A 1-year follow-on study from a randomised, head-to-head, multicentre, open-label study of two pandemic influenza vaccines in children

Article history

The research reported in this issue of the journal was commissioned by the HTA programme as project number 10/111/01. The contractual start date was in November 2010. The draft report began editorial review in May 2011 and was accepted for publication in August 2011. As the funder, by devising a commissioning brief, the HTA programme specified the research question and study design. The authors have been wholly responsible for all data collection, analysis and interpretation, and for writing up their work. The HTA editors and publisher have tried to ensure the accuracy of the authors’ report and would like to thank the referees for their constructive comments on the draft document. However, they do not accept liability for damages or losses arising from material published in this report.

Declared competing interests of authors

Vaccines used in the original study were manufactured by GlaxoSmithKline and Baxter, both of which donated the vaccine, but had no role in study planning or conduct. The vaccine used in this follow-on study was manufactured by GlaxoSmithKline, from whom it was purchased. AJP, AF, PTH and SF act as chief or principal investigators for clinical trials conducted on behalf of their respective NHS trusts and/or universities, sponsored by vaccine manufacturers, but receive no personal payments from them. KH has been an investigator for clinical trials sponsored by vaccine manufacturers, but received no personal payments from them. AJP, AF, PTH and SF have participated in advisory boards for vaccine manufacturers, but receive no personal payments for this work. MDS, PTH, SF, KH and AF have received financial assistance from vaccine manufacturers to attend conferences. All grants and honoraria are paid into accounts within the respective NHS trusts or universities, or to independent charities.

Copyright statement

© Queen’s Printer and Controller of HMSO 2011. This work was produced by de Whalley et al. under the terms of a commissioning contract issued by the Secretary of State for Health. This journal is a member of and subscribes to the principles of the Committee on Publication Ethics (COPE) (http://www.publicationethics.org/). This journal may be freely reproduced for the purposes of private research and study and may be included in professional journals provided that suitable acknowledgement is made and the reproduction is not associated with any form of advertising. Applications for commercial reproduction should be addressed to: NETSCC, Health Technology Assessment, Alpha House, University of Southampton Science Park, Southampton SO16 7NS, UK.

2011 Queen’s Printer and Controller of HMSO

Participants

In the original study, children aged 6 months to 12 years were recruited by five UK sites (Southampton, Oxford, Bristol, London and Exeter). They were randomised (in a 1 : 1 ratio, with assignment by sequentially numbered, identical, opaque sealed envelopes) to receive two doses, 21 days apart, of either a non-adjuvanted whole-virion H1N1 influenza vaccine or an AS03_B-adjuvanted split-virion H1N1 influenza vaccine. Children who completed the original study were invited to participate in this follow-on study, although those who had subsequently received a further dose of the H1N1 vaccine owing to an insufficient response to the original two doses of vaccine were excluded, as were those who had already received a dose of the 2010–11 trivalent seasonal influenza vaccine. Other exclusion criteria were severe allergic reaction following previous vaccination, suspected unexpected severe adverse reaction in the original study, current egg allergy, impaired immunity, receipt of blood products or > 1 week of systemic steroid treatment within the last 3 months, or participation in another clinical trial. Written informed consent was obtained from a parent or guardian, and verbal assent was sought from children aged ≥ 7 years. Enrolment took place in November and December 2010.

Study design

This was a multicentre, open-label, phase IV study, following on from a randomised trial. At the first study visit, a blood sample was taken to assess the persistence of antibody. A single dose of trivalent seasonal influenza vaccine was given by intramuscular injection (into the deltoid muscle). A second blood sample was taken 21 days later (protocol time window 14–35 days). For those who wanted to participate in the study, but who did not wish to receive the trivalent seasonal influenza vaccine, an option was available to consent to only the first blood test.

The study was approved by the UK Medicines and Healthcare products Regulatory Agency (EudraCT number 2010-022817-24), the Oxfordshire Research Ethics Committee A (10/H0604/81) and the local NHS organisations. The study was registered at ClinicalTrials.gov (NCT01239537).

Vaccines

The original study compared two novel H1N1 vaccines: a non-adjuvanted whole-virion H1N1 influenza vaccine (Celvapan) and an AS03_B-adjuvanted split-virion H1N1 influenza vaccine (Pandemrix).

The non-adjuvanted, whole-virion vaccine was derived from Vero cell culture. Each dose (0.5 ml) contained 7.5 µg of haemagglutinin from influenza A/California/07/2009 (H1N1).

The AS03_B-adjuvanted split-virion H1N1 influenza vaccine was derived from egg culture. Each dose (0.25 ml, half the adult dose) contained 1.875 µg of the haemagglutinin antigen and the oil-in-water emulsion-based adjuvant AS03_B (containing 5.345 mg of squalene, 5.39 mg of dl-α-tocopherol, 2.43 mg of polysorbate 80 and thiomersal).

