Age- and fitness-related alterations in vascular sympathetic control

Abstract

In the current study we explored (1) if there were differences in sympathetic activity and baroreflex function by age, sex, or physical activity status, (2) if any aspect of baroreflex function related to differences in resting sympathetic activity, and (3) if mechanical and/or neural baroreflex components related to differences in integrated baroreflex gain. Electrocardiogram, blood pressure, carotid diameter and muscle sympathetic nerve activity were recorded continuously at rest and during sequential bolus injections of sodium nitroprusside and phenylephrine in 22 young, 21 older sedentary and 10 older trained individuals. Analyses of co-variance were used to examine age, sex and training status differences and to explore the explanatory power of integrated baroreflex gain and its mechanical and neural components. Training status and sex influenced neither resting sympathetic outflow nor sympathetic baroreflex gain components. Older subjects had a smaller mechanical component and a strong tendency towards a greater neural component of the sympathetic baroreflex during both pressure falls and pressure rises. Opposing age-related changes in mechanical and neural components resulted in a smaller integrated gain during pressure falls, but a greater integrated gain during pressure rises in older subjects. Thus, in older individuals, compromised sympathetic activation to pressure falls was owing to the stiffening of barosensory vessels, whereas the more sensitive sympathoinhibition to pressure rise was due to an increased neural control. Enhanced neural control with age, however, did not contribute the increased resting sympathetic outflow, which indicates that these two changes are probably driven by distinct neural mechanisms.

Baroreflex control of arterial pressure derives from the interaction of mechanical transduction of pressure into barosensory vessel stretch and neural transduction of stretch into target organ response (Hunt et al. 2001a). The ability to determine mechanical and neural constituents of the arterial baroreflex has provided specific insight to changes in autonomic control with age in humans. For example, the age-related decrease in baroreflex control of cardiac vagal outflow was shown to result from deficits in both mechanical and neural baroreflex components, whereas the lower resting cardiac vagal outflow with age was found to relate only to the decline in the neural component (Kaushal & Taylor, 2002). Moreover, neural plasticity appears to be the main contributor to the maintained baroreflex gain observed in older athletes (Hunt et al. 2001b) and to the maturation of vagal function in children and young adults (Lenard et al. 2004) However, it is more difficult to assess baroreflex control of vascular sympathetic activity and therefore changes with age and/or exercise training in this limb of the arterial baroreflex remain much less explored.

While measurement of vagal baroreflex gain and its components requires the recording of vagal activation to pressure rises (Hunt et al. 2001a), measurement of sympathetic baroreflex gain seems to demand the combined registration of sympathetic activation to pressure falls and sympathoinhibition to pressure rises (Rudas et al. 1999). This latter approach, however, has at least two major limitations. First, potentially different effects of demographic parameters (e.g. age, sex, training status) on sympathetic activation and sympathoinhibition may be concealed. Second, combination of pressure falls and rises may not be applied in the presence of hysteresis. Though sympathetic baroreflex gain was reported to show no hysteresis with pressure falls and rises (Rudas et al. 1999), the carotid artery pressure–diameter relationship (e.g. the mechanical component of baroreflex) exhibits hysteresis in the systolic (Studinger et al. 2007) and probably in the diastolic pressure range. Thus, calculation of sympathetic baroreflex gain components necessitates the separate study of baroreflex function during pressure falls and rises.

To compensate for the dearth of information about human sympathetic baroreflex physiology, we did exploratory work to determine the effects of age, sex and maintained physical activity on vascular sympathetic baroreflex function. We took the direction of pressure changes into account to reveal possibly different effects on sympathetic activation and sympathoinhibition. We also examined the mechanical and neural aspects of sympathetic baroreflex control and explored possible inter-relations between sympathetic baroreflex function and resting sympathetic activity. Our results provide some of the first insight to age-related adaptations in the neural control of vascular sympathetic activity and its role in overall autonomic cardiovascular regulation and challenge the notion of unchanged sympathetic baroreflex gain with age.

