The Relationship Between Autonomic Dysfunction and Mood Symptoms in De Novo Parkinson’s Disease Patients Over Time

Abstract

Background:

Autonomic dysfunction is prevalent in Parkinson’s disease (PD) and can worsen quality of life. We examined: (a) whether specific autonomic symptoms were more strongly associated with anxiety or depression in PD and (b) whether overall autonomic dysfunction predicted mood trajectories over a 5-year period.

Methods:

Newly diagnosed individuals with PD (N = 414) from the Parkinson’s Progression Markers Initiative completed self-report measures of depression, anxiety, and autonomic symptoms annually. Cross-sectional linear regressions examined relationships between specific autonomic subdomains (gastrointestinal, cardiovascular, thermoregulatory, etc.) and mood. Multilevel modeling examined longitudinal relationships with total autonomic load.

Results:

Gastrointestinal symptoms were associated with both higher anxiety (b = 1.04, 95% CI [.55, 1.53], P < .001) and depression (b = .24, 95% CI [.11, .37], P = .012), as were thermoregulatory symptoms (anxiety: b = 1.06, 95% CI [.46, 1.65], P = .004; depression: b = .25, 95% CI [.09, .42], P = .013), while cardiovascular (b = .36, 95% CI [.10, .62], P = .012) and urinary symptoms (b = .10, 95% CI [.01, .20], P = .037) were associated only with depression. Longitudinally, higher total autonomic load was associated with increases in both depression (b = .01, 95% CI [.00, .02], P = .015) and anxiety (b = .04, 95% CI [.01, .06], P < .001) over time, as well as occasion-to-occasion fluctuations (depression: b = .08, 95% CI [.05, .10], P < .001; anxiety: b = .24, 95% CI [.15, .32], P < .001).

Conclusion:

Findings suggest autonomic dysfunction, particularly gastrointestinal and thermoregulatory symptoms, may be an indicator for elevated anxiety/depression and a potential treatment target early on in PD.

Introduction

In this study, we examined how specific autonomic symptoms are differentially linked to depression and anxiety in individuals with Parkinson’s disease and how overall autonomic load might relate to trajectories of mood over time. Parkinson’s disease (PD) is a progressive neurodegenerative disorder with hallmark motor symptoms of tremor, rigidity, bradykinesia, and postural instability.¹ Over the last 30–40 years, there has been increased recognition of the non-motor features of PD, including autonomic symptoms. Autonomic symptoms affect approximately 50%–80% of PD patients^2–5 and can occur at the earliest stages in the disease course.^6–8 In PD, presentations of autonomic insufficiency include gastrointestinal, cardiovascular, thermoregulatory, urinary, and sexual dysfunction. Gastrointestinal dysfunction is particularly common, with constipation being a prodromal indicator for subsequent PD even up to a decade before diagnosis.⁹ Gastrointestinal dysfunction and other early autonomic symptoms are thought to be a result of α-synuclein pathology accumulating in the peripheral sympathetic, parasympathetic, and enteric nervous systems.⁵ Indeed, α-synuclein aggregates can be detected in the gastrointestinal tract years before the onset of motor symptoms.^10,11 This is consistent with Braak’s hypothesis of PD staging, with regard to α-synuclein initially occurring in autonomic peripheral nerves (e.g., enteric nerve cell plexuses, dorsal motor nucleus of the vagus) before gradually progressing up the brain stem, with regions within the central autonomic network affected later on.^12,13 Consequently, key areas within the central autonomic network (e.g., hypothalamus, cortex, insula) have been shown to be susceptible to α-synuclein accumulation in PD and implicated in abnormalities of various systems, including gastrointestinal, cardiovascular, and urinary.^5,14–17