The present study used the 2010–11 trivalent seasonal influenza vaccine (Fluarix^®, GlaxoSmithKline, Rixensart, Belgium). This contained an inactivated, split-virion influenza virus, propagated in fertilised hens’ eggs. Each 0.5-ml dose contained 15 µg of haemagglutinin for each of the three influenza strains [A/California/07/2009 (H1N1) derived strain (NYMC X-181), A/Perth/16/2009 (H3N2)-like strain (NYMC X-187, derived from A/Victoria/210/2009) and B/Brisbane/60/2008].

Laboratory analysis

Microneutralisation and haemagglutination assays were performed at the Centre for Infections, Health Protection Agency (London, UK), using previously described methods. 8 Sera were processed in 1 : 2 serial dilutions. For MN assays, the initial dilution was 1 : 10 and the final dilution 1 : 5120. For HI assays, the initial dilution was 1 : 8 and the final dilution 1 : 16,384.

Safety and reactogenicity assessments

Adverse events of special interest, defined by the European Medicines Agency19 (see Appendix 1 for further details), that had occurred in the year since receipt of the monovalent pandemic influenza vaccine were recorded.

For 7 days after receipt of the trivalent seasonal influenza vaccine, the following information was recorded in a diary card: daily axillary temperature, injection-site reactions, systemic symptoms and the use of antipyretic medication. Different systemic symptoms were solicited for children aged < 5 years than for older children, to accommodate their limited ability to articulate symptoms. All medical consultations occurring between receipt of the trivalent seasonal influenza vaccine and the second study visit were recorded.

Statistical analysis

For antibody persistence, data were analysed according to vaccine received, following a predetermined statistical plan. For response to trivalent seasonal vaccine, a modified intention-to-treat analysis was performed.

To enable an unbiased assessment of the persistence of antibody, a random sample of participants excluded owing to receipt of a third dose were included in the analysis, assuming that their (low) antibody titre was unchanged from its value after two doses of vaccine in the original study. 17,18 The number in this random sample was determined proportionately (the proportion of participants excluded owing to receipt of a third dose who were in the selected sample was equal to the proportion of participants in the original study who enrolled in the follow-on study).

For analysis, HI titres of < 8 were given a value of 4, MN titres of < 10 were given a value of 5 and MN titres of > 5120 were given a value of 10,240.

Comparisons between groups were made using Fisher’s exact test. Geometric means were compared using normal error regression on logged titres.

Additional analyses, examining the effect of different variables, were performed using multivariable logistic and normal error regression.

Statistical significance was set at 5%. Analysis was performed in Stata 10.0 (StataCorp LP, College Station, TX, USA).

Participants

Figure 1 shows the flow diagram for study participants. Eight hundred and ninety-four children completed the original study. Of these, 115 could not be contacted, 49 were excluded and 407 elected not to take part in the follow-on study; 323 children enrolled in the follow-on study (36.1% of those completing the original study). Of these, 19 consented to one blood test only (but not to trivalent seasonal influenza vaccination); 290 children attended the second study visit.

FIGURE 1.

Flow diagram of participants. ^aDid not complete the original study. ^bOne invited in error; one randomised to receive the adjuvanted vaccine, but was given the whole-virion vaccine. ^cDid not fulfil the inclusion criteria because was known to have received a third dose of the vaccine. ^dInvited in error, as did not complete the original study. ^eTwo invited in error; one vaccinated just before third birthday, but assigned to the ≥ 3 years group in the original study in error. ^fOne invited in error; one did not fulfil inclusion criteria, owing to suspected unexpected severe adverse reaction in original study. One randomised to receive the adjuvanted vaccine, but was given the whole-virion vaccine. ^gOne vaccinated just before third birthday, but assigned to the ≥ 3 years group in the original study in error. ^hEight found to have received a third dose of pandemic influenza vaccine; two had already received 2010–11 seasonal influenza vaccine; one had received another vaccine in the exclusion period. ⁱThree found to have received a third dose of pandemic influenza vaccine; nine had already received the 2010–11 seasonal influenza vaccine. ^jTwo had already received the 2010–11 seasonal influenza vaccine; two were too unwell to participate; one was taking part in another clinical trial of an investigation medical product; one was inadvertently excluded in error. ^kFive had already received the 2010–11 seasonal influenza vaccine. ^lWithdrew after consenting to vaccination but before receiving the vaccine. ^mEight extras, randomly chosen from those excluded owing to receipt of a third dose of pandemic influenza vaccine, to reduce bias in persistence analysis. ⁿTwo extras, randomly chosen from those excluded owing to receipt of a third dose of pandemic influenza vaccine, to reduce bias in persistence analysis. ^pWithdrew after vaccination but before the second study visit. ^qIncludes two diary cards returned by participants who withdrew after vaccination. ^rAll diary cards for children < 5 years old. ^s Includes one diary card returned by a participant who withdrew after vaccination. ^tEight diary cards for children < 5 years old, 76 diary cards for children ≥ 5 years old. ^uNine diary cards for children < 5 years old and 82 diary cards for children ≥ 5 years old.

