Effectiveness of Influenza Vaccination During Pregnancy Against Laboratory-Confirmed Seasonal Influenza Among Infants Under 6 Months of Age in Ontario, Canada

Abstract

Routine influenza immunization of all pregnant individuals has been recommended in Canada since 2007 [1]; similar recommendations have been adopted by many countries since the 2009 H1N1 influenza pandemic [2]. Influenza immunization during pregnancy not only directly protects pregnant people, who are considered at high-risk for severe influenza, but also passively protects newborns through transplacental transfer of maternal anti-influenza antibodies to the fetus [3, 4]. As infants <6 months of age are also at high risk for severe influenza and associated complications [5–7], but are ineligible for influenza vaccination [8], immunization during pregnancy is increasingly viewed as an important strategy for protecting young infants [4].

Vaccine efficacy of influenza immunization during pregnancy against laboratory-confirmed influenza infection among infants <6 months of age pooled across 3 randomized controlled trials (RCTs) conducted in South Africa [9], Mali [10], and Nepal [11] was 35% (95% confidence interval [CI], 19%–47%) [12]. Although several observational studies of influenza vaccine effectiveness (VE) of maternal vaccination against laboratory-confirmed influenza infection among infants <6 months are available from high-resource settings [13–22], estimates vary considerably in magnitude (ranging from 31% to 71%), likely due to differences in influenza seasons and study designs. Similar heterogeneity exists for maternal influenza VE against laboratory-confirmed influenza hospitalization among infants (ranging from 37% to 92%) [13–18, 21, 23–27].

Despite the high-quality RCT evidence showing efficacy of influenza vaccination during pregnancy against influenza infection among infants [12], the trials were limited to 1–2 influenza seasons and conducted in low- and middle-income countries with important differences in influenza seasonality, underlying maternal-infant population characteristics, and health care systems compared with higher-resource temperate settings. In this study, we assessed the VE of maternal influenza vaccination during pregnancy against laboratory-confirmed influenza outcomes among infants <6 months of age in Ontario, Canada.

METHODS

Study Design, Setting, and Population

We conducted a test-negative design study among infants <6 months in Ontario—Canada's most populous province with a single-payer publicly funded health care system and approximately 140 000 births each year [28]. This study design is frequently used to evaluate influenza VE because of its ability to reduce biases associated with medical care-seeking behavior compared with other observational designs [29]. We identified all respiratory specimens from infants <6 months collected in any clinical setting (hospitals, emergency departments, physician offices), which were tested for influenza across 9 seasons (2010–2011 to 2018–2019). Only specimens collected during periods of active influenza circulation were included, defined as weeks above a threshold of 5% test positivity for Ontario (dates provided in Supplementary Table 1). For infants with more than 1 influenza test conducted in the same season, we included the earliest positive test (cases) or the earliest test if all results were negative (controls). We excluded infants born outside of Ontario, those not eligible for provincial health insurance on the specimen collection date, and those whose records could not be linked to their mother or whose mother was not continuously eligible for provincial health insurance throughout the entire pregnancy, because this was necessary for ascertainment of influenza vaccination status. Additional exclusions are shown in Figure 1.

Figure 1.

Study flow diagram. ^aUncertain probability of match between the delivery records of the mother and infant. Abbreviation: LMP, last menstrual period.

Open in new tabDownload slide

Data Sources

All data sources were linked using unique encoded identifiers and analyzed at ICES, an independent, nonprofit research institute whose legal status under Ontario's health information privacy law allows it to collect and analyze health care and demographic data, without consent, for health system evaluation and improvement (https://www.ices.on.ca).

