that the majority of patients can undergo PD without
placement of operative drains. We recently retrospectively examined 343
patients that have undergone PD at Yale-New Haven Hospital over a ten-year
period by a single surgeon (RRS). Operative drains were employed in 23
(6.7%) for the reasons described above. Overall mortality was 1.2%, major
morbidity occurred in 16%, and clinically-relevant class B or C POPF
occurred in 9.6%. The sum of the above data suggests that doing so is safe
and may benefit the patient via reduced length of stay and morbidity. Should a
clinically-relevant POPF develop, we have not seen degradation in the
outcomes or limitations in the management of those patients without
operative drains. For surgeons who currently routinely employ operative
drains and are considering omitting them to gain the above potential benefits,
we encourage them to proceed judiciously. Complete familiarity with the
postoperative course is needed and it is imperative to remember that
placement of a surgical drain is necessary using sound surgical judgement in
certain patients. Finally, institutional expertise in the form of excellent house
staff, nursing, and interventional radiology is mandatory to allow efficient
diagnosis and treatment when POPF or other surgical complication ensues.
As additional data accrues and management techniques are refined, the
appropriate usage of operative drains following PD will continue to evolve,
as will the long course towards optimizing surgical outcomes all patients
undergoing pancreatic surgery.
SALIENT POINTS
Operative drainage has historically been an invariable component of
performing pancreaticoduodenectomy but this practice has been
questioned as routine drainage after most elective abdominal surgery
has been abandoned.
The first prospective trial and numerous large retrospective reports
from high-volume institutions suggest that forgoing routine placement of
drains decreases length of stay and the incidence of infectious
complications, delayed gastric emptying, and postoperative pancreatic
fistula formation (although perhaps not clinically-relevant grade B and
C fistulae). These benefits seem to be maximized when compared to
groups with operative drains left in place for a prolonged period of
time.
A single, recent prospective trial suggested that
pancreaticoduodenectomy without routine drain placement increased
operative mortality. However, this trial may not be generalizable to the
actual practice patterns of surgeons who do not routinely employ drains
and efforts are ongoing to objectively stratify risks of placing or
omitting drain placement.
The decreased morbidity associated with forgoing routine drain usage
can be safely achieved if the surgeon assertively investigates any signs
or symptoms of early morbidity and the institutional expertise is
available to assist with management (especially interventional
radiology).
Sound surgical judgement should never be supplanted by inflexible
doctrine regarding drain placement following
pancreaticoduodenectomy.
SELECTED REFERENCES
1. Kunstman JW, Kuo E, Fonseca AL, Salem RR. Evaluation of a recently
described risk classification scheme for pancreatic fistulae
development after pancreaticoduodenectomy without routine post-
operative drainage. HPB (Oxford). 2014; 16(11):987-93.
2. Conlon KC, Labow D, Leung D,Smith, Jarnagin W, Coit DG, et al.
Prospective randomized clinical trial of the value of intraperitoneal
drainage after pancreatic resection. Ann Surg. 2001;234(4):487-93;
discussion 93-4.
3. Mehta VV, Fisher SB, Maithel SK, Sarmiento JM, Staley CA, Kooby
DA. Is it time to abandon routine operative drain use? A single
institution assessment of 709 consecutive pancreaticoduodenectomies. J
Am Coll Surg. 2013;216(4):635-42; discussion 42-4.
4. Correa-Gallego C, Brennan MF, D’Angelica M, Fong Y, Dematteo RP,
Kingham TP, et al. Operative drainage following pancreatic resection:
analysis of 1122 patients resected over 5 years at a single institution.
Ann Surg. 2013; 258(6):1051-8.
5. Van Buren G, 2nd, Bloomston M, Hughes SJ, Winter J, Behrman SW,
Zyomski, et al. A randomized prospective multicenter trial of
pancreaticoduodenectomy with and without routine intraperitoneal
drainage. Ann Surg. 2014;259(4):605-12.

CASE SCENARIO
A 69-year-old man undergoes a pylorus-preserving
pancreaticoduodenectomy for a 2cm ampullary adenocarcinoma. The
pancreas gland is soft, the duct is observed to be 2mm, the
pancreaticojejunostomy is constructed in two interrupted layers with a
pancreatic stent, EBL is 550cc and a drain is left at the
pancreaticojejunostomy. On POD #4 the drain amylase is 167 and drain
output is <100cc per day, and the drain is removed. On POD #7 he becomes
tachycardic and febrile with low urine output. A CT scan is obtained (Figure
1 ) which shows a completely dehisced pancreaticojejunostomy
(International Study Group on Pancreatic Fistulas or ISGPF grade C). He is
booked for an exploratory laparotomy. What is the optimal operative
management?
FIGURE 1: A CT scan of a dehisced pancreatico-jejunostomy with
evidence of local sepsis. Red arrow displays an internalized, short
sialastic pancreatico-jejunostomy stent from the original operation
where the anastomosis was created. This has lost continuity with the
enterotomy on the pancreatobiliary bowel limb.
█ BACKGROUND
Over the past several decades the mortality following
pancreaticoduodenectomy has decreased dramatically and is now reported to
be less than 5% at high volume centers. The overall morbidity is still
reported to be as high as 60%, however. A leak from the pancreatic
anastomosis and subsequent pancreatic fistula remains a major cause of
morbidity following pancreaticoduodenectomy, and can present late in the
postoperative course as sepsis or as hemorrhage, often from the much-feared
complication of gastroduodenal (GDA) artery stump blowout. The incidence
of pancreatic fistula is still reported to be between 9 and 29%, with an
associated mortality of around 10%.
With the overarching aim of improving management and decreasing
morbidity, there has been an effort over the past decade to standardize the
definition and classification of pancreatic fistulae. In 2005 an international
study group comprised of 37 pancreatic surgeons from five continents met
with the aim of standardizing the much-debated definition of pancreatic
fistula. They generally defined a post-pancreaticoduodenectomy pancreatic
fistula as either a failure of healing of a pancreatic-enteric anastomosis or as
a parenchymal leak not directly related to the anastomosis, such as one
originating from the raw pancreatic surface. They suggested a diagnosis
based on any measurable drain output on postoperative day three or onwards
with amylase content greater than three times the upper limit of normal. In
addition, they established a clinical grading system that classified pancreatic
fistula as grade A, B or C based on the clinical severity and impact on
patient outcomes (Table 1 ). In brief, grade A was considered a transient
fistula with no clinical impact and minimal deviation from the expected
postoperative course. Grade B fistula mandates a change in management,
such as disallowing oral intake or treatment with parenteral nutrition, and
often leads to a delayed discharge. Grade C fistulae are defined by clinical
decline or instability, a major delay in discharge, and major deviation from
the expected clinical course including necessity of additional invasive
procedures and reoperation.
TABLE 1: Diagnosis and classification of postoperative
pancreatic fistulas following pancreatic surgery by the
International Study Group on Pancreatic Fistulas (Adapted
with permission from Elsevier in Bassi C, et al. The 2016
update of the International Study Group (ISGPS) definition
and grading of postoperative pancreatic fistula: 11 years
after. Surgery 2017;161(3):584-591).
The management of pancreatic fistula has evolved as the ability to predict

its occurrence has improved. In 2013 Callery et al. evaluated recognized
clinical risk factors for pancreatic fistula after pancreaticoduodenectomy and
constructed a clinical risk score that was then validated prospectively as an
accurate predictor of clinically relevant pancreatic fistula (Table 2 ). The
risk score accounts for excessive blood loss, small duct size, soft gland and
disease pathologies other than pancreatic adenocarcinoma or pancreatitis. In
their review of 445 patients, nearly all of whom had a
pancreaticojejunostomy performed for restoration of enteric continuity,
20.9% had a pancreatic fistula and 13% had a clinically relevant (grade B or
C) fistula. Of these the majority were grade B (54% of all the pancreatic
fistulas) and only 9% were grade C. Eleven patients (2.5%) required
reoperation and one death was directly attributable to pancreatic fistula.
Patients who were considered to be high-risk were those that scored 7 to 10
points on the fistula risk score, and they were not only far more likely to
develop a pancreatic fistula, but when a fistula develop it was much more
likely to be clinically relevant and impact outcomes (6:1 odds). In this
context, a patient who is considered high-risk for fistula development may be
managed more conservatively, with an increased time to drain removal or
slower diet advancement, in an effort to both prevent fistula formation and
optimize management should it occur.
TABLE 2: Fistula Risk Score for Prediction of Clinically
Relevant Pancreatic Fistula after Pancreatoduodenectomy
(Model III) (Reprinted with permission from Elsevier in
Callery MP, et al. A prospectively validated clinical risk
score accurately predicts pancreatic fistula after
pancreatoduodenectomy. J Am Coll Surg. 2013;216(1):1-14)
The incidence of grade C pancreatic fistula requiring reoperation is
reported to be between 5 and 20% and in this instance the subsequent
mortality approaches 39%. In 2009, Fuks et al. reviewed 680 patients who
underwent pancreaticoduodenectomy and five French centers with the aim of
evaluating the incidence, prognosis and associated risk factors for grade C
fistula. Their overall fistula rate was 16.3% (111 patients) and patients were
evenly split between having had pancreaticojejunostomy or
pancreatogastrostomy performed. 32% of patients with fistulas were
classified as grade C (36 patients) and the mortality rate of this group was
about 40%. All grade C patients were symptomatic, with 47% (17 patients)
presenting with sepsis due to an intraabdominal collection, 44% (16
patients) with postoperative bleeding, and 9% (three patients) with
multisystem organ failure. The mean length-of-stay in the hospital for grade C
patients was 46 days, significantly longer than the 29 days for patients with
grade A or B fistula (p<0.001). On univariate analysis the following were
significant risk factors for development of a grade C fistula: soft pancreatic
parenchyma (p=0.011), perioperative blood transfusion (p=0.003) and
postoperative bleeding (p=0.001). On multivariate analysis, none of the risk
factors reached statistical significance.
Nearly all of the patients that presented with grade C fistula underwent
reoperation (35 of 36), with the remaining patient dying of multi-organ
failure before surgery could be performed. Of the 17 patients who presented
with sepsis secondary to intraabdominal collection, 7 underwent attempted
percutaneous drainage. All 7 of these patients progressed and required
operative management, and the other 10 went straight to surgery. Nine
patients (53%) underwent drainage and six of them had the anastomosis
reduced secondary to necrosis and dehiscence, four underwent disconnection
of the anastomosis with resection of the body and conservation of a small
remnant, and two underwent completion pancreatectomy. Data was missing
for the remaining two patients. Six patients (35%) who had undergone
surgery for an intraabdominal collection died with one having undergone
completion pancreatectomy and five having undergone drainage. The 16
patients who presented with bleeding underwent hemostasis – either by
splenectomy or GDA stump ligation to address pseudoaneurysms or by
isolated elective hemostasis – at the time of reoperation. Five of these
patients (31%) developed recurrent bleeding and died.
The management of grade C pancreatic fistula requiring reoperation is a
much-debated topic. The international study group suggested that re-
exploration may be mandated for a grade C fistula in the setting of a
deteriorating clinical status for one of the three reasons: (1) an attempt to
repair the site of leakage with wide peripancreatic drainage, (2) conversion
to alternative means of pancreatic-enteric anastomosis, or (3) completion
pancreatectomy. In a large retrospective review of 669 patients who
underwent pancreaticoduodenectomy over an eight-year period at a high-
volume French center, Balzano et al. aimed to compare operative strategies
for patients requiring surgical management of a pancreatic leak. At initial
operation, the patients were all reconstructed with a two-layer end-to-side
pancreaticojejunostomy, with 50.2% of patients receiving a separate Roux-
en-Y jejunal limb owing to the presence of risk factors for leak. A clinically
relevant (grade B or C) leak occurred in 98 patients (14.6%) with 37 (5.5%)
undergoing repeat laparotomy for either delayed hemorrhage (22 patients,
59.5%) or sepsis (15 patients, 40.5%). The two management strategies
employed were preservation of the pancreatic stump or completion
pancreatectomy. Six patients were excluded because the option did not exist:
two had a necrotic pancreatic stump, two were too unstable for major
surgery and two had an inaccessible pancreatic stump. For the remaining
patients: 17 underwent pancreas preservation, 10 with simple drainage and
seven with pancreatic duct exclusion, and 14 underwent completion
pancreatectomy.
In reviewing their short-term outcomes the authors found that operative
time and intraoperative blood loss were both significantly higher for patients
who underwent completion pancreatectomy. Patients who underwent
pancreas preservation required more subsequent laparotomies (10 of 17
patients, 58.9%) compared to those who underwent up-front completion
pancreatectomy (1 of 14 patients, 7%), usually for relapsing hemorrhage.
Overall mortality was 25.8% and did not differ between groups. Other
complications—including bile leak, delayed gastric emptying and wound
infection—were not significantly different between the two groups. Length-
of-stay was not significantly different, and more patients who underwent
completion pancreatectomy required ICU admission. Overall, the authors
favored completion pancreatectomy and advocated for the use of
intraoperative or postoperative autologous islet transplantation as a rescue
therapy to avoid endocrine insufficiency.
Another operative strategy to address a high-grade
pancreaticojejunostomy leak is to convert the anastomosis to a salvage
pancreatogastrostomy. Bachellier et al. described this procedure and its
outcomes in 2008. They reviewed 403 patients who underwent
pancreaticoduodenectomy over a 17-year period. 318 were reconstructed
with pancreatogastrostomy and 85 with pancreaticojejunostomy. Twelve
patients with pancreaticojejunostomy were treated for a grade C pancreatic
fistula and all 12 underwent repeat laparotomy. Eight patients underwent
completion pancreatectomy and four patients underwent salvage
pancreatogastrostomy. This was performed by taking down the dehisced
pancreaticojejunostomy, debriding the pancreatic stump, mobilizing the
pancreatic remnant to telescope into the gastric lumen, and then fashioning a
two-layered anastomosis. The postoperative mortality was 50% in the
completion pancreatectomy group and 0% in the salvage
pancreatogastrostomy group. Overall rates of postoperative complications
were similar as were postoperative lengths-of-stay. All four surviving
completion pancreatectomy patients were insulin-dependent diabetics as was
one patient following pancreatogastrostomy. One pancreatogastrostomy
patient developed a grade B pancreatic fistula and this was managed non-
operatively. Though the sample number was quite small, the authors
concluded the salvage gastrostomy was a safe and effective alternative
method of operative management that largely avoided the long-term
complications of completion pancreatectomy.
Our group has described a bridge stent technique to intraoperatively
address a disrupted pancreaticojejunostomy. In this technique, the disrupted
pancreaticojejunostomy is exposed and the pancreatic remnant is debrided to
healthy tissue. A 5-Fr or 8-Fr silastic pediatric feeding tube is placed across
the space between the jejunal enterotomy and the pancreatic duct and secured
at both ends with an absorbable suture. Another option is to externalize the
stent through a Witzeled pancreaticobiliary limb and then to the abdominal
wall. In either case, wide drainage is also employed. We reviewed 357
patients who underwent pancreaticoduodenectomy over an eight-year period
at our institution. Seven grade C fistulae occurred (2%). Two patients died
prior to reoperation and neither had a complete anastomotic dehiscence. Five
patients underwent reoperation with bridge stent technique. All survived and
were discharged at a median 41 days (range 8 to 77 days). Median follow-up
was 10.7 months (range 5 to 68 months) and over that time period no patients
developed any recurrent pancreatic fistulae, anastomotic strictures,
pancreatitis, atrophy of the remnant gland, or new-onset diabetes. All five
patients were recognized to have two or three recognized risk factors for
pancreatic fistula including pathology other than pancreatic adenocarcinoma
or pancreatitis, soft gland and small pancreatic duct.
There are many operative strategies to consider for the management of
grade C fistula following pancreaticoduodenectomy and the bridge stent
technique remains our preferred approach. Completion pancreatectomy is
technically difficult and often results in prolonged operative times and
excessive blood loss. Though it is a definitive procedure, it commits the
patient to lifelong total exocrine and endocrine insufficiency. Wide external
drainage often prolongs the hospital course and ultimately requires repeat
operations and interventions. Conversion of a dehisced
pancreaticojejunostomy to a pancreatogastrostomy mandates an often arduous
dissection and mobilization of the pancreatic remnant and is not possible in a
significant proportion of patients. The bridge stent technique successfully
diverts pancreatic secretion and allows a fibrotic wall to develop around the
channel, thereby recreating the disrupted anastomosis.
SALIENT POINTS
Pancreatic fistula remains a significant complication following
pancreaticoduodenectomy, and development of a grade C fistula is a
major source of morbidity and mortality.
The traditional operative management of grade C pancreatic fistula
was completion pancreatectomy, a difficult procedure that commits that
patient to lifelong total exocrine and endocrine insufficiency.
Alternative management strategies include wide drainage, with or
without pancreatic duct occlusion, and conversion of the
pancreaticojejunostomy to a pancreatogastrostomy for restoration of
enteric continuity.
Our preferred approach is the bridge stent technique that avoids
excessive dissection but successfully diverts pancreatic secretion
while allowing for the formation of a ‘neo-anastomosis’.
SELECTED REFERENCES
1. Bassi C, Marchegiani G, Dervenis C, Sarr M, Hilal M, Adham M, et al.
The 2016 update of the International Study Group (ISGPS) definition
and grading of postoperative pancreatic fistula: 11 years after. Surgery
2017;161(3):584-591.
2. Balzano G, Pecorelli N, Piemonti L, Ariotti R, Carvello M, Nano R, et
al. Relaparotomy for a pancreatic fistula after a
pancreaticoduodenectomy: a comparison of different surgical strategies.
HPB (Oxford). 2014;16(1):40-5.
3. Fuks D, Piessen G, Huet E, Tavernier M, Zerbib P, Michot F, et al. Life-
threatening postoperative pancreatic fistula (grade C) after
pancreaticoduodenectomy: incidence, prognosis, and risk factors. Am J
Surg. 2009;197(6):702-9.
4. Callery MP, Pratt WB, Kent TS, Chaikof EL, Vollmer CM, Jr. A
prospectively validated clinical risk score accurately predicts
pancreatic fistula after pancreatoduodenectomy. J Am Coll Surg.
2013;216(1):1-14.
5. Bachellier P, Oussoultzoglou E, Rosso E, Scurtu R, Lucescu I, Oshita A,
et al. Pancreatogastrostomy as a salvage procedure to treat severe
postoperative pancreatic fistula after pancreatoduodenectomy. Arch
Surg. 2008;143(10):966-70; discussion 71.
6. Kent TS, Callery MP, Vollmer CM, Jr. The bridge stent technique for
salvage of pancreaticojejunal anastomotic dehiscence. HPB (Oxford).
2010;12(8):577-82.

