Efficacy and immunogenicity of influenza vaccine in

HIV-infected children: a randomized, double-blind,

placebo controlled trial
Shabir A. Madhia,b,c, Sylvia Dittmerd, Locadiah Kuwandab,c,
Marietjie Ventera, Haseena Cassimd, Erica Lazarusd, Teena Thomasd,
Afaaf Libertyd, Florette Treurnicha, Clare L. Cutlandb,c,
Adriana Weinberge and Avy Violarid
Downloaded from http://journals.lww.com/aidsonline by BhDMf5ePHKav1zEoum1tQfN4a+kJLhEZgbsIHo4XMi0hCywCX1AWnYQp/IlQrHD3i3D0OdRyi7TvSFl4Cf3VC4/OAVpDDa8KKGKV0Ymy+78= on 07/22/2021

Background: HIV-infected children are at heightened risk for severe influenza illness;
however, there is no study on the efficacy or effectiveness of influenza vaccine in these
children. We evaluated the safety, immunogenicity, and efficacy of nonadjuvanted,
trivalent inactivated influenza vaccine (TIV) against confirmed seasonal influenza virus
illness in HIV-infected children.
Methods: A double-blind, placebo-controlled trial was undertaken in Johannesburg in
2009. Four hundred and ten children were randomized to two doses of TIV or placebo 1
month apart. Nasopharyngeal aspirates obtained at respiratory illness visits were tested
by influenza-specific reverse transcriptase-PCR (RT-PCR). Vaccine immunogenicity
was evaluated by hemagglutinin inhibition (HAI) assay. Influenza isolates were
sequenced and evaluated in maximum likelihood phylogenetic analysis.
Results: Overall, the median age of participants was 23.8 months and their median
CD4% was 33.5. Ninety-two percent of enrolees were on antiretroviral therapy. Among
children receiving both doses of vaccine/placebo, confirmed seasonal influenza illness
occurred in 13 (all H3N2) of 205 TIV recipients and 17 (15 H3N2 and two influenza B)
of 200 placebo recipients with vaccine efficacy of 17.7% (95% confidence interval
<0–62.4%). The proportion of TIV recipients who seroconverted after second dose
against vaccine strains of H1N1, H3N2, and influenza B were 47.5, 50.0, and 40.0%,
compared to 4.7, 11.6, and 0%, respectively among placebo recipients. There were no
TIV-related serious adverse events. Sequence analysis of wild-type H3N2 strains
indicated drift from the H3N2 vaccine strain.
Conclusion: Poor immunogenicity of TIV, coupled with drift of circulating H3N2
wild-type compared to vaccine strain, may explain the lack of efficacy of TIV in
young HIV-infected children. Alternate TIV vaccine schedules or formulations warrant
evaluation for efficacy in HIV-infected children.
ß 2013 Wolters Kluwer Health | Lippincott Williams & Wilkins

AIDS 2013, 27:369–379

Keywords: children, efficacy, HIV, immunogenicity, influenza vaccine,

pneumonia, safety

a
National Institute of Communicable Diseases, Centre for Respiratory and Meningitis Diseases, Sandringham, bDepartment of
Science and Technology/National Research Foundation, Vaccine Preventable Diseases, cMedical Research Council: Respiratory
& Meningeal Pathogens Research Unit, dPerinatal HIV Research Unit, University of the Witwatersrand, Johannesburg, South
Africa, and eDepartment of Pediatrics, Medicine and Pathology, University of Colorado, Anschutz Medical Campus, Aurora,
Colorado, USA.
Correspondence to Shabir A. Madhi, National Institute for Communicable Diseases, National Health Laboratory Service, 1
Modderfontein Road, Sandringham, 2131 Johannesburg, Gauteng, South Africa.
Tel: +27 113866137; e-mail: shabirm@nicd.ac.za
Received: 16 August 2012; revised: 21 September 2012; accepted: 24 September 2012.

