This randomized controlled trial with blinded evaluators showed that oral neuromuscular training improved swallowing function in the long term among people with post-stroke dysphagia. While there were no significant between-group differences at the end of treatment in swallowing rate, lip force, or aspiration-penetration on VFS, the intervention group showed a significant improvement in swallowing rate and lip force at 12-month follow-up compared with the control group. These findings indicate the good effects of oral neuromuscular training as a treatment option for post-stroke dysphagia. There is, however, a need for larger randomized controlled trials to confirm our preliminary, positive long-term results.
Both the intervention and control groups showed significant within-group improvements in swallowing rate after the 5-week training period, but there was no significant between-group difference. The non-significant between-group difference may be due to several factors. One is that both groups received active rehabilitation (toothbrush stimulation) during the treatment period and with positive outcome, indicating that both treatments had an immediate effect on swallowing. However, only the neuromuscular training was found to have a long-term effect. Another factor was the small sample size, indicating that there could be a lack of sufficient statistical power to generalise the conclusions of the findings. Yet another factor could be the spontaneous remission of swallowing dysfunction that is usually observed within 2–3 weeks after the stroke incident [15, 16]. In this study the participants were recruited 4 weeks after the stroke event, making it difficult to determine whether all the improvement in the swallowing rate was due to the therapeutic intervention or, to some extent, to spontaneous recovery.
However, the lasting improvement in swallowing rate at 12 months post-treatment in the intervention group compared with the control group indicates that oral neuromuscular training may be more effective in triggering the reorganisation of the brain  than orofacial sensory-vibration stimulation alone. In contrast to orofacial sensory-vibration stimulation, oral neuromuscular training with an oral device aims to both stimulate sensory input and strengthen the facial, oral, and pharyngeal muscles .
Exercised-based approaches to swallowing rehabilitation have been shown to not only improve swallowing function in people with dysphagia, but to improve oral diet and reduce aspiration and dependence on tube feeding [18–21]. Furthermore, exercises to strengthen the swallowing musculature improve swallowing and reduce dysphagia-related comorbidities [18, 20]. Thus, combined sensorimotor exercise (i.e., oral neuromuscular training) might be a more beneficial swallowing rehabilitation method in the long term than sensory stimulation alone.
Improvement in lip force was observed in the intervention group both after 5 weeks of oral neuromuscular training and at the 12-month follow-up, whereas no significant improvement was observed in the control group at any time. A significant between-group difference was observed only at the 12-month follow-up. These findings, combined with the improvements in swallowing rate, seem to support the finding of previous studies [11, 22] in which oral neuromuscular training improved both swallowing rate and lip force.
When used to identify swallowing dysfunction, results from TWST proved to have excellent correlation with findings from VFS, in accordance with a previous study . All participants showed dysfunction, either with penetration/aspiration of contrast medium entering the airway and causing unsafe swallowing at baseline, or with premature spillage and/or pharyngeal residue revealing ineffective pharyngeal swallowing.
The present study is the first randomized controlled trial of oral neuromuscular training with an oral device as post-stroke dysphagia treatment and in which blinded assessors were used at the end of treatment and at 12-month follow-up. However, it does have some methodological issues. A common challenge with clinical trials, including this study, is to recruit enough participants to obtain acceptable statistical power at follow-up, after drop-out. The number of participants included at baseline almost reached the number targeted by the power analyses, but many participants were lost at follow-up due to mortality, PEG, recurrent stroke, or because they declined continued participation (see Fig. 3). The randomization to either the intervention or control group was based on the TWST results and not on the VFS findings. Since the randomization procedure did not take into account important prognostic factors at baseline, an uneven distribution of aspiration at baseline (33% in the intervention group and 57% among controls) might have contributed to the lack of significant outcomes on differences in improvement according to VFS, since it might have been easier to improve from aspiration at baseline in the control group. Thus, an objective finding on the severity of swallowing dysfunction may be an important prognostic factor in the participants’ ability to improve the impaired function, and should be considered at randomisation.