Methods

Subjects

Twenty-two healthy young sedentary (10 women), twenty-one healthy older sedentary (10 women) and ten healthy older trained (all males) volunteers participated in the study. All subjects were non-smokers, normotensive, and free of overt autonomic or cardiovascular disease. Subject characteristics are summarized in Table 1. Sedentary subjects were normally physically active as assessed by the Stanford Physical Activity Questionnaire, and had aerobic capacities below the 80th percentile for age and sex. Older trained subjects performed aerobic exercise for a minimum of 30 min and 5 times per week, had maintained their current training level for at least 10 years, and had maximal aerobic powers above the 90th percentile for individuals of their age and sex. Standard criteria were applied to ensure all subjects achieved maximal aerobic consumption during exercise testing (i.e. respiratory exchange ratio >1.10, heart rate at least 85% of age-predicted, plateau in oxygen consumption with increasing workload, and a maximal rating of perceived exertion). Older volunteers were screened for the presence of coronary and carotid vascular disease. Women were not pregnant, not on oral contraceptives, and not on hormone replacement therapy. The studies conformed to the standards set by the latest revision of the Declaration of Helsinki. All subjects gave written informed consent prior to participation, and the local ethics committee approved the study.

Table 1.

Subject characteristics

	Young females	Young males	Older females	Older males	Masters athletes
Age (years)	25 ± 3	24 ± 3	62 ± 6	63 ± 6	60 ± 4
Body mass index (kg m–2)	21.1 ± 1.8	23.7 ± 2.3	25.2 ± 2.1	26.3 ± 2.8	22.4 ± 1.7
Resting heart rate (beats min−1)	68 ± 8	65 ± 7	63 ± 7	59 ± 7	51 ± 7
Systolic blood pressure (mmHg)	114 ± 14	117 ± 9	123 ± 9	125 ± 10	121 ± 8
Diastolic blood pressure (mmHg)	67 ± 9	64 ± 6	71 ± 8	71 ± 10	69 ± 9
(ml min−1 kg−1)	40.1 ± 7.1	52.1 ± 6.6	28.8 ± 4.2	34.4 ± 6.7	49.2 ± 7.6

Measurements

Arterial blood pressure via finger photoplethysmography (Finapres, Ohmeda, Louisville, CO, USA), standard three-lead electrocardiogram and respiratory excursions (Respitrace) were recorded continuously throughout the study. Multiunit postganglionic muscle sympathetic activity was recorded from the peroneal nerve (Vallbo & Hagbarth, 1967). Neural activity was amplified, band pass filtered, rectified and integrated to create the sympathetic neurogram. During baroreflex testing, ultrasound images of the common carotid artery ∼1.5 cm proximal to the bifurcation were obtained using a 7.5 MHz linear array transducer (Hewlett-Packard, Andover, MA, USA) and a commercially available data acquisition system (PCI DT-3152 Frame Grabber, Data Translation, Marlboro, MA, USA; CVI Acquisition, Information Integrity, Maynard, MA, USA).

Protocol

All subjects were studied in the morning at least 12 h postprandial, at least 24 h since refraining from consumption of caffeine or alcohol, and at least 48 h after their last exercise bout. Supine subjects were instrumented, a sympathetic recording obtained, and data were acquired for a 5 min baseline period to determine resting sympathetic activity. Subsequently sympathetic baroreflex function was tested via two to three trials of the modified Oxford technique. This technique involves bolus injections of 100 μg of sodium nitroprusside followed in 60 s by a bolus of 150 μg of phenylephrine hydrochloride that generally produce an initial ∼15 mmHg arterial pressure fall followed by an ∼15 mmHg pressure rise above resting supine levels. Carotid artery image acquisition started ∼30 s before sodium nitroprusside injection and ended ∼2 min after phenylephrine injection. Triggered from the R wave of the ECG, carotid artery images were acquired with 30 Hz frequency; 15 consecutive carotid images were acquired with each trigger to encompass both end-diastolic and peak-systolic carotid diameters (Hunt et al. 2001a).

Between each modified Oxford trial at least 10 min was allowed for a complete recovery of pressure, carotid diameter and sympathetic activity.

