Reflections on determinants of augmentation index

The changes in arterial pressure, as seen with age, are normally associated with changes in the pressure pulse waveform [1]. The characteristic widening of the pressure pulse and change in wave contour are generally attributed to the effects of loss of elasticity of the aorta and large conduit arteries [2]. Loss of elastic function of the arterial wall can be due to degeneration of the wall load-bearing components as well as increased distending pressure. Since the arterial system is a distributed network, changes in wall stiffness affects wave propagation, such that pulse wave velocity (PWV) is increased with stiffening of the conduit vessels, leading to rapid pulse transit times, and early wave reflection [3]. These general physical principles have led to the notion that information can be extracted from the arterial pulse that can quantify the effects of changes in both central compliant arteries and peripheral resistive vascular beds. Thus, changes in the pulse waveform features can give information on timing and amount of wave reflection and provide a means to assess vasoactive agents as well as a type of integrated parameter related to vascular aging [4]. One parameter related to pulse waveform features, and that has attracted a substantial amount of interest in recent years, is the augmentation index (AIx). Computation of AIx relies on the detection of waveform features in addition to conventional systolic peak. Thus, in the ascending aorta, if an early systolic inflection (Pi) is detected above diastolic pressure (DBP) before the systolic peak (SBP), AIx quantifies the ratio of the augmented pressure (SBP-Pi) above Pi. It follows that if Pi is at the peak, AIx is zero, and if Pi is after SBP, AIx is negative. From simultaneous measurements of pressure and flow in the ascending aorta, peak flow generally coincides with the location of Pi. Hence, if AIx is zero or negative, peak pressure is essentially due to the cardiac ejection into the compliant aorta. If AIx is positive, peak pressure is due to secondary phenomena emanating from the arterial load, such as wave refection, since pressure rises while flow decreases. Clearly, AIx is related to cardiac and arterial function. However, since cardiac ejection is assumed to vary much less than the pressure pulse, the use of AIx is generally made in isolation, such that information is inferred from a single pressure recording. This has led to investigations on how to interpret AIx as an index of arterial stiffness and wave reflection in conditions of concomitant changes in chronotropic and inotropic cardiac properties [5] and vasoactive agents [4,6,7]. Augmentation index can be determined from peripheral pulses [8,9] as well as from central locations at the carotid artery [10]. However, the main determinants of AIx are still being questioned. Recent studies have attempted to provide evidence of determinant factors of AIx and to quantify the contributions from cardiac origin (flow) and from the arterial vasculature (regional compliance, Windkessel pressure and wave reflection) [11–14]. The issue in question is whether the main contributions are from wave propagation phenomena or from the Windkessel effects that determine the so-called reservoir pressure. Some of these studies have been done in the context of ageing, when there are marked changes in AIx and PWV [2,11]. The study by Huijben et al.[15] in this issue of the Journal investigates changes in AIx and PWV in a group of patients with autonomic failure in response to head-up tilt. Their hypothesis is that if wave reflection is dominant, AIx should be stable with tilt in autonomic failure patients, since peripheral vascular tone would not change with fall in blood pressure. Their finding was that AIx in autonomic failure patients showed a large change with increasing tilt angle, following similar changes in mean blood pressure and aortic PWV. In age-matched control patients AIx changed to a lesser degree with tilt, but mean blood pressure and aortic PWV increased. These results were interpreted to suggest that the capacitive or reservoir function of the aorta is a dominant determinant of AIx, since the large fall in AIx would be expected due to the reduced distending pressure in the aorta and decreased left-ventricular ejection due to altered venous return with tilt. The other supporting factor was that no significant relation was found between AIx and PWV in the autonomic failure patients. Whereas there are some potential limitations, the study by Huijben et al. addresses an important issue with respect to elucidating the determinants of AIx as well as the use of the reservoir pressure concept and wave reflection. Whereas the study is well done, neither wave reflection nor reservoir pressure was explicitly measured or calculated. They were inferred from the results obtained and from derived statistical correlation measures. Although AIx in the study was corrected for heart rate, it is not evident that this is a corresponding correction for change in ejection duration during tilt. This of course would affect the timing of the waveform features from which AIx is determined. Other considerations are the effects of the graded hydrostatic pressure component in the aorta with tilt angle. Indeed, this is seen in the increase in mean pressure and aortic PWV in control patients. The hydrostatic effects would be similar for both control and autonomic failure groups, but it is not evident in the autonomic failure group due to the larger change in mean blood pressure. That is, the increase in transmural pressure in the lower abdominal aorta with head-up tilt would have an additional effect on PWV. This was confirmed in a previous study by Xu et al.[16]. An average 26.8% increase in carotid-femoral PWV from the supine to the upright position was observed in a group of 14 healthy male individuals (age 25–37 years) independent of changes in arterial pressure and heart rate. In eight older male individuals (47–67 years) the mean increase was lower (22.1%, P < 0.001), consistent with a reduced response to a similar gravity-related pressure change by a stiffer aorta in older patients. Thus, in both younger and older patients, the tilting effect would result in earlier return of reflected waves. AIx was found to have a similar decrease of 6% per 30 degree of tilt in both young and old groups, whereas AIx corrected for heart rate at 75 beats/min decreased 3% for the old and 1.5% for the young per 30 degree of tilt. If these factors were taken into account, the relation between AI and aortic PWV would be different, and it is not clear whether the P value of 0.07 indicating no association as shown in Fig. 2 in the study by Huijben et al.[15] would increase or decrease with the possibility of altering the conclusions of the study. The other important factor is the interpretation of the reservoir concept in relation to the determinant of AIx. The main changes of AIx occur in systole. However, the underlying assumption of the reservoir concept is that the reservoir pressure changes in a similar manner in both systole and diastole. This has been shown to contain inherent inconsistencies in relation to the separation of the pressure wave into a reservoir pressure and an excess pressure, since in so doing it introduces substantial artefact [17,18]. Correcting for this artefact produces a substantial increase in the amount of wave reflection present in the arterial pressure pulse [17]. The surge of interest in the use of AIx as a surrogate measure of vascular properties has also generated a need for accurate assessment of measurement methodology [19] and interpretation of data in elucidating underlying mechanisms. Studies using postural interventions [20] can provide additional means to interrogate specific components of the arterial system that contribute to waveform features in the arterial pulse that can provide information beyond the conventional measurement of brachial cuff blood pressure. ACKNOWLEDGEMENTS Conflicts of interest There are no conflicts of interest.