Cardiac 18F-Dopamine Positron Emission Tomography Predicts the Type of Phenoconversion of Pure Autonomic Failure

Background Pure autonomic failure (PAF) is a rare disease characterized clinically by neurogenic orthostatic hypotension (nOH) and biochemically by peripheral noradrenergic deficiency. Clinically diagnosed PAF can evolve (“phenoconvert”) to a central Lewy body disease (LBD, e.g., Parkinson’s disease (PD) or dementia with Lewy bodies (DLB)) or to the non-LBD synucleinopathy multiple system atrophy (MSA). We examined whether cardiac 18F-dopamine positron emission tomography (PET) predicts the trajectory of phenoconversion in PAF. Since cardiac 18F-dopamine-derived radioactivity always is decreased in LBDs with nOH and usually is normal in MSA, we hypothesized that PAF patients with low cardiac 18F-dopamine-derived radioactivity may phenoconvert to a central LBD but do not phenoconvert to MSA. Methods We reviewed data from all the patients seen at the National Institutes of Health Clinical Center from 1994 to 2023 with a clinical diagnosis of PAF and data about serial 18F-dopamine PET. Results Twenty patients met the above criteria. Of 15 with low cardiac 18F-dopamine-derived radioactivity, 6 (40%) phenoconverted to PD or DLB and none to MSA. Of 5 patients with consistently normal 18F-dopamine PET, 4 phenoconverted to MSA, and the other at autopsy had neither a central LBD nor MSA. Conclusion In this case series, 40% of patients with nOH and low cardiac 18F-dopamine-derived radioactivity phenoconverted to PD or DLB during follow-up; none phenoconverted to MSA. Cardiac 18F-DA PET therefore can predict the type of phenoconversion in PAF. This capability could refine eligibility criteria for entry into disease-modification trials aiming to prevent evolution of PAF to symptomatic central LBDs.


INTRODUCTION
Pure autonomic failure (PAF), previously called idiopathic orthostatic hypotension [49], asympathicotonic orthostatic hypotension [39], and Bradbury-Eggleston syndrome [1], is a rare form of chronic autonomic failure identi ed by orthostatic hypotension (OH) without a known secondary cause and without clinical evidence of motor or cognitive impairment attributable to central nervous system neurodegeneration [34].
In the general population, Parkinson's disease (PD), the most common neurodegenerative movement disorder, and dementia with Lewy bodies (DLB), the second most common cause of dementia, occur far more frequently than does PAF. Prospective [35] and retrospective [3] longitudinal studies have found, however, that PAF diagnosed based on clinical criteria can evolve ("phenoconvert") to these central Lewy body diseases (LBDs) or to the non-LBD synucleinopathy multiple system atrophy (MSA).
In a multi-center prospective cohort study of 74 patients with clinically diagnosed PAF, 25 (34%) phenoconverted to PD, DLB, or MSA by 4 years of follow-up [35]. A similar rate of phenoconversion (32%) was reported in an Italian cohort of 50 patients with idiopathic autonomic failure after a median of 7 years [14]. A retrospective study of 275 PAF patients evaluated at the Mayo Clinic from 2001 to 2011 reported phenoconversion in 24% to a synucleinopathy with motor or cognitive involvement [3].
Two histopathologically distinct forms of synucleinopathy have been described. LBDs such as PD and DLB entail increased deposition of the protein alpha-synuclein (aS) in neurons and nerve bers (Lewy bodies and Lewy neurites) [45,46], whereas MSA involves aS deposition in glial cytoplasmic inclusions [47]. These differences based on post-mortem analyses seem to have in vivo parallels, in that in LBDs there is increased deposition of native or phosphorylated aS in sympathetic noradrenergic nerves in skin biopsies [7,31], while in MSA there can be increased deposition of aS in non-neuronal Schwann cells [10].
The LB forms of synucleinopathy (PD and DLB) differ from the non-LB form (MSA) in terms of genetics, symptomatic manifestations, progression, management, prognosis, and neuropathology [2,6,23,32,38,47,48]. Accurate prediction of the future phenotype therefore has important implications for clinical diagnosis and treatment and for research. Although retrospective studies have reported an association of certain features such as rapid eye movement (REM) sleep behavior disorder (RBD), bladder symptoms, stridor, and higher supine plasma norepinephrine levels with phenoconversion to MSA [14], these features may not be speci c enough to predict phenoconversion at an individual level.
In a retrospective longitudinal study from the Mayo Clinic, patients who phenoconverted to MSA were reported to have had high levels of neuro lament light chain (NfL) in the cerebrospinal uid (CSF) [43]. In a recent prospective study, protein misfolding cyclic ampli cation (PMCA) detected aS oligomers in the CSF of PAF patients, establishing that PAF entails central neural synucleinopathy [44]. In that study both increased levels of NfL and magnitudes of aS PMCA were associated with phenoconversion of clinically diagnosed PAF to MSA over about 2.5 years of follow-up. In the above-mentioned multi-center study [35] phenoconversion from clinically diagnosed PAF to MSA occurred during the rst 2 years of follow-up, whereas most of the phenoconversion to central LBDs occurred later.
Importantly, the longitudinal studies reported above involved PAF diagnosed without regard to the occurrence of sympathetic noradrenergic de ciency, which is an established hallmark of the disease [22,49]. Using cardiac 18 F-dopamine positron emission tomography (PET), a validated index of myocardial norepinephrine content [36], our group consistently has found that patients with LBDs have profound myocardial noradrenergic de ciency, whereas those with MSA usually have normal cardiac noradrenergic Page 4/19 innervation [16,37]. In a recent cross-sectional study, cardiac 18 F-dopamine PET distinguished PD + OH from the parkinsonian form of MSA with impressive sensitivity (92%) and speci city of (96%) [37]. As cardiac 18 F-dopamine PET has been a highly effective imaging modality for separating LBD forms of nOH from MSA in our retrospective and cross-sectional studies [18,31], we aimed to investigate the utility of 18 F-dopamine PET in the longitudinal assessment of PAF patients, speci cally to explore if this testing modality could predict the trajectory of phenoconversion.

