Clustering analysis of volatile organic compound biomarkers with tobacco exposure and the association with cardiovascular health outcomes in an observation study cohort.

Abstract

Introduction: Volatile organic compounds (VOCs) are toxic compounds found in tobacco smoke. Despite research on cigarette generated single VOCs, scant evidence exists on the mixtures of VOCs associated with different tobacco products. We aimed to explore whether distinct VOC exposure profiles exist among users of combustible cigarettes, e-cigarettes, and non-users, and to assess their associations with cardiovascular (CV) health markers.

Methods: Participants who self-reported use of e-cigarettes, cigarettes, or no tobacco (n=348; mean age 26 ± 7 years) enrolled in The Cardiovascular Injury due to Tobacco Use (CITU) 2.0 study from July 2018 to July 2023 at two US sites (Boston, MA, and Louisville, KY). VOC metabolites were analyzed in urine one-hour post-use of a tobacco product via ultraperformance liquid chromatography. We applied unsupervised K-Means clustering on the creatinine-adjusted VOC metabolite data and explored the association between each cluster and blood pressure, adjusting for age, sex, and race.

Results: The clustering analysis identified two distinct clusters. Cluster 1 (302 individuals, 86.8%) was characterized by low VOC metabolite levels with individuals predominantly e-cigarette users (59.3%), non-users (29.1%), and a smaller proportion of cigarette smokers (11.6%). Cluster 2 (46 individuals, 13.2%) had higher levels of VOC metabolites including CYMA, HPMMA, MHBMA3, and 3HPMA, and included most of the individuals who used cigarettes (91.3%). After adjustment for age, sex, and race, Cluster 2 was associated with a higher heart rate (β=3.29; 95% CI: -0.26-6.84; p<0.05) compared to Cluster 1. No significant differences were observed for systolic (β= -0.66; 95% CI: -4.60-3.28) or diastolic blood pressure (β=0.34; 95% CI: -2.51-3.2) between clusters.

Conclusions: These findings suggest that cigarette-induced VOC exposure may not impact cardiovascular function after acute exposure. Additionally, VOC exposure profiles vary across tobacco product types, suggesting that regulatory assessments of tobacco products could consider exposure patterns rather than product types. Clustering analyses may offer a powerful tool to assess the safety and risks of new and emerging tobacco products based on real-world exposure patterns.