Unraveling Autonomic Dysfunction in GBA‐Related Parkinson's Disease

ABSTRACT Background Patients with Parkinson's disease (PD) and GBA gene mutations (GBA‐PD) develop nonmotor complications more frequently than noncarriers. However, an objective characterization of both cardiovascular and sudomotor autonomic dysfunction using extensive clinical and instrumental measures has never been provided so far. Survival is reduced in GBA‐PD regardless of age and dementia, suggesting that other hitherto unrecognized factors are involved. Objectives To provide instrumental measures of pattern and severity of autonomic dysfunction in GBA‐PD and explore their correlation with other non‐motor symptoms and implications for clinical practice. Methods In this cross‐sectional study, 21 GBA‐PD and 24 matched PD noncarriers underwent extensive assessment of motor and non‐motor features, including neuropsychological testing. Cardiovascular autonomic function was explored through a comprehensive battery of indexes, including power spectral analysis of the R‐R intervals and blood pressure short‐term variability during resting state and active maneuvers. Dynamic Sweat Test was used to assess post‐ganglionic sudomotor dysfunction. Results Despite minimal or absent clinical correlates, cardiovagal and sympathetic indexes, heart rate variability parameters and sudomotor postganglionic function were more severely impaired in GBA‐PD than noncarriers (overcoming relatively preserved compensatory peripheral sympathetic function), suggesting more prominent cardiac sympatho‐vagal demodulation, efferent baroreflex failure and peripheral sympathetic dysfunction in GBA‐PD. Cardiovascular dysautonomia showed marginal correlations with cognitive impairment. Conclusions Compared to PD noncarriers, GBA‐PD display more severe instrumental autonomic abnormalities, which may be underestimated by purely clinical measures, despite their relevance on morbidity and mortality. This supports the necessity of implementing instrumental autonomic assessment in all GBA‐PD, regardless of clinically overt symptoms.

Heterozygous variants in the GBA gene, encoding for the lysosomal enzyme Glucocerebrosidase (GCase), are the most frequent genetic risk factor for Parkinson's disease (PD). 1,25][6][7][8] The increased mortality risk in PD carriers of GBA mutations (GBA-PD) is partially independent from aging and dementia, [4][5][6] suggesting that other yet unrecognized factors are at play.Conceivably, cardiovascular autonomic dysfunction may be one of these factors.
Despite clinical evidence of greater dysautonomia in GBA-PD is well-established, 4,5,9,10 instrumental characterization of pattern and severity of cardiovascular and sudomotor autonomic dysfunction is still insufficient.Nonetheless, several reasons support the importance of assessing autonomic dysfunction instrumentally rather than using scales/questionnaires.
(1) Dysautonomia is under-recognized in 50% of PD 11 and frequently asymptomatic: up to 75% of cases with abnormal instrumental autonomic tests do not report any complaint. 12On the other hand, over 30% of PD patients report orthostatic intolerance without any abnormalities on instrumental testing. 132) Some autonomic dysfunction (eg, cardiovagal modulation) have poor or no clinical correlates, despite its prognostic relevance on morbidity and survival.14,15 (3) Autonomic symptoms could be misdiagnosed on the basis of subjective reporting, which could be misleading.( 4) Both symptomatic and asymptomatic dysautonomia are associated with more rapid disease progression and increased the risk of institutionalization, falls, dementia, and sudden cardiac death in PD. [16][17][18] To date, only a study provided initial evidence of impaired heart-rate variability and lower vagal modulation in GBA-PD compared to PD noncarriers, but instrumental assessment was limited to rest and active standing, and sudomotor dysfunction was not explored.9 Conversely, the use of standardized operative protocols to extensively assess autonomic dysfunction should be addressed to promote reproducibility and comparability of data among centers and enable finer PD non-motor phenotyping and prognostic predictions.
In this scenario, this study aimed to provide an extensive clinical and instrumental characterization of both cardiovascular and sudomotor autonomic dysfunction through standardized broadly available autonomic testing protocols in a cohort of well-characterized GBA-PD versus matched PD noncarriers.In addition, we explored whether greater dysautonomia proceeds in parallel with more severe cognitive impairment in GBA-PD.We expected (1) to confirm GBA-PD patients have more frequent and severe autonomic dysfunction than PD noncarriers 4,5,9,10 and (2) to find that instrumental assessment of autonomic dysfunction is more sensitive than clinical questionnaires, implying that clinical assessment alone may underestimate an underlying cardiovascular autonomic dysfunction, which could contribute to the reduced survival related to GBA mutations. 4e finally expected (3) to find an association between dysautonomia and cognitive dysfunction (reduced frontal-lobe and/or visuo-spatial abilities) among carriers, supporting the concept that GBA-PD present more widespread α-syn pathology than noncarriers. 4,19