METHODS
Data were reviewed from participants in protocols approved by the Institutional Review Board (IRB) of the National Institute of Neurological Disorders and Stroke (NINDS) Clinical Protocols 94N0186, 03N0004, and 18N0140). Written informed consent was obtained from all subjects prior to any research procedures. Data were analyzed in accordance with National Institutes of Health (NIH) Clinical Protocol 000490, which is a secondary research protocol approved by the NIH IRB.
We analyzed clinical and laboratory data from patients classi ed as PAF from 1994 to 2023 at the NIH Clinical Center. Data were evaluated about clinical, laboratory, skin biopsy, or post-mortem ndings from patients who had undergone serial evaluations over years with cardiac sympathetic neuroimaging by 18 Fdopamine PET. In most cases putamen dopaminergic neuroimaging by 18 F-DOPA PET was also done [15].
The presence or absence of neurogenic OH (nOH) was determined based on beat-to-beat blood pressure responses to the Valsalva maneuver and orthostatic fractional increments in plasma norepinephrine levels [22,27]. PAF was diagnosed based on the occurrence of nOH without an identi ed secondary cause and unassociated with clinical evidence of a movement disorder or cognitive dysfunction [13]. Data from patients with nOH in the setting of diabetes [12], amyloidosis [40], multiple myeloma [24], autoimmune autonomic ganglionopathy [17], spinal cord injury [42], or autoimmunity-associated autonomic failure with sympathetic denervation [30] were excluded, since such patients had known or suspected secondary causes. We de ned "phenoconversion" as the evolution of clinical features amounting to clinical diagnoses of PD, DLB, or MSA. 18 F-Dopamine and 18 F-DOPA PET were conducted as reported previously [15]. The lower limit of normal for myocardial 18 F-dopamine-derived radioactivity is 6,000 nCi-kg/cc-mCi [19], and the lower limit of normal for the putamen/occipital cortex (PUT/OCC) ratio of 18 F-DOPA-derived radioactivity is 2.7 [20].
The aS-tyrosine hydroxylase (TH) colocalization index in skin biopsies, a validated, quantitative measure of aS deposition in catecholaminergic nerve bers, was calculated according to previously published methodology [31]. The upper limit of normal is 1.5. Other measures of aS deposition were the aS-protein gene product 9.5 ratio in arrector pili muscle [29] and the intensity of aS signal in arrector pili muscle. For the latter measure the upper limit of normal is 2.0 [31].

Statistics
The frequencies of PAF patients evolving to a central LBD with low vs. normal 18 F-dopamine-derived radioactivity were compared by Fisher's exact test. A p value less than 0.05 de ned statistical signi cance.

