Abstract
Acute kidney injury (AKI) associates with higher in-hospital mortality, but whether it also associates with increased long-term mortality is unknown, particularly after accounting for residual kidney function after hospital discharge. We retrospectively analyzed data from US veteran patients who survived at least 90 d after discharge from a hospitalization. We identified AKI events not requiring dialysis from laboratory data and classified them according to the ratio of the highest creatinine during the hospitalization to the lowest creatinine measured between 90 d before hospitalization and the date of discharge. We estimated mortality risks using multivariable Cox regression models adjusting for demographics, comorbidities, medication use, primary diagnosis of admission, length of stay, mechanical ventilation, and postdischarge estimated GFR (residual kidney function). Among the 864,933 hospitalized patients in the study cohort, we identified 82,711 hospitalizations of patients with AKI. In the study population of patients who survived at least 90 d after discharge, 17.4% died during follow-up (AKI 29.8%, without AKI 16.1%). The adjusted mortality risk associated with AKI was 1.41 (95% confidence interval [CI] 1.39 to 1.43) and increased with increasing AKI stage: 1.36 (95% CI 1.34 to 1.38), 1.46 (95% CI 1.42 to 1.50), and 1.59 (95% CI 1.54 to 1.65; P < 0.001 for trend). In conclusion, AKI that does not require dialysis associates with increased long-term mortality risk, independent of residual kidney function, for patients who survive 90 d after discharge. Long-term mortality risk is highest among the most severe cases of AKI.