Dysphagia treatments in Parkinson's disease: A systematic review and meta‐analysis

The majority of patients with Parkinson's disease (PD) develop oropharyngeal dysphagia during the course of their disease. However, the efficacy of dysphagia treatments for these patients remains controversial. Therefore, we conducted this systematic review and meta‐analysis to evaluate treatment efficacy based on the evidence from randomized controlled trials (RCTs).


| INTRODUC TI ON
Parkinson's disease (PD) is one of the most common neurodegenerative diseases affecting approximately 1903 per 100,000 individuals aged 80 years or older globally. 1 A vast majority of patients develop dysphagia across the course of their disease. The prevalence of dysphagia ranges from 35% to 82%, depending on disease severity and assessment methods. 2,3 Based on objective clinical assessments, it is estimated that dysphagia affects as much as 8 out of 10 PD patients. 2 Dysphagia in PD can lead to prolonged hospitalization and increased mortality rates, secondary to aspiration pneumonia, 4 and reduced quality of life. 5 Due to the progressively degenerative nature of the disease, dysphagia treatments for PD patients aim to maintain functional and safe swallowing to provide sufficient nutrition and hydration for as long as possible. The pathophysiology of PD-related dysphagia remains poorly understood. 6 It is believed to be related to degeneration of the substantia nigra, nondopaminergic systems, and neuronal loss in the medullary swallowing central pattern generator. 7,8 Firstline treatment for PD patients typically involves administration of dopaminergic medications, such as levodopa and apomorphine, to enhance dopaminergic transmission. Although limb motor functions generally respond well to these drugs, their effects on swallowing remain uncertain. 9 Inconsistent outcomes have been reported within individuals 7 and among patients. 10,11 Other studies reported no beneficial effects, 12 or even negative impacts on swallowing. 13 Apart from pharmacological treatments, some studies have argued deep brain stimulation (DBS), which is a common treatment for movement disorders such as tremor, rigidity, and bradykinesia in PD patients, 14 may impact (positively or negatively) swallowing function. 15 Conventional rehabilitative or compensatory swallowing treatments have also been suggested for PD patients. Rehabilitative treatments involve changing the physiology of each swallow to prevent complications in the long-term whereas compensatory treatments adjust for deficits in swallowing through changing posture or food consistency. 16 To date, there is no consensus on the efficacy of dysphagia treatments for PD patients. A number of systematic reviews have generally concluded that the level of evidence for these treatments is low and there is a lack of high-quality randomized controlled trials (RCTs), which are considered as the highest level of evidence in medical research. 9,[15][16][17][18][19][20][21][22][23][24] While these reviews do provide useful information on the treatment effects, they included non-RCT studies, and some of the recent reviews only focused on specific dysphagia treatment for PD patients. 9,23 Furthermore, no meta-analysis has been conducted in this area with the result that the efficacy remains uncertain. Indeed, therefore, the aim of this systematic review and meta-analysis is to comprehensively evaluate the efficacy of all existing dysphagia treatments for PD patients based on the evidence only from well-conducted RCT studies. The findings will provide an update on the current evidence for such treatments to assist decision-making in clinical practice.

| ME THODS
The protocol of this systematic review and meta-analysis was registered with and approved by PROSPERO (registration number: CRD42022320243), which is an international database for prospective systematic reviews. The reporting of this review followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. 25

| Study identification and search method
The search for studies was performed by two reviewers independently. Five electronic databases were systematically searched from their inception until April 2022. These were PubMed, EMBASE (via Ovid), Cumulative Index to Nursing and Allied Health Literature (CINAHL), Web of Science, and Cochrane Library. Citations from identified papers were tracked and bibliographic searches were performed for previously published reviews for relevant references (indicated as "other sources" in Figure 1). Detailed search strategies for each search engine are presented in Appendix S1.

| Eligibility criteria
We only included RCT studies with either parallel group or crossover designs. Case studies, open-label studies, animal studies, observational studies, quasi-experimental studies, retrospective studies, systematic reviews and meta-analyses and studies that did not include original data were excluded. Studies in English or Chinese were eligible for inclusion.

| Participants
We included studies with PD patients with dysphagia as determined clinically regardless of the time of onset and severity. Studies in which dysphagia was evaluated as secondary outcomes were also eligible.
Studies that recruited PD patients without dysphagia were excluded.

