Articles

Comparative efficacy and safety of vaccines to prevent

seasonal influenza: A systematic review and network
meta-analysis
Silvia Minozzi,a* Theodore Lytras,b Silvia Gianola,c Marien Gonzalez-Lorenzo,d Greta Castellini,c Cristina Galli,e Danilo Cereda,f
Stefanos Bonovas,g Elena Pariani,e and Lorenzo Moja e
a
Department of Epidemiology, Lazio regional health Service, Rome, Italy
b
School of Medicine, European University Cyprus, Nicosia, Cyprus
c
IRCCS Istituto Ortopedico Galeazzi, Unit of Clinical Epidemiology, Milan, Italy
d
Laboratory of Clinical Research Methodology, Istituto di Ricerche Farmacologiche Mario Negri IRCCS, Milan, Italy
e
Department of Biomedical Sciences for Health, University of Milan, Milan, Italy
f
Directorate General for Health, Lombardy Region, Milan, Italy
g
Department of Biomedical Sciences, IRCCS Humanitas Research Hospital, Humanitas University, Milan, Italy

Summary
Background Influenza is one of the most common respiratory viral infections worldwide. Numerous vaccines are eClinicalMedicine
used to prevent influenza. Their selection should be informed by the best available evidence. We aimed to estimate 2022;46: 101331
Published online 25
the comparative efficacy and safety of seasonal influenza vaccines in children, adults and the elderly.
March 2022
https://doi.org/10.1016/j.
Methods We conducted a systematic review and network meta-analysis (NMA). We searched the Cochrane Library eclinm.2022.101331
Central Register of Controlled Trials, MEDLINE and EMBASE databases, and websites of regulatory agencies,
through December 15th, 2020. We included placebo- or no vaccination-controlled, and head-to-head randomized
clinical trials (RCTs). Pairs of reviewers independently screened the studies, abstracted the data, and appraised the
risk of bias in accordance to the Cochrane Handbook for Systematic Reviews of Interventions. The primary outcome
was laboratory-confirmed influenza. We also synthesized data for hospitalization, mortality, influenza-like illness
(ILI), pneumonia or lower respiratory-tract disease, systemic and local adverse events (AEs). We estimated summary
risk ratios (RR) using pairwise and NMA with random effects. This study is registered with PROSPERO, number
CRD42018091895.

Findings We identified 13,439 citations. A total of 231 RCTs were included after screening: 11 studies did not provide
useful data for the analysis; 220 RCTs [100,677 children (< 18 years) and 329,127 adults (18−60 years) and elderly
(≥ 61 years)] were included in the NMA. In adults and the elderly, all vaccines, except the trivalent inactivated intra-
dermal vaccine (3-IIV ID), were more effective than placebo in reducing the risk of laboratory-confirmed influenza,
with a RR between 0.33 (95% credible interval [CrI] 0.21−0.55) for trivalent inactivated high-dose (3-IIV HD) and
0.56 (95% CrI 0.41−0.74) for trivalent live-attenuated vaccine (3-LAIV). In adults and the elderly, compared with tri-
valent inactivated vaccine (3-IIV), no significant differences were found for any, except 3-LAIV, which was less effica-
cious [RR 1.41 (95% CrI 1.04−1.88)]. In children, compared with placebo, RR ranged between 0.13 (95% CrI 0.03
−0.51) for trivalent inactivated vaccine adjuvanted with MF59/AS03 and 0.55 (95% CrI 0.36−0.83) for trivalent inac-
tivated vaccine. Compared with 3-IIV, 3-LAIV and trivalent inactivated adjuvanted with MF59/AS03 were more effi-
cacious [RR 0.52 (95% CrI 0.32−0.82) and RR 0.23 (95% CrI 0.06−0.87)] in reducing laboratory-confirmed
influenza. With regard to safety, higher systemic AEs rates after vaccination with 3-IIV, 3-IIV HD, 3-IIV ID, 3-IIV

Abbreviations: AE, adverse event; CI, confidence interval; CrI, credible interval; IIV, inactivated influenza vaccine; ILI, influenza-like
illness; LAIV, live-attenuated influenza vaccine; NMA, network meta-analysis; RCT, randomized controlled trial; RIV, recombinant
influenza vaccine; RR, risk ratio; SUCRA, surface under the cumulative ranking curve; 3-IIV, trivalent inactivated influenza vac-
cine; 3-IIV HD, trivalent inactivated high-dose influenza vaccine; 3-IIV ID, trivalent inactivated intradermal influenza vaccine; 3-
IIV MF59/AS03-adj, trivalent inactivated influenza vaccine adjuvanted with MF59/AS03; 3-IIV vir/lip-adj, trivalent inactivated
influenza vaccine adjuvanted with virosome/liposome; 3-RIV, trivalent recombinant influenza vaccine; 4-IIV, quadrivalent inacti-
vated influenza vaccine; 4-IIV HD, quadrivalent inactivated high-dose influenza vaccine; 4-IIV ID, quadrivalent inactivated intra-
dermal influenza vaccine; 4-IIV MF59/AS03-adj, quadrivalent inactivated influenza vaccine adjuvanted with MF59/AS03; 4-IIV
vir/lip-adj, quadrivalent inactivated influenza vaccine adjuvanted with virosome/liposome; 4-RIV, quadrivalent recombinant influ-
enza vaccine; 3-LAIV, trivalent live-attenuated influenza vaccine; 4-LAIV, quadrivalent live-attenuated influenza vaccine
*Corresponding author.
E-mail address: [email protected] (S. Minozzi).

www.thelancet.com Vol 46 Month April, 2022 1

 Articles

MF59/AS03-adj, quadrivalent inactivated (4-IIV), quadrivalent adjuvanted (4-IIV MF59/AS03-adj), quadrivalent

recombinant (4-RIV), 3-LAIV or quadrivalent live attenuated (4-LAIV) vaccines were noted in adults and the elderly
[RR 1.5 (95% CrI 1.18−1.89) to 1.15 (95% CrI 1.06−1.23)] compared with placebo. In children, the systemic AEs rate
after vaccination was not significantly higher than placebo.

Interpretation All vaccines cumulatively achieved major reductions in the incidence of laboratory-confirmed influ-
enza in children, adults, and the elderly. While the live-attenuated was more efficacious than the inactivated vaccine
in children, many vaccine types can be used in adults and the elderly.

Funding The directorate general of welfare, Lombardy region.

