Article Text

Download PDFPDF

Abnormal baroreceptor-mediated vasopressin release as possible marker in early diagnosis of multiple system atrophy
  1. K Deguchi1,
  2. I Sasaki1,
  3. T Touge1,
  4. M Tsukaguchi1,
  5. H Takeuchi2,
  6. S Kuriyama1
  1. 1Third Department of Internal Medicine, Kagawa Medical University, 1750-1 Ikenobe, Miki-cho, Kita-gun, Kagawa 761-0793, Japan
  2. 2Department of Health Sciences, Kagawa Medical University, 1750-1 Ikenobe, Miki-cho, Kita-gun, Kagawa 761-0793, Japan
  1. Correspondence to:
 Prof. S Kuriyama
 Third Department of Internal Medicine, Kagawa Medical University, 1750-1 Ikenobe, Miki-cho, Kita-gun, Kagawa 761-0793, Japan;


Background: Although autonomic failure (AF) is a critical symptom of multiple system atrophy (MSA), it may not appear until late in the disease process.

Objectives: To clarify whether a detailed investigation of the autonomic nervous system in patients with MSA without overt AF demonstrates latent lesions of central cardiovascular control circuits and facilitates the early diagnosis of MSA.

Methods: Autonomic function tests, and plasma noradrenaline (NA) and vasopressin (AVP) responses to head-up tilt (HUT), were studied in 12 patients with MSA with AF (probable MSA), 12 with MSA without overt AF (possible MSA), and 24 controls.

Results: Abnormalities of cardiovascular autonomic function tests were prominent in the first group but mild in the second. Plasma NA and AVP increments upon HUT differed significantly among all three groups.

Conclusions: These results indicate that probable MSA involves diffuse degeneration of central cardiovascular control circuits. On the other hand, the discrepancies in possible MSA suggest a vulnerability of the noradrenergic (A1) neurones of the caudal ventrolateral medulla that are involved in AVP secretion. This finding also suggests that AVP increment may be useful as a diagnostic tool in the early stages of MSA.

  • Autonomic failure
  • head up tilt
  • multiple system atrophy
  • noradrenaline
  • vasopressin
  • AVP, vasopressin
  • MSA, multiple system atrophy
  • NA, noradrenaline
  • AF, autonomic failure
  • HUT, head up tilt
  • OH, orthostatic hypotension
  • PAF, pure autonomic failure
  • PD, Parkinson’s disease
  • IPD, idiopathic Parkinson’s disease
  • GH, growth hormone
  • PRL, prolactin
  • EPO, erythropoietin
  • SPN, sympathetic preganglionic neurones
  • IML, intermediolateral
  • BP, blood pressure
  • DA, dopamine
  • Ad, adrenaline
  • NMS, neurally mediated syncope
  • VLM, ventrolateral medulla
  • SND, striatonigral degeneration

Statistics from

Multiple system atrophy (MSA) is a neurodegenerative disorder characterised by varying combinations of autonomic failure (AF) / urinary dysfunction, parkinsonism, cerebellar ataxia, and corticospinal dysfunction. The diagnostic criteria for MSA, which were proposed in a recent consensus statement, comprise three categories reflecting differing levels of certainty: definite, probable, and possible.1 A diagnosis of definite MSA requires neuropathological confirmation demonstrating glial cytoplasmic inclusions in association with degenerative changes in the nigrostriatal and olivopontocerebellar pathways.2 However, a diagnosis of probable or possible MSA can be made using purely clinical evidence.

In this study, the degree of systemic autonomic dysfunction in possible MSA was investigated using cardiovascular autonomic function tests, comparing subjects with participants with probable MSA and with healthy controls. In addition, the impairment of the central cardiovascular control circuits was assessed and compared in accordance with the baseline values and responses of plasma noradrenaline (NA) and vasopressin (AVP) upon head up tilt (HUT).


