EDITOR—The DECODE Study Group's paper on the implications of the American Diabetes Association's (ADA) recent changes in diagnostic criteria for diabetes mellitus raises several issues.1

The World Health Organisation's definition of diabetes that was used in the American report included both fasting glucose concentrations and concentrations 2 hours after a glucose load.2 In the DECODE paper, however, WHO criteria were defined by theglucose concentration after challenge alone. This difference in definition of the WHO criteria may help to explain why changes in prevalence estimates of diabetes presented in the DECODE paper differ from findings in the American report. A reanalysis of data using consistent definitions would be desirable.

More important questions stem from the DECODE Study Group's finding that the prevalence of diabetes in older adults will increase substantially with the ADA criteria. We believe that the opposite is likely to happen. Most people who have fasting glucose concentrations that are diagnostic ofdiabetes will also have high glucose concentrations after a glucose load. However, glucose intolerance is a common feature of advancing age and often exists in subjects without a high fasting glucose concentration. As the ADA criteria emphasise fasting glucose concentration over glucose intolerance, fewer older people are likely to be classed as having diabetes overall. We therefore agreethat the ADA criteria will potentially bias prevalence estimates of diabetes relevant to age, but these estimates will underestimate rather than overestimate the prevalence of diabetes in older adults.

The question of whether raised glucose concentrations after a glucose load in the absence of a raised fasting glucose concentration should be considered to be diabetes has been controversial. Arecent study showed that isolated hyperglycaemia after challenge in women aged 50-89 with normal fasting glucose concentrations was associated with a 2.9-fold increase in risk of cardiovascular disease over 7 years …