Background: Breathlessness is a troublesome and prevalent symptom in the general population and there is a need to identify the factors most strongly associated to the symptom.

Aims and objectives: The aim of this study was to identify the factors most strongly associated to breathlessness in the general population.

Methods: A cross-sectional analysis of data from the population-based Swedish CArdioPulmonary bioImage Study (SCAPIS) of adults 50-64 years. Breathlessness was defined as ?2 modified Medical Research Council (mMRC) breathlessness scale. The machine learning algorithm Extreme gradient boosting (XGBoost) was used to classify participants as either breathless or non-breathless using 465 different factors including physiological measurement, blood samples, computer tomography, lifestyle, health conditions, and socioeconomics. Strength of association to breathlessness was measured with SHapley Additive exPlanations (SHAP) scores (comparable to log-odds; increased score means stronger association).

Results: A total of 14801 (52%) women and 13929 men were included in the present analysis in which 4% were breathless. The most strongly associated factors to breathlessness were (SHAP score) body mass index (0.39), forced expiratory volume in 1 second (0.32), physical activity measured by accelerometery (0.27), sleep apnoea (0.22), diffusion capacity for carbon monoxide (0.21), self-reported physical activity (0.17) chest pain when hurrying (0.17).

Conclusions: The identified factors could be important targets for public health interventions aiming to reduce the long-term breathlessness in the general population.