A comparison of perinatal outcomes associated with gestational diabetes mellitus testing practices in British Columbia: A population-based retrospective cohort study.

Abstract

This study sought to compare one-step versus two-step testing approaches for gestational diabetes mellitus (GDM) and to investigate the associations between testing approach, degree of glucose impairment, and perinatal outcomes. A retrospective population-based cohort study was conducted by combining BC's Perinatal Data Registry with laboratory and billing information from 2010 to 2014. Pregnancy characteristics were compared by GDM testing approach. Logistic regression was conducted to determine the association between testing approach, degree of glucose impairment, and outcomes. Approximately 17% of pregnant individuals were diagnosed with GDM using the one-step test, compared to 6% using the two-step test. The odds ratios of adverse outcomes were below 1.75 for GDM pregnancies regardless of testing approach used (compared to the group with negative results on the two-step test). A dose-dependent trend was observed between increasing glucose intolerance and odds of preterm birth. The odds of large for gestational age infants (LGA) and shoulder dystocia were significantly higher when all 75g test values were within one standard deviation below one-step diagnostic thresholds (adjOR 1.94[1.73-2.17] and 1.85[1.55-2.21], respectively). The frequency of GDM was three times higher with the use of the one-step test versus the two-step test. Abnormal results on the two-step test are associated with preterm birth at an odds ratio below 1.75. Pregnant individuals with one-step test results just below diagnostic criteria may be at greater odds for LGA. The benefits of more stringent testing practices need to be weighed against the impact of additional GDM diagnoses.