The ups and downs of assessing baroreflex function

Asymmetry in the baroreflex-mediated cardiovagal response between falling and rising blood pressure is well recognized but the underlying mechanisms have remained incompletely understood. It was initially believed that the less marked R–R interval shortening in response to blood pressure falls, compared with the more pronounced R–R interval lengthening in response to blood pressure rises, were attributable to the baroreflex operating point being located close to the threshold of the sigmoidal stimulus–response curve (Eckberg & Sleight, 1992). However, since a lower baroreflex gain in response to falls in pressure persists even when pressure is first elevated above resting levels, it is likely that the position of the operating point cannot fully explain ‘baroreflex hysteresis' (Rudas et al. 1999). More recently it has been reported that baroreflex hysteresis is due to steeper increases in carotid artery diameter during a hypertensive challenge compared with a hypotensive challenge (Bonyhay et al. 1997; O'Leary et al. 2005), and thus attributable to the viscoelastic properties of the barosensory vessels (i.e. a mechanical event). However, in a recent issue of The Journal of Physiology, Studinger et al. (2007) further advance our understanding of baroreflex hysteresis by beginning to unravel the complex interactions between the underlying components of the baroreflex arc in humans. Using elegant methodologies, Studinger et al. (2007) examined the contributions of both the mechanical (i.e. the transduction of pressure into carotid stretch) and neural (i.e. transduction of carotid stretch into vagal outflow) components of the baroreflex arc to hysteresis. Baroreflex perturbations were achieved by pharmacological manipulation of blood pressure using a bolus intravenous infusion of sodium nitroprusside followed by a bolus infusion of phenylephrine hydrochloride 60 s later (i.e. the modified Oxford technique), while R–R interval (EKG) and common carotid artery diameter (Doppler ultrasound) were continuously monitored. The overall integrated baroreflex responsiveness (systolic blood pressure versus R–R interval) for falls and rises in arterial pressure were separately determined and hysteresis was defined as a difference in gain (i.e. linear slope) or set point (i.e. position), between the downward and upward baroreflex stimulus–response relationships. The pattern of hysteresis derived from the interaction of both the mechanical (systolic blood pressure versus carotid diameter) and neural (carotid diameter versus R–R interval) components was then determined. Overall, Studinger et al. (2007) reported a remarkably complex pattern of baroreflex behaviour in response to hypo- and hypertensive challenges. Although baroreflex hysteresis was demonstrated in all subjects, it was not present in every trial and overall the mean values for integrated baroreflex responsiveness for falling and rising pressures were not significantly different. This intricate pattern of baroreflex behaviour was found to be explained by a rather complex interaction between the mechanical and neural components of the baroreflex arc. Indeed, despite the lack of a difference in the group values, in every trial hysteresis was observed in either the mechanical or neural components. Furthermore, when hysteresis resulted from a difference in gain, both components contributed to this difference. However, when hysteresis was attributable to differences in set point, in the majority of cases the neural component was the primary contributor. Interestingly, when an absence of hysteresis in overall integrated baroreflex responsiveness was noted, the mechanical and neural components were found to be offsetting one another. Thus, contrary to previous results, these data demonstrate that hysteresis is not simply due to the mechanical properties of the carotid vessels, but is driven by the interaction between mechanical and neural aspects of the baroreflex. In addition, the interaction between these components did not contribute to hysteresis in the same manner during consecutive trials in all subjects suggesting that the ‘baroreflex does not generate the same pattern of response each time it is engaged.’ These important mechanistic insights significantly advance our overall understanding of baroreflex hysteresis in humans and in so doing further illustrate the complexity of baroreflex control of the heart. Aside from the demonstration of a neural contribution to baroreflex hysteresis, perhaps a more important contribution made by Studinger et al. (2007) which may be overlooked, is the development of an analytical approach that allows for the objective determination of baroreflex gain. As the pressor responses to the modified Oxford technique normally encompass the non-linear portions of the baroreflex stimulus–response curve, in order to determine a linear gain, data points that fall on the non-linear portion of the curve are commonly eliminated by subjective identification. This increases the potential for error because removal of even a single data point can often dramatically influence baroreflex gain, which can lead to a significant over- or under-representation of baroreflex responsiveness amongst subjects as well as studies. However, by using a statistical technique such as piecewise linear regression, the authors introduce an objective method to identify and eliminate the non-linear portions of the baroreflex curve. This technique searches for breakpoints that occur in the data at the non-linear portions of the curve and allows for the statistical