| Objective The incidence of postoperative complications of cardiac surgery is high,which significantly affects the quality of patient’s life and will lead to unfavorable prognosis.This study aimed to evaluate the ability of biomarkers such as cystatin C(Cys-C)and baseline blood glucose levels to predict short-term complications after cardiac surgery,and establish a predictive model based on these biomarkers.Method A total of 629 patients with heart disease admitted to our hospital between January 2021 and December 2021 were enrolled.Blood samples were collected before,during,and after surgery to determine the concentration of biomarkers.The main clinical endpoints were new postoperative atrial fibrillation,all cause death within 30 days,and postoperative acute kidney injury,and the secondary clinical outcomes were longer ICU stay and longer hospital stay.SPSS and R software were used for statistical analysis.Correlation analysis,single factor analysis and multifactor analysis were used to screen variables affecting outcomes,and prediction models were established and assessed.Result Preoperative levels of baseline blood glucose and cystatin C were independent biomarkers of new-onset postoperative atrial fibrillation and 30-day all-cause mortality.The combined prediction model based on them had good predictive ability,fitting degree,calibration degree and a certain clinical practical value.Cystatin C is an independent prognostic biomarker of postoperative acute kidney injury,prolonged ICU stay and prolonged hospital stay,and the combined prediction model based on cystatin C has good predictive value,discrimination and calibration and clinical practicability.Conclusion We found that CysC and baseline blood glucose were significantly correlated with new-onset atrial fibrillation,death,acute kidney injury,prolonged ICU stay and PLOS after cardiac surgery.The prediction model based on CysC and baseline blood glucose was of certain clinical value.The model was simple,effective and easy to obtain,and could effectively predict short-term postoperative complications.PART 1:Effect of serum cystatin C and baseline blood glucose on postoperative new atrial fibrillation and 30-day mortality and construction of risk modelObjective This study aimed to evaluate the ability of cystatin C(Cys-C)and baseline blood glucose levels to predict new-onset atrial fibrillation(AF)and 30-day mortality after cardiac surgery,and establish a cystatin C-based prediction model for cardiac postoperative outcomes.Method A total of 629 patients with heart disease admitted to our hospital between January 2021 and December 2021 were enrolled.Biochemical indexes including Cys-C and preoperative blood glucose were obtained before surgery,intraoperative indexes including aortic cross-clamp time and cardiopulmonary bypass(CPB)time were recorded.Blood glucose was measured after aorta occlusion and CPB,respectively.Postoperative indicators including cystatin C(Cys-C),high-sensitive cardiac troponin T(hs-cTnT),high-sensitivity C-reactive protein(Hs-CRP),brain natriuretic peptide precursor(BNP)and creatine phosphokinase isoenzymes(CK-MB)were all measured when patients coming into the intensive care unit(ICU).The main clinical end points were postoperative new-onset atrial fibrillation and all-cause death within 30 days after cardiac surgery.The continuous variables except cystatin C were analyzed with the receiver operating characteristic(ROC)by calculating the Jorden index to obtain an optimal truncation value according to the outcome events,and then converted to dichotomous variables.Risk factors for postoperative new-onset atrial fibrillation and 30-day mortality including patients’ age,gender,comorbidities,surgical index,biochemical indicators and other outcome indicators were analyzed by SPSS statistical software,and a prediction model based on these risk factors was established.The discrimination,calibration,and clinical usefulness of the cystatin C and baseline glucose-based model were assessed using ROC curve,calibration plot,and decision curve analysis which were drawn by R software.Result Among the 629 patients undergoing cardiac surgery,new-onset AF occurred in 64 patients(10.2%).Coronary artery disease(CAD),arrhythmia,baseline blood glucose≥4.4mmol/L,preoperative cystatin C level,postoperative BNP≥985.8pg/mL were independent risk factors for postoperative atrial fibrillation.The combined prediction model based on cystatin C and blood glucose had a satisfactory prediction accuracy and moderate clinical prediction value(AUC=0.767,95%CI 0.705-0.828)and good fitting degree and calibration ability(x2=6.968).The decision curve analysis showed that the model was clinical practicability when the probability threshold was between 4.1%and 29.6%.Twenty patients died within 30-day after surgery(3.2%).Preoperative hyperthyroidism,cardiopulmonary bypass time≥121.5min,preoperative cystatin C level and baseline blood glucose≥4.81mmol/L were independent risk factors for 30-day postoperative mortality.The predication ability of the combined