Influenza

Influenza
What You Need to Know

INFLUENZA COLLECTION
Events in 2014 and 2015 engendered intense concern regarding Ebola virus disease, the Middle
East respiratory syndrome, and antibiotic-resistant bacteria. Still, the most worrisome human
the new england pathogens remain the influenza viruses. They not only cause annual epidemics each winter,
journal of medicine
linked to their ability to undergo ongoing antigenic variation, but through genetic reassortment
nejm journal watch with strains that infect birds and swine, new strains highly virulent for humans and capable of
Cardiology
producing severe worldwide pandemics have arisen every several decades.
Dermatology
Emergency Medicine This collection highlights recent work on the prevention and treatment of annual influenza,
Gastroenterology as well as new information on avian influenza and reflections on pandemic preparedness and
General Medicine response.
Hospital Medicine
Infectious Diseases NEJM Group delivers information and services in biomedical research, clinical practice,
Neurology health care delivery, and ongoing medical education. This collection brings together content
Oncology and Hematology from three of its publishing division leaders — the New England Journal of Medicine, NEJM Journal
Pediatrics and Adolescent Medicine
Watch, and Physician’s First Watch. NEJM Journal Watch surveys, summarizes, and provides
Physician’s First Watch
expert commentary on articles from more than 250 medical journals. Its physician-editors select
Psychiatry
Women’s Health the most relevant research studies and clinical practice guidelines, summarize the findings and
recommendations, and place them in context. Physician’s First Watch is our news arm.
nejm careercenter
We hope you find this collection illuminating, useful, and enjoyable.
nejm knowledge+
— Richard T. Ellison III, MD

Professor of Medicine, Microbiology,
and Physiological Systems
Division of Infectious Diseases and Immunology
University of Massachusetts Medical School, Worcester
— Sharon S. Salinger
Executive Editor and Manager, Editorial Operations
Clinical Programs and Product Development,
NEJM Group
October 2015
Clinical Collections — Influenza
Table of Contents
INTRODUCTION
VACCINATION
7 ACIP Issues 2015–2016 Influenza Vaccination Recommendations
Physician’s First Watch Aug 7, 2015
8 P
revention and Control of Influenza with Vaccines: Recommendations
of the Advisory Committee on Immunization Practices, United States,
2015–16 Influenza Season
Morbidity and Mortality Weekly Report (MMWR) Aug 7, 2015
16 Circulating Influenza Viruses Show Good Match with Coming Season’s Vaccine
Physician’s First Watch Sep 18, 2015
17 Vaccine for Prevention of Mild and Moderate-to-Severe Influenza in Children

New England Journal of Medicine Dec 26, 2013
28 Efficacy of Influenza Vaccine Against Severe Disease in Children

NEJM Journal Watch Pediatrics and Adolescent Medicine Apr 29, 2014
30 Efficacy of High-Dose versus Standard-Dose Influenza Vaccine in Older Adults

New England Journal of Medicine Aug 14, 2014
41 High-Dose Influenza Vaccine No More Effective than Standard-Dose in Older Adults

NEJM Journal Watch Infectious Diseases Jul 29, 2015
42 Influenza Vaccination of Pregnant Women and Protection of Their Infants

New England Journal of Medicine Sep 4, 2014
56 2009 H1N1 Influenza and Pregnancy — 5 Years Later

New England Journal of Medicine Oct 9, 2014
59 CDC Advisers Recommend New Flu Vaccine for Adults with Egg Allergy
Physician’s First Watch Jun 24, 2013
60 Do Patients Benefit from Health Care Providers’ Influenza Vaccination?

NEJM Journal Watch Infectious Diseases Jan 23, 2014
DIAGNOSIS AND TREATMENT

62 Preventing and Controlling Influenza with Available Interventions
New England Journal of Medicine Feb 27, 2014
continued
3
Downloaded from collections.nejm.org. For personal use only. No other uses without permission.
Copyright © Massachusetts Medical Society. All rights reserved.
65 Most Influenza Infections Are Subclinical

NEJM Journal Watch Infectious Diseases Mar 20, 2014
66 E
mergency Department Clinicians Struggle to Diagnose Influenza
on Clinical Grounds Alone
NEJM Journal Watch Emergency Medicine Jun 10, 2015
67 Influenza Is More Virulent Among Cancer Patients

NEJM Journal Watch Infectious Diseases Feb 4, 2014
68 New Meta-Analysis Clarifies Oseltamivir’s Benefits and Risks

NEJM Journal Watch General Medicine Feb 5, 2015
69 Oseltamivir for Influenza, Even After 48 Hours of Symptoms?

NEJM Journal Watch Infectious Diseases Dec 13, 2013
70 Effect of Oseltamivir on Viral Shedding, Illness, and Household Transmission of Influenza

NEJM Journal Watch Infectious Diseases Aug 12, 2015
71 FDA Approves the Intravenous Influenza Drug Peramivir

NEJM Journal Watch Infectious Diseases Jan 20, 2015
72 CDC Recommends Antivirals in High-Risk Flu Patients, Even Before Test Results In
Physician’s First Watch Jan 12, 2015
AVIAN INFLUENZA
74 Epidemiology of Human Infections with Avian Influenza A(H7N9) Virus in China
New England Journal of Medicine Feb 6, 2014
87 CDC Issues Health Advisory on Protecting Humans During Bird Flu Outbreak
Physician’s First Watch Jun 4, 2015
88 An Avian Flu Vaccine

NEJM Journal Watch Infectious Diseases Jul 28, 2015
89 Toward Effective Vaccines for Avian Influenza

NEJM Journal Watch Infectious Diseases Oct 15, 2014
90 Avian Influenza Viruses in U.S. Birds — No Surprise

91 Songbirds and Parakeets: Possible Sources of Influenza A(H7N9) Virus Transmission

continued
4
Back to Table of Contents
PANDEMIC INFLUENZA
93 Early Interventions to Truncate Influenza Pandemics
NEJM Journal Watch Infectious Diseases May 20, 2014
94 Pandemic Preparedness and Response — Lessons from the H1N1 Influenza of 2009
New England Journal of Medicine Apr 3, 2014
102 The Next Epidemic — Lessons from Ebola

New England Journal of Medicine Apr 9, 2015
5
VACCINATION
Each year since 1938, a vaccine against the strains of influenza that experts think will be circulating has been
prepared and offered to the public. Although in the past, vaccination was reserved for targeted groups, experts
currently recommend almost universal vaccination. However, in the 2013–2014 flu season, only 58.9% of children
and 42.2% of adults were vaccinated. The vaccines are generally safe. Tenderness at the injection site and nasal
congestion after the nasal spray are the most common effects, but more serious effects have been seen in some
populations (e.g., pregnant women) and with some vaccines (e.g., the 1979 pandemic vaccine). Effectiveness
varies depending on personal characteristics of the vaccinee and the closeness of the match between vaccine
components and circulating strains.
In this section, we present a summary of the 2015–2016 recommendations of the Advisory Committee on
Immunization Practices (ACIP); the full text of the recommendations, including the specific vaccine composition
for this season; and a news report of how well the current vaccine is matched to circulating viruses (good news).
Protection of specific populations is also considered, including in original research articles focused on children,
pregnant women, and older adults. Research summaries address patients with egg allergy and the benefit to
susceptible patients from vaccination of health care providers.
6
Source: Physician’s First Watch
ACIP Issues 2015–16 Influenza Vaccination Recommendations

Physician’s First Watch, published August 7, 2015
NEWS SUMMARY
Updated recommendations on influenza vaccination for the 2015–16 season have been issued by the CDC’s
Advisory Committee on Immunization Practices. As reported in MMWR, changes from the 2014–15 season
include the following:
• Trivalent vaccines will include A/California/7/2009 (H1N1)-like virus, an A/Switzerland/9715293/2013

(H3N2)-like virus, and a B/Phuket/3073/2013-like (Yamagata lineage) virus. These A (H3N2) and B viruses
are different from the previous season.
• The Fluzone Intradermal Quadrivalent vaccine is expected to replace the trivalent Fluzone Intradermal vaccine
for adults aged 18 to 64.
• Children aged 6 months to 8 years who are receiving their first influenza vaccine still need two doses at least
4 weeks apart, but special consideration of influenza A (H1N1)pdm09 vaccination is no longer needed, since
current vaccines incorporate that viral antigen.
• Live-attenuated vaccine (previously preferred for children aged 2–8 years) is no longer recommended over
inactivated vaccine, because observational studies failed to support the superiority of live-attenuated vaccine
seen in randomized trials.
— Christine Sadlowski
7
Morbidity and Mortality Weekly Report
Source: Morbidity and Mortality Weekly Report (MMWR)
Prevention and Control of Influenza with Vaccines: Recommendations

of the Advisory Committee on Immunization Practices, United States,
2015–16 Influenza Season
Lisa A. Grohskopf, MD1; Leslie Z. Sokolow, MSc, MPH1,2; Sonja J. Olsen, PhD1; Joseph S. Bresee, MD1; Karen R. Broder, MD3; Ruth A. Karron, MD4
This report updates the 2014 recommendations of the Information in this report reflects discussions during public
Advisory Committee on Immunization Practices (ACIP) meetings of ACIP held on February 26 and June 24, 2015.
regarding the use of seasonal influenza vaccines (1). Updated Subsequent modifications were made during CDC clearance
information for the 2015–16 season includes 1) antigenic review to update information and clarify wording. Meeting
composition of U.S. seasonal influenza vaccines; 2) informa- minutes, information on ACIP membership, and information
tion on influenza vaccine products expected to be available for on conflicts of interest are available at http://www.cdc.gov/
the 2015–16 season; 3) an updated algorithm for determining vaccines/acip/committee/members.html. Any updates will be
the appropriate number of doses for children aged 6 months posted at http://www.cdc.gov/flu.
through 8 years; and 4) recommendations for the use of live
attenuated influenza vaccine (LAIV) and inactivated influenza Groups Recommended for Vaccination and
vaccine (IIV) when either is available, including removal of the Timing of Vaccination
2014–15 preferential recommendation for LAIV for healthy Routine annual influenza vaccination is recommended for
children aged 2 through 8 years. Information regarding top- all persons aged ≥6 months who do not have contraindica-
ics related to influenza vaccination that are not addressed in tions. Optimally, vaccination should occur before onset of
this report is available in the 2013 ACIP seasonal influenza influenza activity in the community. Health care providers
recommendations (2). should offer vaccination by October, if possible. Vaccination
should continue to be offered as long as influenza viruses are
circulating. Children aged 6 months through 8 years who
Recommendations for routine use of vaccines in chil- require 2 doses (see “Vaccine Dose Considerations for Children
dren, adolescents, and adults are developed by the Advisory Aged 6 Months through 8 Years”) should receive their first dose
Committee on Immunization Practices (ACIP). ACIP is as soon as possible after vaccine becomes available, and the
chartered as a federal advisory committee to provide expert second dose ≥4 weeks later. To avoid missed opportunities for
external advice and guidance to the Director of the Centers vaccination, providers should offer vaccination to unvaccinated
for Disease Control and Prevention (CDC) on use of vac- persons aged ≥6 months during routine health care visits and
cines and related agents for the control of vaccine-preventable hospitalizations when vaccine is available.
diseases in the civilian population of the United States. Antibody levels induced by vaccine decline after vaccina-
Recommendations for routine use of vaccines in children tion (3–5). Although a 2008 literature review found no clear
and adolescents are harmonized to the greatest extent evidence of more rapid decline among older adults (6), a 2010
possible with recommendations made by the American study noted a statistically significant decline in antibody titers
Academy of Pediatrics (AAP), the American Academy of 6 months after vaccination among persons aged ≥65 years (5).
Family Physicians (AAFP), and the American College of A case-control study conducted in Navarre, Spain, during the
Obstetricians and Gynecologists (ACOG). Recommendations 2011–12 influenza season revealed a decline in vaccine effec-
for routine use of vaccines in adults are harmonized with rec- tiveness, primarily affecting persons aged ≥65 years (7). While
ommendations of AAFP, ACOG, and the American College delaying vaccination might permit greater immunity later in
of Physicians (ACP). ACIP recommendations adopted by the season, deferral might result in missed opportunities to
the CDC Director become agency guidelines on the date vaccinate, as well as difficulties in vaccinating a population
published in the Morbidity and Mortality Weekly Report within a more constrained time period. Vaccination programs
(MMWR). Additional information regarding ACIP is avail- should balance maximizing the likelihood of persistence of
able at http://www.cdc.gov/vaccines/acip. vaccine-induced protection through the season with avoiding
missed opportunities to vaccinate or vaccinating after influenza
Originally published on August 7, 2015. virus circulation begins.
8
818 MMWR / August 7, 2015 from
Downloaded / Vol.collections.nejm.org.
64 / No. 30 For personal use only. No other uses without permission.
Influenza Vaccine Composition for the 2. In October 2014, FDA approved an expanded age
2015–16 Season indication for the use of Flublok (Recombinant Influenza
For 2015–16, U.S.-licensed trivalent influenza vaccines will con- Vaccine, Trivalent [RIV3], Protein Sciences, Meriden,
tain hemagglutinin (HA) derived from an A/California/7/2009 Connecticut), which was previously approved for persons
(H1N1)-like virus, an A/Switzerland/9715293/2013 (H3N2)- aged 18 through 49 years. Flublok is now indicated for
like virus, and a B/Phuket/3073/2013-like (Yamagata lineage) persons aged ≥18 years (11). Approval for persons aged
virus. This represents changes in the influenza A (H3N2) virus ≥50 years is based upon studies of immunogenicity and
and the influenza B virus as compared with the 2014–15 sea- safety of the vaccine in three randomized trials (12–14);
son. Quadrivalent influenza vaccines will contain these vaccine data demonstrating a decrease in influenza disease in
viruses, and a B/Brisbane/60/2008-like (Victoria lineage) virus, persons aged ≥50 years after vaccination with Flublok
which is the same Victoria lineage virus recommended for quad- are not available.
rivalent formulations in 2013–14 and 2014–15 (8). 3. In December 2014, FDA approved Fluzone Intradermal
Quadrivalent (Sanofi Pasteur, Inc., Swiftwater,
Available Vaccine Products and Indications Pennsylvania), for persons aged 18 through 64 years (15).
Various influenza vaccine products are anticipated to be It is anticipated that this formulation will replace the
available during the 2015–16 season (Table). These recom- previously available trivalent Fluzone Intradermal for the
mendations apply to all licensed influenza vaccines used within 2015–16 influenza season. In a randomized study of
Food and Drug Administration (FDA)-licensed indications. 3,355 adults aged 18 through 64 years comparing safety
Differences between ACIP recommendations and labeled and immunogenicity of Fluzone Intradermal Quadrivalent
indications are noted in the Table. For persons for whom with two different trivalent intradermal formulations of
more than one type of vaccine is appropriate and available, Fluzone (each one containing one of the two influenza B
ACIP does not express a preference for use of any particular viruses contained in the quadrivalent vaccine), the
product over another. quadrivalent formulation was immunogenically
New and updated influenza vaccine product approvals noninferior to the trivalent formulations for the
include the following: influenza A and matched B viruses, immunogenically
1. In August 2014, FDA approved Afluria (inactivated superior for the unmatched B viruses, and had a similar
influenza vaccine, bioCSL, Inc., King of Prussia, adverse event profile (16). Efficacy data for Fluzone
Pennsylvania) for intramuscular administration via the Intradermal Quadrivalent are not available.
Stratis needle-free jet injector (PharmaJet, Inc., Golden,
Colorado), for persons aged 18 through 64 years (9). Vaccine Dose Considerations for Children Aged
Adults aged 18 through 64 years may receive Afluria 6 Months Through 8 Years
either by the Stratis injector or with a sterile needle and Children aged 6 months through 8 years require 2 doses
syringe. All other inactivated influenza vaccines are of influenza vaccine (administered ≥4 weeks apart) during
approved for administration by sterile needle and syringe their first season of vaccination to optimize response (17–19).
only. The Stratis injector is a reusable spring-powered Since the emergence of influenza A(H1N1)pdm09 (the 2009
device which injects the vaccine through a single-use H1N1 pandemic virus), recommendations for determining
sterile needle-free syringe into the deltoid muscle. In a the number of doses needed have specified previous receipt
prelicensure study of 1,250 adults aged 18 through of vaccine containing influenza A(H1N1)pdm09. In light of
64 years (10), local injection site symptoms were reported the continuing circulation of influenza A(H1N1)pdm09 as
more frequently by persons who received Afluria via the the predominant influenza A(H1N1) virus since 2009, and
Stratis Injector than those who were vaccinated with a the inclusion of an A/California/7/2009(H1N1)-like virus in
sterile needle and syringe; most resolved within 3 days. U.S. seasonal influenza vaccines since the 2010–2011 season,
Those who received Afluria via the Stratis injector had separate consideration of receipt of vaccine doses containing
antibody levels against influenza virus that were this virus is no longer recommended.
noninferior to those who received Afluria by sterile needle Several studies have suggested that for viruses which are
and syringe. Data comparing rates of influenza illness in the same in both doses of vaccine, longer intervals between
persons vaccinated with the Stratis injector versus needle the 2 doses do not compromise immune response (20–22).
and syringe are not available. In a study conducted across two seasons during which the
influenza A(H1N1) vaccine virus did not change but the
B virus did change, children aged 10 through 24 months who
9
MMWR
Downloaded from collections.nejm.org. For personal use only. No other/uses
August 7, 2015
without / Vol. 64 / No. 30
permission. 819
TABLE. Influenza vaccines — United States, 2015–16 influenza season*

Mercury
(from
thimerosal) Ovalbumin
Trade name Manufacturer Presentation µg/0.5 mL µg/0.5 mL Age indications Latex Route
Inactivated influenza vaccine, quadrivalent (IIV4), standard dose
Contraindications*: Severe allergic reaction to any vaccine component, including egg protein, or after previous dose of any influenza vaccine.
Precautions*: Moderate to severe acute illness with or without fever; history of Guillain-Barré syndrome within 6 weeks of receipt of influenza vaccine.
Fluarix Quadrivalent GlaxoSmithKline 0.5 mL single-dose prefilled syringe — ≤0.05 ≥3 yrs No IM†
FluLaval Quadrivalent ID Biomedical Corp. of 5.0 mL multi-dose vial <25 ≤0.3 ≥3 yrs No IM†
Quebec (distributed
by GlaxoSmithKline)
Fluzone Quadrivalent Sanofi Pasteur 0.25 mL single-dose prefilled — § 6 through No IM†
syringe 35 mos
0.5 mL single-dose prefilled syringe — § ≥36 mos No IM†
0.5 mL single-dose vial — § ≥36 mos No IM†
5.0 mL multi-dose vial 25 § ≥6 mos No IM†
Fluzone Intradermal¶ Sanofi Pasteur 0.1 mL single-dose prefilled — § 18 through No ID**
Quadrivalent microinjection system 64 yrs
Inactivated influenza vaccine, trivalent (IIV3), standard dose

Afluria bioCSL 0.5 mL single-dose prefilled syringe — <1 ≥9 yrs†† No IM†
5.0 mL multi-dose vial 24.5 <1 ≥9 yrs†† via No IM†
needle;18
through 64 yrs
via jet injector
Fluvirin Novartis Vaccines and 0.5 mL single-dose prefilled syringe ≤1 ≤1 ≥4 yrs Yes§§ IM†
Diagnostics 5.0 mL multi-dose vial 25 ≤1 ≥4 yrs No IM†
Fluzone Sanofi Pasteur 5.0 mL multi-dose vial 25 § ≥6 mos No IM†
Inactivated influenza vaccine, cell-culture-based (ccIIV3), standard dose

Flucelvax Novartis Vaccines and 0.5 mL single-dose prefilled syringe — ¶¶ ≥18 yrs Yes§§ IM†
Diagnostics
Inactivated influenza vaccine, trivalent (IIV3), high dose

Fluzone High-Dose*** Sanofi Pasteur 0.5 mL single-dose prefilled syringe — § ≥65 yrs No IM†
Recombinant influenza vaccine, trivalent (RIV3), standard dose

Contraindications*: Severe allergic reaction to any vaccine component.
Flublok Protein Sciences 0.5 mL single-dose vial — 0 ≥18 yrs No IM†
Live attenuated influenza vaccine, quadrivalent (LAIV4)

Contraindications*: Severe allergic reaction to any vaccine component, including egg protein, or after previous dose of any influenza vaccine. Concomitant use of aspirin or
aspirin-containing medications in children and adolescents.
In addition, ACIP recommends LAIV4 not be used for pregnant women, immunosuppressed persons, persons with egg allergy, and children aged 2 through 4 years who
have asthma or who have had a wheezing episode noted in the medical record within the past 12 months, or for whom parents report that a health care provider stated
that they had wheezing or asthma within the last 12 months.
LAIV4 should not be administered to persons who have taken influenza antiviral medications within the previous 48 hours.
Persons who care for severely immunosuppressed persons who require a protective environment should not receive LAIV4, or should avoid contact with such persons for 7
days after receipt.
Precautions*: Moderate to severe acute illness with or without fever; history of Guillain-Barré syndrome within 6 weeks of receipt of influenza vaccine; asthma in persons
aged 5 years and older; medical conditions which might predispose to higher risk for complications attributable to influenza.
FluMist Quadrivalent††† MedImmune 0.2 mL single-dose prefilled — <0.24 (per 0.2 mL) 2 through 49 yrs No IN
intranasal sprayer
See table footnotes on page next page.
10
TABLE. (Continued) Influenza vaccines — United States, 2015–16 influenza season*

Abbreviations: ACIP = Advisory Committee on Immunization Practices; ID = intradermal; IM = intramuscular; IN = intranasal.
* Immunization providers should check Food and Drug Administration-approved prescribing information for 2015–16 influenza vaccines for the most complete
and updated information, including (but not limited to) indications, contraindications, warnings, and precautions. Package inserts for U.S.-licensed vaccines are
available at www.fda.gov/BiologicsBloodVaccines/Vaccines/ApprovedProducts/ucm093833.htm.
† For adults and older children, the recommended site for intramuscular influenza vaccination is the deltoid muscle. The preferred site for infants and young children
is the anterolateral aspect of the thigh. Specific guidance regarding site and needle length for intramuscular administration may be found in the ACIP General
Recommendations on Immunization, available at www.cdc.gov/mmwr/preview/mmwrhtml/rr6002a1.htm.
§ Available upon request from Sanofi Pasteur (1–800–822–2463 or MIS.Emails@sanofipasteur.com).
¶ Quadrivalent inactivated influenza vaccine, intradermal: a 0.1-mL dose contains 9 µg of each vaccine antigen (36 µg total).
** The preferred injection site is over the deltoid muscle. Fluzone Intradermal Quadrivalent is administered using the delivery system included with the vaccine.
†† Age indication per package insert is ≥5 years; however, ACIP recommends Afluria not be used in children aged 6 months through 8 years because of increased
risk of febrile reactions noted in this age group with bioCSL’s 2010 Southern Hemisphere IIV3. If no other age-appropriate, licensed inactivated seasonal influenza
vaccine is available for a child aged 5 through 8 years who has a medical condition that increases the child’s risk for influenza complications, Afluria can be used;
however, providers should discuss with the parents or caregivers the benefits and risks of influenza vaccination with Afluria before administering this vaccine.
Afluria may be used in persons aged ≥9 years.
§§ Syringe tip cap may contain natural rubber latex.
¶¶ Information not included in package insert. Estimated to contain <50 femtograms (5x10-8 µg) of total egg protein (of which ovalbumin is a fraction) per 0.5 mL
dose of Flucelvax.
*** Trivalent inactivated influenza vaccine, high-dose: a 0.5-mL dose contains 60 µg of each vaccine antigen (180 µg total).
††† FluMist is shipped refrigerated and stored in the refrigerator at 35°F–46°F (2°C–8°C) after arrival in the vaccination clinic. The dose is 0.2 mL divided equally between
each nostril. Health care providers should consult the medical record, when available, to identify children aged 2 through 4 years with asthma or recurrent wheezing
that might indicate asthma. In addition, to identify children who might be at greater risk for asthma and possibly at increased risk for wheezing after receiving
LAIV, parents or caregivers of children aged 2 through 4 years should be asked: “In the past 12 months, has a health care provider ever told you that your child
had wheezing or asthma?” Children whose parents or caregivers answer “yes” to this question and children who have asthma or who had a wheezing episode
noted in the medical record within the past 12 months should not receive FluMist.
received 1 dose of IIV during the fall of each season had similar FIGURE 1. Influenza vaccine dosing algorithm for children aged
immune responses to the unchanged A(H1N1) virus antigen 6 months through 8 years — Advisory Committee on Immunization
Practices, United States, 2015–16 influenza season
and to the drifted A(H3N2) virus antigen, compared with
children aged 6 through 24 months who received 2 doses of
the same IIV during the latter season. However, the first group Has the child received ≥2
total doses of trivalent or
had significantly lower antibody responses to the B antigen quadrivalent influenza
(20). Since the 2010–11 season, guidance for determining vaccine before July 1, 2015*
the appropriate number of doses has taken strain changes into
account. Because of the change in vaccine composition for
2015–16, children aged 6 months through 8 years will need
No or
to have received ≥2 doses of influenza vaccine previously to Yes
don’t know
require only 1 dose for the 2015–16 season.
For 2015–16, ACIP recommends that children aged
6 months through 8 years who have previously received
1 dose of 2015–16 2 doses† of 2015–16
≥2 total doses of trivalent or quadrivalent influenza vaccine influenza vaccine influenza vaccine
before July 1, 2015, require only 1 dose for 2015–16. The
two previous doses need not have been given during the same
* The two doses need not have been received during the same season or
season or consecutive seasons. Children in this age group who consecutive seasons.
†
have not previously received a total of ≥2 doses of trivalent Doses should be administered ≥4 weeks apart.
or quadrivalent influenza vaccine before July 1, 2015 require
2 doses for 2015–16. The interval between the 2 doses should conducted before the 2009 H1N1 pandemic demonstrated
be at least 4 weeks (Figure 1). superior efficacy of LAIV in children (24–26). A randomized
controlled trial conducted during the 2004–05 season among
Considerations for the Use of Live Attenuated 7,852 children aged 6 through 59 months demonstrated a
Influenza Vaccine and Inactivated Influenza 55% reduction in culture-confirmed influenza among chil-
Vaccine When Either is Available dren who received trivalent LAIV (LAIV3) compared with
Both LAIV and IIV have been demonstrated to be effec- those who received trivalent IIV (IIV3). LAIV3 efficacy was
tive in children and adults. Among adults, most comparative higher than that of IIV3 against both antigenically drifted and
studies have demonstrated that LAIV and IIV were of similar well-matched influenza viruses (24). In a comparative study
efficacy or that IIV was more efficacious (23). Several studies conducted during the 2002–03 season, LAIV3 provided 53%
11
MMWR
August 7, 2015
permission. 821
greater relative efficacy compared with IIV3 in children aged 6 delayed to procure a specific vaccine preparation if an
through 71 months who had previously experienced recurrent appropriate one is already available.
respiratory tract infections (25). 2. For healthy children aged 2 through 8 years who have
In June 2014, following review of evidence on the relative no contraindications or precautions, either LAIV or IIV
efficacy of LAIV compared with IIV for healthy children, is an appropriate option. No preference is expressed for
ACIP recommended that when immediately available, LAIV LAIV or IIV for any person aged 2 through 49 years for
should be used for healthy children aged 2 through 8 years whom either vaccine is appropriate. An age-appropriate
who have no contraindications or precautions. However, data formulation of vaccine should be used.
from subsequent observational studies of LAIV and IIV vaccine 3. LAIV should not be used in the following populations:
effectiveness indicated that LAIV did not perform as well as – Persons aged <2 years or >49 years;
expected based upon the observations in earlier randomized – Persons with contraindications listed in the package
trials (27,28). Analysis of data from three observational studies insert:
of LAIV4 vaccine effectiveness for the 2013–14 season (the ˏ Children aged 2 through 17 years who are
first season in which LAIV4 was available) revealed poor effec- receiving aspirin or aspirin-containing products;
tiveness of LAIV4 against influenza A(H1N1)pdm09 among ˏ Persons who have experienced severe allergic
children aged 2 through 17 years (27). During this season, reactions to the vaccine or any of its components,
H1N1pdm09 virus predominated for the first time since or to a previous dose of any influenza vaccine;
the 2009 pandemic. The reasons for the lack of effectiveness – Pregnant women;
of LAIV4 against influenza A(H1N1)pdm09 are still under – Immunocompromised persons (see also “Vaccine Selection
investigation. Moreover, although one large randomized trial and Timing of Vaccination for Immunocompromised
observed superior relative efficacy of LAIV3 compared with Persons”);
IIV3 against antigenically drifted H3N2 influenza viruses – Persons with a history of egg allergy;
during the 2004–05 season (24), interim analysis of observa- – Children aged 2 through 4 years who have asthma or
tional data from the U.S. Influenza Vaccine Effectiveness (U.S. who have had a wheezing episode noted in the medical
Flu VE) Network for the early 2014–15 season (in which anti- record within the past 12 months, or for whom parents
genically drifted H3N2 viruses were predominant) indicated report that a health care provider stated that they had
that neither LAIV4 nor IIV provided significant protection in wheezing or asthma within the last 12 months (Table,
children aged 2 through 17 years; LAIV did not offer greater footnote). For persons aged ≥5 years with asthma,
protection than IIV for these viruses (28). recommendations are described in item 4 of this list;
In the absence of data demonstrating consistent greater rela- – Persons who have taken influenza antiviral medications
tive effectiveness of the current quadrivalent formulation of within the previous 48 hours.
LAIV, preference for LAIV over IIV is no longer recommended. 4. In addition to the groups for whom LAIV is not
ACIP will continue to review the effectiveness of influenza recommended above, the “Warnings and Precautions”
vaccines in future seasons and update these recommendations section of the LAIV package insert indicates that persons
if warranted. of any age with asthma might be at increased risk for
For children and adults with chronic medical conditions wheezing after administration of LAIV (29). The package
conferring a higher risk for influenza complications, data on insert also notes that the safety of LAIV in persons with
the relative safety and efficacy of LAIV and IIV are limited. other underlying medical conditions that might
In a study comparing LAIV3 and IIV3 among children predispose them to complications after wild-type
aged 6 through 17 years with asthma conducted during the influenza virus infection (e.g., chronic pulmonary,
2002–03 season, LAIV conferred 32% increased protection cardiovascular [except isolated hypertension], renal,
relative to IIV in preventing culture-confirmed influenza; no hepatic, neurologic, hematologic, or metabolic disorders
significant difference in asthma exacerbation events was noted [including diabetes mellitus]) (2), has not been
(26). Available data are insufficient to determine the level of established. These conditions, in addition to asthma in
severity of asthma for which administration of LAIV would persons aged ≥5 years, should be considered precautions
be appropriate. for the use of LAIV.
For 2015–16, ACIP recommends the following: 5. Persons who care for severely immunosuppressed persons
1. All persons aged ≥6 months should receive influenza who require a protective environment should not receive
vaccine annually. Influenza vaccination should not be LAIV, or should avoid contact with such persons for 7 days
12
after receipt, given the theoretical risk for transmission of anaphylaxis) in children with egg allergy was estimated to be
the live attenuated vaccine virus to close contacts. 1.3% (37). Eight children experienced milder, self-limited
symptoms which may have been caused by an IgE-mediated
Influenza Vaccination of Persons With a History of reaction. ACIP will continue to review safety data for use of
Egg Allergy LAIV in the setting of egg allergy.
Severe allergic and anaphylactic reactions can occur in For the 2015–16 influenza season, ACIP recommends
response to various influenza vaccine components, but such the following:
reactions are rare. With the exceptions of recombinant influ- 1. Persons with a history of egg allergy who have experienced
enza vaccine (RIV3, Flublok) and cell-culture based inactivated only hives after exposure to egg should receive influenza
influenza vaccine (ccIIV3, Flucelvax, Novartis, Cambridge, vaccine. Because relatively few data are available for use
Massachusetts), currently available influenza vaccines are pre- of LAIV in this setting, IIV or trivalent recombinant
pared by propagation of virus in embryonated eggs. A 2012 influenza vaccine (RIV3) should be used. RIV3 may be
review of published data, including 4,172 egg-allergic patients used for persons aged ≥18 years who have no other
(513 reporting a history of severe allergic reaction) noted no contraindications. However, IIV (egg- or cell culture-
occurrences of anaphylaxis following administration of IIV3, based) may also be used, with the following additional
though some milder reactions did occur (30). This suggests safety measures (Figure 2):
that severe allergic reactions to egg-based influenza vaccines are – Vaccine should be administered by a health care
unlikely. On this basis, some guidance recommends that no provider who is familiar with the potential manifestations
additional measures are needed when administering influenza of egg allergy; and
vaccine to egg-allergic persons (31). However, occasional cases – Vaccine recipients should be observed for ≥30 minutes
of anaphylaxis in egg-allergic persons have been reported to for signs of a reaction after administration of each
the Vaccine Adverse Event Reporting System (VAERS) after vaccine dose.
administration of influenza vaccine (32,33). IIVs containing as 2. Persons who report having had reactions to egg involving
much as 0.7 µg/0.5 mL have reportedly been tolerated (34,35); such symptoms as angioedema, respiratory distress,
however, a threshold below which no reactions would be lightheadedness, or recurrent emesis; or who required
expected is not known (34). Among IIVs for which ovalbumin epinephrine or another emergency medical intervention,
content was disclosed during the 2011–12 through 2014–15 may receive RIV3 if they are aged ≥18 years and there
seasons, reported maximum amounts were ≤1 µg/0.5 mL dose; are no other contraindications. If RIV3 is not available
however, not all manufacturers disclose this information in or the recipient is not within the indicated age range,
the package inserts. Ovalbumin is not directly measured for IIV should be administered by a physician with
Flucelvax, but it is estimated by calculation from the initial experience in the recognition and management of severe
content in the reference virus strains to contain less than allergic conditions (Figure 2).
5x10-8 µg/0.5 mL dose of total egg protein, of which ovalbu- 3. Regardless of allergy history, all vaccines should be
min is a fraction (Novartis, unpublished data, 2013). Flublok administered in settings in which personnel and
is considered egg-free. However, neither Flucelvax nor Flublok equipment for rapid recognition and treatment of
is licensed for children aged <18 years. anaphylaxis are available (38).
Compared with IIV, fewer data are available concerning the 4. Persons who are able to eat lightly cooked egg (e.g.,
use of LAIV in the setting of egg allergy. In a prospective cohort scrambled egg) without reaction are unlikely to be
study of children aged 2 through 16 years (69 with egg allergy allergic. Egg-allergic persons might tolerate egg in baked
and 55 without), all of whom received LAIV, none of the egg- products (e.g., bread or cake). Tolerance to egg-
allergic subjects developed signs or symptoms of an allergic containing foods does not exclude the possibility of egg
reaction during the one hour of postvaccination observation, allergy. Egg allergy can be confirmed by a consistent
and none reported adverse reactions that were suggestive of medical history of adverse reactions to eggs and egg-
allergic reaction or which required medical attention after containing foods, plus skin and/or blood testing for
24 hours (36). In a larger study of 282 egg-allergic children immunoglobulin E directed against egg proteins (39).
aged 2 through 17 years (115 of whom had experienced ana- 5. For persons with no known history of exposure to egg,
phylactic reactions to egg previously), no systemic allergic reac- but who are suspected of being egg-allergic on the basis
tions were observed after LAIV administration. On the basis of previously performed allergy testing, consultation with
of these data, the upper limit of the 95% confidence interval a physician with expertise in the management of allergic
for the incidence of a systemic allergic reaction (including conditions should be obtained before vaccination
13
MMWR
Downloaded from collections.nejm.org. For personal use only. No other/ uses
August 7, 2015
permission. 823
(Figure 2). Alternatively, RIV3 may be administered if FIGURE 2. Recommendations regarding influenza vaccination of
the recipient is aged ≥18 years. persons who report allergy to eggs*† — Advisory Committee on
Immunization Practices, United States, 2015–16 influenza season
6. A previous severe allergic reaction to influenza vaccine,
regardless of the component suspected of being Can the patient eat lightly Yes Administer vaccine per
responsible for the reaction, is a contraindication to cooked egg (e.g., scrambled
usual protocol
future receipt of the vaccine. egg) without reaction?
Vaccine Selection and Timing of Vaccination for No
Immunocompromised Persons
Administer RIV3, if patient
Immunocompromised states are caused by a heterogeneous After eating eggs or aged ≥18 years
range of conditions. In many instances, limited data are avail- egg-containing foods, does Yes OR
able regarding the use of influenza vaccines in the setting of the patient experience Administer IIV; observe for
ONLY hives? reaction for at least 30
specific immunocompromised states. In general, live virus minutes after vaccination.
vaccines, such as LAIV, should not be used for persons with
most forms of altered immunocompetence (38). The Infectious No
Diseases Society of America (IDSA) has published detailed
guidance for the selection and timing of vaccines for persons After eating eggs or
with specific immunocompromising conditions, includ- egg-containing foods, does Administer RIV3, if
the patient experience patient aged ≥18 years
ing congenital immune disorders, stem cell and solid organ OR
symptoms such as
transplant, anatomic and functional asplenia, and therapeutic • cardiovascular changes If RIV3 is not available,
drug-induced immunosuppression, as well as for persons with (e.g., hypotension) or if patient is aged <18
• respiratory distress (e.g., years, IIV should be
cochlear implants or other conditions leading to persistent administered by a
wheezing) Yes
cerebrospinal fluid-oropharyngeal communication (40). ACIP • gastrointestinal physician with
will continue to review accumulating data on use of influenza symptoms (e.g., nausea experience in the
recognition and
vaccines in these contexts. or vomiting)
management of severe
• reaction requiring
1Influenza Division, National Center for Immunization and Respiratory epinephrine allergic conditions.
Diseases, CDC; 2Battelle Memorial Institute, Atlanta, Georgia; 3Immunization • reaction requiring Observe for reaction for
Safety Office, National Center for Emerging and Zoonotic Infectious Diseases, emergency medical at least 30 minutes after
CDC; 4Johns Hopkins University, Baltimore, Maryland. attention. vaccination.
Corresponding author: Lisa A. Grohskopf, lgrohskopf@cdc.gov, 404-639-2552.
Abbreviations: IIV = inactivated influenza vaccine, trivalent or quadrivalent;
Acknowledgments RIV3 = recombinant influenza vaccine, trivalent.
ACIP members (membership roster for July 2014–June 2015 is * Persons with egg allergy may tolerate egg in baked products (e.g., bread or
cake). Tolerance to egg-containing foods does not exclude the possibility of egg
available at http://www.cdc.gov/vaccines/acip/committee/members. allergy (Erlewyn-Lajeunesse et al., Recommendations for the administration of
html). ACIP Influenza Work Group; Alicia Fry, MD, Brendan influenza vaccine in children allergic to egg. BMJ 2009;339:b3680).
† For persons who have no known history of exposure to egg, but who are
Flannery, PhD, Jessie Clippard, MPH, Influenza Division, National
suspected of being egg-allergic on the basis of previously performed allergy
Center for Immunization and Respiratory Diseases, CDC; Angelia testing, consultation with a physician with expertise in the management of
Cost, PhD, Armed Forces Health Surveillance Center. allergic conditions should be obtained prior to vaccination. Alternatively, RIV3
may be administered if the recipient is aged ≥18 years.
ACIP Influenza Work Group
Ruth Karron, MD, Baltimore, Maryland (Chair); Kevin Ault, MD, References
Kansas City, Kansas; Edward Belongia, MD, Marshfield, Wisconsin; 1. Grohskopf LA, Olsen SJ, Sokolow LZ, et al. Prevention and control of
Henry Bernstein, DO, Hempstead, New York; Jeff Duchin, MD, Seattle, seasonal influenza with vaccines: recommendations of the Advisory
Washington; Janet Englund, MD, Seattle, Washington; Sandra Fryhofer, Committee on Immunization Practices (ACIP) — United States, 2014–15
MD, Atlanta, Georgia; Lee H. Harrison, MD, Pittsburgh, Pennsylvania; Lisa influenza season. MMWR Morb Mortal Wkly Rep 2014;63:691–7.
Ipp, MD, New York, New York; Wendy Keitel, MD, Houston, Texas; Marie‐ 2. CDC. Prevention and control of seasonal influenza with vaccines.
Michèle Léger, MPH, Alexandria, Virginia; Susan Lett, MD, Jamaica Plain, recommendations of the Advisory Committee on Immunization Practices—
Massachusetts; Jamie Loehr, MD, Ithaca, New York; Kathleen M. Neuzil, United States, 2013–14. MMWR Recomm Rep 2013;62(No. RR-7):1–43.
MD, Baltimore, Maryland; William Schaffner, MD, Nashville, Tennessee; 3. Ochiai H, Shibata M, Kamimura K, Niwayama S. Evaluation of the
Robert Schechter, MD, Richmond, California; Kenneth Schmader, MD, efficacy of split-product trivalent A(H1N1), A(H3N2), and B influenza
Durham, North Carolina; Tamara Sheffield, MD, Salt Lake City, Utah; vaccines: reactogenicity, immunogenicity and persistence of antibodies
Nadine Sicard, MD, Montreal, Quebec, Canada; Patricia Stinchfield, MS, following two doses of vaccines. Microbiol Immunol 1986;30:1141–9.
St. Paul, Minnesota; Matthew Zahn, MD, Santa Ana, California.
14
4. Künzel W, Glathe H, Engelmann H, Van Hoecke C. Kinetics of humoral 22. Walter EB, Neuzil KM, Zhu Y, et al. Influenza vaccine immunogenicity
antibody response to trivalent inactivated split influenza vaccine in in 6- to 23-month-old children: are identical antigens necessary for
subjects previously vaccinated or vaccinated for the first time. Vaccine priming? Pediatrics 2006;118:e570–8.
1996;14:1108–10. 23. Ambrose CS, Levin MJ, Belshe RB. The relative efficacy of trivalent live
5. Song JY, Cheong HJ, Hwang IS, et al. Long-term immunogenicity of attenuated and inactivated influenza vaccines in children and adults.
influenza vaccine among the elderly: Risk factors for poor immune Influenza Other Respir Viruses 2011;5:67–75.
response and persistence. Vaccine 2010;28:3929–35. 24. Belshe RB, Edwards KM, Vesikari T, et al.; CAIV-T Comparative Efficacy
6. Skowronski DM, Tweed SA, De Serres G. Rapid decline of influenza Study Group. Live attenuated versus inactivated influenza vaccine in
vaccine-induced antibody in the elderly: is it real, or is it relevant? J infants and young children. N Engl J Med 2007;356:685–96.
Infect Dis 2008;197:490–502. 25. Ashkenazi S, Vertruyen A, Arístegui J, et al.; CAIV-T Study Group.
7. Castilla J, Martínez-Baz I, Martínez-Artola V, et al. Network for Influenza Superior relative efficacy of live attenuated influenza vaccine compared
Surveillance in Hospitals of Navarre. Decline in influenza vaccine with inactivated influenza vaccine in young children with recurrent
effectiveness with time after vaccination, Navarre, Spain, season 2011/12. respiratory tract infections. Pediatr Infect Dis J 2006;25:870–9.
Euro Surveill 2013;18:20388. 26. Fleming DM, Crovari P, Wahn U, et al.; CAIV-T Asthma Study Group.
8. Food and Drug Administration. 138th Meeting of the Vaccines and Comparison of the efficacy and safety of live attenuated cold-adapted
Related Biological Produces Advisory Committee, March 4, 2015 influenza vaccine, trivalent, with trivalent inactivated influenza virus
[Transcript]. Washington, DC: Food and Drug Administration; 2015. vaccine in children and adolescents with asthma. Pediatr Infect Dis J
Available at http://www.fda.gov/downloads/AdvisoryCommittees/ 2006;25:860–9.
CommitteesMeetingMaterials/BloodVaccinesandOtherBiologics/ 27. Advisory Committee on Immunization Practices (ACIP). Summary
VaccinesandRelatedBiologicalProductsAdvisoryCommittee/UCM438843.pdf. Report: October 29–30, 2014 (Meeting minutes). Washington, DC:
9. bioCSL. Afluria: Influenza Vaccine [Package insert]. King of Prussia, US Department of Health and Human Services, CDC; 2014. Available
PA: bioCSL; 2015. Available at http://www.fda.gov/downloads/ at http://www.cdc.gov/vaccines/acip/meetings/downloads/min-archive/
BiologicsBloodVaccines/Vaccines/ApprovedProducts/UCM263239.pdf. min-2014-10.pdf.
10. McAllister L, Anderson J, Werth K, et al. Needle-free jet injection for 28. Advisory Committee on Immunization Practices (ACIP). Summary Report:
administration of influenza vaccine: a randomised non-inferiority trial. February 26, 2015 (Meeting minutes). Washington, DC: US Department
Lancet 2014;384:674–81. of Health and Human Services, CDC; 2015. Available at http://www.cdc.
11. Protein Sciences. Flublok: Influenza Vaccine [Package insert]. Meriden, gov/vaccines/acip/meetings/downloads/min-archive/min-2015-02.pdf.
CT: Protein Sciences; 2015. Available at http://www.fda.gov/downloads/ 29. Medimmune. Flumist Quadrivalent [Draft of package insert]. Gaithersburg,
BiologicsBloodVaccines/Vaccines/ApprovedProducts/UCM336020.pdf. MD: MedImmune; 2015 Available at http://www.fda.gov/downloads/
12. Keitel WA, Treanor JJ, El Sahly HM, et al. Comparative immunogenicity BiologicsBloodVaccines/Vaccines/ApprovedProducts/UCM294307.pdf.
of recombinant influenza hemagglutinin (rHA) and trivalent inactivated 30. Des Roches A, Paradis L, Gagnon R, Lemire C, Begin P, Carr S, et al.
vaccine (TIV) among persons ≥65 years old. Vaccine 2009;28:379–85. Egg-allergic patients can be safely vaccinated against influenza. J Allergy
13. Baxter R, Patriarca PA, Ensor K, Izikson R, Goldenthal KL, Cox MM. Clinical Immunol 2012;130:1213–6 e1.
Evaluation of the safety, reactogenicity and immunogenicity of FluBlok® 31. Kelso JM, Greenhawt MJ, Li JT; Joint Task Force on Practice Parameters
trivalent recombinant baculovirus-expressed hemagglutinin influenza (JTFPP). Update on influenza vaccination of egg allergic patients. Ann
vaccine administered intramuscularly to healthy adults 50–64 years of Allergy Asthma Immunol 2013;111:301–2.
age. Vaccine 2011;29:2272–8. 32. Advisory Committee on Immunization Practices (ACIP). Summary Report:
14. National Institutes of Health. Comparison of the safety of Flublok® June 20–21, 2012 (Meeting minutes). Washington, DC: US Department
versus licensed IIV in healthy, medically stable adults ≥50 years of age of Health and Human Services, CDC; 2012. Available at http://www.cdc.
[Internet]. Bethesda, MD: National Institutes of Health; 2015. Clinical gov/vaccines/acip/meetings/downloads/min-archive/min-jun12.pdf.
study record no. NC0125200. Available from: https://clinicaltrials.gov/ 33. Advisory Committee on Immunization Practices (ACIP). Summary Report:
ct2/show/NCT01825200. June 19–20, 2013 (Meeting minutes). Washington, DC: US Department
15. Sanofi Pasteur. Fluzone Intradermal Quadrivalent vaccine [Package of Health and Human Services, CDC; 2013. Available at http://www.cdc.
insert]. Swiftwater, PA: Sanofi Pasteur; 2015. Available at http://www. gov/vaccines/acip/meetings/downloads/min-archive/min-jun13.pdf.
f d a . g o v / d o w n l o a d s / B i o l o g i c s B l o o d Va c c i n e s / Va c c i n e s / 34. Kelso JM, Greenhawt MJ, Li JT, et al. Adverse reactions to vaccines practice
ApprovedProducts/UCM426679.pdf. parameter 2012 update. J Allergy Clin Immunol 2012;130:25–43.
16. Gorse GJ, Falsey AR, Ozol-Godfrey A, Landolfi V, Tsang PH. Safety 35. Owens G, MacGinnitie A. Higher-ovalbumin-content influenza vaccines
and immunogenicity of a quadrivalent intradermal influenza vaccine in are well tolerated in children with egg allergy. J Allergy Clin Immunol
adults. Vaccine 2015;33:1151–9. 2011;127:264–5.
17. Neuzil KM, Jackson LA, Nelson J, et al. Immunogenicity and reactogenicity 36. Des Roches A, Samaan K, Graham F, et al. Safe vaccination of patients
of 1 versus 2 doses of trivalent inactivated influenza vaccine in vaccine- with egg allergy by using live attenuated influenza vaccine. J Allergy Clin
naive 5–8-year-old children. J Infect Dis 2006;194:1032–9. Immunol Pract 2015;3:138–9.
18. Allison MA, Daley MF, Crane LA, et al. Influenza vaccine effectiveness 37. Turner PJ, Southern J, Andrews NJ, Miller E, Erlewyn-Lajeunesse M,
in healthy 6- to 21-month-old children during the 2003-2004 season. SNIFFLE Study Investigators. Safety of live attenuated influenza vaccine
J Pediatr 2006;149:755–62. in atopic children with egg allergy. J Allergy Clinical Immunol 2015;
19. Ritzwoller DP, Bridges CB, Shetterly S, Yamasaki K, Kolczak M, France pii: S0091–6749(15)00005–6.
EK. Effectiveness of the 2003–2004 influenza vaccine among children 6 38. CDC. General recommendations on immunization—recommendations
months to 8 years of age, with 1 vs 2 doses. Pediatrics 2005;116:153–9. of the Advisory Committee on Immunization Practices (ACIP). MMWR
20. Englund JA, Walter EB, Gbadebo A, Monto AS, Zhu Y, Neuzil KM. Recomm Rep 2011;60(No. RR-2):1–64.
Immunization with trivalent inactivated influenza vaccine in partially 39. Erlewyn-Lajeunesse M, Brathwaite N, Lucas JS, Warner JO.
immunized toddlers. Pediatrics 2006;118:e579–85. Recommendations for the administration of influenza vaccine in children
21. Englund JA, Walter EB, Fairchok MP, Monto AS, Neuzil KM. A allergic to egg. BMJ 2009;339:b3680.
comparison of 2 influenza vaccine schedules in 6- to 23-month-old 40. Rubin LG, Levin MJ, Ljungman P, et al.; Infectious Diseases Society of
children. Pediatrics 2005;115:1039–47. America. 2013 IDSA clinical practice guideline for vaccination of the
immunocompromised host. Clin Infect Dis 2014;58:e44–100.
15
MMWR
August 7, 2015
permission. 825
Circulating Influenza Viruses Show Good Match with Coming Season’s Vaccine
Physician’s First Watch, published September 18, 2015
NEWS SUMMARY
Influenza viruses circulating this summer are similar to the components of the 2015–2016 flu vaccines and are
sensitive to neuraminidase inhibitors, according to CDC and WHO surveillance data in MMWR.
In the U.S. from May 24 to Sept. 5, some 54% of isolates were influenza A (95% subtype H3N2, 5% pH1N1) and
46% were influenza B. Three cases of influenza A variant virus infections (H3N2v and H1N1v) were reported,
but no human-to-human transmission was detected.
Antigenic and genetic analysis showed that all the circulating influenza A and B viruses were similar to compo-
nents of the 2015–2016 vaccines for the Northern Hemisphere. (WHO recommended these components in February.)
All viruses were sensitive to oseltamivir, zanamivir, and peramivir. Influenza A viruses were resistant to amanta-
dine and rimantadine, which are not currently recommended treatments.
— Cara Adler
16
new england
Source: The New England Journal of Medicine
The
journal of medicine ORIGINAL ARTICLE
established in 1812 december 26, 2013 vol. 369 no. 26
Vaccine for Prevention of Mild and Moderate-to-Severe

