Prenatal Tetanus, Diphtheria, Acellular Pertussis Vaccination and Autism Spectrum Disorder

In this retrospective cohort study, we observed the outcome of an ASD diagnosis in children born at all KPSC hospitals between January 1, 2011, and December 31, 2014. KPSC is an integrated health care delivery organization that provides health care in 15 hospitals and ∼220 medical offices to ∼4.4 million members who are broadly representative of the Southern California population.²⁵ KPSC uses electronic medical records (EMRs) to integrate medical information, such as diagnosis and immunizations. All recommended immunizations are free to members regardless of copayment status. This study was approved by the KPSC Institutional Review Board, which waived the requirement for informed consent.

Eligibility was restricted to pregnant women who did not have assisted conceptions (in vitro fertilization) and gave birth to live singleton infants at 22 to 45 weeks’ gestation. Childrens' medical records were linked longitudinally to biological mothers by using unique identifiers. Maternal and child pairs were included if pregnant women had continuous membership since the beginning of pregnancy (allowing for a ≤31-day gap) and infants were members for at least 90 continuous days after turning 1 year of age. Pairs were excluded if children were diagnosed with chromosomal or congenital anomalies.

Maternal vaccination captured in the EMR was defined as receipt of Tdap anytime during pregnancy (yes, no), from the pregnancy start date to the day before the delivery date. The pregnancy start date was calculated by subtracting the gestational age at birth from the child’s birth date. Adacel (Sanofi Pasteur, Swiftwater, PA) was the primary Tdap product used at KPSC facilities during the study period.²⁶ The unvaccinated group was defined by women who were not vaccinated with Tdap anytime during their entire pregnancy.

ASD was defined by a clinical diagnosis any time after turning 1 year of age recorded in the EMR between January 1, 2012, and June 30, 2017 (International Classification of Diseases, Ninth Revision codes: 299.0, 299.8, 299.9; and International Classification of Diseases, 10th Revision codes: F84.0, F84.5, F84.8, F84.9). Although we consider children at risk for ASD from birth, we restricted the minimum diagnosis age on the basis of the 12-month minimum allowable evaluation age of the Autism Diagnostic Observation Schedule, the gold standard diagnostic instrument for identifying ASD.²⁷ 

Maternal and child characteristics were obtained from the EMR. Child data included sex (male or female), birth year (2011–2014), birth season (winter, spring, summer, fall), gestational age at birth (mean, SD; <37 or ≥37 weeks’ gestation), and birth weight (mean, SD). Maternal data at delivery included maternal age (≤25, 26–30, 31–34, ≥35 years), race and/or ethnicity (non-Hispanic white, non-Hispanic African American, Hispanic, non-Hispanic Asian American or Pacific Islander, other or multiracial, unknown), primary language (English, other), interpreter needed (yes, no), educational attainment (less than or equal to high school, some college, bachelor’s degree, greater than or equal to master’s degree, unknown), and medical center of delivery. Additional pregnancy information included Medicaid insurance (yes, no), parity (0, 1, ≥2), receipt of influenza vaccine during pregnancy (yes, no), start of prenatal care (≤3 months gestation, >3 months gestation, no care or missing), and medical and obstetrical complications (pregestational hypertension, preeclampsia or eclampsia, pregestational and gestational diabetes, placenta previa, and placental abruption).

The distributions of maternal and child characteristics by prenatal Tdap vaccination status and either χ² test or Student’s t test for differences in characteristics were calculated. Follow-up characteristics, age of ASD diagnosis, and the incidence rate of ASD diagnosis by Tdap vaccination status were calculated. Children were followed from birth to first ASD diagnosis, end of membership in the health plan, or the end of the study follow-up period (June 30, 2017), whichever came first.

To assess the potential concern that ASD diagnoses were missed because of dual enrollment in another health care system and/or limited health care use within the KPSC system, we also calculated the proportion of children with no outpatient encounters after turning 1 year of age. Proportions were calculated by whether their mothers received the Tdap vaccine during pregnancy.