Data analysis

Arterial pressures were determined using signal processing software (WINDAQ, Dataq Instruments, Akron, OH, USA). Diastolic blood pressures were derived from the minimum of the pressure waveform. Sympathetic bursts were identified using a fully automated program designed and validated in our laboratory (Hamner & Taylor, 2001). The mean voltage neurogram was normalized by calibrating the height of the largest set of bursts during baseline recording to a value of 1000 arbitrary integration units (AIUs). This normalization allowed for calculation of relative changes in sympathetic activity independent of microelectrode distance from the nerve fascicle, and it also accounted for differences in signal-to-noise ratios. Diastolic carotid artery diameters were obtained via image analysis software developed in our laboratory. Sympathetic baroreflex trials were excluded from analysis whenever there was (1) increased noise in the sympathetic nerve signal due to muscle tension, (2) loss of sympathetic nerve recording due to motion of the subject, and/or (3) loss of carotid image due to swallowing, motion of the subject's head or motion of the investigator's hand.

Resting sympathetic activity was expressed as burst frequency (activity per minute), and also as burst incidence (activity per 100 beats) to account for inter-group differences in resting heart rate.

We assessed baroreflex function from the relations between sympathetic activity and arterial pressure, carotid diameter and arterial pressure, and sympathetic activity and carotid diameter. These represent the integrated gain of the vascular sympathetic baroreflex, and the mechanical and neural components of the reflex. We examined responses to nitroprusside (beginning with the onset of the pressure decrease and continuing to the lowest pressure after nitroprusside injection) and to phenylephrine (beginning with the onset of pressure rise and continuing to the highest pressure after phenylephrine injection) separately, to reveal potential age-related differences in responses to pressure falls and rises. Total integrated sympathetic activity (i.e. the product of burst frequency and average burst area in arbitrary units), diastolic pressures, and diastolic diameters were associated within each beat. Raw, beat-by-beat data were used to fully account for cardiac cycles associated with zero sympathetic activity. Theoretically, at high pressures, when baroreflex-mediated sympathoinhibition is complete, no cardiac cycle should display sympathetic activity, whereas at low pressures, when sympathoinhibition is absent, all cardiac cycles should be associated with sympathetic activity. However, the latter is seen in only extreme conditions that result in profound hypotension, such as rapid ventricular tachycardia (when, in fact, sympathetic activity tends to lose its discrete bursting pattern) (Smith et al. 1991). Therefore, pharmacological baroreflex testing tends to show a relationship between sympathetic activity and pressure that moves from cardiac cycles with low pressures and a high, but not uniform association with sympathetic activity to cardiac cycles with high pressures and absent sympathetic activity. With rapid pressure changes, however, the ability to obtain equal samples of sympathetic activity across the observed arterial pressure range is limited, resulting in over-representation of cardiac cycles with zero activity. To compensate for this error, cardiac cycles with zero sympathetic activity were weighted across the observed pressure range. Zeros below the lowest pressure associated with a sympathetic burst were considered as ‘erroneous’ zeros and assigned a weight of 0. Zeros above the highest pressure associated with a sympathetic burst were considered as ‘true’ zeros and assigned a weight of 1. Between the lowest and highest pressures associated with sympathetic bursts, zeros were assigned a weight progressively increasing from 0 to 1, proportional to the range of pressures observed (see Fig. 1). To derive a robust linear gain for baroreflex-mediated sympathoinhibition, we excluded all data 2 mmHg above the greatest pressure associated with a sympathetic burst for integrated gain and all data 0.05 mm above the greatest carotid diameter associated with a sympathetic burst for the neural component. Deriving a linear gain for the mechanical component did not require elimination of any data since the pressure–diameter relation tends to be linear (Sugawara et al. 2000). However, to ensure that threshold or saturation regions were not included, we applied a piecewise linear regression to determine the changes in slope (if any) across the range of observed pressures in the falling and rising phases of drug administration. This approach has been described previously (Studinger et al. 2007). The result of these analyses provided slopes for integrated baroreflex gain (sympathetic activity and diastolic pressure), the mechanical baroreflex component (diastolic diameter and diastolic pressure), and the neural baroreflex component (sympathetic activity and diastolic diameter).

Figure 1. Relation of pressure to sympathetic activity during pressure rise in a young healthy individual.

Size of the circles is proportional with their weight. × represents data with a weight of zero.