Influenza in Children
Varsha K. Jain, M.D., M.P.H., Luis Rivera, M.D., Khalequ Zaman, M.B., B.S., Ph.D., Roberto A. Espos, Jr., M.D., M.H.S.A.,
Chukiat Sirivichayakul, M.D., Beatriz P. Quiambao, M.D., Doris M. Rivera-Medina, M.D., Pirunghul Kerdpanich, M.D.,
Mehmet Ceyhan, M.D., Ener C. Dinleyici, M.D., Alejandro Cravioto, M.D., Ph.D., Mohammed Yunus, M.B., B.S.,
Pornthep Chanthavanich, M.D., Kriengsak Limkittikul, M.D., Zafer Kurugol, M.D., Ph.D., Emre Alhan, M.D.,
Adrian Caplanusi, M.D., Ph.D., Serge Durviaux, B.A., Philippe Boutet, D.V.M., Ph.D., Opokua Ofori-Anyinam, Ph.D.,
Vijayalakshmi Chandrasekaran, M.Sc., Ghassan Dbaibo, M.D., and Bruce L. Innis, M.D.
A bs t r ac t
Background
Commonly used trivalent vaccines contain one influenza B virus lineage and may be The authors’ affiliations are listed in the
ineffective against viruses of the other B lineage. We evaluated the efficacy of a can- Appendix. Address reprint requests to Dr.
Dbaibo at the Department of Pediatrics
didate inactivated quadrivalent influenza vaccine (QIV) containing both B lineages. and Adolescent Medicine, American Uni-
versity of Beirut Medical Center, Rm. C-649,
Methods
Cairo St., Beirut, Lebanon 11-0236, or at
In this multinational, phase 3, observer-blinded study, we randomly assigned children gdbaibo@aub.edu.lb.
3 to 8 years of age, in a 1:1 ratio, to receive the QIV or a hepatitis A vaccine (control).
This article was published on December 11,
The primary end point was influenza A or B confirmed by real-time polymerase chain 2013, at NEJM.org.
reaction (rt-PCR). Secondary end points were rt-PCR–confirmed, moderate-to-severe
N Engl J Med 2013;369:2481-91.
influenza and rt-PCR–positive, culture-confirmed influenza. The vaccine efficacy and DOI: 10.1056/NEJMoa1215817
the effect of vaccination on daily activities and utilization of health care resources Copyright © 2013 Massachusetts Medical Society.
were assessed in the total vaccinated cohort (2584 children in each group) and the
per-protocol cohort (2379 children in the QIV group and 2398 in the control group). Related Material:
Supplementary Appendix
Results
Editorial
In the total vaccinated cohort, 62 children in the QIV group (2.40%) and 148 in the
Correspondence
control group (5.73%) had rt-PCR–confirmed influenza, representing a QIV efficacy
of 59.3% (95% confidence interval [CI], 45.2 to 69.7), with efficacy against culture-
confirmed influenza of 59.1% (97.5% CI, 41.2 to 71.5). For moderate-to-severe rt-PCR–
confirmed influenza, the attack rate was 0.62% (16 cases) in the QIV group and
2.36% (61 cases) in the control group, representing a QIV efficacy of 74.2% (97.5%
CI, 51.5 to 86.2). In the per-protocol cohort, the QIV efficacy was 55.4% (95% CI,
39.1 to 67.3), and the efficacy against culture-confirmed influenza 55.9% (97.5% CI,
35.4 to 69.9); the efficacy among children with moderate-to-severe influenza was
73.1% (97.5% CI, 47.1 to 86.3). The QIV was associated with reduced risks of a body
temperature above 39°C and lower respiratory tract illness, as compared with the
control vaccine, in the per-protocol cohort (relative risk, 0.29 [95% CI, 0.16 to 0.56]
and 0.20 [95% CI, 0.04 to 0.92], respectively). The QIV was immunogenic against all
four strains. Serious adverse events occurred in 36 children in the QIV group (1.4%)
and in 24 children in the control group (0.9%).
Conclusions
The QIV was efficacious in preventing influenza in children. (Funded by
GlaxoSmithKline Biologicals; ClinicalTrials.gov number, NCT01218308.)
n engl j med 369;26 nejm.org december 26, 2013 2481
17
The n e w e ng l a n d j o u r na l of m e dic i n e
T
he incidence of influenza among Appendix) collaborated on writing all drafts of
children is high, and the illness is associ- the manuscript, with assistance from a professional
ated with substantial increases in out- medical writer, who was paid by the sponsor. All
patient visits and hospitalizations during the in- the authors assume responsibility for the accuracy
fluenza season.1-4 Routine vaccination of children and completeness of the data and for the fidelity
against influenza is recommended in the United of the study to the protocol, which is available at
States5 and some other countries, despite limited NEJM.org.
evidence of the efficacy of inactivated influenza
vaccine from randomized, controlled trials in- Participants
volving children.6 We recruited healthy children from 15 centers in
When trivalent influenza vaccines (TIVs) are Bangladesh, the Dominican Republic, Honduras,
used, there is a possibility of a mismatch be- Lebanon, Panama, the Philippines, Thailand,
tween circulating and vaccine B strains, which and Turkey (see the Supplementary Appendix).
results in inadequate protection from the vac- Children were randomly assigned, in a 1:1 ratio,
cine.7-10 A quadrivalent vaccine containing both to receive the QIV (0.5-ml dose) or hepatitis A vac-
B lineages would eliminate B-lineage mismatch. cine (Havrix, 0.5-ml dose), as a control (see the Sup-
This may be particularly important in children, plementary Appendix). Both vaccines were manu-
in whom TIVs elicit weak cross-reactive antibody factured by GlaxoSmithKline Vaccines. The QIV
responses and have reduced efficacy against in- contained 15 μg of hemagglutinin antigen from
fluenza B caused by strains of the lineage that each of four strains: A/California/7/2009 (H1N1),
are not contained in the vaccine.10-12 A/Victoria/210/2009 (H3N2), B/Brisbane/60/2008
In the current study, we evaluated the efficacy (Victoria), and B/Florida/4/2006 (Yamagata). Chil-
of a candidate inactivated quadrivalent influenza dren received one or two vaccine doses depend-
vaccine (QIV) for the prevention of influenza A ing on their vaccine priming status (Fig. S1 in the
or B in children 3 to 8 years of age, using a con- Supplementary Appendix; see the Methods sec-
ventional end point (any influenza) and an addition in the Supplementary Appendix for the crite-
tional end point (moderate-to-severe influenza) ria used to define priming).
that captures the more clinically significant out-
comes of influenza. Study Design
The first children were enrolled in December 2010;
Me thods the exact date varied among countries. Active and
passive surveillance for influenza-like illness were
Study Oversight conducted for at least 6 months (Fig. S1 in the Sup-
We conducted a phase 3, randomized, con- plementary Appendix), continuing until the end
trolled, observer-blind study with funding from of October 2011. An influenza-like illness was
GlaxoSmithKline Biologicals. The sponsor was defined as a temperature of 37.8°C or higher, with
involved in all stages of the study conduct and at least one of the following: cough, sore throat,
analysis of the data. The aim of the study was to runny nose, or nasal congestion. The parents of
assess the efficacy of the QIV for prevention of children with an influenza-like illness completed
influenza A or B in children 3 to 8 years of age. a daily diary for 14 days starting from the onset
The presence of influenza virus was to be con- of the illness (see the Supplementary Appendix).
firmed by means of a real-time polymerase-
chain-reaction (rt-PCR) assay. The trial received Detection of Influenza Virus
ethical approval at each participating center (see The presence of influenza A or B virus in nasal or
the Supplementary Appendix, available with the throat swabs from children with an influenza-
full text of this article at NEJM.org). Parents or like illness was confirmed by means of an rt-PCR
legal representatives provided written informed assay,13 and positive samples were further tested
consent, and children 7 years of age or older pro- by means of cell culturing.14 Influenza A subtyping
vided written assent, according to local standards. was performed with the use of a nested reverse-
The sponsor donated the vaccines. Members of transcriptase (RT)–PCR assay on clinical specimens,
the core writing team (see the Supplementary with primers targeting the hemagglutinin gene
2482 n engl j med 369;26 nejm.org

18 december 26, 2013
Vaccine for Prevention of Influenza in Children
for H1 and H3 (Table S1 in the Supplementary tion; and serious adverse events and adverse
Appendix). Classification of influenza B as Yama- events for which medical care was sought, as-
gata or Victoria lineage was also performed by sessed over the course of the whole study period.
means of a nested RT-PCR assay with primers tar-
geting hemagglutinin (Table S1 in the Supplemen- Statistical Analysis
tary Appendix), followed by sequencing analysis. The principal analysis was an analysis of efficacy
Antigenic matching was performed on cultured in the per-protocol cohort (children who met all
isolates and was defined as a difference by less eligibility criteria, were successfully contacted at
than a factor of 8 in hemagglutination inhibition least once after vaccination, and adhered to the
relative to reference serum and vaccine antigen. protocol); efficacy was also analyzed in the total
vaccinated cohort. Immunogenicity was evaluated
Study End Points in a subgroup of the per-protocol cohort, and
Efficacy safety was evaluated in the total vaccinated co-
The primary end point was influenza A or B of hort (Fig. S2 in the Supplementary Appendix).
any severity that was confirmed by means of The efficacy of the QIV was assessed with the
rt-PCR assay. Secondary end points were rt-PCR– use of time-to-event methods by means of a Cox
confirmed, moderate-to-severe influenza A or B regression model adjusted for age, region, and
and culture-confirmed influenza A or B caused priming status, with 95% confidence intervals
by seasonal strains antigenically matching the for the primary end point and exploratory end
vaccine strains or by any seasonal strain. Moder- points and 97.5% confidence intervals for sec-
ate-to-severe disease was defined as a body tem- ondary end points (see the Supplementary Ap-
perature higher than 39°C, physician-confirmed pendix). The criterion for declaring efficacy was
acute otitis media, lower respiratory tract illness a lower limit of the confidence interval of more
(shortness of breath, pulmonary congestion, than 30% for the primary end point (with the
pneumonia, bronchiolitis, bronchitis, wheezing, criterion affirmed by the Food and Drug Admin-
or croup), or serious extrapulmonary complica- istration) and more than 0% for secondary end
tions such as myositis, encephalitis, seizure, or points. The principal analysis included cases
myocarditis. Exploratory end points were the in- identified from 14 days after vaccination to the
cidence of influenza caused by individual A sub- end of the study; a complementary analysis in-
types or B lineages, the incidence of influenza in cluded cases identified from day 0.
two age subgroups (3 to 4 years and 5 to 8 years), Assuming a QIV efficacy of 60%, we esti-
clinical manifestations of moderate-to-severe in- mated that approximately 194 cases of rt-PCR–
fluenza, and the effect of vaccination on daily confirmed influenza A and B would need to be
activities and utilization of health care resources. documented in order to show a lower limit of
the 95% confidence interval for the vaccine ef-
Immunogenicity and Safety ficacy that was more than 30% at a 5% level of
The hemagglutination-inhibition antibody titer significance. On the basis of an estimated attack
against each vaccine strain was measured with rate of 6%, and assuming that we would not be
the use of standard methods.15 The immune re- able to evaluate approximately 10% of the chil-
sponse was assessed in a randomly chosen im- dren, we calculated that we would need to re-
munogenicity subgroup that comprised 544 chil- cruit approximately 5200 children. The statisti-
dren in the QIV group and 163 in the control cal analysis was performed with the use of SAS
group in the total vaccinated cohort and 457 chil- software, version 9.2 (SAS Institute).
dren in the QIV group and 122 in the control
group in the per-protocol cohort. The following R e sult s
safety end points were evaluated by means of a
review of the diary card: injection-site and sys- Participants
temic symptoms (solicited information), assessed A total of 5220 children were enrolled in the
during the 7-day period after the vaccination; study: 5168 children in the total vaccinated co-
spontaneously reported (unsolicited) symptoms, hort (2584 in the QIV group and 2584 in the con-
assessed during the 28-day period after vaccina- trol group) and 4777 in the per-protocol efficacy
n engl j med 369;26 nejm.org

19 december 26, 2013 2483
cohort (2379 in the QIV group and 2398 in the ond dose (2 children), unavailability during the
control group) (Fig. S2 in the Supplementary Ap- scheduled window for vaccination defined by the
pendix). The mean age was 5.4 years in both the protocol (1 child), parental decision with no rea-
QIV and control groups, with approximately son given (1 child), and investigator decision ow-
equal numbers of boys and girls (Table S2 in the ing to vaccination history (1 child) (Table S3 in
Supplementary Appendix). Five children who had the Supplementary Appendix).
not undergone influenza vaccine priming and
who had received one dose of vaccine during the Influenza-like Illness
study did not receive the second vaccine dose be- In the total vaccinated cohort, 563 influenza-like
cause of illness at the scheduled time of the sec- illnesses occurred in 422 children in the QIV
group and 657 influenza-like illnesses occurred
in 507 children in the control group during the
A Distribution of Influenza Strain According to Country period from day 0 after vaccination until the end
70
B/Yamagata of the study. A swab specimen was obtained in
60 B/Victoria 96% of the cases of influenza-like illness and
50
A/H3N2 was obtained within 7 days after the onset of ill-
31 A/H1N1 ness in all but 1 case. A total of 62 children in the
No. of Children
8
40
23 QIV group (2.4%) and 148 in the control group
30 2 (5.7%) had rt-PCR–confirmed influenza (Fig. 1).
20 4
1
38
In the per-protocol efficacy cohort, 462 influenza-
20
13 like illnesses occurred in 352 children in the QIV
25
10
14
20 group and 531 occurred in 416 children in the
8 1 0
0 4 2 control group during the period from 14 days af-
ter vaccination until the end of the study; 58 chil-
pu an
on c
om h
es
ey
nd
i
ra
m
s
no
bl
Re inic
rk
in
de
du
la
na
ba
Tu
pp
dren (2.4%) and 128 (5.3%), in the two groups,

la
ai
Pa
Le
ng
Th
ili
H
Ph
D
Ba
respectively, had rt-PCR–confirmed influenza.

B Overall Distribution of Influenza Strain Three children in each group in the per-protocol
B/Yamagata cohort had two rt-PCR–confirmed infections
2 (0.9%) each; the second infections were not included in
the analysis of vaccine efficacy.
Efficacy
The study met its primary objective by showing
A/H1N1
66 (30.8%) that the efficacy of the QIV against any case of
B/Victoria
76 (35.5%) rt-PCR–confirmed influenza was 59.3% (95%
confidence interval [CI], 45.2 to 69.7) in the total
vaccinated cohort and 55.4% (95% CI, 39.1 to 67.3)
in the per-protocol cohort (Table 1, and Fig. S3a
A/H3N2
70 (32.7%)
and S4a in the Supplementary Appendix). The at-
tack rate was 2.40% (62 cases) with the QIV and
5.73% (148 cases) with the control vaccine in the
total vaccinated cohort; the corresponding rates
Figure 1. Distribution of Influenza A Subtype and Influenza B Lineage. in the per-protocol cohort were 2.44% (58 cases)
The distribution of four strains of influenza confirmed by means of a real-time and 5.34% (128 cases) (Table 1). Our intention in
polymerase-chain-reaction (rt-PCR) assay is shown according to country evaluating the end point of moderate-to-severe
(Panel A) and overall (Panel B), with data from the total vaccinated cohort influenza was to dichotomize cases into catego-
during the period from day 0 after vaccination until the end of the study. A total ries of mild illness versus illnesses that are as-
of 210 children had 216 cases of rt-PCR–confirmed influenza; 204 children
sociated with the most clinically significant
had a single case and 6 children had two cases each. In 4 of these 6 children,
the two cases were associated with different influenza strains, and in 2 chil- aspects of influenza. We observed higher efficacy
dren only one case could be typed. The graph does not include the two cases of the QIV against moderate-to-severe influenza
that could not be typed, and therefore shows 214 cases. than against influenza of any severity. The effi-
cacy against moderate-to-severe influenza in the
2484 n engl j med 369;26 20

nejm.org december 26, 2013
Table 1. Vaccine Efficacy against rt-PCR–Confirmed and Culture-Confirmed Influenza A or B According to Age and A Subtype and B Lineage.*
Cohort and Influenza Variable QIV Group Control Group QIV Efficacy
Cases Attack Rate Cases Attack Rate

no. % no. % % (95% CI)
Total vaccinated cohort
rt-PCR–confirmed influenza, any severity 62 2.40 148 5.73 59.3 (45.2 to 69.7)
rt-PCR–confirmed influenza, moderate-to-severe 16 0.62 61 2.36 74.2 (51.5 to 86.2)†
Culture-confirmed, rt-PCR–confirmed influenza, 54 2.09 129 4.99 59.1 (41.2 to 71.5)†
any severity, any seasonal strain
any severity, vaccine-matched strain
Per-protocol efficacy cohort
rt-PCR–confirmed influenza, any severity 58 2.44 128 5.34 55.4 (39.1 to 67.3)
rt-PCR–confirmed influenza, moderate-to-severe 14 0.59 52 2.17 73.1 (47.1 to 86.3)†
any severity, any seasonal strain
any severity, vaccine-matched strain
Per-protocol efficacy cohort stratified by age‡
rt-PCR–confirmed influenza, any severity
Children 3–4 yr of age 32 3.78 48 5.69 35.3 (−1.3 to 58.6)
Children 5–8 yr of age 26 1.70 80 5.15 67.7 (49.7 to 79.2)
rt-PCR–confirmed influenza, moderate-to-severe
Per-protocol efficacy cohort stratified by strain§
rt-PCR–confirmed influenza, any severity
Influenza A 37 1.56 85 3.54 56.8 (36.4 to 70.6)
A/H1N1 17 0.71 38 1.58 55.6 (21.3 to 74.9)
A/H3N2 20 0.84 47 1.96 57.6 (28.5 to 74.9)
Influenza B 23 0.97 45 1.88 49.5 (16.6 to 69.5)
B/Victoria 23 0.97 43 1.79 47.2 (12.4 to 68.2)
B/Yamagata 0 0.00 2 0.08 100.0 (— to 100.0)
rt-PCR–confirmed influenza, moderate-to-severe
Influenza A 8 0.34 40 1.67 79.9 (57.1 to 90.6)
A/H1N1 4 0.17 17 0.71 76.5 (30.3 to 92.1)
A/H3N2 4 0.17 23 0.96 82.4 (49.1 to 93.9)
Influenza B 7 0.29 13 0.54 46.5 (34.1 to 78.7)
B/Victoria 7 0.29 12 0.50 42.1 (47.1 to 77.2)
B/Yamagata 0 0.00 1 0.04 100.0 (— to 100.0)
* The total vaccinated cohort included 2584 children in the quadrivalent influenza vaccine (QIV) group and 2584 in the control group. The
incidence of influenza was assessed from day 0 after vaccination to the end of the study. The per-protocol cohort included 2379 children in
the QIV group and 2398 in the control group. The incidence of influenza was assessed from day 14 after vaccination to the end of the study.
Rt-PCR denotes real-time polymerase chain reaction.
† The confidence interval in this category is a 97.5% confidence interval.
‡ In the QIV group, 846 children were 3 or 4 years of age, and 1533 were 5 to 8 years of age; in the control group, 844 children were 3 or 4 years
of age, and 1554 were 5 to 8 years of age. The analysis of vaccine efficacy according to age was a protocol-specified exploratory analysis.
§ The analysis of vaccine efficacy according to A subtype and B lineage was a post hoc exploratory analysis.
n engl j med 369;26 nejm.org december

21 26, 2013 2485
Source: The New England Journal of Medicine T h e n e w e ng l a n d j o u r na l of m e dic i n e
total vaccinated cohort was 74.2% (97.5% CI, 51.5 confirmed, moderate-to-severe influenza, 8 chil-
to 86.2), with attack rates of 0.62% (16 cases) in dren in the QIV group and 35 in the control
the QIV group and 2.36% (61 cases) in the con- group missed school (relative risk with QIV, 0.23
trol group (Table 1, and Fig. S3b in the Supple- [95% CI, 0.11 to 0.49]), and 1 child in the QIV
mentary Appendix). The corresponding efficacy group and 4 in the control group were hospital-
in the per-protocol cohort was 73.1% (97.5% CI, ized (relative risk, 0.25 [95% CI, 0.03 to 2.25])
47.1 to 86.3), with an attack rate of 0.59% (Table 2, and Table S8 in the Supplementary Ap-
(14 cases) in the QIV group and 2.17% (52 cases) pendix). The results were similar in the total
in the control group (Table 1, and Fig. S4b in the vaccinated cohort (Tables S7 and S9 in the Sup-
Supplementary Appendix). The QIV vaccine was plementary Appendix).
also effective against rt-PCR–positive, culture-
confirmed disease caused by any seasonal or Immune Response
vaccine-matched strains (Table 1), with higher In the immunogenicity subgroup of the per-
efficacy against moderate-to-severe disease than protocol cohort, the geometric mean titer after
against influenza of any severity. However, most administration of the QIV was 10 to 20 times the
H3N2 isolates could not be antigenically typed prevaccination titer against all four vaccine
for matching to the QIV (Table S4 in the Supple- strains (Fig. 2). After vaccination, the seroprotec-
mentary Appendix). tion rate against each strain was more than 95%
In an exploratory analysis, we assessed QIV (Fig. S6 in the Supplementary Appendix). Immu-
efficacy according to influenza A subtype and nogenicity was similar in the two age groups. At
influenza B lineage. The QIV was most effective 6 months and later after vaccination, the sero-
against moderate-to-severe influenza A (Table 1). protection rate against the A/H3N2 and B/Yama-
Only two cases of influenza B caused by Yama- gata strains was more than 90%, and the sero-
gata lineage viruses were observed, both in the protection rate against the A/H1N1 and B/Victoria
control group. Efficacy against influenza of any strains was more than 80% (Fig. S6 in the Supple-
severity appeared to be lower among children 3 to mentary Appendix). The findings were similar in
4 years of age than among children 5 to 8 years of the immunogenicity subgroup of the total vacci-
age in the per-protocol cohort (35.3% [95% CI, nated cohort.
−1.3 to 58.6] vs. 67.7% [95% CI, 49.7 to 79.2])
(Table 1). However, efficacy against moderate-to- Safety
severe disease appeared to be similar among There were no notable differences between the
children in the two age groups (67.5% [95% CI, QIV group and the control group with respect to
18.0 to 87.1] and 76.2% [95% CI, 48.5 to 89.0], safety end points, except that pain at the injec-
respectively). An unadjusted per-protocol analysis tion site was reported more frequently in the QIV
and analysis of the total vaccinated cohort showed group (Table 3, and Table S10 in the Supplemen-
similar levels of efficacy (Tables S5 and S6 and tary Appendix). Serious adverse events occurred
Fig. S5 in the Supplementary Appendix). in 36 children in the QIV group (1.4%) and in
High body temperature and lower respiratory 24 children in the control group (0.9%) (Table 3,
tract illness were the most common symptoms as- and Table S11 in the Supplementary Appendix).
sociated with moderate-to-severe influenza; both
symptoms were reported in fewer children in the Discussion
QIV group than in the control group (Table 2, and
Table S7 in the Supplementary Appendix). Few There is limited evidence from randomized trials
children had acute otitis media or pneumonia to support the administration of inactivated in-
(one and three cases, respectively, in the control fluenza vaccine in healthy children.6 This ran-
group), and none had an extrapulmonary com- domized study provides additional evidence of
plication of influenza. the efficacy of QIV against influenza (as con-
The QIV reduced the effect of illness on daily firmed by means of rt-PCR assay). The study was
activities, parental time away from work, and performed to provide direct evidence of the clini-
utilization of health care resources, particularly cal benefit of the vaccine, especially in the pre-
in cases of moderate-to-severe disease. In the vention of moderate-to-severe influenza, an end
per-protocol cohort, of 66 children with rt-PCR– point that captures the most clinically significant

Table 2. Clinical Characteristics of rt-PCR–Confirmed Influenza and Effect on Daily Activities.*
Variable Moderate-to-Severe Disease Mild Disease
QIV Control Relative Risk QIV Control Relative Risk

Group Group with QIV (95% CI) Group Group with QIV (95% CI)
Clinical characteristics
No. of rt-PCR–confirmed cases 14 52
Temperature >39°C with no other 12 41 0.29 (0.16–0.56)
defining symptom (no. of
patients)
Acute otitis media with no other defining 0 1 —
symptom (no. of patients)
Lower respiratory tract illness (no. 2 10 0.20 (0.04–0.92)
of patients)†
With pneumonia‡ 0 3 —
Without pneumonia 2 7 0.29 (0.06–1.38)
Effect on daily activities and utilization
of health care resources
No. of rt-PCR–confirmed cases§ 14 52 45 76
Absence from school
No. of children affected 8 35 0.23 (0.11–0.49)¶ 23 41 0.56 (0.34–0.94)¶
No. of school days missed per illness 4.1±3.1 5.0±3.7 — 3.6±2.5 3.4±1.9 —
Parental absence from work to take care
of the child
No. of parents affected 7 18 0.39 (0.16–0.93)¶ 18 18 1.01 (0.52–1.93)¶
No. of work days missed per illness 3.1±2.0 4.2±3.4 — 2.5±1.6 4.8±4.6 —
Visit to doctor or other medical person
No. of children affected 14 45 0.31 (0.17–0.57) 35 65 0.54 (0.36–0.81)
No. of doctor visits per illness 2.3±1.8 1.8±1.7 — 1.4±1.1 1.4±1.0 —
Hospitalization
No. of children affected 1 4 0.25 (0.03–2.25) 0 1 —
No. of days in hospital per illness 5.0 4.8±2.9 — — 4.0 —
* Plus–minus values are means ±SD. Data are from the per-protocol efficacy cohort, which comprised 2379 children in the QIV group and
2398 in the control group. A total of 12 children (3 in the QIV group and 9 in the control group) were excluded from the analysis because
they received a medication or vaccine that was not allowed by the protocol. The analysis of clinical characteristics of the illness was a post
hoc exploratory analysis of data from day 14 after vaccination until the end of the study. The analysis of the effect on daily activities and
utilization of health care resources was a protocol-specified exploratory analysis of data from day 0 to day 14 of the illness.
† Lower respiratory tract illness included bronchitis, wheezing, shortness of breath, pneumonia, and pulmonary congestion.
‡ Two of the three cases of pneumonia were diagnosed by means of chest radiography, but no causative organism was identified. No infor-
mation is available for the third case, which was a spontaneously reported adverse event, with no explanation of the method of diagnosis
and no description of symptoms.
§ One child had one episode of rt-PCR–confirmed moderate-to-severe influenza and one episode of rt-PCR–confirmed mild influenza and is
included in both categories.
¶ Data were included only when both parents worked outside the home.
outcomes leading to health care consultations or disease, high temperature, or earache.16 Studies
hospitalization. The greatest value of vaccination that do not differentiate these manifestations of
is in the prevention of clinically significant dis- influenza from mild illness cannot assess the ef-
ease rather than mild upper respiratory tract ill- fectiveness of the vaccine in attenuating illness
ness. Parents are most likely to seek medical help and therefore may undervalue its benefit.
for children with a respiratory tract infection Our study showed QIV efficacy of 55% against
who have symptoms of lower respiratory tract influenza of any severity. This is similar to the

23 26, 2013 2487
1000
QIV Control
361.5
318.8
264.7 239.9
Geometric Mean Titer
138.6 157.0
136.5
110.2
100
43.6
28.6 30.3
23.2 24.3 21.8
19.7 18.8 19.2
17.8
16.1 16.1 15.6 16.2
15.3
13.7
10
Baseline 1 End of Baseline 1 End of Baseline 1 End of Baseline 1 End of
Month study Month study Month study Month study
A/H1N1 A/H3N2 B/Victoria B/Yamagata
Figure 2. Immunogenicity According to Strain.

Hemagglutination-inhibition antibody titers against influenza A subtype and B lineage in the per-protocol cohort
are shown for the period before vaccination, 1 month after administration of a single vaccine dose in children who
had influenza priming or 1 month after the second dose in children who did not have influenza priming, and at the
end of the study (6 to 8 months after first dose). (For the definition of priming in the study, see the Supplementary
Appendix.) Data are expressed as geometric mean titers (the antilog of the arithmetic mean of the log10 -transformed
titers). Data were available before vaccination and 1 month after vaccination for 457 children in the group that re-
ceived the quadrivalent influenza vaccine (QIV) and 122 in the group that received the control vaccine (a hepatitis A
vaccine) and were available at the end of the study for 444 children in the QIV group and 115 in the control group.
Children with a hemagglutination-inhibition antibody titer of less than 1:10 (assay cutoff value) were considered to
be seronegative. I bars indicate 95% confidence intervals.
estimates of the efficacy of TIV in other random- the control group), precluding a meaningful es-
ized trials: 43% (among children 6 months to timate of the vaccine efficacy against that strain.
6 years of age),17 51% (among children 18 months However, the immune response to B Yamagata
to 6 years of age),18 and 56% (among children 3 to was as high as the response to the other strains,
9 years of age, against H1N1 disease),19 as well suggesting that the efficacy of the QIV against
as an estimate of 48% from a meta-analysis of B Yamagata may be similar to the efficacy
data for children of all ages.20 In our study, the against the other strains. Vaccine efficacy is sub-
efficacy of QIV was higher against moderate-to- stantially reduced when the vaccine B strain is
severe disease (approximately 70% overall and in mismatched to the circulating strain.7-10 Only in
each of the two age groups). Most breakthrough the Philippines were both B lineages detected,
cases in the QIV group were of mild severity. In reflecting the unpredictable geographic variabil-
a post hoc analysis, the QIV was associated with ity of influenza virus circulation. Introduction of
an 80% reduction in the rate of lower respiratory the QIV is expected to result in a modestly lower
tract illness (the most common serious outcome incidence of influenza-related outcomes than
of influenza) and a 70% reduction in the rate of that seen with TIV, with the net effect varying
body temperature above 39°C, as compared with from one season to another.21 The true value of
the control vaccine. the QIV will be seen in years when the two B
Our study provided evidence that the QIV lineages are cocirculating or if there is a shift
prevents influenza associated with the A/H1N1, from one lineage to another between the time
A/H3N2, and B Victoria strains individually. the vaccine is developed and the beginning of
Only two rt-PCR–confirmed cases associated the influenza season.
with the B Yamagata virus were seen (both in Influenza in children results in increased out-

patient visits, hospitalizations, and days missed

Table 3. Safety Outcomes in the Total Vaccinated Cohort.*
from school.1-3,16,22,23 Among children with
moderate-to-severe disease, the QIV, as compared QIV Group Control Group
with the control vaccine, was associated with Outcome N = 2584 N = 2584
69% fewer medical visits, 75% fewer hospitaliza- no. of children (%)
tions, 77% fewer absences from school, and 61% Injection-site symptom during 7-day
fewer parental absences from work. In all the postvaccination period†
study countries, the school year includes most of Pain 1215 (47.7) 888 (34.8)
the peak influenza season. The age at which Redness 17 (0.7) 5 (0.2)
children begin day care or school varies consid- Swelling 46 (1.8) 10 (0.4)
erably among countries, ranging from 3 to 7 Symptom during 28-day postvaccination
years of age. The proportion of parents working period‡
outside the home is also variable. The estimate Any 843 (32.6) 855 (33.1)
of the effect of vaccination with the QIV on Related to vaccine 30 (1.2) 37 (1.4)
school and workplace absenteeism is thus a Medically attended event during entire
broad approximation. In countries outside the study period§
study where children begin attending day care or Any 792 (30.7) 749 (29.0)
school at an earlier age, a higher proportion of Related to vaccine 6 (0.2) 13 (0.5)
parents work outside the home, or both, the ef- Serious adverse event during entire
fect of vaccination on absenteeism may be study period¶
greater than is indicated here. Any 36 (1.4) 24 (0.9)
The safety profile of the QIV was similar to Related to vaccine‖ 1 (<0.1) 0
that of hepatitis A vaccine, which we used as a Grade 3** 3 (0.1) 1 (<0.1)
control, except that pain at the injection site was
reported more frequently in the QIV group. The * Safety end points were analyzed descriptively. No significant difference was
observed between the incidence of adverse events after the first dose of vac-
experience of injection-site pain with the QIV as cine and the incidence after the second dose.
compared with TIV has been variable, with some † The percentages were calculated on the basis of 2546 children in the QIV
studies suggesting similar levels of pain with the group and 2551 children in the control group whose parents or guardians
completed a symptom diary. Reports of injection-site symptoms were so-
QIV and TIV and others suggesting a modestly licited (i.e., specific questions were included in the daily diary parents were
higher level of pain with the QIV.24-27 A higher to complete). All injection-site symptoms were considered to be related to
incidence of fever resulting from inclusion of a vaccination.
‡ Symptoms during the 28-day period after vaccination were reported sponta-
fourth strain was a theoretical concern because neously (i.e., not in response to specific questions) in the daily diary.
of the high incidence of fever observed with the § Included were hospitalization, emergency department visit, and visit to phy-
TIV Fluvax (CSL) in Australia,28 which is believed sician, nurse practitioner, or other health care worker.
¶ No serious adverse event occurred at an incidence higher than 0.2% (see
to be due to a larger number of viral components Table S11 in the Supplementary Appendix for more details).
in that vaccine than in other TIVs. However, in ‖ There was one serious case of bronchitis in the QIV group.
the current study, the incidence of fever was ** There was one case of bronchitis and one case of convulsion in the QIV
group and one case of drowning in each group.
similar in the QIV and control groups and simi-
lar to the rate observed in previous studies of
TIVs in children (in which fever was defined as circulation of the B Yamagata lineage, as noted
a body temperature of 37.5°C or higher or 38°C above. A major strength is that this was an indi-
or higher).15,29 In the current study involving vidually randomized trial of an inactivated influ-
children residing mainly in tropical countries, enza vaccine in children 3 to 8 years of age. The
the immunogenicity of the QIV was high, simi- study was conducted in three global regions and
lar to that in a study of the same vaccine in used both active and passive surveillance to
children 3 to 17 years of age residing in Canada, identify cases of influenza, with analysis of
Mexico, and the United States.24 more than 95% of samples from children with
A limitation of the study is that it was con- suspected cases. Other important strengths were
ducted during only one season and provides only the end points we selected and the use of an rt-
a snapshot of the situation in that season. For PCR assay to confirm the presence of influenza
example, the findings are limited by the low virus. Selection of appropriate end points is a

25 26, 2013 2489
major challenge for influenza vaccine trials. We shift from one lineage to another. The efficacy
chose rt-PCR–confirmed influenza as the pri- of the vaccine was higher against moderate-to-
mary end point because rt-PCR assay has been severe disease — a potentially important end
shown to be associated with higher detection point associated with the highest clinical, social,
rates than conventional methods such as sero- and economic burden — than against illness of
logic testing and cell culture.13,30-33 As noted any severity. These results highlight the poten-
above, we included the prevention of moderate- tial clinical benefit of administering inactivated
to-severe influenza as an end point because we influenza vaccines in healthy children.
believe that the most important effect of influ-
enza vaccination lies in the prevention of mod- Supported by GlaxoSmithKline Biologicals.
Dr. Jain, Dr. Caplanusi, Mr. Durviaux, Dr. Boutet, Dr. Ofori-
erate-to-severe disease. Ideally, trials of vaccine Anyinam, Ms. Chandrasekaran, and Dr. Innis report being em-
efficacy would capture separately both the most ployees of GlaxoSmithKline, and Drs. Jain, Caplanusi, Boutet,
severe cases (severe lower respiratory tract ill- Ofori-Anyinam, and Innis report holding stock options, restricted
shares, or both in GlaxoSmithKline. Dr. Ceyhan reports being an
ness and serious complications) and cases that are investigator in studies for Sanofi Pasteur, Novartis, Pfizer, Merck
less severe but nevertheless concern parents and Sharp & Dohme, and GlaxoSmithKline and receiving fees for par-
have an effect on utilization of health care re- ticipation on advisory boards from GlaxoSmithKline, Pfizer, No-
vartis, and Merck Sharp & Dohme. Dr. Quiambao reports receiving
sources. Unfortunately, such trials would require grant support and personal fees from Novartis, lecture fees from
the enrollment of impractically large numbers of Sanofi Pasteur, and travel reimbursement from GlaxoSmithKline.
participants. We therefore chose a dichotomous Dr. Dinleyici reports receiving fees for participating on advisory
boards from GlaxoSmithKline and Novartis, and lecture fees from
classification of influenza: mild and not mild (i.e., GlaxoSmithKline. Dr. Alhan reports receiving fees for partici-
moderate-to-severe). We believe that this is a valid pating on advisory boards from GlaxoSmithKline, Pfizer, Sanofi
classification because it distinguishes between Pasteur, and Merck Sharp & Dohme. Dr. Dbaibo reports receiving
grant support from GlaxoSmithKline, Merck Sharp & Dohme,
influenza that causes mild upper respiratory tract and Pfizer. No other potential conflict of interest relevant to this
illness and low-grade fever from more severe ill- article was reported.
ness that may have adverse clinical consequences Disclosure forms provided by the authors are available with
the full text of this article at NEJM.org.
and may increase the utilization of health care We thank the children who participated in the study and their
resources. Moreover, it is a practical classifica- parents; the clinicians, nurses and laboratory technicians at the
tion with respect to the sample size needed for study sites; the project staff of the sponsor for support and contri-
butions throughout the study, especially M. Waleska Cruz, C. Nico-
vaccine trials. Because of this relatively smaller medes Gonzàles Diaz, N. Guevara, L. Mendez Pena, D. Stanziola,
sample size, however, the number of children in J. Tetangco, S. Ghanem, F. Hajar, N. Al-Ayoubi, S. Carlos, A. Aslan,
the study who had lower respiratory tract illness Y. Ozsurekci, and O. Ozgur; all the teams at GlaxoSmithKline Vac-
cines for their contribution to this study, especially A. Amanulah
was small (10 children, all in the control group). who contributed to the writing of the protocol and clinical study
In conclusion, the QIV was shown to be ef- report, P. Li and W. Dewé for their contribution to the statistical
ficacious in preventing influenza A and B in plan, E. Praet, J. McCormick, and J. Torres for clinical study man-
agement, V. Dodeur for data management, K. Walravens, N. Hou-
children 3 to 8 years of age. Given the problem ard, and C. Stalens from the clinical laboratory team, and W. Jiang
of poor vaccine efficacy against mismatched (clinical safety representative); M.L. Greenacre (An Sgriobhadair,
B lineages, the QIV will be of greatest value dur- U.K.) for providing medical writing services; and J. Dedessus Le
Moutier and B. Dumont (Business and Decision Life Sciences, on
ing seasons in which both lineages are circulat- behalf of GlaxoSmithKline Vaccines) for editorial assistance with
ing or in the event that there is an unexpected an earlier version of the manuscript and manuscript coordination.
appendix
The authors’ affiliations are as follows: GlaxoSmithKline Vaccines, King of Prussia, PA (V.K.J., V.C., B.L.I.); Department of Neo-
natal-Perinatal Medicine, Hospital Maternidad Nuestra Señora de la Altagracia, Santo Domingo, Dominican Republic (L.R.); Centre
for Child and Adolescent Health, International Center for Diarrheal Disease Research (ICDDR) (K.Z.), Executive Director’s Office,
ICDDR (A.C.), and Center for Child and Adolescent Health, ICDDR (M.Y.), Dhaka, Bangladesh; Department of Pediatrics, De La
Salle Health Sciences Institute and De La Salle College of Medicine, Dasmarinas City (R.A.E.), and Clinical Research Division,
Research Institute for Tropical Medicine, Muntinlupa City (B.P.Q.), Philippines; Department of Tropical Pediatrics, Mahidol Uni-
versity (C.S., P.C., K.L.), and Department of Infectious Pediatrics, Phramongkutklao Hospital of the Royal Thai Army (P.K.), Bang-
kok, Thailand; Clinical Research Department, Organización para el Desarrollo y la Investigación Salud en Honduras (ODISH),
Tegucigalpa, Honduras (D.M.R.-M.); Department of Pediatrics and Infectious Diseases, Hacettepe University, Ankara (M.C.), De-
partment of Pediatric Intensive Care and Infectious Disease Unit, Eskisehir Osmangazi University Faculty of Medicine, Eskisehir
(E.C.D.), Department of Pediatrics, Ege University Faculty of Medicine, Izmir (Z.K.), and Department of Pediatrics, Division of Pedi-
atric Infectious Disease, Cukurova University, Adana (E.A.) — all in Turkey; GlaxoSmithKline Vaccines, Wavre, Belgium (A.C., S.D.,
P.B., O.O.-A.); and Department of Pediatrics and Adolescent Medicine, American University of Beirut Medical Center, Beirut, Leba-
non (G.D.).

