We assessed influenza vaccine effectiveness and serologic immune correlates in a household cohort during the 2014–2015 influenza season. Vaccines were not effective against antigenically drifted influenza A(H3N2). Among older children and adults, <20% had ≥4-fold antibody titer rise following vaccination.
Keywords: influenza, vaccine effectiveness, household cohort study, vaccine response, serologic susceptibility
Abstract
Background
Antigenically drifted A(H3N2) viruses circulated extensively during the 2014–2015 influenza season. Vaccine effectiveness (VE) was low and not significant among outpatients but in a hospitalized population was 43%. At least one study paradoxically observed increased A(H3N2) infection among those vaccinated 3 consecutive years.
Methods
We followed a cohort of 1341 individuals from 340 households. VE against laboratory-confirmed influenza was estimated. Hemagglutination-inhibition and neuraminidase-inhibition antibody titers were determined in subjects ≥13 years.
Results
Influenza A(H3N2) was identified in 166 (12%) individuals and B(Yamagata) in 34 (2%). VE against A(H3N2) was −3% (95% confidence interval [CI]: −55%, 32%) and similarly ineffective between age groups; increased risk of infection was not observed among those vaccinated in 2 or 3 previous years. VE against influenza B(Yamagata) was 57% (95% CI: −3%, 82%) but only significantly protective in children <9 years (87% [95% CI: 43%, 97%]). Less than 20% of older children and adults had ≥4-fold antibody titer rise against influenza A(H3N2) and B antigens following vaccination; responses were surprisingly similar for antigens included in the vaccine and those similar to circulating viruses. Antibody against A/Hong Kong/4801/14, similar to circulating 2014–2015 A(H3N2) viruses and included in the 2016–2017 vaccine, did not significantly predict protection.
Conclusions
Absence of VE against A(H3N2) was consistent with circulation of antigenically drifted viruses; however, generally limited antibody response following vaccination is concerning even in the context of antigenic mismatch. Although 2014–2015 vaccines were not effective in preventing A(H3N2) infection, no increased susceptibility was detected among the repeatedly vaccinated.
Annual evaluations of influenza vaccine effectiveness (VE) in ambulatory populations have been conducted in an expanding number of countries [1–3]. These studies typically employ the case-test negative study design, which compares influenza vaccination status between reverse-transcription real-time polymerase chain reaction (RT-PCR) positive and negative subjects [4, 5]. A consistent finding has been relatively high VE for influenza A(H1N1)pdm09 and type B viruses but lower VE for A(H3N2) [6]. In some seasons, lower VE has been observed among those vaccinated in both the current and prior seasons compared to those vaccinated only in the current year [1, 7–10].
Since 2010, we have carried out annual influenza VE assessments in a cohort of households with children [7, 8, 12]. This study was originally designed to offer a comparison to VE studies carried out in healthcare settings using the case-test negative design but also to examine illnesses not severe enough to result in an ambulatory visit. Prospective, longitudinal follow-up also allows for collection of blood specimens, enabling the relation of vaccine effects to antibody titers against the hemagglutinin (HA) and neuraminidase (NA).
In February 2014, A/Texas/50/2012 (A/Texas) was chosen as the A(H3N2) component for the 2014–2015 Northern Hemisphere vaccine [13]. However, the A(H3N2) viruses that circulated during the 2014–2015 season were drifted from this vaccine strain [14, 15]. Most circulating A(H3N2) viruses failed to agglutinate red blood cells making antigenic characterization by hemagglutination-inhibition (HAI), the traditional method, difficult [14]. Therefore, genetic characterization of circulating viruses was carried out with most identified as belonging to the variant 3C.2a clade; this clade was genetically distinct from the 3C.1 clade which included the A/Texas vaccine strain [16, 17]. The resulting influenza season was severe, with pneumonia and influenza mortality elevated in the United States for 8 consecutive weeks [14]. In addition to widespread circulation of drifted A(H3N2), both influenza B lineages also circulated.
We carried out 3 parallel influenza VE studies in Southeastern Michigan in 2014–2015; A(H3N2) 3C.2a viruses predominated locally as they did nationally. As part of the US Influenza VE Network we contributed to analyses that found no significant VE against outpatient visits resulting from A(H3N2) 3C.2a virus infection [17]. In contrast, VE was estimated to be 43% in a study carried out in the hospital setting, suggesting that vaccination may have prevented more severe disease [18]. Here we report the results of the third study, a prospectively followed cohort of households with children, including VE against acute respiratory illnesses (ARI) of any severity resulting from infection with A(H3N2) and influenza B viruses. We also characterize preseason susceptibility and response to vaccination in terms of serum antibody against HA and NA of vaccine and circulating strains.
METHODS
Recruitment and Enrollment
Eligible households with ≥3 members, including ≥2 children <18 years, were identified, recruited, and enrolled from June through September 2014, as previously described [7, 8, 12]. Adults provided written informed consent for participation for themselves and their children; children 7–17 years provided oral assent. Demographic data were reported, and access to health system electronic medical records (EMR) was granted. All study visits were carried out at the University of Michigan School of Public Health (UM-SPH). The University of Michigan Medical School institutional review board reviewed and approved the study.
