A Meta-analysis of the predictive effect of neutrophil-lymphocyte ratio on acute kidney injury

Abstract

To systematically evaluate the predictive value of neutrophil-lymphocyte ratio (NLR) in acute kidney injury (AKI). All studies about the predictive effect of NLR on AKI were searched in the National Medical Library of the United States PubMed Database, the Embase database in the Netherlands, the Chinese Biology Medicine disc (CBMdisc) and the Chinese Evidence Based Medicine Cochrane Centre Database (CEBM/CCD). The data updated by October 2020, and regardless of language, region or whether blind method was used. Two authors independently extracted data and evaluated the quality of the studies. Data extracted from the studies were analyzed with RevMan 5.3 to assess the predictive value of NLR on AKI. A subgroup Meta-analysis was conducted to assess the predictive value of NLR on AKI according to different countries, different disease types (cardiovascular surgery, infectious diseases, other diseases including burns, cirrhosis, and emergency), and different sample sizes (≤ 300 cases and > 300 cases). The publication bias of included studies about the predictive effect of NLR on AKI were assessed by funnel plots. A total of 11 studies were included in this Meta-analysis, including 4 997 patients, 1 308 patients in AKI group, and 3 689 patients in non-AKI group. The Meta-analysis results showed that: increased NLR had predictive value for the occurrence of AKI [mean difference (MD) = 2.73, 95% confidence interval (95%CI) was 1.78-3.68, P < 0.000 01]. Subgroup analysis showed that increased NLR had predictive value for the occurrence of AKI in patients from Southeast Asia (MD = 4.04, 95%CI was 1.09-6.99, P = 0.007) and Eurasia (MD = 2.51, 95%CI was 1.12-3.90, P = 0.000 4). Increased NLR had predictive value for the occurrence of AKI in patients undergoing cardiovascular surgery (MD = 0.77, 95%CI was 0.34-1.20, P = 0.000 4), infectious diseases (MD = 4.74, 95%CI was 1.51-7.96, P = 0.004) and other diseases (MD = 8.53, 95%CI was 6.26-10.80, P < 0.000 01). Increased NLR had predictive value for the occurrence of AKI in studies with a sample size of ≤ 300 cases (MD = 6.02, 95%CI was 4.90-7.14, P < 0.000 01) and > 300 cases (MD = 1.32, 95%CI was 0.61-2.03, P = 0.000 3). There was no significant publication bias in the included studies assessed by funnel plots. NLR is an important predictive tool for AKI.