Among the numerous studies concerning contrast media-induced nephropathy (CIN), there was no prospective trial that provided data on the long-term outcomes. To prospectively assess predictors of CIN and long-term outcomes of affected patients. Four hundred twelve consecutive patients with serum creatinine levels of 115 micromol/L to 309 micromol/L (1.3 mg/dL to 3.5 mg/dL) undergoing elective coronary angiography were included. Patients were randomly assigned to periprocedural hydration alone, hydration plus one-time hemodialysis or hydration plus N-acetylcysteine. Multivariate logistic regression identified the following as predictors of CIN within 72 h (equivalent to an increase in creatinine 44.2 micromol/L [0.5 mg/dL] or more) : prophylactic postprocedural hemodialysis (OR 2.86, 95% CI 1.07 to 7.69), use of angiotensin-converting enzyme inhibitors (OR 6.16, 95% CI 2.01 to 18.93), baseline glomerular filtration rate (OR 0.94, 95% CI 0.90 to 0.98) and the amount of contrast media given (OR 1.01, 95% CI 1.00 to 1.01). With regard to long-term outcome (mean follow-up 649 days), multivariate Cox regression models found elevated creatinine levels at 30 days (hazard rate ratio [HRR] 5.48, 95% CI 2.85 to 10.53), but not CIN within 72 h (HRR 1.12, 95% CI 0.63 to 2.02), to be associated with increased mortality. In addition, independent predictors for death during follow-up included left ventricular ejection fraction lower than 35% (HRR 4.01, 95% CI 2.22 to 7.26), serum phosphate (HRR 1.64, 95% CI 1.10 to 2.43) and hemoglobin (HRR 0.80, 95% CI 0.67 to 0.96). From the present prospective trial, performance of postprocedural hemodialysis, use of angiotensin-converting enzyme inhibitors, reduced baseline glomerular filtration rate and amount of contrast media were independent predictors of CIN within 72 h after catheterization. Assessing renal function after 30 days, rather than within 72 h, seemed to be more predictive for patients' long-term survival.