Modeling age and sex variations of arterial pressure and cardiac hemodynamics in a virtual population using a...

Virtual populations offer a promising approach for modeling cardiovascular variability in health and disease, with the potential to support or even replace clinical trials. Although most studies focus on age-related changes in the cardiovascular system, sex remains underrepresented despite its significant influence on cardiovascular physiology. This study aims to generate a healthy virtual population that canvases the full range of cardiac hemodynamics and arterial pressure variability due to age and sex. A zero-dimensional (0-D) elastance model of the left ventricle was coupled with a 0-D and 1-D model of 55 large arteries. Model parameters (cardiac properties, vessel geometry, and vascular bed) were derived from literature and systematically varied using a uniform distribution. This process yielded 972 virtual subjects, representing both sexes and spanning ages from 25 to 75 yr. Six acceptance criteria, based on vascular hemodynamics and cardiac characteristics, were applied to retain only plausible virtual subjects, resulting in 147 males and 180 females. Validation against in vivo literature data showed that the model captures the propagation of pressure waveform from the aorta to its main branches. The virtual population was also compared against a prospective observational study of 63 healthy subjects ranging from 21 to 77 yr. The simulated blood pressure profiles encompassed the physiological range observed in vivo across age and sex, demonstrating the model ability to replicate real-world measurements. This model provides a foundation for generating realistic, demographically diverse cardiovascular populations and can be extended to simulate pathological conditions and support the development of diagnostic and monitoring tools.NEW & NOTEWORTHY This study introduces a 01-1D cardiovascular model and a parameter assignment methodology to reproduce age- and sex-related physiological changes in arterial pressure and cardiac hemodynamics. Uniform sampling of input parameters was conducted to generate a virtual population, encompassing the variability of healthy individuals. A systematic analysis was performed to identify parameter combinations that led to exclusion of implausible virtual subjects. The virtual population was compared with a clinical dataset, demonstrating its potential for realistic physiological simulations.