Autonomic symptoms can negatively impact quality of life and activities of daily living among individuals with PD.¹⁸ Importantly, several studies have demonstrated that autonomic symptoms are associated with depression and anxiety in both PD and non-PD populations. Among the general population, a wealth of literature has examined the relationship between the autonomic nervous system and depression/anxiety, particularly through the lens of cardio-vagal metrics like heart rate variability (HRV)—the change in the time interval between successive heartbeats and an index of the parasympathetic nervous system.^19–21 Research has shown that individuals with depression and/or anxiety have a lower HRV,^22,23 which may be related to lower emotion regulation, social engagement, and flexibility in responding to environmental demands.^24,25 In PD, studies using both self-report measures of autonomic symptoms as well as physiological probes (e.g., heart rate variability, baroreflex-cardio-vagal indicators) have found that autonomic insufficiency is associated with depression/anxiety.^26–34 There are several potential mechanisms for the relationship between autonomic dysfunction and mood in PD. One possibility would be an overlap in the limbic system and central autonomic network brain structures (e.g., pathology in areas such as the hypothalamus and anterior cingulate cortex could give rise to both mood and autonomic symptoms).^30,35 Relatedly, disruptions in neurotransmitter systems may also play a role in both. For example, disruptions in noradrenaline innervation have been linked to both depression and cardiovascular autonomic dysfunction in PD,^36,37 while serotonin systems are implicated in constipation/gastrointestinal motility^38,39 and anxiety symptoms.⁴⁰ A third possibility would be a negative emotional reaction to symptoms, such that increased autonomic symptoms lead to increased depressive and anxiety symptoms due to the negative influence these symptoms have on daily functioning.¹⁸ This relationship could also be bidirectional, with increased depression/anxiety potentially resulting in a higher self-perceived burden of autonomic symptoms.^41,42

Although prior studies have generally shown a relationship between autonomic insufficiency and mood in PD, it is not clear whether certain types of autonomic symptoms have a stronger influence on mood or how this might vary in anxiety vs depression. This is important for treatment purposes, as therapeutics targeting specific autonomic symptoms could in turn lead to improvements in mood (or vice versa). Likewise, there has been limited research on longitudinal relationships to date.^27,30,43 Recently, Sklerov and colleagues,³⁰ using an open-access database of early PD patients, found that depression was significantly associated with autonomic burden across several different time points (i.e., baseline, 2 years, 5 years).

Given the prevalence of anxiety in PD, the present study sought to extend previous observations beyond depression to include anxiety. We had 2 aims. Aim 1 (cross-sectional analysis) addressed whether specific autonomic symptoms were more strongly related to anxiety vs depression. We hypothesized that cardiovascular symptoms would be more strongly linked to depression than other autonomic symptoms, given the potential overlap in neurotransmitter systems (e.g., norepinephrine)^36,37 and a robust literature showing cardiac autonomic imbalance in non-PD populations with depression, such as reduced HRV^22,24 and orthostatic hypotension/lower baroreflex sensitivity.^44–46 Regarding anxiety, we hypothesized gastrointestinal symptoms would be more strongly associated with anxiety than other autonomic symptoms. This hypothesis was based on prior findings linking anxiety with gastrointestinal dysfunction in PD,^27,32,43 potential shared mechanisms such as serotonin dysregulation,^38–40 and gut-brain axis relationships.^12,13 Aim 2 (longitudinal analysis) addressed whether autonomic symptoms more broadly would be associated with worsening of both anxiety and depression over time (i.e., trajectory of change). We predicted that a higher autonomic load overall would be associated with worsening of both depression and anxiety over a 5-year period. This was based on overlap in the central autonomic network/limbic system and neurotransmitter systems, as well as the negative impact of autonomic symptoms on quality of life/functioning.¹⁸

Methods

Data Acquisition

Data were obtained from the Parkinson’s Progression Markers Initiative (PPMI) database (www.ppmi-info.org). The PPMI is a longitudinal multi-site study of newly diagnosed (2 years or less), untreated PD patients. Detailed information about the aims and methodology of the PPMI study is described by Marek et al.⁴⁷ Of note, although individuals were untreated at baseline, most initiated dopaminergic therapy as the study progressed. The PPMI research protocol and data collection methods were approved by the institutional review board at each participating site. All study participants provided written informed consent.

Participants

The current sample included 414 individuals with idiopathic PD who had been diagnosed within the previous 24 months and were not taking dopaminergic medications at the time of enrollment. Individuals who were missing data at baseline (N = 9) were excluded. In brief, participants had an average age of 61.6 years (SD = 9.7) at baseline (range = 33 to 84 years), were mostly male (66.2%), were mostly White (92.5%), and had a high level of education (15.6 years, SD = 3.0; Table 1).

Table 1.

Demographics and Disease Characteristics at Baseline (N = 414).

