Oral medications less effective than insulin for preventing large babies in gestational diabetes study

By Vijay Kumar MalesuReviewed by Danielle Ellis, B.Sc.Jan 9 2025

23.9% of infants were large for gestational age with oral medications compared to 19.9% with insulin, falling short of noninferiority criteria

In a recent study published in the JAMA, a group of researchers evaluated whether a treatment strategy using oral glucose-lowering agents (metformin and glyburide) is non-inferior to insulin in preventing large-for-gestational-age (LGA) infants in gestational diabetes.

Background

Insulin has traditionally been the primary pharmacological treatment for gestational diabetes, improving perinatal outcomes when dietary measures fail. Over the past two decades, oral glucose-lowering agents like metformin and glyburide have gained attention as easier-to-administer, cost-effective alternatives with better patient acceptance.

However, concerns about their placental transfer and long-term safety in offspring persist, leading to varying guidelines globally. While metformin and glyburide are widely used, their effectiveness compared to insulin and the potential benefits of sequential combination therapies remain unclear.

Further research is needed to determine the optimal treatment strategy balancing efficacy, safety, cost, and patient satisfaction.

About the study

The present randomized, open-label trial was conducted across 25 centers in the Netherlands.

Participants were pregnant individuals aged over 18 years with singleton pregnancies between 16 and 34 weeks of gestation who failed to achieve glycemic control after approximately two weeks of dietary intervention.

Glycemic control criteria included fasting glucose levels above 95 mg/dL, one-hour postprandial glucose above 140 mg/dL, or two-hour postprandial glucose above 120 mg/dL, measured via self-testing. Individuals with pre-existing diabetes, severe medical or psychiatric comorbidities, major fetal anomalies, or known chromosomal abnormalities were excluded.

Randomization was performed using a central computerized system with a 1:1 ratio, assigning participants to either metformin or insulin. Metformin was initiated at 500 mg once daily, titrated up to 1000 mg twice daily, or the highest tolerated dose.

If glycemic control remained inadequate, glyburide was added, and insulin was introduced if necessary. Insulin dosing followed local clinical practices. Outcomes were assessed for primary (LGA infants) and secondary measures, including maternal and neonatal health indicators.

Study results

Between April 2017 and November 2022, 820 participants from 1,656 eligible individuals consented to participate and were randomized equally into the intervention group (oral glucose-lowering agents, n=409) and the control group (insulin, n=411).

Three participants in the oral agents group and eight in the insulin group withdrew consent, with an additional five in the insulin group lost to follow-up. Baseline characteristics of the 406 participants in the oral agents group and 398 in the insulin group were similar. The mean age was 33.2 years, mean pre-pregnancy BMI was 30.4, and 35% of participants were nulliparous, with 58% identifying as White.

In the oral agents group, 55% maintained glycemic control with metformin alone, and an additional 24% achieved control with glyburide, resulting in 79% maintaining control without insulin. However, 7.7% required insulin despite oral agents, and adverse effects prompted 3.7% to switch to insulin. Supply issues with glyburide further excluded nine participants from per-protocol analysis. Among insulin-assigned participants, two did not follow protocol, reducing their per-protocol group size slightly.

For the primary outcome, 23.9% of infants in the oral agents group were LGA, compared to 19.9% in the insulin group, with a non-significant absolute risk difference of 4.0% (95% CI, −1.7% to 9.8%). Noninferiority was not established (P=0.09). In per-protocol analysis, the absolute risk difference was 3.4% (95% CI, −2.5% to 9.3%, P=0.06), similarly failing to demonstrate noninferiority.

Secondary outcomes showed significantly more maternal hypoglycemia in the oral agents group (20.9% vs 10.9%; absolute risk difference, 10.0%, 95% CI, 3.7%-21.2%). Other maternal and neonatal outcomes, including cesarean delivery rates, preeclampsia, and neonatal hypoglycemia, were comparable between groups. Adverse effects were more common in the oral agents group, with nausea and diarrhea being the most frequent.

Exploratory analyses indicated no subgroup differences for secondary outcomes. Neonatal intravenous glucose therapy was more frequent in the oral agents group (6.4% vs 3.2%). Patient satisfaction was comparable across groups, though oral agent participants more often recommended their treatment. The time between randomization and delivery was similar. Overall, no significant differences in secondary outcomes were observed between groups in per-protocol analyses.

Conclusions

To summarize, for pregnant individuals with gestational diabetes requiring pharmacological treatment, oral glucose-lowering agents were not non-inferior to insulin in preventing large-for-gestational-age infants, as the confidence interval exceeded the noninferiority margin.

While adding glyburide to metformin reduced the need for insulin compared to prior studies, treatment satisfaction was higher with oral agents despite a higher incidence of adverse effects, including maternal hypoglycemia.

Differences in trial populations and clinician discretion for treatment adaptation may have influenced outcomes.