CASE SCENARIO
The patient is a 60-year-old male who underwent a Whipple procedure for
an ampullary mass. The pancreaticojejunostomy was high risk due to the
combination of a small pancreatic duct and soft gland. On postoperative day
#5, the JP drain output was sinister in appearance and had an amylase value
of 3000, indicating a pancreatic leak. His postoperative course was then
slowed by delayed gastric emptying. On postoperative day 15, the patient
spontaneously drains 400cc of bright red blood from the JP. He is
hemodynamically stable and has no new complaints. What do you do?
█ BACKGROUND
This scenario should be recognized by all experienced pancreatic surgeons
as an ominous finding. Though the patient is stable and the event is tempting
to ignore, it may represent a “sentinel bleed”—bleeding that is relatively
minor and stops without any intervention yet precedes massive hemorrhage.
That this has occurred in delayed fashion relative to the index operation and
in the context of a pancreatic fistula is particularly worrisome.
Multiple algorithms have been developed for the evaluation and treatment
of post-pancreatectomy bleeding. In this chapter, the diagnostic and
therapeutic options used in the evaluation and treatment of post-
pancreatectomy hemorrhage are reviewed. After summarizing the
management options, an algorithm tailored toward delayed post-
pancreatectomy hemorrhage is presented.
DEFINITIONS
For years, the frequency, morbidity, and mortality of post-pancreatectomy
bleeding were difficult to quantify. That is because the term “post-
pancreatectomy bleeding/hemorrhage” was applied heterogeneously, and
standardized definitions were lacking. As a result, reported experiences
were difficult to compare. In 2007, the International Study Group of
Pancreatic Surgery (ISGPS) defined post-pancreatectomy hemorrhage (PPH)
based on a collective review of the literature. Post-pancreatectomy
hemorrhage is now differentiated based on the time of onset, the location and
cause, and the severity (Table 1 ). The time of onset is either early (≤ 24
hours after the end of the index operation) or late (≥ 24 hours after the end of
the index operation). The location is either intraluminal or extraluminal.
Severity is mild (small or medium volume blood loss managed
conservatively) or severe (larger volume blood loss with clinical
deterioration and the need for significant transfusion and/or invasive
treatment). To establish a clinical classification, three different grades of
post-pancreatectomy hemorrhage (grades A, B, C) were defined according to
these parameters and based on the clinical consequence to the patient. Grade
A bleeding is early, mild bleeding without much clinical significance. Grade
B bleeding is early severe or late mild bleeding that may require procedural
interrogation. Grade C bleeding is late, severe hemorrhage of major clinical
impact with potentially life-threatening ramifications. Grade C bleeds
invariably require intervention.
TABLE 1: ISGPS classification of post-pancreatectomy
hemorrhage (Reprinted with permission from Elsevier in
Wente M, Veit J, et al. Postpancreatectomy hemorrhage
(PPH)-An International Study Group of Pancreatic Surgery
(ISGPS) definition. Surgery 2007;142:20-5).
Despite the consensus definition, large series in the literature continue to

define post-pancreatectomy hemorrhage variably. This results from
perceived flaws in the consensus classification. One issue is that the ISGPS
definition includes a large number of patients with mild bleeding of little or
no clinical significance. Given the comorbidities, preoperative
deconditioning, and anemia present in many patients that undergo pancreatic
surgery, most surgeons would regard a short postoperative course that
features mild decreases in hemoglobin concentration, requiring observation
alone or small-extent (< 3 units PRBC) transfusion as essentially
uncomplicated. Furthermore, many patients that receive small-volume
transfusions have no actual documented bleeding or concern for clinically-
significant bleeding. Despite this, the ISGPS consensus definition when
applied rigorously categorizes a large number of patients with uneventful
postoperative courses as having suffered from “post-pancreatectomy
hemorrhage.”
A second issue in applying the consensus definition is the differentiation
of early and late bleeding. Though the 24-hour time point distinguishes early
and late hemorrhage by ISGPS consensus, this may be too stringent. A variety
of time points from 2 days to 30 days have been used elsewhere. The concern
with the consensus definition is that bleeding within 72 hours after surgery
may still result from technical failures or perioperative coagulopathy, but
bleeding that occurs 2-3 weeks after surgery clearly differs in both the
pathophysiology and management options. Any selected time point has
drawbacks, which is why the evaluation and treatment of these patients will
remain individualized.
EVALUATION & TREATMENT

The distinction between early and late hemorrhage is fundamental in
evaluation of post-pancreatectomy bleeding. Early bleeding is most often due
to technical failure. It may originate from anastomotic suture lines,
improperly secured vessels, or the resection cavity. If severe, early operative
re-intervention is the typical strategy and is associated with high rates of
success and low mortality.
Late severe hemorrhage (especially bleeding more than five days after the
index operation) is more likely due to an anastomotic ulcer, vascular
erosion/pseudoaneurysm, or dehiscence of an anastomotic suture line.
Predisposing factors for late severe hemorrhage include intra-abdominal
abscess and leak from the biliary and/or pancreatic anastomoses.
Inflammation and erosive contents in this setting can result in bleeding from
anastomotic suture lines and can disrupt sutures/ligatures. Extensive
dissection done during the operation can also make vessels more vulnerable
to erosion. The most likely etiology of delayed post-pancreatectomy
hemorrhage has consistently been disrupted visceral vessels.
Options consistently applied in the management of PPH include axial
imaging, endoscopy, interventional radiology, and surgery. Choosing the most
applicable technique requires proper assessment of the clinical scenario.
THE “SENTINEL BLEED”

Many patients that develop severe post-pancreatectomy hemorrhage will first
present with a “sentinel” or “herald” bleed. Sentinel bleeding can be defined
as discrete but evident blood loss via abdominal drains, nasogastric tube,
hematemesis, or melena with a decrease in hemoglobin concentration that is
less than 1.5 g/dL. This should cease spontaneously without the need for
transfusion but is associated with repeat hemorrhage after at least 12 hours of
stability. Sentinel bleeds appear to precede at least 30% of late severe post-
pancreatectomy bleeds. In the setting of a concomitant pancreatic fistula,
sentinel bleeds are even more common, with rates as high as 100% reported.
The largest reported series examining post-pancreatectomy hemorrhage was
published by Yekebas, et al. in 2007. In this study, the presence of both a
pancreatic fistula and a sentinel bleed prior to an episode of hemorrhage was
associated with a mortality risk greater than 50%. Notably, sentinel bleeds
are often preceded by an episode of severe abdominal pain.
One difficulty with the recognition of a sentinel bleed can be its subtlety.
A patient that spontaneously drains 250-500cc of bright red blood with
accompanying hemodynamic changes has clearly manifested a likely arterial
issue. By contrast, 100cc of darker blood tainted by an ongoing pancreatic
fistula is less easily explained and easier to disregard. It may simply
represent the drain accessing a pocket of previously loculated or undrained
hematoma.
The most critical aspect of the sentinel bleed is the recognition that any
new bleeding that occurs remote from surgery in a previously stable patient
mandates an urgent evaluation for arterial injury. This is especially true in the
setting of a pancreatic fistula. Diagnostic considerations include a CT scan
or a formal arteriogram, depending on both the institutional preference and
the clinical scenario. In a stable patient without a clear suggestion of arterial
bleeding, I prefer a multiphase CT scan. An arterial phase is mandatory as
the main goal of the study is to rule out arterial pathology. However, while a
CT angiogram is excellent for arterial findings, all other information from the
study is often compromised. Thus I prefer a 3 phase CT with quality arterial
and venous phases.
THE ROLE FOR ENDOSCOPY

Endoscopy is frequently the initial intervention in patients presenting with
intraluminal bleeding. Again, a distinction must be made between early and
late bleeding. Many surgeons are reticent to perform endoscopy within the
first few postoperative days due to concerns over distension of fragile, fresh
anastomoses. This is especially true in the setting of a
pancreaticogastrostomy, which is a far more delicate anastomosis than the
gastrojejunostomy. The alternative is early re-exploration, which is generally
successful. However, endoscopy is used readily in the setting of delayed
intraluminal bleeding and is frequently employed in the early setting as well.
Although endoscopy is used quite readily to triage intraluminal bleeding,
its therapeutic efficacy is somewhat limited. Even in technical “successes”
(as defined by no further episodes of bleeding), there is often no bleeding
site visualized and no intervention done. Rates at which bleeding sites are
seen and treated endoscopically with durable success are quite low. More
commonly, no site of active bleeding is noted. The pattern of bleeding can,
however, be helpful. For instance, a blood-filled afferent loop with little
blood in the stomach and no gastrojejunal ulcer would suggest that the
pancreaticojejunostomy is the most likely site of bleeding. This information
can help the surgeon should the bleeding fail to stop and re-exploration be
necessary. However, these types of endoscopic findings can also mislead.
Bleeding attributed to non-specific endoscopic findings, such as erosive
gastritis, can delay interventions and escalate morbidity after a subsequent
re-bleed. Further, extraluminal bleeding in the setting of an anastomotic
disruption can masquerade as an intraluminal bleed (“pseudo-intraluminal”
bleeding).
Studies have consistently noted that endoscopy has essentially no role in
late extraluminal bleeding. Although “pseudo-extraluminal” bleeding can, in
theory, be treated endoscopically, the success rates are quite low.
Angiographic or surgical measures are far more likely to provide a durable
solution. In their sizable PPH series, Yekebas et al. found that endoscopy
failed to solve any post-pancreatectomy hemorrhage occurring more than 5
days after the index operation.
INTERVENTIONAL RADIOLOGY
Surgical exploration has traditionally been the principle therapeutic venture
in post-pancreatectomy hemorrhage. However, a clear shift has occurred
toward the use of interventional radiology in the diagnosis and treatment of
post-pancreatectomy bleeds. Re-exploration in this group of patients is
fraught with exceedingly high rates of morbidity and mortality. These patients
are often deconditioned, and dense peri-pancreatic adhesions are generated;
visualizing the site of bleeding requires manipulation of tenuous
anastomoses. Catheter-based techniques facilitate the treatment of arterial
injuries without hazardous dissection.
Post-pancreatectomy bleeds can originate from a variety of arterial
sources, including the hepatic arteries, gastroduodenal artery,
superior/inferior pancreaticoduodenal artery, splenic artery, and superior
mesenteric artery. In any endovascular evaluation, both the celiac and
superior mesenteric artery distributions must be interrogated, as multiple
arterial erosions can occur. The most common affected artery across series is
the gastroduodenal. Therapeutic options for treating this and other
pseudoaneurysms include coil embolization or exclusion of the bleeding site
with an endovascular stent. In a systematic review of delayed post-
pancreatectomy hemorrhage, the success of diagnostic angiography in
localizing a bleeding source was 88%, and the success rates of endovascular
measures in achieving hemostasis are essentially equivalent to surgical re-
exploration in modern series.
The interventional approach can fail for multiple reasons. First and
foremost is institutional availability. In many cases, the role for
interventional radiology depends upon patient stability and the time from
referral until the patient is on the table. Assuming this is no obstacle, other
issues can still arise. One is failure to identify the site of bleeding. This
occurs frequently, as bleeding may stop by the time of interrogation, or
originate from as a vascular injury that is too small to identify. Secondly, a
site of origin may be identified but have no treatment option. Since the
dominant regional arteries, including the SMA and hepatic, are skeletonized
in the course of dissection, and branches including the gastroduodenal and
inferior pancreaticoduodenal are often ligated near their origin, there is
concern about the ability to treat vascular issues without compromising
perfusion of the liver and intestine. Multiple complications associated with
endovascular treatment have been described, including re-bleeding, stent
stenosis/thrombosis, hepatic abscess, and bowel/liver infarction. Regardless,
a formal arteriogram that identifies but fails to treat a vascular erosion is still
of value as it can help to direct surgical re-exploration.
A matter of controversy in catheter-based treatment of these bleeds is
empiric embolization. As noted, bleeding in this scenario can originate from
numerous sites. This can create apprehension and skepticism toward empiric
embolization if no one site is suggested. However, empiric embolization
clearly has a role. Intraoperative findings or postoperative studies can
frequently direct attention to a likely bleeding source without absolute
evidence (Figure 1 ). Yekebas et al. reported an 80% success rate for
“blind” embolization based on surgeons’ clinical suspicion. My strategy is to
empirically embolize vessels that can be treated safely in patients with a
clear suggestion of arterial hemorrhage. For instance, empiric embolization
should be strongly considered in the setting of a late severe bleed with
concomitant pancreatic fistula, as the likelihood of an arterial erosion is
quite high. The best candidate for safe empiric embolization is usually the
gastroduodenal artery. This can sometimes be ligated a fair distance (>5mm)
from the hepatic artery, which allows for safe coil embolization. If no such
stump is present, empiric treatment may only be achievable using covered
metal stents. I have opted against empiric treatment with stents unless there is
a very high index of suspicion for a particular site, since stent thrombosis can
have disastrous consequences. If the patient is stable and there is not a high
index of suspicion for arterial disruption, I do not favor empiric embolization
of even safely-treatable vessels, since the results of a complication can be
devastating.
FIGURE 1: Index of suspicion in endovascular assessment. This
patient developed a sentinel bleed several days after a Whipple
procedure in the setting of a concomitant pancreatic fistula.
Although his CT angiogram was negative for a blush or a
pseudoaneurysm, the drain was found to essentially abut the
gastroduodenal artery creating suspicion that the GDA was the
source of the bleed (Figure 1A). The patient ultimately underwent a
formal arteriogram that demonstrated a GDA pseudoaneurysm
(Figure 1B). The drain is seen coursing adjacent to the GDA on the
arteriogram as well.
THE ROLE FOR SURGERY

Though modern approaches are shifting toward less invasive techniques,
large series indicate that the majority of post-pancreatectomy hemorrhage
continues to be managed surgically. Re-exploration best fits the management
of early severe (grade B) bleeds. These are typically the result of technical
failures of hemostasis, and re-exploration is associated with high rates of
success and low mortality in this setting.
Morbidity and mortality escalate significantly in surgical management of
delayed arterial bleeds. This is reflected in the shift from surgery to IR
techniques in the management of these patients. However, surgery remains
pertinent in several specific scenarios. One is as a last resort when
endoscopy and arteriogram fail to demonstrate a bleeding site. A second
basis for surgery is that bleeding often occurs alongside a concomitant septic
complication. This constellation of issues can mandate surgical exploration
as the only measure capable of simultaneously treating both issues. Lastly,
circumstances of profound hemodynamic instability without the time
necessary for IR interrogation are frequently managed in the operating room.
The exact surgical procedure performed depends upon the intraoperative
findings. In general, the goal is to repair the site of bleeding and manage the
concomitant fistula (wide drainage). Completion pancreatectomy should at
least be considered. It is generally difficult to isolate the repaired vessel
from an ongoing pancreatic leak, and thus the risk of rebleeding can be high.
In systematic review, the rate of completion pancreatectomy after laparotomy
done for delayed hemorrhage was 41%.
SUMMARY AND ALGORITHM

The approach to any patient with delayed post-pancreatectomy hemorrhage
must be individualized. Each patient has a unique combination of
intraoperative findings, postoperative course, pattern/magnitude of bleeding,
and hemodynamic parameters that dictate the correct strategy. Any algorithm
that attempts to standardize the management will be imperfect, but general
approaches can be constructed. Figure 2 represents one approach that I
favor.
FIGURE 2: Proposed algorithm for delayed post-pancreatectomy
hemorrhage (psa: pseudoaneurysm).
The first decision for the surgeon is whether a major arterial disruption is
strongly suggested. Either a small volume of bright red blood, or any amount
of venous-appearing blood not accompanied by hemodynamic change, is
ominous but does not clearly suggest arterial erosion. This may, however,
represent a sentinel bleed. Since the patient is hemodynamically stable, and
the problem is not clearly arterial in nature, I would favor a CT angiogram
with delayed images (multiphase CT). This is a good precursor to a formal
arteriogram. If a blush or pseudoaneurysm is noted, the event was clearly a
sentinel bleed and the vessel should be treated interventionally or surgically
with IR strongly favored. If no target is noted (no blush and no
pseudoaneurysm), I would transfer the patient to an intensive care setting and
monitor closely. One of two scenarios is most likely. One is that the bleeding
was not a true sentinel bleed and no major bleeding will occur over the next
few days. The second is a more significant bleeding event some time within
the next 72 hours, which more clearly evidences an arterial etiology.
If bright red blood is noted that is significant in volume and/or
accompanied by hemodynamic changes, a major arterial erosion is very
likely. In this event, I would favor proceeding straight to a formal
arteriogram in interventional radiology if the patient’s hemodynamics allow
for this. While stable patients should certainly be sent to IR, I favor sending
even a somewhat unstable patient as long as the hemodynamics can be
supported with transfusion. Avoidance of the CT scan gives the surgeon the
best chance for identifying a target that can be treated without delay. Ideally,
a target vessel can be identified and treated safely with endovascular coiling
or stenting. If an arterial problem can be identified but not treated, surgery is
the only option. A negative arteriogram (no active extravasation or
pseudoaneurysm) in this setting is not at all reassuring because the likelihood
of an arterial origin remains high. The expectation would be for another
potentially life-threatening event within the next few days. If the arteriogram
is negative but the gastroduodenal artery can be embolized safely,
embolization should be strongly considered. An argument can also be made
for leaving the endovascular sheath in for the next 24 hours in order to
facilitate the next arteriogram. Lastly, I favor obtaining a 3 phase CT in this
setting to determine whether the axial image strongly suggests a target for
empiric embolization (Figure 1 ). A second advantage of the CT is the
ability to visualize a complete anastomotic failure, in which case completion
pancreatectomy becomes a stronger consideration.
It is worth noting that endoscopy is minimized in this algorithm. The
algorithm is tailored toward an intra-abdominal bleed, and sentinel bleeding
that occurs in delayed fashion alongside a concomitant pancreatic leak is far
more likely to be solved in surgical or endovascular fashion. This is true
whether the bleeding is noted intraluminally or extraluminally. Though
endoscopy does have a diagnostic role, this patient is expected to require an
endovascular or surgical approach. I favor proceeding down the
interventional pathway rather than dedicating hours to what is likely to be a
diagnostic procedure at best.
Surgical re-exploration continues to have a pivotal role in the
management of this scenario. However, my preference is to avoid surgery in
the setting of delayed bleeds whenever possible. If the patient’s
hemodynamics do not allow for an arteriogram or there is no treatable target
in IR, surgery is often the only viable strategy. If the site of bleeding cannot at
all be compartmentalized from an ongoing pancreatic leak, the site may
rebleed. In a stable patient that is likely to tolerate completion
pancreatectomy, this should be considered. If the patient remains unstable or
postoperative inflammatory changes are severe, completion pancreatectomy
is fraught with morbidity and wide drainage is favored. Any regional
omental/peritoneal tissue that can be used to protect the ligated vessel from
the leak should be mobilized to accomplish this.
SALIENT POINTS
Massive delayed hemorrhage is often preceded by a sentinel bleed.
The most critical aspect of the sentinel bleed is the recognition that any
bleeding that occurs remote from surgery in a stable patient that has not
previously manifested ongoing bleeding mandates an urgent evaluation
for arterial erosion.
Anastomotic leak and intra-abdominal abscess are risk factors for the
development of arterial erosions and delayed hemorrhage. Sentinel
bleed in the setting of an anastomotic leak is especially ominous and
associated with high rates of mortality.
Although utilized frequently, the therapeutic efficacy of endoscopy in
the setting of post-pancreatectomy hemorrhage and especially late PPH
is low. Endoscopy continues to have a prominent diagnostic role.
Management of delayed post-pancreatectomy hemorrhage continues to
shift from re-exploration to interventional radiology techniques which
are associated with high success rates in some series. Despite this,
reoperation remains a central part of the algorithm.
SELECTED REFERENCES
Pancreatic Surgery (ISGPS) definition. Surgery. 2007;142(1):20-5.
2. Yekebas EF, Wolfram L, Cataldegirmen G, Habermann CR, Bogoevski
D, Koenig AM, et al. Postpancreatectomy hemorrhage: diagnosis and
treatment: an analysis in 1669 consecutive pancreatic resections. Annals
of surgery. 2007;246(2):269-80.
3. Roulin D, Cerantola Y, Demartines N, Schafer M. Systematic review of
delayed postoperative hemorrhage after pancreatic resection. J
Gastrointestinal Surg. 2011; 15(6):1055-62.