Measure	Mean	Range
Age, years (SD)	61.6 (9.7)	33–84
Education, years (SD)	15.6 (3.0)	5–26
Sex, male (%)	274 (66.2)	-
Race, White (%)	383 (92.5)	-
Ethnicity, Hispanic (%)	9 (2.2)	-
MDS-UPDRS part III “on” (SD)	20.9 (8.8)	4–51

MDS-UPDRS, Movement Disorder Society-sponsored revision of the Unified Parkinson’s Disease Rating Scale.

Materials and Procedure

Participants were followed for up to 5 years (baseline and 5 annual follow-ups). Autonomic symptoms were assessed using the Scales for Outcomes in Parkinson’s Disease-Autonomic Dysfunction (SCOPA-AUT). The SCOPAAUT is a validated self-report measure consisting of 25 items developed and recommended for assessing autonomic symptoms in Parkinson’s disease.^48–50 The SCOPAAUT includes a total score, as well as 6 subdomain scores based on symptoms in the following categories: gastrointestinal (7 items), urinary (6 items), cardiovascular (3 items), thermoregulatory (4 items), pupillomotor (1 item), and sexual (2 items for males and 2 items for females). The occurrence of each item over the past month is scored as zero (never), 1 (sometimes), 2 (regularly), or 3 (often). Domains with at least 1 item having a score of ≥ 1 were considered to be “impaired,” in accordance with methods used by Stankovic et al.²⁷

Depressive symptoms were measured using the Geriatric Depression Scale-15 (GDS), a 15-item self-report measure with “yes-no” questions assessing the presence of depressive symptoms over the past week.⁵¹ A 4/5 cutoff score has been previously recommended to identify individuals with possible depression in PD.⁵² The State-Trait Anxiety Inventory (STAI) was used as an index of anxiety.⁵³ The STAI consists of 2 scales—“state” anxiety measuring current symptoms and “trait” anxiety measuring dispositional/enduring symptoms. Higher scores indicate more severe anxiety, and scores of ≥ 40 suggest clinically significant anxiety.⁵⁴ For the present study, we focused on the trait scale given interest in examining enduring symptoms of anxiety rather than situational anxiety. Of note, neither the GDS nor the STAI trait contain items directly assessing autonomic-related symptoms; thus, there was no overlap with the SCOPA-AUT. Motor severity was measured using the clinician-rated Movement Disorder Society-sponsored revision of the Unified Parkinson’s Disease Rating Scale (MDS-UPDRS).⁵⁵ The Part III score was used which assesses common motor features including tremor, rigidity, postural instability, and bradykinesia, with higher scores reflecting greater severity of symptoms. Scores were obtained while participants were taking their normal dose of dopaminergic medications (i.e., “on medication” condition) if they had initiated dopaminergic medication (following baseline).

Statistical Analyses

Statistical analyses were conducted using SPSS version 27.0 software (IBM Corp., Armonk, NY). For Aim 1 (cross-sectional analysis), to evaluate the unique relationships between mood and autonomic symptoms, linear regression analyses were conducted with autonomic (SCOPA-AUT) subdomain scores as predictor variables and separate models for depression (GDS) and anxiety (STAI trait). This analysis was conducted using data from Year 5 given wider distributions of mood and autonomic symptoms (supplemental analyses also examined these relationships at Baseline). Due to the non-normality of the data, bias-corrected bootstrapping using 1000 samples was performed. Bootstrapping is a nonparametric statistical procedure which involves resampling the dataset with replacement to create a multitude of simulated samples (N = 1000 in present analyses) which are then used to calculate standard errors and for hypothesis testing without the assumption of normality.⁵⁶ For Aim 2 (longitudinal analysis), multilevel models (MLM) were used to examine longitudinal relationships between autonomic symptoms and mood. STAI trait and GDS were entered as dependent variables in separate models. Predictors included time, baseline age, sex, MDS-UPDRS Part III motor scale, and SCOPA-AUT total score. In addition to the main effects of autonomic symptoms and motor severity, interaction terms were computed with time-varying autonomic and motor variables centered around each individual’s mean for these variables. These interaction terms examined whether (a) autonomic/motor symptoms explained individual differences in rate of change in mood and (b) occasion-to-occasion fluctuations in autonomic/motor symptoms predicted occasion-to-occasion fluctuations in mood. Random effects were modeled for time and time-varying autonomic and motor variables. Models used bias-corrected bootstrapping with 1000 samples owing to non-normality with full information maximum likelihood estimation. Models estimated homogenous variance and no covariance in repeated measures (scaled identity), and random effects were uncorrelated (variance components). Model estimates presented represent unstandardized regression coefficients controlling for covariates. In addition, given that both dopaminergic medications⁵⁷ and anti-depressants⁵⁸ have been shown to be related to autonomic nervous system variables, analyses were conducted to examine whether average dopaminergic medication burden (L-dopa equivalent daily dose; LEDD) throughout the study duration or psychotropic use (e.g., anti-depressants, antianxiety) at any time during the study duration were related to individuals’ mean SCOPA-AUT Total score using Pearson’s correlations.