Infant laboratory test results were obtained from a network of public health and hospital-based laboratories across Ontario and included single and multiplex reverse-transcriptase polymerase chain reaction (RT-PCR), viral culture, direct fluorescence assays, and rapid antigen tests. For the 2010–2011 to 2015–2016 seasons, we used previously linked data from the Flu and Other Respiratory Viruses Research (FOREVER) cohort [30], and for 2016–2017 to 2018–2019, we used the Ontario Laboratories Information System and the Public Health Ontario Labware data. We linked these infant test results with the MOMBABY database, which contains integrated mother-infant hospitalization records derived from the birth admission, to identify mother-infant dyads and obtain information about pregnancy and birth characteristics. The Canadian Institute for Health Information's Discharge Abstract Database (hospitalizations), National Ambulatory Care Reporting System (emergency department visits), Same Day Surgery database (same-day surgeries), Ontario Health Insurance Plan Database (physician office visits), and Registered Persons Database (demographics) were used to determine the clinical setting where the infant's respiratory test specimen was collected and to obtain other clinical and sociodemographic information. Several ICES-derived registries were used to identify mothers with preexisting medical complications such as asthma, hypertension, and diabetes. Finally, we used Statistics Canada's Postal Code Conversion File to determine the mother's dissemination area of residence, based on the postal code, to obtain data on rural/urban residence and area-based income quintiles. Details on all data sources are provided in Supplementary Table 2.

Ethics Approval

ICES is a prescribed entity under Ontario's Personal Health Information Protection Act (PHIPA). Section 45 of PHIPA authorizes ICES to collect personal health information, without consent, for the purpose of analysis or compiling statistical information with respect to the management of, evaluation or monitoring of, the allocation of resources to, or planning for all or part of the health system. Projects that use data collected by ICES under section 45 of PHIPA, and use no other data, are exempt from Research Ethics Board review. The use of the data in this project is authorized under section 45 and approved by ICES’ Privacy and Legal Office.

Measures

Infant Influenza Outcomes

Respiratory specimen collection dates from infant laboratory records were combined with dates for health services received by the infants to determine the clinical setting where the specimen was collected. If a specimen collection date was associated with more than 1 clinical setting, we used a hierarchy (intensive care unit [ICU] > hospital ward > emergency department > same-day surgery > physician's office). Influenza tests were attributed to hospital/ICU admissions if the date of specimen collection fell within 3 days prior to the admission date through to the discharge date.

For the primary outcome, infants with a respiratory specimen collected in any clinical setting that yielded influenza were considered cases. For the secondary outcome—influenza hospitalization—case specimens were limited to those attributed to a hospital or ICU admission. The control group for both outcomes comprised infants who tested negative for influenza in any clinical setting.

Maternal Influenza Vaccination During Pregnancy

We considered infants born to mothers who had received an influenza vaccine between the estimated date of the last menstrual period up to 14 days before delivery to be exposed. Infants born to mothers who had received the previous season's influenza vaccine or who were vaccinated <14 days before delivery were considered unexposed in the primary analysis. The date of vaccination was used in combination with the infant's date of birth and gestational age to determine the trimester of vaccination. We used billing claims in the Ontario Health Insurance Plan and Ontario Drug Benefit Databases to ascertain seasonal influenza vaccination in primary care and pharmacies, respectively (Supplementary Table 3).

Other Variables

We obtained the following sociodemographic and clinical characteristics for mothers and infants: maternal age at delivery, neighborhood income, rural residence, maternal preexisting medical comorbidities, seasonal timing of conception and birth, multiple gestation, parity, mode of delivery, maternal obstetrical complications, adequacy of prenatal care, nonobstetric health service utilization 1–2 years prior to the index pregnancy, aggregated diagnosis groups (ADGs) calculated using the Johns Hopkins ACG system version 10.0 in a 2-year period prior to pregnancy, gestational age at birth, birth weight, small-for-gestational age birth, infant sex, and infant age at specimen collection. Supplementary Table 3 provides additional information on all measures, including definitions and specific codes.

