Antibody Responses to Influenza A/H1N1pdm09 Virus After Pandemic and Seasonal Influenza Vaccination in Healthcare Workers: A 5-Year Follow-up Study

Abstract

Background: The 2009 influenza pandemic was caused by the A/H1N1pdm09 virus, which was subsequently included in the seasonal vaccine, up to 2016/2017, as the A/H1N1 strain. This provided a unique opportunity to investigate the antibody response to H1N1pdm09 over time.

Methods: Healthcare workers (HCWs) were immunized with the AS03-adjuvanted H1N1pdm09 vaccine in 2009 (N = 250), and subsequently vaccinated with seasonal vaccines containing H1N1pdm09 for 4 seasons (repeated group), <4 seasons (occasional group), or no seasons (single group). Blood samples were collected pre and at 21 days and 3, 6, and 12 months after each vaccination, or annually (pre-season) from 2010 in the single group. The H1N1pdm09-specific antibodies were measured by the hemagglutination inhibition (HI) assay.

Results: Pandemic vaccination robustly induced HI antibodies that persisted above the 50% protective threshold (HI titers ≥ 40) over 12 months post-vaccination. Previous seasonal vaccination and the duration of adverse events after the pandemic vaccination influenced the decision to vaccinate in subsequent seasons. During 2010/2011-2013/2014, antibodies were boosted after each seasonal vaccination, although no significant difference was observed between the repeated and occasional groups. In the single group without seasonal vaccination, 32% of HCWs seroconverted (≥4-fold increase in HI titers) during the 4 subsequent years, most of whom had HI titers <40 prior to seroconversion. When excluding these seroconverted HCWs, HI titers gradually declined from 12 to 60 months post-pandemic vaccination.

Conclusions: Pandemic vaccination elicited durable antibodies, supporting the incorporation of adjuvant. Our findings support the current recommendation of annual influenza vaccination in HCWs.

Clinical trials registration: NCT01003288.

Figures

Figure 1.

The study flow chart. Healthcare workers (HCWs) were recruited to an open-label 5-year extension of a single-arm clinical trial in 2009 (European Clinical Trials Database, EudraCT 2009-016456-43; www.clinicaltrials.gov, NCT01003288) [6]. All HCWs received a single dose of the AS03-adjuvanted monovalent pandemic A/H1N1pdm09 vaccine, except 12 low-responders, who were identified at 3 months post–pandemic vaccination and received a second dose at 5 months after the first dose. During 2010/2011–2013/2014, HCWs voluntarily received the trivalent seasonal inactivated vaccines, containing the A/H1N1pdm09 as the A/H1N1 component during the whole study period, while the A/H3N2 and B viruses changed between seasons. During the course of the study, HCWs did not always provide samples in each season, due to shift work, statutory leave (sick or parental leave), leaving employment at the hospital, or death. The drop out column shows the number of HCWs, irrespective of seasonal vaccination status, who dropped out in each season. At the end of the study, HCWs were divided into 3 groups based on their seasonal vaccination status: repeated (vaccination in all 4 seasons from 2010/2011–2013/2014), occasional (vaccination in <4 seasons), and single (no seasonal vaccination after pandemic vaccination in 2009).

Figure 2.

The preexisting H1N1pdm09-specific hemagglutination inhibition (HI) antibodies, prior to pandemic vaccination in 2009. The HI titers pre–2009 pandemic vaccination in healthcare workers (HCWs) are stratified by (A) previous seasonal vaccination status or (B) working department. Each symbol represents an individual HI titer, with the horizontal lines representing the geometric mean HI titer and 95% confidence interval. The dotted line indicates the protective threshold HI titer of 40. The percentage of HCWs from different working departments with a history of (C) previous seasonal vaccination or (D) no previous seasonal vaccination, stratified by HI antibody levels pre–2009 pandemic vaccination: HI titers <10 (undetectable antibodies), 10–39 (detectable antibodies), or ≥40 (protective antibodies). *P < .01. **P < .001.

Figure 3.

