Oral glucose tolerance test in general practice - when is it worthwhile?

Objective - To examine the probability of impaired glucose tolerance and diabetes, diagnosed with oral glucose tolerance test, at different levels of fasting blood glucose. Design - When indicated, fasting capillary whole blood glucose was measured. Patients with blood glucose ranging from 4.4 to 6.6 mmol/1 had an oral glucose tolerance test. Setting - General practice Subjects - 355 patients had an oral glucose tolerance test. Results - 33% had impaired glucose tolerance and 12% had diabetes. One or more diabetics were found at every level of fasting blood glucose. The probability of finding a diabetic was ten times larger in the fasting blood glucose stratum 6.1 to 6.6 mmol/1 than in the fasting blood glucose stratum 4.4 to 4.9 mmol/1. Likelihood ratio increased 30 times from the lowest to the highest stratum. No cutoff point in the interval 4.4 to 6.6 mmol/1 had both a high sensitivity and a high specificity for the diagnosis of diabetes. Conclusion - If it is of great importance to find every patient with impaired glucose tolerance and diabetes, one has to do an oral glucose tolerance test in every fraction of the fasting blood glucose range 4.4 to 6.6 mmol/l (according to WHO's diagnostic criteria). Using 5.2 mmol/1 as a cutoff point for an oral glucose tolerance test, the sensitivity is still high (0.95), but one would miss a few with diabetes.