Although current prescribing guidelines suggest continuation of psychotropic drugs in pregnant women, population-based evidence supporting their safety is limited.

This study aims to clarify the plausible causal links between maternal psychotropic drug exposures and obstetric complications.

This cohort study investigated all births by Hong Kong residents ≥18 years of age in public hospitals between 2004 and 2022. Birth episodes were classified according to whether they were unexposed to psychotropic drugs, exposed but discontinued before conception or exposed during pregnancy. Firth’s penalised logistic regression was employed in all analysis, and negative control analysis was conducted to assess causality. False discovery rate correction and sensitivity analyses were performed.

Among 587 419 births, 7182 episodes involved psychotropic prescriptions (antipsychotics, antidepressants, anticonvulsants, benzodiazepines) during pregnancy. In broad drug class analysis, all significant associations observed in the exposed group were also observed in negative control analysis (psychotropics discontinued before conception), suggesting that elevated risks could be attributed to unmeasured confounders. Nevertheless, in subclass analyses, certain psychotropic drugs showed increased risks of obstetric complications, i.e. significant associations between atypical antipsychotics and genito-urinary infection (odds ratio 2.70, 95% CI 1.46–4.83), and between valproate and low birth weight (odds ratio 1.68, 95% CI 1.16–2.37). These associations became non-significant in negative control analysis, and the high E-values (atypical antipsychotics and genito-urinary infection, 4.84; valproate and low birth weight, 2.75) suggested that the results were unlikely to have been driven by unmeasured confounders. Maternal diagnoses of schizophrenia and depression were independently associated with increased risk of obstetric complications, after controlling for the effects of psychotropics.

The population-based data and meticulous analyses did not support any clear causal link between broad-class psychotropic exposure during pregnancy and increased risk of obstetric/neonatal complications. However, some psychotropic subclasses may increase obstetric/neonatal complications. The limited number of episodes involving discontinuation of some psychotropic subclasses may have resulted in false negative findings in the negative control analysis.