Abstract: ackground  This study aimed to evaluate the impact of timing of renal replacement therapy (RRT) on prognosis in patients with severe acute kidney injury.  Methods  PubMed, Web of Science, Em-base, CNKI, Wanfang Data Knowledge Service platform, VIP and other databases were searched for literatures related to the influence of different start-up times of renal replacement therapy on patients with severe and complicated acute kidney injury. The retrieval time was limited to create database to December 30, 2021. The article type was limited to randomized controlled study (RCT). Literature screening and data extraction were independently completed by two researchers. The main outcome indicators were 28-day mortality, and the secondary outcome indicators were the need for renal replacement therapy at 90 days and the incidence of related complications. Meta-analysis was performed using RevMan 5.3 software. The Cochrane Risk of Bias Tool was used to assess the risk of bias in all studies.  Results  A total of 5709 severe patients were included in 15 RCTS in this study, including 2851 patients in the early initiation group and 2858 patients in the late initiation group. Meta-analysis showed that there was no significant difference in mortality between the early and late initiation groups [odds ratio (OR) = 0.96, 95%confidence interval (95% CI: 0.86～1.08, P =0.49]. Late initiation reduced RRT dependence within 90 days (OR=1.57, 95% CI:1.13～2.19, P ＜0.05), and the risk of catheter related blood stream infection (CRBSI), RRT-associated hypotension, and hypophosphatemia. (CRBSI:OR=1.4495%, CI:1.06～1.95, P =0.02; Hypotension: OR=1.4395% CI:1.17～1.76, P ＜0.05; Hypophosphatemia: OR=1.85, 95% CI:1.02～3.35, P =0.04).  Conclusions  Early RRT initiation does not reduce mortality in patients with severe AKI, but late RRT initiation can reduce RRT dependence within 90 days and the risk of related complications. 

Key words: Renal replacement therapy, Acute kidney injury；Initiation