Oral, pharyngeal, and esophageal swallowing is a sequential event that transports saliva, ingested solids, and fluid from the mouth to the stomach and protects the airways during swallowing [1, 2]. Pharyngeal function involves numerous interacting control mechanisms that ultimately link pharyngeal contraction patterns to the adjacent oral cavity and esophagus. From a functional point of view, the pharynx can be divided into two parts: the muscle and nerves. Contraction of swallowing muscles occurs as a result of depolarization after acetylcholine release at the endplates. However, contraction of muscles can also occur after direct electrical stimulation. Neuromuscular electric stimulation (NMES) is a special form of neuromuscular stimulation used to re-educate patients to use their pharyngeal muscles in the throat for patterned activity to initiate or re-establish swallowing. NMES is a dual-channel electrotherapy system designed specifically for the treatment of pharyngeal dysfunction from any etiology. NMES electrodes are placed at one or several sites on the throat for stimulation of the underlying musculature. Prior studies have shown the beneficial effect of NMES on the symptoms of dysphagia [3, 4]. Studies have also shown that oral and pharyngeal stimulation may cause reorganization of the human adult motor cortex [5–11]. However, no prior study has addressed objectively how pharyngeal function is affected by NMES.

The purpose of this study was to evaluate and compare the outcome of NMES versus traditional swallowing therapy (TT) in stroke patients using diverse measurements: videoradiographic swallowing evaluation, nutritional status, oral motor function test, and a visual analog scale (VAS) for the patient’s subjective self-evaluation of complaints.

Materials

Subjects

Three different European swallowing centers participated in this randomized trial: Malmö University Hospital, Malmö, Sweden (where the principal investigators are located); University Hospital Maastricht, Maastricht, Netherlands; and University Hospital Toulouse, Toulouse, France. Twenty-five patients (16 men and 9 women) were included in the trial. Twelve patients were randomized for NMES and 13 for TT. Mean age was 70 years for the NMES group and 71 years for the TT group. Four patients in the NMES group were on tube feeding and 7 patients in the TT group.

Inclusion criteria for the trial were as follows: (1) patients had to be 50–80 years old and had cerebrovascular disease (stroke) more than 3 months prior to the study. (2) Patients with hemispheric stroke and without neurologic signs typical for brainstem involvement. (3) Patients could not have nasogastric tubes but could have PEG. (4) Patients had to be able to elicit some pharyngeal swallow as revealed during videoradiographic swallowing evaluation and seen as anterior movement of the hyoid bone and constrictor activity in the pharyngeal wall. (5) They had to be able to communicate. Computed tomography (CT) or magnetic resonance (MR) findings were not used for inclusion or exclusion. A MR examination was not done. The disease process should have been stable. The stroke could have been a first-time stroke or a recurrence. Patients could be on a modified diet and may practice swallowing maneuvers.

Exclusion criteria for the trial were as follows: (1) patients with progressive cerebrovascular disease; (2) patients with other neurologic diseases such as ALS, MS, or Parkinson’s disease; (3) patients with tumors or neoplastic disease of the swallowing apparatus and who had undergone radiotherapy to the neck; (4) patients who had undergone surgery to the swallowing apparatus; (5) patients who were not able to elicit pharyngeal swallow (as described above); (6) patients with a nasogastric tube.

The trial was approved by the ethical committees in Sweden, the Netherlands, and France.

Methods

Randomization

A randomization table was designed and used as an online randomization tool, assigning patients to treatment A or B with a chance of 50% to be assigned to either treatment. This table was used in the online database for the trial in which each patient had a preassigned treatment number that appeared once the eligibility criteria were checked and basic patient information was entered.

Evaluation Measurements

Pre- and Post-treatment Evaluation

Clinical evaluation

A qualitative clinical swallowing assessment was performed and included a visual analog scale (VAS) for the patient’s subjective self-evaluation of complaints (see below), actual nutrition status (Table 1), and oral motor status (Table 2). Also, a quantitative measurement was used, namely, videoradiographic swallowing evaluation (Table 3). The post-treatment evaluation should be performed the same day after the last treatment.