Persistence of antibody

Of the 323 participants enrolled, 15 were excluded from the persistence analysis (five did not have antibody assay results after the second dose of pandemic influenza vaccine in the original study and 10 had insufficient blood for the follow-on assay).

After the original study,17,18 28 children received a third dose of pandemic influenza vaccine, owing to both the MN and HI titres being below the protective threshold after two doses. These children were excluded from the follow-on study, as this study was designed to assess the persistence of antibody after two doses of the pandemic vaccine (not three). This introduced a potential bias, because poor responders to the vaccine in the original study were excluded. 17,18 In order to account for this, and to try to obtain an unbiased assessment of the persistence of antibody in a population representative of the participants in the original study, 10 of these 28 (36%), selected randomly, were included in the persistence analysis. They were not enrolled in (and therefore did not have a blood test in) the follow-on study, but it was assumed that their titres remained unchanged from the low values recorded after the second dose of pandemic influenza vaccine in the original study. Hence, a total of 318 children were included in the persistence analysis. Table 1 shows the demographic characteristics of these children.

Four groups were considered in the analysis, defined by the type of pandemic influenza vaccine received in the original study17,18 (AS03_B-adjuvanted or whole-virion) and age at receipt of the first dose of pandemic vaccine (< 3 years or ≥ 3 years).

The median interval from the second dose of pandemic influenza vaccine in the original study to the blood draw to assess the persistence of antibody was 392 days (range 365–413 days). The median interval from the antibody assay after the second dose of pandemic influenza vaccine in the original study to the blood draw to assess the persistence of antibody was 371 days (range 347–399 days).

Table 2 shows the number (and percentage) of children with an MN titre ≥ 1 : 40 at 3 weeks and at 1 year after the second dose of pandemic influenza vaccine. Table 3 shows the MN geometric mean titre 1 year after the second dose of pandemic influenza vaccine. It was not possible to calculate a valid geometric mean titre for MN at 3 weeks after the second dose of the pandemic influenza vaccine because the upper limit of the MN titre measured in the original study was 1 : 320, and the MN titre was > 1 : 320 for many participants.

Table 4 shows the number (and percentage) of participants with an HI titre ≥ 1 : 32 at 3 weeks and at 1 year after the second dose of pandemic influenza vaccine. Table 5 shows the HI geometric mean titre at these time points. Table 6 shows the distribution of fold changes in the HI titre (change from HI titre 3 weeks after the second dose of pandemic influenza vaccine to HI titre 1 year after vaccination).

Figure 2 shows reverse cumulative distribution curves for HI and MN titres, at 3 weeks and at 1 year after receipt of a monovalent pandemic influenza vaccine, analysed according to vaccine given and age at first dose.

FIGURE 2.

Reverse cumulative distribution curves for HI and MN titres, at 3 weeks and at 1 year after receipt of a monovalent pandemic influenza vaccine, analysed according to vaccine given and age at first dose. MN titres are shown only to > 320, as this was the maximum dilution used in the assay at 3 weeks after vaccination.

The unexpected finding that the HI geometric mean titre was greater 1 year after vaccination than at 3 weeks after vaccination in children who had been given the whole-virion vaccine when < 3 years old prompted further analysis of the changes in titres in whole-virion vaccine recipients. Tables 7–9 compare those whose HI titre was ≥ 1 : 32 at 3 weeks after two doses of whole-virion vaccine with those whose HI titre was < 1 : 32 at this time point. For these two subgroups, Table 7 shows the HI geometric mean titre at 3 weeks and at 1 year after the second dose of the whole-virion vaccine, Table 8 shows the distribution of fold changes in the HI titre (change from HI titre 3 weeks after the second dose of pandemic influenza vaccine to HI titre 1 year after vaccination) and Table 9 shows the numbers and percentages with HI titres of ≥ 1 : 32 at 1 year after vaccination. Table 10 compares those recipients of the whole-virion vaccine whose MN titre was ≥ 1 : 40 at 3 weeks after vaccination with those whose MN titre was < 1 : 40, showing numbers and percentages with MN titre ≥ 1 : 40 at 1 year after vaccination.

