# Enhancing atrial fibrillation risk prediction in an observational cohort of tobacco-exposed individuals: the role of pulmonary function tests, symptom scores, and imaging

## TL;DR

This study finds that adding lung function and imaging data improves predicting atrial fibrillation risk in people with COPD.

## Contribution

The study introduces new cardiopulmonary variables to enhance AFib risk prediction models in COPD patients.

## Key findings

- DLco, heart volume, emphysema percentage, and mMRC scores were linked to AFib in COPD patients.
- Adding these variables improved the CHARGE-AF model's discrimination from 0.53 to 0.63.
- Lung function and CT imaging data may better predict AFib risk in chronic lung disease populations.

## Abstract

COPD is associated with an increased AFib-related morbidity and mortality. There are several AFib risk prediction models available, but none have been validated in the COPD population. Our study aims to identify spirometric and radiographic variables that are associated with an increased risk of AFib. Secondarily, we hope to determine if these associated variables improve the risk discrimination of established AFib risk prediction models in individuals with COPD.

We evaluated 755 participants from a single center tobacco-exposed cohort at baseline. At this study visit, the following were performed: demographic, medical history, and symptom questionnaires, PFT, and CT imaging. We performed logistic regression analysis to determine cardiopulmonary variables associated with prevalent AFib. The multivariable analysis was adjusted for sex, age, number of pack years, BMI, self-reported heart failure, and anti-hypertensive medication use. Exposure variables that were statistically significant in the logistic regression analysis were added in succession to current AFib risk prediction models, CHA2DS2-VASc and CHARGE-AF, to create updated models. C-statistics were calculated for both risk scores alone as well as with each updated model.

DLco (OR 0.40, CI 0.18–0.86), heart volume (OR 13.12, CI 2.32–74.17), percentage of emphysema (OR 2.77, CI 1.04–7.40), and mMRC (OR 1.17, CI 1.02–1.35) were associated with prevalent AFib in the multivariable logistic regression analysis. When conducting the discrimination analysis of the AFib risk prediction scores, the addition of these cardiopulmonary variables improved CHARGE-AF, from C-statistic 0.53 to 0.63 (p < 0.03).

We identified cardiopulmonary factors associated with an increased risk of AFib in a tobacco-exposed cohort. The incorporation of lung function, CT parameters, and symptom scores in validated AFib prediction models may improve AFib risk discrimination in our chronic lung disease populations.

not applicable.

This study was supported by the National Institute of Health (NIH) National Heart, Lung and Blood Institute (NHLBI) grants 1R01HL128289 (J.B.) and P50HL084948 (F.C.S.).

The online version contains supplementary material available at 10.1186/s12931-025-03366-8.