References
1. O’Brien MA, Uyeki TM, Shay DK, et al. Antibody responses to influenza B viruses 24. Langley JM, Martinez AC, Chatterjee
Incidence of outpatient visits and hospi- in immunologically unprimed children. A, et al. Immunogenicity and safety of an
talizations related to influenza in infants Pediatrics 1991;88:1031-6. inactivated quadrivalent influenza vaccine
and young children. Pediatrics 2004;113: 13. Vesikari T, Beran J, Durviaux S, et al. candidate: a randomized trial in children.
585-93. Use of real-time polymerase chain reaction Presented at IDWeek, San Diego, CA, Oc-
2. Izurieta HS, Thompson WW, Kra- (rtPCR) as a diagnostic tool for influenza tober 17–21, 2012. abstract.
marz P, et al. Influenza and the rates of infection in a vaccine efficacy trial. J Clin 25. Tinoco JC, Pavia-Ruz N, Cruz-Valdez
hospitalization for respiratory disease Virol 2012;53:22-8. A, et al. Immunogenicity and safety of an
among infants and young children. N Engl 14. Beran J, Vesikari T, Wertzova V, et al. inactivated quadrivalent influenza vaccine
J Med 2000;342:232-9. Efficacy of inactivated split-virus influenza candidate versus two trivalent influenza
3. Bourgeois FT, Valim C, Wei JC, Mc- vaccine against culture-confirmed influ- vaccines: a randomized controlled trial in
Adam AJ, Mandl KD. Influenza and other enza in healthy adults: a prospective, ran- adults. Presented at IDWeek, San Diego,
respiratory virus-related emergency de- domized, placebo-controlled trial. J Infect CA, October 17–21, 2012. abstract.
partment visits among young children. Dis 2009;200:1861-9. 26. Kieninger D, Lin W-Y, Yu CJ, et al.
Pediatrics 2006;118(1):e1-e8. 15. Hehme N, Künzel W, Petschke F, et al. Immunogenicity and safety profile of a
4. Molinari NA, Ortega-Sanchez IR, Mes- Ten years of experience with the trivalent candidate inactivated quadrivalent influ-
sonnier ML, et al. The annual impact of split-influenza vaccine, Fluarix. Clin Drug enza vaccine in adults and the elderly: a
seasonal influenza in the US: measuring Invest 2002;22:751-69. randomized controlled study. Presented
disease burden and costs. Vaccine 2007;25: 16. Saunders NR, Tennis O, Jacobson S, at Influenza Vaccines for the World (IVW),
5086-96. Gans M, Dick PT. Parents’ responses to Valencia, Spain, October 9–12, 2012.
5. Prevention and control of influenza symptoms of respiratory tract infection in abstract.
with vaccines: recommendations of the their children. CMAJ 2003;168:25-30. 27. Domachowske JB, Pankow-Culot H,
Advisory Committee on Immunization 17. Vesikari T, Knuf M, Wutzler P, et al. Bautista M, et al. A randomized trial of
Practices (ACIP), 2011. MMWR Morb Mor- Oil-in-water emulsion adjuvant with in- candidate inactivated quadrivalent influ-
tal Wkly Rep 2011;60:1128-32. fluenza vaccine in young children. N Engl enza vaccine versus trivalent influenza
6. Jefferson T, Rivetti A, Di Pietrantonj C, J Med 2011;365:1406-16. vaccines in children aged 3-17 years. J In-
Demicheli V, Ferroni E. Vaccines for prevent- 18. Jansen AG, Sanders EA, Hoes AW, van fect Dis 2013;207:1878-87.
ing influenza in healthy children. Cochrane Loon AM, Hak E. Effects of influenza 28. Armstrong PK, Dowse GK, Effler PV,
Database Syst Rev 2012;8:CD004879. plus pneumococcal conjugate vaccination et al. Epidemiological study of severe fe-
7. Skowronski DM, Masaro C, Kwindt versus influenza vaccination alone in pre- brile reactions in young children in West-
TL, et al. Estimating vaccine effectiveness venting respiratory tract infections in chil- ern Australia caused by a 2010 trivalent
against laboratory-confirmed influenza dren: a randomized, double-blind, placebo- inactivated influenza vaccine. BMJ Open
using a sentinel physician network: re- controlled trial. J Pediatr 2008;153:764-70. 2011;1(1):e000016.
sults from the 2005-2006 season of dual A 19. Clover RD, Crawford S, Glezen WP, 29. Baxter R, Jeanfreau R, Block SL, et al.
and B vaccine mismatch in Canada. Vac- Taber LH, Matson CC, Couch RB. Compari- A Phase III evaluation of immunogenicity
cine 2007;25:2842-51. son of heterotypic protection against influ- and safety of two trivalent inactivated sea-
8. Skowronski DM, De Serres G, Dickin- enza A/Taiwan/86 (H1N1) by attenuated sonal influenza vaccines in US children.
son J, et al. Component-specific effective- and inactivated vaccines to A/Chile/83-like Pediatr Infect Dis J 2010;29:924-30.
ness of trivalent influenza vaccine as mon- viruses. J Infect Dis 1991;163:300-4. 30. van Elden LJR, Nijhuis M, Schipper P,
itored through a sentinel surveillance 20. DiazGranados CA, Denis M, Plotkin S. Schuurman R, van Loon AM. Simultane-
network in Canada, 2006-2007. J Infect Dis Seasonal influenza vaccine efficacy and its ous detection of influenza viruses A and B
2009;199:168-79. determinants in children and non-elderly using real-time quantitative PCR. J Clin
9. Belongia EA, Kieke BA, Donahue JG, et adults: a systematic review with meta- Microbiol 2001;39:196-200.
al. Effectiveness of inactivated influenza analyses of controlled trials. Vaccine 2012; 31. Ward CL, Dempsey MH, Ring CJ, et al.
vaccines varied substantially with antigenic 31:49-57. Design and performance testing of quan-
match from the 2004-2005 season to the 21. Reed C, Meltzer MI, Finelli L, Fiore A. titative real time PCR assays for influenza
2006-2007 season. J Infect Dis 2009;199: Public health impact of including two lin- A and B viral load measurement. J Clin
159-67. eages of influenza B in a quadrivalent sea- Virol 2004;29:179-88.
10. Belshe RB, Coelingh K, Ambrose CS, sonal influenza vaccine. Vaccine 2012;30: 32. Gooskens J, Swaan CM, Claas ECJ,
Woo JC, Wu X. Efficacy of live attenuated 1993-8. Kroes ACM. Rapid molecular detection of
influenza vaccine in children against in- 22. Poehling KA, Edwards KM, Weinberg influenza outbreaks in nursing homes.
fluenza B viruses by lineage and antigenic GA, et al. The underrecognized burden of J Clin Virol 2008;41:7-12.
similarity. Vaccine 2010;28:2149-56. influenza in young children. N Engl J Med 33. Liao RS, Tomalty LL, Majury A, Zout-
11. Englund JA, Walter EB, Gbadebo A, 2006;355:31-40. man DE. Comparison of viral isolation
Monto AS, Zhu Y, Neuzil KM. Immuniza- 23. Neuzil KM, Hohlbein C, Zhu Y. Illness and multiplex real-time reverse transcrip-
tion with trivalent inactivated influenza among schoolchildren during influenza tion-PCR for the confirmation of respira-
vaccine in partially immunized toddlers. season: effect on school absenteeism, pa- tory syncytial virus and influenza viruses
Pediatrics 2006;118(3):e579-e585. rental absenteeism from work, and sec- detected by antigen immunoassay testing.
12. Levandowski RA, Regnery HL, Staton E, ondary illness in families. Arch Pediatr J Clin Microbiol 2009;47:527-32.
Burgess BG, Williams MS, Groothuis JR. Adolesc Med 2002;156:986-91. Copyright © 2013 Massachusetts Medical Society.
receive immediate notification when an article

is published online first
To be notified by e-mail when Journal articles

are published Online First, sign up at NEJM.org.
n engl j med 369;26 27

nejm.org december 26, 2013 2491
Source: NEJM Journal Watch
Efficacy of Influenza Vaccine Against Severe Disease in Children

Complete vaccination protects against life-threatening influenza-related illness.
Ferdinands JM et al. Effectiveness of influenza vaccine against life-threatening RT-PCR-confirmed influenza illness in
US children, 2010-2012. J Infect Dis 2014 Mar 26; [e-pub ahead of print]. (http://dx.doi.org/10.1093/infdis/jiu185)
ARTICLE SUMMARY
Influenza vaccine reduces the incidence of influenza and influenza-like illness in all age groups, but efficacy is
greatest in children, and vaccination in children can provide herd immunity to the entire community. Influenza
vaccine has also been shown to reduce hospitalizations, physician visits, and antibiotic prescriptions for otitis me-
dia and pneumonia. Investigators examined whether influenza vaccine reduces the risk for pediatric intensive
care unit (PICU) admission or death in a case-control study during two influenza seasons (2010–2012). Cases
were 44 children with influenza confirmed by polymerase chain reaction at 21 PICUs in the U.S. Two control
groups included 172 PICU children without influenza and 93 children from the same community, matched for
age and chronic medical conditions.
Overall, 34% of PICU patients had an underlying condition, and only 37% of these patients were fully vaccinated.
The efficacy of complete vaccination against life-threatening influenza in PICU cases was 74% compared with
PICU controls, and 82% compared with community controls. Partial vaccination status provided no protective
benefit. A greater proportion of PICU cases than controls died (9.1% vs. 1.7%).
COMMENT
This study suggests that complete vaccination against influenza significantly reduces life-threatening complica-
tions of influenza in children. Comparison of cases with two control groups reduced the risk for confounding that
falsely diminishes vaccine efficacy (because children with high-risk conditions have higher rates of vaccination
and influenza-related complications). Given what we know about risk factors for severe influenza, it is striking
that only one third of admitted children in this study were immunized against influenza. The results reinforce the
importance of complete influenza immunization for all children and remind us that influenza is a serious and
sometimes fatal disease.
— Deborah Lehman, MD
David Geffen School of Medicine, University of California, Los Angeles
NEJM Journal Watch Pediatrics and Adolescent Medicine, published April 29, 2014
28
HIGH-DOSE INFLUENZA VACCINE IN OLDER ADULTS

The following two studies — one in NEJM and one summarized in NEJM Journal Watch— of high-dose influenza
vaccine in older adults led to slightly different conclusions. While DiazGranados and colleagues report signifi-
cantly higher antibody responses and better protection against influenza with higher-dose versus lower-dose
vaccine in persons age 65 and older, Richardson et al. found superiority of higher-dose vaccine only in persons
age 85 and older. However, our summary author warns that the primary study outcome in the latter study —
hospitalization for influenza or pneumonia — may not be a refined enough measure to ascertain the benefit
of the higher-dose vaccine. Authors of the former study assessed results via hemagglutination inhibition titers
and seroprotection rates (the percentage of participants with HAI titers ≥1:40).
29
Original Article
Efficacy of High-Dose versus Standard-Dose

Influenza Vaccine in Older Adults
Carlos A. DiazGranados, M.D., Andrew J. Dunning, Ph.D., Murray Kimmel, D.O.,
Daniel Kirby, B.Sc., John Treanor, M.D., Avi Collins, B.Sc.N.,
Richard Pollak, D.P.M., Janet Christoff, R.N., John Earl, M.D.,
Victoria Landolfi, M.Sc., M.B.A., Earl Martin, D.O., Sanjay Gurunathan, M.D.,
Richard Nathan, D.O., David P. Greenberg, M.D., Nadia G. Tornieporth, M.D.,
Michael D. Decker, M.D., M.P.H., and H. Keipp Talbot, M.D., M.P.H.
A BS T R AC T
BACKGROUND
As compared with a standard-dose vaccine, a high-dose, trivalent, inactivated influ- From Sanofi Pasteur, Swiftwater (C.A.D.,
enza vaccine (IIV3-HD) improves antibody responses to influenza among adults 65 A.J.D., D.K., J.C., V.L., S.G., D.P.G., N.G.T.,
M.D.D.), ReSearch Pharmaceutical Ser
years of age or older. This study evaluated whether IIV3-HD also improves protec- vices, Fort Washington (A.C.), and the
tion against laboratory-confirmed influenza illness. Department of Pediatrics, University of
Pittsburgh School of Medicine, Pitts
burgh (D.P.G.) — all in Pennsylvania; Ac
METHODS celovance, Melbourne, FL (M.K.); Univer
We conducted a phase IIIb–IV, multicenter, randomized, double-blind, active-con- sity of Rochester, Rochester, NY (J.T.);
trolled trial to compare IIV3-HD (60 μg of hemagglutinin per strain) with standard- Endeavor Clinical Trials, San Antonio
(R.P.), and Martin Diagnostic Clinic,
dose trivalent, inactivated influenza vaccine (IIV3-SD [15 μg of hemagglutinin per Tomball (E.M.) — both in Texas; PMG
strain]) in adults 65 years of age or older. Assessments of relative efficacy, effec- Research of Hickory, Hickory, NC (J.E.);
tiveness, safety (serious adverse events), and immunogenicity (hemagglutination- Idaho Falls Infectious Diseases and Snake
River Research, Idaho Falls, ID (R.N.);
inhibition [HAI] titers) were performed during the 2011–2012 (year 1) and the and the Department of Health Policy,
2012–2013 (year 2) northern-hemisphere influenza seasons. Vanderbilt University School of Medicine
(M.D.D.) and Vanderbilt University Medi
cal Center (H.K.T.) — both in Nashville.
RESULTS
Address reprint requests to Dr. Diaz
A total of 31,989 participants were enrolled from 126 research centers in the United Granados at Sanofi Pasteur, 1 Discovery
States and Canada (15,991 were randomly assigned to receive IIV3-HD, and 15,998 Dr., Swiftwater, PA 18370, or at carlos
.diazgranados@sanofipasteur.com.
to receive IIV3-SD). In the intention-to-treat analysis, 228 participants in the IIV3-
HD group (1.4%) and 301 participants in the IIV3-SD group (1.9%) had laboratory- N Engl J Med 2014;371:635-45.
DOI: 10.1056/NEJMoa1315727
confirmed influenza caused by any viral type or subtype associated with a proto- Copyright © 2014 Massachusetts Medical Society.
col-defined influenza-like illness (relative efficacy, 24.2%; 95% confidence interval
[CI], 9.7 to 36.5). At least one serious adverse event during the safety surveillance Related Material:
period was reported by 1323 (8.3%) of the participants in the IIV3-HD group, as Supplementary Appendix
compared with 1442 (9.0%) of the participants in the IIV3-SD group (relative risk,
0.92; 95% CI, 0.85 to 0.99). After vaccination, HAI titers and seroprotection rates
(the percentage of participants with HAI titers ≥1:40) were significantly higher in
the IIV3-HD group.
CONCLUSIONS
Among persons 65 years of age or older, IIV3-HD induced significantly higher
antibody responses and provided better protection against laboratory-confirmed
influenza illness than did IIV3-SD. (Funded by Sanofi Pasteur; ClinicalTrials.gov
number, NCT01427309.)
n engl j med 371;7 nejm.org30August 14, 2014 635

B
etween 1990 and 1999, seasonal in- through October 21, 2012 (year 2). An indepen-
fluenza caused an average of 36,000 deaths dent data and safety monitoring committee as-
and 226,000 hospitalizations per year in sessed the study data.
the United States.1-3 Adults 65 years of age or The study was funded by Sanofi Pasteur. The
older are particularly vulnerable to complications sponsor had primary responsibility for study de-
associated with influenza and account for most sign, protocol development, study monitoring,
seasonal influenza–related hospitalizations and data management, and statistical analyses. The
deaths.2,3 Although vaccination currently repre- investigators at the study centers had primary
sents the most effective intervention against in- responsibility for critical protocol review, study
fluenza and associated complications,3,4 antibody procedures, and data collection. The coordinating
response and protection elicited by the vaccine are investigator (the last author) had a primary role
lower among persons 65 years of age or older than in reviewing and approving the protocol and the
among younger adults.5-7 Strategies to improve clinical study report. The manuscript was draft-
antibody responses to influenza vaccine in the ed by the corresponding author. All the authors
older population, such as increasing the amount critically reviewed, edited, and approved the manu-
of antigen in the vaccine, may improve protec- script and made the decision to submit it for pub-
tion and have a favorable effect on morbidity and lication. No persons other than the named au-
mortality.8 thors played any role in writing the manuscript.
The high-dose, trivalent, inactivated influenza All the authors assume responsibility for the
vaccine (IIV3-HD) contains four times as much accuracy and completeness of the data and for
hemagglutinin (HA) as is contained in standard- the fidelity of the study to the protocol, which is
dose vaccines. On the basis of its safety profile available with the full text of this article at
and superior immunogenicity as compared with NEJM.org.
a standard-dose vaccine,9 IIV3-HD was licensed
for use in the United States in December 2009, Participants and Group Allocation
with a requirement to show clinical benefit. The The study included adults 65 years of age or
primary objective of this study was to show the older without moderate or severe acute illnesses.
efficacy of IIV3-HD as compared with a stan- Details on exclusion criteria are provided in the
dard-dose vaccine for the prevention of laborato- Supplementary Appendix, available at NEJM.org.
ry-confirmed influenza illness in adults 65 years Each study year, participants were randomly as-
of age or older. signed in a 1:1 ratio to receive a single dose of
IIV3-HD or IIV3-SD. Those who were partici-
pants in both years underwent rerandomization
Me thods
the second year. The study used concealed allo-
Study Design and Oversight cation through an interactive voice-response sys-
We conducted a phase IIIb–IV, multicenter, ran- tem that centrally assigned participants on the
domized, double-blind, active-controlled trial com- basis of computer-generated block randomization.
paring IIV3-HD with a standard-dose vaccine Approximately one third of participants were se-
(IIV3-SD) in persons 65 years of age or older at lected randomly by the same system to be in the
126 centers in the United States and Canada immunogenicity subset. Participants, investigators,
from September 6, 2011, through May 31, 2013. and the sponsor’s study staff remained unaware
The study was approved by three institutional of the study assignments.
review boards (Quorum Review IRB, Western
Institutional Review Board, and Vanderbilt Uni- Vaccines
versity Institutional Review Board) and was con- Vaccines were formulated according to Food and
ducted in accordance with the Declaration of Drug Administration (FDA) recommendations.
Helsinki and the International Conference on The standard-dose vaccine (IIV3-SD; Fluzone,
Harmonisation Guidelines for Good Clinical Prac- Sanofi Pasteur) contained 15 μg of HA per strain.
tice. All participants gave written informed con- IIV3-HD (Fluzone High-Dose, Sanofi Pasteur)
sent for study participation. There were two en- contained 60 μg of HA per strain. Both vaccines
rollment periods, from September 6 through were produced in embryonated chicken eggs, were
October 9, 2011 (year 1), and from October 9 inactivated with formaldehyde, were split with a
636 n engl j med 371;7 nejm.org August 14, 2014

31
High-Dose vs. Standard-Dose Influenza Vaccine
nonionic detergent, and contained A/California/ If a participant met the criteria for any respi-
7/2009 (H1N1), A/Victoria/210/2009 (H3N2), and ratory illness, staff members at the study site
B/Brisbane/60/2008 strains for the year 1 season were to collect a nasopharyngeal swab within
and A/California/7/2009 (H1N1), A/Victoria/361/ 5 days after onset of the illness. Laboratory con-
2011 (H3N2), and B/Texas/6/2011 (B/Wisconsin/ firmation of influenza in nasopharyngeal swabs
1/2010-like virus) strains for the year 2 season. was accomplished by a positive result on culture,
The vaccines were provided in ready-to-use 0.5-ml a polymerase-chain-reaction (PCR) assay, or both.
syringes and administered intramuscularly, in the A hemagglutination-inhibition (HAI) assay against
deltoid. a panel of standard ferret antiserum specimens
was performed to determine whether a sample
Surveillance and Ascertainment of Influenza strain was antigenically similar to a vaccine
Participants were instructed to contact their strain.14 Genetic sequencing further evaluated
study site if they had any respiratory symptoms. similarity to vaccine components. The ferret HAI
In addition, participants were contacted by a call assay was the primary method for classification
center twice weekly (between the beginning of of similarity to the vaccine, with genetic sequenc-
January and the end of February) or weekly until ing used only for laboratory-confirmed samples
the end of illness surveillance (April 30 each year). for which ferret HAI assay results were not avail-
Three definitions of clinical illness were evaluat- able. Further details about the laboratory meth-
ed in the study. ods and similarity assessments are provided in
Respiratory illness was defined as the occur- the Supplementary Appendix.
rence of one or more of the following: sneezing, Investigators at the study sites were instructed
nasal congestion or rhinorrhea, sore throat, cough, to make follow-up telephone calls to collect effec-
sputum production, wheezing, or difficulty breath- tiveness information associated with events occur-
ing. This definition provided high sensitivity for ring within 30 days after the start of any respira-
the detection of cases of influenza and triggered tory illness. These events included pneumonia,
key study procedures. cardiorespiratory conditions, health care visits
A protocol-defined influenza-like illness (the (hospitalizations for any cause, visits to the emer-
illness definition for the primary analysis) pro- gency department, and nonroutine medical vis-
vided increased specificity and clinical relevance its), and medication use (restricted to antipyretic
beyond the respiratory illness definition. It was agents, analgesic agents, nonsteroidal antiinflam-
defined as a respiratory illness with sore throat, matory drugs, antiviral agents, and antibiotics).
cough, sputum production, wheezing, or diffi- Safety surveillance extended from vaccination
culty breathing, concurrent with one or more of to approximately May 15 of the following year.
the following: temperature above 37.2°C, chills, Because the side-effect profile of IIV3-HD had
tiredness, headaches, or myalgia. been evaluated previously, in a large-scale pivotal
A modified CDC-defined influenza-like ill- study,9 the safety data that were collected in this
ness was based on the Centers for Disease Con- trial were limited to serious adverse events.
trol and Prevention (CDC) surveillance network
definition of an influenza-like illness and was Immunogenicity
defined as a respiratory illness with cough or Blood samples were collected from participants
sore throat, concurrent with a temperature above in the immunogenicity subset approximately 28
37.2°C. The modified CDC-defined influenza- days after vaccination and were assayed for HAI
like illness incorporated a lower threshold for titers14 by Focus Diagnostics. HAI titers were sum-
the temperature criterion than did the original marized as geometric mean titers and seropro-
CDC definition (≥37.8°C) because of the docu- tection rates (i.e., the percentage of participants
mented low frequency of temperatures of 37.8°C with an HAI titer ≥1:40).
or higher in older adults with confirmed influ-
enza,10,11 because specificity was being provided Measures of Efficacy
by laboratory confirmation in this clinical trial, The primary end point of the study was the oc-
and because of the precedent for a lower thresh- currence, at least 14 days after vaccination, of lab-
old in other influenza efficacy studies in similar oratory-confirmed influenza caused by any influ-
populations.12,13 enza viral types or subtypes, in association with

32 August 14, 2014 637
of 2% for the IIV3-SD group; furthermore, IIV3-

HD would be considered superior to IIV3-SD if
31,989 Patients underwent randomization
the lower bound of the 95% confidence interval
for relative vaccine efficacy exceeded 9.1% for
the primary end point. By agreement with the
FDA, a 9.1% margin for superior vaccine efficacy
15,991 Were assigned to high-dose 15,998 Were assigned to standard-dose was used to provide confidence that the risk of
vaccine vaccine
15,990 Received any vaccine 15,993 Received any vaccine the primary end point was at least 10% higher
15,982 Received high-dose vaccine 15,983 Received standard-dose with the administration of IIV3-SD than with
vaccine
the administration of IIV3-HD.
The efficacy of IIV3-HD relative to IIV3-SD was
734 Discontinued study before final 788 Discontinued study before final estimated as 1 minus the relative risk. The con-
phone call phone call fidence interval for efficacy estimates was calcu-
252 Were lost to follow-up 280 Were lost to follow-up
167 Did not adhere to protocol 195 Did not adhere to protocol lated with the Clopper–Pearson exact method for
210 Withdrew voluntarily 206 Withdrew voluntarily binomial proportions.15
102 Had serious adverse event 106 Had serious adverse event
3 Had other adverse event 1 Had other adverse event Statistical significance was defined by a 95%
confidence interval excluding the null value (0 for
relative vaccine efficacy and seroprotection rate
15,990 Were included in the full 15,993 Were included in the full differences, and 1 for relative risks and the geo-
analysis set analysis set
1 Did not receive any vaccine 5 Did not receive any vaccine metric mean titer ratios [the ratio of the geomet-
and was excluded and were excluded ric mean titer for IIV3-HD to the geometric mean
titer for IIV3-SD]).
Two analysis sets were used. The full analysis
15,892 Were included in the per- 15,911 Were included in the per-
protocol analysis set protocol analysis set
set comprised all participants who received
99 Were excluded 87 Were excluded study vaccine; participants were grouped accord-
9 Did not receive vaccine 15 Did not receive vaccine
per randomization per randomization
ing to their treatment assignment for efficacy
10 Did not satisfy inclusion 7 Did not satisfy inclusion and effectiveness analyses (intention-to-treat)
criteria or met exclusion criteria or met exclusion
criteria criteria
and according to the vaccine actually received for
57 Did not have at least one 36 Did not have at least one safety analyses. Efficacy and effectiveness were
successful surveillance successful surveillance
contact contact
also analyzed in the per-protocol analysis set;
16 Received another seasonal 20 Received another seasonal conditions under which participants were ex-
influenza vaccine during influenza vaccine during
surveillance surveillance
cluded from the per-protocol analysis set are
7 Had other reasons 9 Had other reasons shown in Figure 1. All statistical analyses were
performed with SAS Enterprise Guide 5.1 (SAS
Figure 1. Enrollment and Follow-up of Study Participants. Institute).
Three participants in each group who had serious adverse events were in
stitutionalized and unable to speak on the phone at the final call before
study termination.
R e sult s
Participants
A total of 14,500 participants were enrolled in
a protocol-defined influenza-like illness. Several year 1, and 17,489 in year 2. Of year 1 partici-
secondary efficacy and observational effective- pants, 7645 reenrolled in year 2. Of all 31,989
ness end points were also evaluated, according participant-seasons (termed “participants”) en-
to various clinical illness definitions, methods rolled, 15,991 were randomly assigned to IIV3-
of laboratory confirmation, and levels of simi- HD and 15,998 were randomly assigned to IIV3-
larity to the vaccine. SD (Fig. 1). Of the participants who underwent
randomization, 31,983 (>99.9%) received study
Statistical Analysis vaccine; all 31,983 were included in the full
The total sample size required to provide 80% analysis set, and 31,803 (99.4%) were included in
power to show the superior efficacy of IIV3-HD the per-protocol analysis set. Baseline character-
was 30,000 participants, assuming a relative vac- istics of the recipients of IIV3-HD and the re-
cine efficacy of 30% and an influenza incidence cipients of IIV3-SD were similar (Table 1).

33 August 14, 2014
Table 1. Baseline Demographic and Clinical Characteristics of the High-Dose and Standard-Dose Vaccine Groups.*
IIV3-HD IIV3-SD
Characteristic (N = 15,990) (N = 15,993)
Female sex — no. (%) 9,131 (57.1) 8,963 (56.0)
Mean age — yr 73.3±5.8 73.3±5.8
Racial background — no. (%)†
White 15,103 (94.4) 15,167 (94.8)
Asian 118 (0.7) 105 (0.7)
Black 670 (4.2) 612 (3.8)
Other 97 (0.6) 106 (0.7)
Hispanic ethnic group — no. (%)† 958 (6.0) 982 (6.1)
At least one prespecified chronic coexisting condition — no. (%)‡ 10,750 (67.2) 10,752 (67.2)
At least two prespecified chronic coexisting conditions — no. (%) 5,385 (33.7) 5,403 (33.8)
Cardiac and respiratory disorders — no. (%)
Coronary artery disease 2,735 (17.1) 2,732 (17.1)
Atrial fibrillation 1,103 (6.9) 1,112 (7.0)
Valvular heart disease 744 (4.6) 741 (4.6)
Congestive heart failure 451 (2.8) 446 (2.8)
Chronic obstructive lung disease 1,500 (9.4) 1,495 (9.4)
Asthma 1,415 (8.8) 1,408 (8.8)
Received influenza vaccine the previous season — no. (%) 11,758 (73.5) 11,773 (73.6)
* There were no significant differences between the treatment groups. IIV3HD denotes highdose, trivalent, inactivated
influenza vaccine, and IIV3SD standarddose, trivalent, inactivated influenza vaccine. Plus–minus values are means
±SD.
† Racial background and ethnic group were selfreported. The category “Other” includes American Indian or Alaska
Native, Native Hawaiian or other Pacific Islander, and mixed origin. Information on racial background was missing for
two recipients of the highdose vaccine and three recipients of the standarddose vaccine. Percentages may not total
100.0% because of rounding.
‡ Prespecified chronic coexisting conditions include the listed cardiac and respiratory disorders, as well as sickle cell dis
ease, diabetes mellitus, hypothyroidism, epilepsy, stroke, spinal cord injury, Parkinson’s disease, chronic kidney dis
ease, chronic hepatitis, cirrhosis, human immunodeficiency virus–acquired immunodeficiency syndrome, cancer, long
term systemic glucocorticoid therapy, and other potentially immunosuppressive therapies (pergroup frequencies of
these conditions are provided in the Supplementary Appendix).
Illness Surveillance and Collection for 80.0% of protocol-defined influenza-like ill-

of Nasopharyngeal Swabs nesses, 73.6% of modified CDC-defined influenza-
In the full analysis set (according to treatment like illnesses, and 67.2% of respiratory illnesses.
assignment), 3745 participants in the IIV3-HD In the IIV3-SD group, nasopharyngeal swabs were
group (23.4%) had at least one protocol-defined collected within the protocol-specified time frame
influenza-like illness, 758 (4.7%) had at least one for 79.3% of protocol-defined influenza-like ill-
modified CDC-defined influenza-like illness, and nesses, 73.7% of modified CDC-defined influenza-
8168 (51.1%) had at least one respiratory illness. like illnesses, and 66.8% of respiratory illnesses.
In the IIV3-SD group, 3827 participants (23.9%)
had at least one protocol-defined influenza-like Efficacy
illness, 838 (5.2%) had at least one modified In the full analysis set, 529 participants met the
CDC-defined influenza-like illness, and 8270 primary end point; 228 (1.4%) were in the IIV3-
(51.7%) had at least one respiratory illness. In HD group, and 301 (1.9%) were in the IIV3-SD
the IIV3-HD group, nasopharyngeal swabs were group. The efficacy of IIV3-HD relative to IIV3-
collected within the protocol-specified time frame SD for the primary end point was 24.2% in both

nejm.org August 14, 2014 639
the full analysis set (Table 2) and the per-proto- lated to the study vaccine. No deaths occurred
col analysis set. For both analyses, the lower within 30 days after vaccination in the IIV3-SD
bound of the 95% confidence interval for relative group.
vaccine efficacy was 9.7%, satisfying the prespeci- Three IIV3-HD recipients had serious adverse
fied superiority criterion. In addition, the point events categorized by their site investigators as
estimate for relative vaccine efficacy was consis- related to vaccination: cranial-nerve VI palsy start-
tently positive across influenza types, clinical ill-
ing 1 day after vaccination; hypovolemic shock
ness definitions, methods of laboratory confir- associated with diarrhea starting 1 day after vac-
mation (Table 2), and study years (see the cination; and acute disseminated encephalomyeli-
Supplementary Appendix). tis starting 117 days after vaccination. All three
Overall, relative efficacy estimates were higherevents resolved before study completion; none re-
in analyses restricted to cases caused by vaccine- sulted in discontinuation from the study. No seri-
similar strains (Table 3): relative vaccine efficacy
ous adverse events occurring in IIV3-SD recipi-
against laboratory-confirmed protocol-defined ents were considered to be related to vaccination
influenza-like illness caused by similar strains by the investigators.
was 35.4% (95% confidence interval [CI], 12.5 to A total of 99 participants (0.6%) in the IIV3-HD
52.5). Details on characterization and distribu- group and 103 participants (0.6%) in the IIV3-SD
tion of influenza isolates are available in the Sup-
group discontinued the study owing to serious
plementary Appendix. adverse events, none considered to be related to
vaccination. Cardiac disorders and infections were
Effectiveness the most frequent types of serious adverse events
Per-group rates and corresponding relative risks in both groups (see the Supplementary Appendix
of events occurring within 30 days after a study for serious adverse events according to organ
illness are available in the Supplementary Appen- systems).
dix. Most rates for pneumonia, cardiorespiratory
conditions, hospitalizations, nonroutine medical Immunogenicity
office visits, and medication use were lower for HAI antibody geometric mean titers and sero-
participants in the IIV3-HD group than for those protection rates 28 days after vaccination were
in the IIV3-SD group, with 62 of 66 relative risks significantly higher after vaccination with IIV3-HD
that could be evaluated having a point estimate than with IIV3-SD for all three vaccine strains
below 1. (Table 4).
Safety
Discussion
During the safety surveillance period (approxi-
mately 6 to 8 months after vaccination), 1323 A few randomized, controlled trials have shown
participants (8.3%) in the IIV3-HD group and moderate efficacy of influenza vaccines among
1442 participants (9.0%) in the IIV3-SD group had older adults.5,7,12 However, given the persistently
at least one serious adverse event. The relative high burden of influenza in this population de-
risk for having at least one serious adverse event spite increased vaccination rates,17 improved vac-
with IIV3-HD, as compared with IIV3-SD, was cines are needed.18,19 Some strategies to improve
0.92 (95% CI, 0.85 to 0.99). influenza vaccines for older adults involve higher
During the surveillance period, 83 (0.5%) of the doses of antigen or the use of adjuvants,20-23 as well
participants in the IIV3-HD group died, as did 84 as alternative delivery systems.24
(0.5%) of the participants in the IIV3-SD group. This randomized, double-blind, active-con-
Six recipients of IIV3-HD died within 30 days af- trolled efficacy trial showed that IIV3-HD pro-
ter vaccination. Two deaths were deemed acciden- vided improved protection against laboratory-con-
tal (smoke inhalation and traumatic head injury) firmed influenza illness among adults 65 years
and the other four were caused by heart failure, of age or older. The overall efficacy of 24.2%
cerebral bleeding, pneumonia, and myocardial in- against the primary end point indicates that about
farction and occurred in participants who had one quarter of all breakthrough influenza ill-
established risk factors for those conditions. Site nesses could be prevented if IIV3-HD were used
investigators classified these six events as unre- instead of IIV3-SD. More than a third of break-
640 n engl j med 371;7 nejm.org August 14, 2014

35
Table 2. Efficacy of High-Dose Vaccine Relative to Standard-Dose Vaccine against Confirmed Influenza Caused by Any Viral Type or Subtype.*
Variable Laboratory-Confirmed Influenza† Culture-Confirmed Influenza

IIV3HD IIV3SD Relative Efficacy IIV3HD IIV3SD Relative Efficacy

(N = 15,990) (N = 15,993) (95% CI) (N = 15,990) (N = 15,993) (95% CI)
no. (%) % no. (%) %
Protocoldefined influenzalike illness 228 (1.4) 301 (1.9) 24.2 (9.7 to 36.5)‡ 206 (1.3) 268 (1.7) 23.1 (7.5 to 36.2)
Influenza A 190 (1.2) 250 (1.6) 24.0 (7.8 to 37.4) 170 (1.1) 222 (1.4) 23.4 (6.0 to 37.6)
A/H1N1 8 (<0.1) 9 (0.1) 11.1 (−159.6 to 70.2) 7 (<0.1) 9 (0.1) 22.2 (−134.7 to 75.4)
A/H3N2 171 (1.1) 223 (1.4) 23.3 (6.0 to 37.5) 156 (1.0) 199 (1.2) 21.6 (2.8 to 36.8)
Influenza B 38 (0.2) 51 (0.3) 25.5 (−15.7 to 52.4) 36 (0.2) 46 (0.3) 21.7 (−23.8 to 50.8)
36
n engl j med 371;7
Modified CDCdefined influenzalike illness 96 (0.6) 121 (0.8) 20.6 (−4.6 to 39.9) 84 (0.5) 110 (0.7) 23.6 (−2.4 to 43.2)
Influenza A 86 (0.5) 104 (0.7) 17.3 (−11.1 to 38.6) 75 (0.5) 94 (0.6) 20.2 (−9.3 to 41.9)

A/H1N1 3 (<0.1) 2 (<0.1) −50.0 (−1696.0 to 82.8) 2 (<0.1) 2 (<0.1) 0.0 (−1280.0 to 92.8)
nejm.org
A/H3N2 77 (0.5) 95 (0.6) 18.9 (−10.7 to 40.8) 69 (0.4) 85 (0.5) 18.8 (−12.9 to 41.8)
Influenza B 10 (0.1) 17 (0.1) 41.2 (−36.0 to 75.9) 9 (0.1) 16 (0.1) 43.7 (−35.2 to 78.1)
Respiratory illness 316 (2.0) 387 (2.4) 18.3 (5.0 to 29.8) 277 (1.7) 339 (2.1) 18.3 (3.9 to 30.5)
Influenza A 262 (1.6) 313 (2.0) 16.3 (1.0 to 29.2) 227 (1.4) 272 (1.7) 16.5 (0.1 to 30.3)
August 14, 2014

A/H1N1 14 (0.1) 10 (0.1) −40.0 (−252.4 to 42.2) 13 (0.1) 10 (0.1) −30.0 (−231.3 to 47.33)
A/H3N2 231 (1.4) 281 (1.8) 17.8 (1.8 to 31.2) 205 (1.3) 246 (1.5) 16.6 (−0.7 to 31.1)
Influenza B 54 (0.3) 74 (0.5) 27.0 (−5.1 to 49.6) 50 (0.3) 67 (0.4) 25.4 (−9.3 to 49.3)
* CDC denotes Centers for Disease Control and Prevention.
† Laboratory confirmation of influenza was accomplished by a positive result on culture of a nasopharyngeal swab, a positive polymerasechainreaction assay, or both.
‡ The primary end point of the study was the occurrence, at least 14 days after vaccination, of laboratoryconfirmed influenza caused by any influenza viral types or subtypes, in associa
tion with a protocoldefined influenzalike illness.
641
642
Table 3. Efficacy of High-Dose Vaccine Relative to Standard-Dose Vaccine against Confirmed Influenza Caused by Strains Similar to the Vaccine Components.
Variable Laboratory-Confirmed Influenza* Culture-Confirmed Influenza†
IIV3HD IIV3SD Relative Efficacy IIV3HD IIV3SD Relative Efficacy

(N = 15,990) (N = 15,993) (95% CI) (N = 15,990) (N = 15,993) (95% CI)

no. (%) % no. (%) %
The
Protocoldefined influenzalike illness 73 (0.5) 113 (0.7) 35.4 (12.5 to 52.5) 63 (0.4) 92 (0.6) 31.5 (4.6 to 51.1)
Influenza A 56 (0.4) 82 (0.5) 31.7 (2.9 to 52.3) 46 (0.3) 63 (0.4) 27.0 (−8.5 to 51.2)
A/H1N1 7 (<0.1) 8 (0.1) 12.5 (−176.2 to 73.0) 3 (<0.1) 3 (<0.1) 0.0 (−646.8 to 86.6)
A/H3N2 49 (0.3) 74 (0.5) 33.8 (3.7 to 54.8) 43 (0.3) 60 (0.4) 28.3 (−7.8 to 52.7)
Influenza B 17 (0.1) 31 (0.2) 45.2 (−2.2 to 71.5) 17 (0.1) 29 (0.2) 41.4 (−10.3 to 69.8)

Modified CDCdefined influenzalike illness 26 (0.2) 51 (0.3) 49.0 (16.7 to 69.5) 22 (0.1) 45 (0.3) 51.1 (16.8 to 72.0)
Influenza A 21 (0.1) 36 (0.2) 41.7 (−2.7 to 67.6) 17 (0.1) 31 (0.2) 45.2 (−2.2 to 71.5)
nejm.org
A/H1N1 2 (<0.1) 2 (<0.1) 0.0 (−1280.0 to 92.8) 0 1 (<0.1) 100.0 (−3801.0 to 100.0)

A/H3N2 19 (0.1) 34 (0.2) 44.1 (−0.8 to 69.9) 17 (0.1) 30 (0.2) 43.3 (−6.1 to 70.7)
n e w e ng l a n d j o u r na l
of
Influenza B 5 (<0.1) 15 (0.1) 66.7 (3.5 to 90.5) 5 (<0.1) 14 (0.1) 64.3 (−5.0 to 89.9)
Respiratory illness 106 (0.7) 146 (0.9) 27.4 (6.1 to 44.0) 85 (0.5) 118 (0.7) 28.0 (4.0 to 46.1)
Influenza A 82 (0.5) 101 (0.6) 18.8 (−9.8 to 40.1) 61 (0.4) 75 (0.5) 18.6 (−15.6 to 43.0)
August 14, 2014

A/H1N1 12 (0.1) 9 (0.1) −33.4 (−258.4 to 48.4) 5 (<0.1) 3 (<0.1) −66.7 (−973.5 to 67.6)

m e dic i n e
A/H3N2 70 (0.4) 92 (0.6) 23.9 (−5.0 to 45.0) 56 (0.4) 72 (0.5) 22.2 (−11.9 to 46.1)
Influenza B 24 (0.2) 45 (0.3) 46.7 (10.6 to 68.9) 24 (0.2) 43 (0.3) 44.2 (5.9 to 67.6)
* Laboratory confirmation of influenza was determined by a positive result on culture of a nasopharyngeal swab, a positive polymerasechainreaction assay, or both. For laboratory
confirmed influenza assessments, similarity was determined by ferret antigenicity testing complemented by genetic sequencing.
† For cultureconfirmed influenza assessments, similarity was determined solely by the ferret antigenicity testing method.
Table 4. Hemagglutination Inhibition Immunogenicity of High-Dose Vaccine and Standard-Dose Vaccine against Influenza Viral Types and Subtypes Contained in the Vaccine.*
Viral Type/
Subtype Year 1 Year 2 Combined†
IIV3HD IIV3SD IIV3HD IIV3HD IIV3SD IIV3HD IIV3HD IIV3SD IIV3HD

(N = 2375) (N = 2382) vs. IIV3SD (N = 2879) (N = 2872) vs. IIV3SD (N = 5254) (N=5254) vs. IIV3SD
ratio of geometric ratio of geometric ratio of geometric

geometric mean titer mean titers geometric mean titer mean titers geometric mean titer mean titers
(95% CI‡) (95% CI‡) (95% CI‡) (95% CI‡) (95% CI‡) (95% CI‡)
A/H1N1 481.8 271.8 1.8 (1.6–1.9) 407.0 227.4 1.8 (1.7–1.9) 439.2 246.6 1.8 (1.7–1.9)
(457.7–507.1) (257.4–287.1) (390.2–424.4) (216.8–238.5) (425.1–453.8) (237.9–255.6)
A/H3N2 685.5 349.8 2.0 (1.8–2.1) 460.0 252.8 1.8 (1.7–1.9)
(651.4–721.4) (332.1–368.6) (440.8–480.0) (241.6–264.4)
n engl j med 371;7

B 138.1 97.6 1.4 (1.3–1.5) 98.2 61.8 1.6 (1.5–1.7)
(132.2– 144.2) (93.3–102.0) (94.5–102.0) (59.4–64.2)
percentage-point percentage-point percentage-point
38
nejm.org
% with seroprotection difference % with seroprotection difference % with seroprotection difference
(95% CI§) (95% CI¶) (95% CI§) (95% CI¶) (95% CI§) (95% CI¶)
A/H1N1 98.1 94.2 3.9 (2.8–5.0) 98.8 93.3 5.5 (4.5–6.5) 98.5 93.7 4.8 (4.1–5.5)

(97.5–98.6) (93.2–95.1) (98.3–99.2) (92.3–94.2) (98.1–98.8) (93.0–94.3)
A/H3N2 99.2 96.5 2.7 (1.9–3.5) 98.6 95.0 3.6 (2.7–4.5)
(98.7–99.5) (95.6–97.2) (98.2–99.0) (94.2–95.8)
August 14, 2014