An additional analysis of the persistence of antibody, modelling the logged HI and MN titres 1 year after pandemic influenza vaccination, is shown in Tables 11 and 12. The variables considered were pandemic vaccine received, age at first dose (< 3 years or ≥ 3 years), gender, whether or not seasonal influenza vaccine had been previously given, interval between post-vaccination blood draw in the original study and the first blood draw in the follow-on study, and study site. Gender and receipt of previous seasonal influenza vaccine did not have a statistically significant effect and these variables are not shown in the tables. Table 11 shows the fold effect on HI and MN titres 1 year after pandemic influenza vaccination. Table 12 shows the fold effect on HI titres 1 year after pandemic influenza vaccination, including the HI titre recorded 3 weeks after pandemic vaccination as a covariate.

Safety of pandemic influenza vaccines

There were no reports of clinically significant adverse events related to the original pandemic influenza vaccination in any of the 323 children enrolled {0/157 [95% confidence interval (CI) 0% to 2.3%] in the whole-virion vaccine group and 0/166 (95% CI 0% to 2.2%) in the AS03_B-adjuvanted vaccine group}. Clinically significant adverse events were determined by the investigators, from medical history given by the parents, and included hospitalisations, influenza-like episodes, febrile convulsions and adverse events of special interest.

Immunogenicity of trivalent seasonal influenza vaccine

A total of 302 children received the 2010–11 trivalent seasonal influenza vaccine in the follow-on study. Table 13 shows the demographic characteristics of these children. After vaccination, sufficient blood for analysis was obtained from 282 of these children. MN and HI titres for the H1N1 component of the vaccine were measured.

The median interval from vaccination to post-vaccination blood sampling was 21 days (range 11–44 days). For nine children, this interval was outside the range of 14–28 days (and for two of these children the interval was outside the range of 14–35 days). These children were included in the modified intention-to-treat analysis.

Table 14 shows the number (and percentage) of children with an MN titre ≥ 1 : 40 immediately before and 3 weeks after receipt of the seasonal influenza vaccine. A valid MN geometric mean titre could not be calculated, as titration was not performed beyond 1 : 5120 and the MN titre after vaccination was > 1 : 5120 for many participants.

Table 15 shows the number (and the percentage) of participants with an HI titre ≥ 1 : 32, immediately before and 3 weeks after receipt of the seasonal influenza vaccine. Table 16 shows the HI geometric mean titres at these time points.

Figure 3 shows reverse cumulative distribution curves for HI and MN titres immediately before and at 3 weeks after receipt of the trivalent seasonal influenza vaccine, analysed according to pandemic vaccine previously given and age at first dose.

FIGURE 3.

Reverse cumulative distribution curves for HI and MN titres immediately before (pre) and 3 weeks after (post) a dose of trivalent seasonal influenza vaccine (TIV), analysed according to pandemic influenza vaccine previously given and age at first dose. MN titres are shown to > 5120, as this was the maximum dilution used in the assay.

Table 17 shows the additional analysis, modelling HI titres 3 weeks after receipt of the trivalent seasonal influenza vaccine, split according to pandemic vaccine previously received. The variables of gender, previous receipt of seasonal influenza vaccine, interval from vaccination to blood assay and study site had no statistically significant effect. There was no age effect for the whole-virion vaccine group, but children given the AS03_B-adjuvanted vaccine before 3 years of age had a significantly greater fold rise in HI titre in response to the trivalent influenza vaccine than did children who had received the AS03_B–adjuvanted vaccine when ≥ 3 years old. HI titre 3 weeks after pandemic vaccination had a statistically significant effect on the response to the trivalent influenza vaccine, whereas HI titre immediately before giving the trivalent influenza vaccine did not.

Reactogenicity of trivalent seasonal influenza vaccine

Diary cards were returned by 295 out of the 302 children who received the trivalent seasonal influenza vaccine. Diary cards were completed during the 5 days after vaccination and most were returned at the second study visit. Five children who were vaccinated, but then withdrew from the study before their second visit returned diary cards.

Between vaccination and completion of (or withdrawal from) the study, there were no serious adverse events.

Table 18 shows local reactions reported by children < 5 years old. Table 19 shows local reactions reported by children ≥ 5 years old.

Table 20 shows systemic reactions reported by children < 5 years old. Table 21 shows local reactions reported by children ≥ 5 years old.

Table 22 shows local reactions, severe systemic symptoms and fever ≥ 38 °C reported by children across both age groups.

Redness and severe local symptoms were reported more frequently in the children < 5 years old who had previously received the AS03_B-adjuvanted pandemic influenza vaccine than in those children who had been given the whole-virion vaccine (p < 0.05). For all other solicited local and systemic symptoms, there was no statistically significant difference between the whole-virion and the AS03_B-adjuvanted groups.