RESULTS
Since 1994 we identi ed 41 patients who had nOH, did not have any of the exclusion criteria, and had available 18 F-dopamine PET data (Fig. 1). Of the 41, 20 had clinical follow-up or death data, 15 with low 18 F-dopamine-derived radioactivity and 5 with normal radioactivity. Among the 15 patients with low 18 Fdopamine-derived radioactivity, 6 (40%) phenoconverted during follow-up, all 6 to a central LBD. Among the 5 patients with normal 18 F-dopamine-derived radioactivity, 4 (80%) phenoconverted during follow-up, all 4 to MSA. Therefore, of the 20 patients with clinical follow-up or death data, 10 (50%) phenoconverted to a synucleinopathy with motor or cognitive involvement. The frequency of PAF patients with low radioactivity who phenoconverted to a central LBD (6 of 6 (100%)) differed signi cantly from the frequency of patients with normal radioactivity who phenoconverted to a central LBD (0 of 4 (0%); p = 0.0048).
A patient with normal 18 F-dopamine-derived radioactivity retained a PAF phenotype until his sudden, unexpected death at 61 years old. The patient did not have post-mortem histopathological ndings of either a LBD or MSA [30]. Case 3: PAF to DLB + OH (Fig. A) This woman reported constipation and dream enactment behavior for as long as she could remember. When rst evaluated at the NIH Clinical Center at 76 years old she reported orthostatic intolerance and lack of sense of smell. She had nOH, decreased 18 F-dopamine-derived radioactivity diffusely in the left ventricular myocardium, normal putamen 18 F-DOPA-derived radioactivity, normal CSF DOPA and 3,4dihydroxyphenylacetic acid (DOPAC) levels [26], and increased αS signal in arrector pili muscle in a skin biopsy [31]; she was diagnosed with the LB form of PAF. About 1-2 years later she reported decreased cognitive function, and at 77 years old her MoCA score was 22. Follow-up PET neuroimaging showed low cardiac 18 F-dopamine-derived radioactivity, while the PUT/OCC ratio of 18 F-DOPA was normal. As of this report she is alive at 89 years old but is institutionalized with advanced DLB. Case 4: PAF to PD + OH (Fig. 3B) This man reported dream enactment behavior and ejaculatory dysfunction since he was 39 years old. About 2 years later he developed erectile dysfunction, which persisted. About this time he had an episode of micturition syncope. Upon initial evaluation at the NIH Clinical Center at 48 years old he had decreased 18 F-dopamine-derived radioactivity in the left ventricular free wall, but interventricular septal radioactivity was normal. Subsequent 18 F-dopamine PET scans at 50, 52, and 59 years old demonstrated decreased radioactivity throughout the left ventricular myocardium. Skin biopsy at 56 years old showed an increased αS-TH colocalization index. His initial 18 F-DOPA PET scan at 53 years old revealed a decreased PUT/OCC ratio of 18 F-DOPA-derived radioactivity (2.31), but subsequent scans at 55, 57, 61, and 63 years old disclosed normal PUT/OCC ratios. After 19 years of follow up, at age 72 years, he developed parkinsonian symptoms, had a UPDRS score of 61, and had a low PUT/OCC ratio at 2.2. He is alive with a diagnosis of PD + OH.
Case 5: PAF to PD + OH (Fig. A) This woman developed orthostatic intolerance, palpitations, and impaired sense of smell at 35 years old.
She was diagnosed with PAF at the Mayo Clinic at 51 years old. She subsequently developed dream enactment behavior, and RBD was diagnosed when she was 52 years old. She did not have any motor symptoms or cognitive dysfunction until she was 60 years old. At age 61 years she developed a resting tremor of the hands, micrographia, and hypomimia. Her cardiac 18 F-dopamine PET in 2021 (age 60 years) revealed diffusely decreased radioactivity in the left ventricular myocardium. 18 F-DOPA PET in 2022 (age 61 years) showed a decreased PUT/OCC ratio of 2.2 [15]. As of this report she has not developed visual hallucinations or cognitive dysfunction. Based on these data she is deemed to have phenoconverted to PD + OH.
Case 6: PAF to DLB + OH (Fig. B) This man reported decreased sense of smell for many years. His father had been diagnosed with PD at 55 years old and had died at 72 years old. The patient lost consciousness in about 2009-2010 at an outdoor church service when he got up to go to the bathroom. When rst seen at the NIH in 2011 at 52 years old he did not have OH. Subsequently he developed episodic OH that became persistent and consistent. Dream enactment was rst noted when he was about 56 years old. He developed constipation and in 2020 required disimpaction in an emergency room. In 2021, at 62 years old, he had to retire due to cognitive decline. He lost substantial weight because of decreased appetite. In 2023 he reported the onset of episodic visual hallucinations, brain fog, decreased memory, whispery voice, sloppy handwriting Case 7: PAF to MSA (Fig. 5) This woman had dream enactment behavior (RBD diagnosed by polysomnography), constipation, urinary incontinence, heat intolerance, and chronic orthostatic, post-prandial, and post-exertional dizziness. Her father had had PD. Upon initial evaluation at 56 years old she had nOH, normal olfaction (UPSIT score 40), a decreased PUT/OCC ratio of 18 F-DOPA-derived radioactivity at 2.49, and normal cardiac 18 Fdopamine-derived radioactivity. Subsequent 18 F-dopamine PET and UPSIT scores were consistently normal. At 61 years old she had a UPDRS score of 26 and a low PUT/OCC ratio at 2.33. Based on neurological examination at that time by a Board-certi ed neurologist specializing in movement disorders, she was diagnosed with PD. She died 3 years later. According to the obituary and the Certi cate of Death from the Vital Records Department of her county of residence, she had died of MSA. There was no autopsy.

