Estimation of blood pressure variability from 24-hour ambulatory finger blood pressure

Portapres is a noninvasive, beat-to-beat finger blood pressure (BP) monitor that has been shown to accurately estimate 24-hour intra-arterial BP at normal and high BPs. However, no information is available on the ability of this device to accurately track ambulatory BP variability. In 20 ambulatory normotensive and hypertensive subjects, we measured 24-hour BP by Portapres and through a brachial artery catheter. BP and pulse interval variabilities were quantified by (1) the SDs of the mean values (overall variability) and (2) spectral power, computed either by fast Fourier transform and autoregressive modeling of segments of 120-second duration for spectral components from 0.025 to 0.50 Hz or in a very low frequency range (between 0.00003 and 0.01 Hz) by broadband spectral analysis. The 24-hour SD of systolic BP obtained from Portapres (24+/-2 mm Hg) was greater than that obtained intra-arterially (17+/-1 mm Hg, P<0.01), but the overestimation was less evident for diastolic (3+/-1 mm Hg, P<0.01) and mean (3+/-1 mm Hg, P<0.01) BP. The BP spectral power <0.15 Hz was also overestimated by Portapres more for systolic than for diastolic and mean BPs; similar findings were obtained by the fast Fourier transform, the autoregressive approach, and focusing on the broadband spectral analysis. BP spectral power >0.15 Hz obtained by the Portapres was similar during the day but lower during the night when compared with those obtained by intra-arterial recordings (P<0.01). No differences were observed between Portapres and intra-arterial recordings for any estimation of pulse interval variabilities. The overestimation of BP variability by Portapres remained constant over virtually the entire 24-hour recording period. Thus, although clinical studies are still needed to demonstrate the clinical relevance of finger BP variability, our study shows that Portapres can be used with little error to estimate 24-hour BP variabilities if diastolic and mean BPs are used. For systolic BP, the greater error can be minimized by using correction factors.