Antidepressant use during pregnancy and birth outcomes: analysis of electronic health data from the UK, Norway, and Sweden

ObjectivesTo explore the association between antidepressant use during pregnancy and birth outcomes. DesignCohort study. SettingElectronic health record data. Participants2 528 916 singleton births from the UKs Clinical Practice Research Datalink (1996-2018), Norways Medical Birth Registry (2009-2020), and Swedens Medical Birth Register (2006-2020). Main outcome measuresStillbirth, neonatal death, pre- and post-term delivery, small and large for gestational age, and low Apgar score five minutes post-delivery. ResultsA total of 120 209 (4.8%) deliveries were exposed to maternal antidepressant use during pregnancy. Maternal antidepressant use during pregnancy was associated with increased odds of stillbirth (adjusted pooled OR (aOR) 1.16, 95% CI 1.05 to 1.28), preterm delivery (aOR 1.26, 95% CI 1.23 to 1.30), and Apgar score < 7 at 5 minutes (aOR 1.83, 95% CI 1.75 to 1.91). These findings persisted in the discordant sibling analysis, but with higher uncertainty. The adjusted predicted absolute risk for stillbirth was 0.34% (95% CI 0.33 to 0.35) among the unexposed and 0.40% (95% CI 0.36 to 0.44) in the antidepressant exposed. Restricting to women with depression or anxiety, the association between antidepressant exposure and stillbirth attenuated (aOR 1.07, 95% CI 0.94 to 1.21). Paternal antidepressant use was modestly associated with preterm delivery and low Apgar score. Most antidepressants were associated with preterm delivery (except paroxetine) and Apgar score (except mirtazapine and amitriptyline). ConclusionsMaternal antidepressant use during pregnancy may increase the risk of stillbirth, preterm delivery, and low Apgar score, although the absolute risks remained low. Confounding by severity of indication cannot be ruled out, as the severity of symptoms was not available. The modest association between paternal antidepressant use and both preterm delivery and low Apgar score suggests that residual confounding by familial environment cannot be ruled out.