(A-170) Benzodiazepines Initiation During Pregnancy and the Risk of Adverse Neonatal Outcomes

Background: Benzodiazepines (BZDs) are commonly used during pregnancy to treat anxiety and sleep disorders. Available evidence on the risk of adverse neonatal outcomes associated with BZD use during pregnancy remains limited and conflicting.

Objectives: To evaluate the risk of adverse neonatal outcomes following BZD initiation versus no initiation during pregnancy.

Methods: Using Marketscan (2003-22) and Medicaid (2000-18), we emulated a sequence of target trials examining BZD initiation during each month of pregnancy. For each sequential trial, eligible pregnancies were those who remained pregnant at the beginning of each month and had not previously initiated BZDs. Treatment strategies were defined as filling ≥1 BZD prescription vs not filling a prescription during that month. Propensity score (PS) overlap weights were used to control for demographic characteristics, psychiatric conditions, other comorbid conditions, and healthcare utilization, assessed from 3 months before pregnancy until treatment assignment. The outcomes of interest were preterm birth, small for gestational age (SGA), low birth weight (LBW), neonatal intensive care unit (NICU) admission, respiratory problems and neonatal hypoglycemia. Log-binomial regression was used to estimate the relative risks (RRs) and 95% confidence intervals (CIs). To estimate the effect of the BZD initiation early vs. late in pregnancy (two potential etiologically relevant windows), we pooled RRs of months 1 through 5 and months 6 through 9.

Results: A total of 105,218 (2.2%) women out of 4,847,024 initiated BZD during pregnancy. Before adjustment, the prevalence of psychiatric conditions and healthcare utilization was higher among the BZD-treated pregnancies compared to the untreated. The adjusted RR (95% CI) for preterm birth was 1.03 (1.01-1.05) for early BZD initiation and 1.30 (1.23-1.37) for late initiation. The corresponding RR (95% CI) were 1.08 (1.05-1.11) and 1.14 (1.04-1.24) for SGA; 1.08 (1.05-1.10) and 1.28 (1.19-1.38) for LBW; 1.08 (1.06-1.10) and 1.28 (1.21-1.34) for NICU admission; 1.08 (1.05-1.10) and 1.23 (1.15-1.31) for respiratory problems; 1.04 (0.99-1.08) and 1.18 (1.04-1.31) for neonatal hypoglycemia. Adjustment reduced the estimates by >50%, raising concerns about residual confounding.

Conclusions: While we did not observe meaningful increased risks associated with BZD initiation early in pregnancy, a small to moderate increase was observed for BZD initiation in late pregnancy for all outcomes after adjustment using conventional PS. High-dimensional PS approaches will be explored to evaluate residual confounding as an alternative explanation for these associations, given concerns about complications arising later in pregnancy leading to anxiety and the initiation of BZDs.