Small (autonomic) and large fiber neuropathy in Parkinson disease and parkinsonism

All consecutive patients with the diagnosis of Parkinson’s disease or parkinsonism (secondary to Willson disease N = 4, progressive supranuclear palsy/corticobasal degeneration N = 3, multiple system atrophy N = 3, Lewy body disease N = 1, vascular parkinsonism N = 1 and unestablished etiology N = 4) seen at a newly established tertiary outpatient clinic at the Federal University of Ceará were offered to participate in this study. Patients were enrolled over a period of 1 year and one additional year was necessary to complete the diagnostic testing and laboratory work-up. Very few patients declined to participate in the study (the authors estimate that less than 10 % of the eligible patients) and no patient withdrew from the study after initial enrollment.

The study consisted in a standard evaluation for the presence of large and small (autonomic) fiber peripheral neuropathy. Fifty-four patients were enrolled in approximately 1 year and completed the neuromuscular evaluation. A board-certified neurologist (by both the American Academy of Neurology and Academia Brasileira de Neurologia) conducted the initial PN screening. It consisted of history gathering (e.g. complaints of paresthesias, burning feet, unexplained worsening gait) and detailed neurological exam looking for classicial findings of neuropathy (predominantly distal atrophy, decreased or absent deep tendon reflexes, length-dependent sensory loss). Thereafter, a group of medical students supervised by the corresponding author conducted the skin wrinkling test (SWT), according to the same protocol described by Teoh et al [13] that we had also previously employed [14]. Briefly, the SWT consisted of hand immersion in NaCl at 0.5 mmol/l at 40.5^oC during 30 min. Thereafter, skin wrinkling was visually graded in 4 digits (all but the thumb) and a mean score of wrinkling was created. Patients were considered to have neuropathy if mean SWT score was lower than 2. If a patient had signs or symptoms of peripheral neuropathy and/or an abnormal SWT, this initial evaluation was followed by a nerve conduction study/electromyography (NCS/EMG). The final result of the neuromuscular evaluation was based on the combined results of the neurological exam, SWT and NCS/EMG. For the lower extremities, the following cut-offs were used to diagnose abnormalities in the nerve conduction studies: peroneal compound muscle action potential (CMAP) amplitudes (2.5 mV), tibial CMAP amplitudes (3.5 mV), sural sensory nerve action potential (SNAP) amplitudes (4 uV) and tibial, peroneal and sural nerve conduction velocities (39 m/s). Absent sural responses were not considered to be necessarily abnormal in patients older than 60 years.

The medical charts of all patients were also reviewed, including the risk factors for neuropathy. Basic laboratory tests were requested to pursue the work-up for neuropathy and/or clinical evaluation for the presence of thyroid diseases, diabetes mellitus (DM) or vitamin B12 deficiency. Although most of the patients routinely had prior evaluation for diabetes and thyroid disease by primary care physicians (although not all test results were available), screening for B12 deficiency could only be documented in 63.2 % of the PD and 56.3 % of the parkinsonism group.

This study was approved by the local Institutional Review Board of the Universidade Federal do Ceará (Comitê de Ética em Pesquisa do Hospital Universitário Walter Cantídio). All patients signed an informed consent prior to the enrollment in this study. Descriptive statistics followed by Chi-square and Fischer exact test and parametric (t test) and non-parametric tests (Mann-Whitney test) were then performed (choice according to Gaussian or non-Gaussian distribution). Thereafter, univariate regression analysis was performed to establish whether each variable was a risk factor for neuropathy. One-Way ANOVA followed by the Holm-Sidak test was used to compare the results of the electrodiagnostic tests. Results are detailed as mean ± SEM and differences were considered significant if P < 0.05.

Table 1 compares the demographic characteristics of the 38 patients with PD and 16 patients with other forms of parkinsonism that completed our work-up. Patients with PD were older than patients with parkinsonism (P < 0.05, Mann Whitney test). However, age of onset, disease duration and gender distribution were similar in both groups, despite a trend for higher percentage of women in the parkinsonism group and for older age of onset and longer disease course in the PD group. Mean Hoehn and Yahr scores in the PD group were 2.6 ± 0.1 (score of 2 indicates bilateral involvement with preservation of balance, while score of 3 means bilateral involvement with mild-moderate impairment of postural reflexes). Most of the PD patients were treated with levodopa and 68.8 % of the parkinsonism group were taking levodopa. A third of the PD patients and 31 % of the parkinsonism group were treated with pramipexol while 38.9 % of the PD and none from the parkinsonism group were treated with amantadine and only 8.3 and 18.8 (respectively) were treated with biperiden.