Copyright Ó 2022 The Author(s). Published by Elsevier Ltd. This is an open access article under the CC BY license
(http://creativecommons.org/licenses/by/4.0/)
Keywords: Influenza; Vaccines; Systematic review; Network meta-analysis

Implications of all the available evidence

Research in context Progress in efficacy of influenza vaccines has achieved
consistent improvements in the prevention of influenza.
Evidence before this study Many vaccines can be used with comparable efficacy
The World Health Organization recommends annual profiles. Given similar relative benefits, differences in
influenza vaccination for high-risk groups, but these rec- safety and price might be considered as additional
ommendations do not favor any particular type of vac- important dimensions to select vaccines in preventive
strategies.
cine over others. A large number of studies confirm the
superiority of any type of vaccine against placebo or no
vaccine for several relevant outcomes. To date, an
assessment of the merits of one vaccine over another
using a network meta-analysis approach have been lim-
ited to specific populations, such as HIV patients, but
have never extended to all patients irrespective of age
Introduction
and clinical characteristics. Literature search for this Influenza is a respiratory illness caused by influenza
meta-analysis was performed using PubMed, Embase viruses that are transmitted efficiently from human to
and Cochrane Central from 1991 to December 15th, human.1 Globally, seasonal influenza affects 5−10% of
2020, with no language restrictions. We included RCTs adults and 20−30% of children every year1 and is
that assessed the efficacy and safety of any type of triva- responsible for 3−5 million cases of severe illness and
lent/quadrivalent seasonal influenza vaccine, at the up to 650,000 deaths every year.2 The continuing evolu-
doses licensed by the European Medicines Agency and/ tion of seasonal influenza viruses, which limits the abil-
or the US Food and Drug Administration, on children, ity of our immune system to fight effectively the
adults and the elderly.
infection, is associated with the recurrent burden of sea-
sonal epidemics.3 The socioeconomic costs incurred by
Added value of this study
each influenza season are estimated at USD 87 billion
Our evidence synthesis includes 220 studies and per year.4
429,804 participants and represents to the best of Influenza vaccination is a pillar of public health and
our knowledge the most comprehensive synthesis to is focused on people at high risk of complications - preg-
date on the comparative efficacy and safety of influ-
nant women, children, the elderly and immunocompro-
enza vaccines across age groups. All types of vac-
cines we included in our study (except trivalent
mised, and persons with chronic illnesses - as well as
inactivated intradermal) produce important gains those who live with or care for persons at high risk. The
when compared to placebo in terms of incidence of World Health Organization recommends annual vacci-
laboratory-confirmed influenza in adults and the nation for all high-risk groups.1 The American Advisory
elderly. Evidence for quadrivalent vaccines, intro- Committee on Immunization Practices recommends
duced more recently, is rapidly cumulating achieving that all persons aged ≥ 6 months without contraindica-
benefits similar to well-established trivalent vaccines. tions receive routine annual vaccination with a licensed
Most vaccines had similar efficacy profiles. Vaccines and age-appropriate vaccine.5 These recommendations
were less efficacious and less well tolerated in the do not favor any particular type of vaccine over others.
elderly than in adults and children. In children, triva-
They reflect the results of randomized clinical trials
lent live attenuated vaccine was more efficacious
than trivalent inactivated.
(RCTs), generally designed to test a vaccine against pla-
cebo but limiting the ability of single studies to inform

2 www.thelancet.com Vol 46 Month April, 2022

 Articles

immunization strategies on the most appropriate vac- and high-dose (60 µg of hemagglutinin per strain
cine options. compared with 15 µg per strain of standard-dose vac-
Network meta-analyses (NMAs) can be applied to cines) (HD-IIVs).
estimate comparative efficacy, summarize and interpret  Live-attenuated influenza vaccines (LAIVs) adminis-
the available evidence, and identify the best vaccine tered by intranasal (IN) route.
types for different populations.6 To date, only one NMA  Recombinant influenza vaccines (RIVs) adminis-
on comparative efficacy of influenza vaccines among tered IM.
HIV-positive people has been published.7 To fill this
gap, we performed a systematic review and NMA to We included studies that used placebo, no vaccina-
inform influenza vaccination strategies by comparing tion or no-influenza vaccine as comparator. We
the efficacy and safety of different types of seasonal excluded cluster, crossover RCTs and studies that com-
influenza vaccines. pared the same type of vaccine produced by different
companies; we also excluded studies that analyzed the
2009 pandemic influenza vaccine, as the influenza
Methods manufacturing process and timeline are different.9
We followed the PRISMA Extension Statement for [Details of inclusion criteria are reported in the appen-
Reporting of Systematic Reviews Incorporating Net- dix (p 5)]
work Meta-analyses of Health Care Interventions guide-
lines.8 This review was registered with PROSPERO,
number CRD42018091895.
Outcomes
The primary outcome was laboratory-confirmed influ-
Search strategy and selection criteria enza (i.e., influenza symptoms with a positive laboratory
We conducted a systematic review and NMA. We diagnosis). Secondary outcomes were: hospitalization,
searched the Cochrane Central Register of Controlled overall and influenza-related mortality, influenza-like
Trials (CENTRAL) (The Cochrane Library), MEDLINE illness (ILI), influenza-related pneumonia or lower
(PubMed) and EMBASE from 1991 to December 15th, respiratory-tract disease, systemic and local adverse
2020, with no language restrictions. The full search events (AEs). For safety outcomes we measured the
strategy is reported in the appendix (pp 3−4). We also number of patients with at least one AE (systemic or
manually screened the citation lists from relevant litera- local) in each study arm. When the number of partici-
ture sources (e.g., previously published systematic pants with at least one systemic AE was not reported,
reviews). we used as proxy measures for adults the number of
We included RCTs that assessed the efficacy and participants with malaise as first choice, headache as
safety of any type of trivalent/quadrivalent seasonal second choice, and fever ≥ 37.5 °C as third choice; for
influenza vaccine, at the doses licensed by the European children we used as proxy measures irritability as first
Medicines Agency and/or the US Food and Drug choice, decreased activity/weakness as second choice,
Administration, for the prevention of seasonal influenza and fever ≥ 37.5 °C as third choice. When the number
in any individual irrespective of age and health status, of participants with at least one local AE was not
i.e., healthy children (< 18 years), healthy adults (18−60 reported, we used as a proxy measure nasal congestion
years), and the elderly (age ≥ 61 years), pregnant or rhinorrhea for intranasal vaccines; for intramuscular
women, individuals of any age at risk of influenza- or intradermal vaccines, we used pain as first choice,
related complications due to pre-existing diseases (can- local swelling/induration as second choice, and ery-
cer, immunosuppression, chronic respiratory, cardio- thema/redness as third choice. For pregnant women,
vascular or metabolic diseases). Studies with multiple we examined both pregnancy outcomes (spontaneous
arms comparing different dosages of vaccine were abortion, fetal death, stillbirth, preterm birth < 37 weeks
included only for the arms with licensed doses. We did gestation) and neonatal outcomes (minor and major
not consider vaccination co-interventions, such as topi- congenital malformations and neonatal death). For chil-
cal adjuvants (e.g. topical imiquimod). dren, we examined otitis media and exacerbation of pri-
The following trivalent (3-)/quadrivalent (4-) seasonal mary disease in those with pre-existing respiratory
influenza vaccines were included in the review: disease.

