Using AI tools to predict arterial stiffness in populations with cardiovascular risk factors

Abstract

Introduction: Smoking is a well-established cardiovascular risk factor that contributes to arterial stiffness, accelerating processes such as atherosclerosis. Pulse wave velocity (PWV) is a noninvasive measure that reflects vascular health and helps to detect subclinical damage. Objectives: To describe the variables associated with PWV in the study population, evaluate its association with smoking and ex-smoking, and use machine learning (ML) techniques to identify complex associations not evident in conventional statistical analyses. Materials and Methods: A descriptive cross-sectional study was conducted in the vascular age and smoking cessation clinic of Santojanni Hospital. PWV was measured in 136 patients, analyzing clinical and demographic variables, emphasizing current or past smoking history. Results: PWV showed a positive correlation mainly with age, chronic kidney disease (CKD), and arterial hypertension (AH). Although no significant differences were found between smokers and non-smokers initially, when excluding comorbidities such as hypertension and CKD, smokers and ex-smokers were found to have significantly higher PWV. The use of machine learning identified that, although hypertension and CKD are important factors, the number of pack years and age influenced arterial stiffness more. Conclusions: The study highlights the detrimental effects of smoking on vascular health and the usefulness of PWV in assessing cardiovascular risk. ML techniques allowed the identification of complex interactions, revealing that smoking amplifies the impact of other risk factors. Smoking cessation could improve PWV, reinforcing the importance of continuous monitoring to prevent cardiovascular events.