Risk Factors for Recurrent Acute Kidney Injury in a Large Population-Based Cohort

Rationale & Objective Acute kidney injury (AKI) has numerous sequelae. Repeated episodes of AKI may be an important determinant of adverse outcomes, including chronic kidney disease and death. In a population-based cohort study, we sought to determine the incidence of and predictors for recurrent AKI. Study Design Retrospective cohort study. Setting & Participants 38,659 hospitalized members of Kaiser Permanente Northern California who experienced an episode of AKI from 2006 to 2013. Predictors Demographic, clinical, and laboratory data, including baseline kidney function, proteinuria, hemoglobin level, comorbid conditions, and severity of AKI. Outcomes Incidence and predictors of recurrent AKI. Analytical Approach Multivariable Cox proportional hazard regression. Results 11,048 (28.6%) experienced a second hospitalization complicated by AKI during follow-up (11.2 episodes/100 person-years), with the second episode of AKI occurring a median of 0.6 (interquartile range, 0.2-1.9) years after the first hospitalization. In multivariable analyses, older age, black race, and Hispanic ethnicity were associated with recurrent AKI, along with lower estimated glomerular filtration rate, proteinuria, and anemia. Concomitant conditions, including heart failure, acute coronary syndrome, diabetes, and chronic liver disease, were also multivariable predictors of recurrent AKI. Those who had higher acuity of illness during the initial hospitalization were more likely to have recurrent AKI, but greater AKI severity of the index episode was not independently associated with increased risk for recurrent AKI. In multivariable analysis of matched patients, recurrent AKI was associated with an increased rate of death (HR, 1.66; 95% CI, 1.57-1.77). Limitations Analyses were based on clinically available data, rather than protocol-driven timed measurements of kidney function. Conclusions Recurrent AKI is a common occurrence after a hospitalization complicated by AKI. Based on routinely available patient characteristics, our findings could facilitate identification of the subgroup of patients with AKI who may benefit from more intensive follow-up to potentially avoid recurrent AKI episodes.