Is fasting glucose sufficient to define diabetes? Epidemiological data from 20 European studies

Abstract

Aims/hypothesis. The World Health Organisation Consultation recommended new diagnostic criteria for diabetes mellitus including: lowering of the diagnostic fasting plasma glucose to 7.0 mmol/l and introduction of a new category: impaired fasting glycaemia. The diagnostic 2-h glucose concentrations for diabetes and for impaired glucose tolerance were unchanged. This study identifies fasting plasma glucose concentrations predicting a diabetic 2-h plasma glucose of 11.1 mmol/l or more, analyses the sensitivity and specificity of different screening strategies for diabetes and describes the cardiovascular risk profile in people with impaired fasting glycaemia. Methods. European population based studies (n = 17) or large, representative samples of employees (n = 3) with both fasting and 2-h post load glucose concentrations following 75-g oral glucose tolerance tests were included (18 918 men and 10 190 women). The Iceland study (8881 men and 9407 women) is presented separately as a 50-g glucose load was used. Results. The fasting plasma glucose predicting a 2-h plasma glucose of 11.1 mmol/l or more with optimal sensitivity and specificity was a) 5.8 mmol/l in women and 6.4 mmol/l in men; b) independent of age; c) increased with obesity. Fasting plasma glucose of 7.0/7.8 mmol/l or more predicted a diabetic 2-h plasma glucose with sensitivities of 49.0/29.8 % and specificities of 98.2/99.7 %, respectively. Conclusion/interpretation. If fasting glucose is used alone, the 31 % of diabetic subjects with a non-diabetic fasting glucose but a diabetic 2-h glucose, will not be diagnosed; impaired fasting glycaemia and impaired glucose tolerance do not identify the same people; the risk profile of people with impaired fasting glycaemia depends on 2-h glucose concentrations. Obesity is the main confounder in the association between fasting and 2-h glucose. [Diabetologia (1999) 42: 647–654]