Statistical Analysis

We described baseline characteristics by outcome and exposure using descriptive statistics, including standardized mean differences to compare the distributions (absolute differences >0.10 were considered meaningful) [31]. Multivariable logistic regression was used to compare the odds of maternal influenza vaccination during pregnancy among test-positive infant cases to test-negative infant controls, generating adjusted odds ratios (aOR) and 95% CI. Models were adjusted for infant age at test, seasonal timing of birth, prenatal care adequacy, neighborhood income, and influenza season. VE was calculated as (1 − aOR)×100%. For the primary outcome, we conducted subgroup analyses by trimester of vaccination (first/second trimester combined, third trimester), infant age at test (0 to <2 months, 2 to <6 months), gestational age at birth (term birth ≥37 weeks, preterm birth <37 weeks), and by vaccine matched/mismatched seasons (Supplementary Table 4). We conducted several sensitivity analyses to assess the robustness of our main results. First, we reclassified exposure status to “vaccinated” for mothers who had received the previous season's influenza vaccine during pregnancy (as sometimes occurred if an older infant was tested for influenza at the beginning of an influenza season) and for mothers vaccinated <14 days before delivery. We adjusted for several other potential confounders (gestational age at birth, rural/urban residence, maternal preexisting medical comorbidity) to assess whether there was residual confounding by these variables, and we limited the analysis to infants whose respiratory specimen was collected during an emergency department visit or hospitalization with an International Classification of Diseases-Tenth Revision (ICD-10) code indicating acute respiratory illness. For the secondary outcome, we conducted a sensitivity analysis limiting the control group to infants whose specimen was attributed to a hospital or ICU setting. Analyses were conducted using SAS version 9.4 software (SAS Institute, Inc).

RESULTS

Study Population

Across 9 influenza seasons, 23 860 infants aged <6 months who were tested for influenza virus during the influenza season met study eligibility criteria (Figure 1). Of these, 1783 infants (7.5%) had a respiratory specimen that yielded influenza (Table 1). Infants with influenza were older than test-negative controls (120 to <180 days at testing, 25.8% vs 20.7%, respectively; median age in days at testing, 74 [IQR, 45–123] vs 63 [IQR, 36–109]), more likely to have been conceived in the winter (50.6% vs 38.8%) and born in the fall (56.6% vs 44.5%), and more likely to have had their specimen collected in an emergency department (40.7% vs 34.0%) (Table 1 and Figure 2). Maternal characteristics were similar between test-positive cases and test-negative controls.

Figure 2.

Bar graphs displaying the estimated date of conception (A), specimen collection date (B), and date of birth (C) of influenza positive cases and influenza negative controls, by month.

Open in new tabDownload slide

Overall, 1708 (7.2%) infants were born to mothers who were vaccinated against influenza during pregnancy with the current season's vaccine at <14 days before delivery. Infants born to vaccinated mothers were more likely to be conceived in the spring or summer compared to those born to unvaccinated mothers (84.7% vs 46.7%, respectively), and to be younger at the time of specimen collection (<60 days at testing, 73.5% vs 44.2%; median age in days at testing, 40 [IQR, 22–61] vs 67 [IQR, 38–114]). They were also more likely to be born to older mothers (≥35 years, 27.9% vs 21.7%), mothers living in higher-income neighborhoods (fifth quintile, 20.4% vs 15.0%), and mothers who had more frequent prenatal care visits (intensive/adequate prenatal care, 60.0% vs 49.2%) (Table 1).

Vaccine Effectiveness

Across seasons, estimated effectiveness of influenza vaccination during pregnancy was 64% (95% CI, 50%–74%) against laboratory-confirmed influenza infection among infants (Figure 3 and Supplementary Table 5). VE was similar among infants <2 months of age (63%; 95% CI, 46%–75%) and those 2 to <6 months (64%; 95% CI, 36%–79%); we were unable to separately estimate VE for infants 4 to <6 months due to small numbers. VE was also similar among infants born at term (64%; 95% CI, 50%–75%) and preterm (61%; 95% CI, 4%–86%) gestation, as well as according to vaccination during the first/second trimester (66%; 95% CI, 40%–80%) or third trimester (63%; 95% CI, 46%–74%) of pregnancy. During 7 seasons in which influenza vaccines were well matched to dominant circulating strains, VE was 71% (95% CI, 57%–80%); conversely, in 2 poorly matched seasons (2014–2015 and 2017–2018), it was 37% (95% CI, −9% to 64%). Twenty-four test-positive cases were hospitalized; VE against laboratory-confirmed influenza hospitalization was 67% (95% CI, 50%–78%).

Figure 3.

