Regular Review Impaired glucose tolerance

The classification of abnormalities of glucose tolerance introduced in 1979 by the National Diabetes Data Group and agreed by the World Health Organisation included the category of impaired glucose tolerance.' 2 The diagnosis depends on the blood glucose concentration two hours after a glucose load being above normal (6-7 mmol/l) and below the new diabetic value (10 mmol/l). This category satisfied a need for defining a level of glucose intolerance that was not clearly normal but that was also not sufficiently severe to predict microvascular disease, particularly retinopathy, in prospective studies. As a group, people with impaired glucose tolerance have raised concentrations of insulin both when fasting and after a glucose load3`5 and show insulin resistance when investigated with glucose clamp techniques.46 Several studies of the natural course in people with impaired glucose tolerance have shown that they are at increased risk of developing diabetes (table I),7-2' but there is less consensus about the excess risk of cardiovascular disease.22-27 Recently Saad et al reported an extremely high incidence of non-insulin dependent diabetes in Pima Indians with impaired glucose tolerance,'9 the rate of deterioration to diabetes of 5-6% a year over 10 years exceeding the rates of 1F5-4% reported in most studies in other populations. 10-12 1517 In a separate paper the same authors show that even "transient" impaired glucose tolerance is associated with an increased risk of "deterioration to diabetes."20 Other recent studies have, however, cast doubt on the concept of impaired glucose tolerance, both because of its ephemeral nature (table II)28-30 and as a consequence of doubts about whether it warrants categorisation as a separate entity.3' Stern et al analysed the concept of impaired glucose tolerance, suggesting that the category represented a heterogeneous group of people.32 They used the model of a population with a bimodal distribution of two hour blood glucose concentrations to propose that the category of impaired glucose tolerance will contain some people in the upper tail of normal glucose tolerance ("impaired glucose tolerance normals"), some in the lower tail of diabetic patients ("false negative diabetics"), and some who are truly in the impaired glucose tolerance category. If the distribution oftwo hour blood glucose concentration is bimodal Stern et al suggested that for the nadir to remain apparent few people must have true impaired glucose tolerance, and, therefore, the rate of deterioration through the impaired glucose tolerance category must be fairly rapid. For this reason they defined the third category as "impaired glucose tolerance in transition." Although their model is derived from an analysis of the overlap of two modes of a bimodal population, we believe that it may be more widely applicable and that both the intraindividual variance of the test and the interindividual differences in the population may contribute to blurring the classification. In this article we analyse the concept of deterioration to diabetes and the instability of the impaired glucose tolerance class in the light of data on the variability of the biological response to a glucose load. We conclude that additional criteria are required to categorise people with impaired glucose tolerance and that measures of insulin and of proinsulin like molecules are possible candidates for this task.