Doubts about the true effectiveness of metformin for gestational diabetes

Abstract

TO THE EDITORS: We read with interest the recent article by Picón-César et al 1 concerning patients with gestational diabetes mellitus. The authors reported that metformin was not inferior to insulin therapy in controlling glycemia and reducing the risk of hypoglycemia and showed no difference in terms of obstetrical and perinatal outcomes. Although the results are interesting, there are several concerns that we would like to highlight.Per strict criteria, gestational diabetes mellitus is only diagnosed between 24 and 28 weeks of gestation; however, some patients were diagnosed before this time frame and consequently correspond to the category of diabetes mellitus and pregnancy in which the management and prognosis are quite different. We understand that local guidelines or protocols might be different; however, making a subgroup analysis of patients meeting this criterion would have been valuable for the applicability of the results and would have reduced the risk of misclassification bias.In addition, the sample size was calculated on the basis of a noninferiority margin of the estimated difference in the incidence of macrosomia in newborns. Herein, using an outcome unrelated to glycemic control for this calculation may have limited the statistical power of the analysis.Regarding the data analysis, the need of initiating insulin in patients unable to achieve glycemic control with metformin alone may have overestimated the benefits of this intervention. Thus, the addition of a per-protocol analysis, 2 giving more conservative estimates of effect, could have been useful in exploring the influence of the dual therapy on the results.