(AA) Emulation of a Target Trial of Metformin in Pregnancy in Women with Pre-Gestational Type 2 Diabetes Mellitus

Publication Number: 1025

Background: For women with pre-gestational type 2 diabetes mellitus (T2DM), current guidelines recommend achieving glucose control during pregnancy with diet, exercise, and insulin treatment. There is a lack of evidence on the safety and effectiveness of non-insulin agents for T2DM in pregnancy. The Metformin in Women with Type 2 Diabetes in Pregnancy [MiTy] study evaluated pregnancy outcomes among 502 women with pre-gestational T2DM who were randomized to treatment with insulin plus metformin or insulin plus placebo. There was no difference in the primary outcome (a composite of neonatal events), and women in the metformin group had similar or better glycemic outcomes, compared with those who received placebo.

Objectives: To emulate a target trial in real-world data comparing metformin continuation versus discontinuation among women with T2DM who were using both insulin and metformin during the three months before pregnancy.

Methods: This study was conducted within the IBM MarketScan Database for 2011-2015. Pregnant women with T2DM, with at least one dispensation of both metformin and insulin during the three months before the estimated date of last menstrual period were eligible. Metformin continuation was defined as at least one dispensation of metformin between 6 and 24 weeks of pregnancy (at any time), while metformin discontinuation was defined as no dispensations of metformin during this window. The primary outcome was a composite of pregnancy loss, preterm birth, birth injury, respiratory distress, neonatal hypoglycemia, and neonatal intensive care unit admission. We fit pooled logistic regression models and estimated population risks and risk differences (RDs) by standardizing estimates to the baseline covariates. We used percentile-based bootstrapping with 500 samples to calculate 95% confidence intervals (CIs).

Results: Of 579 eligible women, 373 were treated with only insulin and 206 with both metformin and insulin during gestational weeks 6-24. The unadjusted risk of the primary composite outcome was 29.6% among those who continued metformin and 30.6% among those who discontinued. After adjusting for potential confounders, the risk difference was -1.40% (95% CI: -8.50%, 5.30%).

Conclusions: Preliminary findings are consistent with those reported in the MiTy trial and indicate no increased risk for adverse pregnancy outcomes among women who continue metformin during pregnancy. Next steps include an evaluation of specific outcomes, patient subgroups, and other non-insulin agents, pooling data from other sources.