Background: Previous studies that evaluated the association of kidney function with incident heart failure may be limited by the insensitivity of serum creatinine concentration for detecting abnormal kidney function. Objective: To compare serum concentrations of cystatin C(a novel marker of kidney function) and creatinine as predictors of incident heart failure. Design: Observational study based on measurement of serum cystatin C from frozen sera obtained at the 1992-1993 visit of the Cardiovascular Health Study. Follow-up occurred every 6 months. Setting: Adults 65 years of age or older from 4 communities in the United States. Participants: 4384 persons without previous heart failure who had measurements of serum cystatin C and serum creatinine. Measurements: Incident heart failure. Results: The mean(±SD) serum concentrations of cystatin C and creatinine were 82±25 nmol/L(1.10±0.33 mg/L) and 89±34 μmol/L(1.01±0.39 mg/dL), respectively. During a median follow-up of 8.3 years(maximum, 9.1 years), 763(17%) participants developed heart failure. After adjustment for demographic factors, traditional and novel cardiovascular risk factors, cardiovascular disease status, and medication use, sequential quintiles of cystatin C concentration were associated with a stepwise increased risk for heart failure in Cox proportional hazards models(hazard ratios, 1.0 reference , 1.30 95%CI, 0.96 to 1.75 , 1.44 CI, 1.07 to 1.94 , 1.58 CI, 1.18 to 2.12 , and 2.16 CI, 1.61 to 2.91 ). In contrast, quintiles of serum creatinine concentration were not associated with risk for heart failure in adjusted analysis(hazard ratios, 1.0 reference , 0.77 CI, 0.59 to 1.01 , 0.85 CI, 0.64 to 1.13 , 0.97 CI, 0.72 to 1.29 , and 1.14 CI, 0.87 to 1.49 ). Limitations: The mechanism by which cystatin C concentration predicts risk for heart failure remains unclear. Conclusions: The cystatin C concentration is an independent risk factor for heart failure in older adults and appears to provide a better measure of risk assessment than the serum creatinine concentration. Background: Previous studies that evaluated the association of kidney function with incident heart failure may be limited by the insensitivity of serum creatinine concentration for detecting abnormal kidney function. Objective: To compare serum concentrations of cystatin C (a novel marker of kidney function) and creatinine Design: Observational study based on measurement of serum cystatin C from frozen sera obtained at the 1992-1993 visit of the Cardiovascular Health Study. Follow-up occurred every 6 months. Setting: Adults 65 years of age or older from 4 communities in the United States. Participants: 4384 persons without previous heart failure who had measurements of serum cystatin C and serum creatinine. Measurements: Incident heart failure. Results: The mean (± SD) serum concentrations of cystatin C and creatinine were 82 ± 25 nmol / L (1.10 ± 0.33 mg / L) and 89 ± 34 μmol / L (1.01 ± 0.39 mg / dL), respectively. During a median follow-up of 8.3 years After adjustment for demographic factors, traditional and novel cardiovascular risk factors, cardiovascular disease status, and medical use, sequential quintiles of cystatin C concentration were associated with a stepwise increased risk for heart failure in Cox proportional hazards models (hazard ratios, 1.0 reference, 1.30 95% CI, 0.96 to 1.75, 1.44 CI, 1.07 to 1.94, 1.58 CI, 1.18 to 2.12, and 2.16 CI, 1.61 to 2.91) quintiles of serum creatinine concentration were not associated with risk for heart failure in adjusted analysis (hazard ratios, 1.0 reference, 0.77 CI, 0.59 to 1.01, 0.85 CI, 0.64 to 1.13, 0.97 CI, 0.72 to 1.29, and 1.14 CI, 0.87 to 1.49). Limitations: The mechanism by which cystatin C concentration predicts risk for heart failure remains unclear. Conclusions: The cystatin C concentration is an independent risk factor for heart failure in older adults and appears to provide a better me asureof risk assessment than the serum creatinine concentration.