You are on page 1of 7

ORIGINAL ARTICLE: GASTROENTEROLOGY

Body Positioning and Medical Therapy for Infantile


Gastroesophageal Reflux Symptoms


Clara Loots, yStamatiki Kritas, Michiel van Wijk, yLisa McCall, Laura Peeters,
Peter Lewindon, Rob Bijlmer, Ross Haslam, jjJacinta Tobin, Marc Benninga,
y
Geoffrey Davidson, and yTaher Omari

ABSTRACT
Objective: Proton-pump inhibitors (PPIs) reduce acid gastroesophageal
reflux (GER) and esophageal acid exposure in infants; however, they do
not reduce total GER or symptoms attributed to GER. Reflux is reduced in
the left lateral position (LLP). We hypothesize that the effect of LLP in
combination with acid suppression is most effective in reducing GER
symptoms in infants.
Methods: In this prospective sham-controlled trial, infants (06 months)
with symptoms suggestive of gastroesophageal reflux disease were studied
using 8-hour pH-impedance, cardiorespiratory and video monitoring, direct
nurse observation, and a validated questionnaire. Infants demonstrating a
positive GER symptom association were randomized to 1 of 4 groups; PPI
LLP, PPI head of cot elevation (HE), antacid (AA) LLP, or AA HE.
HE and AA were considered sham therapies. After 2 weeks the 8-hour
studies were repeated on-therapy.
Results: Fifty-one patients were included (aged 13.6 [226] weeks). PPI
LLP was most effective in reducing GER episodes (69 [13] to 46 [10],
P < 0.001) and esophageal acid exposure (median [interquartile range] 8.9%
[3.1%18.1%] to 1.1% [0%4.4%], P 0.02). No treatment group showed
improvement in crying/irritability, although vomiting was reduced in AA
LLP (from 7 [2] to 2 [0] episodes P 0.042). LLP compared with HE
produced greater reduction in total GER (21 [4] vs 10 [4], P 0.056),
regardless of acid-suppressive therapy. Acid exposure was reduced on
PPI compared with AA (6.8 [2.1] vs 0.9 [1.4]%, pH < 4, P 0.043)
regardless of positional intervention. A post-hoc analysis using automated
Received October 24, 2013; accepted April 1, 2014.
From the Pediatric Gastroenterology and Nutrition, Emma Childrens
Hospital, AMC, Amsterdam, The Netherlands, the ySchool of Medicine,
Flinders University, Adelaide, zPediatric Gastroenterology, Royal
Childrens Hospital, Brisbane, the Neonatal Intensive Care Unit,
Womens and Childrens Health Network, Adelaide, and the jjNorth
Western Academic Centre University of Melbourne, Melbourne,
Australia.
Address correspondence and reprint requests to Taher Omari, PhD, Flinders
Medical Science and Technology, School of Medicine, Flinders
University, South Australia, Adelaide, Australia (e-mail: taher.omari
@flinders.edu.au).
This study was supported by grants from the National Health and Medical
Research Council (ID: 508053), the Financial Markets Foundation for
Children (grant no. 2009-112), the Dutch Digestive Disease Foundation
(grant no. WO 07-07), the Channel 7 Childrens Research Foundation,
and the Womens & Childrens Hospital Foundation. Part of the equipment was kindly provided by AstraZeneca. None of these bodies had a
role in the study design; the collection, analysis, and interpretation of
data; the writing of the report; and the decision to submit the article for
publication.
The authors report no conflicts of interest.
Copyright # 2014 by European Society for Pediatric Gastroenterology,
Hepatology, and Nutrition and North American Society for Pediatric
Gastroenterology, Hepatology, and Nutrition
DOI: 10.1097/MPG.0000000000000395

JPGN

Volume 59, Number 2, August 2014

analysis software revealed a significant reduction in crying symptoms in the


PPI LLP group (99 [65103] to 62 [3296] episodes, P 0.018).
Conclusions: Symptomatic gastroesophageal reflux disease implies
disease causation for distressing infant symptoms. In infants with
symptoms attributed to GER, LLP produced a significant reduction in
total GER, but did not result in a significant improvement in symptoms
other than vomiting; however, automated analysis appeared to identify
infants with GER-associated crying symptoms who responded to
positioning therapy. This is an important new insight for future research.
Key Words: diagnosis, gastroesophageal reflux, treatment