CASE SCENARIO
A 73-year-old male underwent pancreaticoduodenectomy for resection of a
pancreatic head adenocarcinoma. His postoperative course was complicated
by a prolonged ileus. On post-operative day 12 the patient developed
worsening abdominal pain, prompting a CT scan of the abdomen revealing a
large upper abdominal fluid collection, which was successfully drained
under CT guidance. Fluid analysis revealed amylase-rich fluid and Gram
stain showed Gram-positive cocci and Gram-negative rods. A percutaneous
drain was left in place and antibiotic therapy was initiated. The patient
continued to improve until postoperative day 18 when he developed bright
red blood per rectum. Prior to any intervention, the patient also developed
hematemesis with associated hemodynamic instability and clinical
deterioration. There was no blood in the drain. What should be the next step
in management?
█ BACKGROUND
A pancreaticoduodenectomy (Whipple procedure) is typically performed to
remove neoplasms involving the head of the pancreas, ampulla, distal
common bile duct, or duodenum while restoring continuity of the biliary-
pancreatic system. Significant improvements in outcomes for this complex
procedure have been achieved since it was first performed in 1935. For
decades after its introduction, mortality hovered at ~25%, while today the
perioperative mortality rate in high-volume centers in the United States is
less than 3%. Despite the improvement in mortality, morbidity remains high,
with the most common postoperative complications being pancreatic
leak/fistula (40%), delayed gastric emptying (18%), wound infection (7%),
and intra-abdominal abscess (6%).
Postpancreatectomy hemorrhage occurs in 1-8% of all pancreatic
resections; however, it accounts for a disproportionate share (11-38%) of
overall mortality. This chapter will focus on the definition, clinical
presentation, and management of postpancreatectomy hemorrhage.
DEFINITION OF POSTPANCREATECTOMY
HEMORRHAGE (PPH)
In 2007, the International Study Group of Pancreatic Surgery (ISGPS)
developed an objective definition of PPH based on literature review and
consensus clinical experience. Since 2007, this definition has gained
widespread acceptance and has been utilized in numerous studies aimed at
guiding appropriate management of PPH. The ISGPS defines PPH based on
three objective parameters: onset, location, and severity.
Onset is defined as either early (<24 hours following completion of the
index operation) or late (>24 hours).
The location of postpancreatectomy hemorrhage is defined as either
intraluminal (gastrointestinal) or extraluminal (intra-abdominal).
The severity of bleeding may be either mild or severe based on volume
of blood loss, degree of clinical impairment, and need for intervention.
PPH is divided into three grades:
Grade A -- early mild hemorrhage
Grade B -- early serious hemorrhage or delayed mild hemorrhage
Grade C – delayed life-threatening hemorrhage
Early PPH is virtually always secondary to a technical error, with failure
to achieve appropriate hemostasis during the index operation. In contrast,
late PPH typically occurs from complications of the operation, with a usual
delay of several days to even weeks postoperatively. There are numerous
causes of late bleeding. Erosion of a peripancreatic vessel secondary to a
postoperative pancreatic fistula associated with intra-abdominal infection is
a common cause. Pancreatic fistula occurs in ~15% of patients after
resection of the pancreatic head and in ~25% of patients after distal
pancreatectomy. The greatest risk factor for fistula development is a normal
pancreas, with soft parenchyma and a small pancreatic duct. More than half
of the mortality after pancreatectomy can be attributed directly to pancreatic
fistula. Other causes include anastomotic ulceration or bleeding in
association with an arterial pseudoaneurysm that has developed as a
consequence of vascular trauma at the time of surgery or chemical injury
(pancreatic enzymes). Leakage of pancreatic enzymes at the
pancreaticojejunal anastomosis can lead to enzymatic digestion of the blood
vessel wall by trypsin, elastase, and other exocrine pancreatic enzymes.
As mentioned above, the location of PPH is defined as either intraluminal
(gastrointestinal) or extraluminal (intra-abdominal). Intraluminal bleeding
can occur from a variety of sources and includes the common sources of
gastrointestinal bleeding seen in the inpatient population, such as peptic ulcer
disease and esophagitis. However, in the post pancreatectomy population,
bleeding from suture lines of the anastomoses (eg, duodenojejunostomy after
pylorus-preserving pancreaticoduodenectomy, gastrojejunostomy after
classical pancreaticoduodenectomy, pancreaticojejunal or hepaticojejunal
anastomoses) must also be considered as possible sources of intraluminal
blood loss. Most commonly, late intraluminal hemorrhage occurs from the
pancreatic anastomosis or marginal ulcers. However, a pseudoaneurysm with
erosion into one of the anastomoses can also present as intraluminal
hemorrhage and must always be considered.
Extraluminal (intra-abdominal) hemorrhage occurs secondary to
involvement of larger vessels such as the hepatic artery, the stump of the
gastroduodenal artery, inferior pancreaticoduodenal artery, or the superior
mesenteric artery. Bleeding secondary to a postoperative pseudoaneurysm is
another important cause of late intra-abdominal hemorrhage, and a high index
of suspicion must be maintained for its presence in the setting of PPH, as the
risk of mortality is high. Patients with a pseudoaneurysm will often present
with a herald bleed, and it is critical to consider this diagnosis and
investigate on initial presentation. The clinical presentation and management
of PPH secondary to a pseudoaneurysm will be discussed below.
CLINICAL ASSESSMENT OF THE

BLEEDING PATIENT
As is the case with any patient in whom hemorrhage is suspected, initial
efforts should be directed towards stabilization and resuscitation. Immediate
establishment of large bore vascular access, securing necessary blood
products, and communication with consultative services is paramount.
Clinical assessment of the patient, including details of the operation
performed, and the presence of postoperative complications can offer clues
helpful in identifying the site of bleeding.
The distinction between intraluminal and intra-abdominal PPH is crucial
as the management options vary significantly depending on the source of
bleeding.
In broad terms, intraluminal bleeding is often managed by endoscopic
approach, and extraluminal bleeding by catheter-directed therapy or surgical
intervention. When bleeding occurs into the gastrointestinal lumen, the
clinical presentation typically includes hematemesis, bloody nasogastric tube
output, or melena. Signs compatible with extraluminal hemorrhage include
abdominal distension and the presence of blood in peritoneal drains. It
cannot be overstated, though, that even when intraluminal hemorrhage is
suspected, it is critical to consider the possibility of intra-abdominal
bleeding with erosion through an anastomosis.
MANAGEMENT OF PPH
Sentinel Bleed
All postoperative bleeding must be taken seriously and evaluated completely
until the etiology has definitively been identified. This is of particular
importance in the setting of PPH.
Even bleeding which is initially felt to be insignificant and self-limited can
be a harbinger of impending, life-threatening hemorrhage. Specifically, a
sentinel bleed is a well described postoperative occurrence in which a
seemingly minor bleeding episode serves as a warning sign of an impending
major hemorrhagic event. Sentinel bleeding has an overall prevalence of
33% and has been associated with higher relaparotomy and mortality rates
compared to patients without a sentinel bleed episode. Sentinel bleeding can
occur days after the index operation, and therefore, bleeding at any time in
the postoperative period warrants a great degree of vigilance and attention.
This includes even a small amount of blood that appears suddenly in the
surgical drains.
In cases of sentinel bleeding, CT angiography is recommended to make an
early diagnosis of a pseudoaneurysm (Figure 1 ). CT angiography is
preferred as the initial diagnostic modality as hemorrhage from ruptured
pseudoaneurysms cannot be completely ruled out based on negative
angiographic findings, and intermittent bleeding of ruptured pseudoaneurysms
is difficult to detect by angiography alone and may be missed. If a bleeding
site is identified, the patient should undergo emergent angiography to provide
endovascular treatment of the pseudoaneurysm. The mainstay of angiographic
treatment is coil embolization; however, in the setting of a short GDA stump,
placement of a covered stent in the common hepatic artery is often required
to maintain hepatic arterial patency.
FIGURE 1: CT angiogram demonstrating pseudoaneurysm of

gastroduodenal artery stump and intra-abdominal hematoma. A)
Axial CT image and B) coronal CT image with pseudoaneurysm
(arrow) and surrounding hematoma (arrow head).
Intraluminal PPH
When an intraluminal source of bleeding is suspected, endoscopy should be
performed. Historically, there has been concern regarding the performance of
endoscopy in the immediate postoperative period secondary to mechanical
forces from torque and air insufflation leading to anastomotic dehiscence. It
is now generally accepted that endoscopic intervention is safe, even in the
immediate postoperative period. When possible, carbon dioxide (CO2),
which is rapidly absorbed by the GI mucosa, should be used as the
insufflation agent for endoscopic procedures. CO2 is absorbed from the GI
tract approximately 160 times faster than nitrogen, the major gaseous
ingredient of ambient air. It is passively absorbed through the mucosal lining
into the bloodstream and eventually exhaled through the lungs. If a bleeding
source is identified, standard endoluminal hemostatic therapies (ie, thermal
and mechanical techniques) are safe in the postoperative period.
False Intraluminal PPH

A critical component in the evaluation of an individual with suspected
intraluminal bleeding is the consideration that this presentation may actually
represent intra-abdominal bleeding with erosion through an anastomosis.
Early recognition is necessary for appropriate management. It is reasonable
to triage a patient with suspected intraluminal bleeding to endoscopy for
diagnosis and treatment if the patient is hemodynamically stable. However, it
is of the utmost importance that unless an endoscopically identifiable source
is recognized, these individuals must proceed to emergent CT angiography
for further investigation of the bleeding source.
Extraluminal PPH
Blood in peritoneal drains is frequently an ominous sign and always warrants
performing an urgent CT angiogram to evaluate the source of bleeding. The
extraluminal sites of bleeding postpancreaticoduodenectomy are numerous
and include the GDA stump, portal vein tributaries, hepatic artery branches,
superior mesenteric vein tributaries (including uncinate vessels), branches of
the superior mesenteric artery (including jejunal mesenteric arterial branches
to the left and inferior pancreaticoduodenal artery to the right), the cut
pancreatic surface, and the gallbladder fossa. It is optimal for the surgical
team to be present in the angiography suite when these procedures are
performed to ensure that there is close communication between the surgeon
and interventional radiologist performing the procedure regarding all of the
potential sources of bleeding and pseudoaneurysm. For example, while
gastroduodenal or common hepatic artery pseudoaneurysms are most
common, a pseudoaneurysm of the inferior pancreaticoduodenal artery can
also occur and careful pull-back angiogram to evaluate the origin of this
vessel from the proximal SMA is critical to avoid missing this potential
bleeding source.
As described above, the most common etiology of extraluminal
hemorrhage in the early (<24 hours) postpancreatectomy population is the
result of technical failure to properly secure the inferior pancreaticoduodenal
artery, or less commonly the gastroduodenal artery stump. Many patients with
early extraluminal bleeding present with acute hemodynamic instability with
associated blood in the operative drains. Historically, these patients have
required emergent reoperation for hemostasis, and this remains the most
widely accepted approach to the management of early PPH.
In contrast to early PPH, a strategy of endoscopic and angiographic
intervention has been utilized in the setting of delayed PPH, with urgent
angiography as the mainstay of treatment. As previously discussed, late PPH
typically results from complications of the operation such as intra-abdominal
abscess, erosion of a peripancreatic vessel secondary to a pancreatic fistula,
ulceration of the site of an anastomosis, or in association with an arterial
pseudoaneurysm. Late PPH frequently presents after hospital discharge,
leading to delays in recognition and treatment, and therefore significant
associated mortality.
SUGGESTED APPROACH
Returning to the case, this patient underwent emergent CT angiography,
which revealed a large, contrast-filled cavity around the
pancreaticojejunostomy. A ruptured pseudoaneurysm from the GDA stump
was suspected. Conventional angiogram of the celiac trunk revealed a
pseudoaneurysm at the GDA stump with active extravasation. Transarterial
coil embolization was successfully performed (Figure 2 ).
FIGURE 2: Celiac angiography performed for evaluation of

suspected gastroduodenal artery pseudoaneurym. A) Digitally
subtracted angiogram of the common hepatic artery with B) active
contrast extravasation for the GDA stump (arrows). C) Coil
embolization (arrow) of the GDA stump and D) postembolization
angiogram with successful occlusion.
The successful management of this patient’s late PPH hinged on a high

index of suspicion on presentation and close communication between
members of multiple specialties caring for this patient. This patient’s initial
presentation of hematemesis and rectal bleeding was suggestive of
intraluminal bleeding. However, in light of the known pancreatic fistula, the
surgical team maintained a high level of suspicion for an extraluminal source
and opted to pursue CT angiography first, given the patient’s hemodynamic
instabililty. The need for rapid consultation and coordination of care between
multiple disciplines in the management of PPH cannot be overstated.
This case highlights an important clinical scenario where an extraluminal
bleeding source masqueraded as an intraluminal source, with blood entering
the lumen of the intestine through the ruptured pancreaticojejunal
anastomosis.
SALIENT POINTS
All PPH must be evaluated urgently and completely until a source is
identified.
Grade B early PPH (<24 hours) is usually best treated by return to the
operating room.
Endoscopy and endoscopic interventions for intraluminal
gastrointestinal sources of bleeding are safe in the early postoperative
period.
Providers must maintain a high level of vigilance for extraluminal
bleeding presenting as luminal blood loss.
The surgical team must remain actively involved in the resuscitation
and management process, including being present in the Interventional
Radiology suite as angiography is taking place.
SELECTED REFERENCES
1. Leichtle SW, Kaoutzanis C, Mouawad NJ, Welch KB, Lampman R,
Hoshal VL Jr, et al. Classic Whipple versus pylorus-preserving
pancreaticoduodenectomy in the ACS NSQIP. J Surg Res
2013;18(1)3:170-6.
2. Cameron JL, Riall TS, Coleman J, Belcher KA. One thousand
consecutive pancreaticoduodenectomies. Ann Surg 2006;244(1):10-5.
Pancreatic Surgery (ISGPS) definition. Surgery 2007;142(1):20-5.
4. Fogel EL, Shahda S, Sandrasegaran K, DeWitt J, Easler JJ, Agarwal, et
al. A Multidisciplinary Approach to Pancreas Cancer in 2016: A
Review. Am J Gastroenterol 2017;112(4):537-54.
5. Khalsa BS, Imagawa DK, Chen JI, Dermirjian AN, Yim DB, Findeiss
LK. Evolution in the Treatment of Delayed Postpancreatectomy
Hemorrhage: Surgery to Interventional Radiology. Pancreas
2015;44(6):953-8.
6. Committee AT, Lo SK, Fujii-Lau LL, Enestvedt BK, Hwang JH, Konda
V, et al. The use of carbon dioxide in gastrointestinal endoscopy.
Gastrointest Endosc. 2016;83(5):857-65.
7. Evans DB, Amini A, Christians KK. Chapter 26. Postpancreatectomy
Hemorrhage: Early and Late. In: Pawlik TM, Maithel SK, Merchant NB
(Eds.) Gastrointestinal Surgery. 271-280. New York, New York,
Springer New York; 2015
Acknowledgments: The authors would like to thank Dr. Sanjeeva Kalva,
Chief and Professor, Interventional Radiology, UT Southwestern Medical
Center, for the CT angiography and angiographic images provided as
figures in this chapter.

CASE SCENARIO
A 71-year-old woman presents with weight loss and obstructive jaundice
secondary to a pancreatic head tumor and undergoes a pylorus-preserving
pancreatoduodenectomy. Histologically, the resected tumor confirmed an
adenocarcinoma with clear resection margins; 2/13 lymph nodes harvested
resulted positive for metastases. Her immediate postoperative course was
uneventful, without evidence of pancreatic fistula. On the eighth
postoperative day she presented with abdominal distention and milky white
discharge draining 600-800 ml per day from each drain site. Biochemical
analysis of the drainage fluid revealed triglycerides of 290 mg/dL. A low-fat
medium-chain triglyceride diet was started and the chylous ascites resolved
completely 4 weeks later.
█ BACKGROUND
Chylous ascites (CA) is a rare complication following abdominal surgery,
with a frequency of only one in 20,464 abdominal operations. The main
cause of postoperative CA is due to injury to the cisterna chyli or its
tributaries. CA has been suggested to be more common following extensive
retroperitoneal dissection for malignant tumors. During pancreatic surgery
the proximity of the pancreas to the cisterna chyli indicates that CA could be
more frequently seen in this type of surgery. However, CA after pancreatic
surgery has been reported in relatively few cases and its incidence and
natural history is poorly known.
Recently, several series of CA in pancreatic surgery have described more
accurately the incidence, diagnosis, and management of this condition. CL
occur after pancreatic surgery about 1.8-13% of the time, between different
series, and a recent increase in the frequency of this complication has been
observed due to more advanced stages of pancreatic cancer that are
intervened upon, possibly with more extensive lymph node resections.
Although their incidence may be low, chyle leaks can be associated with
significant morbidity including dehydration, wound complications,
debilitation, malnutrition, immunosuppression, and delayed delivery of
adjuvant chemotherapy—thus affecting the long-term survival of these
patients.
Distinct definitions have been used to classify CA. It has been described
as the secretion of milky white fluid with high triglyceride content with a
secretion range from above 110 to 130 mg/dL. Another controversy on
defining this complication is the quantity of fluid drained necessary to be
classified as CA; because of this controversy, it is sometimes described
sometimes as a chyle leak (CL). Alternatively, it can be considered chylous
ascites when the volume drained is above 100 to 600 ml and there is diffuse
chyloperitoneum. The International Study Group of Pancreatic Surgery
(ISGPS) defined chyle leak as the output of milky-colored fluid from a drain,
drain site, or wound, on or after postoperative day 3, with a triglyceride
content ≥ 110 mg/dL. This contemporary definition is recommended to
standardize the report of complications after pancreatic surgery between
different centers.
CA is usually recognized until 5th or 6th postoperative day and generally
increases when oral intake is initiated. The majority of patients present with
abdominal distention and discharge of white milky fluid through either the
drain, former drain site, or surgical wound.
Many studies have analyzed factors associated with the formation of CL
after surgery; those found to significantly increase the incidence of CA are
early restoration of enteral feeding, number of lymph nodes harvested,
vascular resection and reconstruction, manipulation of the para-aortic area,
and retroperitoneal tumor invasion.
The type of surgery performed has also been related with the frequency of
CL. During pancreatoduodenectomy, the proximity of the head of the
pancreas to the cisterna chyli promotes the development of a postoperative
CL, although recently some studies have reported a comparable rate of CL
following distal pancreatectomy. The indication of surgery has also been
associated with CL with the majority of the cases relating to malignancy.
This is due to invasion of adjacent structures of malignant tumors, and the
requirement of more aggressive intervention to perform complete resections.
MANAGEMENT OF CHYLE LEAKS

The initial treatment aims at reducing lymph flow through the damaged
lymphatic vessels. This is attempted mainly with conservative measures,
which achieves success rates from 75-100%. The traditional treatment for
CA is restriction of diet to medium-chain triglycerides, with the purpose of
reducing the flow of lymph. This treatment option is widely available and
affordable—avoiding secondary effects due to total parenteral nutrition
(TPN), and is the preferred option for patients who can tolerate oral intake.
Although most of studies have shown success, this treatment is not
effective in every case. TPN is an option for management and has shown
positive outcomes in several reports. TPN can markedly reduce lymphatic
flow and rectify nutritional deficiencies. TPN in and of itself has high
reported rates of success, although combination with somatastatin analogues
(e.g., Octreotide) could further significantly decrease the flow of lymph.
Administration of Octreotide has shown efficacy in the treatment of these
patients, with some studies revealing a significant decrease in the discharge
of fluid after its administration. Although positive results have not been
consistent in other reports, the promising results of this management justify
its use in patients with this condition. In addition, patients treated with
Octreotide resumed oral intake significantly sooner, and abdominal drains
were removed sooner.
The minority of patients fail conservative management, and more invasive
options may become necessary, including laparotomy, peritoneo-venous
shunting. Percutaneous catheter drainage of collections or chylous ascites is
advised only in case of clinical symptoms.
Lymphangiography is used to define the site of CL for operative ligation,
and, in some patients, it can be fully successful therapeutically. Success rates
of lymphangiography have been reported up to 56 to 86%, with those failing
this method undergoing subsequent surgical ligation. Disadvantages of this
management include the difficulty of cannulating the lymphatic vessels, and
its rare application in clinical practice. Other methods that have been used
for detecting CL are lymphoscintigraphyor MR lymphography; however,
these methods have not been associated with resolution of the CL. Another
option for these patients is the use of embolic agents during
lymphangiography or percutaneous procedures; however, they are not widely
used, and results of these studies are not consistently valuable.
Peritoneo-venous shunting helps decompressing chyle from peritoneum
into the systemic circulation. It is usually considered for patients unfit for
surgery. Success rates of this treatment are highly various—reported as poor
in most of studies but up to 100% in some others. Complications associated
with this approach are sepsis, disseminated intravascular coagulation,
hypokalemia, small bowel obstruction, and risk for air embolism. In
addition, it has been reported that the high viscosity of chyle fluid results in
an elevated rate of shunt occlusion in the majority of the cases. In cases of
refractory CL, surgical ligation is indicated; however, experience with
operative management after pancreatic surgery is very limited, and it is not
always satisfying.
Because it is so rare of a condition, there are no studies comparing
different management strategies, and there is no consensus on the appropriate
treatment approach of patients with CL. Furthermore, it is recommended that
a “step-up” approach, starting with dietary restrictions (relying upon a MCT-
diet). If patients fail in this treatment, then opt for TPN. Percutaneous
drainage should be employed in symptomatic patients and, in refractory
cases, more invasive treatment options should be considered. The reported
resolution of CL can be as few as eight days and as long as 232 days, in
some cases.
COMPLICATIONS ASSOCIATED WITH

CHYLE LEAKS
Described complications associated with CA are highly varied including:
abdominal abscess, concomitant pancreatic fistula, malnutrition, peritonitis,
sepsis and increased incidence of portal or mesenteric venous thrombosis.
Chyle leak has been associated with significantly greater duration of
abdominal drainage, longer hospital stay and delayed delivery of adjuvant
chemotherapy, thus affecting long-term survival of these patients.
CONCLUSION
Chyle leak, including chylous ascites, is a clinically relevant complication
after pancreatic surgery, but conservative management is successful in the
majority of cases. The new standardized definition posed by the International
Study Group of Pancreatic Fistula (ISGPS) will allow unified reporting of
this complication and hopefully consensus regarding the appropriate
approach to patients with this condition.
SALIENT POINTS
Chyle leak, as defined by IGSPS, is the output of milky-colored fluid
from a drain, drain site, or wound, on or after postoperative day 3,
with a triglyceride content ≥110 mg/dL.
Chyle leak, including chylous ascites, is a clinically-relevant
complication associated with longer hospital stay and complications
such as malnutrition, inmunosuppression, and generalized debilitation.
The initial treatment approach in these patients should be dietary
restrictions, as these are successful in the majority of cases.
Total parenteral nutrition should be considered when dietary
restrictions fail to reduce the degree of the leak.
Percutaneous catheter drainage of collections of chylous ascites is
advised only in the case of clinical symptoms.
Lymphangiography alone or with embolic agents, peritoneovenous
shunt, and surgical ligation should only be performed in cases of chyle
leak refractory to conservative management.
SELECTED REFERENCES
1. Kuboki S, Shimizu H, Yoshdome H, Ohtsuka M, Kato A, Yoshitomi H,
et al. Chylous ascites after hepatopancreatobiliary surgery. Br J Surg.
2013; 100(4): 522-7.
2. Besselink MG, Van Rijssen LB, Bassi C, Dervenis C, Montorsi M,
Adham M, et al. Definition and classification of chyle leak after
pancreatic operation: A consensus statement by the International Study
Group on Pancreatic Surgery. Surgery 2017;161(2):365-372.
3. Tabchouri N, Frampas E, Marques F, Blanchard C, Jirka A, Regenet N.
Chylous ascites management after pancreatic surgery. World J Surg.
2017; 41(4):1054-1060.
4. Kawasaki R, Sugimoto K, Fujii M, Miyamoto N, Okada T, Yamaguchi
M, et al. Therapeutic effectiveness of diagnostic lymphangiography for
refractory postoperative chylothorax and chylous ascites: correlation
with radiologic findings and preceding medical treatment. AJR Am J
Roentgenol. 2013;201(3):659-66.
5. Yarmohammadi H, Brody LA, Erinjeri JP, Covey Am, Boas FE, Ziv E,
et al. Therapeutic Application of Percutaneous Peritoneovenous
(Denver) Shunt in Treating Chylous Ascites in Cancer Patients. J Vasc
Interv Radiol. 2016;27(5):665–673.

