Development and validation of cardiometabolic risk predictive models based on LDL oxidation and candidate geromarkers from the MARK-AGE data

Abstract

The predictive value of the susceptibility to oxidation of LDL particles (LDLox) in cardiometabolic risk assessment is incompletely understood. The main objective of the current study was to assess its relationship with other relevant biomarkers and cardiometabolic risk factors from MARK-AGE data. A cross-sectional observational study was carried out on 1089 subjects (528 men and 561 women), aged 40-75 years old, randomly recruited age- and sex-stratified individuals from the general population. A correlation analysis exploring the relationships between LDLox and relevant biomarkers was undertaken, as well as the development and validation of several machine learning algorithms, for estimating the risk of the combined status of high blood pressure and obesity for the MARK-AGE subjects. The machine learning models yielded Area Under the Receiver Operating Characteristic Curve Score ranging 0.783-0.839 for the internal validation, while the external validation resulted in an Under the Receiver Operating Characteristic Curve Score between 0.648 and 0.787, with the variables based on LDLox reaching significant importance within the obtained predictions. The current study offers novel insights regarding the combined effects of LDL oxidation and other ageing markers on cardiometabolic risk. Future studies might be extended on larger patient cohorts, in order to obtain reproducible clinical assessment models.

Keywords: Cardiometabolic risk; LDL oxidation; MARK-AGE; Machine learning; Vascular ageing.