Clinical Gastroenterology and Hepatology 2018;16:407–416

Diaphragmatic Breathing Reduces Belching and Proton Pump

Inhibitor Refractory Gastroesophageal Reflux Symptoms
Andrew Ming-Liang Ong,*,‡ Laura Teng-Teng Chua,§ Christopher Jen-Lock Khor,*,‡
Ravishankar Asokkumar,*,‡ Vikneswaran s/o Namasivayam,*,‡ and Yu-Tien Wang*,‡

*Department of Gastroenterology and Hepatology, Singapore General Hospital, Singapore; ‡Duke-NUS Graduate Medical
School, Singapore; and §Department of Speech Therapy, Singapore General Hospital, Singapore

BACKGROUND & AIMS:
In patients with gastroesophageal reflux disease (GERD) and excessive belching, most belches
are supragastric, and can induce reflux episodes and worsen GERD. Supragastric belching (SGB)
might be reduced with diaphragmatic breathing exercises. We investigated whether
diaphragmatic breathing therapy is effective in reducing belching and proton pump inhibitor
(PPI)-refractory gastroesophageal reflux symptoms.

METHODS: We performed a prospective study of 36 consecutive patients with GERD refractory to PPI
therapy and a belching visual analogue scale (VAS) score of 6 or more, seen at a gastroenter-
ology clinic at a tertiary hospital in Singapore from April 2015 through October 2016. Patients
underwent high-resolution manometry and 24-hour pH-impedance studies while they were off
PPIs. Fifteen patients were placed on a standardized diaphragmatic breathing exercise protocol
(treatment group) and completed questionnaires at baseline, after diaphragmatic breathing
therapy, and 4 months after the therapy ended. Twenty-one patients were placed on a waitlist
(control subjects), completed the same questionnaires with an additional questionnaire after
their waitlist period, and eventually received diaphragmatic breathing therapy. The primary
outcome was reduction in belching VAS by 50% or more after treatment. Secondary outcomes
included GERD symptoms (evaluated using the reflux disease questionnaire) and quality of life
(QoL) scores, determined from the Reflux-Qual Short Form and EuroQoL-VAS.

RESULTS: Nine of the 15 patients in the treatment group (60%) and none of the 21 control subjects
achieved the primary outcome (P < .001). In the treatment group, the mean belching VAS score
decreased from 7.1 – 1.5 at baseline to 3.5 – 2.0 after diaphragmatic breathing therapy; in the
control group, the mean VAS score was 7.6 – 1.1 at baseline and 7.4 – 1.3 after the waitlist
period. Eighty percent of patients in the treatment group significantly reduced belching fre-
quency compared with 19% in control subjects (P [ .001). Treatment significantly reduced
symptoms of GERD (the mean reflux disease questionnaire score decreased by 12.2 in the
treatment group and 3.1 in the control group; P [ .01). The treatment significantly increased
QoL scores (the mean Reflux-Qual Short Form score increased by 15.4 in the treatment group
and 5.2 in the control group; P [ .04) and mean EuroQoL-VAS scores (15.7 increase in treat-
ment group and 2.4 decrease in the control group). These changes were sustained at 4 months
after treatment. In the end, 20 of the 36 patients who received diaphragmatic breathing therapy
(55.6%), all with excessive SGB, achieved the primary outcome.

CONCLUSIONS: In a prospective study, we found a standardized protocol for diaphragmatic breathing to reduce
belching and PPI-refractory gastroesophageal reflux symptoms, and increase QoL in patients
with PPI-refractory GERD with belching—especially those with excessive SGB.

Keywords: Involuntary Diaphragmatic Contraction; Acid Reflux; Controlled Trial; Speech Therapy; Supragastric Belching;
Diaphragmatic Breathing; PPI Refractory.

