Use of the 1997 American Diabetes Association diagnostic criteria for diabetes in a Hong Kong Chinese population.

Abstract

Recently, the American Diabetes Association (ADA) has proposed revised diagnostic criteria for diabetes. Lowering of the fasting plasma glucose (FPG) cutoff value is intended to reduce the discrepancy with the 2-h plasma glucose (PG) cutoff value and to encourage the use of FPG. We have applied these new criteria to data collected from a population-based prevalence survey in Hong Kong Chinese subjects of working age. The results of 1,513 oral glucose tolerance tests (OGTTs) from a previously published prevalence survey of glucose intolerance and cardiovascular risk factors in a Hong Kong Chinese working population were reexamined using the new criteria. Of the 1,513 subjects, 27 had a known history of diabetes. Of the remaining 1,486 subjects, 228 were also selected randomly for a second OGTT without prior knowledge of the result of the first test. After exclusion of the 27 subjects with a known history of diabetes, the crude prevalence of diabetes was 2.83% (n = 42) when the World Health Organization's (WHO) criteria were applied. When the criterion of FPG > or = 7.0 mmol/l was used, as recommended by the ADA, the prevalence of diabetes was 1.41% (n = 21). Twenty-nine subjects (1.95%) with FPG < 7.0 mmol/l had a 2-h PG > or = 11.1 mmol/l. Eight subjects (0.53%), previously without a diagnosis of diabetes according to the WHO criteria (FPG < 7.8 mmol/l and 2-h PG < 11.1 mmol/l), had FPG between 7.0 and 7.8 mmol/l and were classified as having diabetes by the ADA criteria. This classification gave a net change of -1.42% in the prevalence of diabetes between the use of FPG > or = 7.0 mmol/l alone and the use of WHO criteria. Among the 1,486 subjects with no known history of diabetes, those classified as having diabetes according to the ADA FPG criterion alone had higher HbA1c and fructosamine levels than diabetic subjects defined by the WHO criteria. Of the 228 subjects for whom two FPG measurements were available, those who had consistent definitions (diabetes, impaired fasting glucose, normal fasting glucose) on both occasions were considered to have reproducible tests, giving an overall reproducibility of 90.8% (207 of 228). Compared with the WHO criteria, the use of FPG to diagnose diabetes, as recommended by the ADA, was a more reproducible test and identified those subjects who had a greater degree of hyperglycemia. Although lowering of the cutoff value from 7.8 to 7.0 mmol/l increased the number of diagnoses among subjects with low FPG, the omission of the 2-h PG would lead to fewer subjects having their diabetes diagnosed.