Exploring cardiac autonomic dysfunction in synucleinopathies with parkinsonism through awake–sleep stages: a possible connection to cardiovascular death

Abstract

Introduction

Alpha-synucleinopathies (ASs), including Parkinson’s disease, multiple system atrophy and dementia with Lewy bodies, are characterised by autonomic dysfunction that may predispose to cardiovascular events and sudden death. Heart rate variability (HRV) is a non-invasive marker of autonomic regulation, yet its modulation across sleep stages in ASs remains unclear.

Methods

We retrospectively analysed 25 patients with ASs and 35 age-matched controls who underwent overnight polysomnography between 2020 and 2023. 5 min ECG segments from wakefulness, non-rapid eye movement (NREM) and rapid eye movement (REM) sleep were used to compute time- and frequency-domain HRV indices. Statistical comparisons and multivariate models were adjusted for sex and REM sleep behaviour disorder.

Results

HRV indices, particularly percentage of successive RR intervals that differ by more than 50 milliseconds (pNN50) and root mean square of successive differences, were significantly reduced in ASs compared with controls. The most pronounced difference occurred during NREM sleep (pNN50, p=0.008), where controls exhibited normal parasympathetic elevation that was absent in ASs.

Conclusion

Patients with ASs demonstrate blunted nocturnal parasympathetic activity, most evident during NREM sleep. Sleep-stage-specific HRV analysis provides sensitive insights into cardiac autonomic dysfunction and may serve as a potential biomarker of disease severity and cardiovascular risk in synucleinopathies.

Introduction

Alpha-synucleinopathies (ASs)—including idiopathic Parkinson’s disease (IPD), dementia with Lewy bodies (DLB) and multiple system atrophy (MSA)—are neurodegenerative disorders marked by misfolded alpha-synuclein (aSyn) aggregates in the central and peripheral nervous systems, leading to progressive neuronal loss. The global prevalence of IPD is approximately 1.5 per 1000 individuals, showing a male predominance.¹ In contrast, DLB accounts for ~5% of dementia cases, and MSA represents a rarer but more aggressive parkinsonian disorder with severe autonomic failure.

The Rochester Epidemiology Project reported a shortened median survival in ASs—about 2 years less than controls—with the highest mortality HR observed in MSA (HR=10.51; 95% CI 2.92 to 37.82).² Deaths were most often attributed to neurodegenerative progression (31.5%), cardiovascular events (15.8%) or pneumonia (15.4%).

A particularly concerning outcome in IPD is Sudden Unexpected Death in Parkinson’s Disease (SUDPAR)—an unexplained fatal event occurring without identifiable cause, even on autopsy. Up to 14% of PD-related deaths may qualify as SUDPAR.³ Proposed mechanisms include cardiac and autonomic abnormalities. Cardiac post-mortem studies revealed aSyn deposition within local cardiac autonomic nerves in 82% of AS patients versus 0% of controls (p<0.001), suggesting that peripheral autonomic nerve involvement may impair cardiac regulation and predispose to arrhythmia or sudden cardiac death.⁴

HRV—the fluctuation in successive heartbeat intervals—is a non-invasive marker of cardiac autonomic control. Time-domain indices such as the SD of normal-to-normal intervals (SDNN), root mean square of successive differences (RMSSD) and percentage of successive RR intervals that differ by more than 50 milliseconds (pNN50) reflect overall and parasympathetic activity, while frequency-domain indices including low-frequency (LF) power, high-frequency (HF) power and ratio of LF to HF (LF/HF) assess sympathetic–parasympathetic balance. HRV follows a circadian rhythm, increasing at night under parasympathetic predominance and decreasing by day when sympathetic tone dominates. During non-rapid eye movement (NREM) sleep, vagal activity and HRV rise, whereas rapid eye movement (REM) sleep enhances sympathetic tone, reducing HRV.⁵ Impaired nocturnal HRV has been linked to sudden nocturnal death post-myocardial infarction⁶ and elevated stroke risk.⁷

In ASs, autonomic dysfunction affects both sympathetic and parasympathetic systems, with patterns varying by subtype. IPD often shows early sympathetic deficits, whereas MSA features early, severe dysautonomia including orthostatic hypotension and supine hypertension.⁸ Both conditions commonly exhibit reduced nocturnal heart rate dipping, indicating diminished vagal tone.

