The effect of 5Hz high-frequency rTMS over contralesional pharyngeal motor cortex in post-stroke oropharyngeal dysphagia: a randomized controlled study

Authors


Address for Correspondence
Jin-Woo Park, Department of Physical Medicine and Rehabilitation, Dongguk University Ilsan Hospital, 27 Dongguk-ro, Ilsandong-gu, Goyang-si, Gyeonggi-do, 410-773, Republic of Korea.
Tel: +(82 31) 961 7484; fax: +(82 31) 961 7488; e-mail: jinwoo.park.md@gmail.com

Abstract

Background  We sought to find the therapeutic effect of 5Hz high-frequency repetitive transcranial magnetic stimulation (rTMS) over the unaffected pharyngeal motor cortex in post-stroke dysphagic patients.

Methods  Eighteen patients with unilateral hemispheric stroke oropharyngeal dysphagia that lasted more than 1 month were randomly divided into two groups. They all performed videofluoroscopic swallowing study (VFSS) before rTMS intervention. The experimental group (EG) received 5Hz rTMS over contra-lesional pharyngeal motor cortex for 10 min per day for 2 weeks. The control group (CG) received sham stimulation under the same condition. Videofluoroscopic swallowing study were performed again just after treatment cessation and 2 weeks afterward. The evaluation was performed using videofluoroscopic dysphagia scale (VDS) and penetration-aspiration scale (PAS).

Key Results  Mean baseline VDS and PAS of EG was 33.6 ± 12.1 and 3.41 ± 2.32 respectively and the scores were reduced to 25.3 ± 9.8 and 1.93 ± 1.52 just after 2 weeks intervention (P < 0.05). This effect lasted for up to 2 weeks after treatment. However, there was no change in the CG. Baseline prevalence of aspiration, pharyngeal residue, delayed triggering of pharyngeal swallowing and abnormal pharyngeal transit time (PTT) in EG was 66.7%, 66.7%, 33.3%, and 44.4%, respectively. After rTMS, the prevalence of aspiration and pharyngeal residue was reduced to 33.3% and 33.3%, respectively. However, the prevalence of delayed triggering and abnormal PTT was not changed.

Conclusions & Inferences  A 5Hz high-frequency rTMS on contra-lesional pharyngeal motor cortex might be beneficial for post-stroke dysphagic patients. This intervention can be used as a new treatment method in post-stroke patients with dysphagia.

Introduction

Oropharyngeal dysphagia is a commonly documented morbidity that follows stroke, but its reported frequency is widely discrepant, ranging between 19% and 81%.1,2 The presence of oropharyngeal dysphagia has been associated with an increased risk for pulmonary complications3 and even mortality.4

The goals in oropharyngeal dysphagia therapy are to reduce the morbidity and mortality associated with chest infections, improve nutritional status, and return patients to a normal diet with a resultant improvement to their quality of life. The therapy can be differentiated into compensatory and rehabilitative strategies; unfortunately, there is a paucity of evidence for oropharyngeal dysphagia therapy.5,6

Recently, a new approach using noninvasive cortical stimulation is being applied to help the neurologic recovery after a stroke such as transcranial magnetic stimulation, which is a well-tolerated procedure that modulates cortical excitability.7,8 A few studies applied repetitive transcranial magnetic stimulation (rTMS) on post-stroke oropharyngeal dysphagia treatment and they reported that rTMS led to a significantly greater improvement in swallowing function.9,10 However, both stimulation frequency (excitatory or inhibitory) and stimulation site (lesion side or intact side) were all different, which added to the uncertainty.

Hamdy et al. reported through a TMS study that swallowing musculature is represented on the motor and premotor cortex of both hemispheres but displays interhemispheric asymmetry, independent of handedness.11 Following a unilateral hemispheric stroke, oropharyngeal dysphagia appeared to be associated with a smaller pharyngeal representation on the intact hemisphere,12 which increases in size with recovery of swallowing.13 This suggested a role for intact hemisphere reorganization in recovery.

