Screening for diabetes mellitus--a two-step approach in individuals with impaired fasting glucose improves detection of those at risk of complications.

AIMS: To compare the new American Diabetes Association (ADA) fasting plasma glucose (FPG) criteria to the 1985 World Health Organization (WHO) 2-h post glucose (2hPG) criteria when used for screening of those with no prior history of diabetes mellitus.

METHODS: The study included 3,407 subjects without a history of diabetes in whom both FPG and 2hPG were available from the 1992 Singapore National Health Survey. The agreement (kappa) between FPG and 2hPG for the diagnosis of DM was assessed. The optimal cut-off of FPG for the detection of individuals with 2hPG > or = 11.1 mmol/l was determined by receiver-operating characteristics analysis.

RESULTS: The prevalence of diabetes diagnosed by FPG alone was 7.3% compared to 8.4% diagnosed by 2hPG. The prevalence of impaired fasting glucose was 8.0%. FPG and 2hPG showed moderate agreement (kappa = 0.646, 95% confidence interval 0.584-0.708). Age, ethnic group and obesity did not affect the degree of agreement. Of those with 2hPG > or = 11.1 mmol/l, 40.8% had FPG in the non-diabetic range while 24.8% of those with FG > or = 7.0 mmol/l had 2hPG in the non-diabetic range. The optimal FPG for the detection of 2hPG > or =11.1 mmol/l was 6.1 mmol/l. Oral glucose tolerance tests (OGTT) in those with 6.0 mmol/ < FPG < 7.0 mmol/l resulted in the diagnosis of diabetes in 90.7% of individuals at risk of microvascular complications.

CONCLUSIONS: FPG provides a simple screening test for diabetes, which shows moderate agreement with the 2hPG. A two-step strategy of OGTT in those with impaired fasting glucose improves the detection of at-risk individuals. However, diabetes should not be diagnosed on a single test. The test should be repeated on another day if an individual tests positive for diabetes.