Statistical analysis

The results for the linear gains from each subject for each trial were analysed to determine: (1) if there were differences in sympathetic activity and baroreflex function by age, sex, or physical activity status, (2) if any aspect of baroreflex function related to differences in resting sympathetic activity, and (3) if mechanical and/or neural baroreflex components related to differences in integrated baroreflex gain. Analyses of co-variance were used to examine age, sex and training status differences and to explore the explanatory power of integrated baroreflex gain and its mechanical and neural components. Main effects and interactions were considered significant when the P value was less than 0.05. Different trials from the same subject were not averaged, which means that the observations in the regression were not completely independent of each other. To correct for this, we used cluster adjusted robust standard errors (Williams, 2000). Hysteresis between relations encompassing falls and rises were compared using the Chow test (Chow, 1960; Neter et al. 1989). All values are expressed as means ± standard deviation.

Results

Recordings for sympathetic baroreflex gain were successful for only one trial in 12 subjects, for two trials in 37 subjects and for three trials in four subjects. Five minutes’ resting sympathetic activity was recorded successfully in every subject.

Resting sympathetic activity data are displayed in Table 2. Resting sympathetic outflow was greater with age, but was not find different between males and females. There was no interaction between age and sex. In addition, older sedentary men and masters athletes differed only in their burst frequency but not in any other measures of resting sympathetic activity. Contrary to our expectations, neither mechanical nor neural components of sympathetic baroreflex gain provided any explanatory power for differences in resting sympathetic activity among our subjects. This was true for components derived from either nitroprusside-induced pressure fall (P= 0.88 and P= 0.89 for mechanical and neural components, respectively) or phenylephrine-induced pressure rise (P= 0.64 and P= 0.12 for mechanical and neural components, respectively).

Table 2.

Effects of age, sex and training status on resting sympathetic outflow

Resting sympathetic activity	Young females	Young males	Older females	Older males	Masters athletes	Age	Sex	Training status
burst min−1	8.7 ± 4.6	12.1 ± 5.1	21.6 ± 7.0	22.4 ± 8.7	17.1 ± 6.6	P < 0.001	P= 0.277	P= 0.044
burst (100 beats)−1	12.9 ± 7.3	18.7 ± 8.2	35.0 ± 12.1	38.6 ± 15.2	33.6 ± 13.4	P < 0.001	P= 0.153	P= 0.312
AIU min−1	291 ± 178	432 ± 199	766 ± 239	824 ± 310	673 ± 297	P < 0.001	P= 0.158	P= 0.151
AIU (100 beats)−1	435 ± 287	669 ± 304	1240 ± 423	1422 ± 542	1339 ± 644	P < 0.001	P= 0.085	P= 0.687

Descriptive data of sympathetic baroreflex function in terms of integrated gain, mechanical and neural components and P values derived from a simple model that contained direction of pressure change, age, sex and training status but no interaction terms as input variables are summarized in Table 3. The hysteresis pattern in mechanical and neural components and in integrated baroreflex gain is displayed in Fig. 2; age-related differences in this example broadly applied to the groups as a whole. The mechanical baroreflex component demonstrated hysteresis in over 90% of trials, and there were smaller slopes during pressure falls as compared to pressure rises regardless of age. Age also had a significant effect on the mechanical component; it was smaller in older subjects whether measured during pressure falls or rises. Sex did not affect the mechanical component, nor did it have any interactions with age or with direction of pressure change (P > 0.5 for both interactions). Likewise, training status had no effect on the mechanical component.

Table 3.

Descriptive data of mechanical and neural components and integrated baroreflex gain during nitroprusside-induced pressure fall (NTP) and phenylephrine-induced pressure rise (PE)

		Young females	Young males	Older females	Older males	Masters athletes	Direction athletes	Ag	Sex	Training status
Mechanical	Fall (NTP)	22.2 ± 10.7	23.8 ± 9.8	14.3 ± 9.3	17.1 ± 9.3	17.2 ± 7.1	P < 0.001	P < 0.001	P= 0.418	P= 0.333
component (μm mmHg−1)	Rise (PE)	28.0 ± 7.7	28.9 ± 10.9	21.2 ± 11.1	21.9 ± 7.5	26.2 ± 10.0
Neural	Fall (NTP)	−545 ± 417	−402 ± 212	−551 ± 366	−646 ± 352	−597 ± 269	P < 0.001	P= 0.124	P= 0.727	P= 0.623
component	Rise (PE)	−351 ± 314	−288 ± 117	−362 ± 147	−403 ± 266	−440 ± 217
(AIU beat−1 mm−1)
Integrated	Fall (NTP)	−16.6 ± 15.4	−17.0 ± 13.3	−11.3 ± 5.0	−14.6 ± 4.7	−13.5 ± 5.1	P= 0.01	P= 0.340	P= 0.361	P= 0.347
baroreflex	Rise (PE)	−10.3 ± 7.4	−10.2 ± 6.7	−9.5 ± 5.3	−12.3 ± 3.9	−15.4 ± 7.2
(AIU beat−1 mmHg−1)