Skin Wrinkling Test (SWT) was performed in 49 patients, 33 with PD and 15 with other forms of parkinsonism. No statistical difference was observed in the percentage of patients with abnormal SWT between the 2 groups: 19 out of 33 of the PD group (57.6 %, mean SWT grade: 1.98 ± 0.2) and 6 out of 15 of the patients with parkinsonism (40 %, mean SWT grade: 2.3 ± 0.3). Among the patients with evidence of small (autonomic) fiber neuropathy in the PD group, 10 % had diabetes, 5 % thyroid disease, 5 % B12 deficiency and 5 % had been treated for leprosy in the past while additional 15 % had B12 levels <300. Among the patients with evidence of small fiber neuropathy in the parkinsonism group, 12.5 % had diabetes mellitus. Therefore, in 65 % of the PD patients and among 87.5 % of the patients with parkinsonism, no common etiology of peripheral neuropathy was identified.

NCS/EMG was performed in 39 patients (26 PD and 13 with other forms of parkinsonism). 12 out of the 26 PD (34.2 % of all PD) and 4 out of the 13 from the parkinsonism group (23.5 % of all patients from the parkinsonism group) had abnormal NCS/EMG results. As expected, patients with abnormalities suggestive of peripheral neuropathy on the screening neurological exam were more likely to have abnormal EMG (P = 0.04, Fischer exact test). Neuropathy prevalence was similar in the groups with PD and parkinsonism (P > 0.05), whether PN was assessed by SWT, NCS/EMG or clinically (final combined neuromuscular evaluation).

Table 1 displays the results of the most important neurophysiological parameters from the lower extremities nerve conduction studies in patients with PD and other forms of parkinsonism. No statistically significant difference was evidenced between PD and other forms of parkinsonism for the tibial and peroneal motor amplitudes and conduction velocities as well as from sural sensory amplitudes and conduction velocities. Figures 1 and 2 further detail the electrodiagnostic findings in patients with PD and parkinsonism with and without large and small fiber neuropathy. As can be seen in Fig. 1a, patients with large-fiber neuropathy and PD had lower peroneal CMAP amplitudes than patients without large-fiber neuropathy (P < 0.01). Figures 1a and 2a also reveal that despite a trend for lower sural SNAP amplitudes in patients with PD and large-fiber neuropathy, there was no statistically significant difference in sural SNAP amplitudes in patients with PD or parkinsonism with and without large-fiber neuropathy. Also, as can be seen in Figs. 1b and 2b, peroneal and sural conduction velocities in patients with PD or parkinsonism were not significantly different in patients with small or large-fiber neuropathy.

Patients with EMG abnormalities were older: 65.5 ± 3.8 vs. 53 ± 3 years (P = 0.01, Mann-Whitney test), but abnormalities on SWT scores were not associated with age. In addition, patients with either SWT or EMG abnormalities were not more likely to be older than normal patients. Patients with EMG abnormalities were more likely to be older at the disease onset (57.2 ± 3.8 versus 45.3 ± 3.1 years, P = 0.01) but there was no significant relationship between EMG abnormalities and disease duration. In addition, there wasn’t any significant relationship between the age of disease onset or disease duration and SWT scores (P = 0.44).

Univariate regression analysis showed a significant correlation between age and EMG abnormalities (P = 0.01) but not with SWT scores (P = 0.53). Univariate regression analysis also showed a significant correlation between age of onset and EMG abnormalities (P = 0.03). Patients with higher Hoehn and Yahr stage scores were more likely to have EMG abnormalities (Fischer exact test, P = 0.02) than patients with lower scores, but no significant difference was found between SWT and Hoehn and Yahr stage scores. Patients taking biperiden were more likely to have EMG abnormalities (P = 0.05, Fischer’s exact test). Patients with neuropathy were not more likely to have diabetes mellitus, hypothyroidism, B12 deficiency, use levodopa, pramipexole or amantadine. Almost all patients from the PD group were taking levodopa but patients with PD were more likely to take levodopa than patients with parkinsonism (P = 0.003). However, the total amount of levodopa exposure on each patient was not quantified, and therefore, the cumulative effect of levodopa on the neuropathy development could not be assessed.

Regression analysis of tibial and peroneal CMAP amplitudes and conduction velocities or sural SNAP and conduction velocities versus the presence of DM or B12 levels was not significant. In addition, regression analysis also did not disclose any significant relationship between SWT scores and electrodiagnostic parameters, i.e. tibial and peroneal CMAP amplitudes and conduction velocities or between sural SNAP and conduction velocities.