Cardiac sympathetic burden reflects Parkinson disease burden, regardless of high or low orthostatic blood pressure changes

Reduced uptake of 123I-meta-iodobenzylguanidine (123I-MIBG) and orthostatic hypotension (OH) are independently associated with worse clinical outcomes of Parkinson’s disease (PD). However, their interactive influence on PD has not been studied. The role of 123I-MIBG myocardial uptake, as a biomarker of PD severity, was investigated, conditional on the mediating effects of OH. A total of 227 PD patients were enrolled. Their motor and nonmotor aspects were assessed with standardized tools. Global disease burden was estimated by averaging the scaled z-scores of the assessment tools. Every patient went through 123I-MIBG scan, and OH was evaluated with the head-up tilt-test. The mediating role of orthostatic blood pressure changes (ΔBP) on the association between cardiac sympathetic denervation and disease burden was investigated. Low heart-to-mediastinum (H/M) ratio with less than 1.78 was seen in 69.6% of the patient population, and 22.9% of patients had OH. Low H/M ratio was associated with OH, and these patients had worse disease burden than subjects with normal 123I-MIBG uptake (global composite z-score: normal 123I-MIBG vs. abnormal 123I-MIBG; −0.3 ± 0.5 vs. 0.1 ± 0.7; p < 0.001). The mediation models, controlled for age and disease duration, revealed that the delayed H/M ratio and global composite score were negatively associated, irrespective of orthostatic ΔBP. Adverse relationship between cardiac sympathetic denervation and disease burden was shown without any interference from orthostatic blood pressure fluctuations. This result suggested that extracranial cardiac markers might reflect disease burden, regardless of labile blood pressure influence.

Cardiac postganglionic sympathetic denervation, reflected more accurately in the decreased delayed heart-tomediastinum (H/M) ratio upon 123 I-meta-iodobenzylguanidine ( 123 I-MIBG) myocardial scintigraphy scan 11 , contributes to orthostatic hypotension (OH) 4 . Reduced myocardial uptake of 123 I-MIBG indicates norepinephrine transporter dysfunction in sympathetic neurons, and may reflect physiological consequences in residual functional cardiac sympathetic axons 7,12 . Reduced uptake of 123 I-MIBG was utilized to discriminate PD from vascular Parkinsonism, drug-induced Parkinsonism, and atypical parkinsonian syndromes (APS) 2 .
Some PD patients have a normal 123 I-MIBG uptake analogous to APS, and this imaging phenotype was previously given the term "scans without evidence of cardiac norepinephrine deficit (SWEND)" 13 . PD patients with SWEND had mild Hoehn and Yahr (H&Y) stage, short disease duration, slow progression of motor dysfunction, a lower incidence of the wearing-off phenomenon, and a lower prevalence of nonmotor symptoms 7,[10][11][12]14,15 .
Most studies have investigated the association of cardiac sympathetic denervation with orthostatic hypotension and the impact on clinical outcomes separately, but the interacting effects of these two biomarkers on PD patient symptoms have seldom been explored 4,16 . As a pathophysiologic contributor, OH is a possible biomarker of pathologic burdens 2,4,7,12,17 , thus the confounding effects of OH need to be considered when assessing the role of 123 I-MIBG myocardial scintigraphy results in PD patients.
In this study, we investigated whether a mediating role for OH could be disproved in the context of a significant association between cardiac sympathetic denervation and clinical disease burden. The refutation of this role would emphasize the sustained value of 123 I-MIBG uptakes as a pathologic biomarker of disease severity despite its contribution to OH.
Orthostatic hypertension (OHT) is also a type of cardiovascular dysregulation which can be anticipated in PD patients 18,19 . This study also aimed to document the nature of OHT in early PD patients.
The mediating effects of orthostatic blood pressure changes on clinical outcomes Table 2  The results were similar when early H/M ratio (a') and washout rate (a") were set as predictors in the mediation analyses (Supplementary Table 1, Supplementary Fig. 1).
The subdomains of global composite score, partitioned by motor, sleep, and autonomic composite scores were analyzed (Table 3). A delayed H/M ratio negatively predicted each domain (total effect), and its inverse influence was not mediated by orthostatic ΔBP (indirect effect).

Bidirectionality of orthostatic blood pressure change
In an observational sub-analysis of orthostatic ΔBP max , OHT group displayed a worsening trend for global composite score as the ΔBP max decreased while the non-OHT group demonstrated a positive association as the ΔBP max increased, especially ΔSBP max ( Supplementary Fig. 2). These associations lacked any statistical significance; however, orthostatic BP rise and drop was accompanied by trends toward adverse disease severity in both directions.

