Calibration of noninvasively recorded upper-limb pressure waves.

Calibration of Noninvasively Recorded Upper-Limb Pressure Waves To the Editor: The article by Verbeke et al1 raises an important question: how does one calibrate the radial artery waveform obtained by applanation tonometry? Does one apply systolic and diastolic brachial values from the Korotkov technique, the “gold standard”, or does one use an oscillometric wrist technique, despite lack of confidence in any available method? Alternatively, as implied by Verbeke et al, should one use brachial tonometry instead, calibrated to brachial-cuff values? The study by Verbeke et al1 suggests (from noninvasive methods alone) that the present technique of calibrating radial tonometry to brachial-cuff pressure may have substantial inaccuracy. The method by Verbeke et al1 depends on a technique originally described by Kelly and Fitchett2 where systolic pressure was extrapolated on the assumption that diastolic and mean pressure were identical throughout the arterial tree. The technique currently recommended in the SphygmoCor process is based on invasive studies of arterial pressure waves, which showed that amplification between brachial and radial artery is small in comparison to that between aorta and brachial artery.3,4 We repeated essentials of the Verbeke et al1 study and confirmed estimation of high pressure amplification between brachial and radial sites. But in the process we became aware of several technical factors that could account for the estimations and, hence, regard these as flawed. First, we could not be confident of successfully applanating the brachial artery. Unlike carotid and radial arteries, the brachial has no support behind so the artery surface cannot confidently be flattened when the tonometer is applied. Second, the artery is deeper than the radial, and often covered by part of the bicipital aponeurosis, so the tonometer had to sense through this. We and others have been able to confirm similarity of tonometric and invasive waveforms at carotid and radial sites,4,5 but there are no such confirmatory data for the brachial site. Our brachial tonometry waveforms were more blunted than when recorded directly by needle or cannula. We did show reasonable accuracy of the Kelly/Fitchett extrapolation technique for systolic pressure when a more blunted systolic peak was calculated from a sharper peak (ie, carotid from radial pulse), but not when the process was reversed, as in noninvasive estimation of radial peak from brachial peak, on which Verbeke et al1 relied.1 We acknowledge inaccuracy of all cuff methods for measuring systolic and diastolic pressure within an artery, but agree that some are better than others. Radial sphygmography and tonometry are supported by precedent, theory, and practice, whereas brachial tonometry is suspect. Radial tonometry can provide useful information on contour of the aortic waveform including ejection duration and augmentation index, even without calibration. If calibration is to be used, we believe that the presently accepted brachial-cuff technique using the Korotkov method is still preferable to any wrist-cuff method.