Over the past three decades the changes in sympathoadrenal function that occur with age in healthy adult humans have been systematically studied using a combination of neurochemical, neurophysiological and haemodynamic experimental approaches. The available experimental evidence indicates that tonic whole-body sympathetic nervous system (SNS) activity increases with age. The elevations in SNS activity appear to be region specific, targeting skeletal muscle and the gut, but not obviously the kidney. The SNS tone of the heart is increased, although this appears to be due in part to reduced neuronal reuptake of noradrenaline (norepinephrine). In contrast to SNS activity, tonic adrenaline (epinephrine) secretion from the adrenal medulla is markedly reduced with age. This is not reflected in plasma adrenaline concentrations because of reduced plasma clearance. Despite widely held beliefs to the contrary, sympathoadrenal responsiveness to acute stress is not exaggerated with age in healthy adults. Indeed, adrenaline release in response to acute stress is substantially attenuated in older men. The mechanisms underlying the age-associated increases in SNS activity have not been established, but our preliminary data are consistent with increased subcortical central nervous system (CNS) sympathetic drive. These changes in sympathoadrenal function with advancing age may have a number of important physiological and pathophysiological consequences for human health and disease.
The sympathetic nervous system (SNS) plays a critical role in the maintenance of physiological homeostasis in general, and arterial blood pressure in particular, under basal (resting) conditions and in response to acute stress. Post-ganglionic sympathetic neurons innervating the heart and resistance vessels help control cardiac output, arterial blood pressure and regional vascular conductance, thus ensuring the proper perfusion of vital organs. SNS stimulation of adrenaline (epinephrine) release from the adrenal medulla contributes importantly to the regulation of cardiovascular function as well as energy metabolism. The SNS also has a key role in the regulation of internal body temperature. In addition to these normal physiological interactions, the SNS has been implicated in a number of common clinical disorders including hypertension, congestive heart failure, sudden cardiac death, the insulin resistance (metabolic) syndrome and obesity.
Adult human ageing is associated with a number of important changes in physiological function and regulation to which the SNS may contribute (Rowe & Troen, 1980; Folkow & Svanborg, 1993; Lakatta, 1993; Seals, 1993). Moreover, the incidence of many chronic disease states, including those mentioned above, increases with advancing age (Biermann & Ross, 1977; DeFronzo, 1979; Schoenberger, 1986; Folkow & Svanborg, 1993; Lakatta, 1993). The changes in the sympathoadrenal system that occur with primary ageing in adult humans, and how such changes may impact important physiological and pathophysiological processes, have been systematically investigated by our laboratories and others over the past three decades. This topical review discusses some of the key experimental observations in the area of human ageing and sympathoadrenal function during this period. The review focuses on the results of studies using neurochemical and/or neurophysiological (microneurographical recordings) measures of sympathoadrenal system function. Investigations utilizing measurements derived from spectral analysis of cardiovascular variability are not included because of the difficulty in properly interpreting such results. For additional information on this topic the reader is referred to prior reviews by the authors and others (Rowe & Troen, 1980; Linares & Halter, 1987; Roberts & Tumer, 1987; Esler et al. 1989; Docherty, 1990; Seals, 1993; Seals et al. 1994; Esler, 1995).
Age and the sympathoadrenal system under basal (resting) conditions
SNS. Historically, methods employed to study age-related changes in the SNS in humans have involved measurements of noradrenaline, the primary neurotransmitter released from post-ganglionic sympathetic nerve endings. Initial approaches focused on measuring noradrenaline concentrations in 24 h urine collections and later in plasma obtained from venous or arterial blood samples, the rationale being that elevations in sympathetic nerve firing rates would be manifest as higher concentrations of noradrenaline and vice versa. In general, based on cross-sectional observations, plasma noradrenaline (PNA) concentrations have been reported to increase 10–15 % per decade over the adult age range (Ziegler et al. 1976; Jones et al. 1978; Goldstein et al. 1983). Age-associated elevations in PNA concentrations appear to be more consistently observed and larger when obtained from arterial rather than venous blood samples. Indeed, some studies on rigorously screened, healthy adults have not found significant increases in venous PNA levels with advancing age (Young et al. 1980; Fleg et al. 1985; Taylor et al. 1992a; Ng et al. 1993).