Highlights
• The efficacy of dysphagia treatments for Parkinson's disease (PD) is controversial.
• We evaluated the treatment effects based on nine randomized controlled trials (RCTs) with 286 PD patients.
• Our results showed that dysphagia treatments, particularly stimulation treatments, may benefit PD patients.
However, the evidence for their efficacy remains weak and uncertain.

| Treatments
We analyzed studies that compared any forms of dysphagia treatments, for example, oromotor exercises, sensory stimulation, brain stimulation treatments, or pharmacological treatments, with sham or no treatments, or conventional dysphagia therapy. Studies without nonactive controls were excluded. Trials with multiple treatments (e.g., co-administration of intraoral treatment and other swallowing rehabilitation strategies) were eligible if the study groups only differed in the use of the target treatment of interest.

| Outcomes
Outcome measures were swallowing function as evaluated by instrumental assessment. For studies with multiple outcomes, the primary outcome or the most clinically important swallowing-related outcome was used. Therefore, outcomes assessed by instrumental assessments were chosen over those based on subjective bedside evaluations, and outcomes representing swallowing safety, such as risk of aspiration and penetration, were chosen over swallowing kinematic measures. Studies that used functional scores only or nonvalidated subjective ratings of swallowing ability as outcome measures were excluded.

| Risk-of-bias assessment
The risks of bias in included studies were evaluated using version 2 of the Cochrane risk-of-bias tool for randomized trials (RoB 2). 26 Several domains of risk of bias were assessed, including bias arising from the randomization process, deviation from intended interventions, missing outcome data, measurement of outcome, and selection of reported results. The assessments were performed by two reviewers independently. Any disagreements on the judgments were discussed and resolved among all authors by consensus.

| Data extraction
Data extraction and synthesis were performed by the first and the second authors independently. Disagreements were resolved by consensus among all authors. The data extracted included the following: demographic information of participants (age and patient characteristics), treatment protocol (content, intensity, and duration), outcomes in terms of mean [standard deviation; SD] or mean [95% confidence interval; 95% CI], and sample sizes. If data were not provided, we attempted to contact the corresponding authors. If data were presented in figures and raw data was not obtainable from the authors, an online plot digitizer program (WebPlotDigitizer 4.3; Reports sought for retrieval (n = 2)

F I G U R E 1 Flow diagram for study identification and inclusion.
Reports not retrieved (n = 0) https://apps.autom eris.io/wpd/; USA) was used to extract graphic data. If data were not obtainable for quantification and analysis despite these attempts, the study was excluded from the review.

| Data synthesis and statistical analysis
Review Manager 5.4 software program (RevMan; Cochrane Collaboration, Oxford) was used to perform the statistical metaanalysis. The effect sizes of treatments were determined by comparing the treatment outcomes against that of the comparators.
Standardized mean differences (SMD) with 95% CI were used as effect measures, which were calculated based on group sizes, group mean differences and pooled SDs. Pooled SD was calculated using the following formula 27 : An SMD of 0.2 represented a small effect, 0.5 a medium effect, and 0.8 a large effect. 27 Heterogeneity was assessed with Cochrane's Q statistic and I 2 test. Heterogeneity was considered substantial when p < 0.05. An I 2 value between 0% to 40% meant that the heterogeneity might not be important; 30% to 60% represented moderate heterogeneity; 50% to 90% represented substantial heterogeneity; and 75% to 100% represented considerable heterogeneity. 28 Subgroup analyses based on the type of treatments were performed to explore the possible causes of heterogeneity. Sensitivity analyses were carried out by eliminating studies with relatively high risks of bias or cross-over study designs to assess the robustness of the synthesized results. The results were presented in the form of forest plots. Potential publication bias was assessed using Egger's test if more than 10 studies were included for the meta-analysis.

