ISSN 0042-9007 (Print) ISSN 1423-0410 (Online) Volume 116 Number 5 MAY 2021

The International Journal of Transfusion Medicine

In this issue
The immune potential of ex vivo stored platelets: a review

Blood supply management in times of SARS-CoV-2 pandemic:

challenges, strategies adopted, and the lessons learned from the
experience of a hospital-based blood centre

Anti-A and SARS-CoV-2: an intriguing association

International Forum on Transfusion Practices in Hematopoietic Stem

Cell Transplantation
Vox Sanguinis
Citation Reports/Science Edition (Clarivate Analytics); MEDLINE/PubMed (NLM);
Nutrition Abstracts & Reviews Series A: Human & Experimental (CABI); Pig News &
Information (CABI); ProQuest Central (ProQuest); ProQuest Health & Medical Complete
(ProQuest); ProQuest Research Library (ProQuest); Protozoological Abstracts (CABI);
PubMed Dietary Supplement Subset (NLM); Review of Medical & Veterinary Entomology
(CABI); Review of Medical & Veterinary Mycology (CABI); Rural Development Abstracts
International Journal of Blood Transfusion (CABI); Science Citation Index (Clarivate Analytics); Science Citation Index Expanded
(Clarivate Analytics); Tropical Diseases Bulletin (CABI); Veterinary Bulletin (CABI).
Aims and Scope
Vox Sanguinis reports on important, novel developments in transfusion medicine. Original
Official Journal of the papers, reviews and international fora are published on all aspects of blood transfusion and
International Society of Blood Transfusion tissue transplantation, comprising five main sections:
1 Transfusion-Transmitted Disease and its Prevention Identification and epidemiology of
Vox Sanguinis has a world wide web site at: infectious agents transmissible by blood; Bacterial contamination of blood components;
Donor recruitment and selection methods; Pathogen inactivation
www.voxsanguinis.com 2 Blood-Component Collection and Production Blood collection methods and devices
(including apheresis); Plasma fractionation techniques and plasma derivatives;
Preparation of labile blood components; Inventory management; Haematopoietic
Editor-in-Chief progenitor cell collection and storage; Collection and storage of tissues; Quality
M. Lozano, Barcelona (Spain) management and good manufacturing practice; Automation and information technology
3 Transfusion Medicine and New Therapies Transfusion thresholds and audits; Haemovigi-
Section Editors lance; Clinical trials regarding appropriate haemotherapy; Non-infectious adverse affects
Transfusion-Transmitted Disease and its Prevention of transfusion; Therapeutic apheresis; Support of transplant patients; Immunotherapy
4 Immunohaematology and Immunogenetics Autoimmunity in haematology; Allo-
C. Seed, Perth (Australia) immunity of blood; Pre-transfusion testing; Immunodiagnostics; Immunobiology;
Blood-Component Collection & Production Complement in immunohaematology; Blood-typing reagents; Genetic markers of blood
D.C. Marks, Sydney (Australia) cells and serum proteins: polymorphisms and function; Genetic markers and disease;
Transfusion Medicine and New Therapies Parentage testing and forensic immunohaematology
P. Tiberghien, Paris (France) 5 Cellular Therapy (including clinical, quality and regulatory aspects) Cell-based therapies;
Stem cell sources; Stem cell processing and storage; Stem cell products; Stem
Immunohaematology & Immunogenetics cell plasticity; Regenerative medicine with cells; Cellular immunotherapy; Molecular
J. R. Storry, Lund (Sweden) therapy; Gene therapy
Cellular Therapy This comprehensive coverage has made the journal essential reading for a wide range of specialists
Z. M. Szczepiorkowski, Lebanon, NH, (USA) interested in the present state of transfusion research and practice.
Donors and Donations Information for Subscribers
K. van den Hurk, Amsterdam (The Netherlands) Vox Sanguinis is published in 8 issues per year. Institutional subscription prices for
2020 are: Print & Online: US$2448 (US), US$2854 (Rest of World), €1680 (Europe), £1325
International Forum Editor (UK). Prices are exclusive of tax. Asia-Pacific GST, Canadian GST/HST and European VAT will be
N. Dunbar, Lebanon, NH, (USA) applied at the appropriate rates. For more information on current tax rates, please go to www.
wileyonlinelibrary.com/tax-vat. The price includes online access to the current and all online
Review Editors backfiles to January 1st 2014, where available. For other pricing options, including access
L. van de Watering, Leiden (The Netherlands) information and terms and conditions, please visit www.wileyonlinelibrary.com/access.
Z. M. Szczepiorkowski, Lebanon, NH, (USA) Delivery Terms and Legal Title
Text Editor Where the subscription price includes print issues and delivery is to the recipient’s address,
delivery terms are Delivered at Place (DAP); the recipient is responsible for paying any
D.W. Huestis, Tucson, AZ (USA) import duty or taxes. Title to all issues transfers FOB our shipping point, freight prepaid.
We will endeavour to fulfil claims for missing or damaged copies within six months of
ISBT Standing Committee on Vox Sanguinis publication, within our reasonable discretion and subject to availability.
G. Clarke, Chairperson, Edmonton (Canada) Back issues: Single issues from current and recent volumes are available at the current single
L. Fung, Brisbane (Australia) issue price from cs-journals@wiley.com. Earlier issues may be obtained from Periodicals
E. Jansen, Amsterdam (The Netherlands) Service Company, 351 Fairview Avenue – Ste 300, Hudson, NY 12534, USA. Tel: +1 518
D. Teo, Singapore (Singapore) 822-9300, Fax: +1 518 822-9305, Email: psc@periodicals.com
M. Lozano, Editor in Chief, Barcelona (Spain) Copyright and Copying (in any format)
P. van der Meer, Amsterdam (The Netherlands) Copyright © 2021 International Society of Blood Transfusion. All rights reserved. No part of
this publication may be reproduced, stored or transmitted in any form or by any means
Observers: without the prior permission in writing from the copyright holder. Authorization to copy
E. Wood, ISBT President items for internal and personal use is granted by the copyright holder for libraries and other
users registered with their local Reproduction Rights Organisation (RRO), e.g. Copyright
J. White, ISBT Executive Director Clearance Center (CCC), 222 Rosewood Drive, Danvers, MA 01923, USA (www.copyright.com),
Journal Customer Services: For ordering information, claims and any enquiry provided the appropriate fee is paid directly to the RRO. This consent does not extend to other
kinds of copying such as copying for general distribution, for advertising or promotional
concerning your journal subscription please go to https://hub.wiley.com/community/ purposes, for republication, for creating new collective works or for resale. Permissions for
support/onlinelibrary or contact your nearest office. such reuse can be obtained using the RightsLink “Request Permissions” link on Wiley Online
Americas: Email: cs-journals@wiley.com; Tel: +1 781 388 8598 or +1 800 835 6770 Library. Special requests should be addressed to: permissions@wiley.com
(toll free in the USA & Canada). Online Open
Europe, Middle East and Africa: Email: cs-journals@wiley.com; Tel: +44 (0) 1865 778315. Vox Sanguinis accepts articles for Open Access publication. Please visit http://olabout.wiley.
Asia Pacific: Email: cs-journals@wiley.com; Tel: +65 6511 8000. com/WileyCDA/Section/id-828081.html for further information about OnlineOpen.
Japan: For Japanese speaking support, Email: cs-japan@wiley.com. Copyright policy
Visit our Online Customer Help at https://hub.wiley.com/community/support/ Papers accepted must be licensed for publication in Vox Sanguinis and a completed Copyright
Transfer Agreement Form must accompany every accepted paper. Authors will be required
onlinelibrary. to license copyright in their paper to John Wiley & Sons Ltd. Upon acceptance of an article,
Production Editor: Shirley Villeza (email: vox@wiley.com). corresponding authors must log into Author Services to complete the licence agreement of
their paper.
Founded 1956 by J.J. van Loghem, L.P. Holländer, J. Dausset,
A. Hässig and J. Julliard (formerly Bulletin of the Central Laboratory Wiley’s Corporate Citizenship initiative seeks to address the environmental, social,
economic, and ethical challenges faced in our business and which are important to our diverse
of the NRC Blood Transfusion Service, founded 1951) stakeholder groups. Since launching the initiative, we have focused on sharing our content
Past Editors-in-Chief: with those in need, enhancing community philanthropy, reducing our carbon impact,
creating global guidelines and best practices for paper use, establishing a vendor code of
J.J. van Loghem, 1956–1960; ethics, and engaging our colleagues and other stakeholders in our efforts. Follow our progress
W.H. Crosby, 1960–1963 (N. and S. America); at www.wiley.com/go/citizenship
L.P. Holländer, 1960–1970 (Europe); Wiley is a founding member of the UN-backed HINARI, AGORA, and OARE initiatives. They are
F.H. Allen, 1963–1977 (N. and S. America); now collectively known as Research4Life, making online scientific content available free or at
M.G. Davey, 1970–1980 (Africa, Asia and Australia); nominal cost to researchers in developing countries. Please visit Wiley’s Content Access –
N.R. Rose, 1977–1980 (N. and S. America); Corporate Citizenship site: http://www.wiley.com/WileyCDA/Section/id-390082.html
C.P. Engelfriet, 1977–1996; The journal to which you are submitting your manuscript employs a plagiarism detection
system. By submitting your manuscript to this journal you accept that your manuscript may be
M. Contreras, 1996–2003; screened for plagiarism against previously published works.
W.R. Mayr, 2003–2011;
Disclaimer
D. Devine, 2011–2020. The Publisher, International Society of Blood Transfusion and Editors cannot be held
responsible for errors or any consequences arising from the use of information contained in this
Abstracting and Indexing Services: The Journal is indexed by Abstracts in Anthropology journal; the views and opinions expressed do not necessarily reflect those of the Publisher,
(Sage); Abstracts on Hygiene & Communicable Diseases (CABI); Academic Search International Society of Blood Transfusion and Editors, neither does the publication of
(EBSCO Publishing); Academic Search Alumni Edition (EBSCO Publishing); Academic advertisements constitute any endorsement by the Publisher, International Society of Blood
Search Premier (EBSCO Publishing); AGRICOLA Database (National Agricultural Transfusion and Editors of the products advertised.
Library); BIOBASE: Current Awareness in Biological Sciences (Elsevier); Biological VOX SANGUINIS, (Online ISSN: 1423-0410 Print ISSN: 0042-9007) is published monthly
except in March, June, September, & December. US mailing agent: Mercury Media Processing,
Abstracts (Clarivate Analytics); BIOSIS Previews (Clarivate Analytics); CAB Abstracts® LLC, 1850 Elizabeth Avenue, Suite #C, Rahway, NJ 07065 USA. Periodical postage paid at
(CABI); CABDirect (CABI); CAS: Chemical Abstracts Service (ACS); CSA Biological Rahway, NJ. Postmaster: Send all address changes to VOX SANGUINIS, John Wiley & Sons
Sciences Database (ProQuest); CSA Environmental Sciences & Pollution Management Inc., C/O The Sheridan Press, PO Box 465, Hanover, PA 17331 USA.
Database (ProQuest); CSA Virology & AIDS Abstracts (ProQuest); Current Contents: Life For submission instructions, subscription and all other information visit: www.
Sciences (Clarivate Analytics); Embase (Elsevier); Global Health (CABI); HEED: Health wileyonlinelibrary.com/journal/vox.
Economic Evaluations Database (Wiley-Blackwell); Index Veterinarius (CABI); Journal Printed in Singapore by C.O.S. Printers Pte Ltd.

116/5/2021
Contents
Review Article
477 The immune potential of ex vivo stored platelets: a review
B. Wood, M. P. Padula, D. C. Marks & L. Johnson
Commentary
489 The interaction between Glycophorin A (GPA) and Band 3 in the
formation of the Wright b (Wrb) antigen S. Ekman, R. T. Barnard,
R. Flower, A. Gould & X. T. Bui
493 Evaluation of SARS-CoV-2 antibody titers and potency for
convalescent plasma donation: a brief commentary E. Wouters,
M. Steenhuis, H. Schrezenmeier, P. Tiberghien, H. Harvala,
H. B. Feys & E. van der Schoot
Original Papers
Donors and Donations
497 Blood supply management in times of SARS-CoV-2 pandemic
– challenges, strategies adopted, and the lessons learned from the
experience of a hospital-based blood centre H. C. Pandey,
P. Coshic, C. C S, P. J. Arcot & K. Kumar
504 Compliance and attitudes of blood donors following transitioning
from permanent to 12-month deferral of men who have sex with
men in Hong Kong J.-C. Lau, C.-K. Lee, C.-P. Chan, J.-S. Leung,
C.-M. Poon & S.-S. Lee
513 Unknown, so also unvalued? Blood donation awareness and
attitudes of potential donors of Dutch and African descent
E. F. Klinkenberg, M. P. Fransen, W. L. de Kort, E. M. Huis in ’t Veld
& J. C. van Weert
Blood component collection and Production
524 Ovine red cell concentrates for transfusion research – is the
storage lesion comparable to human red cell concentrates?
G. Simonova, R. Wellburn, Y. L. Fung, J. F. Fraser & J.-P. Tung
533 The ’rejuvenating factor’ TIMP-2 is detectable in human blood
components for transfusion J. Hoefer, C. Dal-Pont, S. Jochberger,
R. Fantin & H. Schennach
540 A microfluidic analysis of thrombus formation in reconstituted
whole blood samples comparing spray-dried plasma versus fresh
frozen plasma R. S. Bercovitz, C. S. Drew, C. L. Bushee,
M. A. Popovsky, K. D. Friedman & W. Q. Anani
547 Quality assessment of red blood cell suspensions derived from
pathogen-reduced whole blood I. Kumukova, P. Trakhtman,
N. Starostin, D. Borsakova, A. Ignatova & Y. Bayzyanova
Transfusion-transmitted Disease and it Prevention
557 Anti-A and SARS-CoV-2: an intriguing association
V. de Freitas Dutra, C. Bonet-Bub, A. P. H. Yokoyama, R. Achkar,
R. R. G. Machado, M. Assunção, G. Candelária, C. P. Soares,
R. M. Fachini, R. Fontão-Wendel, N. Hamerschlak, L. F. L. Reis,
D. B. Araujo, V. Nudelman, J. R. R. Pinho, L. V. Rizzo, A. M. Sakashita,
P. Scuracchio, E. L. Durigon, S. Wendel & J. M. Kutner
This journal is available online at Wiley Online Library.
Visit wileyonlinelibrary.com to search the articles and
register for table of contents e-mail alerts.
Vox Sanguinis (2021) 116, 475–613
© 2021 International Society of Blood Transfusion
564 An economic reappraisal of hepatitis B virus testing strategy for
blood donors in Taiwan T. B. Matanhire & S.-W. Lin
Transfusion Medicine
574 Blood transfusion activity in a general hospital during the
COVID-19 pandemic K. Marín-Mori, I. González-Gascón y Marín,
M.-Á. Foncillas-García, C. Muñoz-Novas, M. Infante,
J. Churruca-Sarasqueta, E. Landete-Hernández, B. Bueno-García,
M. Duffort-Falco & J.-Á. Hernández-Rivas
581 Costs associated with transfusion therapy in patients with
myelodysplastic syndromes in Sweden: a nationwide
retrospective cohort study J. Zhao, T. Dahlén & G. Edgren
Immunohematology
591 Clinical consequences of the extremely rare anti-PP1Pk
isoantibodies in pregnancy: a case series and review of the
literature P. Di Ciaccio, B. Cutts, T. I. Alahakoon, P. M. Dennington,
L. A. Soo & J. Curnow
601 Red cell genotyping of rare blood donors: donation behaviour and
data visualization W. Q. Anani, J. Gottschall & G. A. Denomme
International Forum
609 International Forum on Transfusion Practices in Haematopoietic
Stem-Cell Transplantation: Summary P. Solves, M. Lozano,
E. Zhiburt, J. Anguita Velasco, A. Maria Pérez-Corral,
S. Monsalvo-Saornil, S. Yamazaki, H. Okazaki, K. Selleng,
K. Aurich, W. Krüger, A. Buser, A. Holbro, L. Infanti, G. Stehle,
L. Pierelli, A. Matteocci, L. Rigacci, K. M. K. De Vooght,
J. H. E. Kuball, K. L. Fielding, D. A. Westerman, E. M. Wood,
C. S. Cohn, A. Johnson, M. B. C. Koh, D. Qadir,
C. Cserti-Gazdewich, E. Daguindau, F. Angelot-Delettre,
P. Tiberghien, S. Wendel-Neto, R.-M. Fachini, S. Morton,
C. Craddock, M. Lumley, J. Antoniewicz-Papis, K. Hałaburda,
M. Łe˛towska & N. Dunbar
e25 International Forum on Transfusion Practices in Haematopoietic
Stem-Cell Transplantation: Responses P. Solves, M. Lozano,
E. Zhiburt, J. Anguita Velasco, A. Maria Pérez-Corral,
S. Monsalvo-Saornil, S. Yamazaki, H. Okazaki, K. Selleng,
K. Aurich, W. Krüger, A. Buser, A. Holbro, L. Infanti, G. Stehle,
L. Pierelli, A. Matteocci, L. Rigacci, K. M. K. De Vooght,
J. H. E. Kuball, K. L. Fielding, D. A. Westerman, E. M. Wood,
C. S. Cohn, A. Johnson, M. B. C. Koh, D. Qadir,
C. Cserti-Gazdewich, E. Daguindau, F. Angelot-Delettre,
P. Tiberghien, S. Wendel-Neto, R.-M. Fachini, S. Morton,
C. Craddock, M. Lumley, J. Antoniewicz-Papis, K. Hałaburda,
M. Łe˛towska & N. Dunbar
613 Diary of Events
t
ea
n lin
a lo vo
x
ou
rn
n al/
i sj jo ur
th m/
w
co
Vie y.
r ar
lib
li ne
on
ley
wi
 Vox Sanguinis (2021) 116, 477–488
© 2020 International Society of Blood Transfusion
REVIEW ARTICLE DOI: 10.1111/vox.13058

The immune potential of ex vivo stored platelets: a review

Ben Wood,1,2 Matthew P. Padula,2 Denese C. Marks1,3 & Lacey Johnson1
1
Research & Development, Australian Red Cross Lifeblood, Alexandria, NSW, Australia
2
School of Life Sciences, University of Technology Sydney, Sydney, NSW, Australia
3
Sydney Medical School, The University of Sydney, Camperdown, NSW, Australia

Received: 18 September 2020,
revised 14 November 2020,
accepted 2 December 2020,
published online 16 December 2020
Abstract
Platelets are now acknowledged as key regulators of the immune system, as they
are capable of mediating inflammation, leucocyte recruitment and activation.
This activity is facilitated through platelet activation, which induces significant
changes in the surface receptor profile and triggers the release of a range of sol-
uble biological response modifiers (BRMs). In the field of transfusion medicine,
the immune function of platelets has gained considerable attention as this may
be linked to the development of adverse transfusion reactions. Further, compo-
nent manufacturing and storage methodologies may impact the immunoregula-
tory role of platelets, and an understanding of this impact is crucial and should
be considered alongside their haemostatic characteristics. This review highlights
the key interactions between platelets and traditional immune modulators. Fur-
ther, the potential impact of current and novel component storage methodolo-
gies, such as refrigeration and cryopreservation, on this functional capacity is
examined, highlighting why further knowledge in this area would be of benefit.
Key words: platelet, storage, leucocyte, function, refrigeration, cryopreservation.

renewed interest in refrigeration and cryopreservation

Introduction
Haemostasis has long been considered the primary role of
platelets. However, contemporary research has highlighted
that platelet function is more complex, as they also play
a significant role in regulating the immune system [1].
Platelets, as immune cells, contribute significantly to
inflammation, leucocyte recruitment and activation [2–4]
(Figure 1). Under normal conditions, the immune function
of platelets is biologically beneficial; however, hyperac-
tivity is linked to a number of pro-inflammatory disorders
[2]. In transfusion medicine, the immune function of pla-
telets has become topical due to its association with
adverse transfusion outcomes [5–8].
Timely transfusion of platelet components decreases the
mortality and morbidity of patients suffering from acute
haemorrhage [9]. Unfortunately, platelet components may
be unavailable in remote or rural medical settings due to
their short shelf-life [10]. However, platelet manufacturing
and storage methodologies are constantly evolving, and a
Correspondence: Lacey Johnson, Research and Development, Australian
Red Cross Lifeblood, 17 O’Riordan St, Alexandria, NSW, Australia, 2015
E-mail: ljohnson@redcrossblood.org.au
could allow an extension of the shelf-life to weeks or
years, respectively [11–13]. Pre-implementation evaluation
of new production methods has primarily focused on
characterizing the haemostatic characteristics of platelets
to ensure that this functional capacity is maintained.
However, considering what is now understood about the
immune function of platelets, a broader understanding of
how alternative storage modes may affect the immune
regulatory aspects of platelet function may also provide
clinically relevant information.
This review outlines the ability of platelets to facilitate
leucocyte activation and inflammation, which is only one
of their immunoregulatory roles. This article also high-
lights the link between platelet immune characteristics
and adverse events and examines the impact that novel
ex vivo storage modes may have on this functionality.
The role of platelets in mediating leucocyte
function
Similar to the initiation of haemostasis, the immune func-
tion of platelets in vivo is mediated through activation.
Platelet activation can be induced through receptor

477
478 B. Wood et al.

Fig. 1 A simplified schematic depicting platelet–leucocyte interactions, which are mediated by the expression of surface receptors and release of soluble
factors. Platelet–leucocyte interaction can be triggered by toll-like receptor (TLR) recognition of 1. pathogen-associated molecular patterns (PAMPs e.g.
bacterial lipids) or 2. damage-associated molecular patterns (DAMPs e.g. HMGB1). This triggers platelet activation and release of 3. a-granules (a), 4.
dense granules (d) and 5. extracellular vesicles (EVs; brown lines). Degranulation of a-granules increases the surface expression of activation markers,
including CD40L (blue rectangle) and P-selectin (green rectangle) and 6. the release of a range of biological response modifiers (BRMs; e.g. HMGB1, IL-
27, RANTES, PF4; black lines) into the plasma which can influence surrounding leucocytes (purple receptors) and have been linked to inflammation and
adverse events. 7. Activated platelets can attach to neutrophils through P-selectin, CD40L and GPIba to PSGL-1, CD40 and Mac-1, respectively. Attach-
ment can also occur through platelet GPIIb/IIIa and leucocyte Mac-1 through a fibrinogen intermediate. This interaction generally increases the activa-
tion of both platelets and leucocytes. 8. The activated platelet can then release internal agonist stores (ADP, calcium) which lead the activation of
platelets in the surrounding area. 9. EVs can activate leucocytes either through transport of their internal contents (e.g. mitochondria, mtDNA, micro-
RNAs, proliferation factors and cytokines) or by receptor signalling. The activation of neutrophils can trigger 10. the generation neutrophil extracellular
traps (NETs) which are associated with the pathogenesis of TRALI. [Colour figure can be viewed at wileyonlinelibrary.com]

recognition of a range of signalling factors, including
pathogen-associated molecular patterns (PAMPs), damage-
associated molecular patterns (DAMPs) and traditional
platelet agonists (Fig. 1) [2,14,15]. The signals are then
transduced across the cell membrane and propagated in
the cytosol by downstream signalling, resulting in signifi-
cant changes to internal cytoskeletal structure and leading
to external morphological changes [2,14]. Cytoskeletal
rearrangement typically results in the release of alpha-
granules (a-granules) and dense granules, dispersing a
range of soluble factors into the circulation (Fig. 1) [16].
Extracellular vesicles (EVs) are also released, which act as
mobile signalling mediators (Fig. 1) [17–19]. In vivo these
changes increase localized inflammation and encourage

© 2020 International Society of Blood Transfusion

Vox Sanguinis (2021) 116, 477–488

the recruitment and activation of leucocytes, escalating
the immune response [20].
Immunoregulatory surface receptors
It has been established that platelets express a variety of
surface molecules that enable them to moderate early
immune responses to both infection and sterile/self-injury.
Platelets express most of the known Toll-like receptors
(TLRs) on their surface membrane (TLR1, 2, 4, 5, 6), or
internally within endosomes or platelet specific T-granules
(TLR3, 7, 8, 9) [2,21,22]. In general, platelet TLRs recog-
nize and bind highly conserved PAMPs, including lipids,
lipoproteins, proteins or nucleic acids derived from bacte-
ria, viruses, fungi or parasites, to initiate anti-microbial
signalling [1,2,15,21,22]. TLRs can also bind a range of
‘self’-derived DAMP molecules that can be secreted or
released from damaged / apoptotic cells or accumulate in
platelet components during storage, which induces pro-
inflammatory signals (Fig. 1; Tables 1 & 2) [2,8,18]. Plate-
let TLR-mediated responses are facilitated through their
interactions with leucocytes [8,15,23]. The binding of
PAMPs and DAMPs to TLRs initiates distinctive signal
transduction responses, which induce platelet activation,
the release of biological response modifiers (BRMs) and
immunomodulatory EVs leading to leucocyte activation
and recruitment [8,15,24].
Release of biological response modifiers
Platelets possess a large internal store of soluble factors
within a- and dense granules [16,25], which are released
following platelet activation (Fig. 1). The platelet releasate
contains a range of BRMs, including adhesion molecules,
cytokines, chemokines and growth factors [16,25,26]
(Tables 1 & 2), which are able to modulate many aspects
of the innate immune response following transfusion
[7,27,28]. Specifically, granule release promotes leucocyte
activation (sCD40L, RANTES, PF4, NAP2), proliferation
(PF4, NAP2), adhesion (RANTES, IL-1b) and inflammation
(IL-1b, RANTES) [23,29,30], as well as amplifying the
activation of the original platelet and any surrounding
platelets (ADP, calcium; Fig. 1).
Platelet extracellular vesicles as immune
regulators
Platelets produce up to 90% of the EVs present in the cir-
culation of healthy subjects [31,32]. Initially, interest in
platelet-derived EVs was due to their haemostatic proper-
ties [33]. However, it is now apparent that platelet EVs are
also key mediators of immune signalling [14,24,31,32,34].
While low numbers of EVs are released from resting
© 2020 International Society of Blood Transfusion
Vox Sanguinis (2021) 116, 477–488
The immune potential of stored platelets 479
platelets, a significant increase is observed following acti-
vation [19,24]. Platelet EVs express a range of surface
receptors including P-selectin and CD40L, which facilitate
attachment and activation of monocytes through Mac-1
and CD40, respectively [35]. Pro-coagulant platelets also
release EVs rich in surface exposed phosphatidylserine,
which attach to monocytes and neutrophils in vivo,
although the specific mechanics of this interaction and its
immunomodulatory effects are still under evaluation
[24,34]. Platelet EVs can also transport and deliver a range
of platelet-derived immunomodulatory compounds to leu-
cocytes including receptors (CD40L), mitochondria and
mitochondrial DNA (mtDNA), microRNAs, cytokines (IL-
1b, MIP-3), proliferation factors (PF4, CXCL-7) and
chemokines (RANTES; Fig. 1) [14,18,31,34,36]. As a result,
platelet EVs exhibit wide-ranging effects and can influ-
ence the activation of most leucocyte populations,
although interactions with neutrophils and monocytes are
more common [14,18,24,34,36]. EV-activated monocytes
exhibit increased expression of several adhesion receptors
(ICAM-1, LFA-1 and Mac-1), increasing their capacity to
bind to endothelial cells [35,37,38]. Additionally, EV-
monocyte interaction causes the release of a range of
monocyte derived pro-inflammatory cytokines (IL-8, IL-
1b, TNF-a, MCP-1 and MMP-9) [35,37,38].
Platelet–leucocyte interactions
The initial response of platelets to tissue damage is the
initiation of haemostasis via adhesion to sub-endothelial
structures and subsequent activation of the coagulation
cascade; however, they also play a key role in recruitment
and activation of leucocytes in order to contain and elim-
inate any potential infiltrating pathogens [3]. Attachment
to leucocytes requires platelet activation and degranula-
tion, which can be induced by stimulatory signals,
including haemostatic agonists (thrombin, collagen and
calcium), PAMPs/DAMPs or ex vivo storage [4,8,15,39,40].
Specifically, platelet activation results in a-granule
release, and the consequent localization of P-selectin,
CD40L and phosphatidylserine on the platelet surface,
which facilitates leucocyte attachment (Figure 1 &
Table 1) [3,41]. P-selectin is often the first point of con-
tact between platelets and leucocytes through binding to
its ligand, PSGL-1 [3,41]. While P-selectin/PSGL-1 bind-
ing alone is relatively weak, co-stimulation with CD40L
induces downstream signalling in leucocytes [3,30,41].
This promotes Mac-1 expression on myeloid cells,
enabling more permanent attachment directly to platelet
GPIba (Fig. 1) [3,30] or through fibrinogen bound to pla-
telet GPIIb/IIIa [42]. Activated platelets also exhibit exter-
nalized phosphatidylserine on the surface membrane,
which has been shown to facilitate platelet attachment to
 Table 1 Summary of the effect of storage methodologies on the surface expression of immune-related platelet proteins
480 B. Wood et al.

Platelet Surface Expression/concentration change over storage

Receptors Role in platelet-mediated immunity References Conventional Storagea Refrigerationb Cryopreservationc References

TLR 1, 2, 4, 5 & 6 • recognizes PAMPs (e.g. proteins and lipids) and [1,2,21,22] na na na na
DAMPs (e.g. HMBG1 and Histone H4)
• triggers platelet activation
• can lead to pro-inflammatory effects
• expressed on the surface membrane
TLR 3, 7 & 9 • recognizes nucleic acids and HMGB1 (TLR9) [1,2,21,22] na na na na
• triggers platelet activation
• expressed internally and on the surface membrane
of activated platelets
Receptors/factors • Role in leucocyte interaction
GPIIb/IIIa (Active • facilitates indirect attachment to leucocytes through [1,42] ↔ or ↓ ↑ ↔ or ↓ [40,48,59–61,70,74,80]
conformation) fibrinogen bound to Mac-1
GPIba • facilitates attachment to leucocytes through Mac-1 [1] ↓ ↓ ↓ [39,58–60,68,70,74,76,77]

P-selectin • facilitates attachment to leucocytes through PSGL-1 [1,3] ↑ ↔ or ↑ ↑ [39,40,48,49,58–60,68,70,74,76]

Phosphatidylserine • anti-inflammatory and immunosuppressive signal [39,75] ↑ ↔ or ↑ ↑ [39,40,48,58–60,68,70,73,74,76,77,80]

• mediates phagocytosis by macrophages
CD40L (CD154) • co-stimulatory molecule [3] ↑ ↑ na [58]
• pro-inflammatory

Changes over storage are described as general trends and vary depending on the method of platelet manufacture, including collection method and plasma content.
Relative changes in factor expression/concentration are expressed as the following: increase = ↑; decrease = ↓; unchanged = ↔; na = no published data available.
DAMPs, damage-associated molecular patterns; PAMPs, pathogen-associated molecular patterns; TLR, toll-like receptor.
a
Conventionally stored platelets (20–24°C, with agitation) at day 5–7 compared to day 1.
b
Refrigerated platelets compared to conventionally stored platelets at same time point of storage.
c
Thawed platelets compared to conventionally stored platelets at day 1.

Vox Sanguinis (2021) 116, 477–488

© 2020 International Society of Blood Transfusion
 Table 2 Summary of the effect of storage methodologies on the release of immune-related soluble factors

Platelet Supernatant Expression/concentration change over storage

Biological Response
b c
Modifier Role in platelet-mediated immunity References Conventional Storagea Refrigeration Cryopreservation References

PF4 [25,81] ↑ ↓ ↔ or ↑ [26,49,68,76]

• promotes haemostasis
• attracts neutrophils and monocytes
TGF-b1 [1,20,81] ↑ ↔ ↑ [49,54,55,68,70]

Vox Sanguinis (2021) 116, 477–488

• context sensitive signal protein
• pro / anti-inflammatory properties
IL-1b [1,3,14,17,20,81] ↑ na ↓ [56,68]
• key mediator or the inflammatory

© 2020 International Society of Blood Transfusion

response
IL-8 [25] ↑ ↓ na [54]
• attracts neutrophils and granulocytes
IL-27 [1,7] ↑ na ↔ [7,77]
• context sensitive signal protein
• pro/anti-inflammatory properties
HMGB1 [1,3,27] na na na na
• pro-inflammatory signalling molecule
• promotes cytokine release through TLRs
RANTES [20,82] ↑ ↓ ↔ or ↑ [26,49,51,56,68,73,76]
• pro-inflammatory
• attracts leucocytes
Soluble CD62P [1,41] ↑ ↓ ↑ [26,49,68,70,76,77]
• promotes leucocyte signalling and
recruitment
Soluble OX40L [7] ↑ na na [7,53]
• promotes endothelial inflammation
and leucocyte activation
Soluble CD40L [1,3,7,20] ↑ ↑ or ↓ ↓ or ↑ [26,28,36,56,58,68,70,76]
• pro-inflammatory
• platelet and leucocyte activator
• facilitates leucocyte differentiation
The immune potential of stored platelets 481
 482 B. Wood et al.

Table 2 (Continued)

Platelet Supernatant Expression/concentration change over storage

Biological Response
b c
Modifier Role in platelet-mediated immunity References Conventional Storagea Refrigeration Cryopreservation References

• co-stimulatory molecule
EVs [1,3,17,19,31,47,72] ↑ ↔ or ↑ ↑ [26,52,59,68,72,73,80]
• remote signalling factors
• facilitate intracellular transfer of receptors,
cytokines and miRNAs
• activate leucocytes
Mitochondrial DNA [53] ↑ na na [53]
• pro-inflammatory signal
• recognized by TLRs

Relative changes in factor expression/concentration are expressed as the following: increase = ↑; decrease = ↓; unchanged = ↔; na = no published data available.
Changes over storage are described as general trends and vary depending on the method of platelet manufacture, including collection method and plasma content.
EV, extracellular vesicles; TLR, toll-like receptors.
a
Conventionally stored platelets (20–24°C, with agitation) at day 5–7 compared to day 1.
b
Refrigerated platelets compared to conventionally stored platelets at same time point of storage.
c
Thawed platelets compared to conventionally stored platelets at day 1.

Vox Sanguinis (2021) 116, 477–488

© 2020 International Society of Blood Transfusion

macrophages and neutrophils through a range of poten-
tial surface receptors [39,43]. Notably, the formation of
platelet–leucocyte aggregates results in the activation and
degranulation of both cell types, increasing the local
haemostatic and inflammatory effects [3].
While platelet–leucocyte aggregates can be beneficial
to the immune response, an overabundance is associated
with pro-inflammatory disorders and adverse transfusion
events such as TRALI [2,3,5,44]. Activated platelets are
able to attach to and circulate with all major leucocyte
subgroups including neutrophils, monocytes, T cells and
B cells [44]. However, association with monocytes and
neutrophils is much more common [30,44]. Notably,
polarized neutrophils, which display asymmetric receptor
expression, can actively search for activated platelets in
the bloodstream and attach through P-selectin/PSGL-1
[45]. The formation of platelet-neutrophil aggregates is
pro-inflammatory and a key step towards the generation
of platelet-induced neutrophil extracellular traps (NETs)
[4,5,45,46]. However, platelet–neutrophil binding alone is
usually insufficient to induce NET formation, which
requires additional secondary signals such as pathogen
recognition via platelet TLR4, signalling from traditional
platelet agonists (thrombin), soluble factors (sCD40L,
IL-1b, HMGB1) or EVs [4,5,8,29,32]. Following receipt of
secondary signals, the neutrophil ejects its nucleus,
releasing a web of DNA coated in a range of anti-
microbial factors (Fig. 1) [4]. The induction of NET for-
mation is tightly controlled due to its inflammatory and
pro-coagulant properties, which can cause significant
collateral tissue damage, particularly to the sensitive
alveoli of the lung [4,5].
As platelet activation is key in mediating platelet
immune function, it is important to note that a progres-
sive degree of platelet activation occurs during ex vivo
storage [40,47,48]. While the haemostatic impact of stor-
age has been well characterized, the storage-related
effects on the immunological functions of platelets are
still being unravelled.
The effect of ex vivo storage on platelet
immune function
Conventional room temperature stored platelets
While specific manufacturing methods vary between
countries, platelet components are produced either by
pooling and isolating platelets from the WB of multiple
donors or by selectively isolating platelets from a single
donor by apheresis [10]. Additionally, the solution in
which the platelets are suspended can vary, consisting of
100% donor plasma or a combination of plasma and up
to 70% platelet additive solution (PAS), which influences
© 2020 International Society of Blood Transfusion
Vox Sanguinis (2021) 116, 477–488
The immune potential of stored platelets 483
the initial composition and concentration of BRMs in the
component [49].
Platelet concentrates are currently stored for 5–7 days
at room temperature (20–24°C), which leads to deleterious
storage-related effects, known as the platelet storage
lesion (PSL) [40,47,50]. Progression of the PSL impacts
the platelet phenotype and function, the composition of
the supernatant of stored components (Tables 1 and 2;
conventional storage), and also affects the immunogenic-
ity of the component, with older platelet units more likely
to induce adverse transfusion events [6].
Room temperature storage induces a gradual shift
towards a more activated phenotype. This is observed by
a loss of certain glycoproteins (GPIba, GPIV, GPVI),
increased surface expression of P-selectin, phos-
phatidylserine and CD40L (Table 1) [40,47,51]. These are
all key mediators of platelet–leucocyte signalling (Fig. 1)
[3,30]. There is also an accumulation of metabolic by-
products, pro-inflammatory cytokines, mitochondrial
DNA (mtDNA) and EVs in the supernatant (Table 2)
[7,49,51–56]. The initial concentration of many soluble
factors is reduced by supplementation with PAS. How-
ever, current data indicate that ex vivo storage can stimu-
late the release of BRMs from platelets several fold higher
than baseline, regardless of the starting concentration in
plasma or plasma / PAS (Table 2) [7,26–28,36,54].
Conventional storage induces changes in platelet
immune function over time, which is directly linked to
the incidence of adverse events. A retrospective study by
Losos et al. examined more than 50 000 leucoreduced
platelet transfusions and identified that the majority of
transfusion reactions occurred after 3 days of storage,
with the likelihood of adverse events increasing per day
(OR, 1
30), 95% CI (1
12, 1
52) [6]. Consequently, the
benefit of extending conventional platelet storage time
must be weighed against the accumulation of BRMs in
components and thus the likelihood of inducing adverse
events. The induction of adverse events has been associ-
ated with platelet activation, high concentrations of
HMGB1, soluble CD40L, OX40L, IL-27, IL-1b, mtDNA and
mitochondria-carrying EVs in platelet components (Fig. 1
& Table 2) [7,18,27,29,53]. Further, DAMPs, such as
HMGB1 and mtDNA, may accumulate in platelet compo-
nents and have been implicated in the induction of
adverse events, including febrile non-haemolytic transfu-
sion reactions (FNHTRs) and transfusion-related acute
lung injury (TRALI) [2,5,8,18,27,53]. Although the mecha-
nisms of TRALI are multi-factorial, transfusion of plate-
lets is thought to trigger leucocyte activation and
increase lung inflammation and NET formation, poten-
tially via the action of activated platelets, TLR signalling,
EVs and BRMs (Fig. 1) [5,8,18,29,45]. Further, platelet
depletion and inhibition by aspirin have been found to
484 B. Wood et al.
have a protective effect against acute lung injury in mice
[45,46].
Refrigerated platelets
There has been a resurgence of interest in platelet refrig-
eration (2–6°C) as a storage methodology since its discon-
tinuation in the 1970s, as refrigeration inhibits bacterial
growth and preserves haemostatic function [12,50,57].
Platelet refrigeration results in significant morphological
change and more pronounced expression of activation
markers in comparison to room temperature stored compo-
nents [58–61]. However, alterations to the immune charac-
teristics of platelets induced by refrigeration are largely
uncharacterized (Tables 1 & 2). Changes in the abundance
of adhesion receptors have been observed during refrigera-
tion, including decreased GPIba, increased CD40L and P-
selectin and phosphatidylserine externalization [59–61]
(Table 1), which could increase the likelihood of platelet
leucocyte–aggregate formation. Refrigeration appears to
slow the rate of release of a-granule-associated soluble fac-
tors, compared to conventionally stored platelets (Table 2)
[26,54]. A study by Johnson et al. demonstrated that refrig-
erated platelets stored for 14 days contained cytokine
levels comparable to day 5 room temperature stored plate-
lets [26]. The release of phosphatidylserine-positive EVs is
also significantly increased during refrigerated storage
compared to conventional storage [52], and while they
have been demonstrated to be haemostatic [26], the
immunologic effects remain unknown.
Research examining the immune effect of refrigerated
platelets in vivo is limited to a single study in a mouse
model, whereby refrigerated platelets increased vascular
leakage compared to conventionally stored components
[62], which could facilitate increased leucocyte infiltration
through endothelial cells. Preliminary safety data suggest
that cold-stored platelets are not associated with more
adverse events than conventionally stored platelets, at
least in patients undergoing cardiac surgery [13]. How-
ever, further studies are warranted, particularly in patient
cohorts where altered platelet-mediated immunological
function may already be a risk factor for adverse events,
such as in trauma [24].
Cryopreserved platelets
Platelet cryopreservation, pioneered by Valeri et al. [63],
was used throughout the 1970s to supply autologous pla-
telets to alloimmunized leukaemia patients [64]. In recent
years, an emphasis has been placed on using cryopre-
served platelets for the treatment of acute bleeding due to
their increased haemostatic effectiveness [11,65,66]. Plate-
let cryopreservation involves adding 5–6% DMSO as a
cryoprotectant, before concentrating the platelets and
removing excess DMSO prior to freezing at -80°C
[63,67,68]. This method allows for long-term storage for
potentially 2–4 years [69]. When required, cryopreserved
platelets can be thawed and reconstituted in saline,
thawed plasma or additive solution [67,70,71].
The cryopreservation process alters the surface phe-
notype and increases degranulation and release of
phosphatidylserine-positive EVs (Table 2) [26,72]. In
terms of membrane changes, 50–70% of cryopre-
served platelets externalize phosphatidylserine (Table 1)
[39,73,74]. There is some evidence that phosphatidylserine
on cryopreserved platelets may be pro-inflammatory,
mediating interactions with macrophages [39,75]. Simi-
larly, GPIba expression is reduced on up to 50% of cry-
opreserved platelets (Table 1) [39,68,74,76], which may
affect their ability to bind to leucocytes. The concentra-
tion of soluble factors in the supernatant of cryopreserved
components is significantly higher than conventionally
stored platelet components (Table 2), including RANTES
and TGF-b1, which have pro- and anti-inflammatory
effects, respectively [26,76,77]. Further, the EV concentra-
tion of cryopreserved platelet components is up to 15-
fold higher than conventionally stored platelets, and
10-fold higher than refrigerated platelets at the end of
storage, which contribute to their haemostatic function
[26,73]. Additionally, EVs from cryopreserved platelets
have been shown to possess an altered surface receptor
profile compared to EVs from fresh platelet components,
demonstrating increased expression of several adhesion
receptors (GPIIb, GPIX and integrin associated protein;
IAP) and exposure of phosphatidylserine [72]. Thus, cry-
opreserved platelets have the potential to be more
immunogenic than conventional platelet units.
To date, relatively few studies have been conducted to
examine the immunogenicity of cryopreserved platelets
(Tables 1 & 2). In an in vitro human whole blood model
of transfusion, cryopreserved platelets suppressed the
responsiveness of BDCA3+ myeloid-derived dendritic cells
to stimulation by LPS and poly(I:C), producing less IL-8,
TNF-a and IP-10, suggesting a potentially immunosup-
pressive effect [78]. In contrast, studies using THP-1 cells
and a murine model of controlled haemorrhage suggest
that cryopreserved platelets may be pro-inflammatory, by
increasing leucocyte release of TNF-a, IL-1b and IL-6
[39,79]. While these contradictory findings are likely due
to differences in experimental models, they highlight the
need for further work in this area. Very few adverse
events have been reported following transfusion of cryop-
reserved platelets [11,64,65,71]. However, larger studies
are required, as the number of cryopreserved platelets
transfused is too low to provide confidence that they are
not associated with adverse events of low incidence.
© 2020 International Society of Blood Transfusion
Vox Sanguinis (2021) 116, 477–488
 The immune potential of stored platelets 485

Future areas of study immune cell. Ex vivo storage conditions have the capacity
to significantly alter the characteristics of platelet compo-
While the immune function of platelets is now acknowl-
nents, which may affect their immune function and the
edged, and studies have begun to consider the effect of
likelihood of adverse events. Consequently, the impact of
conventional and alternative storage methodologies, there
new manufacturing and storage methodologies on the
are significant gaps in our current understanding. Specifi-
immune characteristics of platelets should be considered
cally, the concentration of key BRMs that have been associ-
alongside their haemostatic function prior to implementa-
ated with adverse events (IL-8, IL-27, sCD40L and mtDNA)
tion.
and the expression of immune-related receptors in platelets
stored under novel conditions are largely uncharacterized,
particularly those involved in PAMP / DAMP recognition Acknowledgements
and leucocyte interaction (Table 1) [7,53]. As current data
Australian governments fund Australian Red Cross Life-
demonstrate differences in phenotype and releasate
blood to provide blood, blood products and services to
between platelets stored under different conditions [52,54],
the Australian community. The authors thank Chris Roan
there is the potential that this information could be used to
for his careful and critical reading of the manuscript.
better tailor transfusions to suit the specific condition and
risk factors of the recipients. Additionally, very few studies
have examined how stored platelets interact with recipient Conflicts of interest
leucocytes, which is believed to be a key factor in the inci-
The authors have no conflicts of interest to declare.
dence of adverse events [5,45]. This could be achieved by
the use of animal and in vitro models of transfusion.
Ultimately, clinical trials, which are in progress for both Author contributions
cold-stored and cryopreserved platelets, offer the perfect
All authors contributed to development of the concepts
opportunity to assess the immunological changes occurring
and design of the review article and critically reviewed
in the recipient post-transfusion and should be incorpo-
the manuscript. BW and LJ wrote the manuscript. BW
rated into the design of these trials. This knowledge would
prepared the figure.
aid in characterizing the pathogenesis of adverse events,
support the development of preventative strategies and
inform future changes to transfusion guidelines. Funding
Australian governments fund Australian Red Cross Life-
blood to provide blood, blood products and services to
Concluding remarks
the Australian community. BW is supported by a
Platelets are no longer recognized as simply mediators of Research Excellence Scholarship provided by the Univer-
haemostasis, but also as having a tangible role as an sity of Technology Sydney.

References
1 Stolla M, Refaai MA, Heal JM, et al.:
Platelet transfusion - the new
immunology of an old therapy. Front
Immunol 2015; 6:1–10
2 Cognasse F, Nguyen K, Pauline D,
et al.: The Inflammatory Role of Plate-
lets via Their TLRs and Siglec Recep-
tors. Front Immunol 2015; 6:1–15
3 Kral JB, Schrottmaier WC, Salzmann
M, et al.: Platelet Interaction with
Innate Immune Cells. Transfus Med
Hemother 2016; 43:78–88
4 Clark SR, Ma AC, Tavener SA, et al.:
Platelet TLR4 activates neutrophil extra-
cellular traps to ensnare bacteria in sep-
tic blood. Nat Med 2007; 13:463–9
5 Caudrillier A, Kessenbrock K, Gilliss
BM, et al.: Platelets induce neutrophil
extracellular traps in transfusion-
related acute lung injury. J Clin Invest
2012; 122:2661–71
6 Losos M, Biller E, Li J, et al.: Pro-
longed platelet storage associated with
increased frequency of transfusion-
related adverse events. Vox Sang 2018;
113:170–6
7 Hamzeh-Cognasse H, Damien P,
Nguyen KA, et al.: Immune-reactive
soluble OX40 ligand, soluble CD40
ligand, and interleukin-27 are simulta-
neously oversecreted in platelet com-
ponents associated with acute
transfusion reactions. Transfusion
2014; 54:613–25
8 Land WG: Transfusion-related acute
lung injury: the work of DAMPs.
Transfus Med Hemother 2013; 40:3–
13
9 Cap AP, Spinella PC, Borgman MA,
et al.: Timing and location of blood
product transfusion and outcomes in
massively transfused combat casu-
alties. J Trauma Acute Care Surg
2012; 73:S89–94
10 Keitel S: Guide to the preparation, use
and quality assurance of blood compo-
nents, 19th edn. Strasbourg, Council
of Europe, 2017

© 2020 International Society of Blood Transfusion

Vox Sanguinis (2021) 116, 477–488
486 B. Wood et al.
11 Reade MC, Marks DC, Bellomo R,
et al.: A randomized, controlled pilot
clinical trial of cryopreserved platelets
for perioperative surgical bleeding: the
CLIP-I trial (Editorial, p. 2759). Trans-
fusion 2019; 59:2794-804.
12 Stubbs JR, Tran SA, Emery RL, et al.:
Cold platelets for trauma-associated
bleeding: regulatory approval, accredi-
tation approval, and practice imple-
mentation-just the “tip of the iceberg".
Transfusion 2017; 57:2836–44
13 Strandenes G, Sivertsen J, Bjerkvig
CK, et al.: A pilot trial of platelets
stored cold versus at room tempera-
ture for complex cardiothoracic sur-
gery. Anesthesiology 2020; 133:1173–
83.
14 Brown GT, McIntyre TM: Lipopolysac-
charide signaling without a nucleus:
kinase cascades stimulate platelet
shedding of proinflammatory IL-1b–
rich microparticles. J Immunol 2011;
186:5489–96
15 Hamzeh-Cognasse H, Berthelot P,
Tardy B, et al.: Platelet toll-like recep-
tors are crucial sensors of infectious
danger moieties. Platelets 2018;
29:533–40
16 Blair P, Flaumenhaft R: Platelet alpha-
granules: basic biology and clinical
correlates. Blood Rev 2009; 23:177–89
17 Lindemann S, Tolley ND, Dixon DA,
et al.: Activated platelets mediate
inflammatory signaling by regulated
interleukin 1b synthesis. J Cell Biol
2001; 154:485–90
18 Marcoux G, Magron A, Sut C, et al.:
Platelet-derived extracellular vesicles
convey mitochondrial DAMPs in pla-
telet concentrates and their levels are
associated with adverse reactions.
Transfusion 2019; 59:2403–14
19 Aatonen M, Gr€ onholm M: Siljander
PR-M: Platelet-derived microvesicles:
multitalented participants in intercel-
lular communication. Sem Thromb
Hemost 2012; 38:102–13
20 Ali RA, Wuescher LM, Worth RG: Pla-
telets: essential components of the
immune system. Curr Trends Immunol
2015; 16:65–78
21 Thon JN, Peters CG, Machlus KR,
et al.: T granules in human platelets
function in TLR9 organization and
signaling. J Cell Biol 2012; 198:561–
74
22 Cognasse F, Hamzeh H, Chavarin P,
et al.: Evidence of Toll-like receptor
molecules on human platelets. Immu-
nol Cell Biol 2005; 83:196–8
23 Rossaint J, Margraf A, Zarbock A:
Role of platelets in leukocyte recruit-
ment and resolution of inflammation.
Front Immunol 2018; 9:1–13
24 Vulliamy P, Gillespie S, Armstrong PC,
et al.: Histone H4 induces platelet bal-
looning and microparticle release dur-
ing trauma hemorrhage. PNAS 2019;
116:17444–9
25 Heijnen H, van der Sluijs P: Platelet
secretory behaviour: as diverse as the
granules . . . or not? J Thromb Hae-
most 2015; 13:2141–51
26 Johnson L, Tan S, Jenkins E, et al.:
Characterization of biologic response
modifiers in the supernatant of con-
ventional, refrigerated, and cryopre-
served platelets. Transfusion 2018;
58:927–37
27 Cognasse F, Sut C, Hamzeh-Cognasse
H, et al.: Platelet-derived HMGB1:
critical mediator of SARs related to
transfusion. Ann Transl Med 2019;
8:140–1
28 Cognasse F, Sut C, Fromont E, et al.:
Platelet soluble CD40-Ligand level is
associated with transfusion adverse
reactions in a mixed threshold and hit
model. Blood 2017; 130:1380–3
29 Khan SY, Kelher MR, Heal JM, et al.:
Soluble CD40 ligand accumulates in
stored blood components, primes neu-
trophils through CD40, and is a poten-
tial cofactor in the development of
transfusion-related acute lung injury.
Blood 2006; 108:2455–62
30 Jin R, Yu S, Song Z, et al.: Soluble
CD40 ligand stimulates CD40-depen-
dent activation of the b2 integrin
Mac-1 and protein kinase C zeda
(PKCf) in neutrophils: implications for
neutrophil-platelet interactions and
neutrophil oxidative burst. PLoS One
2013; 8:1–10
31 Baj-Krzyworzeka M, Majka M, Pratico
D, et al.: Platelet-derived microparti-
cles stimulate proliferation, survival,
adhesion, and chemotaxis of
hematopoietic cells. Exp Hematol
2002; 30:450–9
32 Maugeri N, Capobianco A, Rovere-
Querini P, et al.: Platelet microparti-
cles sustain autophagy-associated
© 2020 International Society of Blood Transfusion
activation of neutrophils in systemic
sclerosis. Sci Transl Med 2018; 10:
eaao3089
33 Van Der Meijden PEJ, Van Schilf-
gaarde M, Van Oerle R, et al.: Platelet-
and erythrocyte-derived microparticles
trigger thrombin generation via factor
XIIa. J Thromb Haemost 2012;
10:1355–62
34 Weiss R, Gr€ oger M, Rauscher S, et al.:
Differential interaction of platelet-
derived extracellular vesicles with
leukocyte subsets in human whole
blood. Sci Rep 2018; 8:6598
35 Bei JJ, Liu C, Peng S, et al.: Staphylo-
coccal SSL5-induced platelet
microparticles provoke proinflamma-
tory responses via the CD40/TRAF6/
NFjB signalling pathway in mono-
cytes. Thromb Haemost 2016;
115:632–45
36 Sprague DL, Elzey BD, Crist SA, et al.:
Platelet-mediated modulation of adap-
tive immunity: unique delivery of
CD154 signal by platelet-derived
membrane vesicles. Blood 2008;
111:5028–36
37 Nomura S, Tandon NN, Nakamura T,
et al.: High-shear-stress-induced acti-
vation of platelets and microparticles
enhances expression of cell adhesion
molecules in THP-1 and endothelial
cells. Atherosclerosis 2001; 158:277–
87
38 Barry OP, Pratico D, Lawson JA, et al.:
Transcellular activation of platelets
and endothelial cells by bioactive
lipids in platelet microparticles. J Clin
Invest 1997; 99:2118–27
39 Zhao J, Xu B, Chen G, et al.: Cryopre-
served platelets augment the inflam-
matory response: role of
phosphatidylserine- and P-selectin–
mediated platelet phagocytosis in
macrophages. Transfusion 2019;
59:1799–808
40 Shrivastava M: The platelet storage
lesion. Transfus Apher Sci 2009;
41:105–13
41 Rainger GE, Buckley C, Simmons DL,
et al.: Cross-talk between cell adhesion
molecules regulates the migration
velocity of neutrophils. Curr Biol
1997; 7:316–25
42 Patko Z, Csaszar A, Acsady G, et al.:
Roles of Mac-1 and glycoprotein IIb/
IIIa integrins in leukocyte–platelet
Vox Sanguinis (2021) 116, 477–488

aggregate formation: Stabilization by
Mac-1 and inhibition by GpIIb/IIIa
blockers. Platelets 2012; 23:368–75
43 Lemke G: How macrophages deal with
death. Nat Rev Immunol 2019;
19:539–49
44 Finsterbusch M, Schrottmaier WC,
Kral-Pointner JB, et al.: Measuring
and interpreting platelet-leukocyte
aggregates. Platelets 2018; 29:677–85
45 Sreeramkumar V, Adrover JM, Balles-
teros I, et al.: Neutrophils scan for
activated platelets to initiate inflam-
mation. Science 2014; 346:1234–8
46 Looney MR, Nguyen JX, Hu Y, et al.:
Platelet depletion and aspirin treat-
ment protect mice in a two-event
model of transfusion-related acute
lung injury. J Clin Invest 2009;
119:3450–61
47 Ng MSY, Tung J-P, Fraser JF: Platelet
storage lesions: what more do we
know now? Transfus Med Rev 2018;
32:144–54
48 Thon JN, Schubert P, Devine DV: Pla-
telet storage lesion: A new under-
standing from a proteomic
perspective. Transfus Med Rev 2008;
22:268–79
49 Cardigan R, Sutherland J, Wadhwa M,
et al.: The influence of platelet addi-
tive solutions on cytokine levels and
complement activation in platelet con-
centrates during storage. Vox Sang
2003; 84:28–35
50 Ketter PM, Kamucheka R, Arulanan-
dam B, et al.: Platelet enhancement of
bacterial growth during room temper-
ature storage: mitigation through
refrigeration. Transfusion 2019;
59:1479–89
51 Glenister KM, Payne KA, Sparrow RL:
Proteomic analysis of supernatant
from pooled buffy-coat platelet con-
centrates throughout 7-day storage.
Transfusion 2008; 48:99–107
52 Johnson L, Tan S, Wood B, et al.:
Refrigeration and cryopreservation of
platelets differentially affect platelet
metabolism: a comparison with con-
ventional platelet storage conditions.
Transfusion 2016; 56:1807–18
53 Cognasse F, Aloui C, Anh Nguyen K,
et al.: Platelet components associated
with adverse reactions: predictive
value of mitochondrial DNA relative
© 2020 International Society of Blood Transfusion
Vox Sanguinis (2021) 116, 477–488
The immune potential of stored platelets 487
to biological response modifiers.
Transfusion 2016; 56:497–504
54 Ferrer F, Rivera J, Lozano ML, et al.:
Effect of cold-storage in the accumu-
lation of bioreactive substances in pla-
telet concentrates treated with second
messenger effects. Haematologica
2001; 86:530–6
55 Kanter J, Khan SY, Kelher M, et al.:
Oncogenic and angiogenic growth fac-
tors accumulate during routine storage
of apheresis platelet concentrates. Clin
Can Res 2008; 14:3942–7
56 Sandgren P, Berlin G, Tynng ard N:
Treatment of platelet concentrates
with ultraviolet C light for pathogen
reduction increases cytokine accumu-
lation. Transfusion 2016; 56:1377–83
57 Murphy S, Gardner FH: Platelet
preservation: effect of storage temper-
ature on maintenance of platelet via-
bility—deleterious effect of
refrigerated storage. N Engl J Med
1969; 280:1094–8
58 Reddoch KM, Pidcoke HF, Mont-
gomery RK, et al.: Hemostatic function
of apheresis platelets stored at 4°C and
22°C. Shock 2014; 41:54–61
59 Getz TM, Montgomery RK, Bynum JA,
et al.: Storage of platelets at 4°C in
platelet additive solutions prevents
aggregate formation and preserves
platelet functional responses. Transfu-
sion 2016; 56:1320–8
60 Wood B, Padula MP, Marks DC, et al.:
Refrigerated storage of platelets initi-
ates changes in platelet surface marker
expression and localization of intra-
cellular proteins. Transfusion 2016;
56:2548–59
61 Johnson L, Cameron M, Waters L, et al.:
The impact of refrigerated storage of
UVC pathogen inactivated platelet con-
centrates on in vitro platelet quality
parameters. Vox Sang 2019; 114:47–56
62 Baimukanova G, Miyazawa B, Potter
D, et al.: The effects of 22°C and 4°C
storage of platelets on vascular
endothelial integrity and function.
Transfusion 2016; 56:S52–S64
63 Valeri CR: Hemostatic effectiveness of
liquid-preserved and previously frozen
human platelets. N Engl J Med 1974;
290:353–8
64 Slichter SJ, Jones M, Ransom J, et al.:
Review of in vivo studies of dimethyl
sulfoxide cryopreserved platelets.
Transfus Med Rev 2014; 28:212–25
65 Slichter SJ, Dumont LJ, Cancelas JA,
et al.: Safety and efficacy of cryopre-
served platelets in bleeding patients
with thrombocytopenia. Transfusion
2018; 58:2129–38
66 Bohonek M, Kutac D, Landova L,
et al.: The use of cryopreserved plate-
lets in the treatment of polytraumatic
patients and patients with massive
bleeding. Transfusion 2019; 59:1474–
8
67 Valeri CR, Ragno G, Khuri S: Freezing
human platelets with 6 percent
dimethyl sulfoxide with removal of
the supernatant solution before freez-
ing and storage at -80 degrees C with-
out postthaw processing. Transfusion
2005; 45:1890–8
68 Johnson LN, Winter KM, Reid S, et al.:
Cryopreservation of buffy-coat-derived
platelet concentrates in dimethyl sul-
foxide and platelet additive solution.
Cryobiology 2011; 62:100–6
69 Noorman F, Strelitksi R, Badloe J: Fro-
zen platelets can be stored for 4 years
at -80°C without affecting in vitro
recovery, morphology, receptor
expression, or coagulation profile.
Transfusion 2014; 54:74A
70 Johnson L, Reid S, Tan S, et al.: PAS-
G supports platelet reconstitution after
cryopreservation in the absence of
plasma. Transfusion 2013; 53:2268–77
71 Dumont LJ, Cancelas JA, Dumont DF,
et al.: A randomized controlled trial
evaluating recovery and survival of
6% dimethyl sulfoxide–frozen autolo-
gous platelets in healthy volunteers.
Transfusion 2013; 53:128–37
72 Raynel S, Padula MM, Marks D, et al.:
Cryopreservation alters the membrane
and cytoskeletal protein profile of pla-
telet microparticles. Transfusion 2015;
55:2422–32
73 Johnson L, Coorey CP, Marks DC: The
hemostatic activity of cryopreserved
platelets is mediated by phos-
phatidylserine-expressing platelets and
platelet microparticles. Transfusion
2014; 54:1917–26
74 Six KR, Delabie W, Devreese KM,
et al.: Comparison between manufac-
turing sites shows differential adhe-
sion, activation, and GPIba expression
488 B. Wood et al.
of cryopreserved platelets. Transfusion
2018; 58:2645–56
75 Birge RB, Boeltz S, Kumar S, et al.:
Phosphatidylserine is a global
immunosuppressive signal in efferocy-
tosis, infectious disease, and cancer.
Cell Death Differ 2016; 23:962–78
76 Crimmins D, Flanagan P, Charlewood
R, et al.: In vitro comparison between
gamma-irradiated cryopreserved and
Day 7 liquid-stored buffy coat–derived
platelet components. Transfusion
2016; 56:2799–807
77 Johnson L, Reade M, Hyland R, et al.:
In vitro comparison of cryopreserved
and liquid platelets: potential clinical
implications. Transfusion 2014;
55:838–47
78 Ki KK, Johnson L, Faddy HM, et al.:
Immunomodulatory effect of cryopre-
served platelets: altered BDCA3+ den-
dritic cell maturation and activation
in vitro. Transfusion 2017; 57:2878–
87
79 Zhao J, Sun Z, You G, et al.: Transfu-
sion of cryopreserved platelets exacer-
bates inflammatory liver and lung
injury in a mice model of hemorrhage.
J Trauma Acute Care Surg 2018;
85:327–33
© 2020 International Society of Blood Transfusion
80 Stolla M, Bailey SL, Fang L, et al.:
Effects of storage time prolongation
on in vivo and in vitro characteristics
of 4° C–stored platelets. Transfusion
2020; 60:613–21
81 Rondina MT, Garraud O: Emerging
evidence for platelets as immune and
inflammatory effector cells. Front
Immunol 2014; 5:1–6
82 Morrell CN, Aggrey AA, Chapman LM,
et al.: Emerging roles for platelets as
immune and inflammatory cells. Blood
2014; 123:2759–67
Vox Sanguinis (2021) 116, 477–488

COMMENTARY
The interaction between Glycophorin A (GPA) and Band 3 in
the formation of the Wright b (Wrb) antigen
Serena Ekman,1,2 Ross T. Barnard,1,3 Robert Flower,1,2
1
ARC Training Centre for Biopharmaceutical Innovation, Australian Institute of Bioengineering and Nanotechnology, The University of Queensland,
Brisbane, QLD, Australia
2
Australian Red Cross Lifeblood (Formerly Australian Red Cross Blood Service), Research and Development, Kelvin Grove, QLD, Australia
3
School of Chemistry and Molecular Biosciences, The University of Queensland, Brisbane, QLD, Australia
4
Australian Red Cross Lifeblood (formerly Australian Red Cross Blood Service), Research and Development, Alexandria, NSW, Australia
Glycophorin A (GPA) and Band 3 (also known as Anion
Exchanger 1) are the two most abundant integral proteins in
the red blood cell (RBC) membrane [1]. GPA facilitates the
expression of Band 3 in the cell membrane [2], and the two
proteins are co-precipitated by an anti-Wrb antibody [3].
Wrb, which is antithetical to Wra [4], is a high-frequency
antigen, and antibodies against it can cause haemolytic
transfusion reactions [5]. Although it has been observed that
Glu658 of Band 3 is a required amino acid for the formation
of the Wrb antigen [6], a number of proposals have been put
forward for the GPA site of interaction.
Bruce et al. [6] proposed that the interaction occurs
between residues 80–89 (formerly 61–70) of GPA, with
Arg80 (formerly Arg61) potentially forming a charge-pair
with Glu658 of Band 3. However, Huang et al. [7] have
suggested that Arg80 is not crucial for the formation of
the antigen, and that Wrb is reliant on a longer sequence
in GPA: Val70 to Ile95 (formerly 51–76). This sequence is
encoded by part of exon 3 and extends to the beginning
of the transmembrane domain encoded in exon 5.
Early computational prediction of Gly74 to Phe87 (for-
merly 55-68), using a small GPA peptide, presented as an
a-helical region [8]. Poole et al. [9] showed that the alter-
ation of GPA Ala84 (formerly Ala65)?Pro resulted in
formation of an aberrant Wrb. They concluded that
because Proline is a helix breaker, this polymorphism dis-
rupted the proposed a-helical structure. When modelling
various small GPA peptides in our group, similar a-he-
lices were seen. However, when modelling the complete
extracellular domain of the GPA monomer (performed
using de novo techniques facilitated by the Robetta soft-
ware [10]), it appeared to be comprised of 3 to 5 b-
strands. After 200-ns molecular dynamics simulations
Correspondence: Xuan T. Bui, ARC Training Centre for
Biopharmaceutical Innovation, Australian Institute for Bioengineering
and Nanotechnology, The University of Queensland, Australia, Building
75, Cnr College Rd & Cooper Rd, St Lucia QLD 4072
E-mail: x.bui@uq.edu.au
Vox Sanguinis (2021) 116, 489–492
© 2020 International Society of Blood Transfusion
DOI: 10.1111/vox.13055
Alison Gould4 & Xuan T. Bui1,2
using GROMOS [11], it was determined that although the
majority of the monomeric extracellular domain appeared
to have no stable secondary structure, a stable b-hairpin
was identified at residues Glu64 to His85. The Gly74 to
Phe87 region of GPA had the propensity to form a b-
sheet as one half of this b-hairpin.
To further investigate this, our group produced initial
homology models, based on our GPA model, showing that
GP.Sch (Wrb + ve) displays a b-hairpin secondary struc-
ture, in contrast to the highly homologous GP.Dantu
(Wrb-ve) which lacks the b-hairpin (Figs. 1 & 2). Further-
more, the b-sheet in GP.Sch, formed at 49VQLAH53, and
b-sheet in GPA at 81VQL83 are consistent with the resi-
dues proposed by Huang et al. [7] to be critical for Wrb
formation: 81VQL83 His85 and 87FSEP90 (formerly 62–64,
66 and 68–71). 87FSEP90 is the start of the GPA trans-
membrane domain, where Young et al. [12] showed that
87
FSE89 is associated with enhancement of Band 3 anion
transport capabilities.
Huang et al. [7] further postulated that stabilization of
Wrb is reliant upon parallel packing of the a-helical
transmembrane domains of GPA and Band 3, forcing
proximity between the extracellular domains. Molecular
dynamics simulations of Band 3 and GPA by Kalli and
Reithmeier [13] substantiate this finding. These simula-
tions include only the transmembrane domain of both
proteins, with Arg80 (GPA) anchored to Glu658 (Band 3)
based on the findings by Bruce et al. [6]. Our simulations
placed Arg80 on the end of a central b-sheet (Fig. 3B)
rather than at the end of a flexible loop segment as done
by Kalli and Reithmeier [13]. With the addition of the
extracellular domains and altered presentation of Arg80,
it is possible that the molecular dynamics of GPA/Band 3
interactions would not mimic what is shown by Kalli and
Reithmeier. However, it is possible that the clustering of
Band 3 shown in Kalli and Reithmeier’s simulations, and
predicted by Huang et al. [7,13], is facilitated by residues
in the transmembrane domain as opposed to anchoring
by Arg80 and Glu658.
489
490 S. Ekman et al.

Fig. 1 Initial tertiary structural predictions of GPA, GPB and hybrid glycophorins GP.Sch, GP.Dantu, GP.SAT, GP.Hil and GP.Mur. Models were produced
using comparative modelling techniques under the parameters of the Robetta software [10] using a previously designed GPA extracellular domain mono-
mer model as a template (GPA monomer created using de novo modelling techniques through Robetta software). GPA exon 3 (or exon 3 insert) is
coloured in cyan, GPA exon 4 is coloured in dark blue, GPB exon 2 in orange, GPB pseudoexon 3 segment is coloured in red, GPB exon 4 is coloured in
magenta. GPA or GPB transmembrane (TM) and Wrb status included for each structure. The images were produced using PyMOL [15].

Fig. 2 Comparison of MNS variant allele sequences, presence of Wrb, b-hairpin and GPA transmembrane (TM). GPA exons presented in blue, and GPB
exons presented in green for GPA, GPB and hybrid glycophorins GP.Dantu, GP.Sch, GP.SAT, GP.Hil and GP.Mur. GPA, GPB, GP.Dantu, GP.Sch sequences
adapted from [7,14], GP.SAT sequence from [16], GP.MiV (GP.Hil) sequence from [17], other sequences. Wrb interaction site by Huang et al. and Hsu
et al. outlined in red [7,14].

An alternate site for Wrb expression was proposed by non-expressing cells, whilst expression of Wrb and Band
Hsu et al. [14] as their study showed that GPA expression 3 was increased in GP.Mur-positive cells. They concluded
levels remained consistent in GP.Mur expressing and that the increase in expression of Wrb was due to GP.Mur

© 2020 International Society of Blood Transfusion

Vox Sanguinis (2021) 116, 489–492
 Insight on the formation of the Wright b (Wrb) antigen 491

Fig. 3 Comparison of proposed GPA amino acids required for Wrb formation. GPA and GP.Mur extracellular homodimer created using HADDOCK web
server [18] from GPA and GP.Mur monomers, respectively. GP.Mur monomer created from comparative modelling techniques using Robetta software [10]
with GPA monomer template. GPA transmembrane X-ray crystal structure PDB 5EH4 [19]. GP.Mur transmembrane domain created using comparative
modelling of GPA transmembrane. GPA extracellular domain and transmembrane domain homodimer coloured in green with Wrb interaction sites in red
A) Ridgwell et al. [8], B) Bruce et al. [6] (with essential amino acid Arg80 in blue), C) Huang et al. [7], D) Poole et al. [9] (with essential amino acid
Ala84 in blue), E) Our identified exon 3-4 b-hairpin (residues 64–84). GP.Mur extracellular domain and transmembrane domain homodimer in blue with
interaction sites in red F) Our identified exon 3-4 b-hairpin (residues 64–84) and G) Wrb interaction site proposed by Hsu et al. [14] (with Mur antigen
coloured green). Images were visualized using PyMOL [15].

interacting with Band 3 alongside GPA, hence its facilita-
tion of Wrb production. However, we propose that the
increase in Band 3 expression, facilitated by GP.Mur,
allows a greater proportion of GPA to form a complex
with Band 3, rather than GP.Mur itself forming the Wrb
antigen. This is based on the report by Poole et al. [1]
that orientation plays a crucial role in Wrb expression,
indicating that not all available Band 3 proteins would
interact with all available GPA proteins. This hypothesis
could be tested by repeating the experiments of Hsu et al.
[14] using transfected cells that express GP.Mur and not
GPA, to determine whether there is still Wrb expression.
If GP.Mur is shown to be directly involved in Wrb forma-
tion, this would be in contrast to the mechanism proposed
by Huang et al. [7].
Hsu et al.’s Wrb antigen site was identified as the
region Asp46 to His60 (GP.Hil)/Ser62 (GP.Mur). This was
because Gp.Mur and Gp.Hil differ only in this region,

© 2020 International Society of Blood Transfusion

Vox Sanguinis (2021) 116, 489–492
492 S. Ekman et al.

where GP.Hil has been shown to be Wrb-ve. However, this
does not explain the expression of Wrb on GP.Sch
homozygous cells [7], which lack this region. Moreover,
this region is located in a different equivalent region of
GPA compared to numerous other proposals for the Wrb
interaction site (Fig. 3).
In conclusion, the 87FSE89 segment of the transmem-
brane domain is likely associated with both anion trans-
portation in Band 3 and the formation of the Wrb
antigen. It is also highly likely that an extracellular
domain addition is required for Wrb antigen formation
including the residues 49VQLAH53 which have been pre-
dicted by our molecular dynamic simulations to be one
half of a b-hairpin structure. The requirement for both
the GPA transmembrane domain and the extracellular
domain b-strand would explain why the hybrid gly-
cophorin GP.Dantu (containing GPB extracellular domain
with no b-hairpin, and the GPA transmembrane domain)
can enhance Band 3 expression and increase levels of
anion transport, yet is Wrb-ve [7,12]. Further research will
be necessary to validate these conclusions and to eluci-
date the mechanism by which GP.Mur facilitates elevation
of Wrb antigen expression.
Conflicts of interest
The authors declare no financial or commercial conflict
of interest.
Funding
Elements of this research utilized the facilities, and the sci-
entific and technical assistance of the National Biologics
Facility (NBF) at The University of Queensland. NBF is
supported by Therapeutic Innovation Australia (TIA). Fund-
ing supported by the Australian Research Council through
the Industrial Transformation Research Program (ITRP)
scheme.

