Mf

Ma

b

c

a

ARRAA

KIASA

1

o[nptgmrhoasi

0d

Vaccine 27 (2009) 6959–6965

Contents lists available at ScienceDirect

Vaccine

journa l homepage: www.e lsev ier .com/ locate /vacc ine

F59-adjuvanted versus non-adjuvanted influenza vaccines: Integrated analysisrom a large safety database

ichele Pellegrinia,∗, Uwe Nicolayb, Kelly Lindertc, Nicola Grotha, Giovanni Della Cioppaa

Global Clinical Research & Development, Novartis Vaccines and Diagnostics, Via Fiorentina 1, 53100 Siena, ItalyGlobal Clinical Research & Development, Novartis Vaccines and Diagnostics, Marburg, GermanyGlobal Clinical Research & Development, Novartis Vaccines and Diagnostics, Cambridge, MA, USA

r t i c l e i n f o

rticle history:eceived 7 July 2009eceived in revised form 26 August 2009ccepted 26 August 2009vailable online 12 September 2009

eywords:nfluenza vaccinedjuvantafetyutoimmune disease

a b s t r a c t

Background: Adding adjuvants such as MF59® to influenza vaccines can enhance the immune response.This analysis evaluated the safety profile of MF59-adjuvanted [(+)MF59] compared with non-adjuvanted[(−)MF59] vaccines in a large clinical database.Methods: Safety data were pooled from 64 clinical trials involving (+)MF59 seasonal and pandemicinfluenza vaccines. Safety outcomes were analysed in the overall population and in subjects aged ≥65years, in all clinical trials and in controlled trials only.Findings: Data from 20,447 (+)MF59 and 7526 (−)MF59 subjects were analysed. Overall, (+)MF59 subjectshad lower risks than (−)MF59 subjects of experiencing any unsolicited adverse event (AE) (26.8% vs39.2%; adjusted risk ratio [ARR] 0.65; 95% CI 0.60–0.70), cardiovascular AEs (1.9% vs 5.6%; ARR 0.44; 95%CI 0.35–0.55), new onset chronic diseases (1.3% vs 1.9%; ARR 0.71; 95% CI 0.57–0.87) and death (0.8%vs 1.2%; ARR 0.67; 95% CI 0.51–0.87). Few AEs of potential autoimmune origin were reported: 0.71 and0.67 per 1000 with (+)MF59 and (−)MF59, respectively. As expected, (+)MF59 subjects had a higher riskof solicited local or systemic reactions within 3 days of vaccination (58.5% vs 46.9%, weighted RR 1.34;95% CI 1.28–1.40). Safety outcomes were consistent between total and elderly populations, and between

all trials and controlled trials, although statistical significance was lost for some of the outcomes in thesubgroups.Interpretation: This large-scale analysis supports the good safety profile of (+)MF59 seasonal and pan-

and s

demic influenza vaccines

. Introduction

Influenza epidemics are thought to result in 3–5 million casesf severe illness and 250,000–500,000 deaths worldwide each year1]. In industrialised countries, elderly individuals are most vul-erable to influenza-associated illness and death; therefore, torevent influenza and its complications the World Health Organiza-ion (WHO) recommends annual influenza vaccinations for this ageroup [1,2]. However, conventional influenza vaccinations are onlyodestly protective in the elderly due to the decrease in immune

esponse associated with advancing age [3]. More importantly per-aps, vaccines against pandemic influenza strains (e.g., H5N1) are

nly modestly immunogenic when the antigens are used alone,nd the complexity of pre-pandemic and pandemic vaccinationtrategies requires dose-sparing, cross-reactivity and induction ofmmune memory [4].

∗ Corresponding author. Tel.: +39 0577 243827.E-mail address: michele.pellegrini@novartis.com (M. Pellegrini).

264-410X/$ – see front matter © 2009 Elsevier Ltd. All rights reserved.oi:10.1016/j.vaccine.2009.08.101

uggests a clinical benefit over (−)MF59 influenza vaccines.© 2009 Elsevier Ltd. All rights reserved.

One strategy to enhance the magnitude of the immune responseto influenza vaccination is the addition of adjuvants. MF59®

is an oil-in-water adjuvant emulsion [5]. Trivalent inactivated-subunit influenza vaccine adjuvanted with MF59 has shown ahigher immunogenicity compared with non-adjuvanted vaccinesin the elderly [5–12]. In addition, broader immune responseswere observed with influenza vaccines adjuvanted with MF59,demonstrating the ability to confer higher immunogenicity againstmismatched influenza virus strains than conventional vaccines[6,8,13–15].

In a meta-analysis of 20 clinical trials published in 2001,MF59-adjuvanted seasonal influenza vaccine was found to be welltolerated [5]. Since then, a large number of clinical trials have beenconducted with MF59 adjuvant with both seasonal and pandemicvaccines. The commercial MF59-adjuvanted seasonal influenza

vaccine (FluadTM) was first approved for active prophylaxis ofinfluenza in the elderly in 1997 [16], and is currently licensed in 26countries worldwide, with more than 45 million doses distributed.

The present, integrated database analysis was designed to eval-uate the safety profile of MF59-containing seasonal and pandemic

6 ccine

iacatiolTra

2

fpci(etmviwyytw

wcwmhgp

oiclpgcobf

dtasdc(odfsdjrp

960 M. Pellegrini et al. / Va

nfluenza vaccines — (+)MF59 — compared with non-MF59-djuvanted influenza vaccines — (−)MF59 — using data from 64linical trials. The safety analysis included a detailed analysis ofdverse events (AEs) of potential autoimmune origin, as the ques-ion has been raised if vaccines in general, and adjuvanted vaccinesn particular, can trigger autoimmune diseases [17,18]. The analysisf safety outcomes was performed for the overall enrolled popu-ation, and for the subgroup of elderly subjects (age ≥65 years).he elderly represents the population for which FluadTM is cur-ently licensed. For both the overall and the elderly population, thenalysis was conducted for all trials and for controlled trials only.

