Comparative Reactogenicity of Enhanced Influenza Vaccines in Older Adults

Abstract

Background

We analyzed data from a randomized controlled trial on the reactogenicity of 3 enhanced influenza vaccines compared with standard-dose (SD) inactivated influenza vaccine.

Methods

We enrolled community-dwelling older adults in Hong Kong, and we randomly allocated them to receive 2017–2018 northern hemisphere formulations of SD vaccine (FluQuadri; Sanofi Pasteur), MF59-adjuvanted vaccine (FLUAD; Seqirus), high-dose (HD) vaccine (Fluzone High-Dose; Sanofi Pasteur), or recombinant hemagglutinin vaccine (Flublok; Sanofi Pasteur). Local and systemic reactions were evaluated at days 1, 3, 7, and 14 after vaccination.

Results

Reported reactions were generally mild and short-lived. Systemic reactions occurred in similar proportions of participants by vaccine. Some local reactions were slightly more frequently reported among recipients of the MF59-adjuvanted and HD vaccines than among SD vaccine recipients. Participants reporting feverishness 1 day after vaccination had mean fold rises in postvaccination hemagglutination inhibition titers that were 1.85-fold higher (95% confidence interval, 1.01–3.38) for A(H1N1) than in those who did not report feverishness.

Conclusions

Some acute local reactions were more frequent after vaccination with MF59-adjuvanted and HD influenza vaccines, compared with SD inactivated influenza vaccine, whereas systemic symptoms occurred at similar frequencies in all groups. The association between feverishness and immunogenicity should be further investigated in a larger population.

Clinical Trials Registration

NCT03330132.

The World Health Organization Strategic Advisory Group of Experts on immunization recommend that older adults should be a priority group for annual influenza vaccination [1]. The most widely used influenza vaccines in all ages are “standard-dose” (SD) inactivated influenza vaccines grown in fertilized chicken eggs, administered intramuscularly, and containing 15 µg of hemagglutinin (HA) of each strain. Trivalent vaccines include influenza A(H1N1), A(H3N2), and 1 B strain from either the Victoria lineage or the Yamagata lineage, and quadrivalent vaccines include an additional B strain from the other lineage.

A number of enhanced influenza vaccines have been introduced in recent years, with the aim of providing improved protection in older adults compared with SD vaccines [2–5]. The designers of these enhanced vaccines have taken different approaches to improve immunogenicity and effectiveness. One approach is to increase the amount of antigen in the vaccine. The trivalent high-dose (HD) vaccine (Fluzone High-Dose, Sanofi Pasteur) has 4 times as much HA (60 µg per strain) as the SD vaccine. This vaccine induces stronger humoral immune responses [6] and has improved clinical efficacy [7, 8] compared with an SD vaccine. Another approach is to add an adjuvant to the SD vaccine. The trivalent MF59-adjuvanted vaccine (FLUAD; Seqirus) adds MF59, a squalene-based emulsion. This vaccine produces a stronger immune response than SD vaccine [9, 10].

A third approach, noting that egg adaptations in SD vaccines have been associated with poorer vaccine effectiveness [11], is to use a different production process that could improve the match between vaccine antigen and circulating strains. The Flublok vaccine (Sanofi Pasteur) includes recombinant HA proteins that are produced by baculovirus expression in insect SF9 cell culture [12]. This vaccine also includes an increased amount of antigen (45 µg per strain), and a large trial found that it had improved clinical efficacy compared with SD vaccine [13].

We are currently conducting a comparative randomized controlled trial of 3 enhanced influenza vaccines and SD vaccine in older adults in Hong Kong [14]. We analyzed data from the first year of the trial to compare reactogenicity of the vaccines. In addition, given that prior research suggested that administration of antipyretics at the time of vaccination could be associated with blunted immune response to the vaccine, potentially mediated by suppressing fever [15, 16], we investigated whether there was a correlation between feverishness or other short-term reactions and the strength of the humoral immune response to influenza vaccination.