Acute Respiratory Illnesses Surveillance and Laboratory Testing
Surveillance was carried out October 2014 through May 2015. Households were instructed to report all ARI at illness onset and were sent weekly email queries to identify newly onset ARI. Case definitions for eligible illnesses were defined by symptoms tailored to those ≥3 years (≥2 of: cough, fever/feverishness, nasal congestion, chills, headache, body aches, or sore throat) and, separately, children <3 years (≥2 of: cough, fever/feverishness, runny nose/congestion, difficulty breathing, fussiness/irritability, fatigue or loss of appetite). Subjects with eligible illnesses had combined throat and nasal swab specimens (children <3 years: nasal swab only) collected by study staff ≤7 days from illness onset.
Respiratory specimens were tested for influenza by RT-PCR in the investigators’ laboratory at UM-SPH using primers, probes, and testing protocols developed and provided by the Influenza Division of the Centers for Disease Control and Prevention (CDC). Influenza A(H3N2) positive specimens were genetically characterized by pyrosequencing in the CDC influenza research laboratory [16, 17].
Blood Specimen Collection and Serologic Assays
Blood was collected from willing participants ≥13 years at enrollment and scheduled visits before the influenza season (November–December). Sera from collected specimens were tested in HAI assays using monovalent inactivated influenza vaccine (IIV) subunit material (Sanofi-Pasteur) representing the A/Texas-like virus (3C.1) present in the 2014–2015 Northern Hemisphere vaccine and the A/Hong Kong /4801/2014-like virus (A/Hong Kong; 3C.2a) representing the dominant variant genetic group that circulated in 2014–2015. Circulating 3C.2a viruses were considered antigenically distinct from 3C.1 viruses [19]. Additional HAI testing targeted influenza B strains included in 2014–2015 vaccines—B/Massachusetts/2/2012-like (B/Massachusetts; Yamagata lineage; trivalent and quadrivalent vaccines) and B/Brisbane/60/2008-like (B/Brisbane; Victoria lineage; quadrivalent vaccines); 2014–2015 circulating B viruses were considered vaccine-like [19]. Sera were also tested by neuraminidase-inhibition (NAI) assay, using an inactivated reassortant influenza virus with NA representing the 2014–2015 A(H3N2) vaccine strain, and a mismatched HA (H6 subtype) to avoid HA-specific antibody interference [20].
Statistical Analyses
Households were characterized by size and composition, and subjects by demographic characteristics, EMR documented high-risk health status [21], and influenza vaccination status. Subjects made their own decisions regarding vaccine receipt; documentation of vaccination for current and prior seasons was based on evidence in the EMR or Michigan Care Improvement Registry. Vaccination status was also informed by self-report if a plausible date and location of receipt were reported for the current season; plausible self-report of prior season vaccination was also considered for individuals who participated in previous seasons. Associations between subject characteristics and vaccination status and influenza outcomes were compared by χ2 or Fisher exact tests.
We estimated hazard ratios (HR) using Cox proportional hazard models adjusted for age and high-risk health status; robust variances using sandwich estimators were computed to account for household clustering [22]. The effectiveness of ≥1 dose of influenza vaccine in preventing reverse-transcription polymerase chain reaction (RT-PCR) confirmed influenza was calculated as 100× [1 – HR]. Vaccination status was modeled as time-varying, with subjects considered vaccinated ≥14 days after vaccination. VE by age and in preventing community-acquired and household-acquired outcomes was examined as previously described [7, 8, 12]. VE was also estimated for each combination of current and prior season vaccination (i.e., current only, current and prior, and prior only) with subjects unvaccinated in both seasons referent. VE by prior season vaccination was estimated for those ≥9 years for whom a single dose of vaccine is recommended and separately for those 3–8 years. The effectiveness of IIV and live-attenuated influenza vaccine (LAIV) was also estimated and compared among children 2–17 years.
Preseason susceptibility to A(H3N2) and B(Yamagata) was assessed with HAI and NAI titers measured in sera collected in the fall (or at summer enrollment for subjects without fall specimens and no evidence of vaccine receipt); for vaccinated subjects, these represented post-vaccination titers. HAI and NAI titers were calculated as the reciprocal (e.g., 40) of the highest dilution of sera (e.g., 1:40) that inhibited HA or NA activity. Titers were log-base 2 transformed, and the mean of the transformed values was calculated and then exponentiated to obtain the geometric mean titer (GMT). GMTs were compared by vaccination and infection status using Wilcoxon rank-sum tests. We used multivariable logistic regression models to estimate the odds ratio (OR) of influenza infection associated with a 2-fold increase in preseason HAI and NAI titers, modeled as continuous predictors. The proportion of vaccinated subjects with ≥4 fold rises in titer in paired serum specimens bracketing vaccination was also determined.
All statistical analyses were performed using SAS (release 9.3; SAS Institute). A P-value of <.05 or, for VE estimates, a positive lower bound of a confidence interval (CI) indicated statistical significance.
RESULTS
Households and Participant Characteristics
A total of 1431 participants, including 862 (60%) children <18 years, from 340 households were enrolled and followed in the 2014–2015 study year (Table 1). Household size ranged from 3 to 9 members (median: 4); 73% of households had ≥1 child <9 years. Overall, 195 (14%) subjects had high-risk health conditions, and 982 (69%) subjects received ≥1 dose of 2014–2015 influenza vaccine. Vaccine coverage significantly varied by age and race categories, and was significantly higher (78% vs. 67%, P < .05) among subjects with high-risk health conditions. Among vaccinated subjects, 772 (79%) received IIV and 210 (21%) received LAIV (196 in children 2–17 years); all LAIV and 85% of IIV were quadrivalent.