Although HRV reductions are recognised in ASs, sleep-stage-specific modulation is less understood. Because parasympathetic activity peaks during NREM sleep, sleep-based HRV may serve as a sensitive marker of autonomic dysfunction. This study retrospectively compared HRV across wake, NREM and REM sleep between AS patients with parkinsonism and age-matched controls, hypothesising that blunted nocturnal HRV, particularly in parasympathetic indices, reflects altered circadian autonomic modulation and elevated cardiovascular risk.

Methods

Study design and participants

This retrospective cross-sectional study included consecutive patients who underwent full-night polysomnography (PSG) at the Neurological Institute of Thailand (August 2020–April 2023; IRB 66031). The AS group comprised patients meeting diagnostic criteria for IPD,⁹ either established or probable MSA,¹⁰ or probable DLB.¹¹ Although the diagnostic criteria for MSA and DLB do not require the presence of parkinsonism, only patients who exhibited parkinsonism were included in our study. Controls were age-matched adults (55–70 years) without parkinsonism who underwent PSG during the same period.

PSG acquisition and ECG selection

All participants underwent attended PSG (Nihon Koden) following American Academy of Sleep Medicine Guidelines standards.¹² Standard montage included electroencephalogram, electrooculography, electromyography, ECG, airflow, respiratory effort, oximetry and body position sensors. Sleep staging followed AASM 2012 criteria¹³ and was reviewed by a certified sleep physician (TS).

For HRV analysis, the first 5 min artefact-free ECG segment was manually selected from wakefulness, NREM (stage N2) and REM sleep by TS, excluding epochs with movement, arousals or respiratory events per European Society of Cardiology and The North American Society of Pacing Electrophysiology 1996 HRV guidelines.¹⁴

HRV analysis

Analysis used Python with MNE-Python for signal handling, HeartPy for R-peak detection and Pandas for data management.

• Time-domain indices: SDNN, RMSSD and pNN50.

• Frequency-domain indices: LF (0.04–0.15 Hz), HF (0.15–0.4 Hz) and LF/HF ratio. RMSSD and pNN50 represented parasympathetic activity; LF/HF reflected sympathovagal balance.

Data collection

Demographic and clinical variables (age, sex, body mass index, comorbidities, medications, PSG findings) were extracted. For ASs, additional data included disease duration, age at onset, presenting symptoms, presence of REM sleep behaviour disorder (RBD), cognition, mood, orthostatic hypotension and modified Rankin Scale (mRS) scores. Clinical information, including motor and non-motor features, was obtained through detailed history taking and physical examination during routine evaluation. Sleep-related symptoms were assessed using standardised screening tools: the REM Sleep Behaviour Disorder Single-Question Screen for RBD, the STOP-BANG questionnaire for obstructive sleep apnoea and the Epworth Sleepiness Scale for excessive daytime sleepiness. Cognitive function was evaluated using either the Thai version of the Montreal Cognitive Assessment or the Thai Mental State Examination. These data were recorded and presented in online supplemental table 2). HRV metrics for wake, NREM and REM were entered into Epidata V.3.1.

Statistical analysis

Analyses were conducted using STATA V.18. Continuous variables were summarised as mean±SD or median (IQR); categorical variables as counts (%). Between-group comparisons used t-test/Mann-Whitney U and χ²/Fisher’s exact tests.

Missing HRV data>5% were addressed using multiple imputation (10 datasets) via regression-based prediction models. Linear or quantile regression tested associations between AS diagnosis and HRV across stages, adjusting for sex and RBD. Significance was set at p<0.05.