Table 1 Scores for Actual Nutrition Status (ANS)
Full size table
Table 2 Oral motor function test (OMFT)
Full size table
Table 3 Scores for Videofluoroscopic Examination
Full size table
Visual analog scale (VAS)

The VAS was used to compare differences between pre- and post-treatment. The number of points before and after treatment was also compared; 0 points = no difficulties at all and 10 points = maximum/unable to swallow. A comparison was made between the patient’s subjective feeling (i.e., the VAS) and the objective pharyngeal function (videoradiography). A comparison was also made with the clinical evaluation.

Radiologic Evaluation: Videofluoroscopy

Endpoints

The radiologic evaluation/examination was documented on video or in a digital modality. The frame rate (or equivalent) was 25 frames/s or more. Radiologic evaluation was done within one week before entering the study. The examination covered the oral cavity, including tongue, pharynx, and pharyngoesophageal segment (PES). Patients were examined while they were in a sitting position and thin and thickened liquids were used. The following recipe was used: Thin liquids: 40% w/v barium suspension. Thickened liquid was prepared according to the following recipe: 110 g Fortifresh (Nutricia) raspberry. Add 40 g barium powder. Mix well. Then add 4 g Thicken Up (Resource, Novartis Consumer Health UK, Ltd). Mix it well. Use at room temperature within 10 min.

The patient was given 5-ml boluses. The patient performed five swallows of 5 ml thin liquid and five swallows of 5 ml thick liquid. Three swallows were also performed in an AP view (thin liquid).

To measure distances, a coin (5 eurocent) with a known diameter was placed in the midline of the anterior neck when the patient was imaged in lateral projection. This coin was included in the X-ray field. The assessment focused on (1) dissociation between oral and pharyngeal stage, (2) misdirected swallowing, (3) pharyngeal retention, and (4) width of PES (Table 3).

Dissociation between oral and pharyngeal stage was measured as the start of the anterior movement of the hyoid bone. This correlated with the position of the apex of the bolus. In normal patients the apex of the bolus should not pass beyond the faucial isthmus by more than 0.5 s before the hyoid starts to move anteriorly. This timing was given in milliseconds (Table 3). Leakage was also registered.

Misdirected swallowing was classified as either barium reaching into the laryngeal vestibule or beyond the vocal folds into the trachea (Table 3).

Retention was classified as either minor or major. In the AP or the lateral view, the height of the air fluid level was measured and compared with the 5-eurocent coin (Table 3).

The maximum width of the PES was measured in either the AP or the lateral view (Table 3).

The radiologist was blinded for whether the videoradiographic examination was performed before or after treatment.

Intervention-Procedure

NMES and TT were administered by a speech-language pathologist trained in dysphagia management. The SLPs involved in this study have all attended the VitalStim Certification Program and completed all requirements for certification as a VitalStim™ Therapy Provider [12].

The patients were randomized for either NMES or the control group (TT). Patients referred to the NMES (VitalStim) group were given the following treatment: (1) Placement 3 B; two sets of electrodes were used and located on each side of the midline of the throat. Two electrodes were placed just at or above the level of the thyroid notch over the thyrohyoid muscle. The typical level ranged from 4.5 to 25 mA with a mean level of around 13 mA. (2) One 60-min session every day. (3) Patients were treated 5 days a week over 3 weeks, i.e., 15 treatment sessions. (4) NMES patients could have their diets modified if they were modified when the patient was included in the study. (5) Only spontaneous maneuvers were allowed. No other maneuvers or treatments were added or included. (6) The SLP instructed the patient to swallow hard and fast.

The control group (TT) underwent the following: (1) Diet modification if applicable. (2) A clinician determined appropriate maneuvers or other treatment techniques. (3) Analogous to the NMES group, they were treated for 60 min 5 days a week over 3 weeks, i.e., 15 sessions. If the patient could not participate for 60 min, he/she was told to do a training session on his/her own at home. (4) They received exercise sheets for the specific exercises that they had to perform, equal for all participating clinics.