Table 1.
Characteristics of Participating Household Members [N = 1431] During the 2014–2015 Influenza Season by Documented or Plausible Self-Reported Influenza Vaccine Receipt [n = 982], and Influenza A(H3N2) [n = 166] and Influenza B (Yamagata) [n = 34] Case Status: Household Influenza Vaccine Effectiveness Study, Ann Arbor, Michigan
| Participant Characteristics | All Subjects N (%)a | Documented or Plausible Self-Reported Influenza Vaccinationbc N (%) | Influenza A(H3N2) Casesd N (%) | Influenza B(Yamagata) Cases N (%)e |
|---|---|---|---|---|
| Age category | ||||
| <9 years | 450 (31.4) | 326 (72.4)f | 67 (14.9) | 11 (2.4) |
| 9–17 years | 412 (28.8) | 260 (63.1) | 44 (10.7) | 10 (2.4) |
| 18–49 years | 508 (35.5) | 349 (68.7) | 48 (9.4) | 12 (2.4) |
| ≥50 years | 61 (4.3) | 47 (77.0) | 7 (11.5) | 1 (1.6) |
| Race categories | ||||
| White | 1014 (70.9) | 709 (69.9)f | 118 (11.6) | 28 (2.8) |
| Asian | 128 (8.9) | 101 (78.9) | 15 (11.7) | 1 (0.8) |
| Black | 118 (8.2) | 70 (59.3) | 11 (9.3) | 2 (1.7) |
| Other/Unknown | 171 (11.9) | 102 (59.6) | 22 (12.9) | 3 (1.8) |
| Sex | ||||
| Female | 746 (52.1) | 513 (68.8) | 88 (11.8) | 20 (2.7) |
| Male | 685 (47.9) | 469 (68.5) | 78 (11.4) | 14 (2.0) |
| High-risk health condition | ||||
| Any | 195 (13.6) | 152 (77.9)f | 29 (14.9) | 4 (2.1) |
| None | 1236 (86.4) | 830 (67.2) | 137 (11.1) | 30 (2.4) |
| Documented or plausible self-reported influenza vaccination | ||||
| Yes | 982 (68.6) | — | 106 (10.8) | 17 (1.7)f |
| No | 449 (31.4) | — | 60 (13.4) | 17 (3.8) |
| Total | 1431 (100) | 982 (68.6) | 166 (11.6) | 34 (2.4) |
aDenominator for percentages presented in this column is all subjects (N = 1431).
bAt least one influenza vaccine received during the 2014–2015 vaccination period as documented in the medical record or state immunization registry, or alternatively self-reported with plausible location and date of vaccine receipt; subjects with laboratory confirmed influenza were considered vaccinated if vaccine was administered ≥14 days prior to illness onset.
cDenominator for percentages presented in this column is all subjects (vaccinated and unvaccinated) in the given characteristic row.
dDenominator for percentages presented in this column is all subjects (with and without influenza A[H3N2]) in the given characteristic row.
eDenominator for percentages presented in this column is all subjects (with and without influenza B[Yamagata]) in the given characteristic row.
fPearson’s χ2 or Fisher exact P-value < .05; comparing vaccinated and unvaccinated subjects or subjects with and without laboratory-confirmed influenza.
Acute Respiratory Illnesses Surveillance and Influenza Outcomes
Influenza circulated locally from early-November 2014 through May 2015; A(H3N2) predominated and circulated through February, and influenza B circulated February through May. During surveillance, 730 (51%) participants from 269 (79%) households reported 1434 ARI and 1364 (95%) specimens were collected. Influenza A(H3N2) was identified in 166 (12%) individuals and 94 (28%) households; there were no A(H1N1)pdm09 cases. Influenza B was identified in 44 (3%) individuals (34 B[Yamagata], 10 B[Victoria]) and 31 (9%) households. Because there were few B(Victoria) infections, subsequent analyses were limited to B(Yamagata). Of the A(H3N2) viruses, 118 (71%) were characterized by pyrosequencing; 111 (94%) belonged to the variant 3C.2a clade that predominated nationally, 5 (4%) to the separate variant 3C.3a clade, and 2 (2%) to the vaccine-like 3C.3 clade. Nine participants were infected with both A(H3N2) and influenza B, all in separate illnesses. Risk of A(H3N2) or B(Yamagata) infection did not vary by age category, race, sex, or high-risk health status. Fifty-two A(H3N2) and 5 B(Yamagata) cases were considered household-acquired (secondary infection risk: 17% and 6%, respectively), based on exposure to 114 A(H3N2) and 29 B(Yamagata) community-acquired infections (Table 2).
Table 2.
