Background: Although small changes in creatinine level during hospitalization have been associated with risk of short-term mortality, associations with posthospitalization end-stage renal disease (ESRD) and long-term mortality are unknown. We assessed the relationship between change in serum creatinine levels up to 3.0 mg/dL and death and ESRD among elderly survivors of hospitalization for acute myocardial infarction. Methods: Retrospective cohort study of a nationally representative sample of Medicare beneficiaries admitted with acute myocardial infarction to nonfederal US hospitals between February 1994 and July 1995. Outcomes were mortality and ESRD through June 2004. Results: The 87 094 eligible patients admitted to 4473 hospitals had a mean age of 77.1 years; for the 43.2% with some creatinine increase, quartiles of increase were 0.1, 0.2, 0.3 to 0.5, and 0.6 to 3.0 mg/dL. Incidence of ESRD and mortality ranged from 2.3 and 139.1 cases per 1000 person-years, respectively, among patients with no increase to 20.0 and 274.9 cases per 1000 person-years in the highest quartile of creatinine increase. Compared with patients without creatinine increase, adjusted hazard ratios by quartile of increase were 1.45, 1.97, 2.36, and 3.26 for ESRD and 1.14, 1.16, 1.26, and 1.39 for mortality, with no 95% confidence intervals overlapping 1.0 for either end point. Conclusion: In a nationally representative sample of elderly patients discharged after hospitalization for acute myocardial infarction, small changes in serum creatinine level during hospitalization were associated with an independent higher risk of ESRD and death.