Abstract
Gestational diabetes mellitus (GDM) is among the most common complications of pregnancy and it increases
the risk of perinatal complications and long-term sequelae for both the mother (subsequent diabetes) and the
exposed child (obesity). Up to 50% of patients with GDM require the addition of pharmacologic therapy during
pregnancy. However, the long-term effects of pharmacologic treatment of GDM on maternal and child
outcomes are unknown, making the optimal treatment unclear. We propose a longitudinal cohort study of more
than 44,000 Kaiser Permanente members (>79% from racial/ethnic minority groups) with time-varying
exposure to metformin, glyburide, and insulin during pregnancy. Pregnancies diagnosed with GDM requiring
pharmacologic treatment from 2008 to 2021 at Kaiser Permanente Northern and Southern California will be
included. We will follow pregnant individuals and their children through our comprehensive electronic health
record (EHR) for up to 18 years. We will take advantage of natural variation in clinical practice and changes
over time in prescribing practices for first line medical therapy. The study population will consist of all patients
diagnosed with GDM who - after diet therapy - initiated pharmacotherapy within either insulin, glyburide, or
metformin. We will use state-of-the-art causal and statistical methods that can emulate inferences from
conceptual randomized controlled trials (RCTs) using observational data to inform the choice of pharmacologic
treatment in a large, diverse population of adults with GDM. This will be accomplished by conducting a series
of real-world data cohort studies based on retrospective EHR data with inclusion/exclusion criteria and start of
follow-up that mimic those of the RCTs emulated. More specifically, we plan to conduct analyses for pregnant
individuals diagnosed with GDM to compare the effects of static and dynamic treatment regimens with
glyburide, metformin, and insulin on maternal diabetes risk after pregnancy and child growth and obesity risk.
The static and dynamic treatment regimens will be specified to reflect current clinical practice and directly
inform decisions on, not only, what initial drug type and dose to choose, but also, if and how to intensify
treatment through increased drug dose or initiation of insulin when glycemia control is not reached with an oral
agent. The incidence of maternal diabetes (Aim 1) and child growth and obesity (Aim 2) outcomes will be
contrasted between any two static or dynamic regimens (e.g., sustained exposure to insulin versus exposure to
glyburide unless glycemic control is not reached, at which time insulin is initiated). This study will yield results
immediately applicable to clinical practice by providing head-to-head comparisons of pharmacologic GDM
treatment strategies (metformin, glyburide, insulin) using state-of-the-art analytic methodology