Estimates of Vaccine Effectiveness in Preventing Influenza A(H3N2) and Influenza B(Yamagata) Outcomes, and the Subsets of Community-Acquired, Household-Acquired and Medically-Attended Outcomes, by Age Category, Plus Estimates in Children age 2–17 Years by Vaccine Type, During the 2014–2015 Influenza Season: Household Influenza Vaccine Effectiveness (HIVE) Study, Ann Arbor, Michigan
| Vaccinated N positive / N (%) | Unvaccinated N positive / N (%) | Vaccine Effectivenessa VE % (95% CI)b | |
|---|---|---|---|
| Influenza A(H3N2) | |||
| Overall | 106/982 (10.8) | 60/449 (13.4) | −3 (−55, 32) |
| <9 years | 40/326 (12.3) | 27/124 (21.8) | 20 (−43, 56) |
| 9–17 years | 25/260 (9.6) | 19/152 (12.5) | −1 (−96, 48) |
| ≥18 years | 41/396 (10.4) | 14/173 (8.1) | −61 (−215, 17) |
| Community-acquired | 71/982 (7.2) | 43/449 (9.6) | 0 (−54, 35) |
| Household-acquired c | 35/202 (17.3) | 17/103 (16.5) | 8 (−72, 51) |
| Influenza B(Yamagata) | |||
| Overall | 17/982 (1.7) | 17/449 (3.8) | 57 (−3, 82) |
| <9 years | 3/326 (0.9) | 8/124 (6.5) | 87 (43, 97) |
| 9–17 years | 5/260 (1.9) | 5/152 (3.3) | 44 (−85, 83) |
| ≥18 years | 9/396 (2.3) | 4/173 (2.3) | 1 (−301, 76) |
| Community-acquired | 15/982 (1.5) | 14/449 (3.1) | 53 (−14, 80) |
| Household-acquired c | 2/48 (4.2) | 3/31 (9.7) | 50 (−196, 92) |
Abbreviations: CI, confidence interval; VE, vaccine effectiveness.
aEffectiveness of at least 1 dose of influenza vaccine in preventing laboratory-confirmed influenza A(H3N2) or B(Yamagata). Vaccination status was modeled as time-varying with subjects considered vaccinated 14 days after vaccine receipt. To adjust for correlation of exposures and outcomes among subjects in the same household, robust variances for model parameter estimates were computed using sandwich estimators [22]. Models were adjusted for age in months (natural cubic spline) and medical record documented high-risk health status (present/absent).
bVE% = 100*(1 − hazard ratio).
cHousehold-acquired cases were defined by transmission link to a community-acquired index case if both cases were due to the same influenza type and subtype/lineage and if illness onset in the secondary case occurred 1–7 days after illness onset in the index case.
Influenza Vaccine Effectiveness
Influenza vaccines were not effective in preventing A(H3N2) infections overall and were similarly ineffective across age groups (Table 2). There was no evidence of VE against A(H3N2), or increased risk of infection, for any combination of vaccination across 2 (Supplemental Table 1) or 3 years (Supplemental Table 2). IIV and LAIV were similarly ineffective against A(H3N2) among children 2–8 years and all children 2–17 years (Table 3). VE against B(Yamagata) was higher overall (57%, 95% CI: −3%, 82%), but surprisingly, significant VE was limited to children <9 years (87%, 95% CI: 43%, 97%) with decreasing estimates by age (Table 2). The VE estimate against B(Yamagata) was high, but not significant, for both LAIV and IIV in children age 2–8 years; for all children 2–17 years, only the VE estimate for LAIV was statistically significant (Table 3).
Table 3.
Estimates of Vaccine Effectiveness in Preventing Influenza A (H3N2) and Influenza B (Yamagata) Outcomes in Children Age 2–17 Years by Vaccine Type, During the 2014–2015 Influenza Season: Household Influenza Vaccine Effectiveness (HIVE) Study, Ann Arbor, Michigan
| Vaccinated N positive / N (%) | Unvaccinated N positive / N (%) | Vaccine Effectivenessa VE % (95% CI)b | |
|---|---|---|---|
| Influenza A (H3N2) | |||
| 2–17 years | |||
| LAIV | 19/196 (9.7) | 46/268 (17.2) | 31 (-24,61) |
| IIV | 42/351 (12.0) | 46/268 (17.2) | 5 (-59, 43) |
| 2–8 years | |||
| LAIV | 14/125 (11.2) | 27/116 (23.3) | 31 (-42, 66) |
| IIV | 22/162 (13.6) | 27/116 (23.3) | 24 (-49, 61) |
| Influenza B (Yamagata) | |||
| 2–17 years | |||
| LAIV | 1/196 (0.5) | 12/268 (4.5) | 90 (16, 99) |
| IIV | 7/351 (2.0) | 12/268 (4.5) | 59 (-16, 85) |
| 2–8 years | |||
| LAIV | 1/125 (0.8) | 7/116 (6.0) | 88 (-5, 99) |
| IIV | 2/162 (1.2) | 7/116 (6.0) | 80 (-13, 96) |
Abbreviations: CI, confidence interval; IIV, inactivated influenza vaccine; LAIV, live attenuated influenza vaccine; VE, vaccine effectiveness.
aEffectiveness of at least 1 dose of influenza vaccine in preventing laboratory-confirmed influenza A(H3N2) or B(Yamagata). Vaccination status was modeled as time-varying with subjects considered vaccinated 14 days after vaccine receipt. To adjust for correlation of exposures and outcomes among subjects in the same household, robust variances for model parameter estimates were computed using sandwich estimators [22]. Models were adjusted for age in months (natural cubic spline) and medical record documented high-risk health status (present/absent).
bVE% = 100*(1 − hazard ratio).