Estimated effectiveness of maternal influenza vaccination during pregnancy against laboratory-confirmed influenza among infants <6 months of age (additional data provided in Supplementary Table 5). ^aVaccine effectiveness was estimated as 1 minus the adjusted odd ratio, comparing the odds of maternal influenza vaccination during pregnancy among test-positive infant cases relative to test-negative infant controls. Odds ratios were adjusted for infant age at test, season of birth, prenatal care adequacy, neighborhood income, and influenza season (with the exception of the subgroup analysis by infant age at test, for which odds ratios were adjusted for season of birth, prenatal care adequacy, neighborhood income, and influenza season). ^bResults for first and second trimester vaccinations were combined due to small cell counts. ^cResults for 60 to <120 days and 120 to <180 days were combined due to small cell counts. ^dInfluenza vaccines in 2014–2015 and 2017–2018 were not considered well matched against circulating influenza strains (see Supplementary Table 4). ^eCase specimens were limited to those collected during an admission to a hospital ward or intensive care unit. Controls were infants who tested negative for influenza in any clinical setting. Abbreviation: CI, confidence interval.

Open in new tabDownload slide

Results from sensitivity analyses are provided in Table 2. Estimated VE was somewhat lower than the primary analysis when mothers who received the previous season's vaccine (n = 245) or were vaccinated <14 days before delivery (n = 251) were reclassified as vaccinated. Additional adjustments for infant gestational age at birth, rural/urban residence, and maternal preexisting medical comorbidity did not change the VE estimates. When the study population was limited to records of infants whose specimen was attributed to a hospitalization or emergency department visit coded as an acute respiratory illness, estimated VE was 71% (95% CI, 58%–80%). For the secondary outcome, when the test-negative control group was limited to infants whose specimen was attributed to a hospital or ICU setting, estimated VE did not change.

DISCUSSION

In this test-negative study of 23 860 infants across 9 influenza seasons in Ontario, Canada, we found that influenza vaccination during pregnancy was effective at preventing laboratory-confirmed influenza infection and influenza hospitalization among infants <6 months of age. Taken together with RCT evidence that influenza vaccination during pregnancy reduces the incidence of all-cause severe pneumonia [32] and hospitalization for all-cause lower respiratory tract infection in infants [33], our study supports influenza immunization during pregnancy as an effective strategy to protect young infants from influenza-related morbidity during their first 6 months of life.

Our findings are generally consistent with other observational studies from high-resource settings [13–22]. Nevertheless, estimates across the other studies vary considerably in magnitude—ranging from 31% [20] to 71% [15]—likely due to differences in influenza seasons (ie, match between circulating strains and vaccine strains), testing patterns, study designs (ie, prospective cohort study [13, 19, 20, 22], retrospective cohort study [14, 18, 21], test-negative case-control study [17], and screening study [15, 16]), or possibly due to underlying differences in study setting (United States [13, 14], United Kingdom [15, 16], Greece [19, 20], Denmark [17], Australia [18, 21], and Japan [22]). The most similar study to ours, with respect to study design and time period, was a population-based test-negative study conducted in Denmark from 2010 to 2017, in which the estimated effectiveness of influenza vaccination during pregnancy was 57% (95% CI, 25%–75%) against laboratory-confirmed influenza infection among infants <6 months [17]. Variability in vaccine efficacy against infant influenza infection was also observed across the RCTs conducted in South Africa (49%; 95% CI, 12%–70%) [9], Mali (33%; 95% CI, 4%–54%) [10], and Nepal (30%; 95% CI, 5%–48%) [11], as well as an earlier RCT from Bangladesh (63%; 95% CI, 5%–85%) [34].

To our knowledge, effectiveness of vaccination during pregnancy against influenza infection among preterm infants has not been previously reported. We found similar VE among infants born at preterm and term gestation (61% and 64%, respectively). Although the confidence interval for estimated VE among preterm infants was wide, it is compelling that influenza vaccination in pregnancy has the potential to protect preterm infants from influenza, as has been demonstrated for earlier Tdap vaccination during pregnancy against pertussis infection among preterm infants [35]. We did not observe a qualitative difference in VE according to trimester of maternal influenza vaccination. Transplacental transfer of maternal anti-influenza antibodies is known to start in the early second trimester and increase with advancing gestation [36, 37]. Despite that later gestational timing of influenza vaccination increases the transfer of maternal antibodies to the fetus [37], few studies have evaluated the clinical impact of gestational timing of vaccination on influenza VE among young infants and, those studies that have, were limited by small cell sizes precluding conclusive interpretation of their findings [21, 23, 25]. Unlike the pooled RCT results that demonstrated higher efficacy of vaccination during pregnancy at younger infant ages (0 to <2 months, 56% [95% CI, 28%–73%]; 2 to <4 months, 39% [95% CI, 11%–58%]; 4 to <6 months, 19% [95% CI, −9% to 40%] [12]), we did not observe a difference in estimated VE among infants 0 to <2 months of age and those 2 to <6 months (63% and 64%, respectively), which was unexpected because the concentration of maternal anti-influenza antibodies in infant sera decays over time after birth [38, 39]. This finding may be due to the fact that very few infants 4 to <6 months of age were born to vaccinated mothers, therefore, the estimate for the 2 to <6 month age group primarily reflects VE for infants 2 to <4 months of age, which is a time period during which infants would still be expected to be well protected by maternal immunization [40].

Extant evidence on maternal influenza VE against laboratory-confirmed influenza hospitalization of infants <6 months of age is derived from observational studies [13–18, 21, 23–27], as the incidence was too low to be assessed in the RCTs [24]. Similar to VE against infection, estimates against laboratory-confirmed influenza hospitalization also range considerably in magnitude, from 29% [24] to 92% [13]. Three of the previously published studies used a test-negative design and were conducted in similar years to those in our study, originating from Denmark [17], Australia [23], and South Africa [24]. While the study from Denmark used similar population-based sources of routinely collected laboratory and maternal-newborn data to the sources we used in Ontario and obtained a similar estimate (61%; 95% CI, 26%–80% [17]), the other 2 were prospectively conducted and used clinical criteria for enrollment. The Australian study, conducted across 12 large pediatric hospitals, included infants <6 months admitted with signs of acute respiratory illness and tested for influenza. Estimated effectiveness of influenza vaccination during pregnancy against influenza hospitalization among infants <6 months was 37% (95% CI, 2%–60%) across the 4 influenza seasons, but higher among infants 0 to <2 months (49%; 95% CI, 11%–71%) and in 2018 (98%; 95% CI, 57%–100%), when the vaccine formulation for the Southern hemisphere was well matched to circulating strains [23]. The South African study was conducted in 4 hospital sites from 2015 to 2018. Infants admitted to hospital with an acute respiratory illness were prospectively identified and tested for influenza. Overall VE was 29% (95% CI, −33% to 32%); however, when limited to only vaccine-matched strains, it was 65% (95% CI, 12%–86%) [24]. While we expected higher VE against influenza hospitalization than against any influenza infection in our study, the estimates were similar in magnitude, and had overlapping confidence intervals. However, it is important to note that 95% of testing was performed in the emergency department or after hospital admission and more than 50% occurred in infants who were ultimately admitted, thus, our results reflect VE against relatively more severe infections, and actual VE against any influenza infection may be lower.

