Influenza vaccination during pregnancy and risk of selected major structural congenital heart defects, National Birth Defects Prevention Study 2006–2011
Funding information: Centers for Disease Control and Prevention, Grant/Award Numbers: PA 96043, PA 02081, FOA DD09-001, FOA DD13-003, NOFO DD18-001; Iowa Center for Birth Defects Research and Prevention, Grant/Award Number: U01DD001223
Abstract
Background
Although results from studies of first-trimester influenza vaccination and congenital heart defects (CHDs) have been reassuring, data are limited for specific CHDs.
Methods
We assessed associations between reported maternal influenza vaccination, 1 month before pregnancy (B1) through end of third pregnancy month (P3), and specific CHDs using data from a multisite, population-based case–control study. Analysis included 2,982 case children diagnosed with a simple CHD (no other cardiac involvement with or without extracardiac defects) and 4,937 control children without a birth defect with estimated delivery dates during 2006–2011. For defects with ≥5 exposed case children, we used logistic regression to estimate propensity score-adjusted odds ratios (aORs) and 95% confidence intervals (CIs), adjusting for estimated delivery year and season; plurality; and maternal age at delivery, race/ethnicity, low folate intake, and smoking and alcohol use during B1P3.
Results
Overall, 124 (4.2%) simple CHD case mothers and 197 (4.0%) control mothers reported influenza vaccination from 1 month before through the third pregnancy month. The aOR for any simple CHD was 0.97 (95% CI: 0.76–1.23). Adjusted ORs for specific simple CHDs ranged from 0.62 for hypoplastic left heart syndrome to 2.34 for total anomalous pulmonary venous return (TAPVR). All adjusted CIs included the null except for TAPVR.
Conclusions
Although we cannot fully exclude that exposure misclassification may have masked risks for some CHDs, findings add to existing evidence supporting the safety of inactivated influenza vaccination during pregnancy. The TAPVR result may be due to chance, but it may help inform future studies.
1 INTRODUCTION
Inactivated influenza vaccine has been recommended for pregnant women in any trimester of pregnancy by the Advisory Committee on Immunization Practices (ACIP) and the American College of Obstetricians and Gynecologists since 2004 (ACOG Committee on Obstetric Practice, 2004, 2018; Grohskopf et al., 2021; Harper, Fukuda, Uyeki, et al., 2004). Although results from studies of first-trimester influenza vaccination and congenital heart defects (CHDs) overall have been reassuring (Kharbanda et al., 2017; Louik et al., 2013; Louik, Kerr, Van Bennekom, et al., 2016; Ludvigsson et al., 2016), there are limited data on maternal influenza vaccination and specific CHDs (Louik et al., 2013; Louik et al., 2016).
We previously evaluated associations between early pregnancy influenza vaccination and a spectrum of major structural noncardiac birth defects or groups of defects in the National Birth Defects Prevention Study (NBDPS) and observed adjusted odds ratios (aORs) ranging from 0.53 to 1.74 with corresponding 95% confidence intervals (CIs) that included the null (Palmsten et al., 2022). The objective of our current analysis was to assess the association between early pregnancy influenza vaccination and specific CHDs, including those not previously examined in the literature.
2 METHODS
2.1 NBDPS
The NBDPS was a multisite, population-based case–control study of more than 30 major structural birth defects of unknown etiology that enrolled children from pregnancies with an estimated delivery date (EDD) from October 1997 through December 10, 2011. Control children were live births without major defects randomly selected from birth certificates or hospital records from the same time period and study region as case children. Case children were live births, stillbirths, or from pregnancies electively terminated after prenatal diagnosis of a birth defect. Inclusion criteria for case children have been described (Botto, Lin, Riehle-Colarusso, Malik, & Correa, 2007; Reefhuis et al., 2015). Briefly, clinical geneticists reviewed information abstracted from medical records to confirm eligibility. Case children were assigned a classification based on the presence of extracardiac defects: isolated (no other additional major birth defect[s] in a different organ system); multiple (one or more additional, major unrelated birth defects in a different organ system); or complex (a pattern of major defects that are embryologically related) (Rasmussen et al., 2003; Reefhuis et al., 2015). CHDs were classified by cardiac complexity. Anatomically distinct or well-recognized single entities (e.g., tetralogy of Fallot) were classified as simple CHD; common, uncomplicated combinations of CHDs (e.g., ventricular septal defect and atrial septal defect) were classified as associations; and those that could not be classified as simple or association (e.g., heterotaxy) were classified as complex CHD (Botto et al., 2007).
