David S. Goldstein

Catecholamine Metabolism: A Contemporary View with Implications for Physiology and Medicine

This article provides an update about catecholamine metabolism, with emphasis on correcting common misconceptions relevant to catecholamine systems in health and disease. Importantly, most metabolism of catecholamines takes place within the same cells where the amines are synthesized. This mainly occurs secondary to leakage of catecholamines from vesicular stores into the cytoplasm. These stores exist in a highly dynamic equilibrium, with passive outward leakage counterbalanced by inward active transport controlled by vesicular monoamine transporters. In catecholaminergic neurons, the presence of monoamine oxidase leads to formation of reactive catecholaldehydes. Production of these toxic aldehydes depends on the dynamics of vesicular-axoplasmic monoamine exchange and enzyme-catalyzed conversion to nontoxic acids or alcohols. In sympathetic nerves, the aldehyde produced from norepinephrine is converted to 3,4-dihydroxyphenylglycol, not 3,4-dihydroxymandelic acid. Subsequent extraneuronal O-methylation consequently leads to production of 3-methoxy-4-hydroxyphenylglycol, not vanillylmandelic acid. Vanillylmandelic acid is instead formed in the liver by oxidation of 3-methoxy-4-hydroxyphenylglycol catalyzed by alcohol and aldehyde dehydrogenases. Compared to intraneuronal deamination, extraneuronal O-methylation of norepinephrine and epinephrine to metanephrines represent minor pathways of metabolism. The single largest source of metanephrines is the adrenal medulla. Similarly, pheochromocytoma tumor cells produce large amounts of metanephrines from catecholamines leaking from stores. Thus, these metabolites are particularly useful for detecting pheochromocytomas. The large contribution of intraneuronal deamination to catecholamine turnover, and dependence of this on the vesicular-axoplasmic monoamine exchange process, helps explain how synthesis, release, metabolism, turnover, and stores of catecholamines are regulated in a coordinated fashion during stress and in disease states.

Takotsubo Cardiomyopathy A New Form of Acute, Reversible Heart Failure

Several relatively recent case reports and series have described a condition featuring symptoms and signs of acute myocardial infarction without demonstrable coronary artery stenosis or spasm in which the heart takes on the appearance of a Japanese octopus fishing pot called a takotsubo (Figure 1). In takotsubo cardiomyopathy (also called transient apical ballooning and stress cardiomyopathy), left ventricular dysfunction, which can be remarkably depressed, recovers within a few weeks.1–4 Figure 1 Left ventriculogram (A, end-diastolic phase; B, end-systolic phase) in the right anterior oblique projection. The extensive area around the apex shows akinesis, and the basal segments display hypercontraction, especially in the end-diastolic phase. C, ... Takotsubo cardiomyopathy occurs predominantly in post-menopausal women soon after exposure to sudden, unexpected emotional or physical stress. For instance, the incidence of takotsubo cardiomyopathy increased substantially in elderly women living near the epicenter of the Niigata earthquake.4 Although the left ventricular dysfunction is transient and there is no evidence of obstructive epicardial coronary disease, an increasing number of angioplasty procedures have been performed for presumed acute coronary syndromes. Concepts about the demographics, clinical features, prognosis, and management of this reversible form of left ventricular failure are still evolving. In this brief review, we summarize recent clinical reports and discuss an animal model that may clarify the pathogenesis of this condition.

Recent advances in genetics, diagnosis, localization, and treatment of pheochromocytoma.

