2023 年 8 巻 論文ID: 20230015
Objectives: This study aimed to update the current knowledge on non-invasive brain stimulation (NIBS) effects, such as repetitive transcranial brain stimulation and transcranial direct current stimulation, in patients with post-stroke dysphagia (PSD).
Methods: We summarized the basic principles and therapeutic strategies of NIBS. We then reviewed nine meta-analyses from 2022 that investigated the efficacy of NIBS in PSD rehabilitation.
Results: Although dysphagia is a common and devastating sequela of stroke, the efficacy of conventional swallowing therapies remains controversial. NIBS techniques have been proposed as promising approaches for managing PSD via neuromodulation. Recent meta-analyses have shown that NIBS techniques are beneficial for the recovery of patients with PSD.
Conclusions: NIBS has the potential to become a novel alternative treatment for PSD rehabilitation.
Swallowing abnormalities are one of the most common problems after stroke, with reported frequencies ranging from 29% to 81%.1) Although natural recovery processes occur during the first few weeks after stroke, about 50% of stroke survivors still experience difficulty in swallowing at 6 months after stroke.2) In addition, post-stroke dysphagia (PSD) causes serious complications such as aspiration pneumonia, malnutrition, and dehydration. PSD can be a key risk factor for prolonged hospital stays, poor functional outcomes, and increased mortality.3,4) Therefore, managing PSD would help prevent medical complications and reduce its socioeconomic burden.
Conventional dysphagia treatment involves behavioral interventions, appropriate dietary modifications, and physical sensory stimulation, such as tactile or thermal stimulation.1,5) Although these conventional therapies may have some positive effects, they usually require frequent training over a period of weeks before a good clinical response is observed.6) Furthermore, their efficacy remains uncertain.5)
In recent years, non-invasive brain stimulation (NIBS) has attracted attention as a promising therapeutic tool for stroke rehabilitation.7,8) NIBS modulates human brain function by inducing cortical plasticity. Swallowing is controlled by deep brain areas such as the brain stem, basal ganglia, thalamus, and motor and sensory cortices.9) Therefore, targeting the cerebral cortices involved in the neural control of swallowing can improve recovery from PSD. NIBS has been used for PSD rehabilitation.10) Repetitive transcranial magnetic stimulation (rTMS) and transcranial direct current stimulation (tDCS) are the two major techniques used for NIBS. This review summarizes the basic principles and therapeutic strategies of these techniques and discusses the findings of nine recent meta-analyses (published in 2022) of NIBS for PSD rehabilitation.
TMS is an electromagnetic technique used for painless brain stimulation through an intact skull. An insulated coil is placed directly above the participant’s head and a brief but strong current is passed through the coil. This coil current induces the formation of a transient magnetic field in the brain and generates electric currents in the cortex. Because these currents are sufficient to depolarize neurons, TMS can enhance or decrease cortical activity. Although a single pulse of TMS to the brain does not leave any lasting effects, rTMS can produce neuromodulatory effects that persist after the stimulation period.11) In general, low-frequency rTMS, with stimulus rates of 1 Hz or lower, inhibits cortical excitability, whereas high-frequency rTMS, with stimulus rates higher than 1 Hz, activates cortical excitability12); synaptic changes similar to long-term depression and long-term potentiation have been suggested as possible mechanisms, respectively.
tDCS uses direct electrical current to shift the resting membrane potential of nerve cells. In contrast to TMS, it does not induce neuronal firing by depolarizing neurons. In tDCS, a small electric current (usually 1–2 mA) is applied to the brain through rubber electrodes placed on the participant’s head. Usually, one of the electrodes is placed on the scalp above the targeted cortical region. Depending on the polarity of the stimulation, it inhibits or activates cortical excitability; in general, cathodal tDCS decreases cortical excitability, whereas anodal tDCS increases it.13) Furthermore, tDCS induces prolonged neuromodulatory effects depending on the duration of stimulation.14)
Different forms of reorganization contribute to functional recovery from stroke.7) NIBS protocols in PSD are mainly based on two models: the interhemispheric and compensatory models.15)
The interhemispheric model suggests that unilateral stroke leads to a pathological interaction between the damaged and undamaged hemispheres. In a healthy brain, each hemisphere inhibits the other hemisphere (interhemispheric inhibition, IHI).16) The IHI between the homonymous motor cortex is associated with smooth task performance by suppressing ‘mirrored’ movement. The inhibition is mediated by transcallosal fibers in the corpus callosum.17) In brain stroke, inhibition from the ipsilesional (damaged) hemisphere to the contralesional (undamaged) hemisphere is reduced.18) This results in the overactivation of the contralesional hemisphere, which further leads to abnormal overactivation of inhibition from the contralesional hemisphere to the ipsilesional hemisphere. Based on the assumption that reorganization in the ipsilesional hemisphere is important for functional recovery,19) this interhemispheric imbalance impairs functional recovery in the ipsilesional hemisphere. Therefore, neuromodulation targeting the abnormal imbalance may be a promising method for enhancing stroke rehabilitation.20) These measures could include decreasing the excitability in the contralesional motor cortex by low-frequency rTMS or cathodal tDCS and/or increasing the excitability in the ipsilesional motor cortex21,22) by high-frequency rTMS or anodal tDCS (Fig. 1a).