Interviewers conducted computer-assisted telephone interviews with mothers of case and control children any time between 6 weeks and 24 months after the EDDs to collect information on demographics, pregnancy history, health conditions, and pregnancy exposures (Reefhuis et al., 2015). Each site obtained institutional review board approval for the NBDPS, and participants provided informed consent.
2.2 Exposure
Influenza vaccination status was classified using responses to the question, “During this time period [3 months before pregnancy to the end of pregnancy], did you take any medications, remedies, or treatments that we have not already talked about? For example, flu or allergy shots or medications for asthma, allergies, infections, STDs or HIV/AIDS? What drug?/Any others?” and reported dates of exposure. We defined exposure as report of influenza vaccination from 1 month before pregnancy through end of the third pregnancy month (B1P3). Mothers who did not report influenza vaccination during B1P3 were classified as unexposed. Because of potential error in estimating the date of conception and to allow for exposure effects to carry into early pregnancy, we included the month before pregnancy in the exposure window.
2.3 Inclusion and exclusion criteria
We restricted our analytic sample to mothers who had an EDD on or after January 1, 2006 and responded to the revised computer-assisted telephone interview generally administered to mothers with an EDD in 2006 or later, because it included the question mentioned above that referred to vaccinations. We excluded case children assigned a complex child classification to reduce heterogeneity among case children, as etiologies of complex phenotypes likely differ from those of isolated and multiple phenotypes (n = 119; Figure S1) (Reefhuis et al., 2015). We also excluded mothers (n = 37) missing the timing of their influenza vaccination and mothers (n = 175) who reported type 1 or type 2 pregestational diabetes diagnosed before the index pregnancy, because it is strongly associated with CHDs (Tinker et al., 2020).
2.4 Outcomes
To facilitate comparison of our findings for CHDs with those from the previous literature, we assessed any major structural CHD overall (n = 4,145 case children). To improve homogeneity of CHD phenotypes, we subsequently excluded case children with complex CHDs (n = 196) and CHDs that were part of an association (n = 967). Our analysis of specific CHDs focused on simple major structural CHDs (n = 2,982 case children) with (multiple) or without (isolated) extracardiac defects. We assessed any simple CHD overall and specific simple CHDs if there were at least five exposed case children.
2.5 Covariates
We identified covariates from existing knowledge and literature on CHDs: EDD year (<2010, ≥2010) as influenza vaccination coverage during pregnancy increased after the H1N1 influenza pandemic;(Rasmussen & Jamieson, 2019) EDD season (January–March, April–June, July–September, October–December) to account for seasonality of the exposure; plurality (singleton or multifetal pregnancies); maternal age (<20 years, 20–34 years, ≥35 years), race/ethnicity (non-Hispanic White, non-Hispanic Black, Hispanic, other), and education (≤11, 12, ≥13 years) at delivery; gravidity (0, 1, ≥2); pre-pregnancy body mass index (<18.5 kg/m2, 18.5–24 kg/m2, 25–29 kg/m2, ≥30 kg/m2); cigarette smoking (no active and no passive, active only, passive only, active and passive), any alcohol use, any fever, and any respiratory illness during B1P3; low folate intake (no folic acid supplementation during B1P3 and <600 μg/day dietary folate equivalents in the year before pregnancy; U.S. Food and Drug Administration, 2022); any hypertension during pregnancy, and any antihypertensive medication use during B1P3.
2.6 Statistical analysis
We assessed associations using logistic regression to estimate ORs and 95% CIs adjusting for covariates using propensity scores (Glynn, Schneeweiss, & Sturmer, 2006). To calculate propensity scores, we fit a model for the probability of exposure conditional on the covariates in control mothers only and used the parameter estimates to calculate the predicted probability of exposure in case and control mothers given their covariate values. Covariate distributions among control mothers are expected to approach that of the underlying population given that the outcomes are rare (Mansson, Joffe, Sun, & Hennessy, 2007). In the primary model, the propensity score included a limited set of covariates that we believed to be the most important for confounding (EDD year and season, plurality, maternal age at delivery, race/ethnicity, cigarette smoking, alcohol use, low folate intake). We adjusted exposure-outcome models for quintiles of the propensity score and conducted a complete case analysis. We additionally included the remaining covariates of interest in an expanded model as a sensitivity analysis. We conducted additional sensitivity analyses using the limited set of covariates that: (a) included case children with isolated defects only, as etiology may differ from those with multiple defects, (b) included singletons only, as twinning is associated with birth defects (Dawson et al., 2016), (c) restricted to children without family history of a first-degree relative with a CHD, (d) restricted to children with mothers who did not report influenza vaccination between 2 and 3 months before conception to remove those less likely to be vaccinated during pregnancy, regardless of actual vaccination status during B1P3, and (e) restricted to children with mothers whose first trimester overlapped with typical influenza vaccination months (September through March) to further account for potential seasonal effects.
