3G-01 - GLP-1 receptor agonist continuation into the first trimester of pregnancy and risk of non-live birth and abnormal fetal growth: a target trial emulation

Background: There is limited data on the effect of GLP-1 receptor agonist (GLP-1 RA) use during pregnancy. In some animal studies, administration during pregnancy led to increased early losses and decreased fetal weights, but the applicability to humans of these findings is unclear. Discontinuation during pregnancy might lead to excessive maternal and fetal weight gain.

Objectives: To estimate the association of GLP-1 RA continuation into the first trimester with non-live birth, small for gestational age (SGA), and large for gestational age (LGA) at birth.

Methods: We emulated a target trial among users of GLP-1 RAs of immediate discontinuation of GLP-1 RA following conception vs. at least one further GLP-1 RA dispensation in the first trimester using MarketScan US commercial insurance claims data (2011-2022). Pregnancies among women aged 16-55 with ≥1 GLP-1 RA dispensation in the 90 days prior to LMP were included. To handle indeterminate treatment strategies at baseline, cloning and censoring, with a 12-week grace period, was used. Risks of non-live birth and separately of spontaneous abortion (SAB), elective termination, and stillbirth were estimated using pooled logistic regression and were standardized to baseline covariates. Among live birth pregnancies linked to infants, the standardized prevalence of SGA and LGA was estimated using logistic regression. 95% confidence intervals (CI) were calculated using percentile bootstrapping with 500 iterations.

Results: We included 1,790 pregnancies among 1,716 women. After the 12-week grace period there were 1,275 continuing pregnancies (558 GLP-1 RA continuers), of which 1,231 (534 continuers) resulted in a live birth and 737 (303 continuers) had linked infant data. Only 14.8% of live birth continuer pregnancies (n=43) had GLP-1 RA dispensations beyond the first trimester. The standardized risks among continuers were 29.5% for non-live birth, 23.2% for SAB, 4.3% for elective termination, and 1.1% for stillbirth. The risks among discontinuers were respectively 27.8%, 23.0%, 3.9%, and 0.9%. The corresponding risk ratios (RR) were 1.06 (95% CI 0.97-1.17) for non-live birth, 1.01 (0.94-1.09) for SAB, 1.12 (0.87-1.42) for elective termination, and 1.33 (0.48-3.54) for stillbirth. The standardized prevalence among live births of SGA was 6.1% among continuers vs. 5.8% among discontinuers (RR: 1.04, 0.53-1.92) and of LGA was 10.8% among continuers vs. 11.9% among discontinuers (RR 0.91, 0.62-1.32).

Conclusions: Findings are not indicative of substantially increased risk of non-live birth, SGA, or LGA with GLP-1 RA continuation into the first trimester, giving some reassurance on the safety of early pregnancy exposure with respect to these outcomes.