Keywords
gastroesophageal reflux disease - manual therapy - tuina - massage - systematic review
- meta-analysis
Introduction
Gastroesophageal reflux disease (GERD) refers to an illness in which stomach contents
reflux into the esophagus, causing discomfort or complications like peptic stricture,
esophageal ulceration, Barrett's esophagus (BE), or esophageal adenocarcinoma.[1] Common symptoms of GERD include heartburn and reflux. However, when the reflux spreads
to adjacent tissues such as the mouth, larynx, lungs, and heart, it may cause dysphagia,
dental erosion, laryngitis, chronic cough, asthma, chest pain, or occur in isolation.[2]
[3] Two subgroups of GERD exist, including nonerosive gastroesophageal reflux disease
(NERD) and reflux esophagitis (RE), which vary based on the degree of mucosal damage.[4] GERD is a widespread disorder, with an average global prevalence of approximately
13 and 8.7% in China,[5] and its incidence is increasing yearly.[6]
[7] In the United States, it is the most common gastrointestinal disease with GERD becoming
a risk factor for tumors such as BE and esophageal adenocarcinoma, although the mortality
rate is not high.[8]
[9] Furthermore, the treatment of GERD is expensive, mainly related to the long-term
use of proton pump agents (PPIs).[10]
[11]
[12] PPIs are the recommended first-line therapy for GERD and can heal esophagitis in
72 to 83% of patients (compared with 18–20% for placebo).[13] However, the standard dose of PPI can only resolve the heartburn symptoms in 37–61%
of patients with nonerosive esophagitis, while the treatment rate is low in patients
with atypical GERD. Therefore, the efficacy of PPIs depends on the type of GERD disease
present.[14]
[15] Besides the cost burden, using PPIs involves difficulties in compliance and the
inability to rule out associations with polyps, mucosal degeneration, and osteoporosis.[16]
[17] Other medication options such as H2 receptor antagonists, potassium-competitive
acid blockers, antacids and gastric stimulants, etc., and surgery may have limited
suitability for certain populations, long-term efficacy, or side effects.[15]
[18]
[19]
In the past decade, drug development has experienced a significant decline, while
research on novel nonpharmaceutical therapeutic technologies has increased dramatically.[20] Manual therapy, such as Chinese tuina or therapeutic massage, which is mediated
by the limbs, has been widely used in some gastrointestinal diseases, including GERD,
under the guidance of the basic theory of traditional Chinese medicine (TCM) or the
theory of Western neuromuscular anatomy. Modern medical mechanisms for the effectiveness
of massage for visceral diseases often involve enhancing blood circulation in local
tissues, increasing parasympathetic excitability, decreasing neuromuscular excitability,
and regulating hormone levels.[21] Clinical studies have also revealed that osteopathic visceral therapy can increase
the pressure of the lower esophageal sphincter (LES), which can improve the symptoms
of GERD.[22]
[23] Additionally, in preterm infants with GERD, acupressure has been found to be more
effective in increasing LES pressure and reducing reflux than conventional treatment.[24] Combining acupressure with pharmacological therapy has also been shown to have a
significant advantage over drug therapy alone in terms of symptom improvement and
endoscopic esophageal mucosal repair.[25] Despite these findings, there is no consensus on the efficacy of manual therapy
for the treatment of GERD. Therefore, this study aims to systematically evaluate the
efficacy and safety of manual therapy for GERD by screening relevant RCTs, with the
goal of providing evidence to support medical decisions.
Methods
Search Strategy
Chinese National Knowledge Infrastructure, China Biology Medicine Database, WanFang,
VIP, PubMed, The Cochrane Library, OVID Medline, and EMbase databases were searched
by computer. The search period was from the establishment of the database to November
30, 2022. Chinese search terms included [ “Tui Na” OR “An Mo (massage)” OR “Shou Fa
(manipulation)” OR “Shou Dong Zhi Liao (manual technique)” OR “Xue Wei An Ya (acupressure)”]
AND [(“Wei Shi Guan Fan Liu Bing (gastroesophageal reflux disease)” OR “E Ni (hiccup)”
OR “Shao Xin (heartburn)” OR “Fan Suan (acid reflux)”] AND [ (“Sui Ji Dui Zhao Shi
Yan (randomized controlled trial)” OR “Lin Chuang Yan Jiu (Clinical study)”)], English
search terms include (“massage” OR “tuina” OR “anmo” OR “manipulat*” OR “chiropractic”
OR “manual technique” OR “manual therap*” OR “acupressure”) AND (“gastroesophageal
reflux” OR “GERD” OR “heartburn” OR “hiccup”) AND (“randomized controlled trial” OR
“controlled clinical trial”). The search method was a combination of subject terms
and keywords, which would be moderately adjusted according to the database. [Table 1] displays the full literature retrieval technique with OVID Medline as an example.