B 91.6 83.9 7.7 (5.9–9.6) 86.2 72.8 13.4 (11.4–15.5)
(90.4–92.7) (82.3–85.3) (84.9–87.4) (71.1–74.4)
* The number of participants in the highdose and standarddose categories are those participants in the full analysis set and the immunogenicity subset who had at least one hemagglu

tination inhibition (HAI) assay result for the year. Seroprotection is defined as an HAI titer of at least 1:40.
† The type A (H1N1) virus used to make the 2012–2013 influenza vaccine was the same virus used to make the 2011–2012 vaccine, but the type A (H3N2) and type B viruses used to
make the 2012–2013 influenza vaccine were different from those in the 2011–2012 influenza vaccine, so only the results from the type A (H1N1) strain can be combined.
‡ The geometric mean titer confidence intervals were calculated according to the t distribution and the assumption that log (HAI titer) follows a normal distribution.
§ The confidence intervals were calculated with the use of the Clopper–Pearson exact method.15
¶ The confidence intervals were calculated with the use of the Newcombe–Wilson score method.16
643
through influenza illnesses caused by strains simi- considered more burdensome than other viral
lar to the vaccine could be prevented. types and subtypes in older adults,1,2 it is expected
This study provides an estimate of relative ef- that a benefit of IIV3-HD in this population will
ficacy (i.e., the efficacy of IIV3-HD as compared remain even in the context of quadrivalent stan-
with IIV3-SD). The absolute efficacy of IIV3-HD dard-dose vaccines.
can only be inferred, on the basis of estimates The safety data in this study are consistent
external to the study of the absolute efficacy of with the data in previous studies of IIV3-HD.9,26,31
standard-dose vaccines. Previous studies have sug- Moreover, significantly fewer IIV3-HD recipients
gested that inactivated vaccines similar to IIV3- than IIV3-SD recipients reported serious adverse
SD provide approximately 50% protection against events, which suggests that IIV3-HD may protect
influenza in older adults.5,7 Assuming 50% abso- against the occurrence of influenza-related seri-
lute efficacy for IIV3-SD in older adults, the ab- ous events.
solute efficacy of IIV3-HD would be estimated at This study has several limitations. First, some
62%, a level of protection similar to that seen of the efficacy estimates according to influenza
with standard-dose vaccines in younger adults.19,25 type, definitions of secondary illness, and con-
This estimate is consistent with an immunoge- firmation methods were based on a limited
nicity study that showed that immune responses number of cases and may therefore lack suffi-
induced by IIV3-HD in adults 65 years of age or cient precision. Second, only a minority of influ-
older were similar to those observed with IIV3- enza viruses identified in the study were charac-
SD in younger adults.26 terized as similar to the vaccine. Different results
This study included two heterogeneous influ- might be obtained in years when the relatedness
enza seasons. The first had low influenza activ- of vaccine and circulating strains differed mate-
ity and was characterized by a moderate-to-good rially from that observed in the study seasons.
match between the vaccine and circulating strains27; Third, although the study allowed inclusion of
the second had high influenza activity28,29 and persons with high-risk conditions, participants
was characterized by mismatch between predomi- were excluded if they had moderate or severe
nant circulating strains and egg-propagated vac- acute illnesses or if they were deemed unable to
cines such as those tested in this study.29,30 Despite comply with study procedures. Extrapolation of
the substantial differences in these two influenza study results to such persons should be made with
seasons, IIV3-HD showed significant efficacy as caution.
compared with IIV3-SD against the primary end In conclusion, this study showed that IIV3-
point in each season, a finding that provides reas- HD as compared with IIV3-SD significantly im-
surance that the benefit of IIV3-HD persists despite proved protection against laboratory-confirmed
varying seasonal conditions. influenza illness. It also showed that IIV3-HD
The clinical benefit shown in this study may was associated with superior immune responses
translate into public health benefits. The effec- as compared with IIV3-SD.
tiveness analyses signaled a favorable effect of Supported by Sanofi Pasteur.
Disclosure forms provided by the authors are available with
IIV3-HD on the prevention of hospitalization, the full text of this article at NEJM.org.
pneumonia, cardiorespiratory events, medication We thank the study participants, the investigators from the
126 participating research sites, the members of the indepen-
use, and nonroutine medical visits. Since influ- dent data and safety monitoring committee, and the Sponsor’s
enza infections with type A (H3N2) viruses are Study Team for their contributions to the trial.
References
1. Thompson WW, Shay DK, Weintraub sory Committee on Immunization Prac- The efficacy of influenza vaccination in
E, et al. Mortality associated with influ- tices (ACIP), 2010. MMWR Recomm Rep elderly individuals: a randomized double-
enza and respiratory syncytial virus in the 2010;59 (RR-8):1-62. [Errata, MMWR Re- blind placebo-controlled trial. JAMA
United States. JAMA 2003;289:179-86. comm Rep 2010;59:993, 1147.] 1994;272:1661-5.
2. Thompson WW, Shay DK, Weintraub 4. Palache A. Seasonal influenza vaccine 6. Gross PA, Hermogenes AW, Sacks HS,
E, et al. Influenza-associated hospitaliza- provision in 157 countries (2004-2009) Lau J, Levandowski RA. The efficacy of
tions in the United States. JAMA 2004; and the potential influence of national influenza vaccine in elderly persons: a
292:1333-40. public health policies. Vaccine 2011;29: meta-analysis and review of the literature.
3. Fiore AE, Uyeki TM, Broder K, et al. 9459-66. Ann Intern Med 1995;123:518-27.
Prevention and control of influenza with 5. Govaert TM, Thijs CT, Masurel N, 7. Praditsuwan R, Assantachai P, Wasi
vaccines: recommendations of the Advi- Sprenger MJ, Dinant GJ, Knottnerus JA. C, Puthavatana P, Kositanont U. The effi-

39 August 14, 2014
cacy and effectiveness of influenza vacci- parison of seven methods. Stat Med 1998; 24. Atmar RL, Patel SM, Keitel WA. In-
nation among Thai elderly persons living 17:857-72. tanza: a new intradermal vaccine for sea-
in the community. J Med Assoc Thai 2005; 16. Newcombe RG. Interval estimation sonal influenza. Expert Rev Vaccines
88:256-64. for the difference between independent 2010;9:1399-409.
8. Couch RB, Keitel WA, Cate TR. Im- proportions: comparison of eleven meth- 25. DiazGranados CA, Denis M, Plotkin
provement of inactivated influenza virus ods. Stat Med 1998;17:873-90. [Erratum, S. Seasonal influenza vaccine efficacy and
vaccines. J Infect Dis 1997;176:Suppl 1: Stat Med 1999;18:1293.] its determinants in children and non-
S38-44. 17. Simonsen L, Reichert TA, Viboud C, elderly adults: a systematic review with
9. Falsey AR, Treanor JJ, Tornieporth N, Blackwelder WC, Taylor RJ, Miller MA. meta-analyses of controlled trials. Vac-
Capellan J, Gorse GJ. Randomized, dou- Impact of influenza vaccination on sea- cine 2012;31:49-57.
ble-blind controlled phase 3 trial compar- sonal mortality in the US elderly popula- 26. Tsang P, Gorse GJ, Strout CB, et al.
ing the immunogenicity of high-dose and tion. Arch Intern Med 2005;165:265-72. Immunogenicity and safety of Fluzone in-
standard-dose influenza vaccine in adults 18. McElhaney JE. The unmet need in the tradermal and high-dose influenza vac-
65 years of age and older. J Infect Dis elderly: designing new influenza vaccines cines in older adults ≥65 years of age: a
2009;200:172-80. for older adults. Vaccine 2005;23:Suppl 1: randomized, controlled, phase II trial.
10. Govaert TM, Dinant GJ, Aretz K, S10-S25. Vaccine 2014;32:2507-17.
Knottnerus JA. The predictive value of in- 19. Osterholm MT, Kelley NS, Sommer A, 27. Update: influenza activity — United
fluenza symptomatology in elderly peo- Belongia EA. Efficacy and effectiveness of States, 2011–12 season and composition
ple. Fam Pract 1998;15:16-22. influenza vaccines: a systematic review of the 2012–13 influenza vaccine. MMWR
11. Talbot HK, Falsey AR. The diagnosis and meta-analysis. Lancet Infect Dis Morb Mortal Wkly Rep 2012;61:414-20.
of viral respiratory disease in older adults. 2012;12:36-44. [Erratum, Lancet Infect 28. Estimated influenza illnesses and hos-
Clin Infect Dis 2010;50:747-51. Dis 2012;12:655.] pitalizations averted by influenza vaccina-
12. De Villiers PJ, Steele AD, Hiemstra 20. Banzhoff A, Pellegrini M, Del Giudice tion — United States, 2012–13 influenza
LA, et al. Efficacy and safety of a live at- G, Fragapane E, Groth N, Podda A. MF59- season. MMWR Morb Mortal Wkly Rep
tenuated influenza vaccine in adults 60 adjuvanted vaccines for seasonal and pan- 2013;62:997-1000.
years of age and older. Vaccine 2009;28: demic influenza prophylaxis. Influenza 29. Influenza activity — United States,
228-34. Other Respir Viruses 2008;2:243-9. 2012–13 season and composition of the
13. Mäkelä MJ, Pauksens K, Rostila T, et 21. Mannino S, Villa M, Apolone G, et al. 2013–14 influenza vaccine. MMWR Morb
al. Clinical efficacy and safety of the oral- Effectiveness of adjuvanted influenza vac- Mortal Wkly Rep 2013;62:473-9.
ly inhaled neuraminidase inhibitor zana- cination in elderly subjects in northern 30. Recommended composition of influ-
mivir in the treatment of influenza: a ran- Italy. Am J Epidemiol 2012;176:527-33. enza virus vaccines for use in the 2013–
domized, double-blind, placebo-controlled 22. Skowronski DM, De Serres G, Janjua 2014 northern hemisphere influenza sea-
European study. J Infect 2000;40:42-8. NZ, Hottes TS. Re: “effectiveness of adju- son. Wkly Epidemiol Rec 2013;88:101-14.
14. Kendal AP, Pereira MS, Skehel JJ. vanted influenza vaccination in elderly 31. DiazGranados CA, Dunning AJ, Jor-
Hemagglutination inhibition. In: Kendal subjects in northern Italy”. Am J Epide- danov E, Landolfi V, Denis M, Talbot HK.
AP, Pereira MS, Skehel JJ, eds. Concepts miol 2013;177:593-4. High-dose trivalent influenza vaccine
and procedures for laboratory-based in- 23. McElhaney JE, Beran J, Devaster JM, et compared to standard dose vaccine in el-
fluenza surveillance. Atlanta: Centers for al. AS03-adjuvanted versus non-adjuvant- derly adults: safety, immunogenicity and
Disease Control and Prevention, 1982: ed inactivated trivalent influenza vaccine relative efficacy during the 2009-2010
B19-B35. against seasonal influenza in elderly peo- season. Vaccine 2013;31:861-6.
15. Newcombe RG. Two-sided confidence ple: a phase 3 randomised trial. Lancet Copyright © 2014 Massachusetts Medical Society.
intervals for the single proportion: com- Infect Dis 2013;13:485-96.
n engl j med 371;7 nejm.org August 14, 2014 645

40
High-Dose Influenza Vaccine No More Effective than Standard-Dose

in Older Adults
In a large retrospective cohort study, veterans aged ≥65 receiving high-dose influenza vaccine did not have
lower hospitalization rates for influenza or pneumonia.
Richardson DM et al. Comparative effectiveness of high-dose versus standard-dose influenza vaccination in communi-
ty-dwelling veterans. Clin Infect Dis 2015 Jul 15; 61:171. (http://dx.doi.org/10.1093/cid/civ261)
ARTICLE SUMMARY
Influenza-related morbidity and mortality are high among older adults. Knowledge that antibody response to in-
fluenza vaccination decreases with age led to the development of a high-dose (HD) trivalent inactivated vaccine
that contains four times the standard amount of hemagglutinin antigen per strain of influenza virus. Although
a large-scale randomized trial involving adults aged ≥65 showed that the HD vaccine provided better protection
against influenza than standard-dose (SD) preparations, neither the CDC nor the Advisory Committee on
Immunization Practices has voiced a preference for one product over the other in this population.
To determine the relative effectiveness of the HD vaccine among older adults in a real-world setting, researchers
conducted a retrospective cohort study involving community-dwelling patients aged ≥65 who received influenza
vaccine at a Veterans Health Administration medical center during the 2010–2011 influenza season. They com-
pared hospitalization rates for influenza or pneumonia between 25,714 HD-vaccine recipients and 139,511 SD-
preparation recipients. Hospitalizations occurring ≥2 weeks after vaccination and during the regional influenza
season were included.
The overall hospitalization rate for pneumonia or influenza was 0.3% for both groups of vaccine recipients.
Among patients aged ≥85, however, the hospitalization rate for influenza or pneumonia was significantly lower
with HD than with SD vaccine (0.3% vs. 0.7%; P=0.02).
COMMENT
The primary study outcome — hospitalization for influenza or pneumonia — may not be a refined enough meas
urement to ascertain the benefit of the HD vaccine. Although no overall difference was seen, the HD vaccine’s
greater relative effectiveness for patients aged ≥85 was clear.
— Neil M. Ampel, MD
University of Arizona College of Medicine, Tucson
NEJM Journal Watch Infectious Diseases, published July 29, 2015
41
Original Article
Influenza Vaccination of Pregnant Women

and Protection of Their Infants
Shabir A. Madhi, M.D., Ph.D., Clare L. Cutland, M.D., Locadiah Kuwanda, M.Sc.,
Adriana Weinberg, M.D., Andrea Hugo, M.D., Stephanie Jones, M.D.,
Peter V. Adrian, Ph.D., Nadia van Niekerk, B.Tech., Florette Treurnicht, Ph.D.,
Justin R. Ortiz, M.D., Marietjie Venter, Ph.D., Avy Violari, M.D.,
Kathleen M. Neuzil, M.D., Eric A.F. Simões, M.D., Keith P. Klugman, M.D., Ph.D.,
and Marta C. Nunes, Ph.D., for the Maternal Flu Trial (Matflu) Team*
A BS T R AC T
BACKGROUND
The authors’ affiliations are listed in the There are limited data on the efficacy of vaccination against confirmed influenza
Appendix. Address reprint requests to in pregnant women with and those without human immunodeficiency virus (HIV)
Dr. Madhi or Dr. Nunes at the Respirato-
ry and Meningeal Pathogens Research infection and protection of their infants.
Unit, Chris Hani Rd., Chris Hani–Baragwa-
nath Hospital, New Nurses Residence– METHODS
1st Fl. West Wing, Bertsham, Gauteng We conducted two double-blind, randomized, placebo-controlled trials of trivalent
2013, South Africa, or at shabirm@nicd inactivated influenza vaccine (IIV3) in South Africa during 2011 in pregnant women
.ac.za or nunesm@rmpru.co.za.
infected with HIV and during 2011 and 2012 in pregnant women who were not
*A complete list of investigators in the infected. The immunogenicity, safety, and efficacy of IIV3 in pregnant women and
Matflu Trial is available in the Supple-
mentary Appendix, available at NEJM.org. their infants were evaluated until 24 weeks after birth. Immune responses were
measured with a hemagglutination inhibition (HAI) assay, and influenza was diag-
Drs. Madhi and Nunes contributed equal-
ly to this article. nosed by means of reverse-transcriptase–polymerase-chain-reaction (RT-PCR) assays
of respiratory samples.
N Engl J Med 2014;371:918-31.
DOI: 10.1056/NEJMoa1401480
RESULTS
Copyright © 2014 Massachusetts Medical Society.
The study cohorts included 2116 pregnant women who were not infected with HIV
Related Material: and 194 pregnant women who were infected with HIV. At 1 month after vaccina-
Supplementary Appendix tion, seroconversion rates and the proportion of participants with HAI titers of
Correspondence 1:40 or more were higher among IIV3 recipients than among placebo recipients in
both cohorts. Newborns of IIV3 recipients also had higher HAI titers than new-
borns of placebo recipients. The attack rate for RT-PCR–confirmed influenza among
both HIV-uninfected placebo recipients and their infants was 3.6%. The attack
rates among HIV-uninfected IIV3 recipients and their infants were 1.8% and 1.9%,
respectively, and the respective vaccine-efficacy rates were 50.4% (95% confidence
interval [CI], 14.5 to 71.2) and 48.8% (95% CI, 11.6 to 70.4). Among HIV-infected
women, the attack rate for placebo recipients was 17.0% and the rate for IIV3 re-
cipients was 7.0%; the vaccine-efficacy rate for these IIV3 recipients was 57.7%
(95% CI, 0.2 to 82.1).
CONCLUSIONS
Influenza vaccine was immunogenic in HIV-uninfected and HIV-infected pregnant
women and provided partial protection against confirmed influenza in both groups
of women and in infants who were not exposed to HIV. (Funded by the Bill and
Melinda Gates Foundation and others; ClinicalTrials.gov numbers, NCT01306669
and NCT01306682.)
918 42
n engl j med 371;10 nejm.org September 4, 2014
Source: The New England Journal of MedicineInfluenza Vaccination of Pregnant Women
P
regnant women are designated as a HIV infection and the other involving pregnant
priority group for seasonal influenza vac- women with HIV infection, to evaluate the safety,
cination by the World Health Organization immunogenicity, and efficacy of IIV3 in these
(WHO)1 because of their heightened susceptibil- women and in their infants until 24 weeks post
ity to severe influenza from the second trimester partum.
to the early postpartum period.2,3 Since pregnancy
is associated with immunomodulation, including Me thods
the attenuation of cell-mediated immune re-
sponses,4 the efficacy of inactivated influenza Study Design, Objectives, and Oversight
vaccine (IIV) in pregnant women may differ The two studies were randomized, double-blind,
from its efficacy in healthy nonpregnant women placebo-controlled trials conducted in Soweto,
and in men.5 This difference in vaccine efficacy South Africa, where antenatal HIV testing is rou-
could be further accentuated in pregnant women tine. The enrollment of HIV-uninfected pregnant
infected with the human immunodeficiency vi- women was initiated at four antenatal clinics
rus (HIV), who are at heightened risk for severe before the onset of the 2011 influenza season
influenza illness6-8 because of HIV-related immu- (March 3 through August 4) and the 2012 season
nosuppression.9-14 (March 6 through July 2). Members of the HIV-
Reduced attack rates for all-cause febrile re- infected cohort were enrolled at the same facili-
spiratory illness among women vaccinated against ties (March 3 through June 2, 2011). Eligibility
influenza during pregnancy were observed dur- criteria for both cohorts included an age of 18 to
ing the 1957 Asian influenza pandemic in the 38 years, an estimated gestation of 20 to 36 weeks,
United States and in a later randomized, con- and the ability to understand and comply with
trolled trial in Bangladesh.15,16 A recent case–con- planned study procedures. (The methods used to
trol study in the United States also reported that determine gestational age and the exclusion cri-
pregnant women with influenza confirmed by teria are provided in the Supplementary Appen-
means of reverse-transcriptase–polymerase-chain- dix, which is available with the full text of this
reaction (RT-PCR) assay (hereafter referred to as article, along with the original and final protocols,
confirmed influenza) were 44 to 53% less likely at NEJM.org.) All HIV-infected mother–infant dy-
to have been vaccinated with trivalent IIV (IIV3) ads in 2011 and a subset of HIV-uninfected dyads
than controls who did not have influenza.17 To were included in a nested immunogenicity study
our knowledge, no published randomized, con- that was performed during each year of the trial.
trolled trial has assessed the efficacy of IIV in The primary objectives for the cohort of women
preventing confirmed influenza in pregnant without HIV infection were to evaluate the effi-
women with HIV infection and those without HIV cacy of IIV3 vaccination during pregnancy in
infection.18,19 protecting their infants against confirmed influ-
The vaccination of pregnant women may also enza through 24 weeks of age and to compare
confer partial protection against confirmed in- seroconversion rates between IIV3 recipients and
fluenza in their infants, as reported in the Ban- placebo recipients 1 month after vaccination. Sec-
gladeshi trial (in which infants whose mothers ondary objectives included measuring vaccine ef-
had been vaccinated with IIV3 were 63% less ficacy against confirmed influenza in all women
likely to have influenza [confirmed by means of until 24 weeks post partum. In the HIV-infected
enzyme-linked immunosorbent assay]).16 How- cohort, an additional primary objective was to
ever, observational studies have had conflicting evaluate the immunogenicity of IIV3 in the wom-
results with regard to the efficacy of IIV3 vacci- en. Secondary objectives for this cohort included
nation during pregnancy in protecting infants measuring vaccine efficacy against confirmed
against all-cause respiratory illness.20-23 The pro- influenza in the women and their infants until
tection of infants 6 months of age or younger 24 weeks post partum. Additional secondary ob-
against influenza is a public health priority. These jectives for both cohorts are listed in the Supple-
infants are at high risk for influenza-associated mentary Appendix.
hospitalization, and their immune responses to The studies were approved by the Human
IIV vaccination are poor.24-26 We conducted two Research Ethics Committee of the University of
studies, one involving pregnant women without the Witwatersrand and were conducted in accor-
n engl j med 371;10 nejm.org September 4, 2014 919

43
dance with Good Clinical Practice guidelines. enza virus by means of an RT-PCR assay. The labo-
Written informed consent was obtained from all ratory methods used for sample collection, in-
participants. All the authors vouch for the fidel- fluenza-virus identification, and genotyping are
ity of this report to the protocols and the com- detailed in the Supplementary Appendix.
pleteness of the data and analyses. The funders
did not participate in the conduct of the study, Safety
data collection, analyses of the data, or the writ- Women enrolled in the nested immunogenicity
ing of the manuscript. subsets were provided with diary cards on which
to document possible local and systemic reac-
Randomization and Study Treatment tions to vaccination for 1 week. Digital thermom-
Participants were randomly assigned in a 1:1 ratio eters were provided to measure oral temperature
to receive IIV3 or placebo. Randomization was in women and axillary temperature in infants
performed by the study statistician with the use of after vaccination and during illness. Serious ad-
computer-generated assignments. With the excep- verse events were recorded and graded through-
tion of the statistician and the pharmacist, study out the study period with the use of an estab-
personnel were unaware of the group assign- lished system.29
ments, as were the study participants.
Influenza vaccine (Vaxigrip, lot number G05831 Statistical Analysis
in 2011 and H7221-2 in 2012; Sanofi Pasteur) In the HIV-uninfected cohort, the sample size was
was purchased by the study team. The vaccine based on the primary outcome of vaccine efficacy
contained 15 μg each of A/California/7/2009 in the infants. The sample size for the HIV-unin-
(A/[H1N1]pdm09), A/Victoria/210/2009 (A/H3N2), fected cohort was outcome-driven. We aimed at
and a B/Brisbane/60/2008–like virus (B/Victoria), identifying at least 27 cases of confirmed illness
as recommended by WHO for the Southern Hemi- caused by influenza virus in infants up to 24 weeks
sphere in 2011 and 2012.27,28 The study pharma- of age in order to detect a 70% reduction in con-
cist used a 2-ml syringe to draw 0.5 ml of vaccine firmed influenza among the infants, with 80%
for the women receiving IIV3 and 0.5 ml of sterile power. The sample size required for the HIV-
0.9% normal saline solution for the women re- infected cohort was 180 participants, which pro-
ceiving placebo. The two preparations were mac- vided 90% power to detect a difference of at least
roscopically indistinguishable. The vaccines were 67% in rates of seroconversion to individual vac-
administered into the deltoid muscle by study staff. cine strains between IIV3 recipients and placebo
recipients.
Serologic Efficacy and Immunogenicity The immunogenicity analyses included com-
In 2011 and 2012, all mothers and infants in the parisons of geometric mean titers between study
HIV-infected cohort and a nested group of 180 groups and of the increase in titers from base-
HIV-uninfected mother–infant dyads were enrolled line to 1 month after vaccination. We performed
in an intensive safety and immunogenicity study. a two-sided, two-sample t-test and calculated the
The timing of blood-sample collection and pro- corresponding 95% confidence intervals for the
cessing, the method used for antibody testing titers, using logarithmic transformation for all
(hemagglutination inhibition [HAI]), and the stan- values. Post-vaccination analyses of immune re-
dard criteria used to qualify HAI results as sero- sponse were adjusted for baseline HAI titers and
conversion and as seroprotective or seronegative between-group differences in baseline character-
titers are provided in the Supplementary Appendix. istics. The proportions of participants in each
group who underwent seroconversion and the
Vaccine Efficacy proportions of participants who had local or sys-
Active surveillance for influenza-like illness (as temic reactions were compared by means of chi-
defined in the protocols) was conducted through square or Fischer’s exact tests. Vaccine efficacy
weekly contact with participants (see the Supple- was calculated with the use of the formula (1–IL)/
mentary Appendix for details). Women and infants IP, where IL is the case incidence rate in the vac-
identified as having influenza-like illness, and cinated group and IP is the case incidence rate in
those unexpectedly presenting with or hospitalized the placebo group; 95% confidence intervals were
for any respiratory illness, were tested for influ- calculated and between-group differences were
920 n engl j med 371;10 nejm.org September 4, 2014

44
tested. Estimates of vaccine efficacy were adjusted We enrolled 194 pregnant women who were in-
for differences in maternal age and status with fected with HIV; 100 were randomly assigned to
respect to antiretroviral treatment at enrollment, the group vaccinated with IIV3 and 94 to the
which were prevalent in the HIV-infected cohort. placebo group (Fig. S2 in the Supplementary Ap-
For vaccine efficacy end points, data were cen- pendix). The mean maternal age at enrollment was
sored after the first episode of a specific clinical 28.2 years, and the mean gestational age was
outcome. Between-group differences in the time 27.3 weeks (Table 2). The baseline median CD4+
to a first episode of confirmed influenza were T-cell count in the HIV-infected women was 393.5
compared in survival analyses by means of the cells per cubic millimeter; 12.6% of the women
log-rank test. (24 of 190) had counts of less than 200 cells per
The intention-to-treat analyses included mater- cubic millimeter. The median level of plasma
nal outcomes from receipt of vaccine or placebo HIV-1 RNA was 1067 copies per milliliter; 23.0%
to 175 days after birth and infant outcomes from of the women (43 of 187) had undetectable levels
birth to 175 days of age. The per-protocol analyses of HIV-1 RNA. There were 100 live births in the
included maternal outcomes that occurred 14 or group vaccinated with IIV3 and 88 live births in
more days after receipt of vaccine or placebo; per- the placebo group among the 183 HIV-infected
protocol analyses of outcomes for infants were women who remained in the study until delivery.
limited to those born at least 28 days after their
mother’s vaccination, those whose gestational age Immunogenicity of the Vaccine
at birth was at least 37 weeks, or those who had HIV-Uninfected Cohort
a birth weight of at least 2500 g. The per-protocol Of the 376 participants in the immunogenicity
immunogenicity analysis was limited to women subset, 142 IIV3 recipients and 148 placebo recipi-
from whom a blood sample was obtained 28 to 35 ents were included in the per-protocol analysis
days after vaccination and to infants from whom of immunogenicity (Fig. S3 and Table S1 in the
a blood sample was obtained within 7 days after Supplementary Appendix). Post-vaccination geo-
birth and who had a gestational age of at least 37 metric mean titers increased from baseline by a
weeks at birth or a birth weight of at least 2500 g. factor of 6 to 10 in IIV3 recipients and were sig-
An exploratory analysis was performed with the nificantly higher for all three vaccine strains, as
use of extended windows for obtaining blood compared with titers in placebo recipients (Ta-
samples after vaccination (28 to 42 days) and after ble 3). Seroconversion rates for strains A(H1N1)
delivery (up to 14 days). pdm09, A(H3N2), and B(Victoria) among IIV3 re-
Study data were collected and managed with cipients were 72.5%, 64.8%, and 92.3%, respec-
the use of Research Electronic Data Capture tively; rates among placebo recipients were 8.1%,
(REDCap), version 5.9.13.30 All statistical analyses 2.7%, and 2.0%, respectively (P<0.001 for all com-
were conducted with the use of Stata software, parisons). A greater proportion of IIV3 recipients
version 12.1. All P values were two-sided, and a had seroprotective HAI titers after vaccination,
value of 0.05 or less was considered to indicate as compared with placebo recipients (Table 3).
statistical significance. Newborns of IIV3 recipients had higher HAI
geometric mean titers for all vaccine strains than
did newborns of placebo recipients and were also
R e sult s
more likely to have an HAI titer of 1:40 or higher
Baseline Characteristics of the Study for A/(H1N1)pdm09 (81.1% vs. 34.0%), A(H3N2)
Participants (60.0% vs. 17.5%), and B(Victoria) (82.1% vs.
We enrolled 2116 black African pregnant women 43.7%) (P<0.001 for all comparisons) (Table 3). The
who were not infected with HIV; 1062 were ran- ratios of HAI titers in newborns to maternal ti-
domly assigned to the group vaccinated with IIV3 ters were 0.7 to 1.0 for the three vaccine strains
and 1054 to the group that received placebo. There and were similar between study groups, with the
were 1026 live infants born to IIV3 recipients and exception of higher ratios for B/Victoria in the
1023 live infants born to placebo recipients (Fig. S1 placebo group (Table 3). Similar findings were
in the Supplementary Appendix). The mean ma- observed in the immunogenicity analyses when the
ternal age at enrollment was 26.2 years, and the window for obtaining blood samples was extended
mean gestational age was 26.8 weeks (Table 1). (Tables S2 and S3 in the Supplementary Appendix).

45 September 4, 2014 921
Table 1. Baseline Characteristics of HIV-Uninfected Pregnant Women, Fetal Outcomes, and Newborn Characteristics.*
Characteristic or Outcome Overall IIV3 Placebo

Women
Total no. of women 2116 1062 1054
Mean age — yr 26.2±5.3 26.4±5.3 25.9±5.3
Body-mass index†
Median 27.6 28.0 27.4
Interquartile range 24.3–31.9 24.6–32.1 24.1–31.6
Mean gestational age at enrollment — wk 26.8±4.4 26.7±4.4 26.9±4.4
Gravidity
Median 2.0 2.0 2.0
Parity
Median 1.0 1.0 1.0
Fetuses and newborns
Total no. of known outcomes 2081 1044 1037
Miscarriage, gestational age <28 wk — no. (%)‡ 8 (0.4) 3 (0.3)§ 5 (0.5)
Stillbirth, gestational age ≥28 wk — no. (%)‡ 24 (1.2) 15 (1.4) 9 (0.9)
Live birth — no. (%) 2049 (98.5) 1026 (98.3) 1023 (98.7)
Twin birth — no. (%) 110 (5.3) 60 (5.7) 50 (4.8)
Type of delivery — no. (%)
Normal vaginal delivery 1393 (66.9) 700 (67.0) 693 (66.8)
Cesarian section 656 (31.5) 326 (31.2) 330 (31.8)
Preterm birth, gestational age <37 wk — no. (%)¶ 204 (10.0) 108 (10.5) 96 (9.4)
Overall birth weight — kg¶‖**
Median 3.1 3.1 3.1
Range 0.5–4.8 0.5–4.6 0.7–4.8
Birth weight of infants born during influenza season — kg¶**††
Median 3.1 3.1 3.1
Range 0.4–4.8 0.6–4.6 0.4–4.8
Birth weight of <2500 g — no. (%)¶‖** 255 (12.5) 133 (13.0) 122 (12.0)
Death after delivery — no. (%)¶‡‡ 22 (1.1) 12 (1.2) 10 (1.0)
Admission to neonatal nursery — no. (%)¶ 142 (6.9) 71 (6.9) 71 (6.9)
* Plus–minus values are means ±SD. HIV denotes human immunodeficiency virus, and IIV3 trivalent inactivated influ-
enza vaccine.
† The calculations for body-mass index (the weight in kilograms divided by the square of the height in meters) were
based on 796 women in the IIV3 group and 789 women in the placebo group.
‡ Details on miscarriages and stillbirths are provided in Table S9 in the Supplementary Appendix.
§ This number includes one medically assisted termination of pregnancy.
¶ Numbers and percentages are based on live births.
‖ Numbers are based on 1024 observations in the IIV3 group and 1021 observations in the placebo group.
** No significant difference in birth weight between the study groups was observed, even when weight was stratified
according to sex.
†† Numbers are based on 857 observations in the IIV3 group and 874 observations in the placebo group.
‡‡ Further details on newborn deaths are provided in Table S10 in the Supplementary Appendix.

46 September 4, 2014
HIV-Infected Cohort fants of placebo recipients were A/H3N2 (14 in-

The demographic characteristics of HIV-infected fants [37%]), B/Victoria (11 [29%]), and B/Yamagata
women included in the per-protocol analysis of (12 [32%]). Among infants with confirmed in-
immunogenicity (Fig. S2 in the Supplementary fluenza, the numbers infected with the influen-
Appendix) are shown in Table S4 in the Supple- za virus strain carried by their mothers were as
mentary Appendix. The receipt of IIV3 resulted in follows: 1 of 19 infants in the IIV3 group (5.3%)
increases in geometric mean titers of antibodies and 11 of 37 in the placebo group (29.7%, P = 0.04).
to A/(H1N1)pdm09, A/H3N2, and B/Victoria by The vaccine efficacy estimates in the per-proto-
factors of 2.9, 2.4, and 3.2, respectively, and se- col analysis were similar to the estimates in the
roconversion rates of 42.9%, 35.7%, and 40.0%, intention-to-treat analysis for the women and
respectively. Furthermore, as compared with pla- their infants (Table S8 in the Supplementary Ap-
cebo recipients, a higher proportion of IIV3 recipi- pendix).
ents had seroprotective HAI titers after vaccina-
tion (Table S5 in the Supplementary Appendix). HIV-Infected Cohort
The newborns of IIV3 recipients had higher In the cohort of HIV-infected mothers, the at-
HAI geometric mean titers than did the newborns tack rate was lower among IIV3 recipients than
of placebo recipients. The ratio of geometric mean among placebo recipients (7.0% vs. 17.0%), indicat-
HAI titers ranged between 0.7 and 1.4; ratios ing an adjusted vaccine efficacy of 57.7% (95% CI,
were higher in the placebo group for A/(H1N1) 0.2 to 82.1) (Table 4 and Fig. 1). The attack rates
pdm09 (P = 0.05) and A/H3N2 (P = 0.01). In the IIV3 were 5.0% and 6.8% among infants of IIV3 re-
group, the proportion of newborns with an HAI cipients and infants of placebo recipients, respec-
titer of 1:40 or higher was 60.7% for A/(H1N1) tively (vaccine efficacy, 26.7%; P = 0.60) (Table 4
pdm09, 42.9% for A/H3N2, and 78.6% for B/Vic- and Fig. 1). In 6 of 11 infants of HIV-infected
toria; titers for all three strains were higher than mothers with confirmed influenza (55%), in-
those in the placebo group (Table S5 in the Sup- cluding 2 infants whose mothers received IIV3,
plementary Appendix). Similar findings were ob- the mothers were infected with the same strain
served in the immunogenicity analyses when ex- as their infants. Estimates of vaccine efficacy
tended windows for obtaining blood samples were among HIV-infected women and their infants were
used (Tables S6 and S7 in Supplementary Ap- similar in the per-protocol and intention-to-treat
pendix). analyses (Table S8 in the Supplementary Appendix).
Vaccine Efficacy Safety

HIV-Uninfected Cohort The details of solicited local and systemic reac-
The attack rates for confirmed influenza among tions in the HIV-uninfected cohort and the HIV-
placebo recipients and IIV3 recipients were 3.6% infected cohort are shown in Tables S9 and S10 in
and 1.8%, respectively, indicating a vaccine effi- the Supplementary Appendix. Injection-site reac-
cacy of 50.4% (95% confidence interval [CI], 14.5 tions (mainly mild to moderate) were more fre-
to 71.2) (Table 4 and Fig. 1). The vaccine efficacy quent among IIV3 recipients than among place-
was similar when determined according to the vac- bo recipients in both cohorts, but there were no
cine strain (46.1%; 95% CI, 6.4 to 69.0). For the other significant differences in solicited reac-
2 years of the study, the predominant strain of tions between the two study groups in either
influenza virus circulating among placebo recipi- cohort. Data on serious adverse events in both
ents was A/H3N2 (57.9%), for which an exploratory cohorts, including infant and maternal deaths
analysis identified a vaccine efficacy of 57.5% and hospitalizations, are shown in Tables S11 to
(95% CI, 7.6 to 80.4). S23 in the Supplementary Appendix. There were
The attack rate was lower among infants whose no significant between-group differences with
mothers were IIV3 recipients (1.9%) than among regard to rates of miscarriage, stillbirth, or pre-
those whose mothers were placebo recipients mature birth or birth weight in the HIV-uninfect-
(3.6%), indicating a vaccine efficacy of 48.8% ed cohort (Table 1, and Table S11 in the Supple-
(95% CI, 11.6 to 70.4) (Table 4 and Fig. 1). The mentary Appendix) and in the HIV-infected cohort
predominant strains causing illness in the 38 in- (Table 2). The findings were similar in an analy-

47 September 4, 2014 923
Source: The New England Journal of Medicine The n e w e ng l a n d j o u r na l of m e dic i n e
Table 2. Baseline Characteristics of HIV-Infected Pregnant Women, Fetal Outcomes, and Newborn Characteristics.*

Women
Total no. of women 194 100 94
Mean age — yr 28.2±5.1 27.2±4.9† 29.2±5.1†
Mean body-mass index‡ 28.7±5.3 29.2±5.0 28.3±5.7
Mean gestational age at enrollment — wk 27.3±3.8 27.6±3.9 26.9±3.7
Gravidity
Median 2.0 2.0 2.0
Parity
Median 1.0 1.0 1.0
HAART — no. (%)
At enrollment 60 (30.9) 22 (22.0)† 38 (40.4)†
At delivery 77 (39.7) 33 (33.0) 44 (46.8)
ART for prevention of mother-to-child HIV transmission
— no. (%)
At enrollment 86 (44.3) 54 (54.0)† 32 (34.0)†
At delivery 87 (44.9) 52 (52.0)§ 35 (37.2)§
CD4+ T-lymphocyte count — cells/mm3
At enrollment¶
Median 393.5 410.0 379.0
At post-vaccination visit‖
Median 399.5 371.0 409.0
At delivery**
Median 446.5 441.0 508.0
HIV-1 viral load — copies/ml
At enrollment††
Median 1067.0 1679.0 399.0
Interquartile range 61.0–13,923.0 118.5–15,906.5 39.0–9990.0
At post-vaccination visit‡‡
Median 620.0 812.0§ 292.5§
Interquartile range 39.0–8278.0 130.0–13,221.0 39.0–3056.0
At delivery§§
Median 265.0 222.0 356.0
Fetuses and newborns
Total no. of live-born infants 188 100 88
Twin birth — no. (%) 10 (5.3) 6 (6.0) 4 (4.5)
Type of delivery — no. (%)
Normal vaginal 111 (59.0) 58 (58.0) 53 (60.2)
Cesarian section 77 (41.0) 42 (42.0) 35 (39.8)
Preterm birth, gestational age <37 wk — no. (%) 26 (13.8) 13 (13.0) 13 (14.8)
924 n engl j med 371;10 nejm.org September 4, 2014

48
Source: The New England Journal ofInfluenza
Medicine Vaccination of Pregnant Women
Table 2. (Continued.)

Overall birth weight — kg¶¶
Median 3.0 2.9 3.0
Range 0.8–4.3 0.8–4.3 1.9–4.0
Birth weight of infants born during influenza season — kg‖‖
Median 3.0 3.0 3.0
Range 0.8–4.3 0.8–4.3 1.9–4.1
Birth weight of <2500 g — no. (%)¶¶ 29 (15.6) 14 (14.1) 15 (17.2)
Death after delivery — no. (%)*** 4 (2.1) 2 (2.0) 2 (2.3)
Admission to neonatal nursery — no. (%) 9 (4.8) 6 (6.0) 3 (3.4)
* Plus–minus values are means ± SD. ART denotes antiretroviral therapy, and HAART highly active antiretroviral therapy.
† P<0.01.
‡ Data are based on 81 observations in the IIV3 group and 74 observations in the placebo group.
§ P<0.05.
¶ Numbers are based on 97 observations in the IIV3 group and 93 observations in the placebo group.
‖ Numbers are based on 78 observations in the IIV3 group and 68 observations in the placebo group.
** Numbers are based on 77 observations in the IIV3 group and 69 observations in the placebo group.
†† Numbers are based on 96 observations in the IIV3 group and 91 observations in the placebo group.
‡‡ Numbers are based on 77 observations in the IIV3 group and 64 observations in the placebo group.
§§ Numbers are based on 72 observations in the IIV3 group and 65 observations in the placebo group.
¶¶ Numbers and percentages are based on 99 observations in the IIV3 group and 87 observations in the placebo group.
‖‖ Numbers are based on 88 observations in the IIV3 group and 80 observations in the placebo group.
*** Further details on newborn deaths are provided in Table S17 in the Supplementary Appendix.
sis stratified according to sex and an analysis desh is prolonged and perennial, limiting the
limited to births that occurred during the influ- generalizability of the Bangladesh study to tem-
enza season. perate regions with more discrete influenza sea-
sons, such as South Africa.16 The effectiveness of
IIV3 vaccination in protecting the infants of preg-
Discussion
nant women not infected with HIV is further cor-
The trials of IIV3 efficacy in HIV-infected and roborated by case–control studies in the United
HIV-uninfected pregnant women showed a reduc- States among American Indians (in which most
tion in confirmed cases of influenza in both co- cases were diagnosed through serologic testing)
horts. Furthermore, we observed protection against and by a study of influenza-associated hospital-
confirmed influenza among infants of IIV3 re- ization in a large urban hospital (in which the di-
cipients who were not exposed to HIV, with a agnosis was confirmed through fluorescent anti-
similar trend observed in infants exposed to HIV. body testing).20,22
IIV3 was found to be safe, and the vaccine was It has been proposed that higher HAI titers may
immunogenic in HIV-uninfected women and HIV- be required to provide protection against influ-
infected women, supporting the current WHO enza in children; in adults, an HAI titer of 1:40
recommendations for influenza vaccination dur- or higher is associated with 50% protection.31,32,33
ing pregnancy.19 Nevertheless, in our study the proportion of
The vaccine efficacy against confirmed influ- HIV-unexposed newborns in the IIV3 group with
enza in infants who were not exposed to HIV HAI titers of 1:40 or higher (60 to 82%) suggests
(45.6%; 95% CI, 2.4 to 69.7) was similar to that that this threshold may be associated with ap-
reported in Bangladesh (63%; 95% CI, 5 to 85).16 proximately 50% protection against confirmed
However, direct comparisons with the results of influenza in such infants. Further studies and
the Bangladesh study are limited because of dif- analyses are warranted to establish a serocorrelate
ferences in the assays used for viral detection. In for the protection of HAI antibodies acquired
addition, influenza virus circulation in Bangla- through the placenta. Another way in which the
n engl j med 371;10 nejm.org September

49 4, 2014 925
vaccination of pregnant women may protect their to influenza, even when the HIV infection is man-
infants against influenza involves prevention of the aged with antiretroviral treatment.
transmission of the virus to the infant by reduc- Extrapolation of the potential effect of influ-
ing the mother’s risk of influenza illness. This is enza vaccination in pregnancy solely on the basis
corroborated in part by our study, in which con- of HAI immune responses may be affected by
current confirmed influenza infection by homo- pregnancy-inducing immune tolerance, which
typic strains occurred more frequently in HIV- could influence vaccine effectiveness.4,5,36 The avail-
unexposed infants of placebo recipients than in ability of immunogenicity and efficacy data in our
HIV-unexposed infants of IIV3 recipients; infec- nested immunogenicity study helped to address
tions with homotypic strains also occurred con- this question. The proportions of HIV-uninfected
currently in 54.5% of the mothers of HIV-exposed IIV3 recipients with seroprotective HAI titers for
infants with confirmed influenza illness. A/(H1N1)pdm09, A/H3N2, and B/Victoria were
Unlike the study in Bangladesh, in which rates 93.3%, 78.0%, and 96.0%, respectively, with cor-
of clinical febrile respiratory illness were reduced responding overall vaccine efficacy against con-
by 29% and 36% in infants and mothers, respec- firmed influenza of 54.4%. These findings sug-
tively,16 our study did not show any reduction in gest that the threshold of 1:40 or higher was
the less specific outcomes of clinical influenza- predictive of an anticipated vaccine efficacy of
like illness or respiratory illness in either infants 50%, as proposed for healthy adults.31 Theoreti-
or mothers. The use of pneumococcal polysac- cally, the rate of naturally acquired immunity
charide vaccine as the comparator for the control among placebo recipients (20 to 45%) could have
group in the Bangladeshi study, as well as the use decreased the true efficacy of the vaccine, a con-
of different case definitions, may have contribut- sideration that would not be required in an im-
ed to the discrepant findings in the two studies. munologically naive placebo group of women not
Observational studies in the United States have infected with influenza.
also shown that maternal influenza vaccination Among IIV3 recipients in the HIV-infected co-
was not effective in protecting infants against all- hort, HAI titers of 1:40 or higher were observed
cause respiratory illness, indicating the lack of for A/H3N2 and A/(H1N1)pdm09 (48.6% and
specificity of such end points for the evaluation 68.6%, respectively). Consequently, the HAI titer
of influenza vaccine efficacy in infants.21,23 of 1:40 or higher, as a relative serocorrelate of pro-
The vaccine efficacy against confirmed influ- tection, would have paradoxically underestimat-
enza among HIV-uninfected women in our study ed vaccine efficacy against confirmed influenza
(per-protocol analysis, 54.4%; 95% CI, 19.5 to among the HIV-infected women (reported as
74.2) was similar to that reported in an observa- 70.6%). This observation is consistent with find-
tional study of vaccination against monovalent ings from our previous study involving HIV-infect-
A/(H1N1)pdm09 in pregnant Norwegian women ed adults.34 Among the HIV-infected cohort, IIV3
(70%; 95% CI, 66 to 75)33 and in a more recent did not affect the plasma HIV-1 viral load, mini-
case–control study in the United States (44 to mizing concern about the possibility of an in-
53%).17 The vaccine efficacy against confirmed increased risk of vertical transmission of HIV in
fluenza in HIV-infected pregnant women in our vaccinees.37,38
study (per-protocol analysis, 70.6%) was similar to One of the limitations of our study is the fact
that reported in a trial of IIV3 in South African that it was conducted at a single center; the gen-
HIV-infected men and women during the 2008 eralizability of the findings needs to be corrobo-
influenza epidemic (75.5%).34 The efficacy of the rated. Furthermore, the timing of enrollment was
vaccine in our study was also similar to that re- contingent on the commercial availability of IIV3.
ported in a meta-analysis of HIV-infected adults Although we initiated enrollment within 1 week
in high-income countries (66%; 95% CI, 36 to 82).35 after vaccine availability, vaccination extended
Notably, the attack rate among HIV-infected pla- into the influenza season, with some births oc-
cebo recipients in our study (17.0%) was greater curring after the influenza season. The difference
than that observed among HIV-uninfected plain duration of exposure to the influenza virus be-
cebo recipients (3.6%), a finding that highlights tween the women and their infants limited our
the greater susceptibility of HIV-infected persons ability to make a direct comparison of the attack
926 n engl j med 371;10 50

nejm.org September 4, 2014
Table 3. Immune Responses to IIV3 Vaccine in HIV-Uninfected Pregnant Women and Transplacental Transfer of Antibodies to Newborns.*
Measure A(H1N1)pdm09 A/H3N2 B/Victoria
IIV3 Placebo P Value† IIV3 Placebo P Value† IIV3 Placebo P Value†

Mothers
Total no. of mothers 142 148 142 148 142 148

Geometric mean HAI titer
(95% CI)
Before vaccination 30.7 (25.5–37.0) 27.2 (21.8–34.0) 0.41 14.6 (124–17.0) 14.0 (11.8–16.6) 0.74 17.6 (15.6–19.9) 15.4 (13.4–17.7) 0.15
After vaccination 213.4 (175.7–259.2) 29.1 (23.0–36.8) <0.001 87.8 (70.5–109.2) 16.0 (13.3–19.1) <0.001 175.5 (148.7–207.3) 16.9 (14.6–19.6) <0.001
P value‡ <0.001 0.30 <0.001 0.005 <0.001 0.05
Factor increase in geometric mean 6.9 (5.6–8.6) 1.1 (0.9–1.2) <0.001 6.0 (4.9–7.4) 1.1 (1.0–1.2) <0.001 10.0 (8.4–11.8) 1.1 (1.0–1.2) <0.001
HAI titer (95% CI)
HAI titer ≥1:40 — no. (%);
n engl j med 371;10

(95% CI)
Before vaccination 73 (51.4); 66 (44.6); 0.73 33 (23.2); 37 (25.0); 0.25 29 (20.4); 30 (20.3); 0.97
(42.9–59.9) (36.4–53.0) (16.6–31.1) (18.3–32.8) (14.1–27.7) (14.1–28.0)
51
nejm.org
After vaccination 133 (93.7); 71 (47.8); <0.001 112 (78.9); 40 (27.0); <0.001 138 (97.2); 43 (29.1); <0.001
(88.3–97.1) (40.4–57.0) (71.2–85.3) (20.1–35.0) (92.9–99.2) (21.9–37.1)
P value‡ 0.006 0.37 0.001 0.43 <0.001 <0.001

