[An atrial fibrillation prediction model based on quantitative features of electrocardiogram during sinus rhythm in the Chinese population].

• DOI: 10.12122/j.issn.1673-4254.2025.02.02
• Publication Date: 2025-02-20
• Authors: Xiaoqing Zhu; Yajun Shi; Juan Shen; Qingsong Wang; Tingting Song; Jiancheng Xiu; Tao Chen; Jun Guo
• Keywords: Normal Sinus Rhythm; P wave

To develop an early atrial fibrillation (AF) risk prediction model based on large-scale electrocardiogram (ECG) data from the Chinese population. The of multiple ECG records 30 383 patients admitted in PLA General Hospital between 2009 and 2023 were randomly divided into training set internal testing a 7:3 ratio. predictive factors selected using univariate analysis, LASSO regression, Boruta algorithm. Cox proportional hazards regression was used to establish composite incorporating age, gender, score. discrimination power, calibration, clinical net benefits models evaluated area under receiver operating characteristic curve (AUROC), calibration curves, decision curves. cohort included 51.1% male with median age 51 (36, 62) years AF incidence 4.5% (1370/30 383). In model, parameters related P wave QRS complex identified as significant predictors. set, AUROC for predicting 5-year 0.77 (95% CI: 0.74-0.80), which increased 0.81 0.78-0.83) after reclassification improvement 0.123 integrated 0.04 (P<0.05). close diagonal line. Decision analysis showed that benefit higher than across majority threshold probability. quantitative features during sinus rhythm, along can effectively predict population, thus providing low-cost screening tool assessment management.