Pre- and postpandemic estimates of 2009 pandemic influenza A(H1N1) seroprotection to inform surveillance-based incidence, by age, during the 2013-2014 epidemic in Canada

Journal of Infectious Diseases Advance Access published July 29, 2014 BRIEF REPORT Pre- and Postpandemic Estimates of 2009 Pandemic Influenza A(H1N1) Seroprotection to Inform Surveillance-Based Incidence, by Age, During the 2013–2014 Epidemic in Canada Danuta M. Skowronski,1,2 Catharine Chambers,1 Suzana Sabaiduc,1,2 Naveed Z. Janjua,1,2 Guiyun Li,1 Martin Petric,1,2 Mel Krajden,1,2 Dale Purych,3 Yan Li,4,5 and Gaston De Serres6,7 1 To understand the epidemic resurgence of influenza due to the 2009 pandemic influenza A(H1N1) strain (A[H1N1]pdm09) during the 2013–2014 influenza season, we compared age-related cross-sectional estimates of seroprotection before the pandemic (during 2009) and after the pandemic (during 2010 and 2013) to subsequent surveillance-based, laboratory-confirmed incidence of influenza due to A(H1N1)pdm09 in British Columbia, Canada. Prepandemic seroprotection was negligible except for very old adults (defined as adults aged ≥80 years), among whom 80% had seroprotection. Conversely, postpandemic seroprotection followed a U-shaped distribution, with detection in approximately 35%–45% of working-aged adults but in ≥70% of very old adults and young children, excluding children aged <5 years in 2013, among whom seroprotection again decreased to <20%. The incidence was 5-fold higher during 2013–2014, compared with 2010–2011, and was highest among children aged <5 years and working-aged adults, reflecting a mirror image of the age-based seroprotection data. Keywords. influenza; A(H1N1)pdm09; serosurvey; seroprotection; surveillance; incidence; hemagglutination inhibition. Received 4 May 2014; accepted 21 June 2014. Presented in part: AMMI Canada–CACMID Annual Conference, Victoria, Canada, 2–5 April 2014 [abstract]; ProMED, 1 January 2014 [archive 20140101.2146429]. Correspondence: Danuta M. Skowronski, MD, MHSc, FRCPC, British Columbia Centre for Disease Control, 655 W 12th Ave, Vancouver, British Columbia, Canada V5Z 4R4 (danuta. skowronski@bccdc.ca). The Journal of Infectious Diseases © The Author 2014. Published by Oxford University Press on behalf of the Infectious Diseases Society of America. All rights reserved. For Permissions, please e-mail: journals.permissions@ oup.com. DOI: 10.1093/infdis/jiu366 METHODS A convenience sample of approximately 50 anonymized residual sera per decade of life was obtained in May 2013 from a community laboratory network that provides patient services across southwestern British Columbia. Findings were compared to data from previous cross-sectional serosurveys conducted in the same region by use of similar methods, including collection of approximately 100 sera/decade of life for the prepandemic period (99% of specimens were collected during January–July 2009) [1] and the immediate postpandemic period during May–June 2010 [8]. To enable finer stratification among pediatric ages, oversampling was undertaken for age groups composed of individuals <10 years old during the 2009 prepandemic and 2010 postpandemic periods. Antibody titers were measured against A/California/07/2009 (Supplementary Table 1), using a hemagglutination inhibition (HI) assay, conducted at the British Columbia Public Health Microbiology and Reference Laboratory (PHMRL) according to standard protocols with turkey erythrocytes [1, 8]. Antibody titers were calculated as the geometric mean titer (GMT) of duplicate inverse titers, with titers of <10 assigned a value of 5. BRIEF REPORT • JID • 1 Downloaded from http://jid.oxfordjournals.org/ by guest on March 15, 2016 British Columbia Centre for Disease Control and 2University of British Columbia, Vancouver, 3BC Biomedical Laboratories, Surrey, 4National Microbiology Laboratory, Public Health Agency of Canada and 5University of Manitoba, Winnipeg, and 6 Institut National de Santé Publique du Québec and 7Laval University, Quebec, Canada In April 2009, swine-origin influenza A(H1N1) emerged in humans to cause the first pandemic of the 21st century. The 2009 pandemic influenza A(H1N1) strain (A[H1N1]pdm09) derives from the classic swine lineage, bearing antigenic properties of distant 1918 pandemic A(H1N1)–descendant strains but sharing little or no cross-reactive neutralizing antibody with modern human influenza virus strains of the past several decades [1]. Following the 2009 pandemic, A(H1N1)pdm09 