DOI:10.1097/QAD.0b013e32835ab5b2

ISSN 0269-9370 Q 2013 Wolters Kluwer Health | Lippincott Williams & Wilkins 369
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
370 AIDS 2013, Vol 27 No 3
Background
Antiretroviral therapy (ART)-naive HIV-infected chil-
dren have an eight-fold greater risk of hospitalization due
to influenza virus-associated pneumonia and trend toward
higher case fatality rates (8.0%) compared to HIV-
uninfected children (2.0%) [1,2]. Additionally, HIV
infection is associated with prolonged shedding of
influenza virus [3]. The Advisory Committee on
Immunization Practices (ACIP) recommend annual
immunization of HIV-infected children with seasonal
trivalent influenza vaccine (TIV) [4], premised primarily
on the safety and immunogenicity of TIV in HIV-
infected children [5], without any supporting efficacy or
effectiveness studies. Annual TIV immunization of HIV-
infected children in South Africa and other settings with a
high prevalence of HIV is uncommon. Reasons for this
include lack of access to TIV in public immunization
programs, limited data on the burden of influenza illness
in such countries, and absence of efficacy or effectiveness
studies in HIV-infected children.
The primary objectives of our study were to evaluate the
efficacy and immunogenicity of seasonal TIV in HIV-
infected children. The study was registered under
ClinicalTrials.gov, number NCT00883012.
Methods
Study design and participants
HIV-infected children aged 6–59 months attending the
Perinatal HIV Research Unit Clinic in Johannesburg,
South Africa were approached for study participation.
Study eligibility criteria included CD4% at least 15 or
receiving ART for at least 3 months prior to enrollment;
parent/guardian contactable for weekly visits/calls
throughout the study; and participant available to attend
illness visits at the clinic. Study exclusion criteria included
contraindication to influenza vaccine; presence of
underlying chronic lung disease requiring treatment in
the past 6 months; contraindication to intramuscular
injections; known existing grade 3 or grade 4 laboratory
or clinical toxicity as per Division of Acquired
Imunonediciency Syndrome (Division of AIDS) toxicity
tables [6]; previous history of receiving influenza vaccine;
and systemic steroid treatment for more than 21 days in
the past month.
Study vaccine
Single-dose 0.5-ml vials of TIV (VAXIGRIP; Sanofi-
Aventis; Lyon, France), per WHO recommended strain
selection for the southern hemisphere for 2009, were
procured commercially. Each dose contained 15 mg each
of split influenza A/Brisbane/59/2007(H1N1), A/Uru-
guay/716/2007(H3N2), and B/Florida/4/2006. Chil-
dren were randomized to receive two doses of TIV
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
(0.5 ml if
36 months or 0.25 ml if younger) or equal
volume of matching placebo (0.9% sterile saline) 1 month
apart as recommended [7]. The study pharmacist
decanted 0.25 or 0.5 ml VAXIGRIP or 0.9% saline for
the TIV and placebo group, respectively into 1 ml
syringes, and dispensed according to the computer-
generated randomization list. The preparations were
macroscopically indistinguishable. Study drug was admi-
nistered intramuscularly by the study doctor or nurse
following randomization.
Participant enrollment and randomization
Study enrollment was between 18 February and 22 May
2009, that is, prior to the anticipated start of the influenza
season in South Africa [8]. Randomization lists were
generated by the study statistician using a computer
random number generator (SAS 9.1; SAS Institute Inc.,
Cary, North Carolina, USA) in random permuted group
blocks of 10 with five in each study arm. Participants and
study personnel, excluding the statistician and pharma-
cist, were blinded to treatment assignment. A subset of
participants stratified by age, equal numbers below and
above 36 months of age, was coenrolled in the
immunogenicity cohort. Study nurses and study doctors
who were blinded to the study vaccine allocation were
involved in participant enrollment, vaccination, and
follow-up.
Assessment of vaccine efficacy
Surveillance for influenza-like illness (ILI) was initiated
from enrollment and continued until the end of the
seasonal influenza season based on National Institute for
Communicable Diseases (NICD) surveillance. The child
follow-up time for seasonal influenza was calculated to
have begun on 27 April and ended on 2 August 2009
[8,9]. Weekly short message text (SMS) reminders or
telephonic contact was made with the child’s caregiver to
enquire about ILI symptoms. ILI was defined as an acute
history (<7 days) of at least two of the following signs and
symptoms: fever more than 38.5oC; myalgia or irrit-
ability; sore throat or pharyngitis; rhinorrhea and/or
cough of less than 14 days of duration. Caregivers were
trained to use digital thermometers for recording axillary
temperature and requested to bring the child to the clinic
within 48 h if there were at least two ILI criteria
symptoms or other intercurrent illness.
Nasopharyngeal aspirates (NPAs) were obtained as
described [1], when the child presented with ILI
symptoms or any respiratory tract infection. NPA samples
were immersed into viral transport medium and
transported on wet ice to the NICD for testing within
12 h. Samples were tested for influenza A and influenza B
by shell vial culture and a qualitative one-step reverse
transcriptase-PCR (RT-PCR) assay. Influenza A isolates
were subtyped for seasonal-H1N1 and H3N2. (CDC
RT-PCR protocol for detection and characterization of
influenza, CDC ref# I-007–05). Influenza identification
 Influenza vaccine in HIV-positive children Madhi et al.
either by shell vial culture and/or real-time RT-PCR was
considered a confirmed illness.
Assessment of vaccine safety and
immunogenicity
Blood samples, obtained prior to the first and 1 month
following the second study vaccine dose, were collected
for immunogenicity testing in a subset. Samples were
centrifuged and serum archived at 70oC until shipping
on dry ice to the Molecular and Virology Clinical
Laboratories at the University of Colorado (Denver,
USA) for hemagglutinin antibody inhibition assay (HAI).
Adverse events were actively solicited telephonically for
3 days postvaccination and parents could also contact
study staff to report adverse events. Severe adverse events
(SAEs) were recorded throughout the study. CD4þ cell
counts and HIV PCR viral load (Roche Amplicor
version 1.5; Roche Diagnostics GmbH, Mannheim,
Germany) within 3 months of randomization were
abstracted. CD4þ lymphocyte counts and percentage
were also measured 1 month after the second dose of TIV
by the panleukogating method [10].
Molecular characterization of the hemagglutinin
gene
The HA1 region of the hemagglutinin gene was amplified
by nested reverse transcription PCR assay using the Titan
One Tube RT-PCR System (Roche Diagnostics GmbH)
and PFx Platinum DNA polymerase (Invitrogen
Corporation, Carlsbad, California, USA) reagents.
PCR products were directly sequenced using the BigDye
V3.1 Terminator reagents (Applied Biosystems, Foster
City, California, USA). Sequence alignments were done
in BioEdit Version 7.0.8.0 and evolutionary analysis and
neighbor-joining phylogenetic trees analysis were done in
MEGA 4 using the Maximum likelihood options with
the Tamura-Nei model and 1000 bootstrap iterations as
well as P-distance analysis [11].
Study sample size and statistical analysis
On the basis of an estimated attack rate of 20% for
confirmed influenza illness in the placebo group and a
power of 80% with a a of 0.05 to detect at least a 50%
reduction among vaccinees, the required sample size
calculated was 420. The sample size was adjusted to 550
for possible withdrawals and/or noncompliance with
study protocol. A sample size of 100 was calculated for the
nested immunogenicity study to demonstrate serocon-
version rates of at least 40% among TIV recipients and less
than 5% in placebo recipients. Enrollment into the
immunogenicity subgroup was stratified to include
50 children aged 6–35 months (younger age group)
and 50 aged 36–60 months (older age group).
The primary efficacy endpoint was reduction of
confirmed seasonal influenza illness. Secondary outcome
measures included differences for ILI and any acute
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
371
respiratory illness. Vaccine efficacy was calculated using
the formula of incidence rates as 1–IL/IP, (IL ¼ the case
incidence rate in TIV-vaccinated children; IP ¼ the case
incidence rate in placebo recipients). The 95% CIs were
constructed and differences between the intervention
arms tested at a equal to 0.05 significance level. A per-
protocol analysis included children who received both
doses of correct study intervention per randomization
and with outcomes occurring at least 14 days following
the second dose. Only the first episode of an outcome
following receipt of study vaccine was included in
the analysis.
Immunogenicity analysis included calculation of geo-
metric means of the titers (GMTs) prior to the first dose
and 21–35 days after the second dose of study vaccine.
Treatment groups were compared with respect to the fold
rise by a two-sided, two-sample t-test and the corre-
sponding 95% confidence intervals for the ratio, all using
logarithmic transformation. HAI titers of at least 1:40 if
baseline titers were less than 1:10, or an at least four-fold
increase in those with baseline titers at least 1:10 were
considered as evidence of seroconversion. The primary
immunogenicity endpoint was seroconversion rates to
each of the three vaccine strains. The proportion of
children with HAI at least 1:40 after second dose of
vaccination was a secondary endpoint, and an exploratory
analysis was undertaken to compare the proportion
between groups achieving HAI threshold titers at least
1:110. Differences in proportion between study groups
were compared with Kruskal–Wallis test. All statistical
analyses were done using SAS software V9.1.
Ethical considerations
The study was approved by the Human Research Ethics
Committee of the University of the Witwatersrand and
conducted in accordance with Good Clinical Practice
guidelines. Signed, written informed consent was
obtained from caregiver/parent of study participants.
The study was registered at ClinicalTrial.gov number:
NCT NCT00883012.
Results
Of the 410 children enrolled, 203 (98.5%) TIV recipients
and 200 (98.0%) placebo recipients received both study
vaccine doses (Fig. 1), at a median interval of 28 days
between doses. Overall, 76% of children were aged
6–35 months. The median age, CD4þ lymphocyte
count, HIV viral load, and ART regimens of the overall
efficacy and immunogenicity subset cohorts are shown in
Table 1. Baseline demographic and other clinical
parameters were similar between study arms in the
overall efficacy and immunogenicity subset cohorts
(Table 1). Among TIV recipients, there were no
significant changes in median CD4% or absolute
372 AIDS 2013, Vol 27 No 3

