Introduction

Parkinson’s disease (PD) is commonly defined as a neurodegenerative movement disorder characterized by resting tremor, rigidity and bradykinesia. Patients affected by PD can present different subtypes according to the predominant symptoms (tremor or rigidity/bradykinesia) with clear clinical proof that these different subtypes can have a different clinical course (Jankovic et al. 1990). Slower progression of the disease and less cognitive decline are observed in tremor-dominant patients. These clinical observations have been confirmed by neuroimaging and clinic–pathologic findings describing a more favorable outcome in tremor-dominant patients than in akinetic-rigid subjects, in concordance with different brain biochemical abnormalities (Rossi et al. 2010; Eggers et al. 2011; Lewis et al. 2011). Moreover, PD is often hampered by the appearance of non-motor symptoms (NMS), which in recent years are gaining increasing attention (Langston 2006; Chaudhuri et al. 2006). Among the NMS, autonomic cardiovascular disorders have been associated with significant morbidity and mortality and with variable manifestations in patients affected by PD (Goldstein 2003), although the definite correlations with different subtypes of PD are not clear. In this context, analysis of heart rate variability (HRV) represents a not invasive and established tool in assessing cardiovascular autonomic modulation in different extrapyramidal disorders, especially in PD patients (Friedrich et al. 2008; Valappil et al. 2010; Maetzler et al. 2014).

In this context, it was decided to study cardiovascular autonomic modulation in PD patients with tremor dominant subtype in comparison to akinetic rigid dominant subtype, and compare them with normal subjects using HRV analysis with the aim to investigate the biological relevance of clinical subtypes and possible different pathophysiological process between these subtypes.

Methods

PD patients were selected among individuals attending the Movement Disorders Centre of University of Cagliari, Italy. Inclusion criteria were a diagnosis of probable Parkinson’s disease, defined according to the United Kingdom (UK) brain bank criteria (Hughes et al. 1992) and done by one neurologist specialized in movement disorders, age 40–80 years, and stable medication use.

Patients with moderate-to-severe dementia with the mini-mental state examination (MMSE) (Folstein et al. 1975) of <24 or with incapacity to give an informed consent were excluded from the study. In addition, any patient with atypical Parkinson’s syndromes such as multiple system atrophy (MSA), vascular parkinsonism, drug-induced parkinsonism, and progressive supranuclear palsy (PSP) was excluded.

Twenty-eight consecutive PD patients (17 with tremor dominant subtype and 11 with akinetic rigid dominant subtype) were selected and compared. These two groups were also compared to 17 age and sex-matched healthy controls. Patients with medical disorders that could alter autonomic function (e.g. diabetes, hematological disorders, collagen diseases, malignancies, and neuropathy) were excluded.

Anticholinergic anti-Parkinson’s drugs were withdrawn whenever possible, or reduced to the lowest dosage that allowed appropriate cooperation during the test. In this regard, biperidene, an anticholinergic drug used to decrease tremor in PD patients, was withdrawn in a single patient enrolled in the dominant group. No other patients were taking anticholinergic drugs.

Whenever possible, cardiac drugs that could alter HRV results (i.e. beta-receptor blockers, I f-current blockers, calcium-channel blockers, mineral corticoids) were discontinued or reduced to the lowest dosage that allowed appropriate cooperation during the test. If therapeutic modification was clinically impossible patients were excluded. Medications such as ACE-inhibitors, angiotensin receptor blockers and diuretics were allowed. Cardiac drugs that could alter HRV results were discontinued only in one patient of PD patients with tremor dominant subtype who was taking propranolol. This drug was gradually discontinued before HRV evaluation.

Patients underwent complete neurological examination, including disease duration (time passed from diagnosis), measures of disease severity such as Hoehn and Yahr stage (Hoehn and Yahr 1967), the Unified Parkinson’s Disease Rating Scale (UPDRS) pars-III score (Goetz 2003), and total levodopa equivalent daily dose (LEDD) was also recorded.

LEDD was calculated, as previously described by Evans et al. (2004), with the following formula: levodopa dose + levodopa dose × 1/3 if on entacapone + bromocriptine (mg) × 10 + cabergoline or pramipexole (mg) × 67 + ropinirole or rotigotine (mg) × 20 + pergolide (mg) × 100 + apomorphine (mg) × 8.