With this background information, we hypothesized that if we increased the excitability of the pharyngeal motor cortex on the intact hemisphere, it could help the recovery from oropharyngeal dysphagia following a stroke. Therefore, in this study, we intend to find the therapeutic effect of high-frequency repetitive TMS on a contra-lesional intact pharyngeal motor cortex in post-stroke dysphagic patients.

Materials and methods

Subjects

Forty-five post-stroke patients with oropharyngeal dysphagia were eligible for this study, which was conducted from November 2010 to January 2012. Participants satisfied the following inclusion criteria:1 they had unilateral hemispheric stroke,2 they had oropharyngeal dysphagia confirmed by a videofluoroscopic swallowing study (VFSS), and3 they had oropharyngeal dysphagia that persisted over 1 month after stroke onset. We determined the presence of oropharyngeal dysphagia using our institutional VFSS protocol. Patients were given 2 and 5 mL of diluted barium (35% w/v), curd-type yogurt, and mashed boiled pumpkin. Oropharyngeal dysphagia was defined as any abnormality found in the VFSS. Patients who had metal implants or a pacemaker in their bodies, and had history of seizures were excluded. A physician and a speech language pathologist examined all of the participants. This study was approved by the Institutional Review Board and informed consent was obtained from every patient.

Procedures

This study was designed as a double-blind, randomized, controlled study and was scheduled to last for a total of 4 weeks. Patients were randomly allocated into two groups: real rTMS (experimental) group or a sham rTMS (control) group. A computer-generated randomization sequence was utilized and an automated assignment system ensured allocation concealment. Patients and assessors were unaware of group assignment. Before stimulation therapy, we obtained baseline data for the videofluoroscopic dysphagia scale (VDS)14 and the penetration-aspiration scale (PAS)15 using VFSS. Magnetic stimulation therapy was then given to the patients over the course of 2 weeks. Following this, we repeated the VFSS to evaluate the therapeutic effects of rTMS. During the final 2 weeks of the study, the patients received no treatment after which the last VFSS data were gathered to determine the long-term effects of rTMS therapy.

Magnetic stimulation

Patients were seated in a comfortable reclining chair and a 3 mm diameter pharyngeal catheter (F-EPG-G-15cm-120; Astro-Med, Inc., West Warwick, RI, USA) was inserted nasally and a pair of surface electrodes was placed on the intact side thenar muscle. These were connected with an electromyography (EMG) device (Synergy EMG/EP system; Oxford Instruments Medical Ltd., Surrey, UK) and an EMG signal was obtained. The cranial vertex was identified 16 and marked on the scalp. The pharyngeal and then adjacent thenar motor hot spots were determined by discharging the magnetic stimulator (Magstim Rapid2; Magstim Co. Ltd., Wales, UK) at suprathreshold intensities over intact cortices to identify the site evoking the greatest pharyngeal response and subsequent colocalized thenar response. These sites were marked on the scalp. The pharyngeal and thenar motor thresholds which were defined as the intensity that produced motor evoked potentials of at least 20 μV on 50% of occasions for the pharynx or at least 50 μV on 50% of occasions for the hand were subsequently determined using a single-pulse TMS. The thenar motor threshold was used to define the intensity of rTMS to be delivered so that paradigms complied with current safety guidelines.17

Real rTMS was applied for 10 min every weekday for 2 weeks. A session of stimulation consisted of 10 trains of 5-Hz stimulation, each lasting for 10 s and then repeated every minute given through a 70 mm figure-of-eight coil positioned over the pharyngeal hot spot of the intact hemisphere.18 The intensity of stimulation was set at 90% of the thenar motor threshold of the same hemisphere. Sham stimulation was given using a 90° coil tilt, which produces the same noise as active stimulation, but has been shown not to produce motor cortical stimulation.

Videofluoroscopic swallowing study (VFSS)

Patients were seated upright and a 1.5 m distance from the x-ray tube (Sonialvision-100; Shimadzu Corporation, Kyoto, Japan) to the head was maintained. The patients were asked to swallow 5 mL of 35% barium (w/v) three times. The lateral view of videofluoroscopic images was directly captured by using a digital picture archiving and communication system (PACS). The capture rate was 30 frames s−1 and the frame size was 1021 × 1021 pixels.