The interpretation of increased total PNA spillover rates as experimental support for age-associated elevations in net whole-body SNS activity, however, must be done with an understanding of the corresponding limitations of this measure. NA release from sympathetic nerve endings is modulated pre-synaptically by adrenergic receptor mechanisms (Langer, 1974). Moreover, 80–90 % of neuronally released NA is taken back up by the sympathetic nerve endings through an active reuptake (Reuptake 1) mechanism (Esler et al. 1990). Thus, changes with age in either or both of these modulatory mechanisms could confound the interpretation of PNA spillover measurements. In this context, there is in vitro evidence for an age-associated decrease in α2 pre-junctional inhibition of peripheral NA release in the rat (Daly et al. 1989; Bucholz & Piper, 1990). This would serve to augment the amount of NA released per unit sympathetic nerve discharge with age and result in an overestimation of SNS activity based on PNA spillover. Although reduced neuronal reuptake of NA has been observed in older adult humans in the heart (Esler et al. 1995b, c) (see below), no age-related differences have been observed systemically (Stromberg et al. 1991). Thus, although there is in vitro evidence supporting impaired peripheral α2-adrenergic modulation of NA release with age, this has not been confirmed in vivo in the intact human. Similarly, neuronal reuptake of NA may be reduced with age in specific organs, but currently there is no compelling support for a significant whole-body effect. Given this, age-related elevations in total PNA spillover can be reasonably viewed as experimental evidence for the concept of a net increase in average SNS activity.
In order to: (1) confirm these findings of increased total PNA spillover as indicating increased central nervous system (CNS) sympathetic outflow with age; and (2) provide insight into the specific regions to which SNS activity is increased with age, direct (intra-neural) recordings of post-ganglionic sympathetic nerve activity to skeletal muscle (MSNA) have been obtained in conscious humans using the microneurographic technique (Sundlof & Wallin, 1978; Wallin & Fagius, 1988). It is widely recognized that central SNS outflow can be regulated in an organ-specific manner (Hasking et al. 1986; Esler et al. 1988, 1990). As such, it is possible that SNS activity could be elevated with age to some tissues but not others.
While these neurophysiological data clearly support the idea of increased SNS activity with age, they provide insight into only a single peripheral tissue – skeletal muscle. The microneurographic technique cannot be used to measure SNS activity to internal organs. Accordingly, to gain further insight into other regions to which SNS activity may be elevated with age (and, thus, contribute to the increase in total PNA spillover), we next performed a series of experiments in which PNA spillover was determined for selective internal organs including the heart, gut and kidneys (Esler et al. 1995a, b, c; Mazzeo et al. 1997). Cardiac PNA spillover rate was found to be almost twice as great in healthy older compared with young men (Esler et al. 1995c); however, this appeared to represent not only increased SNS activity to the heart, but also diminished neuronal NA reuptake (Fig. 2). Hepatomesenteric PNA spillover rates also were found to increase with age by 50 % in healthy men (Mazzeo et al. 1997) (Fig. 2). As neuronal reuptake does not influence hepatomesenteric PNA spillover to the same extent as in the heart (Esler et al. 1990), these findings are consistent with elevations in SNS activity. No significant differences were observed for renal NA spillover rates with age (Esler et al. 1995c).
Taken together, the evidence to date supports the view that primary human ageing is associated with a net activation of the SNS (Fig. 3). PNA concentrations are elevated due to a combination of augmented PNA spillover from sympathetic nerve endings and reduced metabolic clearance of NA. Skeletal muscle is a major target of the increased central SNS activity as well as the gut. Sympathetic tone is elevated in the heart with age in humans, apparently due to both reduced neuronal reuptake of NA and increased cardiac sympathetic nerve discharge. Finally, at present there is no compelling evidence that SNS activity to the kidney is elevated in healthy ageing.