Dr. Karel Pacak (Pediatric and Reproductive Endocrinology Branch, National Institute of Child Health and Human Development [NICHD] and Clinical Neurocardiology Section, National Institute of Neurological Disorders and Stroke [NINDS], National Institutes of Health [NIH], Bethesda, Maryland): Pheochromocytomas are chromaffin cell tumors that, although rare, must be considered in patients with hypertension, autonomic disturbances, panic attacks, adrenal incidentalomas, or familial diseases featuring a predisposition to develop pheochromocytoma. Hypertension, whether sustained or paroxysmal, is the most common clinical sign, and headache, excessive truncal sweating, and palpitations are the most common symptoms (1). Pallor is also common, whereas flushing occurs less frequently. Some patients present with severe episodes of anxiety, nervousness, or panic. Patients with a familial predisposition or small incidentally discovered adrenal masses can be normotensive and asymptomatic. The low pretest prevalence of pheochromocytomaclose to 0.5% among those tested because of hypertension and suggestive symptoms (1) and as high as 4% in patients with adrenal incidentaloma (2)together with imperfect sensitivity and specificity of commonly used biochemical and imaging tests, can make diagnosis and localization of pheochromocytoma difficult. Effective methods for diagnosis and localization are important because seemingly mild stimuli can provoke the tumor to release large amounts of catecholamines, with severe or fatal consequences. Moreover, surgical removal can cure pheochromocytoma in up to 90% of cases, whereas if left untreated the tumor can prove fatal. Advances in genetic mutation analysis have greatly improved identification of patients with familial pheochromocytoma, allowing detection of tumors at an early stage, often before typical signs and symptoms occur. These advances provide new insights into the biology and natural history of the disease and highlight inadequacies of commonly used diagnostic tests. In turn, new developments have been made in the management of patients with familial pheochromocytoma and in surgical procedures for preserving normal adrenal cortical tissue in persons with bilateral adrenal tumors. In this paper, we summarize advances in the genetics, biochemical diagnosis, localization, and management of pheochromocytoma and also note key questions that remain unanswered. Molecular Genetic Abnormalities Associated with Pheochromocytoma Drs. W. Marston Linehan and McClellan M. Walther (Urologic Oncology Branch, National Cancer Institute [NCI], NIH, Bethesda, Maryland): Pheochromocytomas may be classified as sporadic or familial. Most pheochromocytomas are sporadic. Familial predisposition is seen mainly in patients with multiple endocrine neoplasia type II (MEN II), von HippelLindau disease, neurofibromatosis type 1, and familial carotid body tumors (Table 1). The exact molecular mechanisms by which the hereditary mutations predispose to tumor development remain unknown. Hereditary forms of pheochromocytoma can differ in rate of tumor growth, malignant potential, and catecholamine phenotype. Table 1. Hereditary Forms of Pheochromocytoma Cancer Genes Identification of a cancer gene can help us understand the origin of cancer, such as pheochromocytoma, and elucidate mechanisms of tumor formation and behavior. Moreover, identification of a disease gene provides a method for genetic diagnosis. Phenotypic manifestations of a hereditary cancer syndrome can vary markedly, and genetic tests can confirm the diagnosis when the clinical presentation is complex. Finally, understanding of cancer genes may provide targets for therapy. The two most studied types of cancer genes are tumor suppressor genes (Figure 1) and oncogenes (7). When mutated, a proto-oncogene becomes activated, resulting in an oncogene. This is referred to as a single hit; that is, the proto-oncogene undergoes a single activating mutation that turns it into an oncogene (8, 9). Familial predisposition to pheochromocytoma in patients with MEN II results from such a mechanism. In contrast, a tumor suppressor gene is a loss-of-function gene, in which inactivation of both copies of the gene causes unregulated cell growth and division. This loss of function can result from mutation of one allele of a tumor suppressor gene and deletion of the second copy (10). Examples of tumor suppressor genes are the retinoblastoma gene, the Wilms tumor gene, the tuberous sclerosis genes, and, in the case of pheochromocytoma, the von HippelLindau gene (11-19). Figure 1. The Knudson two-hit model. Pheochromocytoma in Multiple Endocrine Neoplasia Type II: RETGene Multiple endocrine neoplasia type IIA is characterized clinically by the familial association of medullary thyroid cancer, pheochromocytoma, and parathyroid hyperplasia. Mucosal ganglioneuromas are also found in some patients (MEN IIB). Pheochromocytoma in MEN II is associated with germline mutation