A risk prediction model for repeat revascularization in patients with acute coronary syndrome after per-cutaneous coronary intervention

Wang Guangdong; Yao Qiong; Wang Yuan

doi:10.3969/j.issn.1674-4055.2025.05.11

Abstract：Objective To analyze risk factors for repeat revascularization in patients with acute coronary syndrome(ACS)after percutaneous coronary intervention(PCI),and to establish a risk prediction model.Methods ACS patients received PCI(n=475)were selected and divided,according to repeat revascularization occurrence with 2 y after PCI,into revascularization group(n=70)and non-revascularization group(n=405)in the First People's Hospital of Tianshui City from Dec.2020 to June 2022.The clinical data were collected in all patients.The variables affecting repeat revascularization were screened by using single-factor analysis and LASSO regression analysis,and independent risk factors were determined by using multi-factor Logistic regression analysis.A risk prediction model for post-PCI repeat revascularization ACS patients was established by using RStudio 4.4.1 software,the consistency and net benefit of the risk prediction model were evaluated by using calibration curves,decision curves and C-index,and goodness of fit of risk prediction model was examined by Hosmer-Lemeshow test.The predictive efficiency of the risk prediction model was reviewed by using ROC curve.Results After followed up for 2 y,the rate of repeat revascularization was 14.74％(70/475)in 475 patients.The results of single-factor analysis showed that age,smoking history,number of lesion vessels,hyperlipidemia,type 2 diabetes mellitus(T2DM),Gensini integrals,stent numbers,left ventricular ejection fraction(LVEF),high-density lipoprotein-cholesterol(HDL-C)and low-density lipoprotein-cholesterol(LDL-C)were correlated to post-PCI repeat revascularization(P＜0.05).There were 7 variables with non-zero characteristics by LASSO regression analysis,which were included into multi-factor Logistic regression analysis,and 6 independent risk factors were determined,including smoking history(OR=2.700,95％CI:1.366～5.336),multi-vessel lesions(OR=3.288,95％CI:1.420～7.616),hyperlipidemia(OR=3.538,95％CI:1.745～7.172),T2DM(OR=4.368,95％CI:2.150～8.875)Gensini integrals(OR=1.065,95％CI:1.042～1.088),and LDL-C(OR=3.906,95％CI:2.133～7.154,P＜0.05).A risk prediction model for post-PCI repeat revascularization was established[Logit(P)=-11.931+1.008 × smoking history+1.176 × multi-vessel lesions+1.218 × hyperlipidemia+1.251 × T2DM+0.063 × Gensini integrals+1.371 × LDL-C],and C-index of this risk prediction model was 0.877(95％CI:0.873～0.888).The calibration curve of the risk prediction model closely approximated ideal curve,decision curve was above 2 extreme curves,and Hosmer-Lemeshow test showed P＞0.05.The results of ROC curve analysis showed that AUC of the risk prediction model was 0.877(95％CI:0.834～0.919),sensitivity was 74.29％and specificity was 87.16％in predicting post-PCI repeat revascularization.Conclusion Smoking history,multi-vessel lesions,hyperlipidemia,T2DM,Gensini integrals and LDL-C are independent risk factors of post-PCI repeat revascularization.The risk prediction model established based on these risk factors has a higher predictive efficacy in ACS patients after PIC.