Seroconversion 1 mo after vaccina- 103 (72.5); 12 (8.1); <0.001 92 (64.8); 4 (2.7); <0.001 131 (92.3); 3 (2.0); <0.001
tion — no. (%); (95% CI) (64.4–79.7) (4.3–13.7) (56.3–72.6) (0.7–6.8) (86.6–96.1) (0.4–5.8)
Newborns
September 4, 2014
Total no. of newborns 95 103 95 103 95 103
Influenza Vaccination of Pregnant Women
Geometric mean HAI titer at ≤7 93.3 (71.1–122.4) 17.2 (13.8–22.7) <0.001 41.8 (31.6–55.3) 12.7 (10.6–15.1) <0.001 80.6 (64.9–100.1) 21.4 (17.8–25.7) <0.001
days of age (95% CI)

Ratio of newborn to maternal HAI 0.7 (0.6–0.8) 0.7 (0.6–0.8) 0.78 0.7 (0.6–0.9) 0.7 (0.6–0.8) 0.49 0.8 (0.7.–0.9) 1.0 (0.9–1.2) 0.04
titer (95% CI)§
HAI titer ≥1:40 at <7 days of age 77 (81.1); 35 (34.0); <0.001 57 (60.0); 18 (17.5); <0.001 78 (82.1); 45 (43.7); <0.001
— no. (%); (95% CI)¶ (71.7–88.4) (24.9–44.0) (49.4–69.9) (10.7–26.2) (72.9–89.2) (33.9–53.8)
* HAI denotes hemagglutination inhibition.
† These P values are for the comparison of the IIV3 and placebo groups.
‡ These P values are for the comparison of values within a group before and after vaccination.
§ The ratio of the newborn HAI titer to the maternal HAI titer was based on 93 observations in the IIV3 group and 102 observations in the placebo group.
¶ P values were adjusted for the mean number of days from maternal vaccination to birth.
927
928
Table 4. Efficacy of IIV3 Vaccination in Mothers and Infants until 24 Weeks after Birth, Intention-to-Treat Population.*
Efficacy End Point HIV-Uninfected Cohort HIV-Infected Cohort
IIV3 Placebo Vaccine IIV3 Placebo Vaccine

(N = 1026) (N = 1023) Efficacy P Value (N = 100) (N = 88) Efficacy P Value
Infants
RT-PCR–confirmed influenza —
no. (%); (95% CI)

With inclusion of B/Yamagata 19 (1.9); 37 (3.6); 48.8 (11.6 to 70.4) 0.01 5 (5.0); 6 (6.8); 26.7 (−132.0 to 76.8) 0.60
(1.1 to 2.9)† (2.6 to 5.0)†‡ (1.6 to 11.3)§ (2.5 to 14.3)§
With exclusion of B/Yamagata 13 (1.3); 25 (2.4); 48.2 (−0.8 to 73.3) 0.05 4 (4.0); 3 (3.4); 14.8 (−270.4 to 80.4) 0.57
(0.7 to 2.2) (1.6 to 3.6) (1.1 to 9.9) (0.7 to 9.6)
Influenza-like illness — no. (%); 595 (58.0); 584 (57.1); −1.6 (−9.4 to 5.7) 0.68 66 (66.0); 57 (64.8); −1.9 (−25.5 to 17.3) 0.86
(95% CI) (54.9 to 61.0) (54.0 to 60.1) (55.8 to 75.2) (53.9 to 74.7)
Any respiratory illness — no. (%); 706 (68.8); 697 (68.1); −1.0 (−7.1 to 4.8) 0.74 76 (76.0); 70 (79.5); 4.5 (−11.3 to 18.0) 0.56
The
(95% CI) (65.9 to 71.6) (65.2 to 71.0) (66.4 to 84.0) (69.6 to 87.4)
IIV3 Placebo Vaccine IIV3 Placebo Vaccine
(N = 1062) (N = 1054) Efficacy P Value (N = 100) (N = 94) Efficacy P Value
Women
n engl j med
RT-PCR confirmed influenza —
no. (%)
52371;10
With inclusion of B/Yamagata 19 (1.8); 38 (3.6); 50.4 (14.5 to 71.2) 0.010 7 (7.0); 16 (17.0); 57.7 (0.2 to 82.1) 0.05

(1.1 to 2.8)¶ (2.6 to 4.9)¶ (2.9 to 13.9)‖ (10.1 to 26.2)‖
With exclusion of B/Yamagata 19 (1.8); 35 (3.3); 46.1 (6.4 to 69.0) 0.03 7 (7.0); 13 (13.8); 48.2 (−27.0 to 78.8) 0.15
nejm.org
(1.1 to 2.8) (2.3 to 4.6) (2.9 to 13.9) 7.6 to 22.5
n e w e ng l a n d j o u r na l
Influenza-like illness — no. (%) 175 (16.5); 181 (17.2); 4.0 (−16.0 to 20.6) 0.67 24 (24.0); 27 (28.7); −0.07 (−63.7 to 38.8) 0.99
of
(14.3 to 18.8) (14.9 to 19.6) (16.0 to 33.6) (19.9 to 39.0)

Any respiratory illness — no. (%) 674 (63.5); 687 (65.2); 2.6 (−3.7 to 8.6) 0.41 72 (72.0); 73 (77.7); 5.2 (−12.2 to 19.8) 0.54

(60.5 to 66.4) (62.2 to 68.1) (62.1 to 80.5) (67.9 to 85.6)
September 4, 2014
* Vaccine efficacy in HIV-infected women was adjusted for maternal age and status with respect to antiretroviral treatment at enrollment. RT-PCR denotes reverse-transcriptase–poly-
merase-chain-reaction.
m e dic i n e
† Among HIV-unexposed infants of IIV3 recipients and placebo recipients, there were 0 and 1 cases of infection with A/(H1N1)pdm09, 7 and 14 cases of infection with A/H3N2, 6 and
11 cases of infection with B/Victoria, and 6 and 12 cases of infection with B/Yamagata, respectively.
‡ One baby born to a placebo recipient had a single influenza episode during which both A/(H1N1)pdm09 and A/H3N2 were detected.
§ Among HIV-exposed infants of IIV3 recipients and placebo recipients, infection with A/(H1N1)pdm09 occurred in 2 and 0 infants, respectively, infection with A/H3N2 occurred in
2 infants in each group, infection with B/Victoria occurred in 0 and 1 infants, respectively, and infection with B/Yamagata occurred in 1 and 3 infants, respectively.
¶ Among HIV-uninfected IIV3 recipients and placebo recipients, there were 5 and 6 cases of influenza virus strain A/(H1N1)pdm09, 10 and 22 cases of A/H3N2, 4 and 7 cases of B/
Victoria, and 0 and 3 cases of B/Yamagata, respectively.
‖ Among HIV-infected IIV3 recipients and placebo recipients, there were 6 and 8 cases of influenza virus strain A/(H1N1)pdm09, 1 and 4 cases of A/H3N2, 0 and 1 case of B/Victoria,
and 0 and 3 cases of B/Yamagata, respectively.
Source: The New England Journal ofInfluenza
Medicine Vaccination of Pregnant Women
A HIV-Uninfected Cohort, Mothers B HIV-Uninfected Cohort, Infants

1.00 0.05 1.00 0.10
P=0.01 P=0.01
with Confirmed Influenza Placebo
with Confirmed Influenza

Proportion of Mothers
0.75 0.75
Proportion of Infants
0.03 0.05
Placebo
IIV3
0.50 0.50
IIV3
0.00 0.00
0.25 0 2 4 6 8 10 0.25 0 2 4 6
0.00 0.00
0 2 4 6 8 10 0 2 4 6
Months since Vaccination Age (mo)
No. at Risk No. at Risk
IIV3 1062 1047 1007 985 703 159 IIV3 1026 1004 981 897
Placebo 1054 1023 993 964 685 179 Placebo 1023 993 960 873
C HIV-Infected Cohort, Mothers D HIV-Infected Cohort, Infants

1.00 1.00 0.10
P=0.59
Placebo
P=0.03

Proportion of Mothers
0.75 0.75
Proportion of Infants
0.05 IIV3
0.50 0.50
0.00
0.25 0.25 0 2 4 6
Placebo
IIV3
0.00 0.00
0 2 4 6 8 10 0 2 4 6
Months since Vaccination Age (mo)
No. at Risk No. at Risk
IIV3 100 95 91 86 61 9 IIV3 100 93 86 68
Placebo 94 87 78 74 53 7 Placebo 88 81 75 66
Figure 1. Kaplan–Meier Estimates of Percentages of Confirmed Cases of Influenza According to Cohort and Study Group.
Confirmed influenza was defined as influenza diagnosed by means of reverse-transcriptase–polymerase-chain-reaction assay. The in-
sets show the same data on an expanded y axis. HIV denotes human immunodeficiency virus, and IIV3 trivalent inactivated influenza
vaccine.
rates between them and between these infants protocol analysis. Although 23% of the HIV-in-
and those in the Bangladeshi study, for whom fected and uninfected immunogenicity cohorts
exposure was probably more prolonged.16,39 An- were outside the protocol-defined windows for
other limitation of our study is the fact that the obtaining post-vaccination blood samples, the re-
evaluation of vaccine efficacy among HIV-infect- sults were similar to those included in the ex-
ed women was a secondary objective; neverthe- ploratory immunogenicity analysis with an ex-
less, its efficacy in these women was indicated panded window. An inadvertent occurrence in the
by the high attack rate in the placebo group. study was the difference in HIV disease charac-
However, the study lacked power to evaluate vac- teristics between the placebo group and the IIV3
cine efficacy among the infants of HIV-infected group, including the fact that placebo recipients
women, for whom the point estimate was simi- were more likely to be receiving highly active
lar to that for HIV-uninfected infants in the per- antiretroviral therapy. However, this factor would
n engl j med 371;10 nejm.org September

53 4, 2014 929
have biased the results in the direction of a re- the Department of Science and Technology and National Re-
search Foundation in Vaccine-Preventable Diseases, and the
duced risk of influenza in the placebo group.7 In Respiratory and Meningeal Pathogens Research Unit of the
addition, in HIV-infected women, the lower HAI Medical Research Council.
titers among placebo recipients as compared with Dr. Madhi reports receiving lecture fees and fees for serving
on advisory boards from GlaxoSmithKline, Pfizer, and Sanofi
IIV3 recipients could have been due to differences Pasteur and grant support from Novartis, GlaxoSmithKline,
in HIV disease stage between the groups. Pfizer, Sanofi Pasteur, and MedImmune; and Dr. Weinberg
In conclusion, these data show that IIV3 vac- consulting fees from Merck and grant support from Merck,
MedImmune, Becton Dickinson, Roche, and GlaxoSmithKline;
cination in pregnant HIV-uninfected and HIV- and Dr. Violari fees for serving on a data and safety monitoring
infected African women was immunogenic and board from Janssen and grant support from Janssen, Aeras,
provided protection against confirmed influenza. Gilead, and the Drugs for Neglected Diseases Initiative. No
other potential conflict of interest relevant to this article was
The vaccination was also effective in HIV-unex- reported.
posed infants up until 24 weeks after birth. Disclosure forms provided by the authors are available with
The contents of this report are solely the responsibility of the the full text of this article at NEJM.org.
authors and do not necessarily represent the official views of We thank all the study participants; the staff of the Depart-
their institutions or organizations or of the study sponsors. ments of Obstetrics, Neonatology, and Paediatrics at Chris
Supported by grants from the Bill and Melinda Gates Foun- Hani–Baragwanath Academic Hospital, Soweto, South Africa,
dation (OPP1002747), the National Institutes of Health, Na- for their dedication to their patients, including our trial partici-
tional Center for Advancing Translational Sciences Colorado pants; the study midwives, nurses, laboratory staff, counselors,
Clinical and Translational Sciences Institute (UL1 TR000154, and data capturers; and Niteen Wairagkar, program officer act-
for REDCap), the South African Research Chairs Initiative of ing on behalf of the Bill and Melinda Gates Foundation.
Appendix
From the Medical Research Council, Respiratory and Meningeal Pathogens Research Unit (S.A.M., C.L.C., L.K., A.H., S.J., P.V.A., N.N.,
K.P.K., M.C.N.), the Department of Science and Technology–National Research Foundation, Vaccine-Preventable Diseases (S.A.M.,
C.L.C., L.K., A.H., S.J., P.V.A., N.N., M.C.N.), and the Perinatal HIV Research Unit (A.V.), University of the Witwatersrand, the Na-
tional Institute for Communicable Diseases, the National Health Laboratory Service, Centre for Vaccines and Immunology (S.A.M., F.T.,
M.V.), Johannesburg, and the Department of Medical Virology, University of Pretoria, Pretoria (M.V.) — all in South Africa; the School
of Medicine and Children’s Hospital, University of Colorado (A.W.), the Department of Pediatrics, Medicine and Pathology, University
of Colorado School of Medicine (E.A.F.S.), and the Center for Global Health, Department of Epidemiology, Colorado School of Public
Health (E.A.F.S.) — all in Aurora, Colorado; the Department of Medicine and Department of Global Health, University of Washington
(J.R.O.), and the Vaccine Access and Delivery Global Program, PATH (J.R.O., K.M.N.) — both in Seattle; and the Hubert Department
of Global Health, Rollins School of Public Health, and the Division of Infectious Diseases, School of Medicine, Emory University, At-
lanta (K.P.K.).
References
1. Meeting of the Strategic Advisory 7. Neuzil KM, Coffey CS, Mitchel EF Jr, 12. Fuller JD, Craven DE, Steger KA, Cox
Group of Experts on Immunization, April Griffin MR. Cardiopulmonary hospital- N, Heeren TC, Chernoff D. Influenza vac-
2012 — conclusions and recommenda- izations during influenza season in adults cination of human immunodeficiency vi-
tions. Wkly Epidemiol Rec 2012;87:201- and adolescents with advanced HIV infec- rus (HIV)-infected adults: impact on
16. tion. J Acquir Immune Defic Syndr 2003; plasma levels of HIV type 1 RNA and de-
2. Neuzil KM, Reed GW, Mitchel EF, Si- 34:304-7. terminants of antibody response. Clin
monsen L, Griffin MR. Impact of influ- 8. Archer B, Cohen C, Naidoo D, et al. Infect Dis 1999;28:541-7.
enza on acute cardiopulmonary hospital- Interim report on pandemic H1N1 influ- 13. Benne CA, Kroon FP, Harmsen M, Ta-
izations in pregnant women. Am J enza virus infections in South Africa, vares L, Kraaijeveld CA, De Jong JC. Com-
Epidemiol 1998;148:1094-102. April to October 2009: epidemiology and parison of neutralizing and hemaggluti-
3. Louie JK, Acosta M, Jamieson DJ, et factors associated with fatal cases. Euro nation-inhibiting antibody responses to
al. Severe 2009 H1N1 influenza in preg- Surveill 2009;14:19369. influenza A virus vaccination of human
nant and postpartum women in Califor- 9. Staprans SI, Hamilton BL, Follansbee immunodeficiency virus-infected individ-
nia. N Engl J Med 2010;362:27-35. SE, et al. Activation of virus replication uals. Clin Diagn Lab Immunol 1998;5:
4. Jamieson DJ, Theiler RN, Rasmussen after vaccination of HIV-1-infected indi- 114-7.
SA. Emerging infections and pregnancy. viduals. J Exp Med 1995;182:1727-37. 14. Kroon FP, van Dissel JT, de Jong JC,
Emerg Infect Dis 2006;12:1638-43. 10. Nelson KE, Clements ML, Miotti P, Zwinderman K, van Furth R. Antibody re-
5. Schlaudecker EP, McNeal MM, Dodd Cohn S, Polk BF. The influence of human sponse after influenza vaccination in
CN, Ranz JB, Steinhoff MC. Pregnancy immunodeficiency virus (HIV) infection HIV-infected individuals: a consecutive
modifies the antibody response to triva- on antibody responses to influenza vac- 3-year study. Vaccine 2000;18:3040-9.
lent influenza immunization. J Infect Dis cines. Ann Intern Med 1988;109:383-8. 15. Hulka JF. Effectiveness of polyvalent
2012;206:1670-3. 11. Miotti PG, Nelson KE, Dallabetta GA, influenza vaccine in pregnancy: report of
6. Cohen C, Simonsen L, Sample J, et al. Farzadegan H, Margolick J, Clements ML. a controlled study during an outbreak of
Influenza-related mortality among adults The influence of HIV infection on anti- Asian influenza. Obstet Gynecol 1964;23:
aged 25-54 years with AIDS in South Af- body responses to a two-dose regimen of 830-7.
rica and the United States of America. influenza vaccine. JAMA 1989;262:779- 16. Zaman K, Roy E, Arifeen SE, et al. Ef-
Clin Infect Dis 2012;55:996-1003. 83. fectiveness of maternal influenza immu-

54 September 4, 2014
Source: The New England Journal of Medicine Influenza Vaccination of Pregnant Women
nization in mothers and infants. N Engl J Africa, 2006-2010. J Infect Dis 2012;206: Hemagglutination inhibition antibody ti-
Med 2008;359:1555-64. Suppl 1:S14-S21. ters as a correlate of protection for inacti-
17. Thompson MG, Li DK, Shifflett P, et 25. Poehling KA, Edwards KM, Weinberg vated influenza vaccines in children. Pe-
al. Effectiveness of seasonal trivalent in- GA, et al. The underrecognized burden of diatr Infect Dis J 2011;30:1081-5.
fluenza vaccine for preventing influenza influenza in young children. N Engl J Med 33. Håberg SE, Trogstad L, Gunnes N, et
virus illness among pregnant women: 2006;355:31-40. al. Risk of fetal death after pandemic in-
a population-based case-control study 26. Englund JA, Walter E, Black S, et al. fluenza virus infection or vaccination.
during the 2010-2011 and 2011-2012 in- Safety and immunogenicity of trivalent N Engl J Med 2013;368:333-40.
fluenza seasons. Clin Infect Dis 2014;58: inactivated influenza vaccine in infants: a 34. Madhi SA, Maskew M, Koen A, et al.
449-57. randomized double-blind placebo-con- Trivalent inactivated influenza vaccine in
18. Ortiz JR, Englund JA, Neuzil KM. In- trolled study. Pediatr Infect Dis J 2010;29: African adults infected with human im-
fluenza vaccine for pregnant women in 105-10. munodeficient virus: double blind, ran-
resource-constrained countries: a review 27. Recommended viruses for influenza domized clinical trial of efficacy, immu-
of the evidence to inform policy deci- vaccines for use in the 2011 southern hemi- nogenicity, and safety. Clin Infect Dis
sions. Vaccine 2011;29:4439-52. sphere influenza season. Geneva: World 2011;52:128-37.
19. Vaccines against influenza WHO po- Health Organization, 2010 (http://www 35. Anema A, Mills E, Montaner J, Brown-
sition paper — November 2012. Wkly .who.int/inf luenza/vaccines/virus/ stein JS, Cooper C. Efficacy of influenza
Epidemiol Rec 2012;87:461-76. recommendations/201009_Recommenda- vaccination in HIV-positive patients:
20. Benowitz I, Esposito DB, Gracey KD, tion.pdf?ua=1). a systematic review and meta-analysis.
Shapiro ED, Vázquez M. Influenza vaccine 28. VAXIGRIP. Inactivated influenza vac- HIV Med 2008;9:57-61.
given to pregnant women reduces hospi- cine (Split Virion) product information. 36. Robinson DP, Klein SL. Pregnancy
talization due to influenza in their in- France: Sanofi Pasteur, 2011 (http://www and pregnancy-associated hormones alter
fants. Clin Infect Dis 2010;51:1355-61. .sanofipasteur.com/en). immune responses and disease pathogen-
21. Black SB, Shinefield HR, France EK, 29. Table for Grading the Severity of esis. Horm Behav 2012;62:263-71.
et al. Effectiveness of influenza vaccine Adult and Pediatric Adverse Events Ver- 37. Jamieson DJ, Sibailly TS, Sadek R, et
during pregnancy in preventing hospital- sion 1.0 — December 2004 (clarification al. HIV-1 viral load and other risk factors
izations and outpatient visits for respira- dated August 2009) (http://rsc.tech-res for mother-to-child transmission of HIV-
tory illness in pregnant women and their .com/safetyandpharmacovigilance). 1 in a breast-feeding population in Cote
infants. Am J Perinatol 2004;21:333-9. 30. Harris PA, Taylor R, Thielke R, Payne d’Ivoire. J Acquir Immune Defic Syndr
22. Eick AA, Uyeki TM, Klimov A, et al. J, Gonzalez N, Conde JG. Research elec- 2003;34:430-6.
Maternal influenza vaccination and effect tronic data capture (REDCap) — a meta- 38. Tasker SA, O’Brien WA, Treanor JJ, et
on influenza virus infection in young in- data-driven methodology and workflow al. Effects of influenza vaccination in
fants. Arch Pediatr Adolesc Med 2011;165: process for providing translational re- HIV-infected adults: a double-blind, pla-
104-11. search informatics support. J Biomed In- cebo-controlled trial. Vaccine 1998;16:
23. France EK, Smith-Ray R, McClure D, form 2009;42:377-81. 1039-42.
et al. Impact of maternal influenza vacci- 31. Hobson D, Curry RL, Beare AS, Ward- 39. Influenza surveillance report —
nation during pregnancy on the incidence Gardner A. The role of serum haemagglu- South Africa. National Institute for Com-
of acute respiratory illness visits among tination-inhibiting antibody in protection municable Diseases: a division of Nation-
infants. Arch Pediatr Adolesc Med 2006; against challenge infection with influen- al Health Laboratory Service, May 2014
160:1277-83. za A2 and B viruses. J Hyg (Lond) 1972;70: (http://www.nicd.ac.za/?page=seasonal_
24. Radin JM, Katz MA, Tempia S, et al. 767-77. influenza&id=72).
Influenza surveillance in 15 countries in 32. Black S, Nicolay U, Vesikari T, et al. Copyright © 2014 Massachusetts Medical Society.
posting presentations from medical meetings online

Online posting of an audio or video recording of an oral presentation at
a medical meeting, with selected slides from the presentation, is not considered
prior publication. Authors should feel free to call or send e-mail to the
Journal’s Editorial Offices if there are any questions about this policy.

55 September 4, 2014 931
The NEW ENGLA ND JOURNAL of MEDICINE
Perspective october 9, 2014
2009 H1N1 Influenza and Pregnancy — 5 Years Later

Sonja A. Rasmussen, M.D., and Denise J. Jamieson, M.D., M.P.H.
I n April 2009, a novel influenza A virus, now re-

ferred to as influenza A(H1N1)pdm09 virus (2009
H1N1), was identified in two children in California,
for Disease Control and Preven-
tion (CDC), even though preg-
nant women make up only about
1% of the population. Moreover,
and shortly thereafter, the second U.S. death assoc- the 2009 pandemic virus was
ated with 2009 H1N1 occurred in the United States.1 These se- also bad for babies: infants born
in a previously healthy pregnant vere outcomes among pregnant to women who had been severely
woman. The virus spread rapidly women prompted us to review ill with influenza complications
throughout the United States and lessons learned from the pan- had increased risk for adverse
the world, and on June 11, 2009, demic and ways of reducing in- outcomes such as preterm birth
the World Health Organization fluenza’s effects during pregnan- or small size for gestational age.2
raised the global pandemic alert cy in future influenza seasons. The 2009 H1N1 pandemic
to 6, its highest level. Five years Although data were available brought a change in our approach
have now passed since that pan- before the 2009 pandemic sug- to treating influenza in pregnan-
demic, and in that time, much gesting that pregnant women cy. Previously, pregnant women
has been learned about influen- were at increased risk for influ- with influenza had been treated
za’s effects on pregnant women enza-associated complications, the primarily if they had other high-
and infants. Nevertheless, cases pandemic provided solid data risk medical conditions or severe
of severe influenza illness, hos- on this vulnerability.2 Pregnant disease. During the 2009 pan-
pitalizations, and deaths among women with 2009 H1N1 influ- demic, however, the CDC recom-
young and middle-aged adults, enza were at substantially higher mended that empirical antiviral
including pregnant women, were risk for hospitalization than the therapy be initiated as soon as
reported during the 2013–2014 general population, and they ac- possible during the clinical course
influenza season, when 2009 counted for approximately 5% of if the patient was pregnant or
H1N1 was again the predomi- deaths from 2009 H1N1 influenza had recently delivered. This rep-
nant circulating influenza virus that were reported to the Centers resented a significant shift in
56
n engl j med 371;15 nejm.org october 9, 2014 1369
PE RThe
Source: S PENew England
C T IV E Journal of Medicine
Deep-Brain Stimulation
antiviral treatment guidance: it of adverse outcomes such as ing and after the pandemic, but
was recommended that pregnant small size for gestational age since the 2010–2011 season, cov-
women with suspected 2009 and preterm birth among infants erage has remained relatively
H1N1 influenza receive prompt whose mothers received influen- stagnant at less than 50%. Fewer
antiviral therapy, regardless of za vaccine during pregnancy.2 pregnant women follow the rec-
risk factors, severity of illness, We have also gained knowl- ommendation to receive influen-
history, or the results of diagnos- edge about the barriers to and za vaccine than follow other
tic testing.2 motivators for influenza vacci- health care recommendations; for
Before the pandemic, we had nation during pregnancy. We example, in keeping with CDC
little information on the benefits know that recommendations from guidelines, more than 85% of
of treating pregnant women with health care providers are a criti- pregnant women are screened for
an antiviral medication, since cal motivator for pregnant women group B streptococcus infection.5
pregnant women had been ex- to be vaccinated. We also know Some pregnant women, and even
cluded from clinical trials of that pregnant women are con- a few clinicians, remain con-
these medications. During the cerned about the safety of influ- cerned about vaccine safety, espe-
pandemic, we learned that treat- enza vaccine and are more likely cially regarding vaccination dur-
ing pregnant women with such to be vaccinated if they under- ing the first trimester, even
a medication makes a difference. stand the benefits, especially to though the evidence indicates
A recent systematic review and their infants. The CDC used that influenza vaccination is safe
meta-analysis examining the ef- knowledge about these barriers during pregnancy.2
fects of antiviral medications on and motivators and worked close- Similarly, not all pregnant
mortality due to 2009 H1N1 in- ly with key partners such as the women with suspected influenza
fluenza among hospitalized preg- American College of Obstetricians receive treatment with oseltami-
nant women revealed that women and Gynecologists in order to in- vir, despite the current guide-
who received a neuraminidase crease vaccine coverage. lines. We have four main recom-
inhibitor within the first 2 days Yet reports of pregnant women mendations regarding treatment
after the onset of symptoms were becoming severely ill from or dy- during pregnancy. First, since
about one fifth as likely to die as ing of 2009 H1N1 influenza dur- some women with signs and
women who were treated later or ing the 2013–2014 influenza sea- symptoms of influenza may not
not at all.3 In addition, although son were sadly reminiscent of the seek treatment promptly, health
treatment of influenza has been pandemic 5 years ago, when we care providers need to educate
shown to be most effective if ini- received many calls from clini- pregnant women about these
tiated in the first 48 hours after cians caring for critically ill or signs and symptoms and the need
symptom onset, observational dying pregnant women. Why are for early treatment.
studies have shown that starting pregnant women still getting so Second, treatment should be
treatment even after 48 hours sick and dying from influenza? based on clinical evaluation rather
has clinical benefit.2 With in- In part, the problem may arise than on diagnostic testing, given
creased use of antiviral medica- from misconceptions about influ- the limited sensitivity of rapid in-
tions, it became possible to col- enza or its prevention or treat- fluenza antigen tests and the
lect data regarding their safety ment that lead to missed oppor- time required for more definitive
during pregnancy, and the find- tunities for preventing influenza testing. The sensitivity of the
ings have been reassuring.4 illness or its complications dur- rapid tests is generally 40 to 70%
Since the 2009 pandemic, much ing pregnancy. of that of viral culture or reverse-
more information has also become Despite our knowledge about transcriptase–polymerase chain
available about the safety and the benefits and safety of influ- reaction, and a sensitivity range
benefits of influenza vaccination enza vaccine and recommenda- as wide as 10 to 80% has been
during pregnancy. Receiving an tions from the Advisory Commit- reported — which means that
influenza vaccine reduces the risk tee on Immunization Practices false negative results are common.
of influenza not only for the and the American College of Ob- Treatment should not be delayed
pregnant woman but also for her stetricians and Gynecologists, until the results of more defini-
infant during the first 6 months not all pregnant women get vac- tive testing become available.
of life. In addition, some studies cinated. Substantial increases in Third, pregnant women should
have shown reduced frequencies vaccination coverage occurred dur- be treated regardless of whether

57 october 9, 2014
PER S PE C T IV E Deep-Brain Stimulation
they have been vaccinated, be- some of them die. Research is From the Centers for Disease Control and
Prevention, Atlanta.
cause influenza vaccine is only needed to further elucidate the
about 60% effective. reasons why some pregnant 1. Uyeki TM. Preventing and controlling in-
fluenza with available interventions. N Engl J
Finally, although it is ideal to women are not vaccinated so Med 2014;370:789-91.
begin treatment with antiviral that the problems can be ad- 2. Rasmussen SA, Jamieson DJ. Influenza
medications less than 48 hours dressed. In addition, by imple- and pregnancy in the United States: before,
during, and after 2009 H1N1. Clin Obstet
after symptoms begin, there is menting the current antiviral Gynecol 2012;55:487-97.
still clinical benefit when treat- treatment recommendations, cli- 3. Muthuri SG, Venkatesan S, Myles PR, et al.
ment is begun later. nicians can prevent complications Effectiveness of neuraminidase inhibitors in
reducing mortality in patients admitted to
Five years ago, reports of se- in women with influenza. We hospital with influenza A H1N1pdm09 virus
vere illness and deaths of preg- need to ensure that the informa- infection: a meta-analysis of individual par-
nant women from H1N1 influenza tion about influenza and preg- ticipant data. Lancet Respir Med 2014;2:395-
404.
called attention to questions about nancy that has been gained in 4. Wollenhaupt M, Chandrasekaran A, Tomi-
influenza during pregnancy and the 5 years since the 2009 H1N1 anovic D. The safety of oseltamivir in preg-
led to a large number of publica- pandemic is translated into re- nancy: an updated review of post-marketing
data. Pharmacoepidemiol Drug Saf 2014.
tions and vastly improved knowl- ductions in the number of ill- 5. Prevention of perinatal group B strepto-
edge about influenza and its nesses, hospitalizations, and coccal disease — revised guidelines from
treatment and prevention in this deaths that occur in future in- CDC, 2010. MMWR Recomm Rep 2010;59:
1-36.
population. Unfortunately, preg- fluenza seasons.
nant women continue to become Disclosure forms provided by the authors DOI: 10.1056/NEJMp1403496
seriously ill with influenza, and are available with the full text of this article Copyright © 2014 Massachusetts Medical Society.
at NEJM.org.

nejm.org october 9, 2014 1371
CDC Advisers Recommend New Flu Vaccine for Adults with Egg Allergy
Physician’s First Watch, published August 24, 2013
NEWS SUMMARY
The CDC’s Advisory Committee on Immunization Practices has voted unanimously to recommend use of Flublok
seasonal influenza vaccine in adults aged 18 to 49 with egg allergies during the 2013–2014 flu season. The FDA
approved Flublok in January 2013.
The manufacturer does not use eggs or inactivated influenza viruses; rather, viral hemagglutinins are isolated
and copies of those proteins are made using cell culture. Flublok has a shorter shelf life than traditional vaccines,
expiring 16 weeks after production, so clinicians are advised to check the expiration date before use.
— Joe Elia
59
Do Patients Benefit from Healthcare Providers’ Influenza Vaccination?

A literature review showed a reduction in all-cause mortality and influenza-like illness among the
providers’ patients, but the quality of evidence was moderate to very low.
Ahmed F et al. Effect of influenza vaccination of healthcare personnel on morbidity and mortality among patients: Sys-
tematic review and grading of evidence. Clin Infect Dis 2014 Jan 1; 58:50. (http://dx.doi.org/10.1093/cid/cit580)
Griffin MR. Influenza vaccination of healthcare workers: Making the grade for action. Clin Infect Dis 2014 Jan 1; 58:58.
(http://dx.doi.org/10.1093/cid/cit590)
ARTICLE SUMMARY
Although influenza vaccination is recommended for healthcare providers (HCPs) in the U.S. and many other
countries, whether it benefits patients remains controversial. The debate stems in part from the quality of evi-
dence in previous studies.
To rectify this situation, researchers at the CDC conducted a review of randomized, controlled trials, cohort
studies, and case-control studies published between January 1948 and June 2012 that reported on the association
between influenza vaccination of HCPs and morbidity or mortality among patients of healthcare facilities. They
evaluated evidence quality using the Grading of Recommendations Assessment, Development and Evaluation sys-
tem, a methodology employed by more than 60 organizations worldwide. Out of 8790 articles identified by the lit-
erature search, only 8 (4 cluster-randomized trials and 4 observational studies) met the authors’ inclusion criteria.
In the cluster-randomized trials, HCP vaccination was associated with a 29% (95% confidence interval, 15%–41%)
reduction in all-cause mortality and a 42% (95% CI, 27%–54%) reduction in influenza-like illness among patients.
Results of the observational studies also showed HCP influenza vaccination to be associated with decreased risk
for influenza-like illness in patients. However, the quality of evidence from the randomized trials was only mod
erate for the effect of HCP vaccination on reducing mortality among patients and was low for reducing influenza-like
illness. In addition, the quality of evidence from the observational studies for mortality and influenza-like illness
was very low.
COMMENT
It is sobering that only 8 studies out of more than 8000 were of sufficient relevance and quality to be analyzed
and that the strength of the data was no better than moderate. However, as an editorialist points out, we should
not be slaves to statistical analysis. Given the known safety and efficacy of influenza vaccination in general, the
current study offers additional proof that we have reached the threshold of evidence regarding the benefit of
immunizing healthcare providers to protect susceptible patients.
— Neil M. Ampel, MD
University of Arizona College of Medicine, Tucson
NEJM Journal Watch Infectious Diseases, published January 23, 2014
60
DIAGNOSIS AND TREATMENT

Three FDA-approved antiviral medications are recommended for use in the U.S. in the 2015–2016 influenza season:
oral oseltamivir, inhaled zanamivir, and intravenous peramivir. These neuraminidase inhibitors are active against
both influenza A and B viruses. The adamantanes rimantadine and amantadine are active against influenza A, but
not influenza B, viruses. However, resistance to these medications is high, and their use is not recommended for
treatment or prophylaxis in this 2015–2016 flu season. How and when to prescribe approved treatments is an
ongoing area of debate.
The NEJM Perspective article “Preventing and Controlling Influenza with Available Interventions”, written in
the midst of the 2014–2015 influenza season, provides a well-rounded overview of enduring clinical issues in
treating patients with confirmed or suspected influenza. These issues include treatment of higher-risk patients
and patients with comorbidities; signs, symptoms, and disease course; diagnostic testing; efficacy of drugs and
timing of their administration; and measures to limit contagion. We summarize research on disease severity and
the challenges of diagnosis in the ED and provide further detail on the safety, effectiveness, and administration
of antivirals.
61
Perspective february 27, 2014
Preventing and Controlling Influenza with Available Interventions

Timothy M. Uyeki, M.D., M.P.H., M.P.P.
I nfluenza activity has been surging in the United

States, and there are reports of critical illness
and death in young and middle-aged adults. The
pitalization rates for seasonal in-
fluenza typically occur among
people 65 years of age or older,
followed by children younger than
predominant virus so far this season is influenza 5 years of age; influenza-attribut-
able mortality is highest among
A(H1N1)pdm09, the cause of the fected and unvaccinated persons those 65 years of age or older.1
2009 H1N1 pandemic. Despite or to waning immunity from During the 2009 H1N1 pandemic,
many challenges, there is much prior infection. mortality was high among hospi-
that the public, patients, the pub- Although previously healthy talized middle-aged adults2; since
lic health community, and clini- persons can have severe illness, influenza A(H1N1)pdm09 virus is
cians can do now to reduce influ- certain groups are at increased prevalent this season, young and
enza’s impact. risk for complications that can middle-aged adults, women who
The spread of influenza result in severe illness: children are pregnant or up to 2 weeks
A(H1N1)pdm09 virus suggests younger than 2 years of age, el- post partum, and people with
that despite its ongoing circulation derly people, pregnant women, underlying conditions, including
since 2009, population immunity people with certain chronic con- morbid obesity, may also be at
is not sufficiently high and many ditions (e.g., pulmonary, cardiac particularly high risk for severe
people remain susceptible. To [excluding hypertension], renal, complications.2
date, surveillance data provide no hepatic, metabolic, hematologic, Annual influenza vaccination
evidence of significant antigenic neurologic, or neuromuscular con- is recommended for everyone
drift in the circulating virus ditions, immunosuppression, or 6 months of age or older in the
strains, so susceptibility could be morbid obesity), nursing home United States.1 Prevention strat-
due to the presence of a substan- residents, American Indians, and egies for infants younger than
tial number of previously unin- Alaska Natives.1 The highest hos- 6 months of age include vaccinat-
62
n engl j med 370;9 nejm.org february 27, 2014 789
PER S PE C T IV E Preventing and Controlling Influenza
ing pregnant women and vaccinat- olitis, tracheitis, myositis, rhab- risk persons with influenza signs
ing all household members and domyolysis, encephalopathy, and and symptoms to seek care prompt-
caregivers (“cocooning”). The ef- encephalitis). Influenza virus in- ly for assessment and possible
fectiveness of influenza vaccine fection of the respiratory tract can early antiviral treatment. Early
varies depending on several fac- trigger cytokine dysregulation, re- oseltamivir treatment in young
tors, including the recipient’s age sulting in acute lung injury and children can reduce the risk of
and immune response and the fulminant progression to respira- otitis media, and oseltamivir can
match between circulating virus tory failure, acute respiratory dis- reduce the risk of moderate in-
strains and vaccine strains. In re- tress syndrome, septic shock, fluenza complications requiring
cent years, influenza vaccine has acute kidney injury, and multior- antibiotics. Observational studies
been moderately effective, though gan failure.2 suggest that for the greatest clin-
less so among elderly persons, and Diagnostic testing may inform ical benefit in hospitalized pa-
some vaccinated persons may still clinical decisions. Nasopharyngeal tients, neuraminidase inhibitor
develop influenza.1 To date, cir- swab or aspirate specimens col- treatment should be started early,
culating influenza A(H1N1)pdm09 lected within 3 to 4 days after but there may be some benefit of
virus strains are well matched by illness begins have the highest oral oseltamivir treatment begin-
the H1N1 strains in all available yield for detection of influenza ning more than 2 days after ill-
vaccines this season, and unvac- viruses, but nasal swab specimens ness onset.4 Empirical antiviral
cinated persons should be vacci- may also be acceptable. Antigen- treatment with oral or enteric
nated as soon as possible. detection tests (rapid influenza oseltamivir should be started as
After an incubation period of diagnostic tests [RIDTs] and im- soon as possible in hospitalized
1 to 3 days, most persons with munofluorescence assays) lack patients with suspected influenza,
symptomatic influenza virus in- sensitivity as compared with oth- without waiting for the results of
fection have uncomplicated illness, er assays. When there is influ- RT-PCR assays. For outpatients
with sudden onset of fever, cough, enza activity in the community, who are at a higher risk for compli-
headache, sore throat, rhinor- negative results of antigen test- cations from influenza, neuramin-
rhea, nasal congestion, and mus- ing (especially RIDTs) do not rule idase inhibitor treatment as soon
cle aches, which resolve over 3 to out influenza in a symptomatic as possible is also recommended
5 days, though cough and fatigue patient.2,3 Reverse-transcription– (www.cdc.gov/f lu/professionals/
may persist longer. Children with polymerase-chain-reaction (RT- antivirals/summary-clinicians.htm).
influenza may have diarrhea and PCR) assay is recommended for The standard duration of anti-
abdominal pain with respiratory influenza testing in hospitalized viral treatment is 5 days, with
symptoms; some adults with in- patients. For mechanically venti- longer treatment recommended
fluenza A(H1N1)pdm09 virus in- lated patients, testing an endo- for critically ill patients with re-
fection may also have diarrhea.2 tracheal aspirate or bronchoalve- spiratory failure and prolonged
Signs and symptoms can vary de- olar-lavage fluid by RT-PCR assay viral replication in the lower re-
pending on age and underlying can yield the diagnosis when in- spiratory tract.2
conditions (e.g., elderly people may fluenza virus is not detectable in Resistance of inf luenza
not always have a fever) and can- upper respiratory specimens.2 A(H1N1)pdm09 viruses to osel-
not be easily distinguished clini- Randomized, controlled trials tamivir has been documented.
cally from those of other respira- have established the efficacy of Surveillance indicates that the
tory viral infections. Influenza the neuraminidase inhibitors (oral prevalence of oseltamivir-resistant
complications include exacerbation oseltamivir, inhaled zanamivir) as influenza viruses is low to date,
of chronic conditions (e.g., asth- compared with placebo for early but monitoring is ongoing and
ma, chronic obstructive pulmo- treatment (beginning within 48 must continue. Clinicians should
nary disease, congestive cardiac hours after illness onset) of un- know that the condition of criti-
failure) and pneumonia, but other complicated influenza, showing cally ill patients with influenza
severe complications can occur that they reduce fever and illness may not improve with oseltamivir
(bacterial coinfection, myocardi- duration by approximately 1 day.3 treatment and can deteriorate, but
tis, pericarditis, croup, bronchi- Physicians should encourage high- such deterioration is usually at-

63 february 27, 2014
P E R S PE C T IV E Preventing and Controlling Influenza
tributable not to resistance but to tion, opportunistic infections, and sonnel, and screening of visitors
the natural history of acute lung poor outcomes.5 Glucocorticoid for illness.
injury and progression of pulmo- treatment can be continued for We need more effective influ-
nary damage or other complica- patients receiving long-term glu- enza vaccines and better therapies,
tions (e.g., septic shock, acute re- cocorticoid therapy for other especially for severely ill patients.
nal failure, or ventilator-associated conditions and can be adminis- However, there is much we can
pneumonia). Severely immunosup- tered for certain complications do with available interventions to
pressed patients, however, are at (refractory shock or suspected reduce the spread and burden of
high risk for emergence of osel- adrenal insufficiency). Children influenza this season — and it’s
tamivir resistance during or after younger than 18 years of age not too late to get vaccinated.
oseltamivir treatment due to pro- should not receive aspirin because The views expressed in this article are
longed viral replication.2 For se- of the risk of Reye’s syndrome. those of the author and do not necessarily
reflect the policies of the Centers for Dis-
verely ill patients with influenza Influenza viruses are usually ease Control and Prevention (CDC).
who have strongly suspected or spread from symptomatic persons Disclosure forms provided by the author
documented oseltamivir resistance to susceptible contacts through are available with the full text of this article
at NEJM.org.
or malabsorption, gastric stasis, or respiratory droplets. Persons are
gastrointestinal bleeding, intrave- most contagious during the first From the Influenza Division, National Center
nous zanamivir, an investigational 3 to 4 days of illness, and conta- for Immunization and Respiratory Diseases,
drug, can be considered; it is gion declines with fever resolution. CDC, Atlanta.
available through enrollment in a Infants, immunocompromised peo-

This article was published on January 22,
clinical trial (http://clinicaltrials ple, and critically ill patients may 2014, at NEJM.org.
.gov/show/NCT01231620) or a re- have prolonged viral replication.
quest to the manufacturer (www Among previously healthy persons 1. Prevention and control of seasonal influ-
.clinicalsupporthd.gsk.com). with influenza, staying at home enza with vaccines: recommendations of the
Advisory Committee on Immunization Prac-
For patients with community- away from contacts, covering tices — United States, 2013–2014. MMWR
acquired pneumonia (CAP) dur- coughs and sneezes, and wash- Recomm Rep 2013;62(RR-07):1-43.
ing influenza season, diagnoses ing hands may help to reduce 2. Writing Committee of the WHO Consul-
tation on Clinical Aspects of Pandemic
of both primary influenza viral spread. Infection prevention and (H1N1) 2009 Influenza. Clinical aspects of
pneumonitis and secondary bac- control measures are essential for pandemic 2009 influenza A (H1N1) virus in-
terial pneumonia should be con- preventing nosocomial influenza. fection. N Engl J Med 2010;362:1708-19. [Er-
ratum, N Engl J Med 2010;362:2039.]
sidered. Bacterial pathogens most Hospitalized patients with influen- 3. Harper SA, Bradley JS, Englund JA, et al.
commonly associated with influ- za should be isolated or, if neces- Seasonal influenza in adults and children —
enza include methicillin-suscep- sary, cohorted (grouped together diagnosis, treatment, chemoprophylaxis, and
institutional outbreak management: clinical
tible and methicillin-resistant in the same room), and standard practice guidelines of the Infectious Diseases
Staphylococcus aureus, Streptococcus and droplet precautions should Society of America. Clin Infect Dis 2009;48:
pneumoniae, and Streptococcus py- be implemented (www.cdc.gov/ 1003-32.
4. Muthuri SG, Myles PR, Venkatesan S,
ogenes.2,3 Empirical antimicrobial flu/professionals/infectioncontrol/ Leonardi-Bee J, Nguyen-Van-Tam JS. Impact
therapy should cover these patho- healthcaresettings.htm). For med- of neuraminidase inhibitor treatment on out-
gens, and guidelines for CAP ical procedures during which a comes of public health importance during
the 2009-2010 influenza A(H1N1) pandemic:
should be followed, with appro- patient may generate an aerosol, a systematic review and meta-analysis in hos-
priate deescalation of antibiotic a fit-tested N95 respirator is in- pitalized patients. J Infect Dis 2013;207:553-
treatment based on microbiolog- dicated for health care person- 63.
5. Hui DS, Lee N, Chan PK. Adjunctive ther-
ic testing results. Systemic gluco- nel. Influenza prevention should apies and immunomodulatory agents in the
corticoid treatment should be be strengthened throughout the management of severe influenza. Antiviral
avoided in patients with severe or hospital, with active surveillance Res 2013;98:410-6.
progressive illness because of the for nosocomial influenza, re- DOI: 10.1056/NEJMp1400034
risk of prolonged viral replica- striction of sick health care per- Copyright © 2014 Massachusetts Medical Society.