. Methods

Analyses were performed of all studies using influenza vaccinesormulated with the MF59 adjuvant that were executed in com-liance with good clinical practice. The analysis was based on 64linical trials conducted between the 1992–1993 and 2007–2008nfluenza seasons in Europe, USA, Australia, and South AmericaAppendix A). Of these 64 trials, 53 were parent studies, and 11 werextension studies of a primary study. The studies ranged in dura-ion from 3 weeks to >12 months, and included individuals from 6

onths to 100 years of age. Analyses included antigen and/or adju-ant dose-ranging studies. Thirty-eight of the 64 studies includedn the main analysis were controlled trials, and a subgroup analysis

as carried out in these trials (Appendix A). Results were anal-sed for the overall population and for elderly subjects aged ≥65ears. The study protocols were approved by the relevant institu-ional review boards or ethics committees and all participants gaveritten informed consent before the start of the study.

Inclusion and exclusion criteria were typical of vaccine studies,ith some variation in the different trials, but there were suffi-

iently consistent to allow this analysis. Common inclusion criteriaere that subjects were able to comply with all the study require-ents, were in general good health as determined by medical

istory; physical examination and clinical judgment of the investi-ator; and were willing/able to provide written informed consentrior to study entry.

Common exclusion criteria were the subjects having a knownr suspected impairment of their immune function (exclud-ng that normally associated with advancing age), any serioushronic disease or known allergy to any vaccine components; anyaboratory-confirmed influenza or vaccination against influenzarior to enrollment; or receipt of any other vaccine or investi-ational agent in the months prior to enrollment. Subjects withurrent infectious disease, including those taking systemic antibi-tics or antivirals; women who are pregnant, or women able toear children but not willing to practice acceptable contraceptionor the first 3 weeks of the duration of the trial were also excluded.

In the two studies in which adults with underlying chronic con-itions were specifically studied subjects were willing and ableo give written informed consent prior to study entry; and wereble to comply with all the study requirements. In study V70P3ubjects were suffering from at least one of the following chroniciseases: moderate to severe hypertension, moderate to severeongestive heart failure, chronic obstructive pulmonary diseaseCOPD) or moderate to severe asthma, moderate to severe hepaticr renal insufficiency; atherosclerotic disease or insulin-dependentiabetes mellitus. In study V70P4 subjects with at least one of theollowing chronic diseases were enrolled: heart diseases, hyperten-

ion, COPD or asthma, hepatic or renal insufficiency, arterioscleroticisease or insulin dependent diabetes mellitus. In study V7P39 sub-

ects had a history of successful allogenic renal transplantation,equiring maintenance immunosuppressive therapy for at least therevious 6 weeks.

27 (2009) 6959–6965

Subjects in the (+)MF59 group received one, two or threeinjections of the current formulation of either MF59-adjuvantedseasonal influenza vaccine (FluadTM) or MF59-adjuvanted pan-demic H5N1 (AflunovTM) or H5N3 influenza vaccines, includingboth thiomersal-free and thiomersal-containing formulations. Sub-jects in the (−)MF59 group received conventional non-adjuvantedinfluenza vaccines (AgrippalTM, alpha.RIXTM, InfluvacTM, InflusplitSSWTM, FluarixTM, FluogenTM, FlushieldTM, FluzoneTM, FluvirinTM,Vacuna Antigripal ConnaughtTM or VaxigripTM). The majority ofcomparisons were performed between FluadTM and AgrippalTM,which have the same production process and antigen content anddiffer only in the inclusion of MF59 in the FluadTM formulation.

2.1. Analysis

Safety data from the studies were pooled. Each extension studyand its corresponding parent study were considered as one study.Groups within each study were rearranged according to whether ornot the vaccine contained MF59, combining dose groups in case ofdose-ranging trials. Post-vaccination reactogenicity was assessedby the incidence of: solicited local reactions (e.g., pain, tempera-ture [warmth], induration, erythema at the injection site); solicitedsystemic reactions (e.g., myalgia, headache, fatigue and malaise);and other reactions (reported use of analgesic/antipyretic medica-tions, stayed at home due to the reaction and/or had fever [axillarytemperature ≥38 ◦C]) occurring at days 0–3 after first vaccination.Trials where such reporting was not required by the study protocolwere excluded from this analysis (V7P35 and V87P5).

The following unsolicited outcomes occurring at any time dur-ing the study were analysed: any unsolicited AE, AEs of potentialautoimmune origin, cardiovascular AEs, new onset of chronic dis-eases (NOCDs), serious AEs (SAEs), hospitalisations and deaths.AEs occurring at any time during the study were mapped topreferred terms according to the Medicinal Dictionary for DrugRegulatory Activities (MedDRA). One study (V7P35, n = 13,761) wasexcluded from the analysis of any unsolicited AE, AEs of potentialautoimmune origin, and cardiovascular AEs because this study wasdesigned to collect data on AEs necessitating a physician’s visit onlyoccurring during days 0–7. This study was included in the analysisof SAEs, NOCDs, hospitalisations and deaths, because these eventswere collected throughout the study.

The occurrence of AEs of potential autoimmune origin wasassessed in three different ways:

(1) A list of over 30 MedDRA terms compatible with events ofautoimmune origin was predefined by the authors (AppendixB).

(2) Potential cases based on the above list of MedRA terms wereassessed by three independent external experts blinded to thevaccine group (MF59-adjuvanted vs non-adjuvanted). Autoim-mune disease was confirmed if at least one of the three expertsconfirmed a case.

