Safety of glyburide for gestational diabetes: a meta-analysis of pregnancy outcomes.

BACKGROUND

Gestational diabetes mellitus affects approximately 4% of all pregnancies. As with other oral hypoglycemics, the use of glyburide in pregnancy has been limited by fears of neonatal hypoglycemia following fetal exposure. Recent experimental and clinical studies have suggested, however, that the drug is not detectable in umbilical blood.

OBJECTIVE

To determine the safety of glyburide use in pregnancy in the treatment of gestational diabetes compared with insulin therapy by analyzing all available human studies.

METHODS

We conducted a systematic review and meta-analysis of all randomized and cohort studies that reported on the perinatal complications among women with gestational diabetes who received glyburide versus insulin. MEDLINE, EMBASE, and biological abstracts using BIOSIS previews were searched from inception of these databases through October 2006. The following outcomes were included: macrosomia, birthweight, gestational age, neonatal hypoglycemia, and intensive care unit (ICU) admissions. Odds ratios were calculated by the random effects method using Cochrane's Review Manager version 4.3.

RESULTS

Nine studies met the inclusion criteria, including a total of 745 glyburideexposed pregnancies and 637 insulin-exposed pregnancies, with each adverse perinatal outcome reported by 4-7 studies. The use of glyburide was not associated with risk of macrosomia (OR 1.07; 95% CI 0.78 to 1.47), differences in birth weight (weighted mean difference [WMD] OR 20.46 g; 95% CI -34.90 to 75.82), rate of large for gestational age (OR 1.04; 95% CI 0.75 to 1.43), differences in gestational age at birth (WMD 0.02 wk; 95% CI -0.23 to 0.26), ICU admission (OR 0.95; 95% CI 0.43 to 2.09), or increased risk of neonatal hypoglycemia (OR 1.24; 95% CI 0.91 to 1.69).

CONCLUSIONS

The data shown here do not suggest increased perinatal risks with glyburide. The effectiveness and safety of glyburide require further evaluation, as most studies to date were not randomized.