Coefficient-free blood pressure estimation based on arterial lumen area oscillations in oscillometric methods

Abstract

In this paper we present a novel algorithm destined to estimate systolic and diastolic blood pressures from lumen area oscillations of vessel underneath the cuff by using automated oscillometric method. This algorithm is composed of three procedures which process only the diastolic region information of oscillometric waveform (OMW) from 80 mmHg to 20 mmHg, a domain which is considered a low cuff pressure region. The standard oscillometric methods need inflating the cuff to Supra Systolic Pressure (SSBP) and, as such, require a deflation time proportionally longer than our method. Because of the relative low inflation pressure, our method represents a viable option for patients who need to be monitored continuously, since, in such circumstances, the cuff is not allowed to be inflated to higher pressures for longer periods of time. We developed a unified algorithm composed of three integrated procedures which are all based on the arterial lumen area oscillation model at diastolic region to estimate blood pressure. The first procedure estimates the compliance c of the blood vessel. The second procedure uses c and estimates the maximum lumen area (Am), lumen area at mean arterial pressure — MAP (40) and the systolic arterial pressure (SBP) from the peaks of the OMW pulses. The third procedure uses c, Am, A0 found above to estimate diastolic arterial pressure (DBP) from the troughs of the OMW pulses. The OMW is obtained by filtering the cuff deflation curve (CDC) with a 2nd order Butterworth band pass filter and which has cut-off frequencies of 0.5 to 20 Hz. The proposed method avoids using empirical systolic and diastolic ratios for estimation of blood pressure (like the popular maximum amplitude algorithm — MAA), but rather employs the least square method to optimize the lumen area oscillations model for targeted parameters. We applied this method on 150 oscillometric traces recorded from 10 healthy subjects composed of males and females from 25 to 63 years old, and validated the results with values measured by an Omron device that served as reference for each recording. Results are encouraging as mean absolute errors of estimated values from the Omron references over 150 recordings are 5.13 mmHg in terms of SBP and 3.18 mmHg for DBP with the standard deviation of errors of 3.60 mmHg and 2.58 mmHg respectively.