Meta-analysis on early predictive value of urinary TIMP-2 and IGFBP-7 product for acute kidney injury in cardiac postoperative patients

Objective To assess the early predictive value of urinary TIMP-2 and IGFBP-7 product for postoperative acute kidney injury in cardiac postoperative patients. Methods EMBASE, Pubmed, Cochrane, Web of Science, CNKI, Sinomed and Wanfang database were searched. The limited time is from February 1,2014 to February 1,2018. A total of 138 articles related to "prediction of postoperative acute kidney injury in patients with heart-related surgery" were retrieved and 10 articles were screened out.The required data were extracted. QUADAS-2 quality evaluation criteria were used to evaluate the literature.Meta-DiSc 1.4 software was used for data analysis and publication bias was tested by Stata 12.0. Subgroup analysis, meta-regression, and sensitivity analysis were adopted to identify the sources of heterogeneity. Results Ten articles were screened out eventually and they related to a total number of 659 cases (AKI occurring, 202 cases; and no AKI occurrence, 457 cases). Heterogeneity test showed Spearman correlation coefficient of 0.067 (P=0.855), suggesting that the threshold effect of heterogeneity was not obvious. With meta-analysis it was found that, the pooled diagnostic odds ratio I² was <25%, suggesting that there was no heterogeneity in the non-threshold effect between studies. The combined effect of urine TIMP-2 and IGFBP-7 product for early predictionof AKI in postoperative patients with cardiac surgery showed sensitivity of 0.72 (95% CI 0.66~0.78), specificity of 0.72 (95% CI 0.68~0.76), diagnostic odds ratio of 8.12 (95% CI 5.46~12.08), pooled AUC=0.8021, indicating thaturinary TIMP-2 and IGFBP-7 producthad good predictive value for early acute kidney injury in patients after cardiac surgery.The publication bias was not obvious. Subgroup analysis shows that the sample size might be the source of heterogeneity, the meta-regression results did not reveal significant heterogeneity, and sensitivity analysis showed that this method did not have poor robustness. Conclusions Urine TIMP-2 and GFBP-7 product has important clinical value in early predicting acute renal injury in patients after cardiac surgery, but it still needs further large-scale clinical studies to confirm.