Purpose: To measure the comparative effectiveness of metformin versus insulin for initial pharmacological management of gestational diabetes mellitus (GDM).
Methods: We conducted a population-based retrospective cohort study using administrative claims, maternity care, and laboratory result data from New Zealand. We followed pregnant women aged 15 to 45 from GDM diagnosis through delivery and assessed outcomes using maternity care and hospitalization data. We adjusted for covariates using inverse probability of treatment weights and multiple imputation for missing covariate information. We estimated unadjusted and adjusted risk ratios (RRs), risk differences (RDs) per 100, and 95% confidence intervals (CIs). Linear regression was used to estimate the association of treatment with birthweight. We stratified analyses by ethnicity and infant sex in prespecified sensitivity analyses.
Results: We compared 3818 metformin-treated pregnancies with 3450 insulin-treated pregnancies. We observed differences in treatment initiation by ethnicity, socioeconomic status, region, and calendar year. Treatment groups were similar in age, body mass index (BMI), and timing of diagnosis/treatment initiation. After adjustment, metformin was associated with reduced absolute risk of planned elective c-section (RD = -2.3, 95% CI, -4.3 to -0.3), large for gestational age (RD = -3.7, 95% CI, -5.5 to -1.8), and neonatal hypoglycemia (RD = -5.0, 95% CI, -6.9 to -3.2) compared with insulin. There were no clinically meaningful differences in average birthweight between metformin- and insulin-treated pregnancies. We observed variation in estimates by ethnicity and infant sex for some neonatal outcomes.
Conclusion: Metformin appears to be an effective treatment for women with GDM and may reduce risk of some adverse neonatal outcomes when compared with insulin.
Keywords: New Zealand; birthweight; cohort studies; diabetes, gestational; insulin; metformin; pharmacoepidemiology; pregnancy.
© 2019 John Wiley & Sons, Ltd.