DISCUSSION
Cardiac sympathetic denervation and orthostatic hypotension were encountered in early PD patients. OH was found more frequently in the abnormal 123 I-MIBG group while orthostatic hypertension was more common in SWEND PD patients. PD patients with reduced 123 I-MIBG uptake presented with worse clinical features. Cardiac denervation inversely predicted PDrelated disease burden, irrespective of orthostatic blood pressure changes. The population of this cohort was in the early mild PD disease stage. The prevalence of abnormal 123 I-MIBG uptake was comparable to previous studies 7,12,20 . The frequency of OH in early PD patients was similar to another study 21 , and its association with impaired cardiac sympathetic integrity has been reported 4,22 . In this cohort, PD patients with decreased 123 I-MIBG uptake showed more severe clinical presentations, and they were more susceptible to orthostatic BP drop 7,23 . PD patients with SWEND were associated with orthostatic BP rise, but had less disease burden.
Irrespective of age and disease duration, more severe cardiac sympathetic denervation was related to a larger drop in orthostatic BP (indirect component a → b) and worse disease severity (direct path a → c). Its negative influence was not carried through BP instability to affect PD patients (indirect path a → b → c). Thus, the total effect of sympathetic denervation negatively reflected the global disease burden, and was unaffected by BP lability; more severely impaired cardiac dysautonomia was directly associated with worse clinical outcomes ( Table 2). Cardiac denervation inversely paralleled motor, sleep, and autonomic global severity domains in the subcomponent analyses ( Table 3).
The inverse association between cardiac sympathetic denervation and worse disease severity could be explained by the degree of accumulated pathology. In a cross-sectional study, PD patients with cardiac sympathetic denervation were speculated to bear more pathologic burden that resulted in lack of compensatory reserve compared to patients with less cardiac sympathetic denervation, thus worse clinical manifestations 7 . This phenotype was further researched in a longitudinal study, and these patients were shown to be at a higher risk of developing motor complications 12 . This temporal association between PD with abnormal 123 I-MIBG uptake and wearing-off phenomenon implied that pathologic burden at the periphery could mirror central pathophysiologic disease progression as cardiac sympathetic denervation progressed over time 2,4,10,12 . This centripetal degeneration of cardiac sympathetic response was suggested to represent common degenerative process in PD patients 1 , and suggests multiple origination sites for Lewy pathology spread 24 .
The lack of contribution of blood pressure to clinical burdens amid the paths could suggest myocardial 123 I-MIBG scintigraphy is a potential biomarker that reflects Lewy body pathology 2,17 , in parallel with disease severity. The significant negative association (indirect component a → b), controlled by age and disease duration, confirmed the contributing action of cardiac sympathetic denervation on OH 4 .
Braak's schema was disputed because its model failed to explain all the subtypes of PD, especially cardiac denervation in early PD 24 . Recent hypothesis of "body-first" subtype PD incorporates the cardiac pathobiology which may be the answer to the observed differences between normal and abnormal 123 I-MIBG groups in this study 25,26 . This further enhances the extracranial biomarker role of cardiac denervation in PD.
Early H/M ratio and washout rate which reflect density of the presynaptic cardiac sympathetic nerve endings 2 , its tone 27 , and damaged or failing myocardium 28 , respectively, also manifested similar results that reinforced the hypothesis that 123 I-MIBG   Table 1, Supplementary Fig. 1).
OHT has seldom been investigated in PD patients, and its characteristics are largely unknown. In this study, despite a difference in estimation of ΔBP during head tilt, the prevalence of OH and OHT were comparable. Interestingly, although the analysis did not gain statistical significance, orthostatic BP incremental change was suggested to be related to worse clinical outcomes. Orthostatic BP drop also displayed a trend for worse outcomes, particularly with respect to ΔSBP. OHT is associated with other types of blood pressure lability, and with cardiovascular risk 18 . In accord with our observation, disrupted blood pressure control could be deleterious, regardless of the directionality of the BP changes 19 .
OHT was significantly associated with normal 123 I-MIBG that included higher values of delayed H/M ratios. It also displayed lower and negative washout rate, which apparently signified that delayed H/M ratio was greater than early H/M ratio. In this regard, delayed H/M and WR could denote the same as WR contains the other in its equation. This was in correspondence to other studies indicating that delayed H/M and washout rate represented sympathetic tone 2,27 . Its greater sympathetic tone might have allowed normal 123 I-MIBG group to culminate into higher orthostatic BP rise, particularly SBP, than abnormal 123 I-MIBG group.
This result suggested that OHT could originate by a different cardiovascular dysregulation mechanism since it was related to normal 123 I-MIBG. As OH in PD stems from cardiac, extra-cardiac denervation and arterial baroreflex failure 4 , orthostatic ΔBP rise could be the product of compensation of extant cardiac sympathetic tone amid other dysfunctions, which, when lost, could lead to OH. This may imply that OHT could be a prelude to OH in some PD. Our data did not depict any significant difference of washout rate between non-OHT and OHT among preserved 123 I-MIBG group. Had this been manifested, the sick-but-not-dead phenomenon could have played a role in its pathophysiology; 29 thus, further support the sequential conversion of OHT into OH as the disease progresses. The nature of OHT in PD needs to be further assessed.
There are several limitations in this study. First, because the enrolled PD patients were in the early phase of the disease, we could not completely exclude the possibility that some had atypical Parkinsonism. In particular, the PD with SWEND patient group might have inadvertently included some multiple system atrophy patients 2 . We attempted to reduce selection bias by using strict clinical diagnostic criteria for PD 30,31 and structural neuroimaging. Nevertheless, it can be difficult to differentiate PD from multiple system atrophy from a nosological perspective. Second, we also could not exclude patients who had genetic Parkinsonism. Patients with the Parkin mutation might not present with nonmotor manifestations and abnormal 123 I-MIBG uptake 32,33 . This study did not include familial PD or patients with young-onset PD (≤40 years). Third, direct biomarkers that provide objective measurements to assess PD disease severity are not available at the present; in addition, currently useful biomarkers do not always represent clinical symptoms. A clinical assessment-based approach, as in this study, is an alternative that evaluates disease severity, relevant to clinical practice. This study incorporated validated clinical tools that encompassed comprehensive motor and nonmotor disease domains. Fourth, there was not sufficient evidence that 123 I-MIBG uptake represents myocardial synuclein deposition. Future studies that correlate 123 I-MIBG myocardial scans with a direct synuclein accumulation biomarker are warranted. Moreover, cardiac imaging with 18 F-fluorodopa, which may be a better substance to evaluate cardiac sympathetic innervation, was also investigated in PD 16,34 . Diverse studies with cross-validation between different imaging modalities will strongly enhance the results of this research. Finally, because early patients with mild PD stage were enrolled, the scores of questionnaires tended to center at the left of the scale with relatively large standard deviation. A larger recruitment of PD by engagement of multi-clinics to reduce the bias is warranted.
In this cohort, cardiac sympathetic denervation negatively predicted disease severity, independently of orthostatic blood pressure changes. Even though cardiac denervation is a contributor to orthostatic hypotension, the effect of cardiac denervation on disease burdens was not mediated by orthostatic blood pressure changes. Cardiac sympathetic denervation was suggested to represent the disease burden, in parallel with pathologic severity. Orthostatic hypertension was also observed, and it could potentially contribute to PD symptom clinical severity. Two hundreds twenty-seven de novo and drug-naïve PD patients between April 2014 and January 2020 were enrolled. PD was diagnosed based on the UK Parkinson's Disease Society Brain Bank 30 , and its diagnosis was supported by positron emission tomography imaging studies using 18 F-N-(3-fluoropropyl)-2beta-carbon ethoxy-3beta-(4-iodophenyl) nortropane 31 . All patients had decreased dopamine transporter uptake in the striatum, mainly in the posterior putamen. Patients underwent brain magnetic resonance imaging (MRI) to exclude secondary causes for this finding.