The scales were rated during an “off” phase (12 h off drugs). We defined the clinical subgroups tremor-dominant and akinetic-rigid for each patient in a manner similar to Lewis et al. (2005). The tremor score was obtained from the sum of UPDRS items 20 and 21 divided by 7. The non-tremor score was obtained from the sum of UPDRS items 18, 19, 22, 27, 28, 29, 30, and 31, divided by 12. PD patients were classified as tremor-dominant if the tremor score was at least twice the non-tremor score, as akinetic-rigid patients if the non-tremor score was at least twice the tremor score. ECG 24-h recordings were performed in all subjects at the same time with a digital three-channel recorder (Del Mar-Aria recorder; Del Mar Medical, Irvine, CA). During the ECG recording all patients were free to carry out daily tasks to the extent allowed by hospital regimen. All RR interval time series were automatically edited first. Manual editing was not performed due to the insignificant presence of artifacts and premature beats. For final analysis, 24-h HR-variability data were divided into segments of 3600 ms, and only segments with >85 % sinus beats were included in analysis. Quantitative time domain analysis was performed on normal–normal (NN or RR) interval, and the following data were computed: mean, SDNN (standard deviation of NN intervals), SDNN index (mean of the standard deviations of NN intervals), SDANN (standard deviation of the averaged NN intervals), SDSD (standard deviation of successive NN differences), RMSSD (square root of the mean squared difference of the consecutive NN intervals), pNN50 (proportion of NN intervals differing more than 50 ms to the total number of NN intervals), and TINN (the integral of the density distribution of the number of all NN intervals plotted in a histogram divided by the maximum of the density distribution). An autoregressive model was used to estimate the power-spectrum densities of HR variability (Kay and Marple 1981). Power spectra were quantified by measuring the area in three frequency bands: 0.005–0.04 Hz (very low frequency, VLF), 0.04–0.15 Hz (low frequency, LF), and 0.15–0.4 Hz (high frequency, HF).

Spectral power of HRV was reported in normalized units (representation of the relative value of each power component in proportion to the total power minus the VLF component).

ECG data were transferred for the analysis of HR variability on Impresario Holter Analysis System (Del Mar Reynolds Medical INC, Irvine, CA).

Finally, PD patients were compared to the age and sex-matched healthy controls. Written informed consent was obtained from all participants before screening, consistent with the Declaration of Helsinki.

Statistical analysis

Categorical data are presented as percentages, and quantitative data as mean ± SD. Differences between groups were calculated using a one-way analysis of variance (ANOVA) for repeated measures, and categorical variables were compared with the Fisher’s exact test. Subsequently, values were normalized for age at onset, HY and UPDRS. A two-tailed value of p < 0.05 was considered statistically significant.

Results

Twenty-eight PD patients (17 with tremor dominant subtype and 11 with akinetic rigid dominant subtype) were selected and compared to 17 age and sex-matched healthy controls. Demographic characteristics of patients (subdivided in tremor dominant and akinetic-rigid subtypes) and health subjects involved in the study are summarized in Table 1. The mean age was 65.5 ± 6.5 in PD patients with akinetic rigid dominant subtype, 63.4 ± 8.6 in PD patients with tremor dominant subtype, 65.2 ± 9.8 in control subjects. Both, the control and the PD groups had similar height and weight characteristics. Other clinical characteristics were not significantly different between PD patients groups and healthy control, with the exception of motor disability scores between PD patients. Indeed, HY resulted significantly higher in akinetic-rigid PD patients compared to tremor dominant subtype: 2.7 ± 0.8 vs 2.1 ± 0.4 (p < 0.03), and UPDRS-III was also significantly greater in akinetic-rigid PD patients compared to tremor dominant subtype: 27.5 ± 14.5 vs 18.9 ± 4.4 (p < 0.03).

Table 1 Demographic characteristics of patients (subdivided in tremor dominant and akinetic-rigid subtypes) and healthy subjects involved in the study
Full size table

Among time and geometric domains parameters, summarized in Table 2, SDSD was significantly enhanced in PD patients compared to controls (p = 0.04), while SDNN index and HRV were significantly decreased in PD patients compared to controls (p = 0.03).

Table 2 Heart rate variability measures: time and geometric domains
Full size table

No differences in these parameters were found between akinetic-rigid PD patients compared to tremor dominant subtype.

Among frequency domain parameters, summarized in Table 3, both PD groups and healthy control subjects showed significant differences in LF, LF/HF ratio and HF. In particular, LF values were lower both in akinetic-rigid PD patients (p < 0.0001) and tremor dominant subtype PD patients (p = 0.043) compared to controls, HF values were higher both in akinetic-rigid PD patients (p = 0.006) and tremor dominant subtype PD patients (p = 0.041) compared to controls, while LF/HF ratio were lower both in akinetic-rigid PD patients (p = 0.007) and tremor dominant subtype PD patients (p = 0.018) compared to controls.