composed <20% of influenza virus detections each year in Canada through 2012–2013 [2]. The 2013–2014 influenza season, however, was characterized by a moderately severe epidemic of resurgent and dominant A(H1N1)pdm09 activity in British Columbia, Canada, and other North American jurisdictions, despite minimal genetic or antigenic change in A/California/07/2009-like A(H1N1)pdm09 circulating since 2009 [2–6]. Anecdotal and published reports of a shift in the age distribution to include more working-aged adults accompanied this resurgent activity [2–4, 6, 7]. Conversely, reports of outbreaks in schools and longterm-care facilities were infrequent [4]. To understand the age-related profile of the influenza epidemic due to A(H1N1) pdm09 during 2013–2014, we compared preseason seroprotection to surveillance-based disease patterns in British Columbia during pre- and postpandemic periods. RESULTS Patient Characteristics The age and sex profile of patients included in serosurveys, by study period, is shown in Table 1. Median age was significantly higher in spring 2013 (50 years) than in the 2009 prepandemic period (40 years) or the 2010 postpandemic period (44 years; P < .01), reflecting oversampling of pediatric patients in 2009 and 2010. Females were disproportionately represented each period, particularly among those of childbearing age, but values did not significantly differ overall across study periods (P = .22). Seroprotection and Surveillance-Based Incidence, by Age Mean GMTs and the percentage seroprotected, with 95% CIs and by study period, are displayed by decade of life, as per the original sampling strategy, in Supplementary Table 1, with sensitivity analyses shown in Supplementary Figure 1. The seroprotection estimates in relation to laboratory-confirmed A(H1N1)pdm09 surveillance findings are shown in Figure 1, according to more-meaningful age categories regrouped to reflect other epidemiological considerations pertaining to susceptibility and risk (eg, school/workplace social clustering and age-based immunization recommendations). 2009 Prepandemic Period During the 2009 prepandemic period, overall age-standardized seroprotection was 10% (95% CI, 8%–11%), with broad susceptibility across all age groups (seroprotection <10%) except the elderly among whom seroprotection was 22% for those 65–79 years but approached 80% in very old individuals (ie, those ≥80 years). Consistent with these serosusceptibility patterns, overall laboratory-confirmed A(H1N1)pdm09 incidence was 132 cases/100 000. Values were highest in children aged 5–9 years (375 cases/100 000), <5 years (301 cases/100 000), and 10–19 years (283 cases/100 000) and decreased across successive Table 1. Characteristics of Patients Included in Serial Cross-sectional Serosurveys and Assessed for 2009 Pandemic Influenza A(H1N1) pdm09 Antibody Before the Pandemic (in 2009), After the Pandemic (in 2010), and During Spring 2013, Southwestern British Columbia, Canada 2009 Prepandemic Perioda 2010 Postpandemic Periodb Spring 2013 Patients, No. Median Age, y Female Sex, % Patients, No. Median Age, y Female Sex, % Patients, No. Median Age, y Female Sex, % <5 101 3 43 126 3 45 24 3 50 5–9 95 7 49 98 7 50 21 7 48 10–19 20–29 98 100 15 25 61 71 103 100 16 26 50 73 50 50 16 27 68 70 Age Group, y 30–39 100 33 62 100 35 65 50 34 90 40–49 50–59 98 101 45 54 68 47 100 100 45 54 59 59 50 50 46 56 60 50 60–69 103 65 55 100 64 46 50 66 56 70–79 80–89 97 49 73 84 42 53 100 100 74 83 51 54 50 50 75 83 42 40 51 94 59 100 92 64 51 92 73 993 40 55 1127 44 56 496 50 60 ≥90 Overall a Adapted from [1] by permission of Oxford University Press. b Adapted from [8] with permission of the ©2010 Canadian Medical Association or its licensors (available at: http://www.cmaj.ca). 2 • JID • BRIEF REPORT Downloaded from http://jid.oxfordjournals.org/ by guest on March 15, 2016 Per convention, seroprotection was defined as a titer of ≥40, explored in sensitivity analyses at titers of ≥10 and ≥80. Mean GMTs and the percentage of individuals considered seroprotected, with 95% confidence intervals (CIs), were summarized by age group. Age-standardized estimates of the percentage seroprotected overall were derived by the direct method based on contemporary British Columbia population estimates (Supplementary Table 1). The British Columbia PHMRL conducts the majority of influenza virus testing for British Columbia, using real-time reversetranscription polymerase chain reaction. The following data were calculated