References
1 Poole J: Red cell antigens on band 3
and glycophorin A. Blood Rev 2000;
14:31–43
2 Groves JD, Tanner MJ: Glycophorin A
facilitates the expression of human
band 3-mediated anion transport in
Xenopus oocytes. J Biol Chem 1992;
267:22163–70
3 Telen M, Chasis J: Relationship of the
human erythrocyte Wrb antigen to an
interaction between glycophorin A and
band 3. Blood 1990; 76:842–8
4 Wren MR, Issitt PD: Evidence that Wra
and Wrb are antithetical. Transfusion
1988; 28:113–8
5 Daniels G: Diego blood group system;
human blood groups. Oxford: Wiley-
Blackwell, 2013; 336–53.
6 Bruce L, Ring S, Anstee D, et al.:
Changes in the blood group Wright
antigens are associated with a muta-
tion at amino acid 658 in human ery-
throcyte band 3: a site of interaction
between band 3 and glycophorin A
under certain conditions. [see com-
ments] Blood 1995; 85:541–7
7 Huang C, Reid M, Xie S, et al.: Human
red blood cell Wright antigens: a
genetic and evolutionary perspective
on glycophorin A-band 3 interaction.
Blood 1996; 87:3942–7
8 Ridgwell K, Tanner MJ, Anstee DJ: The
Wrb antigen, a receptor for Plasmod-
ium falciparum malaria, is located on a
helical region of the major membrane
sialoglycoprotein of human red blood
cells. Biochem J 1983; 209:273–6
9 Poole J, Banks J, Bruce LJ, et al.: Gly-
cophorin A mutation Ala65 –>; Pro
gives rise to a novel pair of MNS alle-
les ENEP (MNS39) and HAG (MNS41)
and altered Wrb expression: direct evi-
dence for GPA/band 3 interaction nec-
essary for normal Wrb expression.
Transfus Med 1999; 9:167–74
10 Kim DE, Chivian D, Baker D: Protein
structure prediction and analysis using
the Robetta server. Nucleic Acids Res
2004; 32:W526–W31
11 Schmid N, Eichenberger AP, Choutko
A, et al.: Definition and testing of the
GROMOS force-field versions 54A7
and 54B7. Eur Biophys J 2011; 40:843
12 Young MT, Tanner MJA: Distinct regions
of human glycophorin A enhance human
red cell anion exchanger (Band 3; AE1)
transport function and surface traffick-
ing. J Biol Chem 2003; 278:32954–61
13 Kalli AC, Reithmeier RAF: Interaction
of the human erythrocyte Band 3
anion exchanger 1 (AE1, SLC4A1)
with lipids and glycophorin A:
Molecular organization of the Wright
(Wr) blood group antigen. PLoS Com-
put Biol 2018; 14:e1006284
14 Hsu K, Lin Y-C, Lee T-Y, et al.: Mil-
tenberger blood group antigen subtype
III (Mi.III) supports Wrb expression.
Vox Sang 2011; 100:389–94
15 DeLano WL: Pymol: An open-source
molecular graphics tool. CCP4 News-
lett Protein Crystallogr 2002; 40:82–92
16 Huang C, Reid M, Okubo Y, et al.:
Glycophorin SAT of the human ery-
throcyte membrane is specified by a
hybrid gene reciprocal to glycophorin
Dantu gene. Blood 1995; 85:2222–7
17 Johe K, Vengelen-Tyler V, Leger R,
et al.: Synthetic peptides homologous
to human glycophorins of the Mil-
tenberger complex of variants of
MNSs blood group system specify the
epitopes for Hil, SJL, Hop, and Mur
antisera. Blood 1991; 78:2456–61
18 van Zundert GCP, Rodrigues JPGLM,
Trellet M, et al.: The HADDOCK2.2
web server: user-friendly integrative
modeling of biomolecular complexes.
JMB 2016; 428:720–5
19 Trenker R, Call ME, Call MJ: Crystal
structure of the glycophorin A trans-
membrane dimer in lipidic cubic phase.
J Am Chem Soc 2015; 137:15676–9

© 2020 International Society of Blood Transfusion

Vox Sanguinis (2021) 116, 489–492
 Vox Sanguinis (2021) 116, 493–496
© 2020 The Authors.
COMMENTARY Vox Sanguinis published by John Wiley & Sons Ltd on behalf of International Society of Blood Transfusion
DOI: 10.1111/vox.13060

Evaluation of SARS-CoV-2 antibody titers and potency for

convalescent plasma donation: a brief commentary
Elise Wouters,1,* Maurice Steenhuis,2,* Hubert Schrezenmeier,3,4 Pierre Tiberghien,5,6 Heli Harvala,7
Hendrik B. Feys1 & Ellen van der Schoot 2

1
Transfusion Research Center, Belgian Red Cross-Flanders, Ghent, Belgium
2
Department of Immunopathology, Sanquin Research and Landsteiner Laboratory Academic Medical Centre, Amsterdam, Netherlands
3
Department of Transfusion Medicine, Ulm University, Ulm, Germany
4
urttemberg – Hessen and
Institute for Clinical Transfusion Medicine and Immunogenetics, German Red Cross Blood Transfusion Service Baden-W€
University Hospital Ulm, Ulm, Germany
5
Etablissement Francßais du Sang, La Plaine St-Denis, France
6
UMR1098 RIGHT, INSERM, Etablissement Francßais du Sang, Universite de Franche-Comte , Besancßon, France
7
Microbiology Services, NHS Blood and Transplant, London, UK

Severe acute respiratory syndrome coronavirus 2 (SARS-
CoV-2) is the causative agent of the ongoing COVID-19
pandemic. It is responsible for more than 1 million deaths
worldwide already [1]. Because preventive and anti-viral
treatment options are still limited, COVID-19 convalescent
plasma (CPP) has been suggested as a potential therapy
[2–4].
‘Convalescent’ implies that anti-SARS-CoV-2 antibod-
ies are present in plasma collected from individuals
recovered from COVID-19. However, the dose and nature
of antibodies required to effectively interfere with a
SARS-CoV-2 infection is unclear. Most ongoing observa-
tional studies and prospective clinical trials currently
focus on neutralizing antibodies (nAbs) that interfere with
viral binding to host cells, but non-neutralizing antibod-
ies might mediate a therapeutic effect as well. These and
other unknowns highlight the importance of testing CCP
efficacy in randomized trials. This commentary conse-
quently does not claim to provide evidence on how to
select potent CCP, but does want to provide an opinion-
based discussion on how to investigate CCP potency.
The antibody level in CCP varies greatly between
donors. Therefore, it is required to measure antibody titer
and/or to assess the neutralization potency of CCP. The
current gold standard for the latter is in vitro viral neu-
tralization like in the plaque reduction neutralization test
(PRNT) or microneutralization (MN) assay. Both measure
the ability of nAbs to prevent infection in vitro calculated
either as a reduction in the formation of plaques or as the
inhibition of viral infectivity in a cell monolayer,
Correspondence: Hendrik B. Feys, Transfusion Research Center, Belgian
Red Cross-Flanders, Ghent, Belgium
E-mail: hendrik.feys@rodekruis.be
*These authors contributed equally to this work.
respectively [5,6]. These assays utilize live SARS-CoV-2
virus and, hence, require a biosafety level 3 (BSL-3) facil-
ity. In addition, it is time-consuming (5–7 days). Further-
more, the output data cannot be compared among
laboratories because different assay readouts (e.g. virus
concentration or % inhibition) and protocols are currently
being used. In addition, an international standard is not
yet available. Blood establishments may choose to partner
with a virology laboratory that can perform viral neutral-
ization on donor samples. Alternatively, other assays are
available using pseudoviruses (i.e. a recombinant virus
expressing a SARS-CoV-2 protein) that require lower bio-
safety levels [7].
Anti-SARS-CoV-2 antibody titers can also be measured
using immunoassays such as enzyme-linked (ELISA) and
chemiluminescent immunoassays (CLIA) which are based
on biochemical detection of antibody binding to viral
proteins. Recently, the FDA suggested that all putative
CCP donations should be tested in the Ortho VITROS
SARS-CoV-2 IgG CLIA-based test and donations with a
signal to cut-off of 12 or higher to be qualified as a high
titer plasma [8]. In contrast, European blood establish-
ments are using a variety of commercial immunoassays
(Table 1), making it more difficult to compare data across
the region. Sensitivities and specificities of the commer-
cial assays presented in Table 1 can differ from those
provided by the respective manufacturers. Thresholds,
sensitivities and specificities may change depending on
sample size, the timing post-symptom onset and the sero-
prevalence in the population [9,10].
Immunoassays allow the detection of total or isotype-
specific antibody binding the spike (S), receptor binding
domain of spike (RBD) or nucleocapsid (N) proteins. In
our opinion, immunoassays for IgG targeting RBD are
most likely to be relevant because (i) most potent

This is an open access article under the terms of the Creative Commons Attribution‐NonCommercial‐NoDerivs License,
which permits use and distribution in any medium, provided the original work is properly cited, the use is non‐commercial and 493
no modifications or adaptations are made.
494

neutralizing antibodies are directed towards RBD, (ii)

References
IgG is efficiently transported across the epithelial lung

[22]
[23]
[22]
[24]
[25]
barrier [11] and (iii) IgG has a longer half-life. Finally,
immunoassays are compatible with BSL-1 facilities, do
not require sophisticated technology and may be emu-

10 432/10 453
lated on robots to increase throughput.
Sample size

As viral neutralization assays are not high-throughput

145/146
600/600
156/157
994/995
and thus may become rate limiting for CCP release to
patients, immunoassays may be used to select CCP
donations. However, the immunoassay threshold that
selects a plasma product as CCP then ideally relates reli-
Specificity (% [95% CI])

ably and reproducibly to a corresponding neutralization

titer. Recently, several research groups reported on this
(99.7–99.9)
(96.0–100)
(99.4–100)
(97.0–100)
(99.4–100)

correlation [12–15]. For example, Luchsinger et al. found

correlations between the Ortho IgG (r2 = 0.75), Abbott
IgG (r2 = 0.72) and an in-house IgG ELISA (r2 = 0.69)
99.0
100
99.0
99.9
99.8

with a pseudovirus neutralization assay [12]. Similar

results have been observed for the EUROimmun IgG
ELISA and a microneutralization or pseudotype assay
Sample size

[13]. Another in-house RBD-based IgG ELISA correlated

186/187
140/150

500/540
184/185

well with virus neutralization (r2 = 0.89) [14]. Recently,

61/75

a correlation between anti-spike EUROimmun IgA and

virus neutralization (PRNT) was found, indicating that
also IgA might play a role in virus neutralization [15].
Sensitivity (% [95% CI])

These efforts are at least suggestive for correlation

between certain immunoassays and viral neutralization.
(97.0–100.0)d
(88.2–96.3)b

(90.0–94.7)c
(71.0–89.0)a

The ELISA threshold and/or neutralization titer used

(97.0–100)a

to distinguish CCP from non-CCP plasma remains an

arbitrary choice [16]. For viral neutralization, it ranges
from 1:40 to 1:320 while the FDA recommends 1:160,
99.0
93.3
81.0
92.6
99.5

but without an international standard these titers are not

comparable yet [8]. Note that the consequence of any
threshold for an immunoassay is a shift in the balance
Ortho clinical diagnostics
Beijing wantai biological

bearing a risk of releasing poorly neutralizing CCP units

Table 1 Frequently used serological antibody assays for evaluation of CCP.

on the low end, and restricting release of potentially

neutralizing CCP on the high end (Fig. 1). In England,
Euroimmun
Company

neutralizing antibody titers of 1:100 or higher were

Abbott
Roche

measured in 34% of donations, while using a higher

cut-off would likely have prevented a sufficient supply
of CCP to fulfill trial needs [17].
Of note, unbiased screening of all donors using
Antigen

S-RBD
S-RBD

immunoassays without prior information on SARS-CoV-2

S1
N
N

infection is not advised. As the actual number of seroposi-

2–3 weeks upon respiratory infection.

tive individuals in the population is low, the positive pre-

≥14 days after PCR confirmation.

dictive value (PPV) of any assay that is not 100% specific

13–73 days post-symptom onset.
Platform

≥20 days post-symptom onset.

will unavoidably cause overrepresentation of false posi-

ELISA

ELISA
CLIA

CLIA
CLIA

tives [18]. Therefore, selection of CCP should be based on

laboratory confirmation of SARS-CoV-2 infection plus a
Sample collection:

neutralization assay or a correlating immunoassay. Obser-

EUROimmun IgG
Wantai total Ig

vational studies from Mayo clinic and Salazar et al

Abbott IgG
Elecsys IgG

recently found that CCP is most effective when high

Ortho IgG

amounts of anti-SARS-CoV-2 IgG are present [4,19]. In

Assay

contrast, the multicentre randomized PLACID trial found

b

d
a

© 2020 The Authors.

Vox Sanguinis published by John Wiley & Sons Ltd on behalf of International Society of Blood Transfusion
Vox Sanguinis (2021) 116, 493–496
 495

Fig. 1 Release of neutralizing CCP units using varying assay thresholds. [Colour figure can be viewed at wileyonlinelibrary.com]

no reduction in disease progression nor mortality [20] but
also did not determine nAbs levels upfront. Post hoc analy-
sis showed that the median titer of nAbs in this study was
low. Together with the scientific rationale of biochemical
interference with viral binding, we suggest that CCP selec-
tion is based on medium to high signal thresholds (i.e. the
top 30–40% of donations containing Abs). This selection
strategy may change in the future once the minimal effec-
tive dose of nAbs has been established and high-through-
put standardized assays that can reliably predict viral
neutralization potency are available.
As mentioned previously, standardization or calibration
of these immuno- and neutralization assays to allow
comparison of data across studies has not yet been per-
formed. In this context, the European Commission and
the European Blood Alliance (EBA) recently launched a
joint initiative to support high-quality clinical evaluation
of CCP. This SUPPORT-E consortium (Supporting high-
quality evaluation of COVID-19 convalescent plasma
throughout Europe) [21] will investigate the relationship
between (i) donor and donation parameters, (ii) antibody
content and nature and (iii) clinical outcome of CCP
recipients in EU cohorts. The consortium will also provide
support for testing and distributes calibration standards
among participating blood establishments in the EU to
allow cross border standardization of assays. In addition,
international standards are anticipated to be made avail-
able by the WHO in December 2020, which will facilitate
such direct comparisons [18].
Although the observational studies are suggestive for
CCP efficacy, hard evidence is lacking. Additional studies
are required, but IgG levels obtained by ELISA seem to
correlate well with virus neutralization titers. This indi-
cates that an ELISA/CLIA assay can be used to select CCP
donors, also in the light of the urgency. However, stan-
dardization of ELISAs will be essential.
Acknowledgements
We thank Gaia Mori and Catherine Hartmann for the
expert co-ordination of the SUPPORT-E project.
Conflict of interests
The authors declared no potential conflicts of interest
with respect to the research, authorship and/or publica-
tion of this article.
Funding
This study was supported by the European Commission
(SUPPORT-E, grant number 101015756).

References
1 Coronavirus/COVID-19 Worldwide
Cases Live Data & Statistics. 2020.
https://covidstatistics.org/ (Last
accessed October 22nd 2020).
2 Casadevall A, Joyner MJ, Pirofski LA:
A randomized trial of convalescent
plasma for COVID-19-potentially hope-
ful signals. JAMA 2020; 324:455–7
3 Chai KL, Valk SJ, Piechotta V, et al.:
Convalescent plasma or hyperimmune
immunoglobulin for people with COVID-
19: a living systematic review. Cochrane
Database Syst Rev 2020; 10:CD013600
4 Joyner MJ, Senefeld JW, Klassen SA,
et al.:Effect of Convalescent Plasma on
Mortality among Hospitalized Patients
with COVID-19: Initial Three-Month
Experience. medRxiv 2020.
5 Wolfel R, Corman VM, Guggemos W,
et al.: Virological assessment of hospi-
talized patients with COVID-2019. Nat-
ure 2020; 581:465–9
6 Conzelmann C, Gilg A, Gross R, et al.:
An enzyme-based immunodetection

© 2020 The Authors.

Vox Sanguinis published by John Wiley & Sons Ltd on behalf of International Society of Blood Transfusion
Vox Sanguinis (2021) 116, 493–496
496
assay to quantify SARS-CoV-2 infec-
tion. Antiviral Res 2020; 181:104882
7 Thompson CP, Grayson N, Paton R,
et al.: Detection of neutralising anti-
bodies to SARS-CoV-2 to determine
population exposure in Scottish blood
donors between March and May 2020.
Euro Surveill 2020; 25(42):2000685
8 U.S. Department of Health and Human
Services: Investigational COVID-19
Convalescent Plasma: Guidance for
Industry. September 2020. https://
www.fda.gov/regulatory-information/
search-fda-guidance-documents/inve
stigational-covid-19-convalescent-pla
sma(Last accessed October 22nd 2020).
9 Mathur G, Mathur S: Antibody testing
for COVID-19. Am J Clin Pathol 2020;
154:1–3
10 Theel ES, Harring J, Hilgart H, et al.:
Performance characteristics of four
high-throughput immunoassays for
detection of IgG antibodies against
SARS-CoV-2. J Clin Microbiol 2020;
58:e01243–20
11 Kim KJ, Malik AB: Protein transport
across the lung epithelial barrier. Am
J Physiol Lung Cell Mol Physiol 2003;
284:L247–59
12 Luchsinger LL, Ransegnola B, Jin D,
et al.: Serological assays estimate
highly variable SARS-CoV-2 neutral-
izing antibody activity in recovered
COVID19 patients. J Clin Microbiol
2020; 58:e02005-20–e02020
13 Harvala H, Watkins N, Ijaz S, et al.:
Convalescent plasma therapy for the
treatment of patients with COVID-19:
Vox Sanguinis published by John Wiley & Sons Ltd on behalf of International Society of Blood Transfusion
Assessment of methods available for
antibody detection and their correla-
tion with neutralising antibody levels.
Transf Med 2020
14 Peterhoff D, Gluck V, Vogel M, et al.:
A highly specific and sensitive sero-
logical assay detects SARS-CoV-2
antibody levels in COVID-19 patients
that correlate with neutralization.
Infection 2020; 1–8
15 Jahrsdorfer B, Kroschel J, Ludwig C,
et al.: Independent side-by-side vali-
dation and comparison of four sero-
logical platforms for SARS-CoV-2
antibody testing. J Infect Dis 2020;
16:jiaa656
16 COVID-19 Convalescent Plasma Trans-
fusion. 2020. https://ec.europa.eu/hea
lth/blood_tissues_organs/covid-19_en
(Last accessed October 21th 2020).
17 Harvala H, Mehew J, Robb ML, et al.:
Convalescent plasma treatment for
SARS-CoV-2 infection: analysis of the
first 436 donors in England, 22 April
to 12 May 2020. Euro Surveill 2020;
25(28):2001260
18 FDA calculator to select a COVID-19
Antibody Test for your community.
2020. https://www.cdc.gov/corona
virus/2019-ncov/downloads/lab/fda-ca
lculator.pdf (Last accessed October
22nd 2020).
19 Salazar E, Christensen PA, Graviss EA,
et al.: Treatment of Coronavirus Dis-
ease 2019 Patients with Convalescent
Plasma Reveals a Signal of Signifi-
cantly Decreased Mortality. Am J
Pathol 2020; 190:2290–303
20 Agarwal A, Mukherjee A, Kumar G,
et al.: Convalescent plasma in the
management of moderate covid-19 in
adults in India: open label phase II
multicentre randomised controlled
trial (PLACID Trial). BMJ 2020; 371:
m3939
21 Commission supports crucial research
on convalescent plasma to treat
COVID-19. September 11th 2020.
https://ec.europa.eu/info/news/commis
sion-supports-crucial-research-convale
scent-plasma-treat-covid-19-2020-se
p-11_en (Last accessed October 22nd
2020).
22 GeurtsvanKessel CH, Okba NMA, Igloi
Z, et al.: An evaluation of COVID-19
serological assays informs future diag-
nostics and exposure assessment. Nat
Commun 2020; 11:3436
23 Harritshoej LH, Gybel-Brask M, Afzal
S, et al.: Comparison of sixteen sero-
logical SARS-CoV-2 immunoassays in
sixteen clinical laboratories. medRxiv
2020: 2020.07.30.20165373.
24 Ainsworth M, Andersson M, Auckland
K, et al.: Performance characteristics
of five immunoassays for SARS-CoV-
2: a head-to-head benchmark compar-
ison. Lancet Infect Dis 2020; 20:1390–
400
25 Muench P, Jochum S, Wenderoth V,
et al.: Development and validation of
the Elecsys anti-SARS-CoV-2
immunoassay as a highly specific tool
for determining past exposure to
SARS-CoV-2. J Clin Microbiol 2020;
58:e01694–20
© 2020 The Authors.
Vox Sanguinis (2021) 116, 493–496
 Vox Sanguinis (2021) 116, 497–503
© 2020 International Society of Blood Transfusion
ORIGINAL PAPER DOI: 10.1111/vox.13019

Blood supply management in times of SARS-CoV-2

pandemic – challenges, strategies adopted, and the lessons
learned from the experience of a hospital-based blood
centre
Hem Chandra Pandey, Poonam Coshic, Chippy C S, Priyadarsini Jayachandran Arcot & Karan Kumar
Department of Transfusion Medicine, All India Institute of Medical Sciences, New Delhi, India

Introduction Numerous concerns regarding maintenance of blood inventory have

Received: 8 June 2020,
revised 28 August 2020,
accepted 29 September 2020,
published online 26 October 2020
been raised after SARS-CoV-2 pandemic outbreak. These concerns were based on
the experience of blood centres in previous pandemics where shortage of blood
components was reported. The present study had tried to understand the impact
of SARS-CoV-2 pandemic on blood collection and demand as well as the impact
of disaster planning in maintaining an adequate inventory.
Methods Data related to blood supply and demand were collected retrospectively
using blood bank management software for pre-COVID-19 and COVID-19 time
period and compared. Strategies adopted and effects of changes in existing disas-
ter plans to maintain an adequate inventory were studied.
Results A drastic fall in the red cell inventory was observed as compared to pre-
COVID-19 time period was observed due to disproportionate decrease in blood
collection (1/6 to 1/9 of the previous collection) and demand (1/2 of the previous
demand). The buffer stock fell gradually over a period of three weeks with can-
cellation of planned blood donation drives. A buffer stock equivalent to 2-week
inventory led to adequate inventory in the initial lockdown periods. Similar fall
was observed in the platelet inventory with reduction in the blood collection but
almost a proportionate reduction in the platelet demand led to adequate inven-
tory. No increase in wastage was observed for both red cells and platelets during
this period.
Discussion A buffer stock of blood and blood components, strict adherence to the
transfusion triggers, good coordination with the clinical staff and a prospective
review of blood transfusion requests to ensure rational blood transfusion were
some of the steps which helped us to successfully maintain transfusion require-
ments in the initial phases of the COVID-19 pandemic. Use of first-in-first-out
policy prevented any wastage due to outdating of blood.
Key words: SARS-CoV-2 pandemic, inventory management, buffer stock, disaster
planning.

*Correspondence: Hem Chandra Pandey, Department of Transfusion
Medicine, All India Institute of Medical Sciences, New Delhi,
India-110029.
E-mail: pandeyhemc@gmail.com
Background
Maintaining an adequate and safe blood supply is one of
the primary aims of blood centres across the world. Blood
supply chain is affected by a number of factors which
may operate independently or may influence each other.
Three main factors which govern the blood supply

497
498 H. C. Coshic et al.
include the availability of blood donors, the usage of
blood and blood components in the centre as well as the
availability of critical supplies required for collection and
testing of blood. The extent to which any event affects
blood supply will thus depend on the model of blood
transfusion services being followed and the responses of
the centre/country to such an event.
Blood transfusion services in India are decentralized
with around 3108 licensed blood banks collecting around
12
2 million units against an estimated clinical demand
of 14
6 million units of blood [1, 2]. Most of the blood
centres are hospital based with the blood supply mainly
dependent on replacement donors or blood collected from
donors at off-site voluntary blood donation drives. Each
of these centres maintains their own blood stocks and
processes blood based on daily demands with a minimal
buffer stock to tide over acute disruptions in blood supply
arising out of local epidemics, disasters or other such
local events. Majority of blood transfusions are being
done for correction of anaemia due to nutritional causes
approximating to 39%, for haemato-oncologic needs
approximating to 37% and 3%–4% in thalassaemia
patients [2, 3]. To add to this, most blood banks are
dependent on critical supplies and reagents which origi-
nate in developed countries with only a minority of blood
banking reagents being manufactured in our country.
Despite these limitations, blood centres across the coun-
try fare quite well in situations causing small local disrup-
tions in supply and demand by coordinating with each
other and by following government regulations favouring
transport of blood across different centres with well-de-
fined guidelines [4]. In addition to disrupting the medical
services, the 2019-20 corona virus (SARS-CoV-2) pan-
demic has challenged the blood centres across the world to
maintain a balance between blood supply and demand in
an unprecedented way [5–7]. A significant reduction in
blood donation was also observed during previous SARS
and MERS-CoV outbreaks due to pan-lockdown, panic,
cessation of off-site blood donation drives and disruption
of transport services [8–10]. During the past three months
since the first case of SARS-CoV-2 was reported in India
on 29 January 2020, blood centres have adopted a number
of strategies to tide over this crisis. The purpose of the pre-
sent study is to share our experiences and the effect of var-
ious policies adopted by our blood centre from time to time
and the lessons learned from it for future pandemics.
Material and methods
Study setting
The study was done in Main Blood Bank, Department of
Transfusion Medicine, AIIMS, New Delhi and blood stock
data from the period starting from the day when first case
of COVID-19 was reported in the country (i.e. 29 Jan
2020) till the end of the first lockdown in India (i.e. 14
April 2020) was included. Ethical permission to collect
and analyse the data was obtained from institute ethics
committee.
Our institute is a tertiary level research institute and
receives referral for complex cases from all over the
country. The hospitals attached to the institute provide
services catering to a number of specialities ranging from
basic medical/surgical services to complex procedures
including bone marrow transplants, organ transplants etc.
There are three hospital-based blood banks with the other
two blood banks providing services to trauma patients
and cardio/neuro patients respectively. Our blood bank
supplies blood and blood component to all other speciali-
ties other than those served by the other two blood banks
and a good coordination exists between the three blood
banks in case of shortage of blood in one of the centres.
Study time periods
For the ease of describing the supply and demand
dynamics, we divided the study period into three distinct
periods. The first time period (Time Period A) included
the blood stock data from 29 Jan 2020 to 26 Feb 2020.
During this period, the blood bank supply and demand
were unchanged and not affected by the pandemic with
only a single case reported in the country and thus this
data served as a baseline to compare the stock and
demand during the period of lockdown. The blood centre
used to follow first-in-first-out (FIFO) policy with certain
exceptions such as the provision of blood within 7 days
of collection to thalassaemia patients etc.
The second time period (Time Period B) included blood
stock data from 27 Feb 2020 to 21 March 2020. This time
period differed from the first time period by the fact that
our centre implemented steps to increase the buffer stock
(safety stock) of red cells to approximately two weeks
instead of 1 week. This decision to increase the buffer
stock was an internal administrative decision taken by
the blood centre last year as part of increased readiness
towards acute red cell shortage in summer season and
was not in response to the impending SARS-CoV-2 pan-
demic. Multiple high-volume off-site voluntary blood
donation drives were planned by the blood centre at reg-
ular intervals to achieve the target of maintaining a
higher buffer stock. FIFO policy was continued during
this period also.
The third time period (Time Period C) included blood
stock data from 22 March 2020 to 14 April 2020. This
time period marked the start of the pan-country lockdown
till the intended period of closure of the first lockdown.
© 2020 International Society of Blood Transfusion
Vox Sanguinis (2021) 116, 497–503

Table 1 Existing disaster plan vs revised/additional strategies to maintain adequate blood supply during SARS-CoV-2 lockdown
Existing disaster management plan
• Maintain buffer stock to 1-week level
• Defer elective surgeries depending on the blood need
• Maintain a list of voluntary blood donors
• Maintain reagent/critical supply stock to at least
1 month
• Arrange blood units from centres unaffected by the
disaster
Effect of lockdown on blood supply and demand during
this period and various strategies to maintain an adequate
and safe blood supply were reflected by this time period.
Changes in the existing disaster management plans were
done to cope up with the disrupted supply chain and their
effects on blood supply were analysed (Table 1). FIFO
policy was implemented strictly and no exceptions to this
policy were followed.
Data collection and analysis
The data were collected using blood bank management
software reports for each day of operation. The data was
entered in an excel sheet and analysed to calculate the
daily stock as well as buffer stock. Blood collection, blood
issue and stock were calculated and represented as med-
ian (interquartile range, IQR), and a comparative analysis
of these parameters was done for the three time periods
described above. The effect of implementing different
strategies during the third time period was also studied.
Results
The blood supply and demand dynamics are discussed
under three different time periods as described in the
methods section. This follows a detailed day-by-day
description of the various effects of COVID-19 pandemic
on blood stock dynamics.
Time period A – normal functioning of the blood
centre
The total number of blood units collected during this time
period were 3452 units over a period of 29 days. The
daily red cell stock during time period A was around
1058 units (IQR = 985–1103) with a daily red cell collec-
tion of around 116 units (IQR = 93–159) and a daily red
© 2020 International Society of Blood Transfusion
Vox Sanguinis (2021) 116, 497–503
Inventory management in SARS-CoV-2 pandemic 499
Additional strategies during SARS-CoV-2 pandemic
• Maintain buffer stock to 2-week level
• Defer/limit routine transfusions in addition to postponing
elective surgery
• Maintain a registry of blood donors from within the institute
staff or neighbouring areas
• Maintain inventory of Single Donor Platelet collected from
voluntary donors
• Maintain reagent/critical supply stock to at least 2 months
• Daily prospective audit of blood requests
cell issue of around 133 units (IQR = 124–157) (Fig. 1). A
buffer stock (safety stock) of around 900 red cell units
was available during this time period which was sufficient
to sustain blood centre activities for a period of one week
if blood supply is disrupted and demand remained same.
There was no discard of blood units due to outdate during
this time period.
The daily stock of whole blood derived/ random donor
platelet (RDP) during time period A was around 185 units
(IQR = 156–226) with a collection of around 101 units
(IQR = 87–151) and a daily issue of 128 units
(IQR = 110–137) (Fig. 1b). Thus, on an average a total of
70 platelet concentrate units remained in the buffer stock.
The data does not include apheresis platelet concentrates
as it is prepared on demand and directed to the patient
for whom it was donated.
Time period B – planned increase to counter
summer season decline in inventory
The total number of blood units collected during this time
period were 4350 units over a period of 24 days. As evi-
dent from Fig. 1a, the daily stock of red cell units
increased to 1596 units (IQR = 1480–1765) despite the
daily collection remaining almost same at 106 units
(IQR = 80–140). This was the result of the high-volume
off-site blood donation drive of >2000 blood units during
this time period and is represented as outlier in Fig. 1a.
The daily issue of red cells remained at the same level
132 units (IQR = 118–153) which resulted in an increase
in the safety stock to 1600 red cell units equivalent to a
stock necessary to keep the blood centre functioning for
12 days. Three red cell units were discarded due to out-
date during this period.
Due to the effect of the high-volume drive the RDP
stock also increased to 223 units (IQR = 163–269) though
the RDP prepared daily during this time period remained
500 H. C. Coshic et al.
Fig. 1 Box plot showing a. Red Blood Cell (RBC) stock, collection and issue b. Platelet concentrate (RDP, Random donor platelet) stock, collection and
issue before and during COVID-19 pandemic.
almost same at 98 units (IQR = 76–126). The RDP issued
daily decreased to 106 units (IQR = 85–120) (Fig. 1b). As
a result, the average buffer stock of RDPs almost doubled
to 120 units equivalent to a stock required for 1 day.
Time period C – effect of the COVID-19 pandemic
on the blood supply and demand
The total number of blood units collected during this time
period were 523 units over a period of 24 days (Fig. 2).
There was a drastic fall in the daily red cell collection to
around 26 red cell units (IQR = 12–30) which was 1/6th
to 1/9th of the collection during the time period A and B
respectively. On the other hand, the red cells issued dur-
ing this time period decreased to around 69 red cell units
(IQR = 61–76) (Fig. 1). Thus, the daily red cell stock
dropped to around 654 red cell units (IQR = 498–915).
There was no discard of blood units due to outdate during
this time period.
Similarly, the RDP prepared and issued also decreased
to 23 units (IQR = 10–29) and 21 units (IQR = 16–36)
respectively with a daily stock of 73 (IQR = 61–99) units
(Fig. 3).
The buffer stock of red cells decreased gradually from
1233 red cell units on the day when lockdown started to
317 red cells at the end of study period. Five blood dona-
tion drives needed to be cancelled in view of the lock-
down and no further blood donation drives were
scheduled during the period of study. The collection of
red cells also decreased and as low as eight units was col-
lected on one of the days. Blood was transferred in and
out of the blood centre to/from other centres to maintain
sufficient stock at all times and to prevent the units from
expiring (Fig. 2).
In contrast to the red cell stock, the RDP stock was
maintained only for an initial period of 7 days after
which the stock decreased drastically and was maintained
at around 50 units daily. The reduction in the demand
was proportionate to the reduction in blood collection
which was not the case for red cells. The blood centre
also collected additional single donor platelets from vol-
untary blood donors who were called by the blood centre
in anticipation of the decreased platelet supply (not
shown in diagram).
Discussion
Maintaining an adequate blood inventory is a dynamic
process and requires that blood centres keep an eye on
various aspects of blood inventory management. Blood
being a perishable product with a limited shelf life (for
RBCs with SAGM – 42 days, RDP – 5 days) pose a num-
ber of challenges for inventory management. A low
inventory exposes to an acute shortage during times of
disaster whereas an overfilled inventory may lead to
expiry of the blood components. The risk of expiry is
more pronounced with platelet concentrates as the expiry
period is just 5 days.
© 2020 International Society of Blood Transfusion
Vox Sanguinis (2021) 116, 497–503
 Inventory management in SARS-CoV-2 pandemic 501

Fig. 2 Dynamics of red cell stock during COVID-19 lockdown. Arrows away from the buffer stock represent red cell units issued to other centres, arrows
towards buffer stock represent red cells received from other blood centres and broken arrows towards buffer stock represent cancelled off-site blood
donation drive along with the expected number of units to be collected as strikethrough numbers.

Fig. 3 Dynamics of RDP stock during COVID-19 lockdown. Arrows towards buffer stock represent RDP received from other blood centres with number
of units in boxes.

COVID-19 outbreak resulted in disruption of various resulted in reduced movement of individuals and thus
aspects of the blood supply dynamics. The lockdown reduced the availability of the blood donors. The fear of
imposed by the government to contain the spread of virus acquiring infection during commute to the blood centre

© 2020 International Society of Blood Transfusion

Vox Sanguinis (2021) 116, 497–503
502 H. C. Coshic et al.
and during the process of blood donation also added to
the reduced number of blood donors presenting to the
blood centre in the initial period of lockdown. It also led
to cancellation of the already planned blood donation
drives due to restrictions in movement of blood bank staff
to the site of blood donation drive and restrictions on
public gathering. These all led to a sharp fall in the blood
collection during initial stages of lockdown which
reduced further as the lockdown progressed.
Outpatient services as well as hospital admissions for
non-emergent services were also curtailed to prevent the
spread in community in the initial period of lockdown.
However, the reduction in demand and supply was not
proportionate as evident at our centre. This was due to
the fact that the transfusion requirements of routine
surgical patients form only 20% of our blood requests
and almost 40–50% of patients are being transfused at
our centre at the emergency department or for transfu-
sion dependent anaemia such as thalassaemia or anae-
mia post-chemotherapy. An additional unplanned
increase in transfusion requirement occurred due to
shifting of the patients from trauma care hospital
attached with our institute to the main hospital. The
trauma centre used to have a separate blood centre in
pre-COVID times whose services were now diverted to
complete the transfusion requirements of COVID patients
due to the transformation of trauma centre into a
COVID-care hospital.
An adequate buffer stock equivalent to 1–2 weeks
requirement of red cells helped us in maintaining emer-
gent transfusions despite sharp decrease in blood supply.
To prevent unnecessary wastage of blood due to further
decrease in the requirement in later stages of lockdown
we strictly implemented the policy of FIFO with exception
only for intrauterine transfusions and exchange transfu-
sions. There was no increase in discard of blood due to
outdating despite reduction in the requirement from the
non-COVID-19 times and having a higher buffer stock
due to strict implementation of FIFO policy. This was in
contrast to some reports where blood centres reported
increased discard of blood components due to decreased
requirements during the lockdown [11]. Major user
departments were requested to adhere to the rational use
of the blood components and a prospective review of
each blood request was started. An adequate buffer stock
also provided us enough time to plan activities and mobi-
lize voluntary blood donors from our internal registry
which mainly comprised of institute staff.
In contrast to the red cell inventory, it is the platelet
inventory which pose a major challenge due to a lim-
ited shelf life of 5 days, it is not possible to maintain a
sufficient buffer stock. At the start of lockdown, we had
maintained a stock which was sufficient for a single
day use and we rationed the platelet inventory strictly
to therapeutic use in bleeding patients as well as pro-
phylactically when the platelet counts were <10 000/µl.
All the prophylactic transfusion requests were accepted
when a haematologist had been consulted for the same.
In addition, voluntary donors from the staff were moti-
vated to donate SDP to be used in case RDPs were not
available.
Although maintaining a blood inventory is a dynamic
process and is affected by a number of factors, following
the basic principles of inventory management and disas-
ter planning is a pre-requisite to tide over crises caused
by COVID-19 like situations. Formulating and imple-
menting local mitigation strategies and following them
strictly can go a long way in maintaining blood inven-
tory during a pandemic [12]. The present study demon-
strates that maintaining a buffer stock of blood and
blood components, strict adherence to the transfusion
triggers, good coordination with the clinical staff and a
prospective review of blood transfusion requests to
ensure rational blood transfusion were some of the steps
which helped us to successfully maintain transfusion
requirements in the initial phases of the COVID-19 pan-
demic. In addition to these, our experience also high-
lights the importance of maintaining a registry of
voluntary blood donors from around the neighbouring
areas of blood centre. In our case, it were the staff mem-
bers who were easily accessible, and the registry helped
us in coordinating with them to maintain our inventory
especially for completing the requests of platelet concen-
trates or some of the O negative transfusions required
for intrauterine/ exchange transfusions. To conclude
planning is an important part of blood inventory man-
agement which provides time to adapt and plan further
even in case of future pandemics.
Acknowledgement
None.
Conflicts of interest
The authors declare that they have no conflict of interest
regarding the submitted article.
Author contribution
HCP: Designed the study, analysed data, wrote the manu-
script and approved the final version. PC: Designed the
study and revised the manuscript. CCS: Data collection
and analysis. PJA: Wrote the manuscript and revised the
manuscript. KK: Wrote the manuscript and revised the
manuscript.
© 2020 International Society of Blood Transfusion
Vox Sanguinis (2021) 116, 497–503

References
1 CDSCO, Ministry of Health and Family
Welfare, Government of India. Regula-
tory requirements of blood and/or its
components including blood products.
Available at https://cdsco.gov.in/openc
ms/opencms/system/modules/CDSCO.
WEB/elements/download_file_division.
jsp?num_id=MzY3 [Last accessed on
2020 June 08].
2 National Aids Control Organisation,
Report on National Estimation of Blood
Requirement in India. Available at
http://naco.gov.in/sites/default/files/
Final%20Estimation%20Report%20of%
20Blood%20Requirement%20in%20Ind
ia%20%281%29.pdf [Last accessed on
2020 June 08].
3 Aggarwal R, Prakash A, Aggarwal M:
Thalassemia: an overview. J Sci Soc
2014; 41:3–6
4 MoHFW - Gazette Notification reg.
Transfer of Blood & Blood Compo-
nents to other Blood Banks & Blood
Donation Camps by Private Hospital
© 2020 International Society of Blood Transfusion
Vox Sanguinis (2021) 116, 497–503
Inventory management in SARS-CoV-2 pandemic 503
Blood Banks, available at http://nac
o.gov.in/sites/default/files/MoHFW%
20Gazette%20Notification%20-%
20Dt.%2003rd%20April%202017.pdf
[Last accessed on 2020 June 08].
5 Yahia AIO: Management of blood sup-
ply and demand during the COVID-19
pandemic in King Abdullah Hospital,
Bisha, Saudi Arabia. Transfus Apher
Sci 2020:102836. http://dx.doi.org/
10.1016/j.transci.2020.102836. [Epub
ahead of print].
6 Leung JNS, Lee CK: Impact of the
COVID-19 – a regional blood centre’s
perspective. ISBT Sci Ser 2020. http://
dx.doi.org/10.1111/voxs.12558. [Epub
ahead of print].
7 Pagano MB, Hess JR, Tsang HC, et al.:
Prepare to adapt: blood supply and
transfusion support during the first 2
weeks of the 2019 novel coronavirus
(COVID-19) pandemic affecting Wash-
ington State. Transfusion 2020;
60:908–911
8 Shan H, Zhang P: Viral attacks on the
blood supply: the impact of severe
acute respiratory syndrome in Beijing.
Transfusion 2004; 44:467–9
9 Teo D: Blood supply management dur-
ing an influenza pandemic. ISBT Sci
Ser 2009; 4:293–8
10 Kwon SY, Lee EH, Kim HS, et al.: Mid-
dle East Respiratory Syndrome Coron-
avirus (MERS-COV) outbreak in South
Korea: risk management at the Korean
Red Cross Seoul Nambu Blood Centre
(Oral abstracts). Vox Sang 2015; 109
(Suppl. 2):18
11 Raturi M, Kusum A: The blood supply
management amid the COVID-19 out-
break. Transfus Clin Biol 2020;
27:147–151. [Online ahead of print].
12 Stanworth SJ, New HV, Apelseth TO,
et al.: Effects of the COVID-19 pan-
demic on supply and use of blood for
transfusion. Lancet Haematol 2020; 7
(10):e756–e764.
 Vox Sanguinis (2021) 116, 504–512
© 2020 International Society of Blood Transfusion
ORIGINAL PAPER DOI: 10.1111/vox.13025

Compliance and attitudes of blood donors following

transitioning from permanent to 12-month deferral of men
who have sex with men in Hong Kong
Janice Ying-Chui Lau,1 Cheuk-Kwong Lee,2 Chin-Pok Chan,1 Jennifer Ngar-Sze Leung,2 Chin-Man Poon1 &
Shui-Shan Lee1
1
Stanley Ho Centre for Emerging Infectious Diseases, The Chinese University of Hong Kong, Hong Kong, People’s Republic of China
2
Hong Kong Red Cross Blood Transfusion Service, Hospital Authority, Hong Kong, People’s Republic of China

Received: 28 April 2020,
revised 15 October 2020,
accepted 15 October 2020,
published online 16 November 2020
Background and Objectives Blood safety hinges not just on the scientific ratio-
nale for deferral period but potential donors’ compliance with the prevailing pol-
icy. This study aimed to investigate donors’ awareness, attitudes and compliance
with the two-phased policy implementation of time-limited deferral for men who
have sex with men (MSM) in Hong Kong.
Materials and Methods Three rounds of questionnaire survey were conducted
between July 2017 and June 2019 covering the periods of pre-implementation
(Round A), post-implementation without and with pre-donation questionnaire
revision (Round B and C). Chi-square test and multivariable regression analysis
were performed.
Results Of 3085 donors recruited, 968, 1036 and 1081 completed the surveys in
Round A, B and C, respectively. The non-compliance rate of MSM remained
stable at 0
6% (3/497), 0
4% (2/551) and 0
5% (3/587) among male donors in
Round A, B and C, respectively. Two MSM donors from Round C complying with
the prevailing policy were identified. About two-thirds (60
7%) of respondents
from Round B and C were unaware of the policy change. Overall, over 80% were
either neutral or positive about the change.
Conclusion Our study showed a consistently low non-compliance rate of MSM
over the three periods. The generally high level of acceptance of time-limited
deferral among donors lends support to science-based policy development to pro-
tect blood safety. The identification of compliant MSM donors suggests that the
12-month deferral is effective and acceptable to MSM. With a deferral period far
exceeding the window period, it is a step towards a more equitable policy.
Key words: blood donation, blood safety, donor deferral, men who have sex with
men, risk behaviours, transfusion-transmitted infections.

HIV transmission, accounting for 0
002% of cases in

Introduction
Unlike sexual transmission and injection drug use, trans-
fusion of contaminated blood is an uncommon route of
Correspondence: Shui-Shan Lee, Stanley Ho Centre for Emerging
Infectious Diseases, The Chinese University of Hong Kong, 206
Postgraduate Education Centre, Prince of Wales Hospital, Shatin, Hong
Kong, People’s Republic of China
E-mail: sslee@cuhk.edu.hk
high-income and 0
86% in low-income countries [1].
Nevertheless, safeguarding blood supply, a key compo-
nent in health care services, has continued to be of criti-
cal importance as one means of HIV prevention. Pre-
donation screening aside, donor deferral is a regularly
used monitoring mechanism to exclude prospective
donors whose blood or blood products may affect recipi-
ent safety. With a substantially higher HIV prevalence in
men who have sex with men (MSM), the policy of

504

permanent deferral has been applied globally for decades,
which led to numerous legal and ethical debates on per-
ceived discrimination in the community. Without an
international consensus on safe deferral period, time-lim-
ited deferral policy with varied length was introduced in
some countries, following extensive review on scientific
evidence in conjunction with increasing sensitivity of
blood screening technology in recent years. Countries like
Australia, Belgium, Finland, Germany, Norway, Ireland,
Switzerland and the United States have all implemented
12-month deferral policy. Japan stipulated a 6-month
deferral and Taiwan a 5-year deferral. New Zealand incre-
mentally changed the deferral period from 10 years to
5 years (in 2009) and to 1 year (in 2014). Some nations,
for example, England, Scotland, Wales and Canada, have
further shortened from 12 to 3 months’ deferral, while
the Netherlands, Denmark and France have reduced to
4 months. In 2020, a decision of reducing the 1-year
deferral period to 3 months since the last male-to-male
sex was made both in the United States and Northern Ire-
land. The policy change was a response to blood shortage
caused by the coronavirus disease 2019 (COVID-19) in
the United States [2], and following recommendation by
the Advisory Committee for the Safety of Blood, Tissues
and Organs in Northern Ireland [3].
Despite the availability of highly effective HIV screen-
ing, there remains the probability of infections from
donated blood escaping detection during the window
period. In this connection, self-deferral of donors with
recent risk of HIV exposure could function as a comple-
mentary strategy to protect blood safety. Non-compliance
of HIV-infected donors would potentially lead to
increased risk of transfusion-related infections. The
change of deferral period targeting male donors with his-
tory of sex with another man (MSM) (hereafter referred
as MSM donors) was in line with the strategy of restrict-
ing donation by the length of window periods instead of
permanently. Blood safety hinges however not just on
scientific rationale for selected deferral periods but
potential donors’ compliance with the prevailing policy,
and that the implementation of any new deferral policy
needs to be done with caution to ensure communication
is appropriate and compliance is high. International posi-
tion papers recognize the importance of donor compli-
ance in the efficacy of deferral policy [4], and the use of
donor behavioural-based screening as an additional layer
to protect blood safety [5]. Studies from the United
States examining non-compliance rate in blood donors
showed that a proportion of MSM disclosed donating
blood in spite of the implementation of the permanent
deferral policy [6–7]. Recent study in Australia showed
high compliance of MSM with the time-limited deferral
policy, suggesting that perceived equitability of the new
© 2020 International Society of Blood Transfusion
Vox Sanguinis (2021) 116, 504–512
Transitioning to 1-year deferral of MSM donors 505
policy could minimize non-compliance [8], with similar
results cited in a French study [9]. In Canada, however,
no change of compliance was seen when the deferral
period was reduced to 5-year, then 1-year [10]. In Eng-
land, Scotland and Wales, following implementation of
3-month deferral, assessment of surveillance data showed
that there was no increase in the number of non-compli-
ant HIV positive donors [11]. Understanding factors asso-
ciated with non-compliance could inform strategy to
reduce its occurrence. A number of studies have reported
that non-compliance could arise from the underestima-
tion of risk linked with condom use [9,12], not perceiv-
ing one’s sexual behaviour as risky [8,9,13], confidence
in blood screening in detecting infection [9,12], concerns
about confidentiality [12], and considering it unfair if
excluded from donation [9,12]. In Hong Kong, the per-
manent deferral policy has been in place since 1985,
which was recently replaced by a newly introduced 12-
month deferral period. In this study, we investigated the
non-compliance of MSM donors following the implemen-
tation of the new policy, and assessed donors’ awareness
and attitudes towards the new policy. Results of the
study could be useful for supporting the evaluation of
MSM donor deferral policy, improving its implementa-
tion and informing further policy development for
enhancing blood safety.
Men who have sex with men deferral policy
change in Hong Kong
In Hong Kong, blood donation is exclusively managed
by the Hong Kong Red Cross Blood Transfusion Service
(BTS). All potential donors are requested to complete a
health screening questionnaire (HSQ) with items on
deferrable behaviours including male-to-male sex. The
policy of permanently deferring MSM donors was
replaced by time-limited deferral in two phases. In the
first phase, the change of practice from permanent to
one-year deferral starting from 25 September 2017 was
announced through the media, without alteration of the
administered HSQ. Prospective male donors indicating a
‘Yes’ answer for the question of ‘If you are male, have
you ever had oral or anal sex with a man?’ was inter-
viewed by a nurse in donor centre to confirm if sex had
occurred in the preceding 12 months. An MSM was
allowed to donate blood if his last sexual activity took
place over one year ago. Starting from 1 January 2019,
Phase 2 was implemented with a change of wording in
the HSQ. Eligibility determination has since been framed
by the question of ‘In the past 12 months - if you are
male, have you had oral or anal sex with a man?’ An
MSM would be eligible to donate blood if a ‘No’ answer
is indicated.
506 J. Y. C. Lau et al.
Materials and methods
Survey method
This study was composed of 3 rounds of questionnaire
survey A, B, and C conducted during the following peri-
ods: (i) July to September 2017 (baseline), (ii) December
2017 to February 2018 (shortly after Phase 1 implementa-
tion), and (iii) April to June 2019 (shortly after Phase 2
implementation), respectively. The surveys were adminis-
tered at 7 major donor centres, with varying number of
visitors because of geographical differences. In each
round of the survey, the target sample size of each Centre
was estimated from donor statistics provided by the Cen-
tres, in order to obtain a sample of donors proportional
to the size of the donor population for the Centres. Inclu-
sion criteria were eligibility to donate blood, and ability
to communicate in written and spoken Chinese. At each
donor centre, designated research staff approached eligi-
ble donors consecutively during the recruitment period. A
donor who had just given blood and resting in the
refreshment area was approached and explained about
the study. Following consent obtained, a donor was asked
to complete an anonymous self-administered question-
naire that took about 5–10 min, using a tablet computer.
The survey protocol was approved by the Survey and
Behavioural Research Ethics Committee of the Chinese
University of Hong Kong.
Men who have sex with men survey questions
A structured questionnaire in Chinese language was
developed to collect information on (i) demographics; (ii)
deferrable risk behaviours; (iii) donation history and com-
pliance with deferral; and (iv) attitudes towards the
change of policy from permanent to 12-month deferral in
reference to our previous study [14]. Part (ii) included
items on oral or anal sex with another man; condom use;
number of male sexual partners, sexual partners’ HIV sta-
tus; and sex with bi-sexual persons, injecting illicit drug
users, or commercial sex worker in the last 12 months,
identical to those listed on the HSQ administered by the
donor centres, for both male and female donors. To assess
the reasons for supporting or opposing the policy, part
(iv) listed ‘effectively reduce the number of MSM donors’
(shown in Table 3) as one of the answer options, to gauge
if a period of 12 months is an acceptable policy for defer-
ring higher risk MSM from donation. The same set of
questions was applied in Round A, B and C for purpose
of collecting comparable data. In Rounds B and C, we
included an additional item of: (v) awareness of the 12-
month deferral policy, to assess donors’ knowledge of the
new policy after its introduction.
Data analysis
We used the Statistical Package for Social Studies (SPSS)
program version 25 to analyse the data. Descriptive
statistics and chi-square tests were performed to compare
socio-demographic characteristics of donors across the
three study periods. Multivariable regression analysis was
used to determine factors associated with awareness and
attitudes towards the changing policy. P-values of <0
05
were considered as significant.
Results
Demographics of recruited donors
The total number of donors approached was 3781, of
whom 505 refused to participate, and 158 withdrew with-
out completing the survey. The overall response rate was
82
5% (range among 84
7%, 81
2% and 81
7%, in Round
A, B and C, respectively). During the recruitment period,
the recruited donors represented 39
8%, 34
8%, and
35
5% of all donors giving blood on the days of recruit-
ment in Round A, B and C, respectively. There were
threefold as many donors participating at the large cen-
tres in each study period (74
9%, 71%, 69
2%) compared
to the small centres (25
1%, 29%, 30
8%), respectively.
There was moderately significant variation of sampled
size of donors from donor centres between the 3 rounds
of survey (X2 = 8
407, P = 0
015).
Characteristics of the recruited donors were similar
across the 3 rounds (Table 1). Of a total of 3085 respon-
dents, 968, 1036 and 1081 completed surveys in Round
A, B and C, respectively. Overall, slightly more than half
were males. The distribution of the respondents’ ages var-
ied in the 3 rounds (X2 = 48
105, P = <0
001), with a
consistently higher proportion in the age group of 16–29
(45
1%, 37
2%, 33
1%). Almost all were Hong Kong per-
manent residents, non-student donors, and more than half
of them had attained post-secondary education or above.
The majority of the recruited donors were repeat donors.
Men who have sex with men among blood donors
Out of 1635 male donors, 10 (0
6%) disclosed having had
sex with another man. All were repeat donors, with a major-
ity (8/10) reporting never having been deferred. In Round A
before the 12-month deferral policy was introduced, male
donors who reported male-to-male sex either ‘within
12 months’ or ‘over 12 months ago’ were classified as non-
compliant. Round B and C corresponded to the period of
introduction of the 12-month deferral policy, and thus, only
male donors with male-to-male sex ‘within 12 months’ were
classified as non-compliant. The non-compliance rate
observed remained relatively stable at 0
6% male donors (3/
© 2020 International Society of Blood Transfusion
Vox Sanguinis (2021) 116, 504–512
 Transitioning to 1-year deferral of MSM donors 507

Table 1 Characteristics of donors in Round A, B and C the respondents from these two rounds were not aware of
the policy change that has already been implemented.
Round A Round B Round C
Multivariable regression analysis shows that males (ad-
(N = 968) (N = 1036) (N = 1081)
justed odds ratio [aOR]: 1
37, 95% CI: 1
14–1
63), non-
N % N % N % student donors (aOR: 1
44, 95% CI: 1
05-1
98), having
attained post-secondary education or above (aOR: 1
41,
Male 497 51
3 551 53
2 587 54
3 95% CI: 1
17-1
71), and repeat donors (aOR: 1
83, 95%
Age
CI: 1
31–2
56) were significantly associated with a higher
16–29 437 45
1 385 37
2 358 33
1
level of awareness of policy change. Of the seven MSM
30–39 231 23
9 290 28
0 258 23
9
donors identified in Round B and C, all except one
40–49 176 18
2 201 19
4 247 22
8
50 or over 124 12
8 160 15
4 218 20
2
claimed to have read the HSQ before donation, four
Permanent residents 960 99
2 1023 98
7 1066 98
6 reported that they were aware of the change.
Occupation status Donors’ attitudes towards policy change were evaluated
Student 175 18
1 134 12
9 120 11
1 with data from the surveys as shown in Fig. 1. Overall,
In employment 686 70
9 812 78
4 865 80
0 more than 80% were either neutral or positive about pol-
Not employed 107 11
1 90 8
7 96 8
9 icy change at the three rounds of survey. A higher pro-
Education portion (about half) held positive attitude at Round A, as
Secondary or below 390 40
3 348 33
6 404 37
4 compared with that in Round B and C. Positive attitude
Post-secondary 218 22
5 229 22
1 213 19
7
was significantly associated with age <30 (aOR: 2
31,
Tertiary or above 360 37
2 459 44
3 464 42
9
95% CI: 1
62–3
28) (Table 2). Reasons for supporting or
Repeat donor 881 91
0 944 91
1 972 89
9
opposing the policy change among donors are shown in
Types of donor centrea
Large centresb 725 74
9 736 71
0 748 69
2
Table 3. Two of the most common reasons for supporting
Small centresb 243 25
1 300 29
0 333 30
8 the policy change after implementation were its capability
of ‘reducing stigmatization’ (46
9%), and ‘effectively
a
Donor centres that have enrolled 20 000 or more donors were classified reducing the number of HIV positive donors’ (45
1%).
as ‘large’ contrasting ‘small’ centres with less than 20 000. An apheresis Before policy implementation (Round A), the main rea-
centre was excluded. sons for the opposition were its ‘inability to reduce
b
Four and 3 donor centres were included in the category of ‘larger cen-
stigmatization’ (53
7%), followed by ‘lack of supporting
tres’ and ‘smaller centres’, respectively.
scientific evidence’ (44
3%) and ‘inability to reduce the
*P < 0
05.
number of HIV positive donors’ (40
9%). After implemen-
**P < 0
01.
***P < 0
001; X2 = chi-square statistics.
tation, the main reasons became ‘inability to reduce the
number of HIV positive donors’ (66
2%). Such concern
was more prevalent in females (75
2%) than males
497, 95% CI: 0
00–1
28) in Round A, 0
4% (2/551, 95% CI: (59
2%). Around three-quarters of respondents (75
5%;
0
00–0
87) in Round B and 0
5% (3/587, 95% CI: 0
00–1
09) 800/1060) did not give any comments because of ‘unfa-
in Round C. Two male donors from Round C reported having miliarity with the issue’. Within the subgroup of those
male-to-male sex over 12 months ago and were therefore expressing ‘unfamiliarity with the issue’, 71
2% (274/385)
classified as being compliant with the prevailing policy of in Round B and 76
0 % (315/415) in Round C responded
deferral. Among the eight non-compliant MSM, half (4/8) that they were not aware of the policy change.
reported consistent condom use for sex. Five had only one As shown in Fig. 2, the preferred interval of deferral
male partner in the preceding 12 months, while the rest varied considerably among all donors. The plurality of
reported multiple male partnership (2 with 2–5 partners, and donors did not offer any views (27
7%, 38
7%, 44
5%).
1 with over 10 partners). There were a few inconsistent The most preferred interval of donors was one year
answers regarding sex partnership including 3 who reported (20
2%, 17
4%, 15%), while a consistent proportion of
having male partners which could not be clearly defined as <10% preferred ‘permanent deferral’ (7
9%, 8
7%, 9
1%).
either regular or non-regular. About half of the donors (47%) were uncertain about the
length of the window period for HIV tests, while most
(37
7% of the remainder) perceived that it should be
Donors’ awareness of policy change and their
between 3 and 6 months (Fig. 3). By dichotomising the
attitudes
perceived length of window period and preferred deferral
Donors’ awareness of the change of policy from perma- interval using 3 months and 1 year as the respective cut-
nent to 12-month deferral was examined in Round B and off, a significantly higher proportion of those with a long
C of the surveys. About two-thirds (60
7%; 1286/2117) of perceived window period preferred longer deferral

© 2020 International Society of Blood Transfusion

Vox Sanguinis (2021) 116, 504–512
508 J. Y. C. Lau et al.

Fig. 1 Attitudes towards the policy change.

Table 2 Factors associated with the awareness and attitudes on policy change in Round B & C

Awarenessa (N = 2117) Attitudesb (N = 1057)

Factors Adjusted odds ratio 95% CI Adjusted odds ratio 95% CI

Male 1
37** 1
14, 1
63 0
94 0
72, 1
23

Age < 30 1
12 0
91, 1
39 2
31*** 1
62, 3
28
Non-student 1
44* 1
05, 1
98 1
30 0
81, 2
10
Post-secondary or above 1
41*** 1
17, 1
71 0
91 0
68, 1
22
Permanent resident 1
11 0
49, 2
49 1
00 0
19, 5
17
Repeat donor 1
83*** 1
31, 2
56 0
74 0
45, 1
20
Aware of policy change - - 1
26 0
96, 1
64

a
0 = Unaware, 1 = Aware.
b
0 = Negative, 1 = Positive (excluding the 1060 responses of ‘Neutral’).
*P < 0
05.
**P < 0
01.
***P < 0
001.

interval (X2 = 7
742, P = 0
005). Results of spearman cor-
relation also showed that the perceived length of window
period was only weakly correlated with the preferred
deferral interval (r = 0
183, P < 0
001).
Discussion
In this study, we compared the non-compliance rate of
MSM towards donor deferral before and after the imple-
mentation of the 12-month deferral policy in Hong Kong,
a city in the Asia Pacific region where similar studies are
scarce. Using a unified recruitment method and a stan-
dard questionnaire, we showed a consistently low non-
compliance rate among MSM at <1% both in the pre-im-
plementation and 2 post-implementation periods of the
new policy, compared to previous local studies with
results ranging from 1
2–2
2% [14–16], and similar to
that of 0
23% in Australia [8], and 0
8% to 1
0% in
Canada [17]. The slightly lower non-compliance rate
could be related to the increased awareness of MSM
about the deferral policy, which was increasingly reported
in the media, though the effect of having more men
becoming eligible for blood donation may play a role. A
high level of compliance is important in enforcing donor
deferral policy such that the residual risk of HIV infection
from transfusion can be minimized. A modelling study
predicted a small increase in risk with a change from life-
time to 12-month deferral for MSM in the UK if compli-
ance was unchanged [18]. Studies in United States and
Canada [19], and Italy [13] provided evidence of
increased safety as a result of improved compliance fol-
lowing the adoption of a time-limited deferral policy,
despite the theoretical risk of ‘relaxing’ the criteria. Addi-
tionally, authors of an Australia study concluded that
there was no increase in prevalent infection after compar-
ing the prevalence of HIV from donors before and after
the 12-month deferral policy [20]. Our results gave a
same level of compliance with deferral criteria among
MSM, both before and after the policy change, suggesting
that such change is unlikely to impact blood safety

© 2020 International Society of Blood Transfusion

Vox Sanguinis (2021) 116, 504–512
 Transitioning to 1-year deferral of MSM donors 509

Table 3 Reasons for donors’ attitudes towards the new policy

Pre-implementation Post-implementation
(Round A) (Round B & C)

Male Female Total Male Female Total

Reasons for support (Multiple options), % based on N, support

Effectively reduce the number of suspected/confirmed HIV + donors 65
9 70
8 68
1 44
0 46
4 45
1
Effectively reduce the number of MSM donors 34
9 24
2 30
0 17
5 15
5 16
5
Reduce stigmatization 21
2 19
2 20
3 45
0 49
0 46
9
Supported by new scientific evidence 22
4 17
8 20
3 36
8 28
3 32
8
Desire to follow the implementation as a plan proposed by other 18
4 6
8 13
1 24
7 16
6 20
9
more developed countries
N, support 255 219 474 389 343 732
Reasons for opposition (Multiple options), % based on N, opposition
Unable to reduce the number of suspected/confirmed HIV + donors 46
2 36
9 40
9 59
2 75
2 66
2
Unable to reduce the number of MSM donors 30
8 22
6 26
2 32
1 30
5 31
4
Unable to reduce stigmatization 47
7 58
3 53
7 22
8 19
1 21
2
Lack of supporting scientific evidence 46
2 42
9 44
3 23
4 17
7 20
9
Irrational to follow others’ implementation since it is not applicable 13
8 17
9 16
1 15
2 12
1 13
8
to all countries
N, opposition 65 84 149 184 141 325
Reasons for not giving comments (Multiple options), % based on N,
not giving comments
Not familiar with this issue 78
5 70
8 74
8 133
0 75
4 75
5
Both the new and former policy could effectively reduce the number 8
5 11
3 9
9 21
5 15
2 13
6
of suspect/ confirmed HIV + donors
Both the new and former policy were supported by the same scientific evidence 5
1 4
2 4
6 15
0 5
9 7
3
Both the new and former policy were unable to reduce the number 6
8 10
7 8
7 13
7 4
6 6
3
of suspect/ confirmed HIV + donors
Both the new and former policy lacked scientific evidence 6
8 7
1 7
0 6
2 3
0 3
3
N, not giving comments 177 168 345 321 495 1060

Fig. 2 Preferred deferral interval.