Table 1
Searching strategy
|
Search
|
Query
|
Results
|
|
1
|
exp Massage/
|
6,794
|
|
2
|
tuina.mp.
|
255
|
|
3
|
anmo.mp.
|
7
|
|
4
|
massage therap*.mp.
|
1,599
|
|
5
|
manipulat*.mp.
|
229,989
|
|
6
|
Zone Therap*.mp.
|
50
|
|
7
|
chiropractic.mp.
|
7,363
|
|
8
|
manual technique.mp.
|
524
|
|
9
|
manual therap*.mp.
|
3,365
|
|
10
|
acupressure.mp.
|
1,680
|
|
11
|
1 or 2 or 3 or 4 or 5 or 6 or 7 or 8 or 9 or 10
|
245,018
|
|
12
|
Gastroesophageal Reflux/
|
27,841
|
|
13
|
gastroesophageal reflux disease.mp.
|
13,017
|
|
14
|
GERD.mp.
|
10,099
|
|
15
|
heartburn.mp.
|
6,247
|
|
16
|
hiccup.mp.
|
1,423
|
|
17
|
12 or 13 or 14 or 15 or 16
|
37,360
|
|
18
|
Randomized controlled trial. pt.
|
581,153
|
|
19
|
controlled clinical trial. pt.
|
95,105
|
|
20
|
randomized.ab.
|
583,079
|
|
21
|
placebo.ab.
|
233,428
|
|
22
|
drug therapy.fs.
|
2,549,487
|
|
23
|
randomly.ab.
|
395,949
|
|
24
|
trial.ab.
|
624,493
|
|
25
|
groups.ab.
|
2,437,379
|
|
26
|
18 or 19 or 20 or 21 or 22 or 23 or 24 or 25
|
5,516,267
|
|
27
|
exp animals/ not humans.sh.
|
5,066,999
|
|
28
|
26 not 27
|
4,808,227
|
|
29
|
11 and 17 and 28
|
28
|
Inclusion Criteria
(1) Types of studies: Clinical randomized or quasirandomized controlled trials (RCTs)
on manual therapy for GERD, limited to Chinese and English, and blinded or nonblinded
in article design were eligible. (2) Participants: Studies with a clinical diagnosis
of GERD and stating clear diagnostic criteria were included. Disease types included
but were not limited to NERD, RE (erosive esophagitis), and BE. There were no restrictions
on the patient's gender, age, race, occupation, course of disease, and TCM patterns.
(3) Interventions and comparators: The experimental group underwent tuina or massage
techniques alone (regardless of type, site, duration, frequency, etc.) or in combination
with other therapies such as Western medicine, Chinese medicine, or other TCM external
treatments. The control group used the same Western medicine, Chinese medicine, or
other TCM external treatments as the intervention group except for manual therapy,
or the control group was simply a sham manual therapy group. (4) Outcomes: The primary
outcome indicator was the total treatment effectivity rate, and the secondary outcome
indicators were the Reflux Disease Questionnaire (RDQ) score and adverse events.
Exclusion Criteria
Exclusion criteria were as follows: (1) studies with unreasonable diagnostic methods
or criteria; (2) articles with incomplete data that could not be extracted; (3) repeatedly
published literature; (4) clinical trials that were not RCTs; RCTs using before–after
control in the same patients; and other kinds of literature works such as reviews,
case reports, systematic reviews, animal experiments, and conference papers.
Literature Screening and Data Extraction
Two researchers conducted an independent and consecutive screening of the literature
based on predetermined research criteria and cross-checked their results. Any differences
of opinion were resolved through discussion. In cases where disagreements persisted,
a third evaluator was consulted. The PRISMA 2020 flow diagram was used to document
the literature screening process, and data were extracted independently by both researchers.
Any missing data were obtained from the authors, and if complete data information
could not be obtained, the literature was discarded. Extractions included general
information (such as the title, authors, source, and date of publication of the literature),
study characteristics (such as study site, characteristics of the study population,
method of study design, and specific interventions in the experimental and control
groups), study outcomes (such as outcome indicators, duration of treatment, follow-up
time, adverse effects, and shedding), and other relevant and important variables.
The literature information was collected and presented using a characteristic table.
Risk of Bias Assessment
In this study, two researchers utilized version 5.0 of the Cochrane Collaboration
Risk of Bias Assessment Tool to evaluate the risk of bias in RCTs. The tool includes
seven entries, namely: random assignment method, allocation concealment, blinding
of the participants and doctors who performed the intervention, blinding of the outcome
evaluators, completeness of the data counted and final conclusions, selective outcome
reporting, and other bias. The potential for bias in each item was rated as either
“high” “low” or “unclear”. Two evaluators initially assessed the items, which were
then reviewed by a third evaluator. Any disagreements were resolved through a tripartite
discussion or by the third evaluator. If there was uncertainty regarding the level
of risk, the authors were contacted for clarification.