METHODS Patients
Demographics such as age, sex, body mass index, disease duration, smoking status and history of hypertension, diabetes mellitus, and dyslipidemia were investigated. Disease severity was investigated with the UPDRS and H&Y stage. Global cognition was assessed by the MMSE and CDR.
Patients were disqualified from the study if they had any of the following indications: (1) any symptoms or signs of atypical and/or secondary Parkinsonism, (2) positive family history of Parkinsonism by pedigree analysis, which included first degree relatives, (3) documentation of atrial fibrillation during the head up tilt test, (4) history of diabetic neuropathy, (5) history of symptomatic stroke that could affect general cognition and performance, and (6) history of medications such as tricyclic antidepressants or benzodiazepines that influence autonomic functions or patients who were taking medications at the time of diagnosis known to influence the central dopaminergic, noradrenergic, and/or serotonergic systems.
Patients were followed every 2-6 months for a minimum of 12 months from the time they began taking dopaminergic medication, and their diagnosis was reaffirmed after at least 12 months of follow-up by two neurologists (S.-W.Y., J.-S.K.).

Questionnaires
The patients were evaluated with the following questionnaires: (1) NMSS 35 , (2) MADRS 36 , (3) ESS 37 , (4) PDSS-2 38 , (5) RBDSQ 39 , (6) OHQ 40 , and (7) the SCOPA-AUT 41 . Part I and II of the OHQ were summed separately. Because the SCOPA-AUT sexual dysfunction subsection had too many missing values, this subsection was omitted from that questionnaire's summation. Missing values of sexual subsection was due to patients' reluctance to reveal their sexual activities to the examiner. The sums of each questionnaire were analyzed.