Serologic Antibody Response and Susceptibility
Paired pre- and ≥30 days post-vaccination sera were available from 129 vaccinated subjects (26% of vaccinated ≥13 years). Less than 50% of vaccinated subjects had any titer increase following vaccination and the proportion of subjects with a ≥4-fold titer rise was low (range: 12%–17%) for all measured targets (Table 4). In nearly all cases, a lower proportion of those vaccinated in 2013–2014 responded following 2014–2015 vaccination than those unvaccinated in 2013–2014; however, this was only statistically significant for HAI response to A/Texas (H3N2) and B/Brisbane (Victoria).
Table 4.
Immune Response Following Receipt of 2014–2015 Influenza Vaccine
| GMT Pre-vaccination | GMT ≥ 30 Days Post-vaccination | ≥2-fold risea N (%) | ≥4-fold risea N (%) | |
|---|---|---|---|---|
| Overall (N = 129) | ||||
| HAI A/TX | 2.70 | 3.29 | 54 (41.9) | 19 (14.7) |
| HAI A/HK | 1.24 | 1.75 | 46 (35.7) | 20 (15.5) |
| NAI A/TX | 4.21 | 4.78 | 55 (42.6) | 22 (17.1) |
| HAI B Yam | 4.90 | 5.40 | 58 (45.0) | 16 (12.4) |
| HAI B Vic | 4.73 | 5.19 | 48 (37.2) | 15 (11.6) |
| Vaccinated 2013–2014 (N = 105) | ||||
| HAI A/TX | 2.75 | 3.20 | 42 (40.0) | 10 (9.5)b |
| HAI A/HK | 1.35 | 1.79 | 36 (34.3) | 14 (13.3) |
| NAI A/TX | 4.28 | 4.82 | 45 (42.9) | 16 (15.2) |
| HAI B Yam | 5.02 | 5.41 | 43 (41.0) | 10 (9.5) |
| HAI B Vic | 4.83 | 5.18 | 36 (34.3) | 8 (7.6)b |
| Unvaccinated 2013–2014 (N = 24) | ||||
| HAI A/TX | 2.46 | 3.71 | 12 (50.0) | 9 (37.5) |
| HAI A/HK | 0.75 | 1.58 | 10 (41.7) | 6 (25.0) |
| NAI A/TX | 3.92 | 4.63 | 10 (41.7) | 6 (25.0) |
| HAI B Yam | 4.38 | 5.33 | 15 (62.5) | 6 (25.0) |
| HAI B Vic | 4.29 | 5.25 | 12 (50.0) | 7 (29.2) |
Abbreviations: A/HK, A/Hong Kong /4801/2014 A(H3N2) circulating strain virus; A/TX, A/Texas/50/2012 A(H3N2) vaccine strain virus; B Vic, B/Brisbane/60/2008 Victoria lineage virus; B Yam, B/Massachusetts/2/2012 Yamagata lineage virus; GMT, geometric mean titer; HAI, hemagglutinin-inhibition assay; NAI, neuraminidase-inhibition assay.
aProportions with a ≥2-fold or ≥4-fold rise in HAI or NAI titer against the specified virus strain from pre-vaccination to ≥30 days post receipt of the 2014–2015 influenza vaccine.
b<0.01 Fisher’s exact comparing those vaccinated in 2013–2014 to those unvaccinated in 2013–2014.
Preseason susceptibility to influenza infection was assessed using HAI and NAI antibody titers measured in serum collected from 488 subjects (66% of subjects ≥13 years), including 44 A(H3N2) cases (64% of 69 cases among subjects ≥13 years), and 15 B(Yamagata) cases (83% of 18 cases among subjects >13 years). We did not detect a difference in preseason HAI or NAI GMTs against any of the target antigens between subjects with and without RT-PCR confirmed infections. However, all preseason GMTs were significantly higher among those who were vaccinated compared to those unvaccinated (P < .001 for all comparisons).
For A(H3N2), visual examination of plots of the proportion infected by titer indicated a trend of lower infection risk with increasing preseason HAI titer against A/Hong Kong (3C.2a), but not for HAI or NAI titers against A/Texas (3C.1) (Figure 1). However, a 2-fold increase in HAI titer against A/Hong Kong was not significantly associated with lower odds of A(H3N2) infection in logistic regression models including titers against all 3 targets (OR 0.84 [95% CI: 0.62, 1.14]); for HAI and NAI titers against A/Texas, ORs were near the null value of 1. For B(Yamagata), plots indicated a trend of lower infection risk with increasing HAI titer. In models including HAI titers against both B(Yamagata) and B(Victoria), a 2-fold increase in titer against B(Yamagata) was associated with 0.75-fold (95% CI: 0.56, 1.00) lower odds of infection. A similar 2-fold increase in titer against B(Victoria) virus did not correlate with B(Yamagata) infection (OR 1.10 [95% CI: 0.82, 1.49]).
Figure 1.