Impairment of the autonomic nervous system, with manifestations such as orthostatic hypotension and urological dysfunction, is critical for a diagnosis of probable MSA, but may not appear until late in the disease process. In this case, an accurate early diagnosis is often difficult, and long term follow up may be required to demonstrate the features of multisystem degeneration, including impairment of the autonomic nervous system.3 In fact, a previous report indicates that approximately one quarter of patients presenting with cerebellar ataxia (sporadic olivopontocerebellar atrophy) evolve into MSA with overt AF (probable MSA) within five years.4 In these circumstances, possible MSA, in which only one system is involved with certainty, may be a tentative diagnostic category and may lead to probable MSA.

The practical diagnosis of MSA is based on a number of non-invasive investigations, as well as on clinical evidence. Autonomic screening tests are important in defining the cardiovascular autonomic deficit that is a cardinal feature of MSA.5 Basal levels of NA69 and cardiac nuclear medical techniques (123I-MIBG myocardial scintigraphy1011 and 6-18F fluorodopamine PET12), reflecting postganglionic sympathetic innervation, help differentiate MSA from pure autonomic failure (PAF) and AF with Parkinson’s disease (AF-PD). Differences in cardiac uptake are also useful for differentiating MSA from idiopathic Parkinson’s disease (IPD) without clinical evidence of AF. Neuroimaging studies (MRI1314, MR spectroscopy,15 and PET16) can distinguish MSA from IPD, predominantly by differences in the basal ganglia. Several neuroendocrine approaches have been investigated for mapping central or peripheral autonomic mechanisms. The growth hormone (GH) response to clonidine1718 or apomorphine,19 and basal plasma prolactin (PRL) levels,20 can differentiate MSA from other similar disorders. Measurement of plasma AVP upon HUT is a clear marker that may be used to differentiate between MSA and PAF.21 Serum erythropoietin (EPO) levels demonstrate a significant difference between MSA and IPD.22 However, it remains unclear whether or not these tests are useful for obtaining an accurate early diagnosis of MSA.

The final outputs of the central cardiovascular control circuits are mediated by sympathetic preganglionic neurones (SPNs) of the intermediolateral column cell (IML), cardiovagal neurones of the nucleus ambiguus, and AVP secreting magnocellular neurones of the hypothalamus.23 In contrast to what is known about probable MSA, the areas involved in possible MSA remain uncertain, except for a moderate degeneration of IMLs.24 Assuming that possible MSA may be a transitional stage evolving into probable MSA, a detailed comparison of these stages using neuroendocrine approaches may provide important clues in the early diagnosis of MSA and can help account for the development of lesions in the central autonomic network.


AVP release is triggered either by a rise in plasma osmolarity or by a reduction in blood volume or venous return, leading to reduced low pressure receptor stretching in the atrial and pulmonary vascular circuits.25 Therefore, an assessment of the AVP response to a reduction in central venous pressure and plasma volume upon standing can be a useful test of the integrity of the afferent and central components of the baroregulatory reflex arc.26 In addition, combined with the measurement of plasma NA as an index of postganglionic sympathetic function, the assessment of the AVP response upon HUT may allow for further localisation of an autonomic lesion.26



Twelve patients with probable MSA (mean age 63 (2) years, range 48–73), 12 with possible MSA (64 (2) 51–77) and 24 age matched healthy controls (63 (3) 34–82) participated in this study. The mean duration of illness in the patients with probable MSA (3 years (1) 1–10) and possible MSA (3 (1) 1–15) was equivalent. Probable or possible MSA was diagnosed according to the consensus statement on the diagnosis of MSA.1 The criteria for a diagnosis of probable MSA included autonomic failure / urinary dysfunction together with either cerebellar dysfunction or parkinsonism with poor response to levodopa. For a diagnosis of possible MSA, a criterion had to be met in one of the three domains (autonomic failure / urinary dysfunction, parkinsonism, or cerebellar dysfunction), and two features from other domains also had to be observed. Although the triplet repeats the expansion of SCA 1, 2, 3, 6, and dentatorubral-pallidoluysian atrophy was investigated in these patients, the results were all negative.