 Inactivated influenza vaccines (IIVs) (whole virus,

split or sub-unit) administered intramuscularly Procedures
(IM) or intradermically (ID). IIVs were further Two authors (SM, MGL) independently screened all
grouped in: MF59/AS03 (oil-in-water emulsion) titles and abstracts and assessed the full text for eligibil-
adjuvanted (MF59/AS03-adj-IIVs), virosome/lipo- ity. Any doubt was resolved by discussion; in case of per-
some (microparticule) adjuvanted (vir/lip-adj-IIVs), sisting disagreement, a third author (EP) acted as

www.thelancet.com Vol 46 Month April, 2022 3

 Articles

arbitrator. Two authors independently extracted the data grouped by production method and vaccine characteris-
(SG, GC) and assessed risk of bias (SM, SG) according tics (appendix, p 7), and are summarized in the appen-
to the criteria of the Cochrane Handbook for Systematic dix (Box 1, p 6.).
Reviews of Interventions.10 When outcome data were Separate analyses were performed for children (< 18
available only in graphs, we extrapolated them using years) and adults and elderly (≥ 18years). Subgroup
web plot digitizer application (https://automeris.io/Web analyses were performed for children (0−5 years), the
PlotDigitizer/).11 The intercoder reliability and validity elderly (≥ 61 years), and participants with a comorbidity
of this software for estimating event rates from figures (i.e., cancer, immunocompromised state, and pre-exist-
is high.12 In addition, to increase accuracy, as acceptable ing respiratory diseases).
standards of practice for data extraction in systematic For brevity and consistency, we focus the present
reviews, two independent authors (GC, SG) extracted report on comparisons that included placebo and no
data and reached consensus also for the data extracted vaccination or trivalent inactivated vaccine as compara-
from figures. Any doubt was resolved by discussion; in tors. The Netleague tables and forest plots present the
case of persisting disagreement, a third author (SM) comparisons in the following order: placebo, trivalent
acted as arbitrator. inactivated and recombinant, quadrivalent inactivated
and recombinant, trivalent and quadrivalent live-attenu-
ated vaccines.
Data analysis
We conducted conventional pairwise random-effects
Role of the funding source
meta-analyses for all dichotomous outcomes and com-
The funder of the study had no role in study design,
parisons when at least two studies were available. Fixed
data collection, data analysis, data interpretation, or
effects models were not used as a certain degree of het-
writing of the report. The corresponding author had full
erogeneity was expected among and between the ran-
access to all the data in the study and had final responsi-
domized studies.13 We calculated the risk ratio (RR) for
bility for the decision to submit for publication.
each RCT, with the uncertainty in each result expressed
with a 95% confidence interval (CI). Heterogeneity was
analyzed by means of the I2 statistic (with
I2values ≥ 60% considered as “substantial hetero- Results
geneity”) and the chi-square test (statistically significant We identified 19,997 records through the database
for P value < 0.10).14 search and 74 articles from scanning the reference lists.
NMA was then conducted to estimate the effect size After removing duplicates, 13,439 unique references
for all possible pairwise comparisons between vaccines, remained, 13,046 of which were excluded because of
as well as to rank the efficacy and safety of the vaccines. title and abstract. We retrieved 393 full text studies for
The amount of heterogeneity was assumed to be equal more detailed evaluation, 162 of which were excluded.
across all treatment comparisons in the network.15 To References of excluded studies are reported in the
rank vaccine efficacy with respect to each outcome, the appendix (pp 8−16). A total of 231 studies with 441,093
Surface Under the Cumulative Ranking curve (SUCRA) participants were included (appendix, pp 17−26): 220
and the mean ranks were used.16 We evaluated the tran- studies (519 arms, including multi-arm and multi-
sitivity assumption by visually comparing the distribu- cohort studies) with 429,804 participants (100,677 chil-
tion of clinical and methodological variables (e.g., age, dren and 329,127 adults and elderly) contributed to at
risk of bias) that could act as effect modifiers across least one NMA, whereas 11 studies did not provide use-
treatment comparisons. Incoherence (i.e., agreement ful quantitative data (Figure 1).
between direct and indirect evidence) was evaluated for Out of the 220 studies, 161 arms involved children
laboratory-confirmed influenza and systemic AEs using (age < 18 years), of which children aged ≤ 5 years
the design-by-treatment test17,18 and dividing the direct accounted for 54.7% (n = 68,726); 358 arms involved
and indirect evidence.19 The random-effects model was adults and elderly, of which the elderly (≥ 61 years)
fitted in a Bayesian framework using the RR as effect accounted for 45.4% (n = 149,437). Twenty-six arms
estimate and uncertainty expressed with a 95% credible involved patients with chronic respiratory disease
interval (CrI).16,19−21 (n = 7250; 5471 children and 1779 adults and elderly)
All analyses were performed using the R software and 65 arms involved patients with cancer or immuno-
environment, version 3.5.22 NMA was conducted with compromised status (n = 3865, 1141 children and 2724
JAGS23 and the gemtc package for R. adults and elderly). Sixteen arms (11,130 participants)
We compared any vaccine with placebo or no vacci- included pregnant women. Furthermore, 210 studies
nation, and then any other vaccine with trivalent inacti- (95.4%) were conducted in an outpatient setting; 88
vated vaccine as a common comparator, after which we included one or more placebo arms and five compared
compared each type of vaccine against each other. Net- influenza vaccine with another vaccine. In addition, 93
work nodes were categorized as per licensed vaccines (42.3%) studies were funded by the industry, 76

4 www.thelancet.com Vol 46 Month April, 2022

 Articles

Study Characteristics No. (%) of RCTs (N = 220)

Year of publication
1988−1995 9 (4.1)
1996−2000 18 (8.2)
2001−2005 24 (10.9)
2006−2010 51 (23.2)
2011−2015 56 (25.4)
2016−2020 62 (28.2)
Continent
Europe 47 (21.4)
Africa 10 (4.5)
Asia 42 (19.1)
North America 80 (36.4)
South America 6 (2.7)
Oceania 5 (2.3)
Multi-continent 21 (9.5)
Not reported 9 (4.1)
Setting
Outpatient 210 (95.4)
Inpatient 5 (2.3)
Long-term care facility 5 (2.3)
Funding
Industry 93 (42.3)
Government/ private no profit 76 (34.5)
Not reported 46 (20.9)
Industry and private no profit 5 (2.3)
Figure 1. Study flow from literature search.