Distribution of preseason titers and percent with RT-PCR confirmed influenza infection by titer. Hemagglutination-inhibition (HAI) antibody titers were measured against the influenza A(H3N2) vaccine (A/Texas/50/2012 [A/Texas]) and circulating (A/Hong Kong /4801/2014 [A/Hong Kong]) strains and against the B/Massachusetts/2/2012 (B/Massachusetts) virus that was like the B Yamagata lineage viruses which circulated and were included in the 2014–2015 vaccine. Neuraminidase-inhibition (NAI) antibody titers were measured against the influenza A (H3N2) vaccine (A/Texas) strain. Abbreviation: RT-PCR, reverse-transcription real-time polymerase chain reaction.
For all targets, preseason GMTs of subjects vaccinated both years were not significantly different from those vaccinated only in the current season (Supplemental Figure 1); patterns were similar when vaccination over 3 seasons was considered (Supplemental Figure 2).
DISCUSSION
We found no evidence of VE in preventing infection with antigenically drifted A(H3N2) viruses that predominated during the 2014–2015 season. This is consistent with other studies of mild to moderately severe infection carried out primarily in ambulatory care settings [17, 23, 24].
We also did not detect an association between HAI titer against the A/Texas vaccine strain and risk of A(H3N2) infection, which is not surprising in the context of antigenic mismatch with circulating viruses. Unexpectedly, NAI titers against A/Texas were also not associated with A(H3N2) infection risk. Although NA content is not standardized in currently licensed influenza vaccines, we have previously detected antibody response to NA following IIV receipt in a clinical trial [25]. We speculate that this may indicate that there could also have been a drift in the NA of the A(H3N2) viruses that circulated, which may have contributed to the lack of VE against A(H3N2). HA drift is monitored each year, but NA drift has not been routinely characterized [19]; however, the NA of 2014–2015 circulating viruses differed from that of A/Texas by 2 amino acid substitutions, H150R and E221D, previously implicated in antigenic drift [26].
Although many circulating A(H3N2) viruses did not agglutinate red blood cells [14], we were able to perform HAI assays using, as we have in the past, ether-split monovalent IIV subunit material (Sanofi-Pasteur) representing the target antigens [8, 12, 27]. Using ether split targets has the advantage of measuring antibody to the exact antigen included in the vaccine and may be superior for measuring antibody against influenza B [28]. There was some qualitative suggestion that increasing HAI titers against A/Hong Kong (H3N2 3C.2a) were associated with reduced risk of A(H3N2) infection; however, this was not statistically significant. Given differences in HA activity, and the possibility of other mutations introduced because of egg adaptation, it is possible that the antibody we measured against A/Hong Kong may not have been sufficiently specific to the circulating A(H3N2) 3C.2a viruses. When assay targets are well matched to circulating viruses, HAI titers can be highly correlated with infection risk [29]; therefore, identifying whether antibodies produced by vaccination are specific to circulating viruses is a priority for assessing vaccination programs.
We observed high VE against B(Yamagata) for children <9 years but no VE among older children and adults. This is similar to Canadian studies that only found significant VE against B(Yamagata) in the youngest age group [23] but in contrast to US studies where VE was observed in adults and children [30]. In this study, HAI titers against B(Yamagata) viruses were high among both vaccinated (GMT: 304) and unvaccinated (GMT: 97) individuals ≥13 years. Consistent with this observation, incidence of B(Yamagata) was low among older children and adults making it difficult to detect any effect of vaccination.
For both A(H3N2) and influenza B, we observed minimal antibody response following vaccination. For all targets examined, a lower proportion responded among those vaccinated in the previous season. Minimal response to vaccination was also observed in this cohort during the 2012–2013 and 2013–2014 seasons [8, 12]. Despite this, substantial VE was observed against influenza A(H1N1)pdm09 viruses in 2013–2014 [12]. These findings highlight the work that still needs to be done to further our understanding of biological and methodological issues surrounding the complex annual interactions between preexisting antibody, vaccine antigens, and circulating viruses.
Because antigenically equivalent A(H3N2) viruses had been included in vaccines for 3 consecutive years [13], the 2014–2015 season provided a unique opportunity to examine the interaction between prior vaccination and VE. In this context, Skowronski et al. observed that those vaccinated in all 3 seasons were actually at increased risk of A(H3N2) infection in 2014–2015 compared to those unvaccinated all 3 seasons [23]. In contrast, we found neither significant VE against A(H3N2) nor increased risk of infection for any combination of current and prior vaccine receipt, similar to US VE studies [30].
It had been suggested that LAIV may provide broader protection against antigenically drifted influenza [32]. In this study, neither LAIV nor IIV provided significant protection against drifted A(H3N2) virus infection among children aged 2–17 years. This is consistent with estimates from studies in the ambulatory care setting, which contributed to the decision by the CDC’s Advisory Committee for Immunization Practices to recommend LAIV not be used for the 2016–2017 season [30, 33]. In contrast, VE against B(Yamagata) among children 2–17 years was high for both LAIV and IIV but only statistically significant for LAIV.
Our findings, particularly those from subgroup analyses, should be interpreted in the context of relatively small sample size and limited power. Although we were careful to measure and control for confounding variables, our results could also be affected by other unmeasured factors. In addition, requiring study site illness visits could be expected to delay specimen collection resulting in false negative RT-PCR results. However, we do not expect this to have been a major issue as 82% of illness visits occurred ≤4 days from illness onset.