Strengths of this study include the province-wide linked data sources, enabling us to define a large maternal-infant study population with ascertainment of laboratory-confirmed influenza infections among infants and influenza vaccination during pregnancy. Our study provides estimates of the effectiveness of influenza vaccination during pregnancy against influenza infection and hospitalization of infants <6 months from a Canadian setting; previously, no such estimates had been available for this population and our results are likely generalizable within Canada and possibly to other high-income settings with similar influenza seasonality. Our study also has limitations. First, secondary use of routinely collected laboratory data (as opposed to prospectively collected data with clinical criteria for testing) for test-negative studies can introduce bias if the clinical decision to test is associated with vaccination status and influenza positivity [29]. However, if maternal influenza vaccination status during pregnancy influenced a care provider's decision to test an infant for influenza, a simulation study demonstrated that the probability of being tested would have to be at least twice as high in vaccinated compared with unvaccinated subjects to meaningfully bias VE estimates [41]. There was, however, a higher proportion of test-positive case infants who were hospitalized and/or had a diagnosis code for acute respiratory illness, compared with the test-negative controls, which could suggest variability in testing protocols (eg, some test-negative infants may have been tested for other purposes, such as hospitalization for an unrelated illness or surgery). Next, we used billing claims from primary care and pharmacies to ascertain influenza vaccination during pregnancy and obtained an overall vaccination coverage rate of 7.2%. An Ontario validation study from 2007 to 2009 estimated the sensitivity of primary care claims for influenza vaccination to be 42% among pregnant women [42]. Although we do not know the sensitivity of the vaccination variable derived for our study, it is likely to be somewhat improved due to the addition of claims from vaccinations received in pharmacies. There are limited published coverage data in the Ontario pregnant population against which to compare the vaccination rates from our study―in a survey conducted by the Public Health Agency of Canada, influenza vaccine coverage during pregnancy in Ontario in the 2018–2019 influenza season was estimated to be 43.3% in a sample of 554 pregnant individuals [43]. However, based on other work we have conducted on Tdap vaccination during pregnancy in Ontario, it is possible that the survey methodology and small respondent sample overestimated coverage (ie, 40% Tdap coverage in 2018–2019 estimated by the survey [43] vs 24.9% Tdap coverage estimated in our previous work using Ontario's province-wide health administrative databases [44]). Nevertheless, it is possible that nondifferential misclassification of maternal influenza vaccination status biased our estimates toward the null. Moreover, the low proportion of vaccinated individuals limited the study's power to evaluate VE stratified by smaller gestational age categories or by influenza season. While we had information on many potential confounding variables, we cannot rule out residual confounding by other factors such as maternal smoking or breastfeeding. Finally, we were unable to generate estimates by influenza type/subtype, as this information was not consistently available in the data sources used in this study.

CONCLUSIONS

In this population-based study across 9 recent influenza seasons, maternal influenza vaccination during pregnancy protected infants <6 months from both laboratory-confirmed influenza infection and influenza hospitalization. Given that infants in this age group have a high risk for severe influenza illness and complications, but cannot be vaccinated against influenza themselves, maternal influenza immunization during pregnancy is an important strategy for protecting young infants during their first 6 months of life.

Supplementary Data

Supplementary materials are available at The Journal of Infectious Diseases online (http://jid.oxfordjournals.org/). Supplementary materials consist of 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.

Notes

Acknowledgment. This article used data adapted from the Statistics Canada Postal Code Conversion File, which is based on data licensed from the Canada Post Corporation, and/or data adapted from the Ontario Ministry of Health (MOH) Postal Code Conversion File, which contains data copied under license from Canada Post Corporation and Statistics Canada. Parts of this material are based on data and/or information provided by the Ontario MOH and the Canadian Institute for Health Information. We thank IQVIA Solutions Canada, Inc for use of their Drug Information File.

Author contributions. D. B. F., M. L. R., and J. C. K. conceptualized the study, obtained the grant funding, oversaw the data analysis, interpreted the findings, and drafted the manuscript. H. C. and S. S. conducted the data analysis, and contributed to the study methodology, interpretation, and final manuscript. S. F. assisted with drafting the manuscript. All authors contributed to the interpretation of findings and critical review and revision of the final manuscript, approve the final manuscript as submitted, and agree to be accountable for all aspects of the work.

Data availability. The datasets from this study are held securely in coded form at ICES. While legal data sharing agreements between ICES and data providers (eg, health care organizations and government) prohibit ICES from making the dataset publicly available, access may be granted to those who meet prespecified criteria for confidential access, available at www.ices.on.ca/DAS (email, [email protected]). The full dataset creation plan and underlying analytic code are available from the authors on request, understanding that the computer programs may rely on coding templates or macros that are unique to ICES and are therefore either inaccessible or may require modification.

Disclaimer. The analyses, conclusions, opinions, and statements expressed herein are solely those of the authors and do not reflect those of the funding or data sources; no endorsement is intended or should be inferred.

Financial support. This work was supported by the Canadian Immunization Research Network through a grant from the Public Health Agency of Canada and the Canadian Institutes of Health Research (CNF 151944); and ICES, which is funded by an annual grant from the Ontario Ministry of Health and the Ministry of Long-Term Care. J. C. K. is supported by a Clinician Scientist Award from the University of Toronto Department of Family and Community Medicine. Funding to pay the Open Access publication charges for this article was provided by the Canadian Immunization Research Network.

References

Author notes