The relationship between reactogenicity and immune response was examined using logged HI titres in the multivariable model. Table 23 shows the effects of fever and severe local reactions on HI titre 3 weeks after trivalent seasonal influenza vaccination. Fever and severe local reactions were both associated with a greater HI titre response to the trivalent seasonal influenza vaccine in children who had previously received the AS03_B-adjuvanted pandemic influenza vaccine, but not in those who had been given the whole-virion vaccine. No effect was seen for injection-site redness.

Response to monovalent pandemic influenza vaccine

Nearly all children who received two doses of the AS03_B-adjuvanted split-virion H1N1 monovalent pandemic influenza vaccine had antibody titres deemed protective (HI titre ≥ 1 : 32, MN titre ≥ 1 : 40) 1 year later. Children who received two doses of the whole-virion vaccine had lower titres than recipients of the AS03_B-adjuvanted vaccine, both at 3 weeks after vaccination and at 1 year later.

One year after receipt of the AS03_B-adjuvanted vaccine, the HI geometric mean titre had waned to about 30% of the baseline titre recorded 3 weeks after vaccination. Although AS03_B-adjuvanted vaccine recipients had a greater percentage drop in HI geometric mean titre than children who had been given the whole-virion vaccine, their higher titres 3 weeks after vaccination led to higher titres 1 year later. A similar waning of antibody occurs after the trivalent influenza vaccine – in one study of children < 2 years old, the time taken for antibody to decay to one-half of the post-vaccination titre was calculated to be approximately 126 days for H1N1 and 258 days for H3N2. 20

In the group of children who received the whole-virion vaccine at < 3 years of age, an unexpected finding was that the HI geometric mean titre was higher at 1 year than at 3 weeks after vaccination. In this group of children, one-third (23 out of 68) had a more than twofold rise in HI titre, with 10 out of 68 having a more than eightfold rise in titre (see Table 6). This might be due to asymptomatic natural boosting of antibody following encounter with the virus. Alternatively, it might be due to the occurrence of symptomatic influenza infection (vaccine failure). These children did not report influenza-like illness in the year since receiving the pandemic influenza vaccine, indicating that natural boosting by asymptomatic exposure to the virus is the likely explanation. Subclinical infection with the virus was widespread – a seroepidemiological study in England indicated that many more children were infected during the first wave of 2009 pandemic H1N1 infection than had been estimated from surveillance of clinical cases. 8 We postulated that children with a low HI titre after vaccination were susceptible to asymptomatic infection, with consequent natural antibody boosting, whereas those with high HI titres after vaccination were unlikely to become infected. Further analysis of our results showed that children who had a poor HI titre response to the whole-virion vaccine were those most likely to have a higher HI titre 1 year later. The subgroup of children (of all ages) who had an HI titre < 1 : 32 at 3 weeks after receiving the whole-virion vaccine had a greater geometric mean titre at 1 year than at 3 weeks after vaccination (see Table 7), suggesting that some of them may have had subclinical infection in the intervening period. By contrast, the subgroup who had an HI titre ≥ 1 : 32 at 3 weeks after receiving the whole-virion vaccine, had a lower geometric mean titre 1 year after vaccination, indicative of waning antibody.

Our additional analysis of the persistence of antibody, modelling the HI and MN titres 1 year after pandemic influenza vaccination (see Tables 11 and 12), found that children from the Exeter site had statistically significantly lower HI titres 1 year after pandemic influenza vaccination than children at other sites, even allowing for their lower titres 3 weeks after vaccination. The explanation for these findings is unclear. One hypothesis might be that children in Exeter were less exposed to the virus than children at other sites, therefore having less pre-vaccination natural priming and less post-vaccination natural boosting. However, this hypothesis is not supported by serological evidence. Although it is known that influenza infection rates were highest during the second wave of the pandemic in two large metropolitan areas in England, London and the West Midlands, in other English regions rates did not differ from one another significantly. 8 Another hypothesis might be that the vaccines given in Exeter somehow differed from those given at the other sites. However, the batch numbers did not differ between sites and the Exeter vaccines were not subject to undue temperature deviation.

