Purpose: Patients with severe chronic kidney disease are at substantially higher risk to develop cardiovascular disease, but an increased risk is also evident in individuals with only mild signs of kidney damage in the community. Cardiovascular risk prediction is particularly important in the primary prevention of cardiovascular disease. Nevertheless, data on whether the combined addition of albuminuria and estimated glomerular filtration rate (eGFR) improves cardiovascular risk prediction in individuals without cardiovascular disease in the community is scarce. Methods and results: We investigated associations between urinary albumin excretion rate (UAER), cystatin C-based eGFR and cardiovascular mortality in a community-based cohort of elderly men (ULSAM-study; n=1113, mean age 71 years, 208 cardiovascular deaths, median follow-up 12.9 years) with pre-specified analyses in participants without cardiovascular disease (n=649, 86 cardiovascular deaths). Using multivariable Cox regression, higher UAER and lower eGFR were associated with increased risk for cardiovascular mortality independently of established cardiovascular risk factors in the whole sample and in men without cardiovascular disease at baseline (sub-sample without cardiovascular disease: UAER; hazard ratio per 1-SD 1.26, 95% CI 1.05-1.51, p=0.01; eGFR: hazard ratio per 1-SD 0.74, 95% CI 0.59-0.92, p=0.007). Analyses of model discrimination, calibration, reclassification and global fit suggested that UAER and eGFR add relevant prognostic information beyond established cardiovascular risk factors also in participants without prevalent cardiovascular disease. Interestingly, established cut-offs used to diagnose microalbuminuria (UAER>20µg/min) and chronic kidney disease stage 3 (eGFR<60ml/min/1.73m2), appeared less suitable for cardiovascular risk prediction (integrated discrimination improvement (IDI) 0.006, p=0.11, while cut-offs UAER>6µg/min and eGFR<45ml/min/1.73m2 significantly improved IDI (0.047, p<0.001) Conclusion: UAER and eGFR improved cardiovascular risk prediction beyond established cardiovascular risk factors, suggesting that these kidney biomarkers may be useful in predicting cardiovascular death in the elderly.