Statistical Analysis
The meta-analysis used RevMan 5.4.0 software provided by the Cochrane Collaboration
Network. Dichotomous variables were analyzed using odds ratio (OR) with 95% confidence
intervals (CIs), while continuous variables were assessed using mean difference with
95% CI. A statistical inspection standard of p < 0.05 was employed. The combined effects of outcomes were demonstrated in a forest
plot. Heterogeneity was examined using the χ2 test with a test level of p < 0.1, and I2 was utilized to quantify heterogeneity, with a fixed-effects model utilized for smaller
heterogeneity (p ≥ 0.1 and I
2 ≤ 50%) and a random-effects model used for larger heterogeneity (p < 0.1 and I
2 > 50%). If heterogeneity was still significant after statistical heterogeneity treatment,
further analysis such as subgroup or sensitivity analysis was conducted to identify
the source of heterogeneity.
Publication Bias Assessment
As more than 10 RCTs were included, we drew funnel plots as a means to assess potential
publication bias.
Level of Evidence
The Grading of Recommendations, Assessment, Development and Evaluation (GRADE) tool[26] was used to rate the caliber of the data in this meta-analysis. The degree of the
level of evidence for each outcome indicator was rated as high, moderate, low, and
extremely low. We decreased levels in accordance with the following five criteria:
risk of bias, inconsistency, imprecision, indirectness, and publication bias. The
degree of evidence was evaluated independently by two evaluators. Disagreements were
handled by a third investigator or discussed by both sides.
Results
Study Selections
The initial search yielded 1,565 studies, and 573 were later screened out by the filter.
The software NoteExpress 3.5 was then used to remove 309 duplicates. By browsing through
the titles and abstracts, 40 relevant papers were finally selected. Among these, four
articles were unavailable for full text and could not be read, leaving 36 articles
remaining. However, 5 articles had incomplete data, 1 study had unknown efficacy evaluation
criteria, 13 studies had outcome indicators that were inconsistent, 2 studies had
unclear diagnostic criteria, 1 study had noncompliant intervention, 1 study had unreasonable
control group settings, and 2 studies were non-RCT. Finally, 11 studies that met the
criteria were included in the meta-analysis.[25]
[27]
[28]
[29]
[30]
[31]
[32]
[33]
[34]
[35]
[36]. As shown in [Fig. 1].
Fig. 1 Flowchart of literature screening.
Basic Characteristics of the Included Studies
Types of Participants
A total of 1,107 patients were included in our research, with 558 in the experimental
group and 549 in the control group. The number of patients in each study ranged from
30 to 200, and the age of patients varied from preterm infants to adults up to 70
years old. One of the studies included three groups, but only two groups satisfied
the inclusion criteria, so we extracted data from these two groups only. [Table 2] presents the details.
Table 2
Basic characteristics of included RCTs
|
Study ID
|
Diagnosis
|
Design
|
Sample size
|
Age
|
Duration of disease
|
Sex (Male/Female)
|
Interventions
|
Intervention time
|
Outcome measurements
|
Results
|
Adverse events
|
Droup out
|
Follow-up
|
|
E
|
C
|
E
|
C
|
E
|
C
|
E
|
C
|
E
|
C
|
|
Chen 2018[31]
|
Gastroesophageal reflux disease
|
RCT
|
30
|
30
|
50.96 ± 16.02 y
|
49.00 ± 14.88 y
|
6.72 ± 5.47 y
|
5.88 ± 4.53 y
|
13/17
|
16/14
|
Acupoint finger pressure therapy
|
PPI
|
2 wk
|
① ②
|
The decrease in the scores of RDQ scale of the experimental group was greater than
that in the control group (p = 0.008). For the total effective rate, the experimental group (89.9%) was significantly
higher than the control group (66.7%; p < 0.05).
|
NR
|
Non
|
1 mo
|
|
Yan et al 2015[36]
|
Gastric volvulus (GV) with gastroesophageal reflux (GER)-induced pneumonia
|
RCT
|
E1 = 59
E2 = 29
|
C1 = 56
C2 = 27
|
E1:2.1 ± 1.1 mo
E2:2.0 ± 1.0 mo
|
C1:2.0 ± 1.0 mo
C2:2.0 ± 1.1 mo
|
NR
|
NR
|
E1:38/22
E2:21/9
|
C1:39/21
C2:20/10
|
Basic treatment + massage treatment
|
Basic treatment
|
2 wk
|
①
|
Massage treatment groups showed a significantly higher percentage of cure and total
effect (p < 0.05; p < 0.01) and a lower prevalence of recurrence (but with no statistic difference, p > 0.05) than basic treatment groups.