Results

Table 2 shows the percentages of individuals endorsing at least 1 symptom among the various autonomic symptom subdomains. Autonomic symptoms were prevalent in the sample at both Baseline and Year 5. In particular, urinary symptoms (Baseline = 94.7%, Year 5 = 95.8%) and gastrointestinal symptoms (Baseline = 74.9%, Year 5 = 90.6%) were especially common. With regard to mood symptoms, approximately 1/5^th of the sample was above clinical cutoffs in depression and anxiety at both Baseline and Year 5, though this was slightly lower for depression at Baseline (14%; Table 2). Pearson’s correlations showed weak relationships between LEDD and SCOPA-AUT Total (r = .11, 95% CI [.01, .21], P = .035) as well as use of psychotropics (46.9% of the sample at any time during study) and SCOPA-AUT Total (r = .22, 95% CI [.13, .31], P < .001). As such, these variables were not included as covariates in models.

Table 2.

Proportion of Sample Endorsing Autonomic Symptoms and Clinically Significant Mood Symptoms at Baseline Versus Year 5.

	Baseline (N = 414)	Year 5 (N = 309)
SCOPA-AUT symptoms (% endorsing)
Urinary	94.7	95.8
Gastrointestinal	74.9	90.6
Thermoregulatory	57.0	65.2
Sexual	46.4	58.1
Pupillomotor	34.3	41.9
Cardiovascular	33.8	44.5
Mood symptoms (% above cutoff)
GDS ≥ 5	14.0	20.3
STAI trait ≥ 40	20.0	21.6

Values are presented as percentage of participants with a score ≥ 1 for SCOPA-AUT domains and percentage above clinical cutoffs for mood measures.

SCOPA-AUT, Scales for Outcomes in Parkinson’s Disease-Autonomic Dysfunction; GDS, Geriatric Depression Scale-15; STAI, State Trait Anxiety Inventory.

Aim 1: Relationships Between Specific Autonomic Symptoms And Mood Cross-Sectionally

Aim 1 examined the unique cross-sectional relationships between autonomic symptoms (SCOPA-AUT subdomains) and depression (GDS)/anxiety (STAI trait) at Year 5 using bootstrapped linear regressions (Table 3). For depression, autonomic symptoms explained 28% of the variance. Unique predictors included gastrointestinal (b = .24, 95% CI [.11, .37], β = .23, P = .012, semi-partial r² = .03), thermoregulatory (b = .25, 95% CI [.09, .42], β = .18, P = .013, semi-partial r² = .02), cardiovascular (b = .36, 95% CI [.10, .62], β = .15, P = .012, semi-partial r² = .02), and urinary (b = .10, 95% CI [.01, .20], β = .13, P = .037, semi-partial r² = .01) symptoms. For anxiety, autonomic symptoms explained 28% of the variance, with gastrointestinal (b = 1.04, 95% CI [.55, 1.53], β = .27, P < .001, semi-partial r² = .04) and thermoregulatory (b = 1.06, 95% CI [.46, 1.65], β = .20, P = .004, semi-partial r² = .03) symptoms being the only unique predictors. The pattern of findings was similar using data from Baseline, though with fewer unique relationships identified and less variance explained (Supplemental Table S1).

Table 3.

Regression Analyses for Contributions of Autonomic Symptoms to Mood at Year 5.