3 RESULTS
Overall, 124 (4.2%) mothers of children with simple CHDs and 197 (4.0%) mothers of controls reported influenza vaccination during B1P3. None reported receipt of nasal mist formulations of the vaccine (i.e., live attenuated influenza vaccine), which are contraindicated during pregnancy (Grohskopf et al., 2021). Mothers of controls who reported influenza vaccination more often delivered after 2009 with an EDD in April through September, had at least 13 years of education, were 35 years of age or older, were non-Hispanic White, more often reported alcohol use, and less often reported smoking compared with those who did not report influenza vaccination (Table 1).
| Characteristic | Influenza vaccination during B1P3 (n = 197) | No influenza vaccination during B1P3 (n = 4,740) | ||
|---|---|---|---|---|
| n | % | n | % | |
| Estimated date of delivery year | ||||
| 2006–2009 | 89 | 45.2 | 3,250 | 68.6 |
| 2010–2011 | 108 | 54.8 | 1,490 | 31.4 |
| Missing | 0 | 0 | ||
| Estimated delivery date season | ||||
| Winter (January–March) | 17 | 8.6 | 1,183 | 25.0 |
| Spring (April–June) | 98 | 49.7 | 1,096 | 23.1 |
| Summer (July–September) | 63 | 32.0 | 1,267 | 26.7 |
| Fall (October–December) | 19 | 9.6 | 1,194 | 25.2 |
| Missing | 0 | 0 | ||
| Plurality | ||||
| Multifetal pregnancies | 5 | 2.5 | 141 | 3.0 |
| Singletons | 192 | 97.5 | 4,598 | 97.0 |
| Missing | 0 | 1 | ||
| Maternal age at delivery | ||||
| <20 years | 6 | 3.0 | 427 | 9.0 |
| 20–34 years | 153 | 77.7 | 3,646 | 76.9 |
| ≥35 years | 38 | 19.3 | 667 | 14.1 |
| Missing | 0 | 0 | ||
| Maternal race/ethnicity | ||||
| Non-Hispanic White | 145 | 73.6 | 2,588 | 54.7 |
| Non-Hispanic Black | 13 | 6.6 | 489 | 10.3 |
| Hispanic | 21 | 10.7 | 1,328 | 28.1 |
| Other | 18 | 9.1 | 329 | 6.9 |
| Missing | 0 | 6 | ||
| Education | ||||
| ≤11 years | 10 | 5.2 | 744 | 16.5 |
| 12 years | 23 | 11.9 | 1,031 | 22.8 |
| ≥13 years | 161 | 83.0 | 2,745 | 60.7 |
| Missing | 3 | 220 | ||
| Gravidity | ||||
| 0 | 57 | 28.9 | 1,421 | 30.2 |
| 1 | 59 | 29.9 | 1,272 | 27.0 |
| ≥2 | 81 | 41.1 | 2,020 | 42.9 |
| Missing | 0 | 27 | ||
| Pre-pregnancy body mass index | ||||
| Underweight <18.5 kg/m2 | 9 | 4.6 | 226 | 5.0 |
| Normal weight 18.5–24 kg/m2 | 106 | 54.6 | 2,272 | 50.6 |
| Overweight 25–29 kg/m2 | 34 | 17.5 | 1,046 | 23.3 |
| Obese ≥30 kg/m2 | 45 | 23.2 | 948 | 21.1 |
| Missing | 3 | 248 | ||
| Cigarette smoking B1P3 | ||||
| No active and no passive smoking | 154 | 79.4 | 3,306 | 73.2 |
| Active smoking only | 11 | 5.7 | 354 | 7.8 |
| Passive smoking only | 12 | 6.2 | 455 | 10.1 |
| Active and passive smoking | 17 | 8.8 | 400 | 8.9 |
| Missing | 3 | 225 | ||
| Alcohol B1P3 | ||||
| Drinking | 86 | 44.8 | 1,676 | 37.2 |
| No drinking | 106 | 55.2 | 2,831 | 62.8 |
| Missing | 5 | 233 | ||
| Low folate intakea | ||||
| Yes | 5 | 2.6 | 359 | 7.8 |
| No | 190 | 97.4 | 4,240 | 92.2 |
| Missing | 2 | 141 | ||
| Fever B1P3 | ||||
| Yes | 21 | 11.2 | 399 | 9.1 |
| No | 167 | 88.8 | 3,969 | 90.9 |
| Missing | 9 | 372 | ||
| Respiratory illness B1P3 | ||||
| Yes | 53 | 27.7 | 1,035 | 22.8 |
| No | 138 | 72.3 | 3,510 | 77.2 |
| Missing | 6 | 195 | ||
| Hypertension during pregnancy | ||||
| Yes | 24 | 12.2 | 446 | 9.6 |
| No | 172 | 87.8 | 4,221 | 92.2 |
| Missing | 1 | 73 | ||
| Antihypertensive medication use B1P3 | ||||
| Yes | 5 | 2.6 | 66 | 1.4 |
| No | 189 | 97.4 | 4,557 | 98.6 |
| Missing | 3 | 117 | ||
- Abbreviation: B1P3, 1 month before pregnancy through the third pregnancy month.