Study Population
This cross-sectional study included patients with clinical diagnosis of PD 20 and carrying heterozygous GBA pathogenic mutations or variants.Exclusion criteria were: (1) comorbidities confounding autonomic testing results or precluding their safe execution (ie, diabetes mellitus, unstable cardiovascular or respiratory disorders, renal failure, alcoholic abuse, current use of medications targeting the autonomic system (eg, adrenergic αor β-blockers, anticholinergics, etc.), vagal nerve stimulation, vagotomy); (2) diagnosis of dementia according to MDS criteria 21,22 or DSM-V 23 ; (3) diagnosis of any other neurological disorder; (4) mutations in other PD-related genes (LRRK2, parkin, PINK1, SNCA); (5) inability to attend at our Institute due to restrictions (eg, COVID-19 pandemic).Out of 716 patients with parkinsonism screened for GBA mutations from September 1, 2008 to September 30, 2020, we found 80 GBA-PD individuals.Out of this cohort, 25 GBA-PD were recruited between July 15 to October 20, 2020.Study flow chart is shown in Fig. S1.Twenty-five PD without GBA mutations and other exclusion criteria, carefully matched for age, sex and disease duration at assessment, were consecutively recruited during the same period as control group (PD noncarriers).All patients were of Caucasian ethnicity and Italian origin.Due to incomplete instrumental data in three GBA-PD patients and one noncarrier, only 21 GBA-PD and 24 PD noncarriers were considered for final analysis.
Details of the genetic analysis are available in the Supplementary methods. 24

Clinical Evaluation
All the subjects were evaluated at the Movement Disorder Unit of the Neurologic Institute Carlo Besta (Milan, Italy) by Neurologists experienced in movement disorders.Data were collected on main demographical and clinical information, family history, ongoing drugs and device-aided therapies for PD.Clinical work-up, performed for all patients in morning hours and ONmedication, included assessment of motor and non-motor symptoms, quality of life and functional impairment through the Italian version of the following internationally-validated scales: Movement Disorder Society-Unified Parkinson's Disease Rating Scale (MDS-UPDRS), 25,26 Hoehn & Yahr stage (H&Y), Non-Motor Symptoms Scale (NMSS), 11,27 REM Sleep Behavior Disorder Screening Questionnaire (RBDSQ), 28 Composite Autonomic Rating Scale (COMPASS-31), 29 8-items version of Parkinson's Disease Questionnaire (PDQ-8), 30 Activities of daily Living (ADL) and Instrumental ADL (IADL). 31,32RBD was not confirmed using polysomnography and it is to be considered "probable RBD" (pRBD).Neuropsychological testing is described in detail in the Supplementary methods.
Patients had clinical, neuropsychological, and autonomic testing on the same day.Patients were instructed to take the usual first Levodopa dose in the morning (8.00 am), then the neurologist performed the clinical scales (9.00 am) followed by the neuropsychological assessment (10.30-12.30am); finally, patients underwent autonomic testing (1.00-2.00pm).and all potential confounding medications for at least five half-lives, except for dopaminergic therapy (patients were allowed to take the first morning dose of Levodopa and autonomic testing had been performed at least 5 h away from the last levodopa intake).

