Establishment And Verification Of Prognostic Nomogram For Patients With Acute Coronary Syndrome

Background and objective:Current risk model for long-term prediction in coronary artery disease(CAD)is complicated while a simple useful model is still lacking.In this study,we compared the predictive value of the CHADS2 score and its derived scores in patients with coronary artery disease.Method:We enrolled 3700 patients with CAD between November 2010 and September 2014 at department of cardiology from PLA general hospital.The CHADS2 score,CHA2DS2-VASc score and R2CHADS2 scores were calculated.All cases were followed to track the incidence of composite endpoint consisting of cardiovascular death,non-cardiovascular death,MI,stroke,heart failure.Cox proportional hazard regression analyses were performed to evaluate the independent risk factor of composite endpoint.Furthermore,receiver operator characteristic(ROC)curve was.generated to compare the predictive ability ofCHADS2 score and its derived scores.Results:3633 patients were included in the final analysis.The median age was 61 years and the average age was 61.5 ± 11.7 years.Among them,72.3%of patients were men.Average BMI was 25.6 ± 3.4kg/m2,and more than a half(70.7%)were overweight.65.3%of patients have high blood pressure.During a median 2.9-year follow-up,443 patients experienced at least 1 element of the composite endpoint of cardiovascular death(n=168[4.6%]),MI(n=59[1.6%]),stroke(n=96[2.6%]),heart failure(n=101[2.8%]),and all-cause death(n=240[6.6%]).With the increase of CHADS2 score,CHA2DS2-VASc score and R2CHADS2 score,the composite clinical events,stroke,cardiovascular death,and non-cardiovascular death all gradually increased(p<0.05).Multivariate COX proportional hazards regression analysis showed that after adjusting for other variables,the CHADS2 score was independently associated with composite outcome(HR:2.14,95%CI:1.96,2.34,p<0.001),stroke(HR:8.46,95%CI:6.84,10.47,p<0.001)and cardiovascular death(HR:1.48,95%CI:1.27,1.72).The CHA2DS2-VASc score was independently associated with composite outcome(HR:1.50,95%CI:1.25,1.score was independently associated with composite outcome(HR:1.43,95%CI:1.35,1.53,p<0.001),stroke(HR:1.46,95%CI:1.27,1.67,p<0.001)and cardiovascular death(HR:1.42,95%CI:1.28,1.58,p<0.001).The R2CHADS2 80,p<0.001),stroke(HR:3.61,95%CI:3.12,4.17,p<0.001)and cardiovascular death(HR:1.46,95%CI:1.32,1.62,p<0.001).Receiver operating characteristic analysis showed that the R2CHADS2 score had the highest predictive ability for composite outcome in patients with coronary heart disease(AUC:0.791),followed by CHADS2 score(AUC:0.772),and CHA2DS2-VASc score had the lowest predictive ability(AUC:0.732).The differences among the three scores were statistically significant(p<0.05).Conclusion:1.CHADS2 score,CHA2DS2-VASc scoreand R2CHADS2 score have predictive value for the occurrence of composite clinical events,stroke,and cardiovascular death in patients with coronary heart disease.2.After adjusting for other variables,R2CHADS2 score has the highest predictive abilityfor the occurrence of composite clinical events and cardiovascular death in patients with coronary heart disease,whileCHADS2 score has the highest predictive ability for stroke in patients with coronary heart disease.3.CHADS2 score,CHA2DS2-VASc score and R2CHADS2 score couldpredict the long-term composite events and stroke of patients with stable coronary heart disease and ACS.4.The CHADS2 score and derived scores are simple and easy to apply,and can be used as an effective tool for predicting long-term outcome for patients with coronay artery disease.Objective:Current risk model for long-term prediction in acute coronary syndrome(ACS)is highly variable.In this study,we sort to develop a prognostic nomogram for clinical outcome in patients with acute coronary syndrome.Method:We enrolled 2726 patients with ACS between November 2010 and September 2014 at department of cardiology from our hospital.All cases were followed to track the incidence of composite end point consisting of cardiovascular death,MI,stroke,heart failure,and all-cause death.We used the random forest technique to identify the top 11 predictors of outcome.Then a nomogram predicting clinical outcome was developed based on COX proportional hazards model.To analyze the predictability of nomogram,the C-index was calculated.The calibration plot was also used to determine calibration.Results:During a median 2.9 year follow-up,337 patients experienced at least 1 element of the composite end point of cardiovascular death(n=147[5.4%]),non-cardiovascular death(n=16[0.6%]),MI(n=57[2.1%]),stroke(n=34[1.2%]),and heart failure(n=27[1.0%]).The nomogram include 11 variables:NT-proBNP,hemoglobin,creatinine,D-dimmer,GDF-15,LVEF,LVEDs,age,cTnT,fibrinogen and previous stroke.We performed receiver-operating characteristic analysis to determine the predictability of the nomogram to combined outcome.The c-indexwas 0.825.The nomogram model had higher predictive ability for clinical adverse events in patients with ACS than the GRACE score(AUC:0.825 vs 0.739,p<0.0001).Conclusion:The proposed nomogram provides a useful tool in predicting long-term outcome for patients with acute coronary syndrome.