Establishment And Comparison Of Risk Prediction Models For Patients With Nonvalvular Atrial Fibrillation And Acute Coronary Syndrome Who Underwent Percutaneous Coronary Intervention

Objective:Non valvular atrial fibrillation(NVAF)and acute coronary syndrome(ACS)often coexist.Percutaneous coronary intervention(PCI)was an important treatment for ACS patients.patients with nonvalvular atrial fibrillation and acute coronary syndrome who underwent percutaneous coronary intervention has a higher risk of embolism and bleeding.There were many contradictions in clinical decision-making and poor prognosis.Risk assessment played an important role in management.Many risk scoring systems had been established in the ACS or AF population,but their predictive value in this special population was not clear.We collected a multicenter cohort of this population in Beijing,compared the predictive value of common cardiovascular scoring systems,and tried to establish a new nomograms prediction model.Methods:1.From December 2009 to June 2015,2429 patients with nonvalvular atrial fibrillation and acute coronary syndrome who underwent percutaneous coronary intervention were collected from 12 hospitals in Beijing,The patients were followed up from March 2016 to June 2016.2.The corresponding scores were calculated according to the definitions of CRUSADE and HAS-BLED,the two scores were compared to predict the major bleeding risk in hospital and long-term follow-up.3.TIMI,GRACE,CRUSADE,C-ACS,proACS,CHA2DS2-VASC,HAS-BLED,TIMI-AF scores were calculated according to the definition,and their predictive value for in-hospital and long-term death was compared in this cohort.4.Combination of GRACE Score and TIMI-AF Score to Predict In-Hospital and Long-term Mortality Risk in this cohort.5.Based on the multivariate Cox regression analysis of this cohort,nomogram model was constructed using R language to predict the short-term and long-term mortality risk of this population.Results:1.Hospitalization was 9.35±7.61 days.There were 23 patients with major bleeding events in the hospital,and 34 patients died in the hospital.During the follow-up period of 3.19±1.52 years,45 major bleeding events were observed,and 270 patients died.2.To assess in-hospital major bleeding risk,the AUC of CRUSADE score was 0.737,and the HAS-BLED score was 0.682.To assess long-term major bleeding risk,the AUC of CRUSADE score was 0.693,and the HAS-BLED score was 0.649.3.GRACE,C-ACS,ProACS,and CRUSADE score had similar good predictive performance for in-hospital mortality(C value,0.813-0.868),superior to other scores(C value,0.660-0.780).For long-term all-cause death GRACE,ProACS,CRUSADE,and TIMI-AF had similar predictive performance(C value,0.659-0.698),superior to other scores(C value,0.572-0.635).4.To assess in-hospital death risk,the AUC of combination of GRACE and TIMI-AF score was 0.843.To assess long-term mortality risk,the AUC of combination of GRACE and TIMI-AF score was 0.683.5.Independent factors derived from Cox multivariable analysis of the primary cohort to predict all-cause mortality were age,pattern of ACS,red blood cell distribution width,N-terminal proBNP,and serum creatinine,all of which were assembled into the nomogram.In the modeling cohort,The C-index of the nomogram for predicting mortality was 0.764,which was statistically higher than the C-index values for the current risk scores.In the validation cohort,the C-index of the nomogram for predicting all-cause death was 0.706,with no significant differences compared with GRACE and CRUSADE,but better than that of CHA2DS2-VASC and HAS-BLED.Conclusion:This is a large cohort of patients with atrial fibrillation combined with ACS by PCI intervention in Beijing.in this cohort,CRUSADE score and HAS-BLED score can predict in-hospital and long-term risk of major bleeding,CRUSADE score maybe better than HAS-BLED score.GRACE,ProACS,and CRUSADE score showed good and similar predictive performance for in-hospital mortality and long-term mortality;however,ProACS is relatively simple.The combination of the GRACE and TIMI-AF score can improve the ability to predict the long-term risk of death.For long-term all-cause death,Nomogram model based on this cohort is relatively simple and has good prognostic prediction,it should be further validated.