CASE SCENARIO
A 68-year-old male with an extrahepatic cholangiocarcinoma found on
exploration to extend into the distal common bile duct underwent
pancreatoduodenectomy with a relatively high bilioenteric reconstruction. On
postoperative day two, the output from his surgical drains, which had been
previously serosanguinous, was noted to be overtly bilious. The total drain
output for the past 24 hours was approximately 60 milliliters and had not
increased over that period. He had been afebrile with a normal white blood
cell count. His midline abdominal pain was not atypical.
█ BACKGROUND
POSTOPERATIVE ANASTOMOTIC
FAILURE FOLLOWING
PANCREATODUODENECTOMY
Reconstruction of the foregut during pancreatoduodenectomy (PD) involves
creation of three intestinal anastomoses: pancreatico-jejunostomy or -
gastrostomy, hepaticojejunostomy, and gastro- or duodenojejunostomy. Each
of these anastomoses carries the potential for dehiscence or leak. The suture
line at the blind end of the bilopancreatic limb might also dehisce, increasing
diagnostic and therapeutic difficulty. Leak from any of these sites can be
innocuous and self-limited, or they can be difficult to control, resulting in
clinically relevant morbidity and increased risk of mortality. In patients with
malignancy, anastomotic leak can also result in potential delay to receipt of
adjuvant chemotherapy.
The pancreaticojejunostomy (PJ) is by far the most likely of the three
anastomoses to fail. The technical components of the PJ, the risk factors
associated with clinically-relevant post-operative failure, and the
management strategies for pancreaticoenterostomy failure have been subject
to the most intense study. By comparison, the incidence of a leak or fistula
that does not involve the PJ is much lower. For instance, isolated
hepaticojejunostomy (HJ) leaks are reportedly between 3-8%. The etiology
and best practices concerning the management of leak at the HJ and other
sites of bilious fistula have been less well studied. There is no substantive
body of empiric evidence to guide management of these clinical problems.
As a result, we are left to extrapolate from the management principles of
other types of gastrointestinal fistulas or to rely on personal experience with
bile leak to make treatment decisions. In this chapter, we outline our
approach to the management of patients with early post-PD bile leak as
manifest by bilious output from the surgical drains on postoperative day two.
Figure 1 summarizes our clinical decision-making pathway.
FIGURE 1: Flow diagram summarizing management of early
hepaticojejunostomy leak after pancreatoduodenectomy. (POD:
Postoperative Day; HJ: Hepaticojejunostomy; POPF: Postoperative
Pancreatic Fistula)
CLINICAL DECISION-MAKING IN THE
EARLY UNCONTROLLED BILE LEAK
For postoperative gastrointestinal leak in general, the clinical condition of
the patient dictates early decision-making. This is true for patients following
simple operations, such as enterolysis for adhesive small bowel disease, and
it is true for early biliary leak or fistula following PD.
In rare instances, the presence of enteric organisms in the peritoneal
cavity can result in hemodynamic instability that places the patient at
immediate risk for multisystem organ failure and death. In these cases, the
enteric leak is uncontrolled by definition. The most prudent approach is re-
exploration with the intent to more definitively control the site of the leak.
Control can be achieved upon re-exploration with either immediate repair of
the anastomosis or surgical placement of better drainage. If the patient is in
septic shock as a result of an anastomotic leak two days following PD, urgent
re-exploration is the most appropriate next step in management. In these
circumstances, septic shock will be manifest as progressive hypotension
without another identifiable cause (e.g., no evidence of hemorrhage, acute
heart failure, or pulmonary embolus). Re-exploration will effectively identify
the site of the leak and affords the safest and most efficient method for
obtaining control.
What is done during the re-exploration depends on the location of the
leak. If the leak is from the PJ, it is almost always technically impossible to
improve or revise the prior anastomosis. Instead, repositioning the surgical
drains, or providing additional drainage of the site, is the only viable
management strategy. If the leak is at the HJ, the method of management
depends on the physical appearance of the proximal common hepatic or
common bile duct. If there is a healthy-appearing margin of bile duct above
the site of the dehiscence, the bilioenteric anastomosis could be redone. In
these circumstances, our practice is to place a T-tube in the duct above the
revision and run a short limb of the tube across the revised bilioenteric
anastomosis. A closed suction drain is then placed behind the revised
anastomosis. The T-tube is left to gravity drainage. When the clinical sepsis
has resolved, a cholangiography is performed through the T-tube. If there is
no evidence of contrast extravasation, the T-tube is capped, and the closed
suction drain removed. The T-tube is then routinely left in place for 6 weeks
regardless of drain output to act as a stent for the revised anastomosis. If, at
the re-exploration, there is no healthy-appearing margin of bile duct above
the site of the dehiscence, then the best approach is to reposition the surgical
drains to better control the site of the leak. The intent here is to control the
leak and to pursue a more definitive revision of the anastomosis after the
sepsis has resolved. In these cases, percutaneous placement of transhepatic
catheters may be employed as a second step in effort to avoid interval
surgical revision of the HJ weeks later. Most other sites of early leak that
manifest as hemodynamic instability (e.g., leak from the gastrojejunostomy,
the blind end of the biliopancreatic limb, or a missed enterotomy) may be
effectively revised at the time of the re-exploration.
Overall, an operative intervention is more likely to save the patient with
hemodynamic compromise associated with an early uncontrolled anastomotic
leak after PD than an image-guided percutaneous drainage or an unduly
prolonged resuscitation in the ICU.
CLINICAL DECISION-MAKING IN THE

EARLY CONTROLLED BILE LEAK
Fortunately, early bile leak following PD rarely presents as septic shock with
hemodynamic lability. Biliary leak is rare, in general, and most leaks will
begin as a low-volume fistula that will egress via appropriately placed
surgical drains. More commonly, patients with early bile leak present with
signs of sepsis without multisystem organ failure or cardiovascular collapse.
Patients may have fever, low grade tachycardia, leukocytosis, increased fluid
requirements, and confusion, but will maintain a normal blood pressure and
be stable enough to forgo transfer to the ICU. Frequently, the character of
their surgical drains will also change from serous to bilious. These patients
have a leak that is more well-controlled by the operative drains, and
surgeons will have the luxury of considering non-operative methods for
definitive leak management.
The most effective interventions are determined by two features of the
leak: (1) the location of the leak and (2) the volume of the leak or the degree
of the anastomotic dehiscence. Tendencies to intervene surgically must be
balanced by the understanding that re-exploration always carries potential
risk for further intestinal injury but does not always have the potential to
make a given situation better. In these instances, when one can take the time
to do it, it is prudent to carefully identify the site and to characterize the
degree of the dehiscence to determine as best possible what intervention, if
any, stands to make the patient’s recovery safer and more expeditious.
CONFIRMING CONTROL OF THE LEAK

A minority of patients will tolerate an uncontrolled leak without manifesting
hemodynamic lability. These individuals may have significant anastomotic
disruptions that are not draining completely through the surgical drains and
thus have large intraperitoneal fluid collections that will not resolve without
additional drainage. For this reason, in patients who appear to have a leak
controlled by their surgical drains, confirming control of the leak is the most
sensible first step in management. Confirming control of the leak is best done
by means of axial CT imaging. Magnetic resonance (MR) imaging can be
used; but, it involves leaving a patient at risk for septic decompensation
unsupervised for a long period of time and does not provide any substantive
information beyond CT. If imaging studies demonstrate undrained fluid
collections, a decision must be made between re-exploration versus image-
guided percutaneous drain placement. In general, the decision between these
two options is individualized and based on the size of the collections. In our
opinion, small, single, localized fluid collections can be managed with
image-guided percutaneous procedures, whereas larger, more extensive,
collections likely requiring multiple percutaneous procedures are best
managed with operative re-exploration.
IDENTIFYING POINT OF ORIGIN FOR THE

LEAK
As alluded to above, the site of the leak is relevant to its management. A
dehiscence of the PJ is generally best managed by image-guided
percutaneous drainage and not by operative re-exploration. A dehiscence of
the gastrojejunostomy (GJ) can generally be revised, and is best treated by
early surgical revision. A dehiscence of the HJ, may be amenable to surgical
revision or may be better managed by percutaneous stenting. Unfortunately, it
may not be possible based on physical exam and drain effluent alone to
determine the site of origin of the leak. For patients undergoing distal
pancreatectomy and hepatectomy, leak or fistula will be simple pancreatic
fluid (pale and clear) and bile (usually amber to green and also clear),
respectively. For patients undergoing PD and suffering an anastomotic
dehiscence from any site, that leak will be by definition mixed: containing
some combination of pancreatic fluid, bile and enteric contents from the
jejunal limb used in the reconstruction. Because of this, it may be difficult to
determine by drain character alone which anastomosis has broken down.
Commonly, a dehiscence of the PJ will have more pancreatic fluid in the
effluent than bile, and the drain effluent will appear more ashen grey than
would be the case for a leak from the HJ. By comparison, a dehiscence of the
HJ will have more bile in the effluent and less enteric fluid than a leak from
the GJ (or duodenojejunostomy) and will be yellow-to-yellow-green colored
and thin. However, in many cases, particularly high-volume leaks, the drain
effluent will be dark green and indistinguishable from small intestinal succus
regardless of which of the three anastomoses have dehisced.
BIOCHEMICAL STUDIES OF THE DRAIN

EFFLUENT MAY OFFER INSIGHT INTO THE
SITE OF ORIGIN.
The International Study Group in Pancreatic Surgery (ISGPS) identifies drain
amylase levels greater than three times the upper limit of the normal as a
postoperative pancreatic fistula. Similarly, the International Study Group of
Liver Surgery (ISGLS) identifies drain bilirubin levels greater than three
times the patient’s serum level as a bile leak. Interpreting the drain bilirubin
and amylase levels in tandem can aid in the localization of the leak: leaks
from the PJ will have higher levels of amylase and relatively low levels of
bilirubin, whereas leaks from the HJ will have elevated levels of bilirubin
and lower levels of amylase. A ratio of drain amylase-to-bilirubin has not
been defined to distinguish a PJ from a HJ leak. Nor has consensus been
achieved regarding a grading system. Frequently, biochemical studies will
yield a mixed result with elevated bilirubin and amylase levels. In these
cases, a drain sinogram done by injecting contrast through the surgical drain
under fluoroscopy can aid in clarifying the site of leak. Presence of
pancreatic fistula should take management precedence over potential HJ leak
given that management is similar, and postoperative pancreatic fistulas can
be worsened in the presence of bile.
IDENTIFYING THE MOST EFFICIENT

ROUTE TO RECOVERY
Once a leak from the HJ is confirmed, management is predicated on best
clinical judgment: what method is most likely to result in as efficient a return
to health for the patient as possible? There is little empiric evidence to guide
decision-making. There are also limited numbers of therapeutic strategies
available to the treating surgeon. If the gastro- or pyloro-jejunostomy has
been done as a loop of the pancreaticobiliary limb distal to the HJ, it may be
possible to access the site endoscopically. However, even the most
aggressive endoscopic interventionalists may be reluctant to attempt
endoscopic retrograde cholangiography and stenting of a fresh anastomosis
for reasonable fear of worsening a partial dehiscence.
Beyond endotherapy, there are two interventional options available to
manage the early controlled HJ leak: (1) operative revision and (2)
percutaneous transhepatic cholangiocatheters. In our experience, both
strategies have finite ability to facilitate recovery, but neither approach
results in a rapid resolution of the problem. The trap with operative revision
is that most HJs in PD are constructed by experienced surgeons, and were a
significant leak to happen there is almost always an underlying reason (e.g.,
small friable duct, poor preoperative nutrition, other mechanisms for poor
wound healing) that is not obviated by having the same surgeon reconstruct
the HJ. If the bile duct was thin and susceptible to dehiscence at the time of
the index operation, then it will most certainly be even more susceptible to
dehiscence at the time of re-exploration. Although in rare circumstances a
knot coming undone or a suture rupturing might be the cause of a leak and
appear to warrant re-exploration, these events are also more likely to heal
without re-exploration. While a T-tube can be placed at the time of the
second operation and may act to mitigate risk of re-leak, such a stent can
usually be placed percutaneously through the liver.
In contrast to re-exploration, percutaneous transhepatic catheters can be
inserted without general anesthesia and with limited risk of worsening any
partial anastomotic dehiscence. However, there are several disadvantages of
percutaneous drainage, which can include: (1) not all leaks are amenable
(e.g., some anastomotic disruptions are so large that catheters cannot be
placed across the anastomosis); (2) healing of the leak may be protracted
requiring months of percutaneous drainage; and, (3) the percutaneous drains
must traverse the liver parenchyma and might cause subcapsular hematoma,
abscess, or suprahepatic bile leak.
Given the lack of empiric evidence and limited number of therapeutic
options available, decision-making in cases of controlled HJ fistula are best
managed in an interdisciplinary fashion with collaboration between the
interventional radiologist and the operating surgeon. It is our approach to
base our treatment decisions on what we believe will result in the safest and
fastest road to recovery for the patient. For patients with low-volume leaks
for whom drain sinograms demonstrate a small anastomotic dehiscence, we
avoid additional intervention and optimize nutrition, enterally or
parenterally. For patients with evidence of more significant anastomotic
dehiscence (i.e., drain outputs of >200 milliliters/day, or drain sinograms
with >30% anastomotic dehiscence), we employ an intervention intended to
provide a repair for the leak and/or a transanastomotic stent across the site of
the leak. Our preference in these situations is to use percutaneous
transhepatic catheters. If this cannot be achieved because either the
anastomosis cannot be traversed by percutaneous means, or there is near
complete dehiscence on sinogram, we resort to operative revision using a T-
tube.
NUTRITION IN EARLY BILE LEAK

For all patients undergoing PD, nutrition is an issue as they uniformly require
some time (weeks) to return to adequate oral calorie intake. Many patients
with pancreaticobiliary malignancy are nutritionally depleted at the time of
the attempted resection. Some will have obstructive jaundice and consequent
fat malabsorption in the weeks prior to surgery. Others will receive
chemotherapy prior to surgery, develop impaired appetite, and require
increased caloric needs in the weeks prior to surgery. For those who have a
post-operative bile leak, nutrition is an even more significant issue as sepsis
increases caloric requirements further, and the leak may result in prolonged
ileus. As soon as a leak is identified, nutrition is a foremost concern.
There is no empiric evidence to suggest that oral nutrition worsens a
dehiscence, whereas there is significant evidence to suggest that bacterial
translocation and gut-derived sepsis is a substantial potential cause of
morbidity. For these reasons, our practice is to attempt to feed patients
enterally whenever safe. We do not routinely place feeding tubes at the time
of the PD, however. In the context of an HJ leak, for patients with bowel
function, we attempt an oral diet. If they are able to tolerate it and there is no
evidence of increased output from the leak we continue to let them eat. If they
have associated delayed gastric emptying (DGE) we will attempt nasoenteral
feeding. If they cannot tolerate nasoenteral feeds, we resort to total parenteral
nutrition (TPN) and will usually start that within a week of the diagnosis of
leak. For patients that tolerate nasoenteral feeding but have prolonged DGE
(>4 weeks), we will pursue peroral gastrostomy with subsequent exchange to
jejunostomy tube and use this approach for medium-term nutritional support.
ANTIBIOTIC USE IN EARLY BILE LEAK
As alluded to above, anastomotic failures following PD are by definition
mixed and will result in enteric bacteria spilling into the peritoneum. Most
patients with leak will experience some degree of sepsis. Our practice is to
routinely culture the biliary system at the time of the initial PD and to start
antibiotics appropriate for those cultures at the time that a patient shows
signs of leak-associated sepsis. Additional cultures are submitted from blood
when patients demonstrate hemodynamic compromise and from new
percutaneous drains as they are placed. Antibiotics are tailored to ensure that
organisms grown from cultures are covered and continued until there is no
clinical evidence of sepsis: no fever, normal white blood cell count,
hemodynamically normal.
CASE SCENARIO REVISITED

Our patient was diagnosed with a controlled HJ leak on postoperative day 2
after PD. Biochemical studies from the surgical drains confirmed the bile
leak: drain amylase 100 ng/dL (serum upper limit of normal 54 ng/dL) and
drain bilirubin 15 ng/dL (serum upper limit of normal 1.4 ng/dL). Axial CT
imaging confirmed no undrained intra-abdominal fluid collections, and the
surgical drains appeared to be in appropriate positions. Accordingly, we
elected non-operative measures. He was started on broad spectrum
antibiotics (zosyn) per operative cultures. He was able to tolerate an oral
diet and had normal bowel function. On postoperative day 10, his drain
output increased to approximately 250 mL/day, suggesting a more significant
anastomotic dehiscence that was at long-term risk for persistent fistula or
stricture. A percutaneous transhepatic cholangiocatheter was placed across
his anastomosis (Figure 2 ). By postoperative day 16, output from his
surgical drain had stopped. Cholangiography through the PTD on day 21
demonstrated no significant leak. His PTD was capped and his surgical drain
was removed. He was discharged on postoperative day 24 tolerating an oral
diet with his percutaneous transhepatic cholangiocatheter in place. After
three months post op, his percutaneous transhepatic cholangiocatheter drain
was removed.
FIGURE 2: (A) Percutaneous transhepatic cholangiography in a

patient presenting with bilious output from his surgical drain two
days following pancreaticoduodenectomy for a distal
cholangiocarcinoma. He was initially clinically stable with low
outputs from his drains. We elected conservative management. His
drain outputs increased over time to the point that ten days post op
he had 250 mLs of drain output daily. We elected percutaneous
transhepatic drain (PTD) placement (B). With this intervention he
recovered. His PTD was removed three months following surgery.
SALIENT POINTS
Isolated hepaticojejunostomy (HJ) leaks from pancreatoduodenectomy
(PD) are rarer than leaks from the pancreaticojejunostomy (PJ), with
published rates ranging from 3-8%.
HJ leaks are diagnosed by examining the character of the surgical drain
effluent, measuring the amylase and bilirubin levels from the surgical
drain effluent, and confirming with imaging studies.
Patients with early HJ leaks and clinical instability should be re-
explored in the operating room with the intent to either revise the HJ or
to ensure appropriate drainage of the leak with drains (i.e., obtaining
control).
Use of percutaneous transhepatic cholangiocatheters is our preferred
management strategy for clinically stable patients with HJ leaks.
Inability to place or failure to improve with percutaneous transhepatic
cholangiocatheters typically requires re-exploration using a T-tube.
Adequate nutritional support and appropriate use of antibiotics is
essential to promote healing of the dehiscence.
SELECTED REFERENCES
1. Jester AL, Chung CW, Becerra DC, Molly Kilbane E, House MG,
Zyromski NJ, et al. The Impact of Hepaticojejunostomy Leaks After
Pancreatoduodenectomy: a Devastating Source of Morbidity and
Mortality. J Gastrointest Surg. 2017; 21(6):1017-1024.
2. Malgras B, Duron S, Gaujoux S, Dokmak S, Aussilhou B, Rebours V, et
al. Early biliary complications following pancreaticoduodenectomy:
prevalence and risk factors. HPB (Oxford). 2016;18(4):367-74.
3. Andrianello S, Marchegiani G, Malleo G, Pollini T, Bonamini D,
Salvia R, et al. Biliary fistula after pancreaticoduodenectomy: data from
1618 consecutive pancreaticoduodenectomies. HPB (Oxford).
2017;19(3):264-269.
4. Antolovic D, Koch M, Galindo L, Wolff S, Music E, Kienle P, et al.
Hepaticojejunostomy--analysis of risk factors for postoperative bile
leaks and surgical complications. J Gastrointesl Surg. 2007;11(5):555-
61.
5. Burkhart RA, Relles D, Pineda DM, Gabale S, Sauter PK, Rosato EL, et
al. Defining treatment and outcomes of hepaticojejunostomy failure
following pancreaticoduodenectomy. J Gastrointest Surg.
2013;17(3):451-60.

