Investigating the predictive value of urinary biomarkers in cardiac surgery related acute kidney injury: systematic review and meta-analysis

Introduction Cardiac surgery related acute kidney injury (CS-AKI) develops in 20-50% of patients undergoing cardiac surgery and is responsible for increased postoperative morbidity and mortality. Urinary biomarkers have been investigated to predict and thus prevent acute kidney injury in several critical condition, but their predictive value remains unclear in regards to cardiac surgery related postoperative renal failure. Our aim was to assess the predictive value of urinary biomarkers for prediction of cardiac surgery related acute kidney injury. Methods This study was registered in PROSPERO (CRD42022371166). We conducted our systematic search in 3 databases on 11.11. 2022 (MEDLINE, EMBASE, COCHRANE) without filters or restrictions. We included both randomised and non-randomised studies reporting diagnostic accuracy data. Our primary outcome was the predictive values of individual urinary biomarkers measured at different time-points to identify patients developing acute kidney injury as per Kidney Disease Improving Global Outcomes (KDIGO) criteria and also calculated the performance of combination of urinary biomarkers. We collected the Area Under the Curve (AUC) values and their confidence intervals and performed a meta-analysis using random or mixed effects models. We fitted Summary Receiver Operating Characteristics (SROC) curves using 2 × 2 contingency tables extracted from the studies containing the true positive, false positive, false negative, and true negative values. Risk of bias was assessed by QUADAS-2. Results We screened 10763 records and included 92 articles in the analysis. Predictive value of individual biomarkers was at maximum fair; TIMP2xIGFBP7 measured in the intraoperative and early postoperative period AUC 0.73 (95% CI: 0.65-0.81) for prediction of KDIGO stage 2-3 acute kidney injury; L-FABP measured from 6h till 24h postoperatively AUC 0.75 (95% CI: 0.68-0.81) for prediction of all stages of CS-AKI. There was no significant difference between the AUC-ROC of urinary biomarkers (p=0.0655). Combination of urinary biomarker measurements yielded a good predictive value in identification of both total and only severe cases of acute kidney injury with AUCs 0.82 (95% CI: 0.75-0.88) and 0.85 (95% CI: 0.79-0.91) respectively. The combination of three biomarkers did not provide significant predictive value improvement (p=0.625) compared to the combination of two urinary biomarkers AUC 0.87 (95% CI: 0.78-0.95) vs 0.84 (95% CI: 0.75-0.92) in prediction of severe CS-AKI. Discussion Our study shows that biomarkers identified patients developing CS-AKI. Individual biomarkers performed with fair accuracy, while combination of two biomarkers improved the accuracy. However, combining more than two urinary biomarkers did not result in a better predictive value but it could increase costs unnecessarily.