Botulinum Toxin Treatment for Oropharyngeal Dysphagia Associated With Diabetic Neuropathy

Sixty-eight type 2 diabetic patients (43 insulin dependent and 25 non–insulin dependent) with diabetes-associated peripheral somatic and/or autonomic neuropathy underwent neurological examination for the presence of neurogenic dysphagia. Fourteen of 68 (20.5%) patients complained of dysphagia for both solid and liquid foods. Twelve of them (8 men, 4 women, age range 59–68 years, 9 insulin dependent and 3 non–insulin dependent) with severe dysphagia (mean duration 4.8 ± 1.0 years) for both solid and liquid foods agreed to undergo treatment with BoNT/A. The mean duration of diabetes was 8.3 ± 5.8 years (range 10.7–23.10). All patients were examined by the same neurologist (D.A.R.) and underwent brain magnetic resonance to exclude brain or brainsteam involvement. Clinical data are reported in Table 1. The diagnosis of diabetic neuropathy was based on clinical and neurophysiological criteria. Somatic neuropathy was ascertained by either the modified neuropathy disability score (0–2, no neuropathy; 3–5, mild neuropathy; 6–8, moderate neuropathy; and 9–10, severe neuropathy) (15) or by the vibration perception threshold, which was assessed by a neurotensiometer at the great toe of the dominant foot. Autonomic neuropathy was diagnosed by calculation of the mean expiratory/inspiratory heart rate ratio during deep breathing. All patients gave their written informed consent to the treatment. The study was conducted in accordance with Declaration of Helsinki and approved by the local ethics committee.

Swallowing function was evaluated clinically and by both videofluoroscopy and simultaneous needle electromyography (EMG) of the CP and pharyngeal inferior constrictor (IC) muscles. Clinical evaluation was carried out every other day for the 1st week and every other week thereafter until week 24. Videofluoroscopic and electromyographic examinations were performed before and at week 1, 4, 12, 16, 18, and 24 after BoNT/A injection.

Based on clinical evaluation, four levels of dysphagia were considered: mild dysphagia (dysphagia for liquids), discrete dysphagia (dysphagia for solids or for liquids and solids), severe dysphagia (frequent tracheal aspiration and chocking), and very severe dysphagia (patient fed by nasogastric tube or percutaneous endoscopic gastrostomy. Level zero indicated the absence of dysphagia.

Videofluoroscopy with modified barium swallowing was performed. Three different food consistencies of standardized bolus size were used: thin liquid (equivalent to milk), semisolid (equivalent to jelly), and solid (dry toast coated in barium).

Simultaneous EMG of the CP and IC muscles was performed using concentric needle electrodes according to the protocol proposed by Elidan et al. (16). Briefly, for CP recording, the needle was inserted 1.5 cm lateral to the palpable border of the cricoid cartilage in a postero-medial direction. In normal subjects, a high-frequency tonic electromyographic activity is observed as the needle penetrates the muscle and a 300- to 600-ms pause is seen during voluntary swallowing. For IC muscle recordings, the needle is inserted ∼3 cm above the point of insertion for the CP muscle, lateral to the lateral border of the thyroid cartilage. In normal subjects, no EMG activity is present at rest; whereas a tonic EMG activity is observed during voluntary swallowing (Fig. 1). Therefore, voluntary swallowing is electromyographycally characterized by a burst of activation in the IC muscle and a synchronous EMG pause in the CP muscle, corresponding to UES relaxation.

BoNT/A (Dysport, Speywood, Portons Downs, U.K.) was percutaneously injected into the CP muscle under EMG control using a teflon-coated needle. For each patient, 30 units of botulinum toxin (2 ml dilution, 0.9% saline) were injected in each side of the CP muscle.

Data were evaluated by ANOVA. The level of significance was set at P < 0.05. Likert’s transformation was applied to change the qualitative results of the dysphagia severity score into a score ranging from zero (no dysphagia) to four (very severe dysphagia).

In all patients, brain magnetic resonance ruled out brain and brainsteam ischemic lesions. Electrophysiologic and clinical data revealed the presence of autonomic neuropathy in all patients (heart rate ratio 1.1 ± 0.1, range 0.9–1.1, normal values >1.20), whereas somatic peripheral neuropathy was detected in nine patients (Table 1). Mean neuropathy disability score was 6.6 ± 3.6, range 1–10 (mild neuropathy in one patient, moderate neuropathy in two patients, and severe neuropathy in six patients). Mean vibratory perception threshold was 30.5 ± 4.8 V (range 24–39, normal values <25 V) (Table 1). Before BoNT/A treatment, one patient had very severe dysphagia and all the others had severe dysphagia.