The compensatory model suggests that functional recovery after a stroke is associated with reorganization in the undamaged hemisphere.15) Swallowing is a complex neuromuscular activity that requires appropriate control from the brain stem, basal ganglia, thalamus, limbic system, cerebellum, and cerebral cortices. It involves midline structures innervated bilaterally but controlled asymmetrically by the motor cortices. In humans, the swallowing-dominant hemisphere, which has stronger swallowing representations, is independent of handedness and varies among individuals.23) It also varies among individuals depending on task demands.24) Reorganization in the undamaged hemisphere is associated with the recovery of swallowing in patients with PSD.25) Therefore, accelerating the compensatory process in the undamaged hemisphere can help restore swallowing function in patients with PSD. Stimulating the undamaged hemisphere21,22) with high-frequency rTMS or anodal tDCS may facilitate the compensatory process (Fig. 1b).
The mechanisms of recovery from PSD are more complex than those of the interhemispheric and compensatory models; therefore, these models can only partially explain the results of NIBS treatment for PSD rehabilitation. A recent study reported that the two alternate (bimodal) strategies were needed for functional recovery from brain stroke, depending on severity. In mildly affected patients, inhibition of the contralesional hemisphere improved motor functions, whereas in severely affected patients, facilitation of the ipsilesional hemisphere showed improvement.26) This suggests that the NIBS protocol for PSD rehabilitation should be tailored for each patient.
Given the increasing number of studies investigating the effects of NIBS on PSD rehabilitation, the number of meta-analyses conducted on NIBS studies has also increased. We introduced nine meta-analyses of rTMS and tDCS in PSD rehabilitation published from February 2022 to July 2022.15,27,28,29,30,31,32,33,34) The results of rTMS15,28,30) and tDCS27,29,33) were each summarized by three studies, and a further three studies summarized both.31,32,34) The studies on rTMS included 17 articles (2 articles35,36) were included in all the rTMS studies) and those on tDCS included 26 articles (5 articles37,38,39,40,41) were included in all the tDCS studies) (Table 1).
| Study | Inclusion criteria | Exclusion criteria | Main points |
| Hsiao et al.28) | (1) Ischemic or hemorrhagic stroke confirmed by magnetic
resonance imaging (MRI) or computed tomography (CT) (2) Dysphagia symptoms (3) No neurological diseases other than stroke or swallowing disorders (4) rTMS and sham stimulation (5) Statistical combination of results available for conducting a meta-analysis |
(1) Review articles, case reports, or letters
(2) Insufficient data or results (3) Crossover studies |
(1) High-frequency (≥3 Hz) ipsilesional rTMS and
low-frequency (<1 Hz) contralesional rTMS showed more significant improvement in
SSA scores compared with sham treatments, both immediately and 4 weeks after
treatment. (2) Penetration–Aspiration Scale scores showed greater improvement after low-frequency contralesional rTMS than after sham treatments, immediately after treatment. |
| Qiao et al.30) | (1) RCTs, including crossover and cluster RCTs
(2) PSD confirmed by video fluoroscopic swallowing study (VFSS), fiberoptic endoscopic evaluation of swallowing, swallowing questionnaire, or dysphagic outcome and severity scale (3) ≥18 years (4) rTMS (5) Original data or sufficient information about dysphagia |
(1) No original data | (1) An intervention lasting >5 days and rTMS in the
subacute phase (<60 days) after a stroke had a significantly larger effect size
than control conditions. (2) rTMS with a single stimulation time ≤10 min and >10 min, low-frequency (1 Hz) and high-frequency (≥3 Hz) stimulation, stimulation of the affected hemisphere and that of the unaffected hemisphere, and rTMS in patients ≤65 years and that in patients >65 years had a significant effect size compared with control conditions. (3) High-frequency stimulation of the unaffected hemisphere, high-frequency stimulation of the affected hemisphere, and low-frequency stimulation of the unaffected hemisphere exhibited a significantly larger effect size than control conditions. |
| Xie et al.15) | (1) Ischemic or hemorrhagic stroke confirmed by MRI or
CT (2) Dysphagia (3) No swallowing disorders caused by diseases other than stroke (4) RCTs comparing rTMS with sham stimulation or other routine rehabilitation training |
Not specified | (1) rTMS of the unaffected hemisphere and the bilateral
hemispheres significantly improved overall swallowing function and reduced instances