Abbreviations used in this paper: DB, diaphragmatic breathing; GB,

gastric belching; GERD, gastroesophageal reflux disease; HADS, Hospital Most current article
Anxiety and Depression Scale; LES, lower esophageal sphincter; PPI,
proton pump inhibitor; QoL, quality of life; RDQ, Reflux Disease © 2018 by the AGA Institute
Questionnaire; RQS, Reflux-Qual Short Form; SGB, supragastric belching; 1542-3565/$36.00
VAS, visual analogue scale; WLC, wait-list control subjects. https://doi.org/10.1016/j.cgh.2017.10.038
408 Ong et al
roton pump inhibitor (PPI)-refractory gastro-
P esophageal reflux disease (GERD) is a pressing
clinical problem constituting around 30% of patients
with GERD.1 In patients with GERD, 40%–49% complain
of belching as a predominant symptom.2,3 Previously,
patients with excessive belching were thought to swal-
low air frequently or in large volumes4 resulting in air
venting from the stomach, known as gastric belching
(GB). Recently, it has been shown that supragastric
belching (SGB) is the predominant type of belching in
patients with GERD with excessive belching.3,5,6 The
underlying mechanism of SGB is involuntary diaphrag-
matic contraction resulting in rapid influx of air into the
esophagus, followed by rapid air expulsion.7 It has been
shown that SGB can exacerbate GERD symptoms by
directly inducing reflux episodes.3,5,6 PPIs are used
widely in GERD treatment and have been used for
belching,8–11 but do not directly address the patho-
physiology of SGB.
A single pilot study12 has shown diaphragmatic
breathing (DB) to be useful in SGB treatment but did not
address whether GERD symptoms improved. To our
knowledge, no study has determined whether DB is
useful in patients with GERD and belching. We hypoth-
esize that DB reduces belching and reflux symptoms in
PPI-refractory patients with GERD and excessive belch-
ing. Therefore, we aimed to determine if DB is efficacious
in improving belching and GERD symptoms, and in turn,
improve quality of life (QoL).
Materials and Methods
Patient Selection
The study started in April 2015 when consecutive
patients attending the gastroenterology clinic in a ter-
tiary hospital in Singapore were invited to join after
fulfilling inclusion criteria. Inclusion of patients ended in
October 2016 and follow-up of last patient ended in
April 2017.
The study design was a wait-list controlled cohort
study. We included adult patients (
18 years) with clin-
ical or endoscopic diagnosis of GERD, according to Mon-
treal definition13 and Los Angeles grade greater than or
equal to B esophagitis, respectively. Patients needed to
have troublesome belching over the last 6 months, with a
belching visual analog scale (VAS) score of
6. All patients
had PPI-refractory GERD, defined as having troublesome
GERD (heartburn and/or regurgitation) despite twice
daily PPI therapy for 12 weeks.1 Organic or metabolic
diseases were excluded by routine biochemistry and up-
per gastrointestinal endoscopy. Other exclusion criteria
included pregnancy, patients with extreme body mass
index (<18 or >35), or surgery involving the gastroin-
testinal tract because these may alter anatomy and phys-
iology of the esophagus and stomach.
Clinical Gastroenterology and Hepatology Vol. 16, No. 3
Informed consent was obtained from each participant
and protocol was approved by Singhealth Centralised
Institutional Review Board before the start of study.
Wait-list control subjects (WLC) were not aware of an
active treatment arm being compared with but were told
that their symptoms would be reviewed at the end of the
wait-list period and receive treatment regardless of their
symptoms.
Questionnaires
All patients were given questionnaires at baseline, at
the end of treatment with DB exercises, and 4 months
post-treatment, with WLC given an additional ques-
tionnaire after their wait-list period (Figure 1). Because
there were no validated questionnaires on belching, we
asked patients to quantify their belching severity over
the last 1 week via a 100-mm VAS, along with questions
on frequency, ability to control, and repetitive nature of
belching. To assess frequency and severity of GERD,
patients were given the Reflux Disease Questionnaire
(RDQ), which was validated in evaluating treatment
response in GERD.14 The RDQ consists of heartburn,
regurgitation, and dyspepsia domains and assigns a
score for these domains based on frequency and
severity.15 We measured GERD-specific QoL via Reflux-
Qual Short Form (RQS)16 and general health-related
QoL via the EuroQol VAS.17 Anxiety and depression
levels were assessed using the Hospital Anxiety and
Depression Scale (HADS).18 Severity of somatic
symptoms were assessed using the Patient Health
Questionnaire-15.19
Our primary outcome measured was
50% reduction
in belching VAS at the end of treatment. Our secondary
outcomes included improvements in reflux symptoms
via RDQ and health-related QoL scales via RQS and
EuroQol VAS.
pH-Impedance Measurements
High-resolution esophageal manometry (Medtronic
Inc, Shoreview, MN) was performed to identify the
position of the lower esophageal sphincter (LES) and
rule out significant esophageal dysmotility. A combined
pH-impedance catheter (Sandhill Scientific, Highlands
Ranch, CO) was then placed transnasally at 5 cm above
the LES. All patients underwent these tests and results
were analyzed manually by 2 operators (A.M.-L.O.,
Y.-T.W.), both blinded to patient characteristics and
questionnaire outcomes. All drugs affecting acid sup-
pression, gastrointestinal motility, or sensitivity were
discontinued at least 1 week before.
Gastric belches (Figure 2A) and supragastric belches
(Figure 2B) were defined as previously described.20
Patients were labelled as having excessive SGB if they
had >13 supragastric belches per 24 hours.6 Severe
March 2018
Figure 1. Flow diagram of study. HRM ¼ high-resolution manometry.
SGB was defined as more than 124 episodes per 24
hours.6
Belches were associated with reflux if reflux episodes
occurred within 1 second of a belch (Figure 2A). If a
reflux episode was preceded by multiple belches, the
belch with the closest temporal association to the reflux
episode was considered. A belch was considered symp-
tomatic if it occurred within a 1-minute time window
preceding a patient-reported belch (Figure 2A). If mul-
tiple belches occurred during this window, only the belch
with the closest temporal association to the reported
belch symptom was considered.
Patients were diagnosed as GERD if they had patho-
logic acid exposure (acid exposure time >4% or
Demeester score >14). Reflux hypersensitivity was
diagnosed if they had no pathologic acid exposure but
positive symptom association to reflux episodes (Reflux
Diaphragmatic Breathing Improves Belching and GERD 409
Symptom Index >50% and/or Reflux Symptom Associ-
ation Probability >95%). Functional heartburn was
diagnosed if there was no pathologic acid exposure or
positive symptom association.
Diaphragmatic Breathing Exercises
DB was taught exclusively by 1 trained speech
therapist (L.T.-T.C.). Each session was carried out 1 to 1
over 30 minutes. Patients underwent 4 weekly sessions
over a period of 4 weeks. Adherence to DB exercise
training was measured according to the attendance at
scheduled sessions, and arbitrarily defined as atten-
dance of all scheduled 4 sessions. Compliance to home
exercises was assessed by direct questioning of patient
at 4 months post-treatment. Details and example of the
410 Ong et al Clinical Gastroenterology and Hepatology Vol. 16, No. 3