for the period spanning week 40 (the first week of October) to week 18 (the last week of April) for the 2009–2010 pandemic influenza season, the 2010–2011 influenza season, and the 2013–2014 influenza season: the incidence of laboratoryconfirmed A(H1N1)pdm09 detections per 100 000 population by age group, the percentage distribution of A(H1N1)pdm09 laboratory detections by age group and the percentage distribution of the British Columbia population by age group. The Clinical Research Ethics Board at the University of British Columbia approved the anonymized serosurveys, for which patient consent was not required. Incidence data based on notifiable disease surveillance are exempt from research ethics board approval. Downloaded from http://jid.oxfordjournals.org/ by guest on March 15, 2016 Figure 1. Percentage distribution of A(H1N1)pdm09 laboratory detections, percentage distribution of the British Columbia population, seroprotection estimates, and laboratory-confirmed incidence of infection due to the 2009 pandemic influenza A(H1N1) strain (A[H1N1]pdm09) per 100 000 population, by age group, during the period spanning week 40 (the first week of October) to week 18 (the last week of April) for the 2009 prepandemic period (A), the BRIEF REPORT • JID • 3 adult age groups to <25 cases/100 000 among those ≥80 years. All pediatric age groups were disproportionately affected relative to the British Columbia population, whereas adults were relatively spared. The median age of individuals with laboratory-confirmed A(H1N1)pdm09 infection was 19 years. Spring 2013 During spring 2013, overall age-standardized seroprotection was virtually identical to findings from the 2010 postpandemic period (46%; 95% CI, 42%–51%). The U-shaped age distribution also persisted but no longer included children aged <5 years, among whom seroprotection fell below 20%, significantly lower than the 2010 estimate and lower than all other age groups in 2013 (Supplementary Table 1). Peak seroprotection was still observed in children aged 5–19 years (68%) and very old individuals (age, ≥80 years; 80%). After preschool-aged children, seroprotection was lowest in adults aged 20–39 years (47%) and adults aged 40–64 years (36%). The overall A(H1N1)pdm09 incidence was 30 cases/100 000, >5-fold higher than during the 2010–2011 season. The incidence was a mirror image of findings for seroprotection, by age. Children aged <5 years, among whom the lowest seroprotection frequency was observed, had the highest A(H1N1)pdm09 incidence (76 cases/100 000) and were disproportionately affected. A secondary incidence peak (32 cases/ 100 000) was observed in adults aged 20–64 years, among DISCUSSION By comparing serial preseason serosurvey findings to subsequent surveillance-based incidence in British Columbia, we show that age-dependent serosusceptibility may in part explain the A(H1N1)pdm09 resurgence and the shift in age distribution during the 2013-2014 influenza season. A mirror-image relationship between seroprotection and surveillance-based incidence by age group is evident in all 3 analyses, but findings were more striking in the spring 2013, despite the greater variation by age for both indicators. The consistency of this relationship across age groups and study periods suggests that host susceptibility plays a prominent role in the force of infection. It also suggests that agespecific estimates of seroprotection may inform the risk of an influenza epidemic and the relative disease burden, warranting further exploration in empirical analyses and modeling simulations. The overall postpandemic estimates of A(H1N1)pdm09 seroprotection were identical in spring 2010 and 2013 (46% for each period). However, during the subsequent 2010–2011 season, only minor A(H1N1)pdm09 activity was reported in British Columbia, whereas in 2013–2014 a moderately severe epidemic occurred. The incidence was >5-fold higher in 2013–2014, compared with 2010–2011, and children <5 years of age and workingaged adults were disproportionately affected. Analysis by median age showed a progressive shift in the age distribution of laboratory-confirmed A(H1N1)pdm09 infection toward young and middle-aged adults from 2009–2010 to 2013–2014. A closer examination of age-specific seroprotection and its influences may help to understand these surveillance changes with time. This analysis displays a protective cohort effect introduced in 2009 by high rates of A(H1N1)pdm09 infection and immunization, but it also reveals a negative cohort effect arising from the accumulation of susceptibility in children <5 years of age with the successive addition of vulnerable newborns to the population each year since. A U-shaped age distribution of seroprotection was evident overall in both 2010 [8] and 2013, with the Figure 1 continued. 