(34.5%) by government or private no profit agencies,

while the remaining did not specify the funding No. of No. (%) of
participants compared
source. Finally, 36.4% were conducted in North Amer-
arms
ica, and 21.4% in Europe (Table 1). The appendix (pp (N = 519) *
27−68) reports details of the characteristics of
included studies. Type of vaccine

The results of pairwise meta-analyses are presented placebo/no vaccine/other 75,511 112 (21.6)

in the appendix (pp 77−86). Five vaccines (3-IIV, 3- vaccine

RIV, 4-IIV, 3-LAIV, 3-IIV ID) had at least one placebo- 3-IIV trivalent inactivated 153,885 192 (37.0)

or no vaccination-controlled trial. We use the term 3-IIV HD trivalent inactivated 30,101 21 (4.1)

“placebo-controlled” to mean both placebo, no vaccina- high-dose

tion or other vaccine than the influenza one. The vac- 3-IIV ID trivalent inactivated 15,197 33 (6.3)

cines 3-LAIV, 3-IIV ID, 3-RIV, 3-IIV HD, 3-IIV MF59/ intradermal

AS03-adj, 3-IIV vir/lip-adj and 4-IIV, were compared 3-IIV MF59/AS03-adj trivalent 36,310 33 (6.3)

directly with 3-IIV. For laboratory-confirmed influenza inactivated adjuvanted

and any systemic AE, we found no evidence of incoher- MF59/AS03

ence between direct and indirect evidence (appendix, 3-IIV vir/lip-adj trivalent inacti- 1871 14 (2.7)

pp 87−94). vated adjuvanted virosome/

The NMA for the primary outcome of laboratory-con- liposome

firmed influenza in adults and the elderly included 40 3-RIV trivalent inactivated 3708 7 (1.3)

RCTs (209,095 participants) with eight different types recombinant

of vaccines (Figure 2). All comparisons except one (3- 4-IIV quadrivalent inactivated 52,599 47 (9.1)

IIV ID) showed a significant difference compared 4-IIV HD quadrivalent inacti- 1807 2 (0.4)

against placebo, with RR between 0.33 (95% CrI 0.21 vated high-dose

−0.55) for 3-IIV HD and 0.56 (95% CrI 0.41−0.74) for 4-IIV ID quadrivalent inacti- 1672 1 (0.2)

3-LAIV; little and non-significant differences were vated intradermal

observed on comparison between different vaccines. Table 1 (Continued)

Compared with 3-IIV, the differences were again small

www.thelancet.com Vol 46 Month April, 2022 5

 Articles

No. of No. (%) of

participants compared
arms
(N = 519) *

4-IIV MF59/AS03-adj quadriva- 994 2 (0.4)

lent inactivated adjuvanted
MF59/AS03
Figure 2. Network Geometry of laboratory-confirmed influenza
4-RIV quadrivalent inactivated 17,455 7 (1.3)
Panel a: adults and elderly; Panel b: children. The thickness
recombinant of the line is proportional to the precision of each direct esti-
3-LAIV trivalent live- 36,491 45 (8.7) mate, and the width of each circle is proportional to the num-
attenuated ber of studies included in the treatment. Placebo: placebo/no
4-LAIV quadrivalent live- 2203 3 (0.6) vaccine; 3-IIV: trivalent inactivated intramuscular; 3-IIV HD: triva-
attenuated lent inactivated high dose intramuscular; 3-IIV ID: trivalent inac-
Age (years) tivated intradermal; 3-IIV MF59/AS03-adj: trivalent inactivated
≤5 68,726 88 (17.0) adjuvanted with MF59/ASO3 intramuscular: 3-RIV: trivalent
recombinant intramuscular; 4-IIV: quadrivalent inactivated
6−17 12,512 37 (7.1)
intramuscular; 4-RIV: quadrivalent recombinant intramuscular;
mixed aged children (<18) 19,439 36 (6.9)
3-LAIV: trivalent live attenuated intranasal; 4-LAIV: quadrivalent
18 - 60 156,642 212 (40.9)
live attenuated intranasal.
≥61 149,437 109 (21.0)
adult and elderly (318) 23,048 37 (7.1)
Comorbidity
vaccines (Figure 2). 3-IIV, 3-LAIV and 3-IIV MF59/
none (i.e. healthy participants) 341,853 362 (69.7)
AS03-adj showed significant difference compared
chronic respiratory disease 7250 28 (5.4)
against placebo with RR of 0.55 (95% CrI 0.36−0.83),
multimorbidity 65,071 45 (8.7)
0.28 (95% CrI 0.19−0.41) and 0.13 (95% CrI 0.03
immunodepression (any 3865 65 (12.5)
−0.51) respectively, while 4-IIV and 4-LAIV showed no
cause)/cancer
difference. Comparison of the vaccines with each other
pregnant women 10,936 14 (2.7)
showed that 3-LAIV and 3-IIV MF59/AS03-adj were
pregnant women/ 194 2 (0.4)
more efficacious than 3-IIV with RR of 0.52 (95% CrI
immunodepression
0.32−0.82) and 0.23 (95% CrI 0.06−0.87), respectively
not reported 635 3 (0.6)
(Table 3; appendix, p 101). Ranking of vaccines based on
Table 1: Characteristics of studies included in the review. SUCRAs are summarized in the appendix (p 144).
* the overall number of arms includes multi-arms and multi-cohort In the subgroup analysis of children aged ≤ 5 years,
studies. the 3-LAIV and the 3-IIV MF59/AS03-adj vaccines were
more efficacious than placebo (19 RCTs, 53,973 partici-
pants, 4 vaccines in the network), being RR 0.30 (95%
CrI 0.18−0.46) and 0.14 (95% CrI 0.03−0.68), respec-
and not significant, except for 3-LAIV, which was less tively (appendix, p 101). In children with pre-existing
efficacious [RR 1.41 (95% CrI 1.04−1.88)] (Table 2; respiratory diseases, only the 3-LAIV vaccine was more
appendix, p 95). Ranking of vaccines based on SUCRAs efficacious than placebo (RR 0.09, 95% CrI 0.00−0.54;
are presented in the appendix (p 144). 5 RCTs, 5801 participants, 3 vaccines) (appendix, p 103).
In the elderly subgroup (age ≥ 61 years) (12 RCTs, Due to paucity of data, subgroup analyses of immuno-
107,265 participants, seven vaccines in the network) compromised children and children with cancer were
only 4-RIV showed a significant effect on laboratory- not performed.
confirmed influenza against placebo [RR 0.3 (95% CrI The NMA for hospitalization in all adults and
0.06−0.97)], although point estimates favored all vac- elderly showed that all vaccines but 4-IIV and 3-LAIV
cines over placebo, with RR between 0.4 (95% CrI 0.11 reduced the hospitalization rate compared against pla-
−1.14) for 4-IIV, 0.4 (95% CrI 0.07−2.00) for 3-RIV cebo (21 studies, 59,193 participants, 7 vaccines in the
and 0.63 (95% CrI 0.27−1.44) for 3-LAIV (appendix, p network), with RR between 0 (95% CrI 0−0.1) and
96). In the subgroup of immunocompromised/cancer 0.29 (95% CrI 0.14−0.52) (appendix, p 105). Among
patients (6 RCTs, 1276 participants, 4 vaccines) only 3- children, both 3-IIV vir/lip-adj and 4-RIV reduced hos-
IIV appeared better than placebo [RR 0.19 (95% CrI pitalization rate (13 studies, 50,249 participants, 5 vac-
0.03−0.94)] (appendix, p 98). Because of paucity of cines in the network), RR 0.16 (95% CrI 0.04−0.58)
data, sub group analysis of participants with pre-existing and RR 0.0 (95% CrI 0.0−0.1), respectively, versus
respiratory diseases was not performed. placebo, and RR 0.17 (95% CrI 0.04−0.8) and RR 0.0
The NMA for laboratory-confirmed influenza in chil- (95% CrI 0.0−1.11), respectively, versus 3-IIV (appen-
dren included 24 RCTs (60,502 participants) with 5 dix, p 107).