Outcome measures

Videofluoroscopic dysphagia scale and PAS were determined after observing the entire swallowing action several times in real time and using frame by frame. Conclusions were reached by two analysts’ consensus.

Videofluoroscopic dysphagia scale14 is a 100-point scale for evaluating oropharyngeal dysphagia. It consists of the 14 oral and pharyngeal phase parameters and can produce numerical data on the swallowing functionality of individual stroke patients through the use of the comprehensive VFSS findings. For the oral phase, the examiner assigned lip closure, bolus formation, mastication, and tongue-to-palate contact to one of three levels: grade 0 = normal; grade 1 = inadequate; grade 2 = none. Apraxia, tongue thrust, and piecemeal deglutition were evaluated on 4 levels: grade 0 = none; grade 1 = mild; grade 2 = moderate; grade 3 = severe. The amount of premature bolus loss and the residue of bolus in the oral cavity were graded at 4 levels: grade 0 = none; grade 1 ≤10% of bolus; grade 2 = 10–50% of bolus; grade 3 ≥50% of bolus). The oral transit time was also measured. In the pharyngeal phase, the triggering of pharyngeal swallow, laryngeal elevation, vallecular residue, pyriform sinus residue, coating of pharyngeal wall, repeated swallowing, and pharyngeal transit time were checked. The amounts of vallecular and pyriform sinus residue were classified into four levels: grade 0 = no residue; grade 1 ≤10% of the area of the vallecular or pyriform sinus in two-dimensional view; grade 2 = 10–50% of the area; grade 3 ≥50% of the area). Aspiration was also checked and graded into three levels: grade 1 = no aspiration; grade 2 = penetration; grade 3 = aspiration) (Appendix A).

Penetration-aspiration scale15 is an 8-point clinical scale for rating penetration and aspiration. The severity of the rating depends on the perceived depth of entry of food or fluid material into the airway and clearance of the material from the airway. The mean value among three swallowing tries was used.

Statistical analysis

Statistical analyses were performed using SPSS version 12.0 (SPSS, Inc., Chicago, IL, USA). Quantitative variables are expressed as means ± SD and qualitative variables are expressed as absolute values. Group comparisons on baseline demographics and clinical characteristics were performed with a Mann–Whitney U-test for continuous variables and a χ2 test for categorical variables to test unbalancing between groups. A Wilcoxon signed rank test was used to compare the outcomes (VDS and PAS) between paired variables (before vs after treatment) in each group. The level of significance was set at P < 0.05.

Results

Forty-five patients were assessed for eligibility of this study, and 18 of the 45 patients who completed 10 sessions of stimulation were included in the analysis. Nine patients received real rTMS (experimental group; EG) and nine patients received sham stimulation (control group; CG). Four patients were excluded because of their history of seizures and 23 patients refused to attend this study. There was no complication concerned with rTMS during the study. (Fig. 1) The average age of both groups was 71.3 ± 7.3 years (10 men and 8 women). The demographics and general characteristics of all participants are described in Table 1. In the baseline VFSS, 22% of silent aspirations were observed in each group. There were no complications in the follow-up periods. When we compared the baseline demographic variables, there were no significant differences between the groups (Table 2).

Figure 1.

 Flow diagram of the study.