Criterion	Predictor	F	R2	b [95% CI]	β	Semi-partial r2
Depression		19.51**	.28
	Urinary			.10 [.01, .20]*	.13*	.01
	Gastrointestinal			.24 [.11, .37]*	.23*	.03
	Thermoregulatory			.25 [.09, .42]*	.18*	.02
	Sexual			−.15 [−.31, .00]	−.10	.01
	Pupillomotor			.19 [−.18, .56]	.06	.00
	Cardiovascular			.36 [.10, .62]*	.15*	.02
Anxiety		19.84**	.28
	Urinary			.37 [.02, .71]	.12	.01
	Gastrointestinal			1.04 [.55, 1.53]**	.27**	.04
	Thermoregulatory			1.06 [.46, 1.65]*	.20*	.03
	Sexual			−.33 [−.89, .24]	−.06	.00
	Pupillomotor			.19 [−1.15, 1.53]	.02	.00
	Cardiovascular			.90 [−.04, 1.85]	.11	.01

^*P ≤ .05,

^**P ≤ .001.

Aim 2: Longitudinal Relationships Between Overall Autonomic Dysfunction and Mood

This aim examined longitudinal relationships between overall autonomic insufficiency (SCOPA-AUT Total score) and depression (GDS)/anxiety (STAI trait) using bootstrapped multilevel modeling (Table 4). For the model estimating depression, results revealed that SCOPA-AUT Total showed a significant and positive main effect (b = .14, 95% CI [.11, .18], P < .001), indicating that individuals reporting more autonomic symptoms on average tended to report a higher average level of depression. Overall autonomic dysfunction also moderated individual differences in rate of change: namely, individuals who reported more autonomic symptoms on average showed a slightly greater increase in depression over time (b = .01, 95% CI [.00, .02], P = .015). Figure 1 depicts this interaction term, illustrating that individuals with more frequent autonomic symptoms experienced a greater increase in depressive symptoms relative to their counterparts reporting less frequent autonomic symptoms. Additionally, occasion-to-occasion variations in autonomic symptoms were associated with concurrent variations in depression, such that on those occasions when an individual’s SCOPA-AUT Total score was higher than average, their GDS score was also higher than average (b = .08, 95% CI [.05, .10], P < .001). In contrast, motor severity did not relate to level or rate of change in depression (Ps > .05), though it did predict occasion-to-occasion fluctuations (b = .03, 95% CI [.01, .04], P < .001). This model explained 25% of the within-person variability, 24% of the between-person variability, and 31% of the individual differences in rates of change of depression.

Table 4.

Multilevel Model Analyses Predicting Mood Over a 5-Year Period.

	Depression model estimates [95% CI]	Anxiety model estimates [95% CI]
Initial status
Intercept	1.90 [.54, 3.27]**	36.55 [31.51, 41.59]**
Baseline age	−.02 [−.04, .00]*	−.19 [−.27, −.11]**
Sex	.22 [−.21, .64]*	1.82 [.23, 3.41]**
Mean SCOPA-AUT	.14 [.11, .18]**	.54 [.40, .67]**
Mean MDS-UPDRS part III “on”	.01 [−.02, .03]	.05 [−.04, .14]
Time-varying SCOPA-AUT	.08 [.05, .10]**	.24 [.15, .32]**
Time-varying MDS-UPDRS part III “on”	.03 [.01, .04]**	.08 [.04, .12]**
Rate of change
Time	−.16 [−.33, .00]*	−.53 [−1.02, −.05]*
Time X mean SCOPA-AUT	.01 [.00, .02]*	.04 [.01, .06]*
Time X mean MDS-UPDRS part III “on”	.00 [−.01, .01]	.00 [−.02, .02]
Fit statistics
aΔ −2LL	783.08	1031.13
aΔ AIC	759.08	1007.13
aΔ BIC	691.65	939.72
η2 within	.25	.29
η2 between	.24	.23
η2 time slope	.31	.37

^*P ≤ .05,

^**P ≤ .001.

LL = Log Likelihood; AIC = Akaike Information Criterion; BIC = Bayesian Information Criterion.

^aChanges in fit statistics are decreases relative to a null model with no predictors. These values (i.e., −2LL, AIC, BIC) are standard indicators of fit that are used for model comparison. Smaller values indicate better fit, in this case compared to the null (worst fitting) model. Both AIC and BIC penalize fit based on number of parameters in the model, while −2LL does not.

Figure 1.

Model-implied Autonomic Symptoms × Occasion Interaction. Note: Autonomic symptom characterization (high vs low) was based on median split of SCOPA-AUT Total score. Depression Y-axis reflects GDS score whereas Anxiety Y-axis reflects STAI trait score.

