Development And Validation Of A Prediction Model Of Postoperative Acute Heart Failure After Percutaneous Coronary Intervention

BackgroundThe emergence of acute heart failure(AHF)after percutaneous coronary intervention(PCI)leads to a severely worsened prognosis.However,there are few prediction models for post-PCI AHF among patients undergoing PCI.Therefore,we aim to establish a new prediction model of postoperative AHF for patients following PCI.MethodsPatients:Development group:2091 consecutive patients who underwent PCI in Guangdong Provincial People’s Hospital were enrolled.Validation group:2570 patients in Guangdong,Fujian,Xinjiang and other multi-center hospitals underwent PCI were enrolled.Endpoints:Primary endpoint:Postoperative acute heart failure after PCI.Secondary endpoint:All-cause death in the follow-up.Statistical methods:Normally distributed continuous variables were compared with an unpaired,2-tailed t test and are expressed as the mean ± SD,while nonnormally distributed variables were compared through the Wilcoxon rank-sum test and are expressed as the median± interquartile range.Categorical variables were evaluated using the x2 test or Fisher’s exact test and expressed as percentages.Kaplan-Meier curves were used to explore the association between the postoperative AHF and long-term mortality.The associations between variables were assessed by univariable logistic analysis.Stepwise regression variable selection was conducted using a cutoff of p<0.1 in univariate logistics regression.The strength of the relationship was expressed in logistic regression via odds ratios(ORs)with 95%confidence intervals(95%CIs).Collinearity between variables was also evaluated.A nomogram was then formulated based on the results and by using the rms package of R.The discrimination ability of the nomogram was assessed using the area under the receiver operating characteristic(ROC)curve.The Hosmer-Lemeshow test was conducted to assess the calibration for our model.Its stability is evaluated in an external verification cohort.ResultThe development group finally enrolled 2091 patients who met the admission criteria,with an average age of 63.3 years.Female accounted for 19.2%and acute myocardial infarction accounted for 40.2%.Overall,the incidence of postoperative AHF was 3.67%(n=74).The median follow-up time was 7.3 years(interquartile range:5.5-8.7 years).The long-term all-cause mortality of patients with postoperative AHF was significantly higher than that of patients without AHF(45.9%vs 10.4%,P<0.01).After adjusted for some risk factors,postoperative AHF is the strongest predictor of long-term all-cause death(hazard ratio:3.11;95%confidence interval:1.31-3.88;P=0.01).In the development group,the predictive model included 5 predictors(age,acute myocardial infarction,chronic kidney disease,left ventricular ejection fraction,preoperative ventricular rate).This model had good predictive ability(C=0.85,95%confidence interval:0.78-0.91,P<0.01)and calibration(Hosmer-Lemeshow test:P=0.76).In the external validation group,the nomogram prediction model constructed by the modeling group has good distinguishing ability(C=0.84,95%confidence interval:0.78-0.90,P=0.86)and calibration(Hosmer-Lemeshow test:P=0.10),which shows that the model is stable and suitable for external use.Conclusion1.The occurrence of postoperative AHF after PCI is closely related to the long-term prognosis.2.Based on five common clinical indicators,the simple model for predicting post-PCI AHF exhibited stable performance and good clinical implications in risk stratification for high-risk individuals.