Adrenaline release from the adrenal medulla. Historically, plasma concentrations of adrenaline have been used to determine possible effects of ageing on adrenaline secretion from the adrenal medulla. Generally investigations to date have found that plasma adrenaline concentrations either become slightly lower or do not change across the adult age range (Franco-Morselli et al. 1977; Weidman et al. 1978). As is the case with PNA concentrations, however, the interpretation of plasma adrenaline levels as a measure of secretion from the adrenal medulla is not straightforward given the possibility of age-related changes in clearance. To address this, we employed tracer methodology to study plasma adrenaline kinetics (Esler et al. 1995a). We found that adrenaline secretion from the adrenal medulla was 40 % lower in older compared with young healthy men (Fig. 2). This difference was not reflected in the corresponding arterial plasma adrenaline concentrations, which were not significantly different with age, because plasma clearance was 20 % lower in the older men. In the same investigation, we also examined the possibility that adrenaline is released from the heart, perhaps acting to augment cardiac NA release via stimulation of pre-junctional β-adrenergic receptors. Adrenaline was released from the heart only in the older men, despite the fact that their adrenaline secretion from the adrenal medulla was reduced.
In summary (Fig. 3), in contrast to the increase in SNS activity, adrenaline secretion from the adrenal medulla is markedly reduced with advancing age under resting conditions in healthy humans. The lower secretion in older humans is not apparent from plasma concentrations, which do not change significantly with age, because of a reduction in the rate of clearance of adrenaline from the circulation. Finally, adrenaline is released from the heart at rest in older humans. It is not known if this contributes mechanistically to the aforementioned age-associated increases in cardiac NA spillover via pre-junctional β-adrenergic stimulation.
Mechanisms underlying age-associated changes in the sympathoadrenal system
SNS. Two primary mechanisms have been hypothesized to explain age-related increases in peripheral SNS activity under resting conditions: (1) reduced tonic baroreflex inhibition of ‘normal’ central SNS outflow; and (2) a primary increase in CNS-generated sympathetic nerve discharge.
Both arterial and cardiopulmonary baroreflexes tonically inhibit central SNS outflow in humans (Mancia & Mark, 1983; Mark & Mancia, 1983). Thus one possibility is that this tonic inhibition lessens with advancing age, allowing progressively greater levels of SNS activity to peripheral tissues (Rowe & Troen, 1980). The experimental support for this hypothesis was based largely on results of studies: (1) in humans showing age-related reductions in cardiovagal baroreflex sensitivity (Gribbin et al. 1971; Lindblad, 1977; Cleroux et al. 1989; Ebert et al. 1992); (2) in humans using PNA concentrations as a measure of SNS activity during baroreflex perturbations (Shimada et al. 1985); or (3) in animals in which arterial and/or cardiopulmonary baroreflex control of renal SNS activity was shown to be reduced in senescent animals (Hajduczok et al. 1991a, b). However, in a series of studies performed in healthy humans (Davy et al. 1998a, b; Tanaka et al. 1999), we found that baroreflex control of MSNA was not obviously reduced in older compared with young adults. Other investigators earlier had reported similar findings (Ebert et al. 1992; Matsukawa et al. 1996). In fact, our results indicated that at least one expression of baroreflex control of MSNA (i.e. responses to graded hypovolemia) actually was augmented in older adults (Davy et al. 1998a). There is an age-associated reduction in baroreflex-evoked peripheral vasconstriction (Cleroux et al. 1989; Davy et al. 1998a), but this appears to be due to a decrease in peripheral vascular responsiveness to sympathetic stimulation rather than an inability to evoke the necessary adjustments in SNS activity (Davy et al. 1998a).
The results of earlier studies in young and senescent beagles (Hajduczok et al. 1991a, b) suggested that the marked elevation in basal peripheral SNS activity in the older animals could not be completely explained by a reduction in tonic baroreflex inhibition. Rather, it was postulated that a primary increase in CNS-generated sympathetic outflow also must contribute. Accordingly, we have addressed the potential involvement of this mechanism in preliminary studies measuring brain NA turnover (M. D. Esler & D. R. Seals, unpublished data). In several clinical contexts, most notably cardiac failure and essential hypertension, Esler and colleagues (Ferrier et al. 1993; Lambert et al. 1994, 1995) have demonstrated the possible importance of projections of noradrenergic neurons to the forebrain in generating elevated levels of peripheral SNS activity. Indeed, even in healthy young men in which SNS activity varies only within a narrow normal range, there is a strong and positive relation between NA turnover in the subcortical areas of the brain and peripheral SNS activity (Lambert et al. 1998). Our preliminary findings indicate that subcortical NA turnover under resting conditions is at least 2-fold higher in healthy older compared with young men (317 ± 50 vs. 107 ± 18 ng min−1). This elevation in forebrain NA turnover was positively and significantly related to corresponding age-associated elevations in cardiac NA spillover, which was increased with age. In contrast, cortical NA turnover did not vary with age.