© 2020 International Society of Blood Transfusion

Vox Sanguinis (2021) 116, 504–512
510 J. Y. C. Lau et al.
Fig. 3 Perceived window period.
unfavourably. Despite the small number of MSM donors
identified, their self-reported inconsistency of condom
use pointed to the needs for safer sex promotion.
One small yet notable finding in our third survey fol-
lowing implementation of the 12-month deferral of MSM
was the identification of two male donors with male-to-
male sex over 12 months before donation. This study did
not inquire about the timing of sex before their previous
donations when the permanent deferral policy was in
place, but the fact that both were repeat donors with past
history of multiple donations did imply that they had
escaped the scrutiny of the permanent deferral mecha-
nism in the past. Their detection at the second phase of
policy change could be by chance, but might also be
explained by their awareness of the new policy and there-
fore their willingness to self-declare in confidence.
Should they present for donation during the first phase,
one’s sexual history would be collected at face-to-face
interview after indicating a history of male-to-male sex
on the HSQ. The specific inclusion of the deferral period
on the HSQ in the second phase eliminated the need for
follow-up interview – a procedure that is subject to
stigma. The stigma attached to disclosure of one’s MSM
status might have led to the common practice of non-
compliance [21–22]. The perception of permanent deferral
as unjust discrimination [12] and the need to disclose
male-to-male sex even for a single event [23] could have
led to higher rate of non-compliance with the permanent
deferral than 12-month deferral policy. Enrolling poten-
tial donors who are likely to increase their compliance
because of the new policy provides a basis for policy
enforcement aimed at blood safety.
Clearly, our data shows a lack of awareness among
donors even after the implementation of the 12-month
deferral of MSM in Hong Kong. Some donors, such as
females, those with lower educational levels were less
knowledgeable about the change of deferral policy. The
time period for collecting data for Round B and C was
around 3–4 months after Phase 1 and 2 policy implemen-
tation, which was relatively short in length and may
explain the low level of awareness recorded as it takes
time for potential donors to be aware of the changes.
Although the new policy primarily affects selection of
donors from the MSM community, there is a need for
publicity and education to enhance public confidence on
blood safety and ensuring good communication with
donors to encourage self-deferral. Our results suggested
that females were more likely than males in having con-
cern of the new policy’s inability to reduce the number of
HIV positive donors. One possible explanation may be
that females were less aware of the policy change com-
pared to their counterparts and that may have potentially
led to misconception. Overall, the proportion of respon-
dents holding positive attitudes outnumbered those hold-
ing negative attitudes towards the policy change.
However, we also question whether attitudes might have
shifted over time as indicated by our data, with a slight
reduction in the proportion of those expressing support,
and an increase of those holding neutral attitude after
implementation. The shift of the proportion of positive
attitudes in Round B and C might be due to the smaller
number of young donors, as younger age was associated
with positive attitude towards the policy. As ‘inability to
reduce the number of HIV positive donors’ emerged as
the major concern after implementation of policy change,
some donors remained unassured if blood safety could be
maintained or improved. It is important to explain to the
wider community that the policy change represented only
© 2020 International Society of Blood Transfusion
Vox Sanguinis (2021) 116, 504–512

one aspect of ensuring blood safety, amidst all procedures
including state-of-art donor screening.
In countries like Japan and South Africa, the intervals
for deferring MSM has further been shortened to
6 months [5], and in the UK decreased to 3 months with
results suggesting no additional risk to blood safety [11].
The adoption of a shorter deferral period could be seen as
a realistic direction now that universal application of test-
ing methodology like nucleic acid testing (NAT) has nar-
rowed the window period to a week or less, which has
minimized recipients’ risks of acquiring transfusion-trans-
mitted HIV infection [24–25]. In our study, the preferred
interval of deferral was 1 year. One possible explanation
could be the donors’ lack of awareness of window period
and thus higher acceptance of the length of 1-year for
deferral to protect blood safety. The correlation between
the length of the perceived window period and preferred
deferral interval was however weak. Further investigation
is needed to determine whether the public’s awareness of
the length of the window period would improve the
acceptability of further shortening the deferral period as
regards male-to-male sex. Separately, reduction of defer-
ral period plus improved compliance would be beneficial
in expanding donor pool on top of ensuring safety of
blood transfusion. It is important to conduct further
review of deferral criteria while taking into account the
updated epidemiological evidence both at the global and
local level in the refinement of deferral policy.
Limitations
This study has some limitations. Firstly, the study was
based on a survey of blood donors who were selected
based on their availability and time-location sampling.
Donors recruited by BTS’s mobile services which visit other
locations in the territory have not been included. While
the recruitment of donors was done targeting an estimated
proportion of attenders by centre, the ultimate distribution
remained highly variable by age, employment status and
education level. Generalizability of the results to the entire
blood donor populations in Hong Kong should be cau-
tioned. The lack of standardization by sample weighting
may also explain the differences in attitudes, such as dif-
ferent compliance response proportions, across the three
rounds of the survey. Secondly, all data collected were
self-reported that could have potentially carried socially
desirable bias. The use of tablet computer, however, should
have minimized such bias by the sense of privacy and
confidentiality provided. Donors might feel easier to
answer truthfully when asked questions related to sensitive
topics such as sexual behaviours and illicit drug injection,
as the staff of the donor centres had no access to the com-
pleted questionnaire. Finally, the number of MSM donors
© 2020 International Society of Blood Transfusion
Vox Sanguinis (2021) 116, 504–512
Transitioning to 1-year deferral of MSM donors 511
tested HIV positive per year was small, and there was a
declining trend in the last two years, the phenomenon of
which was probably unrelated to the implementation of
the new donor deferral policy. These small numbers (4 in
2017, 2 in 2018 and 0 in 2019) did not allow us to draw
correlation with the donor deferral policy and its changes.
Tracking the number of HIV positive MSM donors could
nevertheless contribute to evaluate and inform the impact
of policy changes for achieving blood safety.
Conclusion
This study showed a consistently low non-compliance
rate among MSM at <1% both in the pre-implementation
and 2 post-implementation periods of the new time-lim-
ited deferral policy. The survey results showed that the
adopted questions appeared to differentiate non-compli-
ant and compliant donors per the 1-year deferral policy.
The generally high level of acceptance of time-limited
deferral among donors lends support to science-based
policy development to protect blood safety. Communica-
tion of the deferral criteria and its rationale is also related
to public acceptability of the change of deferral policy.
Our study however showed a low level of policy aware-
ness among donors. Improved communication of the pol-
icy and its rationale would be important in increasing
public confidence and sustaining their acceptance towards
policy change. The identification of compliant donors
with history of male-to-male sex suggests that the 12-
month deferral is effective and acceptable to MSM. With
a deferral period far exceeding the window period, it is a
step towards a more equitable policy.
Acknowledgements
The authors thank the staff of the Hong Kong Red Cross
Blood Transfusion Service for supporting the conduction
of the survey. This work was supported by a grant from
the AIDS Trust Fund (ATF MSS261R). Li Ka Shing Insti-
tute of Health Sciences, The Chinese University of Hong
Kong, is acknowledged for rendering technical support in
conducting the studies.
Author contributions
S-SL: conceptualizing the study. JYCL and C-PC: acquisi-
tion and analysis of data. JYCL: drafting the article. All
authors: revising it critically and giving final approval of
the version to be published.
Data Sharing and Data Accessibility
All data generated or analysed during this study are
included in this published article.
512 J. Y. C. Lau et al.
References
1 World Health Organisation (WHO):
Blood safety and availability. Avail-
able from: https://www.who.int/news-
room/fact-sheets/detail/blood-safety-
and-availability. [Last accessed 2
September 2019]
2 U.S. Food & Drug Administration
(FDA). FDA Statement. Coronavirus
(COVID-19) update: FDA provides
updated guidance to address the
urgent need for blood during the pan-
demic. Available from: https://
www.fda.gov/news-events/press-an
nouncements/coronavirus-covid-19-up
date-fda-provides-updated-guidance-
address-urgent-need-blood-during-pa
ndemic [Last accessed 9 June 2020]
3 Department of Health. Rules on blood
donation set to change. Available
from: https://www.health-ni.gov.uk/ne
ws/rules-blood-donation-set-change
[Last accessed 31 Aug 2020]
4 Goldman M, Shih AW-Y, O’Brien SF,
et al.: Donor deferral policies for men
who have sex with men: past, present
and future. Vox Sang 2018; 113:95–103.
5 Benjamin RJ, Bianco C, Goldman M,
et al.: Deferral of males who had sex
with other males. Vox Sang 2011;
101:339–367
6 Belanger GA, McFarland W, Raymond
HF, et al.: If the permanent deferral
were lifted would men who have sex
with men want to donate blood, and if
so, who would be eligible? Transfusion
2013; 53:2729–2733
7 Liszewski W, Terndrup C, Jackson NR,
et al.: The beliefs and willingness of
men who have sex with men to com-
ply with a one-year blood donation
deferral policy: a cross-sectional study.
Transfusion 2017; 57:2234–2239
8 Seed CR, Lucky TT, Waller D, et al.:
Compliance with the current 12-month
deferral for male-to-male sex in Aus-
tralia. Vox Sang 2014; 106:14–22
9 Sauvage C, Spinardi R, Pelat C, et al.:
Noncompliance with blood donor
selection criteria – Complidon 2017,
France. Transfusion 2020a; 60a:73–83.
10 O’Brien SF, Osmond L, Fan W, et al.:
Compliance with time-based deferrals
for men who have sex with men.
Transfusion 2019; 59:916–920
11 GOV.UK. Public Health England. Safe
supplies 2018: Monitor, inform pro-
gress. Available from: https://
www.gov.uk/government/publications/
safe-supplies-annual-review/safe-sup
plies-2018-monitor-inform-progress
[Last accessed 10 June 2020]
12 Grenfell P, Nutland W, McManus S,
et al.: Views and experiences of men
who have sex with men on the ban on
blood donation: a cross sectional sur-
vey with qualitative interviews. BMJ
2011; 343:d5604
13 Suligoi B, Pupella S, Regine V, et al.:
Changing blood donor screening crite-
ria from permanent deferral for men
who have sex with men to individual
sexual risk assessment: no evidence of
a significant impact on the human
immunodeficiency virus epidemic in
Italy. Blood Transfus 2013; 11:441–448
14 Lee SS, Cheung EKH, Leung JNS,
et al.: Non-compliance to infectious
disease deferral criteria among Hong
Kong’s blood donors. Vox Sang 2017;
112:425–433
15 Wong HTH, Lee SS, Lee CK, et al.:
Failure of self-disclosure of deferrable
risk behaviors associated with transfu-
sion-transmissible infections in blood
donors. Transfusion 2015; 55:2175–
2183
16 Lee CK, Lee KC, Lin CK, et al.: Donors’
perspectives on self-deferral of men
having sex with men from blood
donation. Transfusion 2013; 53:2441–
2448
17 Goldman M, Yi QL, Ye X, et al.: Donor
understanding and attitudes about
current and potential deferral criteria
for high-risk sexual behavior. Transfu-
sion 2011; 51:1829–1834
© 2020 International Society of Blood Transfusion
18 Davison KL, Conti S, Brailsford SR:
The risk of transfusion-transmitted
HIV from blood donations of men
who have sex with men, 12 months
after last sex with a man: 2005–2007
estimates from England and Wales.
Vox Sang 2013; 105:85–88
19 Germain M, Remis RS, Delage G: The
risks and benefits of accepting men
who have had sex with men as blood
donors. Transfusion 2003; 43:25–33
20 Seed CR, Kiely P, Law M, et al.: No
evidence of a significantly increased
risk of transfusion-transmitted human
immunodeficiency virus infection in
Australia subsequent to implementing
a 12-month deferral for men who
have had sex with men (CME). Trans-
fusion 2010; 50:2722–2730
21 Duquesnoy A, Danic B, Santos A,
et al.: Context and social perceptions
of blood donation in donors found
positive for human immunodeficiency
virus in France. Transfusion 2017;
57:2240–2247
22 Levy I, Olmer L, Livnat Y, et al.: Atti-
tudes, perceptions and knowledge
among men who have sex with men
towards the blood donation deferral
policy in Israel. PLoS One 2017; 12:
e0170364
23 Sanchez AM, Schreiber GB, Nass CC,
et al.: The impact of male-to-male
sexual experience on risk profiles of
blood donors. Transfusion 2005;
45:404–413
24 Roth WK, Busch MP, Schuller A,
et al.: International survey on NAT
testing of blood donations: expanding
implementation and yield from 1999
to 2009. Vox Sang 2012; 102:82–90
25 Kleinman S, Busch MP, Korelitz JJ,
et al.: The incidence/window period
model and its use to assess the risk of
transfusion-transmitted human
immunodeficiency virus and hepatitis
C virus infection. Transfus Med Rev
1997; 11:155–172
Vox Sanguinis (2021) 116, 504–512
 Vox Sanguinis (2021) 116, 513–523
© 2020 The Authors.
ORIGINAL PAPER Vox Sanguinis published by John Wiley & Sons Ltd on behalf of International Society of Blood Transfusion
DOI: 10.1111/vox.13029

Unknown, so also unvalued? Blood donation awareness and

attitudes of potential donors of Dutch and African descent
Elisabeth F. Klinkenberg,1,2 Mirjam P. Fransen,2 Wim L.A.M. de Kort,1,2 Elisabeth M.J. Huis in ’t Veld1,3 &
Julia C.M. van Weert4
1
Department of Donor Medicine Research, Sanquin Research, Amsterdam, The Netherlands
2
Department of Public Health, Amsterdam UMC, Amsterdam Public Health research institute, University of Amsterdam, Amsterdam, The Netherlands
3
Department of Cognitive Science & Artificial Intelligence, Tilburg University, Tilburg, The Netherlands
4
Amsterdam School of Communication Research / ASCoR, Department of Communication Science, University of Amsterdam, Amsterdam, The
Netherlands

Received: 19 June 2020,
revised 14 October 2020,
accepted 21 October 2020,
published online 7 November 2020
Abstract
Background and objectives Many Western countries face a shortage of African
blood donors, while their specific blood groups are needed to transfuse chronic
transfusion patients of similar ethnic background. Blood donation awareness and
attitudes greatly impact the decision to become a blood donor, but how they are
related and differ across ethnic groups is understudied. This study investigated
blood donation awareness and attitudes of individuals of Dutch and African des-
cent in the Netherlands.
Materials and methods Survey data of 257 African and 152 Dutch non-donors
measuring donation awareness (i.e. being familiar with the Dutch blood bank
organization and knowing others who donated blood), cognitive (evaluative
judgements) and affective (emotional reactions) attitudes were included. t-Tests,
chi-square tests, linear and logistic regressions were conducted to study differ-
ences and associations between donation awareness and attitudes.
Results African individuals were less often aware of the Dutch blood bank orga-
nization (43%; p < 0
05) or others who donated blood (51%; p < 0
05) than
Dutch individuals (55% and 68%, respectively). African individuals had lower
cognitive donation attitudes compared with Dutch individuals (p < 0
001), but no
differences were found for affective attitudes (p = 0
55). High donation awareness
was associated with higher cognitive (p < 0
001) and affective (p < 0
05) dona-
tion attitudes among African minorities, but not among Dutch individuals.
Conclusion The lower donation awareness and cognitive attitudes of African
minorities should be taken into consideration in donor recruitment. Raising
awareness through effective communication strategies might be essential in the
donor decision making process of this target group.
Key words: donor motivation, donor recruitment, donors.

Correspondence: Elisabeth F. Klinkenberg, PhD-student at Department
of Donor Medicine Research, Sanquin Research, Plesmanlaan 125, 1066
CX Amsterdam, The Netherlands
E-mail: l.klinkenberg@sanquin.nl
Introduction
Blood donors of African descent are greatly underrepre-
sented in many Western high-income countries [1]. This
presents an important healthcare problem as blood groups
are different across ethnic groups. Especially, people of
Sub-Saharan African/Black descent have extended

This is an open access article under the terms of the Creative Commons Attribution License, 513
which permits use, distribution and reproduction in any medium, provided the original work is properly cited.
514 E. F. Klinkenberg et al.
antigen-negative blood type variations almost non-exis-
tent in European/White populations [2]. Frequent blood
transfusions are important for treating patients with blood
disorders, including inheritable diseases such as sickle cell
anaemia and thalassaemia, with an estimated 300 thou-
sand affected newborns each year [3]. These kinds of dis-
eases are more prevalent in people from African descent
as compared to people of, for example, European ances-
try. It is essential to adequately match the donor and
recipient blood groups, especially in chronic transfusion
patients, as antibody formation (alloimmunization) can
complicate future transfusions, and future mismatches
can cause severe transfusion reactions [4,5]. Yet, preven-
tive antigen matching and preventing alloimmunization
is a significant challenge when the blood donor popula-
tion is non-diverse [4]. Thus, individuals of African des-
cent are underrepresented in blood donor pools, but are
not underrepresented as recipients for multiple blood
transfusions [6].
Consequently, an ethnically diverse blood donor pool
adequately mirroring the patient population is of great
importance. Although various blood collection agencies
have been focusing and adjusting campaigns to recruit
and retain more African blood donors [6], in the Nether-
lands recruitment efforts of this group in particular have
been small scale, local and incidental, for example at a
few cultural events, mosques and universities. As the
Netherlands has an estimated 500 000 residents with a
sub-Saharan heritage [7], which is growing because of
migration and childbirth, the Dutch national blood collec-
tion agency (Sanquin) now also seeks blood donors of
African descent to better match with the patient popula-
tion [8]. Although the actual donation rates are unknown,
as ethnicity is not registered, the prevalence of the
extended blood groups in the donor pool suggests there is
an underrepresentation [9].
Past studies examining donation barriers and motiva-
tors suggest a lower donation awareness and more nega-
tive donation attitudes towards donating blood in
potential African donors [1,10]. For example, Lemmens,
Abraham [11] found a significant positive correlation
between knowledge and affective and cognitive attitude.
However, knowledge and awareness are related but not
the same concepts [12]. As past research mainly gave
descriptive findings of self-reported barriers and motiva-
tors and often in one group, there is little evidence in the
blood donation context on how awareness and attitudes
relate and differ for ethnic groups. Our aim was to focus
more in depth on this topic by studying the association
with and differences between blood donation awareness
and blood donation attitudes for people of Dutch and
African descent who had no former blood donation expe-
riences in the Netherlands.
Vox Sanguinis published by John Wiley & Sons Ltd on behalf of International Society of Blood Transfusion
Theoretical background
Unawareness has been consistently identified as a main
blood donation barrier for non-donors [13]. Unawareness
of blood donation can be defined as unconsciousness,
which can be caused by never having heard of or com-
municated about blood donation [14,15] or never having
taken in and remembered information about it [16]. It
can be expressed in practical unfamiliarity such as not
knowing the name or the location of the blood bank
organization or not knowing individuals who have
donated blood and not having talked about this subject
[14,17].
Awareness is an important factor in the transtheoretical
model [18]. This model comprises of five stages of
change: Precontemplation, Contemplation, Preparation,
Action and Maintenance. The Precontemplation stage is
characterized by unawareness, and individuals within this
stage do not plan to change behaviour in the near future
because they have not yet contemplated about it [19].
African minorities and migrants generally seem to be less
aware of blood donation than the White majority group
[1]. For instance, Burditt, Robbins [19] found that a
majority of the African American participants were in the
Precontemplation stage [18], indicating a low awareness.
Unawareness of blood donation has been reported as a
major reason why African minority groups are underrep-
resented as blood donors, while on the other hand these
groups seem relatively more motivated by awareness rais-
ing and donation requests than White majority groups
[1]. In an interview study among Ghanaians and African-
Surinamese people in the Netherlands, only 11% of the
participants recognized the name or logo of the Dutch
blood bank organization: ‘Sanquin’ [10]. Participants had
not actively thought of becoming a blood donor, also
because they had never been approached about it. These
findings correspond with a study in Australia, where Afri-
can participants reported that they had never discussed
blood donation before the study [14].
Besides a practical unfamiliarity of blood donation (i.e.
the name of the blood bank, what donating blood entails),
unawareness can also be expressed in cognitive and
affective unfamiliarity, such as not being aware of the
importance of donating blood or overestimating the dis-
advantages of donating blood [13,20].
Attitudes are described in the Theory of Planned Beha-
viour as how positive or negative a specific behaviour is
evaluated by individuals and can be divided into two sub
components; cognitive attitude and affective attitude. The
first comprises the evaluative judgements to the beha-
viour, while the later comprises emotional reactions to
the behaviour [21,22]. Attitudes have been found to be
important determinants of intention and, respectively,
© 2020 The Authors.
Vox Sanguinis (2021) 116, 513–523

behaviour [22–25] and seem to play an important part
particularly in the Precontemplation stage in behaviour
change [18,26]. There is evidence that African minorities
have a slightly more negative attitude towards blood
donation compared to White majority populations in
Western countries [1]. Various studies in the United States
showed that African Americans less often believe it is
important to donate blood (e.g. because of the perception
that their blood is not wanted or will be wasted) and that
they are generally more afraid of needles and pain as
compared to White Americans [1,27,28]. In our qualita-
tive interview study among African migrants living in the
Netherlands, the overall general opinion towards giving
blood was positive, but participants reacted more hesi-
tantly when asked about their own attitude to donate
blood [10].
It is expected that it could be beneficial to design
recruitment campaigns and their target group with the
knowledge from the TTM and TPB in mind. Based on
these models and the literature, we expect that an assess-
ment of the level of and interplay between awareness and
attitudes in our potential donor recruitment target group
will benefit the development of future donor recruitment
strategies. Partly due to being unaware, individuals with
low awareness or in the Precontemplation stage may
experience resistance to behaviour change or believe the
costs of behaviour change are higher than the benefits
[26]. Therefore, when the target audience becomes well
informed, the gained awareness might alleviate cognitive
and affective unfamiliarity surrounding blood donation
and act as a prerequisite for positive attitude change and
a move to the next step in the TTM.
Based on the studies and mechanisms discussed above,
the following hypotheses were developed:
H1: African ethnic minorities have a lower blood dona-
tion awareness compared with people of Dutch descent in
the Netherlands.
H2: African ethnic minorities have a lower blood dona-
tion attitude compared with people of Dutch descent in
the Netherlands.
H3: A higher blood donation awareness is associated
with higher blood donation attitudes for African ethnic
minorities and people of Dutch descent.
Methods
Design and participants
For this study, we made use of various survey data. The
Motivations to Give, Donate and Share study (Motive-
study; https://www.sanquin.org/research/donor-studies-
projects/motive-study/index) consists of stratified online
survey data collected in 2018 among people of African
© 2020 The Authors.
Vox Sanguinis published by John Wiley & Sons Ltd on behalf of International Society of Blood Transfusion
Vox Sanguinis (2021) 116, 513–523
Unknown makes unloved in blood donation 515
background and university students, residing in the
Netherlands [29,30]. Participants in the Motive-study
were recruited among the general population through an
ISO-certified research company (Panelclix) and among
students through Tilburg University. Participants recruited
through Panelclix were compensated with ‘Clix’, which
can be traded for a small amount of money. Participants
recruited through the university were compensated with
course credits. Next to data originating from the Motive-
study, additional data were also collected in 2018 among
social media users (regardless of ethnic background) using
a shortened version of the survey used in the Motive-
study. An overview of the recruitment of the samples can
found in Table 1. Adequate Dutch or English language
proficiency was a requirement for participation. Social
media users were recruited predominantly via Facebook
and WhatsApp, using convenience- and snowball-sam-
pling. The surveys were shared on the social media
account of the Dutch blood bank organization Sanquin
and also via Promoted Posts to reach social media users
who do not ‘like’ or follow this Facebook page. Partici-
pants recruited through social media were compensated
with a voucher for a large online store. To our knowl-
edge, no recruitment campaigns aimed at the target group
that could interfere with the results were rolled out at the
times of data collection.
For the analyses in the present study, we selected par-
ticipants between the ages of 18 and 65, being the age
range individuals can register as a blood donor in the
Netherlands. We only included participants who had
never donated blood but were not definitely excluded for
blood donation and who were either of Dutch or African
ethnic background. In total of the initial 672 participants
combined, 3 (0
4%) were excluded for being either too
young or too old for this study’s purpose, 65 (9
7%) were
excluded for being of different ethnic background than
African or Dutch, 133 (19
8%) for having donated blood
previously, 23 (3
4%) for being definitely excluded to
donate blood and 39 (5
8%) because of missing values on
at least one of the variables included in the analyses of
this study. The final sample included 257 people with a
Sub-Saharan, Afro-Surinamese or Afro-Caribbean (Afri-
can) background (63%) and 152 people of Dutch back-
ground (37%).
Ethical approval for the study was granted by the Ethi-
cal Advisory Board of Sanquin and the Ethics Committee
of Tilburg University. The project protocol this study is
part of was also reviewed by the Medical Ethics Review
Committee of the Academic Medical Center Amsterdam
(now Amsterdam UMC – location AMC), but was waived
from requiring medical ethical approval because the pro-
tocol did not fall under the Medical Research Involving
Human Subjects Act of the Dutch law.
516 E. F. Klinkenberg et al.

Social media users

18 years and older

Additional sample

March–April 2018
Measures

Online, Qualtrics

All ethnicities

Social media
Ethnicity and background characteristics

n = 231
African background in this study refers to individuals of

10 min

n = 97
Dutch

n=2
sub-Saharan descent, which is defined as individuals

Yes
who reported they had at least one parent originating
from sub-Sahara Africa, Surinam or the Caribbean and
who indicated they were of sub-Saharan, African-Suri-
namese or African-Caribbean descent. Dutch background

Psychology lab Tilburg University

was defined as individuals who reported that both par-
ents are Dutch and who were born in the Netherlands.
Additionally, individuals who were born abroad and
University students

whose parents were born abroad were defined as first-

Online, Qualtrics

Dutch & English

April–May 2018

generation migrants, and individuals who were born in

All ethnicities
18–65 years

the Netherlands with at least one parent who was born

n = 141
30 min

n = 64

abroad were defined as second-generation migrants [31].

n=7
Yes

Sociodemographic background variables were age, gen-

der and educational level. Educational level was mea-
sured using the International Standard Classification of
Education (ISCED) adjusted version to the Dutch educa-
tional system of Statistics Netherlands [32,33]. The seven
Only Sub-Saharan African, African-Surinamese and African-Caribbean

answer categories were then divided in low (no educa-

tion till lower secondary education), middle (upper sec-
ondary and post-secondary non-tertiary education) and
high (Bachelor’s degree and higher) educational levels.

Donation awareness
Donation awareness was measured with two separate
questions. ‘Have you ever heard of the Dutch blood bank
organization, Sanquin?’ and ‘Do you personally know
someone who has donated blood or is currently a blood
donor?’ with ‘yes’ (a score of 1) and ‘no’ (a score of 0)
Panelclix and social media

as the possible answer categories (r = 0
30, p < 0
001).

April–September 2018
African background

Both questions were combined to assess the participants’

Online, Qualtrics

Dutch & English

level of awareness with a score of 0 ‘no awareness’, 1

Motive-sample

18–65 years

‘low awareness’ (knows either the blood bank or a blood

n = 300
n = 248
n = 248

donor) or 2 ‘high awareness’ (knows both the blood bank

30 min
Table 1 Overview of survey data incorporated in current study

and a blood donor). The first awareness question had the

Yes

follow-up question ‘How have you heard of Sanquin?

You can select multiple options’ with various peers and
media as sources (i.e. through family members, through
social media) as answer categories. The second aware-
11. Included participants of African descent

ness question was followed up by the question ‘Who has

10. Participants included in current study

donated blood or is currently a blood donor that you

know of? You can select multiple persons.’ with various
9. Total number of participants
6. Residing in the Netherlands

peers presented as options (i.e. colleagues/fellow stu-

5. Eligible ethnic background
2. Time to complete survey

dents, relatives). Both follow-up questions had an ‘other,

1. Data collection period

namely . . .’ option.
4. Eligible age range
3. Type of survey
Type of sample

Donation attitudes
7. Recruitment
8. Language

Donation attitudes were measured using six questions

with a 7-point bipolar statements, based on the measures
Study

validated by France, Kowalsky [26]. The time frame was

© 2020 The Authors.

Vox Sanguinis published by John Wiley & Sons Ltd on behalf of International Society of Blood Transfusion
Vox Sanguinis (2021) 116, 513–523

adjusted from eight weeks in the original scale to
12 months, as the blood donation procedure in the
Netherlands is more lengthy compared to donating at the
blood drives in the United States (e.g. you cannot donate
at your first visit) and a blood donor in the Netherlands
donates on average only once or twice annually [29]. A
factor analysis verified that donation attitudes could be
divided into cognitive attitude (evaluative judgements
towards blood donation) (a = 0
915) and affective attitude
(emotional reactions towards blood donation) (a = 0
863),
which both accounted for 75
0% of the total variance in
the six items. When analysed separately for participants
of African and Dutch background, exactly the same two
factors arose and the Cronbach alphas were still good
(cognitive attitude: a = 0
927 African background,
a = 0
852 Dutch background; affective attitude:
a = 0
888 African background, a = 0
821 Dutch back-
ground). Both cognitive attitude (useless/useful, pointless/
worthwhile and the wrong thing to do/the right thing to
do) and affective attitude (unpleasant/pleasant, unenjoy-
able/enjoyable, frightening/not frightening) are con-
structed from the mean of three 7-point bipolar
statements. On both scales, a score of 0 refers to the low-
est possible attitude, while a score of 6 refers to the high-
est possible attitude.
Analyses
Statistical software (SPSS, version 23, Chicago, IL) was
used to examine the descriptive properties of the data
and test our hypotheses. t-Tests and chi-square tests were
performed to test differences in demographic characteris-
tics between the two ethnic groups. The differences in
donation awareness and donation attitude between partic-
ipants of African and Dutch ethnic background were
tested using logistic and linear regression analyses and
controlled for age, gender and educational level to find
answers for hypotheses H1 and H2. Multivariate linear
regression analyses tested the associations of donation
awareness on cognitive and affective attitude. We made
two separate (stratified) models for those of African and
Dutch descent (H3). These sub-group analyses were con-
trolled for age, gender and educational level, plus migrant
generation in the African sub-sample.
Results
Background characteristics of samples and Dutch
and African individuals
Table 2 shows the sociodemographic background charac-
teristics of the 152 participants of Dutch and the 257 par-
ticipants of African descent. There were more women
© 2020 The Authors.
Vox Sanguinis published by John Wiley & Sons Ltd on behalf of International Society of Blood Transfusion
Vox Sanguinis (2021) 116, 513–523
Unknown makes unloved in blood donation 517
than men in both groups, namely 63% (n = 96) in the
Dutch group and 72% (n = 184) in the African group, but
the difference between the two ethnic groups was not sig-
nificant (X2(1) = 3
1, p = 0
08). The overall age was rela-
tively lower in the Dutch background group (M = 25
7,
SD = 10
5) compared with the African group (M = 35
9,
SD = 12
5; t (361
8) = -8
9, p < 0
001). Also, the Dutch
group was significantly higher educated compared with
the African group (X2(2) = 35
8, p < 0
001). Among the
African individuals, 48% (n = 124) were first-generation
migrants and 52% (n = 133) were second-generation
migrants (see Appendix A for the demographic compar-
ison of the three study samples).
Differences in donation awareness and donation
attitudes
Slightly more than half of the Dutch individuals knew
Sanquin (55%; n = 83) (Table 3). This percentage was sig-
nificantly lower among the African individuals (43%;
n = 111) (adjusted odds ratio; AOR = 0
63, 95% CI [0
40,
0
99], p < 0
05). Of those participants who were familiar
with Sanquin (n = 194), the Dutch participants (n = 83)
most often heard of it through family members (39%;
n = 32) and the African participants (n = 111) most often
heard of it through friends (23%; n = 26) (Figure 1). A
relatively large part of the African individuals also
became familiar with the blood bank through work and/
or colleagues (20%; n = 22). In the ‘Other’ category, indi-
viduals of Dutch background most often mentioned that
knowing Sanquin is part of their general knowledge,
whereas individuals of African background most often
mentioned they knew Sanquin via the Internet. More than
two-thirds of the Dutch individuals knew someone per-
sonally who has donated blood (68%; n = 104), whereas
about half of the African individuals mentioned this
(51%; n = 131) (AOR = 0
56, 95% CI [0
35, 0
90],
p <
05). Of those participants who personally knew one
or more other blood donors (n = 235), this was in most
cases a friend or acquaintance (63% (n = 66) of the Dutch
individuals and 62% (n = 81) of the African individuals)
(Figure 1). In the ‘Other’ category, individuals of African
background mentioned for instance parents in law, the
partner and a neighbour. When combining the two dona-
tion awareness measures, we notice that relatively most
individuals of Dutch background belong to the ‘high
awareness’ category (45%, n = 70), while relatively least
individuals of African background fit in this category
(28%, n = 72). H1 can be accepted as less African indi-
viduals were aware than Dutch individuals.
Dutch individuals showed an overall higher cognitive
attitude (M = 4
7, SD = 1
2) compared with African indi-
viduals (M = 3
9, SD = 1
7). After controlling for age,
518 E. F. Klinkenberg et al.

Table 2 Background characteristics for Dutch and African individuals (N = 409)

Dutch background (n = 152) African background (n = 257) Significance testing

Characteristic n (%)/M (SD) n (%)/M (SD) t (df)/X2 (df)

Gender X2(1) = 3
1

Male 56 (37%) 73 (28%)
Female 96 (63%) 184 (72%)
Age 25
7 (10
5) 35
9 (12
5) t (361
8) = -8
9***
Educational level X2(2) = 35
8***
Low 6 (4%) 27 (11%)
Middle 37 (24%) 124 (48%)
High 109 (72%) 106 (41%)
Migrant generation
First n.a. 124 (48%)
Second n.a. 133 (52%)

***p < 0
001.

gender and educational level, this difference remained
significant (B = -0
90, t(403) = -5
17, p < 0
001).
Regarding affective attitude, we found no significant dif-
ferences between the Dutch (M = 2
7, SD = 1
4) and Afri-
can individuals (M = 2
9, SD = 1
5; B = 0
10, t
(403) = 0
59, p = 0
55). Thus, H2 could only be partly
accepted as there was a significant difference in cognitive
attitude, but not in affective attitude.
The associations between donation awareness on
donation attitudes
Table 4 summarizes the results for the multivariate linear
regression analyses on donation attitudes. Among the
African individuals, high awareness was significantly
associated with a higher cognitive attitude (B = 0
98, t
(249) = 3
61, p < 0
001), whereas this was not the case
for Dutch individuals who showed no significant associa-
tion between awareness and cognitive attitude (B = 0
36,
t(145) = 1
49, p = 0
14).
Regarding the models for affective attitude, scoring
high on awareness was significantly associated with a
higher affective attitude among the African background
group (B = 0
56, t(249) = 2
23, p < 0
05), but not in the
Dutch background group (B = 0
17, t(145) = 1
10,
p =
59).
Based on our results, H3 was accepted for the African
individuals as a higher awareness was associated with
higher cognitive and affective attitudes, but rejected for
the Dutch individuals.

Table 3 Differences in awareness and attitudes between Dutch and African individuals (N = 409)

Dutch background (n = 152) African background (n = 257) B/AOR

Aware blood bank n (%)/M (SD) n (%)/M (SD) AOR = 0
63*

Yes 83 (55%) 111 (43%)
No 69 (45%) 146 (57%)
Aware other donor AOR = 0
56*
Yes 104 (68%) 131 (51%)
No 48 (32%) 126 (49%)
Awareness combined
Not aware 34 (22%) 88 (35%)
Low awareness 51 (33%) 94 (37%)
High awareness 70 (45%) 72 (28%)
Cognitive attitude 4
7 (1
2) 3
9 (1
7) B = -0
90***
Affective attitude 2
7 (1
4) 2
9 (1
5) B = 0
10

All tests are adjusted for gender, age and educational level. Dutch individuals were the reference group.
*p < 0
05.
***p < 0
001.

© 2020 The Authors.

Vox Sanguinis published by John Wiley & Sons Ltd on behalf of International Society of Blood Transfusion
Vox Sanguinis (2021) 116, 513–523

Figure 1 Specified blood bank awareness (n = 194) and knowing another donor (n = 235), divided by ethnic background.
Discussion
In this study, we hypothesized that African migrants have
a lower blood donation awareness and a more negative
attitude towards donating blood than Dutch individuals
and that a higher donation awareness is associated with a
higher donation attitude. In line with our first hypothesis
and the literature, we found that less African individuals
are aware than the Dutch individuals. However, whereas
only 11% of African migrants was aware of the Dutch
blood bank organization in our previous, qualitative
study sample [10], awareness in this current sample was
much higher at 43%. This might be due to a difference in
the composition and recruitment strategies of the study
samples. For example, in the present study sample a
much larger part of the participants are second-genera-
tion migrants. Additionally, they were recruited online
and the participants with an African background most
often mentioned they knew Sanquin via the Internet.
Similar to previous studies and partly in concordance
with our second hypothesis, participants with an African
background had a lower cognitive, but not a lower affec-
tive, attitude than Dutch individuals. This is partly in line
with previous studies, African migrants and minorities
generally report a lower attitude towards blood donation
than the majority population [10,14]. No other studies
that we are aware of studied cognitive and affective blood
donation attitudes separately for different ethnic groups.
© 2020 The Authors.
Vox Sanguinis published by John Wiley & Sons Ltd on behalf of International Society of Blood Transfusion
Vox Sanguinis (2021) 116, 513–523
Unknown makes unloved in blood donation 519
Lastly, we hypothesized that awareness and attitudes
would be positively correlated for all individuals. Interest-
ingly, we only found a positive association between
awareness and attitudes in the African group. Perhaps
other deterrents or motivators, such as self-efficacy or
(in)convenience, might be more impactful for donation
attitudes in this group [34,35].
As this study highlights that more people with an Afri-
can background may be in the ‘precontemplation stage’
of the TTM than people with a Dutch background, it
could be beneficial for Sanquin to design future recruit-
ment campaigns and target strategies such that they can
help African individuals to move to the contemplation or
even preparation stage [18,23,24]. As awareness seems to
increase attitudes, this might eventually lead to behaviour
change according to the TPB [23].
The current study adds to the literature regarding this
topic with insights from a European context, which dif-
fers historically, culturally and socio-economically from
the United States or Australia, where most research is
conducted [1]. However, a limitation of this study is the
recruitment of convenience samples of through a panel, a
university and social media. As the two ethnic groups are
unequally represented in these contexts, the Dutch sample
is of younger age and more highly educated, which also
makes it less generalizable to the Dutch general popula-
tion. Another consideration is that the surveys were
administered in English and Dutch only and also
 520 E. F. Klinkenberg et al.

Table 4 Multivariate linear regressions on cognitive and affective attitude for individuals of Dutch (n = 152) and African background (n = 257)

Cognitive attitude Affective attitude

Dutch African Dutch African

Variables B (SE) t B(SE) t B(SE) t B(SE) t

Intercept 4
57 (0
56) 8
14 3
77 (0
54) 7
02 3
14 (0
72) 4
35 3
03(0
50) 6
12
Blood donation awareness
Not aware Ref. Ref. Ref. Ref.
Low awareness (knows either the blood bank or blood donor) 0
31 (0
25) 1
25 0
28 (0
25) 1
11 0
36 (0
32) 1
10 0
08(0
23) 0
33
High awareness (knows both the blood bank and blood donor) 0
36 (0
24) 1
49 0
98 (0
27)*** 3
61 0
17 (0
31) 0
54 0
56(0
25)* 2
23
Female 0
65 (0
19)*** 3
43 -0
01 (0
24) -0
03 -0
02 (0
24) -0
08 -0
02(0
22) -0
07
Age in years 0
01 (0
01) 0
79 -0
01 (0
01) -0
74 0
004 (0
01) 0
32 0
01(0
01) 0
84
Second migrant generation n.a. -0
11 (0
23) -0
48 n.a. -0
16(0
21) -0
76
Educational level
Low Ref. Ref. Ref. Ref.
Middle -0
45 (0
50) -0
90 -0
02 (0
35) -0
06 -0
81 (0
64) -1
27 -0
55(0
33) -1
69
High -0
85 (0
48) -1
79 0
03 (0
37) 0
07 -0
78 (0
61) -1
26 -0
60(0
34) -1
79

Bold values are statistically significant.

*p < 0
05.
***p < 0
001.

© 2020 The Authors.

Vox Sanguinis (2021) 116, 513–523

Vox Sanguinis published by John Wiley & Sons Ltd on behalf of International Society of Blood Transfusion
 Unknown makes unloved in blood donation 521

distributed online, which can be an obstacle for first-gen-
eration African migrants. We made this choice because
Dutch or English fluency is required to donate blood in
the Netherlands. Additionally, the blood donor registra-
tion process is also online, thus basic Internet competen-
cies are essential for blood donors. The current sample
might not be generalizable to all African migrants living
in the Netherlands, but is generalizable to the potential
donor population of African descent.
The results of this study have valuable implications for
future research and blood donor recruitment. Awareness
raising might be a promising first step to help African
minorities contemplating about blood donation. In the
Netherlands, Sanquin is the only blood bank organiza-
tion. If a person does not recognize the name or logo, he
or she will not notice a blood collection centre nearby
and it also makes it difficult to register as a donor, as the
person does not know where to do so. We did find that a
higher level of awareness – both being familiar with the
blood bank organization and personally knowing a blood
donor – to have a positive association with attitudes. The
importance of a social network needs to be highlighted in
this, as most individuals became familiar of blood dona-
tion through family members and friends, as also was
emphasized in other studies [36,37]. Since ethnic minori-
ties often have less blood donors in their social network,
it is of importance to reach the target communities first
as to enable blood donation awareness to spread through
their social networks [9,10,37]. This is essential informa-
tion for the development of future blood donor recruit-
ment interventions.
Acknowledgements
This work was support by Sanquin Research under the
internal grant: PPOC-14-25. We thank all participants for
their contribution to this study. JvW and EK conceived
the initial idea of the research. EK, MF and EHitV
designed the survey. EK led the survey data collection,
conducted the analyses and wrote the manuscript, while
MF, WdK, EHitV and JvW critically reviewed and revised
the manuscript.
Conflict of interest
The authors declare that they have no competing inter-
ests.
Data availability statement
Data of the Motive-study or the particular data set used
for this study are available upon request by contacting
Elisabeth Klinkenberg (L.Klinkenberg@Sanquin.nl). More
information on the Motive-study can be found here:
https://www.sanquin.org/research/donor-studies-projec
ts/motive-study/index

References
1 Klinkenberg EF, Huis In ‘t Veld EMJ,
De Wit PD, et al.: Blood donation bar-
riers and facilitators of Sub-Saharan
African migrants and minorities in
Western high-income countries: a sys-
tematic review of the literature. Trans-
fusion Med 2019; 29(Suppl 1):28–41
2 Reid ME, Lomas-Francis C, Olsson ML:
The blood group antigen facts book,
3rd edn. San Diego (CA), Academic
Press, 2002
3 Weatherall D: The inherited diseases of
hemoglobin are an emerging global
health burden. Blood 2010; 115:4331–
6 publication/?
4 Zalpuri S, Zwaginga JJ, le Cessie S,
et al.: Red-blood-cell alloimmuniza-
tion and number of red-blood-cell
transfusions. Vox Sang 2012;
102:144–9
5 Yazdanbakhsh K, Ware RE, Noizat-Pir-
enne F: Red blood cell
alloimmunization in sickle cell disease:
pathophysiology, risk factors, and
transfusion management. Blood 2012;
120:528–37
6 Van Dongen A, Mews M, de Kort
WLA, et al.: Missing Minorities – A
survey based description of the current
state of minority blood donor recruit-
ment across 23 countries. Divers Equal
Health Care 2016; 13:138–45.
7 Statistics Netherlands: CBS Statline -
Bevolking; generatie, geslacht, leeftijd
en herkomstgroepering, 1 januari.
2016. http://statline.cbs.nl/Statweb/
DM=SLNL&PA=37325&D1=0&-
D2=a&D3=0&D4=0&D5=0-5,65,75-
76,163,251,253&D6=a&HDR=G5,T,G3,
G2,G4&STB=G1&VW=T (Last accessed
19-06-2020).
8 Van Sambeeck JHJ, De Wit PD, Luken
J, et al.: A conceptual framework for
optimizing blood matching strategies:
balancing patient complications
against total costs incurred. Front Med
2018; 5:199
9 Piersma TW, Klinkenberg EF: The rela-
tion between blood donor recruitment
and donor diversity and loyalty in the
Netherlands. ISBT Sci Series 2018;
13:384–93
10 Klinkenberg EF, Huis In ‘t Veld EMJ,
De Wit PD,, et al.: Barriers and moti-
vators of Ghanaian and African-Suri-
namese migrants to donate blood.
Health Soc Car Community 2019;
27:748–56
11 Lemmens KP, Abraham C, Ruiter RA,
et al.: Modelling antecedents of blood
donation motivation among non-
donors of varying age and education.
Brit J Psychol 2009; 100:71–90
12 Trevethan R: Deconstructing and
assessing knowledge and awareness in

© 2020 The Authors.

Vox Sanguinis published by John Wiley & Sons Ltd on behalf of International Society of Blood Transfusion
Vox Sanguinis (2021) 116, 513–523
522 E. F. Klinkenberg et al.
public health research. Front Public
Health 2017; 5:194
13 Gillespie TW, Hillyer CD: Blood donors
and factors impacting the blood dona-
tion decision. Transfus Med Rev 2002;
16:115–30
14 Polonsky MJ, Renzaho AM, Brijnath
B: Barriers to blood donation in Afri-
can communities in Australia: the role
of home and host country culture and
experience. Transfusion 2011;
51:1809–19
15 Lemmens KP, Abraham C, Hoekstra T,
et al.: Why don’t young people volun-
teer to give blood? An investigation of
the correlates of donation intentions
among young nondonors. Transfusion
2005; 45:945–55
16 Flora JA, Maibach EW, Maccoby N:
The role of media across four levels of
health promotion intervention. Annu
Rev Public Health 1989; 10:181–201
17 James AB, Schreiber GB, Hillyer CD,
et al.: Blood donations motivators and
barriers: a descriptive study of African
American and white voters. Transfus
Apher Sci 2013; 48:87–93
18 Prochaska JO, DiClemente CC: The
transtheoretical approach; in Norcross
JC, Goldfried MR (eds): Handbook of
Psychotherapy Integration (2nd ed).
New York, Oxford University Press,
2005:141-71.
19 Burditt C, Robbins ML, Paiva A, et al.:
Motivation for blood donation among
African Americans: developing mea-
sures for stage of change, decisional
balance, and self-efficacy constructs. J
Behav Med 2009; 32:429–42
20 McVittie C, Harris L, Tiliopoulos N: "I
intend to donate but . . .": non-donors’
views of blood donation in the UK.
Psychol Health Med 2006; 11:1–6
Vox Sanguinis published by John Wiley & Sons Ltd on behalf of International Society of Blood Transfusion
21 Breckler SJ: Empirical validation of
affect, behavior, and cognition as dis-
tinct components of attitude. J Person
Soc Psychol 1984; 47:1191–205
22 France JL, Kowalsky JM, France CR,
et al.: Development of common met-
rics for donation attitude, subjective
norm, perceived behavioral control,
and intention for the blood donation
context. Transfusion 2014; 54:839–47
23 Ajzen I: The theory of planned behav-
ior. Organ Behav Hum Dec 1991;
50:179-211.
24 De Vries H, Dijkstra M, Kuhlman P:
Self-efficacy: the third factor besides
attitude and subjective norm as a pre-
dictor of behavioural intentions.
Health Educ Res 1988; 3:273–82
25 Conner M, Godin G, Sheeran P, et al.:
Some feelings are more important:
cognitive attitudes, affective attitudes,
anticipated affect, and blood donation.
Health Psychol 2013; 32:264–72
26 De Vries H, Mudde AN, Dijkstra A,
et al.: Differential beliefs, perceived
social influences, and self-efficacy
expectations among smokers in vari-
ous motivational phases. Prev Med
1998; 27:681–9
27 Schreiber GB, Schlumpf KS, Glynn SA,
et al.: Convenience, the bane of our
existence, and other barriers to donat-
ing. Transfusion 2006; 46:545–53
28 Mathew SM, King MR, Glynn SA, et al.:
Opinions about donating blood among
those who never gave and those who
stopped: a focus group assessment.
Transfusion 2007; 47:729–35
29 Sanquin.org: Motivations to Give,
Donate and Share (Motive-study).
2019. https://www.sanquin.org/researc
h/donor-studies-projects/motive-study/
index (Last accessed 19-06-2020).
30 Klinkenberg EF, Fransen MP, De Kort
WLAM, et al.: Blood donation among
individuals of African origin in the
Netherlands: How are barriers and
facilitators associated with intention?
Blood Transfus-Italy 2020. Advance
online publication.
31 Statistics Netherlands: Wat verstaat
het CBS onder een allochtoon?, n.d.
https://www.cbs.nl/nl-nl/faq/specifiek/
wat-verstaat-het-cbs-onder-een-alloc
htoon (Last accessed 19-06-2020).
32 Statistics Netherlands: Inpassing van
het Nederlandse onderwijs in ISCED
2011. 2011. https://www.cbs.nl/nl-nl/
onze-diensten/methoden/classificaties/
onderwijs-en-beroepen/isced/niveau-
isced-2011 (Last accessed 19-06-
2020).
33 UNESCO Institute for Statistics: Inter-
national Standard Classification of
Education ISCED 2011. Quebec, Mon-
treal, 2012
34 Veldhuizen I, Ferguson E, De Kort W,
et al.: Exploring the dynamics of the
theory of planned behavior in the
context of blood donation: does dona-
tion experience make a difference?
Transfusion 2011; 51:2425–37
35 Wevers A, Wigboldus DH, de Kort WL,
et al.: Characteristics of donors who
do or do not return to give blood and
barriers to their return. Blood Trans-
fus-Italy 2014; 12(Suppl 1):s37–43
36 Misje AH, Bosnes V, Gasdal O, et al.:
Motivation, recruitment and retention
of voluntary non-remunerated blood
donors: a survey-based questionnaire
study. Vox Sang 2005; 89:236–44
37 Lemmens KP, Abraham C, Ruiter RA,
et al.: Identifying blood donors willing
to help with recruitment. Vox Sang
2008; 95:211–7
© 2020 The Authors.
Vox Sanguinis (2021) 116, 513–523
 Unknown makes unloved in blood donation 523

Appendix A

Background characteristics for the survey data incorporated in current study

Part of Motive-study
Separate study
Study African background (n = 248) University students (n = 64) Social media users (n = 97) Significance testing
Characteristic n (%)/M (SD) n (%)/M (SD) n (%)/M (SD) F (df)/X2 (df)

Gender X2(2) = 3.5

Male 70 (28%) 22 (34%) 37 (38%)
Female 178 (72%) 42 (66%) 60 (62%)
Age 36.4 (12.4) 20.6 (2.5) 28.7 (12.0) F(2, 406) = 55.4***
Educational level X2(4) = 76.5***
Low 27 (11%) 0 (0%) 6 (6%)
Middle 124 (50%) 0 (0%) 37 (38%)
High 97 (39%) 64 (100%) 54 (56%)
Ethnic background X2(2) = 373.9***
Dutch 0 (0%) 57 (89%) 95 (98%)
African 248 (100%) 7 (11%) 2 (2%)

NOTES
*p < 0.05.
**p < 0.01.
***p < 0.001.

© 2020 The Authors.

Vox Sanguinis published by John Wiley & Sons Ltd on behalf of International Society of Blood Transfusion
Vox Sanguinis (2021) 116, 513–523
 Vox Sanguinis (2021) 116, 524–532
© 2020 International Society of Blood Transfusion
ORIGINAL PAPER DOI: 10.1111/vox.13020

Ovine red cell concentrates for transfusion research – is the

storage lesion comparable to human red cell concentrates?
Gabriela Simonova,1,2,3 Rebecca Wellburn,1 Yoke Lin Fung,4 John F. Fraser2,3 & John-Paul Tung1,2,3
1
Research and Development, Australian Red Cross Lifeblood, Brisbane, QLD, Australia
2
Faculty of Medicine, The University of Queensland, Brisbane, QLD, Australia
3
Critical Care Research Group, The University of Queensland and The Prince Charles Hospital, Brisbane, QLD, Australia
4
School of Health and Sports Sciences, University of Sunshine Coast, Sunshine Coast, QLD, Australia

Received: 14 April 2020,
revised 2 September 2020,
accepted 30 September 2020,
published online 26 October 2020
Background and objectives Sheep are increasingly being used as a large in vivo
animal model of blood transfusion because they provide several advantages over
small animals. Understanding the effects of storage duration on ovine (ov) red
cell concentrates (RCCs) and how these changes compare with stored human (hu)
RCCs is necessary to facilitate clinical translation of research findings.
Materials and methods OvRCCs (n = 5) collected and processed in standard
human blood collection packs, and equivalent huRCCs provided by Australian
Red Cross Lifeblood (n = 5), were stored at 2–6°C for 42 days, with samples col-
lected weekly. Haemolysis index was determined by measuring supernatant hae-
moglobin concentration. Biochemical parameters were evaluated using a blood
gas analyser. Energy metabolites and biologically active lipids were measured
using commercial assays. Osmotic fragility was determined by lysis in various
saline concentrations. Extracellular vesicles were characterized by nanoparticle
tracking analysis.
Results Ovine red blood cells (RBCs) are double in number, smaller in size and
more fragile than human RBCs. Haematological values were unchanged through-
out storage. In contrast, biochemical and metabolic values, and haemolysis index
in three of the five ovRCCs exceeded huRCCs licensing criteria by day 42. Accu-
mulation of extracellular vesicles and biologically active lipids was comparable
between huRCCs and ovRCCs.
Conclusion This study documents similarities and differences in the storage
lesion of ovRCCs and huRCCs. This new information will guide the design of
ovine transfusion models to enhance translation of findings to human transfusion
settings.
Key words: sheep, storage lesion, haemolysis, osmotic fragility, extracellular vesi-
cles, biologically active lipids.

transfusion. Under these conditions, huRCCs undergo

Introduction
In Australia, human (hu) red cell concentrates (RCCs) are
routinely stored at 2–6°C in saline–adenine–glucose–man-
nitol (SAGM) storage solution for up to 42 days prior to
Correspondence: Gabriela Simonova, Research and Development,
Australian Red Cross Lifeblood, 44 Musk Avenue, Kelvin Grove, QLD
4059.
E-mail: gsimonova@redcrossblood.org.au
morphological changes and biochemical deterioration col-
lectively referred to as the ‘storage lesion’. While these
changes have been well documented in vitro, their clinical
relevance, especially in huRCCs stored for longer periods,
is still not completely understood [1]. Recent randomized
controlled trials (RCTs) have attempted to address this,
and found no differences in clinical outcomes between
fresh (<8–10 days) and standard-issue (22–28 days)
huRCCs [2–5]. However, the impact of near-expiry

524

huRCCs on patient outcomes remains unknown, high-
lighting the need for more scientific evidence before
changes to transfusion practice are considered or disre-
garded.
Ethical considerations make it difficult to prospectively
investigate the clinical impact of near-expiry blood com-
ponents; however, this is possible in animal models. Fur-
thermore, animal models, unlike RCTs, allow for
investigation of the underlying biological and pathologi-
cal mechanisms underlining any clinical impact observed.
Whether findings from animal models are clinically trans-
latable is uncertain [6], highlighting the importance of
understanding the strengths and limitations of such mod-
els. Several small animals, such as mice [7], rats [8] and
guinea pigs [9], have been used to investigate the effects
of RCC storage lesion. However, some limitations of using
small animals are their small blood volume and the
shorter life span of their red blood cells (RBCs), causing
RCCs to undergo a more rapid deterioration and acceler-
ated storage lesion development. Dogs have been most
commonly used as a larger animal model, as some facili-
ties routinely collect, prepare and store canine RCCs for
veterinary use. While canine RCCs can be stored up to
42 days depending on the additive solution used [10–12],
about half of the canine RCCs stored for longer than
35 days exceed huRCC licensing specifications [13]. A pig
transfusion model also has been described; however, por-
cine RCCs stored for 14 days develop a storage lesion
similar to that of huRCCs stored for 42 days [14].
Sheep are used in transfusion research because of their
similar size and anatomy to humans, and importantly
because ovine (ov) and human RBCs have a similar life
span (~120 days) [15]. However, whether sheep models
can be used to investigate the clinical effects of transfus-
ing near-expiry RCCs remains undetermined. The US
Food and Drug Administration (FDA) has set the maxi-
mum huRCC shelf-life when stored in additive solution
(including SAGM and Adsol (AS-1)) as 42 days [16]. This
is based on two criteria: firstly, for ≥75% of transfused
RBCs to be circulating in vivo 24 hours post-transfusion;
and secondly, for day 42 haemolysis to be <1%. In Aus-
tralia, following the European guidelines by the Council
of Europe (CoE), the upper limit of haemolysis is <0
8%.
Previous research has established that ovRCCs meet these
criteria through to 40 days of storage [17]; however, the
biochemical and metabolic changes, as well as accumula-
tion of biologically active lipids and extracellular vesicles
(EVs) during storage of ovRCCs, have not been well char-
acterized and a direct comparison to the storage lesion of
huRCCs has not been reported. Therefore, this study
aimed to further characterize the storage lesion of
ovRCCs, and to compare this to the storage lesion of
huRCCs.
© 2020 International Society of Blood Transfusion
Vox Sanguinis (2021) 116, 524–532
Storage lesion of sheep red cell concentrates 525
Materials and methods
Whole blood collection, processing and storage
This study received approvals from the Queensland
University of Technology Animal Ethics Committee and
from the Australian Red Cross Lifeblood Human Ethics
Committee. Whole blood (450 – 50 ml) was collected
from five adult (3–4 years old) female Merino sheep into
top-and-bottom collection packs (Fresenius Kabi AG, Bad
Homburg, Germany) containing citrate-phosphate-dex-
trose anticoagulant, and processed into ovRCC using stan-
dard blood bank procedures with minor modifications
described previously [18]. Briefly, sheep whole blood was
centrifuged (50009 g for 30 min at room temperature),
plasma and buffy coat were separated, and ovRCCs were
leucofiltered and preserved in SAGM additive solution for
42 days at 2–6°C. Human whole blood was collected from
volunteer donors and processed into leucoreduced RCCs
stored in SAGM according to standard Australian Red
Cross Lifeblood procedures. HuRCCs (n = 5) were stored
at 2–6°C for 42 days.
Characterization of in vitro storage changes
Samples (15 ml) from both ovRCCs and huRCCs were col-
lected aseptically on days 1, 7, 14, 21, 28, 35 and 42.
Haemolysis, biochemical parameters and osmotic fragility
were determined on the day of sampling. Leftover sam-
ples were double-centrifuged (30009 g for 15 min at
room temperature), and supernatants were stored at -
80°C until required for later analyses.
RBC quality assessment and measurement of
biochemical and metabolic changes
Red blood cell values including RBC count, haemoglobin
(Hb), mean corpuscular volume (MCV), mean corpuscular
Hb (MCHC) and haematocrit (Hct) were determined on the
Coulter AcT diff haematology analyser (Beckman Coulter,
Lane Cove, NSW) on veterinary and human mode, respec-
tively. Haemolysis was determined by measuring super-
natant Hb concentration using a portable HemoCue Hb
201+ Photometer system (HemoCue AB, Angelholm, Swe-
den), and the percentage of haemolysis was calculated as
follows: (100 – Hct) 9 free plasma Hb/total Hb. Biochem-
ical parameters, including pH and concentrations of glu-
cose, lactate, sodium and potassium, were evaluated using
a blood gas analyser (Radiometer ABL System 625,
Copenhagen, Denmark). Concentrations of 2,3-diphospho-
glycerate (2,3-DPG) and adenosine-triphosphate (ATP)
were measured enzymatically from frozen supernatants of
deproteinized RCCs using commercially available non-
526 G. Simonova et al.
primate-specific enzyme-linked immunosorbent assays
(ELISA) and controls (Roche, Mannheim, Germany) as per
the manufacturer’s instruction.
Assessment of osmotic fragility
Osmotic fragility of ovRCCs and huRCCs at day 1 and day
42 was determined by measuring lysis in a range of buf-
fered saline (NaCl) concentrations, as described previously
[19]. Briefly, RCC samples were suspended in water (100%
lysis control) or 1 ml of buffered NaCl solutions ranging
from 0
3–0
9% (w/v), incubated for 30 min at room tem-
perature followed by centrifugation (10009 g for 5 min).
Supernatants were collected, and haemolysis was mea-
sured with a spectrophotometer (Synergy H1 Hybrid
Plate Reader, Millennium Science, BioTek) at a wavelength
of 540 nm. Results were normalized to the 100% lysis
control, and osmotic fragility curves were plotted.
Characterization of extracellular vesicles (EVs)
Concentration and size distribution of EVs in RCC super-
natants were determined by nanoparticle tracking analysis
(NTA) using a NanoSight NS300 (Malvern Technologies,
Malvern, UK). Frozen samples were thawed and diluted in
0
2 µm filtered phosphate-buffered saline adjusted to 80–
150 particles/frame. Ten experimental replicates of each
sample were analysed under constant flow conditions at
room temperature. Concentration and size distribution of
MVs were determined using NTA 3.2 software with detec-
tion threshold of 5.
Quantification of biologically active lipid
mediators
The concentration of 12-hydroxyeicosatetraenoic acid
(HETE) and 15-HETE was determined by using commer-
cially available non-primate-specific ELISAs (Enzo Life
Sciences, Australia) according to the manufacturer’s
instructions. All samples were run in duplicate.
Statistical analysis
Results are reported as the mean – standard deviation
(SD). One-way analysis of variance (ANOVA) with post hoc
Bonferroni multiple comparisons was performed to deter-
mine changes over storage time compared to day 1 for
both huRCCs and ovRCCs, respectively. Two-way ANOVA
was used to assess differences between ovine and human
RCCs throughout storage as well as the osmotic fragility
of RCCs stored for day 1 and/or day 42. All statistical
analyses were performed using GraphPad Prism 7 (version
7.05, GraphPad Software Inc, USA) with significance set
at values of P ≤ 0
05.
Results
RBC quality assessment and biochemical and
metabolic changes
Ovine RBCs were significantly smaller in size and double
the number of human RBCs. The measured haematologi-
cal values in both ovRCCs and huRCCs were unchanged
throughout storage (Table 1).
Haemolysis of both huRCC and ovRCC increased dur-
ing storage (Fig. 1A). In ovRCC, there was a significant
increase on day 21 that continued to day 42. The
haemolysis index of all five huRCCs met the less than
0
8% specification by CoE until the end of shelf-life;
however, this was true for only two of the five ovRCCs.
The pH of huRCCs decreased steadily throughout storage
(Fig. 1B). In contrast, the pH of ovRCCs remained
unchanged. Potassium levels in both human and ovine
RCCs increased during storage, but the increase was
modest in ovRCCs (Fig. 1C). Sodium levels in huRCCs
decreased steadily during storage (Fig. 1D). In contrast,
in ovRCCs there was an initial decrease followed by an
increase on day 14 which remained relatively stable
thereafter (Fig. 1D). At day 1, ovRCCs had higher glu-
cose (Fig. 1E) and lower lactate (Fig. 1F) levels than
huRCCs, but during storage, glucose levels decreased
and lactate levels in both increased. Increased lactate
levels were evident earlier in storage in huRCCs com-
pared to ovRCCs (day 14 vs. day 28). At day 1, ovRCCs
had higher ATP (Fig. 1G) and lower 2,3-DPG (Fig. 1H)
concentrations compared to huRCCs. ATP and 2,3-DPG
concentrations in both huRCCs and ovRCCs decreased
throughout storage. In huRCCs, ATP decreased from day
35 (P = 0
0287) and 2,3-DPG decreased from day 14
(P = 0
0109). In ovRCCs, ATP decreased from day 1
throughout storage (P < 0
0001), while 2,3-DPG
decreased at day 14 (P = 0
0272).
Assessment of osmotic fragility
In huRCCs, the mean corpuscular fragility (defined as
50% haemolysis) occurred between 0
45% and 0
50% and
the maximum haemolysis was evident between 0
35%
and 0
40% NaCl (Fig. 2). In ovRCCs, the osmotic fragility
curve shifted to the right, indicating that ovRCCs are less
resistant to osmotic stress than huRCCs. The mean cor-
puscular fragility of ovRCCs occurred between 0
65% and
0
70% NaCl, and maximum haemolysis occurred between
0
40% and 0
45% NaCl. The osmotic fragility curves of
© 2020 International Society of Blood Transfusion
Vox Sanguinis (2021) 116, 524–532
 Storage lesion of sheep red cell concentrates 527

Table 1 Haematological values of ovine and human RCCs stored for 42 days

Storage time D1 D7 D14 D21 D28 D35 D42 P-value

RBC [91012/l]
ovRCC 12
4 – 0
7 12
6 – 0
6 12
9 – 0
6 12
4 – 0
7 12
5 – 0
9 12
6 – 0
6 12
5 – 0
6 P < 0
0001
huRCC 5
6 – 0
3 5
7 – 0
3 5
8 – 0
3 5
7 – 0
3 5
6 – 0
3 5
8 – 0
3 5
9 – 0
3
Hb [g/l]
ovRCC 153 – 8 155 – 7 160 – 9 153 – 9 154 – 9 153 – 9 153 – 8 P = 0
0449
huRCC 173 – 12 173 – 12 177 – 17 173 – 14 169 – 13 175 – 14 179 – 14
Hct [%]
ovRCC 0
45 – 0
03 0
45 – 0
03 0
46 – 0
03 0
44 – 0
03 0
44 – 0
03 0
44 – 0
03 0
44 – 0
03 P = 0
0091
huRCC 0
50 – 0
03 0
51 – 0
03 0
52 – 0
04 0
51 – 0
03 0
50 – 0
03 0
52 – 0
03 0
53 – 0
03
MCV [fl]
ovRCC 36
2 – 3
3 35
5 – 3
0 35
6 – 3
1 35
2 – 3
0 35
1 – 2
9 35
1 – 2
9 35
2 – 2
9 P < 0
0001
huRCC 89
4 – 2
6 89
3 – 2
8 89
3 – 2
9 89
4 – 2
5 89
5 – 2
4 90
1 – 2
4 90
6 – 3
1
MCH [pg]
ovRCC 12
4 – 0
7 12
3 – 0
7 12
5 – 0
7 12
4 – 0
7 12
4 – 0
8 12
2 – 0
8 12
2 – 0
6 P < 0
0001
huRCC 30
8 – 1
5 30
5 – 1
6 30
6 – 2
0 30
6 – 1
6 30
2 – 1
6 30
4 – 1
9 30
6 – 1
9
MCHC [g/l]
ovRCC 343 – 14 347 – 11 351 – 13 352 – 11 354 – 9 347 – 9 348 – 11 ns
huRCC 345 – 7 342 – 8 343 – 13 342 – 8 337 – 10 338 – 12 337 – 10

Values are shown as mean – SD. P-value obtained using a two-way ANOVA comparing ovRCC vs. huRCC throughout storage.
ovRCC, ovine red cell concentrates; huRCC, human red cell concentrates; D, day; RBC, red blood cell; Hb, haemoglobin; Hct, haematocrit; MCV, mean cor-
puscular volume; MCH, mean corpuscular Hb; MCHC, mean corpuscular Hb concentration; ns, not significant.

day 1 and day 42 of ovRCC and huRCC were similar, Discussion

indicating no storage associated change.
Characterization of EVs and quantification of
biologically active lipid mediators
At day 1, EVs present in both huRCCs and ovRCCs were
similar in size (105
0 – 20
3 nm vs. 120
8 – 13
6 nm;
Fig. 3A). EV size increased in both huRCCs (P < 0
0001)
and ovRCCs during storage (P = 0
0005). There were more
EVs in huRCCs than in ovRCCs at day 1
(1
53 9 1010 – 1
65 9 1010 particles/ml vs.
3
48 9 109 – 1
67 9 109 particles/ml; Fig. 3B). Storage
duration was not associated with increased EV concentra-
tion in huRCCs. In contrast, an increased EV concentra-
tion was clearly evident from day 21 until the end of
storage in ovRCCs (P < 0
0001; Fig. 3B).
Both 12- and 15-HETE were detectable in ovRCC and
huRCC supernatants (Fig. 3C & D). In huRCCs, 12-HETE
concentrations did not change during storage. In ovRCCs,
while there was a trend of increased 12-HETE concentra-
tions, this was not statistically significant due to large
range between the five units (Fig. 3C). In ovRCCs, 15-
HETE concentrations increased from day 21 (P = 0
0336)
up to day 35 (P = 0
0206) but then decreased, whereas in
huRCCs, 15-HETE concentrations only increased at the
end of storage on day 42 (P = 0
0046; Fig. 3D).
Animal models can be used to address the query regard-
ing near-expiry RCC transfusion and clinical outcomes.
Sheep are a potential candidate animal for such a model
because of similarities in blood banking procedures [18],
RBC life spans [15] and in vivo recovery of transfused
RBCs [17]. This study provides a more thorough under-
standing of the similarities and differences in the storage
lesions of ovRCCs and huRCCs.
The accumulation of storage lesion defects results in
breakdown and disruption of the RBC membrane and
haemolysis [20]. Haemolysis in both huRCCs and ovRCCs
increased with storage; however, it was significantly
higher in ovRCCs, with a day 21 ovRCC approximately
equivalent to a day 42 huRCC. By day 42, only two of
the five ovRCCs met the <0
8% haemolysis acceptance
criteria. In this study, the haemolysis in ovRCCs was
higher than previously reported [17]. This is likely due to
differences in manufacturing methods. In the previous
study, ovRCCs were prepared by soft spin (6009 g for
10 min) and resuspension in AS-1 [20]. In contrast, in the
present study, ovRCCs were prepared by hard spin
(50009 g for 30 min) and resuspension in SAGM.
One of the storage effects that contributes to haemoly-
sis is disruption to RBC metabolism. RBC glycolysis con-
sumes glucose and ATP and generates 2,3-DPG and

© 2020 International Society of Blood Transfusion

Vox Sanguinis (2021) 116, 524–532
528 G. Simonova et al.

Fig. 1 Biochemical and metabolic changes of ovine and human RCCs stored for 42 days. Haemolysis [A] in stored ovRCCs and huRCCs was determined
by measuring supernatant haemoglobin concentration, and the percentage of haemolysis was calculated. pH [B], concentrations of potassium [C] and
sodium [D] and levels of glucose [E] and lactate [F] were measured with blood gas analyser. Concentrations of ATP [G] and 2,3-DPG [H] were measured
using ELISA. Results are shown as mean – SD (n = 5, respectively). *P < 0
05 vs. day 1 of ovRCC and #P < 0
05 vs. day 1 of huRCC, using one-way
ANOVA. P-value obtained using a two-way ANOVA comparing ovRCC vs. huRCC throughout storage. ATP, adenosine-triphosphate; 2,3-DPG, 2,3-diphospho-

glyceric acid; K+, potassium; nd Na+, sodium. [Colour figure can be viewed at wileyonlinelibrary.com]

© 2020 International Society of Blood Transfusion

Vox Sanguinis (2021) 116, 524–532
 Storage lesion of sheep red cell concentrates 529

lactate, the latter of which lowers pH in the RCC [21].

Fig. 2 Osmotic fragility of ovine and human RCCs stored for 42 days.
Osmotic fragility curves of RCCs were measured on day 1 and day 42
and plotted as the percentage of haemolysis against variable % NaCl
concentrations. Results are shown as mean – SD (n = 5, respectively).
*P < 0
05 vs. day 1 of ovRCC and #P < 0
05 vs. day 1 of huRCC, using
one-way ANOVA. P-value obtained using a two-way ANOVA comparing
ovRCC vs. huRCC throughout storage. [Colour figure can be viewed at
wileyonlinelibrary.com]
ATP synthesis by RBCs is impaired by this lowered pH
resulting in reduced ATP. Storage-related changes in
huRCCs observed in this study (decreased glucose, ATP,
2,3-DPG and pH, as well as increased lactate) were as
expected. Most of these changes were mirrored in ovRCCs
(decreased glucose and ATP, as well as increased lactate);
however, the increase in lactate was modest, and there
was no change in pH. These results are similar to those
previously reported for ovine whole blood [22] and
ovRCCs [18] and reflect differences in metabolism
between human and ovine RBCs. For example, the activ-
ity of glucose-6-phosphate dehydrogenase, which plays a
major role in glycolysis and for the protection against
oxidative stress, in ovine RBCs appears to be considerably
lower compared to human RBCs [23, 24].
Red blood cells intracellular electrolyte concentrations
are regulated by membrane sodium–potassium pumps
that are temperature-sensitive and pH- and ATP-depen-
dent [19, 25, 26]. As would be expected with decreased
ATP and pH during refrigerated storage, the function of
these pumps is disrupted, resulting in the increased potas-
sium and decreased sodium concentrations evident in

Fig. 3 Accumulation of extracellular vesicles and biologically active lipids in ovine and human RCC supernatants throughout 42 days of storage. Mean
particle size [A] and concentration [B] of extracellular vesicles were determined by nanoparticle tracking analysis. Concentrations of 12-HETE [C] and
15-HETE [D] were measured with ELISA. Results shown are mean – SD (n = 5, respectively). *P < 0
05 vs. day 1 of ovRCC and #P < 0
05 vs. day 1 of
huRCC, using one-way ANOVA. P-value obtained using a two-way ANOVA comparing ovRCC vs. huRCC throughout storage. 12-HETE = 12-hydroxyeicosate-
traenoic acid and 15-HETE = 15-hydroxyeicosatetraenoic acid. [Colour figure can be viewed at wileyonlinelibrary.com]

© 2020 International Society of Blood Transfusion

Vox Sanguinis (2021) 116, 524–532
530 G. Simonova et al.
huRCCs in this study. In contrast, modest increases in
both sodium and potassium were seen during ovRCC stor-
age. This may be in part due to the unchanged pH and
reduced lactate accumulation in ovRCCs. Additionally,
Australian Merino sheep, as were used in this study, have
fewer sodium–potassium pumps on their RBCs [27],
resulting in lower potassium and higher sodium levels
than humans [27, 28].
The osmotic fragility of RBCs is determined by mem-
brane integrity, cell size and shape, with spherical cells
more fragile than normal discocyte cells [29]. Therefore,
as ovine RBCs are smaller and more spherical [30] than
human RBCs, the observation of increased osmotic fragi-
lity of ovRCCs compared to huRCCs was expected, and
consistent with other reports [31–33]. No significant dif-
ferences were found in the osmotic fragility of either
ovine RBCs or human RBCs stored for day 1 or day 42.
This finding was unexpected, as earlier studies reported
association between huRBCs osmotic fragility and pro-
longed storage duration [29, 33]. Whether similar to other
quality parameters it is due to the different manufactur-
ing methods [34] is unclear and needs further investiga-
tion.
Extracellular vesicles are shed from cell membranes
and play a role in adverse transfusion events including
coagulopathies, inflammation and transfusion-related
acute lung injury (TRALI) [35–37]. While the size of
EVs in ovRCCs is similar to that of humans, the con-
centration of EVs differed, with initial concentrations in
huRCCs approximately four times higher than in
ovRCCs. While a number of different factors may have
contributed to this difference, it has been shown that
compared to human RBCs, ovine RBCs produce long
membrane extensions when they undergo stress [38].
These extensions do not result in the budding of the
membrane as seen in human RBCs, rather they retract
then collapse on the membrane surface, and therefore,
ovine RBCs do not release EVs in the same way [38].
Storage duration was associated with increased EV con-
centration in ovRCCs, but not in huRCCs. This contrasts
with previous reports of EV accumulation in huRCCs
[36], but might simply reflect the small sample size in
our study and the fact that one of the five huRCC
units had remarkably different results from the other
four units, resulting in wider error bars and no clear
differences in the post-tests. Another possible explana-
tion for the discrepancy might relate to differences in
the techniques used to measure EVs: nanoparticle track-
ing analysis vs. flow cytometry; with the former pro-
viding data on smaller EVs that could not be
characterized by the latter. It also worth noting that
the smaller EVs reported in our study (100–200 nm)
were more numerous than larger EVs previously
reported using flow cytometry (>~500 nm) [36]. Overall,
while ovRCCs have a lower concentration of EVs than
huRCCs, similarities in the mean size and increased
concentration throughout storage suggest sheep may
still be a viable model to study the impact of EVs in
RCC transfusion. However, EVs may need to be concen-
trated and added to ovRCCs in order to better replicate
the human setting.
Changes to membrane integrity can also influence
the release and production of the biologically active
lipids. HETEs, such as 12-HETE and 15-HETE, accumu-
late during storage of huRCCs as a result of peroxida-
tion of lipids in the RBC cell membrane [39]. Both 12-
HETE and 15-HETE play a role in mediating inflamma-
tion, cell proliferation and apoptosis [40], and have
been associated with adverse transfusion outcomes,
including TRALI [39]. Both 12-HETE and 15-HETE were
detected in ovRCCs and huRCCs at comparable concen-
trations; however, storage-related accumulation was
only observed for 15-HETE. Further investigation is
required before ovine models can be used to study the
effects of these biologically active lipids in pre-clinical
transfusion studies.
In summary, sheep RBCs differ from human RBCs in
size and metabolic characteristics. This study documents
some similarities but importantly also defines the differ-
ences between the storage lesion of human and ovine
RCCs. OvRCCs stored for 21 days develop a storage lesion
similar to that of huRCCs stored for 42 days. These limi-
tations need to be considered when designing ovine
in vivo transfusion models for evaluating the effects of
stored RCC transfusion in various clinical situations such
as trauma or sepsis.
Acknowledgements
The authors thank Kristen Glenister for technical assis-
tance. JFF received funding from the Office of Health and
Medical Research, Queensland Health. Australian govern-
ments fund the Australian Red Cross Lifeblood to provide
blood, blood products and services to the Australian com-
munity.
Conflict of interest
The authors have no conflict of interest.
© 2020 International Society of Blood Transfusion
Vox Sanguinis (2021) 116, 524–532

References
1 Yoshida T, Prudent M, D’Alessandro A:
Red blood cell storage lesion: causes
and potential clinical consequences.
Blood Transfus 2019; 17:27–52
2 Lacroix J, Hebert PC, Fergusson DA,
et al.: Age of transfused blood in criti-
cally ill adults. N Engl J Med 2015;
372:1410–1418
3 Heddle NM, Cook RJ, Arnold DM,
et al.: Effect of short-term vs. long-
term blood storage on mortality after
transfusion. N Engl J Med 2016;
375:1937–1945
4 Cooper DJ, McQuilten ZK, Nichol A,
et al.: Age of red cells for transfusion
and outcomes in critically ill adults. N
Engl J Med 2017; 377:1858–1867
5 Steiner ME, Ness PM, Assmann SF,
et al.: Effects of red-cell storage dura-
tion on patients undergoing cardiac
surgery. N Engl J Med 2015;
372:1419–1429
6 Pavenski K, Saidenberg E, Lavoie M,
et al.: Red blood cell storage lesions
and related transfusion issues: a Cana-
dian Blood Services research and
development symposium. Transfus
Med Rev 2012; 26:68–84
7 Makley AT, Goodman MD, Friend LA,
et al.: Murine blood banking: charac-
terization and comparisons to human
blood. Shock 2010; 34:40–45
8 d’Almeida MS, Jagger J, Duggan M,
et al.: A comparison of biochemical
and functional alterations of rat and
human erythrocytes stored in CPDA-1
for 29 days: implications for animal
models of transfusion. Transfus Med
2000; 10:291–303
9 Baek JH, D’Agnillo F, Vallelian F,
et al.: Hemoglobin-driven pathophysi-
ology is an in vivo consequence of the
red blood cell storage lesion that can
be attenuated in guinea pigs by hap-
toglobin therapy. J Clin Invest 2012;
122:1444–1458
10 Price GS, Armstrong PJ, McLeod DA,
et al.: Evaluation of citrate-phosphate-
dextrose-adenine as a storage medium
for packed canine erythrocytes. J Vet
Intern Med 1988; 2:126–132
11 Wardrop KJ, Tucker RL, Mugnai K:
Evaluation of canine red blood cells
stored in a saline, adenine, and
© 2020 International Society of Blood Transfusion
Vox Sanguinis (2021) 116, 524–532
Storage lesion of sheep red cell concentrates 531
glucose solution for 35 days. J Vet
Intern Med 1997; 11:5–8
12 Solomon SB, Wang D, Sun J, et al.:
Mortality increases after massive
exchange transfusion with older stored
blood in canines with experimental
pneumonia. Blood 2013; 121:1663–
1672
13 Ferreira RRF, Graca RMC, Cardoso IM,
et al.: In vitro hemolysis of stored
units of canine packed red blood cells.
J Vet Emerg Crit Care 2018; 28:512–
517
14 Patel NN, Lin H, Jones C, et al.: Inter-
actions of cardiopulmonary bypass
and erythrocyte transfusion in the
pathogenesis of pulmonary dysfunc-
tion in Swine. Anesthesiology 2013;
119:365–378
15 Mock DM, Lankford GL, Burmeister
LF, et al.: Circulating red cell volume
and red cell survival can be accurately
determined in sheep using the [14C]-
cyanate label. Pediatric Res 1997;
41:916–921
16 Council of Europe: Guide to the prepa-
ration, use and quality assurance of
blood components, 19th edn. Strass-
bourg, France, Council of Europe Pub-
lishing, 2017
17 Baron DM, Yu B, Lei C, et al.: Pul-
monary hypertension in lambs trans-
fused with stored blood is prevented
by breathing nitric oxide. Anesthesiol-
ogy 2012; 116:637–647
18 Simonova G, Tung JP, Fraser JF,
et al.: A comprehensive ovine model
of blood transfusion. Vox Sang 2014;
106:153–160
19 Veale MF, Healey G, Sparrow RL:
Effect of additive solutions on red
blood cell (RBC) membrane properties
of stored RBCs prepared from whole
blood held for 24 hours at room tem-
perature. Transfusion 2011; 51(Suppl
1):25S–33S
20 Hess JR, Sparrow RL, van der Meer
PF, et al.: Red blood cell hemolysis
during blood bank storage: using
national quality management data to
answer basic scientific questions.
Transfusion 2009; 49:2599–2603
21 van Wijk R, van Solinge WW: The
energy-less red blood cell is lost:
erythrocyte enzyme abnormalities of
glycolysis. Blood 2005; 106:4034–
4042
22 Sousa RS, Barreto RA Jr, Sousa IK,
et al.: Evaluation of hematologic,
blood gas, and select biochemical
variables in ovine whole blood stored
in CPDA-1 bags. Vet Clin Pathol 2013;
42:27–30
23 Koch JH: Glucose-6-phosphate dehy-
drogenase and potassium concentra-
tion in sheep erythrocytes. Nature
1963; 200:1334–1335
24 Suzuki T, Agar NS, Suzuki M: Red cell
metabolism: a comparative study of
some mammalian species. Comp Bio-
chem Physiol B 1984; 79:515–520
25 Vraets A, Lin Y, Callum JL: Transfu-
sion-associated hyperkalemia. Transfus
Med Rev 2011; 25:184–196
26 Eaton JW, Brewer GJ, Beck CC, et al.:
ATP and potassium content of sheep
erythrocytes. Biochem Biophys Res
Commun 1967; 28:898–903
27 Dunham PB, Hoffman JF: The number
of Na + : K + pump sites on red blood
cells from HK and LK lambs. Biochim
Biophys Acta 1971; 241:399–402
28 Dunham PB: Ion transport in sheep
red blood cells. Comp Biochem Physiol
Comp Physiol 1992; 102:625–630
29 Blasi B, D’Alessandro A, Ramundo N,
et al.: Red blood cell storage and cell
morphology. Transfus Med 2012;
22:90–96
30 Zura Zaja I, Vince S, Poljicak Milas N,
et al.: A new method of assessing
sheep red blood cell types from their
morphology. Animals (Basel) 2019;
9:1130
31 Perk K, Frei YF, Herz A: Osmotic fra-
gility of red blood cells of young and
mature domestic and laboratory ani-
mals. Am J Vet Res 1964; 25:1241–
1248
32 Coldman MF, Gent M, Good W: The
osmotic fragility of mammalian ery-
throcytes in hypotonic solutions of
sodium chloride. Comp Biochem Phys-
iol 1969; 31:605–609
33 Zehnder L, Schulzki T, Goede JS,
et al.: Erythrocyte storage in hyper-
tonic (SAGM) or isotonic (PAGGSM)
conservation medium: influence on
532 G. Simonova et al.
cell properties. Vox Sang 2008;
95:280–287
34 Shih AW, Apelseth TO, Cardigan R,
et al.: Not all red cell concentrate
units are equivalent: international sur-
vey of processing and in vitro quality
data. Vox Sang 2019; 114:783–794
35 Koshiar RL, Somajo S, Norstrom E,
et al.: Erythrocyte-derived microparti-
cles supporting activated protein C-
mediated regulation of blood coagula-
tion. PLoS One 2014; 9:e104200
36 Aung HH, Tung JP, Dean MM, et al.:
Procoagulant role of microparticles in
routine storage of packed red blood
cells: potential risk for prothrombotic
post-transfusion complications.
Pathology 2017; 49:62–69
37 McVey M, Tabuchi A, Kuebler WM:
Microparticles and acute lung injury.
Am J Physiol Lung Cell Mol Physiol
2012; 303:L364–381
38 Backman L, Jonasson JB, Horstedt P:
Phosphoinositide metabolism and
shape control in sheep red blood cells.
Mol Membr Biol 1998; 15:27–32
39 Silliman CC, Moore EE, Kelher MR,
et al.: Identification of lipids that
© 2020 International Society of Blood Transfusion
accumulate during the routine storage
of prestorage leukoreduced red blood
cells and cause acute lung injury.
Transfusion 2011; 51:2549–2554
40 Powell WS, Rokach J: Biosynthesis,
biological effects, and receptors of
hydroxyeicosatetraenoic acids (HETEs)
and oxoeicosatetraenoic acids (oxo-
ETEs) derived from arachidonic acid.
Biochim Biophys Acta 2015;
1851:340–355
Vox Sanguinis (2021) 116, 524–532
 Vox Sanguinis (2021) 116, 533–539
© 2020 The Authors.
ORIGINAL PAPER Vox Sanguinis published by John Wiley & Sons Ltd on behalf of International Society of Blood Transfusion
DOI: 10.1111/vox.13023

The ‘rejuvenating factor’ TIMP-2 is detectable in human

blood components for transfusion
Julia Hoefer,1 Christian Dal-Pont,2 Stefan Jochberger,3 Raffaella Fantin3 & Harald Schennach2
1
Department of Urology, Medical University of Innsbruck, Innsbruck, Austria
2
Central Institute for Blood Transfusion & Immunological Department, University Hospital of Innsbruck, Innsbruck, Austria
3
Department of Anesthesiology and Critical Care Medicine, University Hospital of Innsbruck, Innsbruck, Austria

Background and Objectives Tissue inhibitor of metalloproteinases 2 (TIMP-2) is a

Received: 3 June 2020,
revised 16 September 2020,
accepted 8 October 2020,
published online 26 October 2020
protein suspected to be crucial in numerous physiological and pathological pro-
cesses such as morphogenesis, tissue remodelling and metastasis suppression. In
animal models, the administration of TIMP-2 to aged mice improved their cogni-
tive functions. Therefore, one can hypothesize that differences in TIMP-2 levels
between blood donors and recipients might influence cognitive functions also in
humans. However, the stability of TIMP-2 during processing and storage of blood
components for transfusion has not been intensively investigated so far. This
study determined TIMP-2 concentrations in fresh-frozen plasma (FFP), erythro-
cyte concentrate (EC) and pathogen-inactivated platelet concentrate (PI-PC)
depending on the donor’s demographic factors age and gender.
Materials and Methods Tissue inhibitor of metalloproteinases 2 was measured in
FFP (n = 30), EC (n = 12) and PI-PC (n = 12) using a Q-Plex single-plex
immunoassay for chemiluminescence-based detection. Absolute quantification of
TIMP-2 was performed with Q-view software. Fresh umbilical cord plasma was
used as a positive control.
Results Tissue inhibitor of metalloproteinases 2 was detected in FFP (30/30 sam-
ples), EC (11/12 samples) and PI-PC (12/12 samples). The median TIMP-2 concen-
tration in EC (17
2 ng/ml; range: 0–26
5 ng/ml) was significantly lower
compared with FFP (63
4 ng/ml; range: 44
4–87
3 ng/ml) or PI-PC (69
9 ng/ml;
range: 39
9–83
6 ng/ml). Across all blood components, TIMP-2 levels are compa-
rable in male and female donors and independent of age.
Conclusion Tissue inhibitor of metalloproteinases 2 is detectable and stable in
FFP, PI-PC and, in low concentration, EC. It can be hypothesized that TIMP-2
will be transmitted to recipients during transfusion.
Key words: blood components, blood transfusion, cognitive function, morphogen-
esis, tissue inhibitors of metalloproteinases.