One year after receiving the whole-virion vaccine, fewer children had an MN titre ≥ 1 : 40 than had an HI titre ≥ 1 : 32, particularly in the younger age group. The HI titre specifically measures antibodies directed against the receptor-binding site of viral haemagglutinin, whereas the MN titre measures a broader range of neutralising antibodies. 21 It is currently unknown which antibody classes are predominantly detected by HI and MN assays, but experience with other sera which show discordance between HI and MN titres suggests that, in addition to detecting different antibody targets, the two tests might detect different antibody classes. This has been observed when comparing results from an HI assay with a single radial haemolysis technique – the latter appears unable to detect immunoglobulin A. 21 Expressed antibody class might affect the level of clinical protection. The HI assay is widely regarded as a surrogate measure for protection and is the assay used for licensure of influenza vaccines. 23,24 Correlation between MN titre and protection is less well documented. The threshold MN titre of 1 : 40 used in our analysis is speculative and was chosen because it is approximately fourfold greater than the MN titre observed in populations that have not been exposed to the virus or vaccinated against it.

Our findings in children are comparable with the results of a study of these two vaccines in adults, which also found the AS03_B-adjuvanted vaccine to be more immunogenic than the whole-virion vaccine. 25 Both vaccines were found to be statistically significantly more immunogenic in adults from 18 to 44 years of age than in older individuals. This study assessed the persistence of antibody 6 months after vaccination. In the 18- to 44-year-old group, 6 months after a two-dose regimen, 98% of the AS03_B-adjuvanted vaccine recipients had an HI titre ≥ 1 : 40, compared with 78% of the whole-virion vaccine recipients. In these adults, 6 months after vaccination, the HI geometric mean titre had declined to 41% of its post-vaccination value in those given the AS03_B-adjuvanted vaccine and to 90% of the post-vaccination geometric mean titre in those given the whole-virion vaccine.

Vaccine effectiveness, assessed shortly after vaccination, for the AS03_B-adjuvanted vaccine given to clinical risk groups in England was 77% (95% CI 11% to 94%) in children < 10 years old and 100% (95% CI 80% to 100%) in 10 to 24-year-olds, but considerably lower in older adults. 26 The vaccine was effective in preventing confirmed cases of pandemic H1N1 influenza infection from 7 days after vaccination. 27 A recent report provided estimates of the effectiveness of vaccination in preventing confirmed influenza A (H1N1) 2009 infection in the UK in the 2010–11 season. 28 The adjusted vaccine effectiveness was 34% (95% CI –10% to 60%) if vaccinated only with monovalent pandemic influenza vaccine during the 2009–10 season, 46% (95% CI 7% to 69%) if vaccinated only with trivalent influenza vaccine in the 2010–11 season and 63% (95% CI 37% to 78%) if vaccinated in both seasons. These data accord with our serological findings of waning antibody titre 1 year after receipt of a monovalent pandemic influenza vaccine, and also with our observation of effective boosting of antibody after a dose of the 2010–11 trivalent influenza vaccine.

Another oil-in-water adjuvant containing squalene, MF59, has been used in the formulation of influenza vaccines (Fluad^® and Focetria^®; Novartis, Marburg, Germany). Two doses of MF59-adjuvanted trivalent influenza vaccine given to children aged from 16 to 48 months resulted in higher HI titres, both at 3 weeks and at 1 year after vaccination, than in children given two doses of non-adjuvanted trivalent seasonal influenza vaccine. 29 An MF59-adjuvanted monovalent pandemic H1N1 influenza vaccine given to 101 children (two-thirds of whom were born at gestational age < 36 weeks) from 6 to 23 months of age, resulted in an HI titre ≥ 1 : 40 in 94% after one dose, and 100% after two doses. 30 Antibody titres induced by a single dose of either AS03_B- or MF59-adjuvanted monovalent pandemic H1N1 influenza vaccine in immunocompetent children (from 6 months to 18 years old) were similar to HI and MN titres in unvaccinated children after natural infection. 31

None of the 323 children enrolled in our study reported clinically significant adverse events related to vaccination with either of the novel pandemic influenza vaccines, although the study was clearly too small to detect very rare adverse reactions. An increase in the incidence of narcolepsy in 4- to 19-year-old children and adolescents was recently reported in Finland, which appears to be associated with previous receipt of the AS03_B-adjuvanted pandemic influenza vaccine. 32 This possible association has now also been reported in Sweden, France and Ireland, and is being investigated further. 33 Squalene-based adjuvants have been associated with autoimmune diseases in newborn rats. 34,35

Response to trivalent seasonal influenza vaccine

Our data show that the 2010–11 trivalent seasonal influenza vaccine, given to children who had received the pandemic influenza vaccine 1 year earlier, produced a marked serological response to the H1N1 component of the vaccine. All children in the study had an HI titre ≥ 1 : 32 and an MN titre ≥ 1 : 40 3 weeks after a dose of trivalent seasonal influenza vaccine. Nearly all had at least a fourfold rise in MN titre and most had at least a fourfold rise in HI titre. All groups showed at least a 10-fold increase in HI geometric mean titre from a high baseline. This is similar to the fold increase in HI geometric mean titre seen after the first dose of trivalent influenza vaccine in children who have not previously received an influenza vaccine, although in these children the baseline titre was low. 36,37

The additional analysis, modelling HI titres 3 weeks after receipt of the trivalent seasonal influenza vaccine (see Table 17), indicated that the HI titre 3 weeks after pandemic vaccination had a statistically significant effect on response to the trivalent influenza vaccine, but the HI titre immediately before giving the trivalent influenza vaccine did not. This might indicate that individuals who had a strong serological response to the monovalent pandemic influenza vaccine also tended to have a similar response to the trivalent seasonal influenza vaccine.

