Reproducibility of radial pulse wave analysis in healthy subjects

Objective To determine the reproducibility of parameters estimated by pulse wave analysis (PWA), mainly of central systolic blood pressure in the aorta (CSP), central systolic pressure-time index (CSPTI, area under the systolic part of the curve in the pressure-time graph per minute), and central augmentation index (CAI, pressure peak caused by wave reflection divided by pressure peak caused by ventricular ejection). Methods In 88 healthy subjects, aged 19-53 years, radial pulse wave was registered by two investigators on separate visits in a random sequence, twice during each visit. Results Variability within and among subjects was significantly different for peripheral systolic pressure (PSP) and for all the above-mentioned parameters (P < 0.0001 by ANOVA for all). Variability within and between observers was tested by Bland-Altman plots; they did not show any trend for the variability to be dependent on the underlying mean value. PSP, CSP and CSPTI decreased significantly during one visit (by 4.6, 4.7 and 4.2%, respectively), PSP and CSP were higher at the first than at the second visit (by 2.2, 2.2%, respectively, and not significant for CSPTI), and there were also significant interobserver differences in all the three parameters as one observer measured higher values (by 2.4, 3.2 and 6.0%, respectively). CAI did not change within and between visits; the same applied to the difference between PSP and CSP since these pressures always changed in parallel. Conclusions PWA gives estimates of several parameters characterizing the pressure load of central circulation and the wave reflection. The reproducibility of CSP and CSPTI is similar to that of PSP, CAI and the difference between PSP and CSP is not influenced by order of measurement, of visit or by investigator. Therefore, CAI is a more stable parameter than PSP measured by an oscillometric device. Since these parameters may contribute to a better assessment of cardiovascular risk, PWA might be used in prospective studies. J Hypertens 2000, 18:1033-1040 (C) Lippincott Williams & Wilkins.