Total enrolled
N = 410

Allocated to TIV group (N = 206) Allocated to placebo (N = 204)

Received TIV (N = 206) Randomized Received placebo (N = 204)
(203 received both doses) (200 received both doses)

Immunogenicity cohort; N = 44 Immunogenicity cohort; N = 52

Loss to follow-up/relocated = 4 Loss to follow-up/relocated=0

Withdrawal: 1 Withdrawal: 0
No post-vaccine sample taken; n = 2 No post-vaccine sample taken; n = 4
Post-immuno after 21–35 days, n = 2 Post-immuno after 21–35 days, n = 5
Evaluable = 40 Evaluable = 43

Loss to follow-up/relocated = 8 Loss to follow-up/relocated = 6

Withdrawal = 2 Withdrawal = 1
Died=1; not HIV-infected = 1 Died = 1
Included in ITT analysis = 206 Included in ITT analysis = 204
Included in per protocol analysis = 203 Included in per protocol analysis = 200

Influenza-like illness: N = 66 Influenza-like illness: N = 78

Seasonal influenza Seasonal influenza
Confirmed influenza: N = 19 Confirmed influenza: N = 21

Fig. 1. Disposition of patients randomized and included in the analysis. ITT, intent-to-treat; TIV, trivalent influenza vaccine.

CD4þ cell counts postvaccination compared to pre-
vaccine levels in younger children (P ¼ 0.88 and 0.11;
respectively). Also, no change in median CD4þ cell count
was observed in the older age group (P ¼ 0.62), but
postvaccination, CD4þ% was higher (P ¼ 0037; Table 1).
Immunogenicity of trivalent influenza vaccine
The immunogenicity per-protocol analysis was limited to
83 (86.5%) of 96 children (Fig. 1). Postvaccination,
seroconversion was observed in 47.5, 50.0, and 40.0% of
TIV recipients for A/Brisbane(H1N1), A/Uruguay-
(H3N2), and B/Florida strains, respectively
(P < 0.0001 for all strains compared to placebo recipients;
Table 2). Among TIV recipients, less than 40% of younger
children seroconverted to any vaccine strain, and this was
significantly lower for A/Uruguay(H3N2) compared to
older age group (Table 2).
The proportion of TIV recipients with postvaccination
HAI at least 1:40 was lower in the younger compared to
older age group for A/Uruguay(H3N2) (52.2 vs. 94.1%;
P ¼ 0.005) and B/Florida (43.5 vs. 88.2%; P ¼ 0.004;
Table 3). Similarly, fewer younger TIV recipients (31.0%)
had HAI at least 1:110 to A/Uruguay(H3N2) compared
to older TIV recipients (52.9%; P ¼ 0.013; Table 2). The
proportion of TIV recipients with HAI at least 1:215 was
less than 30% to A/Brisbane(H1N1), less than 24% to
A/Uruguay(H3N2), and less than 12% to B/Florida in
both age groups (Table 2). Postvaccination, GMTs were
lower in the younger compared to older age group of TIV

Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
 Table 1. Demographic and clinical characteristics for the study cohort evaluating the efficacy of trivalent influenza vaccine included in the intent-to-treat analysis.

Measure Efficacy cohort Immunogenicity cohort subseta

TIV N ¼ 206 Placebo N ¼ 204 TIV N ¼ 40 Placebo N ¼ 43 TIV 6–35 months N ¼ 23 TIV 36–59 months N ¼ 17 P

Median age; months 23.1 (6.0–59.9) 24 (6.0–59.8) 34.0 (11.3–59.1) 41.9 (8.9–56.8) 21.4 (11.3–36.0) 47.1 (36.6–59.1) Not applicable
(range)
6 to <36 months 163 (79.1)b 148 (72.6) 23 (57.5) 20 (46.5) Not applicable
36–59 months 43 (20.9) 56 (27.4) 17 (42.5) 23 (53.5)
Women; n (%) 122 (59.2) 109 (53.4) 28 (70.0) 25 (58.1) 17 (73.9) 11 (64.7) 0.53
Ever breastfed, n (%) 34 (16.5) 41 (20.1) 5 (12.5) 9 (20.9) 1 (4.4) 4 (23.5) 0.14
Median weight kg 10.9 (5.4–23.9) 10.9 (5.8–19.5) 11.9 (6.9–18.9) 12.4 (7.5–19.2) 10.8 (6.8–18.5) 14.5 (10.9–18.9) <0.0001
(range)
Median CD4þ % 33.5 (16.3–59.7) 33.4 (15.2–54) 33.1 (16.7–55.9) 34.7 (17.4–50.5) 37.6 (16.7–55.9) 31.2 (22.5–43) 0.02
(range)
Median CD4þ cell 1770.0 (464–4390) 1648.0 (491–4615) 1614 (1165–2301) 1592 (1156–2849) 1793 (1531–2366) 1218 (1080–2093) 0.08
count (IR)
HIV <400 copies 117/153 (76.5) 113/148 (76.4) 15/20 (75.0) 23/25 (92.0) 7/9 (77.8) 8/11 (72.7) 0.99
per ml (%)
On ART (%) 190 (92.2) 187 (91.7) 36 (90.0) 41 (95.3) 22 (95.7) 14 (82.4) 0.29
D4T-3TC- Kaletrac 128 127 20 29 12 8
ZDV-3TC- Kaletra 50 48 11 7 9 2
Other 12 12 5 5 1 4
Postvaccination
Median CD4% Limited to immunogenicity subset 35.5 (18–55.9) 34.0 (15.4–52) 35.9 (18–55.9) 35.4 (22.8–48.2) 0.50
(range)
Median CD4þ cell 1841 (1377–2406) 1452.5 (1083–2059) 2040 (1677–2632) 1377 (1041–1687) 0.011
count (IR)

3TC, lamivudine; ART, antiretroviral therapy; D4T, stavudine; IR, interquartile range; TIV, trivalent influenza vaccine; ZDV, zidovudine.
a
Children also coenrolled into the nested immunogenicity study.
b
Values in parentheses indicate percentage, unless otherwise indicated in row.
c
Kaletra, lopinavir/ritonavir.
Influenza vaccine in HIV-positive children Madhi et al.

Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
373
 374
AIDS

Table 2. Hemagglutinin inhibition assay titers preintervention and 1 month following two doses of trivalent inactivated influenza or placebo in HIV-infected children.

TIV Placebo TIV, 6–35 months TIV, 36–59 months

2013, Vol 27 No 3

Measure N ¼ 40 N ¼ 43 P N ¼ 23 N ¼ 17 Pa

A/Brisbane (H1N1)
Prevaccine HAI
40; N (%) 25 (62.5) 30 (69.8) 0.48 13 (56.5) 12 (70.6) 0.36
Seroconversionb; N (%) 19 (47.5) 2 (4.7) <0.0001 9 (39.1) 10 (58.8) 0.22
Postvaccination HAI
40; N (%) 34 (85.0) 34 (79.1) 0.48 20 (87.0) 14 (82.4) 0.99
Postvaccination HAI
110; N (%) 18 (45.0) 3 (7.0) <0.0001 9 (39.1) 9 (52.9) 0.39
Postvaccination HAI
215; N (%) 10 (25.0) 0 (0) <0.0001 5 (21.7) 5 (29.4) 0.58
Prevaccine GMT (95% CI)c 33.1 (26.7–40.9) 38.7 (32.8–45.8) 0.24 29.6 (23.4–37.5) 38.4 (25.5–57.7) 0.23
Postvaccine GMT (95% CI) 91.9 (64.9–130.2) 44.8 (37.4–53.6) 0.0003 77.6 (49.2–122.5) 115.5 (64.7–206.0) 0.26
d
Mean GMT fold increase (95% CI) 2.8 (1.9–4.1) 1.2 (1.0–1.3) <0.0001 2.6 (1.7–4.0) 3.0 (1.4–6.3) 0.73
A/Uraguay (H3N2)
Prevaccine HAI
40; N (%) 7 (17.5) 8 (18.6) 0.90 1 (4.4) 6 (35.3) 0.29
Seroconversion (%) 20 (50.0) 5 (11.6) <0.0001 8 (34.8) 12 (70.6) 0.025
Postvaccination HAI
40; N (%) 28 (70.0) 14 (32.6) 0.001 12 (52.2) 16 (94.1) 0.005
Postvaccination HAI
110; N (%) 12 (30.0) 7 (16.3) 0.14 3 (13.0) 9 (52.9) 0.013
Postvaccination HAI
215; N (%) 5 (12.5) 1 (2.33) 0.10 1 (4.4) 4 (23.5) 0.14
Prevaccine GMT (95% CI) 16.8 (12.2–23.1) 16.0 (12.4–20.5) 0.79 12.3 (9.5–16.1) 25.5 (13.4–48.5) 0.020
Postvaccine GMT (95% CI) 57.6 (39.9–82.9) 25.5 (18.4–35.3) 0.001 33.4 (22.2–50.1) 120.3 (72.6–199.3) 0.0002
Mean GMT fold increase (95% CI) 3.4 (2.3–5.1) 1.6 (1.2–2.1) 0.001 2.7 (1.8–4.1) 4.7 (2.1–10.5) 0.17
B/Florida
Prevaccine HAI
40; N (%) 5 (12.5) 2 (4.7) 0.25 1 (4.4) 4 (23.5) 0.14
Seroconversion (%) 16 (40.0) 0 <0.0001 8 (34.8) 8 (47.1) 0.43
Postvaccination HAI
40; N (%) 25 (62.5) 4 (9.3) <0.0001 10 (43.5) 15 (88.2) 0.007
Postvaccination HAI
110; N (%) 7 (17.5) 0 (0) 0.004 4 (17.4) 3 (17.7) 0.99
Postvaccination HAI
215; N (%) 2 (5.0) 0 (0) 0.23 0 (0) 2 (11.8) 0.17
Prevaccine GMT (95% CI) 12.1 (8.7–16.8) 9.8 (8.2–11.9) 0.27 7.9 (6.0–10.2) 21.7 (11.7–40.2) 0.001
Postvaccine GMT (95% CI) 39.3 (27.2–56.8) 11.6 (9.6–14.0) <0.0001 27.9 (17.0–45.7) 62.6 (37.3–105.2) 0.025
Mean GMT fold increase (95% CI) 3.2 (2.3–4.6) 1.2 (1.0–1.3) <0.0001 3.5 (2.3–5.4) 2.9 (1.5–5.6) 0.57
a
P value comparing seroconversion rates in trivalent influenza vaccine (TIV) recipients between children 6–36 months of age compared to older children.
b
Seroconversion defined as hemagglutinin inhibition (HAI) titers of
1:40 if baseline titers were <1:10; or an at least four-fold increase in those with baseline titers
1:10.
c
GMT, geometric mean titers of hemagglutinin inhibition antibody. 95% Confidence intervals (95% CI) shown in parenthesis.
d
Measure fold increase in GMT comparing prevaccine to postvaccine HAI titers.

Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
 Influenza vaccine in HIV-positive children Madhi et al. 375

23.4 (63.0 to 64.7)

24.7 (64.7 to 66.4)

21.2 (80.3 to 66.3)

Table 3. Efficacy of trivalent inactivated influenza vaccine in preventing any respiratory illness, influenza-like illness, and influenza confirmed illness in HIV-infected children during circulation of
recipients for A/Uruguay(H3N2) and B/Florida

Vaccine efficacy

Not calculated

Not calculated
(Table 2).