With regard to anxiety, the pattern of results was similar. Higher overall average autonomic dysfunction was associated with a higher level of anxiety (b = .54, 95% CI [.40, .67], P < .001). Likewise, overall autonomic dysfunction moderated rate of change in anxiety. As Figure 1 shows, individuals who reported more autonomic symptoms had greater increases in anxiety over time (b = .04, 95% CI [.01, .06], P < .001). On occasions where individuals had higher than average autonomic symptoms, their anxiety was also higher than average (b = .24, 95% CI [.15, .32], P < .001). Motor severity also predicted occasion-to-occasion fluctuations in anxiety (b = .08, 95% CI [.04, .12], P < .001), but did not relate to level or rate of change (Ps > .05). This model explained 29% of the within-person variability, 23% of the between-person variability, and 37% of the individual differences in rates of change.

Discussion

Past research investigating the relationship between mood and autonomic functioning in PD has been primarily cross-sectional. Moreover, there has been limited research on how specific autonomic symptoms (e.g., cardiovascular, gastrointestinal) might differentially relate to specific emotional symptoms (e.g., depression, anxiety). The current study expands upon the existing literature by examining differential contributions of autonomic symptoms to mood as well as looking at longitudinal relationships with overall autonomic load.

Our study found that by the fifth annual follow-up, autonomic symptoms explained over 25% of the variance in both anxiety/depression scores. These findings align with prior research which has shown that autonomic insufficiency in PD is associated with both anxiety^26,32,33 and depression.^28–30 Our study adds to the existing literature by addressing which specific autonomic symptoms most strongly contribute to this relationship. Namely, we found that gastrointestinal symptoms were most closely related to both anxiety and depression. This is in line with our hypothesis regarding an anxiety-gastrointestinal link but does not provide support for our hypothesis that the cardiovascular-depression link would be strongest. Although we did not examine biological mechanisms, we hypothesize that relationships between mood and gastrointestinal dysfunction could potentially be related to gut-brain axis interactions, which have received increasing attention in both the pathogenesis of PD⁵⁹ and psychiatric conditions.^60,61 Specifically, changes such as gut microbiota imbalance (dysbiosis), inflammation, and altered intestinal permeability (leaky gut) could lead to increased cytokine production, blood-brain barrier disruptions, and central nervous system inflammation/neuronal dysfunction, which could increase the risk for developing anxiety and depression.^32,59 This relationship is also likely bidirectional. Stress is thought to induce gastrointestinal changes such as altered microbiome composition and increased intestinal permeability,^62,63 and longitudinal studies have shown that higher levels of anxiety are associated with an increased risk of subsequently developing gastrointestinal disorders like irritable bowel syndrome.⁶⁴ Further, it is possible that gastrointestinal symptoms in particular have a significant impact on quality of life, leading to a stronger negative emotional reaction than to other autonomic symptoms (i.e., a primarily psychological relationship). Additionally, thermoregulatory dysfunction emerged as a unique predictor of both anxiety and depression, which is in line with at least 2 prior studies finding similar relationships.^33,34 Finally, we found that urinary symptoms were associated with increased depression, as were cardiovascular symptoms (corresponding to non-PD literature demonstrating cardiac autonomic imbalance in depression).²⁴ Overall, we found a greater number of unique autonomic predictors of depression than anxiety, with similar amount of variance explained in both mood variables.

Our study found that over a 5-year period, overall autonomic dysfunction was associated with higher average depressive and anxiety symptoms and slightly greater rates of increase in both over time. Additionally, on those occasions where one’s autonomic symptom burden was above average, mood symptoms were also more likely to be above average. Of note, motor severity was not associated with level or rates of change in mood. While the clinical relevancy of these findings remains to be ascertained, our results imply that the trajectory of both depression and anxiety in de novo PD patients over the course of several years seems to be related to severity of autonomic symptoms. This is important as intervening and treating autonomic symptoms early on in the course of PD could potentially influence the longer term development of mood symptoms. Future research should examine how autonomic insufficiency may interact with other symptoms in Parkinson’s disease (e.g., sleep, cognitive impairment) to influence trajectories of mood symptoms.