6 www.thelancet.com Vol 46 Month April, 2022

www.thelancet.com Vol 46 Month April, 2022

placebo 0.4 (0.34, 0.46) 0.33 (0.21, 0.55) 0.4 (0.04, 3.89) 0.37 (0.22, 0.64) 0.39 (0.23, 0.61) 0.56 (0.36, 0.83) 0.47 (0.29, 0.7) 0.56 (0.41, 0.74)

2.51 (2.18, 2.95) 3-IIV 0.83 (0.53, 1.37) 1 (0.11, 9.75) 0.94 (0.57, 1.59) 0.97 (0.59, 1.54) 1.41 (0.9, 2.15) 1.18 (0.73, 1.81) 1.41 (1.04, 1.88)
3.01 (1.82, 4.86) 1.2 (0.73, 1.88) 3-IIV HD 1.19 (0.12, 12.08) 1.13 (0.57, 2.18) 1.16 (0.57, 2.19) 1.68 (0.87, 3.07) 1.41 (0.71, 2.56) 1.68 (0.94, 2.85)
2.53 (0.26, 24.11) 1 (0.1, 9.52) 0.84 (0.08, 8.49) 3-IIV ID 0.95 (0.09, 9.45) 0.97 (0.1, 9.55) 1.41 (0.14, 13.77) 1.17 (0.12, 11.47) 1.41 (0.14, 13.68)
2.67 (1.57, 4.5) 1.06 (0.63, 1.74) 0.89 (0.46, 1.75) 1.06 (0.11, 10.86) 3-IIV MF59/AS03-adj 1.03 (0.49, 2.01) 1.49 (0.76, 2.8) 1.25 (0.62, 2.33) 1.5 (0.82, 2.63)
2.59 (1.65, 4.29) 1.03 (0.65, 1.7) 0.86 (0.46, 1.76) 1.03 (0.1, 10.51) 0.97 (0.5, 2.02) 3-RIV 1.44 (0.79, 2.74) 1.21 (0.64, 2.29) 1.45 (0.86, 2.54)
1.79 (1.21, 2.76) 0.71 (0.47, 1.11) 0.59 (0.33, 1.15) 0.71 (0.07, 7.15) 0.67 (0.36, 1.32) 0.69 (0.37, 1.27) 4-IIV 0.84 (0.56, 1.21) 1 (0.61, 1.66)
2.13 (1.42, 3.44) 0.85 (0.55, 1.38) 0.71 (0.39, 1.4) 0.85 (0.09, 8.65) 0.8 (0.43, 1.61) 0.83 (0.44, 1.55) 1.2 (0.83, 1.78) 4-RIV 1.2 (0.72, 2.05)
1.79 (1.35, 2.43) 0.71 (0.53, 0.96) 0.59 (0.35, 1.06) 0.71 (0.07, 7.02) 0.67 (0.38, 1.23) 0.69 (0.39, 1.17) 1 (0.6, 1.64) 0.84 (0.49, 1.38) 3-LAIV

Table 2: Netleague laboratory-confirmed influenza: adults and elderly.

The estimate is located at the intersection of the column-defining vaccine and the row-defining vaccine. Data are RRs (95% CrI). Significant results are in bold.
In the upper triangle, comparison of treatments should be read from right to left. An RR below 1 favors the medication on the bottom right vs. the medication on the top left in the diagonal. E.g., RR 0.40 (95% CrI 0.34−0.46) indi-
cates a significant reduction in the incidence of laboratory-confirmed influenza for the trivalent inactivated vaccine (3-IIV) compared with placebo or no vaccination.
In the bottom triangle, comparison of treatments should be read from left to right. An RR below 1 favors the medication on the top left vs. the medication on the bottom right in the diagonal. E.g., RR 0.84 (95% CrI 0.08−8.49)
indicates a non-significant reduction in the incidence of laboratory-confirmed influenza for the trivalent inactivated high-dose vaccine (3-IIV-HD) compared with the trivalent inactivated intradermal vaccine (3-IIV ID).
Abbreviations: Placebo: placebo or no vaccination; 3-IIV: trivalent inactivated intramuscular; 3-IIV HD: trivalent inactivated high-dose intramuscular; 3-IIV ID: trivalent inactivated intradermal; 3-IIV MF59/AS03-adj: trivalent inacti-
vated adjuvanted with MF59/AS03 intramuscular; 3-RIV: trivalent recombinant intramuscular; 4-IIV: quadrivalent inactivated intramuscular; 4-RIV: quadrivalent recombinant intramuscular; 3-LAIV: trivalent live-attenuated
intranasal.

placebo 0.55 (0.36, 0.83) 0.13 (0.03, 0.51) 0.5 (0.21, 1.19) 0.28 (0.19, 0.41) 0.71 (0.21, 2.4)

1.81 (1.2, 2.77) 3-IIV 0.23 (0.06, 0.87) 0.9 (0.35, 2.4) 0.52 (0.32, 0.82) 1.28 (0.35, 4.7)
7.96 (1.98, 32.92) 4.4 (1.15, 17.03) 3-IIV MF59/AS03-adj 3.97 (0.77, 21.01) 2.27 (0.54, 9.38) 5.64 (0.88, 36.6)
2 (0.84, 4.76) 1.11 (0.42, 2.88) 0.25 (0.05, 1.3) 4-IIV 0.57 (0.22, 1.44) 1.42 (0.32, 6.38)
3.51 (2.46, 5.17) 1.94 (1.22, 3.15) 0.44 (0.11, 1.85) 1.76 (0.7, 4.6) 3-LAIV 2.49 (0.7, 9.08)
1.41 (0.42, 4.81) 0.78 (0.21, 2.82) 0.18 (0.03, 1.13) 0.7 (0.16, 3.15) 0.4 (0.11, 1.42) 4-LAIV

Table 3: Netleague laboratory-confirmed influenza: children.

