A Linear Regression Model for Assessment of Interdependent Parameters of Central Pressure Determined from the Peripheral Pulse

The central aortic pressure waveform (CW) is determined by cardiac and vascular parameters. In the intact circulatory system, the influence of individual parameters on CW cannot be readily assessed because of parameter interdependence. This study presents a linear regression model for estimating CW from a range of measurements of heart rate (HR), brachial systolic pressure (BSP) and frequency components of previous CWs for an individual subject. The model was constructed using data from 34 subjects (age 20-76 years) in whom multiple CWs were determined from the radial pulse waveform over a given range of HR (42-77 bpm) and BSP (90-172 mmHg). For each subject a model was created to estimate the particular CW for any given heart rate and BSP. All models were validated independently using a cross validation technique. The model was able to estimate CW to within a mean error of <1 mmHg. The models were used to compare the influence of BSP on the late systolic pressure augmentation (Augmentation Index (AIx)) for a fixed HR. This study suggests that a linear regression model that estimates the CW can be used to assess the effect of individual parameters on specific waveform features. This allows the characterization of the effect of individual parameters on CW features, such that the response can be quantified independently of other parameters as a family of CWs over a specified range.