of the proto-oncogene RET. This proto-oncogene becomes an oncogene when an activating mutation occurs (20-25). The activating mutation in the RETgene drives the abnormal cellular proliferation that leads to adrenal medullary hyperplasia and pheochromocytoma. Several RETgermline mutations are associated with the development of pheochromocytoma, with some variation dependent on the particular mutation (3-5, 26, 27) (Table 1). Pheochromocytoma in von HippelLindau Disease: the von HippelLindau Gene Patients with von HippelLindau disease have a germline mutation of the von HippelLindau gene (28). Affected persons can develop early-onset bilateral kidney tumors and cysts, pheochromocytomas, cerebellar and spinal hemangioblastomas, retinal angiomas, pancreatic cysts and tumors, epididymal cystadenomas, and tumors in the endolymphatic sac canal of the inner ear (29-31). von HippelLindau disease has marked phenotypic heterogeneity. While patients from some families present with central neural, eye, kidney, and pancreatic tumors, patients in other families present mainly with pheochromocytoma (30, 32, 33). Some reports have described families thought to have familial pheochromocytoma who proved to have von HippelLindau disease (32, 34-37). Missense mutations in the von HippelLindau gene are associated with the development of pheochromocytoma more than twice as often as are other types of mutations (74% vs. 32%) (6, 33). Molecular Genetic Diagnosis von HippelLindau disease and MEN II have a similar prevalence (approximately 1 in 30 000 to 1 in 45 500). Mutations predisposing to pheochromocytoma have greater penetrance in MEN II than in von HippelLindau disease (38, 39). Pheochromocytoma in von HippelLindau families has been reported as familial pheochromocytoma or MEN II (40, 41). Because different kindreds can present with different phenotypes, it can be difficult to distinguish between von HippelLindau disease and MEN II in some patients with familial pheochromocytoma. Patients with bilateral adrenal, recurrent, or multifocal pheochromocytoma should undergo clinical or genetic testing for mutations of the von HippelLindau or RETgenes. The availability of germline testing for both von HippelLindau (42) and RET (15, 20, 23, 40, 43) gene mutations (at OncorMed in Gaithersburg, Maryland, and at the University of Pennsylvania in Philadelphia) has improved the clinical management of patients with hereditary pheochromocytoma. When a patient presents with a family history in which the primary manifestation is pheochromocytoma, the von HippelLindau gene is a likely cause. Some von HippelLindau families present mainly with pheochromocytoma and occult or delayed manifestations in the central nervous system, eye, or other organs. It is less likely that a member of a MEN II family will present predominantly with pheochromocytoma because most of these patients have medullary thyroid carcinoma (44). A small number of families with familial pheochromocytoma have neither von HippelLindau nor RETgermline mutations, and the genetic basis for this is currently being studied. Biochemical Diagnosis of Pheochromocytoma Dr. Graeme Eisenhofer (Clinical Neurocardiology Section, NINDS, NIH, Bethesda, Maryland): Diagnosis of pheochromocytoma usually requires biochemical evidence of excessive catecholamine production by the tumor, usually achieved from measurements of catecholamines or catecholamine metabolites in urine or plasma. These biochemical approaches, however, have several limitations. Since catecholamines are normally produced by sympathetic nerves and by the adrenal medulla, high catecholamine levels are not specific to pheochromocytoma and may accompany other conditions or disease states. In addition, sometimes pheochromocytomas do not secrete enough catecholamines to produce positive test results or typical signs and symptoms. In addition, pheochromocytomas often secrete catecholamines episodically. Between episodes, levels of catecholamines may be normal. Thus, commonly used tests of plasma or urinary catecholamines and metabolites and other biochemical tests, such as measurements of plasma chromogranin A levels, do not always reliably exclude or confirm a tumor (45-55). A recently developed biochemical test, involving measurements of plasma levels of free metanephrines (o-methylated metabolites of catecholamines), circumvents many of the above problems and offers a more effective means to diagnose pheochromocytoma than other tests (46, 56). Sensitivity of Biochemical Tests Measurements of plasma levels of normetanephrine and metanephrine have higher sensitivity than other biochemical tests for diagnosis of both sporadic and familial pheochromocytoma (46, 56). In familial pheochromocytoma, periodic screening can lead to early-stage detection before symptoms and signs, when tumors are small and are not secreting large amounts of catecholamines (6). The difficulty of biochemical diagnosis of familial pheochromocytoma is illustrated

