Nomogram model for screening the risk of frailty in older adult atrial fibrillation patients: a cross-sectional study

Frailty is common in atrial fibrillation (AF) patients, but the specific risk factors contributing to frailty need further investigation. There is an urgent need for a risk prediction model to identify individuals at high risk of frailty.

This cross-sectional study aims to explore the multiple risk factors of frailty in older adult patients with AF and then construct a nomogram model to predict frailty risk.

We recruited 337 hospitalized patients over the age of 60 (average age: 69, 53.1% male) with AF between November 2021 and August 2022. Data collected included patient demographics, disease characteristics, sleep patterns, mental health status, and frailty measures. We used LASSO and ordinal regression to identify independent risk factors. These factors were then incorporated into a nomogram model to predict frailty risk. The model's performance was assessed using the concordance index (C-index) and calibration curves.

Among the AF patients, 23.1% were classified as frail and 52.2% as pre-frail. Six risk factors were identified: age, gender, history of coronary heart disease, number of chronic conditions, sleep disruption, and mental health status. The internal validation C-index was 0.821 (95% CI: 0.778-0.864; bias-corrected C-index: 0.795), and the external validation C-index was 0.819 (95% CI: 0.762-0.876; bias-corrected C-index: 0.819), demonstrating strong discriminative ability. Calibration charts for both internal and external validations closely matched the ideal curve, indicating robust predictive performance.

The nomogram developed in this study is a promising and practical tool for assessing frailty risk in AF patients, aiding clinicians in identifying those at high risk.

This study demonstrates the utility of a comprehensive predictive model based on frailty risk factors in AF patients, offering clinicians a practical tool for personalized risk assessment and management strategies.