Cardiovascular Autonomic Testing
Heart rate (HR) and beat-to-beat systolic blood pressure (sBP) and diastolic blood pressure (dBP) were recorded continuously using the Task Force ® Monitor device (CNSystems Medizintechnik AG, Graz, Austria).Thereafter, a standard Ewing's test battery 33 was used to derive the following parameters: basal cardiovascular parameters during rest; RR Intervals variation and E:I ratio during Deep Breathing (DB); Valsalva Ratio (VR, HR in phase II and IV) and presence of BP overshoot in phase IV during Valsalva Maneuver (VM); variation of sBP, dBP and HR after 3 and 5 min of Hand Grip test (HG); variation from baseline of sBP, dBP and HR during 10 min-Head-Up Tilt Test (HUTT) at each minute and 1 min after tilting down.In addition, the following sympathetic indexes (SI) were calculated according to BP variations during VM: SI-1 (BP Fall during phase 2); SI-2 (entity of BP Recovery late Phase 2); SI-3 (BP Recovery late Phase 2 compared with baseline); SI-4 (Magnitude of phase 4); SI-5 (Pressure recovery time) and BRSa (SI-6). 34Orthostatic hypotension (OH) was defined as a sustained decrease in sBP > 20 mmHg and/or dBP drop>10 mmHg within 3 min of HUTT 35,36 or as fall in sBP > 30 mmHg and/or dBP > 15 mmHg in the case of concomitant supine hypertension. 37The ΔHR/ΔsBP ratio at third minute after tilt-up was calculated. 38pectral analysis in the frequency domain of HR and BP beat-to-beat variability based on current HR variability (HRV) Guidelines were used to derive the high (HF, 0.15-0.40Hz) and low frequency (LF, 0.04-00.15Hz) powers of RR variability, as well as their ratio (LF/HF). 39Autoregressive method was used to obtain SDNN (Standard Deviation of Normal-to-Normal intervals), SDNN Index (reflecting the mean of the standard deviations of all NN intervals for all 5 min segments of the entire recording), SDANN (Standard Deviation of the Average of NN intervals), r-MSSD (root mean square of successive RR interval difference) and pNN50 (Percentage of the adjacent normalto-normal intervals that differed by ≥50 ms), as spectral shortterm measures of HRV in basal resting conditions for 5 min. 39e used the handgrip test (HT) to assess the non-baroreflex mediated α-adrenergic response.The HT was performed by asking patients to firmly squeeze a blood pressure monitor applying an isometric force at the 30% of maximal effort for 3 min.

Statistical Analysis
All the analyses were performed with the software STATA 16 (StataCorp, College Station, TX, USA).The primary aim was to compare PD carriers of GBA mutations to PD noncarriers.Therefore, the sample size was calculated on the primary outcome measure (SDNN index), which is the only index of long-term time-domain analysis (related to a deficit in the sympatho-vagal balance) that significantly differentiated PD patients and healthy control individuals. 43Expecting a clinically meaningful minimum mean difference of 10.5 ms in SDNN (standard deviation 14.0), the sample size sufficient to have a power of 80% with a two-tailed type I error of 10% was estimated to be 42 patients (21 per group).A two-tailed P value < 0.05 indicated statistical significance.The comparison of variables between PD noncarriers and different mutations subgroups was addressed as secondary aim.Details about descriptive statistics, regression and correlations analyses are provided in the Supplementary methods.

Ethics
This study was performed in agreement with the principles of the Declaration of Helsinki and approved by the local Ethics Committee (CE n.85/2021).All patients signed written informed consent.
Demographic, clinical and motor features are reported in Table 1.Patients did not differ for main demographic, clinical and therapeutic features (none was taking any medication to treat OH).No differences emerged according to GBA mutations severity.Motor fluctuations and dyskinesias were more common among GBA-PD than noncarriers.Non-motor Symptoms GBA-PD, mainly GBA-SM, displayed a greater burden of nonmotor symptoms than PD noncarriers according to the NMSS total score, particularly involving sleep (restless legs syndrome, but not pRBD), and pain in GBA-SM.Clinical scales and questionnaires did not reveal major differences on items related to autonomic dysfunction between GBA-PD and noncarriers (MDS-UPDRS-IB on Table 1; NMSS and COMPASS-31 on Table 2).Nevertheless, despite the limited sample size, GBA-SM described significantly more nocturia, pain and sweating dysfunction than noncarriers (Table 2).

