(A-282) Association between continued benzodiazepine use and pregnancy outcomes in women undergoing assisted reproductive technology

Background: Benzodiazepines are commonly prescribed to women with infertility as psychiatric symptoms are likely to reduce the effectiveness of assisted reproductive technology (ART). However, the potential risk of benzodiazepines on ART remains a concern.

Objectives: To investigate the association between continued benzodiazepine use during ART and pregnancy outcomes in women undergoing ART

Methods: We conducted the retrospective cohort study using the nationwide claims database by including women diagnosed infertility with the history of in vitro fertilization (IVF) cycles with an embryo transfer (ET) between 2018 and 2022. The study cohort was restricted to 1) only the first identified cycle of ART and 2) cycles of individuals who had ≥2 benzodiazepine prescriptions from ET-730 days to ET-120 days. We classified continuers and discontinuers whether who had ≥1 benzodiazepine prescriptions or not within exposure assessment window (ET-120 days to ET+120 days). The primary outcome of interest was defined as ART outcomes including clinical pregnancy, live birth, preterm birth, and miscarriage. Relative risks (RRs) and 95% confidence intervals (CIs) were estimated using Poisson regression models with robust standard errors after adopting propensity score overlap weighting to balance covariates. Further, we conducted an additional analysis, restricting the cycles to those with indications for benzodiazepine use within the exposure assessment window, to address confounding by indication.

Results: Among 60,397 women underwent their first IVF cycle with ET, we identified 2,064 benzo-continuers and 2,021 benzo-discontinuers. Compared to discontinuers, continuers had a higher risk of miscarriage (crude RR 1.45, 95% CI 1.20-1.76) and a slightly lower likelihood of live birth (0.82, 0.74-0.90), whereas no significant associations were observed for clinical pregnancy (0.95, 0.87-1.04) or preterm birth (1.01, 0.75-1.37). The results remained consistent after covariate adjustment (live birth: adjusted RR [aRR] 0.85, 0.77-0.94; miscarriage: 1.43, 1.17-1.76; clinical pregnancy: 0.95, 0.87-1.04; preterm birth: 1.01, 0.75-1.37). However, after restricting to those with indications for benzodiazepine use, the risks were reduced for live birth (aRR 0.91, 0.69–1.20) and miscarriage (1.26, 0.69–2.33).

Conclusions: Continuation of benzodiazepine use around the time of embryo transfer (ET) was associated with an increased risk of adverse ART outcomes; however, these associations were attenuated after further accounting for benzodiazepine indications. This suggests that the observed risks may be driven by the underlying indications for benzodiazepine use rather than the continuation of benzodiazepine treatment itself.