(3) The list of MedDRA terms for AEs of potential autoimmune ori-gin as defined by Verstraeten et al. [18] was used (AppendixB).

For solicited local and systemic reactions reported during days0–3, weighted risk ratios (RRs) — (+)MF59/(−)MF59 — were calcu-lated using the pooled Mantel-Haenszel type estimator, adjustedfor study with 2-sided 95% confidence intervals (CIs). The inci-dence of unsolicited AEs occurring at any time during the study

was estimated as crude rate, i.e., irrespective of the study dura-tion and number of vaccinations, and 2-sided 95% CI limits werecalculated according to Clopper and Pearson. A Poisson regressionmodel with a logarithmic link function was applied to calculatean adjusted RR for (+)MF59/(−)MF59 and 2-sided 95% CI, taking

M. Pellegrini et al. / Vaccine 27 (2009) 6959–6965 6961

Table 1Demographics and other baseline characteristics of the enrolled population.

Characteristic Overall populationa (n = 27,998) Elderly subjects (≥65 years)b (n = 19,596)

(+)MF59 (−)MF59 (+)MF59 (−)MF59

All trials (n) 20,466 7532 13,292 6304Mean age (years) 60.2 68.1 74.6 75.0Male (%) 48 48 49 48

Race (%)Caucasian 51 35 29 25Other 3 1 1 <1No data available 46 64 70 75

Controlled trials (n) 13,217 7532 11,468 6304Mean age (years) 70.1 68.1 74.9 75.0Male (%) 48 48 48 48

Race (%)Caucasian 27 35 18 25Other 2 1 <1 <1

( uenzaudedded in

iv

3

3

idc≥(

o1vc(a

it1t

TS

A

No data available 71

+)MF59: MF59-adjuvanted influenza vaccine; (−)MF59: non-MF59-containing infla Includes all 27,998 subjects enrolled, of whom 27,973 were vaccinated and inclb Includes all 19,596 elderly subjects, of whom 19,590 were vaccinated and inclu

nto account the number of days in the study and the number ofaccinations. Significance was claimed if the 95% CI excluded 1.

. Results

.1. Demographics

A total of 27,998 subjects were enrolled in the 64 trials includedn the clinical dataset. Of these, 27,973 subjects received at least oneose of vaccine. A total of 20,447 subjects received (+)MF59 vac-ines and 7526 received (−)MF59 vaccines. Elderly subjects (aged65 years) made up 65% of the population in the (+)MF59 group

n = 13,287) and 84% in the (−)MF59 group (n = 6303).A total of 20,749 subjects were enrolled in controlled trials

f whom 20,730 received at least one dose of vaccine. Of these,3,204 received (+)MF59 vaccines and 7526 received (−)MF59accines. In this subgroup, 87% of subjects receiving (+)MF59 vac-ines (n = 11,465) and 84% of subjects receiving (−)MF59 vaccinesn = 6303) were ≥65 years. Demographic and other baseline char-cteristics were generally balanced between the groups (Table 1).

Periods of adverse event monitoring varied between trials. In allnfluenza trials for the overall population the mean (±SD) observa-ion days were 168.8 (±80.9) for subjects in the (+)MF59 group and78.3 (±72.3) for those in the (−)MF59 vaccine group. In active con-rolled trials in the elderly mean (±SD) observation days were 184.3

able 2ubjects with AEs of potential autoimmune origin occurring at any time during the trial.

Overall population

Subjects with any AE, n (per 1000) Adjuste

(+)MF59 (−)MF59

All trials (n) 11,243 2969Predefined MedDRA terms 8 (0.71) 2 (0.67) 2.26 (0.Cases confirmed by blindedexperts

5 (0.44) 2 (0.67) 1.28 (0.

MedDRA terms as defined byVerstraeten et al. [18]

39 (3.47) 18 (6.06) 0.87 (0.

Controlled trials (n) 4000 2969Predefined MedDRA terms 4 (1.00) 2 (0.67) 1.91 (0.Cases confirmed by blindedexperts

2 (0.50) 2 (0.67) 0.96 (0.

MedDRA terms as defined byVerstraeten et al. [18]

21 (5.25) 18 (6.06) 1.01 (0.

E: adverse event; (+)MF59: MF59-adjuvanted influenza vaccine; (−)MF59: non-MF59-c

64 82 75

vaccine.in the further analyses.the further analyses.

(±60.7) and 185.2 (±59.9), respectively. In uncontrolled studies inthe elderly the observation period was 65.7 (±91.7) days.

3.2. AEs of potential autoimmune origin

A low number of unsolicited AEs of potential autoimmune ori-gin were reported in the analysis of all trials, with 8 cases out of11,243 subjects (0.71 per 1000) in the (+)MF59 group and 2 casesout of 2969 subjects (0.67 per 1000) in the (−)MF59 group (adjustedRR 2.26; 95% CI 0.44–11.47). Similar rates between (+)MF59 and(−)MF59 groups were confirmed in the controlled trial populationand in the elderly subgroups (Table 2).

Three independent, blinded, external experts given full accessto the clinical data profiles of these cases confirmed 5 of the 8 casesof autoimmune disease in the (+)MF59 group (0.44 per 1000): 2cases of Crohn’s disease and single cases of type 1 diabetes mellitus,rheumatoid arthritis and multiple sclerosis. The experts confirmedboth cases in the (−)MF59 group, namely one case of rheuma-toid arthritis and one of temporal arteritis. In the controlled trialdatabase, 2 cases of confirmed autoimmune disease occurred in

the (+)MF59 group and 2 cases in the (−)MF59 group, all of whichwere reported in elderly subjects.