The trivalent seasonal influenza vaccine was well tolerated in this population of children. No serious adverse events occurred in the 3 weeks after vaccination. Reactogenicity to the trivalent seasonal influenza vaccine in our study was similar to that reported when trivalent vaccine was given to children who had not previously received an influenza vaccine. 36–38 In our study, redness and local reactions graded as severe were statistically significantly more frequent in children who had originally been given the AS03_B-adjuvanted pandemic vaccine at < 3 years of age than in those who had received the whole-virion vaccine. In young children, previous receipt of an AS03_B-adjuvanted vaccine seems to enhance local response to a dose of a non-adjuvanted vaccine given 1 year later. Our multivariable model indicated that fever and severe local reactions occurring after trivalent seasonal influenza vaccination were both associated with a greater serological response to the vaccine in children who had previously received the AS03_B-adjuvanted pandemic influenza vaccine, but not in those who had previously been given the whole-virion vaccine (see Table 23). The mechanisms by which oil-in-water adjuvants mediate their immunological effects remain poorly understood. 39

After the trivalent vaccine, fever ≥ 38 °C was more commonly seen in children < 5 years old than in older children. It occurred in 18.3% of < 5 year old children who had previously received the AS03_B-adjuvanted pandemic vaccine and in 13.6% of those who had received the whole-virion vaccine, but in only around 1% of older children. In our study, children < 5 years old developed post-vaccination fever more frequently than the rate reported in one study of influenza vaccine naive children < 3 years old who were given their first dose of trivalent influenza vaccine, and in whom 4/65 developed a fever ≥ 38 °C. 40 In Australia, an increase in the incidence of febrile convulsions occurred in children < 5 years old shortly after being given 2010 seasonal influenza vaccine. 41 This was subsequently found to be associated with one brand of vaccine (Fluvax^®; CSL Biotherapies, Parkville, VIC, Australia), but not with the other brands of trivalent vaccine being used in Australia. 42 Similarly, in the USA, an increased risk of febrile convulsions has been reported with Fluzone^® (Sanofi Pasteur, Swiftwater, PA, USA). 43 No child in our study experienced a febrile convulsion after the trivalent seasonal influenza vaccine.

Study limitations

This study investigates the persistence of putative protective antibody levels. It does not examine vaccine effectiveness.

The recruitment rate for our follow-on study, 36% of those completing the original study, was lower than anticipated (between 40% and 60%), which slightly reduced the power of the study to detect differences between groups. This was in part due to the need to complete recruitment to the study rapidly, prior to the start of the 2010–11 influenza season. The original study recruited participants at a time of very high media and public interest in pandemic influenza. The follow-on study recruited during a quiescent phase, with relatively few new cases, when there was a much lower level of public interest.

One limitation of our study is that a two-dose regimen of pandemic influenza vaccine was used. Our original study was designed when a two-dose schedule was planned for children. However, the majority of children in the UK vaccinated with pandemic influenza vaccine during the 2009–10 campaign were only given one dose. It is likely that these vaccinated children will have somewhat lower residual antibody titres than the children in our study. Our original study did not investigate the serological response after just one dose of pandemic vaccine. In young adults, the HI geometric mean titre was 30% higher 3 weeks after a second dose of AS03_B-adjuvanted vaccine, and 23% higher after a second dose of whole-virion vaccine, than 3 weeks after a single dose of vaccine. 24

The addition of a control group, comprising children not previously vaccinated with pandemic influenza vaccine, would have strengthened the study design. This would have enabled direct comparison of antibody levels in children who had received the pandemic vaccines with children who had not, providing more robust data about the serological effects of the vaccines. A control group was not included because of time and budget limitations.

Implications for health care

Children given two doses of pandemic influenza vaccines still have putative protective titres of antibody 1 year later, although persistence beyond 1 year remains unknown. In these children, administration of the trivalent vaccine, containing the pandemic strain as one component, effectively boosts antibody titre with an acceptable reactogenicity profile. The study provides serological evidence that a two-dose regimen of the AS03_B-adjuvanted pandemic influenza vaccine may be sufficient to maintain protection across two waves of the same strain of virus.