CASE SCENARIO
A 67-year-old woman with type II diabetes was referred to us for evaluation
of a pancreatic mass found incidentally on imaging for other reasons. Her
husband had passed away from pancreatic cancer 12 years earlier. Her MRI
revealed a mass in the proximal body of the pancreas (Figure 1 ). Imaging
was consistent with a hypervascular tumor of the pancreas with features
consistent with a pancreatic neuroendocrine tumor. Decision was made to
proceed with a laparoscopic distal pancreatectomy with splenectomy.
Fourteen days preoperatively, pneumococcal, meningococcal and H influenza
vaccines were administered.
FIGURE 1: Axial MRI image revealing a mass (white arrow) in the
body of the pancreas.
█ BACKGROUND
Distal pancreatectomy is most commonly performed for benign and malignant
disease of the pancreatic tail, body and in some cases, neck. This operation
can be performed open or with a minimally invasive approach, and can be
done with or without removal of the spleen, but is most commonly combined
with a splenectomy. We prefer the minimally invasive approach, and this is
most often done with a concurrent splenectomy after division of the splenic
artery and splenic vein. However, removal of the spleen is not without risk.
Patients who have undergone splenectomy are at increased risk of infectious
and thromboembolic events, some of which can be devastating.
Overwhelming Post Splenectomy Infection (OPSI) results in septicemia,
meningitis, and high rates of mortality. Therefore vaccination of these
patients is extremely important. The decision of when to administer these
vaccines, however, is less certain.
SPLENIC ANATOMY AND PHYSIOLOGY
The spleen lies in the posterior lateral portion left upper quadrant, protected
by the 9th, 10th , and 11th ribs. It is in close proximity to the stomach,
diaphragm, left kidney, and the aptly named “splenic flexure” of the colon.
Most importantly, for the pancreas surgeon, the tail of the pancreas lays in the
hilum of the spleen and is intimately associated with the splenic artery and
vein.
The spleen acts, in many ways, like a filter, to remove old or senescent
red blood cells, clear active red blood cells of nuclear remnants, and the
destruction of platelets or cells that have been coated with antibody. Most
importantly, the spleen is the largest lymphoid organ in the body and the B-
cell populations in the spleen are the main site of production of IgM
antibodies, which are integral to the opsonization and removal of
encapsulated organisms. It is also the main site for synthesis of opsonizing
molecules: opsonin, tuftsin and properdin. Removal of the spleen, therefore,
is not without consequence.
INDICATIONS FOR SPLENECTOMY

There are approximately 25,000 splenectomies performed annually in the
United States, and approximately 1 million asplenic people living in the
United States. In addition to this procedure being performed for benign and
malignant disease of the pancreas, splenectomy is performed in trauma
patients, those with hematologic disorders, and for iatrogenic injury. It was
in trauma patients where OPSI became well known as a complication of
splenectomy, and led to increasing attempts and guidelines for splenic
preservation. Today, medical indications, not trauma, account for the
majority of splenectomies performed. This is followed by trauma and then
associated oncologic resections.
DISTAL PANCREATECTOMY WITH

SPLENECTOMY VERSUS SPLEEN-
PRESERVING DISTAL PANCREATECTOMY
Distal pancreatectomy (DP) is not a common operation. Most patients with
distal pancreatic malignancies have advanced disease at the time of
presentation. In fact, it is estimated that even at high volume centers, less than
30 distal pancreatectomy operations are performed annually. While distal
pancreatectomy is most commonly performed with splenectomy, in cases
where the spleen can be preserved without compromising the outcome for the
patient, splenic preservation is clearly the best means to preventing post-
splenectomy complications. Spleen-preserving distal pancreatectomy can be
done safely, employing either minimally invasive or open techniques, and
with minimal additional operative time and blood loss. In fact, some theorize
that the amount of blood spared within the spleen is more significant than the
amount lost with dissection of the splenic vessels and hilum.
We perform spleen preserving distal pancreatectomy in appropriately
selected patients with either benign disease or in those with lesions of low
malignant potential. We advocate for en bloc splenectomy in patients with
malignancy or in those with benign disease with concurrent splenic vein
thrombosis.
COMPLICATIONS OF SPLENECTOMY
Asplenic patients are more susceptible to infections related to encapsulated
bacteria. Of these, the most devastating is OPSI (Overwhelming Post
Splenectomy Infection) syndrome, which is defined as fulminating sepsis,
meningitis, or pneumonia. The risk of development of OPSI in asplenic
patients is 3.6%, which is 50 times higher than the general population, and is
associated with an overall mortality of 1.4-1.8%. However, these risks are
misleading for the pancreatic surgeon, since the majority of high risk patients
are those with genetic or hereditary disorders, infants, or young children. The
true risk of morbidity and mortality from splenectomy in healthy adults is less
than one percent.
The majority of post-splenectomy complications occur within the first
two years of resection, although there are reports of fulminating infection
occurring more than 20 years after splenectomy. No clear evidence exists
examining the time interval to OPSI for patients who have undergone
splenectomy for malignancy. However, the relative immunocompromised
state associated with such an operation, as well as the use of adjuvant
therapy, can be assumed to increase the risk in these patients.
The organisms implicated in OPSI are S. pneumonia, N. meningitides, and
H. influenza type b. Of these S. pneumoniae is the most common causative
organism, accounting from 50-90% of cases, followed by H. influenza type b
and N. meningitides. Other less frequently seen organisms include
Escherichia coli, Pseudomonas aeruginosa, Capnocytophaga canimorsus,
Enterococcus sp, Bacteroides sp, and Bartonella sp. Symptoms start
insidiously with a mild, influenza like presentation (Table 1 ). Patients report
fever, malaise, myalgias, headache, vomiting, diarrhea and abdominal pain.
This prodrome is quickly followed by septic shock leading to multi-system
organ failure and death. The mortality rate of OPSI ranges from 38-70% even
with treatment. Early institution of empiric antibiotics, followed by tailored
treatment based on culture results, is recommended.
TABLE 1: Signs and symptoms of Overwhelming Post-
Splenectomy Infection
GUIDELINES
Guidelines for prevention and treatment of infection in asplenic and
hyposplenic patients were first prepared by the British Committee for
Standards in Hematology in 1996 and subsequently reviewed in 2002. These
guidelines include three parts: education, vaccination, and antibiotic
prophylaxis. Operative asplenia patients should be aware of the signs and
symptoms of infection and notify their physician of any febrile illness. This
education is especially important for the operating surgeon, as the risk of
post-splenectomy complications continues beyond the follow up period, and
so patients must be taught to be their own advocate. Patients will require
repeat vaccinations and boosters. Travel precautions are also important due
to the risk of bacterial and parasitic infections, and patients must understand
the need for preventative measures prior to travel. There is clear evidence
demonstrating that properly educated patients have the lowest incidence of
post-splenectomy infection. We ascribe to the Advisory Committee for
Immunization Practices (ACIP) recommendations for immunization in
asplenic patients.
VACCINATIONS
Pneumococcus
Currently there are two different pneumococcal vaccines for S. pneumoniae:
the polysaccharide vaccine with 23 different serotypes (PPSV-23) and a 13
valent conjugate vaccine (PSV-13). The PPSV-23 reduces the occurrence of
OPSI in splenectomized patients but only covers for 90% of pneumococcal
infections. PSV-13 protects against fewer organisms, but there is evidence
that it has enhanced immunogenicity in adult patients. In our practice,
otherwise healthy patients receive the PSV13 vaccine 14 days prior to
surgery, followed by PPSV23 at least 8 weeks after PSV13 administration.
The optimal time for pre-operative administration of PPSV-23 is 14 days,
and this is the timing we perform for all our patients undergoing elective
distal pancreatectomy with splenectomy. In patients undergoing urgent or
emergent distal pancreatectomy with splenectomy, we delay administration
of vaccines until at least 14 days after resection, or until once we are
comfortable that they will mount an appropriate immunogenic response to the
vaccines. Vaccinations administered earlier that 14 days post-operatively
have been shows to be less effective. The PPSV-23 vaccination should be
repeated every five years, depending on the age and medical condition of the
patient as the protective effect of the pneumococcal vaccination declines five
years after vaccination.
H. influenza Type b
There are three monovalent Hib conjugate vaccines and three combination
Hib vaccines available. All are equally effective. For our patients, we
follow the ACIP guidelines and give the Hib vaccine as a single
administration 14 days prior to surgery to any patient who has not been
previously vaccinated. While it is true that patients who have been
previously vaccinated and completed the booster schedule are considered
immune and do not need re-vaccination, it is often difficult to document this
pre-operatively, and so if this cannot be adequately documented we give the
vaccine.
Meningococcal Immunization
There are three types of quadrivalent meningococcal vaccines; one
quadrivalent polysaccharide vaccine (MPSV4), three quadrivalent conjugate
vaccines, and serogroup B meningococcal vaccines. The MPSV4 protects
against strains A, C, Y, and W135 capsular polysaccharides. This vaccine is
preferred for adults older than 55 years who require a single dose and who
have not received a previous dose of meningococcal conjugate vaccine. The
majority of patients in the US, however, have already received the MPSV4
vaccine. For these patients, use of a quadrivalent conjugate vaccine
(MenACWY) is recommended, and this is our practice. MenACWY-CRM
and MenACWY-D are the two quadrivalent conjugate vaccines available in
the US. The MenACWY-CRM can be administered concomitantly with
PCV13 and, therefore, this is the vaccine we prefer. MenACWY-D must be
administered at least 4 weeks after PCV13 due to interactions leading to
decreased immune response to PCV13. In addition, there are now serogroup
B meningococcal vaccines (MenB-FHbp or MenB-4C) and patients
undergoing splenectomy should receive this vaccine as well.
The meningococcal immunizations are administered at least 14 days prior
to surgery. Patients should be re-vaccinated every five years.
The adverse effects of any of these vaccinations are minimal and should not
be used as a reason not to vaccinate patients. Local reactions include redness
or swelling and systemic reactions include arthralgia, fatigue, fever,
headache, nausea, myalgia and rash. These local and systemic reactions are
self-limiting and do not have severe sequelae.
COMPLIANCE
However, compliance with vaccination has been a factor in the development
of OPSI. The compliance rate for pneumococcal vaccination has been
reported as low as 42%, and vaccination against meningococcal and H.
influenza type b was even lower. In elective patients undergoing distal
pancreatectomy with splenectomy, pre-operative administration of vaccines
is especially effective. Management of patients undergoing distal
pancreatectomy encompasses the pre-operative decision making for
splenectomy. Indications for splenectomy should be reviewed and individual
patients screened for risks associated with splenectomy. A proposed
‘guideline’ for vaccination in patients with distal pancreatectomy can be seen
in Figure 2 .
FIGURE 2: Guideline for vaccination in patients undergoing distal

pancreatectomy with splenectomy.
SALIENT POINTS
Overwhelming post splenectomy infection syndrome results in
significant morbidity and high mortality in a small but real percentage
of patients undergoing distal pancreatectomy with splenectomy.
In certain patients, spleen preserving distal pancreatectomy can be safe
and effective and avoid risks associated with splenectomy.
We advocate for distal pancreatectomy with splenectomy in cases of
malignancy or in those with benign disease with concurrent splenic
vein thrombosis.
Vaccinations for S. pneumonia, N. meningitides, and H. influenza type
b should be administered for patients undergoing distal pancreatectomy
with splenectomy during pre-operative planning at least 14 days prior
to the operation.
SELECTED REFERENCES
1. Parikh PY, Lillemoe KD. Surgical management of pancreatic cancer--
distal pancreatectomy. Semin Oncol 2015; 42(1):110-22.
2. Bisharat N, Omari H, Lavi I, Raz R. Risk of infection and death among
post-splenectomy patients. J Infect 2001; 43(3):182-6.
3. Theilacker C, Ludewig K, Serr A, Schimpf J, Held J, Bögelein M, et al.
Overwhelming Postsplenectomy Infection: A Prospective Multicenter
Cohort Study. Clin Infect Dis 2016; 62(7):871-8.
4. Vaccines & Immunizations. Centers for Disease Control and Prevention
website. https://www.cdc.gov/vaccines/index.html. Updated April 21,
2017. Accessed June 14, 2017.
5. Shatz DV, Schinsky MF, Pais LB, Romero-Steiner S, Kirton OC,
Carlone GM. Immune responses of splenectomized trauma patients to
the 23-valent pneumococcal polysaccharide vaccine at 1 versus 7
versus 14 days after splenectomy. J Trauma 1998; 44(5):760-5.
CASE SCENARIO
A 68-year-old female with a history of mild congestive heart failure and
obesity, undergoes pancreatoduodenectomy for pancreatic adenocarcinoma.
Her postoperative course is complicated by a pancreatic fistula requiring
prolonged maintenance of a surgical drain. She receives routine prophylactic
subcutaneous heparin perioperatively and during her inpatient recovery
without bleeding event. Should anticoagulation be continued for a prolonged
time-frame following discharge?
█ BACKGROUND
INTRODUCTION
Increasing scrutiny of perioperative morbidity and mortality in recent years
has been paralleled by the recognition that the traditional 30-day
perioperative window is too narrow to capture all surgery-related adverse
events. Indeed, the perioperative period has been redefined as extending 90
days from surgery. For hepatopancreatobiliary (HPB) operations, this
redefinition provides a more meaningful estimate of perioperative events
such as pancreatic fistula, bile leak, and liver insufficiency that take time to
unfold and may lead to morbidity and mortality late in the postoperative
course. Delayed presentation of complications, however, is not limited to
HPB-specific adverse events. Venous thromboembolism (VTE), while
traditionally attributed to intraoperative venous stasis, is increasingly
diagnosed beyond discharge. Scrutiny of larger institutional and nationwide
datasets suggest that a substantial minority of such events are diagnosed after
discharge from the hospital. These data and identification of at-risk
populations have driven consensus recommendations for prolonged VTE
prophylaxis from oncologic and surgical societies. Given the devastating
impact of pulmonary embolism and the risk of VTE in pancreas cancer
patients, adoption of prolonged prophylaxis protocols is gaining acceptance,
but has been slow. This may reflect concern regarding prophylaxis-related
complications, particularly bleeding.
WHAT IS THE RISK OF VTE?

The risk of VTE, including superficial and deep vein thrombosis (DVT) and
pulmonary embolism (PE), is significantly increased in patients with cancer.
This has been recognized at least since the 1860’s when Armand Trousseau
described several cases in which recurrent thrombosis heralded the
diagnosis of visceral malignancy. Modern population-based studies suggest
that cancer substantially increases the risk of VTE. Rates vary across cancer
types with one of the highest rates seen in pancreatic cancer. Presence of
metastatic disease is a particularly powerful predictor of thromboembolism.
Indeed, California Cancer Registry data indicates that symptomatic VTE is
observed in 5% of early stage pancreatic cancer but nearly triples in those
with advanced disease. Patients are at greatest risk during the first 3-6
months after cancer diagnosis, and patients who are being treated with
systemic chemotherapy are at a 6.5-fold increased risk of VTE.
Surgery alone is well-established as an independent risk factor for
thromboembolic disease. In high-risk patients undergoing abdominal surgery,
the incidence of VTE may be as high as 30% in the absence of
thromboprophylaxis. Combining major surgery with a diagnosis of cancer
confers greater risk compared to either surgery or cancer alone. Cancer
patients have a 3-fold increased risk fatal PE following surgery compared to
patients undergoing surgery for non-inflammatory benign conditions. Risk of
VTE in patients undergoing pancreatectomy for pancreatic cancer is
particularly high. Estimates of symptomatic VTE in pancreatectomy patients
may be as high as 5%, making pancreatic cancer the primary cancer with the
second highest risk after lung cancer.
Risk stratification may allow for more rational selection of a VTE
prophylaxis strategy. Large retrospective studies indicate that the greatest
risk factors include a previous history of VTE, anesthesia time greater than
or equal to 2 hours, ASA class 3-4, bedrest for 4 or more days, advanced
cancer stage, sepsis, obesity, and advanced age. Identified risk factors have
been used to create two validated risk factor scoring systems for VTE, the
Rogers and the Caprini Risk Assessment Models. The Rogers scoring system
includes many perioperative risk factors including type of operation, wound
class, ASA classification, and laboratory data such as albumin and bilirubin.
Its major limitation, other than being complicated to compute, is that it does
not allow differentiation between moderate and high risk for VTE.
In contrast, the Caprini scoring method allows high-risk patients to be
readily identified. The Caprini Risk Assessment Model relies primarily on
preoperative and historical patient data including age, BMI, malignancy,
causes of hypercoagulability, and risk factors for venous stasis (Table 1 ). A
total score of five or greater places a patient at high-risk for VTE. Any open
or laparoscopic procedure adds 2 points to the patient score. While the
model initially appears complex, it may be simplified to quickly assign
certain patients to the high-risk category. For example, patients undergoing
surgery for cancer receive 2 points for the operative procedure and 2 points
for malignancy, placing them in the moderate risk category (3-4 points). Any
additional risk factor such as age > 40 years or BMI > 25 kg/m3 will make
the patient “high-risk”.
TABLE 1: Full and simplified Caprini risk assessment

model6
IS THE RISK MODIFIABLE?

