Abstract
BACKGROUND: Stroke volume can be estimated beat-to-beat and non-invasively by pulse wave analysis (PWA). However, its reliability has been questioned during marked alterations in systemic vascular resistance (SVR). We studied the effect of SVR on the agreement between stroke volume by PWA and Doppler ultrasound during reductions in stroke volume in healthy volunteers. METHODS: In a previous study we simultaneously measured stroke volume by PWA (SVPWA) and suprasternal Doppler ultrasound (SVUS). We exposed 16 healthy volunteers to lower body negative pressure (LBNP) to reduce stroke volume in combination with isometric hand grip to elevate SVR. LBNP was increased by 20 mmHg every 6 minutes from 0 to 80 mmHg, or until hemodynamic decompensation. The agreement between SVPWA and SVUS was examined using Bland-Altman analysis with mixed regression. Within-subject limits of agreement (LOA) was calculated from the residual standard deviation. SVRUS was calculated from SVUS. We allowed for a sloped bias line by introducing the mean of the methods and SVRUS as explanatory variables to examine whether the agreement was dependent on the magnitude of stroke volume and SVRUS. RESULTS: Bias ± limits of agreement (LOA) was 27.0 ± 30.1 mL. The within-subject LOA was ±11.1 mL. The within-subject percentage error was 14.6%. The difference between methods decreased with higher means of the methods (-0.15 mL/mL, confidence interval (CI): -0.19 to -0.11, P<0.001). The difference between methods increased with higher SVRUS (0.60 mL/mmHg × min × L-1, 95% CI: 0.48 to 0.72, P<0.001). CONCLUSION: PWA overestimated stroke volume compared to Doppler ultrasound during reductions in stroke volume and elevated SVR in healthy volunteers. The agreement between SVPWA and SVUS decreased during increases in SVR. This is relevant in settings where a high level of reliability is required.