The H1N1pdm09-specific hemagglutination-inhibition (HI) antibody response after pandemic vaccination. A The HI titers of all vaccinated participants are shown pre-vaccination (D0) and at 21 days (D21) and 3, 6, and 12 months (3M, 6M, and 12M, respectively) post-vaccination during 2009. Each symbol represents an individual HI titer, with the horizontal lines representing the geometric mean HI titer (GMT) and 95% confidence interval (CI). The GMT is shown for each time point above the graph. The dotted line indicates the protective threshold HI titer of 40. B Fold-changes in HI titer from pre- (D0) to post–pandemic vaccination (D21, 3M, 6M or 12M) are shown. Each symbol represents an individual fold-change, with the horizontal lines representing the geometric mean ratio (GMR) and 95% CI, generated from all individual fold changes. The GMR is shown for each time point above the graph. The dotted line indicates the fold change of 1, meaning no change in HI titers between pre- and post-vaccination. C The HI titers in 12 low-responders (LRs), who received 2 doses of pandemic vaccine in the 2009 pandemic vaccination. These LRs had HI titers <40 or a <4-fold increase from pre-vaccination titers at 3 months post–pandemic vaccination, and received the second dose at 5 months. The HI titers of these 12 LRs were excluded from the 2009 data of all vaccinated participants at 6 and 12 months after the first dose. Pre- and post-vaccination HI titers of all participants were log-transformed and compared using linear mixed-effect models adjusted for subject variance with repeated measures and demographic factors. Post hoc tests were performed with Bonferroni correction. HI titers of LRs were compared using non-parametric repeated-measure Friedman tests. **P < .01. ***P < .001.

Figure 4.

The adverse events (AEs) after pandemic vaccination. Healthcare workers (HCWs) were provided with a diary card where they self-recorded the incidence and duration of local (erythema/redness, itching, oedema/swelling, pain/tenderness at the injection site, ecchymosis, hardness/induration) and/or systemic (fever, malaise, shivering/chills, fatigue, headache, sweating, myalgia, arthralgia, respiratory symptoms, and diarrhea) AEs during the 21 days after pandemic vaccination. Only 2 HCWs (2/250) did not return their diary cards. A The percentage of HCWs reporting AE as none, local, systemic, or both local and systemic, stratified by the vaccination groups (repeated, occasional, or single group). B The percentage of HCWs reporting the incidence of any AE that resolved in durations of ≤2, 3–5, or ≥5 days, stratified by the vaccination groups.

Figure 5.

The H1N1pdm09-specific hemagglutination-inhibition (HI) antibody titers over 5 years after only pandemic vaccination. As the single group chose not to receive further vaccinations, data are only available after the pandemic vaccination (2009pdm) and prior to each influenza season (24, 36, 48, and 60 months post–pandemic vaccination: 24M, 36M, 48M, and 60M, respectively). A total of 21/65 healthcare workers (HCWs) seroconverted (>4-fold increase in HI titers) during 4 years of follow-up. A The geometric mean HI titers with 95% confidence interval over 5 years are presented, both in the single group and after excluding all HCWs who seroconverted (4-fold increase in titers). The dotted line indicates the protective threshold HI titer of 40. The fold-change in HI titers after 1 season in (B) 2010/2011, (C) 2011/2012, (D) 2012/2013, and (E) 2013/2014, stratified by HCWs with pre-season HI titers <40 or ≥40. F The decline of protective HI titers over time in the single group whose HI titers persisted ≥40 at 12 months (12M) post–pandemic vaccination. HI titers of the single group were log-transformed and compared between years using linear mixed-effect models adjusted for subject variance with repeated measures and demographic factors. Post hoc tests were performed with Bonferroni correction. Fold changes were compared using non-parametric Mann-Whitney tests. A negative correlation between HI titers and years post–pandemic vaccination was identified using linear regression models. *P < .05. **P < .01. ***P < .001.

Figure 6.

The 5-year dynamics of the H1N1pdm09-specific hemagglutination-inhibition (HI) antibody responses after pandemic vaccination and subsequent seasonal vaccinations. A The geometric mean HI titers with 95% confidence interval of vaccinated healthcare workers (HCWs) in the repeated group, the occasional group and in low responders (LRs) pre-vaccination (D0) and at 21 days (D21) and 3, 6, and 12 months (3M, 6M, and 12M, respectively) post-vaccination during the 2009 pandemic vaccination (2009pdm) through 2013/2014 are shown. In the 2009pdm, LRs received the second pandemic vaccination at 5 months (5M) and, therefore, their HI titers at 21 days and 7 months (7M) after the second vaccination are equivalent to the 6 and 12 month measures after the single pandemic vaccination in other participants. For each year/season, the HI titers at 12 months in the former year/season were considered as the pre-vaccination responses only for the HCWs who were vaccinated in the relevant next season. The dotted line indicates the protective threshold HI titer of 40. B The antibody responses of the repeated and occasional groups, including LRs who were either repeatedly vaccinated with the seasonal vaccine (LR repeated group, n = 6) or occasionally vaccinated (LR occasional group, n = 2). For more information on LRs, see Supplementary Figure 2. C The antibody responses of the repeated and occasional groups in comparison with the single group, excluding 21/65 HCWs who seroconverted (>4 fold-increase HI titers) during 2010/2011–2013/2014, at equivalent time points (12 months post–seasonal vaccination or prior to influenza season). HI antibody titers were log-transformed and compared between groups using linear mixed-effect models adjusted for subject variance with repeated measures and demographic factors. Post hoc tests were performed with Bonferroni correction. **P < .01, ***P < .001.