Chemoprophylaxis with either unfractionated or low molecular weight
heparin (LMWH) prior to induction of anesthesia and during the patients’
post-operative hospital stay is widely accepted as the standard of care for
moderate and high-risk surgical patients. Observational data of patients
undergoing surgery for cancer, however, suggests that timing of VTE events
following surgery has a bimodal distribution with 40% of events occurring
more than 21 days after the index opera tion. For this reason, extended
chemoprophylaxis in patients undergoing surgery for cancer has been
explored.
Multiple randomized controlled trials have been performed comparing
rates of VTE, specifically asymptomatic DVT as diagnosed using venography
or duplex ultrasound, in patients treated with 28-31 days of extended
chemoprophylaxis compared to a control group who only received
prophylaxis for one week following surgery. The largest of these were the
ENOXACAN II trial and CANBESURE study, which each enrolled upwards
of 500 patients. Both of these studies utilized bilateral venography at one
month and followed patients for 90 days. These studies show significant
reductions in asymptomatic DVT in the patients receiving extended
prophylaxis compared to the control arm. Other endpoints, such as
symptomatic DVT and PE, were too rare to assess given the limited sample
sizes; however, meta-analysis favors extended prophylaxis for reduction of
all VTE’s. These RCTs also highlight the safety of extended prophylaxis.
Major bleeding risk was not increased in patients receiving prophylaxis both
in the individual studies and on meta-analysis.
While the clinical significance of asymptomatic DVTs is debatable, the
highly thrombogenic state associated with malignancy raises concern for
proximal propagation of clot and subsequent events. In other high-risk
populations, such as patients undergoing major orthopedic surgery, studies
have correlated asymptomatic DVT found on venography and subsequent
later symptomatic DVT. Additionally, reduction of asymptomatic DVT
reduces the risk of symptomatic DVT and post-thrombotic syndrome. While
the association between asymptomatic DVT and PE is unclear, large autopsy
series suggest that fatal PE is seldom preceded by clinically significant DVT.
Based on these data, professional societies including the American
College of Chest Physicians (ACCP), the American Society for Clinical
Oncology (ASCO), the National Comprehensive Care Network (NCCN), and
the European Society of Medical Oncologists (ESMO) have all produced
guidelines supporting at least 28 days of VTE prophylaxis with LMWH
following abdominal or pelvic surgery for cancer (Table 2 ). While
consistent, these guidelines are relatively ambiguous when describing
inclusion and exclusion criteria. All guidelines specify that the
recommendation applies to “high-risk” cancer patients although only the
ACCP seeks to define this term, specifying “high-risk” as a Caprini score ≥
5. The ACCP also lists risk factors for major bleeding complications
including active bleeding, acute stroke, and specifically sepsis, pancreatic
leak, or sentinel bleed following pancreatoduodenectomy.
TABLE 2: Summary of professional society guidelines
SHOULD PATIENTS RECEIVE EXTENDED

PROPHYLACTIC ANTICOAGULATION?
Despite the consistent presentation of data highlighting increased risk of VTE
and PE amongst cancer patients undergoing abdominal surgery, the surgical
oncology community has been slow to adopt protocols employing extended
anticoagulation despite evidence that such protocols are safe in the post-
surgical patient. Prior reservations about shorter duration postoperative
chemical prophylaxis and preoperative induction chemical prophylaxis were
assuaged by studies and experience suggesting safety and efficacy. While
studies on extended prophylaxis specific to pancreatectomy patients remain
elusive, one would expect safety given the wealth of less specific data
available. Late bleeding events are unlikely to become more frequent in the
context of a modest anticoagulation regimen that is already administered
during the in-hospital recovery. Pancreatic fistula has been proposed as a
relative contraindication to prolonged prophylaxis stemming from its
association with bleeding events. Importantly, withholding prophylactic
chemical prophylaxis is unlikely to prevent the rare, devastating ruptured
pseudoaneurysm or massive GI bleed. In the absence of a specific bleeding
concern, our practice has been to proceed with prolonged prophylaxis.
Patient factors may also compel selective omission of prolonged
prophylaxis. While nearly all pancreatectomy patients qualify for prolonged
prophylaxis through some combination of operative duration, age or
underlying malignancy, there may be a subset of patients from whom
prolonged prophylaxis poses more of a risk than benefit (e.g., the very old
unsteady patient at risk for fall). Arguably such patients are similarly unlikely
to benefit from pancreatectomy. As in all aspects of medicine, a non-
algorithmic, nuanced approach is laudable.
IS PROLONGED PROPHYLAXIS
ENOUGH?
In most cases, the benefit of perioperative and prolonged VTE prophylaxis
justify the acceptably low risk of associated complication. Important to note,
however, is that these measures do not provide complete protection against
thromboembolic events. Areas for further study include additional
preoperative scrutiny of especially high-risk patients with duplex
sonography. In our own practice, class III obesity (BMI 40+), immobility,
and prolonged preoperative hospitalization are indications for preoperative
screening. The presence of preoperative deep vein thrombosis may guide
postoperative anticoagulation or selective preoperative placement of an
inferior vena cava (IVC) filter. In the postoperative setting, there may be a
role for the use of screening duplex ultrasonography as well. The
development of a DVT while on prophylactic anticoagulation may be an
indication for more intensive follow-up or for therapeutic dose
anticoagulation. Balancing risk and benefit remains the overarching
challenge.
SALIENT POINTS
Pancreatic cancer is one of the most thrombogenic malignancies.
Most patients who undergo pancreatectomy for pancreatic cancer are at
high-risk for VTE according to validated risk assessment models.
Numerous professional society guidelines recommend at least 28 days
of post-operative VTE prophylaxis for high-risk surgical oncology
patients.
Evidence consistently shows that extended VTE prophylaxis following
surgery for cancer reduces the incidence of asymptomatic DVT. Meta-
analysis supports extended prophylaxis for reduction of all VTE
events.
Extended VTE prophylaxis is safe following pancreatoduodenectomy.
SELECTED REFERENCES
1. Chew HK, Wun T, Harvey D, Zhou H, White RH. Incidence of venous
thromboembolism and its effect on survival among patients with
common cancers. Arch Intern Med. 2006;166(4):458-64.
2. Heit JA, O’Fallon WM, Petterson TM, Lohse CM, Silverstein MD,
Mohr DN, et al. Relative impact of risk factors for deep vein
thrombosis and pulmonary embolism: a population-based study. Arch
Intern Med. 2002;162(11):1245-8.
3. Fagarasanu A, Alotaibi GS, Hrimiuc R, Lee AYY, Wu C. Role of
Extended Thromboprophylaxis After Abdominal and Pelvic Surgery in
Cancer Patients: A Systematic Review and Meta-Analysis. Ann Surg
Oncol. 2016; 23(5)1422-30.
4. Agnelli G, Bolis G, Capussotti L, Scarpa RM, Tonellu F, Bonizzoni E,
et al. A clinical outcome-based prospective study on venous
thromboembolism after cancer surgery: the @RISTOS project. Ann
Surg. 2006;243(1):89-95.
5. Gould MK, Garcia DA, Wren SM, Karanicolas PJ, Arcelus JI, Heit JA,
et al. Prevention of VTE in nonorthopedic surgical patients.
Antithrombotic Therapy and Prevention of Thrombosis, 9th ed:
American College of Chest Physicians evidence-based clinical practice
guidelines. Chest. 2012;141(2 Suppl):e2275-2775.
6. Caprini J. Thrombosis risk assessement as a guide to quality patient
care. Dis Mon 2005;51:70-8.
7. Lyman GH, Bohlke K, Khorana AA, et al. Venous thromboembolism
prophylaxis and treatment in patients with cancer: American Society of
Clinical Oncology clinical practice guideline update 2014. J Clin
Oncol. 2015;33(6):654-656. doi:10.1200/JCO.2014.59.7351.
8. Streiff B, Holmstrom B, Ashrani A, et al. Cancer-Associated Venous
Thromboembolic Disease, Version 1.2015. J Natl Compr Canc Netw.
2015;13(9):1079.
9. Mandalà M, Falanga A, Roila F. Management of venous
thromboembolism (VTE) in cancer patients: ESMO Clinical Practice
Guidelines. Ann Oncol. 2011;22 Suppl 6(Supplement 6):vi85-92.
doi:10.1093/annonc/mdr392.

CASE SCENARIO
J.D. is a 78-year-old female who was diagnosed with stage 1 pancreatic
adenocarcinoma at the head of her pancreas. She was brought to the
operating room for a planned resection, but was found to have tumor
encroachment of her stomach, which necessitated a
pancreaticoduodenectomy with concomitant gastrectomy. Postoperatively,
she required minimal amounts of vasopressor support and was brought to the
intensive care unit for recovery. Subsequently, she recovered well, tolerating
a regular diet and ambulating on postoperative day 7, and was ready to be
discharged. However, given her surgery and postoperative course, she has a
high risk for being readmitted. What should you do?
█ BACKGROUND
In our current healthcare climate, where cost and accountability is
increasingly scrutinized, hospital readmission rates have emerged as an
increasingly popular quality metric. With initiatives such as the National
Surgical Quality Improvement Program (NSQIP), and more public non-profit
organizations such as ProPublika’s surgeon’s report card, these hospital- and
surgeon-specific metrics are increasingly accessible to the public. This in
turn dangerously creates a dilemma for the surgeon when having to make
clinical decisions in the best interest of the patient versus acting due to
pressure from insurance companies and reimbursement policies. In patients
undergoing pancreaticoduodenectomy, 1 in 6 will experience a readmission
within 30-days, posing a significant management issue. In this chapter, we
discuss the causes and predictors of readmission after
pancreaticoduodenectomy, as well as our proposed management strategy for
the patient at high-risk of a readmission. What Are The Causes Of
Readmission After Pancreaticoduodenectomy?
In 2008, the Medicare Payment Advisory Commission reported that the cost
of preventable readmissions was $12 billion a year. However, such
preventable readmissions were in relation to patients with medical
conditions such as congestive heart failure and pneumonia, and it is important
to point out that generalization to the surgical patient is not accurate. In
patients undergoing pancreaticoduodenectomy, the majority of readmissions
occur as a result of procedure-specific complications rather than a lack of
care coordination. In an analysis of 1,173 consecutive patients undergoing
pancreaticoduodenectomy at our institution, the three most common causes of
readmission were intraabdominal abscess, pancreatic fistula, and wound
infection, accounting for 46% of all readmissions. Only 3.5% of patients
were readmitted without an identifiable medical cause, consistent with the
low rates (3-14%) of unrevealing readmission workups reported by other
institutions. As such, surgeons should tread carefully in their efforts to reduce
their readmission rates given that majority of patients requiring readmission
are a vulnerable cohort experiencing clinically significant complications that
require rescue interventions.
WHAT ARE THE PREDICTORS OF
READMISSION AFTER
PANCREATICODUODENECTOMY?
In an attempt to reduce readmissions after pancreaticoduodenectomy, similar
efforts have gone into predicting which patients are at risk of being
readmitted after discharge. In our institutional analysis, predictors of
readmission include patients requiring multivisceral resection (OR 4.0,
p=0.02), intensive care unit admission (OR 2.9, p<0.001), and patients with
pancreatic fistula (OR 1.9, p=0.004). Interestingly, surgeon volume was not a
significant predictor of readmission. In a population-based analysis,
Yermilov et al. demonstrated that increasing age, patient comorbidities, and
tumor T stage were independent predictors of readmission post-PD.
Similarly, in a SEER-Medicare analysis, Hyder et al. demonstrated that the
largest share of variation in readmission is attributable to patient-related
factors such as age and comorbidities (95.4%), as opposed to physician-
(0.3%) and hospital- (4.3%) related factors. This provides further evidence
that readmission after pancreaticoduodenectomy is largely a result of the
complexity of the disease process and surgical procedure necessary, rather
than lack of care coordination. However, it is important to point out that this
does not preclude care coordination efforts and diligent follow-up. At our
institution, a designated nurse practitioner calls every post-PD patient the
day following hospital discharge, and most have visiting nurse services.
Many other high-volume institutions have similar follow-up care mechanisms
in place, likely explaining the studies’ findings demonstrating that
readmissions post-PD are largely due to complexity of the disease and
operation.
TIMELINE OF READMISSION AFTER
PANCREATICODUODENECTOMY
So when do readmissions after pancreaticoduodenectomy actually occur? In
our institutional analysis, 50% of readmissions occur within 7 days after
discharge, and the rest in the subsequent 6 weeks (Figure 1 ). In analyzing
trends across 3 decades, we found that as we decreased the length of hospital
stay from 22 days in 1982 to 7 days in 2012, readmission rates increased
from 3% to 15% (Figure 2 ). Taken together, most readmissions may be a
safety mechanism to address these complications as we continue to reduce
length of hospital stay via enhanced recovery pathways. One argument is to
retain patients for an additional 3-4 days to manage late complications,
which would effectively decrease readmission rates by almost 50%.
However, it should be noted that more than 80% of patients do not require a
readmission, and it would be counterintuitive to be retaining those patients to
prevent readmissions in the 15% to 20% that would need them. Rather, we
should use readmissions as a fallback mechanism to allow us to safely
discharge patients as early as critical pathways dictate.
FIGURE 1: The distribution of readmissions across time after
discharge post pancreaticoduodenectomy (Adapted with permission
from Springer in Fong ZV, Ferrone CR, Thayer SP, et al.
Understanding hospital readmissions after
pancreaticoduodenectomy: can we prevent them? a 10-year
contemporary experience with 1,173 patients at the Massachusetts
General Hospital. J Gastrointest Surg 2014; 18:137-44).
FIGURE 2: Trend lines demonstrating the inverse relationship

between length of hospital stay and readmission rates after
pancreaticoduodenectomy from 1982 to 2012 (Adapted with
permission from Springer in Fong ZV, Ferrone CR, Thayer SP, et al.
Understanding hospital readmissions after
pancreaticoduodenectomy: can we prevent them?: a 10-year
General Hospital. J Gastrointest Surg 2014; 18:137-44).
DESTINATION OF READMISSION – NO
PLACE BETTER THAN HOME?
Given the complexity of PD, many institutions performing the procedure are
tertiary referral centers, servicing a significant proportion of patients who
travel beyond their local counties for the operation. This invariably
translates to a higher likelihood that patients requiring readmissions be
readmitted to a hospital that is different from the one that performed their
procedure (non-index hospital). In an analysis of the California state cancer
registry, Yermilov et al. reported that 49% of all PD readmissions were to
non-index hospitals. This has important implications for two reasons: (1)
there is loss of information when patients are not readmitted to their index
hospitals; (2) most non-index hospitals are not familiar with, or equipped to
care for, the post-PD patient with a complication. This is evident by the
findings of a Medicare study performed by Brooke et al. which demonstrated
that patients readmitted to non-index hospitals after a range of common
operations were associated with a 26% higher risk of 90-day mortality when
compared to patients readmitted to index hospitals. It is also not surprising
that they demonstrated that this effect was most significant for patients
undergoing pancreatectomy; patients readmitted to non-index hospitals were
associated with a 44% increase in risk of mortality.
In our institutional analysis, the mortality rate of the readmitted PD patient
was 5.1%, almost 5-fold higher than the non-readmitted patient (1.1%,
p<0.001). This readmitted PD cohort represents a vulnerable cohort that
requires prompt recognition of complications and timely treatment. It is
imperative that we maintain an open communication with our patients, and
facilitate readmission to patients’ home institutions whenever appropriate.
FUTURE DIRECTIONS
The challenge that remains is identifying high-risk patients that will benefit
from early detection of complications and treatment and low-risk patients
who are safe for early discharge, while minimizing readmission rates. A
promising biomarker to risk-stratify patients that has recently been
investigated is C-reactive protein (CRP). CRP is a marker of inflammation,
including infectious complications, and has been shown to be more sensitive
than the traditional white blood cell count used. In a meta-analysis of 7
cohort studies, Adamina et al. assessed the predictive value of CRP levels
for infectious complications in abdominal surgery across 7 major abdominal
procedures. The authors demonstrated that CRP levels 4 days after surgery
had the highest predictive value, with a sensitivity, specificity, positive
(PPV), and negative predictive value (NPV) of 68.5%, 71.6%, 50.4%, and
84.3% respectively. Taking the immediate postoperative inflammatory state
into account, the cutoff for pancreatic resection on postoperative day 4
specifically is 184 mg/l and had a PPV and NPV of 46% and 79%
respectively. While it is poor in predicting complications, its NPV of 79%
makes it an ideal biomarker to predict low risk patients that has been safely
tracked to enhanced recovery pathways for early discharge. It should be
noted that all the data on CRP’s predictive performance in major abdominal
surgery were from studies done in Europe, and this needs to be validated in a
North American cohort and treatment setting before incorporated in one’s
practice.
CASE SCENARIO – MANAGEMENT

