Breath analysis enables prediction of atherosclerotic cardiovascular disease risk

Atherosclerotic cardiovascular disease (ASCVD) is a leading cause of global mortality and disease burden, necessitating simple, rapid community screening methods. Exhaled breath analysis offers a promising non-invasive approach for identifying high-risk ASCVD patients.

This cross-sectional study involved 1,790 participants from a China-based community cohort. Comprehensive assessments included demographic data, fasting venous blood analysis for glucose and lipids (TC, TG, LDL-C, HDL-C), and stratification into high-risk (10-year ASCVD risk ≥5%) and low-risk (<5%) groups using the China-PAR model. Exhaled breath samples were collected after 12-hour fasting using standardized devices, with volatile organic compounds (VOCs) profiled via high-performance photon ionization time-of-flight mass spectrometry (HPPI-TOFMS). The cohort was randomly split into discovery and tuning sets for feature selection and model testing, respectively, using random forest algorithms with cross-validation.

Of 1,442 eligible participants (1,040 high-risk, 402 low-risk), 402 matched pairs were created via propensity score matching (PSM). The developed ASCVD risk prediction model, integrating clinical risk factors and breath VOCs, showed strong performance in distinguishing high-risk individuals (SEN=78.5%, SPE=79.3%, ACC=78.9%, AUC=0.867). Additionally, ten differentially expressed VOC ions with significant concentration variations between high- and low-risk groups were identified.

This study establishes a novel framework for community-based ASCVD risk screening using exhaled breath biomarkers. While demonstrating breath analysis's potential as a non-invasive tool, further clinical validation is needed to confirm its predictive accuracy for individual-level ASCVD risk assessment.