Quality of Life and Functional Impairment
Quality of life was significantly worse in GBA-PD than PD noncarriers, mainly due to body discomfort and communication issues.ADL and IADL were similarly impaired, GBA-SM had greater functional impairment than noncarriers on daily cognitive tasks at PD-CFRS (Table 2).

Cardiovascular Reflex Responses
Cardiovascular reflex responses are reported in detail in Table 3 and shown in Fig. S2.At basal resting conditions and during parasympathetic tasks, we found no major differences in main cardiovascular parameters between GBA-PD and PD noncarriers.At HUTT, GBA-PD showed greater BP drop at 5 min than noncarriers.OH was observed in two GBA patients.During VM, the SI-2 was relatively more preserved in GBA-PD than PD noncarriers, while no differences were found for other SI.Responses during isometric handgrip test were similar across groups.Exploring main cardiovagal and sympathetic indexes, VR, RR interval variation during DB and E:I ratio showed a progressive decline from PD noncarriers to GBA-SM (Fig. S2).

Time-Domain and Frequency-Domain Analysis of Heart Rate Variability
Spectral analysis of resting HRV demonstrated a high intergroup discrimination ability, as demonstrated by the coefficient of determination (R2) obtained by regression analysis (Table 3, Fig. 1).SDNN, SDANN, SDNN index, r-MSSD, pNN50 indexes of HRV were significantly lower in GBA-PD, GBA-SM and GBA-MM compared to PD noncarriers, while SDNN and r-MSSD significantly differed between GBA-SM and GBA-MM.
Concerning frequency-domain analysis, groups did not differ for LF/HF ratio in basal resting conditions and during DB.During HUTT, the upward trend of LH/HF ratio was less evident for GBA-PD than noncarriers (Fig. S3).Compared to

Sudomotor Autonomic Testing
Dynamic Sweat Testing showed a significantly lower Sweat/ Cm 2 /min in GBA-PD than noncarriers in the most affected body side (Table 3).We found no difference between GBA-SM and GBA-MM compared to PD noncarriers and between GBA-SM vs. GBA-MM.

Relationship Between Autonomic Dysfunction and Cognitive Impairment
In

Discussion
We extensively investigated cardiovascular autonomic function and, for the first time, sudomotor function in a cohort of GBA-PD and matched PD control subjects.In particular, we provided the instrumental correlate of greater cardiovascular and sudomotor dysautonomia in GBA-PD and outlined a specific pattern of sympathetic and cardiovagal dysregulation.Moreover, we demonstrated a clear mismatch between subjective reporting of symptoms and objective instrumental findings, supporting the appropriateness of implementing autonomic testing in clinical practice.

Cardiovagal Function
Parasympathetic function during active maneuvers was similarly affected in GBA-PD and PD noncarriers.Nevertheless, spectral measures of HRV were severely impaired in GBA-PD compared to PD noncarriers.Some spectral indexes (r-MSSD and SDNN) proved so robust to even differentiate GBA-PD according to mutation type (PD noncarriers < GBA-MM < GBA-SM).Normal HRV in long-term (at least 24-h) recordings is an index of good cardiovascular health and overall autonomic integrity. 44,45evertheless, short-term measurements are more standardized due to less susceptibility to environmental influences, so being able to more selectively reflect cardiac sympatho-vagal modulation. 46,47In line with recent evidences, 9 our results confirm a dysfunctional sympatho-vagal cardiac modulation during rest in GBA-PD, appearing before the efficiency of active compensatory responses is lost.The relatively younger age of our GBA-PD cohort further strengthens the relevance of our data and the likelihood of replication in independent cohorts, as an age-related decline of HRV measures is physiologically expected. 44Importantly, HRV carries a remarkable prognostic and predictive role: low HRV is independently associated with increased risk of sudden cardiac death 18 and overall mortality 14,48 and it is associated with an increased risk of PD in the general population. 49Furthermore, reduction of HRV is a relative early finding (being reported in idiopathic RBD 50 and at different PD stages 45,47,51,52 ) that predicts the development and severity of OH. 51,52