|
Some cases with local redness or mild lesions
|
E:2
C:7
|
NR
|
|
Liu 2018[27]
|
Gastroesophageal reflux disease
|
RCT
|
29
|
29
|
48.00 ± 14.10 y
|
46.87 ± 11.27 y
|
6.83 ± 7.27
|
6.66 ± 7.04
|
9/20
|
10/19
|
Acupoint finger pressure therapy
|
PPI and prokinetic agent
|
2 wk
|
① ②
|
The total effective rate of finger acupuncture on Back-Shu Point was similar to that
of Western medicine (p > 0.05), but for the food/acid reflux, heartburn, mouth bitter relief shows obvious
advantages (p < 0.05), and long-term effect (3 mo after the end of therapy) is superior to the
control group (p < 0.05).
|
C:2 cases with low back pain in the early treatment
E:1 case with diarrhea,1 case with constipation,1 case with skin itch
|
E:2
C:1
|
3 mo
|
|
Liu and Liu 2016[28]
|
Gastroesophageal reflux disease
|
RCT
|
40
|
40
|
12. 63 ± 4. 66 d
|
12. 63 ± 4. 66d
|
NR
|
NR
|
24/16
|
22/18
|
Basic treatment +
Tuina
|
Basic treatment
|
7 d
|
①
|
The total effective rate in the observation group was 95%,which was increased significantly
compared to 77. 50% of the controlled group (p < 0. 05).
|
NR
|
Non
|
NR
|
|
Zhang and Wei 2001[25]
|
Gastroesophageal reflux disease
|
RCT
|
100
|
100
|
NR
|
NR
|
NR
|
NR
|
NR
|
NR
|
Acupoint finger pressure therapy + prokinetic agent
|
Prokinetic agent
|
4 wk
|
①
|
The total effective rate of the treatment group was 100%, and the total effective
rate of the control group was 76%, and the difference between the efficacy of the
two groups was significant (p < 0.01). And the degree of endoscopic esophagitis or esophageal reflux and delayed
barium emptying were significantly improved before and after treatment in each group.
|
NR
|
Non
|
NR
|
|
Xie et al 2007[29]
|
Gastroesophageal reflux disease
|
RCT
|
40
|
40
|
45.98 ± 13.1 y
|
8.85 ± 13.1 y
|
7.4 ± 4.8 mo
|
27.4 ± 4.8 mo
|
11/29
|
16/24
|
Medicated acupoint finger pressure therapy
|
PPI and prokinetic agent
|
3 wk
|
①
|
After treatment, the symptom scores of the two groups were remarkably decreased (p < 0.05), but there was no significant differences. The total effective rate was 97.5%
in the treatment group,significantly higher than that in the control group (80.0%;
p < 0.05).
|
C: 4 cases with diarrhea, 3 cases with dizziness.
E: non
|
Non
|
NR
|
|
Li et al 2017[30]
|
Gastroesophageal reflux disease
|
RCT
|
70
|
67
|
Mean 57.1 y
|
Mean 57.4
|
Mean 4.1 y
|
Mean 3.8
|
31/39
|
31/36
|
Medicated acupoint finger pressure therapy
|
Chinese medicine decoction
|
4 wk
|
①
|
Comparison of clinical efficacy of three groups was made twice: the traditional Chinese
medicine group and medicine acupoint pointer therapy group (p < 0.05), the traditional Chinese medicine group and the combined group (p < 0.05), the differences were statistically significant.
|
NR
|
E:1
C:1
|
NR
|
|
Zhong et al 2021[32]
|
Reflux esophagitis and chest pain
|
RCT
|
46
|
46
|
15 ± 5.74yr
|
63 ± 5.38 y
|
49 ± 2.16 y
|
27 ± 2.3 y
|
22/24
|
21/25
|
Acupotomy therapy + Tuina
|
Acupotomy therapy
|
4 wk
|
①
|
The clinical symptom scores in the two groups after treatment were significantly ameliorated
compared with before treatment, which were lower in the treatment group (p <0.05).There was no significant difference in the effective rate between the two
groups (p > 0.05).
|
NR
|
Non
|
NR
|
|
Zhang 2017[33]
|
Non-erosive relux disease
|
RCT
|
30
|
30
|
48.33 ± 15.87 y
|
45.20 ± 15.38 y
|
5.73 ± 5.72 y
|
4.34 ± 4.58
|
13/17
|
15/15
|
Acupuncture + Tuina
|
Acupuncture
|
6 wk
|
① ②
|
The decrease in the scores of RDQ scale of the treatment group was greater than that
in the control group (p < 0.05).The total effective rate was 93.33% in the treatment group and 83.33% in
the control group (p = 0.228), which was not statistically significant.
|
NR
|
Non
|
1 mo
|
|
Liu et al 2015[34]
|
Gastroesophageal reflux disease
|
RCT
|
70
|
70
|
NR
|
NR
|
NR
|
NR
|
38/32
|
36/34
|
Acupoint finger pressure therapy + PPI and prokinetic agent
|
PPI and prokinetic agent
|
8 wk
|
①
|
The total effective rate in the observation group (90.00%) was significantly higher
than that in the control group (81.43%; p < 0.05).
|
E:1case with dizziness and somnolence.