The probable MSA group consisted of eight participants with MSA-C (cerebellar form), two with MSA-P (parkinsonian form), and two with MSA-M (mixed form). The possible MSA group was composed of four subjects with a combination of cerebellar ataxia and parkinsonism without AF, seven with cerebellar ataxia and features of AF, and one with parkinsonism and features of AF. None of the patients was taking levodopa or other anti-parkinsonian medication during the autonomic function tests; however, three patients with parkinsonian features had previously received levodopa, with a suboptimal response. Although two patients with probable MSA were being treated with dihydroxyphenylserine for postural hypotension, this drug was stopped 72 hours before the autonomic function tests. None of the healthy controls had a medical disorder or was receiving treatment that could affect the autonomic nervous system.

Informed consent was obtained from all participants, and the protocol of the study was approved by the local ethics committee.

Cardiovascular autonomic function tests

Autonomic function tests were performed on fasting subjects in the morning in a specially isolated room with stable humidity and temperature conditions. Blood pressure (BP) and pulse rate were recorded using the continuous BP monitoring system CBM 7000 (Nihon Colin, Tokyo, Japan). A cannula with heparinised saline solution was inserted in the antecubital vein for blood sampling, and the subjects were placed on a tilt table for 20 minutes, followed by 60° HUT for 10 minutes, and standing for five minutes. Responses to isometric exercise, cutaneous cold, mental arithmetic, deep breathing, and the Valsalva manoeuvre were measured as previously described.5

Measurement of plasma catecholamines and AVP

Blood samples from probable MSA patients, possible MSA patients, and controls were withdrawn from a cannulated antecubital vein before and 10 minutes after HUT. The samples were centrifuged at 4°C for 20 minutes, and the separated plasma was immediately frozen and stored at −60°C until tested. Plasma adrenaline (Ad), NA, and dopamine (DA) were measured by high performance liquid chromatography using an electrochemical detector, and plasma AVP was measured by a highly specific and sensitive radioimmunoassay. The intra- and interassay coefficients of variance were, respectively, 2.5% and 3.5% for Ad, 0.7% and 2.2% for NA, 0.5% and 1.8% for DA, and 1.7% and 3.9% for AVP.

Data analysis

Data were expressed as mean (SE). The significance of differences was determined by one way analysis of variance (ANOVA) followed by Fisher’s PLSD for multiple comparisons. A p value <0.05 was considered statistically significant. Since AVP secretion is accelerated by BP reduction,2127 analysis of covariance was employed to evaluate the independent secretion of AVP in response to HUT by adjusting the effect of BP change. The post hoc comparison was performed by Fisher’s PLSD. All statistical analyses were carried out using Stat View 5.0 J software on a Macintosh computer.


Cardiovascular autonomic function tests

The mean resting supine BP levels were similar in all groups (controls, 123/70 (4/2) mm Hg; possible MSA, 123/71 (5/3) mm Hg; probable MSA, 127/73 (6/3) mm Hg). There were no heart rate differences among the study groups (controls, 71 (3) beats/minute; possible MSA, 66 (4) beats/minute; probable MSA, 66 (2) beats/minute). Table 1 lists the results of the autonomic function tests.

Table 1

Results of cardiovascular autonomic function tests*

Postural change

For the postural challenge, ANOVA indicated significant differences between BP changes among the study groups (controls, possible MSA, probable MSA) (p<0.001 for systolic/diastolic BP changes during HUT or standing). Patients with probable MSA had a significantly greater fall in systolic and diastolic BP than did those with possible MSA (p<0.001) or control individuals (p<0.001). Participants with possible MSA had a relatively mild but significant reduction in systolic (p<0.01) and diastolic BP (HUT, p<0.05; standing, p<0.01) compared with the controls.

