Validation of a method to estimate stroke volume from brachial-cuff derived pressure waveforms

Abstract

Background: Steady state hemodynamics [Stroke Volume - SV, cardiac output, peripheral resistance (Rp)] have attracted researchers in hypertension for decades. However, due to technical difficulties, they never entered clinical medicine. Objective: To investigate the performance of a model-based algorithm, utilizing brachial pressure waveforms obtained with a cuff, to estimate stroke volume. The model combines a modified three-element like Windkessel system and transmission line theory. Participants and Methods: In study 1, 97 patients with reperfused acute myocardial infarction and two healthy controls underwent cardiac magnetic resonance imaging (1.5 Tesla Magnetom, Siemens, Germany), and SV was measured with standard protocols from short axis cine images (11 slices). In study 2, 19 healthy individuals (12 females) had SVs measured with inert gas rebreathing at rest, and during light exercise (20 watts) on a bicycle ergometer. In both studies, model-based SV, estimated with a brachial cuff, was the comparator. Results: In study 1, both SVs were moderately correlated (r = 0.54, p < 0.001). Using the method of Bland-Altman, mean difference between both methods was 8.7 ml (1.96 limits of agreement were 36.7 and -19.3 ml), with no systematic bias. In study 2, both SVs were moderately correlated at rest (r = 0.63, p = 0.004) and at light exercise (r = 0.70, p = 0.0057). Using the method of Bland-Altman, mean difference between both methods was 8.6 ml (1.96 limits of agreement were 39.2 and -22.0 ml) at rest, and 42.7 ml (1.96 limits of agreement were 95.4 and -10.1 ml) at light exercise. Conclusion: Brachial oscillometry and mathematical modeling provide a reasonable estimate of SV under static conditions, which may be a useful addition to 24-h measurements of blood pressure and pulsatile hemodynamics.