Recently introduced criteria for diagnosing diabetes may not be robust enough, according to a two new studies.

In 1997, the American Diabetes Association (ADA) published new criteria for the diagnosis of diabetes which primarily rely on fasting blood glucose concentration. This new criteria differs from the World Health Organisation (WHO) criteria, which also uses the oral glucose tolerance test.

A major study in Finland, dubbed DECODE, compared the two sets of criteria in more than 25,000 men and women, aged 30 and older. Researchers found that the ADA's recommended test for fasting glucose concentrations did not identify individuals at increased risk of death associated with diabetes.

Furthermore, the different sets of criteria did not identify the same groups of individuals. Only 46 per cent of those identified as diabetic according to the ADA met the WHO criteria for diabetes (Lancet, 1999; 354: 617-21).

A smaller study conducted at Emory University School of Medicine in Atlanta, Georgia, looked at the ability of the two guildelines to detect cardiovascular disease the main complication of diabetes in the elderly. Of more than 4000 participants aged 65 and older, only 22.3 were diagnosed as diabetic on the basis of the new ADA criteria, compared to more than double that with the WHO criteria.

Those diagnosed by the ADA criteria also suffered greater illness over the eight year study period.

If the purpose of screening for diabetes is to identify the maximum number of individuals at risk of heart disease and other types of illness, then it appears that the system needs a major re think (Lancet, 1999; 354: 622-5).