The exact mechanism responsible for the apparent marked increase in subcortical brain NA turnover with age is unclear. In this context, it should be noted that clonidine administration, a central α2-adrenergic receptor agonist that augments pre-junctional inhibition of NA release, evoked similar dose-dependent reductions in PNA concentrations and total PNA spillover rates in young and older men (Featherstone et al. 1987). These results do not support reduced central α2-adrenergic pre-junctional inhibition of NA release as a primary mechanism involved in age-associated increases in forebrain NA turnover. The evidence also is against a reduction in brain neuronal NA reuptake contributing to the increased subcortical NA turnover with age. Reduced neuronal NA disproportionately increases 3-methoxy-4-hydroxy phenylglycol (MHPG) overflow and reduces dihydroxyphenylglycol (DHPG) overflow (Eisenhofer et al. 1991), which was not evident with ageing in our study.
Finally, it is possible that some as yet unidentified humoral signal with either peripheral afferent or direct CNS sympathoexcitatory effects may increase with advancing age and provide a tonic stimulus for the elevation in SNS activity with age. At present, however, there is no compelling experimental support for any such circulating signal.
In summary, the available data do not support the concept that impairments in tonic baroreflex inhibition of central sympathetic outflow play a major role in age-associated increases in peripheral SNS activity in humans. Rather, our preliminary data are consistent with the possibility that elevations in total and/or organ-specific SNS activity may be due, at least in part, to increased activity of noradrenergic neurons in subcortical areas of the brain known to modulate medullary pre-ganglionic sympathetic discharge.
Adrenaline secretion from the adrenal medulla. The mechanism(s) contributing to the reductions in adrenaline secretion from the adrenal medulla with advancing age have not been investigated to date. Possibilities include age-associated: (1) reductions in pre-ganglionic nerve activity to the adrenal medulla; (2) reductions in adrenaline secretion in response to equivalent (or even greater) levels of pre-ganglionic nerve activity; and (3) reduction in adrenaline synthesis and storage in the adrenal medulla.
Age and sympathoadrenal adjustments to acute stress
For purposes of the present review, we define acute stress as a stimulus that requires rapid and, in some cases, marked adjustments in the sympathoadrenal system in order to maintain homeostasis. Over the past 20–30 years there has been a widely held and much emphasized view that even healthy ageing is associated with augmented sympathoadrenal responsiveness to acute stress (Rowe & Troen, 1980). This belief appears to be based largely on early studies showing greater increases in venous PNA concentrations in response to a variety of acute laboratory stressors (Palmer et al. 1978; Young et al. 1980; Barnes et al. 1982; Sowers et al. 1983). These studies had some limitations that could alter, perhaps fundamentally, the interpretation of their results. The first and most obvious are those related to PNA concentrations as a measure of changes in SNS activity with age. For example, reduced neuronal reuptake or systemic plasma clearance of NA, both of which have been reported to occur with age (Esler et al. 1981, 1995c; Veith et al. 1986; Morrow et al. 1987; Marker et al. 1994), would result in greater PNA concentrations in response to a particular stress-evoked increase in SNS activity. Second, in some cases the exact level of stress used was not documented, leaving open the possibility that the older adults may have been subjected to a greater sympathoexcitatory stimulus. Third, the SNS adjustments to stress may be influenced by several factors (disease, obesity, physical activity, gender) that were not always controlled; thus, it is not possible to isolate the effects of the ageing process. Finally, in some of our preliminary studies in this area (Taylor et al. 1991, 1992b; Davy et al. 1995), using well-controlled experimental conditions, we were unable to confirm greater increases in antecubital venous PNA concentrations in response to various forms of laboratory stress in healthy older adults.
Accordingly, we performed a series of investigations aimed at determining if primary ageing is associated with augmented sympathoadrenal adjustments to acute stress (Ng et al. 1994, 1995; Esler et al. 1995a, b; Davy et al. 1997; Mazzeo et al. 1997). An important goal of these studies was to gain insight into possible age-related differences in regional SNS responses as well as adrenaline secretion from the adrenal medulla. Subject characteristics and stress stimuli were carefully controlled in order to experimentally isolate, as much as is possible, the effects of ageing per se. Because changes in sympathoadrenal responsiveness with age could be stimulus specific, we employed several different types of stress including isometric and dynamic exercise, orthostasis, cognitive challenge, local cold stimulation and hypoxia; each of these stimuli produces sympathoexcitation via different afferent and/or CNS pathways.
