A comparative study of the different diagnostic criteria of gestational diabetes mellitus and its incidence.

Background: High prevalence of diabetes and genetic predisposition to metabolic syndrome among Indians places Indian women at risk to develop gestational diabetes mellitus (GDM) and its complications. Literature defines multiple criteria for GDM. This prospective study compares available diagnostic criteria for GDM in Indian women and their correlation with perinatal morbidity.

Method: Nine hundred and forty-eight consecutive voluntary nondiabetic pregnant women were recruited for the study. Seven hundred and twenty-three of these (mean age 23.45 years; 75.7% < 25 years) who reported for the follow-up were screened for GDM at 24-28 weeks gestation by American College of Obstetrics and Gynaecology (ACOG) guidelines and World Health Organization (WHO) criteria. Glycated haemoglobin (HbA1c) and fasting and two-hours postglucose plasma insulin levels were also analysed. Pregnancy outcome was known for 291 of these. Concordance of risk factors and perinatal complications was analysed with respect to GDM.

Results: Prevalence of GDM at 24-28 weeks gestation was found to be 4.8% by WHO criteria, 6.36% by Carpenter and Coustan's criteria, and 3.5% by O'Sullivan's criteria. Prevalence was marginally higher in women of higher age, having past history of abortion or family history of diabetes mellitus (DM) (P > 0.05). None of these women had HbA1c > 6%. Relative risk of abnormal delivery (pregnancy outcome) was 1.93, 1.39, and 1.17 in women with GDM by O'Sullivan's, WHO, and Carpenter's criteria, respectively (P > 0.05). Abnormal deliveries were marginally higher in women with high postglucose load insulin levels. Mean weight of the newborns was essentially the same in GDM and nonGDM women by any of the criteria. One-hour and two-hours postglucose values were more sensitive in diagnosing GDM by O'Sullivan's criteria while fasting plasma glucose value had the poorest specificity with 2.5% of nonGDM women having values above the cut-off. Modifications of these criteria did not im-prove their predictive value for abnormal delivery over that of O'Sullivan's criteria.

Conclusion: Prevalence of GDM and abnormal delivery in women < 35 years of age is low. Therefore, global screening for GDM may not be very useful in women < 25 years of age unless family history of DM or past history of abortion is present. Existing evidence is inadequate to justify the switchover from O'Sullivan's criteria for diagnosis of GDM.