This effect could be attributed to the chemokine eotaxin-1

Introduction
In parabiosis models in which the blood streams of young
mice were surgically connected to those of older animals,
the young mice showed a decline in cognitive functions.
Correspondence: Raffaella Fantin, Department of Anesthesiology and
Critical Care Medicine, University Hospital of Innsbruck, Anichstrasse 35,
6020 Innsbruck, Austria
E-mail: raffaella.fantin@i-med.ac.at
(CCL11), a molecule which has also been negatively corre-
lated with memory function in human Alzheimer’s disease
patients [1,2]. On the other hand, it has been demonstrated
that human umbilical cord plasma revitalized hippocampal
functions of aged mice [3]. The cognitive benefits conferred
by cord plasma were mediated by tissue inhibitor of metal-
loproteinases 2 (TIMP-2), a blood-borne factor abundant in
umbilical cord plasma as well as young mouse plasma and
hippocampi. Taken together, these studies provide evidence

This is an open access article under the terms of the Creative Commons Attribution License, 533
which permits use, distribution and reproduction in any medium, provided the original work is properly cited.
534 J. Hoefer et al.
of the presence of age-dependent, blood-borne factors
which affect neurogenesis, cognitive functions and ageing
when transferred to other individuals.
On the basis of these facts, it can be hypothesized that
such ‘ageing’ or ‘rejuvenating’ factors are present also in
blood components for transfusion and may be transferred
to recipients during transfusion, which might affect the
recipient‘s cognitive capabilities. Indeed, it is known that a
considerable proportion of patients experience a temporary
or permanent decline in cognitive performance after major
surgery, a phenomenon called postoperative cognitive dys-
function (POCD) [4]. The causal relationship of blood trans-
fusion and such disturbances has not been clarified so far.
However, intraoperative transfusion of erythrocyte concen-
trates (>3 units) has been described as an independent risk
factor for the development of POCD [5]. Therefore, it is
essential to investigate whether factors associated with
cognitive function, neurogenesis and ageing are present in
blood components processed and stored for transfusion.
In a previous study, we were able to demonstrate that
the ‘ageing factor’ eotaxin-1 is detectable and stable in
ready-to-use blood products for transfusion. Most impor-
tantly, we discovered a gender-independent increase of
eotaxin-1 with rising age of donors in both fresh-frozen
plasma (FFP) and erythrocyte concentrates (EC), while
eotaxin-1 was subject to only minor fluctuations within
one donor over a longer period of time [6]. This is signifi-
cant as high eotaxin-1 levels are associated with
decreased cognitive functions, although these were not
examined in our previous study [1,2,6]. It might be possi-
ble that besides eotaxin-1 also other factors associated
with cognitive functions and ageing are detectable in
human blood components processed for transfusion.
TIMP-2 has originally been described as an endogenous
inhibitor of matrix-metalloproteinases (MMPs). It is
involved in MMP-2 activation and contributes to the pro-
tection of the extracellular matrix (ECM) from proteolysis
[7]. It has also been reported that TIMP-2 is involved in
neuronal differentiation and can act as a key effector of
the pro-neurogenic response to an inducing stimulus
[8,9]. Moreover, treatment of aged mice with either
umbilical cord plasma or purified TIMP-2 led to signifi-
cant improvements in multiple cognitive measures such
as contextual memory, learning and nesting, as well as
increased synaptic plasticity in the hippocampus [3]. In
addition, TIMP-2 knockout mice display motoric dysfunc-
tions such as dyskinesia, trembling and a splayed, short-
ened gait [10]. This phenotype has mechanistically been
linked to altered formation and maintenance of neuro-
muscular junctions (NMJ), the deterioration of which is a
common sign of ageing [11]. Taken together, these obser-
vations suggest that TIMP-2 influences cognitive as well
as motoric functions, which are both reduced in ageing. It
Vox Sanguinis published by John Wiley & Sons Ltd on behalf of International Society of Blood Transfusion
is therefore possible that TIMP-2 might affect cognitive
capabilities and/or motor functions in recipients if trans-
ferred during blood transfusion.
As stated above, it has been shown that TIMP-2 levels
differ between young and old mice [3]. In humans, Rossig-
nol et al. demonstrated age-dependent changes. It has also
been demonstrated that TIMP-2 is present in freshly drawn
human plasma [12,13]. However, until now, it has neither
been assessed if TIMP-2 is detectable in human blood com-
ponents processed and stored for transfusion, nor if TIMP-2
correlates with age or gender in adult human beings.
In the present study, we assessed for the first time if
blood products for transfusion (FFP, EC, PI-PC) contain
detectable TIMP-2 levels and if TIMP-2 concentration
varies in relation to the donor‘s age or gender.
Material and methods
This trial was performed as a descriptive, single-centre,
non-clinical pilot study. All experiments have been con-
ducted in accordance with the Helsinki Declaration. The
study was approved by the Ethics Committee of the Medical
University of Innsbruck (Ethical vote number: 1083/2018).
Investigational product
For the present study, we analysed leukocyte-depleted
blood products (FFP [n = 30], EC [n = 12], and PI-PC
[n = 12]) from volunteer donors aged 18 to 63. At the
University Hospital of Innsbruck, the suitability for vol-
unteer blood donation is assessed on the basis of strin-
gent, standardized health criteria. After completing a
medical questionnaire on their general health status,
recent diseases, medication and travelling, all persons are
physically examined, including measurements of blood
pressure, heart rate, temperature and haemoglobin con-
centration. The donated blood is further tested for neop-
terin as well as transfusion-transmissible infections such
as HIV, hepatitis, syphilis and parvovirus B19. In case
any disease or abnormality is detected, the donated blood
is discarded. This procedure is to ensure all donors are in
good health at the time of blood donation.
After donation, the blood components used for this
study were stored under standard conditions at the
Central Institute for Blood Transfusion & Immunological
Department at the University Hospital of Innsbruck,
Austria, after having undergone the following procedures:
Preparation of erythrocyte concentrate
Leukocyte-depleted ECs were prepared from single-donor
whole blood donations by cell separation. The preparation
was conducted in accordance with the guidelines of the
© 2020 The Authors.
Vox Sanguinis (2021) 116, 533–539

Council of Europe. One hundred millilitres of additive
solution (SAGM) were added in order to increase shelf
life. Each 280 ml EC package had a haematocrit of 50–70
vol% and contained less than 1*106 leukocytes. ECs are
stored for a maximum of 42 days at 4°C.
Preparation of fresh-frozen plasma
Leukocyte-depleted FFPs were prepared from single-donor
whole blood donation by cell separation. FFPs at a core
temperature of -30°C can be stored up to 2 years.
Preparation of pathogen-inactivated PC (PI-PC)
Leukocyte-depleted PCs were prepared by single-donor
apheresis. Each 300 ml PI-PC contained 2–4*1011 plate-
lets, suspended in approximately 35% plasma and 65%
SSP + platelet additive solution. For safety reasons, patho-
gen inactivation treatment of all PCs was performed with
amotosalen/UVA treatment (INTERCEPT blood system,
CERUS corporation, Amersfoort, the Netherlands). PI-PCs
were stored under continuous agitation in gas-permeable
sterile plastic bags at room temperature (20–24°C). Under
these conditions, the maximum storage time is 7 days.
Sampling for the study
Aliquots from FFP (n = 30), EC (n = 12) and PI-PC
(n = 12) were taken from the respective blood component
using a sterile connecting device.
In detail, samples used for the present study were pre-
pared as follows:
All FFP samples (2 ml) were drawn at the end of the
manufacturing process before freezing within 24 h after
donation. Samples were then aliquoted (3 x 600 µl) and
stored at -20°C until subjected to measurement 2–
6 weeks later. EC samples (2 ml) were taken from the
ready-to-use product within 24 hours after donation and
stored at 4°C until subjected to measurement on the next
day. PI-PC samples (2 mL) were taken from the ready-to-
use product within 24 h after donation and stored at 4°C
until subjected to measurement on the next day.
All measurements of the respective blood product (FFP,
EC, PI-PC) were performed on the same day.
TIMP-2 measurement
A Q-Plex single-plex immunoassay for chemilumines-
cence-based detection of human TIMP-2 was employed
according to the manufacturer’s protocol (Quansys Bio-
sciences, Logan, UT; range: 27
43 to 20 000 pg/ml; lower
limit of detection: 15
9 pg/ml; intra-assay coefficient of
variation: 8%; inter-assay coefficient of variation: 12%).
© 2020 The Authors.
Vox Sanguinis published by John Wiley & Sons Ltd on behalf of International Society of Blood Transfusion
Vox Sanguinis (2021) 116, 533–539
TIMP-2 is detectable in human blood components 535
The kit contained a freeze-dried calibrator for solubiliza-
tion in sample buffer and for generating a seven-point
standard curve. Signals were detected using ChemiDoc
chemiluminescence imaging system (Biorad), and TIMP-2
was quantified with Q-View Software (Quansys) in rela-
tion to an 8-point standard curve obtained by serial dilu-
tion of recombinant TIMP-2. All samples were diluted
1:20 with sample dilution buffer and measured in dupli-
cates. Plasma from fresh residual umbilical cord blood
(n = 3, 2 ml each) was used as a positive control.
Statistical methods
Statistical analyses were performed in SPSS and Graph-
Pad Prism. Differences between groups were analysed
using Mann–Whitney test. All differences highlighted by
asterisks were statistically significant as encoded in fig-
ures (*P < 0
05, **P < 0
01, ***P < 0
001). Data are pre-
sented as Box–Whiskers Plots showing median and
minimum to maximum values. For comparison of TIMP-2
levels in relation to age, donors were divided into two
groups, representing the younger and the older 50% each.
Results
Tissue inhibitor of metalloproteinases 2 is
detectable in blood components for transfusion
Tissue inhibitor of metalloproteinases 2 was consistently
detected in FFP (30/30 samples), PI-PC (12/12 samples) as
well as in EC (11/12 samples). Median TIMP-2 levels were
63
4 ng/ml in FFP (range: 44
4–87
3), 69
6 ng/ml in PI-
PC (range: 39
9–83
6) and 17
2 ng/ml in EC (range: 0–
26
5), as summarized in Table 1. Statistical analysis
revealed significantly higher TIMP-2 concentrations in
FFP and PI-PC, when compared to EC (Fig. 1).
Tissue inhibitor of metalloproteinases 2 levels are
not dependent on donor’s age or gender
Regarding the question whether TIMP-2 levels depend on
demographic factors like age or gender, no significant
Table 1 Summary of TIMP-2 measurement
TIMP-2 (ng/ml) FFP (n = 30) PI-PC (n = 12) EC (n = 12)
Minimum 44
4 39
9 0
25% percentile 60
08 65
65 4
98
Median 63
4 69
6 17
2
75% percentile 68
65 81
38 22
35
Maximum 87
3 83
6 26
5
FFP, fresh-frozen plasma; EC erythrocyte concentrate; PI-PC, pathogen-
inactivated platelet concentrate; TIMP-2 values are presented as ng/ml.
536 J. Hoefer et al.

Fig. 1 TIMP-2 is detectable in fresh-frozen plasma (FFP), pathogen-inactivated platelet concentrate (PI-PC) and erythrocyte concentrate (EC) (a,c,d).
Umbilical cord plasma serves as a positive control. TIMP-2 in FFP is comparable in male and female donors and independent of age (a,b). TIMP-2 in PI-
PC is comparable in male and female donors and independent of age (c,d). TIMP-2 in EC is rather low concentrated but comparable in male and female
donors and independent of age (e,f). Data are presented as Box–Whiskers Plot, showing median and minimum to maximum values. *P < 0
05;
**P < 0
01; ***P < 0
001; Mann–Whitney test.

differences in TIMP-2 in FFP donated by younger (21–
45 years) or older (46–63 years) individuals could be
detected. However, umbilical cord plasma exhibited sig-
nificantly higher TIMP-2 levels compared with plasma
from adult donors (Fig. 1a). Moreover, TIMP-2 levels in
FFP from male and female donors were comparable
(Fig. 1b). Similarly, the median TIMP-2 concentration did
not differ in PI-PC from younger (18–35 years) and older
(36–61 years) donors (Fig. 1c), nor in PI-PC samples from
male or female individuals (Fig. 1f), although variability
was higher in the older age group and in male donors. As
mentioned above, the median TIMP-2 concentration in EC
was significantly lower compared with FFP or PI-PC, and
again, TIMP-2 levels were comparable in EC samples from
younger (24–45 years) and older (46–63 years) individu-
als (Fig. 1e), as well as in samples from male and female
donors (Fig. 1f).
Tissue inhibitor of metalloproteinases 2 is stable
in FFP during storage
Tissue inhibitor of metalloproteinases 2 was detectable
in FFP after 15 days and 42 days of storage at -20°C.
As shown in Fig. 2, no major differences in TIMP-2
levels were found. Median TIMP-2 levels were 65 ng/
mL (range: 44
4–87
3) after 15 days of storage and
60
6 ng/mL (range: 53
5–65
6) after 42 days of storage
(Table 2).

© 2020 The Authors.

Vox Sanguinis published by John Wiley & Sons Ltd on behalf of International Society of Blood Transfusion
Vox Sanguinis (2021) 116, 533–539

Fig. 2 The Box–Whiskers Plot shows median and minimum to maximum
values of TIMP-2 level of FFP samples, which were stored for 15 and
42 days. No major differences in TIMP-2 levels were detected.
Table 2 TIMP-2 is stable during storage
TIMP-2 (ng/ml) 15 days (n = 20) 42 days (n = 10)
Minimum 44
4 53
5
25% percentile 62
95 56
48
Median 65
0 60
6
75% percentile 70
08 62
43
Maximum 87
3 65
6
TIMP-2 is detectable in FFP (fresh-frozen plasma) after 15 days and
42 days of storage at -20°C. TIMP-2 values are presented as ng/ml.
Discussion
Blood transfusion is a common medical procedure used
for patients after serious injuries, certain diseases, or dur-
ing major surgeries. EC, FFP and PI-PC are routinely
transfused without matching recipients and donors for
gender, age and other demographic factors. It is therefore
important to investigate the presence and stability of cru-
cial blood-borne factors in blood products. Moreover, it is
essential to assess differences in the concentration of such
factors in blood products from individuals with different
demographic characteristics as well as the implication of
these differences for the patient‘s outcome after blood
transfusion. TIMP-2 is an interesting candidate for such
an analysis, given that TIMPs are involved in numerous
biological processes that require tissue remodelling and
are thus key players in a range of physiological and
pathological processes such as angiogenesis, wound heal-
ing, inflammation and cancer [14]. In addition, TIMPs
influence hippocampal functions, cognitive measures and
neurite outgrowth, and high TIMP-2 in human umbilical
cord plasma revitalized hippocampal functions of aged
mice [2,3,14–16]. A previous study demonstrated that
TIMP-1 is detectable in platelet-containing blood
© 2020 The Authors.
Vox Sanguinis published by John Wiley & Sons Ltd on behalf of International Society of Blood Transfusion
Vox Sanguinis (2021) 116, 533–539
TIMP-2 is detectable in human blood components 537
preparations [17]. However, the presence of TIMP-2 in
blood components has not been assessed so far.
In this descriptive pilot study, we demonstrated that
FFP and PI-PC contain comparable amounts of TIMP-2 as
reported in plasma concentrations of humans. On the
other hand, TIMP-2 was low to undetectable in EC. It is
known that platelets contain MMPs and their inhibitors –
among them TIMP-2 – and that these factors are released
into the plasma [18]. However, so far it has not been clear
if TIMP-2 is stable during the process of PI-PC and FFP
preparation and storage. Our data clearly show that
TIMP-2 is detectable in PI-PC and FFP and thus will be
transferred to recipients during transfusion. In our study,
FFP exhibits a median TIMP-2 concentration of 63 ng/ml,
ranging from 44 to 87 ng/ml. Notably, in freshly drawn
human plasma, the TIMP-2 concentration has been found
to range from 30 to 200 ng/ml, with a median TIMP-2
concentration of approximately 80 ng/ml in healthy sub-
jects [12,13]. Furthermore, TIMP-2 was detectable in simi-
lar plasma levels after storage of 15 and 42 days at -
20°C. Thus, it can be concluded that TIMP-2 is a rela-
tively stable molecule that is not significantly affected
during processing and long-term storage of FFP.
We were also able to detect TIMP-2 in EC, although
levels were low compared with FFP and PI-PC. Given that
EC contains approximately 4% of residual plasma, in con-
trast to 35% of residual plasma in PI-PC, a part of the
detected amount of TIMP-2 might derive from this source.
Interestingly, our data show that TIMP-2 levels in all
blood components are comparable across male and female
donors and that TIMP-2 does not correlate with the
donor‘s age. This is surprising, given that increased TIMP-
2 levels were found in young mice compared with older
ones [3]. Concerning the dependence of TIMP-2 levels in
human plasma on age, trials showed discordant results.
Tayebjee et al. found a modest negative correlation with
age in healthy humans, whereas Bonnema et al. describe
an increase of plasma TIMP-2 levels with increasing age
[19,20]. However, more data are needed for clarification.
However, it has to be kept in mind that all blood prod-
ucts analysed in this study were of adult donors, with a
minimum age of 18 years. Given that we and others
found TIMP-2 to be enriched in human umbilical cord
plasma, it is possible that TIMP-2 accumulates only dur-
ing embryogenesis and the earliest years of childhood [3].
Nevertheless, we detected considerable differences in
TIMP-2 between individuals, especially in FFP. The rea-
sons for these differences could not be assessed in this
study. Processes that require tissue remodelling or wound
healing might increase levels of MMPs and TIMPs; how-
ever, as mentioned above, one can hypothesize that all
donors were in good health at the time of donation.
Given the revitalizing effects of TIMP-2 on the cognitive
538 J. Hoefer et al.

function of aged mice, it can be expected that a transfu-

sion of FFP or PI-PC containing high levels of TIMP-2 to
recipients with low TIMP-2 might influence their cogni-
tive measures as well [3]. However, this hypothesis has to
be investigated in subsequent studies. Moreover, also
other processes might be influenced by differences in
TIMP-2 in donor and recipient blood. For example, it has
been shown in mice that MMP-2 amplifies the response
of platelets to weak stimuli, thereby promoting arterial
thrombosis and that this process can be abolished by
infusion of TIMP-2 [21].
Taken together, our data clearly provide evidence that
TIMP-2 is present in FFP and PI-PC, but neglectable in EC
processed for transfusion and that TIMP-2 levels vary
between individuals. However, we did not detect any sig-
nificant correlation with age or gender of the donors,
although it has to be stated that the cohort of donors in
this pilot study was small. Subsequent studies have to con-
firm these findings in a large population and further deci-
pher the source of TIMP-2 variability between individuals.
Acknowledgements
JH performed all measurements and analyses and wrote
the manuscript together with HS, SJ and RF. CDP col-
lected samples and data from donors. All authors con-
tributed to study design and interpretation of results. HS
was responsible for the study.
Funding
This work has been supported by the Austrian Science
Fund (FWF) grant T 738-BBL, the Hertha Firnberg fellow-
ship programme to JH and the Austrian Red Cross Dr.
Karl Landsteiner Funding. The funders had no role in
study design, data collection and analysis, decision to
publish or preparation of the manuscript.
Conflict of Interests
All authors declare no conflict of interest.

References
1 Villeda SA, Luo J, Mosher KI, et al.:
The ageing systemic milieu negatively
regulates neurogenesis and cognitive
function. Nature 2011; 477:90–94
2 Bettcher BM, Fitch R, Wynn MJ, et al.:
MCP-1 and eotaxin-1 selectively and
negatively associate with memory in
MCI and Alzheimer’s disease dementia
phenotypes. Alzheimer’s Dementia
Diagnosis Assess Dis Monit 2016;
3:91–97
3 Castellano JM, Mosher KI, Abbey RJ,
et al.: Human umbilical cord plasma
proteins revitalize hippocampal func-
tion in aged mice. Nature 2017;
544:488–492
4 Monk TG, Price CC: Postoperative cog-
nitive disorders. Curr Opin Crit Care
2011; 17:376–381
5 Zhu S-H, Ji M-H, Gao D-P, et al.:
Association between perioperative
blood transfusion and early postopera-
tive cognitive dysfunction in aged
patients following total hip replace-
ment surgery. Upsala J Med Sci 2013;
119:262–267
6 Hoefer J, Luger M, Dal-Pont C, et al.:
The “Aging Factor” Eotaxin-1 (CCL11)
is detectable in transfusion blood
products and increases with the
donor’s age. Front Aging Neurosci
2017; 9:402
7 Clark IM, Swingler TE, Sampieri CL,
et al.: The regulation of matrix metallo-
proteinases and their inhibitors. Int J Bio-
chem Cell Biology 2008; 40:1362–1378
8 P
erez-Mart
ınez L, Jaworski DM: Tissue
inhibitor of metalloproteinase-2 pro-
motes neuronal differentiation by act-
ing as an anti-mitogenic signal. J
Neurosci 2005; 25:4917–4929
9 Sinno M, Biagioni S, Ajmone-Cat MA,
et al.: The matrix metalloproteinase
inhibitor marimastat promotes neural
progenitor cell differentiation into
neurons by gelatinase-independent
TIMP-2-dependent mechanisms. Stem
Cells Dev 2013; 22:345–358
10 Jaworski DM, Soloway P, Caterina J,
et al.: Tissue inhibitor of metallopro-
teinase-2(TIMP-2)-deficient mice dis-
play motor deficits. J Neurobiol 2005;
66:82–94
11 Rudolf R, Khan MM, Labeit S, et al.:
Degeneration of neuromuscular junc-
tion in age and dystrophy. Front
Aging Neurosci 2014; 6:99
12 Hopps E, Presti RL, Montana M, et al.:
Study of the correlations among some
parameters of the oxidative status,
gelatinases, and their inhibitors in a
group of subjects with metabolic syn-
drome. Mediat Inflamm 2014;
2014:510619
13 Rossignol P, Cambillau M, Bissery A,
et al. Influence of blood sampling pro-
cedure on plasma concentrations of
matrix metalloproteinases and their
tissue inhibitors. Clin Exp Pharmacol
Physiol 2008; 35:464–469
14 Lema^ıtre V, D’Armiento J: Matrix met-
alloproteinases in development and
disease. Birth Defects Res Part C
Embryo Today Rev 2006; 78:1–10
15 Baba Y, Yasuda O, Takemura Y, et al.:
Timp-3 deficiency impairs cognitive
function in mice. Lab Invest 2009;
89:1340–1347
16 Ould-yahoui A, Tremblay E, Sbai O,
et al.: A new role for TIMP-1 in mod-
ulating neurite outgrowth and mor-
phology of cortical neurons. PLoS One
2009; 4:e8289
17 Holten-Anderson MN, Brunner N,
Christensen IJ, et al.: Levels of tissue
inhibitor of metalloproteinases-1 in
blood transfusion components. Scand
J Clin Lab Invest 2009; 62:223–230
18 Gresele P, Falcinelli E, Sebastiano M,
et al.: Chapter four matrix metallopro-
teinases and platelet function. Prog
Mol Biol Transl 2017; 147:133–165
19 Tayebjee MH, Lip GY, Blann AD,
et al.: Effects of age, gender, ethnicity,
diurnal variation and exercise on cir-
culating levels of matrix

© 2020 The Authors.

Vox Sanguinis published by John Wiley & Sons Ltd on behalf of International Society of Blood Transfusion
Vox Sanguinis (2021) 116, 533–539
 TIMP-2 is detectable in human blood components 539

metalloproteinases (MMP)-2 and -9,
and their inhibitors, tissue inhibitors
of matrix metalloproteinases (TIMP)-1
and -2. Thromb Res 2005; 115:205–
210
20 Bonnema DD, Webb CS, Pennington
WR, et al.: Effects of age on plasma
matrix metalloproteinases (MMPs) and
tissue inhibitor of metalloproteinases
(TIMPs). J Card Fail 2007; 13:530–540
21 Momi S, Falcinelli E, Giannini S,
et al.: Loss of matrix metalloproteinase
2 in platelets reduces arterial thrombo-
sis in vivo. J Exp Medicine 2009;
206:2365–2379

© 2020 The Authors.

Vox Sanguinis published by John Wiley & Sons Ltd on behalf of International Society of Blood Transfusion
Vox Sanguinis (2021) 116, 533–539
 Vox Sanguinis (2021) 116, 540–546
© 2020 International Society of Blood Transfusion
ORIGINAL PAPER DOI: 10.1111/vox.13027

A microfluidic analysis of thrombus formation in

reconstituted whole blood samples comparing spray-dried
plasma versus fresh frozen plasma
Rachel S. Bercovitz,1,2,3,4 Caleb S. Drew,1 Chana L. Bushee,1 Mark A. Popovsky,5 Kenneth D. Friedman1,4 &
Waseem Q. Anani1,2
1
Blood Center of Wisconsin, Medical Sciences Institute, Medical College of Wisconsin, Milwaukee, WI, USA
2
Department of Pathology, Medical College of Wisconsin, Milwaukee, WI, USA
3
Department of Pediatrics, Children’s Hospital of Wisconsin, Milwaukee, WI, USA
4
Department of Medicine, Medical College of Wisconsin, Milwaukee, WI, USA
5
Velico Medical, Beverly, MA, USA

Received: 4 August 2020,
revised 20 October 2020,
accepted 21 October 2020,
published online 5 December 2020
Background Prompt resuscitation with plasma and other blood products reduces
trauma-related morbidity and mortality. Standard storage and preparation tech-
niques for frozen plasma limit its utility in the pre-hospital setting. Plasma can
be dehydrated using hot air (spray-dried plasma), stored at room temperature and
rehydrated quickly for use. The spray-dry process decreases high-molecular-
weight multimers of von Willebrand factor compared with conventional plasma.
The objective of this study was to compare platelet adhesion and thrombus for-
mation in a microfluidic perfusion assay facilitated by spray-dried compared with
frozen plasma using a non-inferiority design.
Study Design and Methods Whole blood was centrifuged to obtain red cell concen-
trate, and a platelet pellet that was suspended in either spray-dried or frozen plasma to
create recombined whole blood. Platelets were fluorescently labelled, and samples were
flowed through a collagen-coated microchannel. Surface area coverage by platelets and
thrombi was analysed and compared between each spray-dried and frozen plasma pair.
Results Compared with whole blood samples containing frozen plasma, samples
with spray-dried plasma had similar surface area coverage of platelets and
thrombi after 180 s of flow. Even when diluted with von Willebrand factor-free
plasma, there was no reduction thrombus formation.
Conclusion Spray-dried plasma is not inferior in supporting haemostasis com-
pared with fresh frozen plasma in a paired analysis. It offers advantages with
respect to portability and ease of preparation over frozen plasma in the pre-hos-
pital setting. This study supports development of clinical studies to evaluate the
efficacy and safety of spray-dried plasma in trauma patients.
Key words: component processing, plasma, spray drying, transfusion, trauma.

However, early resuscitation with blood products in heav-

Introduction
Traumatic haemorrhage represents a leading cause of pre-
ventable death in military and civilian settings [1,2].
Correspondence: Waseem Q. Anani, Blood Center of Wisconsin, Medical
Sciences Institute; Medical College of Wisconsin, PO Box 2178,
Milwaukee, WI 53201-2178, USA
E-mail: waseem.anani@blood.ca
ily bleeding patients decreases morbidity and mortality
[3,4]. The rapid restoration of coagulation factors is vital
during this process, since coagulopathy is associated with
a fourfold increase in mortality [5,6]. The shift to main-
tain a constant, ready-to-use plasma inventory for rapid
release in the blood bank encouraged the transition of
fresh frozen plasma (FFP) to thawed plasma and more
recently, liquid plasma and anhydrous plasma [7].

540

Unlike plasma in a frozen or liquid state, anhydrous
plasma or plasma stored as a powder has an extended
shelf life, can be reconstituted quickly, tolerates warmer
ambient temperatures and can be easily transported in
the pre-hospital setting [7,8]. Lyophilized plasma was
developed in the 1930s [9] and widely used in World War
II [7]. Its use fell out of favour after infectious disease
risks were identified, and it is not currently FDA
approved. However, the French armed services have a
long history of using lyophilized plasma [10]. While the
infectious disease risk has been mitigated, it still requires
specialized manufacturing at offsite locations.
Single donor spray-dried plasma (SpDP) removes water
by passing plasma microdroplets over heated (up to
150°C) flowing air into a collection bag [8]. It takes
approximately 25 min to spray dry a single donor unit of
plasma. This spray-dried plasma powder can be reconsti-
tuted in five minutes with sterile water. Compared with
FFP, SpDP has ≥80% of all pro- and anti-coagulant fac-
tors with the exception of Factor VIII, which is approxi-
mately 75% [11]. Additionally, endothelial inhibitory
effects on inflammation and vascular permeability were
equivalent in models comparing SpDP with FFP [12].
These results suggest that coagulation is not affected by
the reduced coagulation factor levels, likely because they
are still above the critical level of 30% [13,14]. Addition-
ally, both pro- and anti-coagulation proteins are
decreased, which may enable an appropriate haemostatic
balance.
The coagulation factor that was most affected by the
spray-drying process is von Willebrand Factor (VWF); the
high-molecular-weight multimers (HMWM) are affected
due to the heat drying process and are more susceptible
than smaller-sized multimers. HMWM are the most active
forms of VWF and a lack of HMWM can impair VWF
binding to glycoprotein (GP) Ib and collagen, which is
essential for platelet adhesion [15]. von Willebrand dis-
ease (VWD) is an inherited bleeding disorder character-
ized by qualitative or quantitative defects in VWF, which
is associated with defective platelet adhesion and aggre-
gation [16]. While VWD has an estimated prevalence of
1% in the population, Type 3 VWD (T3VWD), which is
autosomal recessive and characterized by a total or near-
total absence of VWF both in the plasma and cellular
compartments, has an estimated prevalence of 0
5–3/
1 000 000 [17]. Compared with normal controls, patients
with T3VWD have decreased thrombus formation under
flow conditions as measured by the Total Thrombus-For-
mation Analysis System (T-TAS, Fujimori Kogyo, Yoko-
hama, Japan) [18]. After receiving a prophylactic dose of
VWF-Factor VIII concentrate (median dose of VWF: 27
U/kg, range: 15–35 U/kg), thrombus formation improved
with respect to time to thrombus initiation, thrombus
© 2020 International Society of Blood Transfusion
Vox Sanguinis (2021) 116, 540–546
Spray-dried vs. fresh frozen plasma 541
volume at 10 and 30 min and but was still inferior to that
of controls [18].
The impact of the decreased HMWM of VWF in SpDP
on restoring and maintaining platelet adhesion and
thrombus formation is unknown. Therefore, in order to
simulate the worst-case scenario of a potential patient
with VWD being resuscitated with SpDP, we tested its
in vitro efficacy in a subject with T3VWD. The primary
objective of this study was to compare in vitro thrombus
formation in whole blood that has been reconstituted
with either SpDP or FFP. Secondary objectives were to
evaluate SpDP and FFP simulate conditions of VWD by
diluting SpDP and FFP with plasma from a subject with
T3VWD to evaluate efficacy of SpDP and FFP of support-
ing thrombus formation in conditions with decreased
total VWF.
Materials and methods
Replacing whole blood plasma with SpDP or FFP
Using a protocol approved by the local institutional
review board, whole blood from consented healthy adult
volunteers was collected into 3
2% sodium citrate (Bec-
ton, Dickinson and Company, Franklin Lakes, NJ, USA).
Donor FFP was collected and shipped to the Velico (Bev-
erly, MA, USA) for SpDP manufacturing [8,11,19]. Fol-
lowing collection, the whole blood (WB) was centrifuged
at 200 g for 12 min after which the platelet rich plasma
(PRP) was removed from the red blood cell concentrate
(RBCs). The PRP was diluted with 10 ml of 1,4-Piper-
azinediethanesulfonic acid (PIPES, Millipore Sigma, St.
Louis, MO, USA) saline solution, and 10 µl of prosta-
glandin E1 (PGE-1, Santa Cruz Biotechnology, Dallas,
TX, USA) was added (final concentration = 1 µM). The
platelet solution was then centrifuged at 1900 g for
8 min after which the supernatant was completely
removed and discarded. The platelets were resuspended
in 400 µl of either SpDP or FFP to form a platelet con-
centrate (PC). The RBCs, PC and SpDP or FFP were then
mixed together to make recombined whole blood (rWB,
SpDP-WB or FFP-WB). To control for the effect of
haematocrit in the microfluidic model, whole blood
reconstitution was designed to achieve a target haemat-
ocrit and platelet count [20]. The SpDP-WB and FFP-WB
samples were made to have a platelet count 150 000/µl–
275 000/µl and a haematocrit of 33–40%. Platelet count
and haematocrit were similar in the preparation of the
SpDP/FFP pairs.
SpDP was rehydrated with de-ionized water, and
sodium hydroxide was added to bring it to a pH between
6
4 and 6
7. Six SpDP and FFP pairs were created and
represented a batch. In order to create reconstituted whole
542 R. S. Bercovitz et al.
blood, red blood cells and platelets from 16 donors and
one type 3 vWD patient were combined with a batch
(paired SpDP and FFP). Each was tested at least twice
with repeated measures accounted for in the statistical
analysis. Each specific batch was reconstituted per manu-
facturer’s specifications to ensure parity with the FFP unit
that was made from the same donation. The volume of
rehydration fluid was dictated by the manufacturer to
produce a product an osmotically similar to the paired
FFP unit.
To dilute SpDP and FFP with VWF-free plasma, whole
blood was collected from an individual (blood type O)
with T3VWD. The plasma from a subject with T3VWD
(T3VWD-P) was obtained via two centrifugation steps
(300 g and 1900 g) that were performed to isolate platelet
rich and platelet poor plasma, respectively. This T3VWD-
P was subsequently frozen and stored at -20°C and
thawed in a 37°C shaking water bath prior to use. The
SpDP or FFP was diluted 25, 50 or 75% using the
T3VWD-P.
Microfluidic perfusion assay
Platelets in either unmanipulated WB, SpDP-WB or FFP-
WB samples were labelled with a GPIIbIIIa-specific mono-
clonal antibody AP2 (Hybridoma Core, BloodCenter of
Wisconsin [BCW], Milwaukee, WI, USA) conjugated to
Alexa Fluor (AF) 488 (Life Technologies Corp. Carlsbad,
CA, USA). While AP2 can inhibit platelet aggregation at
concentrations between 6–10 µg/ml, [21,22] the concen-
tration used in these experiments was 0
6 µg/ml, which
we found to have no effect on whole blood platelet
aggregometry, light transmission aggregometry or adhe-
sion under flow (data not shown).
The microchannels (Vena8Fluoro + chips, Cellix,
Dublin, Ireland) were coated overnight at 4°C with type I
collagen (Chrono-log Corp, Havertown, PA, USA) 50 µg/
ml diluted in 20 mM glacial acetic acid. WB, SpDP-WB
and FFP-WB were flowed through the channels using the
VenaFlux (Cellix) microfluidic platform at arterial shear
rate of 1500/s for 180 s. Platelet adhesion and aggrega-
tion in traumatic injury models can be confounded by
lower shear rates (capillary and venous) in a microfluidic
device since platelet adhesion may occur independent of
the model. Thus, arterial shear rates were selected to pro-
vide the most representative flow rates in trauma and
decrease the likelihood of platelet adhesion naturally
occurring [23,24]. The surface area (SA) coverage was
evaluated by taking 15 still images over the length of the
microfluidic channel following completion of flow.
Images of adherent platelets and thrombi were captured
in order to calculate SA along the length of the channel.
Images were processed using ImageJ [25]; ImageJ is an
open source software platform as an image-processing
programme designed for scientific multidimensional
images to perform a wide variety of dimensional analy-
ses.
Images from the SpDP-FFP pairs were processed identi-
cally in order to calculate the SA of the platelets and
thrombi (Fig. 1). Fluorescent images were background
corrected, and threshold levels were set for the GPIIbIIIa
positive platelets in each rWB sample using provided
ImageJ thresholding algorithms. Images were then con-
verted to binary masks, which resulted in individual pla-
telets and thrombi. Surface area covered by platelets and
thrombi as well as the number of objects were obtained
from the ‘Analyze Particles’ toolbox in ImageJ. Images
processing and analysis for each image was performed by
at least two researchers blinded to the samples.
Statistical analyses
The geometric mean of the SA from the still images was
calculated, because it is less influenced by outliers and
variability. The geometric mean was also used in the cal-
culation of the ratio paired t test, calculating the differ-
ence between SpDP and FFP as SpDP/FFP as compared
to SpDP—FFP. The pre-determined margin of non-inferi-
ority was 20% (SpDP/FFP >0
8) based on the FDA’s tar-
get of SpDP being – 20% of control. A total of 16 pairs
were needed to have a power of 0
9 at the one-sided
alpha of 0
05. Prism 6
0h (GraphPad, La Jolla, CA, USA)
software was used for the analyses, and PASS15 (NCSS,
LLC, Kaysville, UT, USA) was used for the power calcula-
tion.
Results
SpDP compared with FFP
Six batches of SpDP/FFP were evaluated using 17 sub-
jects. There was no statistical difference between the
SpDP/FFP pairs (P = 0
7558). The mean ratio of SpDP/
FFP was 1
21 with a 95th per cent confidence interval
(CI) of 0
84–1
57, which means that the there is a 95%
chance that the ‘true’ ratio is greater than the 0
8 ratio
that had been a priori determined. Thus, SpDP would be
considered not inferior to the FFP with respect to sup-
porting platelet adhesion and thrombus formation in
whole blood samples (Fig. 2). A two-way ANOVA analysis
demonstrated batch (SpDP and FFP pairs) did not signifi-
cantly affect platelet adhesion and thrombus formation in
the surface area calculation.
Of the 17 samples run, 4 (23
5%) were below the 0
8
ratio (range 0
26–0
66). Two of the four samples less than
0
8 were from the same SpDP sample. Figure 1 is a
© 2020 International Society of Blood Transfusion
Vox Sanguinis (2021) 116, 540–546
 Spray-dried vs. fresh frozen plasma 543

Fig. 1 Representative still images of thrombus formation in the microfluidic perfusion channels after 180 s of flow. (a-c) recombined whole blood (rWB)
made with fresh frozen plasma (FFP). (d-f) rWB made with spray-dried plasma (SpDP). Images (a) and (d) are the originals; images (b) and (e) have been
background corrected using the same Image J algorithm; images (c) and (f) have been converted to a binary mask using the Image J algorithm. Surface
area (SA) coverage by platelets and thrombi computed by Image J using the binary image. SA in rWB-FFP = 136 495 and in rWB-SpDP = 157 682 (ratio
SpDP/FFP = 1.16).

representative sample of the images taken using rWB

made with SpDP and FFP from the same batch.
Dilution with T3VWD-P
Compared with the SA coverage by platelets and thrombi
from whole blood collected from the subject with
T3VWD, SpDP-WB and FFP-WB had more surface area
coverage by platelets and thrombi. SpDP-WB compared
to the type 3 vWD sample had 1
6–4
8-fold higher sur-
face area coverage (median = 3
1-fold) and FFP had 1
9–
4
5-fold higher surface area coverage (median = 3
3-fold).
The small sample size (n = 3) precludes the ability to per-
form a statistical analysis.
In SpDP and FFP samples diluted with 50% T3VWD-P,
median ratio of the diluted sample over the undiluted
sample was 0
89 for both SpDP and FFP (range: 0
16–
0
92, SpDP and 0
10–1
16, FFP) (Fig. 3a). One batch was
responsible for the low ratios in both the diluted SpDP
and FFP samples. Of note, this was the same batch that
was responsible for 50% of the samples where SpDP was
<80% of FFP at 100% concentration. In whole blood sam-
ples reconstituted with SpDP diluted with T3VWD-P to
75%, 50 and 25%, only the 25% SpDP sample had
decreased SA coverage compared with the undiluted
SpDP (Fig. 3b).
Discussion
There are several issues that limit the speed and ease
with which frozen plasma can be administered in the
pre-hospital setting. The weight and size of plasma
units limit transportability and the machines such as
freezers, refrigerators and thawing equipment need to
be kept in a central location, limiting the access of
austere sites to these resources [10]. This study
investigated the in vitro haemostatic potential of SpDP
when compared to FFP, and we have demonstrated that
it is not inferior with respect to supporting platelet
adhesion and thrombus propagation under arterial flow
conditions in a collagen-coated microchannel. On aver-
age, SpDP-WB had 20% higher surface area coverage
compared with FFP-WB. The lower limit of the 95th
per cent CI is a difference of 16%, which is less than
the a priori determined margin of non-inferiority of
20%.
A primary concern that SpDP might be inferior to FFP
is due to decreased HMWM of VWF, likely due to the
temperatures required during the drying process. How-
ever, even when diluted 50% by VWF-free plasma, there
was no reduction in surface area coverage compared to
the undiluted SpDP. These results were similar to diluted
FFP samples. These results suggest that SpDP can support

© 2020 International Society of Blood Transfusion

Vox Sanguinis (2021) 116, 540–546
544 R. S. Bercovitz et al.

Fig. 2 Ratio of SPDP/FFP surface area coverage of platelets and thrombi after flow through microfluidic channels. Symbols represent the different
batches (SPDP-FFP pairs) reconstituted with unique red blood cell and platelet donors. A repeated symbol represents the same batch with a different
donor. The solid grey line indicates ratio of 1, and the grey dashed line is thea prioridetermined level of non-inferiority (0
8). The solid black line repre-
sents the mean of the data and is above the threshold for non-inferiority; SpDP is non-inferior to FFP in platelet adhesion and thrombus formation.

haemostasis in individuals with decreased levels of VWF
as well as FFP.
Lyophilized plasma has been used for decades in the
battlefield; however, it is limited to production by spe-
cialized companies rather than at blood centres. SpDP
can be made rapidly from single donor plasma units
and the spray-drying process can be done at local
blood centres ad hoc. As with other plasma-based and
cellular products, SpDP is subject to the inter-donor
variability, which may have an impact on its efficacy.
One batch of the SpDP (denoted by the ■ in Figs. 2
and 3) was consistently inferior compared to FFP pro-
duced from the same donor. Additionally, FFP from this
same donor demonstrated a minimal support of
thrombus formation when diluted 50% by T3VWD-P.
This donor had normal levels of VWF; however, it is
possible that the VWF multimer distribution made this
donor’s plasma more susceptible to degradation during
the drying process. A potential mitigation for this issue
as production of SpDP moves forward would be the
screening of donor multimers.
In combination with other pre-clinical studies [12,19]
that have shown SpDP to be equivalent to FFP
with respect to thrombin generation and endothelial
interactions, this study supports the development of in-
human studies to evaluate the efficacy and safety of
SpDP in preventing and reversing trauma-related coag-
ulopathy.

© 2020 International Society of Blood Transfusion

Vox Sanguinis (2021) 116, 540–546
 Spray-dried vs. fresh frozen plasma 545

Fig. 3 Thrombus formation in samples using SpDP or FFP diluted with plasma from an individual with Type 3 von Willebrand Disease (VW factor-free
plasma, T3VWD-P). (a) Thrombus formation in samples of recombined whole blood (rWB) made with SpDP or FFP diluted by 50% with plasma T3VWD-P.
The solid grey line indicates ratio of 1, and the grey dashed line is thea prioridetermined level of non-inferiority (0.8). (b) Thrombus formation in rWB
made with SpDP diluted with 25, 50 and 75% T3VWD-P. Both 75 and 50% SpDP had the same surface area coverage as 100% SpDP, whereas 25% SpDP
was decreased (n = 3).

Funding Conflicts of interest

Biomedical Advanced Research and Development Author- RSB, CSD, CLB, WQA and KDF have no financial conflicts
ity provided Velico Medical with funding for this of interest to report. MAP is an employee of Velico Medi-
research. cal, a manufacturer of spray-dried plasma.

References
1 Eastridge BJ, Mabry RL, Seguin P,
et al.: Death on the battlefield (2001–
2011): implications for the future of
combat casualty care. J Trauma Acute
Care Surg 2012; 73:S431–S437
2 Tisherman SA, Schmicker RH, Brasel
KJ, et al.: Detailed description of all
deaths in both the shock and traumatic
brain injury hypertonic saline trials of
the Resuscitation Outcomes Consor-
tium. Ann Surg 2015; 261:586–590
3 Borgman MA, Spinella PC, Perkins JG,
et al.: The ratio of blood products
transfused affects mortality in patients
receiving massive transfusions at a
combat support hospital. J Trauma
2007; 63:805–813
4 Holcomb JB, Tilley BC, Baraniuk S,
et al.: Transfusion of plasma, platelets,
and red blood cells in a 1:1:1 vs a
1:1:2 ratio and mortality in patients
with severe trauma: the PROPPR ran-
domized clinical trial. JAMA 2015;
313:471–482
5 Hess JR, Brohi K, Dutton RP, et al.:
The coagulopathy of trauma: a review
of mechanisms. J Trauma 2008;
65:748–754
6 Brohi K, Singh J, Heron M, et al.:
Acute traumatic coagulopathy. J
Trauma 2003; 54:1127–1130
7 Pusateri AE, Given MB, Schreiber MA,
et al.: Dried plasma: state of the science
and recent developments. Transfusion
2016; 56(Suppl 2):S128–S139
8 Booth GS, Lozier JN, Nghiem K, et al.:
Spray: single-donor plasma product
for room temperature storage. Transfu-
sion 2012; 52:828–833
9 Flosdorf EW, Mudd S: Procedure and
apparatus for preservation in “lyo-
phile” form of serum and other biolog-
ical substances. J Immunol 1935;
29:389–425
10 Sailliol A, Martinaud C, Cap AP, et al.:
The evolving role of lyophilized
plasma in remote damage control
resuscitation in the French Armed
Forces Health Service. Transfusion
2013; 53(Suppl 1):65S–71S
11 Liu QP, Carney R, Sohn J, et al.: Sin-
gle-donor spray-dried plasma. Trans-
fusion 2019; 59:707–713
12 Wataha K, Menge T, Deng X, et al.:
Spray-dried plasma and fresh frozen
plasma modulate permeability and
inflammation in vitro in vascular
endothelial cells. Transfusion 2013; 53
(Suppl 1):80S–90S
13 Rizoli SB, Scarpelini S, Callum J,
et al.: Clotting factor deficiency in
early trauma-associated coagulopathy.
J Trauma 2011; 71:S427–S434
14 Tanaka KA, Key NS, Levy JH: Blood
coagulation: hemostasis and thrombin
regulation. Anesth Analg 2009;
108:1433–1446
15 Schneppenheim R, Budde U: von
Willebrand factor: the complex molec-
ular genetics of a multidomain and
multifunctional protein. J Thromb
Haemost 2011; 9(Suppl 1):209–215

© 2020 International Society of Blood Transfusion

Vox Sanguinis (2021) 116, 540–546
546 R. S. Bercovitz et al.
16 Leebeek FW, Eikenboom JC: Von
Willebrand’s disease. N Engl J Med
2016; 375:2067–2080
17 Rodeghiero F, Castaman G, Dini E:
Epidemiological investigation of the
prevalence of von Willebrand’s dis-
ease. Blood 1987; 69:454–459
18 Agren A, Holmstrom M, Schmidt DE,
et al.: Monitoring of coagulation fac-
tor therapy in patients with von Wille-
brand disease type 3 using a
microchip flow chamber system.
Thromb Haemost 2017; 117:75–85
19 Meledeo MA, Liu QP, Peltier GC, et al.:
Spray-dried plasma deficient in high-
molecular-weight multimers of von
Willebrand factor retains hemostatic
properties. Transfusion 2019; 59:714–
722
20 Bercovitz RS, Brenner MK, Newman
DK: A whole blood model of thrombo-
cytopenia that controls platelet count
and hematocrit. Ann Hematol 2016;
95:1887–1894
21 Sheu JB, Ko WC, Hung WC, et al.: Inter-
action of thrombin-activated platelets
with extracellular matrices (fibronectin
and vitronectin): comparison of the activ-
ity of Arg-Gly-Asp-containing venom
peptides and monoclonal antibodies
against glycoprotein IIb/IIIa complex. J
Pharm Pharmacol 1997; 49:78–84
22 Fukuyama M, Sakai K, Itagaki I, et al.:
Continuous measurement of shear-
© 2020 International Society of Blood Transfusion
induced platelet aggregation. Thromb
Res 1989; 54:253–260
23 Ruggeri ZM, Mendolicchio GL: Adhe-
sion mechanisms in platelet function.
Circ Res 2007; 100:1673–1685
24 Philipose S, Konya V, Sreckovic I,
et al.: The prostaglandin E2 receptor
EP4 is expressed by human platelets
and potently inhibits platelet aggrega-
tion and thrombus formation. Arte-
rioscler Thromb Vasc Biol 2010;
30:2416–2423
25 Rasband WS: ImageJ. Bethesda, MD:
NIH, 1997–2015. https://imagej.nih.
gov/ij/. [Last accessed on 2017
November 15]
Vox Sanguinis (2021) 116, 540–546
 Vox Sanguinis (2021) 116, 547–556
© 2020 International Society of Blood Transfusion
ORIGINAL PAPER DOI: 10.1111/vox.13039

Quality assessment of red blood cell suspensions derived

from pathogen-reduced whole blood
Irina Kumukova,1 Pavel Trakhtman,1 Nicolay Starostin,1 Daria Borsakova,1,2 Anastasia Ignatova1 &
Yana Bayzyanova1
1
Dmitry Rogachev National Medical Research Center of Pediatric Hematology Oncology and Immunology, Moscow, Russia
2
Laboratory of Physiology and Biophysics of the Cell, Center for Theoretical Problems of Physicochemical Pharmacology, Russian Academy of
Sciences, Moscow, Russia

Background We used laboratory indicators to evaluate the quality of pathogen-

Received: 24 July 2020,
revised 6 November 2020,
accepted 9 November 2020,
published online 22 November 2020
reduced red blood cell suspension (RBCS) compared with gamma-irradiated
RBCS.
Materials and methods To determine biochemical and metabolic parameters of
RBCS, we obtained 50 whole blood units from healthy volunteers and random-
ized them into 2 groups: 25 were pathogen-reduced, and then, RBCS prepared
from them. RBCS from the other 25 was gamma-irradiated. Sampling was carried
out on day zero before and after treatment and at 7, 14, 21 and 28 days. To
determine lymphocyte inactivation, we collected another 35 whole blood units.
Each was sampled to form 3 study groups: untreated, gamma-irradiated and
pathogen-reduced. Daily sampling was carried out during 3 days of storage.
Results The quality of RBCS from both groups was largely the same, except for
haemolysis and red blood cell fragility, which were more pronounced in the
pathogen-reduced group. This finding limited the shelf life of pathogen-reduced
RBCS to 14 days. Lymphocyte viability was significantly reduced after both
treatments. Proliferation of lymphocytes after pathogen reduction was reduced to
the detection limit, while low-level proliferation was observed in gamma-irradi-
ated samples.
Conclusion Pathogen-reduced red blood cells have acceptable quality and can be
used for transfusion within 14 days. Results of inactivation of lymphocytes
demonstrate that pathogen reduction technology, applied on WB, can serve as an
alternative to irradiation.
Key words: blood component production, blood processing, blood safety,
haemovigilance.

excluded [4]. There are many reasons for this: the

Introduction
Transmission of pathogens through blood transfusion is
still of great concern [1], and various methods have been
developed to prevent it [2–3]. Despite the successes
achieved, the risk of infection transmission is not
*Correspondence: Irina Kumukova, 13 Kurizova str., ap. 70,
Domodedovo, Russia
E-mail: irina_kumukova@mail.ru
Sets “Mirasol PRT” for the reduction of pathogens were provided by Te-
rumo BCT, Lakewood, CO, USA.
seronegative period, new pathogens, infectious diseases
occurring in non-endemic areas where they are not
screened out, the ever-present threat of bacterial contami-
nation and technical errors [5–7]. According to a mathe-
matical model, every 5 years, ‘new’ infectious agents
appear that can be transmitted by transfusion [8]. The lat-
est such challenge to health security, including, poten-
tially, transfusion safety, is SARS-CoV-2 the causative
agent of COVID-19.
The immune complications of blood products are still a
cause for concern. Leukoreduction reduces, but does not

547
548 I. Kumukova et al.
eliminate donor white blood cell (WBC) side effects [9–
11]. Irradiation prevents transfusion-associated graft ver-
sus host disease, but does not affect other side effects of
WBC presence such as febrile non-haemolytic transfusion
reactions and post-transfusion immunomodulation [12].
The only effective method for the prevention of infec-
tious disease transmission in the seronegative period is
the reduction of pathogens in blood products. Such a sys-
tem is also potentially effective for the prevention of
untestable and new diseases and the adverse effects of
donor WBC [12–14]. Several pathogen reduction tech-
nologies (PRT) are used in routine practice in many coun-
tries for the treatment of platelet products and plasma.
However, it was difficult to develop a PRT for the pro-
cessing of red blood cell blood products, because of the
physical characteristics of red blood cells (RBC) All the
same RBC-containing blood products are the most used in
clinical practice [13]. That is why there is great interna-
tional interest in the area of red cell pathogen reduction.
One PRT for whole blood (WB) can provide three
pathogen-reduced blood products. It is based on the com-
bined action of riboflavin and ultraviolet (RF + UV)
(Mirasol PRT, Terumo BCT, Lakewood, CO, USA).
Studies to evaluate the reduction of pathogen in WB
treated by RF + UV have shown promising results [14–
16], without toxicity in vitro and in animals [17,18].
Studies in humans also showed no safety problem and
transfusion criteria met FDA requirements [19,20].
There has been extensive experience with RF + UV
technology on platelets and plasma. In addition, pub-
lished data describe the components derived from WB
treated by RF + UV [21,22]. Our interest focused on
pathogen-reduced red blood cell suspension (RBCS).
Although some have studied RF + UV-treated whole
blood, our results show how the technology affects red
blood cells and may interest others considering imple-
mentation of the technology. We conducted a randomized
prospective parallel study. The study included 2 stages. In
the first stage, we determined the biochemical and meta-
bolic parameters during the storage of the RBCS obtained
from RB + UV-treated WB (PR-RBCS) and compared these
data with gamma-irradiated RBCS (GI-RBCS), which was
the control group. In the second stage, we examined the
viability and proliferation of lymphocytes in pathogen-re-
duced WB (PR-WB) compared with gamma-irradiated (GI-
WB) and untreated WB (UT-WB).
Materials and methods
The study protocol was reviewed and approved by the
local Ethical Committee of Dmitry Rogachev National
Medical Research Center of Pediatric Hematology Oncol-
ogy and Immunology and the applicable Russian
Government Medical agency. Healthy donors who agreed
to participate in the study gave WB as per standard col-
lection practice (volume of 450 – 20 ml in bags with
63 ml of citrate-phosphate-dextrose anticoagulant (IMU-
FLEX-WB-RP, Terumo, Belgium)). To determine RBCS
storage results, we randomized 50 WB units into two
groups: pathogen-reduced and the gamma-irradiated con-
trol group. All were leukofiltered one hour after the dona-
tion (IMUFLEX-WB-RP, Terumo, Belgium). WB units
assigned to the study group (PR-RBCS; n = 25) were
pathogen-reduced at room temperature with RF + UV
(Mirasol PRT, Terumo BCT, Lakewood, CO, USA) as per
the treatment protocol provided by the manufacturer. WB
was transferred to an illumination bag with 35 ml of
riboflavin solution (500 lM) added and then exposed to
ultraviolet light (wavelength 280–315 nm) for 42–64 min
depending on the volume of the product (total dose 80 J/
ml RBC). WB was then transferred to a storage bag and
separated manually into RBC and plasma after a hard
spin centrifugation at 5000 g for 15 min at 4°C. The stor-
age bags of the PRT system have a non-standard size,
and therefore, they do not fit into the automatic extrac-
tor. One hundred ml of SAGM (Terumo, Tokyo, Japan)
was added to the RBC to prepare RBCS. From the units of
WB assigned to the control group, RBCS was prepared
according to a standard procedure (hard spin 5000 g,
15 min, t = 4°C) with separation using an automatic
plasma extractor (NOVOmatic LMB Technologie, Ger-
many) into RBC and plasma. RBC was suspended in
100 ml of SAGM (saline-adenin-glucose-mannitol) (Ter-
umo, Japan) and immediately subjected to c-irradiation
(Gammacell 3000 Elite, Best Theratronics, Canada) at a
dose of 25 Gy. RBCS were stored at a controlled tempera-
ture of 4 – 2°C for 28 days. Samples were taken from all
WB samples (before treatment) after leukofiltration, as
well as from RBCS on day 0, 7, 14, 21 and 28 of storage.
On day 28, only bacterial culturing was performed.
The residual WBC count in the filtered WB and RBCS
was calculated using the BD LeucocountTM Kit (BD Bio-
sciences, USA) on a BD FACSCantoTM II flow cytometer
(BD Biosciences, USA) with BD FACS Diva Software. Hae-
moglobin concentration (Hb) and haematocrit (Ht) of
RBCS were assessed on the Sysmex KX-21N (Sysmex Cor-
poration Japan). Potassium concentration was determined
on the Cobas c311 (Roche/Hitachi). The pH, glucose and
lactate concentrations were assessed using an
ABL800FLEX (Radiometer, Denmark). The free Hb was
assessed in supernatant, obtained from a sample after
centrifugation at 1400 rpm for 2 min and measured using
a HemoCue Plasma/Low Hb (HemoCue AB, Angelholm,
Sweden). Haemolysis was calculated by:
%Haemolysis = [(1-Ht) 9 Free Hb g/dL) 9 100]/ Total
Hb concentration (g/dL).
© 2020 International Society of Blood Transfusion
Vox Sanguinis (2021) 116, 547–556

Examination for sterility was carried out after treat-
ment on day 0 and day 28 using BACTEC Peds Plus/F
culture vials (Becton Dickinson) and a BACTEC-9050 bac-
teriological analyser (Becton Dickinson).
The osmotic resistance was measured using hypotonic
sodium chloride solutions with different osmolality as
described in Shcherbachenko I [23] on a plate spectropho-
tometer Anthos Zenyth 340rt (Biochrom, United King-
dom). The osmotic resistance (Hc50) was characterized by
the osmolality of the solution, at which 50% of the origi-
nal RBC were lysed. The width of the osmotic fragility
distribution (WD) was characterized by the difference of
the osmolality, at which 10% and 90% of the original
RBC were lysed.
Adenosine triphosphate (ATP) measurement
Perchloric extracts were obtained: 0
25 ml of WB or mix-
ture of 0
125 ml of RBCS and 0
125 ml of PBS were
mixed with 0
5 ml of 0
5 M perchloric acid, and then,
70 ll of a 5M K2CO3 was added. The sample was cen-
trifuged for 5min at 1000g. The supernatant was removed
and frozen at 80°C. ATP was measured as described in
Hans A [24] on the plate spectrophotometer.
Reduced glutathione (GSH) measurement
Samples (WB or RBCS) were fast frozen and stored at -
80°C. Then, samples were thawed and protein-free
extracts were obtained: to 100 ll of a sample, 1 ml of
distilled water and 1 ml of meta-phosphoric acid
(16
7 mg/ml meta-phosphoric acid, 2 mg/ml EDTA,
300mg/ml NaCl) were added. Extracts were mixed thor-
oughly for 2min and centrifuged for 5 min at 1000 g.
The supernatant was removed, and measurement was per-
formed as described in Pradedova E [25] on a plate spec-
trophotometer.
Phosphatidylserine expression (PS) measurement
Samples were diluted 1:1000 with HEPES buffer (10 mM
HEPES; 150 mM NaCl; 2
5 mM CaCl2; pH 7
4; Sigma-
Aldrich, USA)). 2 µL of Annexin V (Sony Biotechnology,
USA) was added to the 40 µL mixture and incubated for
10 min at room temperature. Then, the samples were
diluted by adding a 360 µL buffer and analysed on a
Novocyte (Acea Bioscience, USA).
For the second stage of the study for determination of
the proliferation and viability of lymphocytes, we used
another 35 samples of unfiltered WB obtained from
healthy donors (450 ml of WB (–10%) + 63 ml of CPD).
Leukofiltration was not performed in this stage. Each
sample was divided into three aliquots (10 ml untreated
© 2020 International Society of Blood Transfusion
Vox Sanguinis (2021) 116, 547–556
Pathogen-reduced red blood cells 549
group, 10 ml for gamma irradiation and remaining vol-
ume for pathogen reduction), and thus, three groups were
formed: untreated (UT-WB, n = 35), gamma-irradiated
(GI-WB, n = 35) and pathogen-reduced samples (PR-WB,
n = 35). Gamma irradiation and pathogen reduction were
performed on the day of preparation as described above.
For assessment of proliferative activity, 1 9 107
mononuclear cells were isolated on a density gradient
and treated by CFSE (Carboxy-fluorescein-succinimidyl-
ester) in 1 ml of RPMI (Roswell Park Memorial Institute)
medium, for 5 min at room temperature in the dark. The
amount of the dye was 0
5 ll. Then, the cells were
washed twice with PBS with 10% foetal bovine serum
(FBS) (STEMCELL Technologies, USA). One hundred
microlitres labelled cells were added at a number of
5 9 104 cells with the addition of 100 ll RPMI (negative
control) or 10 lg/ml phytohemagglutinin (PHA).
Untreated WB served as a positive control. The incubation
lasted three days. To determine the lymphocytes viability,
7-AAD was added to the CFSE-labelled cells, which pene-
trates apoptotic cells. The measurements were carried out
on a BD FACSCanto II flow cytometer (Becton Dickinson,
USA) using Diva software.
Statistical analysis was performed using XLSTAT for
Microsoft Excel 2010 and QI Macros 2018. The normality
of the distribution was estimated by the Shapiro–Wilk
test. If the data were not normal distributed, the Box–Cox
transformation was applied. The level of significance in
all calculations was taken to be a = 0
05.
In the first stage of the study, WB and RBCS parame-
ters were compared using the unpaired Student t-test or
the Mann–Whitney z-test, when data cannot be trans-
formed. Comparison of lymphocytes proliferation and
viability was performed using the Kruskal–Wallis H-test.
Results
RBC parameters
Whole blood prior to treatment had comparable values in
both groups (Table 1). The residual WBC count after
leukofiltration in all samples was lower than 1 9 106/
Unit and did not differ between the compared groups
both in WB and in RBCS on day 0. Volume, Hb and Ht of
WB and RBCS met the requirements of the orders of the
Ministry of Health and the Government of the Russian
Federation in relation to donated blood products. The
mean – SD of all the studied values of RBCS at different
storage days are presented in Table 2.
The Hb in the RBCS did not change during storage and
did not differ between groups. The Ht was significantly
increased during storage in the GI-RBCS on day 14
(p < 0
0001) and in the PR-RBCS on day 7 (p = 0
02),
550 I. Kumukova et al.

Table 1 Initial variables of leukofiltered WB samples from study (PR) and control (GI) groups before treatment.1

Variable PR (N = 25) GI (N = 25) p-value Variable PR (N = 25) GI (N = 25) p-value

Volume, ml 509
3 – 8
4 501
2 – 40
2 0
97 pH 7
06 – 0
06 7
05 – 0
04 0
66
Residual WBC, x106/Unit 0
15 – 0
19 0
12 – 0
17 0
47 К+, mmol/L 3
2 – 0
3 3
1 – 0
3 0
34
Hb, g/L 123
8 – 10
6 122
3 – 12
7 0
90 Glucose, mmol/L 20
9 – 0
9 20
8 – 1
3 0
90
Ht, % 37
3 – 2
8 36
9 – 3
1 0
60 Lactate, mmol/L 4
2 – 0
9 4
0 – 0
7 0
26
Free Hb, g/L 0
0 – 0
1 0
0 – 0
1 0
91 ATP, mmol/L RBC 1
02 – 0
23 1
00 – 0
28 0
76
Haemolysis, % 0
01 – 0
02 0
01 – 0
03 1 GSH, mmol/L RBC 2
58 – 0
41 2
49 – 0
49 0
62
H50, mOsm/L 137
1 – 5
9 134
9 – 5
1 0
32 PS expression, % 0
28 – 0
25 0
27 – 0
20 0.99
WD, mOsm/L 41
7 – 14
7 40
7 – 20
9 0
77

p-values were obtained by the Student t-test or Mann–Whitney U-test depending on the distribution law.
WBC = white blood cell, Hb = haemoglobin concentration, Ht = haematocrit, Free Hb = free haemoglobin; Hc50 = osmotic resistance, WD = width of
the osmotic fragility distribution, K+ = concentration of extracellular potassium, ATP = adenosine triphosphate, GSH = reduced glutathione, PS = phos-
phatidylserine.
1
For all parameters, the mean – standard deviations are presented.