of aspiration. (2) Low-frequency rTMS (≤1 Hz) significantly improved overall swallowing functions and reduced aspiration. (3) No significant difference in the dropout rate and adverse effects between rTMS treatments and control treatments. |
| Lin et al.29) | (1) ≥18 years (2) Ischemic or hemorrhagic stroke confirmed by CT or MRI (3) Studies focusing on the effect of tDCS on the recovery of swallowing function (4) RCTs with crossover and parallel designs (5) Standardized, validated dysphagia scales as the outcome measures |
(1) Studies other than RCTs such as reviews,
meta-analyses, or case reports (2) Not published in English or Chinese (3) Unavailable required data |
(1) tDCS was effective in both acute (≤14 days) and
convalescent and chronic (≥15 days) stages after stroke. (2) Although anodal tDCS in the affected, unaffected, and bilateral hemispheres had a significantly larger effect size than the control conditions, the effect size was larger in the unaffected hemisphere than in the other protocols. |
| He et al.27) | (1) RCTs of tDCS for PSD | (1) Studies other than RCTs, such as systematic reviews, letters, case reports, editorials, animal studies, or commentary | (1) Anodal tDCS in both hemispheres was likely to be
superior to that in the unaffected hemisphere. (2) High stimulation intensity (1.6–2 mA) had a larger positive effect than low stimulation intensity (1–1.5 mA). |
| Zhao et al.33) | (1) Ischemic or hemorrhagic stroke confirmed by CT or
MRI (2) tDCS (3) Studies including at least one of the following standardized, validated dysphagia scales: Modified Mann Assessment of Swallowing Ability, SSA, Functional Oral Intake Scale, Dysphagia Outcome Severity Scale, Functional Dysphagia Scale, and VFSS (4) Clinical RCTs of tDCS for the treatment of PSD |
(1) Non-RCTs (2) A repetitive literature, review, and nonpublic literature, such as a conference paper (3) Swallowing dysfunction caused by other diseases, such as craniocerebral trauma or Parkinson disease (4) Poor rating on the Physiotherapy Evidence Database Scale (0–3 of 10) (5) No complete data |
(1) Although both anodal tDCS in the affected and
unaffected hemispheres significantly improved PSD, the effect size in the affected
hemisphere was larger than that in the unaffected hemisphere. (2) tDCS in the convalescent and chronic stages (beyond 15 days) had a significantly large effect size. (3) 1 and 1.6 mA stimulation showed a moderate and significant effect size, respectively. (4) tDCS was effective for PSD in unilateral and brainstem strokes. |
| Speyer et al.31) | (1) Oropharyngeal dysphagia (2) Non-invasive neurostimulation interventions aimed at reducing swallowing or feeding problems (3) Studies including a control group or comparison intervention group (4) Random assignment to one of the study arms or groups (5) Published in English |
(1) Non-electrical peripheral stimulation,
pharmacological interventions, and acupuncture (2) Invasive techniques and/or those that did not specifically target oropharyngeal dysphagia (3) Conference abstracts, doctoral theses, editorials, and reviews |
(1) In rTMS, a longer intervention time (14 days) and
higher pulse number (≥10,000) had increased positive effect sizes. (2) In rTMS, the differences in stimulation sites (bilateral, contralesional, or ipsilesional sites) or stimulation frequencies (1, 5, or 10 Hz) were not statistically significant. (3) In tDCS, a longer intervention time (14 and 28 days), longer stimulation time (200, 300, and 400 min), and higher stimulation currents (1 and 2 mA) were associated with larger positive effect sizes. |
| Tan et al.32) | (1) ≥18 years (2) Ischemic or hemorrhagic stroke confirmed by relevant examinations (CT, MRI) (3) PSD confirmed by VFSS or relevant tests (4) Transcranial stimulation, such as rTMS and tDCS (5) Presence of a control group that received the usual care for dysphagia, sham stimulation, or usual care with sham stimulation (6) Swallowing function evaluated with validated measurement tools (7) RCTs (8) Published in English |
(1) Dysphagia caused by diseases other than stroke,
pre-existing dysphagia, or healthy people (2) Other brain stimulation approaches including transcranial stimulation combined with other brain stimulations (3) Non-RCTs, such as quasi-experimental design, one-group pre-test post-test design, non-experimental studies, qualitative studies, review papers, or case studies |
(1) rTMS had a large effect size and tDCS had a medium
effect size in improving swallowing functions. (2) Stimulation of the esophageal cortical area had a significantly larger effect size than stimulation of the inferior sensorimotor cortex and premotor brain regions, mylohyoid cortical area, and pharyngeal cortex. (3) NIBS treatment had a greater effect on PSD rehabilitation in middle-aged patients. |
| Zhu and Gu34) | (1) RCTs | (1) Studies not exploring the effect of NIBS on PSD
