Estimating central SBP from the peripheral pulse: influence of waveform analysis and calibration error

To compare estimation of central cSBP by application of a generalized transfer function (GTF) to a peripheral arterial waveform and from the late systolic shoulder (SBP(2)) of such a waveform and assess errors introduced by noninvasive calibration of the waveform.

The digital arterial pulse was acquired noninvasively with a servo-controlled finger cuff. A high fidelity pressure tipped catheter was placed in the proximal aortic root. Measurements were made at baseline (n = 40), after nitrovasodilation, handgrip exercise (n = 18) and during pacing (n = 10). Estimates of cSBP obtained using a GTF and from SBP(2) (using an algorithm applied to individual cardiac cycles) of the digital arterial waveform were compared with values measured at the aortic root.

When arterial waveforms were calibrated from aortic intra-arterial mean and DBP there was close agreement between estimated and measured cSBP: mean difference between estimated and measured cSBP (SD): 1.0 (5.7) and -0.7 (5.5) mmHg for GTF and SBP(2), respectively. Noninvasive oscillometric calibration increased variability in estimation of cSBP [mean difference, 1.3 (11) mmHg for SBP(2)] but estimates of the cSBP to peripheral systolic pressure increment from oscillometric calibration of SBP(2) agreed well with those obtained using invasive calibration [mean difference -2.4 (6.1) mmHg].

SBP(2) potentially provides a simple measure of cSBP and is of comparable accuracy to a GTF. Noninvasive calibration increases variability for both methods but has less effect on the cSBP to peripheral SBP increment.