A 2016 systematic review and meta-analysis of short- and long-term outcomes of metformin, compared with insulin, found that metformin did not increase preterm delivery (). And while the 2015 BMJ meta-analysis found that metformin was associated with higher rates of preterm birth (RR, 1.50), the increased risk “was all driven by the Rowan study,” Dr. Landon said. The 2015 meta-analysis also found that metformin was associated with less maternal weight gain and fewer infants who were large for gestational age.
Metformin is also tainted by high rates of failure in GDM. In the 2008 Rowan study, 46% of patients on metformin failed to achieve glycemic control. “But this is a classic half-full, half-empty [phenomena],” Dr. Landon said. “Some people say this isn’t good, but on the other hand, 54% avoided insulin.”
Indeed, the Society of Maternal-Fetal Medicine (SMFM), in its 2018on the pharmacologic treatment of GDM, said that oral hypoglycemic agents that are used as monotherapy work in “more than half” of GDM pregnancies. The need for adjunctive insulin to achieve glycemic control ranges between 26% and 46% for women using metformin, and 4% and 16% for women using glyburide, it says.
In the society’s view, recent meta-analyses and systemic reviews “support the efficacy and safety of oral agents,” and “although concerns have been raised for more frequent adverse neonatal outcomes with glyburide, including macrosomia and hypoglycemia, the evidence of benefit of one oral agent over the other remains limited.”
The society says that the difference between its statement and the ACOG recommendations is “based on the values placed by different experts and providers on the available evidence,” and it adds that more long-term data are needed.
But as Dr. Landon said, the SMFM is “a little more forgiving” in its interpretation of a limited body of literature. And clinicians, in the meantime, have to navigate the controversy. “The professional organizations don’t make it easy for [us],” he said. At this point, “insulin does not cross the placenta, and the oral agents do cross it. Informed consent is absolutely necessary when choosing oral agents for treating GDM.”
Of greater concern than neonatal outcomes are the potential long-term issues for offspring, Dr. Landon said. On the one hand, it is theorized that metformin may protect beta-cell function in offspring and thereby reduce the cross-generational effects of obesity and type 2 diabetes. On the other hand, it is theorized that the drug may cause a decrease in cell-cycle proliferation, which could have “unknown fetal programming effects,” and it may inhibit the mTOR signaling pathway, thus restricting the transport of glucose and amino acids across the placenta, he said. (Findings from in vitro research have suggested that glyburide treatment in GDM might be associated with enhanced transport across the placenta, he noted.)
Long-term follow-up studies of offspring are “clearly needed,” Dr. Landon said. At this point, in regard to long-term safety, he and other experts are concerned primarily about the potential for obesity and metabolic dysfunction in offspring who are exposed to metformin in utero. They are watching follow-up from Dr. Rowan’s MiG trial, as well as elsewhere in the literature, on metformin-exposed offspring from mothers with polycystic ovary syndrome.
A follow-up analysis of offspring from the MiG trial found that children of women with GDM who were exposed to metformin had larger measures of subcutaneous fat at age 2 years, compared with children of mothers treated with insulin alone, but that overall body fat was the same, Dr. Landon noted. The investigators postulated that these children may have less visceral fat and a more favorable pattern of fat distribution ().
A recently published follow-up analysis of two randomized, controlled trials of women with polycystic ovary syndrome is cause for more concern, he said. That analysis showed that offspring exposed to metformin in utero had a higher body mass index and an increased prevalence of obesity or overweight at age 4 years, compared with placebo groups ().
That analysis of metformin-exposed offspring in the context of polycystic ovary syndrome was published after the SMFM statement, as was another follow-up analysis of MiG trial offspring – this one, at ages 7-9 years – that showed an increase in weight, size, and fat mass in one of two subsets analyzed, despite no difference in large-for-gestational age rates between the metformin- and insulin-exposed offspring ().
In 2018, a group of 17 prominent diabetes and maternal-fetal medicine researchers cited these findings in a response to the SMFM statement and cautioned against the widespread adoption of metformin use during pregnancy, writing that, based on “both pharmacologic and randomized trial evidence that metformin may create an atypical intrauterine environment ... we believe it is premature to embrace metformin as equivalent to insulin or as superior to glyburide, and that patients should be counseled on the limited long-term safety data and potential for adverse childhood metabolic effects” ().