II. Validity and reliability of liquid chromatography with electrochemical detection for measuring plasma levels of norepinephrine and epinephrine in man

Abstract Liquid chromatography with electrochemical detection (LCEC) provides a rapid, sensitive, and specific technique for measuring human plasma norepinephrine (NE) and epinephrine (E) levels. We tested the reliability and validity of this technique against that of the catechol-O-methyl-transferase radioenzymatic (COMT-RE) assay. In healthy, resting humans, mean NE and E values were similar using the LCEC and COMT-RE techniques (311 vs. 300 pg/ml for NE; 57 vs. 52 pg/ml for E). In a series of 25 plasma samples obtained from a variety of sources, the correlation between the two methods was 0.99 for both NE and E. Coefficients of variation were similar for catecholamine levels above 100 pg/ml, but below this, the COMT-RE technique appeared to be more reliable. The advantages of the LCEC method are its speed, simplicity of sample preparation, low cost per assay, lack of use of radionuclides, and ease in trouble-shooting. The COMT-RE technique is preferable for small sample sizes or large numbers of samples. LCEC offers a reasonable alternative to the COMT-RE technique for measuring plasma norepineprhine and epinephrine.

Cardiac Sympathetic Nerve Function in Congestive Heart Failure

BACKGROUND Increased availability of norepinephrine (NE) for activation of cardiac adrenoceptors (increased cardiac adrenergic drive) and depletion of myocardial NE stores may contribute to the pathophysiology and progression of congestive heart failure. This study used a comprehensive neurochemical approach to examine the mechanisms responsible for these abnormalities. METHODS AND RESULTS Subjects with and without congestive heart failure received intravenous infusions of [(3)H]NE. Cardiac spillover, reuptake, vesicular-axoplasmic exchange, and tissue stores of NE were assessed from arterial and coronary venous plasma concentrations of endogenous and [(3)H]-labeled NE and dihydroxyphenylglycol. Tyrosine hydroxylase activity was assessed from plasma dopa, and NE turnover was assessed from measurements of NE metabolites. NE release and reuptake were both increased in the failing heart; however, the efficiency of NE reuptake was reduced such that cardiac spillover of NE was increased disproportionately more than neuronal release of NE. Cardiac NE stores were 47% lower and the rate of vesicular leakage of NE was 42% lower in the failing than in the normal heart. Cardiac spillover of dopa and NE turnover were increased similarly in congestive heart failure. CONCLUSIONS Increased neuronal release of NE and decreased efficiency of NE reuptake both contribute to increased cardiac adrenergic drive in congestive heart failure. Decreased vesicular leakage of NE, secondary to decreased myocardial stores of NE, limits the increase in cardiac NE turnover in CHF. Decreased NE store size in the failing heart appears to result not from insufficient tyrosine hydroxylation but from chronically increased NE turnover and reduced efficiency of NE reuptake and storage.

Stress-Induced Norepinephrine Release in the Hypothalamic Paraventricular Nucleus and Pituitary-Adrenocortical and Sympathoadrenal Activity: In Vivo Microdialysis Studies

The hypothalamic-pituitary-adrenocortical (HPA) axis and the autonomic nervous system are major effector systems that serve to maintain homeostasis during exposure to stressors. In the past decade, interest in neurochemical regulation and in pathways controlling activation of the HPA axis has focused on catecholamines, which are present in high concentrations in specific brain areas--especially in the hypothalamus. The work described in this review has concentrated on the application of in vivo microdialysis in rat brain regions such as the paraventricular nucleus (PVN) of the hypothalamus, the central nucleus of the amygdala (ACE), the bed nucleus of the stria terminalis (BNST), and the posterolateral hypothalamus in order to examine aspects of catecholaminergic function and relationships between altered catecholaminergic function and the HPA axis and sympathoadrenal system activation in stress. Exposure of animals to immobilization (IMMO) markedly and rapidly increases rates of synthesis, release, and metabolism of norepinephrine (NE) in all the brain areas mentioned above and supports previous suggestions that in the PVN NE stimulates release of corticotropin-releasing hormone (CRH). The role of NE in the ACE and the BNST and most other areas possessing noradrenergic innervation remains unclear. Studies involving lower brainstem hemisections show that noradrenergic terminals in the PVN are derived mainly from medullary catecholaminergic groups rather than from the locus ceruleus, which is the main source of NE in the brain. Moreover, the medullary catecholaminergic groups contribute substantially to IMMO-induced noradrenergic activation in the PVN. Data obtained from adrenalectomized rats, with or without glucocorticoid replacement, and from hypercortisolemic rats suggest that glucocorticoids feedback to inhibit CRH release in the PVN, via attenuation of noradrenergic activation. Results from rats exposed to different stressors have indicated substantial differences among stressors in eliciting PVN noradrenergic responses as well as of responses of the HPA, sympathoneural, and adrenomedullary systems. Finally, involvement of other areas that participate in the regulation of the HPA axis such as the ACE, the BNST, and the hippocampus and the importance of stress-induced changes in expression of immediate early genes such as c-fos are discussed.