P values are derived from a simple model that contained direction of pressure change, age, sex and training status but no interactions as input variables.

Figure 2. Hysteresis in mechanical and neural components and in integrated sympathetic baroreflex gain in a young and an older individual.

Full circles and continuous lines represent data obtained after nitroprusside (NTP) bolus injection, open circles and dashed lines represent data obtained after phenylephrine (PE) bolus injection.

The neural baroreflex component also demonstrated hysteresis, but this occurred less frequently, in ∼50% of all trials. In addition, the pattern was the opposite of that observed in the mechanical component: the carotid diameter–sympathetic nerve activity relationship was greater (i.e. more negative) during pressure falls both in young and in older subjects. Moreover, in a model that examined interactions between direction of pressure change and age, older subjects had a strong tendency to demonstrate greater responsiveness in the neural component (P= 0.071). Sex had no effect on the neural component and it demonstrated no interaction with age or with direction of pressure change (P > 0.1 for both interactions). Training status had no influence on the neural component.

Baroreflex gain demonstrated hysteresis in ∼40% of all trials. An opposing hysteresis pattern in mechanical and neural components would seem to offset one another; nevertheless, integrated gain differences between falling and rising pressures remained significant. On the other hand, the tendency to a larger neural component with age could minimize the effect of the smaller mechanical component on integrated gain. However, there was a significant interaction between direction of pressure change and age (P= 0.027) such that the magnitude of integrated baroreflex gain hysteresis was significantly smaller in older compared to young subjects. Sex or training status did not have a significant impact on baroreflex gain, and sex demonstrated no interaction with age or with direction of pressure change (P > 0.3 for both interactions).

To explore the contribution of neural and mechanical components to age-dependent changes in baroreflex gain hysteresis, we also created models that contained age, the neural and mechanical components, and an interaction between the two components as input variables and integrated gain during pressure falls and rises as output variables. During pressure falls integrated baroreflex gain was less negative in older compared to young subjects (P= 0.032), even after accounting for differences in the neural component. Thus, despite the fact that the neural component was greater during pressure falls for all subjects and demonstrated a tendency to be greater during both directions of pressure change in older subjects, pressure-induced increases in sympathetic outflow were lesser with age. This may have resulted from a lower stretch-related mechanical component during pressure falls in older individuals. Interestingly, it was the neural (P < 0.001), and not the mechanical, component (P= 0.796) that contributed to integrated gain in our model. Thus, during pressure falls mechanical component determines integrated gain only qualitatively, whereas neural component is an important, quantitative predictor of the integrated gain.

During pressure rises integrated baroreflex gain was more negative in older compared to young subjects (P= 0.002). The interaction between mechanical and neural components was an important determinant of integrated gain for all subjects: those with higher neural and or higher mechanical gains had a greater sympathoinhibitory response (P < 0.001). With regard to age, the interaction between mechanical and neural components means that a greater neural component could sufficiently offset a lower mechanical component such that older individuals demonstrated a greater sympathoinhibition with rises in pressure.

Discussion

Our data demonstrate for the first time that integrated baroreflex control of vascular sympathetic outflow exhibits lesser sympathetic activation and greater sympathoinhibition with age. Integrated sympathetic baroreflex gain is importantly determined by the interaction of central nervous system and structural components of the reflex loop. In older individuals, during pressure falls, a greater neural control of sympathetic outflow could not offset the effects of carotid vascular stiffening, resulting in lesser increases in sympathetic activity. However, during pressure rises, the more sensitive neural control in older individuals was able to surmount structural deficits, effectively enhancing baroreflex mediated sympathoinhibition. Sex and physical activity status had no effect on sympathetic autonomic control and no evidence was found for a clear relationship between neural control and sympathetic outflow; that is, higher neural control of vascular sympathetic outflow was not strictly proportional to higher resting sympathetic outflow with age.

