ABSTRACT
Background
Metformin works by increasing insulin sensitivity, which may aid in the management of insulin-treated type 2 diabetes in pregnancy. However, large trials have found that adjunctive metformin did not reduce composite adverse neonatal outcomes in this population.
Objective
Our objective was to determine if baseline insulin requirements modified the effect of metformin on adverse neonatal and maternal outcomes.
Study Design
We performed a secondary analysis of the Medical Optimization and Management of Pregnancies with Overt Type 2 Diabetes (MOMPOD) trial, a randomized controlled trial of metformin versus placebo in insulin-treated type 2 diabetes (T2D) or diabetes diagnosed in early pregnancy. We included participants who took ≥1 dose of study drug and had baseline insulin data available. Our primary outcome was a composite of adverse neonatal outcomes, as defined by the parent trial. Secondary outcomes included cesarean delivery, maternal gestational weight gain, and change in insulin requirements during pregnancy, among others. Total daily dose (TDD) of insulin at randomization was evaluated as an effect modifier of the relationship between metformin and the primary outcome using a likelihood ratio test with p<0.10 as significant. To illustrate effect modification, we subsequently evaluated the association between metformin and outcomes among subgroups with TDD <30 units (U), >60U, and >90U using multivariable Poisson regression with robust error variance and ordinary least squares regression.
Results
Of the 794 participants included in the parent trial analysis, 785 (99%) met criteria for our study. The median TDD at randomization was 62U (IQR 35, 88) with 155 (20%) using <30U, 225 (29%) using 30-60U, 222 (28%) using 60-90U, 100 (13%) using 91-120U, and 83 (10%) using >120U. There was a significant interaction between metformin and TDD of insulin in relation to the primary composite adverse neonatal outcome (p=0.089). For participants with TDD of insulin <30U, adjunctive metformin was associated with a lower risk of the composite adverse neonatal outcome, compared with placebo (44% vs 55%, adjusted relative risk 0.76, 95% CI [0.59-0.98]). There were no differences in the composite adverse neonatal outcome among participants with TDD of insulin >60U or >90U. However, for participants with TDD of insulin >60U and >90U, adjunctive metformin was associated with a smaller increase in insulin requirements during pregnancy, compared with placebo (TDD >60U: mean difference -13 units, 95% CI -26 to -1; TDD>90U: mean difference -27 units, 95% CI -50 to -5).
Conclusion
The effect of metformin on maternal and neonatal outcomes in the MOMPOD trial cohort differed by baseline insulin requirements. In pregnant women with pregestational type 2 diabetes or diabetes diagnosed early in pregnancy, adjunctive metformin was associated with a lower risk of the composite adverse neonatal outcome, preterm birth, LGA, NICU admission, and less neonatal fat mass among participants with low baseline insulin requirements, when compared with placebo. Conversely, metformin may help mitigate increases in insulin requirements during pregnancy for those with high baseline insulin requirements. Further studies are needed to confirm our findings and investigate the mechanisms underlying the differential effects of metformin based on baseline insulin requirements in pregnancy.