Malignant Pheochromocytoma: Effective Treatment with a Combination of Cyclophosphamide, Vincristine, and Dacarbazine

STUDY OBJECTIVE To determine the efficacy and toxicity of combination chemotherapy in patients with advanced, malignant pheochromocytoma. DESIGN Nonrandomized, single-arm trial. SETTING Governmental medical referral center. PATIENTS Fourteen patients with malignant pheochromocytoma confirmed by histologic tests. All patients had metastatic disease and elevated urinary catecholamine secretion. INTERVENTIONS After optimization of antihypertensive therapy, patients received cyclophosphamide, 750 mg/m2 body surface area on day 1; vincristine, 1.4 mg/m2 on day 1, and dacarbazine, 600 mg/m2 on days 1 and 2, every 21 days. MEASUREMENTS AND MAIN RESULTS Combination chemotherapy with cyclophosphamide, vincristine, and dacarbazine produced a complete and partial response rate of 57% (median duration, 21 months; range, 7 to more than 34). Complete and partial biochemical responses were seen in 79% of patients (median duration, more than 22 months; range, 6 to more than 35). All responding patients had objective improvement in performance status and blood pressure. Toxicity included expected hematologic, neurologic, and gastrointestinal effects of chemotherapy without serious sequelae. There were four minor hypotensive episodes and one minor hypertensive episode. CONCLUSIONS Combination chemotherapy with cyclophosphamide, vincristine, and dacarbazine is effective for advanced malignant pheochromocytoma. Urinary catecholamines are useful to ascertain biochemical response to therapy.

Relationship between plasma norepinephrine and sympathetic neural activity.

For circulating norepinephrine (NE) to reflect sympathetic activity validly, plasma NE should show an intensity-dependent increase during sympathetic stimulation and decrease during sympathetic inhibition, and circulating NE should correlate with more directly obtained measures of sympathetic activity. Review of published evidence indicates that NE in peripheral plasma satisfies these criteria. However, models used to explain the relationship between circulating NE and sympathetic activity must take into account processes intervening between the synaptic cleft and free NE in the circulation and, since sympathetic outflow is regionalized, the contributions of specific vascular beds to circulating NE. In this report a model is presented where removal processes for NE are viewed as acting in series to produce a gradient in NE concentrations from synapse to plasma, and where the relative contributions of specific vascular beds are calculated from the arteriovenous difference in plasma NE across those beds and the percentage of cardiac output distributed to them. In general, venous plasma NE provides a useful estimation of average sympathetic outflow.

Low‐frequency power of heart rate variability is not a measure of cardiac sympathetic tone but may be a measure of modulation of cardiac autonomic outflows by baroreflexes

Power spectral analysis of heart rate variability has often been used to assess cardiac autonomic function; however, the relationship of low‐frequency (LF) power of heart rate variability to cardiac sympathetic tone has been unclear. With or without adjustment for high‐frequency (HF) power, total power or respiration, LF power seems to provide an index not of cardiac sympathetic tone but of baroreflex function. Manipulations and drugs that change LF power or LF:HF may do so not by affecting cardiac autonomic outflows directly but by affecting modulation of those outflows by baroreflexes.

