Evaluation of stroke volume estimation during orthostatic stress: the utility of Modelflow

Abstract

Advanced blood pressure monitoring devices contain algorithms that permit estimation of stroke volume (SV). Modelflow (Finapres Medical Systems) is one common method to non-invasively estimate beat-to-beat SV. However, Modelflow accuracy during profound reductions in SV is unclear. We aimed to compare SV estimation by Modelflow and echocardiography, at rest and during orthostatic challenge. We tested 13 individuals (age 24 ± 2 years; 7 female) using combined head-up tilt and graded lower body negative pressure, continued until presyncope. SV was derived by both Modelflow and echocardiography on multiple occasions while supine, during orthostatic stress, and at presyncope. SV index (SVI) was determined by normalising SV for body surface area. Bias and limits of agreement were determined using Bland-Altman analyses. Two one-sided tests (TOST) examined equivalency. Across all timepoints, Modelflow estimates of SV (73.2 ± 1.6 ml) were strongly correlated with echocardiography estimates (66.1 ± 1.3 ml) (r = 0.56, P < 0.001) with a bias of +7.1 ± 21.1 ml. Bias across all timepoints was further improved when SV was indexed (+3.6 ± 12.0 ml.m-2). Likewise, when assessing responses relative to baseline, Modelflow estimates of SV (−23.4 ± 1.4%) were strongly correlated with echocardiography estimates (−19.2 ± 1.3%) (r = 0.76, P < 0.001), with minimal bias (−4.2 ± 13.1%). TOST testing revealed equivalency to within 15% of the clinical standard for SV and SVI, both expressed as absolute values and relative to baseline. Modelflow can be used to track changes in SV during profound orthostatic stress, with accuracy enhanced with correction relative to baseline values or body size. These data support the use of Modelflow estimates of SV for autonomic function testing.