2010 postpandemic period (B), and spring 2013 (C), southwestern British Columbia, Canada. The incidence of A(H1N1)pdm09 detections per 100 000 British Columbia population, by age group, was based on diagnostic laboratory testing conducted at the British Columbia Public Health Microbiology and Reference Laboratory. Seroprotection was estimated by serial cross-sectional serosurveys and defined as a hemagglutination inhibition titer of ≥40. The overall estimates of seroprotection were age standardized by the direct method to population estimates for the specified year (Supplementary Table 1). Data in panel A are adapted from [1] by permission of Oxford University Press. Data in panel B are adapted from [8] with permission of the ©2010 Canadian Medical Association or its licensors (http://www.cmaj.ca). 4 • JID • BRIEF REPORT Downloaded from http://jid.oxfordjournals.org/ by guest on March 15, 2016 2010 Postpandemic Period After the pandemic, in spring 2010, the overall age-standardized seroprotection was 46% (95% CI, 43%–50%), which was significantly greater than that during the prepandemic period. Seroprotection followed a U-shaped age distribution, with the highest values observed in individuals aged <20 years and very old individuals (ie, those aged ≥80 years; approximately 70% in each group). Among adult age groups from 20 to 79 years, seroprotection gradually decreased with increasing age, from 45% among those aged 20–39 years to 25% among those aged 65–79 years. The overall A(H1N1)pdm09 incidence during the 2010–2011 surveillance period was 6 cases/100 000, which was >20-fold lower than during the 2009 pandemic. The incidence was highest in children <10 years and adults 20–39 years, at 8–10 cases/100 000, and each group was disproportionately affected relative to the British Columbia population. The median age of individuals with laboratory-confirmed A(H1N1) pdm09 infection was 30 years, higher by more than a decade than the median age during 2009–2010. whom the second-lowest seroprotection frequency was observed. Conversely, in school-aged children and very old adults with comparably high seroprotection, A(H1N1)pdm09 incidence was lowest, at approximately 15–20 cases/100 000. The median age of individuals with A(H1N1)pdm09 infection was 38 years, double that for the 2009 pandemic and higher also by nearly a decade, compared with findings from 2010–2011. disease burden [12] and of adult work-related activities in facilitating spread [13]. Mathematical modeling studies from Canada have previously predicted such a demographic shift in disease burden away from school-aged children to adults following the 2009 pandemic [14]. Others have suggested that it may take decades for this gradual shift following pandemics to culminate in the disproportionate mortality burden in the elderly population that is typically associated with seasonal influenza [15]. Limitations to this study include its generalizability to other settings, the inherent variability in laboratory assays, and the definitions of seroprotection, for which HI titers of ≥40 represent the 50% seroprotective threshold and may further vary by age. The smaller sample size per age group included in 2013, particularly among very young individuals, may also be associated with greater variability in seroprotection estimates, as suggested by the 95% CIs. Changes in other immunologic markers, such as cell-mediated, mucosal, or innate immunity, which may also show variation in cross-reactivity and/or duration, have not been addressed here. Patient specimens were obtained before the pandemic and before the influenza seasons as cross-sectional convenience samples of anonymized residual sera; the same individuals were not followed longitudinally, and we are unable to explore prior or subsequent immunization or infection history, waning immunity, comorbidity, healthcare-seeking behaviors, testing proclivities, or other effects on age-dependent susceptibility or surveillance estimates. The influence of seasonal timing (eg, 2009 fall pandemic vs 2013 winter epidemic) on transmission dynamics may also be relevant. Community- versus facility-based influenza virus detections are not