Cox proportional hazards regression was used to estimate the unadjusted and adjusted hazard ratios (HRs) and their 95% confidence intervals (CIs) to evaluate the magnitude of association between maternal Tdap vaccination and ASD diagnosis in children. To evaluate potential variations in results, we also conducted stratified analyses by child’s birth year and by assessing results among women giving birth to their first child (nulliparous). We used propensity score analyses with inverse probability of treatment weighting (IPTW) to adjust for potential confounding. First, we used a logistic regression model including variables associated with vaccination a priori,^28,29 or variables from bivariate results (P < .05) that could reasonably affect Tdap receipt, to estimate the probability of Tdap vaccination. The following variables were included: child’s birth year, gestational age at birth (<37 or ≥37 weeks’ gestation), maternal age, race and/or ethnicity, education, medical center of delivery, Medicaid insurance, parity, start of prenatal care, and influenza vaccination during pregnancy. Second, the weight for each mother was calculated as the inverse of her predicted probability of Tdap exposure and was normalized by dividing by the mean weight of each exposure group. Standardized difference scores were used to assess whether balance of covariates was achieved between the comparison groups. Unlike P values, for which the magnitude is highly related to the sample size, standardized difference is a unified approach to quantifying the magnitude of difference between groups regardless of sample size; an absolute value <0.10 is considered a negligible difference.^30,31 

There were 109 536 women who had continuous enrollment since the beginning of pregnancy with nonassisted conceptions and gave birth to live singletons between January 1, 2011, and December 31, 2014. Of these, 94 400 (86.2%) children were enrolled by 1 year of age. After excluding an additional 9356 (8.5%) children with <90 days of continuous enrollment and 3051 (2.8%) diagnosed with chromosomal or congenital anomalies, the study included 81 993 children (Fig 1).

Prenatal Tdap vaccination coverage ranged from 26% (5407 of 20 553) for the 2012 birth cohort to 79% (16 882 of 21 433) for the 2014 birth cohort. The mean and median gestational age at vaccination was 28 weeks (SD: 7.4) and 29 weeks, respectively (interquartile range: 26–33 weeks). Women vaccinated during pregnancy were more likely to be Asian American or Pacific Islander, have a bachelor’s degree or higher, be nulliparous, have received the influenza vaccine prenatally, and give birth at term (≥37 weeks’ gestation) compared with unvaccinated women (Table 1).

The follow-up time ranged between 1.2 and 6.5 years and was on average longer in the unvaccinated group (Table 2). There was an average difference of 6 months in follow-up time between both groups (unvaccinated, mean: 4.44 years [SD: 1.18]; vaccinated, mean: 3.85 years [SD: 1.29]). Approximately 15% of children in the unvaccinated and 13% in the vaccinated group did not have complete follow-up through the end of the study because of termination of KPSC membership. ASD was diagnosed in 1341 children (1.6%). The proportion of children with ASD across birth years revealed a decline from 2.0% to 1.5% in the unvaccinated and 1.8% to 1.2% in the vaccinated group. Among children diagnosed with ASD, the first diagnosis commonly occurred at 2 years of age in either of the exposure groups (44%) followed by diagnoses at ages 3 and 4 years. Among children diagnosed with ASD before age 2 (n = 213), 93.5% of children had a second confirmatory diagnosis after age 2 years. Among children of unvaccinated mothers and children of vaccinated mothers, there were 187 (0.44%) and 126 (0.32%) children, respectively, who did not have an outpatient encounter after 1 year of age.

The ASD incidence rate was 3.78 per 1000 person-years in the Tdap vaccinated and 4.05 per 1000 person-years in the unvaccinated group (Table 3). The unadjusted HR was 0.98 (95% CI: 0.88–1.09). After propensity score weighting, the standardized difference scores were <0.1 for all variables, suggesting a good balance of covariates (Fig 2). The IPTW-adjusted analyses showed that Tdap vaccination during pregnancy was not associated with increased ASD risk (HR: 0.85, 95% CI: 0.77–0.95). Results were consistent across study birth years and among nulliparous women.

In this large retrospective observational cohort study of 81 993 pairs of diverse pregnant women and their children, we found no evidence of increased risk for ASD diagnosis associated with Tdap vaccination during pregnancy. Subanalyses supported the overall results, revealing minimal variability by year of birth and parity. No study to our knowledge has been published with results examining the risk of ASD after maternal exposure to the Tdap vaccine.

Maternal immune activation during pregnancy is hypothesized to indirectly affect fetal neurodevelopment.^32,–35 Researchers have found infections during pregnancy (eg, rubella and influenza), including prolonged episodes of fever to increase autism risk, hypothesizing that maternal infections, cytokine responses, and proinflammatory pathways are likely to alter fetal brain development.^32,36,–41 Our results potentially indicate that the maternal Tdap vaccine affects immune trajectories protecting infants against infections that would otherwise lead to neurodevelopmental alterations. Alternatively, vaccinations during pregnancy can possibly activate the maternal immune system leading to neurodevelopmental insults. However, except for 1 study in which the association between the influenza vaccine during pregnancy and ASD was examined,⁴² there is an insufficient number of studies that contain examinations of this hypothesis.