Sympathetic Function
Our data suggest a similar cardiac sympathetic function in GBA-PD and PD noncarriers during active maneuvers, along with a relative preservation of baroreflex-mediated vasoconstrictive α-adrenergic responses (SI-2 index) in GBA-PD. 34Despite this, compared to PD noncarriers, GBA-PD showed a clearly distinct pattern of baroreflex modulation during HUTT, characterized by a milder increase of LF/HF ratio.An increase in LF/HF ratio is normally expected upon orthostatic stress 53 indicating a prompt sympatho-vagal responsiveness to standing.In patients with unexplained syncope, decrease in LF/HF ratio in the first 5 min of HUTT predicts syncope versus normal test. 54Accordingly, GBA-PD displayed larger orthostatic sBP drops compared to PD noncarriers.Evidences show that BP compensatory responses in early HUTT rely on sympathetic-induced vasoconstriction, 34 while BP values maintenance over time requires a combination of cardiovascular sympathetic responses and cardiovagal modulation. 30In line with this, our results suggest that a more severe cardiac sympatho-vagal demodulation in GBA-PD, along with relative sparing of peripheral vasoconstrictive responses, is responsible for inadequate increase of LH/HF and poor BP maintenance during HUTT. 55Overall, considering that sympathetic cardiac impairment and reduction in HRV are hallmarks of efferent baroreflex failure in synucleinopathies, 56 our results demonstrate a greater efferent baroreflex insufficiency in GBA-PD compared to PD noncarriers.Some studies suggested a potential effect of Levodopa in lowering BP during HUTT and interfering on sympathetic-parasympathetic regulation during orthostasis, but conclusive results are lacking. 57Despite a confounding effect of dopaminergic therapy is not completely excludible, our study groups were all uniform for total LEDD and autonomic testing was performed for all patients at least 5 h after the last levodopa intake.Moreover, the ONmedication state is more suitable for common clinical practice, as many patients are not able to tolerate a prolonged study protocol in the medication-OFF state.

Sudomotor Autonomic Dysfunction
Post-ganglionic sweat output was significantly more impaired in GBA-PD than PD noncarriers, with a seemingly selective involvement of the most affected body side.In PD, sweating dysfunction displays a length-dependent progression, resulting in hypo-anhidrotic areas at distal sites and compensatory axial sweating. 58,59At skin biopsy, PD displays intraepidermal nerve fiber denervation and loss of autonomic nerves to blood vessels and sweat glands, consistent with somatic and autonomic small fiber neuropathy. 43,60α-syn deposits in autonomic fibers, including around sweat glands, are more abundant in PD with autonomic failure. 41,43Reduced sweat production in GBA-PD could suggest a greater degeneration of autonomic fibers directed to sweat glands on more affected body side.Accordingly, α-syn deposits in autonomic fibers have been detected from skin biopsy in GBA-PD, 61,62 despite the role of skin α-syn deposits is still unclear 63 and further studies are needed.

Is Cardiovascular Dysautonomia Associated with Concomitant Cognitive Dysfunction?
We found a positive association between worse cognitive performances and parasympathetic dysfunction, particularly the E:I ratio.However, this was not specific to GBA-PD, because the E:I ratio was positively associated to visuo-spatial abilities in both carriers and noncarriers.
Nevertheless, only GBA-PD displayed a significant association between the E:I ratio and FAB scores as well as concomitant MCI.This is consistent with previous studies showing an association between reduced parasympathetic measures (E:I ratio, 30:15 ratio and r-MSSD index) and greater risk for MCI in PD. 64,65 Notably, decreased parasympathetic activity is associated with worse cognitive performance both in general population without dementia 66 and in Alzheimer's disease, 67 suggesting a synucleinopathy-independent association between parasympathetic function and cognition.Greater autonomic dysfunction in GBA-PD could actively exacerbate cognitive deficits, with main regards to executive functions.Common underlying pathophysiology should be considered, 68 in terms of hemodynamic-related effects (eg, impaired cerebral autoregulation causing white matter lesions), 68,69 or neuropathological links (eg, brainstem autonomic nuclei degeneration causing noradrenergic and cholinergic denervation of prefrontal and anterior cingulate cortices), 68,70 or both.

