Effects of pharmacological agents for neurogenic oropharyngeal dysphagia: A systematic review and meta‐analysis

Abstract Background This systematic review and meta‐analysis aimed to evaluate the effects of pharmacological agents for neurogenic oropharyngeal dysphagia based on evidence from randomized controlled trials (RCTs). Methods Electronic databases were systematically searched between January 1970 and March 2021. Two reviewers independently extracted and synthesized the data. The outcome measure was changed in (any) relevant clinical swallowing‐related characteristics. Key results Data from 2186 dysphagic patients were collected from 14 RCT studies across a range of pharmacotherapies. The pooled effect size of transient receptor potential (TRP) channel agonists was large compared to placebo interventions (SMD[95%CI] =1.27[0.74,1.80], p < 0.001; I 2 = 79%). Data were limited for other pharmacological agents and the overall pooled effect size of these agents was non‐significant (SMD [95% CI] =0.25 [−0.24, 0.73]; p = 0.31; I 2 = 85%). When analyzed separately, large effect sizes were observed with Nifedipine (SMD[95%CI] =1.13[0.09,2.18]; p = 0.03) and Metoclopramide (SMD[95%CI] =1.68[1.08,2.27]; p < 0.001). By contrast, the effects of angiotensin‐converting enzyme (ACE) inhibitors (SMD[95%CI] = −0.67[−2.32,0.99]; p = 0.43; I 2 = 61%), Physostigmine (SMD[95%CI] = −0.05[−1.03,0.93]; p = 0.92) and Glyceryl Trinitrate (GTN) (SMD [95% CI] = −0.01 [−0.11, 0.08]; p = 0.78) were non‐significant. Within stroke patients, subgroup analysis showed that TRP channel agonists had a moderate pooled effect size (SMD[95%CI] =0.74[0.10,1.39]; p = 0.02; I 2 = 82%) whereas the effects of other agents were non‐significant (SMD[95%CI] =0.40[−0.04,0.84]; p = 0.07; I 2 = 87%). Conclusions & Inferences Our results showed that TRP channel agonists, Nifedipine and Metoclopromide may be beneficial for neurogenic dysphagic patients. Large scale, multicenter clinical trials are warranted to fully explore their therapeutic effects on swallowing.


| INTRODUC TI ON
Dysphagia is a symptom referring to difficulties in the passage of food or liquid from the mouth, through pharynx and esophagus, to the stomach. 1 It can be anatomically classified into oropharyngeal dysphagia and esophageal dysphagia. Dysphagia affects approximately 56 million people worldwide 2 and is prevalent among patients with stroke (8%-80%), Parkinson's disease (11%-81%) and traumatic brain injury (27%-30%), as well as community dwelling elderly people (11%-34%). [3][4][5] Malnutrition, dehydration, aspiration pneumonia, prolonged hospital stay, mealtime anxiety and increased mortality are common physical and psychosocial consequences of dysphagia. [6][7][8][9] Moreover, the cost of healthcare resources is likely to be substantial for patients and to society in general due to their complex nature. 10,11 Dysphagia treatments are generally focused on improving safety and efficiency of swallowing. They can be compensatory, such as modifications of diet consistency or feeding posture, or rehabilitative, such as strength or skill training exercises for swallowing musculature. 12 Rehabilitative interventions also include acupuncture, peripheral sensory stimulation through thermal, tactile or electrical (neuromuscular or pharyngeal) stimulation or non-invasive brain stimulation including repetitive transcranial magnetic stimulation (rTMS) or transcranial electrical stimulation (TES). 12 Of importance to this field, pharmacological agents are a potential management option for dysphagia and yet they have received relatively little attention compared to other treatments. These agents either stimulate swallowing-related neural pathways in the peripheral or central nervous systems or directly modifying muscular function. 13 To date, the drug classes that have been studied in the area of swallowing and oropharyngeal dysphagia include transient receptor potential vanilloid 1 (TRPV1) agonists, [14][15][16][17][18][19][20] transient receptor potential ankyrin 1 (TRPA1) agonists, 21 transient receptor potential melastatin 8 (TRPM8) agonists, 22 levodopa, [23][24][25] other dopaminergic agents, 26 calcium blocking agents, 27,28 dopamine D2 receptor antagonists, 29 angiotensin-converting enzyme (ACE) inhibitors, 30 beta blockers, 31 nitric oxide donors 32 and acetylcholinesterase inhibitors. 33 Studies have suggested that these drugs may improve the swallowing reflex or reduce incidence of aspiration pneumonia in dysphagic patients. However, the underlying therapeutic mechanisms of action of these drugs are poorly understood. One mechanism is stimulation of afferent neural pathways for swallowing, for example receptors (TRPV1, TRPA1 and TRPM8) located in the oropharynx, 34 through TRP channel agonists. [14][15][16][17][18][19][20][21][22] Another mechanism involves increasing the level of or decreasing degradation of substance P, which is a neuropeptide known to enhance the swallow reflex, 35 through capsaicin, ACE inhibitors or beta blockers. 17,30,31,36 Levodopa and dopaminergic agents may improve swallowing through improving dopamine metabolism. [23][24][25][26] Some studies have also suggested that treating coexisting esophageal dysphagia or facilitating stroke recovery may result in overall improvement in swallowing function. [27][28][29]32 Finally, physostigmine may improve swallowing in patients with progressive supranuclear palsy through cholinergic stimulation actions, but no significant effect has been reported. 33 Given the scarce knowledge of the therapeutic potentials of pharmacological agents, this systematic review and meta-analysis aimed to analyze their group effects on swallowing-related outcomes in neurogenic oropharyngeal dysphagia from existing randomized controlled trials (RCTs). Further subgroup analysis was carried out to analyze the effects of these agents on stroke patients as strokes are the commonest cause of neurogenic dysphagia. The findings from our meta-analysis should provide insights into the future research directions on pharmacological interventions for dysphagia.