METHODS

Study Participants

Starting in June 2017, we enrolled community-dwelling older adults who were 65–82 years of age, residing in Hong Kong, and had not already received northern hemisphere 2017–2018 formulation of influenza vaccination. Potential participants were excluded if they (1) showed signs of dementia [17] or had a clinical diagnosis of dementia or other significant cognitive impairment; (2) reported medical conditions making them unsuitable for receipt of an inactivated influenza vaccine [18], such as any documented Guillain-Barré syndrome within 6 weeks of previous vaccination, or any documented allergic reaction to egg protein or previous dose of influenza vaccine; (3) reported medical conditions making them unsuitable for receipt of an intramuscular injection, such as use of anticoagulant medication (other than antiplatelet medication such as aspirin); or (4) had any medical condition rendering them not suitable to receive an inactivated influenza vaccine as determined by a clinician [14]. After screening to determine eligibility, older adults provided signed informed consent to participate. Our trial was registered with ClinicalTrials.gov (NCT03330132).

Participants were randomized to 4 groups. Of 11 participants, we randomized 3 of (27%) to receive SD quadrivalent vaccine (0.5 mL; FluQuadri; Sanofi Pasteur), 3 (27%) to receive trivalent MF59-adjuvanted vaccine (0.5 mL; FLUAD; Seqirus), 3 (27%) to receive trivalent HD vaccine (0.5 mL; Fluzone High-Dose; Sanofi Pasteur), and 2 (18%) to receive quadrivalent recombinant HA vaccine (0.5 mL; Flublok; Sanofi Pasteur). The Flublok vaccine was added to the trial at a late stage and that group was kept smaller, because we planned a factorial design (3 × 3) rerandomizing receipt of the other 3 vaccines in year 2 but did not plan a larger sample size that would be needed if we were to expand to a 4 × 4 design.

All vaccines included the strains recommended for the northern hemisphere 2017–2018 formulation. Vaccines were repackaged into numbered boxes so that the participants and most of the research team, including all members of the team who assessed adverse events and reactions, were blinded to the type of vaccine administered; however, the nurse who removed the vaccine from the numbered box and administered it was not blinded to vaccine type because of differences in the size and shape of the prefilled syringes.

Ethics

Signed informed consent was obtained from all participants. The study protocol was approved by the Institutional Review Board of the University of Hong Kong.

Reactogenicity Assessments

After vaccination, participants were observed for 15 minutes to monitor for acute adverse reactions such as anaphylaxis [14]. Participants were called by telephone 4 times after vaccination (coded as day 0), on days 1–2, 3–4, 7–9, and 14–16, respectively, for recording of any local or systemic reactions. Up to 3 phone calls were made each day at different times of the day. Reporting of day 1 reactions on day 2 was permitted if the participant could not be reached on day 1. For the subsequent 3 time points, calls were permitted up to 2–3 days later than the target day to allow for flexibility in scheduling calls, and to mitigate potential difficulties in contacting participants on weekends. Participants were also encouraged to call the study hotline number listed on the vaccination card to contact our research team for medical advice if they experienced any possible adverse reactions to vaccination. No memory aid was given to record reactions, and participants did not receive any specific education on how to assess reactions.

Similarly to previous studies [19, 20], we monitored local site reactions, including redness, swelling, tenderness, itching, and pain at the injection site; systemic reactions were assessed by asking participants if they had experienced “any flulike symptoms (including fever)”; if so, participants were asked about feverishness, muscle pain, nausea, fatigue, and other symptoms. Participants were asked to rate any local site reaction or symptom as (1) mild (symptom is easily tolerated and does not interfere with any usual activities), (2) moderate (symptom interferes with usual activities), or (3) severe (participant cannot carry out usual activities). As in a previous study [21], we used quantitative classification of mild, moderate, and severe reactions to guide participants in self-rating severity of swelling and erythema as follows: mild, the diameter of the maximum reaction size, <2.5 cm; moderate, 2.5–5 cm, and severe, diameter ≥5 cm. Rulers were not provided to participants, and assessment of reaction size was permitted to be subjective. Feverishness was considered present if a participant reported this symptom; participants were not provided a thermometer or given specific instructions about measuring or recording body temperature. Participants were not given specific instructions about use of antipyretics, and this information was not captured in the study.