Figure 2. (A) Gastric belch.

Characterized by rise in
impedance in distal rings
initially and then moving
proximally, reflective of
retrograde gas movement.
Arrows indicate direction
of airflow. This belch is
symptomatic, evident by
patient-reported symptom
marker. Impedance levels
return to levels lower than
baseline, suggestive of
concomitant reflux
episode, acidic as pH <4.
(B) Supragastric belch.
Characterized by rise in
impedance in proximal
rings initially and then
moving distally. This is fol-
lowed by rapid return to
baseline in distal leads first
and then proximally,
reflective of the quick gas
expulsion. Arrows indicate
direction of airflow.

DB protocol can be found in Supplementary Table 1 and
Supplementary Video 1.
Statistical Analysis
The study was initially designed to compare efficacy
of DB in GB versus SGB. However, early into the study, it
was noted that most patients recruited had SGB and the
study protocol was amended with the new aim to
compare treatment outcomes of post-DB patients with
WLC in all belching patients with GERD. Assuming a
similar standard deviation from the Hemmink et al12
study, our study had at least 90% power to detect a
mean difference of 3.3 in belching VAS between WLC and
post-DB using 15 patients each.
March 2018 Diaphragmatic Breathing Improves Belching and GERD 411

All authors had access to the study data and reviewed
and approved the final manuscript. Data are presented as
mean  standard deviation or median (interquartile
range) after checking for normal distribution. We used
paired and unpaired Student t test for parametric
quantitative data from in-group and between-group
comparisons, and Wilcoxon signed rank test for
nonparametric quantitative data. Fisher exact test was
used to compare categorical variables. Differences were
considered significant when P was < .05.
(14.9%) had functional heartburn. Thirty-two patients
had excessive SGB, of which 5 patients had severe
SGB.6 Median number of belches per 24 hours was 46
(32–80), and 20% of supragastric belches were
symptomatic, whereas 2% of gastric belches were
symptomatic. A total of 44% of all reflux episodes
detected were associated with a belch (Supplementary
Table 2).
Primary Outcome

Results A total of 9 of 15 (60%) in treatment group reduced

Patient Characteristics
Of 64 patients fulfilling the inclusion criteria, 36
patients were enrolled and completed the DB exercise
training (25 women; 32 Chinese; mean age, 45.5  12.7),
with 27 patients excluded because of unwillingness to
participate and 1 patient diagnosed with achalasia
(Figure 1). At baseline, patients had mean belching VAS
of 7.4  1.3 and mean RDQ score of 18.1  14.0. HADS
anxiety and depression scores were 8.9  4.2 and 5.3 
3.1, respectively. An initial 15 patients underwent DB
exercises immediately on enrolment (treatment group),
whereas the subsequent 21 were WLC. These groups did
not differ in demographics, symptom severity, psychiat-
ric comorbidities, Patient Health Questionnaire-15, and
QoL scores (Table 1). Adherence to DB exercise training
was achieved in 33 of 36 (91.7%).
pH-Impedance Analyses
A total of 5 of 36 (13.9%) patients had GERD, 24 of
36 (66.7%) had reflux hypersensitivity, and 7 of 36
Table 1. Comparison of Treatment Group and Wait-List
Control Subjects at Baseline
Treatment group Wait-list control
(n ¼ 15) subjects (n ¼ 21) value
belching VAS by
50% post-DB, whereas 0% of WLC
achieved this primary outcome after their wait list period
(P < .001). A total of 14 of 15 (93.3%) in treatment
group achieved some reduction in belching VAS,
compared with 7 of 21 (33.3%) in WLC (P < .001). Mean
belching VAS of treatment group was significantly
reduced from 7.1  1.5 to 3.5  2.0 compared with WLC,
7.6  1.1 to 7.4  1.3 (Figure 3A and B). A total of 80% in
treatment group had reduction in belching frequency
compared with 19% in WLC (P ¼ .001), whereas 40% in
treatment group improved belch control compared with
4.8% in WLC (P ¼ .01) (Table 2).
Secondary Outcomes
RDQ scores was significantly improved in treatment
group patients compared with WLC (12.2  11.7 vs 3.1
 7.9; P ¼ .01). All 3 domains of RDQ (heartburn,
dyspepsia, regurgitation) showed improvement in post-
DB patients, but only changes in regurgitation domains
had statistical significance (5.1  4.4 vs 0.2  3.5;
P ¼ .001) (Table 2). RQS scores (15.4  15.1 vs 5.1 
12.2; P ¼ .04) and EuroQol VAS (15.7  10.7 vs -2.4 
10.2; P < .001) improved significantly in treatment
group compared with WLC. There were no differences in
improvements of HADS scores between treatment group
P patients and WLC.