Most research to date has found neither a decrement nor an enhancement in arterial baroreflex control of sympathetic outflow with human ageing (Ebert et al. 1992; Matsukawa et al. 1994, 1996; Davy et al. 1998b; Rudas et al. 1999). The only exceptions are studies reporting less sympathoinhibition with Valsalva's manoeuvre (Matsukawa et al. 1998) and more sympathetic activation with lower body suction (Davy et al. 1998a). These stimuli, however, are heterogeneous, unloading cardiopulmonary baroreceptors (Taylor et al. 1995; Smith et al. 1996) as well as differentially unloading aortic and carotid baroreceptors (Smith et al. 1996). Moreover, the use of relative changes instead of absolute values may be deceptive, as a lower relative change in sympathetic traffic in the elderly may simply reflect their higher resting sympathetic activity. Our data differ from the majority of previous results as we demonstrate lesser sympathetic activation yet greater sympathoinhibition in older individuals. Had we combined pressure falls and rises, we probably could have reproduced data that indicate no age-related difference in baroreflex sympathetic control. However, hysteresis in a majority of baroreflex tests did not allow pooling of data across both falling and rising pressures to calculate a single baroreflex sympathetic gain. Interestingly, the only previous study examining baroreflex hysteresis in the sympathetic limb reported no significant differences in baroreflex-mediated sympathetic responses during pressure falls and rises (Rudas et al. 1999) However, the conclusion that hysteresis is not observed in the sympathetic limb of the baroreflex was based on slopes derived from data excluding all pressure levels with zero sympathetic activity, and an approximate 50% difference in slopes obtained from five subjects that failed to achieve significance (P= 0.06). Our analysis of not only integrated gain, but also the pressure–barosensory vessel diameter relationships clearly demonstrates that baroreflex hysteresis is present in not only the cardiac vagal limb of the reflex (Studinger et al. 2007) but also the vascular sympathetic limb. This fact may compel re-examination of prior work basing sympathetic baroreflex gain on data pooled across responses to both pressure falls and rises.

The addition of carotid artery diameter to standard measures of arterial pressure and sympathetic nerve activity allowed us to assess the contributions of carotid elasticity and of afferent–efferent neural control to integrated vascular sympathetic baroreflex gain. We thought it unlikely that these aspects of baroreflex gain remained invariable with age; barosensory vessels demonstrate pronounced age-related stiffening across a wide range of pressures under both static and dynamic conditions (Lenard et al. 2000; Myers et al. 2002), and resting sympathetic activity increases markedly with age (Yamada et al. 1989; Iwase et al. 1991; Ebert et al. 1992; Ng et al. 1993; Narkiewicz et al. 2005). We found that transduction of pressure into carotid stretch was ∼25% lower in our older volunteers, which could contribute to the lesser sympathetic activation during pressure falls. Stiffening of the carotid artery, however, was offset by nervous system adaptations such that during pressure rises, a greater sympathetic baroreflex gain was observed in our healthy older subjects. Increased neural control of sympathetic outflow with age may derive from several sources, among them increased sensitivity of baroreflex afferents. Increased baroreceptor strain sensitivity may be present in conditions characterized by low vessel wall distensibility, such as hypertension and ageing (Andresen et al. 1980). Larger baroreceptor responses to vessel wall strain would result in increased neural control of sympathetic activity. The possibility that increased neural control of sympathetic outflow derives from central neural changes with human ageing is also feasible. For example, data suggest that cerebral noradrenaline turnover increases with age (Esler et al. 2002); this could be a contributor to increased sympathetic sensitivity to baroreflex inputs as well as to increased resting sympathetic outflow. Subcortical noradrenaline turnover and resting sympathetic outflow are correlated, and both increase with hypotensive stimuli (Lambert et al. 1988), implicating noradrenergic forebrain systems in the control of resting levels of and baroreflex control of sympathetic outflow. Furthermore, an indirect influence of vagal deficits may also affect increased neural control of sympathetic outflow with age. The central nervous system integrates autonomic outflows and data suggest a close relation between age-related reductions in vagal outflow and increases in sympathetic activity (Pfeifer et al. 1983; Shimada et al. 1985). According to our previous and current findings that show different neural vagal (Hunt et al. 2001b) and sympathetic adaptations to exercise training, there seems to be no strict interrelation between sympathetic and vagal limbs of the arterial baroreflex at the site of the central nervous system. Nonetheless, we must consider that a potential error in the interpretation of our data might lie with the level at which the output variables were measured. Neurograms allow direct characterization of vascular sympathetic outflow, whereas RR interval is merely a surrogate for cardiac vagal outflow and may not allow accurate assessment of a strict reciprocity between the baroreflex regulation of vagal and sympathetic outflows.