- a No folic acid supplementation during B1P3 and <600 μg/day dietary folate equivalents in the year before pregnancy.
The aOR for any CHD overall was 1.00 (95% CI: 0.81–1.25), and it was 0.97 (95% CI: 0.76–1.23) for any simple CHD overall (Table 2). We assessed eight specific simple CHDs. Three of the aORs were <1.0 and four of the aORs ranged from 1.0 to 1.5 with adjusted CIs including the null. Only one aOR was >1.5: total anomalous pulmonary venous return (TAPVR) with an aOR of 2.34 (95% CI: 1.08–5.07). Sensitivity analyses generally yielded similar results to the primary analysis, although the imprecision of the estimates was greater due to reduced numbers of exposed case children (Tables S1–S3).
| Birth defect | Case childrena,b exposed/unexposed | Unadjusted OR (95% CI) | Adjusted limitedc model OR (95% CI) |
|---|---|---|---|
| Any CHD | 161/3,680 | 1.02 (0.82–1.26) | 1.00 (0.81–1.25) |
| Any CHD simpled | 114/2,650 | 1.00 (0.79–1.27) | 0.97 (0.76–1.23) |
| Anomalous pulmonary venous return | |||
| Total anomalous pulmonary venous return | 8/94 | 1.98 (0.95–4.14) | 2.34 (1.08–5.07) |
| Conotruncal defects | |||
| Tetralogy of Fallot | 13/441 | 0.69 (0.39–1.22) | 0.66 (0.37–1.18) |
| D-transposition of the great arteries | 14/206 | 1.58 (0.90–2.77) | 1.30 (0.73–2.32) |
| Left ventricular outflow tract obstruction defects | |||
| Hypoplastic left heart syndrome | 7/263 | 0.62 (0.29–1.33) | 0.62 (0.29–1.36) |
| Coarctation of the aorta | 16/234 | 1.59 (0.94–2.70) | 1.29 (0.75–2.22) |
| Aortic stenosis | 9/151 | 1.39 (0.70–2.76) | 1.42 (0.70–2.89) |
| Right ventricular outflow tract obstruction defects with restriction | |||
| Pulmonary valve stenosis | 19/427 | 1.04 (0.64–1.68) | 0.96 (0.59–1.57) |
| Septal defects | |||
| Secundum atrial septal defect | 18/427 | 0.98 (0.60–1.61) | 1.05 (0.63–1.74) |
- a Children with missing values of covariates in the limited model have been excluded.
- b 4,447 unexposed controls and 191 exposed controls.
- c Limited model adjusted for estimated date of delivery year and season, plurality, maternal age at delivery, maternal race/ethnicity, cigarette smoking, alcohol use, and low folate intake.
- d Any CHD simple excludes case children with a complex CHD and case children with a CHD association.
4 DISCUSSION
We observed no association between influenza vaccination during the month before pregnancy through the third pregnancy month and any major CHD overall or any simple major CHD overall. Maternal influenza vaccination was associated with an increased risk of simple TAPVR, although there were only eight exposed case children, and the point estimate was imprecise. Furthermore, we observed modest associations (aOR's between 1.3 and 1.4) between maternal influenza vaccination and three specific CHD's (D-transposition of the great arteries, coarctation of the aorta, aortic stenosis) and approximately null or protective associations for four other specific CHD's, all with CIs including the null.
