A meta-analysis of clinical predictors for renal recovery and overall mortality in acute kidney injury requiring continuous renal replacement therapy

Purpose: To determine clinical predictors for continuous renal replacement therapy (CRRT) discontinuation in patients with acute kidney injury (AKI). Materials and methods: Ovid MEDLINE, EMBASE, and Cochrane Library were searched. The protocol is registered on researchregistry.com (reviewregistry909). Our criteria included non-end-stage kidney disease adults who required CRRT for AKI. Renal recovery was defined by CRRT discontinuation. Risk of bias was assessed using ROBINS-I tool. Results: We classified our analyses into renal recovery cohort and overall mortality cohort. All studies were observational. For renal recovery cohort, increasing urine output at time of CRRT discontinuation, elevated initial SOFA score and serum creatinine at CRRT initiation were predictive of renal recovery with OR 1.021 (95%CI = 1.011–1.031), 0.869 (95%CI = 0.811–0.932) and 0.995 (95%CI = 0.996–0.999), respectively. For overall mortality cohort, age and presence of sepsis were significantly associated with overall mortality with OR of 1.028 (95%CI = 1.008–1.048) and 2.160 (95%CI = 0.973–1.932), respectively. Conclusions: Urine output at CRRT discontinuation, lower initial SOFA score, and lower serum creatinine levels at CRRT initiation were associated with higher likelihood of renal recovery. Increasing age and the presence of sepsis were associated with increased overall mortality from AKI on CRRT. However, there were limited data on co-morbidities which might preclude their inclusion in our analysis.