Association of In Utero Antipsychotic Medication Exposure With Risk of Congenital Malformations in Nordic Countries and the US

Abstract

Importance: Psychiatric disorders are common among female individuals of reproductive age. While antipsychotic medication use is increasing, the safety of such medications in pregnancy is an area with large evidence gaps.

Objective: To evaluate the risk of first-trimester antipsychotic exposure with respect to congenital malformations, focusing on individual drugs and specific malformation subtypes.

Design, setting, and participants: This cohort study used data from nationwide health registers from the 5 Nordic countries and the US and spanned 1996 to 2018. The Nordic cohort included all pregnancies resulting in singleton live-born infants, and the US cohort consisted of publicly insured mothers linked to their live-born infants nested in the nationwide Medicaid Analytic eXtract. Data were analyzed from November 2020 to April 2022.

Exposures: One or more first-trimester dispensing of any atypical, any typical, and individual antipsychotic drugs.

Main outcomes and measures: Any major congenital malformation and specific malformation subtypes previously suggested to be associated with antipsychotic exposure in utero: cardiovascular malformations, oral clefts, neural tube defects, hip dysplasia, limb reduction defects, anorectal atresia/stenosis, gastroschisis, hydrocephalus, other specific brain anomalies, and esophageal disorders. Propensity score stratification was used to control for potential confounders. Pooled adjusted estimates were calculated using indirect standardization.

Results: A total of 6 455 324 unexposed mothers (mean maternal age range across countries: 24-31 years), 21 751 mothers exposed to atypical antipsychotic drugs (mean age range, 26-31 years), and 6371 mothers exposed to typical antipsychotic drugs (mean age range, 27-32 years) were included in the study cohort. Prevalence of any major malformation was 2.7% (95% CI, 2.7%-2.8%) in unexposed infants, 4.3% (95% CI, 4.1%-4.6%) in infants with atypical antipsychotic drug exposure, and 3.1% (95% CI, 2.7%-3.5%) in infants with typical antipsychotic drug exposure in utero. Among the most prevalent exposure-outcome combinations, adjusted relative risks (aRR) were generally close to the null. One exception was olanzapine exposure and oral cleft (aRR, 2.1 [95% CI, 1.1-4.3]); however, estimates varied across sensitivity analyses. Among moderately prevalent combinations, increased risks were observed for gastroschisis and other specific brain anomalies after atypical antipsychotic exposure (aRR, 1.5 [95% CI, 0.8-2.6] and 1.9 [95% CI, 1.1-3.0]) and for cardiac malformations after chlorprothixene exposure (aRR, 1.6 [95% CI, 1.0-2.7]). While the association direction was consistent across sensitivity analyses, confidence intervals were wide, prohibiting firm conclusions.

Conclusions and relevance: In this study, considering the evidence from primary and sensitivity analyses and inevitable statistical noise for very rare exposure-outcome combinations, in utero antipsychotic exposure generally was not meaningfully associated with an increased risk of malformations. The observed increased risks of oral clefts associated with olanzapine, gastroschisis, and other specific brain anomalies with atypical antipsychotics and cardiac malformations with chlorprothixene requires confirmation as evidence continues to accumulate.

Figure 1.. Cohort Creation and Antipsychotic Drug Exposure During the First Trimester of Pregnancies in the Nordic Databases and the US Medicaid Analytic eXtract

^aRestriction to singleton births was only applied within the Nordic cohorts. In the US data, multiples remained in the study cohort, with multiple gestation being included as a covariate in the propensity score. NA indicates not available (numbers suppressed due to country-/data-specific suppression policy).

Figure 2.. Standardized Differences of Pregnancies With Atypical or Typical Antipsychotic (AP) Exposure vs No AP Exposure During the First Trimester (Unadjusted and Propensity Score Weighted)

Not all standardized differences for each covariate and country shown due to small cell size suppression policy. ^aOther/unknown: American Indian or Alaska Native, Asian or Pacific Islander, Native Hawaiian or Other Pacific Islander, Hispanic or Latino and ≥1 races, >1 race, and unknown. Hispanic: Hispanic or Latino with no race information available. Other/unknown: Hispanic or Latino with ≥1 races. ^bOther psychoactive drugs include benzodiazepines, barbiturates, anxiolytics, and other hypnotics. ^cEstimated as (X_exp-X_ref)/√((s_exp²+s_ref²)/2). X represents the sample mean and s² the sample variance of the covariate in the AP exposed and reference group.

Figure 3.. Pooled Absolute Risk With 95% CIs of Major Malformations Overall Among Live-born Infants With and Without Antipsychotic Exposure in Utero

Prevalence of major congenital malformations among ziprasidone-, zuclopenthixol-, and clozapine-exposed pregnancies is not presented because the number of cases had to be suppressed due to country-/data-specific small cell suppression policies. Prevalence of major congenital malformations among pimozide- and sertindole-exposed pregnancies was zero and is therefore not shown.

Figure 4.. Pooled Unadjusted Relative Risk and Propensity Score—Weighted Standardized Morbidity Ratio (SMR), Separately for Each Exposure-Outcome Contrast, Sorted by Prevalence of Exposure and Outcome in the Nordic and US Cohorts

Rows are sorted by the most common to least common antipsychotic exposures and columns are sorted by the most common to least common outcomes among the unexposed. The outcome order was determined based on the absolute risk among the unexposed as observed in the US cohort (absolute risk for Nordic cohorts was not available for some outcomes due to country-specific cell suppression policies). Drugs in red belong to the atypical antipsychotic class and drugs in black to the typical antipsychotic class. The blue dotted line connects antipsychotic drugs with ≥100 exposed individuals and malformations with a prevalence of ≥0.5 per 1000 infants among the unexposed (as observed in the US cohort). The orange dashed line connects the least frequent antipsychotic with the least frequent malformation. Cells containing zeros are those with no outcomes of interest among the exposed.