The analysis based on the broader criteria used by Verstraeten etal. [18] confirmed no significant differences in the incidences of AEsof potential autoimmune origin. In the overall population, the fre-

Elderly subjects (≥65 years)

d RR (95% CI) Subjects with any AE, n (per 1000) Adjusted RR (95% CI)

(+)MF59 (−)MF59

4115 175344–11.47) 3 (0.73) 2 (1.14) 1.51 (0.25–9.04)22–7.53) 2 (0.49) 2 (1.14) 1.00 (0.14–7.14)

47–1.61) 24 (5.83) 17 (9.70) 0.95 (0.50–1.82)

2293 175335–10.43) 3 (1.31) 2 (1.14) 1.55 (0.26–9.28)13–6.79) 2 (0.87) 2 (1.14) 1.03 (0.15–7.34)

54–1.90) 21 (9.16) 17 (9.70) 1.13 (0.60–2.16)

ontaining influenza vaccine; RR: relative risk.

6962 M. Pellegrini et al. / Vaccine 27 (2009) 6959–6965

Table 3Subjects with unsolicited AEs occurring at any time during the trial.

Overall population Elderly subjects (≥65 years)

Subjects with any AE, n (per 1000) Adjusted RR (95% CI) Subjects with any AE, n (per 1000) Adjusted RR (95% CI)

(+)MF59 (−)MF59 (+)MF59 (−)MF59

All trials (n) 11,243a or 20,447b 2969a or 7526b 4115a or 13,287b 1753a or 6303b

All AEsa 3010 (267.7) 1163 (391.7) 0.65 (0.60–0.70) 875 (212.6) 617 (352.0) 0.73 (0.66–0.81)Cardiovascular diseasea 219 (19.5) 165 (55.6) 0.44 (0.35–0.55) 177 (43.0) 160 (91.3) 0.58 (0.47–0.73)NOCDb 269 (13.2) 145 (19.3) 0.71 (0.57–0.87) 205 (15.4) 136 (21.6) 0.73 (0.59–0.91)SAEb 1019 (49.8) 528 (70.2) 0.86 (0.77–0.95) 913 (68.7) 509 (80.8) 0.89 (0.80–0.99)Hospitalisationb 929 (45.4) 478 (63.5) 0.88 (0.79–0.99) 827 (62.2) 461 (73.1) 0.91 (0.81–1.02)Deathb 166 (8.12) 88 (11.7) 0.67 (0.51–0.87) 163 (12.3) 88 (14.0) 0.70 (0.54–0.91)

Controlled trials (n) 4000a or 13,204b 2969a or 7526b 2293a or 11,465b 1753a or 6303b

All AEsa 1292 (323.0) 1163 (391.7) 0.85 (0.79–0.92) 671 (292.6) 617 (352.0) 0.91 (0.82–1.02)Cardiovascular diseasea 160 (40.0) 165 (55.6) 0.78 (0.63–0.97) 154 (67.2) 160 (91.3) 0.82 (0.66–1.02)NOCDb 196 (14.8) 145 (19.3) 0.76 (0.62–0.95) 188 (16.4) 136 (21.6) 0.70 (0.61–0.95)SAEb 900 (68.2) 528 (70.2) 0.96 (0.86–1.06) 883 (77.0) 509 (80.8) 0.95 (0.85–1.06)Hospitalisationb 813 (61.6) 478 (63.5) 0.95 (0.85–1.06) 798 (69.6) 461 (73.1) 0.94 (0.84–1.06)Deathb 159 (12.0) 88 (11.7) 1.01 (0.77–1.31) 159 (13.9) 88 (14.0) 0.97 (0.75–1.26)

Shaded adjusted RRs denote that the 95% CI is <1, indicating statistical evidence of a decreased risk with (+)MF59.A vaccind

qpgoss(m0i

3

ewApwAs(a(qh

TS

SRA

E: adverse event; CI: confidence interval; (+)MF59: MF59-adjuvanted influenzaisease; RR: relative risk; SAE: serious adverse event.a Excluding study V7P35 (Overall n = 13,761).b Including study V7P35 (Overall n = 13,761).

uency of events compatible with autoimmune diseases was 3.47er 1000 in the (+)MF59 group and 6.06 per 1000 in the (−)MF59roup (adjusted RR 0.87; 95% CI 0.47–1.61). Similar results werebserved in the controlled trial population, as well as in the elderlyubgroups (Table 2). In the non-elderly population 15 of the 7128ubjects (2.1/1000) in the (+)MF59 group, and 1 of the 1216 subjects0.82/1000) in the (−)MF59 group had an AE of potential autoim-

une origin. The unadjusted RR of 2.56 has a two-tailed 95% CI of.34–19.35, which includes an RR of 1 indicating that there is no

ncreased risk in non-elderly subjects.

.3. Unsolicited AEs

In the analysis of other types of unsolicited AEs, there was a gen-ral tendency towards fewer events rates with (+)MF59 comparedith (−)MF59 (Table 3). A significantly lower risk of any unsolicitedE was reported in subjects who received (+)MF59 vaccines com-ared with (−)MF59 vaccines in the overall population in all trials,hich was confirmed in the analysis of controlled trials (Table 3).lower risk was also observed in the elderly subgroups, although

ignificance was maintained only in the analysis based on all trials

Table 3). The most frequent AEs in the (+)MF59 group of the over-ll population were nasopharyngitis, rhinitis, headache and coughall reported by 2% of subjects). In the (−)MF59 group, the most fre-uent AEs were upper respiratory tract infection (3%), cough (2%),eadache (2%) and rhinitis (2%). The most frequently reported AEs

able 4ubjects with solicited AEs occurring on days 0–3 after first vaccination.

