Derivation And Validation Of Cystatin C In Predicting Postoperative Acute Kidney Injury And Their Outcome In Surgery Patients

Objective: To investigate and verify the value of cystatin C in predicting post-operative acute kidney injury and their outcome in surgical patients.Methods: Retrospective analysis: 225 patients treated by surgery were divided into cardiac surgery group and non-cardiac major surgery group according to the type of surgery.First,collect the serum cystatin C and creatinine levels of the two groups of patients within 24 hours before surgery,compare and evaluate their value in predicting postoperative AKI,and determine the cut-off value of cystatin C in predicting acute kidney injury.The second,cystatin C and creatinine were combined with the American Academy of Anesthesiologists(ASA)classification to form a classification prediction model,and analyze the prediction of cystatin C-ASA classification prediction model and the creatinine-ASA classification prediction model for postoperative acute kidney injury,and determine its best predicted probability.The outcome of postoperative acute kidney injury in the two groups was followed up,and the two groups were divided into recovery and non-recovery of acute kidney injury at 1 week as the end point.We use preoperative cystatin C and creatinine to predict the non-recovery,and finally determine the cut-off value.Prospective study: 65 surgical patients were selected as the validation cohort.The accuracy of cystatin C and cystatin C-ASA in the prediction of post-operative acute kidney injury was evaluated in the validation cohort using the cut-off values and best prediction probability determined by the analysis cohort.Finally,assess the accuracy of the outcome of postoperative acute kidney injury(AKI)non-recovery within 1 week of occurrence.Results: In the cardiac surgery group,the AUC for preoperative cystatin C and creatinine levels to predict the risk of postoperative AKI was 0.804(95% confidence interval 0.726-0.881,P < 0.001)and 0.758(95% CI 0.677-0.840,P < 0.001).In the non-cardiac surgery group,preoperative cystatin C and creatinine levels predicted the risk of postoperative AKI with AUC were 0.802(95%CI 0.660-0.943,P = 0.002)and 0.746(95%CI 0.587-0.905,P = 0.013),respectively.The cut-off values of cystatin C for the diagnosis of AKI after cardiac surgery and non-cardiac surgery were 0.975 mg/L(sensitivity 86.5%,specificity 71.4%)and 0.89 mg/L(sensitivity 94.6%,specificity 62.8%),respectively.In the cardiac surgery group,the AUC of the preoperative cystatin C-ASA classification prediction model and the creatinine-ASA classification model and were 0.801(95%CI 0.720-0.878,P< 0.001)and 0.780(95%CI0.688-0.850,P< 0.001),the best predictive probability of cystatin C-ASA classification prediction model was 0.661(sensitivity 72.2%,specificity 77.8%).The AUC of cystatin C in predicting the risk of non-recovery after AKI in cardiac surgery group was 0.797(95%CI0.702-0.892,P< 0.001),the cut-off value for diagnosis was 1.16 mg/L(sensitivity 81.6%,specificity 73.1%),and the AUC of creatinine for predicting the risk of non-recovery of AKI in cardiac surgery patients was 0.769(95%CI 0.670-0.867,P < 0.001).According to the cur-off value of cystatin C determined in the analysis cohort and validated in the validation cohort,the sensitivity and specificity of cystatin C for predicting postoperative AKI in the cardiac surgery group were 71.9%(55.4-88.3)and 81.8%(54.6-109),respectively.While in the non-cardiac surgery group,the sensitivity and specificity of cystatin C for predicting postoperative AKI were80%(49.8-110.1)and 75%(46.3-103.7),respectively.The sensitivity and specificity of cystatin C-ASA classification in predicting postoperative AKI were 84.4%(71.1-97.7)and 81.8%(54.6-109),respectively.The sensitivity and specificity of cystatin C for predicting non-recovery within 1 week after cardiac surgery in patients with AKI were 78.6%(54-103.2)and 88.9%(72.8-105).Conclusion: Preoperative cystatin C has a predictive value for the occurrence of postoperative acute kidney injury in cardiac surgery and non-cardiac surgery,and the predictive ability is better than that of creatinine.The predictive model combining cystatin C with ASA classification clinical risk factors can slightly improve the predictive ability of postoperative AKI in patients undergoing cardiac surgery,which is better than creatinine-ASA classification.Preoperative cystatin C also has a certain predictive value for the failure to recover within 1week after the occurrence of AKI in patients undergoing cardiac surgery,and its predictive ability is better than that of creatinine.And the predictive ability of cystatin C was verified in the validation cohort.