SAN FRANCISCO — A growing body of evidence suggests that the oral agent glyburide may not be as safe as injected insulin to treat gestational diabetes, and that perceived barriers to women using insulin are unsubstantiated.
“The standard of care was insulin. Then everybody changed to glyburide based on an underpowered study” of 404 women with gestational diabetes who were randomized to glyburide or insulin therapy, Dr. Aaron B. Caughey said at a conference on antepartum and intrapartum management sponsored by the University of California, San Francisco.
The study reported similar glycemic control and neonatal outcomes between groups (N. Engl. J. Med. 2000;343:1134–8). “A lot of people that use glyburide base it on this one prospective, randomized trial” that was too small and showed some worrisome trends, said Dr. Caughey, medical director of the Diabetes and Pregnancy Program at the university.
A closer look at the results reveals multiple trends toward worse neonatal outcomes with glyburide, although none are statistically significant, he said. In the glyburide group, 7% of babies had birth weights above 4,000 grams, compared with 4% of the insulin group. Lung complications were reported in 8% of the glyburide group and in 6% on insulin. Hypoglycemia occurred in 9% on glyburide and 6% on placebo. Hyperbilirubinemia rates were 6% on glyburide and 4% on placebo.
More recently, a retrospective cohort study of 584 women with gestational diabetes in the Kaiser Permanente system found significantly worse rates of preeclampsia (12%) and phototherapy for hyperbilirubinemia (9%) in women treated with glyburide, compared with those given insulin (6% and 5%, respectively). That study also found a nonsignificant trend toward a higher rate of birth injury with glyburide (3%) than with insulin (1%), all of which was “concerning,” he said (Am. J. Obstet. Gynecol. 2005;193:118–24).
Analyses of statewide data from the California Diabetes and Pregnancy Program (known as Sweet Success) in 2007 by Dr. Caughey's associates at the university also found some significantly worse outcomes in women given oral agents compared with insulin for gestational diabetes. They found a 35% higher risk for birth weights above 4,000 g, a 40% higher risk for admission to the neonatal intensive care unit, and a 52% higher risk for preterm delivery before 34 weeks' gestation in women taking oral agents after adjusting for the effects of maternal age, ethnicity, parity, education, gestational age at delivery or at diagnosis of gestational diabetes, body mass index, and gestational weight gain.
Among women diagnosed with gestational diabetes early in pregnancy—at less than 24 weeks' gestation—the increased risk with oral agents was even more pronounced, including more than a threefold higher risk for intrauterine fetal demise compared with the insulin group.
Some clinicians say they prefer to treat gestational diabetes with glyburide because they believe that patients with little formal education cannot understand how to use insulin, or that patients whose primary language is not English will have difficulty, noted Dr. Caughey, who said he had no conflicts of interest related to this topic. When they stratified the Sweet Success data to profile women who had less than 9 years of education or whose primary language was Spanish, those on oral agents still had significantly worse outcomes.
“In my practice, we have one person a year [who will] not be able to use insulin and [who has] to use an oral agent. It's pretty rare.” At UCSF, gestational diabetes management begins with nutritional counseling and a prescription to take a short walk after each meal. If borderline or mild glucose elevations persist, they offer the patient insulin and will consider the alternative of oral metformin, but only in patients who are unlikely to have pre-existing diabetes mellitus, who have been diagnosed with gestational diabetes at 26–32 weeks' gestation, and who have been counseled about an increased preterm delivery risk associated with metformin.