Cardiac Sympathetic Denervation in Parkinson Disease

Orthostatic hypotension is common in Parkinson disease (1). Although earlier studies implicated L-dopa treatment as the cause (2), more recent studies have shown that orthostatic hypotension may result from deficient cardiovascular reflexes that depend on release of the sympathetic neurotransmitter norepinephrine in the heart and blood vessels (3-5). We call this phenomenon sympathetic neurocirculatory failure. Several recent studies have reported decreased myocardial concentrations of radioactivity after injection of the sympathoneural imaging agent 123I-metaiodobenzylguanidine (123I-MIBG) in patients with Parkinson disease (6-13). This finding is consistent with but does not prove cardiac sympathetic denervation. In addition, studies have not specifically considered the possible association between cardiac sympathetic denervation and sympathetic neurocirculatory failure in Parkinson disease. Measures of autonomic function have included blood pressure during tilt-table testing (abnormalities of which can have several causes), heart rate responses to the Valsalva maneuver (which are determined mainly by changes in parasympathetic cholinergic outflow to the heart), or skin conductance or sweating responses (which are determined mainly by alterations in sympathetic cholinergic outflow to the skin). These measures may not allow assessment of sympathetic noradrenergic function. One way to detect sympathetic neurocirculatory failure in a patient with orthostatic hypotension is by analyzing beat-to-beat blood pressure associated with performance of the Valsalva maneuver (Figure 1). In patients with sympathetic neurocirculatory failure, blood pressure decreases progressively during phase II of the maneuver, whereas normally blood pressure plateaus or increases at the end of phase II (phase II-L). In patients with sympathetic neurocirculatory failure, phase IV blood pressure increases slowly back to baseline after release of the maneuver, whereas normally blood pressure overshoots. These abnormalities are a direct result of deficient cardiovascular reflexes that depend on sympathetically mediated release of norepinephrine. In our study, we defined sympathetic neurocirculatory failure as chronic, reproducible orthostatic hypotension associated with abnormal blood pressure responses in both phase II-L and phase IV of the Valsalva maneuver. Figure 1. Heart rate and blood pressure responses to the Valsalva maneuver in a control patient with a history of neurocardiogenic syncope ( left ) and a patient with Parkinson disease and orthostatic hypotension ( right ). Previous studies also have not independently confirmed that a low myocardial concentration of 123I-MIBGderived radioactivity actually reflects cardiac sympathetic denervation in Parkinson disease. Neurochemical findings indicating decreased norepinephrine release, neuronal uptake, turnover, and synthesis in the heart could provide such confirmation. In humans, 6-[18F]fluorodopamine can be used to visualize cardiac sympathetic innervation by positron emission tomographic (PET) scanning (14), which provides excellent spatial and temporal resolution. Since 6-[18F]fluorodopamine is a catecholamine handled in the heart in a manner similar to the way in which norepinephrine is handled (15), PET scanning may allow functional and anatomic assessments of sympathetic cardiac innervation (16). We used PET scanning after injection of 6-[18F]fluorodopamine and neurochemical measurements during cardiac catheterization to answer the following questions: 1) What proportions of patients with Parkinson disease, with or without sympathetic neurocirculatory failure, have decreased myocardial 6-[18F]fluorodopaminederived radioactivity? 2) Does decreased myocardial 6-[18F]fluorodopaminederived radioactivity in Parkinson disease actually reflect cardiac sympathetic denervation, as identified by indices of cardiac norepinephrine release, neuronal uptake, turnover, and synthesis? 3) Does the frequency of cardiac sympathetic denervation differ between groups of patients with Parkinson disease who have sympathetic neurocirculatory failure and those who do not? 4) Does cardiac sympathetic denervation also occur in patients with multiple-system atrophy, a progressive neurodegenerative disease of adults that features autonomic dysfunction and has parkinsonian, cerebellar, or mixed forms (17)? [The diagnosis of multiple-system atrophy is clinical and, except for a typically poor response to L-dopa treatment, can be difficult to distinguish from Parkinson disease.] 