Table 1. Demographics and general characteristics of the participants in the study
PatientsGenderAgeStroke typeLesion locationLesion sideNIHSSBarthel indexCharlson indexInitial feeding statusInitial VFSS findingsSilent aspirationComplication in the follow-up periods
Experimental group
 1Male71InfarctionMiddle cerebral artery territoryRight19364Modified oral dietPremature bolus loss, vallecular and pyriform sinus residue
 2Male70InfarctionMiddle cerebral artery territoryLeft19495Levin tube dietInadequate tongue-to palate contact, vallecular and pyriform sinus residue, aspiration
 3Female76InfarctionStriatocapsularRight3904Levin tube dietPremature bolus loss, delayed swallowing reflex and aspiration+
 4Male69HemorrhageBasal gangliaRight43310Modified oral dietDelayed swallowing reflex and penetration
 5Male80InfarctionMiddle cerebral artery territoryRight4606Regular oral dietTongue hesitation, vallecular and pyriform sinus residue
 6Female74InfarctionMiddle cerebral artery territoryLeft1654Levin tube dietDelayed swallowing reflex and aspiration
 7Female70InfarctionMiddle cerebral artery territoryRight11155Modified oral dietInadequate bolus formation, vallecular and pyriform sinus residue
 8Female77InfarctionMiddle cerebral artery territoryRight1805modified oral dietInadequate bolus formation, vallecular and pyriform sinus residue and penetration
 9Male76InfarctionStriatocapsularLeft12264Levin tube dietInadequate bolus formation, vallecular and pyriform sinus residue and aspiration+
Control group
 AMale66inf$arctionMiddle cerebral artery territoryRight1535Modified oral dietInadequate bolus formation, tongue hesitation, delayed swallowing reflex, penetration
 BFemale74InfarctionStriatocapsularRight10404Modified oral dietDelayed swallowing reflex and aspiration
 CMale52InfarctionMiddle cerebral artery territoryLeft10474Levin tube dietPyriform sinus residue, delayed swallowing reflex and penetration
 DFemale68HemorrhageBasal gangliaLeft13496Modified oral dietDelayed swallowing reflex
 EMale74HemorrhageBasal gangliaRight3534Levin tube dietInadequate tongue-to palate contact, vallecular and pyriform sinus residue, aspiration+
 FMale72InfarctionMiddle cerebral artery territoryRight12394Levin tube dietInadequate bolus formation, delayed swallowing reflex, aspiration
 GFemale64InfarctionBasal gangliaLeft7555Regular oral dietDelayed swallowing reflex
 HMale64InfarctionMiddle cerebral artery territoryLeft14124Modified oral dietDelayed swallowing reflex and penetration
 IFemale86InfarctionMiddle cerebral artery territoryRight5115Levin tube dietInadequate tongue-to palate contact, vallecular residue and aspiration+
Table 2. Comparisons of basic characteristics of the participants in the study
 Experimental group (n = 9)Control group (n = 9) P-value
  1. *Mean ± SD (minimum–maximum).

Age(y)73.7 ± 3.8(69–80)*68.9 ± 9.3(52–86)*ns
Gender
 Male55ns
 Female44
Etiology
 Hemorrhage21ns
 Infarction78
Lesion side
 Right65ns
 Left34
Duration of illness (from onset to study enroll) (days)59.9 ± 16.3(45–91)*63.9 ± 26.8(36–114)*ns
Barthel Index34.9 ± 28.4(0–90)34.3 ± 20.1(3–55)ns
VDS (baseline)33.6 ± 12.1(17.5–53.0)*23.4 ± 14.6(4.5–53.5)*ns
PAS (baseline)3.4 ± 2.3(1–8)*3.3 ± 2.0(1–8)*ns

Mean baseline VDS of EG was 33.6 ± 12.1 and the score was reduced to 25.3 ± 9.8 just after 2 weeks intervention. This score was maintained until 2 weeks after stopping treatment. Through the Wilcoxon signed rank test, statistical difference between baseline VDS and week 2 VDS was confirmed. In addition, there was a significant difference between baseline VDS and week 4 VDS (Fig. 2A). When we analyzed VDS divided into oral phase and pharyngeal phase scores, the significant improvement was only observed in the pharyngeal phase, not in the oral phase (Fig 2CE). Mean baseline VDS of CG was 23.4 ± 14.6, which was changed to 21.2 ± 15.6 at week 2 and 20.4 ± 15.3 at week 4. However, there was no significant difference among them (Fig. 2B).

Figure 2.