Recommendations for future research

The inclusion of AS03_B adjuvant has resulted in an antigen-sparing vaccine producing a marked antibody response, which persists a year after vaccination. The inclusion of this adjuvant in future seasonal influenza vaccines might enhance immunogenicity, particularly in children < 3 years old, and this warrants further investigation. It would be interesting to assess whether or not previous receipt of the AS03_B-adjuvanted pandemic vaccine affected the serological response to the other two strains in the 2010–11 seasonal influenza vaccine. We propose to investigate this using stored serum. Further research is required to gain greater understanding of the immune response to AS03_B adjuvant at a cellular level.

Assessment of the total duration of effective immunity after vaccination with either AS03_B-adjuvanted or whole-virion pandemic influenza vaccines will require further study. It would be useful to assess the persistence of antibody after a single dose of these vaccines. There should be continuing surveillance of the long-term safety profile of these novel vaccines, especially in view of recent concerns regarding an association with narcolepsy in some countries. It is still unknown why two particular brands of trivalent seasonal influenza vaccine appeared to increase the risk of febrile convulsions in children < 5 years old in Australia and the USA, whereas other brands have not been implicated. It would be valuable to obtain further information on vaccine effectiveness, derived from cohort studies. Another priority should be the elucidation of the correlation between MN titre and protection from disease.

Contribution of authors

Dr Philip de Whalley (Clinical Research Fellow) contributed to design of the study, preparation of regulatory submissions, participant enrolment, data collection and interpretation. He drafted this report which has been reviewed by all the authors.

Dr Woolf Walker (Wellcome Trust Clinical Research Fellow, Paediatrics) contributed to the design of the study, participant enrolment and data interpretation.

Dr Matthew Snape (Consultant Paediatrician and Vaccinologist) contributed to the design of the study, preparation of regulatory submissions, participant enrolment and data interpretation. He was the principal investigator at the Oxford site.

Dr Clarissa Oeser (Clinical Research Fellow, Vaccines) contributed to participant enrolment and data collection.

Michelle Casey (Senior Research Nurse) contributed to participant enrolment and data collection.

Phoebe Moulsdale (Clinical Research Nurse) contributed to participant enrolment and data collection.

Caroline Harrill (Clinical Research Nurse) contributed to participant enrolment and data collection.

Nick Andrews (Senior Statistician) was responsible for statistical analysis and contributed to the design of the study and to data interpretation.

Katja Hoschler (Advanced Healthcare Scientist/Clinical Scientist) was responsible for laboratory analysis.

Ben Thompson (Project Manager) was responsible for project management and contributed to preparation of regulatory submissions.

Claire Jones (Postdoctoral Research Assistant) contributed to project management and was the laboratory liaison.

Jem Chalk (Software Developer/Computer Officer) was responsible for the computer database and contributed to data collection and analysis.

Simon Kerridge (Quality Assurance Manager) was responsible for data monitoring.

Dr Richard Tomlinson (Consultant Paediatrician) contributed to participant enrolment. He was the principal investigator at the Exeter site.

Dr Paul Heath (Executive Officer Paediatric Studies, Reader and Honorary Consultant Paediatrician) contributed to the design of the study, participant enrolment and data interpretation. He was the principal investigator at the St George’s, London site.

Professor Adam Finn (David Baum Professor of Paediatrics) contributed to the design of the study, participant enrolment and data interpretation. He was the principal investigator at the Bristol site.

Dr Saul Faust (Senior Lecturer in Paediatric Immunology and Infectious Diseases) contributed to the design of the study, participant enrolment and data interpretation. He was the principal investigator at the Southampton site.

Professor Elizabeth Miller (Head, Immunisation Department) contributed to the design of the study.

Professor Andrew Pollard (Professor of Paediatric Infection and Immunity) contributed to the design of the study, participant enrolment and data interpretation. He was the chief investigator.

Publication

Walker WT, de Whalley P, Andrews N, Oeser C, Casey M, Michaelis L, et al. H1N1 Antibody Persistence 1 Year After Immunization With an Adjuvanted or Whole-Virion Pandemic Vaccine and Immunogenicity and Reactogenicity of Subsequent Seasonal Influenza Vaccine: A Multicenter Follow-on Study. Clin Infect Dis 2012; in press. DOI: 10.1093/cid/cir905.

Disclaimers

The views expressed in this publication are those of the authors and not necessarily those of the HTA programme or the Department of Health.

Health Technology Assessment programme

1. Director, NIHR HTA programme, Professor of Clinical Pharmacology, University of Liverpool

2. Professor of Dermato-Epidemiology, Centre of Evidence-Based Dermatology, University of Nottingham