Immunogenicity Assessments

We collected a 9-mL blood sample from each participant immediately before vaccination, and approximately 30 days (range, 26–35 days) after vaccination. We randomly selected a subset of 200 pairs of day 0 and day 30 serum samples from each of the 4 vaccine groups to evaluate vaccine immunogenicity, for a total of 800 paired serum samples evaluated [14]. The serum samples were tested using hemagglutination inhibition (HAI) assays against egg-propagated vaccine strains A/Singapore/GP1908/2015 (A/Michigan/45/2015(H1N1)-like virus), A/Hong Kong/4801/2014(H3N2), and B/Brisbane/60/2008 (Victoria lineage). Because some of the vaccines used in our trial were trivalent and included only a B/Victoria strain [14], we did not analyze B/Yamagata titers here.

Statistical Analysis

We used χ ² tests to compare the prevalence of reactions between groups at the various time points. We plotted duration of symptoms by symptom severity within 1 day after vaccination and compared the durations, using asymptotic log-rank tests allowing for the interval censoring in the data on time to cessation of symptoms. We estimated the mean fold rise (MFR) in HAI titers from day 0 to day 30 for each strain. We compared MFRs between participants who reported feverishness on the day after vaccination and those who did not, using Wilcoxon rank sum tests. We used log-linear regression models to quantify the association between the MFR in antibody titer and the presence of feverishness on day 1, adjusting for vaccine type, prevaccination antibody titer, and age as potential confounding factors. In a sensitivity analysis, we tested for possible differences in effect by vaccine type by adding an interaction term for feverishness by vaccine type to the models. Differences were considered statistically significant at P < .05. All statistical analyses were conducted using R software, version 3.4.3 (R Foundation for Statistical Computing).

The sample size of the trial was designed to have power for primary comparisons of vaccine immunogenicity between groups [14]. For this analysis of adverse reactions, with about 500 participants per group and adverse reaction frequencies of 15%, we would have 80% power to detect differences in proportions by a factor of ≥1.5. Power would be reduced for adverse reactions of lower frequency, and for comparisons with the recombinant vaccine group with 330 participants.

RESULTS

The trial included 1861 participants randomly allocated to the 4 vaccines, and participant characteristics were comparable across vaccine groups (Table 1). During the 15 minutes after vaccination, there were no adverse reactions. In the days after vaccination, we captured information on reactions from the majority of participants. We successfully contacted 1575 of 1861 participants (85%) on days 1–2 about reactions on day 1, 1513 of 1861 (81%) on days 3–4, 1688 of 1861 (91%) on days 7–9, and 1665 of 1861 (89%) on days 14–16. The patterns in reported reactions on these days are shown in Figure 1. Only 21 (1.1%) of all participants could not be reached within 14 days after vaccination, but they were reached on day 30 during the follow-up appointment for blood collection without reporting any experience of adverse reactions, except 3 participants who did not show up for follow-up and could not be reached after repeated attempts. However, we did not include them in the analysis to avoid possible recall bias. Among all participants, we received reports of ≥1 adverse event from 44% of the recipients of the SD vaccine, 51% of the MF59-adjuvanted vaccine recipients, 53% of the HD vaccine recipients, and 40% of the recombinant HA vaccine recipients (P < .01; χ ² test).

Reactions were reported more frequently on days 1–2 (for day 1) than on subsequent days, and 614 of 1575 participants (39%) in all vaccine groups reported ≥1 adverse event on day 1. The most frequently reported reactions for day 1 were local reactions, such as tenderness (range across vaccine groups, 15%–24%), pain (8%–14%), and swelling (4%–12%) (Supplementary Table 1). For day 1, tenderness was reported more frequently by recipients of the MF59-adjuvanted (21%) and the HD (24%) vaccine than by recipients of the recombinant HA (15%) or SD (17%) vaccines.