In the case scenario presented, patient J.D. underwent a concomitant extra-
pancreatic organ resection with her PD and was admitted to the intensive
care unit postoperatively for recovery. On postoperative day 7, she is
tolerating a regular diet and ambulating well and ready for discharge.
However, based on prediction models, her age, operation, and postoperative
course puts her at a high risk for readmission. Based on our experience and
interpretation of the literature, we would still elect to discharge the patient.
The patient should be educated on her risk of a subsequent readmission and
awareness heightened on potential signs and symptoms that should prompt a
notification to the physician’s office. Finally, it should be emphasized to the
patient that if she would need to present to an emergency room for
assessment, she should, to her best attempt, return to our hospital or the
closest emergency room with plans set in place for her transfer back here if
appropriate after initial triage. If the patient came from far away, we would
suggest that she stay local for at least 3 days before making the trip back
home.
SALIENT POINTS
After PD, 1 in 6 patients will experience a readmission within 30 days
after discharge from their index hospitalization.
Predictors of readmissions after PD are largely patient-related factors,
such as increasing age, patient comorbidities, multivisceral resection,
postoperative pancreatic fistulas and intensive care unit admission.
The most common causes for readmission after PD were
intraabdominal and wound infections, while only 3-14% of
readmissions were unrevealing with no identifiable cause.
Half of all readmissions after PD occurred within 7 days after
discharge, and is a result of efforts aimed at decreasing length of
hospital stay over the years.
High-risk patients who are clinically ready for discharge should be
discharged, and readmission reserved for the 15% to 20% of patients
that will need it.
Patients requiring a readmission after PD should be readmitted to their
index hospital whenever appropriate, as it is associated with a 44%
decrease in mortality rates as compared to patients readmitted to non-
index hospitals.
SELECTED REFERENCES
1. Fong ZV, Ferrone CR, Thayer SP, Wargo JA, Sahora K, Seefeld KJ, et
al. Understanding hospital readmissions after
General Hospital. J Gastrointest Surg. 2014; 18(1):137-44; discussion
144-5.
2. Yermilov I, Bentrem D, Sekeris E, Jain S, Maggard MA, Ko CY, et al.
Readmissions following pancreaticoduodenectomy for pancreas cancer:
a population-based appraisal. Ann Surg Oncol. 2009; 16(3):554-61.
3. Hyder O, Dodson RM, Nathan H, Schneider EB, Weiss MJ, Cameron
JL, et al. Influence of patient, physician, and hospital factors on 30-day
readmission following pancreatoduodenectomy in the United States.
JAMA Surg. 2013; 148(12):1095-102.
4. Brooke BS, Goodney PP, Kraiss LW, Gottlieb DJ, Samore MH,
Finlayson SR. Readmission destination and risk of mortality after major
surgery: an observational cohort study. Lancet. 2015; 386(9996) 884-
95.
5. Adamina M, Steffen T, Tarantino I, Beutner U, Schmied BM,
Warschkow R. Meta-analysis of the predictive value of C-reactive
protein for infectious complications in abdominal surgery. Br J
Surg.2015; 102(6):590-8.
CASE SCENARIO
A 74-year-old female undergoes pylorus-preserving Whipple for
cholangiocarcinoma. Her surgery is uneventful as is her postoperative
recovery. Her drain is removed on POD#4. She is cleared by physical
therapy and is discharged to home with her husband on POD#7. Three days
later, on POD#10, she presents to the emergency department with vague
complaints of not doing well at home and overall failure to thrive. Evaluation
fails to identify any postoperative complication to account for her situation.
What do you do?
█ BACKGROUND
Readmission following pancreatic resections and pancreaticoduodenectomy
(PD) in particular remains a significant problem with multifactorial
causality. It is an especially important topic given the increasingly prevalent
use of 30-day readmission rates as a surrogate for quality care, and the
creation of mandates by the Center of Medicare and Medicaid services that
potentially affect hospital payments if patients are readmitted for “a
condition that could have reasonably been prevented through the application
of evidence-based guidelines.” Recent publications have reported 30-day
readmission rates ranging from 15 to 23%. At 90 days, the readmission rate
increases to between 19 and 29%. Most major pancreatic surgery centers
have large and often geographically disparate referral bases, and it is
therefore likely that published readmission rates underestimate actual
readmission rates by failing to capture patients that are readmitted at outside
hospitals. The majority of readmissions occur within the first 30
postoperative days, with two studies reporting median times to readmission
of 8 and 9 days from initial discharge, and another reporting that 50% of
readmissions occurred within 7 days. Two studies reported that multiple
readmissions were ultimately required for 21% of those initially readmitted.
The total estimated additional cost for readmitted patients averaged $16,000
in one study, owing to index admission costs that were almost $6,000 more in
readmitted patients and readmission costs that exceeded $10,000.
The most common reasons for readmission vary depending on the time
frame from initial discharge. In a review of 1,173 patients who underwent
PD over a 10-year study period, Fong et al. reported that the most common
reasons for early (within 7 days) readmission were ileus, delayed gastric
emptying (DGE), and pneumonia. They reported that after 7 days, wound
infection, failure to thrive, and intra-abdominal hemorrhage were the
predominant reasons (6). Sadot et al. reviewed 490 patients who underwent
PD as well as distal and central pancreatectomies, and found that the main
reasons for readmission within 30 days were procedure-related infections
(58%), failure to thrive (21%), and non-surgical infections (7%). Between
31 and 90 days, the most common reasons were failure to thrive and
chemotherapy-related symptoms (38%), leak, fistula, abscess or wound
infection (28%), non-surgical infections (21%), and drain malfunctions
(14%). Ahmad et al. compiled medical records from six high-volume
institutions for 1302 patients who underwent PD over a five-year period and
found that the most common reasons for readmission within 30 days were
infectious complications (19.9%) and DGE (14.3%). After 90 days, the most
common reasons were wound infection (14.2%), intraabdominal abscess
(12.5%), and failure to thrive (13.5%). Kent et al. reviewed 578 pancreatic
resections over an eight-year period. Their most common reasons for
readmission were procedure-specific complications such as DGE,
pancreatic fistula, bile leak, and portal vein thrombosis (48%), followed by
general postoperative complications including wound infection and J-tube
damage (18%) and also failure to thrive (18%). In our experience,
postoperative complications, including pancreatic fistula and DGE, account
for many readmissions and can be obvious, or more subtle, presenting with
vague complaints like our patient above. Therefore, patients presenting for
readmission must be thor oughly evaluated for such postoperative
complications prior to attributing their problem to failure to thrive. However,
failure to thrive does comprise an important number of readmissions.
One way to approach the problem of readmission after pancreatic
resection is by attempting to identify which patients are at high risk for
readmission. Fong et al. found that pancreatic fistula (OR 1.86, p=0.004),
multivisceral resection at the time of PD (OR 4.02, p=0.02), length of
hospital stay (LOS) >7 days (OR 1.57, p=0.01), and admission to the
intensive care unit (ICU) (OR 2.9, p=0.0005) were independently associated
with readmissions. When Ahmad et al. performed a multivariate analysis,
they identified preoperative diagnosis of chronic pancreatitis (OR 2.356,
p=0.007), high transfusion requirement (OR 1.875, p=0.007) and
postoperative complications including wound infection (OR 2.2, p=0.005),
intra-abdominal abscess (OR 1.96, p=0.02), and pancreatic fistula (OR 2.4,
p=0.02) to be independently associated with readmission. Kent et al. found
that small pancreatic duct size (<3mm) was the only preoperative or
intraoperative factor associated with readmission (p=0.03). On multivariate
analysis, the following factors independently contributed to readmission: any
postoperative complication (OR 2.24, p=0.006), major postoperative
complication (OR 2.2, p=0033), clinically relevant pancreatic fistula (OR
5.03, p=0.0001), and latent fistula (OR 4.29, p=0.0001).
Of note, most previous publications relating to readmission after
pancreatectomy focus on identifying risk factors for readmission that have to
do with patient or disease factors but rarely take into account other
potentially important and modifiable factors such as disposition environment
and the true impact/availability of the patient’s family or friends in helping
care for the patient. For example, many institutions use fairly routine physical
therapy evaluation such that patients are “tested” at walking and stair
climbing. However, evaluation of completion of most activities of daily
living, like bathing or cooking, is extremely rare. Additionally, many studies,
as indicated above, report risk factors for readmission but do not report on
efforts to incorporate this knowledge into formal disposition planning.
Glass et al. reviewed 658 patients who underwent pancreatectomy over a
10-year period at a single institution and performed a root cause analysis to
identify reasons for readmission: whether or not the readmission was
potentially avoidable, the adequacy of initial discharge, and where failures
occurred. Their readmission rate was 18% at a median of 9 days following
discharge and the median length-of-stay once readmitted was 7 days. Most
common reasons for readmission were the development of postoperative
complications (64%), failure to thrive (14%), and need for diagnostic
evaluation (13%). They examined the disposition of readmitted patients and
found that the majority were discharged to home with services (48%),
followed by discharge to home without services (36%) and then discharge to
rehabilitation (12%). Following readmission, 21% of patients required an
escalation of care on their subsequent discharge. On reviewing the records of
patients requiring readmission, they noted a paucity of information regarding
the ability of the patients and families to manage care at home. In reviewing
the patients for whom these data were available, they concluded that 26% of
patient readmissions were potentially avoidable and that 11% of readmitted
patients had problems that may have been able to be managed on an
outpatient basis with appropriate resources.
Recurring themes were inadequate nutrition or oral intake prior to
discharge, and the inability of patients and families to access adequate help
and resources in managing new adjuncts to their care at home, such as drains
and medications, and in resuming baseline activities of daily living at the
preoperative level. Interventions, such as improved communication regarding
postoperative expectations, better assessment of support systems at home,
increased emphasis on nutritional support, and developing ways to facilitate
patient evaluation aside from inpatient admission may reduce the rate of
avoidable readmissions. Other authors suggest that shorter interval follow-up
after discharge, ideally within 7 days and therefore prior to the time that
readmission rates peak, may better allow for detection and management of
problems that would otherwise lead to readmission.
On the other hand, it has also been suggested that readmission following
pancreatic resection is unavoidable owing to the magnitude of the operation
and not necessarily always predictable. Gawlas et al. reviewed 787
pancreatic resections performed at a single institution over a five-year
period. They reported a 30-day readmission rate of 11.6% and on
multivariate analysis they identified young age (age ≤65 years, OR 1.86,
p=0.007) and development of a postoperative complication, either major or
minor, (Grade I or II, OR 6.5, p<0.001; Grade III or IV, OR 11.99, p<0.001)
to be independent predictors of readmission. The most common reasons for
readmission were leak, fistula, abscess and/or wound infection (45.1%) and
DGE (12.1%). Median time to readmission was 8 days and median length-
of-stay once readmitted was 6 days. When they reviewed their 91 readmitted
patients, they found that the majority of readmissions were related to
development of complications directly attributable to the initial operation.
They did not find any that correlate to the length of the initial hospital stay or
to the discharge disposition, and they did not find evidence that readmissions
were due to poor coordination of care or discharge planning. Only 12
patients were admitted for concerns or symptoms that resolved without an
identifiable cause, and the median time from discharge to readmission for
this patient group was 8 days, a time frame that does not suggest premature or
inappropriate discharge was a likely cause. The authors concluded that
because, to a certain extent, the development of these postoperative
complications is expected, and because there are no evidence-based
practices to prevent or decrease their rate of occurrence that readmissions
required for management should not be used as an indicator of quality care. It
has also been suggested that pushing to decrease short-term readmission rates
may lead to increased length of initial hospital stays and that denying hospital
coverage for reimbursement may limit access to care for patients with
complications and ultimately lead to increased mortality.
Failure to thrive is a significant and particularly challenging problem
following PD and other pancreatic resections. It is reported as the main
reason for readmission in both the short and long-term postoperative period
(12-21% at 30 days, 13.5-38% at 90 days). In these patients, an acute
process justifying in-hospital level of care is not found, yet some studies
suggest that readmissions for this reason trend toward longer lengths-of-stay
than patients with an acute postoperative complication (median 5 versus 3
days respectively, p=0.051). Contributing factors may include patient age,
cancer aggressiveness, and the initiation of adjuvant chemotherapy and
nutrition. Several studies have identified increased age as a predictor of
readmission, raising the possibility that older patients may be prone to
develop failure to thrive and require readmission. The correlation of
readmission with cancer aggressiveness is suggested by the findings of Sadat
et al., which found that elevated preoperative CEA is independently
predictive of late (31 to 90 day) readmission, and that postoperative CA19-9
levels were two-fold higher in the late readmission group than in non-
readmitted patients. In their study, the majority of patients with
adenocarcinoma received postoperative chemotherapy, which was a factor in
late readmission.
SUGGESTED APPROACH
The patient initially presented in the clinical scenario will be readmitted to
an inpatient service where she will undergo a thorough evaluation, including
laboratory studies and imaging, to identify or exclude the development of any
new postoperative complications or infectious issues. In the absence of
identification of postoperative complications to which her complaints could
be attributed, she can be diagnosed as failure-to-thrive. Subsequently, her
admission should be geared towards reevaluating her needs and supports at
home, as well as whether an alternate disposition (facility or visiting nurse)
might be safer for her. A careful assessment should be made of her nutritional
status and oral intake; this assessment should include investigation into what
food is available at home, whether she can prepare it herself, and/or whether
she has someone who can help to prepare it. Additionally important is
follow-up education on nutritional recommendations, such as for small,
frequent meals that may be better tolerated in the short-term. It is also
important to reassess for evidence of pancreatic insufficiency and need for
pancreatic enzymes. Both the patient and her family (and/or relevant others)
should be involved in discussions regarding the resources, should have help
available for managing care at home, including an understand of, and
compliance with, all new medications and other recommended treatments.
Their expectations regarding her continued postoperative recovery should be
reviewed, and it is possible that her disposition would need to be changed to
either home with services or rehabilitation. Safeguards should be put in
place so that any difficulty or failure to improve can be identified prior to
reaching a severity where readmission would be again required. It is
possible that factors such as high preoperative CEA or small pancreatic duct
size that would predict readmission could be identified; however it is
equally possible that none exist. While continued efforts should be made to
optimize patient recovery and disposition following pancreatic resection in
order to lower readmission rates overall, it is unlikely and ultimately
unrealistic to expect that they will ever cease entirely.
SALIENT POINTS
The majority of readmissions following pancreatic resection occur
early, with reported rates of 15-23% in the first 30 days and 19-29% in
the first 90 days. The major reasons for readmission vary depending on
the time frame, with infectious complications and DGE generally
accounting for more of the early readmissions, and failure to thrive, as
well as infectious complications, contributing to those that occur later.
On discharge following pancreatic resection, special attention should
be paid to the adequacy of nutrition and to the ability of patients and
families to manage new medical therapies as well as carry out baseline
activities of daily living.
Clarifying and communicating postoperative expectations to patients,
families, visiting nurses and other providers is crucial and may lead to
a decrease in avoidable readmissions.
Even with appropriate and well-coordinated discharge, the
development of postoperative complications following initial
discharge and leading to readmission is to be expected to some degree.
Therefore, the use of readmission following pancreatic resection as an
indicator of quality care or as a tool for decreasing hospital
reimbursement should be discouraged.
SELECTED REFERENCES
1. Gawlas I, Sethi M, Winner M, Epelboym I, Lee JL, Schrope BA, et al.
Readmission after pancreatic resection is not an appropriate measure of
quality. Ann Surg Oncol. 2013;20(6):1781-7.
2. Ahmad SA, Edwards MJ, Sutton JM, Grewal SS, Hanseman DJ,
Maithel SK, et al. Factors influencing readmission after
pancreaticoduodenectomy: a multi-institutional study of 1302 patients.
Ann Surg. 2012;256(3):529-37.
3. Sadot E, Brennan MF, Lee SY, Allen PJ, Gonen M, Groeger JS, et al.
Readmission after pancreatic resection: causes and causality pattern.
Ann Surg Oncol. 2014;21(13):4342-50.
4. Glass CC, Gondek SP, Vollmer CM, Jr., Callery MP, Kent TS.
Readmission following pancreatectomy: what can be improved? HPB
(Oxford). 2013;15(9):703-8.
5. Kent TS, Sachs TE, Callery MP, Vollmer CM, Jr. Readmission after
major pancreatic resection: a necessary evil? J Am Coll Surg.
2011;213(4):515-23.
6. Fong ZV, Ferrone CR, Thayer SP, Wargo JA, Sahora K, Seefeld KJ, et
al. Understanding hospital readmissions after
General Hospital. J Gastrointest Surg. 2014;18(1):137-44; discussion
44-5.
CASE SCENARIO
A 50-year-old male, with a peri-ampullary adenocarcinoma, underwent a
pancreatoduodenectomy (PD). Intra-operatively, there was a soft and friable
pancreas with a posteriorly placed 3 mm duct. Duct-to-mucosa pancreato-
jejunostomy was performed with 5-0 polydiaxonone sutures. During the
anastomosis, two sutures cut through the pancreas making the operating
surgeon not very confident of the integrity and the ultimate outcome of the
anastomosis. Apart from taking all possible care and measures to ensure a
good outcome (i.e. pancreatic stents / specific drainage / falciform ligament
wrap / tissue glue etc), is there anything else which can be done in this high
risk situation?
█ BACKGROUND
Apart from an experienced surgeon and his / her team taking a fresh guard to
perform a flawless anastomosis appropriately tailored to the situation, in the
particular situation mentioned above, we would prophylactically administer
100 mcg of intravenous octreotide, even before commencing the anastomosis.
This is in view of the anastomosis being at high risk for a post-operative
pancreatic fistula (POPF). Although, there is much conflicting evidence on
this practice, and its routine use for all PDs has now been strongly debated,
we propose its benefit in some select scenarios, like the one described
above.
INTRODUCTION
Morbidity after PD persists at 20-40% despite a reduction in peri-operative
mortality of < 5% across continents. POPF is one of the major causes of
morbidity, prolonged hospitalization, and even mortality, after PD for
pancreatic and peri-ampullary cancer. The inherent corrosive nature of
activated pancreatic enzymes, often compounded by secondary infection, in
the vicinity of multiple vulnerable anastomoses makes early resolution of
POPF highly desirable in this life threatening situation. Apart from the
clinician monitoring the sick patient with a hawk’s eye, appropriate drainage,
antibiotics and nutrition, synthetic somatostatin analogues (octreotide,
lanreotide, vapreotide and pasireotide) have been used in clinical practice
with the sole goal of preventing POPF or its attending complications to
hasten post-operative recovery.
WHY USE SOMATOSTATIN

ANALOGUES?
A reduction of pancreatic secretions should theoretically hasten the
spontaneous resolution of POPF. This was historically achieved by
nasogastric aspiration and restricted oral intake. The advent of several
pharmacological agents like loperamide, atropine, pirenzepine, and
omeprazole all led to attempts to biochemically reduce the quantity of
pancreatic secretions, but their effectiveness remained questionable. One of
the most pursued agents, somatostatin analogues, have been used over the
years in different scenarios of POPF with varying results. Their use has
consistently shown reduced biochemical POPF but has had questionable
benefit in clinically-relevant POPF (CR-POPF). It has been noted in the mid
1990’s that octreotide has more benefit in reducing pancreatic secretions in
high volume POPF (>500 ml) vis-à-vis low volume POPF, results of which
have not been reproduced consistently.
WHAT ARE SOMATOSTATIN

ANALOGUES?
Somatostatin analogues are peptides with potent inhibitory effect on
pancreatic exocrine secretions. Naturally-acting somatostatin (t1/2 1-3 mins),
discovered in 1972, required con tinuous infusion in order to obtain
desirable effects, leading to reduced patient mobility and compliance. This
led to formulation of longer acting analogues like octreotide, which binds to
somatostatin receptor subtypes 2 and 5. Their role in reducing complications
after PD was a concept originally attributed to Klempa et al. Table 1 enlists
commonly used somatostatin analogue compounds. Although the most notable
of these is pasireotide, which has a longer half-life, broader binding profile
(somatostatin receptor subtypes 1, 3 and 5), and 5-40 times more binding
affinity than octreotide, we focus our discussion on octreotide, the most
extensively used somatostatin analogue world-wide over the past two
decades.
TABLE 1: Somatostatin and its analogues
TIMING AND DURATION OF

ADMINISTRATION
The timing of octreotide administration remains crucial and seems more
effective if administered before surgery, which is then continued intra-
operatively and post-operatively. Pancreatic secretions peak approximately
6-7 hours following administration of octreotide, and this rebound increase
in enzyme concentration may have deleterious effect on POPF healing.
Effective inhibition of pancreatic secretions is diminished after 7 days of
continuous octreotide administration due to downregulation of somatostatin
receptors (tachyphylaxis); hence, its use beyond 7 days, even in well-
established POPF, is questionable.
DO SOMATOSTATIN ANALOGUES
REALLY MAKE A DIFFERENCE?
Apart from the role of octreotide in influencing overall morbidity, its ability
to specifically reduce the POPF output, rate to spontaneous closure of POPF,
and impact on hospital stay have all been debated over the last few decades.
POPF
They have been shown to reduce overall POPF rates, but data is pre-
dated to the currently practiced standardized definition of clinically-
relevant POPF (Table 2: studies 1 to 4). A Cochrane review has shown
overall number of participants with postoperative complications was
significantly lower in the somatostatin analogue group (RR 0.70; 95%
CI 0.61 to 0.80; n = 1903). Incidence of POPF was lower in the
somatostatin analogue group (RR 0.66; 95% CI 0.55 to 0.79; n = 2206),
but on inclusion of trials that clearly distinguished clinically relevant
fistulas (CR-POPF), there was no significant difference between the
two groups (RR 0.69; 95% CI 0.38 to 1.28; n = 292).
The newer pasireotide (first US randomized controlled trial, albeit
single center, to show benefit of somatostatin analogue in impacting
POPF) has been found to significantly reduce Grade III or higher POPF,
leak or abscess (9 vs. 21%, p=0.006).
TABLE 2: Role of Somatostatin Analogues
Rate of Fistula Closure

There is some evidence favoring somatostatin analogues since they
seem to facilitate higher percentage of fistula closure, i.e. 65% with
octreotide, 84% with somatostatin, and 27% in the control group
(p=0.003), but this has not been consistently reported across studies
(Table 2: study 8).
Fistula Output
Some studies show that the use of octreotide or soma tostatin analogues
reduce pancreatic exocrine secretions and hence POPF output to about
50-70% after 24 hours of administration. (Table 2: study 9).
Peri-operative Mortality
There is unequivocal evidence to suggest that somatostatin and its
analogues do not decrease peri-operative mortality (Table 2: studies 1
and 4).
Re-operation Rates
Somatostatin and its analogues have no effect on re-operation rates (RR
1.26; 95% CI 0.58 to 2.70; n = 687).
Hospital Stay
Shorter hospital stay was noted with use of somatostatin analogues in
malignancy subgroups in initial studies but this effect was negated by
subsequent literature by the same investigators (MD -1.29 days; 95% CI
-2.60 to 0.03; n = 1314) between the groups.
Theoretically, there should be a bigger benefit for octreotide in high-risk

patients who were noted to have a soft or normal gland with preserved
secretory function, in view of its strong in vivo inhibitory effect on
pancreatic secretions. This was also noted in a retrospective analysis of
patients with high-risk glands, where octreotide prophylaxis was associated
with a decreased incidence (20% versus 35%) as well as decreased
morbidity of CR-POPF. These improved clinical outcomes were associated
with reduced resource use, translating to considerable cost savings ($11,849)
per high-risk patient. Also, as somatostatin analogues do not address the
underlying disease, their use is unlikely to affect survival outcomes.
DRAWBACKS OF EXISTING LITERATURE

The majority of the evidence regarding octreotide leading to reduction in
pancreatic leaks was generated prior to the current, universal, standardized
definition of POPF, and hence, there remains heterogeneity in the results even
amongst randomized controlled trials and meta-analyses.
As discussed earlier, octreotide found to be most beneficial in the pre-
operative period is difficult to evaluate solely for high-risk glands, as
pancreatic consistency is largely an intra-operative finding, and pre-
operatively, objective determination of this is either radiologically
challenging or remains a clinical assumption.
There is a large body of evidence regarding the use of somatostatin
analogues, and octreotide per se, and most of these lack unequivocal results,
which leaves us in a clinical dilemma as to what to follow in day-to-day
practice.
There is no recent level I evidence to support or refute the selective use
of somatostatin analogues. Even historically, there has been no level I
evidence to show potential clinical harm from the use of somatostatin
analogues, for us to strongly discourage their use.
WHEN WOULD WE USE SOMATOSTATIN

ANALOGUES TODAY?
A soft pancreas usually lacks parenchymal fibrosis and is known to have
well-preserved exocrine function, more often associated with a small, non-
dilated, main pancreatic duct. These glands are at high risk for occurrence of
pancreatic complications after PD, including POPF, and prophylactic use of
octreotide may not only benefit patients by decreasing incidence and
morbidity of clinically-relevant POPF but also help the health care system
with optimal utilization of resources and decreased cost of treatment. Other
anastomoses at high risk may be those that involve considerable blood loss
and those that are done for pathologies other than pancreatic ductal
adenocarcinoma, e.g. peri-ampullary and islet cell tumors. When PD is
performed at select high volume centers, having low morbidity and POPF
(Table 2, study 7), octreotide perhaps has no bearing on an already excellent
postoperative outcome. It is, perhaps, the more common less-experienced
and less-skillful surgeon who may benefit from the potential POPF-mitigating
effect of a somatostatin analogue in the high risk anastomosis/gland, in the
want of better evidence.
WHAT ARE THE DISADVANTAGES OF