Age 44.9  15.7 46  10.44 NS

Follow up
Gender (female), % 73.3 61.9 NS
Race (Chinese), % 86.7 85.7 NS All 21 WLC underwent DB eventually and improved
Alcohol (yes), % 6.7 4.8 NS their belching VAS, from 7.4  1.3 to 4.8  1.8 (P < .001).
Smoking (yes), % 13.3 14.3 NS Nine of 21 WLC (42.9%) achieved the primary outcome of
Belching VAS 7.1  1.5 7.6  1.1 NS
RDQ heartburn 4.7  4.7 5.3  6.1 NS
reducing belching VAS by
50% post-DB. Nineteen of 21
RDQ regurgitation 7.9  6.5 6.8  5.1 NS patients (90.5%) of WLC showed significant improve-
RDQ total 17.5  13.4 18.5  14.8 NS ments in their belching frequency and 9 of 21 (42.9%)
RQS 52.3  25.8 39.4  16.3 NS improved belch control post-DB (Supplementary Table 3).
HADS anxiety 8.4  4.9 9.2  3.7 NS Overall, 20 of all 36 patients (55.6%) who underwent DB
HADS depression 5.4  3.6 5.3  2.9 NS
Patient Health 10.1  5.5 11.8  6.2 NS
achieved the primary outcome.
Questionnaire-15 We followed up 30 out of 36 all post-DB patients at 4
EuroQol VAS 57.7  20.6 62.2  13.5 NS months post-treatment. All clinical outcomes (primary
and secondary) remained significantly improved
compared with baseline, but deteriorated from end of
HADS, Hospital Anxiety and Depression Scale; NS, non-significant; QoL,
quality of life; RDQ, Reflux Disease Questionnaire; RQS, Reflux-Qual Short treatment (Figure 4). A total of 27 of 30 (90.0%) were
Form; VAS, visual analog scale. compliant to home exercises when asked during their
412 Ong et al
follow-up visit. There were no differences in clinical
outcomes between those compliant and those non-
compliant to home exercises.
Treatment Response Based on
Type of Belching
All 20 patients who achieved the primary outcome
had excessive SGB, whereas none of the patients without
excessive SGB achieved this (P ¼ .03). 78% of those with
excessive SGB reduced belching frequency compared
with 25% of those without excessive SGB (P ¼ .06), and
also had greater improvement in RDQ (12.2  12.1 vs 3.8
 4.0; P ¼ .18) and greater improvement in RQS (17.4 
13.6 vs 10.2  20.0; P ¼ .35) compared with patients
without excessive SGB, but these were not statistically
significant.
No differences were found between responders and
nonresponders with regards to gender, pH-impedance
Clinical Gastroenterology and Hepatology Vol. 16, No. 3
Figure 3. (A) Mean belch-
ing VAS pre- and post-DB
in treatment group.a (B)
Mean belching VAS pre-
and post-wait-list period in
WLC. aP value of compar-
ison of belching VAS
improvements between
patients in A and B < .001.
diagnoses, severity of SGB, HADS scores, or somatiza-
tion scales (Supplementary Table 4).
Discussion
Our study showed that in a group of PPI-refractory
patients with GERD, standardized DB exercise training
significantly reduced belching severity and GERD
symptoms, and improved patient QoL. These benefits
were greater in patients with excessive SGB, and sus-
tainable at 4 months post-treatment. Although belching
is common in GERD,3 little research has been done in
this, and there are no studies looking at treatment
strategies for these patients. Ours is the first such study
examining the efficacy of DB exercises in patients with
GERD and belching.
DB may improve belching through several mecha-
nisms. Early case studies21 had shown belching reduc-
tion in children with developmental disabilities by
March 2018 Diaphragmatic Breathing Improves Belching and GERD 413

Table 2. Clinical Outcomes of Post-DB Treatment Group and Post-Wait-List WLC

Post-DB treatment Post-wait-list P All post-DB

group (n ¼ 15) WLC (n ¼ 21) valuea patients (n ¼ 36)

Reduce belching VAS
50% 9 (60%) 0 (0%) <.001 20 (55.6%)