Overall population

Subjects with any solicited AE, n (per 1000) Weighted RR (95

(+)MF59 (n = 10,238) (−)MF59 (n = 2961)

Any reaction 5986 (584.7) 1390 (469.4) 1.34 (1.28–1.40)Local reactiona 4815 (470.3) 842 (284.4) 1.71 (1.61–1.82)Systemic reactionb 3096 (302.4) 581 (196.2) 1.33 (1.22–1.46)Other reactionsc 1110 (108.4) 423 (142.9) 1.23 (1.10–1.37)

haded weighted RRs denote that the 95% CI is >1, indicating statistical evidence of an inRs were identical for the analysis of all trials and controlled trials only.E: adverse event; CI: confidence interval; (+)MF59: MF59-adjuvanted influenza vaccinea Including pain, warmth, induration and erythema at the injection site.b Including myalgia, headache, fatigue and malaise.c Where subjects reported the use of analgesic/antipyretic medications, stayed at hom

e; (−)MF59: non-MF59-containing influenza vaccine; NOCD: new onset chronic

in the elderly (+)MF59 group were bronchitis (54 subjects, <1%),injury, myalgia, headache and injection site erythema (36 sub-jects each, <1%). The most frequently reported AEs in the elderly(−)MF59 group were bronchitis, hypertension, injury, urinary tractinfections and respiratory disorders (2–3%).

(+)MF59 vaccines were also associated with a lower risk ofcardiovascular AEs compared with (−)MF59 in the analysis ofthe overall integrated database. Hypertension, cerebral ischaemia,congestive cardiac failure and haemorrhage were the most com-monly reported AEs in both groups. The majority of cardiovasculardiseases occurred in elderly subjects. Results were similar whenconsidering controlled trial data only (Table 3).

NOCD occurred in 13.2 per 1000 (+)MF59 subjects comparedwith 19.3 per 1000 (−)MF59 subjects in the overall population inall trials: this difference was significantly in favour of the (+)MF59group (Table 3). A significantly lower incidence with (+)MF59was consistently shown in the controlled trials and elderly sub-groups (Table 3). The most commonly reported NOCDs includedosteoarthritis, atrial fibrillation and hypertension.

SAEs occurred less frequently with (+)MF59 compared with(−)MF59 in the overall population (Table 3). The trend was sim-ilar in the elderly subgroup. In the analysis of controlled trials a

significant difference was not confirmed in the overall populationor the elderly.

Hospitalisation rates were lower for (+)MF59 subjects comparedwith the (−)MF59 group in the overall population of all trials.

Elderly subjects (≥65 years)

% CI) Subjects with any solicited AE, n (per 1000) Weighted RR (95% CI)

(+)MF59 (n = 4109) (−)MF59 (n = 1750)

1823 (443.7) 684 (390.9) 1.32 (1.23–1.41)1285 (312.7) 382 (218.3) 1.74 (1.57–1.94)

725 (176.4) 212 (121.1) 1.29 (1.10–1.52)535 (130.2) 279 (159.4) 1.08 (0.95–1.23)

creased risk with (+)MF59. As a result of the statistical analysis used, the weighted

; (−)MF59: non-MF59-containing influenza vaccine; RR: relative risk.

e due to the reaction and/or had fever (axillary temperature ≥38 ◦C).

ccine

Hi

jat<idti

3

oigos(wTwr

4

piorc

atvfitiiod

rdamcjrnbtrcswpfv

iuc

M. Pellegrini et al. / Va

ospitalisation rates were similar between (+)MF59 and (−)MF59n the analysis of controlled trials, and in elderly subjects (Table 3).

In the analysis of all trials, death occurred in 8.12 per 1000 sub-ects administered (+)MF59 vaccines and 11.7 per 1000 subjectsdministered (−)MF59 vaccines. The majority of deaths occurred inhe elderly population, only three deaths being reported in subjects65 years. In the overall analysis, 12.3 per 1000 elderly subjectsn the (+)MF59 group and 14.0 per 1000 in the (−)MF59 groupied (adjusted RR 0.70; 95% CI 0.54–0.91). In the analysis of con-rolled trial data only, all deaths occurred in the elderly at a similarncidence between (+)MF59 and (−)MF59 vaccines (Table 3).

.4. Solicited AEs

The analysis of solicited local and systemic reactions thatccurred 0–3 days after first vaccination revealed a higher raten subjects who received (+)MF59 vaccines compared with thoseiven (−)MF59 vaccines (Table 4). This increase was seen in theverall study population and the elderly subgroups in the analy-is of all studies and controlled trials only (Table 4). In both the+)MF59 and (−)MF59 groups, the majority of solicited reactionsere local, and mild or moderate in intensity (results not shown).

he most commonly reported AEs included pain, injection-sitearmth, induration and erythema. The most common systemic

eactions were myalgia, headache, fatigue and malaise.

. Discussion

This large-scale integrated safety analysis of data from 64 trialsrovides evidence for the good safety profile of MF59-adjuvanted

nfluenza vaccines. Trends in results were consistent between theverall study population and the elderly subgroup. In addition,esults derived from all trials in the safety database were generallyonsistent with results using only data from controlled trials.

Results from the present analysis are consistent with those frompublished meta-analysis of over 10,000 subjects in 20 clinical

rials, which concluded that MF59-adjuvanted seasonal influenzaaccines are well tolerated [5]. The current analysis extends thesendings by including data from an additional 44 trials and morehan 20,000 people vaccinated with both seasonal and pandemicnfluenza vaccines. In addition, it provides information on thempact of MF59-adjuvanted influenza vaccines on a wider rangef safety outcomes including a detailed analysis of autoimmuneiseases.