Table 2 Variables changes during storage in study (PR-RBCS) and control (GI-RBCS) groups.1

Day 0 Day 7 Day 14 Day 21

Variable GI-RBCS PR-RBCS GI-RBCS PR-RBCS GI-RBCS PR-RBCS GI-RBCS PR-RBCS

Volume, ml 307
2 – 16
7 300
0 – 14
4 297
2 – 16
7 290
8 – 14
8 292
2 – 16
7 285
8 – 14
8 287
2 – 16
7 280
8 – 14
8
Hb, g/L 197
8 – 11
7 197
9 – 9
8 198
3 – 11
7 198
6 – 9
5 199
0 – 10
4 198
8 – 9
8 199
2 – 11
7 198
3 – 9
4
Ht, % 59
0 – 2
1 59
5 – 2
3 59
5 – 2
3 61
2 – 2
7* 61
9 – 2
3 64
4 – 3
0* 63
3 – 2
4 66
5 – 3
1*
pH 7
00 – 0
05 7
00 – 0
05 6
77 – 0
04 6
76 – 0
21* 6
67 – 0
03 6
63 – 0
04* 6
57 – 0
05 6
55 – 0
03
К+, mmol/L 2
0 – 0
2 2
0 – 0
3 45
0 – 4
1 45
2 – 4
8 51
6 – 5
3 51
7 – 5
1 65
1 – 4
6 64
7 – 4
2
Free Hb, g/L 0
3 – 0
2 0
1 – 0
1 0
9 – 0
3 0
9 – 0
6 1
5 – 0
4 2
2 – 1
0* 2
0 – 0
6 4
4 – 2
2*
Haemolysis, % 0
08 – 0
13 0
02 – 0
03 0
19 – 0
06 0
17 – 0
09 0
28 – 0
07 0
38 – 0
12* 0
36 – 0
10 0
72 – 0
37*
ATP, mmol/L RBC 1
12 – 0
38 1
19 – 0
33 1
09 – 0
26 1
14 – 0
23 1
08 – 0
24 1
02 – 0
21 0
90 – 0
26 0
81 – 0
24
GSH, mmol/L RBC 2
33 – 0
35 2
34 – 0
35 2
38 – 0
52 2
27 – 0
35 2
16 – 0
39 2
10 – 0
30 1
98 – 0
34 1
96 – 0
33
PS expression, % 0
22 – 0
11 0
23 – 0
17 0
29 – 0
22 0
24 – 0
16 0
45 – 0
42 0
32 – 0
27 0
21 – 0
14 0
51 – 0
5*
Glucose, mmol/L 28
3 – 1
2 26
9 – 1
1* 23
7 – 1
4 22
0 – 1
5* 20
4 – 1
4 19
0 – 1
5* 17
9 – 2
0 16
8 – 1
8*
Lactate, mmol/L 3
4 – 0
5 3
1 – 0
5 12
3 – 1
5 9
5 – 2
0* 18
8 – 2
2 15
4 – 2
9* 22
0 – 1
9 19
9 – 3
1*
H50, mOsm/L 138
8 – 4
3 140
3 – 5
1 136
4 – 4
9 142
2 – 7
1* 138
6 – 4
9 148
4 – 8
2* 141
2 – 5
4 155
3 – 8
4*
WD, mOsm/L 32
8 – 9
5 36
3 – 9
1 60
4 – 19 56
3 – 22
1 64
5 – 15
4 68
8 – 22
3 73
7 – 23
3 74
5 – 17
7

p-values were obtained by the Student t-test or Mann–Whitney U-test depending on the distribution law.
Hb = haemoglobin concentration, Ht = haematocrit, K+ = concentration of extracellular potassium; Free Hb = free haemoglobin, ATP = Adenosine
triphosphate, GSH = reduced glutathione, PS = phosphatidylserine, Hc50 = osmotic resistance, WD = width of the osmotic fragility distribution.
1
For all parameters, the mean – standard deviations are presented.
*Differences between the study and the corresponding control group are statistically significant.

compared with on day 0. Significant differences between
the groups were detected from day 7 (p = 0
03) and per-
sisted until the end of the observation period (on day 14
p = 0
002, on day 21 p = 0
0002).
The free Hb and the haemolysis level (Fig. 1A) sig-
nificantly increased over storage in both groups and
significantly differed between the groups from the
14th day of storage (for free Hb, p = 0
0003, for
haemolysis, p = 0
0004). Following 21 days of storage
in 9 PR-RBCS (36%), percentage of haemolysis was
equal or exceeded 0
8%. In the GI-RBCS, haemolysis
was below the threshold level for the entire observa-
tion period.
The extracellular potassium concentration increased
significantly over storage, but did not differ between
groups. The pH progressively decreased in both groups.
Statistically significant differences between groups in pH
were found on days 7 (p = 0
02) and 14 (p = 0
002) of
storage with lower level in the PR-RBCS, but by the 21st
day these differences were levelled (p = 0
1).

© 2020 International Society of Blood Transfusion

Vox Sanguinis (2021) 116, 547–556
 Pathogen-reduced red blood cells 551

Figure 1 Changes during the storage of the haemolysis level (A), osmotic resistance (B) and phosphatidylserine expression (C) of pathogen-reduced (PR)
and gamma-irradiated (GI) samples in the first stage of the study. The boundaries of the box are the 25th and 75th percentiles, respectively. The line in
the middle of the box is the median. The ends of the whiskers are the edges of a statistically significant sample (no outliers). Crosses indicate outliers.
*Differences between the groups are statistically significant (p < 0
05). p-values were obtained by the Student t-test or Mann–Whitney U-test depend-
ing on the distribution law. [Colour figure can be viewed at wileyonlinelibrary.com]

The extracellular glucose concentration at day 0 was
significantly higher in the control group (p = 0
0002).
These differences persisted throughout storage (day 7
p < 0
0001; day 14 p = 0
001; day 21 p = 0
03). Lactate
concentration was significantly lower in the PR-RBCS
from day 7 (p < 0
0001), and these differences persisted
until the end of the follow-up period (day 14 p < 0
0001;
day 21 p = 0
006).
Immediately after treatment, the Hc50 (Fig. 1B) of
RBCS was significantly worse in both groups compared to
WB (in the study group, p = 0
046; in the control group,
p = 0
006). In GI-RBCS, changes in the Hc50 during the
storage did not reach statistical significance compared
with day 0 after treatment. In the PR-RBCS, Hc50 signifi-
cantly worsened on day 14 compared to days 0 and 7
(p = 0
005), and between 14th and 21st days of storage
(p = 0
005). The Hc50 in the PR-RBCS were significantly
worse than Hc50 in the GI-RBCS from day 7 (p = 0
002).
However, the WD did not differ significantly between
groups in any of the points. Immediately after treatment,
the WD did not change in either WB or RBCS, although
on the day 7 a significant increase in the WD occurred

© 2020 International Society of Blood Transfusion

Vox Sanguinis (2021) 116, 547–556
552 I. Kumukova et al.
(p < 0
0001) in both groups of RBCS. Subsequently, the
WD in the PR-RBCS increased significantly between the
7th and 14th day of storage, and in the GI-RBCS between
7th and 21st day.
Immediately after treatment, the GSH concentration
decreased and the ATP concentration increased in both
groups RBCS compared to WB. In the PR-RBCS, these
changes reached statistical significance (for ATP p = 0
05,
for GSH p = 0
03), but not in the GI-RBCS. ATP and GSH
concentrations did not differ between groups at any
point. Interestingly, a statistically significant decrease of
the ATP and GSH compared with the values on day 0 in
the PR-RBCS was observed from day 14 of storage (for
ATP p = 0
02; for GSH p = 0
014) and in the GI-RBCS
from day 21 (for ATP p = 0
007, for GSH p = 0
001).
PS expression (Fig. 1C) did not significantly change
after both treatments compared to WB. In the GI-RBCS, a
significant increase in PS expression was observed on the
14 day, compared to day 0 (p = 0
016), but in the PR-
RBCS only on 21 days (p = 0
008). Statistically signifi-
cant intergroup differences were found only on the 21st
day of storage with higher level in the PR-RBCS
(p = 0
003).
All samples of both groups were negative for the bac-
terial culture on day 0 and day 28.
The results of the lymphocytes viability and
proliferation
Lymphocyte viability was significantly reduced during
storage in all three groups (Fig. 2). PR-WB and GI-WB
showed significantly lower viability of lymphocytes com-
pared to UT-WB (p < 0
05). Although throughout the fol-
low-up period the lymphocytes viability after the
treatment by PRT was lower than after gamma irradia-
tion, no significant differences between the treatment
methods were found (p > 0
05).
The spontaneous proliferative activity of untreated
lymphocytes (Fig. 3) was 1
2%, which could be sporadic
proliferation. After stimulation by PHA, the number of
proliferating cells increased to 46
0%. In gamma-irradi-
ated samples, spontaneous proliferation of lymphocytes
did not significantly differ from untreated samples
(p > 0
05) and amounted to 1
0%. PHA stimulation
induced proliferation in only 5
3% of the cells in this
group. In the pathogen-reduced samples, spontaneous and
stimulated proliferation of lymphocytes did not exceed
0
1% of the cells.
Discussion
The use of additional methods for the processing of blood
products such as irradiation and reduction of pathogens
enhances damage to blood components [26]. All changes
in RBCS found in our study are typical of the red cell
storage lesion.
The reason for the initially low concentration of glucose
in the PR-RBCS was the use of manual separation, which
uses less preservative solution, although the differences in
the volume of RBCS units did not reach statistical signifi-
cance (p = 0
1) due to the small sample size. Our study, as
well as others [19,22], showed a decrease in glucose con-
sumption in PR-RBCS, in contrast to control. The lactate
production in the control was higher. The high production
of lactate should theoretically lower the pH in this group
(control) lower than in the experimental group. But that
did not happen. It is not just lactate that affects pH. In the
experimental group, the pH was lower at the middle of
storage (day 7 and day 14), but the differences between the
groups disappeared towards the end of storage.
All anticoagulants and preservative solutions are acidic
(CPD pH = 5
6; SAGM pH = 5
7). Therefore, after dona-
tion, the pH of venous blood decreases which leads to
increase in ATP production [27]. This explains the higher
ATP concentration in RBCS of both groups on day 0
compared to WB. During storage, ATP production
decreases, and its concentration in the RBC progressively
drops. Although the intergroup differences in ATP con-
centration did not reach statistical significance, the study
group had a more pronounced reduction in ATP than the
control group. The probable reason is decreased glycoly-
sis. Faster depletion of ATP in the study group is impor-
tant: as the concentration of ATP affects the recovery and
survival of RBC after transfusion [19].
Storage lesion leads to the spherocytic and echinocytic
morphology of RBC. These changes showing resistance of
RBC packaging in microhematocrit tube during centrifuga-
tion and does not reflect actual haematocrit of the donor
on the day of blood donation [28]. Thus, the increase of the
haematocrit found in our study may indicate more pro-
nounced changes in the morphology of PR-RBCS. We do
not believe that the introduction of a slightly lower volume
of SAGM significantly affected the haematocrit, if only in
part, because significant differences in Ht between groups
were found from 7 days of storage.
Hc50 demonstrates RBC resistance to osmotic stress.
The WD characterizes homogeneity of the RBC population
by its ability to withstand hypo-osmotic stress. The
greater WD the less homogeneity of the RBC in osmotic
resistance. Both treatments significantly influenced the
Hc50, but in the PR-RBCS the deterioration was more
pronounced, which indicates less stability and a greater
tendency to haemolysis pathogen-reduced RBC in the
early days of storage. The results of the changes in the
WD demonstrate a deeper damaging effect of PRT on the
mechanisms of cell protection from osmotic stress.
© 2020 International Society of Blood Transfusion
Vox Sanguinis (2021) 116, 547–556

Figure 2 The viability of lymphocytes at 0, 1, 2, 3 days of storage, depending on the type of treatment. Values indicate the mean percentage of viable
lymphocytes. The ends of the whiskers show standard deviation. After both treatments (gamma irradiation and pathogen reduction) viability of lympho-
cytes statistically significantly lower than in untreated samples (p < 0
05). No significant differences between the treatment methods in lymphocyte via-
bility (p > 0
05)). p-values were obtained by the Kruskal–Wallis H-test.
GSH performs an antioxidant and detoxification func-
tion in RBC [29]. Both treatments (gamma irradiation and
PRT) increase peroxidation and decrease glycolysis, which
leads to a drop in the concentration of GSH. In the PR-
RBCS, the peroxidation processes are more pronounced
than in untreated RBCS [30]. But in our study, as in Qadri
et al. 2017 [31], differences in GSH concentration
between groups did not reach statistical significance,
although the dynamics of decline was more pronounced
in PR-RBCS.
The mean binding of annexin V did not exceed 1% in
both groups. Compared with the day 0 in the GI-RBCS,
increase of the PS expression on RBC was observed ear-
lier than in the PR-RBCS (day 14). Also, on the 14th day
of storage, significant differences were found in haemoly-
sis, which was higher in the PR-RBCS. Possibly, the dif-
ferences in the PS expression was due to pathogen-
reduced RBC (probably more fragile populations) under-
going more destruction while the remaining more stable
cells had a lower expression of PS than the gamma-irra-
diated RBC, in which more cells survived to day 14.
According to our data, the incidence of post-transfu-
sion hyperkalemia and its adverse effects after transfusion
© 2020 International Society of Blood Transfusion
Vox Sanguinis (2021) 116, 547–556
Pathogen-reduced red blood cells 553
of PR-RBCS will be no worse than that after GI-RBCS of
corresponding shelf life.
The greater haemolysis in the PR-RBCS compared with
untreated, and GI-RBCS has been described in many stud-
ies [30,31]. It was the indicator that limited shelf life in
our study. We used a haemolysis level below 0
8% of
RBC to define the shelf life of PR-RBCS as 14 days.
Our study showed that the PR-RBCS can be used for
transfusions for 14 days of storage, during which it
demonstrates quality.
Suppression of lymphocytes proliferation and viability
demonstrate that this PRT can serve as an alternative to
irradiation of blood products.
Perhaps modification of the RF + UV system will lead
to an extension of the storage time and the quality of the
PR-RBCS [19]. If the manufacturer could produce storage
bags to fit automatic fractionation devices, this would be
helpful to transfusion personnel.
Conflicts of interest
The authors declare that they have no conflicts of
interest.
554 I. Kumukova et al.

Not-stimulated PHA-stimulated
(b)

(a)
Untreated WB

(d)
(c)
Gamma-irradiated WB

(e)
(f)
Pathogen-reduced WB

Figure 3 Proliferation of lymphocytes isolated from untreated WB (A, B), gamma-irradiated WB (C, D) and pathogen-reduced WB (E, F), non-stimulation
(A, C, E) and stimulated by PHA (B, D, F), after incubation for three days. In gate, P2 indicates the percentage of proliferating cells. The spontaneous pro-
liferative activity of lymphocytes in untreated (A) and gamma-irradiated (C) samples did not significantly differ (p > 0
05). PHA-stimulated lymphocytes
proliferation was significant higher in untreated samples (B) than in gamma-irradiated (D) (p < 0
05). Spontaneous (E) and PHA-stimulated (F) lympho-
cytes proliferation in RF + UV-treated samples were significant lower than in untreated and in gamma-irradiated samples (p < 0
05). p-values were
obtained by the Kruskal–Wallis H-test.

© 2020 International Society of Blood Transfusion

Vox Sanguinis (2021) 116, 547–556

References
1 Stramer SL, Dodd RY: Transfusion-
transmitted emerging infectious dis-
eases: 30 years of challenges and pro-
gress. Transfusion 2013; 53:2375–83
2 Busch MP: Transfusion-transmitted
viral infections: building bridges to
transfusion medicine to reduce risks
and understand epidemiology and
pathogenesis. Transfusion 2006;
46:1624–1640
3 Pruett CR, Vermeulen M, Zacharias P,
et al.: The use of rapid diagnostic tests
for transfusion infectious screening in
Africa: a literature review. Transfus
Med Rev 2015; 29:35–44
4 Glynn SA, Busch MP, Dodd RY, et al.:
Emerging infectious agents and the
nation’s blood supply: responding to
potential threats in the 21st century.
Transfusion 2013; 53:438–54
5 Stramer SL, Hollinger FB, Katz LM,
et al.: Emerging infectious disease
agents and their potential threat to
transfusion safety. Transfusion 2009;
49(Suppl 2):1S–29S. https://doi.org/10.
1111/j.1537-2995.2009.02279.x
6 Lanteri MC, Kleinman SH, Glynn SA,
et al.: Zika virus: a new threat to the
safety of the blood supply with world-
wide impact and implications. Trans-
fusion 2016; 56:1907–14. https://doi.
org/10.1111/trf.13677
7 Harrington T, Kuehnert MJ, Kamel H,
et al.: West Nile virus infection trans-
mitted by blood transfusion. Transfu-
sion 2003; 43:1018–22. https://doi.
org/10.1046/j.1537-2995.2003.00481.x
8 Gallagher L, Ganz P, Yang H: Advanc-
ing risk assessment for emerging
infectious diseases for blood and blood
products: proceedings of a public
workshop. Transfusion 2013; 53:455–
463.
9 Akahoshi M, Takanashi M, Masuda M,
et al.: A case of transfusion-associated
graft-versus host disease not prevented
by white cell-reduction filters. Trans-
fusion 1992; 32:169–72
10 Fisher M, Chapman JR, Ting A, et al.:
Alloimmunisation to HLA antigens
following transfusion with leucocyte-
poor and purified platelet suspensions.
Vox Sang 1985; 49:331
11 Paglino JC, Pomper GJ, Fisch GS,
et al.: Reduction of febrile but not
© 2020 International Society of Blood Transfusion
Vox Sanguinis (2021) 116, 547–556
allergic reactions to RBCs and platelets
after conversion to universal prestor-
age leukoreduction. Transfusion 2004;
44:16–24
12 Ohto H: Gamma radiation does not
prevent transfusion induced HLA
alloimmunization. Transfusion 1997;
37:878–9
13 Seltsam A: Pathogen inactivation of
cellular blood products - an additional
safety layer in transfusion medicine.
Front Med (Lausanne) 2017; 4:219.
https://doi.org/10.3389/fmed.2017.00219
14 Yonemura S, Doane S, Keil S, et al.:
Improving the safety of whole blood-
derived transfusion products with a
riboflavin-based pathogen reduction
technology. Blood Transfus 2017;
15:357–364. https://doi.org/10.2450/
2017.0320-16
15 Goodrich RP, Doane S, Reddy HL:
Design and development of a method
for the reduction of infectious patho-
gen load and inactivation of white
blood cells in whole blood products.
Biologicals 2010; 38:20–30
16 Cap AP, Pidcoke HF, Keil SD, et al.:
Treatment of blood with a pathogen
reduction technology using ultraviolet
light and riboflavin inactivates Ebola
virus in vitro. Transfusion 2016; 56
(Suppl 1):S6–15. https://doi.org/10.
1111/trf.13393
17 Doane SK, Yonemura S, Hovenga N,
et al.: Evaluation of the acute toxicity
of red blood cells derived from ribo-
flavin and UV light-treated whole
blood in a canine red blood cell
exchange model. Transfusion 2016;
56:193A
18 Goodrich RP, Murthy KK, Doane SK,
et al.: Evaluation of potential immune
response and in vivo survival of ribo-
flavin-ultraviolet light-treated red
blood cells in baboons. Transfusion
2009; 49:64–74. https://doi.org/10.
1111/j.1537-2995.2008.01940.x
19 Cancelas JA, Slichter SJ, Rugg N,
et al.: Red blood cells derived from
whole blood treated with riboflavin
and ultraviolet light maintain ade-
quate survival in vivo after 21 days
of storage. Transfusion 2017; 57:
1218–1225. https://doi.org/10.1111/trf.
14084
Pathogen-reduced red blood cells 555
20 Cancelas JA, Rugg N, Fletcher D,
et al.: In vivo viability of stored red
blood cells derived from riboflavin
plus ultraviolet light-treated whole
blood. Transfusion 2011; 51:1460–8
21 Hervig T, Braathen H, Jaboori AA,
et al.: Platelet recovery and survival
after whole blood treated with mirasol
pathogen reduction. Transfusion 2016;
56:3A–262A
22 Schubert P, Culibrk B, Karwal S, et al.:
Whole blood treated with riboflavin
and ultraviolet light: quality assess-
ment of all blood components pro-
duced by the buffy coat method.
Transfusion 2015; 55:815–23. https://
doi.org/10.1111/trf.12895
23 Shcherbachenko I, Lisovskaya I,
Tikhonov V: Oxidation-induced cal-
cium-dependent dehydration of nor-
mal human red blood cells. Free
Radical Res 2007; 41:536–45
24 Hans A: Adenosine-5 -triphosphate:
determination with phosphoglycerate
kinase; in Bergmeyer H-U (ed): Meth-
ods of Enzymatic Analysis (Second
Printing, Revised), Academic Press,
1965:539–558, ISBN 9780123956309.
Available at: https://doi.org/10.1016/
B978-0-12-395630-9.50112-2.
25 Pradedova E, Nimaeva O, Putilina T,
et al.: Determination of glutathione
and its redox status in isolated vac-
uoles of red beetroot cells. J Stress
Physiol Biochem 2016; 12:87–107
26 Hess J: Red cell storage. J Proteom
2010; 73:368–373
27 Simon TL, McCullough J, Snyder EL,
et al.: Strauss RG Rossi’s Principles of
Transfusion Medicine. Oxford, UK:
Wiley-Blackwell, 2016:760
28 Mustafa I, Marwani A, Mamdouh N,
et al.: Time dependent assessment of
morphological changes: leukodepleted
packed red blood cells stored in
SAGM. Biomed Res Int 2016;
2016:4529434. https://doi.org/10.
1155/2016/4529434
29 Dumaswala U, Wilson M, Wu Y, et al.:
Glutathione loading prevents free radi-
cal injury in red blood cells after stor-
age. Free Radical Res 2000; 33:517–
529
30 Chen D, Schubert P, Devine D: Identi-
fication of potential protein quality
556 I. Kumukova et al.

markers in pathogen inactivated and 31 Qadri S, Chen D, Schubert P, et al.: lesion and promotes eryptosis. Trans-
gamma - irradiated red cell concen- Pathogen inactivation by riboflavin fusion 2017; 57:661–673
trates. Clin Appl Thromb Hemost and ultraviolet light illumination
2017; 11:7–8 accelerates the red blood cell storage

© 2020 International Society of Blood Transfusion

Vox Sanguinis (2021) 116, 547–556
 Vox Sanguinis (2021) 116, 557–563
© 2021 International Society of Blood Transfusion
ORIGINAL PAPER DOI: 10.1111/vox.13044

Anti-A and SARS-CoV-2: an intriguing association

Valeria de Freitas Dutra,1 Carolina Bonet-Bub,1 Ana Paula H. Yokoyama,1 Ruth Achkar,3 Rafael Rahal Guaragna Machado,2
Murilo Assuncß~ao, Gabriela Candelaria, Camila Pereira Soares,2 Roberta Maria Fachini,3 Rita Font~
1 3
ao-Wendel,3
Nelson Hamerschlak, Luiz Fernando Lima Reis, Danielle Bastos Araujo, Victor Nudelman, Joao R. R. Pinho,1 Luiz V. Rizzo,1
1 3 1,2 1

Araci M. Sakashita,1 Patrıcia Scuracchio,3 Edison Luiz Durigon,2,4 Silvano Wendel3 & Jose M. Kutner1
1
Hospital Israelita Albert Einstein, S~ao Paulo, Brazil
2
Department of Microbiology, Institute of Biomedical Sciences, University of Sao Paulo, S~
ao Paulo, Brazil
3
Hospital Sırio Liban^e s, S~ao Paulo, Brazil
4
Scientific Platform Pasteur USP, S~ao Paulo, Brazil

Background Blood groups and anti-A isohemagglutinin may be involved in sus-

Received: 29 July 2020,
revised 13 November 2020,
accepted 15 November 2020,
published online 2 March 2021
ceptibility to SARS-CoV-2 infection.
Materials and Methods We retrospectively studied 268 COVID-19 convalescent
plasma donors and 162 COVID-19 inpatients (total 430 subjects, confirmed by
RT-PCR) and 2,212 healthy volunteer first-time blood donors as a control group.
These were further divided into two groups: those with anti-A (blood types O
and B) and those without it (types A and AB). Titres of nucleoproteins, and neu-
tralizing SARS-CoV-2 antibody were measured in the convalescent plasma
donors and inpatients. Multivariate logistic regression and non-parametric tests
were applied.
Results Persons having types O or B showed less infection prevalence than those
of types A or AB (OR = 0
62, 95% CI 0
50–0
78; P < 0
001), but there was no
difference when COVID-19 inpatients were analysed. Immunoglobulins M, G and
A were lower in COVID-19 subjects of types O or B group than those of A or AB
(0
16 vs. 0
19; P = 0
03, 2
11 vs. 2
55; P = 0
02, 0
23 vs. 0
32; P = 0
03, respec-
tively).
Conclusion In this retrospective cohort, COVID-19 individuals were less likely to
belong to blood types O and B, and also had lower SARS-CoV-2 antibody titres
than A and AB individuals. COVID-19 severity did not associate with the blood
groups.
Key words: ABO groups, anti-A, SARS-CoV-2.

and lymphopenia [2,3], have already been associated with

Background
Since December 2019, when the first outbreak of novel
coronavirus disease (COVID-19) occurred in Wuhan,
China, over 33 million people have been diagnosed, and
over one million people have died worldwide [1]. It is
unclear which individual characteristics determine sus-
ceptibility and intensity of symptoms. However, age, sex,
ethnicity, hypertension, body mass index and haemato-
logical biomarkers, such as D-dimer, thrombocytopenia
Correspondence: Jose M Kutner, Hospital Israelita Albert Einstein – Av.
Albert Einstein, 627, 3o andar, Bloco E, S~ao Paulo 05652-900, SP, Brazil
E-mail: kutner@einstein.br
a worse outcome.
Recently blood types and anti-A isohemagglutinin have
been associated with susceptibility to the severe acute res-
piratory syndrome coronavirus 2 (SARS-CoV-2) infection
[4,5]. Considering that the first studies reported type A as
a risk factor and O as protection, some authors have sug-
gested that anti-A and not the blood type itself could be
responsible for the findings [6].
The human ABO histo-blood group system has a single
gene located on the terminal portion of the long arm of
chromosome 9 (9q34.2), with three main alleles; one
recessive O and two co-dominant A and B [7]. Differences

557
558 V. de Dutra et al.
in blood group antigen expressions and presence or
absence of anti-A or -B provide strong defensive lines
against infection [8]. In persons with O blood type, char-
acterized by the absence of A or B antigens, stimulation
by microbiota having glycan motifs similar to A or B
antigens, leads to a natural production of anti-A and
anti-B [9]. Type B individuals also produce anti-A but in
lower titres [10]. Interestingly, in 2003, during the SARS-
CoV outbreak in China, the O blood group was considered
protective against infection—OR = 0
18 (0
04–0
81) [11].
Later, a cell-binding assay showed that either a mono-
clonal or human natural anti-A could inhibit the SARS-
CoV S protein/ACE2 interaction [12]. At a cellular level,
this supports the idea that the type O protection against
SARS-CoV involves the antibodies rather than the anti-
gens.
Based on these findings, our study aimed to analyse
the association of SARS-CoV-2 infection with the pres-
ence of anti-A (In types O and B) or its absence (in types
A and AB), related to the production of antibodies to
SARS-CoV-2 nucleoprotein (NP; IgA, IgM and IgG) and
neutralizing antibodies (nAb).
Methods and materials
Ethics statement
The study was approved by both hospitals’ Institutional
Review Boards (IRB) and the Brazilian Commission on
Ethics and Research (CONEP) under requests CAAE
32558220.0.0000.0071 and CAAE 30259220.4.2001.5461.
All patients and COVID-19 convalescent plasma donors
provided written informed consent.
Subjects
We analysed a retrospective cohort of 430 people with
COVID-19 [268 convalescent plasma donors (CCPD) and
162 inpatients (CIP)] from both hospitals. All patients and
plasma donors had a previous diagnosis confirmed by
RT-PCR.
The CCPD group comprised convalescent patients who
had had mild symptoms (no hospitalization during their
COVID-19 evolution). Eligibility criteria required a posi-
tive diagnostic test by naso-oropharyngeal swab (NOS)
RT-PCR and resolution of symptoms for at least 14 days.
The candidates were then tested for SARS-COV-2 by RT-
PCR either on peripheral blood or NOS swab. If the RT-
PCR was negative, the plasma was collected, and antibod-
ies to SARS-CoV-2 nucleoprotein (anti-NP - IgA, IgM and
IgG) and neutralizing antibodies (nAb) were measured.
The CIP group was composed of patients with a posi-
tive SARS-CoV-2 RT-PCR, who had moderate to severe
symptoms and needed hospitalization. Blood type infor-
mation was available in the electronic chart. Samples for
anti-NP and nAbs were drawn by the time of admission.
In order to avoid blood group bias from repeat donors,
given that blood type O (‘universal donor’) is usually
over-represented, our control group (CG) comprised 2212
first-time voluntary healthy blood donors from Hospital
Israelita Albert Einstein blood bank database, who
donated whole blood from August to October 2019,
before the COVID-19 outbreak in Brazil.
Subjects of blood types A and AB were grouped as
‘without anti-A’ (A/AB group), whereas those with O and
B were named ‘with anti-A’ (O/B group).
Samples tests
Blood typing was done by the automated analyser, gel
technique, Erytra Eflexis (Grifols, Barcelona, Spain) and
IH-platform (Biorad, Creisser, Switzerland).
RT-PCR: A real-time reverse transcriptase polymerase-
chain-reaction technique confirmed SARS-CoV-2 diagno-
sis from naso-oropharyngeal swab specimens. Molecular
tests were based on Corman et al. [13], with five copies/
reaction of sensitivity.
Neutralizing antibodies (nAbs) and anti-nucleocapsid
(anti-NP) antibodies (IgM, IgG and IgA): We used the
cytopathic effect-based virus neutralization test (CPE-
based VNT), which was carried out with SARS-CoV-2
(GenBank: MT350282.1), as previously described [14,15].
Anti-NP was determined according to Wendel et al. [15].
Both methods have been described in more detail else-
where [15].
Data analysis
Age was compared between control, CCPD and CIP
groups using the ANOVA test. For sex and blood type
analysis we used the chi-square test. CCPD and CIP
groups were classified as COVID-19 individuals’ group
and compared to the CG.
For further analysis regarding the presence or absence
of ‘circulating anti-A’ and its association with COVID-19,
we merged our study population into two groups: one
‘with circulating anti-A’; including types O and B (O/B
group)” and another ‘without circulating anti-A’; includ-
ing types A and AB (A/AB group).”
For adjusted models, we applied multiple logistic
regression. The Mann–Whitney test and Spearman’s test
were used to compare anti-A presence/absence and anti-
NP (IgM, IgG and IgA) among groups.
Data were organized in Microsoft Excel 2010 and were
analysed using Statistical Package for Social Sciences
(SPSS) (Chicago, IL) or GraphPad Prism version 8.0 for
© 2021 International Society of Blood Transfusion
Vox Sanguinis (2021) 116, 557–563
 Association between SARS-CoV-2 and blood groups 559

Windows, GraphPad Software (La Jolla, CA, USA). A P- Table 2 Multivariate logistic regression for COVID-19 individuals
value <0
05 was considered significant. (n = 430) and blood donors (CG) (n = 2212)

95% CI
Results
Variable OR Lower Upper P
We had age, sex and blood type of all 430 COVID-19
individuals (268 CCPD and 162 CIP) and 2212 healthy Age (years) 1
06 1
05 1
06 <0
001
Gender (male) 1
27 1
02 1
59 0
035
volunteer blood donors (control group: CG) However, as
Anti-A (O/B) 0
62 0
50 0
78 <0
001
they were from a retrospective cohort, the anti-NP and
nAbs were available in only 295 of the COVID-19 per- CG, control group; CI, confidence interval, OR, odds ratio.
sons. Table 1 shows the distribution, mean age, sex and Age and male gender were positively related to COVID-19 (OR = 1
06,
blood type among groups. Although blood type O most 95% CI: 1
05–1
06; P < 0
001 and OR = 1
27, 95%CI: 1
02–1
59;
frequent among blood donors, type A was more common P = 0
035 respectively). The presence of circulating anti-A (O/B group)
in the COVID-19 group. There was no statistical differ- showed a protective factor to COVID-19 (OR = 0
62, 95% CI; 0
50–0
78;
ence in blood type distribution between CCPD and CIP. P < 0
001).
Age and male sex were positively related to COVID-19
(OR = 1
06, 95% CI: 1
05–1
06; P < 0
001 and OR = 1
27,
95%CI: 1
02–1
59; P = 0
035 respectively). However, the Additionally, we analysed whether anti-NP was associ-
presence of circulating anti-A (O/B group) showed a pro- ated with the presence of anti-A. We included 295 sub-
tective effect against COVID-19 (OR = 0
62, 95% CI: jects from the COVID-19 group (148 from O/B group and
0
50–0
78; P < 0
001), as shown in Table 2. 147 in the A/AB group). Figure 2 shows the distribution
In order to evaluate the association of circulating anti- of anti-NP between the groups. The O/B group showed
A with COVID-19 severity, we compared CCPD and CIP to lower median IgM, IgG and IgA levels than for A/AB
the CG. Belonging to the O/B group was protective only (0
16 vs. 0
19; P = 0
03, 2
11 vs. 2
55; P = 0
02, 0
23 vs.
for CCPD, as shown in Fig. 1 (OR = 0
60, 95% CI: 0
45– 0
32; P = 0
03, respectively).
0
75 P < 0
001). Figure 3 shows the nAb distribution analysis. As the
As type O persons usually have higher anti-A titre, we samples’ value varies from <1:20 to >1:5120, we divided
also performed a sub-analysis for O vs. B blood, as shown the nAb titres into two groups: <320 and ≥320 (320 as
in Table 3. the middle cut-off point for the reaction range). Patients
Age was positively related to COVID-19 (OR = 1
05, 95% of types O or B showed a lower trend of neutralizing anti-
CI: 1
04–1
06; P < 0
001). Sex did not show any association body value and lower frequencies when the titres were
to COVID-19 between these groups. However, a higher titre higher than 320 (chi-square test = 6
99, P = 0
008). The
of circulating anti-A (O group) was protective against results are shown in Table 4.
COVID-19 (OR = 0
66, 95% CI: 0
46–0
95; P = 0
026). There was an evident linear correlation between anti-
NP (IgA, IgG and IgM) and nAbs according to the Spear-
man’s correlation test (all P < 0
0001), as shown in Fig. 4.
Table 1 Demographic data from subjects included in the study
We identified a better correlation between IgG and nAbs
Blood donors (CG) CCPD CIP for both O/B and A/AB groups (r = 0
687 in O/B and
Variable (N = 2212) (N = 268) (N = 162) P r = 0
640 in A/AB). IgM and IgA did not show such good
correlations (O/B group IgA r = 0
593; IgM r = 0
430 and
Agea (years; 37
5 – 12
0 36
8 – 8
1 69
3 – 15
7 <0
001
A/AB group IgA r = 0
555; IgM r = 0
457).
Mean – SD)
Gender, n (%)b
Female 1004 (45
4) 103 (38
4) 59 (36
4) 0
012 Discussion
Male 1208 (54
6) 165 (61
6) 103 (63
6)
Blood group, n (%)b The ABO system can be associated with the inflammatory
A 785 (35
5) 128 (47
8) 70 (43
2) <0
001 response and has a varied geographical frequency, with
AB 84 (3
8) 11 (4
1) 7 (4
3) growing evidence that it can affect the predisposition to
O 1117 (50
5) 103 (38
4) 59 (36
4) certain diseases, such as thrombosis or H. pylori infection
B 226 (10
2) 26 (9
7) 26 (16
1) [7]. A meta-analysis recently determined the odds of
SARS-CoV-2 positive individuals of having a specific
CCPD, COVID-19 convalescent plasma donors; CG, control group; CIP,
blood type compared with controls. The association of
COVID-19 inpatients; N, number of participants.
a
ANOVA.
SARS-CoV-2 with blood type A was significant with a
b
Chi-Square. pooled OR of 1
23 (95%CI: 1
09–1
40), although the

© 2021 International Society of Blood Transfusion

Vox Sanguinis (2021) 116, 557–563
560 V. de Dutra et al.
Fig. 1 Odds ratio (OR) for CIP and CCPD compared to blood donors (CG),
considering O/B vs. A/AB groups. Legend: O/B group was identified as a
protective factor only for CCPD (OR = 0
60, 95% CI: 0
45–0
75), whereas
for CIP this effect was not identified (OR = 0
85, 95% CI: 0
45–1
25).
Multivariate logistic regression; CCPD, COVID-19 convalescent plasma
donors; CIP, COVID-19 inpatients; *P < 0
001.
Table 3 Multivariate logistic regression analysis for O vs. B blood groups,
comparing COVID-19 patients and blood donors (CG)
95% CI
Variable OR Lower Upper P
Age (years) 1
05 1
04 1
06 <0
001
Gender (male) 1
34 0
99 1
82 0
062
Anti-A (O/B) 0
66 0
46 0
95 0
026
CI, confidence interval; OR, odds ratio.
COVID-19 patients O blood group (n = 103), COVID-19 patients B blood
group (n = 26), Blood donors O group (n = 1117), Blood donors B group
(n = 226). Age was positively related to COVID-19 (OR = 1
05, 95% CI:
1
04–1
06; P < 0
001) and the presence of a higher titre of circulating
anti-A (O group) showed a protective factor to COVID-19 (OR = 0
66,
95% CI: 0
46–0
95; P = 0
026).
Bold indicates statistical significant values (P < 0
05).
random-effect meta-analysis revealed a considerable
heterogeneity among studies [16]. Despite our first-time
volunteer blood donors being mainly of type O, this was
not the most prevalent type among COVID-19 patients.
This finding is similar to previous studies that reported
ABO blood group association with SARS-CoV in 2003
[11] and with SARS-CoV-2, in 2020 [4,17,18]. One possi-
ble mechanism is that blood types A and B have sugars
crucial to the O-glycosidic target, functioning as a possi-
ble site for binding the virus to the host [19].
As previously proposed, the presence of anti-A, and not
the blood type, could be associated with the susceptibility to
SARS-CoV-2 infection [6], although future studies must
prove this assumption. Our results are similar, with a statisti-
cal difference between groups of patients with anti-A (B/O)
Fig. 2 COVID-19 individuals (n = 295) IgM, IgG, and IgA distribution
between A/AB and O/B groups. Legend: O/B group showed an IgM, IgG
and IgA median level lower when compared to A/AB (0
16 vs. 0
19;
*P = 0
03, 2
11 vs. 2
55; #P = 0
02, 0
23 vs. 0
32; &P = 0
03, respec-
tively).Mann–Whitney test.
and without anti-A (A/AB). Moreover, we showed a greater
protective effect of type O compared with B subjects, sug-
gesting that anti-A from the former might be more effective.
Our data suggest that higher anti-A titres could improve the
host defence. It is a plausible argument, since type B pro-
duces lower titres of anti-A [10].
Our findings are similar to previously reported studies
that group O was associated with a lower risk of infection
than non-O blood types [4]. We did not find any statisti-
cally significant difference between CCPD donors and
COVID inpatients (CIP), like other studies [18] that did not
identify any relationship between blood type and intuba-
tion or death rate. In contrast, one study in China found
fewer cardiovascular diseases and lesser severity in the O
group [17].
Our study’s primary limitation is the control group
used: voluntary first-time blood donors from a single
hospital. In Brazil, data regarding blood type belong to
local centres, and data for the Brazilian population as a
whole are unsatisfactory. Few papers show the prevalence
of type O among Caucasians, mixed race and Afro-de-
scendants (46
5%, 53
2% and 47
9%, respectively) fol-
lowed by A-type (39
4%, 29
6% and 31
9% respectively)
[20]. Our control group was similar to the ABO blood type
distribution in Brazil, supporting our findings. Moreover,
only voluntary first-time blood donors were included in
the control group to avoid possible bias for over-repre-
senting any blood type in repeat donors.
The ABO type could be involved in the S protein
SARS-CoV and SARS-CoV-2 cell-binding mechanism
© 2021 International Society of Blood Transfusion
Vox Sanguinis (2021) 116, 557–563
 Association between SARS-CoV-2 and blood groups 561

Fig. 3 Distribution of the frequencies of COVID-19 individuals (n = 295) neutralizing antibodies between A/AB and O/B groups. Legend: O/B group
showed a neutralizing antibody lower value trend and lower frequencies when the titres were higher than 320.

Table 4 Distribution of COVID-19 individuals (n = 295) neutralizing anti- respiratory failure in COVID-19, due to the rs657152 A/C
body (nAb) titres in two groups: <320 and ≥320
single nucleotide polymorphism at 9q34.2. Also, some
nAb titres
procoagulant markers are associated with genetic varia-
tion at the ABO locus, and this region could have a role
ABO Group <320 ≥320 v2 P OR (95% CI) in modifying genes [21].
Curiously, when we evaluated specific immunoglobulin
O/B 97 (65
5) 51 (34
5) 6
99 0
008 0
53 (0
33- 0
85)*
production, we found a statistically significant difference
A/AB 74 (50
3) 73 (49
7)
for IgM, IgG and IgA results, with median values lower in
O/B COVID-19 individuals had a lower chance of having values ≥320. the O/B group. Similarly, neutralizing antibody titres were
*0
54 (0
34; 0
87), when adjusted by age and gender. lower in the O/B group.
This is one of the few studies that analyse a possible
[12,19]. Genetic research also has increased in this area. correlation between humoral response for SARS-CoV-2
Genome-wide association analysis in an Italian-Spanish and ABO group. In that way, even though we did not
group showed that A-positive people are at higher risk of observe a statistical difference between the groups

Fig. 4 Linear correlation among COVID-19 individuals (n = 295) immunoglobulins and neutralizing antibody titres (Log2), into the group A/AB (a) and
O/B (b). Legend: Spearman correlation for A/AB group (left) and for O/B group (right). A better correlation between IgG and neutralizing antibodies was
found in both O/B and A/AB (r = 0
687 in O/B and r = 0
640 in A/AB). IgM and IgA did not show such good correlations (O/B group IgA r = 0
593; IgM
r = 0
430 and A/AB group IgA r = 0
555; IgM r = 0
457).

© 2021 International Society of Blood Transfusion

Vox Sanguinis (2021) 116, 557–563
562 V. de Dutra et al.

CCPD and CIP concerning severity, the profile of
increased humoral immune response in A/AB patients is
an intriguing question and may be related to other fac-
tors, such as the clinical evolution and progression of
the disease. Our study did not correlate clinical status
and level of IgM, IgG, IgA anti-NP or nAb titres; thus,
this parameter correction for ABO status may be biased.
IgA is an immune barrier and can probably neutralize
SARS-CoV-2 before the virus reaches and binds to the
epithelial cells. This has taken on increased importance
recently. IgM and IgG levels have a potential role in the
evaluation of severity and prognosis of COVID-19 [22]. In
37 patients, IgA and IgG levels were markedly higher
(P < 0
001) in patients with severe disease compared with
mild disease, while there was no difference in IgM level
[23]. Additionally, in a two-year prospective study, IgG
antibody and NAb titres were positively correlated in
SARS-CoV [24]. However, we did not find any research
correlating IgM and IgA, especially in SARS-CoV-2 infec-
tions. Future longitudinal studies can show if blood
groups interfere with protection antibodies, including the
long-term immune response.
The impact of anti-A isohemagglutinin titres in the
SARS-CoV-2 infection and its association with neutraliz-
ing antibodies is not clear [25]. In our retrospective study,
we could not measure the quantitative effect of anti-A.
Further studies are required to evaluate both anti-A iso-
hemagglutinin and neutralizing antibody titres, and their
role in SARS-CoV-2 infection.
Conclusion
Blood types O and B, which produce anti-A, showed a
protective effect against SARS-CoV-2 infection. There
was no statistical difference between COVID-19 inpatients
and COVID-19 convalescent plasma donors, suggesting
that blood types do not associate with COVID-19 severity.
Moreover, COVID-19 individuals from types O and B had
lower titres of neutralizing antibodies and lower levels of
IgM, IgG and IgA anti-nucleocapsid antibodies than did
the types A and AB.
Acknowledgements
C.P.S. is funded by Grant 2018/23680-0 (Fundacß~ao de
Amparo a Pesquisa do Estado de S~ao Paulo); D.B.A. by
Grant 88 887.131387/2016-00 (Coordenacß~ao de Aper-
feicßoamento Pessoal de Nıvel Superior - CAPES), R.R.G.M.
by Grant 2017/24769-2 (Fundacß~ao de Amparo a Pesquisa
do Estado de S~ao Paulo) and E.L.D. by Grants 2016/
20045-7 and 2020/06409-1 (Fundacß~ao de Amparo a Pes-
quisa do Estado de S~ao Paulo). This project was partially
supported by the initiative ‘Todos Pela Saude - Fundacß~ao
Itau para Educacß~ao e Cultura’.
Conflict of interests
The authors declare no conflict of interests.
Author contributions
Conceptualization – VFD, CBB, APHY, SW, JMK; Investi-
gation – VFD, CBB, APHY, NH, JRRP, RFW, RMF, RRGM,
GC, AS, RA, PS, DBA, CPS, ELD, MA, VN, SW, JMK. For-
mal analysis – VFD, CBB, APHY, SW, JMK; Resources –
LFLR, LVR; Writing – VFD, CBB, APHY, SW, JMK. Project
administration and funding acquisition – LFR, LVR.

References
1 Dong E, Du H, Gardner L. An interac-
tive web-based dashboard to track
COVID-19 in real time. Lancet Infect
Dis 2020;20:533–4.
2 Frater JL, Zini G, d’Onofrio G, et al.
COVID-19 and the clinical hematology
laboratory. Int J Lab Hematol
2020;2:11–8.
3 Caussy C, Pattou F, Wallet F, et al.
Prevalence of obesity among adult
inpatients with COVID-19 in France.
Lancet Diabetes Endocrinol
2020;8:562–4
4 Zhao J, Yang Y, Huang H-P, et al.
Relationship between the ABO Blood
Group and the COVID-19 susceptibil-
ity. medRxiv. Published online
2020;2020.03.11.20031096. https://doi.
org/10.1101/2020.03.11.20031096
5 Zaidi FZ, Zaidi ARZ, Abdullah SM,
et al. COVID-19 and the ABO blood
group connection. Transfus Apher Sci
2020;59:102838.
6 Gerard C, Maggipinto G, Minon J-M.
COVID-19 & ABO blood group:
another viewpoint. Br J Haematol
2020;39:62–3.
7 Franchini M, Bonfanti C. Evolutionary
aspects of ABO blood group in humans.
Clin Chim Acta 2015;444:66–71.
8 Yamamoto F, Cid E, Yamamoto M,
et al. An integrative evolution theory
of histo-blood group ABO and related
genes. Sci Rep 2014;4:1–12.
9 Breiman A, Ruv€en-Clouet N, Le Pendu
J. Harnessing the natural anti-glycan
immune response to limit the trans-
mission of enveloped viruses such as
SARS-CoV-2. PLoS Pathog 2020;16:
e1008556.
10 Kang SJ, Lim YA, Baik SY.
Comparison of ABO antibody titers on
the basis of the antibody detection
method used. Ann Lab Med
2014;34:300–6.
11 Cheng Y, Cheng G, Chui CH, et al.
ABO blood group and susceptibility to
severe acute respiratory syndrome.
JAMA 2005;293:1450–1.
12 Guillon P, Clement M, Sebille V, et al.
Inhibition of the interaction between

© 2021 International Society of Blood Transfusion

Vox Sanguinis (2021) 116, 557–563

the SARS-CoV Spike protein and its
cellular receptor by anti-histo-blood
group antibodies. Glycobiology
2008;18:1085–93.
13 Corman VM, Landt O, Kaiser M, et al.
Detection of 2019 -nCoV by RT-PCR.
Euro Surveill 2020;25:1–8.
14 Araujo DB, Machado RRG, Amgarten
DE, et al. SARS-CoV-2 isolation from
the first reported patients in Brazil
and establishment of a coordinated
task network. Mem Inst Oswaldo Cruz.
2020;115. https://doi.org/10.1590/
0074-02760200342
15 Wendel S, Kutner JM, Machado R,
et al. Screening for SARS-CoV-2 anti-
bodies in convalescent plasma in Bra-
zil: preliminary lessons from a
voluntary convalescent donor pro-
gram. Transfusion 2020. https://doi.
org/10.1111/trf.16065
16 Golinelli D, Boetto E, Maietti E, et al.
The association between ABO blood
group and SARS-CoV-2 infection: a
© 2021 International Society of Blood Transfusion
Vox Sanguinis (2021) 116, 557–563
Association between SARS-CoV-2 and blood groups 563
meta-analysis. PLoS One 2020;15(9):
e0239508.
17 Dai X. ABO blood group predisposes
to COVID-19 severity and cardiovas-
cular diseases. Eur J Prev Cardiol
2020;27:1436–7.
18 Zietz M, Tatonetti NP. Testing the
association between blood type and
COVID-19 infection, intubation, and
death. medRxiv 2020;2020.04.08.
20058073. https://doi.org/10.1101/
2020.04.08.20058073
19 Arend P. How SARS-CoV-2 (COVID-
19) may invade the human body via
ABO(H) blood group-determining car-
bohydrates: a. perspective. 2020;2
(May):17. https://doi.org/10.20944/PRE
PRINTS202005.0097.V1
20 Novaretti MCZ, Dorlhiac-Llacer PE,
Chamone DAF. Estudo de grupos
sang€uıneos em doadores de sangue
caucas oides e negr
oides na cidade de
S~ao Paulo. Rev Bras Hematol Hemoter
2000;22:23–32.
21 Ellinghaus D, Degenhardt F, Bujanda
L, et al. Genomewide association study
of severe Covid-19 with respiratory
failure. N Engl J Med 2020;383:1522–
34.
22 Hou H, Wang T, Zhang B, et al. Detec-
tion of IgM and IgG antibodies in
patients with coronavirus disease
2019. Clin Transl Immunol 2020;9:1–
8.
23 Yu H, Sun B, Fang Z, et al. Distinct
features of SARS-CoV-2-specific IgA
response in COVID-19 patients. Eur
Respir J 2020;56:2001526.
24 Liu W, Fontanet A, Zhang PH, et al.
Two-year prospective study of the
humoral immune response of patients
with severe acute respiratory syn-
drome. J Infect Dis 2006;193:792–5.
25 Daniele F. Anti-A Isohemagglutinin
titers and SARS-CoV2 neutralization:
implications for children and conva-
lescent plasma selection. Br J Haema-
tol 2020;190:e148–50.
 Vox Sanguinis (2021) 116, 564–573
© 2020 International Society of Blood Transfusion
ORIGINAL PAPER DOI: 10.1111/vox.13045

An economic reappraisal of hepatitis B virus testing strategy

for blood donors in Taiwan
Tapiwa Blessing Matanhire1 & Shi-Woei Lin1,2
1
Department of Industrial Management, National Taiwan University of Science and Technology, Taipei, Taiwan
2
Artificial Intelligence for Operations Management Research Center, National Taiwan University of Science and Technology, Taipei, Taiwan

Received: 30 June 2020,
revised 11 November 2020,
accepted 15 November 2020,
published online 5 December 2020
Abstract
Background and objectives Taiwan is among the few hepatitis B virus (HBV)
high-endemic countries that implement universal mini-pool nucleic acid testing
(MP-NAT) and hepatitis B surface antigen (HBsAg) testing together with confir-
matory individual donor nucleic acid testing (ID-NAT) for its blood supply since
2013. The aim of this study was to reappraise the value of HBsAg test in Tai-
wan’s HBV testing strategy.
Materials and methods A Markov model was constructed, and cost-effectiveness
analysis was conducted in order to reappraise the existing HBV screening strat-
egy in Taiwan.
Results The incremental cost-effectiveness ratio (ICER) for the current testing
strategy in Taiwan was estimated to be $US 443 154 per quality-adjusted life
year (QALY) gained. This is almost six times the willingness-to-pay (WTP) thresh-
old that reflects local preferences.
Conclusion Universal HBsAg and MP-8-NAT together with confirmatory ID-NAT
testing prevents a significant amount of HBV infections from entering the Taiwan
blood supply. However, this comes at a disproportionate increase in cost.
Key words: hepatitis B virus, blood safety, HBsAg test, NAT testing, cost-effec-
tiveness analysis, Markov model.

as HBV from entering into the blood supply, blood opera-

Introduction
HBV is a transfusion-transmitted infection (TTI). It is
prevalent worldwide but has an incidence which varies
across different countries and territories. In highly ende-
mic areas, 70–95% of the population exhibit past or pre-
sent serological evidence of HBV infection [1,2].
Although Taiwan can be categorized as a highly endemic
area, very good progress has been made in fighting
against viral hepatitis since launch of the universal HBV
mass vaccination programme in 1984. At the end of year
2011, it was estimated that at least 700 000 HBV carriers
had been prevented [3]. In an effort to prevent TTIs such
Correspondence: Shi-Woei Lin, Department of Industrial Management,
National Taiwan University of Science and Technology, No. 43, Sec. 4,
Keelung Rd., Taipei 10607, Taiwan
E-mail: shiwoei@mail.ntust.edu.tw
tors around the world implement various safety measures
such as pre-donation selection and laboratory screening
tests [4,5]. Regardless of these safety measures, there is
always some residual risk of infection to transfusion
patients. For example, in 2017, a case of posttransfusion
HBV was reported in a low-endemic country [6].
Blood transfusion operators endeavour to eradicate
residual risk of infection by implementing diagnostic tests
that have a higher degree of sensitivity and specificity
[7]. However, these are expensive which explains why
most developed countries can afford universal HBV-NAT,
whilst the majority of developing countries can only
afford using serologic tests such as HBsAg and anti-HBc
to detect HBV in blood donors [8].
Over the past two decades, a number of countries have
effectively introduced universal HBV-NAT in their testing

564

algorithms [9]. The current practice for the detection of
HBV in Taiwanese blood donors involves the use of uni-
versal HBV-NAT and HBsAg testing. In Taiwan, anti-HBc
testing is not universal because deferring anti-HBc-posi-
tive units critically affects blood supply at a high cost in
medium to high-endemic areas [10]. When the first gen-
eration of HBV-NAT technology was licensed, questions
were raised whether it should replace HBsAg for screen-
ing blood donors [11]. Yet, most developed countries that
added universal HBV-NAT to their testing strategies
retained the HBsAg test. Lack of relevant scientific evi-
dence made it difficult for the regulatory bodies and
experts to support its discontinuation at the time [12,13].
More recently, the value of retaining HBsAg screening
where universal HBV-NAT and anti-HBc tests are used
has become a subject of active discussion by several
researchers, with current evidence indicating diminutive
value in its continued use. Although discontinuing such a
seemingly redundant screening test might result in cost
reduction, the existence of credible ethical and scien-
tific objections makes dropping it highly challenging
[10,14–16].
In spite of the considerations to discontinue HBsAg
screening test as discussed above, the test remains valu-
able in high-endemic areas where the use of anti-HBc is
not universal. Taiwan is among the few high-endemic
countries that implement universal HBV-NAT and HBsAg
testing for its blood supply since 2013 [17]. In a previous
study, a majority of Taiwanese blood donors were found
to be under 30 years old [18] and according to another
study, the anti-HBc-positive rate in this population
appears to have decreased over the past 30 years due to
the extensive infant vaccination programme [19]. There-
fore, giving rise to the question whether Taiwan could
adopt the blood donor policy for low-endemic countries
(excluding the blood donors who are anti-HBc-positive in
order to eliminate the potential sources of occult HBV
infection).
This study aimed to substantiate the value of the
HBsAg test in Taiwan’s HBV blood donor testing strategy.
More specifically, an HBV Markov model was constructed
and cost-effectiveness analysis was conducted in order to
reappraise the existing HBV screening strategy in Taiwan.
The integrated decision analytic model presented in this
paper is adjustable and can readily be reused to perform
economic evaluations of blood donor diagnostic tests for
other related TTIs (see Supporting Information).
First, it is expected that the results of this study can
provide feedback on the performance of the existing HBV
screening strategy in Taiwan. Second, provide recommen-
dations for maintaining, modifying, or changing the
existing strategy. Finally, provide useful insights, guiding
© 2020 International Society of Blood Transfusion
Vox Sanguinis (2021) 116, 564–573
Economic reappraisal of HBV blood donor tests 565
principles or policy recommendations for transfusion
practitioners especially in high-endemic areas.
Materials and methods
Blood donor examination
Taiwan has two centralized testing laboratories in
charge of the nationwide blood examination operations,
at Taipei and Kaohsiung blood centre. Whole blood
donations were screened for HBV between 1 June and
31 November 2017 (six months) at Taipei blood centre.
These donations were tested for HBsAg using Freedom
EVOlyzer and HBV-DNA using Procleix Ultrio Plus
assay (Grifols, Emeryville, CA, USA) on the fully auto-
mated TIGRIS system.
Presently, every donated whole blood unit is univer-
sally tested for HBsAg and HBV-DNA using the testing
algorithm (see Fig. 1) similar to the one described in the
pilot study for introducing NAT for screening blood
donors in Taiwan [20].
Primarily:
(i) HBsAg seronegative and seropositive samples are both
tested for HBV-DNA.
(ii) MP-8-NAT is universally used to test for HBV-DNA,
and reactive samples are confirmed with ID-NAT.
(iii) HBsAg seropositive samples that are unreactive to
MP-8-NAT are tested for HBV-DNA with ID-NAT.
(iv) ID-NAT reactive samples that test positive for both
HBV-DNA and HBsAg are considered as confirmed
cases of HBV infection.
(v) ID-NAT reactive samples that test positive for HBV-
DNA and negative for HBsAg are considered as con-
firmed cases of HBV infection.
(vi) Donors with samples that test HBsAg-positive and
are nonreactive to MP-NAT and ID-NAT are consid-
ered infectious (according to the blood bank policy
since there is no universal anti-HBc testing).
(vii) HBsAg seronegative and MP-NAT nonreactive dona-
tions are classified as negative, and their blood is
permissible for donation.
In order to substantiate the value of HBsAg in the
existing testing algorithm, four blood donation screening
strategies were considered as follows:
(1) No intervention (reference/base case)
(2) HBsAg (EIA)
(3) NAT (MP-8-NAT and ID-NAT)
(4) HBsAg (EIA) plus NAT (MP-8-NAT and ID-NAT)
To begin with, the residual risk of posttransfusion
infection for each screening strategy was determined and
then cost-effectiveness analysis was performed to deter-
mine the optimal strategy. Readers are referred to [21] for
566 T. B. Matanhire and S.-W. Lin
Figure 1 General schematic diagram of the HBV blood screening algorithm at Taipei blood centre.
a concise summary of health economic study types rele-
vant to blood safety intervention assessments.
Threat of infection and recipient exposure
Due to the transient nature of HBV, residual risk estima-
tion for HBV is more intricate as compared to that of HIV
and HCV [22]. This value is critical in the analysis of the
cost-effectiveness model and it mainly depends on the
incidence rate, prevalence and window period estimates.
The residual risk estimates were obtained using the inci-
dent rate-WP risk day equivalent model [22,23] (see Sup-
porting Information for more details). Over 80% of the
donations in Taiwan are from repeat donors with an aver-
age annual donation frequency of 1
75.
A recent population-based cohort study estimated
mean age of blood transfusion patients in Taiwan to be
58
4 years [24]. We considered the cohort that included
this demographic group and other limited populations of
susceptible potential recipients of blood products in
which the vaccine nonresponse rate is substantial.
Disease progression model
In order to estimate the costs and effects of an HBV transfu-
sion-transmitted infection, a Markov disease model was
constructed. Stages that have the most significant impact on
the disease progression were incorporated to resemble the
natural history of an HBV infection [25]. The Markov model
consists of seven states as depicted in Fig. 2.
All blood donors are tested in the initial state. It is
assumed that diagnostic window period transmissions to
transfusion recipients can occur from this initial state. It
is important to note two things for this Markov model.
First, the acute infection state is temporary because it has
short-term effects. In that regard, it is a tunnel state with-
out a self-loop arrow, which means that a patient cannot
spend no more than a single cycle in the acute infection
state [26]. Second, the Markovian memoryless property is
assumed to those recipients whose acute infection
resolves and acquires lifelong immunity. State transition
probabilities for the model were synthesized from litera-
ture and are provided in Table 1.
Testing costs, disease burden costs and health
utilities
Each state of the Markov model is associated with its
own costs and health utilities. The cost for HBV blood
screening comprised of reagent, equipment, personnel
and other ancillary costs. Costs were estimated at Taipei
blood centre. Disease burden costs for the chronic phases
of hepatitis B were obtained from a Taiwanese study
which was conducted from a national health insurance
© 2020 International Society of Blood Transfusion
Vox Sanguinis (2021) 116, 564–573
 Economic reappraisal of HBV blood donor tests 567

Figure 2 Markov state diagram for the natural history of HBV infection (for state transition probabilities, see Table 1).

Table 1 State transition probabilities.

From To Probability Range References

Acute Infection Chronic hepatitis B (CHB) 0
05 (0
02–0
10) [27]

CHB Compensated cirrhosis (CC) 0
0488 (0
02–0
09) [28]
CHB Hepatocellular carcinoma (HCC) 0
0115 (0
008–0
015) [28]
CHB Death 0
0014 (0
0011–0
0017) [28]
CC Decompensated cirrhosis (DC) 0
0572 (0
031–0
099) [28]
CC HCC 0
0391 (0
02–0
071) [28]
CC Death 0
05 (0
049–0
051) [28]
DC HCC 0
0652 (0
047–0
084) [28]
DC Death 0
2506 (0
144–0
39) [28]
HCC Death 0
3713 (0
233–0
433) [28,29]

perspective [30]. The disease burden costs were adjusted
for inflation using the consumer price index for Taiwan.
Due to the spontaneous nature in which acute hepatitis B
resolves, no healthcare costs were attached to the acute
state [31]. The health-related quality of life (HRQoL) of
the recipients who get posttransfusion chronic hepatitis B
is an essential parameter in determining the cost-effec-
tiveness of the blood safety interventions. The HRQoL
values used for the Markov model were obtained from
related literature. The testing costs, disease burden costs
and health utilities are summarized in Table 2. All costs
are denominated in United States dollars.
Economic evaluation
Cost-effectiveness analysis was performed using the Mar-
kov model to compare the four blood donor screening
strategies under consideration. The estimated residual risk
value for each strategy was used to determine the number
of recipients who could get infected with hepatitis B from
transfusion. This study considered HBV transmissions
from the blood donors to the blood recipients but did not
consider potential transmission from the blood recipients
to their partners.
To ensure that the analysis captured important differences
in costs and outcomes, the model was simulated using a life-
time horizon for 35 cycles, with each cycle analogous to a
year. The simulation was run for a cohort of 288 947. Half-
cycle correction was not applied to the model as it often
produces wrong results for discrete Markov models [32].
Costs and health effects were discounted at a rate of 3
5%
per annum for all the Markov state values.
Deterministic sensitivity analysis was conducted to
investigate the effect of varying specific input parameters.
Probabilistic sensitivity analysis (PSA) was also performed
in order to ascertain the overall parameter uncertainty.
Appropriate distributions were assigned for all the input
parameters and fitted using the mean and standard devia-
tion (Table 2). The PSA was repeated for 1000 simulations.
Computational techniques
The health economic model for donor blood testing was
implemented in RStudio using the ‘heemod’ package [33],

© 2020 International Society of Blood Transfusion

Vox Sanguinis (2021) 116, 564–573
568 T. B. Matanhire and S.-W. Lin

Table 2 Input variables.

Parameter Value (95% CI) Distribution Reference

Testing costs (US$)

HBsAg 2
96 (2
03–3
89) Gamma (l, r)* Estimated
NAT 9
80 (7
11–12
49) Gamma (l, r)* Estimated
Drug fees (US$) Gamma (l, r)*
CHB 65
80 (57
21–74
39) Gamma (l, r)* [30]
CC 92
40 (81
57–103
23) Gamma (l, r)* [30]
DC 422
40 (376
8–468
0) Gamma (l, r)* [30]
HCC 704
90 (625
7–784
1) Gamma (l, r)* [29,30]
Hospital fees (US$) Gamma (l, r)*
CHB 192
90 (171
2–214
6) Gamma (l, r)* [30]
CC 265
60 (238
3–292
9) Gamma (l, r)* [30]
DC 1440
30 (1279
7–1600
9) Gamma (l, r)* [30]
HCC 3765
40 (3399–4131) Gamma (l, r)* [29,30]
HRQOL
Acute infection 0
91 (0
89–0
93) Beta (29
81, 2
95) [28]
CHB 0
68 (0
66–0
70) Beta (10
28, 4
84) [28]
CC 0
69 (0
66–0
71) Beta (6
56, 2
95) [28]
DC 0
35 (0
32–0
37) Beta (1
99, 3
70) [28]
HCC 0
38 (0
36–0
41) Beta (1
06, 1
74) [28]
Residual risk probability
No intervention 2
59 9 10-4 (0
002159–0
003025) Binomial (8/5973, 5973) Estimated
HBsAg 17
6 9 10-6 (0
000011–0
000019) Binomial (1/56788, 56788) Estimated
NAT 7
79 9 10-6 (0
000001–0
000003) Binomial (1/128269, 128269) Estimated
HBsAg plus NAT 3
84 9 10-6 (0
000001-0
000003) Binomial (1/260198, 260198) Estimated
Transition probabilities
Acute Infection to CHB 0
05 (0
02–0
10) Binomial (0
05, 100) (see Table 1)
All Markov other states Multinomial (see Table 1)

*Gamma ~ (cost, √cost).

and the results were post-processed to give graphical
summaries by interfacing model results with the global
environment of the ‘BCEA’ package [34].
Results
for the current strategy is almost six times more than soci-
etal WTP threshold. The addition of MP-NAT to HBsAg sig-
nificantly increased the estimated total cost by over 70%.
However, there was slight increase in the estimated net
effectiveness.

Residual risk and infectious Window Period days
The residual risk estimates were found to be 17
6, 7
79
and 3
84 per million donations for ‘HBsAg’, ‘NAT’ and
‘HBsAg plus NAT’, respectively. The lengths of the
infectious window periods are presented in Table 3.
Cost-effectiveness summary
The cost-effectiveness results for the evaluated strategies
are summarized in Table 4. The incremental cost-effective-
ness ratio (ICER) for the current testing strategy in Taiwan
was estimated to be $US 443 614 per QALY, whilst the
willingness-to-pay (WTP) threshold that reflects local pref-
erences is $US 70 000 per QALY [35]. Therefore, the ICER
Deterministic sensitivity analysis
Deterministic sensitivity analyses showed that ICER values
were more sensitive to the residual risk estimates, transition
probability from acute infection to chronic hepatitis B, test-
ing costs and the discount rate, respectively. The tornado
plot for the variables which had the most impact on the
ICER values is shown in Fig. A1 (see Appendix 1). Very
low residual risk estimates significantly increased the ICER
value, whilst higher estimates decreased the ICER value. On
the other hand, very high discount rates significantly low-
ered the ICER value, whilst lower discount rates increased
the ICER value. For example, increasing the discount rate
to 8
5% per annum decreased the ICER value twofold and
this could lead to bias against the HBsAg intervention.

© 2020 International Society of Blood Transfusion

Vox Sanguinis (2021) 116, 564–573

Table 3 Residual risk and infectious WP days
Screening strategy Length of infectious WP (days)
HBsAg 41
3
NAT 20
8
HBsAg plus NAT 20
8
Table 4 Cost-effectiveness results
Interventions Estimated total costs (US$) Cost difference (US$)
No Intervention 2 111 712
HBsAg 18 123 911 55
42
NAT 58 023 825 138
10
HBsAg plus NAT 76 115 984 62
61
A cohort of 288 947 was simulated for 35 cycles.
Probabilistic sensitivity analyses
Figure 3 depicts the cost-effectiveness acceptability
curves (CEACs) that quantify the uncertainty in the eco-
nomic re-evaluation model of HBV testing strategies for
blood donors in Taiwan. Implementing the HBsAg strat-
egy is less costly, less effective and over 90% of its joint
density involves cost savings and minimal health gains
when compared to NAT strategies. On the other hand,
both NAT and HBsAg plus NAT strategies are effective to
a certain extent. None of their joint densities are cost sav-
ing but some of them are characterized by some health
gains. Consequently, their probability of cost-effective-
ness is an increasing function of the societal WTP and it
asymptotes to a value around the value of 0
3. It can be
seen that adding NAT to HBsAg slightly increases the
probability of cost-effectiveness as the societal willing-
ness to pay increases. However, the curves for strategies
with NAT flatten as the WTP increases. The overlaid grey
line on the CEACs depicts the cost-effectiveness accept-
ability frontier curve (CEAF) which depicts the range of
WTP values when each of the blood screening interven-
tions is optimal (also see Appendix 2).
Discussion
Taiwan is determined to halt viral hepatitis transmission
by 2030 in line with the vision of the World Health Orga-
nization (WHO) [36]. Hence, ensuring the safety of the
blood supply from TTIs like HBV remains of great impor-
tance in order to keep abreast with that vision. The cur-
rent strategy utilized by Taiwan interdicts a significant
© 2020 International Society of Blood Transfusion
Vox Sanguinis (2021) 116, 564–573
Economic reappraisal of HBV blood donor tests 569
Predicted incidence/million donations
17
6
7
79
3
84
Effectiveness (QALYs) Effectiveness difference (QALYs) ICER (US$/QALY)
10 098 275
10 112 957 0
05 1090
10 113 101 0
00036 386 584
10 113 124 0
00014 443 614
number of HBV infections from entering its blood supply.
However, this significant safety benefit comes at a dispro-
portionate increase in costs. This is evident from the
baseline ICER value of HBsAg plus NAT strategy
($443 614 per QALY) which is significantly higher as
compared to that for the HBsAg strategy ($1090 per
QALY). In light of the existing literature, this result com-
pares to the one reported for a related study conducted in
the Netherlands [37]. However, upon considering the ref-
erence case scenario of our study and that of the Nether-
lands, the screening strategy in Taiwan appears to be
more cost-effective. This is most likely as a result of the
high incidence of HBV in Taiwanese blood donors.
Our economic analysis showed that the value of
HBsAg was a decreasing function of the societal WTP
when compared with NAT. When HBsAg was combined
with NAT, it added minimal value as the societal WTP
increased (Fig. 3). Here, it is important to highlight the
issue of units that test positive for HBsAg only and neg-
ative for other markers such as anti-HBc and HBV-DNA.
These could be as a result of false positives, transient
positivity after vaccination or early stage acute HBV
[14,38,39]. However, to the best of our knowledge, no
such case in Taiwan has been reported as being early
acute HBV.
The scope of this article did not include follow-up
study on the test samples. Therefore, occult HBV (OBI)
cases were not classified. It was evident that use of MP-
8-NAT alongside ID-NAT helps overcome the safety gap
that results from not using the anti-HBc test in Taiwan.
Nevertheless, cases of occult OBI could still occur due to
the dilution factor from pooling the samples. We
570 T. B. Matanhire and S.-W. Lin
Figure 3 Cost-effectiveness acceptability curves and frontier. The overlaid grey line shows the effectiveness frontier curve.
speculate that implementing universal ID-NAT could fur-
ther reduce the risk of OBI in Taiwan with very limited
added efficiency and at a much higher cost.
There are some limitations to this study and to our
model. The donor test results that were utilized for analy-
sis were from a 6-month period. Obtaining test data that
covers a longer period would be beneficial in refining the
residual risk estimates. Similar to previous studies [37],
some of the input parameters that populate the model are
single study-based estimates (e.g. HRQoL values). There
are limited published data on region-specific health state
utilities for HBV. This could result in the increase of
uncertainty in such parameter estimates. Despite these
limitations, our model utilized some specific inputs based
on Taiwanese published data, for example the disease
burden costs. Previous studies in Taiwan have mostly
focused on efficacy research of adding NAT to HBsAg
screening in order to increase the safety against TTIs
[20,40]. The results of our study further demonstrate the
benefit of implementing such a screening strategy in
regions having significant OBI carriers such as Taiwan by
conducting cost-effectiveness analysis.
Conclusion
To conclude, our study showed that the strategy of using
universal HBsAg and MP-8-NAT together with confirma-
tory ID-NAT prevents a significant amount of HBV infec-
tions from entering the Taiwan blood supply. However, this
comes at a disproportionate increase in cost. We also found
that HBsAg is the preferred choice as indicated on the
CEAF. Additionally, occult hepatitis B remains the main
safety risk for blood transfusion in Taiwan. Excluding anti-
HBc-positive donors might help to eliminate potential
sources of occult HBV infection; however, the main draw-
back for anti-HBc testing is that it results in the loss of a
high percentage of non-infectious carriers thereby seri-
ously affecting the blood supply. Therefore, the HBsAg test
remains valuable in high-endemic areas, especially in those
with blood operators that do not afford to use NAT.
Acknowledgements
This study was partially supported by the Ministry of
Science and Technology of Taiwan under the grant num-
ber MOST 103-2410-H-011-012-MY3 and MOST 106-
2410-H-011-004-MY3. Any opinions, findings, and con-
clusions or recommendations expressed herein are those
of the authors and do not necessarily reflect the views of
the sponsors.
Conflicts of interest
The authors declare that they have no conflicts of interest
relevant to the manuscript submitted to Vox Sanguinis.
Authors contributions
All authors listed meet the authorship criteria and con-
tributed to research design, data acquisition, analysis and
© 2020 International Society of Blood Transfusion
Vox Sanguinis (2021) 116, 564–573
 Economic reappraisal of HBV blood donor tests 571

interpretation, writing, critical review and approval of the
manuscript. Here we indicate the specific contributions
made by each author. Conceptualization/Methodology:
Shi-Woei Lin, Tapiwa Blessing Matanhire. Acquisition,
Analysis and/Interpretation of data: Tapiwa Blessing
Matanhire, Shi-Woei Lin. Writing-Original Draft
Preparation: Tapiwa Blessing Matanhire. Review: Shi-
Woei Lin, Tapiwa Blessing Matanhire. Visualization:
Tapiwa Blessing Matanhire. Final approval of version for
submission: Shi-Woei Lin, Tapiwa Blessing Matanhire.
Supervision/Project Administration/Funding Acquisi-
tion: Shi Woei Lin.

