The Predictive Value Of Three Scoring Systems On The Risk Of In-hospital Hemorrhage After PCI For ACS And Discussion On Influencing Factors Of Potential Hemorrhage

ObjectiveTo evaluate the value of three systems,namely CRUSADE,PARIS,and PRECISE-DAPT,in predicting the risk of in-hospital hemorrhage in ACS patients receiving PCI,and discuss the the potential predictors of bleeding by assessing the baseline characteristics and operation related factors.MethodsWe retrospectively analyzed 1 384 consecutive patients with ACS received PCI attending.In this study,End point of bleeding event was defined as BARC non–CABG-related bleeding≥type 2(except type 4).The baseline characteristics of all patients and the occurrence of bleeding during hospitalization were recorded,and the predictive value of three scoring systems for nosocomial bleeding events after intervention was evaluated by drawing a ROC curves.Logistic multivariate regression model was used to determine predictors of perioperative bleeding events in ACS patients.Results1.In the selected 2 278 patients diagnosed with ACS,1 384 were treated with PCI and finally enrolled after applying the exclusions.252 patients of the 1384 patients had non-CABG-related bleeding events of type BARC≥2(except type 4).There were 128 cases of type 2,114 cases of type 3 and 10 cases of type 5.There were 62 cases with gastrointestinal hemorrhage(24.60%),52 cases with puncture site hemorrhage and hematoma(20.63%),48 cases with blood loss ≥30g/L(19.05%),38 cases with cutaneous mucosal hemorrhage(15.08%),26 cases with respiratory tract hemorrhage(10.32%),14 cases with urinary tract bleeding(5.56%),6 cases with pericardial tamponade(2.38%),and bleeding resulted in 2 deaths(0.79%).2.Comparion of bleeding scores among three series in bleeding groupWhen BARC≥type 2(except type 4)was used as the end point of hemorrhage,CRUSADE score(32.78±13.89),PARIS score(5.02±2.26),and PRECISE-DAPT score(18.88±10.86)of patients with hemorrhage were significantly higher than those without hemorrhage(22.36±12.45),(3.82±1.72),and(12.69±7.51)(all P<0.001),respectively.3.Univariate analysis of factors related to bleeding eventsIn the bleeding group,the majority of patients were female,older,with lower average blood pressure,lower creatinine clearance rate,higher white blood cell count,and lower baseline hemoglobin.More patients were associated with decreased ejection fraction,and most had a history of hypertension and diabetes.The incidence of bleeding events in the patients was associated with a number of therapeutic and perioperative factors,such as the use of tirofiban,trans-femoral artery approach,intraoperative intra-aortic balloon retropulsation,intraoperative hemorrhage hypotension,and the duration of operation > for 1 hour,all of which had statistically significant differences(P<0.05 or P<0.001).4.Logistic regression analysis of the bleeding event as the dependent variableAccording to the CRUSADE score,the patients were divided into five groups.Logistic regression analysis showed that the risk of bleeding increased along with the increase of the CRUSADE score.The risk of bleeding in group B was 2.031 times higher than in group A.Group C was 2.047 times higher than in group A.Group D was 2.441 times higher than in group A,and group E was 4.885 times higher than in group A.According to the PARIS score,the patients were divided into three groups.Logistic regression analysis showed that the risk of bleeding increased with the increase of PRECISE-DAPT score.The risk of bleeding in group B was 1.801 times higher than in group A,and the risk of bleeding in group C was 5.053 times higher than in group A.According to the PRECISE-DAPT score,the enrolled patients were divided into four groups.Logistic regression analysis showed that the risk of bleeding increased with the increase of PRECISE-DAPT score.The risk of bleeding in group B was 1.434 times higher than in group A.The risk of bleeding in group C was 2.452 times higher than in group A,and the risk of bleeding in group D was 4.267 times higher than in group A.5.ROC curve analysis on predictive ability for bleeding eventsUsing BARC≥type 2(except type 4)as the bleeding criteria,the area under the curve of the CRUSADE,PARIS and PRECISE DAPT score was 0.694(95%CI: 0.669-0.718),0.660(95%CI: 0.634-0.685),0.679(95%CI: 0.654-0.704),respectively.The results showed that all three scores are valuable on the assessment of in-hospital hemorrhage in patients with ACS after PCI.No differences were observed in the ability of the three scores to predict nosocomial bleeding events in patients with ACS.6.Logistic regression analysis of bleeding eventsPredictors of P<0.05 in univariate analysis were included in the multivariate regression model,and potential predictors of bleeding were selected(P<0.05).Binary Logistic regression was used for backward selection(P<0.05)and hosmer-lemeshow fitting degree was used.Ultimately,there were 10 bleeding predictors that predicted the risk of bleeding during hospitalization,namely,history of hypertension,diabetes,hypotension,white blood cell count,hemoglobin,creatinine clearance,the use of tirofiban,femoral artery entry,intra-aortic balloon retropulsation(IABP),length of procedure>1h.HGB and Ccr are protective factors for in-hospital hemorrhage.Conclusions1.All three scores have predictive value for nosocomial bleeding events in the patients of our study.They can be used for hazard layering.No differences were observed in the ability of the three scores to predict nosocomial bleeding events in patients with ACS.2.Based on population analysis in this study,predictors of in-hospital hemorrhage included history of hypertension,history of diabetes,hypotension state,white blood cell count,hemoglobin,creatinine clearance,use of tirofiban,femoral artery entry,IABP,and length of procedure>1h.HGB and Ccr are protective factors for in-hospital hemorrhage.