The estimate is located at the intersection of the column-defining vaccine and the row-defining vaccine. Data are RRs (95% CrI). Significant results are in bold.
In the upper triangle, comparison of treatments should be read from right to left. An RR below 1 favors the medication on the bottom right vs. the medication on the top left in the diagonal. E.g., RR 0.55 (95% CrI 0.36−0.83) indi-
cates a significant reduction in the incidence of laboratory-confirmed influenza for the trivalent inactivated vaccine (3-IIV) compared with placebo/ no vaccination.
In the bottom triangle, comparison of treatments should be read from left to right. An RR below 1 favors the medication on the top left vs. the medication on the bottom right in the diagonal. E.g., RR 0.25 (95% CrI 0.05−1.3) indi-
cates a non-significant reduction in the incidence of laboratory-confirmed for the trivalent inactivated adjuvanted with MF59/AS03 intramuscular vaccine (3-IIV MF59/AS03-adj) compared with the quadrivalent inactivated intra-
muscular (4-IIV).
Abbreviations: Placebo: placebo/ no vaccination; 3-IIV: trivalent inactivated intramuscular; 3-IIV MF59/AS03-adj: trivalent inactivated adjuvanted with MF59/AS03 intramuscular; 4-IIV: quadrivalent inactivated intramuscular; 3-
LAIV: trivalent live-attenuated intranasal; 4-LAIV: quadrivalent live-attenuated intranasal.

Articles
7
 Articles

In adults and the elderly (31 RCTs, 174,705 partici- p 125). In the subgroup with immunocompromised or
pants, 10 vaccines in the network), all but two vaccines cancer patients (19 RCTs, 2212 participants, 8 vaccines) 3-
3-LAIV and 4-IIV-HD) were associated with a significant RIV was associated with less systemic AEs compared
reduction in mortality compared against placebo with against placebo and 3-IIV, but data are sparse and non-
RR between 0.0 (95% CrI 0.0−0.3 and 0.37 (95% CrI informative (appendix, p 128).
0.16−0.78) (appendix, p 110). In children (15 studies, In children, the NMA of any systemic AEs included
42,834 participants, 6 vaccines in the network), 3-IIV 59 RCTs (77,208 participants, 9 vaccines). None of the
MF59/AS03-adj, 4-RIV and 4-LAIV were associated vaccines were associated with a significantly higher per-
with a reduction in mortality with RR of 0.04 (95% CrI centage of children reporting at least one systemic AE
0.0−0.96), 0.0 (95% CrI 0.0−0.4) and 0.0 (95% CrI compared against placebo. Compared with 3-IIV, only 3-
0.0−0.43). However, the data were sparse and the IIV MF59/AS03-adj had a higher percentage of children
results largely imprecise (appendix, p 112). reporting at least one systemic AE (RR 1.23, 95% CrI
In adults and the elderly, only 3-IIV was shown to be 1.02−1.49) (appendix, p 131).
efficacious in reducing influenza-like illness (ILI) com- In children aged ≤ 5 years, (33 RCTs, 54,146 partici-
pared with placebo (30 RCTs, 83,537 participants, 7 vac- pants, 7 vaccines) only 3-LAIV was associated with a
cines in the network), RR 0.8 (95% CrI 0.72−0.89) higher percentage of children reporting at least one sys-
(appendix, p 114). In children, both 3-IIV and 3-LAIV temic AE compared against placebo with RR of 1.17
were efficacious in reducing ILI compared against pla- (95% CrI 1.0−1.37). Compared against 3-IIV, only 3-IIV
cebo (16 RCTs, 37,165 participants, 4 vaccines in the net- MF59/AS03-adj was associated with more AEs with a
work), with RR 0.59 (95% CrI 0.42−0.80) and 0.64 RR of 1.24 (95% CrI 1.0−1.54) (appendix, p 134). Due to
(95% CrI 0.44−0.84), respectively (appendix, p 116). paucity of data, we were unable to perform subgroup
None of the vaccines was found to be more effica- analyses in immunocompromised children and chil-
cious than placebo (12 RCTs, 22,690 children, 5 vac- dren with cancer.
cines in the network) in reducing the incidence of acute The NMA for any local AEs in adults and the elderly
otitis media in children (appendix, p 118). included 123 RCTs (223,093 participants, 13 vaccines).
Six RCTs (11,130 participants) included pregnant All vaccines showed higher frequency of participants
women. Four studies compared 3-IIV against placebo/ reporting at least one local AE compared against pla-
no intervention or other non-influenza vaccines, one cebo, with RR from 4.47 (95% CrI 1.96−10.15) for 4-IIV
compared 3-LAIV versus 3-IIV, and one 4-IIV versus 3- ID to 2.23 (95% CrI 1.18−4.17) for 4-LAIV. In compari-
IIV. No significant differences were found for any preg- son with 3-IIV, only 3-IIV HD, 3-IIV ID, and 3-IIV
nancy or neonatal outcome (appendix, p 120). MF59/AS03-adj were associated with more participants
Due to paucity of data, we were unable to perform reporting local AEs (appendix, p 137).
analyses for influenza-related pneumonia/lower respira- The NMA for any local AE in children included 55
tory diseases and influenza-related mortality. RCTs (64,004 participants, 9 vaccines). Only 3-LAIV
The NMA for systemic AEs in adults and elderly showed significantly more children reporting at least
included 121 RCTs (220,595 participants, 13 vaccines). one local AE compared with placebo (RR 1.22, 95% CrI
Compared against placebo, after vaccination with 3-IIV, 1.08−1.38) and 3-IIV (RR 1.25, 95% CrI 1.01−1.55)
3-IIV HD, 3-IIV ID, 3-IIV MF59/AS03-adj, 4-IIV, 4-IIV (appendix, p 140).
MF59/AS03-adj, 4-RIV, 3-LAIV or 4-LAIV, more partici- The appendix (p 145) reports the outcomes and sub-
pants reported at least one systemic AE, with a RR group analyses that were planned but not conducted
between 1.5 (95% CrI 1.18−1.89) for 4-IIV MF59/AS03- because of insufficient data.
adj and 1.15 (95% CrI 1.06−1.23) for 3-IIV. Compared More than a half (57.6%) of the studies had low risk of
with 3-IIV, only the 3-IIV HD, the 3-IIV MF59/AS03-adj bias for random sequence generation, only one study had
and the 4-IIV MF59/AS03-adj vaccines were associated high risk and the remaining (41.9%) had an unclear risk.
with at least one systemic AE in significantly more par- Allocation concealment was judged at low risk in 39.8% of
ticipants, with a RR of 1.16 (95% CrI 1.04−1.29), 1.2 studies, unclear in 57.7%, and at high risk in 2.6%. Fur-
(95% CrI1.09−1.32) and 1.31 (95% CrI 1.05−1.63), thermore, 45.1% were judged at low risk for performance
respectively (appendix, p 122). The appendix (p 144) bias, 11.2% at unclear risk, and 43.7% at high risk; 54.6%
presents the vaccine ranking based on SUCRAs. were judged at low risk of detection bias, 18.6% at unclear
In the subgroup analysis of the elderly (age ≥ 61 risk, and 26.8% at high risk. The majority of studies were
years) (44 RCTs, 107,701 participants, 11 vaccines), none at low risk of attrition bias (87%) and selective reporting
of the vaccines showed significant differences compared bias (93.0%) (appendix, pp 69−76).
against placebo. In comparison with 3-IIV, the subgroup
analysis of elderly participants yielded results similar to
the overall analysis, with the 3-IIV HD and 3-IIV MF59/ Discussion
AS03-adj vaccines associated with a higher percentage of Our NMA quantifies the progress achieved in the pre-
participants reporting at least one systemic AE (appendix, vention of influenza with different types of vaccines