Reduce belching VAS 14 (93.3%) 7 (33.3%) <.001 31 (86.1%)
Improvement in belching VAS 3.5  1.8 0.2  1.2 <.001 3.3  1.8
Reduce belching frequency 12 (80.0%) 4 (19.0%) .001 26 (72.2%)
Frequency of belching/day Baseline Post-DB Baseline Post-wait-list
>10 10 1 15 11
5–10 2 5 3 8
1–5 2 3 3 2
Occasional 1 6 0 0
Change in belching control 6 (40.0%) 1 (4.8%) .01 14 (38.9%)
Improvement in RDQ (total) 12.2  11.7 3.1  7.9 .01 11.3  11.8
Improvement in RDQ (heartburn) 4.1  4.6 1.5  3.6 .07 3.7  4.6
Improvement in RDQ (dyspepsia) 2.8  5.0 1.0  3.9 .22 3.4  4.9
Improvement in RDQ (regurgitation) 5.1  4.4 0.2  3.5 .001 3.9  4.5
Improvement in RQS 15.4  15.1 5.2  12.2 .04 16.6  14.3
Improvement in EuroQol VAS 15.7  10.7 2.4  10.2 <.001 10.9  11.7
Improvement in HADS (total) 3.3  4.3 2.5  7.3 .71 3.7  6.3
Improvement in HADS (anxiety) 1.7  3.2 1.5  4.3 .91 2.1  4.4
Improvement in HADS (depression) 1.7  2.9 0.5  3.6 .3 1.6  3.0

DB, diaphragmatic breathing; HADS, Hospital Anxiety and Depression Scale; QoL, quality of life; RDQ, Reflux Disease Questionnaire; RQS, Reflux-Qual Short
Form; VAS, visual analogue scale; WLC, wait-list control subjects.
a
P value for comparison between treatment group and WLC.