distinguished in the provincial laboratory data, and we acknowledge that patient specimens may be skewed toward those with more severe illness and high-risk conditions. In summary, in a population that was substantially immune following the 2009 pandemic, the addition of just a few extra cohorts of susceptible children <5 years of age may have been sufficient to tip the epidemic threshold, enabling a resurgence in A(H1N1)pdm09 incidence during the 2013-2014 season despite antigenic conservation of the circulating strain. Other variation in age-dependent serosusceptibility may account for greater involvement of young and middle-aged adults, compared with school-aged children or elderly individuals. Serosurveys may inform risk assessment and relative disease burden by age, but other sociobehavioral and immunoepidemiologic factors must also be considered. Finally, in the context of accumulated and selective immune pressure on virus evolution, ongoing monitoring for antigenic changes in A(H1N1)pdm09 and corresponding changes in age-related seroprotection is warranted. Supplementary Data Supplementary materials are available at The Journal of Infectious Diseases online (http://jid.oxfordjournals.org). Supplementary materials consist of BRIEF REPORT • JID • 5 Downloaded from http://jid.oxfordjournals.org/ by guest on March 15, 2016 notable exception of children aged <5 years in 2013, most of whom were born after the 2009 pandemic. As such, children <5 years of age lacked the substantial frequency of seroprotection in 2013 (<20%) that was evident in 2010 (approximately 70%). Beginning in late October 2009, approximately 45% of young British Columbia children overall ultimately received the monovalent A/California/07/2009-like A(H1N1)pdm09 vaccine (>95% of vaccine distributed was the AS03-adjuvanted formulation), for which vaccine effectiveness of >90% was estimated overall [9]. Advancing birth-cohort effects resulting from high infection and immunization rates during the 2009 pandemic may still be evident in the seroprotection estimate of >65% among school-aged children in 2013 but now substantially reduced in the very young. Similar cohort effects may also be reflected in the high seroprotection observed both before and after the pandemic in very old individuals (age, ≥80 years), among whom robust childhood priming with related 1918-like viruses and cross-reactive lifetime boosting exposures have been hypothesized. Annual seasonal influenza vaccine coverage estimates are approximately 30% overall in British Columbia, with higher values in elderly individuals (approximately 60%) and those of any age with comorbidity (approximately 40%) [10]. Infants aged 6–23 months have been eligible for publicly funded vaccine since 2004, but healthy preschool-aged children (age, 2–4 years) have only been eligible since 2012–2013, and healthy school-aged children are not eligible. High levels of repeat immunization with (nonadjuvanted) seasonal vaccine containing the same A/California/07/2009-like antigen [2–6] since 2009 may have reinforced A(H1N1)pdm09 seroprotection in elderly individuals, as suggested by comparison of adults aged 65–79 years in 2013 (52%; 95% CI, 18%–32%) with those in 2010 (25%; 95% CI, 41%–63%), but with lesser contribution to seroprotection in other age groups. Beyond the high levels of serosusceptibility found among children <5 years of age in 2013, a residual pocket of persistent low-level seroprotection was also present among working-aged adults, including both young parental-aged individuals (approximately 45%) and middle-aged individuals (approximately 35%). The addition of just a few more cohorts of susceptible newborns between 2010 and 2013 to the preexisting vulnerability in parents with whom they have close interpersonal contacts may have been sufficient to reignite disease propagation through the population, despite substantial seroprotection in other age groups. Propagation at the workplace and through adult social networks may also have accounted for the shift in age distribution to include more young and middle-aged adults in 2013–2014. Higher rates of comorbidity in the latter groups may also explain their overrepresentation in severe outcome surveillance reported elsewhere [7]. Although some studies have emphasized the pivotal role of school-aged children in amplifying influenza risk [11], others have suggested a greater role of preschool-aged children in predicting the population-based data provided by the author that are published to benefit the reader. The posted materials are not copyedited. The contents of all supplementary data are the sole responsibility of the authors. Questions or messages regarding errors should be addressed to the author. 