Autoimmune diseases have a low prevalence, and the potentialisk of autoimmune disease after vaccination is a topic of someebate [17], particularly with reference to the impact of noveldjuvants [18]. In the current analysis, AEs compatible with autoim-une diseases were analysed in three different ways. Potential

ases identified with predefined MedDRA terms in 11,243 sub-ects vaccinated with MF59-containing influenza vaccines wereare (0.71 per 1000) and showed no significant difference fromon-adjuvanted controls. Autoimmune disease cases confirmed bylinded independent experts occurred in 0.44 per 1000 subjectsreated with MF59-adjuvanted influenza vaccines; and again, thisate did not differ significantly from controls. Finally, potentialases were identified with MedDRA terms used in the paper by Ver-traeten et al. [18] to allow better comparison with results obtainedith a different adjuvant (AS04). Using their criteria for AEs ofotential autoimmune origin, we also showed no significant dif-erences between MF59-adjuvanted and non-adjuvanted influenza

accines.

MF59-adjuvanted influenza vaccines are not associated with anncreased incidence of unsolicited AEs in the overall study pop-lation or elderly, whether extended to all trials or restricted toontrolled trials. Instead, generally lower rates of unsolicited AEs

27 (2009) 6959–6965 6963

were observed. Surprisingly, MF59-adjuvanted vaccines are asso-ciated with a lower rate of NOCD and cardiovascular disease thannon-MF59-containing vaccines. These results add to the previousliterature, which contains three case–control studies (n = 105,454)which suggest that MF59-adjuvanted subunit influenza vaccinereduces the risk of hospitalisation for acute coronary syndromes(adjusted odds ratio [OR] 0.13; 95% CI 0.03–0.65), cerebrovascu-lar accidents (adjusted OR 0.07; 95% CI 0.01–0.48) and pneumonia(adjusted OR 0.31; 95% CI 0.14–0.71) in the elderly compared withno influenza vaccination [19].

A 30% risk reduction in mortality with MF59-adjuvantedinfluenza vaccines was suggested in the complete analysis of thesafety database, but was not shown in the pooled analysis ofonly controlled trial data. A beneficial effect of MF59-adjuvantedinfluenza vaccines on mortality compared with non-adjuvantedvaccines has been suggested previously in a study of elderly sub-jects (n = 1807) who were immunised with adjuvanted vaccine.They had a 60% lower mortality rate than individuals who receivednon-adjuvanted vaccines (8 vs 17 deaths; RR 0.39; 95% CI 0.17–0.91)[20].

Although the underlying reason for a potential relationshipbetween MF59-adjuvanted vaccines and cardiovascular diseaseand mortality remains to be elucidated, it can be hypothesised thatthe higher immunogenicity of MF59-adjuvanted influenza vaccines[4–12] strengthens the suggested benefit of influenza vaccinationcompared with no vaccination in the elderly [21,22]. This study alsoindicates a lower rate of NOCD with MF59-adjuvanted influenzavaccines, but there is little information about a potential relation-ship between influenza and NOCD.

Administration of MF59-adjuvanted vaccines was associatedwith an increased incidence of solicited local and systemic reac-tions within 3 days of vaccination, which were generally mild andtransient in nature. The higher frequency of local and systemic reac-tions was expected, and was consistent with previous studies ofMF59-adjuvanted vaccines, validating the outcome of this analysis[4–8].

The analysis has some limitations associated with the poolingof a number of individual studies, particularly regarding inter-trial variability. For example, the safety monitoring periods ofthe included studies varied from 3 weeks to approximately 20months. In addition, the pooled studies used different antigens,different amounts of MF59, and different numbers of vaccina-tions (1–3). The number of vaccinations and the number of studydays were used as quantitative covariates in the Poisson regres-sion model to calculate adjusted RRs. Data on baseline healthconditions were limited. Although the majority of studies werecontrolled trials, the primary analysis also included several uncon-trolled open-label trials. Analysis of pooled individual data, asopposed to a study-based meta-analysis, may skew the outcomestowards the results of the studies with the largest number ofsubjects. Despite these limitations, the results of the presentanalysis provide strong evidence for the good safety profile ofMF59-adjuvanted influenza vaccines in >20,000 subjects world-wide.

To conclude, this large-scale integrated analysis supports thegood safety profile associated with MF59-adjuvanted influenzavaccines and suggests there may be a clinical benefit overnon-MF59-containing vaccines. Further, the demonstrated goodtolerability of MF59-containing vaccines removes a major concernover the use of this adjuvant in large-scale immunisation cam-paigns, such as would be required in the event of an influenza

pandemic. In these circumstances adjuvantation will be requiredto overcome the problems of limited supply of antigen withinherent low immunogenicity, and the difficulties in havingan exact match with the infecting strain, so requiring cross-reactivity.

6 ccine

A

TttsN

Aa

964 M. Pellegrini et al. / Va

cknowledgments

All authors are employees of Novartis Vaccines and Diagnostics.he authors received editorial/writing support in the preparation ofhis manuscript, funded by Novartis Vaccines and Diagnostics. Hes-er van Lier, PhD of Excerpta Medica provided the editorial/writingupport. The authors would like to thank Stefan Hofmann, Pantaleoacci, and Michael C. Penlington for their contributions.

ppendix A. Details of the clinical trials included in thenalysis

Study Influenza vaccines used Sample sizea

Controlled trialsV7P37, V7P37E1b Fluad-H5N3,

non-adjuvanted H5N365

V25P1 Fluad, Agrippal 74V57P1 Adjuvanted

H5N3/H3N2/B, intranasalvaccine

100

V95P1 Agrippal, Fluad,Fluad + CpG7909

60

V7P3, V7P3X1b, V7P3X2b Fluad, Agrippal 92V7P5, V7P5X1b, V7P5X2b Fluad, Agrippal 317V7P6 Fluad, Agrippal, Influvac 143V7P7, V7P7X1b Fluad, Agrippal 214V7P8, V7P8X1b Fluad, Agrippal 308V7P9 Fluad, Agrippal 209V7P12 Fluad, Influvac/MF59,

