Feasibility of waveform separation of central aortic pressure pulse based on lognormal flow wave approximation

Abstract

Wave reflections affect the aortic pressure and flow waves and may yield effective indices for monitoring cardiovascular health and disease. More accurate measures of wave reflection are obtainable after separation of aortic pressure waveform into its forward and backwards components. This paper presents a novel method for waveform separation of aortic pressure using a single pressure pulse measurement together with a lognormal flow wave approximation. The separated forward and backward waves were used to calculate the wave reflection parameters including reflection amplitude (RM) and reflection index (RI), and derive the aortic pulse transit time (PTT). The feasibility of the proposed method was verified using the central aortic pressure estimated from a measured radial pressure pulse by an aortic pulse analyzer and aortic flow waveform (39 data sets). Wave separation analysis (WSA) was assessed using the pressure and approximated flow waveforms based on two methods respectively, the proposed lognormal flow wave and the existing triangular flow wave. The results were compared with that of the measured flow wave. In time domain, for RM, RI, and PTT, the relations between the values calculated from the estimated flow and that from the measured flows can be characterized accurately by the linear regression models with the following coefficients of determination: R2 = 0.87 (RM), 0.86 (RI), 0.79 (PTT) using the lognormal flow; R2 = 0.84 (RM), 0.83 (RI), 0.77 (PTT) using the triangular flow. Agreement between reference and approximated values was increased when the approximated flow waveform was obtained using lognormal approximation rather than triangular estimation. The lognormal flow wave, a more physiological waveform, for approximating the flow improved the results of WSA in both time and frequency domains.