Liu R; Beijing Friendship Hospital, China and other centres in USA and Brazil
Arena R; Santos-de-Araujo AD; Borghi-Silva A; Popovic D

Current Problems in Cardiology. 51(10):103399, 2026 Oct.

BACKGROUND: Atherosclerotic cardiovascular disease remains a leading cause
of morbidity and mortality worldwide. Established risk equations guide
prevention but rely mainly on static, clinic-based variables and
incompletely capture physical activity, sedentary behavior, and
cardiorespiratory fitness.

METHODS: This narrative review synthesizes evidence from epidemiological
studies, clinical trials, and methodological frameworks on the roles of
physical activity, sedentary behavior, and cardiorespiratory fitness in
cardiovascular risk prediction, with emphasis on digital phenotyping,
mechanistic exercise physiology, and multimodal clinical prediction
models.

RESULTS: Cardiorespiratory fitness reflects integrated physiological
reserve and is strongly associated with cardiovascular and all-cause
outcomes. Cardiopulmonary exercise testing extends fitness assessment by
identifying mechanisms of exercise limitation, including cardiac,
ventilatory, autonomic, pulmonary vascular, and peripheral contributors.
Wearable technologies provide longitudinal, real-world measures of
physical activity and sedentary behavior that complement static clinical
risk factors. Multimodal models integrating electronic health records,
electrocardiography, imaging, wearable signals, and exercise testing may
support more personalized and actionable risk stratification. However, the
current evidence base remains limited by insufficient external and
prospective validation, incomplete calibration reporting, limited
decision-analytic evaluation, and inadequate assessment of subgroup
performance and equity.

CONCLUSIONS: Physical activity and cardiorespiratory fitness should be
considered clinically relevant and modifiable phenotypes in contemporary
cardiovascular risk prediction. Future models should prioritize
standardized measurement, mechanistic validation of digital phenotypes,
calibration, external validation, subgroup evaluation, workflow
feasibility, and prospective evidence of clinical utility before
widespread implementation.