Influvac312

V7P15 Fluad, Influvac/MF59,Vacuna AntigripalConnaught

291

V7P17 Fluad, Agrippal 310V7P18, V7P18X1b Fluad, Fluzone 301V7P24 Fluad, Flushield 301V7P27 Fluad, Flushield, Fluvirin 308V7P29 Fluad, Fluogen/Flushield 216V7P30 Fluad, Agrippal 451V7P34 Fluad, Agrippal 560V7P38 Fluad, Agrippal 487V7P39 Fluad, Agrippal 113V87P2 Aflunov, non-adjuvanted

A/H5N140

V70P2, V70P2E1b Fluad, Vaxigrip 269V7P16 Influvac/MF59, Vaxigrip 279V7P25, V7P25X1b Fluad, Vaxigrip 283V7P26 Fluad, �.Rix 142M63P1 Fluad, Fluarix 350V7P35 Fluad, Influvac 13,761V70P3 Fluad, Agrippal 361

Uncontrolled/Ongoing controlled (blinded) trialsV87P3 Aflunov 58V104P1 Fluad, Fluad + IC31 72V7P32 Fluad 50V7P1S Fluad 111V7P2S Fluad 108V7P3S Fluad 108V7P4S Fluad 63V7P5S Fluad 54V63P1S Fluad 81V70P1S Fluad 53V70P2S Fluad 61V70P3S Fluad 60V70P4S Fluad 57V70P5S Fluad 56V70P7S Fluad 64V70P4 Fluad 52

V87P5 Aflunov, Agrippal 405V7P31 Fluad 129V63P1 Fluad, preservative-free

Fluad295

V87P6 Aflunov, Fluad 469V70P6 Fluad, Fluzone 350

27 (2009) 6959–6965

V70P1 Fluad, thiomersal-freeFluad

299

V87P1, V87P1E1b Aflunov 486V101P1 Aflunov 600V70P5 Fluad, Agrippal,

INFLUSPLIT SSW3300

V87P4 Aflunov, Fluad 3516

MF59-adjuvanted influenza vaccines: Aflunov (pandemic H5N1), Fluad-H5N3 (pan-demic H5N3), Fluad (seasonal influenza vaccine), Fluad + CpG7909, Fluad + IC31,MF59-adjuvanted Influvac, MF59-adjuvanted H5N3/H3N2/B influenza vaccine.Conventional non-adjuvanted influenza vaccines: �.RIX, Agrippal, Influvac, Fluarix,Fluogen, Flushield, Fluzone, Fluvirin, Vacuna Antigripal Connaught, Vaxigrip, Influs-plit SSW.

a Overall sample size including subjects exposed to influenza vaccines (±MF59)and placebo/non-influenza vaccines.

b Extension study.

Appendix B. MedDRA preferred terms and SMQs used toidentify adverse events of potential autoimmune origin

Predefined MedDRA terms byauthors of this paper

MedDRA terms defined by Verstraeten etal. [18]

MedDRA terms that overlap between the two definitionsLeukoencephalomyelitis LeukoencephalomyelitisSMQ: non-infectious

encephalitisaEncephalitisb, encephalitispost-immunisationb

Addison’s disease Addison’s diseaseAntiphospholipid syndrome Antiphospholipid syndromeWarm type haemolytic

anaemiaWarm type haemolytic anaemia

SMQs: haematopoieticcytopaeniasa, haemolyticdisordersa

Anaemia haemolytic autoimmuneb, coldtype haemolytic anaemiab, Coombspositive haemolytic anaemiab, haemolyticanaemiab, thrombocytopeniab,autoimmune thrombocytopenia

Autoimmune hepatitis Autoimmune hepatitisCoeliac disease Coeliac diseaseCrohn’s disease Crohn’s disease, inflammatory bowel

disease, proctitis ulcerative, ulcerativecolitis

Type I diabetes mellitus Diabetes mellitus, diabetes mellitusinsulin-dependent

Basedow’s disease Basedow’s diseaseGuillain-Barré syndrome,

SMQ: Guillain-Barrésyndromea

Guillain-Barré syndrome

Autoimmune thyroiditis Autoimmune thyroiditisIdiopathic thrombocytopenic

purpuraIdiopathic thrombocytopenic purpura

Systemic lupuserythematosus

Systemic lupus erythematosus

SMQ: systemic lupuserythematosusa

Systemic lupus erythematosus rashb, lupusnephritisb, cutaneous lupuserythematosusb

Multiple sclerosis Multiple sclerosisMyasthenia gravis Myasthenia gravisOptic neuritis Optic neuritis, Optic neuritis retrobulbarRheumatoid arthritis Rheumatoid arthritis, arthritis reactive,

arthritis, juvenile arthritis,Sjogren’s syndrome Sjogren’s syndrome

Different MedDRA terms used in the two definitionsAnkylosing spondylitis,

oophoritis, Goodpasture’ssyndrome, Kawasaki’s disease,mixed connective tissuedisease, opsoclonusmyoclonus, pemphigus, biliarycirrhosis, Reiter’s syndrome,Takayasu’s arteritis, giant cellarteritis, Wegener’sgranulomatosis, aplasticanaemia, pernicious anaemia

Myelitis, myelitis transverse,demyelination, dermatomyositis, psoriasis,vitiligo, psoriatic arthropathy, erythemanodosum, Stevens-Johnson syndrome,Raynaud’s phenomenon, Behcet’ssyndrome, nephritis, nephritisautoimmune, glomerulonephritis, uveitis,leukocytoclastic vasculitis, vasculitis,sarcoidosis, hyperthyroidism, thyroiditis,goiter, thyroiditis acute, hypothyroidicgoiter, thyroiditis subacute, scleroderma

MedDRA: Medicinal Dictionary for Drug Regulatory Activities; SMQ: StandardisedMedDRA Query.

a SMQ includes a list of MedDRA preferred terms.b MedDRA term used by Verstraeten et al. [18] that is included in the MedDRA

term list of the corresponding SMQ defined by the authors of this paper.

ccine

R

[

[

[

[

[

[

[

[

[

[

[

M. Pellegrini et al. / Va

eferences

[1] World Health Organization. Influenza fact sheet no. 211 (March 2003).Available at: http://www.who.int/mediacentre/factsheets/fs211/en/ (accessed31.01.09).

