A Cluster Randomized Noninferiority Field Trial of Gestational Diabetes Mellitus Screening

Abstract

Context: Although it is well-acknowledged that gestational diabetes mellitus (GDM) is associated with the increased risks of adverse pregnancy outcomes, the optimal strategy for screening and diagnosis of GDM is still a matter of debate.

Objective: This study was conducted to demonstrate the noninferiority of less strict GDM screening criteria compared with the strict International Association of the Diabetes and Pregnancy Study Groups (IADPSG) criteria with respect to maternal and neonatal outcomes.

Methods: A cluster randomized noninferiority field trial was conducted on 35 528 pregnant women; they were scheduled to have 2 phases of GDM screening based on 5 different prespecified protocols including fasting plasma glucose in the first trimester with threshold of 5.1 mmol/L (92 mg/dL) (protocols A, D) or 5.6 mmol/L (100 mg/dL) (protocols B, C, E) and either a 1-step (GDM is defined if one of the plasma glucose values is exceeded [protocol A and C] or 2 or more exceeded values are needed [protocol B]) or 2-step approach (protocols D, E) in the second trimester. Guidelines for treatment of GDM were consistent with all protocols. Primary outcomes of the study were the prevalence of macrosomia and primary cesarean section (CS). The null hypothesis that less strict protocols are inferior to protocol A (IADPSG) was tested with a noninferiority margin effect (odds ratio) of 1.7.

Results: The percentages of pregnant women diagnosed with GDM and assigned to protocols A, B, C, D, and E were 21.9%, 10.5%, 12.1%, 19.4%, and 8.1%, respectively. Intention-to-treat analyses satisfying the noninferiority of the less strict protocols of B, C, D, and E compared with protocol A. However, noninferiority was not shown for primary CS comparing protocol E with A. The odds ratios (95% CI) for macrosomia and CS were: B (1.01 [0.95-1.08]; 0.85 [0.56-1.28], C (1.03 [0.73-1.47]; 1.16 [0.88-1.51]), D (0.89 [0.68-1.17]; 0.94 [0.61-1.44]), and E (1.05 [0.65-1.69]; 1.33 [0.82-2.00]) vs A. There were no statistically significant differences in the adjusted odds of adverse pregnancy outcomes in the 2-step compared with the 1-step screening approaches, considering multiplicity adjustment.

Conclusions: The IADPSG GDM definition significantly increased the prevalence of GDM diagnosis. However, the less strict approaches were not inferior to other criteria in terms of adverse maternal and neonatal outcomes.

Keywords: cluster randomized non-inferiority field trial; gestational diabetes mellitus; screening.