Influencing factors for atrial fibrillation in elderly patients with heart failure and construction of a nomogram model
Su Yakun
Lin Tingting
Wang Qiong
Li Fengde
Song Junying
Abstract：Objective To elucidate the independent risk factors for AF in elderly patients with HF,develop a nomogram prediction model,and assess its predictive value for AF in elderly HF pa-tients.Methods A retrospective study was conducted on 299 patients of first HF hospitalization in Hengshui People's Hospital from June 2019 to December 2023,and based on the presence or absence of AF,they were divided into an AF group(133 cases)and a non-AF group(166 cases).The patients admitted from June 2019 to December 2022 were assigned into a modeling cohort(206 cases),while those from January 2023 to December 2023 into a validation cohort(93 cases).The general data were collected.Multivariate logistic regression analysis was performed on the modeling cohort to identify the independent predictors of AF in elderly HF patients,and a nomo-gram prediction model was constructed.ROC curve,calibration curve and clinical decision curve analyses were applied to evaluate the discrimination,calibration and clinical practicability of the prediction model.Results Multivariate logistic regression analysis revealed that the etiology of HF(CHD vs HP,OR=0.610,95％CI:0.301-1.250,P=0.178;CMD vs HP,OR=0.213,95％CI:0.052-0.883,P=0.033),LAD(OR=1.081,95％CI:1.020-1.152,P=0.015),HF classifi-cation(HFmrEF vs HFrEF,OR=5.897,95％CI:2.448-14.201,P=0.000;HFpEF vs HFrEF,OR=7.211,95％CI:2.522-20.640,P=0.001),pre-albumin(PALB)(OR=0.438,95％CI:0.217-0.901,P=0.030),UAC(OR=2.186,95％CI:1.075-4.430,P=0.025),and direct biliru-bin(DBi)(OR=4.531,95％CI:2.052-9.990,P=0.000)were independent risk factors for AF in the elderly HF patients.ROC curve analysis showed the AUC value of the prediction model based on these factors in the modeling cohort was 0.831,and the AUC value in the validation cohort was 0.840.Decision curve analysis indicated that the model possessed clinical practicability within a probability threshold range of 10％-82％for the modeling cohort and 12％-100％for the vali-dation cohort.Conclusion Our nomogram prediction model based on multivariate logistic regres-sion analysis exhibits good predictive value for the occurrence of AF in elderly HF patients,and can facilitate clinical decision-making on diagnosis and treatment.
Keywords：agedheart failureatrial fibrillationroot cause analysisnomograms
Publication Date：2025-08-15
Online Publishing Date：2025-08-21(First online date of this platform, not the publication date of the document)
Pages：6( 1010-1015 )