In general, we found that the absolute increases in measures of net whole-body SNS activity (i.e. arterial PNA concentrations and total PNA spillover) in response to these stressors were not different in young and older healthy adults (Esler et al. 1995a, b; Mazzeo et al. 1997) (middle panel, Fig. 4). With regard to regional SNS activity, the absolute unit increases in MSNA (Ng et al. 1994, 1995; Davy et al. 1997) and hepatomesenteric PNA spillover (Mazzeo et al. 1997) were similar in young and older subjects. In fact, the relative (percentage) increases in these measures of SNS activity actually were smaller in the older adults because of their elevated baseline (resting) levels. In contrast, the increases in cardiac PNA spillover were consistently greater, in some cases markedly so, in older compared with younger men in response to a variety of acute stressors (Esler et al. 1995b) (bottom panel, Fig. 4). As was the case at rest, lower neuronal reuptake of NA appeared to contribute significantly to the greater increases in cardiac PNA spillover rates in response to acute stress with age (Esler et al. 1995b).
We also measured the magnitude of increase in adrenaline secretion from the adrenal medulla in response to several types of stress using isotope dilution methodology (Esler et al. 1995a). Unlike the SNS responses, the absolute stress-evoked increases in adrenaline secretion were markedly attenuated with age (top panel, Fig. 4). Specifically, the augmentation in adrenaline release in response to stress in the older men was only 33–44 % of that observed in the young controls.
In summary, in striking contrast to the prevailing view, the results of our systematic investigations overwhelmingly support the concept that primary human ageing is not associated with exaggerated sympathoadrenal responsiveness to acute stress. The increase in cardiac PNA spillover with stress does appear to be augmented in older adults, but this may be due largely to faulty neuronal uptake of NA. Importantly, the ability of the adrenal medulla to secrete adrenaline in response to stress is markedly impaired even in healthy older adults.
In conclusion, experimental data from our laboratories and others support the view that chronic (basal) SNS activity increases with advancing age in healthy adult humans (Fig. 3). The elevations in SNS activity appear to be region specific, targeting skeletal muscle and the gut, but not obviously the kidney. The SNS tone of the heart is increased, although this appears to be due at least in part to reduced neuronal NA reuptake. In contrast to SNS activity, basal adrenaline secretion from the adrenal medulla is markedly reduced with primary ageing in humans. This is not reflected in plasma adrenaline concentrations because of reduced plasma clearance.
The mechanisms underlying these age-associated changes in sympathoadrenal function have not been established. Our preliminary results suggest that the increase in basal peripheral SNS activity with age is associated with elevated forebrain noradrenergic activity. These data are consistent with the hypothesis that increased CNS sympathetic drive may be a key mechanism involved. In contrast to studies in experimental animals, currently there is little or no evidence in humans to support a role for reduced baroreflex inhibition in the increases in SNS activity with age. The mechanism(s) underlying blunted adrenaline secretion from the adrenal medulla remain to be investigated.
Finally, despite widely held beliefs to the contrary, the results of our systematic investigations demonstrate that sympathoadrenal responsiveness to acute stress is not exaggerated with age, at least in healthy adults. Indeed, as observed under resting conditions, adrenaline release in response to acute stress is substantially attenuated in older men.
The work from the laboratory of Professor Seals cited in this review was supported by National Institutes of Health award AG06537. Professor Seals wishes to acknowledge the important contributions of trainees and/or colleagues Robin Callister, Kevin Davy, Frank Dinenno, Pamela Parker Jones, Mary Beth Monroe, Alexander Ng, Mary Jo Reiling, Hirofumi Tanaka and J. Andrew Taylor. The work from the laboratory of Professor Esler cited in this review was supported in part by an institute grant from the National Health and Medical Research Council of Australia to the Baker Medical Research Institute, and in part by National Institutes of Health award AG06537 (USA). Professor Esler wishes to acknowledge the important contributions of his colleagues David Kaye, Gavin Lambert, Gary Jennings, Mario Vaz, Claudia Ferrier and Graeme Eisenhofer.