Statistical Analyses

Combined group differences between post-therapy and baseline data (effect data) were tested for significance by a Mann-Whitney test for each parameter of the videoradiography of swallowing, the Actual Nutrition Scale (ANS), and the total score of the Oral Motor Function Test (OMFT). The data of the VAS scale were sufficiently normally distributed to allow an independent-samples t test. Differences in effect data between the NMES group and the TT group were tested for significance by a Wilcoxon signed rank test, except for the VAS scale for which data distribution allowed the use of a paired-samples t test. The relationship between the baseline data and the therapy effects on all evaluation tools was studied by nonparametric Spearman’s correlation coefficients. A Bonferroni correction has been applied (p < 0.05/18). All computations were performed using SPSS v12.0.1 (SPSS Inc., Chicago, IL).

Results

Effects of Therapy

Visual Analog Scale (VAS)

Table 4 presents the median and the 25th and 75th percentiles of the patient’s self-evaluation of dysphagia. Using a paired-samples t test, statistically significant therapy effects were found for the total group of patients. No statistically significant differences were found between NMES and TT (independent-samples t test).

Table 4 Visual analog scale (VAS), comparisons pre- and post-treatment
Full size table

Actual Nutrition Scale (ANS)

Table 5 contains descriptive statistics of the pretherapy data and the effect data for the ANS. A Wiloxon signed rank test was used to test for significant changes between baseline and post-therapy data. When combining both groups, a statistically significant group change after therapy could be demonstrated. No statistically significant differences were found between NMES and TT (Mann-Whitney test).

Table 5 Actual nutrition scale (ANS), comparisons pre- and post-treatment
Full size table

Oral Motor Function Test (OMFT)

Table 6 shows descriptive baseline and effect data for each item. The results of a Wilcoxon signed rank test indicate that positive significant therapy effects are present for the total score of the OMFT. No statistically significant differences were found between NMES and TT (Mann-Whitney test).

Table 6 Oral motor function test (OMFT), comparisons pre- and post-treatment
Full size table

Videoradiographic Evaluation of Swallowing

Table 7 shows the results of Wilcoxon signed rank tests per variable. In contrast to the results of the other evaluation tools, no statistically significant therapy effects are found for any of the variables of videoradiography. In particular, the width of the PES and the amount of retention were unchanged. No statistically significant differences were found between NMES and TT (Mann-Whitney test).

Table 7 Videofluoroscopy (VFS), comparisons pre- and post-treatment
Full size table

Relationship Among Evaluation Tools

The relationship among the evaluation parameters was studied by nonparametric Spearman’s correlation coefficients (Table 8). The correlations between the baseline data were low. The mean correlation was −0.06 (range = −0.30 < R < 0.16). The correlations between the effect data were low also, except for the correlation between the effect data of the total score of the videoradiography and the patient’s self-evaluation (R = 0.72). The mean correlation of the effect data was 0.10 (range = −0.12 < R < 0.72). These poor correlations between the baseline data and between the effect data imply that the outcome of the diverse evaluation tools represents independent aspects of swallowing.

Table 8 Correlation between measurement instruments
Full size table

Discussion

Neuromuscular electrical stimulation (NMES) is widely used in the United States but not known in Europe until lately. The scientific evidence of its efficacy is unclear and further research is needed [13]. The goal of this randomized trial was to compare differences in outcome between NMES (VitalStim™) and traditional swallowing therapy, and also to compare differences between oral and pharyngeal function and the patient’s symptoms. Our results show that there were no differences in outcome between NMES and TT, which is in accordance with another newly published study on this topic [14]. Ludlow et al. [15] concluded that surface electrical stimulation in some dysphagic patients could interfere with hyolaryngeal elevation which is required for airway protection during swallowing. Shaw et al. [16] found that VitalStim therapy helped patients with mild to moderate dysphagia. Therefore, they concluded that VitalStim therapy has a place in the management of dysphagia but that the most severely injured patients did not improve. Referring to Logemann [17]: “We still need much more research to determine whether VitalStim or other methods of neuromuscular electrical stimulation have any role to play in the management of oro-pharyngeal swallowing difficulties.”

