Oral Antihyperglycemic Agents and Diabetes in Pregnancy

There were no significant differences between the groups in the rate of infants who were large for gestational age or who had macrosomia, a ponderal index greater than 2.85, lung complications, hypoglycemia, or fetal anomalies.

We also tested the cord serum at delivery and found similar cord-serum insulin concentrations in the two groups. Glyburide was not detected in the cord serum above the level of 10 ng/mL.

Since 2000, more than 20 studies (4 of them randomized) have been published that show similar success rates with glyburide and insulin in achieving good glycemic control in gestational diabetes as well as similar perinatal outcomes. Most of the studies have been small and not randomized. Oftentimes, however, well-designed retrospective or case-control studies can be just as reliable. In this case, the studies collectively provide a solid basis for evaluation.

In a meta-analysis published last year, investigators concluded that the studies suggest there are no increased perinatal risks with glyburide compared with insulin for the treatment of GDM (Ann. Pharmacotherapy 2008;42:483-90).

Nine studies met the inclusion criteria for the analysis, which totaled 745 glyburide-exposed pregnancies and 637 insulin-exposed pregnancies. Women were typically treated starting at 24 weeks of gestation.

The use of glyburide was not associated, the investigators said, with risk of macrosomia, differences in birth weight, rate of large-for-gestational-age births, differences in gestational age at birth, ICU admission, or risk of neonatal hypoglycemia.

Metformin as an Option

Glyburide and metformin have different mechanisms of action. Glyburide works on the pancreas to stimulate insulin secretion. Metformin, which belongs to the class of oral antihyperglycemic agents known as the biguanides, lowers glucose levels by decreasing hepatic glucose production and decreasing peripheral insulin resistance.

Some have suggested that because metformin does not stimulate insulin secretion, it is less likely than glyburide to cause hypoglycemia and may be the preferable choice for treating diabetes in pregnancy.

While we have not directly compared metformin and glyburide in this regard, our data and data from other studies demonstrate that the rate of maternal hypoglycemia is significantly higher with insulin than with glyburide therapy. In one study using continuous blood glucose measurements, we showed that the maternal rate of hypoglycemic episodes was five times higher in insulin-treated patients than in glyburide-treated patients (Obstet. Gynecol. 2004;104:88-93).

Earlier findings suggesting the opposite—that glyburide is more likely to cause hypoglycemia than is insulin therapy—were from studies in much older, nonpregnant women. Diabetes in patients who are in their 50s through their 80s cannot be compared, in general, to the less severe disease in younger women of reproductive age.

Metformin, like glyburide, has been shown in numerous studies to have no adverse effect in pregnancy in terms of anomalies. The first large randomized, controlled trial to assess the safety and efficacy of metformin versus insulin—published last year—found similar efficacy in achieving target levels of glucose control and no difference in perinatal outcomes among 751 women randomized to one of the two groups (N. Engl. J. Med. 2008;358:2003-15).

Like glyburide, metformin is a class B drug. Because metformin crosses the placenta, physicians must take this into consideration when deciding which oral antihyperglycemic agent to choose. Even if a drug crosses the placenta, however, it should not automatically be considered contraindicated for use in pregnancy because the majority of drugs used in pregnancy cross the placenta without adverse effect to the fetus.

Also of possible concern is the fact that the rate of large-for-gestational-age infants in the New England Journal of Medicine (NEJM) metformin-versus-insulin study was twice the rate of large-for-gestational-age infants in our NEJM study comparing glyburide with insulin. This suggests that the rate of success in achieving glycemic control in pregnancy may be lower with metformin than with glyburide.

We need other studies, however, that directly compare glyburide with metformin (rather than comparing each with insulin), and the resultant perinatal outcomes and glycemic control, in order to address this issue.

Metformin is a popular drug for the treatment of polycystic ovary syndrome (PCOS), which presents the question of whether patients on metformin for PCOS should conceive while on the drug, or halt the drug if they unexpectedly conceive.

The answers in these cases call for individual judgment. In my opinion, metformin is a drug that can be used in pregnancy, as long as one keeps in the back of one's mind the fact that it does cross the placenta. One must also consider that although recent retrospective and prospective trials have shown no adverse effects of metformin in terms of anomalies, no published randomized study has evaluated pregnancy outcomes when patients were treated with the drug from preconception throughout gestation.