OBJECTIVES: We examined whether the finding of glycosuria and its level in themselves give information of clinical relevance, apart from being an unreliable indicator of glycemic control. METHODS: Subjects were a population-based sample of 1,284 newly diagnosed type 2 diabetic patients. Median age was 65.2 years. Urinary glucose concentration (UGC) was determined quantitatively in a freshly voided morning urine specimen. RESULTS: The over-all prevalence of peripheral vascular disease (PVD) was 16.5%. Bivariately, high values of UGC were associated with low prevalence of PVD (p<0.001, chi2-test). The predictive value of PVD--together with HbA1c, glomerular filtration rate (GFR) and 10 other possible predictors--was confirmed in a logistic regression analysis with glycosuria (Y/N) as outcome variable (p=0.0004). CONCLUSION: Surprisingly, type 2 diabetic patients with PVD tend not to have glycosuria as compared to patients without PVD. PVD may be indicative of generalized atherosclerotic lesions in the major vessels, including the renal arteries. This could lead to a lowering of GFR and thereby of the amount of glucose filtered. Assuming no, or only a minor direct effect of PVD on tubular function, this would lead to an increased renal threshold for glucose in patients with PVD.