No data are available on the absolute risk of either venous thromboembolism (VTE) or arterial thromboembolism (ATE) in patients with nephrotic syndrome. Reported risks are based on multiple case reports and small studies with mostly short-term follow-up. We assessed the absolute risk of VTE and ATE in a large, single-center, retrospective cohort study and attempted to identify predictive factors in these patients. A total of 298 consecutive patients with nephrotic syndrome (59% men; mean age, 42+/-18 years) were enrolled. Mean follow-up was 10+/-9 years. Nephrotic syndrome was defined by proteinuria > or =3.5 g/d, and patients were classified according to underlying histological lesions accounting for nephrotic syndrome. Objectively verified symptomatic thromboembolic events were the primary study outcome. Annual incidences of VTE and ATE were 1.02% (95% confidence interval, 0.68 to 1.46) and 1.48% (95% confidence interval, 1.07 to 1.99), respectively. Over the first 6 months of follow-up, these rates were 9.85% and 5.52%, respectively. Proteinuria and serum albumin levels tended to be related to VTE; however, only the predictive value of the ratio of proteinuria to serum albumin was significant (hazard ratio, 5.6; 95% confidence interval, 1.2 to 26.2; P=0.03). In contrast, neither the degree of proteinuria nor serum albumin levels were related to ATE. Sex, age, hypertension, diabetes, smoking, prior ATE, and estimated glomerular filtration rate predicted ATE (P< or =0.02). This study verifies high absolute risks of symptomatic VTE and ATE that were remarkably elevated within the first 6 months. Whereas the ratio of proteinuria to serum albumin predicted VTE, estimated glomerular filtration rate and multiple classic risk factors for atherosclerosis were predictors of ATE.