(95% CI) Efficacy of trivalent influenza vaccine against

seasonal influenza illness
Investigation for influenza virus occurred in 123 children
who were investigated on 144 illness occasions, including
74 (51.4%) which fulfilled ILI criteria (Table 3). Multiple
N ¼ 200
Placebo

(2.9)
(1.2)
(0.7)
(0.6)

17 (0.6)

15 (0.6)
Per-protocol

2695.4

illness visits spaced at least 7 days apart were undertaken in

78
31
18
17
20 children, including one with three illness visits. The
median age of children was 18.8 months (range 6–54.6)
for ILI and 20.9 months (range 6–54.6) for confirmed
influenza illness. There were 19 episodes of confirmed
N ¼ 203

(2.4)
(0.7)
(0.5)
(0.5)

13 (0.5)

12 (0.4)
2736.4
TIV

influenza illness among TIV recipients and 21 in placebo

66
19
14
14

recipients. The dominant strain of influenza virus

identified was H3N2 (37 of 40 cases). Two TIV
recipients, aged 13.5 and 21.7 months with CD4þ% of
27.6 and 36.8 and HIV viral load of less than 40 and more
(2.7)
(0.9)
(0.6)
(0.6)

30 (0.6)

27 (0.5)
N ¼ 403
5431.8
Total

than 750 000 copies/ml at enrollment, respectively,

50
32
31
144

experienced two episodes of confirmed H3N2 illness

spaced 13 and 38 days apart.

The overall incidence (per 100 child-weeks) was 2.6 (95%

22.6 (36.2 to 56.6)

19.1 (61.0 to 59.9)

15.9 (72.8 to 59.5)

Vaccine efficacy

CI 2.2–3.1) for medically attended respiratory illness,

Not calculated

Not calculated
(95% CI)

including 1.3 (95% CI 1.1–1.7) for ILI and 0.7 (95% CI

0.5–1.0) for confirmed influenza illness (Table 3). The
incidence rates of ILI or confirmed influenza illness did
not differ between study arms in the intent-to-treat or
per-protocol analysis (Table 3). There were three TIV
recipients in the immunogenicity subset with confirmed
H3N2 influenza illness, all of whom received both TIV
Intent-to-treat

N ¼ 204
Placebo

(2.8)
(1.6)
(1.1)
(0.8)

21 (0.8)

19 (0.7)

doses and were on ART. These children were 21.4, 25.0,

2749.8

and 54.6 months of age and had CD4þ% of 24.3, 31.3,

CI, confidence interval; ILI, influenza-like illness; TIV, trivalent influenza vaccine.
78
45
31
21

and 31.4, respectively. Two children less than 36 months

failed to mount an immune response to any vaccine strain,
although postvaccination HAI titers were 1:40 for each
N ¼ 206

(2.4)
(1.1)
(0.9)
(0.7)

17 (0.6)

16 (0.6)
2752.3

vaccine strain in the older child.

Total follow-up time during circulation of seasonal influenza virus.
TIV

66
29
24
19

There were one TIV and two placebo recipients

hospitalized for pneumonia during the influenza season,
none of whom tested positive for influenza virus. One
(2.6)
(1.3)
(1.0)
(0.7)

38 (0.7)

35 (0.6)
N ¼ 410
5502.1

death occurred in a placebo recipient outside of a health

Total

facility during the seasonal influenza period, which was

74
55
40
144

attributed to pneumonia but not tested for influenza

virus.
Seasonal influenza follow-up period

Phylogenetic analysis on 21 randomly selected wild-type

1st episode of confirmed influenza
1st episode of ILI; n (incidence)b

Incidence per 100 child-weeks.

H3N2 viruses from study participants formed two clusters

1st episode of confirmed H3N2
All confirmed influenza illness;

with a mean protein distance of 1.2 and 2.6%,

Illness visits; n (incidence)b
ILI episodes; n (incidence)b

respectively, to the vaccine strain (Supplemental digital

seasonal influenza virus.

illness; n (incidence)b

illness; n (incidence)b

material, http://links.lww.com/QAD/A267). Site-by-

site amino acid analysis of the H3N2 viruses (Fig. 2)
in child-weeksa

n (incidence)b

showed evidence of drift in reference to the

A/Uruguay(H3N2) vaccine strain, as indicated by mis-
matched amino acid changes mapped to the immuno-
Measure

dominant epitopes A (S138A), B (K158N; N189K;

P194L), and D (K173Q; T212A).
b
a

Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
376 AIDS 2013, Vol 27 No 3

130 140 150 160 170 180 190 100 210 220

A/Uruguay/716/2007
30324
19226
23857
50993
45538
43766
26582
51757
34187
43717
52343
51042
47112
55427
51255
24860
44648
54226
54565
45606
43075
A/Perth/16/2009

Epitope A Epitope B Epitope D Epitope B Epitope D

Fig. 2. Multiple sequence alignment showing amino acid (aa) mismatches in the immunodominant epitope positions (aa120 to
aa225) in reference to the A/Uruguay/716/2007 (H3N2) vaccine strain. X denotes a heterozygous aa position namely 124(S/N).

Safety of trivalent influenza vaccine Immunodominant epitopes A and B induce high

There were no grade 3 or above solicited adverse events
following either the first or second dose of study vaccine.
The only solicited adverse events documented among
TIV recipients within 3 days of either dose of vaccine
were pain at site of injection (n ¼ 1; 2.3%), fever at least
37.5oC (n ¼ 2; 4.2%), induration at injection site (n ¼ 1;
2.3%), and weakness (n ¼ 2; 4.5%).
Discussion
To our knowledge, this is the first randomized-controlled
study which evaluated the efficacy of TIV in HIV-
infected children. In our study of children mainly 6–35
months of age, who are at heightened risk of severe
influenza illness independent of HIV infection [12], TIV
vaccination did not prevent confirmed seasonal influenza
illness or ILI. The lack of efficacy was in part corroborated
by the poor immunogenicity of TIVagainst the dominant
wild-type circulating strain, particularly in the younger
age group (32.0% seroconversion).
Possible reasons for the lack of efficacy in our study
include those which are inherent to any study evaluating
TIV vaccine efficacy. These include the unpredictability
of the severity of the influenza virus which emerges and
possible vaccine strain mismatch to the wild-type strain.
Genotypic analysis of H3N2 isolates from our study
revealed only a 1.2% change in mean protein distance
compared to A/Uruguay-like virus. These amino acid
changes, however, affected key positions in three
hemagglutinin epitopes (A, B, and D). The changes in
epitopes A (S138A) and B (P194L) corresponded to
positions that form part of the receptor-binding site [13].
efficiency neutralizing antibodies. Whereas single epitope
changes for A or B are associated with minor antigenic
drift, mutations resulting in two or more changed
epitopes as in our study are considered as major antigenic
drift which necessitates selection of a new vaccine strain
[14]. HAI analysis of the wild-type H3N2 strains was not
possible due to the phenotypic characteristic of recent H3
strains which do not hemagglutinate red blood cells from
various species [15]. A reference serum to A/Uru-
guay(H3N2)(2007), however, was negative in HAI assays
against A/Perth/16/2009, suggesting a similar result
could be expected for wild-type strains with the same
mutations at receptor-binding site positions 138 and 194
[13].
The lack of efficacy may also be attributed to the poor
immunogenicity of TIV, particularly in younger HIV-
infected children. Although we did not enroll a control
group of HIV-uninfected children, the seroconversion
rates in the younger age group were lower compared to
the 77.8% (H3N2) to 86.1% (Influenza/B) among
healthy Canadian children aged 6–35 months [16], in
whom an identical TIV formulation and dosing
concentration was evaluated. The seroconversion rates
in the Canadian study were also generally higher
compared to our study’s older age group. The impaired
immune responses to TIV in our study may be due to
HIV-induced impaired B-lymphocyte function, required
for immune responses to TIV, which persist even
following immune reconstitution with ART [17–19].
Impaired memory B-cell response to TIV in HIV-
infected individuals has been attributed to deficiencies in
serum and B-cell activating factor and APRIL (a
proliferating-inducing ligand) and alteration in their
receptors [19].

Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
 Influenza vaccine in HIV-positive children Madhi et al.
Traditionally HAI titers at least 1:40 have been used as a
putative measure of relative protection against influenza
illness in adults [20,21]. The relevance of this threshold in
children is unclear [22,23]. Additionally, a HAI titer of
1:40 is associated with only 50% protection against
influenza illness in adults [20,23], despite the likely
presence of underlying cell-mediated immunity to
influenza induced by previous exposure to influenza
antigens from vaccination or from circulating wild-type
influenza-virus. This underlying cell-mediated immunity
is associated with cross-protection against different
influenza strains [24], and boosting thereof by vaccination
may contribute towards enhancing the efficacy of TIV in
adults, which may also in-part mitigate the effect of any
mismatch between the vaccine and circulating wild-type
influenza strains. Although cell-mediated immune
responses are induced in young children immunized
with live-attenuated influenza vaccine (LAIV), such
responses are less evident with TIV in the absence of
previous exposure to related vaccine strains [24]. In
addition, young children are more likely to be influenza-
naive and consequently may require higher HAI titers,
although vaccine strain mismatch may be more relevant in
them compared to that in adults. Black et al. [23] modeled
that in children, HAI titers of at least 1:110 and at least
1:215 were predictive of 50 and 70% protection,
respectively. In our study, an exploratory analysis
identified only 13% of younger and 52.9% of older
TIV recipients achieved HAI titers at least 110 to
A/Uruguay (H3N2), and less than 30% in both age
groups had titers at least 1:215 for any vaccine strain. The
one vaccine failure due to H3N2 involved a 54.6-month-
old child in our study, who had postvaccination H3N2-
HAI titer of 1:40.
Most other immunogenicity studies on influenza vaccine
in HIV-infected children have involved older HIV-
infected children and adolescents [25–30]. Some of these
studies preceded management of HIV-infected children
with combination, triple ART [25,26,28], and more
recently have focused on children/adolescents on ART
who were immunologically reconstituted and virologi-
cally suppressed [29–31]. Among HIV-infected children
on ART, immune responses to TIV correlated with
immunological status prior to vaccination, with moder-
ate-to-severe immunocompromised children benefiting
from a second dose of vaccine [29]. In addition, the
magnitude of immune response and proportion of
children with presumed seroprotective levels were lower
than in healthy controls for at least two vaccine strains
[30].
The lack of efficacy in our study was surprising,
considering the 75% efficacy reported in the same setting
among HIV-infected adults despite only modest ser-
oconversion rates to TIV (47.4–60.8%) [32]. In addition
to the slightly lower seroconversion rates to TIV in this
study (40–50%), other reasons for the difference in TIV
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
377
efficacy between HIV-infected adults and children may
include that the adults were possibly more likely to have
established underlying cell-mediated immunity to influ-
enza virus because of more frequent exposure to wild-
type virus over their life time. This would have possibly
enhanced their protection against influenza illness
because of the role of cell-mediated immunity in clearing
influenza infection as well as conferring some protection
against influenza strains which are heterologous to the
vaccine strain [24].
Recent studies have evaluated alternate influenza vaccine
formulations, such as virosomal adjuvant vaccines [33,34],
MF59-adjuvant and ASO3-adjuvant influenza vaccine,
and LAIV [35–40], for safety and immunogenicity in HIV-
infected children. Palma et al. [40] reported immunogeni-
city of MF59-adjuvanted monovalent H1N1 influenza
vaccine in HIV-infected children following a single dose of
vaccine, in whom seroconversion rate was 60%, but less
than that in healthy controls (82%). In healthy children
aged 6–72 months, the MF59-adjuvanted influenza
vaccine was 81% more efficacious than unadjuvanted
TIV. This included 79% efficacy against confirmed
influenza illness in children aged 6–35 months and 92%
efficacy in children aged 36–72 months, which was
approximately double the efficacy observed for non-
adjuvanted TIV [41].
Limitations of our study include being a single-center
study because of it primarily been investigator-funded.
Furthermore, the evaluable sample size (n ¼ 403) only
had 40% power to detect a 50% reduction in confirmed
influenza based on the 10% attack rate observed among
placebo recipients. The sample size was, however,
sufficient to detect at least 73% vaccine efficacy at 80%
power based on the observed attack rate. Although the
study did not complete enrollment of the targeted 550
participants prior to the peak of the influenza season, this
was offset by only 4.1% (n ¼ 17) of enrollees, rather than
the projected 20%, being lost to follow-up or withdrawn.
A further limitation of our study is that it was conducted
over a single influenza season. The study is, therefore,
unable to definitively conclude the relative roles of poor
TIV immunogenicity in younger children or antigenic
drift of the circulating wild-type H3N2 away from the
vaccine strain as being responsible for the lack of efficacy
observed against confirmed influenza illness. Also, very
few (<18%) of the children in our study were breastfed at
any stage. It is possible that predominantly breastfed
children may have mounted better immune responses, as
breastfeeding contributes to boosting of the infant
adaptive and innate immune system [42].
The results from our study do not support current
recommendations for immunization of young HIV-
