Design
Editorial
Answers are grounded in this paper and cite their sources.
Pulse wave analysis and derived central aortic pressures may be useful for evaluating differential drug effects, but the augmentation index (AIx) is heavily confounded by factors like heart rate and cannot be assumed to be a direct measure of aortic stiffness.
One hundred and thirty years ago, long before the routine use of blood pressure measurement, Mahmoud [1] advocated pulse wave analysis as a means of evaluating the functional integrity of the circulation. Enthusiasm for this approach subsequently diminished when sphygmomanometry emerged as a simple and reliable means of recording peripheral blood pressure, which was in turn validated as a potent predictor of future cardiovascular morbidity and mortality. In recent years, there has been a resurgence of interest in the possibility of using alternatives to simple brachial blood pressure measurements to non-invasively evaluate circulatory function and predict cardiovascular risk. These indices, amongst others, include pulse pressure, pulse wave velocity and pulse wave analysis. Pulse wave analysis is a simple procedure most commonly undertaken by applanation tonometry over the radial artery [2]. The tonometer generates a radial artery wave form which can then be transformed using a mathematical radial-to-aortic transfer function to derive an aortic wave-form. By entering data from simultaneous, non-invasive, oscillometric measurement of brachial blood pressure, the wave forms and transfer function have been used to derive central aortic pressure indices [2]. Pulse wave pressure is generally amplified in transit from the aorta to the radius; thus, the derived central pressures are generally lower than the peripheral pressures, although this relationship is variable. In addition to deriving pressures, this methodology has been utilized to derive additional indices, most notably the augmentation index (AIx). This is taken to represent augmentation of systolic pressure within the aorta as a consequence of early pulse wave reflection. The timing of wave reflection is in part determined by pulse wave velocity, which is in turn influenced by aortic stiffness [2]. This AIx has been suggested by some to be a non-invasive index of arterial stiffness [3]. As arterial stiffness increases with ageing and cardiovascular disease states, it has been suggested that AIx could be used as a simple non-invasive surrogate index of cardiovascular risk. This concept has some support from the observation that AIx was strongly predictive of cardiovascular mortality in patients with endstage renal disease [4]. Others have questioned the validity and clinical utility of pulse wave analysis and, in particular, the use of a computed transfer function to derive central aortic wave forms and other related indices [5–9]. There is no doubt that this field is now bedevilled by competing forces of evangelism and scepticism. The truth usually lies somewhere between the two. Some important questions remain. Is the transfer function valid, accurate and necessary? Are measurements of AIx clinically useful? How accurate are central blood pressure measurements derived from the peripheral circulation? In what circumstances, if any, might data derived from pulse wave analysis prove to be clinically useful? Two principal sets of data have been derived using the transfer function: (i) aortic AIx and (ii) central aortic pressures. The accuracy of the transfer function has been questioned and discussed in recent editorials [5,6]. Recent studies have highlighted the fact that the use of this function usually leads to under-estimation of aortic systolic blood pressure and over-estimation of aortic diastolic pressures when the derived pressures are compared with those measured directly at cardiac catheterization [7,8]. This may in part relate to the fact that the original transfer function was validated using invasively measured peripheral pressures and that non-invasive oscillometric measurement of brachial artery pressure may introduce this error [7]. Another possibility is that it is simply not possible to develop a single mathematical function that can accurately account for the many variables introduced by human variation and the impact of disease states. The transfer function has been particularly criticized with regard to the derivation of AIx. Hope et al. [8] recently showed that there were very wide margins of error comparing centrally measured and centrally derived AIx, so much so, that the measured and derived measurements were not statistically significantly correlated. Others have gone further and questioned whether transformation of the peripheral radial wave form is required to estimate AIx by demonstrating that the data obtained directly from the peripheral wave form was just as useful without the need for transformation to assess AIx [9]. The next question is what useful information can be gained by measuring AIx? As indicated above, it has been suggested that AIx can be used as a surrogate for vascular stiffness [3] and there is consensus that increased vascular stiffness is a marker of arterial disease and cardiovascular risk. AIx is a complex composite measurement, derived from many potentially important dynamic variables. AIx, is influenced by ventricular ejection characteristics, pulse wave velocity and the timing of the reflected pulse wave, and arterial tone and structure at peripheral reflection sites. In this issue of the journal, Lemogoum et al. [10] reveal the many factors that can independently influence AIx, including blood pressure, age, gender, body height and especially heart rate. Increased blood pressure usually increases AIx, as does shorter stature and female gender. Increasing heart rate dramatically reduces AIx, presumeably due to altered timing of the reflected wave in relation to systole [3]. Their study also highlights an important dissociation between AIx and pulse wave velocity measured simultaneously in the same patients [10]. The infusion of isoprenaline increased peripheral blood pressure, stroke volume and heart rate. There was no effect of isoprenaline on pulse wave velocity (an accepted index of arterial stiffness); however, there was a dramatic heart-rate dependent fall in AIx [10]. In this context, as in many other circumstances, AIx cannot be assumed to be a measure of aortic stiffness. It is significantly confounded by changes in heart rate. When considering the aforementioned additional potential confounders and the concern over the accuracy and validity of central AIx derivation from pulse wave analysis, it is difficult to see how it can provide clinically useful additional data in the assessment of individuals with hypertension or cardiovascular comorbidity. Radial artery pulse wave analysis and its computed transformation also provide derived indices for central aortic pressures. Although studies have suggested that central systolic blood pressure in particular may be underestimated using this technique [7,8], the margins of error are less than that for AIx, and the data are much less confounded by the aforementioned variables that hinder the usefulness of AIx. This may also relate to the fact that measurement of AIx is dependent on higher frequency signals than blood pressure measurement and the transfer function appears to be less accurate and to show greater between-subject variability at high frequencies. Thus, the estimation of central blood pressures using the radial–aortic transfer function could have value. Moreover, unlike AIx, conceptually, the estimation of central aortic pressure has more obvious relevance to the pathogenesis of cardiovascular disease, especially stroke. The main potential use of pulse wave analysis and central aortic pressure derivations might come from an examination of the differential effects of blood pressure lowering and other cardiovascular drugs on the relationship between peripheral and central blood pressures. In this context, the absolute accuracy of the measurement is not essential to demonstrate hitherto unrecognized and important qualitative differences in the pulse wave form and central pressure indices between treatments. Such evaluations could provide important insights into mechanisms of cardiovascular protection that go beyond the impact of seemingly similar peripheral blood pressure levels on different treatments. Thus, such technology may prove most useful in the non-invasive evaluation of differential drug effects on central blood pressures. To go beyond this into routine clinical practice would require the demonstration that central pressures derived from pulse wave analysis provide more useful prognostic information with regard to cardiovascular outcome than conventional blood pressure measurement alone. ‘They will’ says the evangelist, ‘they will not’ says the sceptic – I think they might.
No takes yet. Share an insight, caveat, or question.
Bryan Williams (2004) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: