Perinatal outcomes in gestational diabetes: A comparison of criteria for diagnosis

OBJECTIVE: The purpose of this study was to compare the perinatal outcomes of women after diagnosis of gestational diabetes by the current American Co llege of Obstetricians and Gynecologists-National Diabetes Data Group recom mendations with outcomes after diagnosis by the American Diabetes Associati on criteria. STUDY DESIGN: We identified records of 242 women who had had the standard 3 -hour oral glucose tolerance test between 1995 and 1999 at the Regional Med ical center in Memphis. Patients were categorized into 1 of 3 groups as fol lows. euglycemic control subjects (n = 69), subjects with gestational diabe tes diagnosed by the National Diabetes Data Group criteria (n = 130), and s ubjects with gestational diabetes diagnosed by the American Diabetes Associ ation criteria (n = 43). Maternal and infant charts were reviewed. Primary outcomes included frequency of cesarean delivery, preeclampsia, and macroso mia. In univariate analysis the chi (2) test was used to compare group diff erences, and in multivariate analysis we used stepwise logistic regression and controlled for confounding factors. RESULTS: No differences existed among the 3 groups regarding maternal race, body mass index, history of preeclampsia, or family history of diabetes. T he frequency of overall cesarean delivery, of cesarean delivery for macroso mia or arrest disorder, of preeclampsia, and of macrosomia did not differ s ignificantly among the 3 groups. Neonatal hypoglycemia was more frequent in the groups with a diagnosis by the American Diabetes Association criteria (23.3%) and by the National Diabetes Data Group criteria (16.2%) than in th e control subjects (7.2%), reaching near significance (P = .057). in the mu ltivariate analysis, cesarean delivery for macrosomia or an arrest disorder correlated negatively with parity and positively with body mass index. Pre -eclampsia was associated with African American race and body mass index; m acrosomia correlated with a history of macrosomia and familial diabetes. Ne onatal hypoglycemia was more common in the American Diabetes Association gr oup (odds ratio, 2.45; 95% confidence interval, 1.004-5.97) and in the insu lin-requiring National Diabetes Data Group category (odds ratio, 3.71; 95% confidence interval, 1.20-11.44). CONCLUSION: The benefits of defining an additional high-risk population of women with gestational diabetes by the American Diabetes Association criter ia are unclear Further large-scale prospective clinical trials are required .