5) Is cardiac sympathetic denervation in patients with Parkinson disease related to L-dopa treatment or to disease duration or severity? Methods The Intramural Research Board of the National Institute of Neurological Disorders and Stroke approved the study protocol. All participants provided written informed consent. Participants We included patients with Parkinson disease or multiple-system atrophy who were studied at the National Institutes of Health Clinical Center in Bethesda, Maryland. Twenty-nine patients had Parkinson disease, including 10 who were not receiving or had never received L-dopa. Twenty-four patients had multiple-system atrophy, including 8 who were taking L-dopa at the time of evaluation. For comparison, we used 6-[18F]fluorodopamine PET scan data and, in most cases, cardiac neurochemical data from 7 patients with pure autonomic failure (5 men, 2 women [mean age SE, 60 6 years]) and 33 controls. Of these 33 controls, 22 had a history of neurocardiogenic syncope (4 men, 18 women [mean age, 35 3 years]) and 11 had a history of postural tachycardia syndrome (1 man, 10 women [mean age, 42 4 years]). 6-[18F]fluorodopamine PET scan data were also obtained from 19 normal volunteers. All patients with Parkinson disease were referred by neurologists or movement disorder clinics and fulfilled accepted clinical criteria (18). Parkinson disease was staged by using the HoehnYahr classification. All affected patients had bradykinesia, cogwheel rigidity, and one or more additional parkinsonian features (pill-roll tremor, stooped posture, festinating gait, difficulty initiating movement, masklike face, micrographia, or marked improvement in motor function during treatment with L-dopa). Patients with multiple-system atrophy had at least two parkinsonian features but were not classified in terms of cerebellar, parkinsonian, or mixed subtypes (17). All had gradually progressive parasympathetic failure (manifested by impotence in men, urinary retention or incontinence, constipation, or constant pulse rate) and had one or more additional features of multiple-system atrophy (heat or cold intolerance and decreased sweating, intention tremor or other evidence of cerebellar dysfunction, slurred speech or a history of aspiration, or no or only slight improvement during an adequate trial of L-dopa treatment). Sympathetic neurocirculatory failure was defined as reproducible, chronic orthostatic hypotension (decrease in diastolic pressure of at least 10 mm Hg and in systolic pressure of at least 20 mm of Hg after 3 to 5 minutes of standing), coupled with abnormal responses of beat-to-beat blood pressure associated with the Valsalva maneuver (19). As noted previously, patients with sympathetic neurocirculatory failure usually exhibit a progressive decrease in blood pressure in phase II-L of the maneuver and an absence of a pressure overshoot in phase IV after release of the maneuver. Valsalva Maneuver For the Valsalva maneuver, the patient lay supine with his or her head on a pillow and blew into a plastic or rubber tube connected to a sphygmomanometer, keeping a pressure of 30 mm Hg for 10 to 12 seconds. The response of beat-to-beat blood pressure during phase II-L of the Valsalva maneuver was considered to be normal if the diastolic and mean arterial pressure increased before the end of the straining and abnormal if they decreased. The response during phase IV was considered to be normal if the systolic blood pressure increased progressively to a value exceeding the baseline (measured just before the patient inhaled and then began straining) and abnormal if the systolic pressure did not exceed the baseline. Sympathetic Neuroimaging Patients were positioned in a GE Advance scanner (General Electric, Milwaukee, Wisconsin), with their thoraxes in the gantry. 6-[18F]fluorodopamine (specific activity, 7.4 to 37 MBq/mmol; dose in most cases, 0.037 MBq) was dissolved in approximately 10 mL of normal saline and infused intravenously at a constant rate for 3 minutes. Thoracic PET scanning was performed for up to 3 hours. The tomographic data were divided into intervals of 5 to 30 minutes. Data acquisition was not gated to the electrocardiogram. In most patients, PET scanning was also used to delineate the left ventricular myocardium and assess myocardial perfusion after administration of the perfusion imaging agent 13 N-ammonia. Intravenously injected 13 N-ammonia exits the bloodstream rapidly and enters cells nonspecifically. A few minutes after the injection, the concentration of 13 N-ammoniaderived radioactivity in the left ventricular myocardium exceeds that in the left ventricular chamber, enabling visualization of the myocardium. Myocardial tissue concentrations of 13 N-ammoniaderived radioactivity depend on local perfusion (20). Neurochemical Testing Patients underwent right-heart catheterization for measurements of norepinephrine spillover into coronary sinus plasma and of venousarterial differences in plasma levels of dihydroxyphenylglycol (DHPG) and L-dopa. After placement of arm and right internal jugular venous catheters (the latter advanced into the coronary sinus), a tracer amount of [3H]norepinephrine (levo- [2, 5, 6] [3H]norepinephrine, New England Nuclear, Boston, Massachusetts) was infused intravenously. Coronary sinus blood flow was measured by thermodilution, and arterial and great cardiac venous or coronary sinus blood was sampled