Our findings are consistent with other longitudinal investigations of autonomic dysfunction and anxiety/depression in PD.^27,30,43 It is likely that these relationships can be explained by overlap in neural substrates for depression/anxiety and autonomic dysfunction. Key areas in the central autonomic network (e.g., hypothalamus, anterior cingulate cortex, amygdala, insula) involved in regulating the balance of sympathetic vs parasympathetic activity⁶⁵ also play important roles in regulating emotion.^30,66–69 Several of these regions are also susceptible to α-synuclein accumulation in PD.^14,15,70 Additionally, dysregulation of neurotransmitter systems such as serotonin, noradrenaline, and adrenaline has been found to correlate with both mood and autonomic symptoms.^36–40 Our findings may also be influenced by heterogeneity in PD clinical presentations and rates of progression. For example, research by Fereshtehnejad and colleagues (2017) using the PPMI database identified a “diffuse malignant” phenotype of PD which included a subset of patients (12% of sample) who had more severe motor and non-motor features (e.g., cognitive deficits, autonomic symptoms, anxiety/depression, rapid eye movement sleep behavior disorder).⁷¹ This subgroup had greater rates of progression over time, more pronounced dopaminergic deficits, and greater atrophy in PD-specific brain networks (e.g., substantia nigra). As such, the findings of the present study may, in part, be influenced by individuals within this subgroup who experience more severe disease burden/progression and a wider range of non-motor symptoms. It is also possible that worsening autonomic symptoms may be associated with increasing anxiety/depression due to a negative emotional reaction, particularly given prior findings linking worsening of autonomic symptoms with decline in quality of life and activities of daily living.¹⁸ Individuals with anxiety/depressive symptoms may also self-perceive their autonomic symptoms as being more burdensome or severe.^41,42 Regardless of the mechanism, the present study raises an important question about potential clinical interventions—whether treating autonomic symptoms (particularly gastrointestinal, urinary, thermoregulatory, and cardiovascular dysfunction) leads to improvements in mood, or at the very least a slower rate of worsening of mood symptoms.

There are a few limitations of the current study. Notably, autonomic functioning was measured using a self-report measure, rather than physiological probes (e.g., heart rate variability). While this allowed individuals to report frequency of symptoms they experienced across a variety of domains, we were not able to confirm findings using objective measures of autonomic functioning. Prior studies using objective physiological measures of autonomic functioning to examine relationships with mood have had mixed findings in PD. For example, in a sample of levodopa-naïve patients, Park and colleagues (2016) monitored blood pressure for a 24-hour period and found that depression was associated with greater measured drop in blood pressure during orthostasis and other measures of baroreflex-cardio-vagal functioning. Similarly, in a sample of patients with Lewy body disease (the majority of whom had idiopathic PD), Del Pino et al. (2020) found that depression was associated with certain cardiovascular autonomic parameters such as reduced heart rate variability during deep breathing. In contrast, other studies have failed to identify relationships between cardiovascular probes and depression in PD.^72,73 Further research using objective autonomic measures is needed, including in other autonomic domains (e.g., gastrointestinal functioning). Additionally, while the SCOPA-AUT is a validated and common tool for measuring autonomic symptoms in PD,^48–50 certain domains may be overrepresented. For example, there are 7 items regarding gastrointestinal symptoms, while there is only 1 item for pupillomotor symptoms. Moreover, some items from this scale could be potentially related to primary motor dysfunction rather than autonomic-related causes (e.g., difficulty swallowing, difficulty passing urine). Another limitation is that our study focused on the contributions of autonomic symptoms to mood, but did examine contributions from other factors which may also influence mood in PD (e.g., cognitive impairment, pain, fatigue). Lastly, our sample was limited in terms of both the range of demographic characteristics (majority were highly educated and White) and distribution of mood symptoms (most participants did not report clinically-significant mood symptoms), which may limit generalizability of the results to other populations.

In conclusion, our study found that autonomic symptoms are strongly related to mood symptoms, explaining over 1/4^th of the variance in both anxiety and depression after a 5-year follow-up in patients with PD who were levodopa-naïve at time of enrollment. Gastrointestinal symptoms in particular seem to be most closely related to both anxiety and depression, though limitations exist with the self-reported measure used for measuring autonomic functioning in the present study. Finally, higher overall autonomic load, and not motor severity, was associated with greater increases in both depression and anxiety over time. Future studies should incorporate objective measures of autonomic dysfunction, as well as examine potential mechanisms underlying the association between autonomic and mood symptoms. This could inform future treatment interventions for non-motor conditions in PD (e.g., therapeutic targets for both the treatment of anxiety/depression and autonomic insufficiency).