Table 3 Heart rate variability measures: frequency domain
Full size table

Moreover, it was observed that LF values were significantly lower in the akinetic rigid dominant subtype than in the tremor dominant group (LF 41.4 ± 13.6 vs 55.5 ± 11.6; p = 0.007), while the LF/HF ratios tended to be lower in akinetic rigid dominant subtype, although no statistical differences were found. To exclude a possible bias due to HY, UPDRS and age at observation (although the last difference was not significant), low-frequency values were normalized for these variable. Normalized LF were significantly lower in the akinetic rigid dominant subtype than in the tremor dominant group: normalized LF for HY were 16.2 ± 7.6 vs 26.9 ± 7.0, p = 0.001; normalized LF for UPDRS were 2.0 ± 1.4 vs 3.1 ± 0.9; p = 0.021; normalized LF for age at observation were 0.6 ± 0.3 vs 0.9 ± 0.3; p = 0.023. Thus, normalization of LF value for these variables did not change the significant difference between tremor dominant subtype and akinetic rigid dominant subtype subjects.

Discussion

In PD patients two different clinical phenotypes are commonly described, namely a tremor dominant subtype and an akinetic-rigid subtype. The first subtype of patients is usually characterized by slower disease progression, a lower degree of cognitive impairment and minor incidence of neuropsychiatric disorders compared with the latter group (Lewis et al. 2005; Oh et al. 2009).

These data reflect pathological differences where tremor dominant subtype patients presented with more severe impairment of the middle part of the substantia nigra pars compacta and lesion of retrorubral area A8, while patients with akinetic-rigid subtype presented with a major degeneration in the ventrolateral part of the substantia nigra pars compacta (Jellinger 1999).

However, neurodegeneration in PD not only involves brain areas, but can also be present externally of central nervous system with the occurrence of severe non-motor features (Langston 2006; Chaudhuri et al. 2006). Among them, symptoms of cardiovascular dysautonomia are a common occurrence in Parkinson’s disease, often occurring prior to motor impairment (Mathias 1998; Goldstein 2003). Previously, SPECT studies with 123I-metaiodobenzylguanidine (MIBG) indicated a cardio-selective, postganglionic denervation in PD patients (Taki et al. 2004; Mitsui et al. 2006).

This data explicates how the damage to the postganglionic sympathetic efferences in PD patients may be the main cause of dysautonomia (Ziemssen and Reichmann 2010).

HRV parameters reflect cardio-vagal modulation, sympathetic excitation (Pagani et al. 1997) and sympathovagal balance (Eckberg 1997) and can be quantified and displayed using spectral analysis (Cadeddu et al. 2010). Among the power spectrum frequency bands, the HF band is thought to reflect the parasympathetic activity mediated by the vagus nerve, whereas the LF band is thought to reflect the baroreflex modulation of the autonomic outflow mediated by both sympathetic and parasympathetic systems (Goldstein et al. 2011).

Given that, this study shows LF values significantly lower in the akinetic rigid dominant subtype compared to tremor dominant control group, indicating that the disease led to a more evident impairment of the baroreflex modulation of the autonomic outflow mediated by both sympathetic and parasympathetic systems in the first class of patients. Thus, there was a significant alteration in autonomic parameters in PD patients with akinetic rigid dominant compared to tremor dominant subtype.

Moreover, using time domain analysis of HRV, we found that despite the normal mean resting heart rate (as measured by RR interval) in PD subjects, there was a notable reduction in HRV and SDNN (estimate of overall HRV) when compared with normal controls. Total HRV and SDNN index correlate to the total power of the analysis when applied on a 24-h measurement (Camm et al. 1996) underlining the value of the reduction of the LF components of the power spectrum. On the other hand, the increase of SDSD, which correlates to the HF components of the power spectrum when applied on a 24-h measurement (Camm et al. 1996), confirms the imbalance of the autonomic components evidenced with the power spectrum analysis.

These findings support the biological relevance of clinical subtypes backing the idea of a different pathophysiological process between these subtypes. These differences are not only important for a better classification between PD patients, but also suggest that different subtypes may also result in different responses to therapy or in the possible development of cardiovascular side effects of dopaminergic drugs in these different populations.

This study also confirms previous results on significant abnormalities in sympathetic function of PD patients compared to controls and suggests the presence of an enhanced autonomic dysfunction in PD patients, especially in patients with akinetic-rigid PD subtypes. The presence of cardiovascular autonomic dysfunction might represent an important cause of comorbidity in these specific forms of PD, with possible effects on cardiovascular symptoms.