C:1case with thirst and vertigo.
|
Non
|
NR
|
|
Martínez et al 2019[35]
|
Gastroesophageal reflux disease
|
RCT
|
15
|
15
|
49.9 ± 14.4yr
|
46.9 ± 14.8 y
|
NR
|
NR
|
3/12
|
4/11
|
Myofascial release (MFR) techniques
|
Sham myofascial release (MFR) techniques
|
1 wk
|
②
|
At week 4, patients receiving MFR showed significant improvements in symptomatology
(mean difference−1.1; 95% CI: −1.7 to −0.5), gastrointestinal quality of life (mean
difference 18.1; 95% CI: 4.8 to 31.5), and PPIs use (mean difference−97 mg; 95% CI:
−162 to −32), compared to the sham group, indicating that the application of the MFR
protocol used in this study decreased the symptoms and PPIs usage and increased the
quality of life of patients with nonerosive GERD up to 4 wk after the end of the treatment.
|
No adverse events.
|
C:1
|
4 wk
|
Abbreviations: C, control group; CI, confidence interval; E, experimental group; GERD,
gastroesophageal reflux disease; NR, not reported; PPI, proton pump inhibitor; RCT,
randomized controlled trial; RDQ, reflux disease questionnaire.
Types of Interventions and Controls
In terms of the experimental group, five studies implemented individual manual therapy,
which included acupressure, medicated acupoint finger pressure therapy, and myofascial
release techniques.[27]
[29]
[30]
[31]
[35] Two combined manual therapy and Western medicine (primarily proton pump inhibitors
and/or gastroprokinetic agents),[25]
[34] while two other studies combined tuina and conventional therapy (mostly feeding
style modifications, such as posture and diet).[28]
[36] Finally, two studies utilized chiropractic manipulation and acupotomy therapy.[32]
[33] Regarding the control group, five were treated with Western medicine,[25]
[27]
[29]
[31]
[34] one used herbal medicine,[30] two used conventional treatment,[28]
[36] and two used acupuncture therapy.[32]
[33] Only one study involved sham technique therapy.[35]
Types of Outcome Measures
Ten RCTs assessed total treatment effectiveness,[25]
[27]
[28]
[29]
[30]
[31]
[32]
[33]
[34]
[36] and four RCTs scored the RDQ.[27]
[31]
[33]
[35] Five RCTs mentioned adverse effects.[27]
[29]
[34]
[35]
[36]
Risk of Bias Assessment
(1) Randomization: Of the included literature, five studies were randomized by random
number table,[28]
[30]
[31]
[32]
[36] one study utilized the dynamic random Taves minimization method,[27] one study was randomized by calculator random numbers,[29] and two used random assignment software.[33]
[35] So they were judged as “low risk.” The other two studies only mentioned the word
“random” without specifying the grouping method and were assessed “high risk”.[25]
[34] (2) Allocation hiding: Allocation concealment was only applied in one study which
was judged “low risk”.[35] (3) Blind method: Most of the studies could not be blinded due to the specificity
of manual therapy, but one study had a sham technique group, so it was judged as low
risk.[35] (4) Completeness of outcome data: one study lacked a description of data processing
for dislodged cases and was judged to be “unclear risk”.[36] (5) Selective reporting: one study was deemed “high risk” because it lacked the
protocol-required planned outcome markers.[34] (6) All other biases were not described in detail and were considered as unknown
risk. As shown in [Figs. 2] and [3].
Fig. 2 Risk of bias graph.
Fig. 3 Risk of bias summary. Notes: 
: low risk of bias; 
: unclear risk of bias; 
: high risk of bias.