Pressor tests

For the pressor tests, ANOVA indicated significant differences between BP changes among the study groups (p<0.05 for diastolic BP change during isometric handgrip; p<0.001 for systolic and diastolic BP changes during mental arithmetic; p<0.05 for systolic BP change during cutaneous cold). The mean increases in BP during isometric handgrip, mental arithmetic and cutaneous cold were significantly lower in patients with probable MSA than in the controls (p<0.05 for diastolic BP change during handgrip; p<0.01 for systolic BP change during cutaneous cold; p<0.001 for systolic and diastolic BP changes during mental arithmetic). During mental arithmetic, patients with possible MSA showed a significantly different systolic BP change from the controls (p<0.05) and a significantly different diastolic BP change from the patients with probable MSA (p<0.05).

Respiratory stimuli

ANOVA indicated significant differences in heart rate variability among the study groups (p<0.05 for Δheart rate during deep breathing; p<0.001 for Valsalva ratio). The heart rate variability during deep breathing and the Valsalva ratio during and following a Valsalva manoeuvre were significantly lower in subjects with probable MSA than in control individuals (p<0.05 for Δheart rate; p<0.001 for Valsalva ratio). Participants with possible MSA had a significantly lower Δheart rate during deep breathing compared with the controls (p<0.05), and a significantly higher Valsalva ratio compared with the probable MSA patients (p<0.05).


Plasma catecholamine levels

Plasma catecholamine concentrations were assessed before and after HUT in the three study groups. A significant difference in plasma NA level was observed in baseline values (p<0.05) and in response to HUT (p<0.01) (table 2). Subjects with probable MSA had a significantly lower NA level compared with controls (baseline, p<0.01; HUT, p<0.001) or possible MSA patients (baseline, p<0.05; HUT, p<0.05). The plasma NA level at baseline or in response to HUT in possible MSA participants was equivalent to that of controls. Plasma Ad and DA levels at baseline, as well as those in response to HUT, did not reveal significant differences among the study groups.

Table 2

Plasma catecholamines* at rest and in response to HUT

Plasma AVP levels

Plasma AVP at supine rest was not significantly different among the study groups. After adjusting for the BP reduction during HUT using analysis of covariance, there were significant differences in AVP increment among the study groups (p<0.01) (table 3). The rise in AVP was significantly lower in subjects with probable MSA (p<0.05) and possible MSA (p<0.01), in comparison with the controls. The response of plasma AVP to BP changes is shown in fig 1, which relates the plasma concentration of AVP to the decline in BP during HUT. The regression line and 99% confidence limits of the data from all healthy controls are shown. Unlike the plasma NA in response to HUT, almost all participants with probable and possible MSA had an inappropriate AVP response to the stimulus of BP alteration.

Table 3

Plasma AVP* at rest and in response to HUT

Figure 1

Relationship between the plasma vasopressin (AVP) increment and the decline in mean blood pressure (MBP) during head up tilt. The regression line and 99% confidence limits of the data from healthy controls (n  =  24) are shown. The squares and circles represent probable and possible multiple system atrophy, respectively.


This study of subjects with probable MSA demonstrated prominent OH and blunted responses in cardiovascular autonomic reflexes, reflecting failure of the sympathetic and parasympathetic nervous systems. Eight out of 12 subjects with possible MSA satisfied a feature of OH (a fall in BP of >20 mm Hg systolic or >10 mm Hg diastolic within three minutes of standing from the recumbent position) as described in the consensus guidelines,1 so that the participants with possible MSA showed a mild but significant fall in both systolic and diastolic BP upon postural challenge in comparison with the healthy controls. All of the cardiovascular autonomic responses in the possible MSA patients showed intermediate values between the probable MSA patients and the healthy controls. The latent and subtle impairment shown in cardiovascular autonomic function tests is consistent with the hypothesis that possible MSA corresponds to a transitional stage that evolves into probable MSA.