 The change of the videofluoroscopic dysphagia scale (VDS). (A) total scores of experimental (real rTMS) group (EG). (B) total scores of control (sham rTMS) group (CG). (C) oral phase scores of EG. (D) oral phase scores of CG. (E) pharyngeal phase scores of EG. (F) pharyngeal phase scores of CG. There was a significant difference between baseline VDS and week 2 VDS and the difference was maintained until week 4 in the EG. This change was only seen in the pharyngeal phase scores, not in the oral phase scores.

Similar results were observed in PAS. A mean baseline PAS of EG was 3.41 ± 2.32, which was reduced to 1.93 ± 1.52 at week 2 and 1.37 ± 0.87 at week 4. The Wilcoxon signed rank test showed the difference between PAS at baseline and week 2 and between PAS at baseline and week 4 (Fig. 3A). The mean PAS at baseline, week 2, and week 4 in CG were 3.30 ± 2.03, 3.00 ± 2.17 and 3.11 ± 2.15, respectively. There was no statistically significant difference among them (Fig. 3B).

Figure 3.

 The change of the penetration-aspiration scale (PAS). (A) experimental (real rTMS) group. (B) control (sham rTMS) group. There was a significant difference between baseline PAS and week 2 PAS and the difference was also maintained until week 4 in the experimental group.

The baseline prevalence of aspiration and penetration, vallecular and pyriform sinus residue, delayed triggering of pharyngeal swallowing and abnormal pharyngeal transit time (PTT) in EG was 66.7%, 66.7%, 33.3%, and 44.4%, respectively. After rTMS treatment, the prevalence of aspiration and penetration, and vallecular and pyriform sinus residue was reduced to 33.3% and 33.3%, respectively. However, the prevalence of delayed triggering of pharyngeal swallowing, and abnormal PTT in EG was not changed. The baseline prevalence of aspiration and penetration, vallecular and pyriform sinus residue, delayed triggering of pharyngeal swallowing, and abnormal PTT in CG was 77.8%, 22.2%, 77.8%, and 33.3%, respectively. After sham rTMS, the prevalence of aspiration and penetration, vallecular and pyriform sinus residue, delayed triggering of pharyngeal swallowing, and abnormal PTT in CG was recorded as 66.7%, 22.2%, 77.8%, and 22.2%, respectively.

Discussion

Two weeks of a 5Hz excitatory rTMS over the intact pharyngeal motor cortex gave clinical improvement to dysphagic patients with unilateral hemispheric stroke and the effect lasted up to 2 weeks after a treatment cessation. This application might increase the cortical excitability of the unaffected hemisphere, which enhanced the stimulation of bulbar motor neurons projecting to the pharynx and could facilitate swallowing function. It was suggested that 5 Hz rTMS may have produced a stronger inhibition of GABAergic circuits in the pharyngeal cortex, allowing a clear buildup of long-term potentiation and possibly brain glutamate, and a large and prolonged increase in corticobulbar excitability.19 Multi-session cumulative rTMS treatment might provide a long-lasting effect on patients.20

Our results suggested that the addition of cortical stimulation increased the speed of recovery from oropharyngeal dysphagia and supported a previous research finding13 that the reorganization of the intact swallowing cortex might have an important role in recovery from oropharyngeal dysphagia after a unilateral hemispheric stroke.

However, this hypothesis was completely contrary to Verin and Leroi’s hypothesis.10 They used inhibitory 1Hz rTMS for 20 min every day for 5 days to induce suppression in the healthy hemisphere of seven stroke patients with oropharyngeal dysphagia targeting decreased transcallosal inhibition. This resulted in an improvement in behavioral markers for swallowing impairment, penetration-aspiration scale, and swallow reaction times as observed with videofluoroscopy. We thought that up-regulation of an intact hemisphere was more effective than down-regulation because the return of swallowing after a dysphagic stroke was associated with an increased pharyngeal representation in the unaffected hemisphere.13

Khedr et al.9 delivered excitatory rTMS (300 pulses at 3 Hz – 120% hand motor threshold) to 26 unilateral hemispheric stroke patients with swallowing problems for 10 min per day for five consecutive days. The application of the excitatory rTMS was to the affected stroke hemisphere and the intervention resulted in a bilateral increase in brain excitability 1 and 2 months after treatment and recovery from oropharyngeal dysphagia. This was thought to be a good method. However, practically, it is not easy to find a ‘hot spot’ in the injured hemisphere because the corticobulbar tract might be disrupted.