The 2014–2015 influenza season was unique in the scale and severity of the epidemic, antigenic mismatch between vaccine and circulating viruses, and an A(H3N2) vaccine component that was consistent over 3 consecutive years. VE against A(H3N2) has been low even in well-matched years, and issues regarding repeated vaccination have been primarily observed in A(H3N2) predominant seasons. Given these issues, more effective and broadly protective vaccines are needed, as well as further study into optimal strain selection strategies for currently available vaccines.
Supplementary Data
Supplementary materials are available at Clinical Infectious Diseases online. Consisting of data provided by the authors to benefit the reader, the posted materials are not copyedited and are the sole responsibility of the authors, so questions or comments should be addressed to the corresponding author.
Supplementary Material
Notes
Acknowledgements. We thank the HIVE Study staff for their hard work and dedication to the project: Barbara Aaron, Anne Kaniclides, and EJ McSpadden University of Michigan School of Public Health (UM-SPH). We thank Dr. Vasiliy Mishin Centers for Disease Control and Prevention (CDC) for performing pyrosequencing analysis on collected influenza A(H3N2) viruses. We are indebted to St Jude Childrens’ Research Hospital for providing plasmids that were used in to prepare reassortant viruses by reverse genetics. The reassortant viruses were generated and qualified by Jin Gao and Laura Couzens, Food and Drug Administration (CBER, FDA).
Disclaimer. The findings and conclusions in this report are those of the authors and do not necessarily represent the official position of the Centers for Disease Control and Prevention or funding agency.
Financial support. This work was supported by the CDC (U01 IP000474) and the National Institute of Allergy and Infectious Diseases (R01 AI097150).
Potential conflicts of interest. S. E. O. has received grant support from Sanofi Pasteur for work unrelated to this report. E. T. M. has received grant support from Merck for work unrelated to this report. A. S. M. has received grant support from Sanofi Pasteur and consultancy fees from Sanofi, GSK, and Novavax for work unrelated to this report. All other authors report no potential conflicts of interest. 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.
Footnotes
Presented in part: IDWeek 2016 meeting, New Orleans, Louisiana, October 26–30 2016. Presentation 799.
References
- 1. Ohmit SE, Thompson MG, Petrie JG et al. Influenza vaccine effectiveness in the 2011–2012 season: protection against each circulating virus and the effect of prior vaccination on estimates. Clin Infect Dis 2014; 58:319–27. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Skowronski DM, Janjua NZ, De Serres G et al. A sentinel platform to evaluate influenza vaccine effectiveness and new variant circulation, Canada 2010–2011 season. Clin Infect Dis 2012; 55:332–42. [DOI] [PubMed] [Google Scholar]
- 3. Valenciano M, Ciancio B; I-MOVE study team I-MOVE: a European network to measure the effectiveness of influenza vaccines. Euro Surveill 2012; 17:20281. [DOI] [PubMed] [Google Scholar]
- 4. Jackson ML, Nelson JC. The test-negative design for estimating influenza vaccine effectiveness. Vaccine 2013; 31:2165–8. [DOI] [PubMed] [Google Scholar]
- 5. Foppa IM, Haber M, Ferdinands JM, Shay DK. The case test-negative design for studies of the effectiveness of influenza vaccine. Vaccine 2013; 31:3104–9. [DOI] [PubMed] [Google Scholar]
- 6. Belongia EA, Simpson MD, King JP et al. Variable influenza vaccine effectiveness by subtype: a systematic review and meta-analysis of test-negative design studies. Lancet Infect Dis 2016; 16:942–51. [DOI] [PubMed] [Google Scholar]
- 7. Ohmit SE, Petrie JG, Malosh RE et al. Influenza vaccine effectiveness in the community and the household. Clin Infect Dis 2013; 56:1363–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Ohmit SE, Petrie JG, Malosh RE, Fry AM, Thompson MG, Monto AS. Influenza vaccine effectiveness in households with children during the 2012–2013 season: assessments of prior vaccination and serologic susceptibility. J Infect Dis 2015; 211:1519–28. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Thompson MG, Li DK, Shifflett P et al. Effectiveness of seasonal trivalent influenza vaccine for preventing influenza virus illness among pregnant women: a population-based case-control study during the 2010–2011 and 2011–2012 influenza seasons. Clin Infect Dis 2014; 58:449–57. [DOI] [PubMed] [Google Scholar]
- 10. Skowronski DM, Janjua NZ, De Serres G et al. Low 2012–13 influenza vaccine effectiveness associated with mutation in the egg-adapted H3N2 vaccine strain not antigenic drift in circulating viruses. PLoS One 2014; 9:e92153. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Smith DJ, Forrest S, Ackley DH, Perelson AS. Variable efficacy of repeated annual influenza vaccination. Proc Natl Acad Sci U S A 1999; 96:14001–6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Ohmit SE, Petrie JG, Malosh RE et al. Substantial influenza vaccine effectiveness in households with children during the 2013–2014 influenza season, when 2009 pandemic influenza a(H1N1) virus predominated. J Infect Dis 2016; 213:1229–36. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. World Health Organization (WHO). Recommended composition of influenza virus vaccines for use in the 2014–2015 northern hemisphere influenza season Available at: http://www.who.int/influenza/vaccines/virus/recommendations/201402_recommendation.pdf?ua=1. Accessed 16 November 2016.