The observation that both resting vascular sympathetic outflow and sympathetic neural control are higher in the elderly suggests a potential link. Elevated sympathetic outflow relates to cardiovascular risk (Leenen, 1999), and possible mechanisms underlying increased resting sympathetic activity with age would have clinical relevance. Although we found that the sympathetic neural component and resting sympathetic activity were equally high in the older subjects, they were not related. This may indicate that these changes derive from different compensatory neural mechanisms or that there is not a simple linear relationship between reflex neural control and resting levels of sympathetic activity.

In contrast to the age-related adaptations, we did not observe any effect of training status on sympathetic baroreflex function. Regarding resting sympathetic outflow, training status had a marginal effect on burst frequency, which might have been owing to the lower heart rate of masters athletes as compared to sedentary individuals. Burst incidence, a heart rate independent measure of resting sympathetic outflow did not differ between trained and sedentary groups. Our findings are inconsistent with the observations of Ng et al. (1994) who reported increased resting sympathetic nerve activity in older athletes. This difference might be owing to the sex distribution of the studies. We studied only males in our endurance-trained group of subjects, while Ng et al. examined both males and females, and group differences were due primarily to the higher sympathetic activity in endurance trained females.

One of the potential limitations of our study is that vasoactive drugs may have an age-dependent effect on vascular smooth muscle, contributing to age-related differences observed in mechanical component. In our study the effect of the vasoactive drugs was not affected qualitatively by ageing as the injection of nitroprusside resulted in decreased carotid diameter, whereas the injection of phenylephrine resulted in increased carotid diameter both in young and older individuals. These findings are consistent with a prior work suggesting that during drug-induced changes in blood pressure, baroreceptor activity in humans is influenced more by passive stretch than by local smooth muscle contraction (Bonyhay et al. 1997). There is no evidence that greater diameter changes observed in the young subjects would be due to a less pronounced direct drug effect on carotid smooth muscle rather than the direct effects of arterial pressure. Moreover, it has been shown that both sodium nitroprusside-mediated vasodilatation (Taddei et al. 1995) and α₁ adrenergic responsiveness (Seals & Dinenno, 2004) decrease with ageing, and therefore a less pronounced direct drug effect would be expected in older, not in younger individuals.

Due to technical limitations, arterial dynamics were observed only in the carotid artery and we did not consider aortic and cardiopulmonary baroreceptor populations. Though aortic baroreflex was suggested to be the dominant regulator of sympathetic outflow during steady-state elevations of arterial pressure (Sanders et al. 1988), its relative contribution to the regulation of sympathetic outflow during acute pressure changes has never been studied. Cardiopulmonary input into systemic baroreflex control has been suggested to remain unchanged during a modified Oxford trial (O’Leary et al. 2005). Therefore, we consider the lack of measurement of aortic diameter or central venous pressure as a limitation though we think it did not influence the main conclusions of our study.

Our data suggest that ageing affects sympathetic responses to pressure falls and rises differently: while sympathetic activation to pressure fall decreases with age owing to stiffening of the barosensory vessels, sympathoinhibition to pressure rise increases due to a more sensitive central neural control. These findings extend previous studies that suggested a link between increased vascular stiffness and the more prevalent orthostatic hypotension in the elderly (Mattace-Raso et al. 2006) by demonstrating an actual mechanism – a decreased sympathetic response to pressure drop – through which arterial stiffening may act on orthostatic challenges. A more sensitive sympathoinhibition with age may be a possible compensatory mechanism for the lesser capacity to augment cardiac vagal outflow to provide adequate pressure control against hypertensive challenges with age.