Serum N-terminal Pro-b-type Natriuretic Peptide And Cystatin C Predict Acute Kidney Injury And Adverse Outcomes In Critically Ill Patients:A Prospective,observational Study

Background and objectiveAcute kidney injury(AKI)is a common complication of patients in intensive care unit(ICU).Both serum N-terminal pro-B-type natriuretic peptide(NT-proBNP)and cystatin C(CysC)are available clinically and beneficial in AKI detection and prediction.However,for AKI in critically ill patients,their combined diagnostic accuracy and prognostic valve are still unknown.Our purpose is to identify the performance of NT-proBNP and CysC levels at ICU admission for AKI and ICU adverse outcomes prediction,and build a predictive model for AKI and prognosis,so as to improve the prognosis of critically ill patients with AKI.MethodsSection 1:A prospective observational study was executed on a large,heterogeneous cohort of patients who were admitted to a mixed medical-surgical ICU.Serum NT-proBNP and CysC were detected at ICU admission.To evaluate the predictive value of NT-proBNP combined with CysC for AKI prediction,we analyzed the risk factors ofA KI,and constructed the prediction models.The area under the receiver operating characteristic curve(AUC-ROC),category-free net reclassification index(cNRI),and incremental discrimination improvement(IDI)were utilized to compare the discriminative powers of NT-proBNP,CysC and their combination for predicting AKI with a clinical model.Section 2:Aprospective observational study was executed,enrolling adults admitted to ICU,excluding patients who had developed AKI at ICU admission Serum NT-proBNP and CysC were detected at ICU admission.To evaluate the predictive value of NT-proBNP combined with CysC for later-onset AKI prediction,we analyzed the risk factors of later-onset AKI,and constructed the prediction models.TheAUC-ROC,cNRI,and IDI were utilized to compare the discriminative powers of NT-proBNP,CysC and their combination for predicting later-onset AKI with a clinical model.Section 3:A prospective observational study was executed,enrolling adults admitted to ICU.Serum NT-proBNP and CysC were detected at ICU admission.To evaluate the predictive value of NT-proBNP combined with CysC for prediction of ICU adverse outcomes,we analyzed the risk factors of ICU adverse outcomes,and constructed the prediction models.The AUC-ROC,cNRI,and IDI were utilized to compare the discriminative powers of NT-proBNP,CysC and their combination for predicting ICU adverse outcomes with a clinical model.ResultsSecction 1:AKI was detected in 256 out of 1222 included patients(20.9%),117(9.6%)were severe AKI.The AUC-ROC of the combination of NT-proBNP and CysC in predicting total AKI and severe AKI was significantly higher than that of the two alone(P<0.05).The predictive model containing NT-proBNP and CysC further increased the AUC-ROC to 0.859 beyond that of the clinical model without these two biomarkers(P<0.05).Moreover,the predictive model significantly improved risk reclassification of total AKI beyond that of the clinical model alone or with single biomarker(P<0.05),as measured by NRI and IDI.Section 2:AKI was detected in 184 out of 1150 included patients(16%).The AUC-ROC of the combination of NT-proBNP and CysC in predicting later-onset AKI was significantly higher than CysC(P<0.05).Compared to the clinical model without NT-proBNP and CysC,the new predictive model containing these two biomarkers for later-onset AKI prediction had the highest predictive ability with improved AUC-ROC,cNRI,IDI.The predictive model significantly improved risk reclassification of later-onset AKI beyond that of the clinical model(P<0.05).Section3:Of the 1222 patients enrolled,74(6.1%)had adverse outcomes.The AUC-ROC of the combination of NT-proBNP and CysC in predicting ICU adverse outcomes was significantly higher than any individual biomarker(P<0.05).Compared to the clinical model without NT-proBNP and CysC,the new predictive model containing these two biomarkers for ICU adverse outcomes had not improve the predictive ability and risk reclassification,measured by AUC-ROC,cNRI,IDI.(P>0.05).ConclusionsIn critically ill patients,the combination of serum NT-proBNP,CysC and clinical risk factors revealed significantly superior discriminative performance for AKI prediction and yielded not additional prognostic information in adverse outcomes.