Although much emphasis has been placed on screening for albuminuria in type II diabetic patients, less attention has been focused on the role of glomerular filtration rate (GFR) in the assessment of risk. Herein, we examined the association between GFR and vascular complications in a consecutive cohort of 5174 type II diabetic patients between 1995 and 2000. Renal function was assessed by GFR (estimated by Modification of Diet in Renal Disease equation). The frequency of chronic kidney disease (CKD) as defined by GFR <60 ml/min/1.73 m(2), micro- and macrovascular complications, and their associations were analyzed. In this study cohort, 6% had serum creatinine > or =150 micromol/l and 15.8% had CKD. After adjustment for potential confounders, including urinary albumin excretion, odds ratios [95% confidence interval (CI)] across different stages of estimated GFR (> or =90, 60-89, 30-59, 15-29, <15 ml/min/1.73 m(2)) for macrovascular disease were 1.00, 1.42 [1.12-1.80], 1.80 [1.32-2.45], 2.74 [1.64-4.56], and 4.05 [1.77-9.26], respectively (P for trend <0.001); for retinopathy were 1.00, 1.23 [1.04-1.46], 1.80 [1.40-2.30], 2.05 [1.25-3.37], and 4.12 [1.56-10.90], respectively (P for trend <0.001); for sensory neuropathy were 1.00, 1.53[1.27-1.85], 2.09 [1.58-2.76], 4.32 [2.41-7.77], and 3.16 [1.25-8.02], respectively (P for trend <0.001); and for microalbumuria (with GFR <15 ml/min/1.73 m(2) excluded from the analysis) were 1.00, 1.51 [1.30-1.75], 5.80 [4.52-7.44], and 52.5 [16.4-168.2] respectively (P for trend <0.001). Measurement of serum creatinine alone without GFR may underestimate renal impairment in type II diabetic patients. Decreasing GFR was significantly associated with increasing frequency of micro- and macrovascular complications.