identification of the linear portion of the blood pressure–heart period relationship, thus eliminating subjectivity and investigator bias. As such, this methodology offers a significant advantage in the analysis of the modified Oxford assessment of baroreflex control. In addition to providing an objective method to identify linear portions of the baroreflex stimulus–response relationship, the piecewise linear regression technique employed by Studinger et al. (2007) may also help facilitate the future consideration of separate upward and downward baroreflex slopes. This is important because the significant intricacies in baroreflex function arising from hysteresis may be overlooked if only a single measure of overall cardiovagal baroreflex gain is reported. Although in a previous methodological evaluation of the modified Oxford technique and hysteresis, Rudas et al. (1999) concluded that the ‘separate responses to falling and rising pressures yield more information about vagal baroreflex gain than lumped results’, few studies have fully incorporated this approach. Indeed, in a recent review of 23 studies that have used the modified Oxford technique in humans, and cited the classic study of Rudas et al. (1999), we found that only four articles presented both the response to hypotensive and hypertensive stimuli. More commonly, the cardiac responses to nitroprusside and phenylephrine were combined, or just the hypertensive response to phenylephrine considered. Although the latter approach obviates the problems of hysteresis, not assessing baroreflex responses to hypotension may overlook important information regarding the directional sensitivity of the cardiac response to changes in arterial blood pressure. This may be very important from a clinical perspective because of the high prevalence of orthostatic intolerance in several pathophysiological conditions as well as with normal ageing (Grubb & Karas, 1999). Also, overall baroreflex gain may not always reflect an alteration in baroreflex responsiveness to falls in pressure. For example, we recently studied a diabetic patient who had an impaired cardiac response to hypotension, but a normal response to a hypertensive challenge. If in this patient we only considered the overall integrated baroreflex responsiveness, which also appeared normal, the impaired cardiac baroreflex response to hypotension would have been missed (see Fig. 1). Thus, a separate evaluation of baroreflex responsiveness to hypotension and hypertension is important. Analysis of overall baroreflex gain alone may mask an impaired ability to respond to hypotension This figure illustrates an impaired cardiac response to a nitroprusside (NTP)-induced hypotensive challenge in a diabetic patient compared with a control subject whereas the response to a phenylephrine (PE)-induced hypertensive challenge is similar. In this example, consideration of overall integrated baroreflex gain alone (inset) would lead to impairments in baroreflex regulation of blood pressure being overlooked. This illustrates the importance of separately examining the cardiac responses to falls and rises in blood pressure in order to fully assess baroreflex function using the modified Oxford technique. Given the methodological limitations associated with human experimentation, the authors should be applauded for their insightful mechanistic approach. Nevertheless, several points are worthy of consideration. The results do not encompass the aortic baroreceptors and although these typically work in parallel with the carotid baroreceptors, this may not always be the case. As such, examination of the mechanical and neural components of the aortic baroreflex, using the methodologies employed by the authors, may provide further insight into baroreflex hysteresis. In addition, although a clear neural contribution to baroreflex hysteresis was described, the precise mechanism responsible (e.g. baroreceptor afferent output and conduction, central integration, efferent autonomic activity and/or cardiac responsiveness to autonomic adjustments) for the pattern of hysteresis observed remains unknown. This will likely require integration with more invasive animal studies. One final point worthy of mention is that although the authors have elucidated the underlying mechanisms responsible for hysteresis in the cardiovagal limb of the baroreflex, the complex interplay between mechanical and neural components of the sympathetic baroreflex arc remains undefined. Indeed, despite previous reports suggesting a lack of hysteresis in the sympathetic arm of the baroreflex (Rudas et al. 1999; O'Leary et al. 2005), whether an interaction exists between the mechanical and neural components underlying the baroreflex–sympathetic stimulus–response relationship remains to be elucidated. In summary, hysteresis in the cardiac vagal limb of the baroreflex is well recognized (Eckberg & Sleight, 1992); however, prior to the recent work of Studinger et al. (2007) it was commonly believed that baroreflex hysteresis was due to the mechanical properties of the blood vessels in which the baroreceptors are situated (Bonyhay et al. 1997; O'Leary et al. 2005). The contribution of Studinger et al. (2007) is to reveal the complexity of both mechanical and neural contributions of the baroreflex arc to hysteresis in the cardiovagal control of heart period. These findings further suggest that separate consideration of gains for both the ups and downs of baroreflex regulation of blood pressure may provide more specific insight to alterations in autonomic control in physiological and pathophysiological conditions.