It has previously been suggested that several hormonal substances might provide useful means of differentiating between the various forms of MSA and disorders with similar neurological features. A pharmacological challenge with clonidine, an alpha2 adrenoceptor agonist, demonstrated a blunted GH response in patients with MSA, but not in those with IPD, progressive supranuclear palsy (PSP), or PAF.171828 However, administration of the GH secretagogue, levodopa,17 or apomorphine,19 to MSA patients induced a rise in GH releasing hormone and also in GH. On the other hand, participants with MSA had reduced suppression of PRL secretion following levodopa administration, when compared with control subjects29; furthermore, patients with MSA showed higher base PRL levels when compared with both IPD patients and healthy controls.2029 Thus, these results may represent a selective dysfunction of the alpha2 adrenoceptor in the hypothalamus and of the tubero-infundibular dopaminergic in the hypothalamic dopaminergic pathways in patients with MSA.1729 In addition, the GH response to a low dose of apomorphine was significantly increased in patients with IPD but not in those with MSA, suggesting a hypersensitivity of the hypothalamic dopamine receptors in IPD.19 Unlike subjects with IPD without sympathetic AF, a proportion of the subjects with MSA with sympathetic AF demonstrated a deficiency of EPO, and therefore had normochromic anaemia.22 Although these neuroendocrine indices could provide positive evidence of probable MSA, and such indices could also be used to map lesions in the central or peripheral autonomic mechanisms of persons with probable MSA, their usefulness for the early diagnosis of MSA remains unclear.

It is known that individuals with MSA manifest a near normal baseline value and an inadequate increase in plasma NA upon standing. The former reflects a peripherally intact sympathetic nervous system.689 The latter is caused by the depletion of adrenergic C1 neurones in the rostral ventrolateral medulla (VLM),3031 which provide a major input to the SPNs and are critically involved in the maintenance of tonic sympathetic vasomotor tone and the integration of baroreceptor reflexes.32–34 However, this study of subjects with probable MSA showed a substantially low supine plasma NA level. Considering that similar data were obtained for persons with MSA-M,17 this finding may reflect an anterograde or transsynaptic degeneration of the peripheral sympathetic nervous system secondary to the depletion of SPNs in advanced stages of MSA.

The present study of possible MSA showed a nearly normal plasma NA increment and subtle BP reduction upon HUT. Although no quantitative assessment of the supraspinal system has been undertaken in cases of possible MSA, it has been suggested that retrograde degeneration secondary to SPN depletion may induce the depletion of C1 neurones of the rostral VLM.30 Since the SPN depletion in possible MSA appears to be less severe than in probable MSA,24 the residual supraspinal mechanisms could be effective in maintaining the baroreflex responses in cases of possible MSA. In addition, a compensatory mechanism, such as collateral sprouting of residual preganglionic and postganglionic nerves, may be effective in individuals with possible MSA.35

An increase in plasma osmolarity,7 dopaminergic,36 or cholinergic37 stimulation, and the occurrence of neurally mediated syncope (NMS)2138 induce a significant increase in the plasma concentration of AVP. However, except for the reduction in central blood volume caused by HUT, these factors are not likely to have contributed significantly in this study. First, no significant alterations of plasma osmolarity has been observed in MSA patients in an upright position.3941 Secondly, neither a marked increment in plasma DA resulting in dopaminergic stimulation, nor hypoglycaemia resulting in cholinergic stimulation, appeared in any of the subjects. Thirdly, none of the participants showed features of neurally mediated syncope (NMS)—for example, precipitating events (such as emotional distress),42 prodromal symptoms (such as nausea and vomiting),42 or marked increment of plasma Ad.38

