In Computers in biology and medicine
BACKGROUND : Chronological age (CA) has been adopted as an important independent risk factor in cardiovascular risk assessment. However, different individuals with same CA may have distinct actual vascular aging due to various lifestyles. Therefore, it is difficult to fully describe the difference of actual vascular aging by CA.
OBJECTIVE : This study proposes a new index vascular age (VA) to avoid the limitations of CA.
METHOD : In this work, VA refers to the sum of CA and lifestyle impact (AgeLI). Firstly, we take the pulse signal features and CA as independent variables and dependent variable respectively, and adopt cross validation to train Support Vector Regression model. Then we acquire the predicted chronological age (PA) of all subjects with the model. Secondly, we obtain the function model between CA and PA, and calculate the expectation of PA (ePA) for each subject. Simultaneously, we take the difference between PA and ePA as the estimated value of AgeLI to further calculate VA. Finally, in order to evaluate the effectiveness of VA, we compare the correlations between CA, PA, VA and 8 objective indices such as augmentation index, pulse transit time, diastolic augmentation index, etc. RESULTS: In general, VA and PA are closer to these 8 objective indices than CA. Moreover, VA is also superior to PA in vascular aging evaluation.
CONCLUSION : The VA suggested in this study emphasizes the difference of vascular aging in same CA group, which can better reflect the actual vascular aging than CA and PA.
Tang Qingfeng, Pan Zhiqiang, Tao Changlong, Jiang Jing, Su Benyue, An Hui, Liu Guodong
2022-Nov-26
Chronological age, Lifestyle impact, Pulse signal, Support vector regression, Vascular age
