Comparison of ADA and WHO Criteria in Detecting Pattern of Glucose Disorders in a Population-Based Study: Tehran Lipid and Glucose Study

Many studies have been performed to compare WHO and ADA criteria for detection of diabetic patients. This study aims to compare these two criteria in a community-based epidemiological survey in an urban population of Tehran, Iran. Subjects were chosen from among 15005 urban individuals, 3 years old and over, selected by cluster random sampling in the cross-sectional phase of a longitudinal study conducted in the east of Tehran; there were 3870 men and 5359 women aged 20 years and over. Those with known diabetes were excluded. Blood samples were taken after 12-14 hours overnight fast and 2 hours post 75gr glucose taken orally. Prevalence of glucose tolerance categories and the level of agreement (ĸ statistic) were obtained using WHO and ADA criteria. Based on WHO criteria 6.0% (0.95 CI, 5.5-6.5) had type-2 diabetes and 13.0% (12.3-13.7) had IGT (Impaired glucose tolerance). Using ADA, 3.3% (2.9-3.7) had type-2 diabetes and 4.8% (4.4-5.2) had IFG (Impaired fasting glucose); p<0.05 and <0.001 for diabetes and IGT, respectively and 7355 (84%) had concordance with both criteria. Among 7105 normal subjects classified according to WHO criteria, 153 (2.2%) had IFG or diabetes by ADA, whereas from 8068 normal subjects according to ADA criteria, 1116 (13.8%) had IGT or diabetes based on WHO criteria. The level of agreement (ĸ statistic) between the two criteria was 35% (p<0.001). Sensitivity and specificity of ADA criteria were 45.5 and 100%, respectively, considering WHO as the gold standard. Our data shows a low level of agreement between WHO and ADA diagnostic criteria for detection of diabetes. Patients with unknown diabetes, glucose disorders are detected more frequently using WHO criteria.