When we analyzed VDS divided into pharyngeal and oral phase, significant improvement was observed only in the pharyngeal phase. It means rTMS did not affect oral phase function but pharyngeal phase function. Because rTMS enhanced the excitability of pharyngeal motor cortex, it may increase the pharyngeal motor performance and reduce penetration or aspiration. There would be ‘non-responders’ in real rTMS group and maybe they had a main problem in their oromotor functions.

One of the most difficult things is to determine the treatment parameters. Fortunately, Hamdy18,19 already revealed several important rTMS parameters, which gave the most impact on the brain plasticity in his previous studies. We chose the 5Hz frequency, which was the optimal frequency to increase the excitability of the corticobulbar projection to the pharynx.19 In addition, we chose 90% thenar motor threshold as the stimulation intensity and 500 pulses as stimulation duration.18 However, 10 times of treatment session for 2 weeks were arbitrarily determined by us because the optimal number of sessions was not known. This point should be further researched.

We selected the patients with persistent oropharyngeal dysphagia that lasted longer than 1 month after stroke onset to rule out natural recovery because most patients regain their swallowing abilities within 1 month after stroke.21,22 However, we did not think that 1 month of evolution is completely enough to prevent an effect due to spontaneous recovery. We only followed up patients for 2 weeks after the cessation of an intervention and it was somewhat short to say the long-term effect of this intervention. Besides, the sample size was too small. We randomly assigned the participants and used a statistical analysis for nonparametric valuables, but the limitation still remained. The two groups of patients seem to have seemingly different baseline VFS scores and the control group had visibly lower scores. Indeed, the difference between baseline VFS for active and control was larger than the significant change in the active group after treatment. A lower baseline score for the control group has consequences for the data – as it would be more difficult to improve an already less dysphagic population to something better – the window of effect size would be limited by a flooring effect. These factors should be addressed in future research.

Conclusion

A 5Hz high-frequency rTMS on contra-lesional pharyngeal motor cortex might be beneficial for dysphagic patients with unilateral hemispheric stroke and mainly pharyngeal phase dysfunction. This intervention can be used as a new treatment method in post-stroke patients with oropharyngeal dysphagia.

Funding

This research was supported by Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Education, Science and Technology (grant number: 2010-0003964).

Disclosure

No competing interests declared.

Author Contribution

RKH, YJS, OJC, and LJH performed the research; PJW designed the research, analyzed the data, and wrote the paper.

Appendix

Appendix A. Videofluoroscopic Dysphagia Scale (VDS)

ParameterCoded valueScore
Lip closure
 Intact04
 Inadequate2
 None4
Bolus formation
 Intact06
 Inadequate3
 None6
Mastication
 Intact08
 Inadequate4
 None8
Apraxia
 None04.5
 Mild1.5
 Moderate3
 Severe4.5
Tongue-to palate contact
 Intact010
 Inadequate5
 None10
Premature bolus loss
 None04.5
 <10%1.5
 10–50%3
 >50%4.5
Oral transit time
 ≤1.5 s03
 >1.5 s3
Triggering of pharyngeal swallow
 Normal04.5
 Delayed4.5
Vallecular residue
 None06
 <10%2
 10–50%4
 >50%6
Laryngeal elevation
 Normal09
 Impaired9
Pyriform sinus residue
 None013.5
 <10%4.5
 10–50%9
 >50%13.5
Coating of pharyngeal wall
 No09
 Yes9
Pharyngeal transit time
 ≤1.0 s06
 >1.0 s6
Aspiration
 None012
 Supraglottic penetration6
 Subglottic aspiration12
Total 100

Ancillary