| RE SULTS
The identification and screening of studies is represented in Figure 1.
We identified 187 studies after an initial search of the five electronic

| Study characteristics
The included studies evaluated two categories of dysphagia treatments, including behavioral treatments and stimulation treatments.
No eligible studies for pharmacological treatments were identified.
Although most studies reported the immediate treatment effects, the duration of treatment varied across studies, ranging from a single session to 6 weeks. Most studies had some risk of bias. Table 1 summarizes the characteristics of included studies.

| Participants
The total number of patients included in this meta-analysis was 286.

| Behavioral treatments
Among all included studies, three investigated the effects of behavioral treatments, including expiratory muscle strength training (EMST) 30,31 and video-assisted swallowing therapy (VAST). 32 EMST is a device-facilitated behavioral exercise that aims to increase the force generation capacity of pharyngeal muscles. 33 Patients were asked to blow against resistance into a handheld device with built-in one-way spring-loaded valve. Video-assisted

| Stimulation treatments
Six included studies evaluated various forms of stimulation. [34][35][36][37][38][39] In general, these treatments can be classified into central and peripheral stimulation. Four studies investigated the effects of central stimulation, which stimulates different parts of the brain. Two of these studies evaluated the effects of deep brain stimulation (DBS), 35,38 whereas the other two studied the effects of repetitive transcranial stimulation (rTMS). 34 rTMS is a noninvasive brain stimulation that modulates brain activities through electromagnetic induction. 41 Some studies suggested that rTMS can induce long-lasting neural plasticity changes, which are critical for functional recovery following brain disruptions, in  the swallowing (pharyngeal) motor cortex and improve swallowing function. 42,43 Three studies investigated different forms of peripheral stimulation, which deliver stimuli directly to peripheral nerves or structures that are involved in swallowing, 36,37,39  It is believed that acupuncture can improve "Qi," which is the pathway of energy flow, throughout the body and cure diseases caused by disruptions of "Qi." Acupuncture may be considered a form of neurostimulation as functional neuroimaging studies have reported regionally specific and quantifiable effects on brain structures. 49 However, its underlying mechanisms and effects on swallowing functions remain controversial.

| Pharmacological treatments
No eligible RCTs were identified from the systematic search and assessment. Therefore, meta-analysis on these treatments was not carried out.

| Duration of treatment
The duration of treatment ranged from a single session to 6 weeks.
Except for deep brain stimulation and one study on 2 forms of rTMS and PES, 36 all treatments employed a 5-day-per-week schedule, with a median duration of 4 weeks.

| Outcome measures
Two studies used fibreoptic endoscopic evaluation of swallowing (FEES) scores as an outcome measure. 30 the Penetration-Aspiration Scale (PAS), which is a clinical rating scale used to describe the passage of bolus and hence the occurrence of penetration or aspiration. 50 The remaining study measured the timings of swallowing events in terms of swallowing response time. 37

| Adverse effects
One study on DBS reported that four patients dropped out prematurely due to side effects of simultaneous stimulation of the subthalamic nucleus and substantia nigra. 35 The side effects included worsening of motor function, a lack of beneficial effects of levodopa, akathisia, a general uncomfortable feeling, aggressiveness, and increased confusion and hallucinations. One study on rTMS reported that a patient in the experimental group developed headaches and insomnia. 34 Other studies on rTMS, PES, and EMST reported no adverse effects. 30,36 The remaining studies did not document any adverse effects.

