Derivation and validation of urinary TIMP-1 for the prediction of acute kidney injury and mortality in critically ill children

Background Acute kidney injury (AKI) is associated with high morbidity and mortality. Multiple urinary biomarkers have been identified to be associated with the prediction of AKI and outcomes. However, the accuracy of these urinary biomarkers for AKI and associated outcomes has not been clearly defined, especially in heterogeneous populations. The aims of the study were to compare the ability of 10 existing or potential urinary biomarkers to predict AKI and pediatric intensive care unit (PICU) mortality and validate urinary tissue inhibitor of metalloproteinases-1 (uTIMP-1) as a better biomarker for early prediction in heterogeneous critically ill children. Methods A derivation-validation approach with separate critically ill cohorts was designed. We first conducted a prospective cohort study to determine the ability of 10 urinary biomarkers serially measured in 123 children during the first 7 days of PICU stay to predict AKI and PICU mortality (derivation study) and further validated the better biomarker of uTIMP-1 in a separate cohort of 357 critically ill children (validation study). AKI diagnosis was based on KDIGO classification with serum creatinine and urine output. PICU mortality was defined as all-cause mortality. Results In the derivation cohort, 17 of 123 (13.8%) children developed AKI stage 3 or died during the PICU stay, and both the initial and peak uTIMP-1 displayed the highest AUCs of 0.87 (0.79–0.94) and 0.90 (0.84–0.96), respectively, for predicting AKI stage 3 or death. In the validation cohort, 78 of 357 (21.8%) developed AKI during the first week after admission, and 38 (10.6%) died during the PICU stay. The initial uTIMP-1 level was validated to be independently associated with AKI (AOR = 2.88, 95% CI 1.97–4.21), severe AKI (AOR = 2.62, 95% CI 1.78–3.88), AKI stage 3 (AOR = 2.94, 95% CI 1.84–4.68) and PICU mortality (AOR = 1.92, 95% CI 1.11–3.30) after adjustment for potential confounders. The predictive values of uTIMP-1 for AKI, severe AKI, AKI stage 3 and PICU mortality were 0.80 (0.74–0.86), 0.83 (0.77–0.89), 0.84 (0.77–0.92) and 0.83 (0.76–0.89), respectively. Conclusions Urinary TIMP-1 levels have been identified and validated to be independently associated with AKI and PICU mortality in independent prospective cohorts and may be an early potential indicator of AKI and PICU mortality in critically ill children.