5. Notes References 1. Skowronski DM, Hottes TS, McElhaney JE, et al. Immuno-epidemiologic correlates of pandemic H1N1 surveillance observations: higher antibody and lower cell-mediated immune responses with advanced age. J Infect Dis 2011; 203:158–67. 2. Public Health Agency of Canada. FluWatch: weekly reports 2013-2014 season. Ottawa: Public Health Agency of Canada, 2014. http://www. phac-aspc.gc.ca/fluwatch/index-eng.php. Accessed 14 July 2014. 3. Skowronski D, Chambers C, Sabaiduc S, et al. Interim estimates of 2013/14 vaccine effectiveness against influenza A(H1N1)pdm09 from Canada’s sentinel surveillance network, January 2014. Euro Surveill 2014; 19:pii:20690. 4. BC Centre for Disease Control. BC influenza surveillance bulletins: 2013-2014. Vancouver: BC Centre for Disease Control, 2014. 6 • JID • BRIEF REPORT 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. Downloaded from http://jid.oxfordjournals.org/ by guest on March 15, 2016 Disclaimer. The funding organization did not have a role in study design, data collection, analysis, decision to publish, or preparation of the manuscript Financial support. This work was supported by the Michael Smith Foundation for Health Research (grant OT-GIA-00012091). Potential conflicts of interest. Within 36 months of manuscript submission, M. K. received research grants from Roche, Merck, Siemens, Hologic (Gen-Probe), and Boerhinger Ingelheim. G. D. S. received research grants from GlaxoSmithKline (GSK) for unrelated vaccine studies and travel fee reimbursement, without honorarium, to attend an ad hoc GSK advisory board meeting. D. P. was a shareholder with BC Biomedical Laboratories, which contributed anonymized sera included in serosurveys, but he did not receive direct compensation for contributions to this analysis. S. S. was partially funded by the Michael Smith Foundation for Health Research grant that supported this study. 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. http://www.bccdc.ca/dis-cond/DiseaseStatsReports/influSurveillance Reports.htm. Accessed 11 April 2014. World Health Organization (WHO). Recommended composition of influenza virus vaccines for use in the 2014–2015 northern hemisphere influenza season. Geneva: WHO, 2014. http://www.who.int/influenza/ vaccines/virus/recommendations/201402_recommendation.pdf?ua=1. Accessed 11 April 2014. Centers for Disease Control and Prevention (CDC). Past weekly surveillance reports. 2013-2014. Atlanta: CDC, 2014. http://www.cdc.gov/flu/ weekly/pastreports.htm. Accessed 11 April 2014. McNeil SA, Shinde V, Andrew M, et al. Interim estimates of 2013-14 influenza clinical severity and vaccine effectiveness in the prevention of laboratory-confirmed influenza-related hospitalization, Canada, February 2014. Euro Surveill 2014; 19:pii=20729. Skowronski DM, Hottes TS, Janjua NZ, et al. Prevalence of seroprotection against the pandemic (H1N1) virus after the 2009 pandemic. CMAJ 2010; 182:1851–6. Skowronski DM, Janjua NZ, De Serres G, et al. Effectiveness of AS03adjuvanted pandemic H1N1 vaccine: case-control evaluation based on sentinel surveillance system in Canada, autumn 2009. BMJ 2011; 342: c7297. BC Centre for Disease Control. Influenza immunization coverage. Vancouver: BC Centre for Disease Control, 2013 [updated 22 August 2013]. http://www.bccdc.ca/imm-vac/BCImmunizationCov/flucoverage/default. htm. Accessed 27 April 2014. Rudenko LG, Slepushkin AN, Monto AS, et al. Efficacy of live attenuated and inactivated influenza vaccines in schoolchildren and their unvaccinated contacts in Novgorod, Russia. J Infect Dis 1993; 168:881–87. Brownstein JS, Kleinman KP, Mandl KD. Identifying pediatric age groups for influenza vaccination using a real-time regional surveillance system. Am J Epidemiol 2005; 162:686–93. Viboud C, Bjornstad ON, Smith DL, Simonsen L, Miller MA, Grenfell BT. Synchrony, waves, and spatial hierarchies in the spread of influenza. Science 2006; 312:447–51. Bansal S, Pourbohloul B, Hupert N, Grenfell B, Meyers LA. The shifting demographic landscape of pandemic influenza. PLoS One 2010; 5: e9360. Simonsen L, Clarke MJ, Schonberger LB, Arden NH, Cox NJ, Fukuda K. Pandemic versus epidemic influenza mortality: a pattern of changing age distribution. J Infect Dis 1998; 178:53–60.