Results

Participant characteristics

60 participants were included: 25 with ASs (13 IPD, 10 MSA, 2 DLB) and 35 controls. The mean ages did not differ significantly between groups (67.4±7.9 vs 63.1±10.4 years; p=0.09). However, the AS group had a higher proportion of males (76% vs 45.7%; p=0.019). Comorbidities and general medication profiles were comparable between groups; however, antiparkinsonian medications were present only in the AS group (online supplemental table 1).

RBD occurred more frequently in ASs (32%) than controls (8.6%) (p=0.039). Periodic limb movement and obstructive sleep apnoea rates were comparable (online supplemental table 1). Among ASs, mean age at onset was 61.9±9.2 years; presenting symptoms included bradykinesia (52%), tremor (48%) and gait disturbance (44%). Sleep complaints were common in the AS group (88%), and among those with sleep symptoms, RBD was the most frequent (72.7%). Cognitive impairment was present in 36% of AS patients, orthostatic hypotension in 28% and mood symptoms in 32%. The median levodopa equivalent daily dose was 695 mg/day (IQR 557.5). Over half (52%) had mRS=3, reflecting moderate disability (online supplemental table 2).

HRV across wake and sleep stages

Across all states, HRV indices—RMSSD, pNN50, LF and HF—were lower in ASs versus controls, with the greatest reduction during NREM sleep (table 1).

Table 1. Heart rate variability measures across wakefulness, NREM and REM sleep in alpha-synucleinopathy patients and controls with corresponding regression analysis p values.

HRV parameters	Synucleinopathies with parkinsonism (n=25)	Controls (n=35)	P value
Time domain analysis
HR (beats/min), mean (SD)	66.56 (12.16)	66.16 (11.60)	0.821
Wakefulness	70.21 (14.91)	69.64 (13.13)	0.875
NREM	63.47 (9.98)	64.48 (11.48)	0.725
REM	66.01 (10.57)	64.36 (9.42)	0.528
RR interval (ms), mean (SD)	930.16 (155.58)	931.39 (144.29)	0.957
Wakefulness	887.96 (167.35)	886.82 (136.65)	0.977
NREM	967.50 (148.93)	956.35 (154.46)	0.781
REM	935.01 (145.01)	951.01 (134.40)	0.662
SDNN (ms), median (IQR)	37.40 (33.47)	39.22 (26.38)	0.685
Wakefulness	42.54 (35.38)	39.91 (26)	0.815
NREM	34.45 (19.64)	36.97 (28.06)	0.727
REM	39.26 (30.98)	39.92 (27.80)	0.938
RMSSD (ms), median (IQR)	24.52 (18.30)	32.14 (32.76)	0.084
Wakefulness	23.92 (23.16)	32.14 (35.96)	0.342
NREM	24.52 (18.30)	38.14 (32.94)	0.066
REM	25.99 (15.01)	23.52 (30.84)	0.801
pNN50 (%), median (IQR)	4.43 (9.9)	9.61 (20.93)	0.044 *
Wakefulness	5.39 (12.09)	9.33 (14.17)	0.325
NREM	3.93 (9.12)	15.99 (20.93)	0.008 *
REM	3.92 (8.12)	6.37 (21.36)	0.585

Frequency domain analysis
LF (ms2), median (IQR)	210.92 (323.98)	250.53 (342.30)	0.467
Wakefulness	278.28 (539.65)	266.27 (257.38)	0.912
NREM	171.66 (264.7)	226.85 (426.37)	0.609
REM	246.52 (302.1)	172.90 (326.09)	0.46
HF (ms2), median (IQR)	206.66 (518.55)	249.60 (530.01)	0.633
Wakefulness	212.91 (499.65)	253.92 (493.05)	0.809
NREM	204.16 (489.76)	348.63 (733.90)	0.422
REM	204.42 (482)	175.76 (545.99)	0.851
LF/HF ratio, median (IQR)	0.95 (1.74)	0.76 (1.61)	0.445
Wakefulness	1.13 (1.57)	0.71 (1.47)	0.395
NREM	0.81 (1.34)	0.46 (1.14)	0.322
REM	0.99 (1.77)	1.35 (1.84)	0.528

Bolded numbers accompanied by asterisks (*) indicate statistical significance (p<0.05).