64 february 27, 2014 791
Most Influenza Infections Are Subclinical

During six influenza seasons in England, only 25% of individuals with serologically confirmed influenza
had polymerase chain reaction–confirmed influenza, and only 17% of the latter group sought medical
attention.
Hayward AC et al. Comparative community burden and severity of seasonal and pandemic influenza: Results of the
Flu Watch cohort study. Lancet Respir Med 2014 Mar 17; [e-pub ahead of print]. (http://dx.doi.org/10.1016/
S2213-2600(14)70034-7)
ARTICLE SUMMARY
Most viral infections — including influenza — cause symptoms in only a fraction of infected individuals, but
those with subclinical infection might also transmit the disease. Knowing the proportion of such individuals
in a cohort is essential for epidemiologic predictions.
To estimate the burden and severity of influenza during recent years in England, researchers conducted a
household-level community cohort study of acute respiratory illnesses during the 2006–2007, 2007–2008, and
2008–2009 periods of seasonal influenza circulation and the first (spring and summer, 2009), second (autumn
and winter, 2009), and third (winter, 2010–2011) waves of the 2009 H1N1 influenza pandemic. Participants reported
their health status weekly and if they experienced respiratory illness, completed symptom diaries and provided
nasal swabs for viral isolation. In addition, serum samples taken before and after each influenza season were
tested for antibodies against circulating influenza strains.
Overall, the study involved 5448 person-seasons. Influenza infection occurred on average in 18% of the unvac
cinated population each season, based on serologic testing. The rate of respiratory illnesses was 69 per 100 person-
seasons among those with serologically confirmed influenza and 44 per 100 person-seasons in individuals with-
out serologic evidence of influenza, for an age-adjusted influenza-attributable respiratory infection rate of 23 per
100 person-seasons.
Only 25% of individuals with serologically confirmed influenza also had polymerase chain reaction (PCR)-
positive nasal swabs. And only 46% of those with PCR-confirmed influenza met the CDC definition of influenza-
like illness. Symptoms of 2009 pandemic H1N1 influenza were milder than symptoms of seasonal H3N2 influenza,
but more severe than those associated with noninfluenza respiratory viruses. Seventeen percent of the individuals
with PCR-confirmed influenza (and 21% those with infections caused by other respiratory viruses) consulted a
physician; the hospital admission rate for influenza was <1%.
COMMENT
Reported influenza cases likely represent just the tip of a large clinical and subclinical iceberg that is mainly
invisible to routine, primary care–based surveillance systems. It appears that the extent of influenza infection
and illness in the community has been greatly underestimated so far.
— Thomas Glück, MD
University of Regensburg, Bavaria, Germany
NEJM Journal Watch Infectious Diseases, published March 20, 2014
65
Emergency Department Clinicians Struggle to Diagnose Influenza

on Clinical Grounds Alone
With PCR as the gold standard, clinical diagnosis had a sensitivity of only 36%.
Dugas AF et al. Clinical diagnosis of influenza in the ED. Am J Emerg Med 2015 Jun; 33:770. (http://dx.doi.org/
10.1016/j.ajem.2015.03.008)
ARTICLE SUMMARY
The CDC recommends antiviral treatment for patients with influenza who have a severe or complicated course
of illness, are at high risk for complications, or require hospitalization. In a prospective observational cohort study
conducted at a single U.S. emergency department (ED), investigators assessed clinicians’ ability to diagnose influ-
enza based on history and physical exam alone, with blinded polymerase chain reaction (PCR) results as the gold
standard. All participants met CDC criteria for antiviral treatment if diagnosed with influenza.
Of the 270 patients, 16% had influenza by PCR. Clinical diagnosis of influenza had a sensitivity of 36% and a spec-
ificity of 78%. Only 36% of the patients with PCR-confirmed influenza received antiviral medication, whereas 11%
of those without influenza received it.
COMMENT
This study adds to a body of literature showing that history and physical exam are not accurate for influenza
treatment decisions. We should advocate for rapid PCR assays in our EDs. More controversially, I would suggest
that we alter our thinking and first consider whether a patient with a respiratory illness would be eligible for treat-
ment if she had influenza. If so, either test or treat for influenza; if not, don’t worry about the etiology of the respi-
ratory illness, and treat symptomatically as required. Even if results were delayed, we could call the patient back as
we do for other infectious diseases, such as chlamydia and gonorrhea. However, bear in mind that relatively few
patients with influenza are actually eligible (generally patients with comorbidities) for treatment.
— Daniel J. Pallin, MD, MPH

Harvard Medical School, Boston
NEJM Journal Watch Emergency Medicine, published June 10, 2015
66
Influenza Is More Virulent Among Cancer Patients

Yet many of the most common flu symptoms are less common in this group.
Memoli MJ et al. The natural history of influenza infection in the severely immunocompromised vs nonimmunocompro-
mised hosts. Clin Infect Dis 2014 Jan 15; 58:214. (http://dx.doi.org/10.1093/cid/cit725)
ARTICLE SUMMARY
Influenza is particularly virulent among immunocompromised people, but studies have been largely retrospective,
without point-by-point comparisons with immunologically normal people during the same flu season. Now, NIH
investigators have prospectively evaluated 32 immunocompromised patients (25 with hematologic malignancies,
7 with solid tumors) who presented with culture-confirmed influenza between 2008 and 2011. Compared with
54 nonimmunocompromised influenza patients (a group that included patients with well-controlled HIV infec-
tion) treated during the same years, immunocompromised patients were less likely to be overweight and to be
current smokers and also less likely to be vaccinated (25% vs. 60%).
Fever was reported by about 80% of both groups, but most other flu symptoms were significantly less common
among the immunocompromised, including dry cough (78% vs. 96%), chills (52% vs. 81%), shortness of breath
(37% vs. 62%), and myalgias (40% vs. 68%). Conversely, radiologic abnormalities were more common among
immunocompromised patients, more of whom also required intensive care unit admission, and the single fatality
in the study was an immunocompromised patient. Average durations of viral shedding were significantly longer
in the immunocompromised group.
COMMENT
No surprises here regarding the increased virulence of the flu and the higher likelihood of pneumonia in severely
immunocompromised people. The relative dearth of common symptoms in this group is interesting and suggests
that clinicians should consider flu in these patients even in the absence of many of the usual clinical indicators.
— Abigail Zuger, MD
Columbia University College of Physicians and Surgeons, New York
NEJM Journal Watch Infectious Diseases, published February 4, 2014
67
New Meta-Analysis Clarifies Oseltamivir’s Benefits and Risks

A patient-level meta-analysis confirms this drug’s modest antiviral activity and substantial side effects.
Dobson J et al. Oseltamivir treatment for influenza in adults: A meta-analysis of randomised controlled trials.
Lancet 2015 Jan30; [e-pub ahead of print]. (http://dx.doi.org/10.1016/S0140-6736(14)62449-1)
ARTICLE SUMMARY
In a 2014 Cochrane meta-analysis of eight manufacturer-conducted, placebo-controlled, influenza treatment
trials in adults, researchers concluded that oseltamivir (Tamiflu) “modestly reduces” duration of symptoms but
causes nausea and vomiting and, possibly, adverse psychiatric effects. The report pointedly questioned the anti
viral activity of oseltamivir and “its use in clinical practice as an anti-influenza drug.” However, the analysis was
based on aggregate rather than patient-level data.
Researchers now report a meta-analysis of patient-level data from the same trials plus one recent addition (total
of 4328 adult patients with flu-like illness). In all trials, patients with fever and at least two influenza symptoms
of shorter than 36 hours’ duration were treated with oseltamivir (75 mg) or placebo twice daily for 5 days. Roche
provided all data and unrestricted funding for the analysis.
Among 2893 patients with confirmed influenza, median time to alleviation of all symptoms was significantly
shorter in those who received oseltamivir than in those who received placebo (98 vs. 123 hours), and oseltami-
vir patients experienced significantly fewer lower respiratory complications and hospitalizations (1 fewer per
100 treated patients). Among all patients with flu-like illness, the difference between groups in time to symptom
alleviation was attenuated but still statistically significant. Among patients who were found to be uninfected with
influenza, outcomes were similar for the oseltamivir and placebo groups. Patients taking oseltamivir experienced
significantly more nausea and vomiting (5 more patients with vomiting per 100 treated patients) but no excess
psychiatric symptoms.
COMMENT
This analysis confirms oseltamivir’s antiviral activity, its modest effect on duration of influenza symptoms, and
its substantial side effects. Ideally, treatment with oseltamivir should be limited to patients with confirmed
influenza, with careful evaluation of potential benefits and harms.
— Bruce Soloway, MD
Albert Einstein College of Medicine and Montefiore Medical Center, Bronx, New York
NEJM Journal Watch General Medicine, published February 5, 2015
68
Oseltamivir for Influenza, Even After 48 Hours of Symptoms?

The drug reduced symptom duration and virus shedding among patients in Bangladesh who started
treatment up to 5 days after symptom onset.
Fry AM et al. Efficacy of oseltamivir treatment started within 5 days of symptom onset to reduce influenza illness dura-
tion and virus shedding in an urban setting in Bangladesh: A randomised placebo-controlled trial. Lancet Infect Dis
2013 Nov 22; [e-pub ahead of print]. (http://dx.doi.org/10.1016/S1473-3099(13)70267-6)
Ison MG. Optimum timing of oseltamivir: Lessons from Bangladesh. Lancet Infect Dis 2013 Nov 22; [e-pub ahead of
print]. (http://dx.doi.org/10.1016/S1473-3099(13)70287-1)
ARTICLE SUMMARY
Oseltamivir has been shown to reduce the duration of influenza illness and to limit virus shedding when started
within 48 hours after symptom onset. Now, researchers have tested the drug in a trial involving 1163 urban resi-
dents in Bangladesh (minimum age, 1 year; 51% <5 years old) who had influenza symptoms lasting ≤5 days and
who had a confirmed positive finding on a rapid influenza test of nasal wash specimens. Participants were ran-
domly assigned to receive oseltamivir (donated by the manufacturer) or placebo twice daily for 5 days. Random-
ization was separate for participants with symptoms <48 hours (67% of the cohort) and those with symptoms
≥48 hours (33%). Specimens were taken at baseline and on days 2, 4, and 7.
Overall, the median duration of post-enrollment major symptoms was significantly shorter in the oseltamivir
group than in the placebo group (3 vs. 4 days). Among participants with pre-enrollment symptoms of ≥48 hours,
oseltamivir significantly outperformed placebo in reducing virus isolation on days 2 and 4, but not day 7. All 124
seasonal influenza A H1N1 isolates that had been tested before treatment were resistant to oseltamivir. (Seasonal
influenza A H1N1 was found in 11% of all specimens.) Resistance during drug treatment emerged in <1% of osel-
tamivir recipients. No severe adverse events were associated with oseltamivir, but vomiting was significantly more
common in the treatment group than the placebo group (5.9% vs. 3.2%).
COMMENT
This study documents a modest benefit of oseltamivir, even when started ≥48 hours after symptom onset. Few
studies have assessed antiviral treatment of influenza in low-income populations, such as this one in Bangladesh,
where an editorialist notes that malnutrition and other comorbidities might contribute to more-severe disease.
The study’s authors caution that use of a rapid diagnostic test to determine eligibility might have selected for pa-
tients with high viral loads.
— Mary Wilson, MD
Harvard T.H. Chan School of Public Health, Boston
NEJM Journal Watch Infectious Diseases, published December 13, 2013
69
Effect of Oseltamivir on Viral Shedding, Illness, and Household Transmission

of Influenza
Oseltamivir reduced the duration of symptoms but not the duration of viral shedding or the risk for
household transmission of influenza.
Cheung DH et al. Association of oseltamivir treatment with virus shedding, illness, and household transmission of
influenza viruses. J Infect Dis 2015 Aug 1; 212:391. (http://dx.doi.org/10.1093/infdis/jiv058)
ARTICLE SUMMARY
Oseltamivir has become a mainstay in influenza treatment and prevention. To better characterize its effects when
prescribed to patients in community settings, investigators conducted an observational study that evaluated not
only clinical response to therapy but also the drug’s effects on duration of viral shedding and risk for household
transmission. Enrollees, recruited from outpatient clinics in Hong Kong, had real-time reverse-transcriptase
polymerase chain reaction (PCR)–confirmed influenza virus infection and resided in households with ≥2 other
members, none of whom reported acute illness in the preceding 14 days. Respiratory symptoms, total symptoms,
and time to symptom alleviation were assessed using a scoring system, and secondary cases were defined by PCR
positivity.
A total of 582 index patients were included, of whom 223 (38%) received oseltamivir. Initiation of oseltamivir
≤24 hours after symptom onset — compared with no antiviral treatment — was associated with a 56% reduction
in time to cessation of all influenza symptoms, a 47% reduction in duration of fever, and a 56% reduction in time
to resolution of respiratory symptoms (acceleration factor, 0.44; 95% CI, 0.33–0.59). Duration of viral shedding
was similar between groups. Analysis of household transmission involving 1420 contacts, 139 (10%) of whom
developed influenza, showed no significant association between oseltamivir treatment and risk for transmission.
COMMENT
Unfortunately, oseltamivir did not reduce the rate of secondary cases among household contacts, as one would
have hoped. Although this shortcoming could be related to the lack of effect on viral shedding, the data need to
be examined in the context of multiple earlier studies that, based on viral culture, have shown decreased duration
of viral shedding with oseltamivir.
— Larry M. Baddour, MD
Mayo Clinic College of Medicine, Rochester, Minnesota
NEJM Journal Watch Infectious Diseases, published August 12, 2015
70
FDA Approves the Intravenous Influenza Drug Peramivir

Peramivir is comparable to but much more expensive than other antivirals for uncomplicated influenza;
its role in treating hospitalized patients is unclear.
FDA approves Rapivab to treat flu infection [press release]. Silver Spring, MD: U.S. Food and Drug Administration;
December22, 2014. (http://www.fda.gov/NewsEvents/Newsroom/PressAnnouncements/ucm427755.htm)
Rapivab [prescribing information]. Durham, NC: BioCryst Pharmaceuticals; 2014. (http://rapivab.com/wp-content/

uploads/2014/12/Rapivab-PI.pdf)
ARTICLE SUMMARY
On December 22, 2014, the FDA announced the approval of peramivir (Rapivab) for treatment of uncomplicated
influenza in adults who have been symptomatic for ≤2 days. The drug is the third neuraminidase inhibitor (NAI)
to be approved for this indication but the first for intravenous use. It is administered as a single 600-mg infusion,
with dose reductions in patients with renal dysfunction.
In a randomized, controlled trial involving 297 adults with uncomplicated influenza in Japan, a single 600-mg
intravenous dose of peramivir decreased the time to symptom alleviation by 21 hours compared with placebo —
comparable to what has been seen with oral oseltamivir and inhaled zanamivir. In a phase III study involving
patients with uncomplicated influenza in South Korea, Japan, and Taiwan, peramivir proved noninferior to
5 days of oral oseltamivir.
Efficacy of peramivir could not be established, however, in patients with serious influenza requiring hospitaliza-
tion. In a randomized, controlled, multicenter trial involving 398 such patients, 600 mg of peramivir administered
daily for 5 days did not improve time to clinical resolution of symptoms compared with placebo. Patients in both
study arms also received standard of care, which could include an oral or inhaled NAI.
COMMENT
Unfortunately, the price of a single dose of peramivir is currently $1000 or more, depending on the supplier. Given
that the drug costs much more than oral oseltamivir but has similar efficacy, its role in uncomplicated influenza
will likely be very limited; for serious influenza, efficacy has not been shown. These concerns notwithstanding,
peramivir might find use in patients for whom there are strong concerns about absorption of oral oseltamivir.
Questions remain about duration of therapy in this scenario.
— Lynn L. Estes, PharmD
Mayo Clinic, Rochester, Minnesota
NEJM Journal Watch Infectious Diseases, published January 20, 2015
71
CDC Recommends Antivirals in High-Risk Flu Patients,

Even Before Test Results in
Physician’s First Watch, published January 12, 2015
NEWS SUMMARY
High-risk patients suspected of having influenza should be given neuraminidase inhibitors before confirmatory
test results are in, the CDC reiterated to healthcare providers. This includes patients who are hospitalized; have
severe, complicated, or progressive illness; or have other high-risk factors, like being younger than 2 years or older
than 64, living in a chronic care facility, or being pregnant.
The agency said that the antivirals oseltamivir (Tamiflu), zanamivir (Relenza), and peramivir (Rapivab) are being
underused. Regional shortages of antivirals have been reported, but there is no overall shortage.
Officials made these remarks in response to the heavy 2014–2015 influenza season, when vaccine was not well
matched to the predominant H3N2 strain.
— Kelly Young
72
AVIAN INFLUENZA
Avian influenza, the influenza A H5N1 and H7N9 viruses, occasionally jump from birds to humans. The World
Health Organization records 844 confirmed cases of H5N1 in humans and 449 related deaths reported to WHO
in the period 2003–2015. Human disease due to avian influenza A H7N9 virus has emerged in Southeast Asia. A
total of 571 laboratory-confirmed cases of human infection with H7N9 virus, including 212 deaths, have been
reported to WHO.
This section opens with an epidemiologic study of H7N9 cases identified in humans in China as of December 1,
2013. We also include a news report of CDC recommendations for protecting humans during outbreaks and
report encouraging news about possibly effective vaccines. Other research summaries provide insight into the risk
for transmission of avian viruses in the U.S., as well as identification of the types of birds that could carry these
viruses.
73
original article
Epidemiology of Human Infections with

Avian Influenza A(H7N9) Virus in China
Qun Li, M.D., Lei Zhou, M.D., Minghao Zhou, Ph.D., Zhiping Chen, M.D., Furong Li, M.D.,
Huanyu Wu, M.D., Nijuan Xiang, M.D., Enfu Chen, M.P.H., Fenyang Tang, M.D.,
Dayan Wang, M.D., Ling Meng, M.D., Zhiheng Hong, M.D., Wenxiao Tu, M.D.,
Yang Cao, M.D., Leilei Li, Ph.D., Fan Ding, M.D., Bo Liu, M.D., Mei Wang, M.D.,
Rongheng Xie, M.D., Rongbao Gao, M.D., Xiaodan Li, M.D., Tian Bai, M.D.,
Shumei Zou, M.D., Jun He, M.D., Jiayu Hu, M.D., Yangting Xu, M.D.,
Chengliang Chai, M.D., Shiwen Wang, M.D., Yongjun Gao, M.D., Lianmei Jin, M.D.,
Yanping Zhang, M.D., Huiming Luo, M.D., Hongjie Yu, M.D., M.P.H.,
Jianfeng He, M.D., Qi Li, M.D., Xianjun Wang, M.D., Lidong Gao, M.D.,
Xinghuo Pang, M.D., Guohua Liu, M.D., Yansheng Yan, M.D., Hui Yuan, M.D.,
As of February 2015, the WHO reports Yuelong Shu, Ph.D., Weizhong Yang, M.D., Yu Wang, M.D., Fan Wu, M.D.,
571 confirmed cases of influenza A(H7N9) Timothy M. Uyeki, M.D., M.P.H., M.P.P., and Zijian Feng, M.D., M.P.H.
virus infection in humans (see WHO report). A bs t r ac t
Background
The authors’ affiliations are listed in the The first identified cases of avian influenza A(H7N9) virus infection in humans oc-
Appendix. Address reprint requests to curred in China during February and March 2013. We analyzed data obtained from
Dr. Feng at the Public Health Emergency
Center, Chinese Center for Disease Con- field investigations to describe the epidemiologic characteristics of H7N9 cases in
trol and Prevention, Beijing, 102206, China identified as of December 1, 2013.
China, or at fengzj@chinacdc.cn; or to Methods
Dr. Fan Wu at Shanghai Municipal Center
for Disease Control and Prevention, 1380 Field investigations were conducted for each confirmed case of H7N9 virus infec-
Zhongshan W. Rd., Shanghai, 200336, tion. A patient was considered to have a confirmed case if the presence of the H7N9
China, or at fwu@scdc.sh.cn.
virus was verified by means of real-time reverse-transcriptase–polymerase-chain-
Drs. Qun Li, L. Zhou, M. Zhou, Z. Chen, reaction assay (RT-PCR), viral isolation, or serologic testing. Information on demo-
and F. Li contributed equally to this article. graphic characteristics, exposure history, and illness timelines was obtained from pa-
A preliminary version of this article was
tients with confirmed cases. Close contacts were monitored for 7 days for symptoms
published on April 24, 2013, at NEJM.org. of illness. Throat swabs were obtained from contacts in whom symptoms developed
and were tested for the presence of the H7N9 virus by means of real-time RT-PCR.
N Engl J Med 2014;370:520-32.
DOI: 10.1056/NEJMoa1304617
Results
Copyright © 2013 Massachusetts Medical Society. Among 139 persons with confirmed H7N9 virus infection, the median age was 61 years
(range, 2 to 91), 71% were male, and 73% were urban residents. Confirmed cases oc-
Related Material:
curred in 12 areas of China. Nine persons were poultry workers, and of 131 persons with
Supplementary Appendix
available data, 82% had a history of exposure to live animals, including chickens (82%).
Correspondence A total of 137 persons (99%) were hospitalized, 125 (90%) had pneumonia or respira-
tory failure, and 65 of 103 with available data (63%) were admitted to an intensive care
unit. A total of 47 persons (34%) died in the hospital after a median duration of illness
of 21 days, 88 were discharged from the hospital, and 2 remain hospitalized in critical
condition; 2 patients were not admitted to a hospital. In four family clusters, human-
to-human transmission of H7N9 virus could not be ruled out. Excluding secondary
cases in clusters, 2675 close contacts of case patients completed the monitoring period;
respiratory symptoms developed in 28 of them (1%); all tested negative for H7N9 virus.
Conclusions
Most persons with confirmed H7N9 virus infection had severe lower respiratory
tract illness, were epidemiologically unrelated, and had a history of recent exposure
to poultry. However, limited, nonsustained human-to-human H7N9 virus transmis-
sion could not be ruled out in four families.
520 74 nejm.org
n engl j med 370;6 february 6, 2014
Source: The New England Journal ofEpidemiology
Medicine of avian Influenza (H7N9) Outbreak in China
T
he first identified cases of human investigations and obtained respiratory specimens,
infection with a novel influenza A(H7N9) which were shipped to the National Influenza
virus occurred in eastern China during Feb- Center of the China CDC in Beijing for H7N9 labo-
ruary and March 2013 and were characterized by ratory testing. The China CDC was involved in field
rapidly progressive pneumonia, respiratory failure, investigations of the first 82 confirmed cases of
the acute respiratory distress syndrome (ARDS), H7N9 virus infection; local CDCs conducted field
and fatal outcomes.1 We analyzed available data investigations of subsequent confirmed cases.
from field investigations to characterize the de-
scriptive epidemiology of laboratory-confirmed Data Collection
cases of avian influenza A(H7N9) virus infection The National Health and Family Planning Com-
in humans that were reported to the Chinese Cen- mission (NHFPC) determined that the collection
ter for Disease Control and Prevention (China of data from H7N9 case patients and their close
CDC) as of December 1, 2013. In this report, we contacts was part of a continuing public health
summarize the epidemiologic findings of case in- investigation of an outbreak and was exempt
vestigations and follow-up monitoring of close from institutional review board assessment. Data
contacts of persons with confirmed cases of H7N9 were collected through a review of medical rec-
virus infection who have been identified to date. ords and interviews with relatives, contacts, and
Investigations are ongoing. health care workers who provided medical care
for the case patients. We collected information on
Me thods the dates of illness onset, visits to clinical facili-
ties, hospitalization, and clinical outcomes. Epi-
Case Definitions demiologic data were collected through interviews
The case definitions of suspected and confirmed and field observations and were reported to the
human infection with H7N9 virus were based on China CDC. Investigators interviewed the relatives
the H5N1 case definitions, as recommended by of each patient with a confirmed case of H7N9
the World Health Organization (WHO) in 2006 virus infection to determine exposure histories
(Section S1 in the Supplementary Appendix, during the 2 weeks before the onset of the ill-
available with the full text of this article at NEJM ness, including the dates, times, frequency, and
.org).2 A patient was considered to have a confirmed patterns of exposures to poultry or other animals
case if the presence of the H7N9 virus was veri- such as swine and wild birds. All epidemiologic
fied by means of real-time reverse-transcriptase– information that was collected during the field
polymerase-chain-reaction assay (RT-PCR), viral investigations, including exposure history, time-
isolation, or serologic testing. The laboratory test lines of events, and identification of close con-
assays for H7N9 virus that we performed have tacts, was cross-validated, since we were unable
been described previously (Section S2 in the Sup- to interview any critically ill H7N9 case patients.
plementary Appendix).1 Households and places known to have been vis-
ited by the case patients in the 2 weeks before the
Identification of Cases onset of illness were investigated to assess expo-
Suspected cases of H7N9 virus infection among sures to poultry and swine, as well as environ-
hospitalized patients with pneumonia were iden- mental exposures. We also collected information
tified through the Chinese surveillance system on control measures implemented by national,
for pneumonia of unknown origin, which was provincial, and municipal governments.
established in 2004.3,4 Beginning on April 3, 2013,
enhanced surveillance was implemented for sus- Identification and Follow-up of Close
pected cases of H7N9 virus infection among per- Contacts
sons with mild or moderate illness.5 Persons with We defined close contacts of patients with con-
suspected cases of H7N9 virus infection with mild firmed H7N9 virus infection as described previ-
or moderate illness were identified from the Chi- ously for H5N1 field investigations7 (Section S3
nese sentinel surveillance system for influenza-like in the Supplementary Appendix); once we identi-
illness, which has been described previously.6 Once fied the close contacts, we monitored them daily
each suspected case of H7N9 virus infection was for 7 days for symptoms of illness and collected
identified, the local CDCs, including prefecture throat swabs from contacts in whom symptoms
and provincial CDCs, conducted the initial field developed to test for the presence of H7N9 virus.

75 february 6, 2014 521
Antiviral chemoprophylaxis was not provided to firmed cases (71%) occurred in males, 73% of
close contacts. Paired serum samples were ob- the case patients were urban residents, and 79 of
tained for H7N9 serologic testing from patients 108 patients with available data (73%) had under-
with suspected H7N9 virus infection who did not lying medical conditions (Table 1). Among 70 of
have respiratory specimens available. Oseltamivir 79 case patients with sufficient data for a more
treatment was recommended for close contacts in specific classification of underlying conditions,
whom symptoms developed (Section 3 in the Sup- 54 (77%) were considered to be at increased risk
plementary Appendix). Data on demographic char- for influenza complications owing to age (<5 years
acteristics and exposure were collected for close or ≥65 years) or prevalence of certain underlying
contacts. medical conditions.9 Nine of the patients with
confirmed cases (6%) worked as poultry work-
Statistical Analysis ers: 6 slaughtered poultry at a live poultry mar-
We used descriptive statistics to summarize the ket, 2 transported live poultry, and 1 raised and
epidemiologic characteristics and H7N9 testing traded pigeons.
results for persons with suspected cases of H7N9 A total of 137 of 139 patients with confirmed
virus infection, for those with confirmed cases, H7N9 virus infection (99%) were hospitalized
and for close contacts of those with confirmed (Table 2), and 125 (90%) had pneumonia or re-
cases. The methods we used for estimating the in- spiratory failure. Among the 103 patients with
cubation period have been described previously.8 confirmed cases for whom data were available,
isolation precautions were instituted for 65 (63%)
R e sult s in an intensive care unit (ICU) because of severe
lower respiratory tract disease. Among patients
Epidemiologic Characteristics with cases identified as of December 1, a total of
of Confirmed Cases 47 with confirmed H7N9 virus infection (34%)
From March 25 through December 1, 2013, respira- and 1 with suspected infection had died of the
tory specimens from 1372 hospitalized patients acute respiratory distress syndrome (ARDS) or
with pneumonia of unexplained origin were tested, multiorgan failure in the hospital, 2 critically ill
and 104 patients (7.6%) were confirmed to be in- patients with confirmed cases remained hospi-
fected with the H7N9 virus. Of 2,130,049 respira- talized, and 88 had been discharged. One 2-year-
tory specimens obtained from outpatients with an old boy and one 25-year-old woman with mild
influenza-like illness through the sentinel surveil- upper respiratory tract illness were not hospital-
lance system for influenza-like illness, 8 (0.0004%) ized. Among 139 confirmed cases of H7N9 virus
tested positive for the H7N9 virus. As of Decem- infection, 49 (35%) were confirmed by means of
ber 1, 2013, a total of 139 laboratory-confirmed virus isolation, 1 by means of serologic testing,
cases of H7N9 virus infection (Fig. 1 and 2) and and the remaining 89 (64%) by means of nucleic
1 suspected case had been identified; cases were acid detection. Five of 139 persons with virologi-
identified in the following 10 provinces and 2 mu- cally confirmed cases also tested seropositive for
nicipalities: Zhejiang (50 confirmed cases), Shang- H7N9 virus antibodies with the use of a turkey
hai (33 confirmed cases and 1 suspected case), Ji- red-cell hemagglutinin inhibition assay (Table 1).
angsu (28 confirmed cases), Jiangxi (6 confirmed Data on recent exposure to animals were avail-
cases), Fujian (5 confirmed cases), Anhui (4 con- able for 131 of the 139 patients with confirmed
firmed cases), Henan (4 confirmed cases), Hunan H7N9 virus infection. Of these, 107 (82%) reported
(2 confirmed cases), Beijing (2 confirmed cases), a history of recent exposure to animals (Table 1):
Shandong (2 confirmed cases), Guangdong (2 con- 88 (82%) to chickens, 24 (22%) to ducks, and 6 (6%)
firmed cases), and Hebei (1 confirmed case). to swine; the exposures occurred either while
The median age of patients with confirmed they were working at or while they were visiting
H7N9 virus infection was 61 years (range, 2 to a live animal market. Other animals that these
91; interquartile range, 46 to 73); 58 cases (42%) 107 patients reported having been exposed to in-
occurred in persons 65 years of age or older, and cluded pigeons, geese, quail, wild birds, pet birds,
4 (3%) were in children younger than 5 years of cats, and dogs. Information on a history of ex-
age, all of whom had clinically mild upper respi- posure to live animals is unclear for 8 patients
ratory illness (see Fig. S1 in the Supplementary with confirmed H7N9 virus infection, either be-
Appendix for the age distribution). Most con- cause the investigations are still ongoing or be-
522 n engl j med 370;6 76

nejm.org february 6, 2014
Live poultry markets closed in 1 municipality
and 9 provinces
Hospitalized
April 3 Recovered
National JPCM established, guideline (version 1) released, and
Died
Source: The New England Journal
enhanced surveillance implemented in ILI surveillance system

April 7
8 Local JPCM established in Shanghai and Jiangsu,
Zhejiang, and Anhui provinces
Epidemiology
March 28
7 NIC of China CDC
isolated virus
April 19
6
H7N9-specific Internet-based reporting
n engl j med 370;6

surveillance system launched
March 26
5 NIC of China CDC
finished gene
nejm.org
sequencing
4
May 10
Number of Cases
H7N9 guideline (version 2) released
3 March 8

Shanghai family
cluster detected
77 february 6, 2014
2

of Medicine of avian Influenza (H7N9) Outbreak in China
0
2/1 2/17 3/5 3/21 4/6 4/22 5/8 5/24 6/9 6/25 7/11 7/27 8/12 8/28 9/13 9/29 10/15 10/31 11/16 12/2
Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
Date of Onset of Illness
Figure 1. Date of Onset of Illness in 139 Patients with Confirmed H7N9 Virus Infection in China.
Detailed information on closures of live poultry markets and emergency responses in 10 provinces and 2 municipalities is provided in Figure S2 in the Supplementary Appendix.
CDC denotes Center for Disease Control and Prevention, ILI influenza-like illness, JPCM Joint Prevention and Control Mechanism, and NIC National Influenza Center.
COLOR FIGURE
Draft 6 1/16/2014
Author Feng_rr1304617
Fig # 1
523
Title Preliminary Report:
Epidemiology of the Initial
H7N9 Influenza Outbreak in China
DE Baden
ME Moskowitz
Artist Williams
Pub Date 2/6/2014
Source: The New England Journal of Medicine The n e w e ng l a n d j o u r na l of m e dic i n e
suspected infection. The index case patient and his

father (confirmed case 2) lived in the same house,
Beijing and the index patient’s brother (confirmed case 1)
(2/0/2)
N lived with his wife nearby. After the index case pa-
tient (suspected case 1) became ill, his brother (con-
firmed case 1) and his father (confirmed case 2)
had prolonged, close, unprotected contact with
him, including eating together, providing care, and
Hebei accompanying him to seek medical care before his
(1/1/0) Yellow
Sea hospitalization (Fig. 3A, and Section S4 in the Sup-
Shandong plementary Appendix). None of the three members
Jiangsu
(2/0/2)
(28/9/19)
of the cluster raised poultry or other animals, none
Shanghai
brought live poultry into their home, and none had
(33/18/15) direct contact with sick or dead poultry. The index
Henan case patient had visited a live poultry market,
(4/1/3) purchased a chicken, observed the slaughtering
Anhui East China process, brought the freshly killed chicken home,
China (4/2/2) Sea and prepared, cooked, and ate the chicken within
2 weeks before the onset of his illness.
Zhejiang
(50/13/35)
The second family cluster included two persons
with confirmed H7N9 virus infection (Fig. 3B, and
Section S4 in the Supplementary Appendix). After
Jiangxi Sizes of circles are the father (confirmed case 1) became ill, his
Hunan proportional to the number
(6/1/5)
(2/1/1)
Fujian
of cases
Currently
daughter (confirmed case 2) had prolonged, close,
(5/1/4) hospitalized unprotected contact with him, including eating
Died together, providing care, and accompanying him
Discharged or to seek medical care before his hospital admission.
not admitted
She also provided unprotected bedside hospital
Name of province
Beijing or municipality care for her father during the period from March
Guangdong
(2/0/2)
(2/0/2) Discharged or 11 through March 15. Diarrhea developed in the
not admitted father, and the daughter washed her father’s diar-
0 250 500 km
Died rhea-soiled underwear on March 18 while wearing
Total number of cases gloves. The father had visited a live poultry mar-
Figure 2. Geographic Distribution of 139 Confirmed Cases of H7N9 Virus ket 7 days before the onset of his illness. The
Infection in China, as of December 1, 2013. daughter did not raise poultry or animals at
home and had not had any exposures to animals
(i.e., had not brought live poultry into the home
cause data on exposure could not be obtained or visited a live poultry market or had any direct
from patients owing to critical illness resulting in or indirect contact with poultry or pigs).
death. The estimated median incubation period The third family cluster included a wife and
in 23 patients with confirmed cases for whom husband with confirmed H7N9 virus infection
detailed data on animal and environmental ex- (Fig. 3C, and Section S4 in the Supplementary Ap-
posures were available was 6 days (range, 1 to 10) pendix). After the wife (confirmed case 1) became
(Table S1 in the Supplementary Appendix). ill, her husband (confirmed case 2) had prolonged,
close, unprotected contact with her, including eat-
Family Clusters ing with her, providing care for her, and accompa-
As of December 1, four family clusters had been nying her to seek medical care before her admis-
identified in three areas. Detailed information on sion to the hospital. He also provided unprotected
exposure, date of illness onset, and follow-up for bedside care for his wife in the hospital on April 2
these family clusters is provided in Figure 3, and in and April 3. The wife had visited a live poultry
Section S4 in the Supplementary Appendix. The market within 7 days before the onset of her ill-
first family cluster comprised two persons with ness. The husband did not raise poultry or animals
confirmed H7N9 virus infection and one with a at home and had not had any exposures to animals
524 n engl j med78

370;6 nejm.org february 6, 2014
Source: The New England Journal of Epidemiology
Medicine of avian Influenza (H7N9) Outbreak in China
Table 1. Epidemiologic Characteristics of 139 Patients with Confirmed H7N9 Virus Infection in China.
Characteristic Value
Age — yr
Median 61
Interquartile range 46–73
Age <5 yr — no. (%) 4 (3)
Age ≥65 yr — no. (%) 58 (42)
Male sex — no. (%) 98 (71)
Type of residence — no. (%)
Urban 101 (73)
Rural 38 (27)
Poultry worker — no. (%)* 9 (6)
Presence of underlying medical conditions — no./total no. (%)† 79/108 (73)
Exposure to symptomatic case patient within 2 wk before illness onset 5/120 (4)
— no./total no. (%)‡
History of exposure to animals — no./total no. (%)§ 107/131 (82)
Chickens 88/107 (82)
Ducks 24/107 (22)
Pigeons 13/107 (12)
Quail 2/107 (2)
Geese 2/107 (2)
Pet birds 3/107 (3)
Wild birds 7/107 (7)
Swine 6/107 (6)
Cats 2/107 (2)
Dogs 3/107 (3)
Type of exposure to animals — no./total no. (%)
Direct contact with poultry 63/107 (59)
Direct contact with swine 4/107 (4)
Visit to live poultry market 70/107 (65)
Method used for diagnosis of H7N9 — no. (%)
Virus isolation 49 (35)
Nucleic acid detection 89 (64)
Serologic testing¶ 6 (4)
* Six poultry workers slaughtered poultry at a live poultry market, two in Nanjing, Jiangsu province, and one each in Bozhou,
Anhui province; Zhoukou, Henan province; Zhengzhou, Henan province; and Huizhou, Guangdong province. Two poultry
workers transported live poultry, one each in Shanghai and Xuzhou, Jiangsu province. One poultry worker raised and traded
pigeons in Hangzhou, Zhejiang province.
† A total of 79 patients with confirmed H7N9 virus infection reported at least one underlying medical condition, including hyperten-
sion (32 patients), diabetes (14), heart disease (12), chronic bronchitis (7), hepatitis (4), smoking (4), rheumatic arthritis (4), rhi-
nitis (2), and obstructive pulmonary emphysema, pulmonary emphysema, asthma, silicosis, cerebral infarction, lymphoma, thy-
roid cancer, rectal carcinoma, breast cancer, chronic nephropathy, gout, prostatitis, pregnancy, and alcoholism (1 patient each).
‡ Four family clusters were reported (see Section S4 in the Supplementary Appendix). One cluster included a father (with a con-
firmed case) and his two sons (one with a confirmed case and one with a suspected case) in Shanghai, one cluster included a
father (with a confirmed case) and his daughter (with a confirmed case) in Jiangsu, one cluster included a woman (with a
confirmed case) and her husband (with a confirmed case) in Shanghai, and one cluster included a father (with a confirmed
case) and his son (with a confirmed case) in Shangdong.
§ Among 107 case patients who reported recent exposure to animals, 62 (58%) reported a single exposure to bird or poultry,
including chickens, ducks, geese, quail, pigeons, wild birds, and pet birds. Of the remaining 45 patients, 42 (39% of the
case patients with recent exposure to animals) had multiple exposures to poultry and swine and 3 (3%) had exposure to
only swine. Investigation of the history of exposure to animals is ongoing for 8 of 139 case patients (6%).
¶ Serum specimens tested seropositive for H7N9 virus antibodies with the use of a turkey red-cell hemagglutinin inhibition assay,
with an increase by a factor of 4 or more in antibodies in paired serum samples obtained during the acute and convalescent stages.
Among six seropositive cases, one, in Shanghai family cluster 1 (see Section S4 in the Supplementary Appendix), was confirmed
only by means of serologic testing; the remaining five seropositive cases were also confirmed by means of virus isolation or nucleic
acid detection, including one in Fujian province, two in Shandong province, one in Shanghai province, and one in Hunan province.