(2) Meta-analyses, reviews, and case reports (3) Insufficient information regarding pre- and post-treatment swallowing-associated scores |
(1) Both rTMS and tDCS of the pharyngeal motor cortex showed a significant improvement in swallowing function compared with sham treatments. |
Hsiao et al. analyzed six randomized controlled trials (RCTs) to evaluate the effects of rTMS parameters, including stimulation frequency and site, on PSD.28) High-frequency (≥3 Hz) ipsilesional rTMS and low-frequency (<1 Hz) contralesional rTMS showed significant improvement in Standardized Swallowing Assessment (SSA) scores when compared with sham treatments, both immediately and 4 weeks after treatment. When assessed immediately after treatment, the Penetration–Aspiration Scale scores showed greater improvement after low-frequency contralesional rTMS than after sham treatments. The authors concluded that the rTMS protocol based on the interhemispheric model could be a therapeutic option for PSD.
Qiao et al. evaluated the efficacy of rTMS in PSD rehabilitation and explored optimal stimulation parameters.30) Twelve RCTs were included in this meta-analysis. They reported that rTMS treatment had a positive effect on PSD. The subgroup analyses showed that an intervention lasting for at least 5 days and rTMS in the subacute phase (<60 days) after a stroke had a significantly larger effect size than control conditions, whereas an intervention lasting for less than 5 days and rTMS in the recovery phase (>60 days) after a stroke did not have a significant effect. In addition, the following rTMS treatment groups showed significant effect size when compared with control conditions: single stimulation time of 10 min or less, single stimulation time of greater than 10 min, low-frequency (1 Hz) and high-frequency (≥3 Hz) stimulation, stimulation of the affected hemisphere and that of the unaffected hemisphere, and rTMS in patients younger than 65 years and that in patients over 65 years. Furthermore, high-frequency stimulation of the unaffected hemisphere, high-frequency stimulation of the affected hemisphere, and low-frequency stimulation of the unaffected hemisphere exhibited a significantly larger effect size than control conditions.
Xie et al. pooled data from ten RCTs.15) The meta-analysis showed that rTMS improved overall swallowing function, reduced instances of aspiration, and improved daily living activities significantly when compared with control treatments. Subgroup analyses revealed that stimulation of the unaffected hemisphere and the bilateral hemispheres significantly improved overall swallowing function and reduced instances of aspiration, whereas stimulation of the unaffected hemisphere did not. They also revealed that low-frequency rTMS (≤1 Hz) significantly improved overall swallowing function and reduced aspiration, whereas high-frequency stimulation (>1 Hz) had no greater effect in reducing aspiration than the control treatments. Furthermore, they observed no significant difference in the dropout rate or adverse effects between rTMS treatments and control treatments.
Lin et al. performed a meta-analysis of ten RCTs to evaluate the efficacy of tDCS in PSD rehabilitation.29) They found that tDCS was effective in both acute (≤14 days) and convalescent and chronic (≥15 days) stages after stroke. Although anodal tDCS in the affected, unaffected, and bilateral hemispheres had a significantly larger effect size than the control conditions, the effect size was larger in the unaffected hemisphere than in the other protocols.
He et al. conducted a meta-analysis that included 15 RCTs on tDCS in PSD rehabilitation.27) They confirmed that tDCS had a positive effect on PSD. They revealed that anodal tDCS in both hemispheres was likely to be superior to that in the unaffected hemisphere. They also demonstrated that high stimulation intensity (1.6–2 mA) had a larger positive effect than low stimulation intensity (1–1.5 mA).
Zhao et al. analyzed 16 RCTs to evaluate the effects of tDCS on PSD rehabilitation.33) Overall, the results demonstrated that tDCS facilitated the recovery of PSD, with a significantly large pooled-effect size. Although both anodal tDCS in the affected and unaffected hemispheres significantly improved PSD, the effect size in the affected hemisphere was larger than that in the unaffected hemisphere. In subgroup analyses, tDCS in the convalescent and chronic stages (beyond 15 days) had a significantly large effect size. According to the stimulation intensity, 1 and 1.6 mA stimulation showed a moderate and significant effect size, respectively. According to the stroke location, tDCS was effective for PSD in unilateral and brainstem strokes, but not in cerebellar and basal ganglia strokes.