References
1 Hou J, Liu Z, Gu F: Epidemiology and
prevention of hepatitis B virus infec-
tion. Int J Med Sci 2005; 2(1):50–7
2 Domanovic D: Assessing the risk of
transfusion-transmitted emerging
infections. ISBT Sci Ser 2016; 11
(Suppl. 1):68–75
3 Chen DS: Fighting against viral hep-
atitis: Lessons from Taiwan. Hepatol-
ogy 2011; 54(2):381–92
4 Lin CK, Leung JNS, So BKL, et al.:
Donor selection for blood safety: is it
still necessary? ISBT Sci Ser 2014; 9
(1):26–9
5 Lieshout-Krikke RW, Zaaijer HL, Van
DeLaar TJW: Predonation screening of
candidate donors and prevention of
window period donations. Transfusion
2015; 55(2):373–8
6 O’Flaherty N, Ushiro-Lumb I, Pomeroy
L, et al.: Transfusion-transmitted hep-
atitis B virus (HBV) infection from an
individual-donation nucleic acid (ID-
NAT) non-reactive donor. Vox Sang
2018; 113(3):300–3
7 Galel SA, Simon TL, Williamson PC,
et al.: Sensitivity and specificity of a
new automated system for the detec-
tion of hepatitis B virus, hepatitis C
virus, and human immunodeficiency
virus nucleic acid in blood and plasma
donations. Transfusion 2018; 58:649–
659
8 Kupek E: Residual risk of hepatitis-B-
infected blood donations. ISRN Infect
Dis 2013; 2013:1–15
9 Vermeulen M, Lelie N: The current sta-
tus of nucleic acid amplification tech-
nology in transfusion-transmitted
infectious disease testing. ISBT Sci Ser
2016; 11(Suppl. 2):123–128
10 Candotti D, Laperche S: Hepatitis B
virus blood screening: Need for reap-
praisal of blood safety measures?
Front Med 2018; 5:1–10
11 Busch MP, Should HBV: DNA NAT
replace HBsAg and/or anti-HBc
screening of blood donors? Transfus
Clin Biol 2004; 11(1):26–32
12 Coste J, Reesink HW, Engelfriet CP,
et al.: Implementation of donor
screening for infectious agents trans-
mitted by blood by nucleic acid tech-
nology : update to 2003. Vox Sang
2005; 88:289–303
13 Kuhns MC, Busch MP: New strategies
for blood donor screening for hepatitis
B virus: nucleic acid testing versus
immunoassay methods. Mol Diagn
2006; 10(2):77–91
14 Seed CR: Value of retaining HBsAg
donor screening where HBV NAT and
anti-HBc donor screening apply. ISBT
Sci Ser 2018; 13(1):70–5
15 Dodd RY: Epidemiology, performance
characteristics, or both? Transfusion
2017; 57(1):1–2
16 Kramer K, Verweij MF, Zaaijer HL: Are
there ethical differences between stop-
ping and not starting blood safety
measures? Vox Sang 2017; 112
(5):417–24
17 Roth WK: History and future of
nucleic acid amplification technology
blood donor testing. Transfus Med
Hemother 2019; 46(2):67–75
18 Tsai SL: Challenge of donor recruit-
ment. ISBT Sci Ser 2009; 4:302–6
19 Ni YH, Chang MH, Jan CF, et al.: Con-
tinuing decrease in hepatitis B virus
infection 30 years after initiation of
infant vaccination program in Taiwan.
Clin Gastroenterol Hepatol 2016; 14
(9):1324–30
20 Li L, Chen PJ, Chen MH, et al.: A pilot
study for screening blood donors in
Taiwan by nucleic acid amplification
technology: detecting occult hepatitis
B virus infections and closing the
serologic window period for hepatitis
C virus. Transfusion 2008; 48:1198–
2106
21 Custer B, Janssen MP: Health eco-
nomics and outcomes methods in risk-
based decision-making for blood
safety. Transfusion 2015; 55(8):2039–
47
22 Lelie N, Vermeulen M, Van Drimmelen
H, et al.: Direct comparison of three
residual risk models for hepatitis B
virus window period infections using
updated input parameters. Vox Sang
2020; 115(3):133–45.
23 Weusten J, Vermeulen M, DrimmelenH
Van, et al.: Refinement of a viral
transmission risk model for blood
donations in seroconversion window
phase screened by nucleic acid testing
in different pool sizes and repeat test
algorithms. Transfusion 2011; 51
(1):203–15
24 Lin SY, Hsu WH, Lin CC, et al.: Asso-
ciation of transfusion with risks of
dementia or Alzheimer’s disease: a
population-based cohort study. Front
Psychiatry 2019; 10:1–10
25 Pan CQ, Zhang JX: Natural history
and clinical consequences of hepatitis
B virus infection. Int J Med Sci 2005;
2(1):36–40
26 Sonnenberg FA, Beck JR: Markov
models in medical decision making: a
practical guide. Med Decis Mak 1993;
13:322–38
27 Liaw YF, Chu CM: Hepatitis B virus
infection. Lancet 2009; 373:582–592
28 Wiens A, Lenzi L, Venson R, et al.:
Economic evaluation of treatments for
chronic hepatitis B. Brazilian J Infect
Dis 2013; 17(74):418–26
29 Lang HC, Wu JC, Yen SH, et al.: The
lifetime cost of hepatocellular carci-
noma: A claims data analysis from a
medical centre in Taiwan. Appl Health
Econ Health Policy 2008; 6(1):55–65
30 Hsieh CR, Kuo CW: Cost of chronic
hepatitis B virus infection in Taiwan. J
Clin Gastroenterol 2004; 38(10 Suppl
3):S148–52
31 Jindal A, Kumar M, Sarin SK: Man-
agement of acute hepatitis B and

© 2020 International Society of Blood Transfusion

Vox Sanguinis (2021) 116, 564–573
572 T. B. Matanhire and S.-W. Lin

reactivation of hepatitis B. Liver Int
2013; 33(SUPPL. 1):164–75
32 Barendregt JJ: The half-cycle correc-
tion: Banish rather than explain it.
Med Decis Mak 2009; 29(4):500–2
33 Filipovic-Pierucci A, Zarca K, Durand-
Zaleski I: Markov models for health
economic evaluation: the R Package
heemod. R package version 0.12.0.
34 Baio G, Berardi A: BCEA: Bayesian
cost effectiveness analysis. R package
version 2.3-1.1.
35 Shiroiwa T, Sung YK, Fukuda T, et al.:
International survey on willingness to
pay (WTP) for one additional QALY
gained: what is the threshold of cost
effectiveness? Health Econ 2010;
19:422–37
36 WHO. Global Health Sector Strategy
on Viral Hepatitis 2016–2021. 2016:1–
56 Available from: http://www.who.
int/hepatitis.
37 Borkent-Raven BA, Janssen MP, Van
DerPoel CL, et al.: Cost-effectiveness of
additional hepatitis B virus nucleic acid
testing of individual donations or mini-
pools of six donations in the Nether-
lands. Transfusion 2009; 49(2):311–9
38 Yuen MF, Wong DKH, Lee CK, et al.:
Transmissibility of hepatitis B virus
(HBV) infection through blood trans-
fusion from blood donors with occult
HBV infection. Clin Infect Dis 2011;
52(5):624–32
39 Dodd RY, Nguyen ML, Krysztof DE,
et al.: Blood donor testing for hepatitis
B virus in the United States: is there a
case for continuation of hepatitis B
surface antigen detection? Transfusion
2018; 58(9):2166–70
40 Yang MH, Li L, Hung YS, et al.: The
efficacy of individual-donation and
minipool testing to detect low-level
hepatitis B virus DNA in Taiwan.
Transfusion 2010; 50:65–74

Supporting Information
Additional Supporting Information may be found in the online version of this article:

ICER (US$/QALY)

HBsAg residual risk 343,750 1,150,000

Acute infection to Chronic HBV (transition probability) 237,500 581,250

NAT cost 225,000 513,260

Discount rate 166,250 392,600

NAT residual risk 281,750 386,000

Figure A1 Tornado plot. The chart shows the sensitivity of the parameters around the ICER value of 386 585 between the HBsAg and the NAT strate-
gies.

Table A1 Net monetary benefit.

Societal WP value

$80 000 $160 000 $240 000 $320 000 $6 40,000 $1 280 000 $2 560 000 $5 120 000

No Intervention 0 0 0 0 0 0 0 0
HBsAg 4009 8074 12 138 16 203 32 462 64 981 130 017 260 091
NAT 3899 7993 12 086 16 179 32 553 65 300 130 794 261 781
HBsAg plus NAT 3848 7953 12 057 16 162 32 581 65 418 131 092 262 441

The number in bold shows the highest net monetary value at that particular WTP threshold value, and the corresponding strategy is optimal at that
threshold.

© 2020 International Society of Blood Transfusion

Vox Sanguinis (2021) 116, 564–573
 Economic reappraisal of HBV blood donor tests 573

Supplementary Material R code for the economic evaluation of hepatitis B virus testing strategies

Table S1 Input parameters and risk calculations results.

Table S2 Test yield and result classification.
Table S3 Parameters for fitting the Beta distribution.
Table S4 Multinomial distribution for state transition probabilities.
Figure S1 Pdf chart for CHB HRQoL (Beta).
Supplementary Material #Model Parameters

© 2020 International Society of Blood Transfusion

Vox Sanguinis (2021) 116, 564–573
 Vox Sanguinis (2021) 116, 574–580
© 2020 International Society of Blood Transfusion
ORIGINAL PAPER DOI: 10.1111/vox.13024

Blood transfusion activity in a general hospital during the

COVID-19 pandemic
Karen Mar
ın-Mori,1,† Isabel Gonz
alez-Gasc


on y Mar
ın,1,† Mar
ıa-Angeles Foncillas-Garc
ıa,1 Carolina Mu~
noz-Novas,1
Mar
ıa Infante, Juan Churruca-Sarasqueta, Elena Landete-Hern

1 1 1
andez, Bego~ na Bueno-Garc
ıa, Mercedes Duffort-Falco4
3

& Jos
e-Angel Hern
andez-Rivas 1,2

1
Transfusion Service and Hematology Department, Hospital Universitario Infanta Leonor, Hospital Virgen de la Torre, Madrid, Spain
2
Universidad Complutense de Madrid, Madrid, Spain
3
Intensive Care Unit, Hospital Universitario Infanta Leonor, Hospital Virgen de la Torre, Madrid, Spain
4
Internal Medicine Department, Hospital Universitario Infanta Leonor, Hospital Virgen de la Torre, Madrid, Spain

Received: 27 May 2020,
revised 24 September 2020,
accepted 13 October 2020,
published online 2 November 2020
Background The COVID-19 outbreak has affected almost all hospital depart-
ments, including transfusion services. However, the demand for transfusions in a
general hospital designated to deal with COVID-19 patients has not been anal-
ysed before.
Study Design and Methods A retrospective study was conducted to evaluate
blood transfusion practices from 15 March to 14 April 2020 at Hospital Universi-
tario Infanta Leonor (Madrid, Spain). During this month, with few exceptions, the
hospital became a ‘COVID-19’ centre. In addition, transfusion rates during this
time frame and the same period over the last 4 years were compared.
Results From 15 March to 14 April 2020, only 254 blood components were
transfused, resulting in a 49
3% reduction over the previous year. Interestingly,
in critically ill patients, the red blood cell (RBC) transfusion/bed ratio signifi-
cantly decreased during this period (0
92) compared to the same ratio over the
past 4 years (2
70) (P = 0
02). Of note, 106 blood components (95 RBC; 11 plate-
let concentrates) were transfused to only 36 out of 1348 COVID-19 patients
(2
7%). The main reason for RBC transfusion in COVID-19 patients was a previ-
ous underlying disease (44%) followed by bleeding (25%) and inflammatory
anaemia (25%).
Conclusion This is the first study to report a decrease in blood transfusions dur-
ing the COVID-19 pandemic in a general hospital and especially in the intensive
care unit. The results of this study suggest that COVID-19 does not generally
induce transfusion requiring anaemia, being the main causes for transfusion in
these patients underlying conditions or bleeding.
Key words: COVID-19, patient blood management, transfusion.

China. At the beginning of January 2020, a new beta-

Introduction
In late December 2019, various severe pneumonia cases
of unknown cause emerged in Wuhan, Hubei Province,
Correspondence: Karen Mar
ın-Mori, Haematology Department, Hospital
Universitario Infanta Leonor. Hospital Virgen de la Torre, C/Gran V
ıa del
Este 80, Madrid 28031. Spain
E-mail: karen.marin@salud.madrid.org
† of 13 August, the disease has affected more than 20 M
K M-M and I G-G y M contributed equally to this article.
coronavirus was identified, which was called severe acute
respiratory syndrome coronavirus-2 (SARS-CoV-2) and
was related to Wuhan pneumonia patients [1].
During January 2020, isolated imported cases began to
appear in several countries around the world. On 11
March, the WHO declared the disease as a pandemic, and
by 14 March the state of emergency was declared in
Spain with severe restrictions on social relationships. As

574

patients causing more than 740 000 deaths. Thus, Spain
has been severely hit by the pandemic and harbours one
of the highest disease burden worldwide with more than
300 000 affected patients and more than 28 000 deaths
[2]. Madrid has been the epicentre of the Spanish out-
break, with around 30% of the confirmed cases in Spain
[3].
The new SARS-CoV-2 produces mild-to-moderate
symptoms in approximately 85–90% of cases that do not
require hospital admission. Accordingly, most COVID-19
patients do not develop significant anaemia nor thrombo-
cytopenia [4,5]. However, 10% of patients may suffer a
more severe clinical course (bilateral pneumonia, respira-
tory failure, septic shock, coagulopathy and multiple
organ failure, among others). This phase can lead to
patient’s death in the context of systemic inflammatory
response syndrome associated with cytokine storm [6].
Among critical COVID-19 patients, a decrease in haemo-
globin and platelet levels is common. Indeed, thrombocy-
topenia has been associated with a worse outcome [4].
Furthermore, this patients might suffer a coagulation dis-
order that results in a high incidence of thrombotic
events, which is also related to prognosis and mortality
[7]. This coagulopathy is characterised by hypercoagula-
bility with high levels of D-dimers and fibrinogen and
might resemble, in some cases, disseminated intravascular
coagulopathy (DIC). However, based on currently avail-
able knowledge, an increment in haemorrhagic risk has
not been reported before [8].
In our region, the blood collection centres collect and
distribute blood to hospital transfusion departments. One
of the problems associated with the COVID-19 crisis has
been the decrease in blood donations, due to the large
number of infected people and the social restrictions.
Nowadays, scarce data are available about the amount of
blood components consumed during the pandemic [9–12]
and the transfusion requirements for COVID-19 patients
[13]. Very recently some studies pointed out a decrease of
blood transfusion among COVID-19 patients in compar-
ison with non-COVID-19 patients and a decrease in blood
components usage during the pandemic relative to the
weeks before and after [13–15].
In this setting, the aims of this study are (1) to evaluate
the transfusion of blood components at our Hospital
between 15 March and 14 April 2020; (2) to compare the
results with the same period of time during the last
4 years and (3) to analyse the transfusion requirements in
COVID-19 patients.
Material and methods
We conducted a single-centre retrospective study evaluat-
ing blood transfusion requirements from 15 March to 14
© 2020 International Society of Blood Transfusion
Vox Sanguinis (2021) 116, 574–580
Transfusion during COVID-19 crisis 575
April 2020 at the Transfusion Department of the Hospital
Universitario Infanta Leonor (Madrid, Spain). This is a
second-level general hospital with 362 beds and teaching
activities in some departments. The haematology service
treats patients with acute leukaemia and performs, in col-
laboration with a third level hospital, autologous bone
marrow transplantations (patient stays al our centre since
day + 1 until the discharge date). The intensive care unit
(ICU) under normal conditions is a general facility in
charge of non-surgery critically ill patients and has a
capacity of eight beds. Cardiovascular surgery and venti-
lation with extracorporeal oxygenation membranes
(ECMO) is not available at this centre.
Our centre has a disaster programme with a decision-
making committee. However, a specific COVID-19 surge
plan for the pandemic was established from 30 January.
This plan has been evolving over time. As the pandemic
began, all surgeries and regular medical appointments
were cancelled or referred to other centres, including
deliveries, with the exception of oncological and haema-
tological treatments and a non-COVID-19 ward with 30
beds. In addition, general measures to prevent SARS-coV-
2 were adopted, such as the creation of COVID-19 free
circuits and the restriction of visitors, among others.
The hospital has an active patient blood management
(PBM) programme with trimestral meetings composed of
surgical, medical and intensive care unit specialists and
nurses. Indications for blood transfusion are in accor-
dance with the recommendations of the Spanish blood
transfusion society, based on patient´s clinical condition
and suggesting the threshold for transfusion of red blood
concentrates (RBC) in a haemoglobin level of 7 g/dl. Dur-
ing the pandemic, the transfusion service did not imple-
ment a specific blood conservation strategy.
We then described the blood usage for this 4-week per-
iod (15 March to 14 April 2020) and the same 4-week
period on the last four years in all hospitalized patients.
We selected that month since it reflects the period with
largest number of COVID-19 patients admitted to the hos-
pital (Fig. 1), and at that point, the centre became a
‘COVID-19’ Hospital.
We also compared transfusion rates at the ICU during
this period with the rates of the last four years. As
COVID-19 brought a high demand for ICU and our ICU
triplicated its capacity for non-surgery patients, we calcu-
lated the transfusion ratio dividing the number of prod-
ucts used during this month by the median of occupied
beds per day.
Clinical data were obtained from electronic medical and
hospital records (SELENE SP12) and the transfusion data
from the e-PROGESA computer system (version 5.0.3.).
The local Ethics Committee at Hospital Universitario
Infanta Leonor approved the study.
576 M.-M. Karen et al.

Fig. 1 Number of patients with COVID-19 diagnosis admitted from 1 March to 30 April 2020 in bars chart. [Colour figure can be viewed at wileyonline
library.com]

Statistical analysis was performed using SPSS Table 1 Distribution of transfused blood components per year at the
21.0 software package (SPSS, Chicago, IL, USA). Mann– hospital in the last 4 years (2016–2019)
Whitney tests were used to compare quantitative
Blood Components 2016 2017 2018 2019 Total (mean)
variables. Statistical significance was defined as a
<0
05. RBC 5109 5159 5209 5079 5139
PC 543 513 437 458 488
FFP 308 329 422 251 328
Results Total 5960 6001 6068 5788 5983
By 14 April, 1348 adult patients with a confirmed SARS-
FFP, fresh-frozen plasma; PC, platelet concentrate; RBC, red blood cell.
COV-2 infection by a pharyngeal smear RT-PCR had been
admitted to our centre in the previous 30 days. Forty-
nine of them were transferred to the ICU (3
6%). Mortal- Table 2 Comparative transfused blood components: 15 March—14 April
ity among critically ill patients with COVID-19 was during the previous 4 years and the COVID-19 2020 outbreak
65
3% (32/45). In addition, 210 (15
58%) deaths occurred
due to COVID-19 during this month of follow-up for the Blood Components 2016 2017 2018 2019 2020
other hospitalized patients. Median length of stay was
RBC 453 391 436 385 225
7 days for the whole cohort and 9 days for the ICU PC 51 54 34 33 29
patients. FFP 32 32 18 83 0
Total 536 477 488 501 254

The use of blood products during the COVID-19
pandemic
Around 6000 blood components (5000 red blood concen-
trates [RBC]; 650 platelet concentrates [PC]; 350 fresh-
frozen plasma [FFP] units) are transfused annually at our
hospital. The distribution of transfusions in the past four
years is shown in Table 1. Between 15 March and 14
April 2020, the total number of blood components trans-
fused decreased dramatically compared to the same per-
iod of previous years. As Table 2 illustrates, only 254
blood components were transfused, which results in a
49
3% reduction over the previous year: 225 RBC (41
6%
reduction); 29 PC (12
12% reduction); and 0 FFP units
(100%).
FFP, fresh-frozen plasma; PC, platelet concentrate; RBC, red blood cell.
Transfusion in critically ill patients during the
COVID-19 pandemic
From 15 March to 14 April, the ICU tripled its capacity
with around 26 admitted patients, when in normal condi-
tions approximately 8 beds are available.
All patients admitted at the ICU at this moment were
COVID-19 patients. We observed that the transfusion ratio
during the evaluated period was 0
92 for RBC (24 RCB/26
beds), 0
04 for PC (1 PC/26 beds) and 0 for FFP units. The
ratio of transfused components during the last 4 years in

© 2020 International Society of Blood Transfusion

Vox Sanguinis (2021) 116, 574–580

the same period was significantly higher for RBC (ra-
tio = 2
7, 100 RBC/37 beds) (P = 0
02), and for FFP units
(ratio = 1
35, 50 FFP/37 beds) (P = 0
012). It was also
higher for PC (ratio = 0
49, 18 PC/37 beds) although not
reaching significance (P = 0
1; Table 3).
COVID-19 patients who required transfusion
From 15 March to 14 April, 106 blood components (95
RBC; 11 CP) were transfused to only 36 of 1348 COVID-
19 patients (2
7%). Characteristics of these 36 patients are
shown in Table 4. Median age was 75 years (34–94), and
61
1% of the patients were male (22/36). Interestingly, all
of them had underlying chronic medical conditions
including 47% (17/36) of the patients with cancer (8/36
onco-haematological; 9/36 solid tumour). The indications
for transfusion were underlying medical condition
(N = 16, 44%), bleeding (N = 10, 28%) and inflammatory
anaemia (N = 10, 28%). Detailed information about
underlying medical conditions is summarized in Table 4.
Of the 10 patients transfused due to inflammatory anae-
mia, 5 of them had a documented bacterial infection
which led to septic shock.
Out of the 36 transfused patients, 23 (64%) died due to
COVID-19 complications.
Discussion
The COVID-19 outbreak has become the most important
health, economic and social problem in recent decades.
Spain is now probably the European country in the world
with more COVID-19 cases related to its population [3].
Since its onset in December 2019, some groups have
reported a decline in the number of donations, blood sup-
plies and concerns about blood safety [9–11]. However, to
date, no evidence of SARS-CoV-2 transmission by blood
products has been published [16]. On the other hand,
many of the hospitals in countries particularly affected
by the pandemic have become ‘COVID-19’ centres, with a
Table 3 Number of transfused blood components at the Intensive Care
Unit in the last four years vs 2020 (period 15 March—14 April)
Ratio last Ratio
four years: N N Blood
Blood Blood Components/N Components/N
Components Beds ICU Beds ICU 2020
RBC 2
70 0
92
PC 0
49 0
04
FFP 1
35 0 0
01
Total 4
68 1
00
FFP, fresh-frozen plasma; ICU, Intensive Care Unit PC; platelet concen-
trate; RBC, red blood cell.
© 2020 International Society of Blood Transfusion
Vox Sanguinis (2021) 116, 574–580
Transfusion during COVID-19 crisis 577
drastic decline or absence of elective surgeries, paediatric
care, other medical admissions and obstetric care. Conse-
quently, blood use decreased at these institutions
[13,17,18]. The development and implementation of
patient blood management programmes during the pan-
demic might have also have played an important role in
the decline in transfusions [19].
In this study, we confirm a significant decrease of
around 50% in total blood components transfusion during
the pandemic in a ‘COVID-19 Hospital’. It should be noted
that in the period of time studied, our institution multi-
plied by 2
5 the number of conventional beds and by 3
the number of ICU beds compared with usual occupation.
Similar results have been observed in recent reports from
Maryland, Singapore, China or Chicago with a decrease
on blood product transfusion that ranged between 12 and
42% [13–15,20]. We observed that RBC transfusion
decreased around 41%, platelet transfusion decreased
around 12% and plasma units were not transfused during
this period, as compared to the same period in previous
years. Other groups have also showed a minimal use of
FFP among COVID-19 patients [14,15,20]. Even though
specially critically ill COVID-19 patients develop coagu-
lopathy, it is usually associated with thrombotic events,
elevation of D-dimer and fibrinogen and usually does not
impact in transfusion needs [8].
Furthermore, it is worth emphasizing that most of the
blood component transfusions performed to non-COVID-
19 patients were requested by Haematology and Oncology
Departments. Due to the SARS-CoV-2 complications, ICUs
have been adapted to increase their usual capacity to
address the need for invasive mechanical ventilation [21].
Similarly, from 15 March to 14 April, the ICU of our hos-
pital tripled its capacity. We observed that, despite the
increase in critically ill patients, the ICU RBC transfusion/
bed ratio was significantly lower than it was on the pre-
vious 4 years (P = 0
02, Table 3). Prior studies have
shown that 95% of ICU patients are found to have anae-
mia and 85% of those with a 7-day ICU admission or
longer received at least 1 RBC unit [22–24]. However, we
hypothesize that the shift in practice (almost all admitted
and all critically ill patients were COVID-19 at this time,
cancellation of elective surgery and non-essential care)
could have been involved in the decline in blood transfu-
sion. One limitation of our study is that at our hospital,
P Value ventilation with extracorporeal oxygenation membranes
(ECMO) is not available, and patients who require this
0
02
modality are referred to other centres. This ventilation
0
1
modality is known to increase bleeding, with the need for
around 2–4 units of packed red blood cells and 4–8 units
of platelet concentrate daily [25,26]. Therefore, the results
of our work cannot be generalized to other centres that
use ECMO ventilation.
578 M.-M. Karen et al.

Table 4 Description of transfused confirmed COVID-19 patients

Blood component
Transfusion (During
Patient Sex/Age Previous Disease ICU ABO/Rh Group Reason for transfusion hospitalization) Death

1 M/87 AML No O Rh+ Underlying disease 5 RBC No

2 M/84 AML No O Rh+ Underlying disease 5 RBC No
2 PC
3 M/85 HTN Yes O Rh+ Inflammatory anaemia 1 RBC Yes
DM
4 M/74 HTA Yes A Rh+ Inflammatory anaemia 1 RBC Yes
CRF
5 M/72 None Yes A Rh- Septic shock due to bacterial infection 2 RBC Yes
6 F/58 HTN Yes A Rh- Bleeding 6 RBC Yes
7 F/72 SLE No A Rh+ Septic shock due to bacterial infection 2 RBC Yes
APS
8 F/54 HL No O Rh- Underlying disease 3 RBC Yes
9 F/62 CLL AA No A Rh- Underlying disease 4 RBC Yes
5 PC
10 F/82 MDS No O Rh+ Underlying disease 2 RBC Yes
11 M/87 Prostate Cancer No A Rh+ Bleeding 1 RBC No
12 M/87 AF No A Rh+ Underlying disease 3 RBC No
CRF
13 F/70 HTND No B Rh+ Underlying disease 4 RBC Yes
MCRF
14 M/83 HTNAF No O Rh+ Bleeding 2 RBC Yes
15 M/82 AF No O Rh+ Bleeding 2 RBC Yes
Gastric Cancer
16 F/84 HTN No B Rh+ Bleeding 1 RBC No
Humerus Fracture
17 M/66 AML No O Rh+ Underlying disease 2 RBC Yes
18 M/72 Iron Deficiency No O Rh+ Underlying disease 1 RBC No
Anaemia
HTN
19 F/34 Gestational No A Rh+ Septic shock due to bacterial infection 2 RBC Yes
Trophoblastic 1 PC
Neoplasia
20 M/88 HTN No O Rh+ Bleeding 5 RBC Yes
Urothelial Cancer
21 F/86 MDS No A Rh+ Underlying disease 2 RBC Yes
22 F/72 Colon Cancer No A Rh+ Septic shock due to bacterialinfection 4 RBC No
23 M/75 HTNAF No A Rh+ Bleeding 2 RBC No
24 F/91 MDS HTN No B Rh+ Underlying disease 1 RBC No
25 M/84 Lung Cancer No O Rh+ Underlying disease 2 RBC Yes
26 M/75 HTN Yes O Rh+ Inflammatory anaemia 1 RBC Yes
27 F/45 Iron Deficiency No O Rh+ Underlying disease 2 RBC No
Anaemia
28 M/72 HTN Yes A Rh+ Bleeding 3 RBC Yes
29 M/80 Lung Cancer No B Rh+ Underlying disease 4 RBC No
2 PC
30 M/81 Bladder Cancer No A Rh+ Bleeding 2 RBC Yes
31 M/94 HTN DM No O Rh+ Underlying disease 3 RBC No
32 M/71 ET Yes O Rh+ Bleeding 4 RBC No
33 F/54 Crohn´s disease Yes AB Rh+ Septic shock due to bacterial infection 4 RBC Yes
HTN
34 M/62 No AB Rh+ Underlying disease 4 RBC Yes

© 2020 International Society of Blood Transfusion

Vox Sanguinis (2021) 116, 574–580
 Transfusion during COVID-19 crisis 579

Table 4 (Continued)

Blood component
Transfusion (During
Patient Sex/Age Previous Disease ICU ABO/Rh Group Reason for transfusion hospitalization) Death

Metastatic Prostate
Cancer
35 M/68 HTN Yes A Rh+ Inflammatory anaemia 1 RBC Yes
AF
36 F/56 DM Yes A Rh + Inflammatory anaemia 2 RBC Yes
Hypothyroidism 1 PC

AA, aplastic anaemia; AF, atrial fibrillation; AML, acute myeloid leukaemia; APS, Antiphospholipid Syndrome; CLL, chronic lymphocytic leukaemia, CRF,
chronic renal failure; DM, diabetes; ET, essential thrombocythemia; HL, Hodgkin lymphoma; HTN, hypertension; MDS, Myelodysplastic Syndrome; SLE, sys-
temic lupus erythematosus.

Next, we observed that only 2
7% (36) of the patients
admitted with COVID-19 were transfused. Of note, all of
them had previous medical conditions including a high
percentage of patients with an active neoplasm (44%).
When we analysed the reason for transfusion, we observed
that a very low proportion of patients were transfused due
to the virus-induced inflammatory state, which seems very
interesting, and has not been reported so far. Only 10
patients received RBC due to their critically ill condition,
and in 5 of them, another bacterial infection was docu-
mented. The need for blood transfusion in COVID-19
patients might be also considered a poor prognostic factor,
as 64% (23/36) of the COVID-19 transfused patients died.
This is not surprising, as the need of blood components
might reflect a serious underlying condition [27].
In this study, we analysed blood usage during the per-
iod of 30 days in which the pandemic most hardly
affected our Hospital. However, as shown in Fig. 1
COVID-19 patients were admitted to the centre before and
after those dates. Therefore, another limitation of the
study is that some patients, especially those with a pro-
longed critical condition might be under-represented.
In conclusion, during the first wave of COVID-19 pan-
demic, a very significant decrease in transfused blood
products has been observed, which can help us to design
our transfusion policy (blood procurement, processing
and use) and be prepared to undertake possible further
outbreaks of the disease in the next few months. Most of
COVID-19 patients did not require blood products despite
their critical condition.
Acknowledgements
We thank the technicians of the Transfusion Service for
their technical assistance.
Conflicts of interest
The authors declare no conflict of interests.
Author contributions
KMM, JAH and IGGM conceived the idea, collected, anal-
ysed the data and wrote the manuscript. MAFG, CMN, MI,
JCHS, ELH, BBG, MDF collected, analysed the data and
reviewed the manuscript. JAH provided critical revisions
to the manuscript.

References
1 Lai C-C, Shih T-P, Ko W-C, et al.: Sev-
ere acute respiratory syndrome coron-
avirus 2 (SARS-CoV-2) and
coronavirus disease-2019 (COVID-19):
the epidemic and the challenges. Int J
Antimicrob Agents 2020; 55:105924
2 WHO: Coronavirus Disease (COVID-19)
Dashboard [Internet]. Available from:
https://covid19.who.int [Last accessed
on 2020 August 13]
3 Ministerio de Sanidad, Consumo y
Bienestar Social: Professionals -
Situaci

on actual Coronavirus [Inter-
net]. Available from: https://www.
mscbs.gob.es/en/profesionales/salud
Publica/ccayes/alertasActual/nCov-
China/situacionActual.htm [Last
accessed on 2020 August 13]
4 Liao D, Zhou F, Luo L, et al. Haemato-
logical characteristics and risk factors
in the classification and prognosis
evaluation of COVID-19: a retrospec-
tive cohort study. Lancet Haematol
2020; 7:e671–e678
5 Fan BE, Chong VCL, Chan SSW, et al.:
Hematologic parameters in patients
with COVID-19 infection. Am J Hema-
tol 2020; 95:E131–E134
6 Zhou M, Zhang X, Qu J: Coronavirus
disease 2019 (COVID-19): a clinical

© 2020 International Society of Blood Transfusion

Vox Sanguinis (2021) 116, 574–580
580 M.-M. Karen et al.
update. Front Med 2020; 14:
126–135
7 Jin S, Jin Y, Xu B, et al.: Prevalence
and impact of coagulation dysfunction
in COVID-19 in China: a meta-analy-
sis. Thromb Haemost 2020. https://doi.
org/10.1055/s-0040-1714369
8 G€orlinger K, Dirkmann D, Gandhhi A,
et al.: COVID-19 associated coagulopa-
thy and inflammatory response: what
do we know already and what are the
knowledge gaps? Anesth Analg 2020;
131:1324–1333
9 Raturi M, Kusum A: The blood supply
management amid the COVID-19 out-
break. Transfus Clin Biol 2020;
27:147–151
10 Pagano MB, Hess JR, Tsang HC, et al.:
Prepare to adapt: blood supply and
transfusion support during the first 2
weeks of the 2019 novel coronavirus
(COVID-19) pandemic affecting Wash-
ington State. Transfusion 2020;
60:908–911
11 Garraud O: COVID-19: is a paradigm
change to be expected in health care
and transfusion medicine? Transfus
Clin Biol 2020; 27:59–60
12 Chang L, Yan Y, Wang L: Coronavirus
disease 2019: Coronaviruses and blood
safety. Transfus Med Rev 2020;
34:75–80
13 Barriteau CM, Bochey P, Lindholm PF,
et al.: Blood transfusion utilization in
hospitalized COVID-19 patients. Trans-
fusion. 2020. https://doi.org/10.1111/
trf.15947
14 Murphy C, Jackson B, Fontaine M:
Tools for rapid analysis of blood usage
and inventory during the COVID-19
pandemic. Transfusion 2020. https://d
oi.org/10.1111/trf.15996
15 Fan BE, Ong KH, Chan SSW, et al.:
Blood and blood product use during
COVID-19 infection. Am J Hematol
2020; 95:E158–E160
16 Cai X, Ren M, Chen F, et al.: Blood
transfusion during the COVID-19 out-
break. Blood Transfus. 2020; 18:79–82
17 Goh KJ, Wong J, Tien J-CC, et al.:
Preparing your intensive care unit for
the COVID-19 pandemic: practical
considerations and strategies. Crit Care
Lond Engl 2020; 24:215
18 Stanworth SJ, New HV, Apelseth TO,
et al.: Effects of the COVID-19 pan-
demic on supply and use of blood for
transfusion. Lancet Haematol 2020; 7:
e756–e764
19 Shander A, Goobie SM, Warner MA,
et al.: Essential role of patient blood
management in a pandemic: a call for
action. Anesth Analg 2020; 131:74–85
20 Wang Y, Han W, Pan L, et al.: Impact
of COVID-19 on blood centres in
© 2020 International Society of Blood Transfusion
Zhejiang province China. Vox Sang
2020; 115:502–506
21 Baron DM, Franchini M, Goobie SM,
et al.: Patient blood management dur-
ing the COVID-19 pandemic: a narra-
tive review. Anaesthesia 2020;
75:1105–1113
22 Corwin HL, Parsonnet KC, Gettinger A:
RBC transfusion in the ICU. Is there a
reason? Chest 1995; 108:767–771
23 Vincent JL, Baron J-F, Reinhart K,
et al.: Anemia and blood transfusion
in critically ill patients. JAMA 2002;
288:1499–1507
24 Jandu AS, Vidgeon S, Ahmed N:
Anaemia and transfusion triggers in
critically ill patients - What we have
learnt thus far. J Intensive Care Soc
2019; 20:284–289
25 Koeckerling D, Pan D, Mudalige NL,
et al.: Blood transfusion strategies and
ECMO during the COVID-19 pandemic.
Lancet Respir Med 2020; 8:e40
26 Kowalewski M, Fina D, Słomka A,
et al.: COVID-19 and ECMO: the inter-
play between coagulation and inflam-
mation-a narrative review. Crit Care
Lond Engl 2020; 24: 205
27 Kamper-Jørgensen M, Ahlgren M,
Rostgaard K, et al.: Survival after
blood transfusion. Transfusion 2008;
48:2577–2584
Vox Sanguinis (2021) 116, 574–580
 Vox Sanguinis (2021) 116, 581–590
© 2020 The Authors.
ORIGINAL PAPER Vox Sanguinis published by John Wiley & Sons Ltd on behalf of International Society of Blood Transfusion
DOI: 10.1111/vox.13034

Costs associated with transfusion therapy in patients with

myelodysplastic syndromes in Sweden: a nationwide
retrospective cohort study
Jingcheng Zhao,1 Torsten Dahl
en1,2 & Gustaf Edgren1,3
1
Department of Medicine Solna, Clinical Epidemiology Division, Karolinska Institutet, Stockholm, Sweden
2
Hematology Center, Karolinska University Hospital, Stockholm, Sweden
3
Department of Cardiology, S€odersjukhuset, Stockholm, Sweden

Received: 1 September 2020,
revised 24 October 2020,
accepted 2 November 2020,
published online 18 November 2020
Abstract
Background and objectives Blood transfusion is a cornerstone therapy for many
patients with myelodysplastic syndromes (MDS), but ranges from few to no trans-
fusions to intensive transfusion therapy. To date, no large studies have described
transfusion use or costs for patients with MDS, accounting for the range of dis-
ease severity.
Materials and methods A nationwide cohort study was conducted with all
patients diagnosed with MDS in Sweden between 2008 and 2017, based on the
Swedish MDS register and the Swedish part of the Scandinavian Donations and
Transfusions Database 3 (SCANDAT3-S). Patients were followed from diagnosis
until death, emigration, allogeneic hematopoietic stem cell transplantation or end
of follow-up. Average cumulative transfusion count and costs over time were
calculated, stratified by the revised international prognostic scoring system
(IPSS-R) and age at diagnosis. Costs calculations used data on incident transfu-
sions and laboratory testing and were divided into: direct material costs, direct
labour costs and laboratory costs.
Results In total, 2311 patients were included in the cohort. In the first four years
after diagnosis, patients in the very low IPSS-R category received on average 25
red cell (95% confidence interval, 20–32) and 4 (3–7) platelet transfusions. Con-
versely, patients in the very high-risk category received on average 171 (135–
200) red cell and 66 (51–78) platelet transfusions. Correspondingly, transfusion
costs ranged from $8805 ($6482–$11 625) to $80 106 ($61 460–$95 792).
Conclusion Transfusion count and costs for patients with MDS increased mark-
edly with IPSS-R risk category, but were similar across age groups. Transfusion
costs were considerable for the highest IPSS-R risk categories.
Key words: myelodysplastic syndromes, blood transfusion, transfusion costs.

variable clinical presentation [1,2]. Indolent cases of MDS

Background
Myelodysplastic syndromes (MDS) are a heterogeneous
group of chronic haematological disorders with highly
Correspondence: Jingcheng Zhao, Department of Medicine Solna,
Karolinska Institutet, Clinical Epidemiology Division T2, Karolinska
University Hospital, SE-171 76 Stockholm, Sweden
E-mail: Jingcheng.zhao@ki.se
typically present with mild anaemia, often not requiring
treatment at all, whilst fulminant cases present with pan-
cytopenia that require intensive transfusion therapy and
are associated with high mortality [3]. Using the revised
international prognostic scoring system (IPSS-R), patients
with MDS can be divided into categories with vastly
different prognosis, requiring different therapeutic
modalities [4].

This is an open access article under the terms of the Creative Commons Attribution-NonCommercial License,
which permits use, distribution and reproduction in any medium, provided the original work is properly cited and 581
is not used for commercial purposes.
582 J. Zhao et al.
Whilst many patients with MDS respond favourably to
erythrocyte stimulating agents, for example erythropoi-
etin analogs, as well as to hypomethylating agents, or to
the thalidomide analog lenalidomide, regular blood trans-
fusion remains a cornerstone of therapy [5].
We have previously reported that, on average, blood
product costs alone for patients with MDS can exceed
20,000 USD in the first two years after diagnosis [6].
However, prior data did not allow the stratification of
patients into different prognostic categories, did not take
patient mortality into consideration, and cost estimates
only included blood product costs. As patients with MDS
span from requiring little to no transfusion therapy to
intensive transfusion therapy over multiple years, more
detailed characterization of transfusion patterns and costs
is needed for transfusion planning, as well as health eco-
nomic analyses of treatment alternatives.
Therefore, we used nationwide registers in Sweden to
investigate patterns of transfusion use in MDS patients,
stratified by the revised international prognostic scoring
system, IPSS-R. Furthermore, using incident data on
transfusions and laboratory results, we produced compre-
hensive estimates of transfusion costs, including costs of
blood products, disposables, laboratory testing and labour
costs.
Methods and materials
Setting, data sources and record linkages
All patients diagnosed with MDS in Sweden were iden-
tified through the Swedish MDS register, a nationwide
quality-of-care register, which also included details on
date of diagnosis, and disease characteristics, and the
IPSS-R score [7]. Data linkages to other registers and
databases were conducted using national registration
numbers (NRNs) [8]. Data on all administered blood
transfusions were ascertained from Swedish part of the
recently updated Scandinavian Donations and Transfu-
sions (SCANDAT3-S) database [9]. The SCANDAT3-S
database included the date and type of all transfusions
as well as results from blood typing, antibody screening
and antibody identification panels for all patients in
the MDS cohort. Dates for hospital stays were extracted
from the National Patient Register, to differentiate
between inpatient and outpatient transfusions, and
admissions for allogeneic hematopoietic stem cell trans-
plantations were identified using procedure codes
(DR041, DR042, DR044, DR045, DR046 and DR047).
Vital status and any dates of emigration and death
were determined through linkages to national popula-
tion registers.
Vox Sanguinis published by John Wiley & Sons Ltd on behalf of International Society of Blood Transfusion
Study design
Analyses were restricted to patients diagnosed with MDS
between 2008 and 2017. We excluded patients without a
known IPSS-R category, as well as patients who were not
recorded to be living in Sweden at the time of their diag-
nosis. Patients were followed from the date of their MDS
diagnosis until death, emigration or end of follow-up
(December 31, 2017), whichever occurred first. Follow-up
was also terminated upon being admitted to hospital for
hematopoietic stem cell transplantation, using date of
admission as the date of transplantation.
Calculation of costs
Transfusion costs were divided into three groups: direct
material costs for blood products and disposables, direct
labour costs and additional transfusion-related laboratory
costs.
Direct material costs were calculated using data on
incident transfusions based on the type of blood product.
Cost for intravenous infusion sets was added for each day
of transfusion, as they are typically reused for multiple
transfusions for the same patient in the same day. Costs
for both blood products and disposables were determined
from publicly available regional price lists from Stock-
holm, Sweden (Appendix 1).
Direct labour costs were calculated based on a previ-
ously published approach, which was modified to allow
differential costs for inpatient and outpatient transfusions
[10]. Labour costs were calculated as the product of the
median hourly cost to employ a registered nurse, based
on publicly available data from the Stockholm Regional
Council in Sweden, and the approximate time taken to
administer the transfusion. Based on previously published
data and supplemental interviews with head nurses in
haematological inpatient and outpatient clinics at a major
hospital in Stockholm, Sweden, four different administra-
tion times were estimated depending on whether it was
the first transfusion on a given day or the second, or later
unit, and whether they were administered in an inpatient
or outpatient setting. Due to variation and uncertainty in
administration time, a lower and upper range was estab-
lished for each of the four administration scenarios. For
details, see Appendix 2.
Additional transfusion-related laboratory costs were
calculated based on incident data on laboratory tests,
including blood typing and antibody screening (Appen-
dix 3). Positive antibody screening incurred a one-time
cost for an antibody identification panel, and cross-
matching costs for subsequent red cell transfusions. Costs
were determined from public price lists. Currency
© 2020 The Authors.
Vox Sanguinis (2021) 116, 581–590

conversion from SEK to USD was done using mid-2018
conversion rate of 8
998 SEK to 1 USD.
Statistical analysis
For both transfusions and transfusion costs, the number
of persons in the study was calculated per day, to
account for the high mortality of the MDS population.
Patients who emigrated, underwent hematopoietic stem
cell transplantation or died were excluded at the end of
the day. The daily mean number of transfusions was
calculated as the number of transfusions in each day
divided by the number of people alive at the beginning
of the day; these were accrued to form the expected
cumulative number of transfusions. The expected cumu-
lative transfusion cost was calculated in a similar fash-
ion. Both the cumulative number of transfusions and
the ensuing cumulative cost were calculated per day
since diagnosis, stratified by IPSS-R category and age
at diagnosis (<65 years, 65–80 years, >80 years). Costs
were also calculated separately for only red cell trans-
fusions.
For all analyses, 95% confidence intervals were calcu-
lated using bootstrapping, with 10,000 runs. The pre-
sented confidence intervals for transfusion costs were
modified in such a way that the lower range for nurse
administration times was used for the lower bound of the
confidence interval and the upper range for nurse admin-
istration times for the upper bound of the confidence
interval. The figures are censored when there were less
than 10 persons in study.
The proportion of transfusions conducted inpatient,
including confidence intervals, was calculated using
logistic regression, with time since diagnosis modelled as
a restricted cubic spline (with knots at 0, 3, 7, 180, 360,
720, 1460 days).
All data processing and statistical analyses were per-
formed using SAS Statistical Analysis Software, version
9.4 (Cary, NC, USA). The creation of the SCANDAT3-S
database and the conduct of this study was approved by
the regional ethics committee in Stockholm, Sweden and
the Swedish Ethics Review Authority (ref. nr. 2018/167-
31, 2019-02636).
Results
We identified a total of 2858 patients with a diagnosis of
MDS between 2008 and 2017 from the Swedish MDS reg-
ister. After excluding patients who had an unknown date
of diagnosis (N = 1), no recorded IPSS-R score (N = 449),
a diagnosis date after date of death (N = 1), a prior allo-
genic hematopoietic stem cell transplantation (HSCT)
(N = 3), or who were not recorded to be living in Sweden
© 2020 The Authors.
Vox Sanguinis published by John Wiley & Sons Ltd on behalf of International Society of Blood Transfusion
Vox Sanguinis (2021) 116, 581–590
Transfusion costs in patients with MDS 583
at the time of their diagnosis (N = 93), a total of 2311
patients remained for analysis.
Baseline characteristics of the study cohort, stratified
by IPSS-R score, are presented in Table 1. There were 423
(18
3%), 741 (32
1%), 463 (20
0%), 364 (15
8%) and 320
(13
8%) patients in the very low, low, intermediate, high
and very high prognostic categories, respectively. The
proportion of females ranged from 34
8% in the very
low-risk group to 46
3% in the very high-risk group, and
the mean age from 74
4 years (standard deviation [SD],
9
5) in patients with low risk, to 71
0 years (SD, 12
8) in
patients with very high risk. The median length of fol-
low-up was progressively shorter with higher IPSS-R
score, ranging from 2
4 years (interquartile range [IQR],
1
2–4
5) in patients with very low-risk disease, to
0
5 years (IQR, 0
3–1
0) in patients with very high-risk
disease.
The average cumulative number of transfusions strati-
fied by IPSS-R at diagnosis is presented in Table 2. Fur-
ther stratification by age at diagnosis, is presented in
Figure 1. Frequency of red cell and platelet transfusions
increased with higher IPSS-R category, whereas plasma
transfusions were rare in all IPSS-R categories. Amongst
patients in the very low category, those >80 years at
diagnosis received more red cell transfusions but fewer
platelet transfusions compared to those <65 years, on
average. Interestingly, the number of transfusions was
otherwise similar between age groups within the same
IPSS-R prognosis category. In the first 4 years after diag-
nosis, patients in the very low-risk category overall
received 25 red cell (95% confidence interval [CI], 20–32),
and 4 platelet transfusions (95% CI, 3–7), whereas
patients younger than 65 years at diagnosis received on
average 16 red cell (95% CI, 9–25) and 9 platelet (95% CI,
3–16) transfusions. Conversely, patients in the very high-
risk category had on average 171 red cell (95% CI, 135–
200) and 66 platelet transfusions (95% CI, 51–78) 4 years
after their diagnosis. More than half of patients younger
than 65 years of age in the intermediate, high and very
high categories underwent HSCT (53%, 53% and 52%,
respectively), and very few remained in the study after
2 years since diagnosis (22, 5 and 2 patients, respec-
tively). Amongst patients older than 80 years at diagnosis
in the high and very high-risk categories, few remained
alive after two years after diagnosis; however, none
underwent HSCT.
The average cumulative cost of transfusions stratified
by IPSS-R category is presented in Table 3 and further
stratified by age group at diagnosis in Figure 2. Trends in
transfusion costs are congruent with transfusion trends
and increased with increasingly severe IPSS-R category at
diagnosis. Across all age groups, transfusion costs were
highest on average for the very high-risk category,
584 J. Zhao et al.

Table 1 Baseline characteristics of study population

IPSS-R Risk score

Very low risk Low risk Intermediate risk High risk Very high risk

No. subjects (% of total) 423 (18
3) 741 (32
1) 463 (20
0) 364 (15
8) 320 (13
8)
Females, N (%) 147 (34
8) 310 (41
8) 179 (38
7) 149 (40
9) 148 (46
3)
Age at diagnosis, N (%)
<65 years 63 (14
9) 116 (15
7) 93 (20
1) 92 (25
3) 81 (25
3)
65–79 years 232 (54
8) 398 (53
7) 267 (57
7) 188 (51
6) 165 (51
6)
≥80 years 128 (30
3) 227 (30
6) 103 (22
2) 84 (23
1) 74 (23
1)
Mean, years (SD) 74
2 (9
5) 74
4 (9
5) 71
9 (10
6) 70
3 (12
9) 71
0 (12
8)
Length of follow-up, N (%)
<1 year 169 (40
0) 346 (46
7) 319 (68
9) 321 (88
2) 297 (92
8)
1–4 years 164 (38
8) 274 (37
0) 120 (25
9) 39 (10
7) 21 (6
6)
≥5 years 90 (21
3) 121 (16
3) 24 (5
2) 4 (1
1) 2 (0
6)
Median, years (IQR) 2
4 (1
2–4
5) 2
1 (1
1–3
8) 1
2 (0
4–1
4) 0
8 (0
4–1
4) 0
5 (0
3–1
0)
Hematopoietic transplant during follow-up, N (%) 11 (2
6) 41 (5
5) 69 (14
9) 62 (17
0) 53 (16
6)

Table 2 Cumulative average number of transfusions over time, stratified by IPSS-R category at diagnosis.

Duration of follow-up

IPSS-R score 90 days 1 year 2 years 3 years 4 years

Average no. of red cell transfusions (95% confidence limit)

Very low 1 (1–1) 4 (3–5) 10 (8–13) 17 (14–21) 25 (20–32)
Low 3 (2–3) 11 (10–13) 26 (23–29) 41 (36–45) 56 (50–63)
Intermediate 4 (4–5) 20 (18–22) 39 (35–45) 61 (54–70) 85 (73–99)
High 9 (9–11) 35 (32–39) 72 (65–81) 102 (90–117) 136 (111–165)
Very high 13 (12–14) 44 (40–50) 82 (74–93) 135 (119–153) 171 (135–200)
Average no. of platelet transfusions (95% confidence limit)
Very low 0 (0–0) 1 (0–1) 2 (1–3) 3 (2–5) 4 (3–7)
Low 0 (0–1) 2 (1–2) 4 (3–5) 7 (5–10) 10 (7–14)
Intermediate 1 (1–2) 6 (5–7) 12 (10–16) 17 (14–22) 22 (17–28)
High 3 (2–4) 12 (10–15) 24 (20–30) 34 (26–44) 45 (31–63)
Very high 5 (5–6) 20 (17–24) 34 (28–41) 62 (49–77) 66 (51–78)
Average no. of plasma transfusions (95% confidence limit)
Very low 0 (0–0) 0 (0–0) 0 (0–0) 0 (0–0) 0 (0–1)
Low 0 (0–0) 0 (0–0) 0 (0–1) 1 (0–2) 1 (0–2)
Intermediate 0 (0–0) 0 (0–1) 1 (0–1) 1 (1–1) 1 (1–2)
High 1 (0–1) 1 (1–2) 1 (1–2) 2 (1–3) 2 (1–4)
Very high 0 (0–0) 0 (0–1) 0 (0–1) 1 (0–1) 1 (0–1)

amounting to $80 106 (95% CI, $61 460–95 792), and
lowest in the very low-risk category, amounting to $8805
(95% CI, $6482–11 625) after four years. Costs for red
cell transfusion support only is presented in Appendix 4.
The proportion of transfusions given in an inpatient
setting is shown in Figure 3. Trends were similar for all
IPSS-R groups, with a higher proportion of transfusions
given inpatient during the first months after diagnosis
that stabilized to around 30% after about half a year.
Discussion
In this nationwide cohort study of transfusion patterns
and costs for patients with HSCT-na€ıve MDS, average

© 2020 The Authors.

Vox Sanguinis published by John Wiley & Sons Ltd on behalf of International Society of Blood Transfusion
Vox Sanguinis (2021) 116, 581–590
 Transfusion costs in patients with MDS 585

Figure 1 Cumulative mean number of RBC and PLT transfusions over time, stratified by IPSS-R and age at diagnosis. [Colour figure can be viewed at
wileyonlinelibrary.com]

Table 3 Total costs of transfusion therapy during follow-up, stratified by IPSS-R category at diagnosis

Duration of follow-up

90 days 1 year 2 years 3 years 4 years

IPSS-R score Average total cost in USD* (95% confidence limit)

Very low 337 (186–538) 1335 (947–1810) 3600 (2567–4798) 6191 (4451–8201) 8805 (6482–11 625)
Low 864 (671–1088) 3941 (3231–4728) 8937 (7402–10 632) 14 631 (11 882–17 764) 20 184 (16 127–25 098)
Intermediate 1894 (1481–2368) 8341 (6863–9997) 17 072 (13 888–20 631) 25 457 (20 562–30 888) 34 068 (27 334–41 877)
High 4074 (3320–4941) 15 850 (13 252–18 720) 31 964 (26 371–38 486) 45 629 (36 925–56 061) 60 514 (44 936–80 410)
Very high 6121 (5226–7094) 22 560 (18 929–26 616) 39 862 (33 155–47 863) 69 034 (55 711–84 334) 80 106 (61 460–95 792)

*Using mid-2018 SEK to USD exchange rate of 8
9984.

costs for transfusions in the first 4 years after diagnosis As far as we know, this is the largest and most detailed
were found to be as high as $80 106 for patients in the study of transfusion costs in this patient group. Our
very high IPSS-R prognostic category and as low as method for calculating costs expands on previously
$8805 for patients in the very low prognostic category. reported methods, by differentiating between inpatient

© 2020 The Authors.

Vox Sanguinis published by John Wiley & Sons Ltd on behalf of International Society of Blood Transfusion
Vox Sanguinis (2021) 116, 581–590
586 J. Zhao et al.

Figure 2 Cumulative mean cost of transfusions in USD over time, stratified by IPSS-R and age at diagnosis. *Upper and lower confidence interval uses
the upper and lower bound of nurse administration time, respectively.

Figure 3 Proportion of transfusions in an inpatient setting, stratified by IPSS-R.

and outpatient transfusions, as well as incorporating incorporate the high mortality of the MDS patients by
results from antibody screening to include costs for cross- calculating transfusion counts and costs daily. Interest-
matching and antibody panels. Furthermore, we ingly, transfusion incidence and cost were mainly driven

© 2020 The Authors.

Vox Sanguinis published by John Wiley & Sons Ltd on behalf of International Society of Blood Transfusion
Vox Sanguinis (2021) 116, 581–590

by IPSS-R risk category, whereas it was generally similar
across age groups within the same IPSS-R category.
In a previous study, blood product costs for MDS
patients were calculated to exceed 20 000 USD in the first
2 years after diagnosis [6]. However, there are several dif-
ferences that lead to differential results in this study.
Firstly, the previous study included only blood product
costs and did not include costs for laboratory testing,
labour or materials. Secondly, the previous study did not
stratify on disease severity and was an average across dif-
ferent age groups. Thirdly, the previous study calculated
average costs based on the number of patients at baseline
which may lead to a downward bias in costs, particularly
if mortality is high. In this study, we compensated for
mortality by using the number of patients at risk of trans-
fusions, updated daily, as the denominator. Finally, the
previous study used data from 2000 to 2010, whereas this
study uses more contemporary data from 2008 to 2017.
As such, the figures are not directly comparable, and this
study offers a more accurate characterization of transfu-
sion incidence and costs for patients with MDS. Further-
more, whilst we have earlier reported that transfusion
costs were markedly higher in older patients above
65 years at diagnosis, we show here that this is no longer
the case after incorporating patient mortality and disease
severity [6].
The major strengths include the use of incident trans-
fusions, as well as transfusion laboratory tests and results
from a nationwide transfusion database, together with
data on IPSS-R prognosis categories from nationwide
quality registers, to calculate transfusion costs. Further-
more, using national health and population registers, we
were able to identify the date for migration, HSCT and
death. This allowed us to use the average transfusion
count and cost on a daily basis, as weekly or monthly
time-periods would likely underestimate the results due to
the high mortality.
There are several limitations to the study. Firstly, costs
are limited to blood product costs, direct labour costs and
additional transfusion-related laboratory costs. We did
not incorporate societal costs in a wider sense (e.g. costs
for donor transportation and absence from work), or costs
for possible transfusion complications. Costs for relevant
medications, such as erythropoietin analogs and iron
chelation, which may affect or be necessitated by long-
term transfusions therapy, are also not included. Follow-
ing Nordic MDS group guidelines, iron chelation is
References
1 Cancer TIAfRo: WHO Classification of Lymphoid Tissue. Lyon, France, World
Tumours of Haematopoietic and Health Organization, 2008
© 2020 The Authors.
Vox Sanguinis published by John Wiley & Sons Ltd on behalf of International Society of Blood Transfusion
Vox Sanguinis (2021) 116, 581–590
Transfusion costs in patients with MDS 587
typically indicated if ferritin levels are above 1500 lg/L,
or after approximately 25 red cell transfusions [11]. More
indirect costs such as cost of hospital facilities or equip-
ment were not included as they are largely determined by
local circumstance, such as cost of rent, resource utiliza-
tion rates, and alternative costs for resources and assets.
Secondly, the cost algorithms assume that all first-time
positive antibody screens lead to one antibody panel and
cross-matching for subsequent units, disregarding more
complex scenarios with positive direct antiglobulin tests,
or need for additional antibody panels or blood group
genotyping. Thirdly, we do not account for the need of
phenotype-matched blood products, which are typically
at least twice as expensive. Our projected costs therefore
likely underestimate true costs, but the extent of the
underestimation is difficult to assess without more
detailed clinical data and may vary according to local cir-
cumstances.
At the same time, we were not able to differentiate
transfusions necessitated by the disease per se from trans-
fusions driven by other indications such as high-dose
chemotherapy. Whilst we did censor follow-up upon
undergoing transplantation, the available data do not reli-
ably capture other instances of chemotherapy and there-
fore does not allow the same strategy to remove
transfusions due to cytopenia following chemotherapy,
including patients that may have transformed to acute
myeloid leukaemia. We speculate that transfusions given
for such indications would predominantly be seen in the
higher risk categories and that interpretation of blood use
and costs in these groups must consider such effects,
whilst likely having less impact on lower risk categories.
In conclusion, in this nationwide study on costs associ-
ated with blood transfusions in patients with MDS, higher
risk IPSS-R prognosis category at diagnosis was associ-
ated with significant costs. On average in the first
4 years, transfusion costs ranged from $8805 to $80 106.
Acknowledgements
We are greatly indebted to all the blood banks in Sweden
for both collecting and contributing data to this study.
Furthermore, we would like to thank Roland Fiskesund
MD PhD (Region Stockholm) for information on costs
associated with transfusions. JZ and GE performed the
data collection. JZ, TD, and GE designed the study, per-
formed the analyses, and wrote the manuscript.
2 Greenberg PL, Attar E, Bennett JM,
et al.: Myelodysplastic syndromes:
588 J. Zhao et al.
clinical practice guidelines in oncol-
ogy. J Natl Compr Cancer Netw 2013;
11:838–74
3 Haase D, Stevenson KE, Neuberg D,
et al.: TP53 mutation status divides
myelodysplastic syndromes with com-
plex karyotypes into distinct prognostic
subgroups. Leukemia 2019; 33:1747–58
4 Greenberg PL, Tuechler H, Schanz J,
et al.: Revised international prognostic
scoring system for myelodysplastic
syndromes. Blood 2012; 120:2454–65
5 Malcovati L, Hellstr€ om-Lindberg E,
Bowen D, et al.: Diagnosis and treat-
ment of primary myelodysplastic syn-
dromes in adults: recommendations
from the European LeukemiaNet. Blood
2013; 122:2943–64
Appendix 1 Cost modelling inputs
Item
Blood products and disposables
Red cells
Platelets
Plasma
IV administration sets
Labour
Nurse median cost per hour**
Laboratory
Blood typing and antibody screening
Antibody identification panel
Cross test per unit
Activity
Inpatient setting
First administered unit
Additional units
Outpatient setting
First administered unit
Additional units
NOTES
*Using mid-2018 SEK to USD exchange rate of 8
9984.
**Sum of salary employer and employers’ social security contribution.
Vox Sanguinis published by John Wiley & Sons Ltd on behalf of International Society of Blood Transfusion
6 Zhao J, Ryd
en J, Wikman A, et al.:
Blood use in hematologic malignan-
cies: a nationwide overview in Sweden
between 2000 and 2010. Transfusion
2018; 58:390–401
7 Moreno Berggren D, Folkvaljon Y,
Engvall M, et al.: Prognostic scoring
systems for myelodysplastic syn-
dromes (MDS) in a population-based
setting: a report from the Swedish
MDS register. Br J Haematol 2018;
181:614–27
8 Ludvigsson JF, Otterblad-Olausson P,
Pettersson BU, et al.: The Swedish per-
sonal identity number: possibilities
and pitfalls in healthcare and medical
research. Eur J Epidemiol 2009;
24:659–67
9 Zhao J, Rostgaard K, Hjalgrim H,
et al.: The Swedish Scandinavian
donations and transfusions database
(SCANDAT3-S) – 50 years of donor
and recipient follow-up. Transfusion;
n/a.https://onlinelibrary.wiley.com/d
oi/full/10.1111/trf.16027.
10 Stokes EA, Wordsworth S, Staves J,
et al.: Accurate costs of blood transfu-
sion: a microcosting of administering
blood products in the United Kingdom
National Health Service. Transfusion
2018; 58:846–53
11 Kittang AO, Cavelier L, Dybedal I,
et al.:Guidelines for the diagnosis and
treatment of Myelodysplastic Syn-
drome and Chronic Myelomonocytic
Leukemia. Nordic MDS Group, 2019.
Costs (SEK/USD)*
1523 SEK/$169
4751 SEK/$528
846 SEK/$94
100 SEK/$11
254 SEK/$28
390 SEK/$43
803 SEK/$89
282 SEK/$31
Time (lower-upper limits)
60 min (30–90 min)
24 min (18–30 min)
90 min (60–120 min)
60 min (30–90 min)
© 2020 The Authors.
Vox Sanguinis (2021) 116, 581–590
 Transfusion costs in patients with MDS 589

Appendix 2 Algorithm for direct labour costs

Appendix 3 Algorithm for transfusion laboratory costs.

© 2020 The Authors.

Vox Sanguinis published by John Wiley & Sons Ltd on behalf of International Society of Blood Transfusion
Vox Sanguinis (2021) 116, 581–590
590 J. Zhao et al.

Appendix 4 Total costs of red cell transfusion therapy during follow-up, stratified by IPSS-R
category at diagnosis.
Duration of follow-up

90 days 1 year 2 years 3 years 4 years

IPSS-R score Average total cost in USD* (95% confidence limit)

Very low 200 (186–215) 1020 (944–1096) 2597 (2402–2792) 4376 (4049–4704) 6413 (5934–6891)
Low 652 (604–700) 2924 (2708–3141) 6584 (6095–7074) 10 455 (9676–11 233) 14 401 (13 328–15 474)
Intermediate 1126 (1045–1207) 4980 (4615–5345) 9948 (9211–10 685) 15 508 (14 352–16 663) 21 515 (19 903–23 128)
High 2340 (2170–2511) 8884 (8235–9533) 18 351 (17 004–19 698) 26 055 (24 131–27 980) 34 570 (32 013–37 127)
Very high 3116 (2892–3340) 11 010 (10 212–11 808) 20 683 (19 172–22 194) 33 909 (31 423–36 395) 42 958 (39 794–46 122)

NOTES
*Using mid-2018 SEK to USD exchange rate of 8
9984.

© 2020 The Authors.