8 www.thelancet.com Vol 46 Month April, 2022

 Articles

over the last 30 years. Based on the cumulative data magnitude for laboratory-confirmed influenza cases.26
from 220 clinical trials of influenza vaccines involving This and our review are comprehensive for the number
429,804 children, adults, and elderly people, we found of studies and range of outcomes covered, the study
that, in adults and elderly, all influenza vaccines, except population (not restricted to any particular population
for the trivalent inactivated intradermal vaccine, con- or clinical setting), and the sophisticated statistical anal-
ferred protection against laboratory-confirmed influ- yses (including consistency tests of direct and indirect
enza compared against placebo or no vaccination. evidence). Still, there might be clinically important dif-
Cumulative results for benefits such as reduction in ferences among influenza vaccines that have not
mortality or hospitalization are less certain: while some emerged so far. Given the modest rates of infection in
vaccines reduced the relative risk of hospitalization or low-incidence seasons, and the occurrence of complete
death by one half or more, the CIs around effect esti- or partial antigenic mismatch between circulating
mates were imprecise, limiting our confidence. In viruses and vaccine strains, a definitive assessment of
adults and the elderly, evidence for newer quadrivalent the merits of one vaccine over another would require
vaccines is rapidly cumulating achieving benefits simi- very large pragmatic RCTs. Differences in vaccine per-
lar to well-established trivalent inactivated vaccine, formance with different vaccines might be assessed by
which is probably the most common in the world. In means of serum antibodies analyses However, serocon-
children, trivalent live-attenuated vaccines were signifi- version and seroprotection may not represent a valid
cantly superior to trivalent inactivated vaccines against surrogate of clinical protection, particularly in frail
laboratory-confirmed influenza, with the exception of patients (e.g., the elderly).27
MF59-adjuvanted vaccines which seem more effective Our review has some limitations. First, analyses
but also associated with more AEs. were limited by the amount of data in the studies.
In the elderly subgroup when compared to placebo Although patient-relevant outcomes were included, not
vaccines protection against laboratory-confirmed influ- all studies reported them, and most had few events,
enza is less pronounced, with only the quadrivalent especially with regard to key events such as pneumonia
recombinant intramuscular vaccine reaching a statisti- and influenza-related mortality. Trials were of short
cally significant result. This discrepancy between age duration. Some vaccines based on virosomes were
groups might reflect a heterogeneous immunological included in our network but are no longer available on
response, a higher susceptibility to disease, or some het- the market28; pediatric influenza vaccine adjuvanted
erogeneity among studies. A Cochrane review address- with MF59 was included but it has not been authorized
ing the role of influenza vaccination in the elderly for use in children in most countries.29,30 Since we ana-
concluded that there is low-certainty evidence of efficacy lyzed only average treatment effects and did not investi-
of influenza vaccines.24 Observational studies found gate potentially important clinical and demographical
that MF59-adjuvanted vaccines are more efficacious modifiers of vaccination response at the individual
than unadjuvanted vaccines in the elderly.25 Moreover, patient level (e.g., age, sex, immune status), our sub-
less data have been collected on newer vaccines involv- group analyses might have an aggregation bias. We
ing the very elderly (85+). Key target populations of vac- were unable to assess variation in vaccine efficacy by
cine campaigns for which there is at present little or no influenza type/subtype, circulating strains and degree
data should be the focus of future investigations so that of match as only few studies reported type/subtype-spe-
more direct evidence about the relative merits of these cific estimates together with an overall efficacy estimate
vaccines can be obtained. However, these comparisons and provided data on individual years, a precondition to
should privilege head-to-head vaccine comparisons. explore the effect of match/mismatch. This modifier
In terms of safety, RCTs provided a wealth of indica- could also have impact on the meta-analysis transitivity
tive data. In adults and the elderly, all vaccines but triva- assumption. Finally, we did not perform a formal cost-
lent inactivated adjuvanted with virosome/liposome, effectiveness analysis.
trivalent recombinant, quadrivalent inactivated high In conclusion, influenza vaccines were associated
dose and intradermal caused significantly more sys- with a lower risk of laboratory-confirmed influenza
temic AEs than placebo. The largest differences were compared with placebo or no vaccination. Across the
limited to frequent but not clinically relevant reactions. head-to-head comparisons, there were no significant dif-
In children, only the trivalent live-attenuated vaccine ferences in the associations between any of the vaccines
was associated with more systemic and local AEs than considered here and the risk of influenza among adults
placebo. When compared with trivalent inactivated vac- and the elderly. Short-term safety outcomes were more
cine, only the 3-IIV MF59/AS03-adj was associated with heterogenous. In children, the trivalent live-attenuated
significantly more children reporting at least one sys- vaccine was more efficacious in preventing laboratory-
temic AE. confirmed influenza than the trivalent inactivated vac-
An overview of systematic reviews concluded that cine, though it showed a greater number of minor AEs.
most seasonal influenza vaccines show statistically sig- In adults and the elderly, the choice of vaccine may
nificant efficacy, with an overlapping degree of depend on patient and caregiver values and preferences