Only a handful of researchers have investigated adverse events potentially associated with Tdap receipt during pregnancy that may point to possible intermediate physiologic pathways to ASD. Two studies at sites participating in the Vaccine Safety Datalink project and a third study of >1 million pregnant women revealed a small but statistically significant increased risk of chorioamnionitis after maternal Tdap vaccination (6.1% in vaccinated, 5.5% in unvaccinated).^43,–46 Although Tdap increased the risk for chorioamnionitis, the investigators in the referenced studies found that the resulting chorioamnionitis was not associated with a preterm birth, a potential risk factor for ASD.⁴⁷ Given that maternal Tdap vaccination alone or mediated by chorioamnionitis did not translate to a preterm birth or any clinically relevant adverse event in newborns,^9,46 a direct or mediated relationship between maternal Tdap and ASD is not supported.

Our findings have some strengths and limitations. ASD was determined by recorded diagnoses in the EMR and was not validated by a study-specific standardized assessment. In addition, misclassification of ASD diagnosis could have occurred among children with dual insurance or limited health care use, if they sought care for ASD elsewhere. However, incident ASD diagnoses were likely captured consistently during most of the study years because of a California law enacted in 2012 (Senate Bill 946) requiring health plans to cover ASD-associated health costs, such as diagnostic and behavioral health treatment,⁴⁸ prompting the implementation of systematic procedures for screening and diagnosing within the KPSC health care system. Thus, a diagnosis of ASD can only be made by qualified mental health professionals (child and/or adolescent psychiatrists, developmental and/or behavioral pediatricians, child psychologists, or neurologists). In addition, the combined prevalence of ASD in this study was 1.6%, comparable to the estimated 1.7% prevalence among 8-year-old children reported in the United States.⁴⁹ In addition, vaccinated and unvaccinated women may be different in aspects that may also be associated with ASD diagnosis. Because study follow-up time was dependent on the child’s birth year, children in this study had different opportunities for receiving an ASD diagnosis (ie, children born in 2011 had maximum follow-up potential of 5.5–6.5 years, whereas children born in 2014 had 2.5–3.5 years). Based on US representative data, 3 to 5 years of age is the period when most children with an ASD are diagnosed.⁵⁰ Thus, we likely did not capture some children with ASD born in later study years considering that some children with milder ASD would not be diagnosed until they reach school age. Similarly, maternal Tdap vaccine uptake varied by birth year because of evolving ACIP recommendations and pertussis epidemics in California.^2,51,52 However, because we identified minimal variability in study results when stratifying by birth year, these variations likely did not affect our results. In addition, although we controlled for major known confounding factors related to Tdap vaccination and ASD, our results could have been affected by residual confounding, such as behavioral or clinical factors not captured by the EMR. For example, we did not control for having an older sibling with ASD, although the sibling recurrence risk is between 10% and 20%, a 10- to 20-fold increase over the general population’s prevalence.^53,54 However, in our analysis restricted to mothers giving birth to their first child, the results were consistent with overall findings.

Our study has additional strengths. Maternal Tdap vaccination and ASD information was not subject to recall bias, and our weighting procedures enabled us to balance the Tdap exposed and unexposed groups to compare 2 populations that were comparable in important measured confounding factors. In addition, we included children diagnosed with ASD starting at 1 year of age in consideration of the latest evidence in ASD screening and diagnosis research and practice.^55,–57 Still, we were able to confirm that the majority (>93%) of children with their first diagnosis of ASD before age 2 also had a diagnosis after age 2, which aligns with the American Academy of Pediatrics autism screening recommended schedule of 18- and 24-months.⁵⁸ 

Our findings suggest that getting vaccinated with Tdap during pregnancy is not associated with an increased risk of ASD in children. Future studies with additional birth years and longer follow-up can add to the scientific evidence about Tdap vaccination during pregnancy and ASD in offspring. We provide evidence supporting the ACIP’s recommendation to vaccinate pregnant women to protect vulnerable infants, who are at highest risk of hospitalization and death after pertussis infection.

ACIP

Advisory Committee on Immunization Practices

ASD

autism spectrum disorder

CI

confidence interval

EMR

electronic medical record

HR

hazard ratio

IPTW

inverse probability of treatment weighting

KPSC

Kaiser Permanente Southern California

Tdap

tetanus toxoid, reduced diphtheria toxoid, and acellular pertussis, adsorbed