1349428 - Simultaneous systemic and pulmonic blood pressure modeling

Pressure generation in the ventricular chambers is known to track a single sinusoidal lobe which is interrupted by opening of the outflow valves (aortic and pulmonic for left and right ventricles, respectively).  We hypothesized that a sinusoidal-based transform of key cardiac variables could model both systemic systolic blood pressure (SBP) and pulmonic arterial systolic blood pressures (PAPs).

Methods:

Data from patients (n=95) with SBP and cardiac measured variables were available. In a sub set of 41 patients (43%) with pulmonary arterial hypertension (PAH) right heart catheterization measures of PAPs were available. Sinusoidal transforms of key cardiac measures obtained using CMR were separately fit to SBP and PAPs. Importantly, the frequency of each cardiac variable was constrained to be the same for the SBP and PAPs, while the phase of the transform was allowed to be set separately. The cardiac variables entered into the model were ventricular ejection fraction (EF), ventricular end systolic volume (ESV) and ventricular-outflow impedance (Imp), each measured separately for left and right ventricles. In addition to sinusoidal transformed variables, the following variables were included in the model without transformation; left and right ventricular mass and age. The sinusoidal transformed and linear variables were included in two linear regression models, one for SBP and one for PAPs. The fit of each model was optimized by maximizing the summation of the r² fit for each model.

Results:

The model fit to SBP had an r² of 0.59, and the model fit to PAPs had an r² of 0.75, figure 1. The constraint that the same sinusoidal transform frequency be applied to each cardiac variable emphasizes the physiologic nature of the relationship to SBP and PAPs. Further, since both left and right ventricular variables contributed to SBP and PAPs is an acknowledgement of the interaction of left and right circuits. While disease of the left side is not responsible for elevation of PAPs in PAH patients, the interaction of the ventricles is not necessarily pathological but a consequence of the two ventricles being in parallel with each other, being constrained in total volume (at least in the short term) by the pericardial sac, and being in physical contact for direct pressure interaction via the septum. The model for both systemic and pulmonic pressures, for both PAH and non-PAH patients was fit by a common set of variables applied to left and right ventricular variables.

Conclusion:

A model was fit to SBP and PAPs using CMR derived data and patient age. The model employed several variables that were sinusoidally transformed, and importantly the same frequency was applied to each variables when modeling SBP and PAPs, indicating the physiological basis of the model.