Vox Sanguinis published by John Wiley & Sons Ltd on behalf of International Society of Blood Transfusion
Vox Sanguinis (2021) 116, 581–590
 Vox Sanguinis (2021) 116, 591–600
© 2020 International Society of Blood Transfusion
ORIGINAL PAPER DOI: 10.1111/vox.13042

Clinical consequences of the extremely rare anti-PP1Pk

isoantibodies in pregnancy: a case series and review of the
literature
Pietro Di Ciaccio,1 Briony Cutts,2 Thushari Indika Alahakoon,3 Peta M. Dennington,4 Luke A. Soo4 & Jennifer Curnow1
1
Department of Haematology, Westmead Hospital, Sydney, NSW, Australia
2
Department of Haematology, The Royal Women’s Hospital, Melbourne, VIC, Australia
3
Westmead Institute for Maternal Fetal Medicine, Westmead Hospital, Sydney, NSW, Australia
4
Australian Red Cross Lifeblood, Sydney, NSW, Australia

Received: 28 September 2020,
revised 12 November 2020,
accepted 15 November 2020,
published online 16 December 2020
Background The absence of the red cell antigens P, P1 and Pk, known as ‘p’, rep-
resents an extremely rare red cell phenotype. Individuals with this phenotype
spontaneously form anti-PP1Pk isoantibodies, associated with severe haemolytic
transfusion reactions, recurrent spontaneous abortion and haemolytic disease of
the fetus and newborn (HDFN).
Methods We report a series of four successful pregnancies in three women with
anti-PP1Pk isoantibodies, one complicated by HDFN, another by intrauterine
growth restriction, all managed supportively. We also review the literature
regarding the management of pregnancy involving anti-PP1Pk isoimmunization.
Results The literature surrounding anti-PP1Pk in pregnancy is limited to a very
small number of case reports. The majority report management with therapeutic
plasma exchange (TPE) with or without intravenous immunoglobulin. The rela-
tionship between titre and risk of pregnancy loss remains unclear, though a his-
tory of recurrent pregnancy loss appears important. Although a positive cord
blood direct antiglobulin test is frequently noted, clinically significant HDFN
appears uncommon, though possible.
Conclusion Early initiation of TPE in high risk patients should be strongly con-
sidered. If possible, pregnancies should be managed in a high-risk obstetric or
maternal fetal medicine service. The fetus should be monitored closely with inter-
val fetal ultrasound and middle cerebral artery peak systolic volume Doppler to
screen for fetal anaemia. Timely sourcing of compatible blood products is likely
to be highly challenging, and both directed and autologous donation should be
contemplated where appropriate. The International Red Cell Donor Panel may
also provide access to compatible products.
Key words: haemolytic disease of the fetus and newborn, immunohaematology,
patient blood management, RBC antigens and antibodies, transfusion strategy.

cells of both the P and Pk surface antigens is virtually ubiq-

Introduction
The red cell antigens P, P1 and Pk are closely related gly-
cosphingolipid structures, and the presence on red blood
Correspondence: Pietro Di Ciaccio, Department of Haematology, Level 5,
The Kinghorn Cancer Centre, Darlinghurst, New South Wales, Australia
E-mail: Pietro.diciaccio@health.nsw.gov.au
uitous in human populations [1]. Their absence, together
with absence of P1, is known as the extremely rare ‘p’ phe-
notype. The European prevalence of this phenotype is
approximately 5
8 per million, with marginally higher inci-
dences in Japan, Israel and Amish populations [2–4].
Individuals with the p phenotype spontaneously form
anti-PP1Pk antibodies (formerly known as anti-Tja), which

591
592 P. Di Ciaccio et al.
are a mixture of IgG and IgM isoantibodies against P1, P
and Pk. This occurs as early as infancy [5,6]. Although
anti-P1 is generally a weak cold antibody without clinical
significance, anti-P and anti-Pk are potent haemolysins
associated with severe haemolytic transfusion reactions
[7].
Anti-PP1Pk antibodies are also associated with recur-
rent spontaneous abortion, particularly in the first trime-
ster [3,8–10]. P and Pk in particular are expressed at high
levels on the placenta, and antibodies to both of these
antigens are deleterious to placental trophoblasts. It is
likely that both the IgG and IgM components are patho-
logical [3,9,11–15]. Although occurring only in a minor-
ity of cases, haemolytic disease of the fetus and newborn
(HDFN) has also been reported [16,17].
The synthesis of both P1 and Pk is dependent upon the
action of alpha 1-4-galactosyltransferase, encoded by the
A4GALT gene which catalyses the addition of galactose
to lactosylceramide glycosphingolipid. The P antigen
belongs to the separate globoside (GLOB) family,
Fig. 1 Formation of the P1, Pk and P antigens by sequential additions of monosaccharides, with catalysing enzymes. Simplified from Kazcmarek et al.
(2014) [42]. [Colour figure can be viewed at wileyonlinelibrary.com]
dependent on the transferase encoded by the B3GALNT1
gene (Fig. 1). The p phenotype arises in the context of
homozygous mutation of the A4GALT gene, which apart
from coding for P1 and Pk, is also involved in upstream
construction of the P antigen (Fig. 1). Over 30 different
polymorphisms in A4GALT resulting in the p phenotype
have been reported, the majority being single nucleotide
variants or small indels causing a variety of missense or
nonsense mutations [18]. Defects in B3GALNT1 cause the
P1k and P2k phenotypes, which are even rarer than p
(Table 1) [1,19–22].
We present a series of four successful pregnancies in
three patients whose management at our institutions was
complicated by the presence of anti-PP1Pk isoantibodies.
Case 1
A 22-year-old woman of Indian descent (Tamil Nadu
region) was referred at five weeks of gestation due to the
presence of anti-PP1Pk antibodies, detected during
© 2020 International Society of Blood Transfusion
Vox Sanguinis (2021) 116, 591–600

Table 1 Phenotypes and corresponding antibodies. Adapted from Hell-
berg et al. 2013 [1]
Antigens Present
Phenotype Prevalence on red cells Isoantibodies
P1 20–90% P1, Pk, P - during week 39 of gestation, the patient entered sponta-
P2 10–80% Pk, P Anti-P1
p null Rare - Anti-PP1Pk
P1k Rare P1, Pk Anti-P
P2k Rare Pk Anti-P, P1
routine antenatal screening. This was her first known
pregnancy, and there was no history of blood product
transfusion. Her red cell phenotype was A1, R1R1, K-k+,
Fy(a + b-), Jk(a-b+), MSs, p, Le(a-b-). The paternal pheno-
type was P1+, and the couple were first cousins. There
was no family history of the p phenotype, transfusion
reactions or miscarriages.
Anti-PP1Pk antibodies were detected by room tempera-
ture saline indirect antiglobulin testing (IAT), polyethy-
lene glycol (PEG) IAT and papain IAT. The initial
anti-PP1Pk titre was 1:8 by saline IAT tested against a
pool of PP1Pk cells at the Australian Red Cross Lifeblood
(Lifeblood) Red Cell Reference Laboratory. All future titra-
tions were performed centrally at this laboratory and
tested in parallel with the previous sample with titres
peaking at 1:64. Anti-K could not be definitively
excluded due to the lack of K + p reference cells.
Genotyping revealed homozygosity for the
NM_017436.7:c.752C> T point mutation in the A4GALT
gene. This mutation results in the missense mutation
p.Pro251Leu. This is a previously reported mutation in
A4GALT, known to result in the p phenotype in the
homozygous state [23].
Given the presence of anti-PP1Pk, the decision was
made to observe the pregnancy in a high-risk maternal
fetal medicine setting, with close monitoring every
4 weeks of the antibody titre, as well as fetal well-being
by serial ultrasound, including middle cerebral artery
peak systolic velocity (MCA PSV), which was suggestive
of mild fetal anaemia (MCA Vmax 62
6 cm/s). The patient
had a normal full blood count. Vitamin B12 and folate
stores were replete. There was a mild iron deficiency (fer-
ritin 14 µg/l; normal range 20–220 µg/l), corrected with
oral supplementation.
An exhaustive search for potential p blood donors was
undertaken. There were no compatible donors in Australia
or New Zealand. Compatible donors from the patient’s
immediate family could not be identified. Autologous col-
lection was not offered due to safety concerns during
pregnancy. An international search via the International
Rare Donor Panel (IRDP) located two potential p donors,
© 2020 International Society of Blood Transfusion
Vox Sanguinis (2021) 116, 591–600
Anti-PP1Pk isoantibodies in pregnancy 593
and four cryopreserved compatible red cell units in Japan.
Arrangements were made for the collection and transfer
of fresh red cells from these two donors prior to the
anticipated time of delivery.
Although elective induction of labour was planned
neous labour at 38 + 0 weeks, delivering by emergency
caesarean section (CS) for fetal compromise demonstrated
by tachycardia on cardiotocography. The estimated blood
loss was 300 ml, and maternal transfusion was not
required. The newborn had Apgar scores of nine at 1 min
and nine at 5 min, and weighed 3140 g (within normal
range) [24]. The labour and delivery occurred prior to the
arrival of the p red cells in Australia.
The newborn’s haemoglobin was initially normal at
165 g/l. The cord blood group was B positive and the
direct antiglobulin test (DAT) with monospecific antihu-
man globin reagents was positive for IgG and negative
for complement, using column agglutination. A panag-
glutinin was eluted, reacting against all cells, including
group O cells. This was consistent with anti-PP1Pk,
though a concurrent anti-B could not be definitively
excluded due to insufficient sample for further testing.
The newborn was moderately jaundiced with a serum
bilirubin (SBR) of 74 µmol/l at 6 h of age, initially man-
aged with phototherapy. A fall in haemoglobin (133 g/L;
normal range 142–240 g/l) and rise in SBR to a peak of
119 µmol/l was consistent with evolving haemolysis and
managed with IVIg 1 g/kg and ongoing phototherapy.
Both mother and newborn were discharged six days after
birth and remained well six months later.
Case 2
A 31-year-old woman of Burmese descent presented at
6 weeks of gestation in the context of known anti-PP1Pk
antibodies. She had two previous pregnancies. The first
pregnancy ended in miscarriage early in the first trimester
with anti-PP1Pk first detected at the time of fetal loss. For
the second pregnancy, thrice weekly therapeutic plasma
exchange (TPE), in conjunction with intravenous
immunoglobulin (IVIg), was administered; however, the
pregnancy miscarried at eight weeks. At the time of fetal
demise, the titre was 1:4. The maternal phenotype was B,
R1R2, K-k+, Kp(a-b+), Fy(a + b-), Jk(a + b-), Mss, p, Le(a-
b+), Lu(a-). The paternal phenotype was P1+, with no
consanguinity.
The initial titre of anti-PP1Pk in this third pregnancy
was 1:2, and only trace after treatment with dithiothreitol
(DTT), suggesting a mix of IgG and IgM. No other anti-
bodies were detected after alloadsorption, by either saline
IAT or PEG IAT. Maternal full blood count and haema-
tinic indices were normal.
594 P. Di Ciaccio et al.
The pregnancy was managed expectantly, with serial
antibody titres, as well as serial fetal ultrasound, includ-
ing MCA PSV, which remained within normal limits. The
patient was offered TPE and IVIg, however declined. She
was prescribed aspirin 100 mg daily until week 36 for
placental protection. The anti-PP1Pk titre showed minimal
fluctuation, peaking at 1:8 at week 24, falling back to 1:2
at week 36.
No compatible donors or cryopreserved units in Aus-
tralia or New Zealand were identified, nor were there any
eligible directed family donors. A search via the IRDP
located two potential p donors in Japan. Arrangements
were made for the collection and importation of fresh red
cells from these two donors prior to the anticipated date
of delivery.
The patient underwent elective CS at 37 + 6 weeks for
a breech presentation. There was a mild postpartum
haemorrhage, with 500 ml estimated blood loss. The new-
born weighed 2668 g (normal for gestational age), with
Apgar scores of eight at 1 min and nine at 5 min [24].
Cord blood showed haemoglobin of 162 g/L, blood group
B positive and a negative DAT. There was no neonatal
jaundice. Both mother and newborn were discharged well
after routine postoperative care.
Nine months following this first successful pregnancy,
the patient re-presented for antenatal care with her fourth
pregnancy at 16 weeks of gestation. Anti-PP1Pk was
detected by saline IAT at 22°C and 37°C and tube low-
ionic IAT at her initial attendance, but its titre was not
able to be estimated due to insufficient blood sample vol-
ume. No additional alloantibodies were detected by saline
or low-ionic IAT following alloadsorption.
The patient attended several different institutions for her
ongoing antenatal care. Once again TPE was offered and
declined by the patient. She was prescribed aspirin 100 mg
daily until week 36. The anti-PP1Pk titre was 1:2 after treat-
ment with DTT at 25 weeks and was unchanged at 33 weeks.
Fetal development was normal throughout the pregnancy.
As was the case with the patient’s previous pregnancy,
no compatible donors in Australia or New Zealand were
identified. However, the two red cell units originally
imported fresh from p donors in Japan and frozen upon
arrival in Australia prior to the delivery date for her pre-
vious pregnancy, remained available in the cryopreserved
rare blood inventory at Lifeblood.
The patient was scheduled for a repeat elective CS in
week 38 of gestation for breech presentation; however,
she entered spontaneous labour and underwent emer-
gency CS at 35 + 0 weeks for breech presentation and
placental abruption. Her preoperative haemoglobin was
124 g/l, and perioperative blood loss was minimal. The
newborn weighed 2740 g (normal for gestational age)
[24]. Apgar scores were nine at 1 min and nine at 5 min.
The DAT was negative. There was mild jaundice with the
SBR rising to 220 µmol/l; however, treatment with pho-
totherapy was not required.
Case 3
A 28 year-old-lady of Burmese descent, the sister of Case
2, presented at 12 weeks of gestation. This was her first
pregnancy, and anti-PP1Pk antibodies were detected on
routine antenatal screening. The maternal red cell pheno-
type was O, R1R1, K-k+, Fy(a+,b-), Jk(a + b+), p, MSs.
The paternal phenotype was P1+, with no consanguinity.
The initial titre of anti-PP1Pk was 1:4. No other anti-
bodies were detected by saline IAT or PEG IAT after
alloadsorption, though definitive exclusion of anti-K was
not possible due to the absence of K + p reference cells.
The antibody titre was monitored throughout the preg-
nancy, remaining at 1:4.
The patient was offered TPE and IVIg, however
declined. No compatible donors or cryopreserved units in
Australia or New Zealand were available, nor were there
any suitable family donors. An international search iden-
tified a potential donor in the United States; however,
donation could not be arranged prior to delivery.
Although MCA PSVs remained within normal limits
throughout pregnancy and there was no evidence of pre-
eclampsia or infection, the fetus demonstrated intrauter-
ine growth restriction (IUGR) and was consequently
induced at 36 + 6 weeks and delivered by vaginal ven-
touse delivery. There was no antenatal or postnatal haem-
orrhage. The newborn was small for gestational age at
2248 g and demonstrated Apgar scores of nine at 1 min
and nine at 5 min [24]. Cord blood showed a normal hae-
moglobin of 156 g/l, blood group B positive and a nega-
tive DAT, suggesting placental toxicity rather than
haemolysis as the cause of IUGR. The newborn developed
mild jaundice and was treated successfully with pho-
totherapy alone.
Review of the literature
A review of the literature was undertaken to evaluate the
available evidence regarding the impact of anti-PP1Pk
antibodies in pregnancy. A keyword search of the MED-
LINE database was undertaken, limiting to English lan-
guage. Further, we interrogated the references of these
articles to identify further additional relevant articles.
The identified literature was limited to case reports and
one case series of two patients (Table 2). Cases were
included if authors reported a case of the management of
a pregnancy with anti-PP1Pk, or a case of presentation
with fetal demise. Overall, 19 pregnancies complicated by
anti-PP1Pk, in 16 different women, were identified.
© 2020 International Society of Blood Transfusion
Vox Sanguinis (2021) 116, 591–600
 Table 2 Summary of published literature regarding anti-PP1Pk isoantibodies in pregnancy

Miscarriage Maternal
Reference Year Maternal Age Ethnicity history Phenotype Titre Management Fetal Outcome Complications

Rock [22] 1985 28 Kuwaiti 13 (1st P1k 1:32 (1:2 to 1:4 on TPE Delivered 33 weeks
Shirey [9] trimester) TPE) BW 1930 g
Positive DAT Jaundice
(phototherapy)
Hanafusa [32] 2006 36 Japanese 2 (2nd p 1:256 (<1:16 on TPE (double Term TPE catheter
trimester) TPE) filtration TPE) BW 2946 g infectionOccluded

Vox Sanguinis (2021) 116, 591–600

Negative DAT arteriovenous
fistula
Taniguchi [43] 2003 31 Japanese 1 (1st p 1:512 (1:32 to TPE (double Term Pre-eclampsia
trimester) 1:128 on TPE) filtration TPE) BW 1978 g

© 2020 International Society of Blood Transfusion

Negative DAT
Placental microinfarctions
Schecter [10] 1987 36 (two 7 (1st p TPE catheter-
consecutive trimester) (1) (1) (1) related
pregnancies) Not reported TPE Miscarriage 13 weeks bacteraemia
(2) (2) (2)
1:256 (1:32 on TPE, TPE. Withheld in Delivered 36 weeks BW 2500 g
rebound to 1:128 third trimester due Positive DAT (anti-PP1Pk
after TPE cessation) to recurrent Sta- eluted) Jaundice (phototherapy)
phylococcal bacter-
aemia.
Strowitzki [27] 1991 Turkish p TPEIVIg Delivered 31 weeks
Fetal distress
Weiss [25] 1975 39 Moroccan 5 (1st and 2nd p 1:64 Presented with pregnancy loss
trimesters)
Yoshida [4] 1994 35 (two Japanese 6 (1st and 2nd p
consecutive trimesters) (1) (1) (1)
pregnancies) 1:128 (1:32 on TPE Miscarriage 13 weeks
TPE) (2) (2)
(2) TPE with intermit- Delivered 34 weeks
1:128 (1:4 to 1:8 tent flow cell sepa- (3)
on TPE) rator BW 2554 g Jaundice (pho-
totherapy)
Heim [29] 1990 19 2 (1st p TPE IVIg
trimester)
Anti-PP1Pk isoantibodies in pregnancy 595

(Continues)
 Table 2 (Continued)

Miscarriage Maternal
Reference Year Maternal Age Ethnicity history Phenotype Titre Management Fetal Outcome Complications

Levene [17] 1977 32 Israeli 2 (1st p Supportive Term Severe postpartum

trimester) BW 2500 g Positive DAT (weak) haemorrhage
Anti-PP1pk eluted Severe
596 P. Di Ciaccio et al.

anaemia (87 g/l) and severe

jaundice (exchange
transfusion)
Fernandez-Jimenez [30] 2001 28 7 (1st p 1:256 (1:8 on TPE) TPE IVIg Term BW 3000 g Positive DAT
trimester) Anti-PP1Pk eluted Not
anaemic IVIg given
Mohd Azri [31] 2014 27 Indian- Malay 1 (1st trimester) p Dydrogesterone Term BW 2860 g
Benidt [14] 2010 22 Amish p 1:16 Supportive Delivered 34 weeks Positive
DAT Anti-PP1Pk eluted
Benidt [14] 2010 22 Amish p 1:16 Supportive Term Negative DAT
Haentjens-Verbeke [16] 1993 19 Unknown p TPE CS 36 weeks due to anaemia
Haentjens-Verbeke [26] 1996 20 (two p Plasma infusion CS in prematurity due to IUGR
pregnancies) and fetal distress (both
pregnancies
Conlan [28] 2004 25 Chinese 3 in 1st and p 1:64 (1:16 on TPE) TPE Term Postpartum
2nd trimester Positive DAT haemorrhage
(titres 1:128 Anti-PP1pk eluted Jaundice
to 1:256) (phototherapy)
This series 2020 22 Tamil Indian p 1:8 (peak 1:64) Supportive Term
BW 3140 g Positive DAT with
panagglutinin
Mild-moderate HDFN
(phototherapy and IVIg)
31 (first live Burmese 2 (1st p 1:2 Supportive Aspirin Term BW 2668 g Negative DAT
birth) trimester)
(second live 1:2 Supportive Aspirin Term Emergency CS for
birth) placental abruption
BW 2740 g Negative DAT
28 Burmese p 1:4 Supportive Term
Low BW 2248 g
Fetal distress and IUGR
Negative DAT

BW, birth weight; CS, caesarean section; DAT, direct antiglobulin test; HDFN, haemolytic disease of the fetus and newborn; IVIg, intravenous immunoglobulin; IUGR, intrauterine growth restriction; TPE, thera-
peutic plasma exchange.

Vox Sanguinis (2021) 116, 591–600

© 2020 International Society of Blood Transfusion

Reports of women or families simply with history of
early-term miscarriage and anti-PP1Pk were not included.
One case concerned anti-P in a P1k woman.
Ten of the 16 women had a history of recurrent mis-
carriage (defined as two or greater), typically in the first
trimester, though occasionally in the second. This history
was unreported in three women. One patient had experi-
enced 13 consecutive first-trimester miscarriages [22].
Three fetal losses in total were reported amongst the
reviewed cases. One presented for the first time with a mis-
carriage with a titre of 1:64, another patient miscarried
despite TPE from week nine and a third miscarried despite
TPE commenced at week five [4,10,25]. Of note, either fetal
distress, anaemia or intrauterine growth restriction (IUGR)
was noted in four pregnancies that did not end in demise
[16,26,27]. A positive neonatal DAT was noted in five preg-
nancies, two of which occurred in jaundiced but not anae-
mic neonates. Two cases of severe HDFN were noted [17].
One woman suffered severe postpartum haemorrhage due
to placenta accreta, and required transfusion of autologous
and directed red cells [28].
There was high variability in whether titres were
reported before or after treatment with DTT, or whether
this was performed at all. Given this, as well as the fact
that both IgG and IgM isotypes are likely pathological,
titres before DTT treatment, where stipulated by the
authors, are referred to [14].
The peak titre, which was at the time of diagnosis in
all cases, ranged between 1:16 and 1:256, with a median
titre of 1:128. There was no significant correlation
between miscarriage and titre, though not all cases
reported titres. For the two fetal losses where a titre was
reported, they were 1:64 and 1:128, respectively.
A variety of management strategies were employed,
most consistently TPE, which was administered in 12 of
19 pregnancies. Most commonly, the replacement fluid
was human albumin, with approximately one total
plasma volume (TPV) exchanged in each session. Consis-
tently, TPE was started early in the first trimester, com-
monly the fifth or sixth week of gestation, at a frequency
between once to five times per week.
Intravenous immunoglobulin was given in three preg-
nancies, in all cases concurrently with TPE [27,29,30]. One
pregnancy was managed with oral dydrogesterone, due to
the resource-poor setting, with delivery of a normal infant
[31]. One patient was managed with random plasma infu-
sions, presumably in an effort to adsorb anti-PP1Pk; how-
ever, both her pregnancies were complicated by IUGR and
fetal distress, requiring emergent CS in prematurity [26].
Two pregnancies in two members of an Amish family
were managed conservatively, both resulting in term
deliveries of a normal infant. One infant had a positive
DAT and the other did not, though there were no signs of
© 2020 International Society of Blood Transfusion
Vox Sanguinis (2021) 116, 591–600
Anti-PP1Pk isoantibodies in pregnancy 597
HDFN in either. The anti-PP1Pk titre in both of these
pregnancies was 1:16 at maximum [14].
Discussion
Anti-PP1Pk isoimmunization in pregnancy is very rare,
and the associated published literature very limited. It is
important that a multidisciplinary approach is provided
to such patients, ideally within a specialist maternal fetal
medicine service.
Even though a positive DAT is frequently noted on
newborn cord blood, clinically significant HDFN appears
uncommon, with only two definite cases reported, in
addition to Case 1 in our series [16,17]. Nevertheless, it is
wise to monitor for fetal anaemia with serial MCA PSV.
There is, however, a clear link between anti-PP1Pk and
recurrent miscarriage, due to direct placental toxicity [12,32].
The cornerstone of management appears to be TPE. Where
there is a history of recurrent first-trimester miscarriage, the
literature suggests that early initiation of TPE between weeks
five and eight is important, with continuation at least into
the third trimester, and possibly up to delivery [30]. Aggres-
sive regimes, typically three times a week, have been most
commonly employed. It may be possible to taper the therapy
if antibody titres are low and well-controlled.
The American Society for Apheresis (ASFA) supports
the use of TPE for alloimmunization in pregnancy,
though does note that decision-making should be highly
individualized based on the low-quality of evidence.
ASFA Guidelines suggest commencement early in gesta-
tion, exchanging one to 1
5 TPVs [33].
It should be noted, however, that TPE is highly
resource intensive, may not be available in all settings,
and is associated with certain risks, including catheter
infection, hypotension and hypocalcaemia [34]. Our series
of patients demonstrates four successful pregnancies
without TPE; however, one newborn had mild-to-moder-
ate HDFN, and another IUGR and fetal distress.
It is likely relevant that our patients presented with low
titres of anti-PP1Pk relative to those reported in the litera-
ture. Of the reported cases involving the use of TPE, base-
line titres ranged from 1:32 to 1:512. The majority resulted
in viable births where the titre was maintained at <1:32. It
is difficult, if not impossible, to be confident that there is a
‘safe’ titre. Notably, Case 2 in our series had previously
miscarried despite TPE and IVIg, with a titre of 1:4 at that
time. Conversely, the titre in Case 1, a successful pregnancy
without specific intervention, incremented as high as 1:64
during the pregnancy. Conceivably, a history of early mis-
carriage is more relevant in identifying the women who
will benefit most from TPE. In our series, cases 1 and 3 had
no history of pregnancy loss, and Case 2 had only two
early-term miscarriages, the second despite TPE therapy.
598 P. Di Ciaccio et al.

The role of IVIg is poorly defined in anti-PP1Pk-associ-
ated pregnancies, and its antenatal use has only been
reported in conjunction with TPE. IVIg has been used for
several decades in cases of severe alloimmunization in
pregnancy [35]. Proposed mechanisms include inhibition
of endogenous IgG production, saturation of Fc receptors
on effector cells and placental tissue, and increased cata-
bolism of endogenous IgG [36–38]. Its use in the context
of anti-PP1Pk should be individualized.
Another challenge is the identification and supply of
compatible red cells for transfusion. Management of a
pregnant patient with a rare red cell phenotype requires
careful assessment of the risks of antenatal, perinatal and
postpartum haemorrhage, the anticipated need for trans-
fusion and availability of compatible red cell components.
Ideally this information should be documented in a trans-
fusion plan communicated to all members of the treating
team. If the neonate is considered at risk of HDFN, plans
to access compatible red cells or IVIg may be required.
If available, fresh or thawed deglycerolized phenotype-
matched red cells can be used. If fresh phenotype-matched
red cells are not available, management of potential massive
postpartum haemorrhage should include consideration of
both transfusion and obstetric issues. Confirmation of a
planned delivery time during working hours is sensible so
that specialist services are readily available. Patient blood
management strategies during pregnancy including iron
repletion should be encouraged [39]. If no blood is available,
patients should be managed as per protocols developed for
patients who decline transfusion [40].
Autologous collection during pregnancy is not rou-
tinely available in Australia but may be available in cen-
tres elsewhere. Typing of first-degree relatives should be
considered; phenotype-matched siblings should be con-
sidered for directed donations and future autologous and/
or allogeneic donation. In our series, directed donation
was precluded in cases 1 and 3 by ABO incompatibility
and in Case 2 by low donor body weight (42 kg).
Rare red cell units can also be accessed by the Interna-
tional Rare Donor Panel (IRDP) [41]. Lifeblood currently
has one group O positive donor on its panel who is p.
Both fresh or thawed deglycerolized red cells may be
accessed from overseas. However, obtaining frozen red
cells is often not feasible due to difficulties ensuring the
cold chain during long-distance transport, as well as
regional variability in packs used for storage and
consequent incompatibility with local thawing equipment.
Postpartum, patients with rare red cell phenotypes should
be encouraged to become blood donors.
Autologous donation and cryopreservation postnatally,
in anticipation of future possible pregnancies, should be
strongly considered. Family members should be screened
as potential directed donors, and the paternal phenotype
defined.
Limitations of the literature may include a publication
bias of cases where interventions such as TPE were suc-
cessfully utilized, and of women with more extensive his-
tories of miscarriage. Another difficulty in interpreting
the literature is inconsistency in whether and how anti-
body titres are reported.
Conclusion
We present a series of three patients with the very rare p
phenotype, complicated by anti-PP1Pk alloimmunization
in pregnancy. We describe a total of four successful
pregnancies in these patients with a conservative
approach to management, which may be possible in
cases with very low antibody titre. Treatment should be
individualized, however, and strong consideration be
given to early initiation of TPE, especially in women
with a previous history of miscarriage. Management
within a high risk fetal maternal medicine unit is recom-
mended, where possible, and monitoring for IUGR and
fetal anaemia is essential.
Acknowledgements
Australian Federal and State Governments fund Aus-
tralian Red Cross Lifeblood to provide blood products and
services to the Australian community. The authors would
like to thank the staff of the Lifeblood Red Cell Reference
Laboratories for their valuable assistance with the testing
of these samples.
Funding
There are no sources of specific funding for this research.
Conflict of interests
The authors declare no conflict of interests.

References
1 Hellberg A, Westman JS, Thuresson B,
et al.: P1PK: the blood group system
that changed its name and ex-
panded. Immunohematology 2013;
29:25–33
2 Race RR, Sanger R: Blood Groups in
Man. 6th edn. Oxford, Philadelphia,
PA: Blackwell Scientific Publica-
tions; distributed by J. B. Lippincott;
1975
3 Cantin G, Lyonnais J: Anti-PP1Pk and
early abortion. Transfusion 1983;
23:350–351
4 Yoshida H, Ito K, Kusakari T, et al.:
Removal of maternal antibodies from

© 2020 International Society of Blood Transfusion

Vox Sanguinis (2021) 116, 591–600

a woman with repeated fetal loss due
to P blood group incompatibility.
Transfusion 1994; 34:702–705
5 McNeil C, Trentelman EF, Thomas M:
Anti-Tja-another family example. Vox
Sang 1957; 2:114–116
6 Reeves HM, Cary V, Mino MA, et al.:
Unexpected non-maternally derived
anti-PP1P(k) in an 11-week-old
patient. J Pediatr 2017; 181:302–305
7 Poole J, Daniels G: Blood group anti-
bodies and their significance in trans-
fusion medicine. Transfus Med Rev
2007; 21:58–71
8 Levine P, Koch EA: The rare human
isoagglutinin anti-Tja and habitual
abortion. Science 1954; 120:239–241
9 Shirey RS, Ness PM, Kickler TS, et al.:
The association of anti-P and early abor-
tion. Transfusion 1987; 27:189–191
10 Shechter Y, Timor-Tritsch IE, Lewit N,
et al.: Early treatment by plasma-
pheresis in a woman with multiple
abortions and the rare blood group p.
Vox Sang 1987; 53:135–138
11 A4GALT alpha 1,4-galactosyltrans-
ferase (P blood group). Gene. https://
www.ncbi.nlm.nih.gov/gene/53947
[Last accessed 28 April 2020]
12 Lindstrom K, Von Dem Borne AE,
Breimer ME, et al.: Glycosphingolipid
expression in spontaneously aborted
fetuses and placenta from blood group
p women. Evidence for placenta being
the primary target for anti-Tja-anti-
bodies. Glycoconj J 1992; 9:325–329
13 Kojima Y, Fukumoto S, Furukawa K,
et al.: Molecular cloning of globo-
triaosylceramide/CD77 synthase, a gly-
cosyltransferase that initiates the
synthesis of globo series glycosphin-
golipids. J Biol Chem 2000; 275:
15152–15156
14 Benidt GR, Jaben EA, Winters JL,
et al.: Identification of anti-PP1P(k) in
a blood donor and her family: a case
report following her pregnancy and
review. Transfus Apher Sci 2010;
43:369–374
15 Lopez M, Cartron J, Cartron JP, et al.:
Cytotoxicity of anti-PP1Pk antibodies
and possible relationship with early
abortions of p mothers. Clin Immunol
Immunopathol 1983; 28:296–303
16 Haentjens-Verbeke K, Dufour P, Vina-
tier D, et al.: Anti-Tja (PP1Pk) isoim-
munization. A case, a review of the
© 2020 International Society of Blood Transfusion
Vox Sanguinis (2021) 116, 591–600
Anti-PP1Pk isoantibodies in pregnancy 599
literature. J Gynecol Obstet Biol Reprod
(Paris) 1993; 22:393–397
17 Levene C, Sela R, Rudolphson Y, et al.:
Hemolytic Disease Of The Newborn
Due To Anti-PPIPk (Anti-Tja). Trans-
fusion 1977; 17:569–572
18 ISBT: Names for P1PK (ISBT 003)
Blood Group Alleles. ISBT. https://
www.isbtweb.org/fileadmin/user_upload/
files-2015/red%20cells/blood%20group
%20allele%20terminology/allele%20
tables/003_P1PK_Alleles_v3.1.pdf [Last
accessed 28 April 2020]
19 Kaczmarek R, Buczkowska A, Mikola-
jewicz K, et al.: P1PK, GLOB, and
FORS blood group systems and GLOB
collection: biochemical and clinical
aspects. Do we understand it all yet?
Transfus Med Rev 2014; 28:126–136
20 Steffensen R, Carlier K, Wiels J, et al.:
Cloning and expression of the histo-
blood group Pk UDP-galactose:
Ga1beta-4G1cbeta1-cer alpha1,
4-galactosyltransferase. Molecular
genetic basis of the p phenotype. J
Biol Chem 2000; 275:16723–16729
21 Suzuki N, Yamamoto K: Molecular
cloning of pigeon UDP-galactose:
beta-D-galactoside alpha1, 4-galac-
tosyltransferase and UDP-galactose:
beta-D-galactoside beta1,4-galactosyl-
transferase, two novel enzymes catalyz-
ing the formation of Gal alpha1-4Gal
beta1-4Gal beta1-4GlcNAc sequence. J
Biol Chem 2010; 285:5178–5187
22 Rock JA, Shirey RS, Braine HG, et al.:
Plasmapheresis for the treatment of
repeated early pregnancy wastage
associated with anti-P. Obstet Gynecol
1985; 66:57S–60S
23 National Center for Biotechnology
Information. ClinVar; 2019. [Last accessed
2 May 2020] https://www.ncbi.nlm.nih.
gov/clinvar/variation/VCV000002693.1
24 Joseph FA, Hyett JA, Schluter PJ,
et al.: New Australian birthweight cen-
tiles. Med J Aust 2020; 213:79–85
25 Weiss DB, Levene C, Aboulafia Y, et al.:
Anti-PP1Pk (anti-Tja) and habitual abor-
tion. Fertil Steril 1975; 26:901–903
26 Haentjens-Verbeke K, Dufour P, Vina-
tier D, et al.: Anti-TJa alloimmunization
(anti-PP1Pk): two consecutive pregnan-
cies of an anti-TJa-carrying patient.
Fetal Diagn Ther 1996; 11:120–125
27 Strowitzki T, Wiedemann R, Heim MU,
et al.: Pregnancy with an extremely
rare P blood group with anti-PP1Pk.
Geburtshilfe Frauenheilkd 1991;
51:710–713
28 Conlan S: Managing a pregnancy in
the presence of the rare blood group
antibody PP1Pk. Aust N Z J Obstet
Gynaecol 2004; 44:479–480
29 Heim MU, Weidemann R, Brehm G,
et al.: Plasmapheresis and/or inter-
mediate high dosage immunoglobulin
therapy–effective measures in threa-
tened premature labor and intrauterine
hemolysis in anti-PP1Pk (-Tja)? Beitr
Infusionsther 1990; 26:302–306
30 Fernandez-Jimenez MC, Jimenez-
Marco MT, Hernandez D, et al.:
Treatment with plasmapheresis and
intravenous immunoglobulin in preg-
nancies complicated with anti-PP1Pk
or anti-K immunization: a report of
two patients. Vox Sang 2001; 80:117–
120
31 Mohd Azri MS, Kunasegaran K, Azrina
A, et al.: Successful pregnancy out-
come in malaysian woman with rare p
phenotype and anti-PP1P(k) antibody.
Indian J Hematol Blood Transfus
2014; 30:405–408
32 Hanafusa N, Noiri E, Yamashita T,
et al.: Successful treatment by double
filtrate plasmapheresis in a pregnant
woman with the rare P blood group
and a history of multiple early miscar-
riages. Ther Apher Dial 2006; 10:498–
503
33 Padmanabhan A, Connelly-Smith L,
Aqui N, et al.: Guidelines on the Use
of Therapeutic Apheresis in Clinical
Practice - Evidence-Based Approach
from the Writing Committee of the
American Society for Apheresis: The
Eighth Special Issue. J Clin Apher
2019; 34:171–354
34 Shemin D, Briggs D, Greenan M: Com-
plications of therapeutic plasma
exchange: a prospective study of
1,727 procedures. J Clin Apher 2007;
22:270–276
35 Scott JR: Revisiting the use of intra-
venous immune globulin (IVIG) for
Kell alloimmunization. Am J Obstet
Gynecol 2018; 219:223–224
36 de la Camara C, Arrieta R, Gonzalez
A, et al.: High-dose intravenous
immunoglobulin as the sole prenatal
treatment for severe Rh immunization.
N Engl J Med 1988; 318:519–520
600 P. Di Ciaccio et al.
37 Urbaniak SJ, Duncan JI, Armstrong-
Fisher SS, et al.: Transfer of anti-D anti-
bodies across the isolated perfused
human placental lobule and inhibition
by high-dose intravenous immunoglob-
ulin: a possible mechanism of action. Br
J Haematol 1997; 96:186–193
38 Yu Z, Lennon VA: Mechanism of intra-
venous immune globulin therapy in anti-
body-mediated autoimmune diseases. N
Engl J Med 1999; 340:227–228
39 National Blood Authority: Patient
Blood Management Guidelines: Module
5 Obstetrics and Maternity. Canberra:
National Blood Authority (Australia),
2015
40 Kidson-Gerber G, Kerridge I, Farmer S,
et al.: Caring for pregnant women for
whom transfusion is not an option. A
national review to assist in patient
care. Aust N Z J Obstet Gynaecol
2016; 56:127–136
41 Nance S, Scharberg EA, Thornton N,
et al.: International rare donor panels:
a review. Vox Sang 2016; 110:209–
218
© 2020 International Society of Blood Transfusion
42 Kaczmarek R, Buczkowska A,
Mikołajewicz K, et al.: P1PK, GLOB,
and FORS blood group systems and
GLOB collection: biochemical and
clinical aspects. Do we understand it
all yet. Transfus Med Rev 2014;
28:126–136
43 Taniguchi F, Horie S, Tsukihara S,
et al.: Successful management of a
p‐incompatible pregnancy using dou-
ble filtration plasmapheresis. Gynecol
and Obstet Invest 2003; 56:117–
120.
Vox Sanguinis (2021) 116, 591–600
 Vox Sanguinis (2021) 116, 601–608
© 2020 International Society of Blood Transfusion
ORIGINAL PAPER DOI: 10.1111/vox.13047

Red cell genotyping of rare blood donors: donation

behaviour and data visualization
Waseem Q. Anani,1,2 Jerome Gottschall1,2,3 & Gregory A. Denomme1,3
1
Diagnostic Laboratories, Versiti, Milwaukee, WI, USA
2
Department of Pathology, Medical College of Wisconsin, Milwaukee, WI, USA
3
Blood Research Institute, Versiti, Milwaukee, WI, USA

Received: 10 August 2020,
revised 27 October 2020,
accepted 9 November 2020,
published online 5 January 2021
Abstract
Background The continual identification of rare blood among donors is critical
to support national programs like the American Rare Donor Program (ARDP).
Some blood centres require consent from donors to be registered with a national
registry. This situation provides an opportunity to determine whether a donor’s
willingness to register is associated with a change in donation behaviour.
Methods Rare donors were identified by molecular typing. The average number
of donations per year was compared for each donor prior to and after receiving a
consent letter. Donors were categorized as either accepting or declining the
request. Non-parametric t tests compared the statistical significance within and
between categories. Rare types were overlaid with consensus data to look for
trends using data visualization techniques.
Results A total of 270 molecularly typed rare donors received letters over
4 years. Half of the donors (132, 49%) agreed to participate in the ARDP. Overall,
donation frequency increased after the letter when enrolled. Both Caucasian and
non-Caucasian donors increased their donations after enrolling providing an
additional 159 red blood cell units over 3 years. Declining participation did not
change donation frequency. Data visualization showed that enrolled donors were
more affluent, high school and college educated, and lived in their home for
longer periods of time.
Conclusion A donor’s willingness to enrol in the ARDP was associated with a
post-response increase in donation frequency. New interventions to reach non-
Caucasian donors may be a prerequisite to increase donation frequency and a
willingness to be a rare blood donor.
Key words: donor recruitment, rare blood, mass-scale red cell genotyping, donor
geo mapping.

disconcerting for both uncommon and rare blood types.

Introduction
Red blood cell use has been and continues to contract in
recent years [1]. However, the demand for antigen-
matched blood rare blood types remains a continual need.
The unmet need for antigen-negative blood is
Correspondence: Waseem Anani, MD, Blood Research Institute, Versiti,
Inc., 638 N 18th Street, Milwaukee, WI 53233, USA
E-mail: Waseem.anani@blood.ca
Between 2005 and 2013, 9
7% of high-prevalence anti-
gen-negative and combination antigen-negative Ameri-
can Rare Donor Program (ARDP) requests went unfilled
or partially filed [2]. Currently, the ARDP enrols donors
that are negative for high-prevalence antigens (<1:1000)
and only group O and A donors negative for multiple
common antigen-negative types that are deemed rare [3].
The active donor database is predominately comprised of
donors negative for multiple common antigens (91
5%).

601
602 W. Q. Anani et al.
Yet disproportionately two-thirds of rare blood requests
are units negative for high-prevalence antigens. A 2003
survey identified 52 patients with antibodies to high-fre-
quency antigens and found one-third of the patients had
unsatisfactory transfusion support including incompatible
transfusions [4]. Thus, screening, recruiting and maintain-
ing active rare donors are especially important to bridge
the disparity between the donor database and requests for
rare blood. In addition to the need for rare blood, recent
recommendations for the transfusion of Rh and Kell anti-
gen-matched blood for patients with sickle cell disease,
common among African Americans but not so common
among the general donor population, may rapidly out-
weigh the supply [5,6].
Some information on the success of donor recruitment
can be gleaned from the American Rare Donor Program
(ARDP). Donor recruitment strategies applied to the gen-
eral population were extended to rare donors and include
phone calls [7–9], letters [7–11], and incentives [12–14].
Once contacted, blood centres rely on donors remaining
committed to return when recruited. Some blood centres
have noted difficulty in recruiting donors including those
with uncommon and rare blood types using standard
practices and have opted for more personalized communi-
cations [10,15,16]. The resource-intense processes of iden-
tifying, recruiting and maintaining a donor databases and
inventory rely on the premise that the standard recruiting
methods are sufficient to meet transfusion recipient
demand.
However, overall, donor behaviour is changing due to
a diversification of demographics. The bulk of donors
have shifted due to health issues, travel deferral and age-
ing-out for the historically large baby boomer donor pop-
ulation. The broad behaviour change of today’s donors
likely follows generational trends, most notedly Genera-
tion Y and Millennials [17–19]. The long-term outcome of
affecting donor behaviour using standard methods is
sparse for these two populations [10,15].
The aim of this study was to evaluate the change in
donation frequency of rare donors when contacted by the
accepted method of communication for recruitment into
the ARDP: letters and phone calls. We collated routinely
gathered comprehensive donor data and used geo map-
ping and data visualization tools to broadly identify loca-
tions with an over-representation of certain types, which
would assist in smart donor drives resulting in better suc-
cess rates at finding and fulfilling future recruitment
needs. With such a strategy, future recruitment into
uncommon and rare donor databases would be driven
through the identification of first-time donors with in-de-
mand blood types using strategies tailored to today’s dif-
ferent donor demographics.
Material and methods
Genotyping and donor recruitment strategies
This study was a retrospective review of rare donors
derived from an internal database (LifeTrak, Mediware
Information Systems, Oak Brook, IL). Rare donors were
defined by the American Rare Donor Program: antigen-
negative requests for antigens found in <1:1000 or rare
units secondary to combinations of multiple antigen-neg-
ative requirements. All rare donors were identified by red
cell genotyping per a previously described method [20].
Once identified, rare donors received a letter describing
their rare blood group antigen(s) with a request to enrol
in the American Rare Donor Program and later, at least
two standard (not tailored to their rare type) phone calls
were made to set up additional future donations, regard-
less of response to the letter.
Data collection and analysis
The data were collected to represent only molecularly
typed blood donors from the start of genotyping at
BloodCenter of Wisconsin, 1 January 2010 to 31 Decem-
ber 2013. We collected data pertaining to donor address,
age, gender, ethnicity, dates of donation, response to the
rare donor letter and rare antigen(s) identified. The donors
were given the option to accept or decline the request to
enrol in the American Rare Donor Program; donors not
responding 6 months after the letter mailing were consid-
ered to decline participation. Such a policy affords the
opportunity to examine trends in donation among those
donors willing to register with a national rare donor pro-
gram.
The mean number of donations before and after con-
tacting a donor was compared to account for the year-to-
year variations in donation frequency. Donor baseline
donation rate was established by averaging the number
of red blood cell units donated per year in the three pre-
ceding years. Following the rare donor letter mailing, red
blood cell donations were averaged over the subsequent
three years. Donors were categorized by their response to
the rare donor letter (yes or no). The mean donations per
year for each donor prior to receiving the rare donor let-
ter and after the letter was mailed were compared with a
Wilcoxon matched-pairs signed rank test (two-tailed,
P < 0
05) with GraphPad Prism (v8.0.2, La Jolla, Califor-
nia, USA). The difference in donation frequency before
and after the mailing for each donor was compared
between the two groups with a Mann–Whitney test (two-
tailed, P < 0
05) with GraphPad Prism. Statistical consid-
erations about donor subgroups were limited to Caucasian
© 2020 International Society of Blood Transfusion
Vox Sanguinis (2021) 116, 601–608

and non-Caucasian donors secondary to sample size of
minority donors. Instead, all minority donors were
grouped as non-Caucasians and analysed as one group.
Donor geolocation and visualization
Distance to donation sites and socio-economic status are
cited as some of the barriers to donate for minority donors
based on prior surveys [21,22]. We applied geolocation and
data visualization of our rare donors, for whom minorities
are integral, to semi-quantitatively assess whether these
factors affect our local rare donors. To visualize and calcu-
late the distance of a donor to the closest donor site, a
Microsoft Excel (2016, version 1803) Visual Basic for
Applications script was used to calculate the driving dis-
tance of donors to physical donor sites in the Milwaukee
region. The per cent of donors within five and 10 miles of
physical donor sites was calculated. Mobile donor sites
were not included since fixed donation sites represent a
static location where donors can always visit. Second, data
were categorized by age and ethnicity and were visualized
by the sum of donors in each zip code to identify any local-
ized regions or trends of rare blood donors to include med-
ian household income, education (Bachelor’s degree or
higher) and per cent of individuals living in their homes
less than 4 years compared with the 2015 5-year census
data [23]. The 2015 census data were chosen to best fit the
range of donations during the study period. Data were
Fig. 1 Rare donors displayed a variety of desirable antigen-negative genotypes sorted by prevalence.
© 2020 International Society of Blood Transfusion
Vox Sanguinis (2021) 116, 601–608
Geo mapping of rare blood donors 603
visualized with 3D Map (Microsoft Excel 2016, version
1803) [24]. The donor and extracted population data were
superimposed on geographic maps by zip code.
Results
A total of 309 rare donors were identified with some
donors lacking more than one desirable antigen during
this time (Figs 1 and 2). Only 270 molecularly typed,
newly identified rare donors (0
6% of all donors geno-
typed) were found to fulfil the three preceding and subse-
quent three year history after received letters and phone
calls for the study. The mean age for a rare donor was 41
(Table 1) with non-Caucasian donors being younger on
average than Caucasians. African Americans (53%, 142/
270) and Caucasians (37%, 100/270) were the most com-
mon ethnicities of rare donors. Age and sex were not sta-
tistically significant between the two groups.
Donation frequency preceding and following rare
donor letter
A total of 132 (49%) donors agreed to participate in the
ARDP with a median of 2
0 (range 0
40–4
5) donations
per year before and 2
0 (range 0
50–5
7) donations per
year after the letter was sent (P < 0
0001, Fig. 3a). The
median number of donations per year did not change for
donors declining ARDP participation before (1
0, range
604 W. Q. Anani et al.
Fig. 2 Rare donors were found in all donor age groups by decade. The largest single group of donors were ages 20–29 with a predominance of Afri-
can-American donors. Overall, rare donors were most likely African American.
0
60–4
0) and after (1
0, range 0
60–3
3) the letter was
sent. The number of donations before and after respond-
ing was statistically insignificant (P = 0
52). In total, the
two groups’ donation patterns were statistically different
after receiving the ARDP letter (P = 0
0004).
The data were then categorized by Caucasian and non-
Caucasian groups to look for further trends. Of Caucasian
donors, 83 (83%) agreed to participate in the ARDP with
a median of 2
1 (range 0
16–4
5) donations per year
before and 2
5 (range 0
60–5
7) donations per year after
the letter was sent (Fig. 3b). The 17 Caucasian donors
declining ARDP participation had a median of 1
3 (range
0
60–3
7) and 1
2 (range 0
60–3
3) donations per year
before and after the letter was sent, respectively. The
number of donations before and after responding within
each group was only statistically significant for donors
participating in the ARDP (P = 0
050 vs <0
0001). The
two groups’ donation patterns were statistically different
after receiving the ARDP letter (P = 0
0024).
Of non-Caucasian donors, 49 (29%) agreed to participate
in the ARDP with a median of 1
3 (range 0
40–3
0) dona-
tions per year before and 1
7 (range 0
50–4
8) donations
per year after the letter was sent (Fig. 3b). The 121 non-
Caucasians declining ARDP participation had a median of
1
0 (range 0
67–4
0) and 1
0 (range 0
60–3
0) donations
before and after the letter was sent, respectively. The num-
ber of donations before and after responding within each
group was statistically significant for donors declining
(P = 0
015) and participating (P = 0
0090) in the ARDP.
The two groups’ donation patterns were statistically differ-
ent after receiving the ARDP letter (P < 0
0001).
Overall, irrespective of ethnicity, donors enrolling in
the ARDP resulted in an increase in the median donation
rate by 0
4. Compared with donors not enrolling in the
rare donor program, the median donation rate increase in
Caucasian donors increased red blood cells collected
from 523 to 623 units, leading to an additional 100 red
blood cell units over three years. Non-Caucasian donors
© 2020 International Society of Blood Transfusion
Vox Sanguinis (2021) 116, 601–608
 Geo mapping of rare blood donors 605

Table 1 Rare donor demographics categorized by response to the ARDP

letter Data visualization
Overall, the median household income of donors in zip
Rare donor demographics
codes enrolling in the ARDP was $57 268 and $40 453 for
ARDP letter response donors declining participation. Education and length occu-
pancy in homes also differed. Approximately 92% ARDP-
Yes No enrolled donors had a high school education and 32% had a
Mean Age (range) 41 (18–92) 41 (16–92)
bachelor’s degree whereas 87% donors declining enrolment
Sex had a high school education and 26% had a bachelor’s
Male 56 70 degree. Home occupancy less than four years in a single res-
Female 76 68 idence for enrolled donors was 4% and 6% for donors
Ethnicity declining participation. Most rare donors lived within a 5
African American 39 104 mile driving radius of a fixed donation site (95%), and
Asian 2 2 almost all donors lived within 10 miles (99%) (Fig. 4).
Caucasian 83 17 A disparity also existed between Caucasian and non-
Hispanic 6 8
Caucasian donors within each category based on zip code
Other 2 7
census data. Caucasian donors enrolling in the ARDP
Blood Group
made a median salary of $61 151, had a 93% high school
O 64 66
A 49 45
graduation rate, 33% with a bachelor’s degree, and only
B 16 23 3% lived in their homes less for than 4 years compared
AB 3 4 with non-Caucasians with a median salary of $40 852, an
Total 132 138 85% high school graduation rate, 27% with a bachelor’s
degree and 7% living in their homes less for than 4 years.
Caucasian donors declining ARDP participation had a
median salary of $63 713, 94% high school graduation
responding favourably to the ARDP letter had an addi- rate, 38% with a bachelor’s degree and 4% in their homes
tional 191–250 red blood cell units compared with donors less for than 4 years. Non-Caucasian donors had a med-
declining participation, leading to an additional 59 red ian salary of $35 230, 85% with a high school diploma,
blood cell units over three years. Just over half (138/270) 23% with a bachelor’s degree, and 7% lived in their
of donors failed to enrol in the ARDP and did not change homes less than 4 years. Rare non-Caucasian donors were
donation frequency, despite a statistically significant younger and clustered in specific zip codes, with a pau-
change in non-Caucasians. city of rare Caucasian donors.

Fig. 3 (a) Overall donation patterns in response to a rare donor letter. Donors not registering to be a rare donor did not change their frequency signifi-
cantly. In contrast, donors that responded favourably increased the number of donations per year significantly. (b) When the donation patterns were sep-
arated, Caucasian that declined participation (n = 17) in the rare donor program was not significant in comparison with non-Caucasian donors
(n = 121) that decreased their donation frequency. Both groups had statistically significant increases in donations following enrolment in the rare donor
program (Caucasian n = 83 and non-Caucasian n = 49).

© 2020 International Society of Blood Transfusion

Vox Sanguinis (2021) 116, 601–608
606 W. Q. Anani et al.

(a) (b) (c)

(d) (e) (f)

Fig. 4 Rare donors were mapped by zip code with and overlaid with census data to look for commonalities among donors. Reference maps (a, b, c)
display donors by ethnicity in pie charts, age in bar charts and census data in colour gradients. Greater than 98% of the donor data were captured in
the map screenshots. Non‐Caucasian donors lived in defined zip codes and were younger on average than Caucasian donors. Non‐Caucasian donors had
lower incomes, lived in their homes for less time and were less likely to hold a high school education or Bachelor’s degree (d, e, f). The size of the pie
chart represents the total number of donors in the zip code. [Colour figure can be viewed at wileyonlinelibrary.com]

A personalized method could result in a more favourable

Discussion
A donor’s willingness to enrol in the ARDP was associ-
ated with a post-response increase in donation frequency
regardless of ethnicity. Communication with donors
would have resulted in an additional 159 red blood cell
units over three years based on the median donation
rates. Both Caucasians and non-Caucasians declining or
not responding to the ARDP letter did not change the fre-
quency of their donations. Although the trends in beha-
viour were similar between the groups, the majority of
Caucasians enrolled in the ARDP, while the opposite was
noted for non-Caucasians. The median income, the aver-
age education and home occupancy duration were higher
for Caucasians than non-Caucasians.
Although rare blood is ‘rarely’ needed, the donor–recip-
ient relationship is not 1:1. Transfusion recipients often
require more than one unit, and the preference is to have
a liquid unit. As high-throughput genotype screening for
rare donors [25] and the number of diseases benefiting
from extended antigen matching becomes more common
[26–29], finding and maintaining rare donors is essential.
The biggest gain in rare donors could be made among
non-Caucasians that typically do not join the ARDP.
response rate compared with letters and phone calls [30].
For example, first-time African-American donors
increased from 12
2% to 60% with community engage-
ment and personal stories from the community. While the
increase in units over three years appeared modest, rare
units can be exceedingly difficult to find depending on
the phenotype. Increasing donations in this group could
lead to reductions in organizational resources to obtain
red blood cells and with time, an excess of rare units that
could be frozen for future use.
Letters and phone calls were least effective with non-
Caucasians and may be partly explained by disparities in
socio-economic factors. Milwaukee has an urban region
of significant racial segregation like other cities in the
United States with many living in socio-economic stress
predominately in the zip codes identified in this study for
non-Caucasian donors [31]. African-American blood
donors are more likely to be young – an attractive trait
for potential long-term donations. However, the non-Cau-
casian donors are also more likely to occupy a residence
for less time. And failure to respond was a significant
factor since donor engagement is key to a successful rare
donor recruitment program. Thus, letters mailed to an

© 2020 International Society of Blood Transfusion

Vox Sanguinis (2021) 116, 601–608
 Geo mapping of rare blood donors 607

address cannot reach the intended recipient and donors
are lost to follow-up. Engaging younger donors with
phone applications [32] and improved marketing strate-
gies [33] can improve the conversion of first-time to
repeat donors. Our data suggest frequent donors are more
likely to consent when approached to participate in the
ARDP. The overlying census data with our donor geogra-
phies do not provide causality of why a donor may not
respond. These data instead highlight areas for further
investigation (e.g. messaging donors from a particular
community or testing new methods of communication to
effect a change in donor frequency). The retrospective
nature of this study has some limitations. The baseline
donation frequency of donors enrolling in the ARDP was
higher than donors declining participation. This study
was not designed to ascertain differences in donor char-
acteristics. Due to the limited number of rare donors, eth-
nicities and lifetime number of donations could not be
completely stratified. The passive nature of collecting data
precluded varying the communication method to observe
an effect. Thus, we cannot conclude which contact
method would be appropriate for unresponsive donors.
Regardless, the financial contraction of the blood
industry requires more efficient use of the current
resources and optimizing collection strategies. Direct costs
to recruit rare red blood cell donors are unknown, but in
2007, it was estimated that selective recruitment of family
members for HLA matched haematopoietic stem cell
screening cost approximately €7 per new donor found for
only mailing letters [34]. By identifying donors less likely
or unfavourably respond to standard methods, such as
relatively short-term geographic residence, the monetary
donor acquisition cost can be redirected to a more appro-
priate method. Improving the first interaction with a rare
donor can be crucial to encouraging life-long donation
frequency [35]. Once established, a program could easily
expand to identify and target uncommon donors such as
R0R0/R0r to donate more frequently, which are also in
high demand due to guideline recommendations.5
We demonstrated that rare donors enrolling in the
ARDP increase their average number of donations per
year. The increased donations would result in a meaning-
ful rise in rare units available for a blood centre. Addi-
tionally, rare donors not responding as favourably to
letters and phone calls were more likely non-Caucasian.
These donors tended to be less wealthy, less educated
and more transient making contact potentially more dif-
ficult. However, contact with these donors did not result
in a decrease in donations. By applying geo mapping
more broadly, other valuable donors could be located
and more effectively, targeted with messaging that res-
onates better based on various donor groups and demo-
graphics.
Conflict of interest
The authors declare that they have no conflicts of interest
related to this manuscript.

References
1 Yazer MH, Anani WQ, Denomme GA,
et al.: Trends in antigen-negative red
blood cell distributions by racial or
ethnic groups in the United States.
Transfusion 2018; 58:145–50
2 Flickinger C: REGGI and the American
rare donor program. Transfus Med
Hemother 2014; 41:342–5
3 Meny GM, Flickinger C, Marcucci C:
The American rare donor program. J
Crit Care 2013; 28: 110 e9–e18
4 Seltsam A, Wagner FF, Salama A,
et al.: Antibodies to high-frequency
antigens may decrease the quality of
transfusion support: an observational
study. Transfusion 2003; 43:1563–6
5 Compernolle V, Chou ST, Tanael S,
et al.: Red blood cell specifications for
patients with hemoglobinopathies: a
systematic review and guideline.
Transfusion 2018; 58:1555–66
6 Fasano RM, Meyer EK, Branscomb J,
et al.: Impact of red blood cell antigen
matching on alloimmunization and trans-
fusion complications in patients with
sickle cell disease: a systematic review.
Transfus Med Rev 2019; 33:12–23
7 Thornton NM: The United Kingdom
national rare donor panel. Immunohe-
matology 2016; 32:67–9
8 Goldman M, St Croix L: Rare donor
program: Canadian Blood Services.
Immunohematology 2016; 32:15–6
9 Peyrard T: The French national rare
blood program. Immunohematology
2016; 32:23–5
10 Yahalom V, Finkel L, Shinar E, et al.:
The Israeli rare donor blood program.
Immunohematology 2016; 32:11–2
11 Flickinger C: The American rare donor
program. Immunohematology 2016;
32:71–3
12 Paccapelo C, Truglio F, Villa MA,
et al.: Rare donor program in Italy.
Immunohematology 2016; 32:47–8
13 Soeker R: Rare donor program at Wes-
tern Province Blood Transfusion Service
in South Africa. Immunohematology
2016; 32:57–8
14 Moghaddam M: National rare donor
program in Iran. Immunohematology
2016; 32:27–8
15 Price CL, Boyd JH, Watkins AR, et al.:
Mailing of a sickle cell disease educa-
tional packet increases blood donors
within an African American com-
munity. Transfusion 2006; 46:1388–
93
16 Bachegowda LS, Timm B, Dasgupta P,
et al.: Impact of predictive scoring
model and e-mail messages on African
American blood donors. Transfusion
2017; 57:1515–21

© 2020 International Society of Blood Transfusion

Vox Sanguinis (2021) 116, 601–608
608 W. Q. Anani et al.
17 Ertas N: Millennials and volunteering:
sector differences and implications for
public service motivation theory. Pub-
lic Adm Quarterly 2016; 40:517–58
18 Paulin M, Ferguson RJ, Schattke K,
et al.: Millennials, social media, proso-
cial emotions, and charitable causes:
the paradox of gender differences. J
Nonprofit Public Sector Market 2014;
26:335–53
19 Scholz C: Generation Y and blood
donation: the impact of altruistic help
in a darwiportunistic scenario. Trans-
fus Med Hemother 2010; 37:195–202
20 Flegel WA, Gottschall JL, Denomme
GA: Implementing mass-scale red cell
genotyping at a blood center. Transfu-
sion 2015; 55:2610–5; quiz 09
21 Shaz BH, James AB, Demmons DG,
et al.: The African American church as
a donation site: motivations and barri-
ers. Transfusion 2010; 50:1240–8
22 Shaz BH, Demmons DG, Hillyer KL,
et al.: Racial differences in motivators
and barriers to blood donation among
blood donors. Arch Pathol Lab Med
2009; 133:1444–7
23 Bureau USC: American Community
Survey, 2015 American Community
Survey 5-Year Estimates. http://factf
inder.census.gov (Last accessed 1/27/
2019 2019).
24 Au C, Rischpater R: Power Map for
Excel; Microsoft Mapping: Geospatial
Development in Windows 10 with Bing
Maps and C#. Berkeley, CA, Apress,
2015:159–65
25 Wagner FF, Doescher A, Bittner R,
et al.: Extended donor typing by
pooled capillary electrophoresis:
impact in a routine setting. Transfus
Med Hemother 2018; 45:225–37
26 Yee MEM, Josephson CD, Winkler AM,
et al.: Red blood cell minor antigen
mismatches during chronic transfusion
therapy for sickle cell anemia. Trans-
fusion 2017; 57:2738–46
27 Chapuy CI, Nicholson RT, Aguad MD,
et al.: Resolving the daratumumab
interference with blood compatibility
testing. Transfusion 2015; 55:1545–54
28 Anani WQ, Marchan MG, Bensing KM,
et al.: Practical approaches and costs
for provisioning safe transfusions dur-
ing anti-CD38 therapy. Transfusion
2017; 57:1470–9
29 Shirey RS, Boyd JS, Parwani AV,
et al.: Prophylactic antigen-matched
donor blood for patients with warm
autoantibodies: an algorithm for
© 2020 International Society of Blood Transfusion
transfusion management. Transfusion
2002; 42:1435–41
30 Price CL, Johnson MT, Lindsay T, et al.:
The Sickle Cell Sabbath: a community
program increases first-time blood donors
in the African American faith commu-
nity. Transfusion 2009; 49:519–23
31 Cooper RA, Cooper MA, McGinley EL,
et al.: Poverty, wealth, and health care
utilization: a geographic assessment. J
Urban Health 2012; 89:828–47
32 Gemelli CN, Carver A, Garn A, et al.:
Evaluation of the impact of a personal-
ized postdonation short messaging ser-
vice on the retention of whole blood
donors. Transfusion 2018; 58:701–9
33 Boenigk S, Leipnitz S: Acquiring
potential blood donors in large cities:
a preference-based donor segmenta-
tion study. J Nonprofit Public Sector
Market 2016; 28:364–93
34 Schmidt AH, Stahr A, Baier D, et al.:
Selective recruitment of stem cell
donors with rare human leukocyte
antigen phenotypes. Bone Marrow
Transplant 2007; 40:823
35 Sojka BN, Sojka P: The blood dona-
tion experience: self-reported motives
and obstacles for donating blood. Vox
Sang 2008; 94:56–63
Vox Sanguinis (2021) 116, 601–608
 Vox Sanguinis (2021) 116, 609–612
© 2021 International Society of Blood Transfusion
INTERNATIONAL FORUM DOI: 10.1111/vox.13061

International Forum on Transfusion Practices in

Haematopoietic Stem-Cell Transplantation: Summary
Pilar Solves, Miquel Lozano, Eugene Zhiburt, Javier Anguita Velasco, Ana Maria P
erez-Corral, Silvia Monsalvo-Saornil,
Sho Yamazaki, Hitoshi Okazaki, Kathleen Selleng, Konstanze Aurich, William Kr€ uger, Andreas Buser, Andreas Holbro,
Laura Infanti, Gregor Stehle, Luca Pierelli, Antonella Matteocci, Luigi Rigacci, Karen M. K. De Vooght, Jurgen H. E. Kuball,
Katherine L. Fielding, David A. Westerman, Erica M. Wood, Claudia S. Cohn, Andrew Johnson, Mickey B. C. Koh,
Dara Qadir, Christine Cserti-Gazdewich, Etienne Daguindau, Fanny Angelot-Delettre, Pierre Tiberghien,
Silvano Wendel-Neto, Roberta-Maria Fachini, Suzy Morton, Charles Craddock, Matthew Lumley,
Jolanta Antoniewicz-Papis, Kazimierz Hałaburda, Magdalena Łez towska & Nancy Dunbar

Then, hospital transfusion departments must assure an

Introduction
Haematopoietic stem-cell transplantation (HSCT) is per-
formed mainly in patients diagnosed with malignancies
or some severe congenital diseases. The two main
modalities of HSCT are autologous in which stem cells
are harvested from the patient and allogeneic in which
the cells are harvested from a donor. Autologous HSCT
is performed mainly in patients who have multiple mye-
loma and lymphoma, while allogeneic HSCT is per-
formed for acute leukaemia and myeloproliferative and
myelodysplastic syndromes. The improvement in the
procedures and clinical outcome has led to a progressive
increase in the number of transplantations performed
around the world [1,2]. New modalities such as hap-
loidentical HSCT, that have considerable transfusion
requirements, have also contributed to the increasing
transplantation activity [3].
Blood product transfusion support, mainly in the form
of red blood cells (RBCs) or platelet (PLT) support, is
necessary for the prolonged cytopenia period that occurs
after transplantation and also for some complications
that can arise during the procedure such as graft-versus-
host disease (GVHD), viral infections or immune haemol-
ysis [4].
Platelet transfusion can be used with therapeutic and/
or prophylactic approaches. Platelet transfusion policy is
usually based on guidelines summarizing available scien-
tific evidence. While the prophylactic transfusion strategy
is well established for allogeneic HSCT, there is not
enough evidence to support this practice in the autolo-
gous transplantation setting [5]. Prophylactic transfusion
strategy using low-dose platelets (threshold of 10 9 109/
L) is the current recommendation for haematological
patients undergoing chemotherapy or allogeneic stem cell
transplantation. This threshold should be increased to
20 9 109/L if there are some additional risk factors for
bleeding [6].
adequate transfusion management for each type and
moment of transplantation, paying special attention to the
ABO group mismatch between donor and recipient [7].
Major incompatibility is defined when the recipient plasma
has isohaemagglutinins against donor RBC and minor
when the donor has isohaemagglutinins against recipient
RBC. Immunohaematological complications of ABO incom-
patibility HSCT should be detected in time to be treated.
There are some guidelines about transfusion management
of patients receiving HSCT; however, there is little scien-
tific evidence supporting them [8–10]. Despite the high
number of transplantations performed and the long experi-
ence in transfusion practices in this subset of patients, there
are still controversial and unclear aspects.
This Vox Sanguinis International Forum aims to assess
current transfusion practices in the field of haematopoi-
etic stem cell transplantation.
Summary of responses
This survey provides an interesting overview of transfu-
sion practices and immunohaematological follow-up in
the HSCT setting around the world. Data from 13 coun-
tries and 14 centres from North America, South America,
Asia, Australia and Europe, performing both autologous
and allogeneic HSCT, have been collected. There is a wide
variability in the transplantation activity of centres rang-
ing from 20 to 250 autologous HSCT and from 10 to 190
allogeneic HSCT, and also in the patient transfusion man-
agement. The characteristics of centres and the summary
of the answers to most issues are shown in Tables 1 and
2, respectively.
In patients undergoing autologous HSCT, the time
receiving irradiated blood product ranged from 3 months
(Birmingham) to lifelong. More than half of the centres
do not delete the blood product irradiation condition in
patients who underwent allogeneic HSCT. Different

609
610 International Forum

practical reasons are given as patients having other or de

Number of allogeneic HSCT

novo indications to receive irradiated blood products (flu-
darabine during conditioning), disease relapse and diffi-
culty of getting follow-up information.
Most centres transfuse 1 red blood cell (RBC) unit per
episode, and RBC transfusion thresholds are 80 or less in

20–30

60–70
all hospitals. In fact, the restrictive blood transfusion

125

190
100

100

116
60
33

15

30

55
32
10
strategy is widely extended among centres. Only hospitals
from Japan, Australia, Germany and France use the 2
RBC units transfusion policy. The 1 RBC transfusion pol-
icy has been implemented in many transplantation units

Number of autologous HSCT

around the world since the published reports provide
noninferior outcome and blood savings in patients receiv-
ing 1 versus 2 RBC units [11]. The haemoglobin thresh-
olds are highly variable; they are as low as 57 g/L in the
centre from Germany to 80 g/L in 8 centres (Table 1). The
centres from Switzerland and both from UK follow the

20–30

60–80
most restrictive transfusion policy: a haemoglobin thresh-

250

120

158
350
110
22

45
20

30

32
24
21
old of 70 g/L and 1 RBC unit.
Platelet prophylactic transfusion threshold is mostly
10 9 109/L in the absence of bleeding risk factors, and

Institut fur Immunologic and Transfusionsmedizin, Universitatsmedizin Greifswald

only the centre from France uses 20 9 109/L. In actual
practice many, patients undergoing HSCT have fever, infec-

Department of Laboratory Medicine and Pathology, University of Minnesota

tion or other factors that decrease the platelet yields and
increase the transfusion threshold to 20 9 109/L. Most cen-
tres do not perform a routine platelet count shortly after
platelet transfusions to evaluate yields, but if they detect a Regional Blood Transfusion Service, University Hospital Basel
low count after transfusion in routine blood cell count
Department of Blood Transfusion, University of Tokyo

(usually the next day), then they perform the platelet count

University Health Network, University of Toronto

between 15 and 60 minutes after platelet transfusion. All
centres except two use mostly platelet products suspended
Pirogov National Medical Surgical Center

in additive solution. Haemolytic transfusion reactions are a

well-known but rare complication of ABO plasma minor
University Medical Center Utrecht

incompatible platelet transfusions [12]. Because of lack of

Peter MacCallum Cancer Centre

University Hospital of Besancon

St George’s University Hospital
San Camillo Forlanini Hospital

ABO compatible platelets, up to 40% of platelet transfu-

on
n

Queen Elizabeth Hospital

Sirio-Libanes Blood Bank

Hospital Gregorio Mara~

sions in current practice are ABO incompatible [13]. Six

centres have established different approaches to reduce pla-
telet product isoagglutinin content, when plasma incom-
patible platelets must be transfused. The centre from
Germany pays special attention to platelet ABO plasma
Centre

compatibility in low-weight children and patients with high

transfusion frequency. The centre from Minnesota has a
maximum volume limit (1000–1500 ml) of ABO incompati-
ble plasma to be transfused per week, when it is reached,
Birmingham
Table 1 Characteristics of centres.

Minneapolis

they perform a direct antiglobulin test (DAT) and if positive

Melbourne
Greifswald

Sao Paulo
Besancon
Moscow

Toronto

the volume of incompatible plasma of platelets is reduced.

Utrecht

London
Madrid
Tokyo

Rome
City

Basel

The remaining four centres (Birmingham, Canada, France

and Brazil) take into account the isoagglutinin titres in pla-
telet products, transfusing unmodified platelet products if
the titre is low and performing plasma volume reduction if
Netherlands
Switzerland

the titre is high. HLA-matched single donor platelets are

Country

Germany

Australia

Canada
France

mostly used in patients with platelet transfusion refractori-

Russia

Japan

Brazil
Spain

Italy

USA
UK

ness. Two centres (Russia and Switzerland) manage this

© 2021 International Society of Blood Transfusion

Vox Sanguinis (2021) 116, 609–612
 Table 2 Summary of responses.

