Since the oral glucose tolerance test is not applicable in large surveys for diabetes mellitus, economic tests or clinical criteria are required to select subjects needing a complete glycometabolic assessment. This study evaluated, in a selected high risk population (110 subjects; 71 males; median age: 53 years, range 30-70), the utility of glycosuria after a meal rich in carbohydrates in the detecting of diabetes, compared to other tests for glycosuria (fasting glycosuria, glycosuria after a glucose load) and clinical criteria (diabetic family history, overweight and their combination). For each test and, retrospectively, for clinical criteria, sensitivity, sensibility, predictive value positive (PVP) and predictive value negative (PVN) were calculated. Postprandial glycosuria resulted to be an effective approach (sensitivity: 100%; specificity: 44%; PVP: 23%, PVN: 100%) in people over 50 years old, in which both clinical criteria (sensitivity: 22-78%; PVN: 81-90%) and fasting glycosuria (sensitivity 11%; PVN: 86%) proved defective. In younger subjects both the presence of fasting glycosuria (sensitivity: 100%; specificity: 96%; PVP: 33%; PVN: 100%) and of relevant diabetic family history (sensitivity: 100%; specificity: 85%; PVN: 100%) proved more adequate as screening procedures.