DISCUSSION
In this study we reviewed longitudinal follow-up data from PAF patients who had undergone comprehensive clinical laboratory testing including cardiac 18 F-dopamine PET and during follow-up phenoconverted to a symptomatic central synucleinopathy. Six patients phenoconverted to a central LBD; all 6 had low 18 F-dopamine-derived radioactivity initially or during follow-up. Four patients phenoconverted to MSA; all 4 had persistently normal 18 F-dopamine-derived radioactivity. Previous crosssectional studies have shown that among patients with nOH those with central LBDs have low 18 Fdopamine-derived radioactivity, whereas those with MSA usually have normal radioactivity [16,18,25,37]. Here we report longitudinal data supporting the view that in PAF low 18 F-dopamine-derived radioactivity predicts phenoconversion to a central LBD and not to MSA.
All the PAF patients who phenoconverted to a central LBD also had evidence for increased aS deposition in sympathetic noradrenergic nerves in skin biopsies, another aspect that separates central LBDs from MSA [8-10]. Whether elevated aS in sympathetic nerves predicts the pattern of phenoconversion of PAF remains unknown.
Based on previous reports, the type of phenoconversion (to a central LBD or to MSA) depends importantly on the duration of follow-up. The longer PAF patients are followed, the greater the prevalence of phenoconversion to a central LBD [35]. Since the disease process in MSA progresses quickly [23], the latency of phenoconversion from PAF to MSA is relatively short. Rapid progression of MSA can explain the relatively small number of patients in the present study who phenoconverted to MSA, since such patients would be unlikely to return for follow-up testing.
Similar to PAF, patients with isolated RBD are known to be at risk for phenoconversion to a central synucleinopathy [4]. A recently published study used cardiac 123 I-metaiodobenzylguanidine ( 123 I-MIBG) scintigraphy, which is analogous to cardiac 18 F-dopamine PET, for predicting phenoconversion of patients with isolated RBD [41]. In that study RBD patients who phenoconverted to a central LBD had signi cantly lower heart-to-mediastinum ratios of 123 I-MIBG-derived radioactivity than did other RBD patients.
In the present study there was a wide range of latency from the diagnosis of PAF to the diagnosis of a central LBD, from about 3 years to about 19 years of follow-up. A key task for the future is to develop means to identify when the disease process is transitioning from a peripheral to a central LBD. One possibly informative biomarker is 18 F-DOPA PET, since patients with PAF have normal PUT/OCC ratios of 18 F-DOPA-derived radioactivity, whereas patients with PD have low PUT/OCC ratios [21]. PAF also entails normal CSF levels of DOPAC, the main neuronal metabolite of dopamine, whereas in PD and MSA CSF DOPAC is subnormal [26]. Under the presumption of an ascending spatio-temporal process within the brainstem with earlier involvement of the pontine locus ceruleus than of the midbrain substantia nigra [5,28], neuromelanin magnetic resonance imaging might also help identify PAF evolving to a central LBD [11]. As of this writing, no studies have reported on whether assays of NfL or aS seeding identify the timing of phenoconversion from PAF to a central LBD vs. MSA. Future studies should also attempt to identify factors that might distinguish phenoconversion to PD from phenoconversion to DLB [33].
Although 40% of our patients with nOH and low 18 F-dopamine-derived radioactivity phenoconverted to a central LBD, 60% were lost to follow-up or are still being followed. The 40% gure therefore is likely to be an underestimate.

Limitations
The number of patients reported here was small. PAF and MSA are rare disorders, and following such patients longitudinally using sophisticated objective measures over years requires substantial time, effort, and expense. Nevertheless, the clinical and laboratory ndings of the patients described here support the utility of cardiac 18 F-dopamine PET for predicting the phenoconversion trajectory of PAF. 18 F-Dopamine PET is available only at the NIH Clinical Center, which makes it di cult to replicate the ndings. 123 I-MIBG scintigraphy is widely available but is not FDA approved for distinguishing LB from non-LB forms of nOH. No study to date has compared 18 F-dopamine vs. 123 I-MIBG imaging in the same patients.
In the 4 PAF patients who phenoconverted to probable MSA there was no post-mortem neuropathologic con rmation of the diagnosis.

Implications
The best way to gain insights into the natural course of PAF is through longitudinal studies. We illustrate in this longitudinal study that cardiac 18 F-dopamine PET is an effective modality for predicting the trajectory of phenoconversion of PAF. This capability could facilitate enrollment into appropriate diseasemodifying clinical trials.