Country Russia Spain Japan Germany Switzerland Italy Netherlands Australia USA UK Canada France Brazil

Irradiation of blood products

137
Method X-ray NR X-ray c-ray (Cs137) c-ray (Cs137) c-ray (Cs137) X-ray c-ray X-ray X-ray c-ray c-ray (Cs ) X-ray c-ray (Cs137)
(X-ray on
occasion)
Dose 25–35 Gy 25 Gy 15–50 Gy 30 Gy 30 Gy 25 Gy 25–50 Gy 25–50 Gy 25 Gy 25 Gy 25–50 Gy 25 Gy 25–45 Gy 25 Gy

Vox Sanguinis (2021) 116, 609–612

Time after autologous/ NR 6 months/1 year Lifelong 3–6 months Lifelong NR 1 year/5 years Lifelong 1 year/Lifelong 1 year/Lifelong Lifelong Lifelong 3 months/1 year Lifelong
allogeneic HSCT (minimum)
RBC transfusion policy
Haemoglobin threshold in 80 g/L 80 g/L 70 g/L 57 g/L 70 g/L 80 g/L 80 g/L 80 g/L 80 g/L 70 g/L 70 g/L 70–75 g/L 80 g/L 70–80 g/L
stable patients
RBC units 2 vs 1 1 1 2 2 (mostly) 1 1 1 1–2 1 1 1 1 2 (90%) 1
RBCs are changed Yes No Yes Yes Yes Yes No Yes Yes NR Yes No No Yes

© 2021 International Society of Blood Transfusion

to blood (for platelets
donor group when is only)
engrafted?
PLT transfusion policy
Prophylactic (1)/ 1 1 1 1 1 1 1 1 1 1 1 1 1 1
therapeutic for (2 in a clinical
autologous HSCT (2) trial)
PLT threshold for 10 x 109/L 10 9 109/L 10–20 9 109/L 10 9 109/L 10 9 109/L 10 9 109/L 10 9 109/L 10 9 109/L 10 9 109/L 10 9 109/L 10 9 109/L 10 9 109/L 20 9 109/L 10 9 109/L (in absence
prophylactic 20 9 109/L 20 9 109/L 20 9 109/L 20 9 109/L (in absence of bleeding)
transfusion (if bleeding (if bleeding (if fever or (if bleeding of risk
risk factors) risk factors) bleeding) risk factors) factors)
Is a post-transfusion PLT Next day No 60 min (only in Next day 15–60 min Next day 60 min Next day 10–60 min Next day Next day Next day Next day Next day
count refractoriness)
systematically performed?
PLT refractoriness 24-h continuous HLA-matched NR HLA-matched HLA-matched Fresh RP/Cross- HLA-matched HLA-matched Cross-matched HLA-matched HLA/HPA- HLA-matched HLA/HPA- HLA/HPA-matched SDP
management infusion SDP SDP SDP/24-h matched SDP SDP PLT assay/HLA- SDP matched SDP compatible
continuous matched trial SDP SDP
infusion
PRCA monitoring and management
Is routine Isoagglutinin Weekly or twice No, only when No No Yes Weekly until No Yes (+100 No Every year No No Yes (+1 month) Weekly
determination a week if initial haemolysis (+30, +180 engraftment days)
after HSCT performed? titre >1/128 suspicion days)
Incidence* 0% 15% 10–15% 1 case/year 20% 5% 3 cases/year 1–3% Rare 10% NR NR 5% NR
Treatment NR Transfusion/Rtx Transfusion Transfusion, Transfusion/taper Transfusion Transfusion, Transfusion/ Conservative/ Transfusion EPO Transfusion/IS/ Transfusion/Rtx NR
selected cases CS, IS IS/Dara/Rtx and CS others EPO CS/Rtx/PE apheresis/Dara

CS, steroids; Dara, daratumumab; EPO, erythropoietin; HSCT, haematopoietic stem cell transplantation; IS, immunosuppressive agents; NR, not reported; PE, plasma exchange; PLT, platelets; PRCA, pure red cell
aplasia; RP, random platelets in plasma; RP-AS, random platelets in additive solution; RP-AS-PR, random platelets in additive solution, pathogen reduced; RTX, rituximab; SDP, single donor platelets in plasma;
SDP-AS, single donor platelet in additive solution; SDP-AS-PR, single donor platelet in additive solution pathogen reduced.
*Incidence is provided as percentage of major/bidirectional ABO incompatible HSCT.
International Forum 611
612 International Forum
complication with 24-h continuous infusion of platelets, as
previously published [14].
Most centres do not apply any special measures to
detect early immunohaematological complications in
HSCT. Only 5 centres perform DAT systematically in the
post-transplantation period: weekly (Russia); two times a
week for 3 weeks 1 time a week for 2 weeks and monthly
in ABO incompatible transplantations (Italy); at day +
100 after HSCT (Australia); once a week in minor ABO
compatible HSCT (Brazil); and unspecified in ABO minor
and mixed HSCT (Madrid).
Pure red cell aplasia (PRCA) is related to ABO major
incompatible HSCT and characterized by anaemia, reticulo-
cytopenia and the only absence or near absence of erythroid
progenitors, having excluded other causes, such as infec-
tions, haemolysis, disease relapse or drug toxicity. Patients
with PRCA can become RBC transfusion dependent for a
long period, suffering an important long-term iron over-
load, alloimmunization and transfusion reactions [15]. Pure
red cell aplasia incidence ranges from 0% (Russia) to 15%
(Switzerland) of ABO incompatible HSCT, and this variabil-
ity is also shown in the literature and could be explained by
the use of different transplantation platforms, among others
[16]. Treatment options are variable, with RBC transfusion
being the main support approach. Six centres also consider
immunosuppressive treatment, and apheresis and daratu-
mumab are proposed by one centre (Canada).
References
1 D’Souza A, Lee S, Zhu X, et al. Current use and trends in
hematopoietic cell transplantation in the United States. Biol
Blood Marrow Transplant 2017;23:1417–21.
2 Passweg JR, Baldomero H, Basak GW, et al. The EBMT activ-
ity survey report 2017: a focus on allogeneic HCT for nonma-
lignant indications and on the use of non-HCT cell therapies.
Bone Marrow Transplant 2019;54:1575–85.
3 Yuan S, Yang D, Nakamura R, et al. RBC and platelet transfu-
sion support in the first 30 and 100 days after haploidentical
hematopoietic stem cell transplantation. Transfusion
2019;59:3371–85.
4 Cohn CS. Transfusion support issues in hematopoietic stem
cell transplantation. Cancer Control 2015;22:52–9.
5 Kaufman RM, Djulbegovic B, Gernsheimer T, et al. Platelet
transfusion: a clinical practice guideline from the AABB. Ann
Inter Med 2015;162:205–13.
6 Estcourt LJ, Birchall J, Allard S, et al. Guidelines for the use
of platelet transfusions. Br J Haematol 2017;176:365–94.
7 Negrin RS. Hematopoietic support after hematopoietic cell
transplantation. UpToDate2019. Available from: https://www.
uptodate.com/contents/hematopoietic-support-after-hema
topoietic-cell-transplantation. [Last accessed 10 December
2020].
8 Booth GS, Gehrie EA, Bolan CD, et al. Clinical guide to ABO-
incompatible allogeneic stem cell transplantation. Biol Blood
Marrow Transplant 2013;19:1152–8.
9 Raimondi R, Soli M, Lamparelli T, et al. ABO-incompatible
bone marrow transplantation: a GITMO survey of current
practice in Italy and comparison with the literature. Bone
Marrow Transplant 2004;34:321–9.
10 Worel N ABO-mismatched allogeneic hematopoietic stem cell
transplantation. Transfus Med Hemother 2016;43:3–12.
11 Berger MD, Gerber B, Arn K, et al. Significant reduction of
red blood cell transfusion requirements by changing from a
double-unit to a single-unit transfusion policy in patients
receiving intensive chemotherapy or stem cell transplanta-
tion. Haematologica 2012;97:116–22.
12 Karafin MS, Blagg L, Tobian AA, et al. ABO antibody titers
are not predictive of hemolytic reactions due to plasma
incompatible platelet transfusions. Transfusion
2012;52:2087–93.
13 Josephson CD, Castillejo M, Grima K, et al. ABO-mismatched
platelet transfusions: Strategies to mitigate patient exposure
to naturally occurring hemolytic antibodies. Transfus Apher
Sci 2010;42:83–8.
14 Cid J, Guijarro F, Carbass
e G, et al. 24-h continuous infusion
of platelets for patients with platelet transfusion refractori-
ness. Br J Haematol 2018;181:386–9.
15 Bolan CD, Leitman SF, Griffith LM, et al. Delayed donor red
cell chimerism and pure red cell aplasia following major
ABO-incompatible nonmyeloablative hematopoietic stem cell
transplantation. Blood 2001;98:1687–94.
16 Hirokawa M, Fukuda T, Ohashi K, et al. PRCA Collaborative
Study Group. Efficacy and long-term outcome of treatment
for pure red cell aplasia after allogeneic stem cell transplan-
tation from major ABO-incompatible donors. Biol Blood Mar-
row Transplant 2013;19:1026–32.
Guest Editors
Dr. Pilar Solves
Hospital Universitari i Politècnic La Fe
Valencia, Spain
Email: solves_pil@gva.es
Dr. Miquel Lozano
Hospital Clı́nic Universitari
Barcelona, Spain
Email: mlozano@clinic.cat
International Forum Editor
Dr. Nancy Dunbar
Dartmouth Hitchcock Medical Center
Lebanon, NH, USA
Email: nancy.m.dunbar@hitchcock.org
© 2021 International Society of Blood Transfusion
Vox Sanguinis (2021) 116, 609–612
 Vox Sanguinis (2021) 116 e25–e43
© 2020 International Society of Blood Transfusion
INTERNATIONAL FORUM DOI: 10.1111/vox.13021

International Forum on Transfusion Practices in

Haematopoietic Stem-Cell Transplantation: Responses
Pilar Solves, Miquel Lozano, Eugene Zhiburt, Javier Anguita Velasco, Ana Maria P
erez-Corral, Silvia Monsalvo-Saornil,
Sho Yamazaki, Hitoshi Okazaki, Kathleen Selleng, Konstanze Aurich, William Kr€ uger, Andreas Buser , Andreas Holbro,
Laura Infanti, Gregor Stehle, Luca Pierelli, Antonella Matteocci, Luigi Rigacci, Karen M. K. De Vooght,
Jurgen H. E. Kuball, Katherine L. Fielding, David A. Westerman, Erica M. Wood, Claudia S. Cohn, Andrew Johnson,
Mickey B. C. Koh, Dara Qadir, Christine Cserti-Gazdewich , Etienne Daguindau, Fanny Angelot-Delettre,
Pierre Tiberghien, Silvano Wendel-Neto, Roberta-Maria Fachini, Suzy Morton, Charles Craddock, Matthew Lumley,
Jolanta Antoniewicz-Papis, Kazimierz Hałaburda, Magdalena Łez towska & Nancy Dunbar

Eugene Zhiburt - The selection of red blood cells (RBCs) according to

Russia
Question 1
250 autologous and 10 allogeneic.
Question 2
We use X-ray as the source of irradiation.
Following are the main technical parameters:
- the exposure time to ensure the absorbed dose
from 25 to 35 Gy is 45 min;
- absorbed dose rate level: (in the centre of the cam-
era and tray is 0.62 Gy/min, 2) at the upper and
lower levels of the tray 0.72 Gy/min;
- the dimensions of the tray allow irradiation at the
same time up to 3 litres of donated blood;
- the tray is oscillated at a frequency of 60 cycles
per minute at an angle of up to –15°;
- the size of the focal spot of the X-ray tube is 1.4 *
1.4 mm;
- the power of the equivalent dose of ionizing radiation
at any accessible point at a distance of 0.1 m from the
surface of the apparatus does not exceed 1.0 lSv/h.
Question 3
We perform a complete blood count every day during HSCT.
Question 4
Following is our blood transfusion protocol:
- 80 g/l is the haemoglobin threshold for red blood cell
transfusion in stable patients during the aplasia
period.
- We use 1 red blood cell unit transfusion policy.
the ABO system depends on the period of
haematopoietic stem cells (HSC) transplantation.
Three periods are distinguished: the period of prepa-
ration of the recipient for HSC transplantation (period
I), the early period after HSC transplantation until
complete donor chimerism occurs, when red blood
cells of the host and donor (period II) are determined,
and the complete engraftment of the HSC donor (pe-
riod III). In the first period, transfused RBCs that are
identical or compatible with the patient’s phenotype.
In the second period, starting from +4 day, it is nec-
essary to carry out weekly monitoring of the anti-A
and anti-B IgM class, as well as the IgG class: fixed
on the surface of the patient RBCs (in direct Coombs
test) and circulating in the serum (indirect Coombs
test). If the titre of anti-A and/ or anti-B in the pre-
transplantation period was more than 128 (dilution 1:
128), then IgM and IgG antibodies are measured twice
a week after HSC transplantation until the titre
decreases to less than 16 (dilution 1: 16), then once a
week until the antibodies completely disappear within
the next 2 weeks, except for transfusion recipients.
Period II lasts until the complete change of blood
group, the absence of antidonor antibodies in two
sequential blood tests of the recipient for 2 weeks, or
the development of transplant rejection. The end of
period II means the start of the use of RBCs only with
the phenotype of the donor. In the third period, trans-
fused RBCs that are identical in phenotype with an
HSC donor [1].
Question 5
Regarding platelet transfusion protocol:
- We use the prophylactic strategy for autologous
HSCT.

e25
e26 International Forum

- As platelet target, we consider 10 9 109/l for platelet Spain

prophylactic transfusion in patients undergoing allo-
geneic and autologous HSCT.
- We use single donor platelets, pathogen reduced
(amotosalen + UVA) in additive solution (SSP+).
Pooled random platelets are used widely in Russia
but still inaccessible for our hospital at local market
- In case of a minor ABO-incompatible platelet trans-
fusion, we do not apply any measure (volume reduc-
tion, isoaglutinins titration) due to plasma
replacement by additive solution.
- We systematically monitor the efficacy of platelet
transfusions by a post-transfusion platelet count at
20 h after transfusion.
- In case of platelet transfusion refractoriness, we man-
age it with 24-h continuous infusion of platelets [2].
Question 6
We measure the isoagglutinin titres routinely in patients after
allogeneic HSCT weekly. If the titre of anti-A and/ or anti-B
in the pre-transplantation period was more than 128 (dilution
1: 128), then the isoagglutinin titres are measured twice a
week after HSC transplantation until the titre decreases to
less than 16 (dilution 1: 16), then once a week until the anti-
bodies completely disappear within the next 2 weeks?
Question 7
See Question 4.
Question 1
We perform 81 transplants: 21 autologous, 60 allogeneic.
Question 2
For autologous: 6 months, if patient has immunologic
recovery.
For allogeneic: 1 year at least. Immunospresors have
been withdrawn and total immunologic recovery. If there
is doubt, we maintain irradiation.
• I do not know the source of irradiation.
• 25 Gy.
Question 3
Yes, we perform a complete blood count every day for
autologous and allogeneic.
Question 4
Protocol for red blood cell transfusion:
• The haemoglobin threshold is <8 g/dl.
• We use 1 unit.
• No. Donor/recipient match units are always used.
Engraftment is defined by clinical and complete ana-
lytic criteria.

Question 5
Question 8
Regarding platelet transfusion protocol:
We did not have pure red cell aplasia.
• Prophylactic.
• 10 000/µL if stable and 20 000 if bleeding risk fac-
References tors.
1 Akselrod BA, Balashova EN, Bautin AE, et al. Clinical guideli- • Indistinctly. HLA single donor to alloimmunization
nes for red blood cell transfusion. Russian Journal of Hematol- pts. Yes, the platelets are suspended.
ogy and Transfusiology (Gematologiya i Transfuziologiya). • No.
2018; 63:372-435.https://doi.org/10.25837/HAT.2019.62.39.006
• No.
2 Cid J, Guijarro F, Carbass
e G, et al. 24-h continuous infusion of
platelets for patients with platelet transfusion refractoriness. Br
• HLA match single donor.

J Haematol 2018;181:386–9. https://doi.org/10.1111/bjh.14572

Question 6
Eugene Zhiburt
No, not routinely. Abo major incompatibility. Haemolysis
Pirogov National Medical Surgical Center
suspicion. No specific time is defined.
Moscow, Russia
Email: ezhiburt@yandex.ru
Question 7
Javier Anguita Velasco, Ana Maria P
erez-Corral & Silvia
Yes. After abo minor and mixed transplants. No Hb responds
Monsalvo-Saornil
after transfusion. Haemolysis Suspicion (analytical).

© 2020 International Society of Blood Transfusion

Vox Sanguinis (2021) 116, e25–e43
 International Forum e27

Question 8
Question 5
Approximately 15%, depending on the number of ABO
We transfuse random platelets prophylactically (transfu-
transplants with major incompatibility. Periodic transfu-
sion threshold is 10 000–20 000/ll). Platelets are sus-
sion mainly. In some very complicated and individual
pended in plasma and ACD-A liquid. In case of platelet
cases, we have used rituximab and also careful review of
transfusion refractoriness, we monitor corrected count
immunosuppression.
increment (about 60 min after transfusion). Additionally,
anti-HLA antibodies and anti-HPA antibodies are mea-
Javier Anguita Velasco
Hospital General Universitario Gregorio Mara~
no
n sured.
Madrid, Spain
Email: javier.anguita@salud.madrid.org Question 6
Ana Maria P
erez-Corral No.
Hospital General Universitario Gregorio Mara~
no
n
Madrid, Spain
Email: apcorral@salud.madrid.org Question 7
No.
Silvia Monsalvo Saornil
Hospital General Universitario Gregorio Mara~
no
n
Madrid, Spain Question 8
Email: silvia.monsalvo@salud.madrid.org
About 10–15% of recipients who underwent major ABO-
mismatched HSCT developed PRCA. We continue red
blood cell transfusion until recovery.
Sho Yamazaki & Hitoshi Okazaki
Sho Yamazaki
Tokyo University of Tokyo Hospital
Tokyo, Japan
Email: shiyamazaki-ky@umin.ac.jp
Question 1
We performed 22 autologous and 33 allogeneic HSCT in Hitoshi Okazaki
University of Tokyo Hospital
2019 (approximately 50–60 per year).
Tokyo, Japan
Email: okazakih-ky@umin.ac.jp
Question 2
In Japan, all red blood cells and platelets are irradiated.
- X-ray. Kathleen Selleng, Konstanze Aurich & William Kr€
uger
- 15–50 Gy.
Germany
Question 3
Question 1
We perform a CBC count every day until engraftment in
both cases. Our centre is a University hospital including a haema-
tology department with an outpatient clinic and a spe-
cial unit for transplant patients. The collection of
Question 4
autologous and related donor stem cells is performed in
- We transfuse two units of red blood cells to maintain collaboration of the haematology clinic, the paediatric
a haemoglobin level of 7g/dl. haematology clinic and the transfusion medicine depart-
- One unit of red blood cells is prepared from 200 ml ment. Allogeneic stem cell transplants from unrelated
of whole blood in Japan. donors are imported. We perform between 20–30 autol-
- We switch to donor type blood group when anti- ogous and 20–30 allogenic stem cell transplantations
donor isoagglutinins are undetectable and RBCs of per year.
donor type are present.

© 2020 International Society of Blood Transfusion

Vox Sanguinis (2021) 116, e25–e43
e28 International Forum
Question 2
We follow the guidelines by the British committee of
standards in haematology [1]. The guidelines recommend
irradiation of cellular blood products after autologous
HSCT for 3 months and for 6 months if the patient was
treated with irradiation of his body. After allogeneic
transplantation, cellular blood products for transfusion
are irradiated 3–6 months. It depends on lymphocyte
counts and the duration of immunosuppressive therapy.
- Irradiation of RBCs and platelets is performed by cae-
sium
- 30 Gy.
Question 3
During HSCT period, the patient blood count is performed
every day until discharge. After discharge, patients are
monitored once a week until day 100 after transplanta-
tion when no complications occur.
Question 4
Transfusion thresholds for transplant patients do not differ
from other patients. Physicians follow the German guideli-
nes for hemotherapy [2]. Transfusions of RBCs are abso-
lutely indicated if the haemoglobin value falls below
3.7 mmol/l (5.7 g/dl). Between 3.7 and 6.0 mmol/l (5.6–
9.6 g/dl), RBC transfusions are indicated if the patient
shows symptoms of a non-appropriate oxygen supply, for
example tachycardia, tachypnea and angina pectoris [2].
Typically, 2 RBCs are transfused in anaemic non-bleeding
patients, but since the introduction of patient blood man-
agement measures, the number of 1-RBC transfusions
became more frequently. One reason for the 2-RBC transfu-
sion strategy is to prolong the period between transfusions,
especially for outpatients for increasing their quality of life
and to reduce the visits in the outpatient clinic.
Our transfusion strategy for patients after allogeneic
HSCT depends on the ABO type of the donor red cells
and the anti-A and anti-B antibodies in the patient
plasma. For transfusion, we choose red cells which are
compatible to the antibodies in the patient plasma, for
instance for a patient with blood type A and a transplant
of blood type B we further transfuse RBCs of blood type
A or O at least as long as anti-B antibodies in the plasma
are detectable.
After ABO-incompatible allogeneic transplantation, we
change the ABO and RhD blood type of the patient when the
majority of the circulating red cells express the blood type
of the donor and the serum group is compatible with the
donors’ red cells. We determine the red blood cell group
serologically. Every blood sample that is sent to our labora-
tory is tested for the ABD type. In case of an engraftment
failure, we will recognize changes in the expression of ABD
antigens and adapt our RBC transfusion policy individually.
Question 5
The indication for prophylactic platelet transfusions
depends on the bleeding risk assessment of the individual
patient. In our hospital, the prophylactic platelet transfu-
sion strategy is absolutely indicated in patients with pla-
telet counts lower than 10 G/L without other bleeding
risks. In the majority of cases, platelets are transfused for
treatment of bleeding. Our transfusion medicine depart-
ment provides predominantly platelet concentrates pooled
from buffy coats of 4 donors of the same blood type sus-
pended in additive solution. Platelet apheresis is per-
formed for patients with rare blood types or platelet
transfusion refractoriness because of HLA class I antibod-
ies. In these products, platelets are suspended in plasma
of the donor. No special procedures are performed to
detect or reduce isoagglutinins in the donors’ plasma. To
decrease the risk for haemolytic episodes and clinical
symptomatic transfusion reactions after minor incompati-
ble platelet transfusions, we specially consider ABO
plasma compatibility in children with low body weight
and patients with high transfusion frequency. Platelet
transfusion efficacy is monitored typically on the next
day after transfusion. If there is no appropriate increment
in the platelet count, laboratory testing for HLA class I
antibodies is initiated. When the antibody screening is
positive, only antibodies with lymphocytotoxic effects are
considered as clinically relevant and HLA class I compati-
ble donors are chosen for platelet apheresis. Other reasons
than HLA antibodies can induce platelet transfusion
refractoriness in non-bleeding patients. To distinguish
immunological from non-immunological reasons, we
transfuse 2 platelet concentrates at the same time and
measure the platelet increment 1 h after transfusion. If
there is no increment, an immunological reason is likely.
Question 6
We defined a standard for immunohematology laboratory
testing before HSCT including the patients ABO and Rhesus
type (CcDEe), K antigen, red cell antibody screening, isoag-
glutinin IgM and IgG titration performed by a microcolumn
gel centrifugation technique and a direct antiglobulin test.
After HSCT, we clearly state the transfusion strategy in the
patient database that the laboratory assistants have a
guideline for choosing blood products for this patient.
When we observe changes in the ABD antigen expression
© 2020 International Society of Blood Transfusion
Vox Sanguinis (2021) 116, e25–e43
 International Forum e29

of the patient, we repeat isoagglutinin testing of the patient Kathleen Selleng

but without titration. If ABD antigen expression has chan- Institut f€
ur Immunologie und Transfusionsmedizin
ged to the donors group and the presence of isoagglutinins Greifswald, Germany
is compatible to the donors ABO type, we change the blood Email: kathleen.selleng@med.unireifswald.de
type of the patient in the patient database.
Konstanze Aurich
Institut f€
ur Immunologie und Transfusionsmedizin
Question 7 Greifswald, Germany
Email: konstanze.aurich@med.unireifswald.de
This is already stated in answer 6 we perform a standard of
immunohematology investigations before HSCT including William Kr€uger
the direct antiglobulin test. The result before HSCT is the Klinik f€
ur H€amatologie und Onkologie
benchmark for results after HSCT but we do not perform Greifswald, Germany
the antiglobulin test routinely after HSCT in a defined fre- Email: william.krueger@med.unireifswald.de
quency. As long as the red cell antibody screen and cross-
matches are negative, we do not repeat the direct antiglob-
ulin test except there is a clinical suspicion for haemolysis. Andreas Buser, Andreas Holbro, Laura Infanti & Gregor Stehle
If the antibody screen is positive, all further investigations
including the direct antiglobulin test and elution proce-
dures of the autologous red cells are directed to detect or Switzerland
exclude clinically relevant alloantibodies. But this is the
same for all patients not only after HSCT. Question 1
100 allogeneic, 45 autologous.
Question 8
The incidence of pure red cell anaemia (PRCA) is very low Question 2
with less than one case per year in our institution. When Autologous: principally for 1 year, in clinical practice
the diagnosis of a PRCA is suspected, a complete haemato- lifelong.
logical work-up is carried out. It includes the physical Allogeneic: as long as there are signs of GvHD, or if
examination, peripheral blood smear, reticulocyte count, patients are under immunosuppressive treatment, in clini-
iron parameters, vitamin B12 and folic acid levels and a cal practice lifelong.
metabolic panel. Additionally, antibodies against viruses
are determined to exclude viral infections associated with • Caesium 137.
anaemia. A bone marrow examination is mandatory with • 30 Gy.
smear, histology, cytogenetics and molecular genetics to
exclude a myelodysplastic syndrome. Therapeutic options
include red cell transfusions, steroids and immunosuppres- Question 3
sive agents, when necessary. The choice of therapy in Complete blood count is performed every day, once a
absence of associated causes depends upon severity of the week there is also a microscopic blood film analysis, in
disease and patients’ age and constitution. Allografts for certain circumstances additional blood smears at the dis-
PRCA were not used in our institution so far. cretion of the treating physician.

References
1 Treleaven J, Gennery A, Marsh J, et al. Guidelines on the use
of irradiated blood components prepared by the British Com-
mittee for Standards in Haematology blood transfusion task
force. Br J Haematol 2011;152:35–51
2 Bundes€arztekammer. Querschnitts-Leitlinien zur Therapie mit
€
Blutkomponenten und Plasmaderivaten Deutscher ArzteVerlag;
2009. Available from: https://www.bundesaerztekammer.de/
fileadmin/user_upload/downloads/QLL_Haemotherapie_2014.
pdf.
Question 4
Transfusion protocol:
- 70 g/l.
- 1 RBC unit.
- We change RBC blood groups to donor type usually
after 6 months. AB0 antigens must be donor type, and
also RH, K, JK, FY and MNS (extended phenotype). The
DAT should be negative. Criteria for RBC engraftment

© 2020 International Society of Blood Transfusion

Vox Sanguinis (2021) 116, e25–e43
e30 International Forum
are an absolute reticulocyte count of >30 9G/l and the
absence of RBC transfusion requirements. The detec-
tion of donor AB0 blood group is not part of the crite-
ria. (Neutrophil engraftment is defined as the first of
three consecutive days of an absolute ANC count of
>0.5 G/l; PLT engraftment is defined as PLT > 20 G/l
without transfusions in the last 7 days).
Question 5
Regarding platelet transfusion protocol:
- Prophylactic.
- 10 G/l in stable clinical situation, 20 G/l when there
are risk factors for bleeding such as fever, aGvHD,
30 G/l in case of systemic anticoagulation due to
thrombosis.
- We use both single donor and buffy coat pooled PLT,
all are pathogen reduced using the Intercept Blood
system from CERUS (since 2011) all are in platelet
additive solution (Intersol). No, considering also the
suspension in platelet additive solutions (PAS).
- Yes. 15–60 min post-transfusion.
- Assess if non-immunological reason for refractoriness,
check for HLA –Antibodies (Luminex). If HLA –Ab are
present, we try to prevent transfusion of single donor
PLT bearing the relevant antigens, in special cases we
transfuse HLA-matched platelet concentrates.
- In case of no success, we use PLT drip infusion (de-
scribed by J.Cid and M. Lozano): [1] adult doses are
divided into 4 split doses, and each dose will be con-
tinuously infused over a 4–5 h period of time in
order to have a constant (small) amount of PLT in
the circulation supporting endothelial integrity.
Question 6
Yes.
- Before HSCT, at day 30 and at day 180.
- More frequently in case of pure red cell aplasia or
relapse.
Question 7
Yes, we use DAT and Elution techniques (incl. elution
against A and/or B panel cells).
Question 8
The PRCA incidence is about 20% of the HSCT with
major or bidirectional AB0 incompatibility (about 3% of
all allogeneic HSCT).
Management information:
(i) red blood cell transfusion (Hb threshold 70 g/l).
(ii) we try to taper immunosuppression as fast as possi-
ble.
(iii) daratumumab and/or rituximab in cases not
resolving spontaneously or not responding to the
reduction of immunosuppression.
Reference
1 Cid J, Guijarro F, Carbass
e G, et al. 24-h continuous infusion
of platelets for patients with platelet transfusion refractori-
ness. Br J Haematol 2018;181:386–9.
Andreas Buser
Regional Blood Transfusion Service, Swiss Red Cross
Basel, Switzerland
Email: andreas.buser@usb.ch
Andreas Holbro
Regional Blood Transfusion Service, Swiss Red Cross
Basel, Switzerland
Email: andreas.holbro@usb.ch
Gregor Stehle
Regional Blood Transfusion Service, Swiss Red Cross
Basel, Switzerland
Email: gregor.stehle@usb.ch
Luca Pierelli, Antonella Matteocci & Luigi Rigacci
Italy
Question 1
We perform 20 autologous and 15 allogeneic haemopoi-
etic transplantations per year.
Question 2
We irradiate all blood products for autologous and allo-
geneic recipients except platelet concentrates (both from
single donors and from pooled buffy coats) for which
irradiation has been replaced by Amotosalen/UV patho-
gen inactivation.
Irradiation is applied using a dedicated Caesium-137
source and the energy of 2500 cGy.
© 2020 International Society of Blood Transfusion
Vox Sanguinis (2021) 116, e25–e43
 International Forum e31

Question 3
Question 7
All patients, after autologous or allogeneic transplanta-
The basic immunohematologic monitoring of ABO-in-
tion, have a daily complete blood count and leucocyte
compatible recipients includes also direct/indirect
differential till full haemopoietic recovery.
antiglobulin tests and isoagglutinin titration. The direct
antiglobulin test is performed 2 times a week for 3 weeks,
Question 4 1 time a week for 2 weeks and then monthly until the
DAT will be negative. We also test LDH and haptoglobin.
The haemoglobin threshold to perform red blood cell trans-
fusion in stable allografted or autografted patients is 80 g/
l. In routine practice, we use a single-unit red blood cell Question 8
transfusion policy. In ABO-incompatible transplants, we
In our practice, the incidence of pure red cell aplasia in ABO-
change the red blood cell group of RBC concentrates to
incompatible haemopoietic transplantation is around 5%. In
donor cell group when stable full engraftment is obtained
general, patients are supported with carefully cross-matched
and monitored: engraftment is in general defined by
red cell transfusions and corticosteroids. During treatment
molecular full-donor chimerism and, for red blood cells, by
and support, immunohematologic monitoring includes
globular/plasmatic tests (including titrations) combined
direct/indirect antiglobulin tests, serological and genomic
with negative results of direct/indirect antiglobulin tests
blood group typing as well as isoagglutinin titration.
and, finally, with the ABO genotyping confirmation.

Luca Pierelli
Question 5
Sapienza University-San Camillo Forlanini Hospital
We use a prophylactic strategy for platelet transfusion. In Rome, Italy
stable patients, the prophylactic threshold is 10 9109/L in Email: luca.pierelli@uniroma1.it
both autologous and allogeneic transplantation. We trans-
Antonella Matteocci
fuse both single donor and pooled buffy-coat platelet con-
San Camillo Forlanini Hospital
centrates. Our platelet concentrates are produced in platelet
Rome, Italy
additive solution with a poor plasma content. In any case, Email: matteocci31@gmail.com
in most of our transfusion episodes, we release ABO identi-
cal/compatible platelet concentrates for all patients, includ- Luigi Rigacci
ing haemopoietic transplant recipients. We monitor platelet San Camillo Forlanini Hospital
transfusion efficacy with a complete blood count the morn- Rome, Italy
ing after transfusion. We do not perform routinely blood Email: lrigacci@scamilloforlanini.rm.it
counts 15–60 min after transfusion.
In recipients with serologically detected immune plate-
let refractoriness by ELISA (anti-HPA platelet antibodies) Karen M. K. De Vooght & Jurgen H. E. Kuball
and Luminex (anti-HLA platelet antibodies) tests, if the
patient is bleeding and the transfusion is urgent, we
assign fresh pooled buffy-coat platelet concentrates, and The Netherlands
otherwise, we perform platelet cross-match with SPRCA
technology on ABO, rarely HPA/HLA, compatible single Question 1
donor concentrates. For these patients, we monitor rou-
In our academic hospital, we perform about 60–80 autol-
tinely blood counts 60 min after transfusion.
ogous and 60–70 allogeneic HSCT’s in adults per year.

Question 6 Question 2
We routinely monitor isoagglutinin titres during ABO-in-
After an autologous HSCT, patients have an indication for
compatible haemopoietic transplantation. Isoagglutinin
irradiated blood products for a period of 1 year. Patients
titres are performed prior transplantation, before and after
undergoing an allogenic HSCT receive irradiated blood
plasma-exchange in case the recipient undergo isoagglu-
products 4 weeks before until 5 years after transplanta-
tinin removal. After transplantation, isoagglutinin titres
tion. In our country, institutes normally not irradiate
are performed every week until engraftment and then every
products themselves. This is performed by Sanquin, the
15 days, with ABO blood group checked every month.

© 2020 International Society of Blood Transfusion

Vox Sanguinis (2021) 116, e25–e43
e32 International Forum
only institute (foundation) in the Netherlands allowed to
and responsible for the blood supply.
-X-ray is used as source of irradiation
-at an energy dose of 25–50 Gy.
Question 3
We perform a complete blood count in autologous and
allogeneic HSCT every 2nd day, until repopulation. In
case of pre-treatment with anti-thymocyte globulin
(ATG), a complete blood count is performed daily.
Question 4
In stable patients, an haemoglobin threshold of 8 g/dl is
used for red blood cell transfusion. We use an one red
blood cell transfusion policy, in the outpatient ward occa-
sionally a two red blood cell unit transfusion policy.
Patients receiving a non ABO identical transplant, receive
both patient and donor ABO compatible red cell units start-
ing from two weeks before HSCT. We do not change to
donor ABO compatible units after engraftment. However,
we do change the patient’s ABO blood group to the donor
blood group, once we are sure engraftment is reached. For
that, the patient must be transfusion independent for three
months and his/her blood group must match the donor
ABO blood group, based on antigen typing, without dis-
crepancies. For treatment purposes, engraftment is deter-
mined by white cell donor chimerism detection.
Question 5
In our institution, a prophylactic platelet transfusion proto-
col is used in case of an autologous HSCT. The prophylactic
platelet transfusion trigger is >10 9 109/l both in patients
undergoing allogeneic as autologous HSCT. The standard
platelet product for this purpose is a 5-donor random plate-
let product containing PAS-E/plasma. In case of transfu-
sion refractoriness with proven HLA class 1 antibodies,
there is an indication for an HLA-A and HLA-B typed
(identical) single donor platelet product, again suspended
in PAS-E/plasma. Because we use platelets suspended in
PAS-E/plasma, iso-agglutinins concentrations are lowered.
Therefore, we don’t apply any measure in case of a minor
ABO-incompatible platelet transfusion. We systematically
monitor the efficacy of platelet transfusions in patients
undergoing HSCT by a post-transfusion platelet count
about an hour after finishing the transfusion. In case of
transfusion refractoriness, we first check for the presence
of HLA class 1 antibodies (by use of a Luminex technique).
In case they are present, the patient has an indication for
HLA-A and HLA-B typed (identical) single donor platelet
products. Waiting for the result of the HLA class 1 antibody
screening, and in case of bleeding, the patient will receive
HLA-A and HLA-B typed (identical) single donor platelet
product(s) on a trial basis. In case of severe platelet refrac-
toriness and absence of HLA class 1 antibodies, we screen
for HPA antibodies (following an HPA typed single donor
platelet product transfusion).
Question 6
Isoagglutinin titres are not determined routinely in patients
after allogeneic HSCT. They do are determined before HCST
in case of major ABO blood group antagonism. Further-
more, we perform ABO typing (antigen and antibody) after
HSCT on a regular basis, and anti-A and/or anti-B aggluti-
nation strengths are sometimes used to evaluate aplastic
anaemia due to major ABO blood group antagonism.
Question 7
After HSCT, ABO blood group typing (antigen and anti-
body) and an irregular antibody screen is performed on a
regular basis. At first every three days, thereafter with
each visit to the outpatient department, until engraftment
is reached. The direct antiglobulin test is not performed
routinely, however, is done whenever there are discrepan-
cies in the ABO blood group typing or when the irregular
antibody screening is positive. Of course, the direct
antiglobulin test is also performed in case of clinical signs
of (auto-immune) haemolytic anaemia.
Question 8
We see about three cases of pure red cell aplasia in HSCT
patients per year. In case of major ABO blood group, antago-
nism anti-A and/or anti-B agglutination strengths can be
used for monitoring. Patients receive red blood cell and plate-
let transfusions when needed. Further treatments of AA are in
line with the Dutch guideline for AA (ATGAM, allo SCT).
Karen M. K. De Vooght
University Medical Center Utrecht
Utrecht, the Netherlands
Email: k.devooght@umcutrecht.nl
Jurgen H. E. Kuball
University Medical Center Utrecht
Utrecht, the Netherlands
Email: J.H.E.Kuball@umcutrecht.nl
Katherine L. Fielding, David A. Westerman & Erica M. Wood
© 2020 International Society of Blood Transfusion
Vox Sanguinis (2021) 116, e25–e43

Australia
Question 1
120 autografts and 100 allografts for adult patients.
Question 2
Irradiated cellular blood products are provided indefi-
nitely after both autologous and allogeneic HSCT.
Gamma irradiation is performed by the Australian Red
Cross Lifeblood prior to product issue to hospitals. This is
undertaken according to Australian and New Zealand Soci-
ety of Blood Transfusion ‘Guidelines for Prevention of
Transfusion-Associated Graft-Versus-Host Disease’, 2011.
A minimum 25 Gy with no part of the component
receiving more than 50 Gy. [1]
Question 3
Full blood counts are performed every day after both allo-
geneic and autologous HSCT until count recovery/discharge.
Question 4
The typical threshold for red cell transfusion in a stable
post-transplant patient is Hb < 80 g/l. We note the recent
publication by Tay and colleagues reporting similar qual-
ity of life and other outcomes in HSCT patients using
restrictive (Hb 70 g/l) compared with liberal (Hb 90 g/l)
thresholds for RBC transfusion [2].
If red cell transfusion is required on the combination
of clinical features and haemoglobin result, 1–2 units of
RBCs are usually transfused.
Allogeneic HSCT recipients requiring RBC transfusion
receive RBCS which are compatible with both donor and
recipient blood groups until engraftment. Following engraft-
ment, according to our protocol, the patient’s blood group is
changed to donor type in the laboratory information system
when, after 3 months without RBC transfusion, testing is per-
formed that shows their forward group is consistent with
donor group with no mixed field, the reverse group demon-
strates no anti-donor isoagglutinins, the DAT is negative,
and the RBC phenotype is consistent with donor phenotype.
From this point, if the patient requires further transfusions,
this will be with RBCs of donor blood type.
Question 5
A prophylactic platelet transfusion strategy (i.e. transfusion
in the setting of thrombocytopenia but without bleeding, in
© 2020 International Society of Blood Transfusion
Vox Sanguinis (2021) 116, e25–e43
International Forum e33
an effort to prevent bleeding) is still routine. A few consul-
tants (10%) follow the TOPPS trial findings [3] and transfuse
only for clinically significant bleeding rather than using a
prophylactic strategy in the stable autograft cohort.
For both autograft and allograft, a platelet threshold of
<10 9 109/L routinely is used, or <20 9 109/L if febrile
and/or bleeding.
The majority (~70% of units transfused) of platelets
supplied by our local blood service are derived from
pooled random platelets (pool of 4) suspended in additive
solution. Single donor/apheresis units are provided in
~30% cases and on request, for example platelet transfu-
sion reactions or poor incrementation. One to three adult
doses of platelets may be prepared from a single apheresis
platelet donation. Since March 2019, these platelets are
suspended in an additive solution.
No measure is applied for a minor ABO-incompatible
platelet transfusion.
The efficacy of platelet transfusions is not routinely
measured by a post-platelet transfusion increment (PPI),
however if there are any concerns regarding the daily
platelet counts (e.g. persisting low counts after prior
transfusion such that the patient is requiring very fre-
quent transfusions), we would perform platelet increment
testing at 30–60 min post-transfusion.
For platelet transfusion refractoriness, post-platelet
increment measurements are undertaken to demonstrate
potential refractoriness. Subsequently, a trial of single
donor/apheresis platelets with increment testing is per-
formed, and if there are continued concerns about refrac-
toriness, then anti-HLA/HPA antibodies are tested. If
HLA-antibodies are demonstrated, then HLA-matched pla-
telets are trialed and further PPI testing undertaken.
Question 6
Isoagglutinin titres are routinely measured pre-transplant,
as well as at day 100, and other time points (if clinical
concerns) for ABO-mismatched HSCT recipients.
Question 7
Red blood cells phenotype and DAT are performed rou-
tinely at day 100 for all allogeneic HSCT recipients. They
are also performed at other time points as needed,
according to risk assessment and clinical requirements.
Question 8
Based on a definition of complete absence of erythroid pre-
cursors in the bone marrow in the presence of donor
engraftment (as determined by myeloid chimerism) and
e34 International Forum
persistence of granulopoiesis and thrombopoiesis, our inci-
dence of post-transplant red cell aplasia is approximately
1–3%.
Our general approach to treatment is supportive care
with transfusions and EPO and waiting for resolution.
Acknowledgements
We acknowledge valuable contributions from Mr Michael
Haeusler, Dr Giles Kelsey and Professor David Ritchie to
preparation of these answers.
References
1 Australian and New Zealand Society of Blood Transfusion.
“Guidelines for prevention of transfusion-associated graft-
versus-host disease”, 2011. Available at https://anzsbt.org.au/
wp-content/uploads/2018/06/PreventionofTA-GVHD.pdf.
2 Tay J, Allan DS, Chatelain E, et al. Liberal versus restrictive
red blood cell transfusion thresholds in hematopoietic cell
transplantation: A randomized, open label, phase III, noninfe-
riority trial. J Clin Oncol. 2020: JCO1901836. https://doi.org/
10.1200/JCO.19.01836. [Epub ahead of print].
3 Stanworth SJ, Estcourt LJ, Powter G, et al. A no-prophylaxis
platelet-transfusion strategy for hematologic cancers. N Engl
J Med. 2013;368:1771–80.
Katherine L. Fielding
Royal Melbourne Hospital and Peter MacCallum Cancer Centre
Parkville, Victoria, Australia
Email: kate.fielding@petermac.org
David A. Westerman
Peter MacCallum Cancer Centre
Parkville, Victoria, Australia
Email: david.westerman@petermac.org
Email: david.ritchie@mh.org.au
Erica M. Wood
Peter MacCallum Cancer Centre
Parkville, Victoria,
Email: erica.wood@monash.edu
Claudia S. Cohn & Andrew Johnson
United States
Question 1
We performed 110 autologous HSCT and 125 allogeneic
HSCT in the most recent calendar year. The allogeneic
transplants could be further broken down to include 54
marrow transplants, 44 cord blood transplants and 27
peripheral blood stem cell transplants.
Question 2
We continue to irradiate cellular blood components for one
year after an autologous stem cell transplant. After an allo-
geneic transplant, we continue to irradiate for the life of the
patient.
- X-ray.
- 25 Gy.
Question 3
Yes, a CBC is completed each day for an inpatient under-
going HSCT. Many autologous HSCT are done as outpa-
tients.
Question 4
Regarding the red blood cell transfusion protocol:
- 8 g/dl.
- 1 RBC.
- We do change the blood group of the patient
based on the following: (1) 120 days without a RBC
transfusion; (2) molecular studies showing full
engraftment; (3) back type is compatible with the
front type.
Question 5
Regarding platelet transfusion protocol:
- Prophylactic.
- 10 000 plts/µl.
- Single donor. We use both plasma and additive solu-
tion interchangeably.
- We have a 1000 ml limit (recently raised to 1500 ml,
but not fully implemented) per week. If we reach the
incompatible plasma limit, we perform a DAT. If DAT
is negative, the medical director may allow more
incompatible plasma. If DAT is positive or medical
director does not waive, then we volume reduce
incompatible units.
- We try but often the nurses do not draw the post-
count. Ideally, it would be drawn 10–60’ after each
transfusion.
- First, we try to rule out non-immune causes for refrac-
toriness. If we cannot find other reasons for refractory
state, then we order a cross-match assay. If approxi-
mately >1/3 are incompatible, we order compatible
units and monitor the next three transfusions closely
(if <1/3 incompatible we stay with random units.) to
© 2020 International Society of Blood Transfusion
Vox Sanguinis (2021) 116, e25–e43
 International Forum e35

see if CCI improves and is >5000. If either trial of ran-

dom or of XM-compatible fails, then we test for HLA United Kingdom–London
antibodies against class I HLA. We then talk to our
local blood centre to find either 4/4 direct HLA match
or we try to avoid antibodies. When we get HLA- Question 1
matched (or antibody avoidance) units, we do a three Our centre performs 30 autologous and 30 allogeneic.
transfusion trial. If this also does not increase the CCI
>5000, we return to using random units.
Question 2

Question 6 1 year after autologous and lifelong for allogeneic

-X-RAY.
We do not measure isoagglutinin titres. We look at the -25 Gy.
back type and grade the strength of the reaction 0-4 and
follow trends.
Question 3

Question 7 Daily FBC is routine for both autologous and allogeneic

We perform a DAT after a transfusion reaction (which are
frequently reported in this population due to neutropenic
fevers) and if we suspect antibody-mediated haemolysis
after a drop in haemoglobin.
Question 8
I do not know the exact incidence, but it is rare. Most
frequently (and classically), we see PRCA after a group O
patient receives a group A transplant. Initially, PRCA is
managed by the clinicians conservatively, with the expec-
tation that it will ‘burn itself out’. The clinician might try
a trial of steroids or rituximab. If the patient has an
extended course, they will come to transfusion medicine
for apheresis to reduce the titre of isohemagglutinins. We
have done this aggressively on some patients and it has
always worked.
Claudia S. Cohn
University of Minnesota
Minneapolis, MN, USA
Email: cscohn@umn.edu
Andrew Johnson
University of Minnesota
Minneapolis, MN, USA
Email: john4613@umn.edu
Mickey B. C. Koh & Dara Qadir
transplants.
Question 4
Regarding the red blood cell transfusion protocol:
- 70 g/l.
- 1 unit.
- This statement is confusing. Is this asking for engraft-
ment of the stem cells (I assume not) or whether this is
full donor blood group change (we do not term this
engraftment). In this case, engraftment is by full
grouping and antibody screen which will include a
antiglobulin test.)
Question 5
Regarding platelet transfusion protocol:
- Prophylactic.
- In both 109 109/l unless other circumstances (bleed-
ing, fever, on other anti-coagulants, etc).
- Single donor in additive solution.
- None.
- Not routinely unless platelet refractory. Then, a count
check is done 60 min post-transfusion.
- HLA-matched platelets.
Question 6
Only prior to transplant and before stem cell infusion.
Not routinely after except once every year to detect blood
group change.

© 2020 International Society of Blood Transfusion

Vox Sanguinis (2021) 116, e25–e43
e36 International Forum
Question 7
Generally not unless there is a clinical indication – high
bilirubin, delayed engraftment.
Question 8
Transient hypoplasia post-major ABO-incompatible trans-
plant more common (about 10%). Actual red cell aplasia
much less common. Managed with transfusion support
until the aplasia/hypoplasia recovers.
Mickey B. C. Koh
St George’s University Hospital
London, UK
Email: mickey.koh@stgeorges.nhs.uk
Dara Qadir
St George’s University Hospital
London, UK
Email: Dara.qadir@stgeorges.nhs.uk
Christine Cserti-Gazdewich
Canada
Question 1
The University Health Network performs HSCT at Princess
Margaret Cancer Centre, one of its three main (acute care)
hospitals. Annually, 350 autologous and 190 allogeneic
HSCTs are performed (540).
Question 2
Though we are in medical agreement with the principles
of Canada’s National Advisory Committee on Blood and
Blood Products (https://www.nacblood.ca/resources/guide
lines/downloads/Recommendations_Irradiated_Blood_Com
ponents.pdf_), which specify 3–6 months after autologous
HSCT and for as long as immune suppressant therapies
and/or (chronic) graft versus host disease may be present
after autologous HSCT (with annual reviewer thereof). At
our institution, the instruction to irradiate invariably
remains in place indefinitely for the following reasons:
(a) Disease relapses or de novo indications, which
themselves are indications for irradiation, have
arisen after the NAC term conditions or intervals
have elapsed, and
(b) it has not been feasible to expeditiously obtain suf-
ficiently indubitable case details on the NAC term
conditions or intervals, as patients may have health
information in more locations than we are able to
access, interpret and verify
We use Cesium-137 as our irradiation source at a mini-
mum central dose of 25 Gray.
Question 3
A complete blood count occured daily during HSCT
admissions, with the exception of reduced-frequency test-
ing scenarios (to every other day, in lower-volume tubes),
in patients with stronger blood conservation mandates
(i.e. transfusions relatively or absolutely contraindicated).
More recently, with the shift of autologous SCT to the
outpatient setting, CBC frequencies have decreased. At
least one count check is prescribed in the first week post-
discharge, and patients are seen 3–10 times in the first
30 days after discharge with laboratory testing at those
times.
Question 4
The haemoglobin threshold for red blood cell transfusion
in stable, non-bleeding inpatients during the aplasia per-
iod is single unit red blood cell orders at ≤70–75 g/l.
We are able to change the red blood cell group of RBC
concentrates to the donor’s group when engraftment
occurs. This occurs when the forward group reflects
exclusively donor erythropoiesis, without confounding by
transfusion, and by demonstration of the absence of anti-
graft isoagglutinin activity (on reverse type or by DAT).
However, we tend not to encode the group change for a
variety of reasons:
(a) Grouping samples in the post-engraftment period,
after sufficiently lengthy transfusion-free periods,
may not be drawn (due to the associated absence of
need for transfusion); and
(b) Experience with a number of ‘impermanence anec-
dotes’ (i.e. disease relapses and/or evidence of
delayed recurrences of host erythropoiesis), perhaps
as a function of the increased proportion of reduced
intensity/non-myeloablative conditioning regimens,
with reversion to the hybrid type designation and
permissible component-specific ABO exposure truth
tables.
Question 5
In autologous HSCT, the platelet transfusion protocol is
permissive on the use of prophylactic platelet transfu-
sions, that is it is not restricted to treatment only. The
Platelet Transfusion Requirements in Hematopoietic
Transplantation (PATH) trial (https://clinicaltrials.gov/ct2/
© 2020 International Society of Blood Transfusion
Vox Sanguinis (2021) 116, e25–e43

show/NCT02650791), which is investigating outcomes by
either strategy, is active at our centre. Both approaches
can be seen in audits by virtue of enrolments or clinician
equipoise (with the institutional policy permitting pro-
phylaxis at platelet counts of ≤10 for auto- or allo-HSCT).
In times of (and in accordance with extent of) platelet
shortage, stricter triggering is enforced.
Our adult dose platelet concentrates are predominantly
buffy coat pools (four whole blood donor derived) with
single donor apheresis platelets. The ratio delivered to our
centre by Canadian Blood Services is presently 10:1,
which has been rising from previous ratios (3:1) over the
last several years [1].
Apheresis platelet stocks are enriched for those that
have been selected by HLA- (and/or HPA-)-type in the
care of sensitized and otherwise platelet transfusion
refractory patients.
Our platelet concentrates are suspended in the plasma
of a contributing male donor in pooled concentrates.
In the case of minor ABO-incompatible platelet trans-
fusion, a room temperature (immediate spin) 1:50 titre is
performed on group O platelets if this is the only avail-
able type left in inventory for non-O recipients. The pro-
duct is unmodified if titre negative. If titre positive, the
platelet undergoes a plasma volume reduction (PVR).
We advise monitoring of post-transfusion platelet
counts (15–60 min post) in ‘high-use alert cases’ (three
consecutive days with ≥1 platelet transfusion per day), or
in the evaluation of immune-selected single donor
apheresis platelets (to assess the in vivo efficacy of prod-
ucts which had been selected to circumvent the last
known array of HLA- and/or HPA-specific antibodies).
Our hospital transfusion laboratory receives its testing
and product support through Canadian Blood Services,
with a local (Toronto-based) HLA Histocompatibility labo-
ratory, a national (Winnipeg-based) Platelet and Leuko-
cyte Reference laboratory, and the associated
apheresis/donor management programmes.
Platelet transfusion refractoriness is managed by iden-
tifying suspicious cases (with the aforementioned [auto-
mated] high-use alert placed in the patient’s electronic
medical record), and a clinical policy which supports
immune refractoriness testing and the coordinated provi-
sion of selected platelets. Suitable options for the latter
reflect a hierarchy of potentially available ‘matches,’ (i.e.
HLA-identical, HLA-‘similar’ [software-modelled/struc-
turally-predicted tolerance], HLA-‘negative’ [with respect
to the patient’s last-determined specificities] or ‘low MFI’
[HLA-mismatched with respect to the lowest mean fluo-
rescence intensity antibody specificities on PRA]).
Due to the preparation time entailed in sourcing and
delivering preferred platelets (especially for the lower fre-
quency profiles commanded by certain cases), the
© 2020 International Society of Blood Transfusion
Vox Sanguinis (2021) 116, e25–e43
International Forum e37
anticipated needs are reported a week in advance by the
most responsible physician. An active list of sensitized
patients with platelet transfusion needs is coordinated in
the blood transfusion laboratory and amended with clini-
cian input and pertinent emerging laboratory results.
For the latter, serologic profiles predicting for refrac-
toriness come to the transfusion physician’s attention for
clinical review. Retesting the antibody repertoire is
encouraged quarterly and/or when patients exhibit refrac-
toriness to matched products (should this be sooner [or
later/after missed rescreening opportunities]).
Question 6
We do not routinely measure isoagglutinin titres routinely
in patients after allogeneic HSCT, though we provide this
service on request. Requests are usually limited to those
cases with suspected pure red cell aplasia/delayed red cell
engraftment from ABO major mismatch.
Question 7
The transfusion laboratory does not prescribe or enforce a
prospective schedule of monitoring by direct antiglobulin
tests or other measures to detect early immunohemato-
logic complications after ABO-incompatible HSCT. How-
ever, the transplant teams consult closely with the
transfusion laboratory medical directors to obtain inves-
tigative advice and interpretation.
Question 8
We do not have information on our incidence of pure red
cell aplasia. Management is on a case-by-case basis, from
antigen-negative chronic transfusion support to apheresis,
intensified immune suppressant therapy and off-label
daratumumab in rare instances.
Reference
1 Cserti-Gazdewich C, Escorcia A, Pendergrast J, et al. Platelet
transfusion reactions do not occur more often in recipients
transfused with apheresis versus buffy coat platelet concen-
trates. Transfusion 2016;56:3144–6.
Christine Cserti-Gazdewich
University Health Network/University of Toronto
Toronto, Ontario, Canada
Email: Christine.Cserti@uhn.ca
Etienne Daguindau, Fanny Angelot-Delettre & Pierre Tiberghien
e38 International Forum
France
Question 1
In 2019, the Besancon University Hospital transplantation
unit performed 32 autologous and 55 allogeneic
hematopoietic cells transplantation, figures similar to pre-
vious years regarding allogeneic transplantation but num-
ber of auto HSCT is slightly decreasing.
Question 2
Up to 3 months (1 year in case of total body irradiation)
after autologous HSCT and 1 year after allogeneic HSCT,
with the exception of patients per.
The platelets units are no longer irradiated since they
undergo amotosalen plus UVA illumination treatment.
- X-ray.
- As recommended by our National Health Authorities,
the energy applied to the unit has to be between 25
and 45 Gy. In our institution, the device is set up for
35 Gy.
Question 3
A complete blood count every day during HSCT along
with a white blood cell differential twice a week, and
not differently between autologous or allogeneic trans-
plantation.
Question 4
Regarding the red blood cell transfusion protocol:
- The national recommendation for transfusion in stem
cell transplantation is to keep a level of haemoglobin
superior to 7 g/dl [1]. In fact, 70% of the French cen-
tres participating to the survey published in 2019 [1],
including ours, use a 8 g/dl threshold for red blood
cell transfusion. This cut-off is adapted to the clinical
tolerance of anaemia.
- It is recommended to use one rather than 2 RBC for
transfusion [1]. In practice, more than 90% of RBC
transfusion include 2 units because of the convenient
set-up of such a policy (one appointment rather than
2 for outpatients, reduction of nursing care plan. . .)
- Starting the day of transplantation, RBC concen-
trates are ABO compatible with both the donor and
recipient ABO groups. In case of Rh and Kell
incompatibility between donor and recipient, donor
compatibility is privileged. We provide to the recipi-
ent a document containing a personalized ‘transfu-
sion policy’. After 3 months, this policy is adapted
to the engraftment only for platelet transfusion. If
engraftment confirmed, dual (prior) recipient and
donor ABO compatibility is no longer required while
the recipient remains transfused with donor-compat-
ible platelets units. The engraftment is determined
by both forward (cellular) and reverse (plasmatic)
group typing.
Question 5
Regarding platelet transfusion protocol:
- We use prophylactic strategy for autologous HSCT.
- Our threshold is 20 G/l. The recent national survey
shown that one third of participating centres have a
threshold at 10 G/l for platelet transfusion and two-
third at 20 G/l. The national recommendation for
platelet transfusion [2] determines different thresh-
olds considering risks factor for bleeding: 10 G/l for
patients without risk factors and 20 G/l for patients
with fever, infection, mucositis, arterial hypertension
or fast decrease of platelet level in the last 72 h.
Most of the patients with hematopoietic cell trans-
plantation are in the second group, explaining our
policy.
- Approximately two thirds of platelets units trans-
fused are random platelets. The platelets are sus-
pended in additive solution.
- In case of minor ABO-incompatible platelets transfu-
sion, we transfuse platelets with low titre A/B Ab.
- We rarely monitor the efficacy of platelets transfu-
sions immediately after transfusion. However, platelet
counts are verified the day after (daily CBC count).
Earlier monitoring (60 min) may occur in cases of
(anti-HLA or anti-HPA) alloimmunization and after
transfusion of compatible platelets units.
- We verify HLA or HPA alloimmunization, if so com-
patible units are selected for further transfusions. In
the absence of haemorrhagic features, we change our
policy from ‘prophylactic’ to ‘curative’.
Question 6
In case of ABO mismatch transplantation, isoagglutinin
titre is systematically measured before transplantation and
one month after transplantation. After this delay, isoagglu-
tinin titre is measured only in case of transfusion depen-
dency (pure red cell aplasia). In other cases (ABO
compatibility), isoagglutinin titre is not routinely assessed.
Question 7
We do not perform direct antiglobulin test after ABO-in-
compatible HSCT. We do, however, undertake graft
© 2020 International Society of Blood Transfusion
Vox Sanguinis (2021) 116, e25–e43

plasma depletion or RBC depletion (in case of minor or
major ABO mismatch, respectively).
Question 8
The incidence of pure red cell aplasia (PRCA) after allo-
geneic HSCT in our centre is around 5%. If PRCA persists
for more than 6 months, we consider the use of rituximab
or more recently daratumumab. Otherwise, we wait for
spontaneous improvement.
References
1 Giraud C, Thibert JB, Desbrosses Y, et al. Transfusion in
autologous and allogenic hematopoietic stem cell transplant:
Guidelines from the Francophone Society of Bone Marrow
Transplantation and Cellular Therapy (SFGM-TC). Bull Cancer
2019;106:S52–s8.
2 Transfusion de plaquettes: produits, indications. Recommanda-
tions de la Haute Autorit
e de Sant
e et de l’AgenceNationale
de S
ecurit
e du M
edicament. 2015:(accessed 30/04/20).https://
www.has-sante.fr/upload/docs/application/pdf/2015-11/rec
ommandations_-_transfusion_de_plaquettes.pdf
Etienne Daguindau
University Hospital of Besancon
Besancon, France
Email: edaguindau@chuesancon.fr
Fanny Angelot-Delettre
Etablissement Francais du Sang Bourgogne Franche-Comt
e
Besancon, France
Email: fanny.delettre@efs.sante.fr
Pierre Tiberghien
Etablissement Francais du Sang
La Plaine St-Denis, France
Email: pierre.tiberghien@efs.sante.fr
Silvano Wendel-Neto & Roberta Maria Fachini
Brazil
Question 1
The S
ırio-Liban^es Hospital performs an average of 24
autologous HSCT and 32 allogeneic HSCT per year.
Question 2
All patients undergoing autologous or allogeneic HSCT
receive irradiated blood components from the time of ini-
tiation of conditioning chemotherapy and we continue to
© 2020 International Society of Blood Transfusion
Vox Sanguinis (2021) 116, e25–e43
International Forum e39
provide this special care throughout the life of these
patients.
To red blood cell concentrate (RBC), the source of irra-
diation is caesium-137 and the dose is of at least
2500 cGy into the centre of the component.
To platelet components and fresh frozen plasma, since
March 2017, all of these components are treated to patho-
gen inactivation by INTERCEPT (from CERUS) and we
considered that the UV irradiation of this procedure is an
effective prophylaxis for Transfusion-Associated Graft-vs-
Host Disease (TA-GVHD) too.
Question 3
The routine is to perform a complete blood count every
day during HSCT until the platelet and neutrophil
engraftments are established. No differentiation is made
in monitoring patients undergoing autologous or allo-
geneic HSCT.
The platelet engraftment is considered when the
patient’s count is at least 20 000/µl after three consecu-
tive days without platelet transfusion, and the neutrophil
engraftment is defined when the patient presents an abso-
lute neutrophil count of more than 500/µl across three
consecutive days too.
Question 4
Regarding the red blood cell transfusion protocol:
- For stable and hospitalized patients, the red blood cell
transfusion has been considered to a haemoglobin
threshold of 7.0–8.0 g/dl, unless the patient is symp-
tomatically anaemic or has other comorbidities.
- We are adhering to a restrictive strategy and our pol-
icy is to transfuse one red blood cell unit and evaluate
the clinical and laboratory response of the patient, as
long as he is stable, before indicating the second unit.
- The transfusion strategy in ABO-mismatched cases
must consider both the blood group systems of the
recipient and the donor.
In case of major or bidirectional ABO-mismatched trans-
plants, transfusion of blood group O RBCs is necessary until
anti-donor isohemagglutinins (IHA) are undetectable in
two consecutive blood samples of the recipients.
At our institution, if a patient is independent of RBC
transfusion for 90 days and no incompatible isohemagglu-
tinins against the new RBC phenotype are detected in two
consecutive blood samples, then the patient’s native blood
type is switched to the donor type for future transfusion.
e40 International Forum
Question 5
Our institutional protocol recommends prophylactic plate-
let transfusion at a platelet count of 10 9 109 cells/l or
less in a nonbleeding patient.
The recommendation of a 50 9 109 cells/l platelet
transfusion threshold is considered for a major invasive
procedure and 30 9 109cells/l for patients having elective
central venous catheter placement.
In the presence of active bleeding, the platelet count is
evaluated in conjunction with clinical signs, and the
transfusion is considered therapeutic.
We transfuse pools of random platelets and single donor
platelets by apheresis devices with no distinction, unless to
patients with specific antibodies identified. Both of them
are doses of around 3 9 1011 cells suspended in plasma.
ABO-compatible platelet transfusions are always pre-
ferred. When it is not available and the patient needs to
receive ABO-incompatible plasma via platelet transfusion,
the product with a titre of anti-A and/or Anti-B iso-
hemagglutinins less than 100 is selected. Furthermore,
when only products with high titres are available, a
plasma volume reduction is performed.
The efficacy of platelet transfusions by a post-transfu-
sion platelet count at 15–60 min are not systematically
monitor. This procedure is adopted when the patient
shows clinical signs or a 24-h platelet count suggestive of
low transfusion increment.
Initially, the non-immune conditions, such as consump-
tive coagulopathy, sepsis and splenomegaly are evaluated.
The investigation of alloimmune refractoriness in a
patient with thrombocytopenia due to bone marrow fail-
ure is considered when a 1-h increment is less than
7.5 9 109 cells/l on two consecutive occasions, using
ABO-identical platelets and the absence of predominantly
non-immunological factors.
To manage these situations, we perform different methods
to screening (the indirect Platelet Immunofluorescence Test
– PIFT) and identification (Monoclonal Antibody Immobi-
lization of Platelet Antigen – MAIPA) of platelet antibodies.
The patients who are refractory to platelet transfusion
and have class I HLA antibodies and/or HPA antibodies
identified, have been transfused preferably with class I
HLA and/or HPA-selected platelet components.
In our Institution, approximately 2000 donors are
already typed by class I HLA and HPA 1, HPA 5. And,
additionally, 500 donors are typed to all other HPAs.
Question 6
In cases of major ABO incompatibilities, our centre moni-
tor post-transplantation anti-A and anti-B
isohemagglutinins IgG and IgM titres weekly, especially
when pre-transplant levels are high.
After transplantation, the intention is to achieve titres
below 1:16.
We believe that this is particularly important to allow
detection of an increase in titre when this coincides with
the appearance of donor-derived RBCs.
Hemagglutinin titre quantification is followed until it is
undetectable for two consecutive weeks, except in
patients with persistent RBC transfusion requirements.
Question 7
In the case of minor ABO incompatibility, we perform a direct
antiglobulin test (DAT) and antibody screening test periodi-
cally, at least once a week, until negative results are obtained.
Question 8
Unfortunately, we do not have this data. The clinical
management of these patients is done by the physician
directly responsible for the indication of the transplant.
Silvano Wendel-Neto
S
ırio-Liban^es Blood Bank
S~ao Paulo, SP -Brazil
Email: snwendel@terra.com.br
Roberta-Maria Fachini
S
ırio-Liban^es Blood Bank
S~ao Paulo, SP -Brazil
Email: fachinir@ihsl.com.br
Suzy Morton, Charles Craddock & Matthew Lumley
United Kingdom–Birmingham
Question 1
In 2019, 116 allogeneic HSCT (16 sibling donor, 3 hap-
loidentical, 95 unrelated and 2 umbilical cord) were per-
formed at University Hospitals Birmingham across two
inpatient facilities. 22 patients had myeloablative condi-
tioning. 158 autologous HSCT were performed.
Question 2
Use of irradiated cellular blood components is in keeping
with British Society for Haematology (BSH) guideline [1]
which states that irradiated cellular blood components
should be given to recipients of autologous HSCT from
initiation of conditioning chemo/radiotherapy until
© 2020 International Society of Blood Transfusion
Vox Sanguinis (2021) 116, e25–e43

3 months post-transplant (6 months if total body irradia-
tion was used in conditioning). After allogeneic stem cell
transplant, irradiated components should be given from
the time of initiation of conditioning chemoradiotherapy
and continued for the duration of graft-versus-host dis-
ease (GvHD) prophylaxis, that is usually for 6 months
post-transplant, or until lymphocytes are >1 9 109/l. If
chronic GvHD is present or if continued immunosuppres-
sive treatment is required, irradiated blood components
should be given indefinitely.
In practice, patients are identified as requiring irradi-
ated components and we do not proactively remove this
requirement. Few autologous recipients go on to have
significant transfusion requirements, and many allogeneic
SCT recipients have other indications to need lifelong
irradiated components, for example anti-thymocyte glob-
ulin, alemtuzumab or fludarabine given during condition-
ing.
Irradiation is undertaken by NHS Blood and Transplant
usually using a gamma irradiator or on occasion an x-ray
irradiator, if the component is irradiated at a site other
than our regional blood centre. These meet the specifica-
tion defined by the Joint United Kingdom (UK) Blood
Transfusion and Tissue Transplantation Services Profes-
sional Advisory Committee, that is a minimum of 25 Gy,
but with no part receiving more than 50 Gy [2].
Question 3
Platelet counts are performed daily for inpatients. Patients
remain in hospital until they demonstrate marrow
engraftment. Following engraftment, counts are done at
least weekly until the 100th day post transplant, or on
every patient contact, if required more frequently for
clinical reasons.
Autologous transplants may be performed as an outpa-
tient but all patients are admitted at the onset of neu-
tropenia until neutrophil recovery at which stage
neutrophile and platelet counts are measured until neu-
trophil engraftment documented according to EBMT crite-
ria. Following discharge from hospital, blood counts are
done at every clinical review, many of which are per-
formed at referring centres, the frequency of which is
guided by clinical circumstances.
Question 4
The haemoglobin threshold is 70 g/l for stable patients
during the period of aplasia, in the absence of significant
cardiac disease. Single unit transfusions are used for
inpatients unless they are scheduled for imminent dis-
charge in which case two units may be transfused. For
outpatients, two units may be given depending on the
© 2020 International Society of Blood Transfusion
Vox Sanguinis (2021) 116, e25–e43
International Forum e41
trajectory of the count, frequency of visits and comor-
bidities of the patient.
Donor group red cells are transfused following full
conversion to the donor group. This is established by the
(1) Disappearance of any dual population on forward and
reverse ABO grouping by indirect antiglobulin testing
(2) Negative direct antiglobulin test (DAT) with
polyspecific anti-human globulin
As an additional measure, at one of the hospital sites,
donor group components are not transfused for the first
year after transplant, and until full or near full chimerism
is confirmed.
Question 5
A prophylactic treatment strategy for both autologous
and allogeneic HSCT is used, with a threshold of
10 9 109/l in the absence of fever, bleeding or planned
procedures. Thresholds are in keeping with BSH guideli-
nes for platelet transfusion [3]. Single donor apheresis
platelets and pooled platelets are used interchangeably.
Apheresis platelets are collected solely from male donors,
or female donors who have been tested and are negative
for human leucocyte antigen (HLA) and human platelet
antigen (HPA) antibodies. Apheresis platelets are sus-
pended in plasma. Pooled platelets are suspended in 60–
65% additive solution and 30–35% plasma (from a male
donor) [4].
For minor ABO-incompatible platelet transfusions,
‘high titre negative’ units are given (high titre is defined
as >1/128) [4]. Pooled platelets are given in this scenario
where possible due to their lower plasma volume.
Post transfusion increments are only undertaken if
there is a concern regarding poor increments, or after
transfusion of HLA-selected platelets in order to guide
further donor selection. Platelet counts are routinely
taken within 24 h of most transfusions by virtue of the
patient being in hospital during the hypoplastic phase.
When increments are checked, these are performed within
15–60 min of the end of the transfusion.
Platelet refractoriness is managed by first demon-
strating true refractoriness, investigating and then pro-
viding appropriate support. In the first instance,
refractoriness is demonstrated with increments taken
within 24 h of a single unit of ABO identical platelet
transfusion on two occasions. If the rise in platelet
count is <10 9 109/l on both occasions, the patient is
considered refractory. Causes of platelet refractoriness
(including immune and non-immune) are considered
and addressed where possible. Patients with no reversi-
ble cause and clinical suspicion of immune refractori-
ness are tested for HLA antibodies. If no HLA
antibodies are detected, HPA antibodies are tested for.
e42 International Forum
If alloantibodies are detected, apheresis platelets from
donors negative for the corresponding antigen are
selected, and post-transfusion increments are taken
with every bag to guide further donor selection. If no
HLA or HPA antibodies are detected but there is
strong clinical suspicion of alloantibody destruction of
transfused patients, apheresis platelets are selected
based on the HLA type, to closely match the patient
as much as possible and avoiding potential sensitisa-
tion to high frequency antigens.
Question 6
Isoagglutinin titres are not routinely measured following
HSCT.
Question 7
Direct antiglobulin tests are not performed routinely fol-
lowing ABO-incompatible HSCT, except when the forward
and reverse groups cease to demonstrate a dual popula-
tion as described in question 4.
Question 8
Pure red cell aplasia is only rarely observed after allo-
geneic HSCT. Its incidence is not monitored routinely
in our institution. There are some patients with major
ABO incompatibility who have a significant red cell
requirement for a prolonged period; we investigate in
order to exclude PRCA and parvovirus infection, and to
confirm trilineage engraftment, and then treat with
erythropoietin. Patients with evidence of immune
haemolytic anaemia are treated with steroids and ritux-
imab.
References
1 Treleaven J, Gennery A, Marsh J, et al. Guidelines on the use of
irradiated blood components. Br J Haematol 2010;152:35–51.
2 Joint United Kingdom (UK) Blood Transfusion and Tissue
Transplantation Services Professional Advisory Committee:
Guidelines for the Blood Transfusion Services in the United
Kingdom, ed 8. London, TSO, 2013.
3 Estcourt L, Birchall J, Allard S, et al. Use of platelet transfu-
sions. Br J Haematol 2016;176:365–94.
4 NHSBT Portfolio of Blood Components and Guidance for
their Clinical Use: SPECIFICATION SPN223/10. Available at:
https://nhsbtdbe.blob.core.windows.net/umbraco-assets-corp/
16494/spn223v10.pdf Last accessed 10/05/2020.
Suzy Morton
NHS Blood and Transplant
Birmingham, UK.
Email: suzy.morton@nhsbt.nhs.uk
Charles Craddock
University Hospitals Birmingham NHS
Foundation Trust
Birmingham, United Kingdom
Email: charles.craddock@uhb.nhs.uk
Matthew Lumley
University Hospitals Birmingham NHS
Foundation Trust
Sutton Coldfield, UK
Email: matthew.lumley@uhb.nhs.uk
Jolanta Antoniewicz-Papis, Kazimierz Hałaburda & Magdalena
Łez towska
Poland
Question 1
At the Institute of Hematology and Transfusion Medicine
(Institute), approximately 60–65 autologous and 65–70
allogeneic transplants are performed per year.
Question 2
At the Institute, every patient who requires multiple trans-
fusions, every potential transplant recipient as well as
transplant recipient is transfused with irradiated blood
components until transfusions are required. In the case of
HSCT patients, cellular blood components are irradiated for
at least 12 months after autologous donation and indefi-
nitely following allogeneic transplantations.
Currently, caesium (Cs) is used as source of irradiation.
The dose delivered to the blood component is no less
than 25 Gy at each site of the component, but does not
exceed 40 Gy. Once in 3 years, the distribution of the
delivered dose is mapped. Mapping shows that the central
dose is 32 Gy.
For microbiological safety and prophylaxis against
graft-versus-host disease, most platelet concentrates are
pathogen reduced instead of being subjected to irradia-
tion. If FFP is indicated, it is also administered after
pathogen reduction or after quarantine.
© 2020 International Society of Blood Transfusion
Vox Sanguinis (2021) 116, e25–e43
 International Forum e43

Question 3 transplants. If antibodies are detected, we limit pro-

phylactic transfusions of platelets and in case of
Yes, complete blood count (CBC) is performed during
bleeding – in principle, we administer platelets with
pancytopenic phase in all transplant patients. During the
no antigens to the antibodies detected in patients.
regeneration phase – when regular transfusion support is
not required and the patients are clinically stable – CBC
is performed every other day. Question 6
Isoagglutinin titre is routinely measured before transplan-
Question 4 tation; after transplantation – only fort special indications
resulting from post-transplant complications.
Red blood cell transfusion protocol:
Chimerism is monitored with measurements performed
- The threshold is 8 g/dl.
monthly and as required. If the results are indeterminate,
- We use 2 RBC unit transfusion policy.
antibody tests and titre measurements are performed, usu-
- At the Institute, the red blood cell group of RBC con-
ally 1 month after transplant and then at 3 months fol-
centrates is changed to donor cell group usually
lowing transplant and if required.
within 6 months following transplantation, providing
no transfusions were performed and no relapse
occurred. At the same time, chimerism is tested and Question 7
additional immunohematology tests are performed
At the Institute, prior to transplant, we perform direct and
(including – among others – DAT, IAT, blood group-
indirect antiglobulin test, blood grouping as well as IgG
ing) for assessment of engraftment.
and IgM antibodies. In the post-transplant period, antibody
monitoring is performed, as well as direct antiglobulin test,
Question 5 Rh phenotype as well as other phenotypes, as required.

Regarding platelet transfusion protocol:

- At the Institute, we use the prophylactic platelet Question 8
transfusion strategy for autologous transplantations.
We report pure red cell aplasia in 1–2 patients per year
- The threshold is 20 G/l in patients with grade III
and solely after reduced intensity conditioning. If the
mucositis, petecchiae or fever ≥38°C. In asymp-
patient does not require multiple transfusions, whenever
tomatic patients, the threshold is 10–15 G/l.
possible, we prefer early start of immunosuppression
- Preferably, single donor platelet transfusions are used.
taper to other therapeutic measures.
If unavailable, which does not often happen, random
Our choice of management procedure is justified by the
donor platelets are acceptable. Platelets suspended in
fact that the majority of such patients is in advanced age
additive solution and in plasma are in use at the
and have low-risk diseases. Early immunosuppression
Institute.
taper decreases the risk of relapse. If necessary, we per-
- As a general rule, minor ABO platelet incompatibility
form plasmaphereses and apply/administer rituximab
is not accepted at the Institute. On rare occasions
(especially in case of concomitant graft-versus-host dis-
when minor ABO-incompatible platelets must be
ease) and erythropoiesis-stimulating agents.
used, washed platelets suspended in saline or additive
solution are transfused.
Jolanta Antoniewicz-Papis
- There is no systematic monitoring of the efficacy of Institute of Hematology and Transfusion Medicine
platelet transfusions by a post-transfusion platelet Warsaw, Poland
count. Efficacy of PLT transfusion is monitored at Email: jpapis@ihit.waw.pl
60 min only if ineffectiveness is suspected, for exam-
ple in patients who require daily transfusion support. Kazimierz Hałaburda
- In such case, the diagnostic procedure is performed Institute of Hematology and Transfusion Medicine
which consists in detection of anti-HPA and anti- Warsaw, Poland
HLA antibodies as well as evaluation of symptoms of Email: khalaburda@ihit.waw.pl
post-transplant thrombotic microangiopathy. In cases
Magdalena Łez towska
with better potential, it is recommended to discon-
Institute of Hematology and Transfusion Medicine
tinue calcineurin inhibitor administration and switch
Warsaw, Poland
to regime of mycophenolate and steroids for graft- Email: letowska@ihit.waw.pl
versus-host disease prophylaxis in allogeneic

© 2020 International Society of Blood Transfusion

Vox Sanguinis (2021) 116, e25–e43
 Vox Sanguinis (2021) 116, 613

DIARY OF EVENTS © 2021 International Society of Blood Transfusion

DOI: 10.1111/vox.12956

See also http://www.isbtweb.org/congresses/

4.5.2021 IPFA/PEI – The International Workshop on Surveillance and Screening of Blood-borne Pathogens

13–15.5.2021 The Canadian Society for Transfusion Medicine (CSTM) are holding their annual scientific conference virtually in 2021.

26–27.05.21 21st Congress of the European Society for Hemapheresis

5–9.6.2021 ISBT In Focus, the 31st regional congress of the ISBT, will be a virtual event in 2021

17.9.2021 11th BIC International Conference – Advances in Haemostasis and Bleeding Disorders

22–24.9.2021 Deutsche Gesellschaft f€ur Transfusionsmedizin und Immunh€amatologie e.V.

23–26.9.2021 16th International Congress on Myelodysplastic Syndromes (MDS 2021)

13–16.11.2021 32nd Regional congress of ISBT, Brisbane, Australia

613