Differences were statistically significant between MF59-adjuvanted and SD vaccines (P = .05), and between HD and SD vaccines (P = .04). In addition, for day 1, the recipients of the HD vaccine reported significantly more swelling than recipients of the SD vaccine (12% vs 8%, respectively; P < .01), whereas recipients of the recombinant HA vaccine reported significantly less swelling than SD recipients (4% and 8%, respectively; P < .01). On days 3–4, recipients of the MF59-adjuvanted vaccine reported significantly less pain (1.1%) and swelling (0.9%) than recipients of the SD vaccine (4.8% and 4.6%, respectively; both P < .01) (Supplementary Table 2). On days 7–9 the recipients of the MF59-adjuvanted vaccine reported significantly less tenderness (0.4%) than recipients of the SD vaccine (2%; P = .03) (Supplementary Table 3). No other statistically significant differences in responses were reported on days 7–9 (Supplementary Table 3) or days 14–16 (Supplementary Table 4).

Among all reported reactions, most (91%) were mild, 8% were graded as moderate, and 1% were graded as severe (Figure 1). The final column in Figure 1 shows “other” reactions, which were mostly respiratory symptoms (Supplementary Table 1). Whereas the local and systemic reactions were reported most frequently on days 1–2 (for day 1) and less frequently on subsequent days, the “other” reactions were consistently reported by 5%–9% of participants at each of the 4 time points and were similar across vaccine groups. We did not identify any potentially vaccine-related serious adverse events in the 30 days after vaccination, as judged by the study investigator.

For participants who reported tenderness, pain, swelling, and feverishness on day 1, we examined the duration of these reactions. Figure 2 shows the time to resolution of each reaction among those who reported having the reaction on day 1, stratified by the reported severity on day 1. More than half of these reactions had ceased within 2 days after vaccination, but some persisted for up to 6 days in a minority of participants. Compared with participants reporting mild swelling on day 1, those reporting moderate swelling on day 1 had significantly longer times to resolution of swelling (P = .04). There was no significant difference in the duration of pain, tenderness. or feverishness based on the reported severity on day 1.

In a subset of 800 participants (200 in each vaccine group), we measured antibody titers before vaccination (day 0) and at day 30. Among these 800 participants, we collected reports of reactions on day 1 from 688 (86%). Feverishness on day 1 was reported by 15 (2.1%) of these 688 participants, with no difference between vaccine groups. We used the MFRs in antibody titers from day 0 to day 30 as measures of the immune response to vaccination. For each of the strains tested, we examined whether there was a difference in MFR between participants (across vaccine types) who did or did not report feverishness on day 1 (Figure 3). We found significantly higher MFRs in HAI titers against influenza A(H1N1) and B/Victoria, and an increase that was not statistically significantly different in A(H3N2) titers, in participants who reported feverishness on day 1. In a sensitivity analysis, the associations between self-reported feverishness and MFRs in HAI titers were similar in magnitude across vaccine types and were not statistically significant (Supplementary Figure 1). Among the 24 combinations of acute reactions (apart from feverishness) and strains examined, there was 1 statistically significant association between reported local redness and a decrease in the MFR in HAI titer against influenza A(H3N2).

In a regression analysis that adjusted for vaccine type and prevaccination titer, we found that MFR in antibody titer was statistically significantly associated with reported feverishness on day 1 for A(H1N1). The model estimated that participants reporting feverishness on day 1 had, on average, MFRs 1.85-fold higher (95% confidence interval, 1.01–3.38) than those in participants without feverishness (Table 2). Adjusted associations between fever and MFR were not statistically significant for influenza A(H3N2) and B/Victoria. To consider whether these associations differed by vaccine type, we added interaction terms for feverishness by vaccine types to the models, but none were statistically significant.

Discussion

In summary, we observed that generally mild and local adverse events were common with all the injected influenza vaccines used, including enhanced vaccines, among older adults aged 65–82 years in Hong Kong. The reactions to vaccine typically lasted only a few days, and more than half had resolved within 2 days after vaccination. Other adverse reactions less likely to be related to vaccination were reported at all time points by a similar proportion of older adult participants receiving each of the different vaccines (Figure 1), which served as a potential negative control. Participants who reported feverishness on day 1 had higher MFR in HAI titers against all vaccine strains than those who did not report feverishness; this was statistically significant for A(H1N1), suggesting that fever after influenza vaccination may be linked to a better immune response to vaccine in older adults.

