Transcutaneous vagal nerve stimulation for treating gastrointestinal symptoms in individuals with diabetes: a randomised, double-blind, sham-controlled, multicentre trial

Abstract

Aims/hypothesis: Diabetic gastroenteropathy frequently causes debilitating gastrointestinal symptoms. Previous uncontrolled studies have shown that transcutaneous vagal nerve stimulation (tVNS) may improve gastrointestinal symptoms. To investigate the effect of cervical tVNS in individuals with diabetes suffering from autonomic neuropathy and gastrointestinal symptoms, we conducted a randomised, sham-controlled, double-blind (participants and investigators were blinded to the allocated treatment) study.

Methods: This study included adults (aged 20-86) with type 1 or 2 diabetes, gastrointestinal symptoms and autonomic neuropathy recruited from three Steno Diabetes Centres in Denmark. Participants were randomly allocated 1:1 to receive active or sham stimulation. Active cervical tVNS or sham stimulation was self-administered over two successive study periods: 1 week of four daily stimulations and 8 weeks of two daily stimulations. The primary outcome measures were gastrointestinal symptom changes as measured using the gastroparesis cardinal symptom index (GCSI) and the gastrointestinal symptom rating scale (GSRS). Secondary outcomes included gastrointestinal transit times and cardiovascular autonomic function.

Results: Sixty-eight participants were randomised to the active group, while 77 were randomised to the sham group. Sixty-three in the active and 68 in the sham group remained for analysis in study period 1, while 62 in each group were analysed in study period 2. In study period 1, active and sham tVNS resulted in similar symptom reductions (GCSI: -0.26 ± 0.64 vs -0.17 ± 0.62, p=0.44; GSRS: -0.35 ± 0.62 vs -0.32 ± 0.59, p=0.77; mean ± SD). In study period 2, active stimulation also caused a mean symptom decrease that was comparable to that observed after sham stimulation (GCSI: -0.47 ± 0.78 vs -0.33 ± 0.75, p=0.34; GSRS: -0.46 ± 0.90 vs -0.35 ± 0.79, p=0.50). Gastric emptying time was increased in the active group compared with sham (23 min vs -19 min, p=0.04). Segmental intestinal transit times and cardiovascular autonomic measurements did not differ between treatment groups (all p>0.05). The tVNS was well-tolerated.

Conclusions/interpretation: Cervical tVNS, compared with sham stimulation, does not improve gastrointestinal symptoms among individuals with diabetes and autonomic neuropathy.

Trial registration: ClinicalTrials.gov NCT04143269 FUNDING: The study was funded by the Novo Nordisk Foundation (grant number NNF180C0052045).

Keywords: Autonomic neuropathy; Diabetic gastroenteropathy; Gastrointestinal dysmotility; Gastrointestinal symptoms; Gastroparesis; Heart rate variability; Vagal nerve stimulation.

Fig. 1

Flowchart detailing screening, randomisation and participation in both study periods

Fig. 2

Compliance profiles (a, c) and stimulation intensity (b, d) for study period 1 (four daily stimulations for 1 week) (a, b) and study period 2 (two daily stimulations for 8 weeks) (c, d). Data are mean values with 95% CI

Fig. 3

Mean values for the GCSI (a, c) and the GSRS (b, d) visualised before and after study period 1 (a, b) and study period 2 (c, d). The thicker lines represent the mean values for each treatment arm

Fig. 4

The mean GCSI (a) and GSRS (b) for each week in study period 2. Data are mean values with 95% CI

Fig. 5

Median gastric (a), small-bowel (b), colonic (c) and whole-gut (d) transit times before and after the 8 weeks of two daily stimulations in study period 2. The thicker lines represent the median values for each treatment arm