Meta-analysis
Total Effective Rate
Ten RCTs reported this outcome measure, and the forest plot demonstrated that the
total effective rate of the manual therapy group was superior to that of the control
group.[25]
[27]
[28]
[29]
[30]
[31]
[32]
[33]
[34]
[36] Heterogeneity among studies was low (I
2 = 40%, p = 0.09), and the combined effect sizes showed statistically significant differences
(OR = 4.63, 95% CI [3.01, 7.14], p < 0.00001) using a fixed effects model. The subgroup analysis based on intervention
modality revealed that the total effective rate was significantly higher in the manual
therapy group versus the Western medicine group (OR = 3.32, 95% CI [1.34, 8.26], p = 0.01), the manual therapy group versus the Chinese medicine group (OR = 3.06, 95%
CI [1.18, 7.95], p = 0.02), manual therapy + the Western medicine group versus the Western medicine
group (OR = 5.94, 95% CI [2.63, 13.41], p < 0.0001), and manual therapy + the conventional treatment group versus the conventional
treatment group (OR = 11.13, 95% CI [3.26, 38.02], p = 0.0001). However, there was no statistically significance in the manual therapy + acupuncture
treatment versus the acupuncture treatment group (OR = 2.43, 95% CI [0.71, 8.29],
p = 0.16). The results showed that manual therapy was more effective than Western medicine
or TCM decoction alone, and the combined effect of manual therapy + Western medicine/conventional
treatment was superior to that of Western medicine or conventional treatment alone.
However, there was no significant difference in the efficacy of the combination of
manual therapy + acupuncture compared to acupuncture alone. As shown in [Fig. 4].
Fig. 4 Forest plots of total effective rate.
Reflux Disease Questionnaire Scores
Four studies reported this outcome indicator.[27]
[31]
[33]
[35] Heterogeneity among studies was large (p = 0.02, I
2 = 70%), so a random effects model was employed and the final combined effect size
revealed a significant difference in RDQ scores {weighted mean difference (WMD) = −1.59,
95% CI [−2.85, −0.33], p = 0.01}. These findings suggested that the manual therapy group demonstrated a greater
improvement in RDQ scores compared to the control group. As shown in [Fig. 5].
Fig. 5 Forest plots of RDQ score.
Adverse Events
Out of eleven trials, five of them reported on adverse events.[27]
[29]
[34]
[35]
[36] According to Martínez's report,[35] there were no significant adverse reactions in the subjects during the trial. Yan's
report mentioned that some patients experienced skin redness or minor injuries during
treatment, but there was no comment on how they were managed.[36] In the Liu's report, two patients in the trial group experienced low back muscle
pain during treatment with acupoint finger pressure, while one patient in the control
group developed diarrhea, one developed constipation, and one developed pruritus during
the administration of lansoprazole enteric tablets, but all were able to continue
to participate in the study after the symptoms resolved on their own.[27] Xie's report showed that during the administration of omeprazole enteric tablets
and mosapride tablets, four patients in the control group developed diarrhea (10.0%)
and three patients developed dizziness (7.5%), but the symptoms were not significant
and did not affect the continuation of treatment.[29] Liu's report[34] recorded drowsiness and dizziness in one patient in the treatment group during acupressure
therapy combined with oral esomeprazole and mosapride tablets and dry mouth and vertigo
in one patient in the control group after oral Western medicine. However, adverse
events in both groups were not treated specifically and resolved after rest, without
affecting the completion of the course of treatment.
Subgroup Analysis
In the earlier parts of the article, a subgroup analysis of intervention modalities
was conducted. Although the heterogeneity for the entire group was only 40%, the within-group
heterogeneity was still high in some groups, indicating that the interventions were
not the source of heterogeneity. Now, subgroup analyses were conducted again, based
on age (young children vs. adults), as displayed in [Fig. 6]. The findings demonstrated that the differences in age could significantly influence
heterogeneity.
Fig. 6 Subgroup analyses of the total effective rate.
Publication Bias
[Fig. 7] presents a funnel plot analysis of studies that utilized the total effective rate
for outcome measurements. The asymmetrical shape of the graph from left to right suggested
the possibility of publication bias in the included RCTs, as indicated by the results.
Fig. 7 Funnel plot of the total effective rate.
Level of Evidence
The GRADE analysis revealed that the overall quality of evidence for all outcome indicators
was low and did not support our recommendation for the outcome. We lowered the standards
mostly due to the possibility of bias, inconsistency, and imprecision. As shown in
[Table 3].
Table 3
Level of evidence
|
Outcomes
|
Sample size (studies)
|
Effects (95% CI)
|
Quality of evidence
|
Comments
|
|
Total effective rate—
manual therapy vs.Western medicine[a]
|
198 (3)
|
OR 3.32 (1.34, 8.26)
|
⊕○○○
Very low
|
Serious
risk of bias[b], inconsistency[c], imprecision[d]
|
|
Total effective rate—
manual therapy vs. Chinese medicine decoction
|
137 (1)
|
OR 3.06 (1.18, 7.95)
|
⊕⊕○○
Low
|
Serious
risk of bias[b], imprecision[d]
|
|
Total effective rate—
manual therapy + Western medicine vs. Western medicine
|
340 (2)
|
OR 5.94 (2.63, 13.41)
|
⊕○○○
Very low
|
Serious
risk of bias[b],
Inconsistency[e], imprecision[d]
|
|
Total effective rate—
manual therapy + basic treatment[f] vs. basic treatment
|
251 (2)
|
OR 11.13 (3.26, 38.02)
|
⊕⊕○○
Low
|
Serious
risk of bias[b], imprecision[d]
|
|
Total effective rate—
manual therapy + acupuncture therapy vs. acupuncture therapy
|
152 (2)
|
OR 2.43 (0.71, 8.29)
|
⊕⊕○○
Low
|
Serious
risk of bias[b], imprecision[d]
|
|
RDQ score
|
207 (4)
|
MD −1.59 (−2.85, −0.33)
|
⊕○○○
Very low
|
Serious
risk of bias[b], inconsistency[g], imprecision[d]
|
Abbreviations: CI, confidence interval; OR, odds ratio; ⊕, evidence quality level,
+1 score; ○, evidence quality level, + 0 score; PPI, proton pump inhibitor.