| MATERIAL S AND ME THODS
This review of data followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. Two reviewers performed the search for studies, data extraction and risk of bias assessment independently. Data synthesis was carried out by one reviewer and verified by a second reviewer. Disagreements were resolved by consensus among all authors.

| Participants
Studies with adult patients with neurogenic oropharyngeal dysphagia (ie, dysphagia resulted from damage or deterioration of the central or peripheral nervous system) as determined clinically or through validated self-report questionnaires regardless of the time of onset were included. Studies with healthy volunteers, patients without dysphagia or patients with esophageal dysphagia only were not considered. For studies that included both patients with and without dysphagia, only data from patients who were considered dysphagic, based on modified diet or at an elevated risk of aspiration pneumonia were extracted and analyzed.

| Interventions
We included studies that compared pharmaceutical interventions with placebo intervention. Trials with multiple interventions (eg, co-administration of pharmacological agents and other swallowing therapies) were eligible if the study groups only differed by the use of the target pharmaceutical intervention of interest.

| Outcomes
Study outcomes related to swallowing, which included swallowing physiology measurement, clinical swallowing function ratings, functional dysphagia symptom scales or health outcomes related to swallowing functions, for example incidence of aspiration pneumonia, were included for comparisons. Studies that used non-validated subjective rating of swallowing ability as an outcome measure were excluded.

| Data extraction
The data extracted included: demographic information of participants (age and patient characteristics), intervention protocol (drug strength and dosage regimen), outcomes (mean [standard deviation; SD] or mean [95% confidence interval; 95% CI]) and sample sizes. For studies with multiple outcome measures, the most relevant primary swallowing-related outcome was used. 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 digitalizer program (WebPlotDigitizer 4.3; 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.

| Risk of bias assessment
Seven domains of risk of bias of RCTs were evaluated using the Cochrane Collaboration's tool for assessing risk of bias. 37 These include random sequence generation, allocation concealment, blinding of participants and personnel, blinding of outcome assessment, incomplete data, selective reporting and other sources of bias. Two reviewers rated the risk of bias of the included studies independently. Any disagreement on the judgements was discussed and resolved among all authors.   Pooled SD was calculated using the following formula 38 :

| Statistical analysis
Confidence intervals (CIs) were converted to SDs as suggested in the Cochrane Handbook. 37 For outcome measures that increase with disease severity, the mean values were multiplied by −1. Treatment effects for continuous outcomes were analyzed as standardized mean difference (SMD) with 95% CI. A weighted average of SMD across studies was computed using a random effects model analysis. The significance level was set at p < 0.05 and the effect sizes were presented as SMD [95% CI]. For the interpretation of effect sizes, SMD of 0.2 represented a small effect, 0.5 a moderate effect, and 0.8 a large effect. 38 Heterogeneity was assessed with Cochrane's Q statistic and I 2 test in which heterogeneity was considered substantial with p < 0.05 and I 2 higher than 50%.

| Study characteristics
The included studies were all published between 1998 and 2020.

| Risk of bias assessment
The risk of bias assessment result is presented in Figures 2 and   3. Most studies had a low risk of selection and detection bias.
Approximately half of the included studies had a high risk of performance bias due to the lack of blinding of personnel or participants. Attribution bias was high in 25% of the studies because of dropouts or deaths during the studies. Reporting bias was low for all but one study 18 which did not report the outcomes of their control group in one of their sub-studies. There was insufficient information to determine other risks so these could not be further quantified.

| Outcome measures
The

| Adverse events
Regarding serious adverse events, one study reported significantly higher mortality in the intervention (Lisinopril) group, which led to the premature termination of the study. 30 Worsening of heart failure, flushing and giddiness were reported in the study with Nifedipine, 27 although the relationships between these events and Nifedipine were not discussed by the authors. The GTN study reported that patients in the intervention group were more likely to have headache or clinical hypotension than the control group. 32 No serious adverse events were reported with other pharmacological agents.

| Dosage
The daily dosage ranged from once to six times whereas the overall course of intervention ranged from one to 28 days.