Speyer et al. assessed the effects of rTMS and tDCS on PSD rehabilitation.31) A total of 24 RCTs (11 studies on rTMS, 9 on tDCS, and 4 on another type of neurostimulation, including neuromuscular electrical stimulation in addition to rTMS) were included in their meta-analysis. They confirmed the significant beneficial effects of rTMS and tDCS. The subgroup analyses for rTMS showed that a longer intervention time (14 days) and higher number of pulses (≥10,000) had increased positive effect sizes. Differences in stimulation sites (bilateral, contralesional, or ipsilesional sites) or stimulation frequencies (1, 5, or 10 Hz) were not statistically significant. The subgroup analyses for tDCS showed that a longer intervention time (14 and 28 days), longer stimulation time (200, 300, and 400 min), and higher stimulation currents (1 and 2 mA) were associated with larger positive effect sizes.
Tan et al. reviewed 16 RCTs studying the effects of NIBS on PSD rehabilitation (9 studies on rTMS and 7 on tDCS).32) In their meta-analysis, rTMS had a large effect size and tDCS had a medium effect size in improving swallowing functions. Subgroup analyses on the results of both rTMS and tDCS showed that stimulation of the esophageal cortical area had a significantly larger effect size than stimulation of the inferior sensorimotor cortex and premotor brain regions, mylohyoid cortical area, and pharyngeal cortex. This study also reported that NIBS had a greater effect on PSD rehabilitation in middle-aged patients than in older patients.
Zhu and Gu explored the therapeutic effects of rTMS and tDCS on PSD in RCTs.34) A total of 14 RCTs (7 studies each on rTMS and tDCS) were included. Both rTMS and tDCS of the pharyngeal motor cortex showed a significant improvement in swallowing function compared with sham treatments.
All the meta-analyses mentioned above showed that NIBS could improve swallowing functions compared with the control conditions. In their recent guideline,42) the European Stroke Organisation and the European Society for Swallowing Disorder recommended neurostimulation treatment including rTMS and tDCS as an adjunct to conventional dysphagia treatments to improve swallowing function. Despite these promising findings, there is no consensus regarding the optimal stimulation parameters. A past meta-analysis reported that NIBS in the unaffected hemisphere yielded better outcomes than treatment in the affected hemisphere,43) whereas recent meta-analyses reported that NIBS in the bilateral hemispheres yielded better outcomes than treatment in the unilateral hemisphere.15,44) However, another recent meta-analysis showed that anodal tDCS in the unaffected hemisphere yielded better outcomes than treatment in the affected hemisphere or the bilateral hemispheres.29) It is possible that these varied conclusions were caused by variations in the characteristics of the RCTs, such as inclusion/exclusion criteria (Table 1) and outcome measures, and the limitations of NIBS.45)
First, the response variability of NIBS is related to genetic predisposition and synaptic activity before NIBS.45) Some types of single nucleotide polymorphism are known to influence the effects of NIBS.46,47,48) Given that the homeostatic mechanism of the brain plays an important role in regulating synaptic plasticity,49,50) the threshold for inducing plasticity and its direction may depend on the history of synaptic activation.
Second, the lack of valid sham conditions leads to difficulty in blinding the participants.45) Without adequate control conditions, the treatment effects of NIBS may have been underestimated.
Third, the safety data for NIBS in PSD is insufficient.27) Although reports of adverse events are infrequent,42) this may be a result of the small sample sizes.27) Patients with PSD are likely to have comorbidities or to take multiple medications. It remains unclear whether such conditions will increase the risk of adverse events.45) Researchers and clinicians must continue to collect clinical evidence regarding the safety of NIBS in the treatment of PSD.
The aforementioned meta-analyses published in 2022 have shown promising results for NIBS in PSD rehabilitation. Considering the high variability of the response to NIBS and the mechanisms of stroke recovery, treatment protocols may need to be individualized. Future research investigating the physiological mechanisms of PSD recovery and the development of new stimulation protocols will accelerate the adoption of NIBS for PSD rehabilitation.
This study was supported by Grants-in-Aid for Scientific Research (KAKENHI) [grant number 20K21770 (S.K.), 21H03308 (S.K.), 21K17671 (S.S.), 21K19745 (T.M.), and 22H04788 (T.M.)] from the Japan Society for the Promotion of Science.
The Department of Regenerative Systems Neuroscience, Graduate School of Medicine, Kyoto University, is an endowed department through funding by the Kodama Foundation.
