Validity of Fasting Plasma Glucose in Reference to HbA1c for the Detection of Diabetes in an Overweight and Obese Population in Sri Lanka

Objective: The validity of fasting plasma glucose (FPG) measurements for detecting diabetes was evaluated in reference to HbA1c among a 35-64-year-old overweight and obese population. Methods: Data were collected from 1200 adults aged 35-64 with a body mass index (BMI) of 25 kg/m(2) or above using a questionnaire-based interview, basic clinical measurements, and biochemical analyses. Sensitivity, specificity, and ROC analyses were performed on 1121 persons for FPG (the index test) and HbA1c (the reference standard). HbA1c >= 6.5%was considered to indicate diabetes. Results: Approximately 64% were female, 73% were employed, 31% had BMI >= 30 kg/m(2), and 70% lived in urban settings. The mean age (SD) was 47.8 (+/- 7.5) years. The prevalence of diabetes was 25.7% according to HbA1c >= 6.5%, and 44% of them were undiagnosed. A significant correlation was found between FPG values and HbA1c (r = 0.863; P < .001). At the FPG cutoff point of 126 mg/ dL, the sensitivity was 60% (95% CI: 53.6, 65.5), and the specificity was 99.8% (95% CI: 99.4, 100.1), with high agreement (kappa = 0.69; P <.001).The positive and negative predictive values were 98.7% (96.9, 100.5) and 89.3% (87.4, 91.3), respectively. At FPG 100 mg/dL, the sensitivity increased to 82.0% (77.2, 86.7), and specificity reduced to 88.5% (86.2, 90.7), with a marginal change in kappa (0.66; P < .001). Receiver operating characteristic analysis found an area of 0.91 (P < .001) under the curve at this optimum cutoff point. Conclusion: A lower threshold of FPG (>= 100 mg/dL) may achieve a maximum yield of diabetes during the screening of overweight or obese persons in the age group 35-64 years.