infected children with nonadjuvanted TIV, even if on
ART. We propose that further multicentered controls
studies, which include a placebo arm and which are
378 AIDS 2013, Vol 27 No 3
powered to investigate efficacy against confirmed
influenza illness and not only immunogenicity, are
required. Such studies should consider evaluating higher
strength doses of TIV in young children, as well as the
possible roles of LAIV and adjuvanted TIV in protecting
HIV-infected children against influenza illness.
Acknowledgements
The authors would like to thank all the children and
their parents for their participation in this study. In
addition, the support of the nursing, pharmacy, and
general staff at Perinatal HIV Research Unit and data
clerks at the Respiratory and Meningeal Pathogens
Research is acknowledged.
S.A.M. and A.V. contributed to the study conceptualiz-
ation and design. First draft of the article was cowritten by
S.A.M. and A.V. S.D., H.C., E.L., and T.T. were involved
in participant follow-up and management. M.V. and F.T.
were responsible for laboratory testing of influenza virus.
A.W. was responsible for doing the hemagglutinin
inhibition tests. C.L.C. was involved in protocol devel-
opment and study logistics. M.V. and F.T. made input on
the article sections on genotypic characteristic of the wild-
type influenza virus, and A.W. on the immunogenicity
data. All the authors critically reviewed the article and
contributed to the final draft.
The study was funded in part through an unrestricted
grant from Secure the Future Fund. The rest of the study,
including vaccine procurement, was investigator-funded.
Secure the Future did not have any input into the study
design, data analysis, or write-up of the article.
Conflicts of interest
There are no conflicts of interest.
References
1. Madhi SA, Schoub B, Simmank K, Blackburn N, Klugman KP.
Increased burden of respiratory viral associated severe lower
respiratory tract infections in children infected with human
immunodeficiency virus type-1. J Pediatr 2000; 137:78–84.
2. Madhi SA, Ramasamy N, Bessellar TG, Saloojee H, Klugman
KP. Lower respiratory tract infections associated with influen-
za A and B viruses in an area with a high prevalence of pediatric
human immunodeficiency type 1 infection. Pediatr Infect Dis J
2002; 21:291–297.
3. Evans KD, Kline MW. Prolonged influenza A infection respon-
sive to rimantadine therapy in a human immunodeficiency
virus-infected child. Pediatr Infect Dis J 1995; 14:332–334.
4. Fiore AE, Shay DK, Broder K, Iskander JK, Uyeki TM, Mootrey G,
et al. Prevention and control of influenza: recommendations of
the Advisory Committee on Immunization Practices (ACIP).
MMWR Recomm Rep 2008; 57 (RR-7):1–60.
5. Prevention and control of influenza: recommendations of the
Advisory Committee on Immunization Practices (ACIP). Morb
Mortal Wkly Rep 1997; 46:1–25.
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
6. 2004. DAIDS Toxicity Tables. http://rcc.tech-es.com/Docu-
ment/ safetyandpharmacovigilance/Toxicity Tables_Pediatri-
c_Over3MonthsAge_v03.pdf. [Accessed 26 June 2012].
7. Sanofi. PRODUCT MONOGRAPH VAXIGRIP: inactivated in-
fluenza vaccine trivalent types A and B (Split Virion) http://
wwwnovaccinecom/pdffiles/vaxigrip_package_insertpdf. [Ac-
cessed 21 May 2012].
8. National Institute for Communicable Diseases. Consolidated
Influenza Surveillance Weekly Reports. http://www.nicd.ac.za.
[Accessed 26 June 2012].
9. Van Kerkhove MD, Mounts AW. 2009 versus 2010 comparison
of influenza activity in southern hemisphere temperate coun-
tries. Influenza Other Respir Viruses 2011; 5:375–379.
10. Glencross D, Scott LE, Jani IV, Barnett D, Janossy G. CD45-
assisted PanLeucogating for accurate, cost-effective dual-
platform CD4R T-cell enumeration. Cytometry 2002; 50:69–
77.
11. Tamura K, Dudley J, Nei M, Kumar S. MEGA4: Molecular
Evolutionary Genetics Analysis (MEGA) software version
4.0. Mol Biol Evol 2007; 24:1596–1599.
12. Neuzil KM, Mellen BG, Wright PF, Mitchel EFJ, Griffin MR. The
effect of influenza on hospitalizations, outpatient visits, and
courses of antibiotics in children. N Engl J Med 2000; 342:225–
231.
13. Ndifon W, Wingreen NS, Levin SA. Differential neutralization
efficiency of hemagglutinin epitopes, antibody interference,
and the design of influenza vaccines. Proc Natl Acad Sci U S A
2009; 106:8701–8706.
14. Huang JW, Yang JM. Changed epitopes drive the antigenic drift
for influenza A (H3N2) viruses. BMC Bioinformatics 2011; 12
(Suppl 1):S31.
15. Medeiros R, Escriou N, Naffakh N, Manuguerra JC, van der Werf
S. Hemagglutinin residues of recent human A(H3N2) influenza
viruses that contribute to the inability to agglutinate chicken
erythrocytes. Virology 2001; 289:74–85.
16. Langley JM, Scheifele DW, Quach C, Vanderkooi OG, Ward B,
McNeil S, et al. Safety and immunogenicity of 2010-2011
H1N12009-containing trivalent inactivated influenza vaccine
in children 12-59 months of age previously given AS03-adju-
vanted H1N12009 pandemic vaccine: a PHAC/CIHR Influenza
Research Network (PCIRN) study. Vaccine 2012; 30:3389–
3394.
17. Cagigi A, Nilsson A, Pensieroso S, Chiodi F. Dysfunctional
B-cell responses during HIV-1 infection: implication for influ-
enza vaccination and highly active antiretroviral therapy.
Lancet Infect Dis 2010; 10:499–503.
18. Cagigi A, Palma P, Nilsson A, Di Cesare S, Pensieroso S,
Kakoulidou M, et al. The impact of active HIV-1 replication
on the physiological age-related decline of immature-transi-
tional B-cells in HIV-1 infected children. AIDS 2010; 24:2075–
2080.
19. Pallikkuth S, Kanthikeel SP, Silva SY, Fischl M, Pahwa R, Pahwa
S. Innate immune defects correlate with failure of antibody
responses to H1N1/09 vaccine in HIV-infected patients.
J Allergy Clin Immunol 2011; 128:1279–1285.
20. Hobson D, Curry RL, Beare AS, Ward-Gardner A. The role of