Across the current analysis of CHDs and our previous analysis of maternal influenza vaccination and non-CHDs using NBDPS data (Palmsten et al., 2022), only the adjusted CIs for simple TAPVR excluded the null. The biologic plausibility of this association is unclear (Alvarado-Terrones et al., 2018). Based on the number of comparisons from our CHD and non-CHD analyses (i.e., 29 overall), we expected one or two statistically significant associations due to chance. Regarding the existing literature, previous analyses from the Slone Epidemiology Center Birth Defects Study, also a multisite case–control study, reported no association between pandemic H1N1 influenza vaccination (2009–2010, 2010–2011 seasons; aOR: 1.09 [95% CI: 0.24–5.07]) and seasonal influenza vaccination (2011–2012, 2012–2013, 2013–2014 seasons; aOR: 0.64 [95%: 0.21–1.94]) and anomalous pulmonary venous return; information on TAPVR was not reported (Louik et al., 2013; Louik et al., 2016). Given only eight exposed cases of TAPVR were available in the current analysis and <5 exposed cases of anomalous pulmonary venous return were available in each of the analyses mentioned above, additional studies of maternal influenza vaccination and TAPVR are justified.
Other results from our study were similar to findings from previous studies. In the Slone Epidemiology Center Birth Defects Study, aORs for the associations between first-trimester seasonal flu vaccination and pandemic H1N1 influenza vaccination and CHDs, including any CHD overall, were <1.25 (Louik et al., 2013, Louik et al., 2016). Using automated health care data, the Vaccine Safety Datalink reported no association between first-trimester inactivated influenza vaccination and risk of any major CHD overall (adjusted prevalence ratio [aPR]: 1.00 [95% CI: 0.89–1.10] or any severe CHD overall [aPR: 0.99 [95% CI: 0.76, 1.30]) (Kharbanda et al., 2017). A study using registry data from Sweden reported an aOR of 1.10 (95% CI: 0.96–1.26) for first-trimester pandemic H1N1 influenza vaccination and any CHD overall (Ludvigsson et al., 2016).
Strengths of our study include the population-based data source, standardized interview protocols with collection of information on many potential confounders, detailed protocols to ensure accuracy of case classification (Rasmussen et al., 2003), and the ability to assess several specific simple CHDs that had not previously been examined in relation to maternal influenza vaccination.
Limitations of our study include imprecise estimates for specific CHDs, inability to study vaccine type (seasonal or pandemic) due to lack of detail, inability to study season of influenza vaccination due to small numbers, and potential selection bias due to early pregnancy loss. Also, misclassification of influenza vaccination is an important limitation and could have resulted from the lag between exposure, especially among mothers recalling vaccination status from more than one influenza season, underreporting due to indirect querying about influenza vaccination status, and our inability to confirm self-reported vaccination status. Differential exposure misclassification could bias results in either direction; nondifferential misclassification would tend to bias results downward, potentially masking positive associations (i.e., harmful effects).
5 CONCLUSION
Although we cannot entirely rule out increased risks for the specific CHDs analyzed due to the potential for exposure misclassification and imprecise estimates, our results add to the existing evidence supporting the safety of inactivated influenza vaccination during pregnancy. The observed TAPVR association may be due to chance, but it may help inform future studies. Together with other studies, our study strengthens the evidence that inactivated influenza vaccination in early pregnancy, including the month before conception, is not strongly associated with the risk for major structural CHDs.
ACKNOWLEDGEMENTS
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 the California Department of Public Health. KP conducted the analysis and JS independently replicated the analysis.
FUNDING INFORMATION
This project was supported through Centers for Disease Control and Prevention (CDC) cooperative agreements under PA #96043, PA #02081, FOA #DD09-001, FOA #DD13-003, and NOFO #DD18-001 to the Centers for Birth Defects Research and Prevention participating in the National Birth Defects Prevention Study (NBDPS) and/or the Birth Defects Study to Evaluate Pregnancy exposureS and the Iowa Center for Birth Defects Research and Prevention U01 DD001223 (PAR).
CONFLICT OF INTEREST
KP receives research grants from AbbVie, GSK, and Sanofi that are unrelated to this study.
ETHICS STATEMENT
Each National Birth Defects Prevention Study site obtained institutional review board approval and participants provided informed consent.
Open Research
DATA AVAILABILITY STATEMENT
The study questionnaires and process for accessing the data used in this study are described at https://www.cdc.gov/ncbddd/birthdefects/nbdps-public-access-procedures.html. The code book and analytic code may be made available upon request.