www.thelancet.com Vol 46 Month April, 2022 9

 Articles

or costs and feasibility. The variable mismatch between References

vaccine strains and circulating viruses is an important 1 World Health Organization (WHO). Influenza (Seasonal). Geneva:
WHO, 2018. https://www.who.int/news-room/fact-sheets/detail/
confounder that we could not control. This and other influenza-(seasonal). Accessed 14 March 2022.
uncertainties that might result from different treatment 2 Iuliano AD, Roguski KM, Chang HH, et al. Estimates of global sea-
sonal influenza-associated respiratory mortality: a modelling study.
settings, temper any strong conclusions that can be Lancet. 2018;391(10127):1285–1300.
drawn from the present findings. Nonetheless, we 3 Petrova VN, Russell CA. The evolution of seasonal influenza
believe that health systems can exploit the prevention viruses. Nat Rev Microbiol. 2018;16(1):47–60.
4 European Commission. Proposal for a council recommendation on
benefits conferred by similarly effective vaccines in large seasonal influenza vaccination. Brussels: Commission of the Euro-
vaccination programs, selecting products based on pean Communities, 2009. https://op.europa.eu/en/publication-
price, safety and local factors (e.g., national production). detail/-/publication/99fdfdb9-ce99-4064-9711-44385d555c1b/lan
guage-en. Accessed 14 March 2022.
5 Grohskopf LA, Sokolow LZ, Broder KR, Walter EB, Fry AM, Jerni-
gan DB. Prevention and control of seasonal influenza with vac-
Funding cines: recommendations of the advisory committee on
immunization practices-United States, 2018-19 influenza season.
Directorate general of welfare, Lombardy region. MMWR Recomm Rep Morb Mortal wkly Rep Recomm Rep. 2018;67
(3):1–20.
6 Higgins JP, Welton NJ. Network meta-analysis: a norm for compar-
ative effectiveness? Lancet. 2015;386(9994):628–630.
Author contributions 7 Zhang W, Sun H, Atiquzzaman M, Sou J, Anis AH, Cooper C.
Minozzi, Pariani and Moja conceived and designed the Influenza vaccination for HIV-positive people: systematic review
study. Minozzi, Castellini, Gonzalez-Lorenzo and Gia- and network meta-analysis. Vaccine. 2018;36(28):4077–4086.
8 Hutton B, Salanti G, Caldwell DM, et al. The PRISMA extension
nola selected studies and extracted data. Lytras analysed statement for reporting of systematic reviews incorporating net-
data. Minozzi, Moja, and Lytras wrote the first draft of work meta-analyses of health care interventions: checklist and
the manuscript. Bonovas, Castellini, Cereda, Galli, Gon- explanations. Ann Intern Med. 2015;162(11):777–784.
9 Dormitzer PR, Tsai TF, Del Giudice G. New technologies for influ-
zalez-Lorenzo, Gianola and Pariani critically reviewed enza vaccines. Hum Vaccin Immunother. 2012;8(1):45–58.
the manuscript for important intellectual content. All 10 Higgins JP, Green S. Cochrane Handbook for Systematic Reviews of
Interventions Version 5.1.0. (Editors). The Cochrane Collaboration;
authors agreed with the results and conclusions of this 2011. (updated March 2011). www.handbook.cochrane.org.
article. All authors had full access to all the data in the 11 Rohatgi A. WebPlotDigitizer (Version 4.5). Pacifica, California
study and accept responsibility to submit for publication. (US), 2021. https://automeris.io/WebPlotDigitizer/citation.html.
Accessed 14 March 2022.
12 Burda BU, O'Connor EA, Webber EM, Redmond N, Perdue LA.
Estimating data from figures with a web-based program: considera-
Data sharing tions for a systematic review. Res Synth Methods. 2017;8(3):258–262.
13 DerSimonian R, Laird N. Meta-analysis in clinical trials. Control
Raw data from studies included in this review are partly Clin Trials. 1986;7(3):177–188.
available in supplementary data files. Additional data, 14 Higgins JPT, Thomas J, Chandler J, Cumpston M, Li T, Page MJ,
such as meta-analyses originally planned and subse- et al. Cochrane Handbook for Systematic Reviews of Interventions. (edi-
tors). 2nd ed. Chichester (UK): John Wiley & Sons; 2019.
quently not performed because of paucity of studies, 15 Rhodes KM, Turner RM, Higgins JP. Predictive distributions were
can be made available upon request. developed for the extent of heterogeneity in meta-analyses of con-
tinuous outcome data. J Clin Epidemiol. 2015;68(1):52–60.
16 Salanti G, Ades AE, Ioannidis JP. Graphical methods and numeri-
cal summaries for presenting results from multiple-treatment
Declaration of interests meta-analysis: an overview and tutorial. J Clin Epidemiol. 2011;64
EP declares support for attending meeting and travel by (2):163–171.
17 Higgins JP, Jackson D, Barrett JK, Lu G, Ades AE, White IR. Con-
Sanofi and Seqirus and, participation in Advisory Board sistency and inconsistency in network meta-analysis: concepts and
of Sanofi. CG declares participation in Advisory Board models for multi-arm studies. Res Synth Methods. 2012;3(2):98–110.
of Sanofi. SM, TL, SG, MGL, GC, DC, SB and LM 18 Dias S, Welton NJ, Caldwell DM, Ades AE. Checking consistency
in mixed treatment comparison meta-analysis. Stat Med. 2010;29
declare no competing interests. (7−8):932–944.
19 Caldwell DM, Ades AE, Higgins JP. Simultaneous comparison of
multiple treatments: combining direct and indirect evidence. BMJ.
2005;331(7521):897–900. (Clinical research ed).
Acknowledgments 20 Salanti G, Kavvoura FK, Ioannidis JP. Exploring the geometry of
The study was funded by the Directorate General of Wel- treatment networks. Ann Intern Med. 2008;148(7):544–553.
21 Cipriani A, Higgins JP, Geddes JR, Salanti G. Conceptual and tech-
fare, Lombardy Region. Lombardy Region [Project nical challenges in network meta-analysis. Ann Intern Med.
“Vaccination strategies in Lombardy and use of the Evi- 2013;159(2):130–137.
dence to Decision (EtD) framework”]. The views expressed 22 R Core Team (2020). R: A language and environment for statistical
computing. R Foundation for Statistical Computing, Vienna, Aus-
are those of the authors and not necessarily those of the tria. URL https://www.R-project.org/. Accessed 14 March 2022.
Directorate General of Welfare, Lombardy Region. 23 Plummer M. JAGS: a program for analysis of Bayesian graphical
models using Gibbs sampling. In: Proceedings of the 3rd International
Workshop on Distributed Statistical Computing. Vienna; 2003125.
24 Demicheli V, Jefferson T, Di Pietrantonj C, et al. Vaccines for pre-
Supplementary materials venting influenza in the elderly. Cochr Datab Syst Rev. 2018;2:
Supplementary material associated with this article can CD004876.
25 Domnich A, Arata L, Amicizia D, Puig-Barbera J, Gasparini R, Pan-
be found in the online version at doi:10.1016/j. atto D. Effectiveness of MF59-adjuvanted seasonal influenza
eclinm.2022.101331.

10 www.thelancet.com Vol 46 Month April, 2022

 Articles

vaccine in the elderly: a systematic review and meta-analysis. Vac- 28 Moser C, Amacker M, Zurbriggen R. Influenza virosomes as a vac-
cine. 2017;35(4):513–520. cine adjuvant and carrier system. Expert Rev Vaccines. 2011;10
26 Manzoli L, Ioannidis JP, Flacco ME, De Vito C, Villari P. Effective- (4):437–446.
ness and harms of seasonal and pandemic influenza vaccines in 29 EMA. Fluad paediatric. Influenza vaccine, surface antigen, inacti-
children, adults and elderly: a critical review and re-analysis of 15 vated, adjuvanted with mf59c.1. withdrawal assessment report:
meta-analyses. Hum Vaccin Immunother. 2012;8(7):851–862. European medicines agency, 2012.
27 Wijnans L, Voordouw B. A review of the changes to the licensing of 30 US Food and Drug Administration (FDA). Fluad. Washington, DC:
influenza vaccines in Europe. Influenza Other Respir Viruses. FDA, 2021. https://www.fda.gov/vaccines-blood-biologics/vac
2016;10(1):2–8. cines/fluad. Accessed 14 March 2022.

www.thelancet.com Vol 46 Month April, 2022 11