serum haemagglutination-inhibiting antibody in protection
against challenge infection with influenza A2 and B virus.
J Hyg (Lond) 1972; 70:767–777.
21. Coudeville L, Bailleux F, Riche B, Megas F, Andre P, Ecochard
R. Relationship between haemagglutination-inhibiting anti-
body titres and clinical protection against influenza: develop-
ment and application of a bayesian random-effects model.
BMC Med Res Methodol 2010; 10:18.
22. Fox JP, Cooney MK, Hall CE, Foy HM. Influenzavirus infections
in Seattle families, 1975-1979. II. Pattern of infection in
invaded households and relation of age and prior antibody
to occurrence of infection and related illness. Am J Epidemiol
1982; 116:228–242.
23. Black S, Nicolay U, Vesikari T, Knuf M, Del Giudice G, Della
Cioppa G, et al. Hemagglutination inhibition antibody titers as
a correlate of protection for inactivated influenza vaccines in
children. Pediat Infect Dis J 2011; 30:1081–1085.
24. Forrest BD, Pride MW, Dunning AJ, Capeding MR, Chotpitaya-
sunondh T, Tam JS, et al. Correlation of cellular immune
responses with protection against culture-confirmed influenza
virus in young children. Clin Vaccine Immunol 2008; 15:1042–
1053.
 Influenza vaccine in HIV-positive children Madhi et al.
25. Lyall EG, Charlett A, Watkins P, Zambon M. Response to
influenza virus vaccination in vertical HIV infection. Arch
Dis Child 1997; 76:215–218.
26. Jackson CR, Vavro CL, Valentine ME, Pennington KN, Lanier ER,
Katz SL, et al. Effect of influenza immunization on immunologic
and virologic characteristics of pediatric patients infected with
human immunodeficiency virus. Pediatr Infect Dis J 1997;
16:200–204.
27. Ramilo O, Hicks PJ, Borvak J, Gross LM, Zhong D, Squires JE,
et al. T cell activation and human immunodeficiency virus
replication after influenza immunization of infected children.
Pediatr Infect Dis J 1996; 15:197–203.
28. Chadwick EG, Chang G, Decker MD, Yogev R, Dimichele D,
Edwards KM. Serologic response to standard inactivated influ-
enza vaccine in human immunodeficiency virus-infected chil-
dren. Pediatr Infect Dis J 1994; 13:206–211.
29. Kosalaraksa P, Srirompotong U, Newman RW, Lumbiganon P,
Wood JM. Serological response to trivalent inactive influenza
vaccine in HIV-infected children with different immunologic
status. Vaccine 2011; 29:3055–3060.
30. Machado AA, Machado CM, Boas LS, Lopes MC, Gouvea Ade F,
Succi RC, et al. Immunogenicity of an inactivated influenza
vaccine and postvaccination influenza surveillance in HIV-
infected and noninfected children and adolescents. AIDS Res
Hum Retroviruses 2011; 27:999–1003.
31. Montoya CJ, Toro MF, Aguirre C, Bustamante A, Hernandez M,
Arango LP, et al. Abnormal humoral immune response to
influenza vaccination in pediatric type-1 human immunodefi-
ciency virus infected patients receiving highly active anti-
retroviral therapy. Mem Inst Oswaldo Cruz 2007; 102:501–
508.
32. Madhi SA, Maskew M, Koen A, Kuwanda L, Besselaar TG,
Naidoo D, et al. Trivalent inactivated influenza vaccine
in African adults infected with human immunodeficient
virus: double blind, randomized clinical trial of efficacy,
immunogenicity, and safety. Clin Infect Dis 2011; 52:128–
137.
33. Vigano A, Zuccotti GV, Pacei M, Erba P, Castelletti E, Giacomet
V, et al. Humoral and cellular response to influenza vaccine in
HIV-infected children with full viroimmunologic response to
antiretroviral therapy. J Acquir Immune Defic Syndr 2008;
48:289–296.
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
379
34. Tanzi E, Esposito S, Bojanin J, Amendola A, Trabattoni D,
Pariani E, et al. Immunogenicity and effect of a virosomal
influenza vaccine on viral replication and T-cell activation
in HIV-infected children receiving highly active antiretroviral
therapy. J Med Virol 2006; 78:440–445.
35. Vigano A, Giacomet V, Pariani E, Giani E, Manfredini V,
Bedogni G, et al. Long-term immunogenicity after one and
two doses of a monovalent MF59-adjuvanted A/H1N1 influ-
enza virus vaccine coadministered with the seasonal 2009-
2010 nonadjuvanted influenza virus vaccine in HIV-infected
children, adolescents, and young adults in a randomized con-
trolled trial. Clin Vaccine Immunol 2011; 18:1503–1509.
36. Phongsamart W, Sirisanthana V, Wittawatmongkol O, Malee-
satharn A, Sudjaritruk T, Chearskul P, et al. Immunogenicity and
safety of monovalent influenza A (H1N1) 2009 in HIV-infected
Thai children. Vaccine 2011; 29:8705–8711.
37. Weinberg A, Song LY, Fenton T, Nachman SA, Read JS, Patter-
son-Bartlett J, et al. T cell responses of HIV-infected children
after administration of inactivated or live attenuated influenza
vaccines. AIDS Res Hum Retroviruses 2010; 26:51–59.
38. Levin MJ, Song LY, Fenton T, Nachman S, Patterson J, Walker R,
et al. Shedding of live vaccine virus, comparative safety, and
influenza-specific antibody responses after administration of
live attenuated and inactivated trivalent influenza vaccines to
HIV-infected children. Vaccine 2008; 26:4210–4217.
39. Weinberg A, Song LY, Walker R, Allende M, Fenton T, Patter-
son-Bartlett J, et al. Antiinfluenza serum and mucosal antibody
responses after administration of live attenuated or inactivated
influenza vaccines to HIV-infected children. J Acquir Immune
Defic Syndr 2010; 55:189–196.
40. Palma P, Romiti ML, Bernardi S, Pontrelli G, Mora N, Santilli V,
et al. Safety and immunogenicity of a monovalent MF59(R)-
adjuvanted A/H1N1 vaccine in HIV-infected children and
young adults. Biologicals 2012; 40:134–139.
41. Vesikari T, Knuf M, Wutzler P, Karvonen A, Kieninger-Baum D,
Schmitt HJ, et al. Oil-in-water emulsion adjuvant with influ-
enza vaccine in young children. N Engl J Med 2011; 365:1406–
1416.
42. M’Rabet L, Vos AP, Boehm G, Garssen J. Breast-feeding and its
role in early development of the immune system in infants:
consequences for health later in life. J Nutr 2008; 138:1782S–
1790S.