- 14. Appiah GD, Blanton L, D’Mello T et al. ; Centers for Disease Control and Prevention (CDC). Influenza activity—United States, 2014–15 season and composition of the 2015–16 influenza vaccine. MMWR Morb Mortal Wkly Rep 2015; 64:583–90. [PMC free article] [PubMed] [Google Scholar]
- 15. D’Mello T, Brammer L, Blanton L et al. ; Centers for Disease Control and Prevention (CDC). Update: Influenza activity–United States, September 28, 2014–February 21, 2015. MMWR Morb Mortal Wkly Rep 2015; 64:206–12. [PMC free article] [PubMed] [Google Scholar]
- 16. Mishin VP, Baranovich T, Garten R et al. A pyrosequencing-based approach to high-throughput identification of influenza A(H3N2) virus clades harboring antigenic drift variants. J Clin Microbiol 2017; 55:145–54. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Flannery B, Zimmerman RK, Gubareva LV et al. Enhanced genetic characterization of influenza a(H3N2) viruses and vaccine effectiveness by genetic group, 2014–2015. J Infect Dis 2016; 214:1010–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Petrie JG, Ohmit SE, Cheng CK et al. Influenza vaccine effectiveness against antigenically drifted influenza higher than expected in hospitalized adults: 2014–2015. Clin Infect Dis 2016; 63:1017–25. [DOI] [PubMed] [Google Scholar]
- 19. World Health Organization (WHO). Recommended composition of influenza virus vaccines for use in the 2015–2016 northern hemisphere influenza season Available at: http://www.who.int/influenza/vaccines/virus/recommendations/201502_recommendation.pdf?ua=1. Accessed 16 November 2016.
- 20. Couzens L, Gao J, Westgeest K et al. An optimized enzyme-linked lectin assay to measure influenza A virus neuraminidase inhibition antibody titers in human sera. J Virol Methods 2014; 210:7–14. [DOI] [PubMed] [Google Scholar]
- 21. Grohskopf LA, Shay DK, Shimabukuro TT et al. Prevention and control of seasonal influenza with vaccines. Recommendations of the Advisory Committee on Immunization Practices—United States, 2013–2014. MMWR Recomm Rep 2013; 62:1–43. [PubMed] [Google Scholar]
- 22. Wei LJ, Lin DY, Weissfeld L. Regression analysis of multivariate incomplete failure time data by modeling marginal distribution. J Am Stat Assoc 1989; 84:1065–73. [Google Scholar]
- 23. Skowronski DM, Chambers C, Sabaiduc S et al. A perfect storm: impact of genomic variation and serial vaccination on low influenza vaccine effectiveness during the 2014–2015 season. Clin Infect Dis 2016; 63:21–32. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Valenciano M, Kissling E, Reuss A et al. Vaccine effectiveness in preventing laboratory-confirmed influenza in primary care patients in a season of co-circulation of influenza A(H1N1)pdm09, B and drifted A(H3N2), I-MOVE Multicentre Case-Control Study, Europe 2014/15. Euro Surveill 2016; 21:30139. [DOI] [PubMed] [Google Scholar]
- 25. Monto AS, Petrie JG, Cross RT et al. Antibody to influenza virus neuraminidase: An independent correlate of protection. J Infect Dis 2015; 212:1191–9. [DOI] [PubMed] [Google Scholar]
- 26. Air GM. Influenza neuraminidase. Influenza Other Respir Viruses 2012; 6:245–56. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27. Ohmit SE, Petrie JG, Cross RT, Johnson E, Monto AS. Influenza hemagglutination-inhibition antibody titer as a correlate of vaccine-induced protection. J Infect Dis 2011; 204:1879–85. [DOI] [PubMed] [Google Scholar]
- 28. Monto AS, Maassab HF. Ether treatment of type B influenza virus antigen for the hemagglutination inhibition test. J Clin Microbiol 1981; 13:54–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29. Petrie JG, Parkhouse K, Ohmit SE, Malosh RE, Monto AS, Hensley SE. Antibodies against the current influenza A(H1N1) vaccine strain do not protect some individuals from infection with contemporary circulating influenza a(H1N1) virus strains. J Infect Dis 2016; 214:1947–51. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30. Zimmerman RK, Nowalk MP, Chung J et al. 2014–2015 influenza vaccine effectiveness in the United States by vaccine type. Clin Infect Dis 2016; 63:1564–73. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31. Petrie JG, Ohmit SE, Johnson E, Truscon R, Monto AS. Persistence of antibodies to influenza hemagglutinin and neuraminidase following one or two years of influenza vaccination. J Infect Dis 2015; 212:1914–22. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32. Belshe RB, Edwards KM, Vesikari T et al. ; CAIV-T Comparative Efficacy Study Group. Live attenuated versus inactivated influenza vaccine in infants and young children. N Engl J Med 2007; 356:685–96. [DOI] [PubMed] [Google Scholar]
- 33. Grohskopf LA, Sokolow LZ, Broder KR et al. Prevention and control of seasonal influenza with vaccines: recommendations of the Advisory Committee on Immunization Practices, United States, 2016–17 Influenza Season. MMWR Recomm Rep 2016; 65:1–54. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.