Bolded numbers without asterisks (*) indicate a trend toward statistical significance.

.HF, high frequency; HR, heart rate; HRV, heart rate variability; LF, low frequency; NREM, non-rapid eye movement; pNN50, the percentage of RR intervals differing by more than 50 ms; REM, rapid eye movement; RMSSD, root mean square of successive differences; SDNN, SD deviation of RR intervals.

• pNN50 showed significant reduction in ASs across all stages (p=0.044), most markedly during NREM (p=0.008).

• RMSSD trended lower (p=0.066).

In controls, RMSSD, pNN50 and HF increased during NREM relative to wakefulness, consistent with normal nocturnal parasympathetic dominance. This physiological elevation was absent in ASs (online supplemental figure 1).

Multivariate regression confirmed lower pNN50 in ASs after adjustment for RBD (p=0.029) and remained marginally significant after adjustment for RBD and gender (p=0.048 for NREM; p=0.088 overall). These findings indicate blunted parasympathetic modulation during NREM sleep in ASs, which remained evident after adjustment for RBD and gender.

Discussion

Patients with ASs showed significant reductions in HRV, particularly in time-domain indices reflecting parasympathetic activity, with pNN50 emerging as the most sensitive measure. The greatest impairment occurred during NREM sleep, indicating disrupted circadian and sleep-stage modulation of cardiac autonomic control. These findings align with previous studies reporting reduced nocturnal HRV in IPD and MSA. Under normal physiology, NREM sleep enhances vagal tone, leading to increased HRV; this expected rise was evident in controls but blunted in AS patients, suggesting parasympathetic dysfunction. This impairment likely reflects degeneration of brainstem autonomic nuclei, including the dorsal motor nucleus of the vagus and nucleus ambiguus—regions implicated early in Braak’s staging of Lewy pathology. Peripheral involvement is also suggested by cardiac aSyn deposition in autonomic nerves.⁴ While overall HRV reductions in ASs have been described, stage-specific sleep modulation remains insufficiently characterised.¹⁵ Our results add new evidence demonstrating that HRV impairment is most pronounced during NREM sleep, a stage typically dominated by parasympathetic activity. This highlights the value of sleep-stage-specific HRV as a sensitive marker of autonomic dysfunction beyond waking or aggregated nocturnal measures.

RBD, common in ASs, has been linked to altered nocturnal autonomic tone. In our study, HRV remained reduced after adjusting for RBD, supporting the interpretation that parasympathetic dysfunction primarily reflects underlying neurodegeneration rather than sleep disturbance. Disease duration also did not influence HRV outcomes; median duration was 4 years (IQR 5), and regression analysis showed no significant association with pNN50 (p=0.482), indicating that HRV reductions were not driven by disease chronicity.

Clinically, HRV—particularly pNN50 during NREM—may serve as a non-invasive biomarker of parasympathetic dysfunction, may help identify individuals at higher cardiovascular or SUDPAR risk^{6 7} and may support autonomic risk stratification. Therapeutically, interventions that enhance vagal tone, including slow breathing, biofeedback or vagus nerve stimulation, may offer benefit, and emerging wearable technologies allow long-term HRV monitoring.

Limitations include modest sample size, short ECG segments, medication effects and limited diabetes data, which prevented detailed assessment of diabetic autonomic neuropathy. Nonetheless, the consistency of findings across analyses supports their robustness.

Conclusions

Patients with ASs show reduced HRV, especially pNN50, across sleep–wake states, with the greatest impairment during NREM sleep. These results highlight parasympathetic dysfunction as a core feature and support sleep-stage HRV as a sensitive biomarker requiring validation in larger longitudinal studies.

Data availability statement

Data are available upon reasonable request.