79 february 6, 2014 525
Table 2. Clinical Characteristics and Medical Care Timelines for 139 Patients Figure 3 (facing page). Timeline of Pertinent Exposures
with Confirmed H7N9 Virus Infection in China.* and Dates of Illness Onset in Four Family Clusters of
Cases of H7N9 Virus Infection in China.
Patients with Confirmed
ICU denotes intensive care unit, and rRT-PCR real-time
Variable Cases (N = 139)
reverse-transcriptase–polymerase-chain-reaction assay.
Clinical outcome — no./total no. (%)†
Required hospitalization 137/139 (99) son (confirmed case 2) had direct and very close
Currently hospitalized 2/139 (1) contact with him, including hugging, sitting, eat-
Admitted to ICU 65/103(63) ing, and sleeping together. The family did not raise
Required mechanical ventilation 56/91 (62) poultry or other animals and did not bring live
Had ARDS 48/83 (58)
poultry into their home. Neither the father nor the
Died in hospital‡ 47/139 (34)
son had any direct contact with poultry and neither
Recovered 90/139 (65)
had visited a live poultry market within 2 weeks
Oseltamivir treatment — no./total no. (%) 79/109 (72)
before the onset of illness. H7N9 virus infection in
Time from illness onset to oseltamivir treatment
— days
the son was identified through monitoring and
Median 6.0
testing of 12 close contacts of confirmed case 1.
Interquartile range 5.0–9.0
Medical Care Timelines
Time from illness onset to first medical care — days
Median 1.0 Among the 137 patients with confirmed H7N9 vi-
Interquartile range 0–3.0 rus infection for whom data were available (99% of
Time from illness onset to hospitalization — days the 139 patients with confirmed cases), the median
Median 4.0 time from the onset of illness to the first medical
Interquartile range 3.0–6.0 visit was 1 day, and patients were hospitalized a
Time from illness onset to ICU admission — days median of 4 days after the onset of illness (Table 2).
Median 7.0 Among 109 of the patients with available data on
Interquartile range 5.0–9.0 oseltamivir administration, 79 (72%) received osel-
Time from illness onset to development of ARDS tamivir treatment beginning a median of 6 days
— days (range, 0 to 15) after the onset of illness. ARDS
Median 7.0 developed during the course of hospitalization in
Interquartile range 5.0–9.0 48 of 83 patients with confirmed H7N9 virus infec-
Time from illness onset to death — days tion for whom data were available (58%) after a
Median 21.0 median of 7 days, and 47 patients died a median of
Interquartile range 12.5–36.0 21 days after the onset of illness.
* In the case of some characteristics, complete data were not available for all
Close Contacts
139 patients with confirmed H7N9 virus infection. Case investigations are on-
going. ARDS denotes acute respiratory distress syndrome, and ICU intensive As of December 1, excluding data on patients with
care unit. confirmed cases who were close contacts of index
† A total of 87 patients with confirmed cases had been discharged as of December 1,
2013. One 2-year-old boy with a confirmed case in Shanghai was not hospital- patients in the four family clusters, data were avail-
ized, because his illness was mild; he recovered 5 days after the onset of illness. able for 2675 close contacts of the 139 patients with
A 25-year-old woman with a confirmed case in Jiangsu province was not hospitalized, confirmed cases, in Zhejiang (1133 contacts), Ji-
because her illness was mild, and she recovered 18 days after the onset of illness.
‡ One additional death occurred after hospital discharge and is not included as angsu (552), Shanghai (378), Anhui (135), Fujian
a death in the hospital. This was a case patient from Jiangsu province who re- (111), Hunan (91), Guangdong (103), Beijing (43),
covered and was discharged from the hospital in August but died at home in Henan (42), Jiangxi (46), Hebei (30), and Shandong
September from an underlying medical condition.
(11). Among the 1953 close contacts of the patients
with confirmed cases for whom demographic in-
(i.e., had not brought live poultry into the home or formation was available, 1039 (53%) were health
visited a live poultry market or had any direct or care workers, 515 (26%) were family members or
indirect contact with poultry or pigs). relatives, and 399 (20%) were social contacts. As
The fourth family cluster included a father and of December 1, all 2675 contacts had been followed
son with confirmed H7N9 virus infection (Fig. 3D, for up for 7 days; among these close contacts, respi-
and Section S4 in the Supplementary Appendix). ratory symptoms developed during the 7-day sur-
After the father (confirmed case 1) became ill, his veillance period in 28 (1%): 9 household members,
526 n engl j med80

370;6 nejm.org february 6, 2014
Epidemiology of avian Influenza (H7N9) Outbreak in China
A Shanghai Family Cluster 1

Throat-swab collection;
Admitted to H7N9-negative
hospital by rRT-PCR
Onset ICU Died
Index case patient
(Suspected case)
2/5 2/11 2/20 2/21 2/26 2/28
Convalescent
Provided care for serum collection
index case patient Throat-swab and
Lived Admitted to acute serum collection H7N9 confirmed
together Onset hospital Discharged by serology
Case patient 2
(Confirmed case 1) 2/5 2/11 2/19 2/25 2/26 3/31 4/6 4/11
Provided care for Throat-swab H7N9 confirmed by

index case patient Admitted to collection virus isolation
Onset hospital ICU Died
Case patient 3
(Confirmed case 2) 2/5 2/11 2/19 2/25 2/26 2/27 3/4 3/28
B Jiangsu Family Cluster

Transferred Endotracheal aspirate collection;
to ICU of H7N9 confirmed by rRT-PCR
Onset hospital A and virus isolation
Transferred Throat-swab
Visited Admitted to to ICU of collection; H7N9-negative
market hospital A hospital B by rRT-PCR Died
Index case patient
(Confirmed case 1)
3/1 3/8 3/11 3/15 3/18 3/27 3/31 5/4
Visited ICU of hospital A Throat-swab collection;
Onset
Provided bed- H7N9 confirmed by Endotracheal aspirate
Provided care side care for Visited rRT-PCR collection; H7N9 confirmed by
at home for index index case ICU of Admitted to rRT-PCR and virus isolation
case patient patient hospital B hospital B Died
Case patient 2
(Confirmed case 2)
3/1 3/8 3/11 3/15 3/18 3/21 3/24 3/27 3/31 4/24
C Shanghai Family Cluster 2

Respiratory tract aspirate collection;
Visited hospital Admitted H7N9 confirmed by rRT-PCR
Onset to hospital Died
Index case patient
(Confirmed case 1)
3/23 3/27 3/29 4/2 4/3 4/4
3/28 3/30 Respiratory tract–aspirate collection;
Throat-swab
H7-positive and N9-negative by rRT-PCR
collection; H7N9-negative
Visited by rRT-PCR
Provided care for Throat-swab collection; Throat-swab collection; H7N9
index case patient hospital confirmed by rRT-PCR
Admitted H7N9 and H1N1
Onset to hospital confirmed by rRT-PCR Died
Case patient 2
(Confirmed case 2)
3/23 3/27 4/2 4/3 4/4 4/5 4/10 4/11 4/12 6/26
D Shandong Family Cluster

Traveled by car at least twice Visited clinic and Admitted
per day and parked near received intravenous to hospital Throat-swab collection;
chicken coop antibiotic treatment H7N9 confirmed by rRT-PCR
Onset ICU Discharged
Index case patient
(Confirmed case 1)
4/1 4/18 4/20 4/21 4/22 4/23 5/16
4/16
Lived together in same house, hugging, sitting,
eating, and sleeping together
4/1 4/18 4/20 4/27 4/28 5/6
Case patient 2
(Confirmed case 2)
Medical Onset Discharged
Accompanied ill father observation Throat-swab collection;
to clinic visits
H7N9 confirmed by rRT-PCR
COLOR FI G URE
Draft 9 1/22/2014
Author Feng_rr1304617
n engl j med 370;6 nejm.org february 6, 2014 527
Fig #
Title
3
81
Preliminary Report:
Epidemiology of the Initial
H7N9 Influenza Outbreak in China
Copyright © Massachusetts
DE Medical Society. All rights reserved.
Baden
ME
Back toMoskowitz
Artist
Table of
Williams
Contents
Pub Date 2/6/2014
AUTHOR PLEASE NOTE:
1 medical intern, 1 patient who shared a room with Discussion

a confirmed case patient, and 17 health care work-
ers (Section S5 in the Supplementary Appendix). An outbreak of human infections with a novel
Throat swabs were collected from these 28 ill avian influenza A(H7N9) virus began in Febru-
contacts a median of 1 day (range, 0 to 8) after ary 2013 and as of December 1 had resulted in
the onset of illness. All 28 ill close contacts were 139 confirmed cases identified in 10 provinces
negative for the H7N9 virus, as assessed by means and 2 municipalities, primarily in eastern China.
of real-time RT-PCR. Although persons in a wide age spectrum were
affected (age range, 2 to 91 years), the majority of
Control Measures patients with confirmed H7N9 virus infection
On April 3, 2013, the NHFPC established the Joint were older (median age, 61 years), male (71%),
Prevention and Control Mechanism (JPCM) to lead and urban residents (73%), and most were con-
the national response to H7N9, issued the first na- sidered to be at increased risk for complications
tional H7N9 technical guideline, and implemented from influenza owing to age (<5 years or ≥65 years)
enhanced surveillance for H7N9 virus infection or the prevalence of certain underlying medical
among persons with influenza-like illness in the conditions. Most patients were hospitalized with
existing sentinel surveillance system (Fig. 1). A local severe lower respiratory tract illness,10 with a
JPCM was established in four provinces — Shang- case fatality proportion of 34% to date. Except
hai, Jiangsu, Zhejiang, and Anhui — on April 7. The for four case clusters among family members,
China CDC established an Internet-based national most patients with confirmed H7N9 virus infec-
H7N9 case-reporting system for provinces and mu- tion were epidemiologically unrelated. Two ad-
nicipalities on April 19. The national H7N9 techni- ditional cases in critically ill patients were not
cal guideline was revised on May 10. Of the 10 af- included in our analyses; one patient had traveled
fected provinces and 2 municipalities, all except from eastern China to Taiwan, where illness on-
Guangdong province launched emergency re- set occurred and H7N9 virus infection was con-
sponses immediately after detection of a confirmed,11 and one had traveled from southern
firmed H7N9 case (Fig. S2 in the Supplementary China to the Hong Kong Special Administrative
Appendix) and closed live poultry markets for 1 to Region (HKSAR), where illness onset occurred
2 months after an H7N9 case occurred. (Live poul- and H7N9 virus infection was confirmed in early
try markets have been forbidden in Beijing since December. Since December 1, other confirmed
2005.) Live poultry markets in Guangdong were cases of H7N9 virus infection have been identi-
monitored after an H7N9 case was identified, and fied in China, HKSAR, and Taiwan.
daily poultry deliveries were limited to ensure that Human infections with avian influenza A (H7)
the unsold poultry that was kept overnight did not viruses have been reported sporadically and are
exceed 10% of the poultry delivered daily. usually associated with exposures to poultry.12-14
The outbreak can be characterized by five Previous cases of H7 virus infection in humans
phases (Fig. 1, and Fig. S2 in the Supplementary have been characterized by mild illness (conjuncti-
Appendix): the initial phase from February 17 vitis or uncomplicated influenza) or by moderate
through March 27, with sporadic cases identified illness (lower respiratory tract disease) that results
in Shanghai, Zhejiang, and Jiangsu provinces in in hospitalization.12,13,15,16 Only one fatal case of
eastern China; the epidemic peak from March 28 H7 virus infection has been reported previously;
through April 18, during which 98 of the 139 total that case occurred in an adult with a highly patho-
confirmed cases (71%) were reported, in two mu- genic avian influenza A (H7N7) virus infection.17
nicipalities and eight provinces; a declining Many of the confirmed H7N9 case patients had
phase from April 19 through May 31, with a sharp critical and fatal illness, suggesting that the H7N9
reduction in cases after closure of live poultry mar- virus is more virulent in humans than are other H7
kets in one municipality and eight provinces; an viruses. The age distribution among H7N9 case
interepidemic phase from June 1 through Septem- patients is older and wider than that among H5N1
ber 30, with only 2 confirmed cases reported; and case patients in China (Fig. S1 in the Supplemen-
a reemergence phase beginning October 1, with tary Appendix), whereas the case fatality propor-
5 new confirmed H7N9 cases as of December 1. tion and the risk of death to date are lower than

82 february 6, 2014
those for reported cases of H5N1 virus infection in suggest that early oseltamivir treatment has the
China and worldwide.18,19 greatest clinical benefit but that starting treatment
Early surveillance for H7N9 cases was focused up to 5 days after the onset of illness may still
on case finding for severe lower respiratory tract reduce the risk of critical illness and death.27-30
illness, and since April 3, expanded testing of out- Reports suggest that the H7N9 viruses isolated
patients with influenza-like illness has identified from humans and analyzed to date are resistant
some mild cases of illness with H7N9 virus infec- to adamantane antiviral agents and are suscep-
tion.20-22 Enhanced surveillance for less severe ill- tible to neuraminidase inhibitors.3,31,32 Early clin-
ness with H7N9 virus infection can help to deter- ical suspicion of H7N9 virus infection and early
mine the clinical spectrum of the illness and the administration of oseltamivir may help to reduce
total number of cases of H7N9 symptomatic ill- the severity of the disease. However, emergence
ness and to inform an understanding of the true of the R292K mutation in viral neuraminidase,
case fatality proportion. Since this H7N9 virus which confers in vitro resistance to neuramini-
appears to have emerged recently to infect hu- dase inhibitors, has been documented in H7N9
mans, population immunity is expected to be low, viruses during or after oseltamivir treatment in
and persons of any age may be susceptible to infec- some critically ill patients.3,33 Therefore, to in-
tion. Owing to limitations of surveillance and test- form clinical management, clinical trials are
ing, the number of patients with confirmed cases needed to define efficacious therapy for H7N9
of H7N9 virus infection is likely to be an underes- virus infection, including investigational agents
timate of all symptomatic cases that have oc- and combination antiviral treatment.
curred. One modeling study estimated that 27,000 Although the source of H7N9 virus infection in
cases of symptomatic H7N9 virus infection (95% patients with confirmed cases who had exposure
confidence interval, 9530 to 65,000) had occurred to animals cannot be verified without extensive
in China as of May 28.23 Very limited data are avail- H7N9 testing of animals, we suspect that it is
able from small seroprevalence studies.24,25 Larger likely to be infected poultry. H7N9 virus is a low
seroepidemiologic studies among different ex- pathogenic avian influenza A virus and does not
posed groups can help define the true denomina- cause identifiable illness or death in poultry; there-
tor of H7N9 virus infections. fore, only laboratory testing can identify poultry
The median time from the onset of illness to infections. Among cases for which data are avail-
hospitalization among the 133 of 137 patients with able, 82% occurred in patients who had exposure
confirmed H7N9 virus infection for whom data on to live animals such as poultry or swine, including
hospitalization were available was 4 days, and the during visits to live animal markets. This raises the
median time from the onset of illness to the devel- possibility of zoonotic H7N9 virus transmission
opment of ARDS among the 48 case patients with from healthy-appearing poultry or swine to
ARDS (of 83 patients for whom data on ARDS were humans through direct or close contact or through
available) was 7 days; the corresponding median exposure to environments that are contaminated
times among patients with H5N1 virus infection with infected poultry or swine. For example, visit-
were 7 days and 7.5 days.18,26 The median duration ing a live poultry market, where avian influenza A
from the onset of illness to death among the 47 viruses can be maintained and amplified, has been
persons with confirmed cases who died was 21 identified as a risk factor for H5N1 virus infection
days, which is much longer than the correspond- in Hong Kong34 and urban China.35,36 H7N9 virus
ing median of 11 days among 30 persons with fatal has been isolated from poultry or environmental
H5N1 cases in China. specimens collected at live poultry markets that
Patients with confirmed H7N9 infection re- were visited by some patients with confirmed cases
ceived oseltamivir antiviral treatment a median before the onset of illness, and genetic sequencing
of 6 days after the onset of illness (the median results revealed the same H7N9 virus strain in vi-
before April 3 was 9 days), probably owing to ruses isolated from humans and from poultry or
delays in seeking medical help and consequent environmental samples from live poultry mar-
delays in the identification of suspected influen- kets.37-40 Nine cases of H7N9 virus infection (6.5%)
za. Retrospective observational studies of influ- occurred in poultry workers, including seven work-
enza A(H1N1)pdm09 and H5N1 virus infections ers in live poultry markets.

nejm.org february 6, 2014 529
Further evidence implicating live poultry mar- hospitals; and active monitoring of close contacts
kets in urban areas as the source of zoonotic H7N9 for illness. In addition, national guidelines recom-
virus transmission is derived from case–control mend that antiviral treatment with oseltamivir
studies and the observed reduction in confirmed should be administered as soon as possible in pa-
H7N9 cases after closure of live poultry markets. tients with suspected or confirmed cases of H7N9
One case–control study identified direct contact virus infection.
with poultry or birds in the 2 weeks before the Our study had several limitations. First, we were
onset of illness, chronic medical conditions exclud- not able to collect detailed information from all
ing hypertension, and exposure to an environment patients on exposures, such as the times, frequen-
with poultry, including visiting a live poultry mar- cy, intensity, and duration of exposures. Informa-
ket, as risk factors associated with H7N9 virus in- tion on exposures is useful for estimating the in-
fection.41 A larger study identified visiting live cubation period after possible exposure to animals
poultry markets as a risk factor, and contact with or live-animal markets and for evaluating risk
poultry as an independent risk factor, for H7N9 factors for H7N9 virus infection. Second, we may
virus infection (unpublished data from the China not have identified all the close contacts of case
CDC). One ecologic modeling study estimated that patients. Third, we did not have a standard proto-
closure of live poultry markets reduced the mean col and questionnaire to collect information from
daily number of H7N9 virus infections in the four all contacts of the 139 patients with confirmed
most affected cities by 97 to 99%.42 cases. However, the China CDC issued a guideline
To date, follow-up prospective investigations and protocol for field investigations of case pa-
of close contacts of patients with confirmed tients and close contacts and since April 1 has
H7N9 virus infection have not conclusively estab- provided training for personnel at all 31 provincial
lished human-to-human H7N9 virus transmis- CDCs. A structured Internet-based reporting sys-
sion. However, in four family clusters, limited, tem was used to collect information on confirmed
nonsustained human-to-human transmission of cases. This has helped to standardize data collec-
H7N9 virus after close, prolonged, unprotected tion. Fourth, specimens were not available for
contact with a symptomatic patient with H7N9 H7N9 testing from some patients with suspected
virus infection remains a possibility. Confirmed cases. It is also possible that by obtaining a single
cases in three of the four clusters were identified throat swab to test for the presence of the H7N9
in blood-related family members. Similar family virus, as compared with obtaining multiple speci-
clusters of H5N1 cases that occurred after com- mens from different respiratory sites on different
mon exposures to poultry or limited human-to- days, we could have missed detection of the virus
human transmission have been identified.8,43,44 in symptomatic contacts. Paired serum samples
Paired serum samples were obtained during the have not been obtained from some of the contacts
acute and convalescent stages of illness from because they have declined to have serum collect-
contacts of case patients for further assessment ed. Fifth, in-hospital mortality may be an under-
of the potential for secondary human-to-human estimate of H7N9-associated mortality; one patient
H7N9 virus transmission, including the potential with a confirmed case was hospitalized, recovered,
identification of asymptomatic infections when was discharged, and later died of an underlying
testing is completed. Several studies have shown medical condition but was not included as one of
that H7N9 virus can bind to receptors in both the the 47 deaths among hospitalized patients with
upper and lower airway in humans and can be confirmed cases. Owing to surveillance limita-
transmitted variably from experimentally infected tions, some H7N9 virus infections have probably
ferrets to susceptible ferrets by means of drop- been missed.23 Finally, since the outbreak is ongo-
lets.31,37,45-49 These studies support the observa- ing, we were unable to include in our analyses data
tion that limited, nonsustained human-to-human from the patients with confirmed H7N9 cases that
transmission of H7N9 virus might occur. have been identified since December 1, 2013, or
Although the risk of human-to-human H7N9 from their close contacts.
virus transmission appears to be low, Chinese na- In summary, a novel avian influenza A(H7N9)
tional guidelines recommend implementing con- virus has caused severe and fatal lower respira-
trol measures, such as prompt isolation of H7N9 tory tract illness in persons in 12 different areas
patients; implementation of standard, contact, and of China. Some clinically mild cases have been
droplet precautions by health care personnel in identified since the surveillance was widened,

84 february 6, 2014
suggesting that there is a wide clinical spectrum and control efforts and assessing the pandemic
of H7N9 virus infection. The epidemiologic potential of this H7N9 virus. Prompt closure of live
findings suggest that most confirmed H7N9 poultry markets is indicated in areas where new
cases were epidemiologically unrelated. On the human cases of H7N9 virus infection are detected.
basis of data from various studies to date, most Editor’s note: As of January 21, 2014, more than
case patients were probably exposed to H7N9 200 confirmed cases of H7N9 had been identified
virus and infected during visits to live poultry since February 2013, including more than 65 since
markets, and the decline in cases during May December 1, 2013, in this ongoing outbreak.
followed the closure of live poultry markets in The views expressed in this article are those of the authors
Shanghai municipality and nine provinces. Fol- and do not represent the official policy of the Chinese Center for
Disease Control and Prevention or the U.S. Centers for Disease
low-up investigations of contacts of patients Control and Prevention.
with confirmed H7N9 virus infection suggest Supported by the China–U.S. Collaborative Program on
that the risk of secondary transmission of H7N9 Emerging and Re-emerging Infectious Diseases, a grant
(201202006-01) from Research and Promotion of Key Technol-
virus, including to health care personnel, is low ogy on Health Emergency Preparation and Dispositions, a grant
at this time. However, in four family clusters, (No. KJYJ-2013-01-02) from the National Ministry of Science and
limited, nonsustained human-to-human trans- Technology Emergency Research Project on human infection
with avian influenza H7N9 virus, a grant (No. ZX201109) from
mission of H7N9 virus could not be ruled out Jiangsu Province Health Development Project with Science and
and may have occurred. Education, and a grant (No. 2009GG10002054) from Science
On the basis of the experience with increased and Technology Development Plan of Shandong Province.
Disclosure forms provided by the authors are available with
circulation of H5N1 virus among poultry during the full text of this article at NEJM.org.
periods of cooler temperatures, the spread of We thank staff members of the Bureau of Disease Control and
H7N9 virus among poultry is likely to increase Prevention and Health Emergency Response Office of the National
Health and Family Planning Commission and provincial and local
during the winter and spring months, with the departments of health for providing assistance with administra-
potential for increased transmission to exposed tion and data collection; staff members at county, prefecture and
persons. Therefore, enhanced surveillance for provincial government offices, CDCs, and hospitals in Shanghai
and Beijing municipalities and Zhejiang, Jiangsu, Anhui, Henan,
H7N9 virus among poultry and people, investi- Hunan, Shangdong, Fujian, Jiangxi, Hebei, and Guangdong prov-
gations of contacts of confirmed cases, and vi- inces for providing assistance with field investigation administra-
rologic analyses to assess genetic changes that tion and data collection; Jeffrey McFarland and Dale Hu from the
U.S. CDC for comments on the manuscript; and Liangliang Cui
might suggest increased transmissibility among from the Jinan Prefecture CDC and Jian Zhao from the Chinese
humans are all critical to informing prevention CDC for assistance with the preparation of figures.
Appendix
The authors’ affiliations are as follows: the Public Health Emergency Center (Q.L., L.Z., N.X., L.M., Z.H., W.T., Y.C., L.L., F.D., B.L.,
M.W., Y.G., L.J., Y.Z., Z.F.) and National Institute for Viral Disease Control and Prevention (D.W., R.G., X.L., T.B., S.Z., S.W., Y.S.),
Chinese Center for Disease Control and Prevention (H.L., H.Y., W.Y., Y.W.), Chinese Field Epidemiology Training Program (R.X.), and
Beijing Municipal Center for Disease Control and Prevention (X.P.), Beijing, Jiangsu Provincial Center for Disease Control and Preven-
tion (M.Z., F.T.) and Nanjing Prefecture Center for Disease Control and Prevention (Y.X.), Nanjing, Zhejiang Provincial Center for
Disease Control and Prevention, Hangzhou (Z.C., E.C., C.C.), Anhui Provincial Center for Disease Control and Prevention, Hefei (F.L.,
Jun He), Shanghai Municipal Center for Disease Control and Prevention, Shanghai (H.W., J. Hu, F.W.), Guangdong Provincial Center
for Disease Control and Prevention, Guangzhou (Jianfeng He), Hebei Provincial Center for Disease Control and Prevention, Shiji-
azhuang (Q.L.), Shandong Provincial Center for Disease Control and Prevention, Jinan (X.W.), Hunan Provincial Center for Disease
Control and Prevention, Changsha (L.G.), Henan Provincial Center for Disease Control and Prevention, Zhengzhou (G.L.), Fujian Pro-
vincial Center for Disease Control and Prevention, Fuzhou (Y.Y.), and Jiangxi Provincial Center for Disease Control and Prevention,
Nanchang (H.Y.) — all in China; and Influenza Division, Centers for Disease Control and Prevention, Atlanta (T.M.U.).
References
1. Gao R, Cao B, Hu Y, et al. Human infec- 4. Xiang N, Havers F, Chen T, et al. Use al. Transmission dynamics, border entry
tion with a novel avian-origin influenza A of national pneumonia surveillance to de- screening, and school holidays during the
(H7N9) virus. N Engl J Med 2013;368: scribe influenza A(H7N9) virus epidemi- 2009 influenza A (H1N1) pandemic, Chi-
1888-97. ology, China, 2004-2013. Emerg Infect na. Emerg Infect Dis 2012;18:758-66.
2. World Health Organization. WHO Dis 2013;19:1784-90. 7. Wang H, Feng ZJ, Shu YL, et al. Prob-
case definitions for human infections with 5. National Health and Family Planning able limited person-to-person transmis-
influenza A (H5N1) virus (http://www Commission, People’s Republic of China. sion of highly pathogenic avian influenza
.who.int/influenza/resources/documents/ Guidelines of case clinical intervention A (H5N1) virus in China. Lancet 2008;
case_definition2006_08_29/en). and disease control and prevention of hu- 371:1427-34.
3. Yu H, Shu Y, Hu S, et al. The first con- man infection with avian influenza H7N9 8. Huai Y, Xiang N, Zhou L, et al. Incu-
firmed human case of avian influenza A virus (in Chinese) (http://www.moh.gov bation period for human cases of avian
(H5N1) in Mainland China. Lancet 2006; .cn/mohwsyjbgs/fkzs/list.shtml). influenza A (H5N1) infection, China.
367:84. 6. Yu H, Cauchemez S, Donnelly CA, et Emerg Infect Dis 2008;14:1819-21.

nejm.org february 6, 2014 531
9. Prevention and control of influenza ness in avian influenza A(H7N9) virus- factors for human illness with avian in-
with vaccines: recommendations of the infected poultry worker, Huzhou, China, fluenza A (H5N1) virus infection in Chi-
Advisory Committee on Immunization April 2013. Emerg Infect Dis 2013;19: na. J Infect Dis 2009;199:1726-34.
Practices (ACIP), 2010. MMWR Recomm 1885-8. 37. Chen Y, Liang W, Yang S, et al. Human
Rep 2010;59(RR-8):1-62. [Errata, MMWR 23. Yu H, Cowling BJ, Feng L, et al. Hu- infections with the emerging avian influ-
Recomm Rep 2010;59:993, 1147.] man infection with avian influenza A enza A H7N9 virus from wet market poul-
10. Gao HN, Lu HZ, Cao B, et al. Clinical H7N9 virus: an assessment of clinical se- try: clinical analysis and characterisation
findings in 111 cases of influenza A verity. Lancet 2013;382:138-45. of viral genome. Lancet 2013;381:1916-
(H7N9) virus infection. N Engl J Med 24. Bai T, Zhou J, Shu Y. Serologic study 25.
2013;368:2277-85. [Erratum, N Engl J Med for influenza A (H7N9) among high-risk 38. Wang C, Wang J, Su W, et al. Relation-
2013;369:1869.] groups in China. N Engl J Med 2013; ship between domestic and wild birds in
11. Chang SY, Lin PH, Tsai JC, Hung CC, 368:2339-40. live poultry market and a novel human
Chang SC. The first case of H7N9 influ- 25. Xu W, Lu L, Shen B, Li J, Xu J, Jiang S. H7N9 virus in China. J Infect Dis 2014;
enza in Taiwan. Lancet 2013;381:1621. Serological investigations of subclinical 209:34-7.
12. Koopmans M, Wilbrink B, Conyn M, influenza A(H7N9) infection among 39. Bao CJ, Cui LB, Zhou MH, Hong L,
et al. Transmission of H7N7 avian influ- healthcare and non-healthcare workers in Gao GF, Wang H. Live-animal markets
enza A virus to human beings during a Zhejiang Province, China. Clin Infect Dis and influenza A (H7N9) virus infection.
large outbreak in commercial poultry 2013;57:919-21. N Engl J Med 2013;368:2337-9.
farms in the Netherlands. Lancet 2004; 26. Yu H, Gao Z, Feng Z, et al. Clinical 40. Han J, Jin M, Zhang P et al. Epidemio-
363:587-93. characteristics of 26 human cases of high- logical link between exposure to poultry
13. Tweed SA, Skowronski DM, David ST, ly pathogenic avian influenza A (H5N1) and all influenza A(H7N9) confirmed
et al. Human illness from avian influenza virus infection in China. PLoS One 2008; cases in Huzhou city, China, March to
H7N3, British Columbia. Emerg Infect 3(8):e2985. May 2013. Euro Surveill 2013;18(20):
Dis 2004;10:2196-9. 27. Yu H, Feng Z, Uyeki TM, et al. Risk pii:20481.
14. Belser JA, Bridges CB, Katz JM, factors for severe illness with 2009 pan- 41. Ai J, Huang Y, Xu K, et al. Case-con-
Tumpey TM. Past, present, and possible demic influenza A (H1N1) virus infection trol study of risk factors for human infec-
future human infection with influenza in China. Clin Infect Dis 2011;52:457-65. tion with influenza A(H7N9) virus in Ji-
virus A subtype H7. Emerg Infect Dis 28. Muthuri SG, Myles PR, Venkatesan S, angsu Province, China, 2013. Euro
2009;15:859-65. Leonardi-Bee J, Nguyen-Van-Tam JS. Im- Surveill 2013;18:20510.
15. Ostrowsky B, Huang A, Terry W, et al. pact of neuraminidase inhibitor treat- 42. Yu H, Wu JT, Cowling BJ et al. Effect
Low pathogenic avian influenza A (H7N2) ment on outcomes of public health im- of closure of live poultry markets on poul-
virus infection in immunocompromised portance during the 2009-2010 influenza try-to-person transmission of avian influ-
adult, New York, USA, 2003. Emerg Infect A(H1N1) pandemic: a systematic review enza A H7N9 virus: an ecological study.
Dis 2012;18:1128-31. and meta-analysis in hospitalized pa- Lancet 2013 October 30 (Epub ahead of
16. Avian influenza A/(H7N2) outbreak in tients. J Infect Dis 2013;207:553-63. print).
the United Kingdom. Euro Surveill 2007; 29. Louie JK, Yang S, Acosta M, et al. 43. Kandun IN, Wibisono H, Sedyaning-
12(5):E070531.2. Treatment with neuraminidase inhibitors sih ER, et al. Three Indonesian clusters of
17. Fouchier RA, Schneeberger PM, for critically ill patients with influenza A H5N1 virus infection in 2005. N Engl J
Rozendaal FW, et al. Avian influenza A (H1N1)pdm09. Clin Infect Dis 2012;55: Med 2006;355:2186-94.
virus (H7N7) associated with human con- 1198-204. 44. Ungchusak K, Auewarakul P, Dowell
junctivitis and a fatal case of acute respi- 30. Adisasmito W, Chan PK, Lee N, et al. SF, et al. Probable person-to-person
ratory distress syndrome. Proc Natl Acad Effectiveness of antiviral treatment in hu- transmission of avian influenza A (H5N1).
Sci U S A 2004;101:1356-61. man influenza A(H5N1) infections: anal- N Engl J Med 2005;352:333-40.
18. Cowling BJ, Jin L, Lau EH, et al. Com- ysis of a Global Patient Registry. J Infect 45. van Riel D, Leijten LM, de Graaf M, et
parative epidemiology of human infec- Dis 2010;202:1154-60. al. Novel avian-origin influenza A (H7N9)
tions with avian influenza A H7N9 and 31. Zhou J, Wang D, Gao R, et al. Biologi- virus attaches to epithelium in both upper
H5N1 viruses in China: a population- cal features of novel avian influenza A and lower respiratory tract of humans.
based study of laboratory-confirmed cas- (H7N9) virus. Nature 2013;499:500-3. Am J Pathol 2013;183:1137-43.
es. Lancet 2013;382:129-37. 32. Watanabe T, Kiso M, Fukuyama S, et 46. Belser JA, Gustin KM, Pearce MB, et
19. World Health Organization. Cumula- al. Characterization of H7N9 influenza A al. Pathogenesis and transmission of avi-
tive number of confirmed human cases viruses isolated from humans. Nature an influenza A (H7N9) virus in ferrets
for avian influenza A(H5N1) reported to 2013;501:551-5. and mice. Nature 2013;501:556-9.
WHO, 2003-2013 (http://www.who.int/ 33. Hu Y, Lu S, Song Z, et al. Association 47. Zhang Q, Shi J, Deng G, et al. H7N9
influenza/human_animal_interface/ between adverse clinical outcome in hu- influenza viruses are transmissible in fer-
EN_GIP_20131008CumulativeNumber man disease caused by novel influenza A rets by respiratory droplet. Science 2013;
H5N1cases.pdf). H7N9 virus and sustained viral shedding 341:410-4.
20. Xu C, Havers F, Wang L, et al. Moni- and emergence of antiviral resistance. 48. Xu L, Bao L, Deng W, et al. Novel avi-
toring avian influenza A(H7N9) virus Lancet 2013;381:2273-9. an-origin human influenza A (H7N9) can
through national influenza-like illness 34. Mounts AW, Kwong H, Izurieta HS, et be transmitted between ferrets via respi-
surveillance, China. Emerg Infect Dis al. Case-control study of risk factors for ratory droplets. J Infect Dis 2013 Septem-
2013;19:1289-92. avian influenza A (H5N1) disease, Hong ber 24 (Epub ahead of print).
21. Ip DK, Liao Q, Wu P, et al. Detection Kong, 1997. J Infect Dis 1999;180:505-8. 49. Richard M, Schrauwen EJ, de Graaf M,
of mild to moderate influenza A/H7N9 35. Yu H, Feng Z, Zhang X, et al. Human et al. Limited airborne transmission of
infection by China’s national sentinel sur- influenza A (H5N1) cases, urban areas of H7N9 influenza A virus between ferrets.
veillance system for influenza-like illness: People’s Republic of China, 2005-2006. Nature 2013;501:560-3.
case series. BMJ 2013;346:f3693. Emerg Infect Dis 2007;13:1061-4. Copyright © 2013 Massachusetts Medical Society.
22. Lv H, Han J, Zhang P, et al. Mild ill- 36. Zhou L, Liao QH, Dong LB, et al. Risk
532 n engl j med 370;6 86

nejm.org february 6, 2014
CDC Issues Health Advisory on Protecting Humans During Bird Flu Outbreak
Physician’s First Watch, published June 4, 2015
NEWS SUMMARY
The CDC is warning clinicians that humans could be infected by highly-pathogenic avian influenza (HPAI) A H5
viruses that have been circulating among U.S. poultry flocks.
The U.S. Department of Agriculture estimates that 40 million poultry across 20 states have been infected with or
exposed to HPAI H5N2, H5N8, or H5N1. It’s the first time these strains have been identified in U.S. flocks. “Their
appearance in North American birds may increase the likelihood of human infection in the United States,” the
CDC reports.
The CDC recommends the following:
• Clinicians should consider these HPAI H5 viruses in people with new-onset influenza-like illness or acute re-
spiratory infection (which could include conjunctivitis), who’ve also had close contact with birds that may be
infected or surfaces contaminated by potentially-infected birds within 10 days of illness onset.
• After a potential exposure, people should be monitored for 10 days for signs of illness.
• Antiviral prophylaxis may be considered after exposure.
There are currently no human vaccines available against these strains.
— Kelly Young
87
An Avian Flu Vaccine

A potentially protective immune response can be induced with two doses of an adjuvanted H7N9
influenza vaccine.
Jackson LA et al. Effect of varying doses of a monovalent H7N9 influenza vaccine with and without AS03 and MF59
adjuvants on immune response: A randomized clinical trial. JAMA 2015 Jul 21; 314:237. (http://dx.doi.org/10.1001/
jama.2015.7916)
ARTICLE SUMMARY
Since 2013, when human infection with the novel influenza A(H7N9) virus was first recognized in China, more
than 550 cases and 210 deaths have been confirmed. The infections have generally developed following poultry
exposure, but limited person-to-person transmission seems to have occurred. An initial trial involving an inacti-
vated monovalent H7N9 vaccine suggested that the unadjuvanted preparation was poorly immunogenic but that
potentially protective antibody titers could be induced by two doses of the vaccine with MF59 adjuvant.
In a second trial, 946 healthy adults were randomized to receive various intramuscular doses of the vaccine, with
and without AS03 or MF59 adjuvants, administered in two injections 21 days apart. After the initial injection,
participants showed no significant antibody responses. Among those who received two doses of an AS03-adju-
vanted preparation, >80% achieved hemagglutination inhibition antibody (HIA) titers ≥1:40 (believed to be indic-
ative of protective immunity), regardless of preparation’s hemagglutinin antigen content. Lower HIA titers were
achieved with two doses of the MF59-adjuvanted vaccine, with one dose of adjuvanted and one of unadjuvanted
vaccine, and with one dose given with AS03 and one with MF59. Any schedule involving at least one adjuvanted
dose achieved higher HIA titers than schedules with two unadjuvanted doses. Older age and prior receipt of sea-
sonal influenza vaccine were associated with lower vaccine response. Although the vaccines were generally well
tolerated, local reactions were significantly more common with adjuvanted preparations; systemic reactions, after
a first dose that included an adjuvant.
COMMENT
These results confirm previous observations that an effective H7N9 vaccine will require administration in two
doses with an adjuvant. Equally problematic, people likely at highest risk during an H7N9 pandemic will probably
have low vaccine response because of older age and prior receipt of seasonal influenza vaccine.

NEJM Journal Watch Infectious Diseases, published July 28, 2015
88
Toward Effective Vaccines for Avian Influenza

Two studies provide clarity on the feasibility of avian influenza vaccines.
Mulligan MJ et al. Serological responses to an avian influenza A/H7N9 vaccine mixed at the point-of-use with MF59
adjuvant: A randomized clinical trial. JAMA 2014 Oct 8; 312:1409. (http://dx.doi.org/10.1001/jama.2014.12854)
Belshe RB et al. Immunogenicity of avian influenza A/Anhui/01/2005(H5N1) vaccine with MF59 adjuvant: A random-
ized clinical trial. JAMA 2014 Oct 8; 312:1420. (http://dx.doi.org/10.1001/jama.2014.12609)
Treanor JJ. Expanding the options for confronting pandemic influenza. JAMA 2014 Oct 8; 312:1401. (http://dx.doi.org/
10.1001/jama.2014.12558)
ARTICLE SUMMARY
Although world attention is focused on Ebola, concern persists over future influenza pandemics caused by the
novel avian influenza A H7N9 and H5N1 strains recently found in Asia. The development of effective vaccines
against these strains would be of great benefit, but progress has been limited by relatively low immunogenicity.
Two groups report new strategies to enhance the effectiveness of such vaccines.
In a randomized, controlled trial involving 700 adults, Mulligan and colleagues assessed the safety and efficacy
of various doses of a split virus inactivated monovalent H7N9 vaccine preparation — with or without the oil-in-
water MF59 adjuvant — administered intramuscularly on days 0 and 21. Although almost no participants had
seroconverted by day 21, even when given doses containing 45 μg of hemagglutinin, those who received any adju-
vanted priming dose from 3.75 to 15 μg of hemagglutinin had high seroconversion rates 21 days after the second
dose. Both older age and receipt of seasonal influenza vaccine in the 2012–2013 or 2013–2014 season were asso
ciated with lower seroconversion rates. Adjuvant use was associated with a higher likelihood of local reactions.
In a similar controlled trial, Belshe and colleagues studied the safety and efficacy of various doses of a monovalent
inactivated surface antigen H5N1 influenza vaccine, with or without the MF59 adjuvant. A total of 565 vaccine-
naive adults received immunizations on days 0 and 28; in addition, 72 individuals who had received one or two
doses of an older H5N1 influenza vaccine 1 year earlier received a single dose of the new vaccine. By day 28 after
immunization, 21% to 50% of the previously vaccinated participants had high hemagglutinin inhibition assay
antibody titers, regardless of adjuvant use; the neutralizing antibody response was similarly high. Two doses of
the older vaccine seemed to offer no advantage over a single dose. In the vaccine-naive population, there was no
significant response to a single immunization, but a demonstrable dose response occurred following the second,
with or without adjuvant use. In all groups, the antibody response lasted <180 days. The vaccine was generally
well tolerated, but among participants who received adjuvanted vaccine, systemic events (primarily mild malaise)
were more common in those who had previously been immunized.
COMMENT
As noted by an editorialist, these complementary approaches provide hope that effective avian influenza vaccines
can be developed. However, the need for multiple immunizations and the use of adjuvants, as well as the apparent
effect of previous seasonal influenza immunizations, are substantial practical and logistical hurdles.

NEJM Journal Watch Infectious Diseases, published October 15, 2014
89
Avian Influenza Viruses in U.S. Birds — No Surprise

Three such pathogens were recently isolated from domestic and wild birds in the northwestern U.S.
Jhung MA and Nelson DI. Outbreaks of avian influenza A (H5N2), (H5N8), and (H5N1) among birds — United States,
December 2014–January 2015. MMWR Morb Mortal Wkly Rep 2015 Feb 6; 64:111. (http://www.cdc.gov/mmwr/
preview/mmwrhtml/mm6404a9.htm?s_cid=mm6404a9_w)
ARTICLE SUMMARY
Highly pathogenic avian influenza (HPAI) viruses, which can cause severe — sometimes lethal — disease
in humans, were recognized in Hong Kong in 1997 and spread rapidly to other Asian countries. The animal
reservoirs for these viruses include several species of wild and domestic birds, including migratory sea birds.
Now, infections with these viruses are emerging in U.S. birds. Between December 15, 2014, and January 16, 2015,
there were 14 reports of HPAI H5 isolation — 7 H5N2, 6 H5N8, and 1 H5N1 — from domestic and wild birds in
five northwestern states (California, Idaho, Oregon, Utah, and Washington). All domestic birds with known in-
fection were destroyed. Despite 24 reported human exposures, only one person developed influenza-like illness,
and that person tested negative for influenza.
COMMENT
Given the migratory bird reservoir of these infectious agents, it was only a matter of time until HPAI viruses ap-
peared in North America, and the West Coast was the most probable initial site. Although bird-to-human trans-
mission did not occur in these instances, it is highly likely to happen at some point, as has been seen in Asia. What
is not quite so predictable is the most serious of possible scenarios, human-to-human transmission. Fortunately,
such transmission has been rare in Asia. Practitioners encountering severe influenza cases should inquire about
recent bird exposure and notify the CDC of any suspected cases of avian influenza.
— Stephen G. Baum, MD
Albert Einstein College of Medicine, Bronx, New York
90
Songbirds and Parakeets: Possible Sources of Influenza A(H7N9) Virus

Transmission
Influenza A(H7N9) viruses from humans can replicate in songbirds and parakeets, be shed into
the environment, and infect other birds.
Jones JC et al. Possible role of songbirds and parakeets in transmission of influenza A(H7N9) virus to humans.
Emerg Infect Dis 2014 Feb; [e-pub ahead of print]. (http://dx.doi.org/10.3201/eid2003.131271)
ARTICLE SUMMARY
Avian-origin influenza A(H7N9) virus continues to cause sporadic human infections in China; older men have
experienced disproportionate rates of illness and death. To assess whether wild and domesticated small birds can
be infected with and transmit this pathogen, researchers inoculated finches, sparrows, and parakeets with fluid
containing H7N9 virus originating from a patient in China. A sample of birds tested preinoculation were influenza
antibody negative.
All inoculated birds shed virus via the oropharynx. Most showed no clinical signs of disease, although two died;
all surviving birds cleared the virus by day 8 postinoculation. Virus was detected on multiple days in the com
munal water trough shared by the birds. Among inoculated birds, all society finches and sparrows, 75% of zebra
finches, and 80% of parakeets seroconverted. Among birds that had direct contact with the inoculated birds,
a smaller proportion seroconverted, and some shed virus.
COMMENT
This study confirms that songbirds and parakeets are susceptible to influenza H7N9, contaminate water, and can
transmit the virus to uninfected birds. The authors speculate that these birds could transmit virus to domesticated
poultry or to other wild birds if they share common water sources.
The authors comment that keeping pet birds in China is associated with luck and is common among older men,
who often carry caged birds with them on strolls. They also describe religious ceremonies that involve purchase
of a songbird that is held to the face, kissed, and released. They conclude that such birds may be intermediate hosts
and sources of virus for wild birds, domestic poultry, and humans. Future studies should focus on the role of
songbirds and parakeets in the ecology, maintenance, and transmission of H7N9 virus.
— Mary E. Wilson, MD
Harvard T.H. Chan School of Public Health, Boston
91
PANDEMIC INFLUENZA
The dire effects of the Spanish influenza pandemic of 1918–1919, responsible for the deaths of an estimated
50 million to 100 million people, continue to echo in the collective consciousness. Pandemic influenza occurs
when a new viral strain is passed from another animal species to humans, who have no immunity to the
new virus. Many discussions of the Ebola outbreak of 2014–2015 refer to lessons to be learned by the global
community in how to prepare for and respond to epidemic disease.
In this section, we summarize the findings of a model created to examine the effects of vaccination rates on a flu
pandemic. A Global Health Review article looks at lessons to be learned from the H1N1 outbreak of 2009, and a
Perspective article from Bill Gates includes a global call to action and recommendations for preparing for future
epidemics.
92
Early Interventions to Truncate Influenza Pandemics

Earlier use of vaccines and nonpharmaceutical interventions would save lives and money in a flu
pandemic.
Khazeni N et al. Health and economic benefits of early vaccination and nonpharmaceutical interventions for a human
Influenza A (H7N9) pandemic. Ann Intern Med 2014 May 20; 160:684. (http://dx.doi.org/10.7326/M13-2071)
ARTICLE SUMMARY
The emergence of new influenza strains with high case-fatality rates and the bioengineering of strains with in-
creased potential for person-to-person spread have increased concern about our preparedness for an influenza
pandemic.
During the 2009 influenza A(H1N1) pandemic, there was a 9-month lag in instituting mass vaccination. Now,
researchers have created a mathematical model to assess the value of earlier vaccination. The model assumed a
population of 8.3 million people with no immunity from prior exposure to the influenza A(H7N9) strain; an
urban setting resembling New York City; a 30% vaccination rate; 1000 people infected at the start of the pandemic,
of whom 67% were symptomatic and 50% were isolated; a reproductive number (the number of secondary
infections caused by each infected person) of two, similar to that found in the 1918 pandemic; and a 2.5%
case-fatality rate.
The model suggested that without vaccination, 1.92 million persons would be symptomatically infected and
48,254 would die. Completion of vaccination at 9 months would avert 85,930 infections and 2365 deaths. Comple-
tion at 6 months instead of 9 months would prevent an additional 230,321 cases and 5775 deaths. Completion at
4 months would yield the most substantial preventive effect, with 225,992 fewer cases and 5633 fewer deaths than
in the 6-month model. Treatment costs would be saved proportionately. Adherence to nonpharmaceutical inter-
ventions such as hand washing, sneeze etiquette, face masks, and self-isolation if ill would have a substantial effect
on mortality and spread (similar to vaccinating 30% of the population within 4 months) and are seen as crucial
measures if vaccine development or distribution is delayed.
COMMENT
Like all models, this one has significant limitations and is based on many assumptions that may not be borne out.
For example, the assumed mortality rate of 2.5% is far lower than that in the 2009 pandemic. Nonetheless, this
model does support efforts to accelerate vaccine production. The discussion of increased long-term medical costs
for people who are spared by vaccination probably will not significantly affect decision-making about saving lives.
— Stephen G. Baum, MD
Albert Einstein College of Medicine, Bronx, New York
NEJM Journal Watch Infectious Diseases, published May 20, 2014
93
review article
global health
Pandemic Preparedness and Response —

Lessons from the H1N1 Influenza of 2009
Harvey V. Fineberg, M.D., Ph.D.
A
number of viruses have pandemic potential. For example, the From the Institute of Medicine, Washing-
coronavirus responsible for the severe acute respiratory syndrome (SARS), ton, DC. Address reprint requests to Dr.
Fineberg at the Institute of Medicine, 500
which first appeared in southern China in November 2002, caused 8096 Fifth St. NW, Washington, DC 20001-2721,
cases and 774 deaths in 26 countries before coming to a halt by July 2003 mainly or at fineberg@nas.edu.
owing to isolation and quarantine.1 In terms of persistence, versatility, potential
This article was updated on December
severity, and speed of spread, however, few viruses rival influenza virus. Endemic 22, 2014, at NEJM.org.
in a number of species, including humans, birds, and pigs, influenza virus causes
N Engl J Med 2014;370:1335-42.
annual outbreaks punctuated by occasional worldwide pandemics, which are char- DOI: 10.1056/NEJMra1208802
acterized by sustained community spread in multiple regions of the world. Copyright © 2014 Massachusetts Medical Society.
Beyond spread, the degree to which a pandemic is defined according to the

severity of the disease, or whether it may be simply described as often producing
many illnesses and deaths, remains ambiguous.2 At its worst, pandemic influenza
can be catastrophic: the great influenza pandemic of 1918–1919 is estimated to
have infected 500 million persons worldwide and to have killed 50 to 100 million
persons.3 In a typical year of seasonal outbreaks in the Northern and Southern
Hemispheres, influenza virus causes as many as 5 million cases of severe illness
in humans and 500,000 deaths.4
Over the past decade, sporadic cases of severe influenza and deaths in humans
have been caused by a number of avian influenza A viruses, including the H5N1
virus, first detected in 1997, and the H7N9 and H10N8 viruses, first reported in
2013. Such sporadic cases may be harbingers of a gathering pandemic, but the
likelihood is difficult to judge because it is not known how frequently similar zoo-
notic episodes occurred silently in the past, when surveillance was more limited,
and did not cause pandemics.
The most recent global pandemic was caused by the influenza A (H1N1) strain,
which was first detected in North America in 2009 (influenza A[H1N1]pdm09).
This event prompted the first activation of provisions under the 2005 Interna-
tional Health Regulations (IHR), which went into effect in 2007.5 Deliberations
that led to the 2005 IHR revisions were shaped by experience in the SARS outbreak
of 2003. The regulations delineate the responsibilities of individual countries and
the leadership role of the World Health Organization (WHO) in declaring and
managing a public health emergency of international concern.
The 2009 H1N1 pandemic presented a public health emergency of uncertain
scope, duration, and effect. The experience exposed strengths of the newly imple-
mented IHR as well as a number of deficiencies and defects, including vulnerabil-
ities in global, national, and local public health capacities; limitations of scien-
tific knowledge; difficulties in decision making under conditions of uncertainty;
complexities in international cooperation; and challenges in communication
among experts, policymakers, and the public.
n engl j med 370;14 nejm.org 94

april 3, 2014 1335
At the request of the WHO, an international (Fig. 1). An interactive graphic showing the time-
committee, which I chaired, reviewed the experi- line of the 2009 H1N1 pandemic is available with
ence of the pandemic, with special attention the full text of this article at NEJM.org.
An interactive
timeline of the given to the function of the 2005 IHR and the Evidence from the first outbreak in Mexico
2009 H1N1 performance of the WHO.6 Since this was the was alarming. An observational study of 899
influenza first time that the 2005 IHR was tested in a real- hospitalized patients showed that 58 (6.5%) be-
pandemic
world situation, it was inevitable that aspects of came critically ill, and of those, 41% died.7 Dur-
is available
at NEJM.org the response to the series of outbreaks and sub- ing the course of the pandemic, mortality among
sequent pandemic could have been improved. children, young adults, and pregnant women
Even though there were areas of outstanding was much higher than in a typical influenza
performance, such as the timely identification of season, and there was substantial variation in
the pathogen, the development of sensitive and severity among different regions of the world.8
specific diagnostics, and the creation of highly In general, older adults fared relatively well, and
interactive networks of public health officials, the the total number of influenza-related deaths
most fundamental conclusion of the committee, worldwide (estimated ranges of 123,000 to
which applies today, is not reassuring: “The world 203,000 deaths8 and 105,700 to 395,600 deaths9)
is ill prepared to respond to a severe influenza proved similar to the number in a relatively mild
pandemic or to any similarly global, sustained year of seasonal influenza. However, because of
and threatening public-health emergency.” 6 the proportionately higher mortality among chil-
In this article, I focus on lessons from the dren and young adults, the severity in terms of
global response to the 2009 H1N1 pandemic. I years of life lost was greater than in a typical
identify some of the key successes and short- year of seasonal influenza.10
comings in the global response, on the basis of
the findings and conclusions of the review com- 2005 INTERNATIONAL HEALTH REGULATIONS
mittee. The article concludes by pointing to steps A number of provisions of the 2005 IHR proved
that can improve global readiness to deal with helpful in dealing with the 2009 H1N1 pandemic.
future pandemics. For example, the 2005 IHR established system-
atic approaches to surveillance, early-warning
systems, and response in member states and pro-
T IME C OUR SE OF THE 20 0 9 H1N1
PA NDEMIC moted technical cooperation and sharing of logis-
tic support. Communication among countries and
The first laboratory-confirmed cases of H1N1 the WHO was strengthened by the establishment
influenza appeared in Mexico in February and in each member state of National Focal Points
March of 2009. Cases that were detected in Cali- — national offices that would be responsible for
fornia in late March were laboratory-confirmed rapid collection and dissemination of emerging
by mid-April. By the end of April, cases had been data and guidance.
reported in a number of U.S. states and in coun- A static and potentially outdated list of notifi-
tries on various continents, including Canada, able diseases in previous regulations was replaced
Spain, the United Kingdom, New Zealand, Israel, by a more flexible flow diagram and decision
and Germany. On April 25, invoking its authority tool that identified conditions warranting public
under the 2005 IHR, the WHO declared a public health action. The 2005 IHR required, for the
health emergency of international concern and first time, that member states implementing uni-
convened the emergency committee called for in lateral measures that interfere with international
the regulations. The WHO also established a traffic and trade inform the WHO and that they
dedicated internal group to coordinate the re- also provide a public health rationale and scien-
sponse to the widening outbreaks. As of June 9, tific justification for those measures. Most im-
2009, a total of 73 countries had reported more portant, the 2005 IHR formally assigned to the
than 26,000 laboratory-confirmed cases, and the WHO the authority to declare a public health
WHO declared on June 11 that the situation met emergency of international concern and take a
the criteria for phase 6 — that is, a full-fledged leading role in the global response.
pandemic (Table 1). By the time the pandemic Despite these positive features, many member
had waned, in August 2010, virtually all coun- states did not have in place the capacities called
tries had reported laboratory-confirmed cases for in the IHR, nor were they on a path to meet

95 april 3, 2014
Source: The New England Journal of Medicine global health
Table 1. World Health Organization (WHO) Pandemic-Phase Descriptions and Main Actions According to Phase.
Estimated Probability Main Actions in Affected Main Actions in Nonaffected

Phase of Pandemic Description Countries Countries
1 Uncertain No animal influenza virus circulating Developing and implementing na- Same as in affected countries
among animals has been reported tional pandemic-influenza pre-
to cause infection in humans paredness and response plans
and harmonizing them with
national emergency prepared-
ness and response plans
2 Uncertain An animal influenza virus circulating Same as phase 1 Same as phase 1
in domesticated or wild animals is
known to have caused infection in
humans and is therefore considered
a specific potential pandemic threat
3 Uncertain An animal or human–animal influenza Same as phase 1 Same as phase 1
reassortant virus has caused spo-
radic cases or small clusters of dis-
ease in people but has not resulted
in a level of human-to-human
transmission sufficient to sustain
community-level outbreaks
4 Medium to high Human-to-human transmission of an Rapid containment Readiness for pandemic
animal or human–animal influenza response
reassortant virus that is able to sus-
tain community-level outbreaks
has been verified
5 High to certain The same identified virus has caused Pandemic response: each country Readiness for imminent
sustained community-level out- implements the actions called pandemic response
breaks in at least two countries in for in its national plans
one WHO region
6 Pandemic in In addition to the criteria for phase 5, Same as phase 5 Same as phase 5
progress the same virus has caused sus-
tained community-level outbreaks
in at least one other country in an-
other WHO region
their obligations by the 2012 deadline specified or punitive trade sanctions are called for under
in the document. Of the 194 eligible states, 128 the 2005 IHR.
(66%) responded to a WHO questionnaire on
their state of progress in 2011. Only 58% of the WORLD HEALTH ORGANIZATION
responding member states reported having de- The WHO is an indispensable global resource for
veloped national plans to meet their core capac- leading and coordinating the response to a pan-
ity requirements, and only 10% claimed to have demic. In the 2009 H1N1 pandemic, the WHO
fully established the capacities called for in had many notable achievements. The organization
the IHR.6 provided guidance to inform national influenza-
The IHR fails to specify a basis for virus shar- preparedness plans, which were in place in 74%
ing and vaccine sharing. This has been partially of countries at the time of the first outbreak in
ameliorated in a framework for pandemic-influ- North America, and helped countries monitor
enza preparedness, adopted in 2011, that calls their development of IHR core capacities. The
on member states to encourage vaccine manu- WHO Global Influenza Surveillance Network de-
facturers to set aside a fraction of their pandem- tected, identified, and characterized the virus in
ic-vaccine production for donation and for dis- a timely manner and monitored the course of the
counted pricing in developing countries.11 A pandemic.
glaring gap in the IHR, which has not been Within 48 hours after the activation of provi-
remedied, is its lack of enforceable sanctions. sions in the 2005 IHR, the WHO convened the
For example, if a country fails to explain why it first meeting of the emergency committee of ex-
restricted trade or travel, no financial penalties perts who would advise the WHO on the status
n engl j med 370;14 nejm.org 96

april 3, 2014 1337
A Extent of H1N1 Influenza Worldwide by Late April 2009
Number of cumulative positive H1N1 samples by country
0 50,000–74,999 Not available

1–24,999 75,000–99,999
B Extent of H1N1 Influenza by Late July 2010
25,000–49,999 ≥100,000
Figure 1. H1N1 Influenza Pandemic.