Our findings regarding the frequencies of adverse events with enhanced vaccines compared with SD vaccine are consistent with the frequencies reported in previously published studies among older adults [6, 7, 12, 13, 22–31]. Although we could not compare every adverse event reported by participants in our trial with other studies, we were able to compare some adverse events among participants receiving HD, adjuvanted, and recombinant HA vaccines with findings of other studies. Overall, previous studies reported similar patterns in adverse events, with higher rates of local and systemic reactions reported by recipients of the HD vaccine and the MF59-adjuvanted vaccine compared with recipients of SD vaccines [6, 7, 22–31], but with similar or lower rates of local reactions among recipients of the recombinant HA vaccines and SD vaccine in some studies [12, 13].

Feverishness was not commonly reported among older adults in our study on day 1 after vaccination; however, our study strengthens a hypothesis that postvaccination feverishness, and potentially fever, may be associated with more robust immune responses after influenza vaccination. Analyses from 3 pediatric clinical trials also observed an association between postvaccination measured fever and immunogenicity [32], and there could be at least 2 interpretations. It is possible that a febrile response after vaccination is indicative of a healthy innate immune system, which in turn is associated with good adaptive immune response and production of influenza-specific antibodies. An alternative hypothesis is that fever is a causal mediator of immunogenicity; specifically, a systemic fever response and activation of innate immune mechanisms may facilitate adaptive immune engagement and subsequent antibody response.

One way to potentially distinguish between the role of fever as an incidental correlate versus a causal mediator of immunogenicity is to experimentally disrupt a febrile response after vaccination, but to the best of our knowledge studies have not directly assessed this question. Pediatric studies have shown that antipyretic use after vaccination can reduce the strength of the antibody response to vaccination [32–34], though this effect may be limited to the initial immunization with a novel antigen, and not to booster doses [15]. In the single small randomized controlled trial of which we are aware, older adults who received prophylactic acetaminophen had similar and low rates of fever and no difference in antibody response after influenza vaccination compared with that in placebo recipients [35].

Our study had some limitations. First, we relied on self-reporting of adverse reactions, which may underestimate the occurrence of some adverse events. Hospitalizations in participants were not uncommon, but the rates of postvaccination hospitalizations did not differ significantly among the 4 vaccine groups and are not studied here [14]. Given the complexity and challenges of documenting elevated temperature [36] we recorded feverishness in our study; having information on measured temperature in the days after vaccination would have aided in interpretation of our findings and allowed further investigation of the association between height of fever and immune response. Serum samples were tested in a subset of 200 participants in each of the 4 vaccine groups, and thus our study was not necessarily powered to assess smaller differences in adverse reactions among the 4 vaccines used or to assess associations with adverse reactions and immune response. Finally, even though we had a thorough protocol for postvaccination follow-up, contacting and reaching participants after vaccination was challenging, and some participants were not reachable. Nevertheless, we do not believe that follow-up challenges affected our findings overall, because the proportion of unreachable participants was small.

In conclusion, we found that the frequencies of systemic reactions after vaccination were similar for enhanced and SD vaccines, and the patterns of injection-site reactions were consistent with those found in other studies. The association between feverishness and immunogenicity should be further investigated in a larger population. Most reactions that occurred were mild and short-lived in study participants. Findings from this study as a whole will inform future vaccine trials and will contribute to decisions on the use of enhanced vaccines among older adult populations. Furthermore, this work indicates a need for continued monitoring through vaccine adverse events reporting systems, including studies of infrequent but more serious potential adverse reactions, to adequately evaluate the safety profile of enhanced vaccines.

Notes

Acknowledgments. The authors thank Julie Au for administrative support.

Disclaimer. The sponsor had no role in data collection and analysis, or the decision to publish, but was involved in study design and preparation of the manuscript. The contents of this report are solely the responsibility of the authors and do not necessarily represent the official views of the Centers for Disease Control and Prevention or the Department of Health and Human Services.

Financial support. This study was supported by the Centers for Disease Control and Prevention (cooperative agreement IP001064-02).

Potential conflicts of interest. B. J. C. has received honoraria from Sanofi and Roche for advisory committees. All other authors report no potential conflicts. All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.

References