GRADE Working Group grades of evidence:
Note: High quality: Further research is very unlikely to change our confidence in
the estimate of effect.
Note: Moderate quality: Further research is likely to have an important impact on
our confidence in the estimate of effect and may change the estimate.
Note: Low quality: Further research is very likely to have an important impact on
our confidence in the estimate of effect and is likely to change the estimate.
Note: Very low quality: We are very uncertain about the estimate.
a Western medicine: PPI or prokinetic agent or PPI + prokinetic agent.
b Allocation concealment report is insufficient, or blinding of participants and personnel
is missing, or blinding of outcome assessment is unclear.
c The test for heterogeneity is significant (I
2 = 52%).
d The OIS (optimal information size) is not satisfied.
e The test for heterogeneity is significant (I
2 = 86%).
f Basic treatment: posture, diety.
g The test for heterogeneity is significant (I
2 = 70%).
Discussion
The pathophysiology of GERD is multifactorial, often associated with an imbalance
of aggressive and defensive factors, where the impaired function of anatomical structures
is of much concern. The antireflux barriers comprise the LES complex, including the
LES, the esophagogastric junction (EGJ), and the crura of the diaphragm (CD). These
structures play a crucial role in swallowing and compress the esophagus to prevent
reflux.[37]
[38] GERD can be caused by a variety of factors that lead to functional or structural
impairment of the digestive system. These factors can include conditions like esophageal
hiatal hernia, increased intra-abdominal pressure due to obesity or pregnancy, prolonged
hypergastric pressure due to delayed gastric emptying, certain hormones like cholecystokinin
and glucagon, as well as high-fat, high-sugar foods or medications such as calcium
channel blockers and diazepam that can cause temporary relaxation of the LES.[39] Moreover, cognitive and emotional changes can also play a role in GERD by increasing
an individual's sensitivity to esophageal sensation known as visceral hypersensitivity,
or affecting CD motor and pain modulation via neurohumoral-endocrine pathways.[40]
[41]
[42]
The direct damage caused by reflux to the esophagus is attributed to gastric acid
and pepsin. Therefore, PPIs have become the preferred choice of medication to relieve
GERD symptoms. However, studies have shown that these drugs do not significantly reduce
the incidence of reflux events, with 10 to 40% of patients not responding to PPI treatment.
Moreover, some symptoms of refractory GERD, including heartburn, improve by less than
50% even after 12 weeks of double-dose PPI treatment.[43] Consequently, the most important aspect of treating GERD is repairing the damaged
antireflux barrier. Antireflux surgery, such as fundoplication, can restore LES and
EGJ function and halt reflux. However, the stringent requirements of the procedure,
as well as potential adverse events such as postoperative bloating, diarrhea, or gastrointestinal
dysfunction, coupled with the uncertainty of the procedure's long-term efficacy, have
led physicians and patients to seek better alternatives.[43]
[44]
The manipulative treatments utilized in this study possess the capability to impact
the visceral state from numerous angles. The advantages conferred by manipulative
treatments, including acupressure, Chinese tuina, fascial release, and osteopathy,
concerning the pathophysiology of internal organs have been established in various
illnesses.[14]
[45]
[46]
[47]
[48]
[49] According to the theory of TCM, as outlined in the ancient Chinese medicine book
Yellow Emperor's Inner Cassic (Huang Di Nei Jing), the internal organs are linked to the body surface through meridians. The acupoints
on the body surface denote the points where the meridians pass through, which can
be utilized for palpation, visualization, and treatment purposes. Acupuncture, moxibustion,
massage techniques, and other stimuli on the acupoints can be implemented to rectify
and cure internal ailments through meridians.[50] The governor vessel has the ability to govern the spinal column and spinal cord,
which is also known as the middle line of the back. The first lateral line of the
bladder meridian of foot-taiyang (located 1.5 cun lateral to the governor vessel)
is associated with the path of the sympathetic nerve. The “Back-Shu acupoints” on
the bladder meridian of foot-taiyang correspond to the locations of the sympathetic
ganglia as well.[51] Therefore, manipulation or stimulation of the spine-related acupoints can indirectly
influence the somatic or visceral nerve fibers that travel through the intervertebral
foramina of the spine via the skin, muscles, or bones, thus improving the functional
status of the gastrointestinal tract. Additionally, myofascial release can directly
act on the anatomical structures associated with the antireflux barrier to alleviate
the underlying cause of reflux.[35] Warm or mechanical stimulation by abdominal manipulation can accelerate peristalsis
to reduce gastric hypertension or modulate gastrointestinal hypersensitivity through
vegetative reflexes.[52] Thus, multipath manipulations are more effective in eliminating the root cause of
reflux and achieving long-term relief.