| Risk-of-bias assessment
The results of the risk-of-bias assessment are presented in Figure 2.
All but one study had some concerns as regard the overall risk of bias. The most common concern was related to the selection of reported results. This arose from the lack of information regarding the prespecified outcome measurements and analyses for all studies. Other concerns with respect to potential bias included inadequate details of randomization method, no or insufficient detail regarding allocation concealment procedures, missing outcome data and no or unclear blinding of outcome assessors. One study had a high risk of missing data because the results from VFSS (PAS scores) were available from only half of the participants.

| Meta-analysis
Overall, dysphagia treatments for PD patients showed a significant Sensitivity analyses were carried out by eliminating (1) a study with high risk of bias associated with missing data 35

and (2) studies
with cross-over designs. No substantial differences in the results were observed. Therefore, the primary analysis was considered robust. Egger's test was not conducted because <10 studies were included.

| DISCUSS ION
To our knowledge, this is the first systematic review and meta- Similarly, in our review, only nine RCT studies fulfilled our inclusion and exclusion criteria. Some treatments (VAST, PES, NMES, and acupuncture) only had one eligible RCT for meta-analysis. Surprisingly, although PD patients are often given dopaminergic medications for their movement disorders, and some case-controlled and nonrandomized trials have suggested that these drugs may impact swallowing function, 9 we found that no RCTs have been conducted on their effects on swallowing. This lack of high-quality evidence on pharmacological treatments for PD-related dysphagia has also been recognized in previous systematic reviews. 9,20 Study design, treatment durations, patient characteristics, and disease severity varied across studies. Moreover, the quality of included studies was low with concerns regarding the risk of bias resulting from randomization, blinding of participants or study personnel, or missing data.
Most studies did not adequately describe the procedures for ran-  Parkinson's disease is a neurodegenerative disease, in which the progression of pathology can last for more than 15 years. 54 In this review, we found that all included studies evaluated only shortterm (up to 2 months) effects. It is not yet known whether these dysphagia treatments could provide a longer-term benefit, or the swallowing function of PD patients will still deteriorate eventually.
Moreover, there is limited research around combination of treatments, and repeated courses of treatments over of a few months to investigate the maintenance of any benefits observed with one course of treatment. These areas should be explored in future studies to better understand the long-term and maintenance effects of these treatments.
Our review has several limitations. We limited our analyses to outcomes that were based on instrumental assessments only. This may lead to underestimation of effects reported by subjective clinical evaluation or self-reports of dysphagia. However, studies have suggested that PD-related dysphagia is more frequently diagnosed with objective than subjective assessments, as objective assessments can detect subclinical dysphagia. 2 Therefore, it would be reasonable to analyze outcomes based on more sensitive evaluation methods. Another limitation is the small number of included studies, often with very small numbers of patients recruited. Some treatments only had one eligible RCT for analysis, which makes it difficult to draw definitive conclusions on the efficacy of these treatments.
However, despite these limitations, our meta-analysis was still able to find treatment effects, which are an important step in generating new questions and future research.

| CON CLUS IONS
In summary, our systematic review and meta-analysis revealed that dysphagia treatments, particularly stimulation treatments, may potentially improve swallowing efficiency and safety in PD patients.
However, the level of evidence for their efficacy is low with a lack of high-quality RCTs in the literature. Future large-scale, multicenter, and well-designed RCTs are warranted to evaluate their benefits for PD patients with dysphagia.

AUTH O R CO NTR I B UTI O N S
IC performed data extraction, synthesis, analysis, and interpretation, and drafted and critically revised the manuscript. AS performed data extraction, synthesis, analysis, and interpretation, and critically revised the manuscript. SH conceptualize the study, interpreted the data, and critically revised the manuscript. All authors approved the final version of the manuscript before submission.

FU N D I N G I N FO R M ATI O N
This research did not receive any funding from agencies in the public, commercial, or not-for-profit sectors.

D I SCLOS U R E S
SH is a board director and chief scientific officer of Phagenesis Ltd.,