teaching a response incompatible with swallowing, such
as biting on forefingers. Other case reports in adults22,23
have shown that eliminating subconscious repetitive
swallowing via keeping the mouth open combined with
teaching slow DB was effective in reducing belching.
However, reducing air swallowing is unlikely to be the
sole mechanism for improvement as other studies20 have
shown that patients with excessive belching did not
exhibit excessive air swallows. Kessing et al24 showed
that most supragastric belches are initiated by creation
of subatmospheric pressure in the thoracic cavity by
movement of the diaphragm in the aboral direction.
Because the diaphragm consists of striated muscle under
voluntary contraction, patients theoretically can be
taught how to control these contractions. However, most
of our patients with improvement in belching were still
not able to control their belch voluntarily, so perhaps it is
a combination of both previously mentioned factors
mediating the reduction of belching frequency.
It is likely that reduction in SGB frequency accounted
for most symptom improvement. Our subgroup analysis
showed that only patients with excessive SGB achieved
the primary outcome, whereas those without excessive
SGB did not do so. GB has a completely different mech-
anism involving transient LES relaxations triggered by
proximal stomach distention,7 and therefore these
patients may not respond to DB exercises. There is also a
second less common mechanism of SGB24 that involves
an increase in pharyngeal pressure causing air influx into
the esophagus as the initiating mechanism. Perhaps some
of our patients not responding to DB exercises either had
mostly GB or had SGB associated with changes in
pharyngeal pressure. We were not able to distinguish
which SGB mechanisms were predominant in our
patients because we did not perform prolonged
post-prandial high-resolution esophageal manometry
monitoring to evaluate the characteristics of belches, but
this would be interesting to explore.
GERD symptoms significantly improved after un-
dergoing DB exercises, and this is largely mediated by
improvement in regurgitation reflected by the changes
in RDQ domains. Regurgitation scores significantly
improved from baseline but not heartburn and
dyspepsia scores. Hemmink et al3 showed that SGB can
occur either at the onset or during a liquid reflux
episode and our study further confirms the close rela-
tionship between GERD and SGB. In our patients, 44%
of all reflux events were associated with a belch, and
most of these were supragastric belches. It is likely
these events were contributing to the patient’s symp-
tom burden. In our patients, supragastric belches were
more likely to be symptomatic compared with gastric
belches (20% vs 2%) and this was also shown by
Kessing et al.5 We postulate that DB exercises improved
GERD symptoms via reducing SGB frequency, which
then reduces the frequency of SGB-associated GERD
symptoms, especially regurgitation. Although we did
not repeat pH-impedance studies post-DB, it could be
inferred that because patients reported reduced belch-
ing frequency in their post-treatment questionnaires,
the number of reflux events were possibly reduced.
Further studies looking at number of reflux episodes
and acid exposure time pre- and post-DB exercise
training would support this theory.
414 Ong et al
There are, however, other possible mechanisms
accounting for improvement in belching and GERD
symptoms after DB. The importance of the crural dia-
phragm is recognized in maintaining the antireflux bar-
rier at the LES, and abdominal breathing exercises have
been shown to improve GERD symptoms independent of
belching.25 Because SGB can be a response from an
unpleasant sensations, such as reflux episodes,3 DB exer-
cises could potentially improve the antireflux barrier by
strengthening the tone at the crural diaphragm resulting
in less frequent reflux episodes and hence, less SGB
episodes. Another possible mechanism could be related to
psychological factors, which have been shown to be
important in the initiation and perpetuation of SGB.26 DB
exercises is a relaxation technique shown to reduce anx-
iety27 and belching may be reduced by modulating
attention via distraction,26 so perhaps these exercises
distracted the patient from belching, or had a relaxation
effect to reduce their anxiety. Our patients had a mean
HADS anxiety score of 8.9, reflecting a high prevalence of
anxiety, but in our analysis, although there were signifi-
cant reductions in both HADS anxiety and depression
scores pretreatment and post-treatment, there were no
differences between treatment group and WLC (Table 2).
Furthermore, it is not clear if improvements in symptoms
Clinical Gastroenterology and Hepatology Vol. 16, No. 3
Figure 4. Clinical out-
comes pre-DB, post-DB,
and 4 months post-DB. (A)
Mean belching VAS, (B)
mean RDQ, (C) mean RQS,
(D) mean EuroQol VAS.
led to improvements in mood, or if reductions in anxiety
levels mediated the improvement in belching.
PPI-refractory GERD has many possible mechanisms,
and there is a subgroup of patients whose reflux symp-
toms are precipitated by events such as SGB. We found
that most of our patients did not have GERD on pH-
impedance, but had been diagnosed as having GERD
largely based on the Montreal Classification. Interest-
ingly, our pH-impedance studies showed that 66.7% and
19.4% (Supplementary Table 2) had reflux hypersensi-
tivity and functional heartburn, respectively, but sub-
group analysis did not show any differences in treatment
response between pH-impedance diagnoses, likely
because of our small sample size (Supplementary
Table 4). All 36 patients were on stable twice-daily
dosing of PPI during the intervention, but only 10 of
36 (27.8%) of these patients were still on PPI post-DB.
Because DB exercises are safe, unlike PPIs where there
are concerns on long-term safety,28 DB may be an
attractive therapeutic option. However, based on our
results, recommendations for using DB for GERD can
only be made for those with excessive belching, espe-
cially SGB.
There are several limitations to our study. The study
protocol was amended early in the study. Therefore, we
March 2018
could not randomize the allocation of patients to treat-
ment and WLC groups, although we showed in Table 1
that groups were similar in demographics and baseline
symptom severity. WLC studies may overestimate
treatment effects29 but we believed that our primary
outcome of reducing belching VAS by
50% had been set
at a high and clinically meaningful threshold to mitigate
this. Regardless, we would still need to confirm our
findings via a randomized controlled trial. We did not use
Rome III criteria for belching disorders4 but used inclu-
sion criteria of troublesome belching for last 6 months
with VAS
6. There was no available validated measure
at the time of study, but we believe that belching VAS