In all patients, videofluroscopy showed the reduction of pharyngeal clearance and the incomplete CP opening. EMG showed a remarkable decrease of the maximal IC muscle activation with an increased discharge frequency and an excessive CP tonic hyperactivity leading to reduced or no relaxation (Fig. 2). Moreover, EMG of the IC muscle disclosed fibrillation potentials and/or motor unit action potential abnormalities in all subjects, reflecting active and chronic denervation.

After a single BoNT/A injection in each side of the CP muscle, dysphagia completely disappeared in 10 patients (severity score zero) and significantly improved in the remaining 2 patients (severity score “mild dysphagia”). Mean dysphagia severity score significantly ameliorated compared with the pretreatment value (score before BoNT/A, 3.1 ± 0.2; week 1, 0.2 ± 0.3; P = 0.0001) and all patients were able to voluntarily swallow. Figure 3 shows the mean and SD changes in dysphagia severity score before and at 1, 4, 12, 16, 18, and 24 weeks after botulinum toxin injection. Videofluoroscopic and electromyographic studies confirmed normal swallowing and correct IC/CP muscle coordination. The beneficial effect of BoNT/A was evident after 4 ± 1.1 days (range 3–7). At week 1, the improvement was evident in all patients and remained unchanged up to week 12. It then decreased but remained significant (P = 0.04) at weeks 16 and 18. The mean effect duration was 14.6 ± 2.1 weeks (range 12–18). At week 24, all patients complained of the reoccurrence of dysphagia, the severity of which did not differ from the pretreatment value (mean dysphagia severity score 2.9 ± 0.5, P = 0.18). All patients underwent a further BoNT/A treatment at week 24 and experienced the same improvement observed at the first injection. Therefore, treatment was repeated every 3–4 months according to the response of each patient. Overall, all patients remained able to voluntary swallow for the whole period of treatment. Treatment with BoNT/A was well tolerated, and no relevant local or systemic side effects were reported. Only four patients complained of slight pain in the injection site. No adverse side effects were observed during and after the further follow-up treatments.

In diabetic patients, oropharyngeal dysphagia associated with autonomic and/or somatic peripheral neuropathy is likely more frequent than previously presumed, and its pathophysiology has not been completely elucidated (3,4,6). Although all phases of swallowing can be involved, the oropharyngeal phase is the most frequently impaired phase in diabetic patients (1,3,4,6). Control of oropharyngeal swallowing is mediated by voluntary and involuntary (reflexive) mechanisms. Although voluntary swallowing is not an autonomic function, it is allowed by the coordination of two muscles: the IC muscle, which controls the voluntary component, and the CP muscle of the UES, which controls the involuntary (reflexive) component. During voluntary swallowing, the activation of the IC muscle is synchronous with the relaxation of CP muscle, thus allowing the bolus to transit into the upper esophageal tract. In patients with dysphagia due to diabetic neuropathy, the coordination between the two muscular components (CP and IC) is impaired. This can be assessed by dynamic EMG analysis. In all our diabetic patients, the IC muscle showed neuropathic changes and reduced activity during swallowing. The denervation of the IC muscle induces the reduction of its voluntary activation with consequent impairment to the physiological balance between these two components. In fact, the reduction of the voluntary activation (contraction) may lead to a tonic hyperactivation of the CP muscle that cannot be inhibited and causes a defective control of voluntary and reflexive mechanisms during voluntary swallowing. A long-lasting block of the parasympathetic fibers innervating the CP muscle might be useful for treating UES hyperactivity.

Percutaneous injection of BoNT/A has been already used to successfully treat CP muscle hyperactivity associated with various neurological diseases (10–14). Our study demonstrated that BoNT/A is a safe and very effective treatment also for dysphagia associated with diabetic neuropathy.

The advantage of BoNT/A treatment is that it can be performed in an outpatient clinic, needing neither hospitalization nor anesthesia. It can be repeated when the symptoms reappear, retains the same efficacy, and requires no specific follow-up. However, this treatment may have potential risks. The diffusion of BoNT/A into the nearby laryngeal muscles might lead to laryngeal spasm or to worsening of the preexisting dysphagia. For this reason, the treatment must be carried out under electromyographic guidance by an expert operator in order to identify the target muscle and to rule out the possible diffusion of the botulinum toxin into the nearby laryngeal muscles.