In this study, a brisk rise in plasma AVP upon HUT occurred in the healthy controls but not in the participants with probable MSA, a finding that is consistent with published findings.21263943 Of the structures involved in AVP secretion, the hypothalamo-neurohypophyseal system responded normally to osmotic stimuli, even in subjects with probable MSA.3944 On the other hand, clonidine failed to suppress AVP release in subjects with probable MSA.45 These results could well be due to a marked depletion of noradrenergic (A1) neurones of the caudal VLM,3032 which activate the AVP secreting magnocellular hypothalamic neurones and may mediate, at least in part, the reflex AVP responses to haemodynamic stimuli.34 In addition, the possibility of secondary denervation of the hypothalamus, to which a significantly reduced number of A1 and C1 neurones project, has been pointed out.30

Whereas the participants with possible MSA had a nearly normal plasma NA increment upon HUT, the response of their plasma AVP was abolished, a finding that corresponds to the data in several cases in a previous report.39 Although a precise mechanism for the blunted response of AVP remains uncertain, it has been demonstrated that a marked depletion of A1 neurones in the VLM exists, even in persons with striatonigral degeneration (SND) without AF (possible MSA), as well as in persons with SND and AF (probable MSA).46 To clarify the mechanism for this blunted AVP response in cases of possible MSA, plasma AVP secretions should be investigated by various neuroendocrine approaches, such as hypertonic saline infusion for the hypothalamo-hypophyseal system,3944 clonidine administration for the noradrenergic pathways,45 metoclopramide47 or physostigmine37 administration for the cholinergic pathways, and levodopa43 or apomorphine36 administration for the dopaminergic pathways.

AF-PD and IPD with mild OH (IPD-OH) may frequently be difficult to differentiate from probable and possible MSA. Participants with AF-PD showed subnormal plasma NA levels and supersensitivity to NA infusion; in addition, HUT tests resulted in a significant increase in the plasma AVP levels, but not in the plasma NA levels.48 The features determined by the neuroendocrine approach suggest that AF-PD is similar to PAF but distinct from probable MSA. IPD-OH was associated with normal plasma NA increments and with blunted but significant plasma AVP increments upon HUT.49 One explanation for the differing AVP responses in participants with AF-PD or IPD-OH and those with probable or possible MSA may be that there is a marked depletion of A1 neurones in persons with MSA, but not in persons with PD.50 Under these circumstances, the blunted AVP secretion upon HUT may provide a valuable clue for the early diagnosis of MSA.

One possible limitation of this study was that the HUT might have been insufficient for AVP secretion, because the plasma AVP response is less sensitive to postural challenges of short duration or in cases involving mild OH. However, it seems unlikely that the AVP response in the subjects with probable and possible MSA was affected by those factors. Although the rise in plasma AVP upon HUT in the healthy controls was lower than in a previous report,21 the plasma AVP levels showed a significant increase 10 minutes after HUT (p<0.001; paired samples t-test). In addition, the participants with possible MSA, as well as those with probable MSA, showed a significant reduction in BP compared with the healthy controls. If the structures involved in AVP secretion were intact, a greater degree of plasma AVP release would be expected in cases of probable and possible MSA than in healthy controls.

In summary, possible MSA showed a nearly normal response of plasma NA and a subtle BP reduction upon HUT. This finding suggests that individuals with possible MSA have a substantial degree of residual function or compensatory mechanism in the C1 neurones in the rostral VLM, the descending catecholaminergic fibre projection to the IML, or SPNs in the IML. In contrast, the plasma AVP response upon HUT was abolished in persons with possible MSA as well as in participants with probable MSA. On the assumption that possible MSA represents an early stage of the process of MSA, we suspect that impairment of the A1 neurones in the afferent pathway involved in AVP secretion may predominantly occur in the early stages of MSA. Since a blunted AVP response upon HUT may be a useful index for the early diagnosis of MSA, further analyses should be performed on a large sample of subjrcts with possible MSA, and responses should be compared with those of patients with related disorders such as IPD, PSP, and cerebellar degeneration.


Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.