was clinically meaningful because it encompasses fre-
quency, severity, and QoL, and furthermore was used in
the study by Hemmink et al.12 This likely excluded
patients with mild belching and hence the results may
not be generalizable to all belching patients. Our sample
size was relatively small and hence underpowered to
detect differences in subgroup analysis of responders,
but despite these limitations, this is the largest study
evaluating treatment in patients with belching and the
first looking at treatment options in patients with GERD
and belching.
In conclusion, we showed that a standardized DB
exercise training protocol improves belching, GERD
symptoms, and QoL in patients with PPI-refractory
GERD and belching, especially those with excessive
SGB. DB exercise training can be a useful adjunct to
current pharmacologic options in treating these
patients.
Supplementary Material
Note: To access the supplementary material accom-
panying this article, visit the online version of Clinical
Gastroenterology and Hepatology at www.cghjournal.org,
and at https://doi.org/10.1016/j.cgh.2017.10.038.
References
1. Sifrim D, Zerbib F. Diagnosis and management of patients with
reflux symptoms refractory to proton pump inhibitors. Gut 2012;
61:1340–1354.
2. Klauser AG, Schindlbeck NE, Müller-Lissner SA. Symptoms in
gastro-oesophageal reflux disease. Lancet 1990;335:205–208.
3. Hemmink GJM, Bredenoord AJ, Weusten BL, et al. Supragastric
belching in patients with reflux symptoms. Am J Gastroenterol
2009;104:1992–1997.
4. Tack J, Talley NJ, Camilleri M, et al. Functional gastroduodenal
disorders. Gastroenterology 2006;130:1466–1479.
5. Kessing BF, Bredenoord AJ, Velosa M, et al. Supragastric
belches are the main determinants of troublesome belching
symptoms in patients with gastro-oesophageal reflux disease.
Aliment Pharmacol Ther 2012;35:1073–1079.
6. Koukias N, Woodland P, Yazaki E, et al. Supragastric belching:
prevalence and association with gastroesophageal reflux dis-
ease and esophageal hypomotility. J Neurogastroenterol Motil
2015;21:398–403.
Diaphragmatic Breathing Improves Belching and GERD 415
7. Kessing BF, Bredenoord AJ, Smout AJPM. The pathophysi-
ology, diagnosis and treatment of excessive belching symp-
toms. Am J Gastroenterol 2014;109:1196–1203.
8. Kahrilas PJ, Miner P, Johanson J, et al. Efficacy of rabeprazole
in the treatment of symptomatic gastroesophageal reflux dis-
ease. Dig Dis Sci 2005;5:2009–2018.
9. Miner P, Orr W, Filippone J, et al. Rabeprazole in nonerosive
gastroesophageal reflux disease: a randomized placebo-
controlled trial. Am J Gastroenterol 2002;97:1332–1339.
10. Gerson LB, Kahrilas PJ, Fass R. Insights into gastroesophageal
reflux disease–associated dyspeptic symptoms. Clin Gastro-
enterol Hepatol 2011;9:824–833.
11. Kriengkirakul C, Patcharatrakul T, Gonlachanvit S. The
therapeutic and diagnostic value of 2-week high dose proton
pump inhibitor treatment in overlapping non-erosive gastro-
esophageal reflux disease and functional dyspepsia patients.
J Neurogastroenterol Motil 2012;18:174–180.
12. Hemmink GJM, Ten Cate L, Bredenoord AJ, et al. Speech
therapy in patients with excessive supragastric belching: a pilot
study. Neurogastroenterol Motil 2010;22:24–30.
13. Vakil N, Van Zanten SV, Kahrilas P, et al. The Montreal definition
and classification of gastroesophageal reflux disease: a global
evidence-based consensus. Am J Gastroenterol 2006;
101:1900–1920.
14. Shaw M, Dent J, Beebe T, et al. The Reflux Disease Question-
naire: a measure for assessment of treatment response in
clinical trials. Health Qual Life Outcomes 2008;6:31.
15. Shaw MJ, Talley NJ, Beebe TJ, et al. Initial validation of a
diagnostic questionnaire for gastroesophageal reflux disease.
Am J Gastroenterol 2001;96:52–57.
16. Amouretti M, Nalet B, Robaszkiewicz M, et al. Validation of the
short-form REFLUX-QUAL (RQS), a gastro-esophageal reflux
disease (GERD) specific quality of life questionnaire. Gastro-
enterol Clin Biol 2005;29:793–801.
17. Rabin R, de Charro F. EQ-5D: a measure of health status from
the EuroQol Group. Ann Med 2001;33:337–343.
18. Zigmond AS, Snaith RP. The hospital anxiety and depression
scale. Acta Psychiatr Scand 1983;67:361–370.
19. Kroenke K, Spitzer RL, Williams JB. The PHQ-15: validity of a
new measure for evaluating the severity of somatic symptoms.
Psychosom Med 2002;64:258–266.
20. Bredenoord AJ, Weusten BLAM, Sifrim D, et al. Aerophagia,
gastric, and supragastric belching: a study using intra-
luminal electrical impedance monitoring. Gut 2004;
53:1561–1565.
21. Flaisher GF. The use of suggestion and behavioral methods in
the treatment of aerophagia: two case reports. Child Fam Behav
Ther 1994;16:24–32.
22. Cigrang JA. Behavioral treatment of chronic belching due to
aerophagia in a normal adult. Behav Modif 2006;
30:341–351.
23. Katzka DA. Simple office-based behavioral approach to patients
with chronic belching. Dis Esophagus 2013;26:570–573.
24. Kessing BF, Bredenoord AJ, Smout AJPM. Mechanisms of
gastric and supragastric belching: a study using concurrent
high-resolution manometry and impedance monitoring. Neuro-
gastroenterol Motil 2012;24:573–579.
25. Eherer AJ, Netolitzky F, Hogenauer C, et al. Positive effect of
abdominal breathing exercise on gastroesophageal reflux
disease: a randomized, controlled study. Am J Gastroenterol
2012;107:372–378.
416 Ong et al
26. Bredenoord AJ, Weusten BLAM, Timmer R, et al. Psychological
factors affect the frequency of belching in patients with
aerophagia. Am J Gastroenterol 2006;101:2777–2781.
27. Chen Y-F, Huang X-Y, Chien C-H, et al. The effectiveness of
diaphragmatic breathing relaxation training for reducing anxiety.
Perspect Psychiatr Care 2017;53:329–336.
28. Freedberg DE, Kim LS, Yang Y-X. The risks and benefits of
long-term use of proton pump inhibitors: expert review and best
practice advice from the American Gastroenterological
Association. Gastroenterology 2017;152:706–715.
29. Cunningham JA, Kypri K, McCambridge J. Exploratory
randomized controlled trial evaluating the impact of a waiting list
control design. BMC Med Res Methodol 2013;13:150.
Clinical Gastroenterology and Hepatology Vol. 16, No. 3
Reprint requests
Address requests for reprints to: Andrew Ming-Liang Ong, MBChB, MRCP,
Department of Gastroenterology and Hepatology, Singapore General Hospital,
20 College Road, Level 3, Academia Building, Singapore 169856. e-mail:
Andrew.ong.m.l@singhealth.com.sg; fax: (þ65) 6227 3623.
Conflicts of interest
The authors disclose no conflicts.
Funding
This study was funded by SingHealth Duke-NUS Academic Medical
Centre, through the Medicine Academic Clinical Programme Grant: Pitch
for Funds Programme. The manuscript is the original work of the authors.
All data, tables, figures, etc. used in the manuscript are prepared originally
by authors, otherwise the sources are cited and reprint permission should
be attached.
March 2018 Diaphragmatic Breathing Improves Belching and GERD 416.e1

