Supplementary Material for: Predicting Risk of Recurrent Acute Kidney Injury: A Systematic Review

<b><i>Background:</i></b> Although the epidemiology of acute kidney injury (AKI) has been well described, less is known about recurrent AKI (r-AKI). We undertook a systematic review to identify incidence, risk factors, and outcomes of r-AKI. <b><i>Methods:</i></b> MEDLINE, EMBASE, CINAHL, Cochrane, Web of Science were searched, from inception to December 2017, for quantitative studies on adults with AKI, where follow-up included ­reporting of r-AKI. Two reviewers independently identified studies and assessed study quality. <b><i>Summary:</i></b> From 2,824 citations, 10 cohort studies met inclusion criteria (total patients <i>n</i> = 538,667). There were 2 distinct set of studies; 4 studies assessed r-AKI within the same hospital admission (most were intensive care unit [ICU] patients) and 6 studies assessed postdischarge r-AKI. The median percentage of people developing r-AKI within the same hospital admission was 23.4% (IQR 20.3–27.2%) and postdischarge r-AKI was 31.3% (IQR 26.4–33.7%). A higher Acute Physiology and Chronic Health Evaluation score was associated with increased risk of r-AKI within the same hospital admission in ICU patients. Cardiovascular disease and AKI severity were associated with increased risk of postdischarge r-AKI. R-AKI (within same admission and postdischarge) was associated with worse survival. It was not possible to pool results due to methodological differences across studies, such as varying definitions for AKI and r-AKI, varying length of follow-up and effect measures. <b><i>Key messages:</i></b> More representative population-based studies with robust assessment of predictors and consensus definition of r-AKI are needed to identify risk factors and develop risk stratification tools to reduce recurrence and improve outcomes. <b><i>Systematic Review Registration:</i></b> CRD42017082668.