One-step or 2-step testing for gestational diabetes: which is better?

Abstract

In the United States, the common approach to detecting gestational diabetes mellitus is the 2-step protocol recommended by the American College of Obstetricians and Gynecologists. A 50 g, 1-hour glucose challenge at 24 to 28 weeks' gestation is followed by a 100 g, 3-hour oral glucose tolerance test when a screening test threshold is exceeded. Notably, 2 or more elevated values diagnose gestational diabetes mellitus. The 2-step screening test is administered without regard to the time of the last meal, providing convenience by eliminating the requirement for fasting. However, depending upon the cutoff used and population risk factors, approximately 15% to 20% of screened women require the 100 g, 3-hour oral glucose tolerance test. The International Association of Diabetes and Pregnancy Study Groups recommends a protocol of no screening test but rather a diagnostic 75 g, 2-hour oral glucose tolerance test. One or more values above threshold diagnose gestational diabetes mellitus. The 1-step approach requires that women be fasting for the test but does not require a second visit and lasts 2 hours rather than 3. Primarily because of needing only a single elevated value, the 1-step approach identifies 18% to 20% of pregnant women as having gestational diabetes mellitus, 2 to 3 times the rate with the 2-step procedure, but lower than the current United States prediabetes rate of 24% in reproductive aged women. The resources needed for the increase in gestational diabetes mellitus are parallel to the resources needed for the increased prediabetes and diabetes in the nonpregnant population. A recent randomized controlled trial sought to assess the relative population benefits of the above 2 approaches to gestational diabetes mellitus screening and diagnosis. The investigators concluded that there was no significant difference between the 2-step screening protocol and 1-step diagnostic testing protocol in their impact on population adverse short-term pregnancy outcomes. An accompanying editorial concluded that perinatal benefits of the 1-step approach to diagnosing gestational diabetes mellitus "appear to be insufficient to justify the associated patient and healthcare costs of broadening the diagnosis." We raise several concerns about this conclusion. The investigators posited that a 20% improvement in adverse outcomes among the entire pregnancy cohort would be necessary to demonstrate an advantage to the 1-step approach and estimated the sample size based on that presumption, which we believe to be unlikely given the number of cases that would be identified. In addition, 27% of the women randomized to the 1-step protocol underwent 2-step testing; 6% of the study cohort had no testing at all. A subset of women assigned to 2-step testing did not meet the criteria for gestational diabetes mellitus but were treated as such because of elevated fasting plasma glucose levels, presumably contributing to the reduction in adverse outcomes but not to the number of gestational diabetes mellitus identified, increasing the apparent efficacy of the 2-step approach. No consideration was given to long-term benefits for mothers and offspring. All these factors may have contributed to obscuring the benefits of 1-step testing; most importantly, the study was not powered to identify what we understand to be the likely impact of 1-step testing on population health.