Composite scores
Motor severity (motor composite score) was gauged by averaging the z-scores of the scaled UPDRS II and UPDRS III. The sums of each individual questionnaire for nonmotor features were standardized to z-scores. Nonmotor features were also divided into the sleep domain (sleep composite score: average z-scores for the ESS, PDSS-2, RBDSQ), and the autonomic domain (autonomic composite score: average z-scores for the OHQ and SCOPA-AUT), and were further analyzed separately. The affection domain was not analyzed because the MADRS data were not normally distributed, and the NMSS was excluded from subcomponent analysis due to its inclusive psychometric properties.
Parkinson's disease overall burden (the global composite score) was estimated by averaging the scaled z-scores of all motor and nonmotor assessments. Higher z-scores indicated worse severity. All questionnaires were evaluated by investigators blind to patient clinical information.

Head-up tilt test
All patients were tested in the full resting state. Continuous electrocardiograph leads and non-invasive BP monitoring equipment were applied to the patients (YM6000, Mediana Tech, Redmond, WA, USA). A supine position was maintained for 20 min during recording of BPs and heart rates every 5 min, before tilting to 60 degrees (ENRAF NONIUS, Rotterdam, The Netherlands). At the tilted position, the same measurements were taken at 0, 3, 5, 10, 15, and 20 min.
After excluding the first supine BP at 0 min, average supine systolic and diastolic BPs were estimated from the measurements at 5, 10, 15, and 20 min. SH was defined if the average supine systolic and/or diastolic BPs (SBP/DBP) were ≥140/90 mmHg 42 .
The lowest SBP/DBP at 3 or 5 minutes during the tilted position were selected for the diagnosis of OH. The orthostatic BP changes in systole (ΔSBP min ) and diastole (ΔDBP min ) were also calculated (supine average BP minus lowest orthostatic BP). When patients were hypertensive with ≥140/ 90 mmHg in the supine position, ΔSBP min and/or ΔDBP min ≥30/15 mmHg within 5 min was applied to define OH; otherwise, ΔSBP min and/or ΔDBP min ≥20/10 mmHg were adopted 43 .
The highest SBP/DBP among tilted measurements at 3, 5, 10, 15, and 20 min were re-selected, and the orthostatic ΔBPs were calculated from the average BPs (supine average BP minus highest orthostatic BP). PD patients with SH were defined as having orthostatic hypertension (OHT) if ΔSBP max and/or ΔDBP max was ≤ −20/10 mmHg. PD patients without SH were categorized as OHT when their orthostatic BP max was ≥140/90 mmHg or ΔBP max was ≤ −20/10 mmHg 19 .
Positive orthostatic ΔBP indicated a drop in standing BP, and the negative ΔBP an increase in standing BP. 123 I-metaiodobenzylguanidine ( 123 I-MIBG) scintigraphy 123 I-MIBG scintigraphy was performed using a dual-head camera equipped with a low-energy, high-resolution collimator (Siemens), and data were collected at 30-min (early) and 2-h (delayed) time points after the injection of 111 MBq of 123 I-MIBG. A static image was obtained with a 128 × 128 matrix. Regions of interest were manually drawn around the heart and mediastinum. Tracer uptake was measured within each region of interest to calculate the H/M ratio. The lower limit of the reference value for delayed H/M ratio was calculated to be 1.78 12,13 . A delayed H/M ratio < 1.78 was defined as abnormal. Washout rate was calculated as the following: [(early H/M ratio − late H/M ratio)/early H/M ratio] x 100 13 .

Statistical analyses
All statistical analyses were conducted with Jamovi software (version 1.6; retrieved from https://www.jamovi.org) for the Mac, a graphical user interface for R, with the additional jAMM module. The jAMM module provides a GLM mediation model that utilizes the lavaan R package 44 . Descriptive statistics, and independent or Welch's t-tests, the Mann-Whitney U test or Fisher's exact test were performed as appropriate. Mediation models, partialized by the covariates of age and disease duration, were manipulated to assess whether there was a mediating role for orthostatic ΔBP min (b) between cardiac sympathetic denervation (a) and disease burden (c). The direct association between delayed H/M ratio and composite scores (a → c) was established within the context of an indirect effect of orthostatic ΔBP min (a → b → c). Their relative roles were estimated using the maximum likelihood method 45 and statistical significance was defined as a two-tailed p value < 0.05.

Reporting summary
Further information on research design is available in the Nature Research Reporting Summary linked to this article.

DATA AVAILABILITY
Anonymized data generated during the current study are available from the corresponding author on reasonable request from individuals affiliated with research or health care institutions.

CODE AVAILABILITY
Jamovi is a statistical spreadsheet and graphical user interface (GUI) for R. All the analyses were performed basically using jamovi software as mentioned in the method, statistical analysis section. The R package mentioned was utilized within jamovi GUI.