The vagus nerve may play a crucial role in the pathogenesis of Parkinson's disease (PD). It has been hypothesized that α-synuclein aggregates form in the enteric nervous system, and spread via the autonomic nervous system to the central nervous system (1). In this context, vagal parasympathetic neurons that project to most of the gastrointestinal tract serve as entrance to the dorsal motor nucleus of the vagus (DMV). In support, truncal vagotomy decrease the risk of developing PD (2, 3). The DMV and the vagus nerve itself are among the most frequently affected structures in postmortem PD studies (4–7). This is accompanied by up to 50% DMV motor neuron loss in deceased patients (8–10).

Using ¹¹C-donepezil PET/CT, we have previously demonstrated decreased gastrointestinal acetylcholinesterase density in patients with PD (11–13). This is believed to reflect parasympathetic denervation, mainly derived from the vagus nerve. Furthermore, subjects with isolated REM-sleep behavior disorder, the strongest marker of prodromal PD (14), also exhibit decreased gastrointestinal acetylcholinesterase density (15). Hence, the vagal parasympathetic system is affected years before diagnosis in some patients. Therefore, precise evaluation of vagal degeneration could serve as tool to diagnose PD patients, even during the premotor phase.

High-resolution ultrasound is a non-invasive, radiation-free alternative to assess vagal degeneration. The method can precisely measure the caliber of superficial nerves. Six previous studies tested the hypothesis that vagal neuron loss is detectable by ultrasound, but reported conflicting results. Four studies found significant atrophy of both left and right vagus nerves in patients with PD (16–19). In contrast, three studies reported no difference between patients with PD and healthy controls (HCs) (20–22). Thus, whether ultrasound can detect vagal neuronal loss in patients with PD remains unresolved.

In the present study we determined the vagus cross-sectional area (CSA) in the largest cohort of PD patients studied so far and compared to HCs. Five images of each nerve were obtained to optimize accuracy, and all analyses were performed offline for optimal precision. We compared the vagus CSA with measures of gastrointestinal acetylcholinesterase density (parasympathetic innervation) and retained radiopaque markers (estimate of gastrointestinal transit time) in subgroups. These measures should theoretically be altered in parallel to vagal parasympathetic depletion.

Mean vagus nerve CSAs are presented in Table 1, and depicted graphically in Figure 2. There was no significant difference in the crude comparison between patients with PD and HCs. We observed a tendency toward higher CSAs in males than females, most pronounced in the right vagus (Table 1). Thus, we used multiple linear regression to determine sex-adjusted CSA of the vagus. First, we applied a model allowing different slopes in each group. We found no significant slope interactions (P > 0.5). Thus, we applied a model with one common slope. Sex-adjusted difference in mean left vagus CSA was 0.12 mm² (95% CI: −0.06 to 0.29 mm², P = 0.19) and right vagus was 0.21 mm² (95% CI: 0.01–0.41 mm², P = 0.041). In this model, sex was a significant predictor of right vagus CSA (P = 0.003), but not of the left vagus CSA (P = 0.06). No difference in symmetry index was observed (P = 0.69).

In the HC group, mean right vagus CSA was 0.4 mm² larger than mean left vagus CSA (95% CI: 0.24–0.57 mm², P < 0.0001). In the PD group, mean right vagus CSA was 0.34 mm² larger than left vagus CSA (95% CI: 0.20–0.48 mm², P < 0.0001). Left and right vagus CSA did not correlate with age, BMI, time since diagnosis, disease severity (H&Y), ROM (estimate of colonic transit time), nor colon acetylcholinesterase density measured by ¹¹C-donepezil PET (P > 0.05).

Mean coefficient of variation of vagus CSAs was 6.7%.

In the present study, we found no difference in vagus nerve CSA between patients with PD and HCs, but female vagus CSA was smaller than male vagus CSA in both groups. One previous report found the opposite pattern; vagus CSA in females were larger than males (23). More females were included in our HC group compared to the PD group so we determined the sex-adjusted differences between patients with PD and HCs. We found a statistically significant reduction of the right vagus CSA (P = 0.041) but not of the left vagus CSA (P = 0.19).

Although the mean right vagus CSA was significantly reduced in PD, the effects size of ~10% is small. The effect size in other studies, which reported statistically significant CSA reduction, range from ~10–30% (16–18). Neuronal loss in DMV has been reported to be up to 50% in postmortem studies (8–10). Other nuclei related to the vagus nerve (nucleus ambiguus, solitary nucleus, and spinal trigeminal nucleus) are less affected by α-synuclein aggregates (24). Also, no motor neurons are lost in nucleus ambiguus (9, 25). The axonal diameter of somato-motor and somato-sensory myelinated fibers is generally much larger than that of autonomic neurons, and must necessarily occupy a substantial fraction of the midcervical vagus CSA. Cardiac sympathetic neurons also reside in the vagus nerve and occupy up to 5% of the CSA (26, 27). Cardiac sympathetic neurons are particularly vulnerable in PD (28) and may also contribute to slight vagal atrophy. However, if vagal atrophy is mainly caused by sympathetic and parasympathetic (DMV) axonal loss, the reduction in CSA is probably very small and could explain why three studies failed to show vagal atrophy in PD (20–22). Additionally, the vagus nerve shows frequent branching (up to 29%) (29). Branched nerves have smaller CSA at the midcervical level (29). Although we have no reasons to believe that branching should differ between patients with PD and HC, it may have contributed to increased data variance.