prediction model based on cystatin C and blood glucose(AUC=0.827,95%CI 0.725-0.929)was significantly higher than that with single prediction index,and it had a good fitting degree and calibration ability(x2=13.501).The decision curve analysis showed that the model was clinical practicability when the probability threshold was between 1.1%and 60%.Conclusion Preoperative cystatin C and baseline glucose values were independently associated with new-onset atrial fibrillation and 30-day mortality in patients undergoing heart surgery.The model based on preoperative CysC and baseline blood glucose improved the prediction accuracy of postoperative new AF and 30-day all-cause mortality.The model was simple,effective and easy to obtain.It can effectively predict short-term postoperative complications.PART 2:Effect of serum cystatin C on postoperative AKI and length of hospital stay after cardiac surgery and establishment of prediction modelObjective This study intended to evaluate the relationship between postoperative cystatin C level and AKI,extended ICU stay,and length of hospital stay after cardiac surgery,and establish a prediction model and ROC curve based on cystatin C.Method A total of 629 patients with heart disease admitted to our hospital between January 2021 and December 2021 were enrolled.Serous cystatin C,glucose and creatinine concentration were measured prior to surgery.The aortic cross-clamp time,cardiopulmonary bypass time and operation time were recorded during surgery,and blood glucose value of arterial blood was also detected.Blood glucose was measured after aorta occlusion and CPB,respectively.Postoperative indicators including Cys-C,hs-cTnT,BNP,Hs-CRP and CK-MB the time patients coming into the ICU were all measured.Serum creatinine was measured 7 days after operation and urine volume and hemodialysis were also recorded.Primary clinical endpoints included acute renal injury(AKI),extended ICU stay,and extended hospital stay.The continuous variables except cystatin C were analyzed with the ROC by calculating the Jorden index to obtain an optimal truncation value according to the outcome events,and then converted to dichotomous variables.Risk factors including patients’ age,gender,comorbidities,surgical indicators,biochemical indexes and other outcome indicators for postoperative AKI,extended ICU stay,and extended hospital stay were screened by SPSS software,and a prediction model based on these risk factors was established.The discrimination,calibration,and clinical usefulness of the model were assessed using ROC curve,calibration plot,and decision curve analysis which were drawn by R software.Result Postoperative AKI occurred in 436(69.3%)of 629 patients.Male,hypertension,preoperative cystatin C,and postoperative cystatin C were independent predictors for developing postoperative AKI.The predictive power of the postoperative cystatin C-based combined prediction model(AUC=0.797,95%CI 0.762-0.831)was significantly higher than that of the single-prediction index,and it had a good fitting degree and calibration ability(x2=3.217).The decision curve analysis showed that the model was clinical practicability when the probability threshold was>41.6%.Prolonged postoperative ICU stay was observed in 268 patients(42.6%).Alcohol consumption,coronary heart disease,cardiopulmonary bypass time≥117.5min,postoperative cystatin C value and postoperative BNP≥830.3pg/mL were independent risk factors for prolonged ICU stay.The predictive ability of the postoperative cystatin C-based combined prediction model(AUC=0.781,95%CI 0.745-0.917)was significantly higher than that of the single prediction index and it had a good fitting degree and calibration ability(x2=10.061).The decision curve analysis showed that the model was clinical practicability when the probability threshold was between 10.2%and 88.5%.The duration of postoperative hospital stay was prolonged in 77 patients(12.2%).Euro SCORE Ⅱ≥4.5,hypothyroidism,CPB time≥123.5min,and postoperative cystatin C value were independently correlated with duration of hospital stay after cardiac surgery.The combined prediction model,with good predictive accuracy and clinical predictive value,which based on postoperative cystatin C(AUC=0.792,95%CI 0.738-0.846)was significantly higher than that of the independent prediction index and it had a good fitting degree and calibration ability(x2=9.975).The decision curve analysis showed that the model was clinical practicability when the probability threshold was between 21.4%and 58.8%.Conclusion The incidence of postoperative AKI was high in patients with heart disease.Gender,hypertension,preoperative and postoperative serum cystatin C level affected the incidence of postoperative AKI,and the cystatin C-based prediction model had a high accuracy in AKI prediction.Postoperative serum cystatin C were independently associated with longer ICU treatment and longer hospital stay.In addition,the predictive value of the model based on postoperative cystatin C is higher than that of cystatin C alone. |
PDF Full Text Request