This was a prospective cohort that included 450 women seeking prenatal care at the Mexican Social Security Institute (IMSS). Subjects were ≥19 years old, in early pregnancy, without previous dysglycemia or hypertension, and with fasting glycemia <126 mg/dL. Insulin and HOMA-IR were measured. At 24–36 weeks of gestation, a 3-h 100 g oral glucose tolerance test (OGTT) was performed, applying the modified Carpenter-Coustan criteria. Multiple logistic regression models including an offset term as indicator of time in risk allowed us to estimate the risk of developing glucose intolerance (GI), GD, or hyperglycemia 1-h after the glucose load (HG-1). Areas under the ROC curve allowed comparison of the models’ predictive capability.
Incidence of dysglycemic events was 20.7 % . The risk of GD was higher for the highest glycemia (RR = 4.1, 95 % CI 1.6–10.6), insulin (RR = 4.1, 95 % CI 1.2–14.2), and HOMA-IR (RR = 6.4, 95 % CI 1.9–21.9). A higher risk of GI was found for the highest values of glycemia (RR = 2.6, 95 % CI 1.3–5.3). Higher glycemia (RR = 3.9, 95 % CI 1.6–9.2), insulin (RR = 2.8, 95 % CI 1.0–7.5) and HOMA-IR (RR = 3.8, 95 % CI 1.4–10.2) were associated with HG-1. Areas under the ROC curve for adjusted models with glycemia, insulin, or HOMA-IR were 0.749, 0.715, and 0.747, respectively (p = 0.4).
Fasting glycemia was the best adjusted predictor of GD in early pregnancy, having equal predictive capability compared to insulin and HOMA-IR.