Subjects

The experiences from all three centers involved in this study revealed great difficulties in identifying patients who met the inclusion criteria. Several potential subjects had a spontaneous recovery within the inclusion time of 3 months or were so severely injured that it was impossible for them to participate in the trial. Some of them could not initiate a pharyngeal swallow. Therefore, we had to close the study after 2 years.

One of the inclusion criteria was hemispheric stroke, while patients with brainstem lesions were excluded. The rationale for this was that the NMES is supposed to stimulate efferent nerves in the neck. In patients with brainstem lesions, the lower motor neuron is supposed to be nonfunctioning or degenerated. Therefore, we assumed that this type of patient would not be suitable for NMES.

The patients for this trial were not stratified by hemispheric lesions, severity of CVA, or time post-onset. The participants were specifically not identified by lesion or severity of CVA because this was a preliminary study with small numbers. To stratify by lesion or isolate just one particular type of lesion would have lengthened participant enrolment past what was of interest for this study. Larger, more controlled studies have to stratify their participants according to these variables. Future studies may also stratify individuals with CVA by lesion location, time post-onset, and severity to determine the effectiveness of NMES as a function of CVA-related variables as well.

Evaluation Measurements

VAS

The outcome from the VAS showed that baseline data of the NMES group were more severe than those of the TT group. It must be considered an accident that the subjects in the NMES group seemed to be more severely impaired by their swallowing disorders than the subjects in the TT group.

Actual Nutrition

When combining both groups, statistically significant group changes in actual nutrition status after therapy were demonstrated. Also, an improvement in nutrition for both groups combined was found. We may conclude that swallowing treatment will improve the awareness of how to eat and drink. From earlier studies we have learned that the traditional swallowing techniques and diet modification could decrease misdirected swallowing to the trachea and pharyngeal retention [18].

Oral Motor Function

Positive significant therapy effects were present for the total score of the OMFT. Oral motor range-of-motion exercises in some cases are an option for dysphagic patients [19], and in this study the individually recommended oral motor exercises seem to have improved the actual swallowing dysfunction.

Videofluoroscopy

In contrast to the results of the other evaluation tools, no statistically significant therapy effects were found for any of the variables of videoradiography. In particular, the width of the PES and the amount of retention were rather unchanging variables in this study. However, clinically we often experience obvious differences in follow-up patients. Based on the literature, videoradiography is considered to be the most objective and reliable measurement in the evaluation of swallowing [20]. Perhaps when using larger numbers of subjects or defining parameters used for the evaluation of videoradiography that are robust and more sensitive to changes resulting from therapy, statistically significant therapy effects might be found in future studies.

Correlation Between Measurement Instruments

Statistically significant positive therapy effects for both groups (NMES and TT) combined were found but there was no statistically significant difference in therapy effect between the groups. The correlations between the different measurements were low. Therefore, the patient’s subjective experience of improvement did not correlate with the objective videoradiography. A false-positive experience of improvement may lead to serious swallowing complications such as choking attacks. At least two of the patients included in this study (randomized for NMES) were a couple of months after they had completed the trial treated for severe aspiration pneumonia. Because they had a very positive post-treatment VAS score and felt that they had recovered from their swallowing difficulties, they did not follow the recommendations regarding diet modification. Elmståhl et al. [21] concluded that “assessment of subjective swallowing complaints is not sufficient to evaluate the clinical course.” A quality of life measurement such as SWAL-QOL would probably have been a better tool to estimate how the patient’s experienced quality of life after swallowing treatment.

Conclusion

In this study no statistically significant differences were found in therapy effect between the NMES group and the TT group. The correlations between the different measurements were low. The subjective improvement (VAS) experienced by several patients did not correlate with the objective videoradiographic outcome. We still need to know more before we can claim that NMES is valuable as a treatment option for dysphagic patients.