WASHINGTON, DC — Research recently published on diabetes in pregnancy has offered insight on issues such as earlier pharmacologic treatment of gestational diabetes (GDM) and combination treatment with both metformin and insulin for GDM and type 2 diabetes mellitus (T2DM).

Still, uncertainty about the roles of insulin and metformin in treating diabetes in pregnancy — particularly relating to potential risks for offspring — persists. “The wires constantly cross in my mind about whether we’re balancing efficacy and risk in a meaningful way,” said Maisa N. Feghali, MD, MS, assistant professor in the Department of Obstetrics, Gynecology & Reproductive Sciences and Magee Women's Research Institute at the University of Pittsburgh, during a discussion of medication management during the biennial meeting of the Diabetes in Pregnancy Study Group of North America.

Clinical practice guidelines offer differing advice. The American College of Obstetricians and Gynecologists strongly recommends insulin as the preferred, first-line pharmacologic therapy for GDM when such therapy is needed, with metformin as an option when patients decline or cannot safely use insulin. The Society of Maternal-Fetal Medicine accepts metformin as a “reasonable and safe” first-line alternative to insulin.

Feghali said that, right now, there’s a role for both metformin and insulin. In a few years, she and others hope decision-making will be informed by findings from a multi-center, multi-million dollar study — coined the DECIDE trial — that is currently recruiting more than 1500 pregnant individuals with GDM in the United States.

Participants will be randomized to either metformin or insulin, and mothers and children will be followed for 2 years after delivery. Funded by the Patient-Centered Outcomes Research Initiative, the study aims to determine whether metformin is inferior to insulin in reducing adverse pregnancy outcomes and if it is comparably safe for patients and their children. 

Current Practice and Recent Trials on Metformin and Insulin

A shift toward metformin as the preferred oral option — and away from glyburide — began 10 years ago when a large retrospective population-based study and a meta-analysis showed more frequent neonatal hypoglycemia, macrosomia and other negative outcomes in offspring of individuals treated with glyburide than insulin and metformin. 

The findings on glyburide, which had been commonly used for GDM prior to 2015, had a broad effect on practice, leading some clinicians to shift to metformin but the vast majority to shift to insulin, said Feghali in an interview after the meeting.

She estimates, based on professional conversations, that metformin is utilized less than 20% of the time for GDM. “That’s partly a backlash against oral medications and the reaction toward glyburide…People became a lot more cautious with oral medications [which cross the placenta] than they’d been before,” she said, noting that the increase in insulin options and technology also played a role in the shift.

At the meeting, describing the benefits of insulin for GDM, Feghali said that “there’s quite a short time frame between diagnosis and the ability to make a difference [for neonatal outcomes], and it can be a lot harder to do with medications that work more mildly than insulin.” Insulin, she noted, “allows for more rapid up-titration.”

GDM accounts for the vast majority of diabetes in pregnancy, but the occurrence of severe insulin resistance is growing and prevalence of pregestational T2DM is rapidly increasing, she noted. 

Whether metformin has a role as an add-on to insulin was examined in two recent studies: the MiTy trial published in 2020 and the MOMPOD study published in 2023. The latter also included GDM detected in early pregnancy. 

The MiTy trial randomized 502 women in Canada and Australia to either metformin or placebo added to insulin and found no significant difference in a composite of neonatal morbidity and mortality (pregnancy loss, preterm birth, birth injury, respiratory distress syndrome, neonatal hypoglycemia, and NICU admission > 24 hours). (40% in the metformin group vs 40%, P = .86).

However, patients in the metformin group had slightly better glycemic control, lower insulin requirements, less gestational weight gain, and fewer cesarean births. In addition, metformin-exposed neonates had lower birthweights (by just over 218 g), were less likely to be extremely large for gestational age (> 97^th percentile), and had less macrosomia and reduced adiposity measures. The proportion of small-for-gestational-age infants was higher. 