| Effects of pharmacological agents compared to placebo interventions
Among all drug classes, TRP channel agonists were studied most extensively with 8 RCTs. Therefore, a pooled effect size was com-  For the other pharmacological agents ( Figure 5), only single or dual studies were evaluable, making interpretation less meaningful.
Overall, the pooled effect size for these agents was non-significant

| Effects of pharmacological agents on poststroke dysphagia
Given that stroke was the most studied disease group among all included studies (67%), a further analysis was carried out ( Figure 6).

| DISCUSS ION
This systematic review and meta-analysis evaluated the effects of pharmacological agents on swallowing-related outcomes in (neurogenic) dysphagic patients. Among all drug classes, TRP channel agonists, predominantly capsaicin (TRPV1 agonist), were most extensively studied. We found that overall, TRPV1, TRPA1 and TRPM8 agonists are superior to placebo interventions with large effect sizes. Preliminary neurophysiological evidence appears to support the hypothesis that functional changes induced by TRP agonists are cen- Apart from sending sensory impulses to the central nervous system, TRPV1 agonists may modulate swallowing through releasing substance P, which is a neuropeptide that enhances cough reflex. 53 Studies have found that reduced levels of substance P are associated with an increased risk of aspiration pneumonia in elderly patients, 35 stroke patients 54,55 and patients with Parkinson's disease. 56 Given that an increase in serum substance P level after capsaicin treatment has been reported in some RCTs, 14,15 it is possible that this neuropeptide may play a role in the observed improvements in swallowing function. The mechanisms of TRPV1 agonists on the release of substance P and the relationship between substance P and swallowing function are not fully understood. In healthy volunteers, Suntrup-Krueger et al. 52 found that the effects on salivary substance P level are dose dependent, where an increase was only detected with high dose (50μM) but not low dose (10μM) oral capsaicin. In elderly patients with dysphagia, a recent RCT found that increased levels of substance P is associated with improvement in swallowing efficiency following capsaicin treatment. 17 Some studies have explored the relationship between substance P and the physiology of swallowing.
Tomsen et al. 57 found that elderly patients with oropharyngeal dysphagia showed impaired pharyngeal sensitivity compared to healthy volunteers and substance P level was negatively correlated with pharyngeal sensory threshold. Moreover, in acute stroke patients, low substance P level was associated with low frequency of spontaneous swallowing and increased incidence of pneumonia. 55 These findings suggested that substance P level is closely related to swallowing performance and may be a potential marker for pharyngeal sensitivity or stroke-related aspiration pneumonia.
Previous reports have suggested that ACE inhibitors may be beneficial to dysphagic patients by reducing degradation and inactivation of substance P. 58 Arai et al. 54 suggested that Imidapril hydrochloride may increase substance P and reduce the risk of silent aspiration in stroke patients, although the effect size was nonsignificant in our meta-analysis. In contrast, Lee et al. 30  for the former two agents, but underlying mechanisms remain largely speculative. 27,29 Nifedipine is a calcium blocking agent that can be used to alleviate chest pain and rapidly lower blood pressure. 59 Perez et al. 27 postulated that Nifedipine may improve pharyngeal dysphagia through reducing coexisting esophageal spasm or through global enhancement on stroke recovery. Metoclopramide is a dopamine antagonist used to reduce nausea and vomiting. 60 Warusevitane et al. 29 suggested several possible mechanisms of Metoclopramide in reducing incidence of aspiration pneumonia.
These include reduced regurgitation through increasing the tone of lower esophageal sphincter and accelerating gastric emptying in patients with nasogastric tube-feeding. although the effects of these agonists appeared to be smaller in stroke patients than in patients with neurogenic dysphagia, the difference may be attributed to the smaller number of trials in stroke patients. Moreover, a mixed population of stroke patients with different severity and chronicity were studied in these RCTs. A recent meta-analysis showed that the effects of neurostimulation treatments varied according to the chronicity of stroke. 65 Therefore, it is plausible that the stroke characteristics may have influenced the responsiveness to TRP channel agonists, hence limiting their treatment efficacy in stroke patients.
The quality of studies included in our meta-analysis was considered moderate due to the high risk of performance bias.
Approximately half of the included studies did not have reliable blinding of participants or personnel. These were primarily studies with TRP channel agonists. While blinding is ideally done by delivering a placebo treatment that appears identical to the active treatment, it can be challenging for some compounds with strong, distinctive taste and smell such as TRP channel agonists.
Moreover, placebo treatment may not be available from manufacturers 62 such that a control condition that resembles the active treatment needs to be made from other materials, which may influence its validity. The use of an active control may minimize performance bias, but in some cases, single-blinded designs may be unavoidable.
Our review is limited by the small number of studies. For some drug classes, only one RCT was eligible for analysis, making it difficult to draw any definitive conclusions regarding their efficacies.
Given the small number, the risk of publication bias cannot be evaluated. Moreover, only English studies were included for analysis.
Lastly, patient characteristics, outcome measures and intervention protocols of included studies were highly heterogeneous. Therefore, our results must be interpreted with some caution.
In conclusion, our systematic review found that TRPV1, TRPA1 fully explore the potential of these agents.