Data are from the World Health Organization and http://fluNet.org.
of the pandemic. Within 32 days after the WHO strains and control reagents were made available
had declared a public health emergency of inter- within a few weeks. The Strategic Advisory Group
national concern, the first candidate reassortant of Experts on immunization at the WHO provided
vaccine viruses were developed, and vaccine seed early recommendations on vaccine target groups
1338 n engl j med

97370;14 nejm.org april 3, 2014
and dose. The WHO provided prompt and valu- may vary according to age and past exposure to
able field assistance to affected countries and influenza viruses. Although a catastrophic pan-
efficiently distributed more than 3 million cours- demic probably depends on the emergence of a
es of antiviral drugs to 72 countries. new antigenic type of influenza virus, it does
Against this backdrop of accomplishment, not follow that every newly emerging influenza
the WHO confronted systemic difficulties and virus will produce an especially severe burden of
made a number of missteps in the course of cop- influenza. For example, in the 40 years between
ing with the unfolding pandemic. Although the the mid-1930s and mid-1970s, the 5 years of
WHO is the only global agency with legitimate greatest excess mortality from influenza in the
authority to lead the response to a pandemic, it is United States were 1937, 1943, 1953, 1957, and
burdened by a number of structural impediments. 1960, but among these years, only 1957 was
First, the WHO is simultaneously the moral voice marked by a new antigenic type (H2N2), and
for health in the world and the servant of its 1968 (the year when H3N2 appeared) did not
member states, which authorize the overall pro- rank in the top five for severity.13 The expecta-
gram and budget. National interests may con- tion of a very severe pandemic was understand-
flict with a mandate to equitably protect the able in the context of H5N1 but not necessarily
health of every person on the planet. Second, for every new antigenic type.
the budget of the WHO is incommensurate with Since the formal criteria for advancing from
the scope of its responsibilities. Only approxi- one phase to the next higher phase in an emerg-
mately one quarter of the budget comes from ing pandemic were based entirely on the extent
member-state assessments, and the rest depends of spread and not on severity, this led to public
on specific project support from countries and confusion about exactly what the WHO meant by
foundations. These budget realities and the per- a pandemic. The WHO lacked a consistent, mea-
sonnel-management requirements inherent in be- surable, and understandable depiction of the
ing a United Nations agency constrain flexibility. severity of a pandemic. This situation was prob-
Third, the WHO is better designed to respond lematic because, regardless of the definition of
to focal, short-term emergencies, such as investi- a pandemic, the decisions about response logi-
gating an outbreak of hemorrhagic fever in sub- cally depend on both spread and severity. In
Saharan Africa, or to manage a multiyear, steady- addition, the defining phase structure that was
state disease-control program than to mount and based on spread was needlessly complex in that
sustain the kind of intensive, global response it defined more stages than there were differen-
that is required to deal with a rapidly unfolding tiated responses, and the structure that seemed
pandemic. Finally, the regional WHO offices are suitable for planning proved less suited to op-
autonomous, with member states of the region erational management.
responsible for the election of the regional direc- The weekly requests by the WHO for data
tor, budget, and program. Although this system were overwhelming for some countries, particu-
allows for regional variation to suit local condi- larly those with limited epidemiologic and labo-
tions, the arrangement limits the ability of the ratory capacity. As the epidemic progressed, it
WHO to direct a globally coherent and coordi- was not always evident to country officials that
nated response during a global health emergency. the data they submitted were being analyzed and
In anticipation of a possible pandemic before used. Rather than focus on laboratory-confirmed
2009, public health authorities had focused on cases, a surveillance model that relied on syn-
the threat of avian H5N1 influenza, and a signal dromic surveillance and selective, systematic viro-
feature among recognized cases of H5N1 influ- logic testing might have been more revealing.14
enza in humans was mortality exceeding 50%.12 Public health officials in some countries, such
Hence, it was expected that a newly emerging as the U.K. Health Protection Agency, produced
pandemic virus would cause many deaths as well weekly summaries that tracked domestic indica-
as widespread disease, and the WHO said as tors of influenza spread and severity while not-
much on its website on pandemic preparedness ing pertinent global influenza activity, and this
in advance of the 2009 H1N1 pandemic. approach could hold lessons for other countries
The prospects of a pandemic depend on the as well as for the WHO.15
transmissibility and virulence of the virus and When the WHO convened an expert group,
on the susceptibility of the population, which typically for a 1- or 2-day consultation, the prac-

98 april 3, 2014 1339
tice of the organization was not to disclose the media by discontinuing routine press conferences
identities of the experts until the consultation on the pandemic.
was concluded. Similarly, the WHO kept confi- The most serious operational shortcoming,
dential the identities of emergency-committee however, was the failure to distribute enough
members convened under the provisions of the influenza vaccine in a timely way. Ultimately,
IHR, who would advise the WHO on the status 78 million doses of vaccine were sent to 77 coun-
of the emerging pandemic. Although the intent tries, but mainly long after they would have done
was to shield the experts from commercial or the most good. At its root, this reflected a short-
political influences, the effect was to stoke sus- fall in global vaccine-production capacity and
picions about the potential links between indi- technical delays due to reliance on viral egg
vidual members of the emergency committee cultures for production, as well as distributional
and industry.16 Although the review committee problems. Among the latter were variation among
uncovered no evidence of inappropriate influence wealthier countries and manufacturers in their
on the emergency committee, the decision to willingness to donate vaccine, concerns about
keep the members’ identities secret fostered sus- liability, complex negotiations over legal agree-
picions about WHO decision making, which were ments with both manufacturers and recipient
exacerbated by the failure to apply systematic countries, a lack of procedures to bypass national
and open procedures for disclosing, recognizing, regulatory requirements for imported vaccine,
and managing conflicts of interest. A practice of and limited national and local capacities to trans-
confidentiality that was arguably fitting for a port, store, and administer vaccines. Some re-
1-day consultation was ill-suited to an advisory cipient countries thought that the WHO did not
function that extended over a period of months. adequately explain that the liability provisions
The failure to acknowledge legitimate criti- included in their recipient agreements were the
cisms, such as inconsistent descriptions of the same as the provisions accepted by purchasing
meaning of a pandemic and the lack of timely countries.
and open disclosure of potential conflicts of
interest, undermined the ability of the WHO to L O OK ING A HE A D
respond effectively to unfounded criticisms. For
example, the WHO was wrongly accused of In light of these structural impediments and op-
rushing to declare phase 6, or a full-fledged erational deficiencies, the world was very fortu-
pandemic, because such action would trigger nate that the 2009 H1N1 influenza pandemic
vaccine orders sought by manufacturers. This was not more severe. On the basis of its analysis,
kind of suspicion proved hard for the WHO to the review committee offered 15 recommenda-
dispel, despite the fact that the declaration of tions to the WHO and the member states (Table 2).
phase 6 was delayed until the sustained com- The report and recommendations were endorsed
munity spread in multiple countries in multiple by the member states at the 64th World Health
WHO regions was incontrovertible. Assembly in May 2011, and the relevant WHO
The WHO made a number of operational mis- departments incorporated the recommendations
steps, including conferring with only a subset of into their biennial work plans.17 Some recom-
the emergency committee, rather than inviting mendations, such as improved protocols for vac-
input from the full group, at a crucial point of cine sharing, have been carried out, some are
deciding to declare progression from phase 4 to within the power of the WHO to implement, and
phase 5. Throughout the pandemic period, the others depend on the actions and resources of
WHO generated an unmanageable number of the member states, which have yet to be commit-
documents from multiple technical units within ted to this purpose.
the organization and lacked a cohesive, over- Beyond institutional, political, and manage-
arching set of procedures and priorities for pro- rial difficulties, the most fundamental con-
ducing consistent and timely technical guidance. straints on pandemic preparedness are the lim-
In addition, after the declaration of phase 6, a its of scientific understanding and technical
time when public awareness of the evolving pan- capacity. Perhaps because only three or four in-
demic was especially important, the WHO chose fluenza pandemics tend to occur each century,
to diminish proactive communication with the at least in recent centuries, the annals of influ-
1340 n engl j med 370;14 99

nejm.org april 3, 2014
enza are filled with overly confident predictions

Table 2. Recommendations of the WHO Review Committee on the Functioning
based on insufficient evidence.18 Studies de- of the 2005 International Health Regulations (IHR) in Relation to the 2009
signed to select for avian-origin viruses that can H1N1 Influenza Pandemic.
be transmitted more readily than the original
Accelerate the implementation of the core capacities required by the IHR
virus in mammalian species (gain-of-function
studies) may arguably help predict the pandemic Enhance the WHO Event Information Site*
potential of naturally occurring viruses but have Reinforce evidence-based decisions on international travel and trade
raised concerns about the possibilities of inten- Ensure necessary authority and resources for all National Focal Points†
tional misuse and unintended consequences.19,20 Strengthen the internal capacity of the WHO for sustained response
In the current state of scientific knowledge, Improve practices for the appointment of an emergency committee
however, no one can predict with confidence
Revise pandemic-preparedness guidance
which influenza virus will become dangerous to
Develop and apply measures to assess the severity of a pandemic
human health and to what degree. The only way,
potentially, to reduce this uncertainty is through a Streamline the management of guidance documents
deeper biologic and epidemiologic understanding. Develop and implement a strategic, organization-wide communications policy
Disease detection, surveillance, and labora- Encourage advance agreements for vaccine distribution and delivery
tory capacity are improving in many countries. Establish a more extensive public health reserve workforce globally
The new techniques of Web-based field reports Create a contingency fund for public health emergencies
and analysis of Web-based search patterns can
Reach an agreement on the sharing of viruses, access to vaccines, and other
yield valuable intelligence that can give the world benefits
a head start on the next emerging pandemic.21
Pursue a comprehensive influenza research and evaluation program
In addition to superior surveillance and agree-
ments on virus and vaccine sharing, the world * The Event Information Site is a WHO website that, in the event of a pandemic,
needs better antiviral agents and more effective would serve as an authoritative resource to disseminate reliable, up-to-date,
influenza vaccines, greater production capacity, and readily accessible information related to the pandemic.
† National Focal Points are national offices that are responsible for the rapid
and faster throughput. One comprehensive as- collection and dissemination of emerging data and guidance.
sessment showed that the effectiveness of cur-
rent influenza vaccines in practice is lower than
is typically asserted, especially among elderly Pandemics will challenge national authorities
persons.22 The traditional methods of influenza- and the WHO to function more efficiently and
vaccine production, which rely on egg cultures, effectively with insufficient resources. Prepara-
are often too slow to keep up with a first wave tion beyond planning, with advance protocols and
of pandemic spread, and in total, the annual agreements, the commitment of ready reserves
capacity of influenza-vaccine production covers of public health experts and a financial line of
less than one third of the global population. credit, and the fulfillment of the IHR require-
In early 2013, the Food and Drug Administra- ments can all help. Whenever the next influenza
tion approved the first trivalent influenza vac- pandemic arises, many more lives may be at risk.
cine produced with the use of recombinant By heeding the lessons from the 2009 H1N1
technology,23 and other production methods are pandemic, the international community will be
under active research and development. At least able to cope more successfully the next time.
four lower-income countries have their own
The views expressed in this article are those of the author and
influenza-vaccine manufacturing facilities, and do not necessarily represent the views of the Institute of Medicine.
more are on the way. Most important, if research Disclosure forms provided by the author are available with the
could yield a universal (non–strain-specific), full text of this article at NEJM.org.
Members of the World Health Organization committee for the
long-lasting, safe, and effective vaccine against review of the 2009 H1N1 pandemic and 2005 International Health
influenza, the annual frenzy of action against Regulations, on whose work this article is largely based, include
influenza would be transformed into a proac- Preben Aavitsland (Norway), Tjandra Y. Aditama (Indonesia), Sil-
via Bino (Albania), Eduardo Hage Carmo (Brazil), Martin Cetron
tive, long-term prevention program.24,25 (United States), Omar El Menzhi (Morocco), Yuri Fedorov (Russia),
In the meantime, influenza outbreaks and Andrew Forsyth (New Zealand), Claudia Gonzalez (Chile), Moham-
pandemics will continue to challenge policy- mad Mehdi Gouya (Iran), Amr Mohamed Kandeel (Egypt), Arlene
King (Canada), Abdulsalami Nasidi (Nigeria), Paul Odehouri-
makers and public health leaders to make deci- Koudou (Ivory Coast), Nobuhiko Okabe (Japan), Mahmudur
sions under conditions of stress and uncertainty. Rahman (Bangladesh), Palliri Ravindran (India), José Ignacio
n engl j med 370;14 nejm.org100

april 3, 2014 1341
global health
Santos (Mexico), Palanitina Tupuimatagi Toelupe (Samoa), Patri- Chamberland, Nadia Day, Alice Ghent, Sue Horsfall, Janet Kin-
cia Ann Troop (United Kingdom), Kumnuan Ungchusak (Thai- caid, Phillip Lambach, Linda Larsson, Fabienne Maertens, Joan
land), Kuku Voyi (South Africa), Yu Wang (China), and Sam Ntabadde, Les Olson, Magdalena Rabini, Sarah Ramsay, Mick
Zaramba (Uganda). The committee secretariat was led by Nick Reid, Chastine Rodriguez, Alexandra Rosado-Miguel, and Na-
Drager and included Dominique Metais, Faith McLellan, Mary tasha Shapovalova.
REFERENCES
1. World Health Organization. Summa- 687-95. [Erratum, Lancet Infect Dis 2012; zation Department of Global Capacities,
ry of probable SARS cases with onset of 12:655.] Alert and Response. World Health Orga-
illness from 1 November 2002 to 31 July 10. Viboud C, Miller M, Olson D, Oster- nization perspective on implementation of
2003 (based on data as of the 31 Decem- holm M, Simonsen L. Preliminary esti- International Health Regulations. Emerg
ber 2003) (http://www.who.int/csr/sars/ mates of mortality and years of life lost Infect Dis 2012;18:1041-6.
country/table2004_04_21/en). associated with the 2009 A/H1N1 pan- 18. Neustadt RE, Fineberg HV. The epi-
2. Doshi P. The elusive definition of pandemic in the US and comparison with demic that never was: policy-making and
demic influenza. Bull World Health Or- past influenza seasons. PLoS Curr 2010; the swine flu scare. New York: Vintage
gan 2011;89:532-8. 2:RRN1153. Books, 1983.
3. Taubenberger JK, Morens DM. 1918 11. Pandemic influenza preparedness 19. Herfst S, Schrauwen EJA, Linster M,
Influenza: the mother of all pandemics. framework for the sharing of influenza et al. Airborne transmission of influenza
Emerg Infect Dis 2006;12:15-22. viruses and access to vaccines and other A/H5N1 virus between ferrets. Science
4. Lozano R, Naghavi M, Foreman K, et al. benefits (http://www.ip-watch.org/weblog/ 2012;336:1534-41.
Global and regional mortality from 235 wp-content/uploads/2011/04/PIP-Framework 20. Imai M, Watanabe T, Hatta M, et al.
causes of death for 20 age groups in 1990 -16-April_2011.pdf). Experimental adaptation of an influenza
and 2010: a systematic analysis for the 12. World Health Organization. Cumulative H5 HA confers respiratory droplet trans-
Global Burden of Disease Study 2010. number of confirmed human cases for mission to a reassortant H5 HA/H1N1
Lancet 2012;380:2095-128. avian influenza A(H5N1) reported to WHO, virus in ferrets. Nature 2012;486:420-8.
5. International health regulations (2005). 2003-2013 (http://www.who.int/influenza/ 21. Brownstein JS, Freifeld CC, Madoff
2nd ed. Geneva: World Health Organiza- human_animal_interface/EN_GIP_2013 LC. Digital disease detection — harness-
tion, 2008. 1008CumulativeNumberH5N1cases.pdf). ing the Web for public health surveil-
6. Implementation of the International 13. Dowdle WR. Influenza: epidemic pat- lance. N Engl J Med 2009;360:2153-5,
Health Regulations (2005): report of the terns and antigenic variation. In: Selby P, 2157.
Review Committee on the Functioning ed. Influenza: virus, vaccine and strategy. 22. Osterholm MT, Kelley NS, Sommer A,
of the International Health Regulations New York: Academic Press, 1976:17-21. Belongia EA. Efficacy and effectiveness of
(2005) in relation to pandemic (H1N1) 14. Lipsitch M, Hayden FG, Cowling BJ, influenza vaccines: a systematic review
2009. Geneva: World Health Organiza- Leung GM. How to maintain surveillance and meta-analysis. Lancet Infect Dis 2012;
tion, May 5, 2011 (http://apps.who.int/gb/ for novel influenza A H1N1 when there 12:36-44. [Erratum, Lancet Infect Dis
ebwha/pdf_files/WHA64/A64_10-en.pdf). are too many cases to count. Lancet 2012;12:655.]
7. Domínguez-Cherit G, Lapinsky SE, 2009;374:1209-11. 23. FDA approves new seasonal influenza
Macias AE, et al. Critically Ill patients 15. Weekly epidemiological updates archive vaccine made using novel technology. Press
with 2009 influenza A(H1N1) in Mexico. (http://www.hpa.org.uk/Topics/Infectious release of the Food and Drug Administra-
JAMA 2009;302:1880-7. Diseases/InfectionsAZ/PandemicInfluenza/ tion, Bethesda, MD, January 16, 2013 (http://
8. Simonsen L, Spreeuwenberg P, Lustig H1N1PandemicArchive/SIEpidemiological www.fda.gov/NewsEvents/Newsroom/
R, et al. Global mortality estimates for Data/SIEpidemiologicalReportsArchive/ PressAnnouncements/ucm335891.htm).
the 2009 influenza pandemic from the influswarchiveweeklyepireports). 24. Treanor J. Influenza vaccine — out-
GLaMOR project: a modeling study. PLoS 16. Flynn P. The handling of the H1N1 pan- maneuvering antigenic shift and drift.
Med 2013;10(11):e1001558. demic: more transparency needed. Council N Engl J Med 2004;350:218-20.
9. Dawood FS, Iuliano AD, Reed C, et al. of Europe Parliamentary Assembly, 2010 25. Kanekiyo M, Wei C-J, Yassine HM, et al.
Estimated global mortality associated with (http://assembly.coe.int/CommitteeDocs/ Self-assembling influenza nanoparticle
the first 12 months of 2009 pandemic in- 2010/20100329_MemorandumPandemie vaccines elicit broadly neutralizing H1N1
fluenza A H1N1 virus circulation: a mod- _E.pdf). antibodies. Nature 2013;499:102-6.
elling study. Lancet Infect Dis 2012;12: 17. Hardiman MC, World Health Organi- Copyright © 2014 Massachusetts Medical Society.
receive the journal’s table of contents each week by e-mail
To receive the table of contents of the Journal by e-mail

every Wednesday evening, sign up at NEJM.org.
1342 n engl j med 370;14 101

nejm.org april 3, 2014

Perspective
Perspective april 9, 2015
april 9, 2015
The Next Epidemic — Lessons from Ebola

Bill Gates
The Next Epidemic — Lessons from Ebola
P
Bill Gates
erhaps the only good news from the tragic Ebola though the system is not perfect,
epidemic in Guinea, Sierra Leone, and Liberia is NATO countries participate in joint
P
that it maythe
erhaps
for epidemic
only
serve as good
in Guinea,
future epidemics
news call:
a wake-up fromwe
Sierra Leone,
of diseases
themust
that and
tragic
mayLiberia
that it may serve as a wake-up call: we must prepare
Ebola
prepare
spread is
exercisesthe
though
logistics
NATO
in system
which is
such as
countries
food will in
exercises
guage they
bewhich
theynotwork
how fuel
participate
provided,
they what
willasspeak,
out
perfect,
and
in joint
work lan-out
logistics such how and fuel what
and
for
morefuture epidemics
effectively of diseases
than Ebola. There that maydirect
than requiring spread
contact. radio will
food frequencies
be provided,will what
be used.lan-
is a significant chance that an People may not even be aware Few, if they
guage any, will
suchspeak,
measures andarewhat in
epidemic
more of a substantially
effectively than Ebola. There more that they
than are infected
requiring direct orcontact.
infec- place for
radio response towill
frequencies an epidemic.
be used.
infectious disease
is a significant chance that anwill occur tious. Since
People may a notpersoneven carrying one
be aware The world
Few, if any,does
such notmeasures
fund anyare orga-in
sometimeofina the
epidemic next 20 years;
substantially more of these
that they are pathogens
infectedcan infect
or infec- nization
place for to managetothe
response broad set
an epidemic.
after all, we saw
infectious disease willmajor epidemics
occur many Since
tious. strangers in a carrying
a person marketplace
one of coordinated
The world does not activities
fund any required
orga-
during the in20th
sometime the century,
next 20 includ-
years; or on an airplane, the
of these pathogens can infect number of in an epidemic.
nization to manage Thethelast serious
broad set
ing the Spanish influenza
after all, we saw major epidemics epidem- cases can escalate very
many strangers in a marketplace quickly. simulation
of coordinatedof anactivities
epidemic in the
required
ic of 1918–1919
during the 20th and the ongoing
century, includ- As an
or on theairplane,
Ebola epidemic
the number fades
of United
in States, the
an epidemic. TheDark Winter
last serious
pandemic of human
ing the Spanish influenza immunode-
epidem- cases can escalate very quickly. we
from the world’s attention, exercise, took
simulation of anplace in 2001.in And
epidemic the
ficiency
ic virus. Inand
of 1918–1919 fact,
theofongoing
all the riskAsmissing
the Ebola the epidemic
opportunity to
fades few countries
United States,havethemetDarktheirWinter
com-
things that could kill
pandemic of human immunode- more than learn from
from it. Even attention,
the world’s if the systemwe mitmentstook
exercise, under theinInternational
place 2001. And
10 million people around
ficiency virus. In fact, of all the the we have
risk missingtodaythehad worked per-
opportunity to Health
few Regulations,
countries have metwhich were
their com-
world, the most likely
things that could kill more than is an epi- fectly from
learn for Ebola,
it. Evenit would fail to
if the system adopted by
mitments the the
under United Nations
International
demic
10 stemming
million peoplefromaround
either nat-
the contain
we havea today
more infectious
had worked disease.
per- after theRegulations,
Health 2002–2003 which outbreak wereof
ural causes or bioterrorism.
world, the most likely is an epi- It’s for
fectly instructive
Ebola, ittowouldcomparefailour
to the severebyacute
adopted the respiratory
United Nations syn-
Ebola
demic is far from
stemming from theeither
mostnat-in- preparations
contain a morefor epidemics
infectious with
disease. dromethe
after (SARS) and were
2002–2003 intended
outbreak of
fectious known disease.
ural causes or bioterrorism. Other ourIt’s
preparations
instructive for another sort
to compare our to improve
the the world’s
severe acute abilitysyn-
respiratory to
disease 1
Ebolaagents (measles
is far from and in-
the most in- of global threat
preparations — war. with
for epidemics The prevent(SARS)
drome and contain
and wereoutbreaks.
intended
fluenza, for example)
fectious known disease. Other are far more North
our Atlantic Treaty
preparations Organiza-
for another sort to Because
improve the there was ability
world’s so little to
infectious
disease because
agents they and
(measles can in-be tionglobal
of (NATO) has a—mobile
threat war. unit
The preparation,
prevent the world
and contain lost time
outbreaks. 1
spread through the air,
fluenza, for example) are far more rather that is ready
North Atlanticto deploy
Treaty quickly.
Organiza-Al- in the current
Because epidemic
there was trying
so little to
infectious because they can be tion (NATO) has a mobile unit preparation, the world lost time
n engl j med 372;15 nejm.org april 9, 2015 1381
spread through the air, rather that is ready 102to deploy quickly. Al- in the current epidemic trying to
Downloaded from collections.nejm.org. For personal use only. No other uses without permission. 1381
n engl j med 372;15 nejm.org april 9, 2015
TheRNew
Source: PE England Journal of Medicine
S PE C T IV E The Next Epidemic — Lessons from Ebola
Health Systems and Surveillance linked with national public health

Recommendations for Preparing
for Future Epidemics First, there is a critical need to laboratories to enable robust mon-
reinforce basic public health sys- itoring and response. And the data
The world needs to build a warning and response tems, including primary health derived from such testing need to
system for outbreaks. This system should
care facilities, laboratories, sur- be made public immediately. Many
• be coordinated by a global institution that is
given enough authority and funding to be ef-
veillance systems, and critical laboratories in developing coun-
fective, care facilities, among other com- tries have been financed by the
• enable fast decision making at a global level, ponents. As many commentators polio-eradication campaign, so we
• expand investment in research and develop-
have noted, Ebola has spread will have to determine what capac-
ment and clarify regulatory pathways for de- much faster and more widely in ities will be needed once that cam-
veloping new tools and approaches, countries whose health systems paign is over.
• improve early warning and detection systems, — and especially whose primary
including scalable everyday systems that can care systems — were severely Human and Other Resources
be expanded during an epidemic,
weakened by years of armed con- Once it became clear that a seri-
• involve a reserve corps of trained personnel
and volunteers,
flict and neglect. ous emergency was under way in
Strengthening health care sys- West Africa, many local clini-
• strengthen health systems in low- and
middle-income countries, and tems not only improves our abil- cians should have been recruited,
• incorporate preparedness exercises to identify
ity to deal with epidemics, but it and trained personnel should
the ways in which the response system needs also promotes health more have flowed rapidly into the af-
to improve. broadly. Without a functioning fected countries. That didn’t hap-
health system, it is very hard for pen. Some countries stepped for-
answer basic questions about a country to end the cycle of dis- ward with volunteers within 2 to
combating Ebola. In the next epi- ease and poverty. Health is so 3 months, but they were needed
demic, such delays could result fundamental to development that within days. It was fortunate that
in a global disaster. even if there were no chance of Médecins sans Frontières could
The problem is not the fault another epidemic, building and mobilize volunteers more quickly
of any single institution — it re- improving health systems would than any government.
flects a global failure. The world be a worthwhile — and lifesav- We need trained personnel
needs a global warning and re- ing — investment. The fact that ready to confront and contain an
sponse system for outbreaks. they also bolster our ability to epidemic quickly: incident manag-
(Though the World Health Orga- confront epidemics is all the ers; experts in epidemiology,
nization [WHO] has a Global more reason to invest in them. disease surveillance, and other
Outbreak Alert and Response In addition, there is no sys- relevant fields who can provide
Network, it is severely under- tematic disease-surveillance pro- surge capacity; respected com-
staffed and underfunded.) Such a cess in place today in most poor munity leaders who can lead lo-
system could enable us to man- countries, which is where a natu- cal engagement efforts; and
age not only a naturally occur- rally occurring epidemic seems community workers who speak
ring epidemic, but also one ig- most likely to break out. Even local languages. Ideally, we
nited by a bioterror attack.2 once the Ebola crisis was recog- would have updated lists of such
Although I have not seen a rigor- nized last year, there weren’t re- personnel indicating their avail-
ous estimate of the cost of build- sources to effectively map where ability and capabilities. There
ing such a system, World Bank cases were occurring and in what would also be standby training
projections give a sense of the quantity. centers and an explicit under-
cost of inaction: a worldwide in- We need to invest in better standing regarding compensa-
fluenza epidemic, for example, disease-surveillance and labora- tion and insurance for volun-
would reduce global wealth by an tory-testing capacity, for normal teers. Each country could commit
estimated $3 trillion.3 situations and for epidemics. Rou- to managing a pool of volunteers
I hope the following sketch of tine surveillance systems should and to sending a certain number
what such a warning and re- be designed in such a way that of people with various skills and
sponse system might look like they can detect early signs of an equipment within a week after an
will spark action to prepare for an outbreak beyond their sentinel emergency began, with plans for
epidemic that could have global sites and be quickly scaled up evacuating any who were exposed
consequences (see box). during epidemics. They should be to the epidemic pathogen.
1382 n engl j med 372;15

103 nejm.org april 9, 2015
PE R S PE C T IV E The Next Epidemic — Lessons from Ebola
Transportation and equipment and other key elements was en- Other than watching for symp-
are also key. When an epidemic tered into a digital database that toms, the diagnostic approach
strikes, roads and airports in af- was instantly accessible to the rel- used during the Ebola epidemic
fected areas are overwhelmed by evant organizations and agencies. has involved sending blood sam-
people trying to get out. Volun- The groups working on the Ebola ples for quantitative polymerase-
teers will be more likely to sign data — including the WHO, the chain-reaction (qPCR) analysis.
up if they know they will be able U.S. Centers for Disease Control But qPCR machines are expen-
to leave if they get sick or when and Prevention, and others — sive and not widely available, so
their duty is done. Few organiza- could recommend specifications, on average it has taken 1 to 3 days
tions are capable of moving and some combination of founda- to get test results. For the next
thousands of people — some of tions and technology companies epidemic, an adequate number of
them infected — to various loca- could build such a system within qPCR machines should be made
tions around the world at a the year. available while novel diagnostic
week’s notice. The Ebola epidem- Experts will also need com- methods are rapidly developed.
ic might have been much worse puter models to predict what We also need a clear process for
if the U.S. and U.K. governments might happen and which inter- developing and manufacturing
had not used military resources ventions should be prioritized. accurate diagnostic tests rapidly.
to fly people in and out of the af- With access to satellite photogra- A focused effort to accelerate
fected countries. All countries phy and cell-phone data, they this process and establish a rapid
could identify trained military re- could understand the movement approval and procurement pro-
sources that would be available for of populations and individuals in cess would be worthwhile.
epidemics; in a severe epidemic, the affected region. But Internet On the therapeutics front,
the military forces of many or all and cell-phone capacity need to there are drugs that work against
middle- and high-income coun- be improved. We should be able viruses similar to Ebola, and some
tries might have to work together. to use cell-phone systems to con- of them have been shown in test
During severe epidemics, re- tact the public and to poll people assays to have an effect against
sponders also need tents, porta- about what they are seeing and Ebola. Unfortunately, they were
ble power sources, medical sup- experiencing. Key centers should not tested in patients with Ebola
plies, and other materials. A list have high-bandwidth Internet ca- until after the epidemic had
of the supplies that would be pacity through satellite, and Wi- peaked — in part because there
needed to stop an epidemic af- Fi capacity should be added in was no clear process for approv-
fecting 10 million people — 100 key areas so that digital tools ing a novel trial format or for
times the population affected by can help with reporting data and providing indemnity against le-
the Ebola epidemic — could be coordinating personnel. gal liability. We will need to de-
developed, and experts could develop a clear set of guidelines
termine which items would need Medical and Public Health Tools (and testing and regulatory path-
to be stockpiled or be subject to It should be possible to make di- ways) for determining whether
commandeering. agnostic tests, drugs, and vaccine existing drugs could be repur-
It is also critically important platforms that could be adapted posed to help stop a particular
to have good data about what’s for use against various pathogens. epidemic.
going on. Unfortunately, during Today, with the possible exception We also need to invest in more
the Ebola epidemic, the case da- of influenza vaccines, we do not research on antiviral drugs, anti-
tabase has not always been accu- have nearly enough capacity for body treatments, and RNA-based
rate or up to date — partly be- developing adaptable platforms, constructs. We should have either
cause of the chaotic situation, partly because there are opportu- stockpiles or manufacturing ca-
but also because good technolo- nity costs for private-sector orga- pacity for therapies that might be
gy and training have not been nizations in shifting resources effective in an epidemic.
available and there are no clear away from more commercially Plasmapheresis should have
rules regarding making data ac- viable projects to work on tools been used in the Ebola epidemic,
cessible. For future epidemics, it for epidemics that may not hap- but its application wasn’t ap-
should be possible to have a sys- pen. We may need an interna- proved and scaled up until it was
tem in which information on sus- tional funding system that fac- too late for this intervention to
pected cases, locations, survivors, tors in these opportunity costs. have a large impact. Plasmaphe-
n engl j med 372;15

104nejm.org april 9, 2015 1383
PER S PE C T IV E The Next Epidemic — Lessons from Ebola
resis is quite effective for a num- ing an outbreak, a vaccine could empowered and funded to coordi-
ber of diseases (including small- be designed, tested for safety, nate the system. The United Na-
pox and viral hemorrhagic fevers and ready for manufacture at tions and the WHO are studying
such as Lassa fever) and has a scale within a few months. There the lessons from the Ebola epi-
reasonable chance of working for is no guarantee of success, but I demic and ways to improve inter-
Ebola as well. The Gates Founda- believe that given enough time and national crisis management; these
tion started working to establish resources, such efforts could pro- evaluations can provide a start-
plasmapheresis units in early Sep- duce an invaluable contribution for ing point for discussions of ways
tember 2014 and quickly found epidemics and overall health. to strengthen the WHO’s capaci-
partners ready to take them into Given Ebola’s limited infec- ty and about which parts of the
the affected countries. Unfortu- tiousness in the early stages of process it should lead and which
nately, the effort was hampered the disease, most of the quaran- ones others (including the World
by the lack of a clear process for tine policies that were proposed Bank and the G7 countries) should
approving new approaches. We would have been counterproduc- lead in close coordination. The
should develop rules now to ex- tive. But when a far more infec- conversation should include mili-
pedite drug approvals in future tious agent comes along, quaran- tary alliances such as NATO,
epidemics and establish clear tine may be one of the few tactics which should make epidemic re-
guidelines for approving studies that can reduce its spread in the sponse a priority. The final ar-
and treatments, including experi- early stages of disease. Because rangement should include a re-
mental ones. A global epidemic- democratic countries try to avoid serve corps of experts with the
drug–approval process could avert abridging individuals’ rights to broad range of skills needed in
long delays by indemnifying travel and free assembly, they an epidemic.
companies working on new ap- might be too slow to restrict ac- An epidemic is one of the few
proaches. tivities that help spread disease. catastrophes that could set the
Three different Ebola vaccine Part of the process should in- world back drastically in the next
constructs were being developed clude a plan for effective public few decades. By building a global
in the summer of 2014. Although communications, including coor- warning and response system, we
all were in early stages, this work dination of the messages conveyed can prepare for it and prevent
made us more prepared for Ebola by all the different voices people millions of deaths.
than we would be for an entirely will hear, from governments, to Disclosure forms provided by the author
are available with the full text of this arti-
new pathogen, for which vaccine United Nations agencies, to news cle at NEJM.org.
development could take 2 or more media, to bloggers. Digital com- A more detailed version of this article is
years. Moreover, it is not clear how munication can be used to great provided in the Supplementary Appendix,
available with the full text of this article at
quickly vaccine developers could or advantage, but unless a plan is in NEJM.org.
would move or who should fi- place, it will only spread confusion
From the Bill and Melinda Gates Foundation,
nance the final research and man- and panic faster. Seattle.
ufacturing of a new vaccine. This article was published on March 18,
Among known pathogens, in- A Global Call to Action 2015, at NEJM.org.
fluenza is the one most likely to Despite efforts by the United States 1. Implementation of the International
cause a large epidemic; even sea- and a few other countries, there Health Regulations (2005): report of the Re-
view Committee on the Functioning of the
sonal influenza variants probably are still big holes in the world’s International Health Regulations (2005) in
cause several hundred thousand ability to respond to an epidem- relation to pandemic (H1N1) 2009. Geneva:
excess deaths each year. So it’s ic. Other countries may be more World Health Organization, May 5, 2011
(http://apps.who.int/gb/ebwha/pdf_files/
An audio interview disappointing that likely to step up if they see an WHA64/A64_10-en.pdf).
with Dr. Nicole Lurie we don’t have a vac- overall plan and understand their 2. Myhrvold N. Strategic terrorism: a call to ac-
is available at NEJM.org cine for all influen- role in it. We need a rigorous tion. Lawfare research paper no. 2-2013. July 3,
2013 (http://papers.ssrn.com/sol3/papers
za strains. There is work being study of the cost of building a .cfm?abstract_id=2290382&download=yes).
done toward this goal, but it has global warning and response sys- 3. The World Bank. Pandemic risk and One
garnered nowhere near the re- tem and a plan for contributions Health. October 23, 2013 (http://www
.worldbank.org/en/topic/health/brief/
sources that it deserves. from various countries. pandemic-risk-one-health).
Ideally, vaccine research would Through the United Nations, DOI: 10.1056/NEJMp1502918
be funded in such a way that dur- some global institution could be Copyright © 2015 Massachusetts Medical Society.

105 april 9, 2015
NEJM Group combines gold standard content with new technology to deliver innovative
products and services that advance knowledge, learning, practice, and professional
development. NEJM Group products include the New England Journal of Medicine, NEJM
Journal Watch, and NEJM Knowledge+, and are leaders in providing the information
health care professionals need to improve health care quality and patient outcomes.
Visit https://collections.nejm.org for more Clinical Collections from NEJM Group.
© 2015 Copyright Massachusetts Medical Society. NEJM Group is a division of the Massachusetts Medical Society.
All rights reserved.

Influenza - NEJMGroup - Collection PDF

Uploaded by

Copyright:

Available Formats

You might also like

Influenza - NEJMGroup - Collection PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Influenza - NEJMGroup - Collection PDF

Uploaded by

Copyright:

Available Formats

What You Need to Know

— Richard T. Ellison III, MD

17 Vaccine for Prevention of Mild and Moderate-to-Severe Influenza in Children

28 Efficacy of Influenza Vaccine Against Severe Disease in Children

30 Efficacy of High-Dose versus Standard-Dose Influenza Vaccine in Older Adults

41 High-Dose Influenza Vaccine No More Effective than Standard-Dose in Older Adults

42 Influenza Vaccination of Pregnant Women and Protection of Their Infants

56 2009 H1N1 Influenza and Pregnancy — 5 Years Later

60 Do Patients Benefit from Health Care Providers’ Influenza Vaccination?

DIAGNOSIS AND TREATMENT

65 Most Influenza Infections Are Subclinical

67 Influenza Is More Virulent Among Cancer Patients

68 New Meta-Analysis Clarifies Oseltamivir’s Benefits and Risks

69 Oseltamivir for Influenza, Even After 48 Hours of Symptoms?

70 Effect of Oseltamivir on Viral Shedding, Illness, and Household Transmission of Influenza

71 FDA Approves the Intravenous Influenza Drug Peramivir

88 An Avian Flu Vaccine

89 Toward Effective Vaccines for Avian Influenza

90 Avian Influenza Viruses in U.S. Birds — No Surprise

91 Songbirds and Parakeets: Possible Sources of Influenza A(H7N9) Virus Transmission

102 The Next Epidemic — Lessons from Ebola

ACIP Issues 2015–16 Influenza Vaccination Recommendations

• Trivalent vaccines will include A/California/7/2009 (H1N1)-like virus, an A/Switzerland/9715293/2013

Prevention and Control of Influenza with Vaccines: Recommendations

TABLE. Influenza vaccines — United States, 2015–16 influenza season*

Inactivated influenza vaccine, trivalent (IIV3), standard dose

Inactivated influenza vaccine, cell-culture-based (ccIIV3), standard dose

Inactivated influenza vaccine, trivalent (IIV3), high dose

Recombinant influenza vaccine, trivalent (RIV3), standard dose

Live attenuated influenza vaccine, quadrivalent (LAIV4)

TABLE. (Continued) Influenza vaccines — United States, 2015–16 influenza season*

Vaccine Selection and Timing of Vaccination for No

Vaccine for Prevention of Mild and Moderate-to-Severe

n engl j med 369;26 nejm.org december 26, 2013 2481

2482 n engl j med 369;26 nejm.org

n engl j med 369;26 nejm.org

dren (2.4%) and 128 (5.3%), in the two groups,

respectively, had rt-PCR–confirmed influenza.

2484 n engl j med 369;26 20

Cases Attack Rate Cases Attack Rate

n engl j med 369;26 nejm.org december

2486 n engl j med 369;26 nejm.org

Table 2. Clinical Characteristics of rt-PCR–Confirmed Influenza and Effect on Daily Activities.*

Variable Moderate-to-Severe Disease Mild Disease

QIV Control Relative Risk QIV Control Relative Risk

n engl j med 369;26 nejm.org december

Figure 2. Immunogenicity According to Strain.

2488 n engl j med 369;26 nejm.org

patient visits, hospitalizations, and days missed

n engl j med 369;26 nejm.org december

2490 n engl j med 369;26 nejm.org

receive immediate notification when an article

To be notified by e-mail when Journal articles

n engl j med 369;26 27

Efficacy of Influenza Vaccine Against Severe Disease in Children

HIGH-DOSE INFLUENZA VACCINE IN OLDER ADULTS

Efficacy of High-Dose versus Standard-Dose

IIV3HD IIV3SD Relative Efficacy IIV3HD IIV3SD Relative Efficacy

IIV3HD IIV3SD Relative Efficacy IIV3HD IIV3SD Relative Efficacy

IIV3HD IIV3SD IIV3HD IIV3HD IIV3SD IIV3HD IIV3HD IIV3SD IIV3HD