Supplementary Table 1. Standardized Protocol for Administering Diaphragmatic Breathing Exercises

1. Belching mechanism and the purpose of breathing exercises explained to patient.

2. Patients started the exercise in a supine position. One hand was placed on the chest and the other hand on the abdomen above the navel.
3. Instructions were given to inhale through the nose, and exhale with the mouth open, moving only the abdomen. The metaphor of inflating
and deflating a balloon in the abdomen was given to aid patient in performing this maneuver.
4. The goal was to feel the abdomen rise and fall with each breath, while the chest and shoulders remained still. Importance was placed on
complete inflation and deflation of the abdomen.
5. Inhalation and exhalation were made to be slow, with the patient counting to 4 with each inhalation and exhalation. Inhalation was then
maintained at 4 counts, and exhalation gradually extended to 8 counts. If a participant had difficulty prolonging exhalation, breathing out
was done through pursed lips instead.
6. When assessed by the speech therapist to be competent in performing the above maneuvers, participants progressed to doing the same
exercise while sitting and then standing.
7. Patients were given homework to practice, and was recommended to perform 30 breaths or for 5 minutes 3 times daily, and for 5 minutes
when symptomatic.
8. Each participant was seen for either up to 4 sessions, or until the speech therapist judged that the participant could perform diaphragmatic
breathing as taught in supine and upright positions.

Supplementary Table 2. pH-Impedance Analysis of all 36

Patients
Mean AET 2.0  2.5
Mean Demeester score 6.5  6.7
AET >4% or Demeester score >14 5 (13.9%)
(GERD diagnosis)
Reflux hypersensitivity diagnosis 24 (66.7%)
Functional heartburn diagnosis 7 (19.4%)
Total belches per 24 h 46 (32–80)
% with excessive SGB 88.9 (32/36)
% with severe SGB 13.9 (5/36)
% symptomatic SGB 20 (4–34)
% symptomatic GB 2 (0–8)
% belches associated with reflux episode 37 (26–63)
Total reflux episodes per 24 h 38 (25–58)
% reflux episodes associated with belch 44 (33–71)
Positive symptom association belch, % of patients 80
Positive symptom association 60
heartburn, % of patients
Positive symptom association 56
regurgitation, % of patients

AET, acid exposure time; GB, gastric belch; GERD, gastroesophageal reflux
disease; SGB, supragastric belch.
416.e2 Ong et al Clinical Gastroenterology and Hepatology Vol. 16, No. 3

Supplementary Table 3. Outcomes of Wait-List Control Subjects Post-DB

Post-wait-list period (n ¼ 21) Post-DB exercises (n ¼ 21) P value

Belching VAS 7.4  1.3 4.8  1.8 <.001

Reduce belching by VAS
50%, n (%) 0 (0) 9 (42.9) .001
Reduce belching VAS, n (%) 7 (33.3) 19 (90.5) <.001
Reduce belch frequency, n (%) 4 (19.0) 16 (76.2) <.001
Change in burp control, n (%) 1 (4.8) 9 (42.9) .004
RDQ heartburn 3.8  4.9 1.9  3.6 .06
RDQ regurgitation 6.6  5.0 3.6  3.7 .007
RDQ dyspepsia 5.1  5.6 2.3  4.0 .009
RDQ total 15.4  13.1 7.9  7.8 .004
RQS 39.4  16.3 50.7  18.2 .002
HADS anxiety 7.7  4.1 6.5  3.4 .14
HADS depression 4.8  3.8 3.8  2.8 .23
HADS total 12.0  6.7 10.3  5.2 .23
EQ5D VAS 60.0  15.8 67.7  15.3 .03

DB, diaphragmatic breathing; EQ5D, EuroQol 5 dimensions; HADS, Hospital Anxiety and Depression Scale; RDQ, Reflux Disease Questionnaire; RQS, Reflux-Qual
Short Form; VAS, visual analog scale.

Supplementary Table 4. Subgroup Analysis of Responders

Proportion achieving primary outcome (responders) P value

Gender, n (%)
Female (n ¼ 25) 14 (56) NS
Male (n ¼ 11) 6 (54.5)
pH-impedance diagnosis, n (%)
GERDa (n ¼ 5) 3 (60) NS
Reflux hypersensitivity (n ¼ 24) 15 (62.5)
Functional heartburn (n ¼ 7) 2 (28.6)
Severity of SGB, n (%)
Severe (>124/24 h) (n ¼ 5) 4 (80) NS
Nonsevere (n ¼ 31) 16 (51.6)
Somatization, n (%)
PHQ15 >9b (n ¼ 22) 14 (63.6) NS
PHQ15 9 (n ¼ 14) 6 (42.9)
Psychiatric comorbidities, n (%)
HADS anxiety >7 (n ¼ 19) 9 (47.4) NS
HADS anxiety 7 (n ¼ 17) 11 (64.7)
HADS depression >7 (n ¼ 10) 5 (50) NS
HADS depression 7 (n ¼ 26) 15 (57.7)

GERD, gastroesophageal reflux disease; HADS, Hospital Anxiety and Depression Scale; NS, NS, non-significant; PHQ15, Patient Health Questionnaire-15; SGB,
supragastric belch.
a
GERD diagnosis ¼ positive acid exposure time and/or Demeester score >14.
b
PHQ15 >9 suggests significant somatization burden.19