[2] World Health Organization. Weekly Epid Rec 2005;80:277–88.[3] Goodwin K, Viboud C, Simonsen L. Antibody response to influenza vaccination

in the elderly: a quantitative review. Vaccine 2006;24:1159–69.[4] Treanor JJ, Campbell JD, Zangwill KM, Rowe T, Wolff M. Safety and immuno-

genicity of an inactivated subvirion influenza A (H5N1) vaccine. N Engl J Med2006;354:1343–51.

[5] Podda A. The adjuvanted influenza vaccines with novel adjuvants: experiencewith the MF59-adjuvanted vaccine. Vaccine 2001;19:2673–80.

[6] De Donato S, Granoff D, Minutello M, Lecchi G, Faccini M, Agnello M, et al. Safetyand immunogenicity of MF59-adjuvanted influenza vaccine in the elderly. Vac-cine 1999;17:3094–101.

[7] Gasparini R, Pozzi T, Montomoli E, Fragapane E, Senatore F, Minutello M,et al. Increased immunogenicity of the MF59-adjuvanted influenza vaccinecompared to a conventional subunit vaccine in elderly subjects. Eur J Epidemiol2001;17:135–40.

[8] Minutello M, Senatore F, Cecchinelli G, Bianchi M, Andreani T,Podda A, et al. Safety and immunogenicity of an inactivated subunitinfluenza virus vaccine combined with MF59 adjuvant emulsion in elderlysubjects, immunized for three consecutive influenza seasons. Vaccine 1999;17:99–104.

[9] Squarcione S, Sgricia S, Biasio LR, Perinetti E. Comparison of the reactogenicityand immunogenicity of a split and a subunit-adjuvanted influenza vaccine inelderly subjects. Vaccine 2003;21:1268–74.

10] Baldo V, Menegon T, Bonello C, Floreani A, Trivello R, Collaborative Group.Comparison of three different influenza vaccines in institutionalised elderly.Vaccine 2001;19:3472–5.

11] Baldo V, Baldovin T, Floreani A, Minuzzo M, Trivello R. Response to influenzavaccine in people with non-protective HI antibody titers. Eur J Epidemiol2006;21:843–5.

[

[

27 (2009) 6959–6965 6965

12] Banzhoff A, Nacci P, Podda A. A new MF59-adjuvanted influenza vac-cine enhances the immune repose in the elderly with chronic diseases:results from an immunogenicity meta-analysis. Gerontology 2003;49:177–84.

13] Baldo V, Baldovin T, Floreani A, Fragapane E, Trivello R, Family Medicine Group.Response of influenza vaccines against heterovariant influenza virus strains inadults with chronic diseases. J Clin Immunol 2007;27:542–7.

14] Ansaldi F, Baciliari S, Durando P, Sticchi L, Valle L, Montomoli E,et al. Cross-protection by MF59-adjuvanted influenza vaccine: neutralizingand haemagglutination-inhibiting antibody activity against A(H3N2) driftedinfluenza viruses. Vaccine 2008;26:1525–9.

15] Ansaldi F, Canepa P, Parodi V, Baciliari S, Orsi A, Compagnino F, et al. Adju-vanted seasonal influenza vaccines and perpetual viral metamorphosis: theimportance of cross-protection. Vaccine 2009;27:3345–8.

16] Podda A, Del Guidice G, O’Hagan D. MF59: a safe and potent adjuvant for humanuse. In: Schijns D, O’Hagan D, editors. Immunopotentiators in modern vaccines.London: Elsevier Academic Press; 2006. p. 149–59.

17] Wraith DC, Goldman M, Lambert PH. Vaccination and autoimmune disease:what is the evidence? Lancet 2003;362:1659–66.

18] Verstraeten T, Descamps D, David MP, Zahaf T, Hardt K, Izurieta P, et al. Analysisof adverse events of potential autoimmune aetiology in a large integrated safetydatabase of AS04 adjuvanted vaccines. Vaccine 2008;26:6630–8.

19] Puig-Barberà J, Diez-Domingo J, Varea AB, Chavarri GS, Rodrigo JA, Hoyos SP,et al. Effectiveness of MF59-adjuvanted subunit influenza vaccine in prevent-ing hospitalisations for cardiovascular disease, cerebrovascular disease andpneumonia in the elderly. Vaccine 2007;25:7313–21.

20] Martin JT. Development of an adjuvant to enhance the immune response toinfluenza vaccine in the elderly. Biologicals 1997;25:209–13.

21] Jansen AG, Sanders EA, Nichol KL, van Loon AM, Hoes AW, Hak E. Decline ininfluenza-associated mortality among Dutch elderly following the introductionof a nationwide vaccination program. Vaccine 2008;26:5567–74.

22] Ortqvist A, Granath F, Askling J, Hedlund J. Influenza vaccination and mortality:prospective cohort study of the elderly in a large geographical area. Eur RespirJ 2007;30:414–22.