Clinical Study On Early Diagnosis Of Acute Kidney Injury Associated With Cardiac Surgery

Part Ⅰ:Analysis of risk factors for cardiac surgery-related acute kidney injuryBackground and purpose: The incidence and risk factors(especially laboratory indicators)of acute kidney injury(AKI)following primary cardiovascular surgery remain unclear in developing countries.The present study was designed to identify the incidence and risk factors predictive of AKI in patients undergoing primary cardiovascular surgery.Methods: A retrospective analysis of the records was performed in a single teaching hospital in Nanjing,China between December 1,2019,and December 31,2020.Each patient was categorized within the first week after surgery according to the maximal Kidney Disease Improving Global Outcomes(KDIGO)clinical practice guidelines.Univariate and multivariate logistic regression analyses were carried out to stratify postoperative AKI-related risk factors.Furthermore,the patients were also investigated according to their need for renal replacement therapy,as well as their short-term adverse reactions.Results: Of the 227 patients involved in the final study,66 patients(29.07%)developed AKI.Incidence of AKI was significantly correlated with age,type 2 diabetes mellitus(T2DM),anemia,albumin,type of surgery,duration of surgery,cardiopulmonary bypass(CPB)time,erythrocyte transfusions on the day of surgery,intensive care unit(ICU)length of stay,mechanical ventilation time,and hospitalization time.Following the primary cardiovascular surgery,only 2(3.03%)patients required continuous renal replacement therapy,and both recovered before discharge from hospital.Of the 4 patients(1.76%)who died during hospitalization,3 patients had developed postoperative AKI.Moreover,multivariate analysis verified that T2DM(OR=3.318;95%CI: 1.327–8.297),anemia(OR=2.576;95%CI: 1.017–6.524),albumin(OR=3.393;95%CI: 1.509–7.629),erythrocyte transfusions(OR=4.506;95%CI: 1.587–12.794),and CPB time(OR=3.125;95%CI: 1.044–9.357)were predictors independently associated with the development of AKI.Conclusion: Due to the high incidence of AKI,understanding the independent risk factors associated with the development of AKI as well as early detection of AKI are helpful during the clinical management of patients undergoing primary cardiovascular surgery.Part Ⅱ: Machine learning for the prediction of acute kidney injury in patients after cardiac surgeryBackground and purpose: Cardiac surgery-associated acute kidney injury(CSA-AKI),the most prevalent major complication of cardiac surgery,exerts a negative effect on a patient’s prognosis,thereby usually leading to mortality.Although several risk assessment models have been developed for patients undergoing cardiac surgery,their performance is unsatisfactory.In this study,a machine learning algorithm was employed to obtain better predictive power for CSA-AKI outcomes relative to statistical analysis.Methods: Patients admitted to the ICU(aged 18 years or above)after cardiac surgery by cardiopulmonary bypass(CPB)at the Department of Cardiothoracic Surgery of Nanjing First Hospital between December 1,2019,and April 30,2020 were enrolled in this study.In addition,random forest(RF),logistic regression with LASSO regularization,extreme gradient boosting(Xgboost),and support vector machine(SVM)methods were employed for feature selection and model training.Moreover,the calibration capacity and differentiation ability of the model were assessed using net reclassification improvement(NRI)along with Brier scores and receiver operating characteristic(ROC)curves,respectively.Continuous variables were compared by two-tailed t-tests,whereas Fisher’s exact test was used for categorical data.The significance level was set at P < 0.05 unless specified otherwise.Results: A total of 215 patients underwent cardiac surgery under CPB between December 1,2019,and April 30,2020,of which 135 were enrolled in this study.According to the KDIGO clinical practice guidelines,44 patients suffered from hospital-acquired AKI within 1 week of cardiac surgery.Fatty acid-binding protein(FABP),hemojuvelin(HJV),neutrophil gelatinase-associated lipocalin(NGAL),mechanical ventilation time,and troponin I(Tn I)were significantly correlated with the incidence of AKI.RF was the best model for predicting AKI(Brier score: 0.137,NRI: 0.221),evidenced by an AUC value of 0.858(95% confidence interval[CI]: 0.792–0.923).Conclusion: We developed a model that integrated advanced machine learning algorithms and easily accessible patient characteristics to predict the risk of developing CSA-AKI among patients undergoing cardiac surgery.The model may provide powerful assistance to clinicians in identifying patients with a higher risk of AKI early in the postoperative period.Overall,the findings of this study can assist in developing timely diagnostic and treatment strategies for the clinical management of patients undergoing cardiac surgery.Part Ⅲ: Early diagnostic value of urinary NGAL,HJV,and Dickkopf-3 combined with Cleveland score in CSA-AKIBackground and purpose: Cardiac surgery-associated acute kidney injury(CSA-AKI)is the second most common cause of AKI,and a novel biomarker can detect tubular damage earlier than serum creatinine when it occurs.In this study,by observing the dynamic changes of NGAL,HJV,and DKK-3 in urine after cardiac surgery,combined with the Cleveland scoring system,the early predictive value of NGAL,HJV,DKK-3,and related factors in urine for CSA-AKI were explored along with the diagnostic efficacy.Then,the diagnostic value of these three novel predictive biomarkers was evaluated to identify the occurrence of CSA-AKI determined earlier.Validation of a multivariate model for predicting CSA-AKI,based on preoperative and combined preoperative and intraoperative clinical information,was then developed.Methods: Patients admitted to the ICU(aged 18 years and above)after cardiac surgery with cardiopulmonary bypass(CPB)in the Department of Cardiothoracic Surgery of Nanjing First Hospital from December 1,2019 to April 30,2020 were included in this study.Urine samples were collected at 5 time points: 0 h,3 h,6 h,12 h,and 24 h,and the contents of NGAL,HJV,and DKK3 in the urine samples were detected by ELISA.Cleveland score combined with urinary biomarkers(NGAL,HJV,and DKK3)to build a predictive model,and the predictive power of the CSA-AKI risk model was evaluated by applying the area under the receiver operating characteristic curve(AUROC)and Hosmer-Lemeshow goodness-of-fit test(H-L Test)to evaluate the fitting effect of the prediction model.Results: Between December 1,2019,and April 30,2020,a total of 215 patients underwent cardiac surgery under CPB,of whom 135 were enrolled in this study.According to the KDIGO clinical practice guidelines,44 patients developed hospital-acquired AKI within 1 week after cardiac surgery.Three urinary marker levels(DKK3,HJV,and NGAL)can serve as reliable risk predictors for developing CSA-AKI in cardiac surgery patients.NGAL combined with theCleveland model has both good sensitivity(0.432)and satisfactory specificity(0.890),and it can be used as a clinical prediction model for CSA-AKI in patients undergoing cardiac surgery.Conclusions: From this study,we concluded that,unlike traditional markers of kidney injury(such as creatinine),DKK3,HJV,and NGAL were able to identify CAS-AKI earlier.Among the15 predictive models,after excluding overfitting models,we found that NGAL combined with the Cleveland scoring model has both good sensitivity and satisfactory specificity,and can be used as a clinical predictive model for CSA-AKI in cardiac surgery patients.This model is based on reliable clinical information,which can be used in CSA-AKI risk assessments to guide the selection of clinical treatment strategies.However,external validation of this predictive model in other cohorts is still required before large-scale application is possible.