Time since the diagnosis could be another factor explaining some incongruence between studies. In the present study, the median duration was short (8 months), as many of the patients were newly diagnosed. The four previous studies that demonstrated significant CSA reduction included patients 5–10 years post-diagnosis (16–19) (Table 2). Thus, the rather small difference demonstrated in the present study might be partly caused by the shorter duration of motor symptoms. However, we found no correlation between vagus CSA and time since diagnosis in the present data, and the three studies that failed to show a significant difference also investigated moderately affected patients, 4–8 years postdiagnosis (20–22) (Table 2).

If a biological difference in right vagus CSA truly exists between patients with PD and HCs, two aspects may have enabled us to identify this small difference in the present study: (1) our two- to three-times larger PD cohort than previous studies and (2) offline measurements as advised for optimal precision (30). However, our study does not support that this method can be applied for meaningful interpretation at the individual level, because the average between-group difference of 0.21 mm² is hardly discerned, given that the normal vagal CSA ranges from 1 mm² to nearly 4 mm² (Figure 2).

Right vagus CSA was significantly larger than left vagus CSA as previously reported (16–18, 20, 22, 30). This also corresponds to the side-to-side difference of the cervical vagus neuronal area defined on postmortem (31). Further, the right midcervical vagus contains ~20% more nerve fibers than the left (32). We believe this side difference may be the main reason for the larger CSA reduction of the right vagus, that is, the potential absolute reduction must increase with larger nerve CSA. The mean CSA values were in quite close agreement with six previous studies that estimated vagus CSA by tracing the epineural rim, as in the present study (16, 18, 20, 22, 23, 30). Thus, in studies using the tracing method, a reasonable interstudy reproducibility of the vagus CSA seems to exist.

Vagus CSA did not correlate with the number of retained ROM, nor with colon acetylcholinesterase density measured by ¹¹C-donepezil PET. If vagus CSA atrophy occurs in parallel to parasympathetic denervation, we would expect these parameters to be correlated. One explanation could be that ROM and colon acetylcholinesterase density reflect global colonic function/innervation, and vagal parasympathetic projections end close to the left colonic flexure. However, α-synuclein aggregates are also found in spinal parasympathetic neurons that innervate the descending and sigmoid colon and the rectum (33). Therefore, it is likely that these parasympathetic neurons degenerate in parallel to those in the vagus, and the correlation between vagus CSA and these global colonic measurements should still be apparent. Thus, we argue that the most likely cause for the lack of correlation is the inability of the ultrasound method to detect vagal parasympathetic loss.

The precision of the three middle values was evaluated for each nerve by calculating the coefficient of variation. The average coefficient of variation was acceptable (6.7%). We also calculated the average distance from the three middle values to the mean. The average distance was 4.9% of the mean, for example, values defining a CSA of 2 mm² on average were 0.1 mm² from the “exact” value. Therefore, we argue that obtaining five ultrasound images of each nerve and using the three middle values to determine the CSA yields precise results.

This study has some limitations. First, the ultrasound investigator was not blinded to clinical status during examination. However, the offline CSA estimation was performed blinded. Second, our two groups were matched on age but not sex, forcing us to do linear adjustment in the comparison analyses.

In conclusion, we found a small but significant reduction in sex-adjusted right vagus CSA in patients with PD. No difference was found in left vagus CSA. Sex was a significant predictor for the right but not left vagus caliber. This study does not support the use of vagus ultrasound as a diagnostic marker for PD, but there may be applications for the method in a research setting. Other methods to assess vagal atrophy, including MR neurography, should be explored in the future.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

The studies involving human participants were reviewed and approved by Science Ethical Committee of Central Denmark Region. The patients/participants provided their written informed consent to participate in this study.

JH: acquisition, analysis, interpretation of the data, and drafted the manuscript. TF: analysis and interpretation of the data and critical revision the manuscript. KA, CS, KK, NO, PvW-M, SD, and BB: acquisition of the data and critical revision the manuscript. UW: design and concept of the study, analysis and interpretation of data, and critical revision the manuscript. PB: design and concept of the study, interpretation of data, and critical revision the manuscript. All authors have approved the final manuscript and agree to be accountable for the content of the work and made substantial contribution to the following parts of the study.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.