(BSPOT19) Estimating the Risk of Gestational Diabetes Mellitus Associated With Antidepressant Continuation During Pregnancy Using a Treatment Decision Study Design

Background: Antidepressants (AD) are routinely used in the management of depression and anxiety during pregnancy. While AD use has been linked to increased risk of type 2 diabetes, inconsistent results have been observed with AD use in pregnancy and the risk of gestational diabetes mellitus (GDM).

Objectives: To examine the association between AD continuation during pregnancy and the risk of GDM.

Methods: This population-based cohort study of privately insured pregnant persons used data from Merative MarketScan between 2000 and 2021. We constructed a cohort of pregnant persons aged 12-55 years whose pregnancies ended in livebirth, stillbirth, induced abortion, or spontaneous abortion. Study participants were required to be continuously enrolled in their insurance plan, including prescription coverage from 90 days before their estimated last menstrual period (LMP) through 1-month post-partum. We further restricted the study population to include pregnancies that had a first prenatal care (PNC) visit between 5 to 20 weeks post LMP, had an ICD-10 diagnosis code for depression in the 120 days before the first PNC visit, and filled a prescription for an AD in the same period. Pregnancies with evidence of pregestational diabetes were excluded. We estimated the crude and inverse probability of treatment (IPT) weighted risk-ratios (RR) of GDM associated with AD continuation comparing participants who continued (continuers) filling prescriptions for any AD following their first PNC visit to participants who discontinued (discontinuers) after their first PNC visit allowing for a 30-day grace period.

Results: Of the 30,188 pregnancies included in this analysis, 15,569 (51.6%) filled a prescription for any AD following their first PNC visit. While AD continuers had a higher prevalence of anxiety at cohort entry, the prevalence of other comorbidities and medication use were similar between the two groups. The unadjusted risk for GDM was 8.7% among continuers of AD and 5.7% among discontinuers of AD. The unweighted RR for GDM comparing continuers of any AD to discontinuers was (RR = 1.44, 95% CI: 1.36 – 1.52), while that for SSRI was (RR=1.43. 95% CI: 1.34-1.53) and for SNRI was (RR= 1.46, 95% CI: 1.28-1.66). After weighting, the risk of GDM associated with continuation was similar for any AD (RR = 1.40, 95% CI: 1.32-1.48), SSRI (RR = 1.38, 95% CI: 1.29-1.47), and SNRI (RR= 1.40, 95% CI: 1.23-1.60).

Conclusions: Our results suggests any AD continuation was associated with an increase in the risk of GDM. While associations may be driven by residual confounding due to pre-pregnancy weight, quantitative bias analysis of the association suggested that this did not fully explain the relationship. Further studies examining AD class-specific risk of GDM associated with AD classes are warranted.