Ultrasound measurement of central pulse pressure from carotid diameter

The use of ultrasound for imaging of carotid arteries is directed principally at detection of atherosclerotic plaques causing luminal obstruction. The imaging resolution of conventional ultrasound devices is now sufficiently high for reliable detection of subclinical atherosclerosis through changes in intima–medial thickness (IMT) determined from analysis of two-dimensional M-mode images of the blood vessel. This allows accurate measurement of vessel diameter and wall thickness by obtaining average values over a defined spatial region of interest. However, additional processing is required for fiducial detection of temporal change in vessel dimensions using enhanced edge-detection algorithms applied to M-mode derived images [1] or analysis of radiofrequency signals reflected from the moving vessel walls by means of multigated Doppler systems [2,3]. The early seminal work in ultrasound by Robert Reneman and Arnold Hoeks and colleagues in Maastricht in the Netherlands provided the first major impetus for noninvasive measurement of arterial diameter for estimation of vessel distensibility using pressure/diameter relationships [2–4]. The hardware and software technology of arterial wall tracking based on analysis of radiofrequency signals, which was usually the domain of research laboratories, is now incorporated into conventional commercial devices [5] used routinely in hospital and clinical departments. The noninvasive registration of the motion of the artery wall by interrogating reflected ultrasound signals has enabled the continuous measurement of arterial diameter, and by implication, the arterial distension associated with the arterial pressure pulse during the cardiac cycle [2]. The relation between pressure and diameter is defined by the nonlinear properties of the arterial wall – a fundamental property of arterial design [6]; that is, arterial distention decreases with increasing pressure, resulting in the pressure dependence of arterial stiffness [7]. The nonlinearity is inherent in the full range of the pressure–diameter curve, and the phase plots of pressure–diameter over the pressure excursion range of the cardiac cycle are used to assess the viscoelastic properties of the artery wall quantified by the amount of hysteresis [8,9]. However, this effect is relatively small, and a single (generally exponential) relationship is usually assigned to the pressure–diameter curve during both systolic and diastolic excursion phases [10]. Notwithstanding the inherent nonlinearity in the pressure–diameter relationship, the arterial distension waves have a strong morphological resemblance to the intra-arterial pressure waves, and this has suggested the possibility of determining carotid pressure from noninvasive distension waves [11,12]. The study by Kozakova et al.[13] in this issue of the Journal of Hypertension evaluates the measurement of carotid pressure by use of radiofrequency wall tracking techniques and calibrating the distension curve using values of diastolic and mean pressure determined from systolic and diastolic pressure measured by the brachial cuff sphygmomanometer. The study employs a conventional ultrasound device (MyLab70; Esaote, Genova, Italy) with a 10-MHz probe for carotid examination and equipped with the additional modules for processing radiofrequency data. Measurements were taken in 346 patients, comprising a cohort of healthy controls and patients with diabetes (21.7%) and hypertension (57.8%). Systolic and diastolic pressure obtained by means of the carotid distension curve were compared with systolic and diastolic pressure estimated by carotid applanation tonometry using the PulsePen device (Diatecne, Milan, Italy) calibrated with the brachial cuff pressure values [14]. Additional measurements of IMT were obtained in the whole cohort and left ventricular mass (LVM) in a smaller number of participants (n = 253). Pulse pressure obtained by wall tracking of carotid distension and by tonometry showed good agreement with a relatively high correlation (r = 0.87) and a mean difference of 3.1 ± 6.8 mmHg, and both were associated with similar predictive variables (age, heart rate, antihypertensive therapy). Carotid pulse pressure obtained by both methods was more strongly correlated with IMT and LVM compared to brachial pulse pressure. This agrees with a recent meta-analysis of previous studies showing a higher association of central aortic systolic pressure with target organ damage compared to brachial pressure [15]. The study by Kozakova et al.[13] suggests that carotid pulse pressure obtained by carotid wall tracking techniques can be used as a reliable surrogate for central aortic pressure. As presented, the study needs to be qualified in the context of some of its limitations. The method of calibration inherently assumes a linear relationship between pressure and diameter, whereas the relationship is inherently not linear. A study by Vermeersch et al.[16] in a large cohort (N = 2026) has shown that an exponential pressure–diameter calibration produces substantial improvement in carotid systolic pressure estimation from distension curves compared to a linear calibration. A second limitation of the study by Kozakova et al.[13] is the suboptimal comparison between the carotid pressure values obtained by distension curves and by applanation tonometry using different form factors in each case for calibration. For both measurements, calibration assumed invariant diastolic and mean pressure between brachial and carotid arterial sites. However, for tonometry measurements mean pressure was calculated as [diastolic pressure] + 0.33 × [pulse pressure], whereas for the distension measurements it was calculated as [diastolic pressure] + 0.4 × [pulse pressure] [17]. The difference in calibration form factor is undoubtedly responsible for the average difference found for the two methods of 3.1 ± 6.8 mmHg. However, for this cohort studied, this differences is within the ‘substantial equivalent’ limits of 5 ± 8 mmHg as prescribed by the American Association for the Advancement of Medical Instrumentation (AAMI) [18]. Notwithstanding the ‘substantial equivalence,’ an important caveat is that the pressure calibration form factor in the commercial device used in this study cannot be altered, as is also the case for the linear pressure–diameter relationship. Due to change in pulse wave shape with age and level of arterial pressure, there is no guarantee that the AAMI ‘substantial equivalence’ would be consistent across all cohorts. It is of interest to question why such a simple software operation would need to be closed and not be made available to the operator to select a desired form factor for mean pressure calculation from measured systolic and diastolic pressures. For example, some brachial oscillometric devices provide an estimation of true mean arterial pressure and other devices measuring pressure waveform shape allow calculation of true mean pressure. These values would be more accurate within individuals as points of calibration for carotid pulse pressure than any form factor estimation of mean pressure. Notwithstanding the caveats and limitations, the study by Kozakova et al.[13] has provided robust evidence that if routine ultrasound examination of carotid arteries is to be conducted, a measure of central pulse pressure is obtainable without any substantial extra effort and it could provide added value for assessment of pressure-related effects on target organ damage [15]. The modern technical advances in image acquisition and analysis of radiofrequency signals have provided a clinical translational platform for the early breakthrough research in ultrasound by Hoeks et al.[3].