Limitations and Strengths
We acknowledge that our study has limitations, the most relevant being the small sample size.
Therefore, further study in larger cohorts of GBA-PD versus PD noncarriers, additionally including patients with dementia with Lewy bodies and non-PD healthy control groups, is needed to replicate the present findings.Moreover, autonomic testing had been performed without stopping DA agonists and Levodopa for at least five half-lives as we did for other concomitant medications potentially impacting autonomic testing but limited to 24 h after the last DA agonists intake and at least 5 h away from the last levodopa intake.Levodopa was not stopped because its pharmacodynamic effects (long-duration response) last much longer than five half-lives. 71Some nonmotor and autonomic symptoms were not objectively assessed, and we did not provide data on MIBG myocardial scintigraphy and skin biopsy.Finally, we acknowledge that the DST has been used on a limited cohort of PD patients so far as compared to other tests used to investigate the sudomotor function, such as the "QSART" and the "TST".The DST has been recently applied to differentiate PD from MSA, 41 and it may be more widely used in clinical practice because its equipment is the least expensive.
Despite such limitations, there are strengths worth emphasizing.First, this is the first study providing a comprehensive instrumental assessment of both cardiovascular and sudomotor autonomic function in a well-characterized cohort of GBA-PD, studied with standardized operative protocols and in reproducible conditions and compared to carefully matched PD controls.In confirmation, despite small sample size, highly significant differences (P < 0.001) were obtained for all time-domain spectral data analyses, even when limiting analysis to severe versus mild GBA mutations, emphasizing the strong effect of GBA mutations on cardiovascular autonomic function, particularly on HRV.

Conclusions
The present study demonstrates that GBA-PD display more severe cardiovascular autonomic impairment, mainly due to greater cardiac sympathetic and parasympathetic demodulation and impaired efferent baroreflex, along with a more pronounced sudomotor dysfunction.Critically, clinical correlates of these autonomic abnormalities could be minimal or absent, despite their potential impact on morbidity (eg, falls and fractures) and mortality.Translated into clinical practice, we suggest assessing cardiovascular autonomic reflexes instrumentally in all GBA-PD patients (eg, cardiovascular autonomic tests, prolonged ambulatory test for delayed OH, 24 h ECG monitoring and/or shortrecordings of HRV whenever available), even in the absence of clinical complaints and particularly in younger individuals, whose compensatory mechanisms could mask or delay any symptom.We also encourage to use standardized autonomic protocols to promote comparability of data among centers.Finally, in a "precision medicine" perspective, cardiovascular autonomic measures (mainly HRV) should be included as outcome measures in disease-modifying clinical trials and in future prospective studies investigating survival in GBA-PD.

Figure 1 .
Figure 1.Scatter plots for spectral analyses (r-MSSD and pNN50) for PD noncarriers (PD) and GBA-PD (A-C) and for PD noncarriers, GBA-SM and GBA-MM (B-D), showing 25th and 75th percentile (lower and upper side of the box respectively), the median (middle line) and the minimum to the maximum values (lines extending from box ranges).**P < 0.001.

TABLE 2
Significant values (P < 0.05) are shown in bold.Percentage of successive RR intervals that differ by more than 50 ms; sBP, Systolic blood pressure; SDNN, Standard deviation of NN intervals; SDANN, Standard deviation of the average NN intervals; SI, Sympathetic Index; SM, severe mutations; r-MSSD, Root mean square of successive RR interval differences; VR, Valsalva Ratio.
All scores are corrected for age and education level and reported as mean AE SD.All P-values are calculated according to linear (continuous variables) regression models adjusted for disease duration.Significant values (P < 0.05) are shown in bold.P-values under 0.1 are considered as trends for significance and are shown in italics.