Galectin-3,PGC-1α,high sensitivity C-reactive protein and senile coronary heart disease complicated by atrial fibrillation:a multi-model prediction study

Xie Hui; Cai Yongfang; Song Wei; Xiangchang Guangzi; Xu Dejun

doi:10.3969/j.issn.1674-4055.2025.06.09

Abstract：Objective To study a multi-model prediction of Galectin-3(Gal-3),peroxisome proliferator-activated receptor gamma coactivator 1α(PGC-1α),high sensitivity C-reactive protein(hs-CRP)and senile coronary heart disease(CHD)complicated by atrial fibrillation(AF).Methods CHD patients(n=120)were selected from Affiliated Hospital of Hubei Three Gorges Polytechnic from Feb.2022 to Feb.2023.According to results of electrocardiogram(ECG)examination,CHD patients complicated by AF were selected into AF group(n=45),and those without AF were selected into non-AF group(n=75).Clinical data was collected and compared in 2 groups.Single-factor screening was performed for AF occurrence of in elderly CHD patients.After multi-factor Logistic regression analysis,a nomogram prediction model and a ROC curve were conducted.Results Age,levels of Gal-3 and hs-CRP and percentages of hypertension and hyperlipidemia were significantly higher,and PGC-1α was significantly lower in AF group than those in non-AF group.The results of multi-factor Logistic regression analysis showed that age≥65,higher Gal-3,lower PGC-1α and higher hs-CRP were risk factors for AF occurrence in elderly CHD patients(P<0.05).The results of ROC curve analysis showed that AUC of age,Gal-3,PGC-1α,hs-CRP and combined indexes was,respectively,0.684,0.745,0.816,0.881 and 0.974 in predicting AF occurrence,which indicated that above indexes had some predictive value to AF occurrence in elderly CHD patients.Based on the Bootstrap method,internal validation with 1000 resamplings of the original data showed that Bias-corrected prediction curve closely aligned with Ideal line,indicating that the model has good predictive and application value.The decision curve analysis of the model demonstrated a net benefit>0 across a threshold probability range of 0.01～1.00.Conclusion The risk factors for AF occurrence in elderly CHD patients include age≥65,higher Gal-3,lower PGC-1α and higher hs-CRP.The multi-model construction demonstrates that the aforementioned biomarkers have certain application value in predicting such diseases,which can provide new insights for the prediction and intervention of CHD combined by AF in elderly patients.