The MOMPOD trial of almost 800 pregnant adults across the United States similarly found that the addition of metformin did not significantly reduce a composite of neonatal adverse outcome (71% with metformin vs 74% with placebo; adjusted odds ratio, 0.86). As in MiTy, there was a reduction in large-for-gestational infants and birthweight in the combination group. But unlike MiTy, investigators found no difference in neonatal fat mass or in insulin dosing. 

“There had been a lot of optimism that [combination treatment] would mitigate some of the insulin resistance, reduce the insulin dose potentially, and minimize adverse [neonatal] outcomes,” Feghali said. “I think everyone was a little surprised. To us [the results of the trials] suggested that, especially in individuals who have more advanced hyperglycemia, the addition of metformin won’t be that impactful.”

The studies covered different populations, she noted. More than 50% of patients in the US MOMPOD study identified as Hispanic, compared to a small minority in the MiTy trial, and Hispanic individuals have been shown to have greater insulin resistance and less response to metformin. 

Anna Palatnik, MD, of the Medical College of Wisconsin, Milwaukee, called attention to a large observational study that used two healthcare claims databases in the United States (2000-2020) — one commercial and one public — to identify women with T2DM who were treated with both metformin and insulin prior to pregnancy. Some continued metformin during pregnancy; others (72%) did not. 

The neonatal composite adverse outcome did not differ between the two groups. Of note, she said, metformin continuation was associated in the commercially insured cohort with increased risk of infants being small-for-gestational age. “So there is some signal for high risk of [small for gestational age] in those with type 2 diabetes mellitus exposed to metformin,” Palatnik said. 

Regarding the long-term safety of intrauterine exposure to metformin, there is conflicting data, Palatnik and Feghali said. Overall, offspring of women who participated in early trials of metformin vs insulin for GDM have been shown in follow-up studies to have similar body composition and metabolic outcomes. Researchers want additional data, however, to further investigate previous findings of increased BMI and/or higher adiposity in children who had intrauterine exposure to metformin (in diabetic pregnancies and pregnancies in which mothers took metformin for polycystic ovary syndrome).

Signs that immediate pharmacologic treatment for GDM could be beneficial came from the EMERGE randomized trial published in 2023. Researchers found that early treatment with metformin (in addition to standard advice on lifestyle modification) was not superior to placebo (usual care only) for the primary composite outcome of insulin initiation or a fasting glucose level of 5.1 mmol/L or greater at gestation weeks 32 or 38. 

“But there were some interesting secondary findings,” including longer time to insulin initiation, lower rates of insulin initiation (38% vs 51%), less maternal weight gain, and fewer large-for-gestational age infants (6.5% vs 14.9%), Feghali pointed out. 

“Depending on where you might fall on the spectrum of being for or against oral agents, you might see these secondary findings as being very favorable, or you might consider them as delaying the inevitable,” she said at the meeting. And while the findings were specific to metformin, “maybe the thought here is just to start treatment early.” 

Patients’ Preferences

Research in the last several years supports earlier findings on the need for adjunctive insulin in patients who initiate metformin for GDM. Depending on the study, Feghali said, trials published recently show between 16% and 46% of patients required adjunctive insulin for adequate glycemic control. 

Some research has also shown strong patient preference for metformin due to its oral administration, convenience, and lower treatment burden, Palatnik pointed out. Its lower cost may improve adherence and sustained glucose control, and in low-resource or rural settings, metformin-based protocols can be implemented in primary care or community-level programs without the need for specialized diabetes educators, she noted. 
The DECIDE trial, for which Feghali is a co-investigator, will capture patient experiences and preferences. “And I’m hopeful, given the size of the study and the heterogeneity of the population with GDM, we’ll be able to look at patient characteristics and define early responders [to metformin] versus non-responders,” Feghali said. “Right now, I have a hard time at an individual level being able to identify who’s going to respond…It’s still very much a trial-and-error approach with every patient.”

And given that the metabolic state of patients who become pregnant has worsened overall recently, “it’s become very hard to project the results of studies done in the early 2000s to our patient population right now,” she noted.

Feghali and Palatnik reported having no relevant financial disclosures.