To the best of our knowledge, this is the first systematic review and meta-analysis
aimed at evaluating the effectiveness and safety of manual therapy in treating GERD.
Our findings suggested that, when considering the total effective rate, manual therapy
was more significant compared to Western medicine or Chinese medicine decoction used
alone. Furthermore, manual therapy combined with either Western medicine or conventional
treatment was more effective than either of these treatments alone. However, when
compared to acupuncture treatment alone, our results showed no significant change
in efficacy after manual therapy was added. Subgroup analysis of the total effective
rate according to age of the participants revealed reduced heterogeneity among groups,
indicating a significant difference in the response of infants and adults with GERD
to manual therapy. We also observed heterogeneity among studies in terms of sample
size, intervention site and modality, and duration of treatment, which may serve as
sources of heterogeneity for future studies. Although we found a significant difference
in RDQ scores (WMD = −1.59, 95% CI [−2.85, −0.33], p = 0.01), the intervention modality and symptom indicators and scoring criteria of
the questionnaire varied among studies. Therefore, it is still premature to confirm
this result. Regarding adverse events, only a few instances of small skin rashes and
back pain have been reported, making it difficult to determine whether other unpleasant
effects are caused by the massage. In summary, the meta-analysis revealed that manual
therapy treatment for GERD has positive effects and minimal adverse effects for various
age groups, indicating that it could be used as an alternative or complementary therapy
to traditional pharmacological treatments to alleviate negative effects. Despite methodological
variations, these findings are consistent.
This study has several limitations. First, the number of literature sources included
in the study was not sufficient. Although we searched for eight reputable databases,
the scope needs to be expanded to gather more data. Moreover, there were not many
high-quality RCTs of manual therapy for GERD, and some clinical studies were excluded
because they failed to meet the inclusion criteria due to reasons such as the absence
of valid randomization and incomplete results data. Second, the intervention modalities
in the included literature were too diverse and not entirely independent, causing
limitations. While the heterogeneity of the combined effect size only amounted to
40%, the tuina or massage methods varied between studies, and the number of literature
sources accompanied by equivalent manipulations were inadequate. Furthermore, manipulation
is mostly used with Western medicine, Chinese medicine decoction, or acupuncture for
cotreatment purposes, which makes it hard to control variables, resulting in a final
conclusion lacking relevance and caution. Third, there is a lack of sufficient outcome
indicators. Outcome measurements commonly utilized in GERD-related studies, such as
LES pressure measurement, 24-hour esophageal pH monitoring, gastroscopy grading, and
GERD scale (Gerd Q), were excluded from this study due to differences in intervention
modality, study type, and study quality. On the contrary, the total effective rate
and RDQ score were commonly used, but their high subjectivity may have affected the
credibility and generalizability of the meta-analysis results. Fourth, the quality
of the included studies is low. To date, clinical studies investigating manual therapy
for GERD have been inadequate in terms of randomization methods, blinding, and allocation
concealment. Some sites may provide subjects with information about the entire trial
process and other related information before conducting clinical trials to protect
both doctors and patients, which might affect blinding and allocation concealment
implementation. Therefore, it is essential to explore how to standardize clinical
trial procedures in the current health care environment.
Conclusion
This meta-analysis suggested that manual therapy has a positive impact on the relief
of symptoms such as reflux and heartburn. Manual therapy, as a standalone treatment,
is more beneficial than Chinese medicine decoction or Western medicine in the management
of GERD, and combination therapy is often superior to single therapy. However, the
conclusions should be viewed with caution due to the lack of methodological quality,
limited literature, and low sample size. For manual therapy practitioners in clinical
settings, there is evidence to support the adjunctive use of manual therapy in GERD
treatment. In the future, well-designed RCTs with larger samples and multiple centers
are necessary, utilizing tuina or massage as the primary and independent interventions,
and employing more objective and varied evaluation indicators to establish more clearly
the effectiveness, in both the short and long-term, of manual therapy in treating
GERD and which techniques are more advantageous.