USING SOMATOSTATIN ANALOGUES?
The side effects are hyperglycemia and pain at the injection site, as well as
rash, hot flashes, nausea, vomiting, and asymptomatic biliary sludge.
Practically, despite being historically known for their relatively expensive
nature, the inconvenience of parenteral administration and, most of all,
conflicting evidence discourage their routine use. However, cost-benefit
analyses have begun to show that expenditure on somatostatin analogues is
sometimes prudent to cut resource expenditure attributed to CR-POPF and its
implications.
A large retrospective analysis concluded that octreotide may potentiate
CR- POPF development in the presence of risk factors, but this study used
octreotide at the surgeon’s discretion and had significantly higher high-risk
pathology and soft gland-texture in the ‘octreotide cohort’. These features
may serve as confounding factors in themselves, leaving us to wonder
whether it’s the inherent gland risk factor which led to the POPF or the use of
octreotide, as these were more likely to leak even without administration of
octreotide. As mentioned earlier, there has been no level I evidence to show
potential clinical harm from the use of somatostatin analogues for us to
strongly discourage their use.
CONCLUSION
Though some randomized controlled trials convincingly prove its advantage
(Table 2, studies 1 to 4), other trials have presented equivocal results (Table
2, studies 6 and 7 ), and meta-analyses advise adequately powered trials
with low risk of bias. This leads one to believe that evidence, Level I or
otherwise, can be practice-changing only to some extent, and, in the
remaining situations, it is left to the discretion of the surgeon on whether
somatostatin analogues should be used, requiring individualized treatment
decisions more than protocols and evidence-based practices.
In our experience of over a thousand Whipple resections while
documenting a CR-POPF rate of 13.4% makes us believe that selective use
of octreotide retains its place in the management of a high risk pancreatic
anastomosis. Although a Cochrane review suggests that based on the current
available evidence, somatostatin and its analogues are recommended for
routine use in people undergoing PD, we advocate its selective use in high-
risk anastomoses as described above.
SALIENT POINTS
The use of octreotide to reduce pancreatic exocrine secretions and
post-operative pancreatic fistula output after Whipple resection (PD)
remains a matter of discussion and debate.
We have convincing level I evidence to show that somatostatin
analogues reduce post-operative complications, but this may have
included biochemical POPF as well.
We also have level I evidence from highly specialized centers, which
showed no-benefit no-harm, but this may not be applicable to the
relatively lesser-experienced relatively-lower volume centers who
still document POPF rates in the range of 15-20%.
We already have retrospective evidence to show that the use of
octreotide in the high risk gland decreased incidence and morbidity of
clinically relevant POPF and translated into lower resource-usage and
cost-saving.
Until it is proven by further adequately powered trials with low risk of
bias, selective use of octreotide in the high risk anastomosis is our
choice, as its potential benefits may outweigh any otherwise noticed
incidental harm.
SELECTED REFERENCES
1. Shrikhande SV, Sivasanker M, Vollmer CM, Friess H, Besselink MG,
Fingerhut A, et al. Pancreatic anastomosis after
pancreatoduodenectomy: A position statement by the International Study
Group of Pancreatic Surgery (ISGPS). Surgery 2017; 161(5):1221-
1234.
2. Klempa I, Schwedes U, Usadel KH. Prevention of postoperative
pancreatic complications following duodenopancreatectomy using
somatostatin. Chirurg. 1979;50(7):427–31.
3. Gurusamy KS, Koti R, Fusai G, Davidson BR. Somatostatin analogues
for pancreatic surgery. Cochrane Database Syst Rev. 2013 ;
(4):CD008370.
4. Vanounou T, Pratt WB, Callery MP, Vollmer CM. Selective
administration of prophylactic octreotide during
pancreaticoduodenectomy: a clinical and cost-benefit analysis in low-
and high-risk glands. J Am Coll Surg. 2007; 205(4):546–57.
5. McMillan MT, Christein JD, Callery MP, Behrman SW, Drebin JA,
Kent TS et al. Prophylactic octreotide for pancreatoduodenectomy:
more harm than good? HPB (Oxford). 2014; 16(10):954-62.

CASE SCENARIO
A 68-year-old woman with a 4 cm well-differentiated neuroendocrine tumor
of the pancreatic head presents for pancreaticoduodenectomy. At surgery, her
pancreatic duct is 3-4 mm, and the parenchymal texture is soft. A two-layer
end-to-side pancreaticojejunostomy anastomosis is performed. On post-
operative day number 5, she is noted to have increased nausea associated
with a mild leukocytosis. The next day, she is noted to have a fever of 101.7
°F. On evaluation, her vital signs are normal, she complains of mild
abdominal pain and nausea, and her incision is intact, but the character of her
drain output has changed. The drain fluid is sent for an amylase level, which
returns elevated at >7500 U/L. A CT scan is then performed, which reveals a
peri-pancreatic fluid collection not in communication with the surgical drain.
A percutaneous drain is placed, which also reveals amylase-rich fluid. The
patient is made NPO and started on intravenous antibiotics. When her nausea
resolves, she is started on a diet and advanced as tolerated. Over the course
of the next week, her fever and leukocytosis resolve, and the remainder of
her recovery proceeds uneventfully. At a follow-up appointment two weeks
post-operatively, the drain outputs have steadily declined to nearly zero, and
the drains are removed.
█ BACKGROUND
The case history as presented is all too common to even the most technically
gifted pancreatic surgeons. When performed at high-volume institutions, the
mortality rate following pancreatic resection is typically in the range of 1-
4%. Despite these low mortality rates, morbidity from the operation remains
high, generally in the range of 35-50%, a large portion of which is
attributable to post-operative pancreatic fistulas (POPF). In 2005, the
International Study Group on Pancreatic Fistula (ISGPF) established criteria
to define and grade POPF to facilitate the study of this common clinical
problem. That definition specified a POPF as present when drain output on
or after post-operative day 3 has an amylase content greater than three times
the serum level. Fistulas were further classified into three grades based on
clinical significance, as indicated in Table 1 , with those that prolong
recovery time, or require specific treatment, considered clinically relevant
(grades B and C).
TABLE 1: Post-operative pancreatic fistula grading
system according to the International Study Group on
Pancreatic Fistulas (Adapted with permission from Elsevier
in Bassi C, et al. The 2016 update of the International Study
Group (ISGPS) definition and grading of postoperative
pancreatic fistula: 11 years after. Surgery 2017;161(3):584-
591).
The incidence rates of POPF have been reported to be between 3-40%,

varying primarily based on the type of resection performed, size of the main
pancreatic duct, and the character of the gland. To better predict the
likelihood of a patient developing a clinically relevant post-operative
pancreatic fistula, Callery, et. al used the ISGPF criteria to establish a
Fistula Risk Score (FRS), as outlined in Table 2 , based on four main patient
and operative characteristics. Even prior to the development of this FRS, it
was widely held that because a small duct makes the anastomosis more
technically challenging, and normal soft pancreatic parenchyma is less
capable of holding sutures and secretes more digestive enzymes, these
characteristics were associated with higher rates of POPF. As POPF
frequently results in significant morbidities such as intra-abdominal
abscesses, hemorrhage, and/or sepsis, there has been enormous attention paid
to its prevention. Many operative techniques and adjuncts have been studied
in an attempt to eliminate or at least minimize the incidence and severity of
POPF, including stenting the duct, fibrin glues, and patches and the use of
octreotide and other somatostatin analogues; however, none has yet
demonstrated a consistent, clinically meaningful risk reduction in POPF.
TABLE 2: Fistula Risk Score used to predict the

development of clinically-relevant post-operative
pancreatic fistulas following pancreaticoduodenectomy
(Reprinted with permission from Elsevier in Callery MP, et
al. A prospectively validated clinical risk score accurately
predicts pancreatic fistula after pancreatoduodenectomy. J
Am Coll Surg. 2013;216(1):1-14).
A number of large prospective randomized controlled trials, as well as

meta-analyses and systematic reviews, have now been completed, evaluating
the role of somatostatin analogues, primarily octreotide, in the prevention of
POPF. The conduct of so many studies in this area attests to the importance of
the problem, the complexity of studying it, and the fact that the data has been
conflicting as to the value of octreotide. Upon close examination of the
available data, there is a trend over the last two decades toward studies
demonstrating less of a benefit, with some even suggesting potentially
detrimental effects.
Of the studies that support the prophylactic use of octreotide to prevent
fistula formation, most were published prior to the ISGPF’s establishment of
a consensus definition of post-operative pancreatic fistula, while more recent
studies indicate no significant improvement in incidence rates with
prophylactic use. Prior to this definition, the diagnosis of fistula was
inconsistent and hence less reliable, making comparisons between
institutions and studies challenging. However, even if this trend is
discounted, there remains no clear signal from the literature suggesting
octreotide’s use is beneficial following pancreaticoduodenectomy in
particular.
A Cochrane review conducted in 2013 analyzed the results from 19
randomized controlled trials performed between 1991 and 2012, comparing
the use of a somatostatin analogue following pancreatic resections to placebo
or no treatment. Though the review noted a small decrease in the overall rate
of POPFs, it is critical to note there was no significant difference in the
incidence of those that were clinically significant (grade B or C).
Additionally, despite a decrease in the overall rate of post-operative
complications—including sepsis, associated with the use of a somatostatin
analogue—there was no improvement specifically in the incidence of
anastomotic leaks, renal failure, bleeding, re-operation, intra-abdominal
fluid collections or abscesses, delayed gastric emptying, pulmonary
complications, or shock. Furthermore, this review revealed no significant
difference between the two groups in mortality rates, length of hospital stay,
or length of ICU stay, suggesting that any difference in overall complication
rates was not clinically significant. Importantly, the use of somatostatin
analogues was associated with a higher rate of treatment-related adverse
effects, further arguing against its routine use.
Another recent review used the incidence of clinically relevant post-
operative pancreatic fistula, based on the ISGPF criteria, as an indicator of
performance quality in surgeons and institutions performing
pancreaticoduodenectomy. Aside from an expected correlation with higher
Fistula Risk Scores, increased rates of clinically relevant fistulas were also
associated with the surgeons and institutions that routinely used prophylactic
octreotide. So, this analysis failed to demonstrate any benefit to routine
prophylactic use and actually revealed a trend toward higher rates of POPF.
While it is not clear why octreotide would lead to potentially detrimental
effects, the authors suggested the contribution of decreased pancreatic and
gastroduodenal perfusion leading to poor anastomotic healing as a likely
cause. After 20 years of study, if there has not been a clear signal that
octreotide is of benefit, we would argue the answer is in fact clear; it is not
of sufficient benefit to warrant routine use.
A recent randomized controlled trial compared the use of prophylactic
pasireotide, a long-acting somatostatin analogue, to placebo following
pancreatic resections at a single institution. Though the authors did find a
reduction in clinically significant pancreatic fistulas, the primary outcome
was based on a definition and grading system that were different from the
ISGPF criteria, making comparison of their results to the rest of the literature
difficult. As a secondary outcome, they did also find a slight improvement in
the incidence of ISGPF grade B/C fistulas; however, this was far less
significant than seen in the primary outcome. That being said, if these results
are reproducible by others in the context of the standardized criteria used in
the literature, then perhaps this agent may make its way to the clinic as a
standard of care, particularly in higher risk patients. In general, we view this
controversy as somewhat misguided as it generally ignores the basic tenets of
fistula risk; i.e., we know who the higher risk and lower risk patients
(pancreata) are for POPF, yet we continue to focus discussions of use of
prophylaxis on all patients or none. We would venture that in the hands of an
experienced pancreatic surgeon, the risk of POPF in a patient with a 1 cm
pancreatic duct and an atrophic, firm gland associated with a diagnosis of
IPMN or pancreatic cancer is well under 5 %. Thus, the use of any
prophylactic agent against POPF in this population would needlessly expose
at least 95 patients out of 100 to the cost and adverse effects associated with
a somatostatin analogue. In 2017, it makes sense to us that decisions
regarding use of a somatostatin analogue be based on a general understanding
of a reasonable expected risk reduction accounting for the expected risk of
POPF for an individual patient—lo and behold, personalized medicine!
Unfortunately, too few of the studies conducted on this topic have focused
specifically on high risk strata or at least have provided enough data to begin
to crunch numbers for patient subsets. This clearly should be a focus of the
field going forward.
At present, given the available evidence and the available agents
(octreotide for most institutions), there are few circumstances in which the
prophylactic use of a somatostatin analogue is sensible. Our experience
echoes the cumulative findings in the literature as we have found no
substantial value in its prophylactic use. That said, we may be faced with a
patient in whom the anatomy suggests a very high risk of POPF, and in
particular, in the presence of other co-morbidities, which would make a
POPF of greater threat to a complete recovery than normal; in this
circumstance, though the effect may be weak and the benefit small, the upside
of even modest risk reduction becomes reasonable. Though there is a paucity
of literature available on the role of somatostatin analogues in the
management of already-established pancreatic fistulas, our experience
suggests that there are also limitations to its therapeutic use.
Where we have found the most value is in the management of rare high-
output fistulas that are less likely to resolve quickly or those that may cause
significant problems, such as electrolyte disturbances or external
communications leading to skin breakdown. The use of octreotide in and of
itself does not appear to decrease the amount of time it takes for a fistula to
close; rather, it may decrease the output volume. So, for low output fistulas
or those expected to close quickly, the reduction in adverse effects and
discomfort associated with octreotide is unlikely to be worthwhile.
However, for those fistulas with high volume output, by decreasing the output
volume, the fistula may become easier to manage for both the patient and the
provider and less uncomfortable for the patient. Furthermore, by decreasing
the output, the potential for electrolyte abnormalities, skin breakdown, and
other fistula-related complications can be reduced.
The potential adverse effects of octreotide must always be recognized.
Typically, octreotide is administered through subcutaneous injections, which
are often associated with pain, bruising, and scarring, all of which can limit
compliance with the regimen. Beyond the minor discomforts, we have
observed a number of more consequential side effects, including
hyperglycemia, bradycardia, malaise, nausea, diarrhea, and depending on the
length of therapy, even cholestasis and malabsorption.
With the absence of a clear improvement in outcomes, in the setting of an
increase in treatment-associated adverse effects, we do not support the
routine use of octreotide following pancreaticoduodenectomy. Furthermore,
we do not advocate its routine use even in all patients who develop post-
operative pancreatic fistulas. We do believe it is reasonable to consider its
use in the highest risk patient subgroups (i.e. soft gland, small duct, obese
patients, and/or the setting of a technically unsatisfactory anastomosis) and in
those clinically relevant, high-output fistulas not expected to close quickly.
Hopefully, this debate will soon become moot if other somatostatin
analogues, such as pasireotide, are validated as beneficial by ongoing Phase
III trials.
SALIENT POINTS
Post-operative pancreatic fistulas are a significant source of morbidity
associated with pancreatic resections.
There is no clear signal in the literature to suggest a benefit to the
routine prophylactic use of somatostatin analogues following
pancreaticoduoedenectomy to prevent fistula formation.
By decreasing the high output fistula volume, therapeutic somatostatin
analogues may make fistulas easier to manage and limit their
associated complications.
SELECTED REFERENCES
1. Kitahata Y, Kawai M, Yamaue H. Clinical trials to reduce pancreatic
fistula after pancreatic surgery— review of randomized controlled
trials. Transl Gastroenterol Hepatol 2016;1:4.
2. Bassi C, Dervenis C, Butturini G, Fingerhut A, Yeo C, Izbicki J, et al.
Postoperative pancreatic fistula: An international study group (ISGPF)
definition. Surgery 2005; 138(1):8-13.
3. Hackert T, Werner J, Buchler MW. Postoperative pancreatic fistula.
Surgeon 2011; 9(4):211-7.
4. Callery MP, Pratt WB, Kent TS, Chaikof EL, Vollmer CM Jr. A
prospectively validated clinical risk score accurately predicts
pancreatic fistula after pancreaticoduodenectomy. J Am Coll Surg.
2013;216(1):1-14.
5. McMillan MT, Soi S, Asbun HJ, Ball CG, Bassi C, Beane JD, et al.
Risk-adjusted Outcomes of Clinically Relevant Pancreatic Fistula
Following Pancreaticoduodenectomy: Model for Performance
Evaluation. Ann Surg. 2016; 264(2):344-52.
6. Gurusamy KS, Koti R, Fusai G, Davidson BR. Somatostatin analogues
for pancreatic surgery. Cochrane Database of Syst Rev. 2013;
(4):CD008370.
7. Allen PJ, Gonen M, Brennan MF, Bucknor AA, Robinson LM, et al.
Pasireotide for postoperative pancreatic fistula. N Engl J Med 2014;
370(21):2014-22.
8. Gans SL, van Westreenen HL, Kiewiet JJ, Rauws EA, Gouma DJ,
Boermeester MA. Systematic review and meta-analysis of somatostatin
analogues for the treatment of pancreatic fistula. Br J Surg.
2012;99(6):754-60.


Hígado Páncreas y Transplante Ahpba Capitulo Pancreas

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Hígado Páncreas y Transplante Ahpba Capitulo Pancreas

Uploaded by

Copyright:

Available Formats

BEAUX BOOKS PUBLISHING

P.O. BOX 53364

Hepato-Pancreato-Biliary and Transplant Surgery: Practical Management of Dilemmas

Mallory Williams, MD, MPH

Quyen D. Chu, MD, MBA

Jad Abou-Khalil, MD, MSc (Epi) FRCSC

Marwan S. Abouljoud, MD, FACS, CPE, MMM

Peter Abrams, MD, FACS

Anya Adair, MBBS, PhD, FRCS (Eng)

David B. Adams, MD, FACS

Syed A. Ahmad, MD, FACS

Joseph Ahn, MD, MS, FAASLD

Okan Akhan, MD, FCIRSE, FESGAR

Devrim Akinci, MD, FCIRSE

Farzad Alemi, MD FACS

Peter J. Allen, MD, FACS

John D. Allendorf, MD, FACS

Thomas A. Aloia, MD, FACS

Eisar Al-Sukhni, MD, MSc, FRCSC

Fernando Andrés Alvarez, MD

Christopher D. Anderson, MD, FACS

Wellington Andrus, MD, PhD

Anand Annamalai, MD, FACS

Avo Artinyan, MD, MS, FACS

Horacio J. Asbun, MD, FACS

Reza Askari, MD, FACS

Elizabeth A. Bailey, MD, MEd

Emily J. Baird, MD, PhD

Marshall S. Baker, MD, MBA, FACS

Chad G. Ball, MD, MSc, FRCSC, FACS

Todd H. Baron, MD, FASGE

Savio George Barreto, MD

Alfred E. Baylor, III, MD, FACS

Eliza W. Beal, MS, MD, FACS

Stephen W. Behrman, MD, FACS

Jacques Belghiti, MD, HACS

Matthew Benns, MD, FACS

Jennifer Berumen, MD, FACS

Marc G.H. Besselink, MD, PhD, MSc

Sylvester M. Black, MD, PhD

Adam S. Bodzin, MD, FACS

Christoph E. Broelsch, MD, FACS

Kimberly M. Brown, MD, FACS

David S. Bruce, MD, FACS

Joseph F. Buell, MD, MBA, FACS

Mark P. Callery, MD, FACS

Andrew M. Cameron, MD, PhD, FACS

Andre Campbell, MD, FACS, FACP, FCCM

Jamie Case, PhD

Eugene P. Ceppa, MD, FACS

John A. Chabot, MD, FACS

William C. Chapman, MD, FACS

Kenneth D. Chavin, MD, PhD, FACS

Edward E. Cho, MD, ScM

Michael A. Choti, MD, FACS

John D. Christein, MD, FACS

Quyen D. Chu, MD, MBA, FACS

Francisco G. Cigarroa, MD, FACS

Sean Cleary MD, MSc, MPH, FRCSC, FACS

Robert J.S. Coelen, MD, PhD

Felipe José Fernández Coimbra, MD, FACS

Waldo Concepcion, MD, FACS

Claudius Conrad, MD, PhD, FACS