(JPGN 2014;59: 237243)

astroesophageal reflux (GER) is the retrograde flow of


gastric contents into the esophagus. GER is a physiological
mechanism that occurs in all infants (1). Some infants present with
troublesome symptoms, such as irritability and crying, with or
without excessive regurgitation. These infants are often diagnosed
as having GER disease (GERD) once all other possible causes (eg,
infection) have been excluded.
It has been well established that the underlying mechanism of
nearly all GER episodes in infants is transient lower esophageal
sphincter relaxation (TLESR) (27). TLESRs allow acidic (<pH 4)
or nonacidic liquid, gas, or mixed GER episodes to occur. In older
children and adults most GER symptoms and complications are
believed to be the result of excessive acid liquid GER. In infants,
GER is more frequently not acidic owing to regular feeds with
potent buffers of gastric contents such as breast milk and formula;
however, GER episodes in the shorter infant esophagus project
more proximally, and evidence suggests that such GER episodes do
also cause symptoms (810).
The diagnosis of GERD in infants is fraught with difficulty
owing to the wide range of clinical presentations and the lack of
diagnostic modalities and appropriate diagnostic criteria. These
issues lead to failure of accurate GERD diagnosis.
The first steps in the therapeutic process are parental reassurance and conservative nonpharmacological measures. When
conservative therapy fails, often infants are empirically prescribed
drugs that suppress gastric acid production. Despite widespread use
(11), there is a paucity of evidence-based guidance on efficacy
including appropriate dosing levels (1214).
Although the therapeutic efficacy of proton-pump inhibitors
(PPIs) in controlling symptoms has not been demonstrated, the
pharmacodynamic effect of PPIs on intragastric pH and acid GER
has been well characterized (15,16). PPIs effectively reduce the
proportion of reflux episodes that are acidic; however, PPIs do not
decrease the overall number of GER episodes (15,17). This may
explain the lack of symptom efficacy, because experience suggests

237

Copyright 2014 by ESPGHAN and NASPGHAN. Unauthorized reproduction of this article is prohibited.

Loots et al

JPGN

that occurrence of GER rather than acidity causes symptoms.


Presently there are no safe pharmacological approaches to reduce
TLESRs and GER in children. We have shown in infants that
TLESRs and GER are significantly reduced in the left lateral
position (LLP) compared with right lateral position (RLP)
(1820). LLP has the ability to reduce all types of reflux during
the entire postprandial period but does not alter the proportion of
acidic reflux episodes (pH < 4) (21).
The present study evaluates an LLP regimen as a concept test
for the efficacy of therapies designed to treat GER-related symptoms by reducing the rate at which reflux is triggered following
feeding. We hypothesized that infants with a GER-symptom association (proven by impedance) demonstrate a symptom response to
LLP owing to reduced GER episodes. The effect of PPI was also
investigated as potential adjunct therapy.

METHODS
Infants referred for evaluation of typical GERD symptoms
such as excessive vomiting, crying, irritability, feeding difficulties,
failure to thrive, and sleep disturbance were assessed for eligibility to
participate in the study. Patients were referred to the Department of
Gastroenterology of the Womens and Childrens Health Network
(WCHN) in Adelaide, Australia; the Royal Childrens Hospital in
Brisbane, Australia; and the Emma Children Hospital, Academic
Medical Centre in Amsterdam, the Netherlands. Parents/guardians
provided written consent before study procedures were started.
Ethical approval was provided by the ethical committees of all
participating hospitals. The study was registered with the Australia
New Zealand Clinical Trials Registry (ACTRN12609000988257).
All of the authors had access to the study data and had reviewed and
approved the final article.

Inclusion and Exclusion Criteria


Term and preterm infants up to 6 months postnatal age were
enrolled in the study if the following inclusion criteria were met:
parental/guardian written informed consent was given; any 2 of
apnea/bradycardia/oxygen desaturations, vomiting/gagging, and
irritability/pain were observed at least every second feed or at least
twice every 8 hours; and clinical symptoms were present for at least
5 days or increasing in frequency or severity for 3 days. Patients
with a history of acute life-threatening events, environmental
exposure to cigarette smoke, a history or a present need for resection
or reconstructive surgery of the gastrointestinal tract or with any
condition that may require surgery during the course of the study,
any condition that would make performance of the study procedures
unsafe, or that would make it unlikely that the patient would
complete the study procedures to the final day were excluded.

Having satisfied inclusion criteria, infants underwent a baseline assessment with combined pH-impedance, physiological
monitoring, an 8-hour video study, direct nursing observation,
and a gastric emptying breath test. The primary caregiver completed
a validated infant GERD questionnaire (I-GERQ-R questionnaire),
which has been used to quantify symptomatic responses of infants to
pharmacological and nonpharmacological reflux therapies (13)
(Fig. 1).
At baseline, infants were required to stop all of the medication known to affect acid secretion or gastric motility for at least
48 to 72 hours. Infants were studied in the supine nonelevated
position. The occurrence of symptoms was recorded on the tracing
in real time.
For diagnosis, we relied on a positive symptom association
probability (SAP) as recommended by present North American/
European Society for Pediatric Gastroenterology, Hepatology,
and Nutrition guidelines for diagnosis of infantile GERD (22).
Infants who demonstrated a positive GER-symptom association
(SAP > 95%) to any symptom occurring >5 times were randomized
to undergo 14 days of treatment.
Patients were randomized to 1 of 4 treatment groups using a
randomization schedule generated by an independent monitor.
Patients were usually discharged home for the treatment period
and received a minimum once per week visit by a clinical trial nurse
to ensure compliance with the treatment protocols. Treatment
groups were as follows:
1. Group 1: Left-side positioning for 2 hours postprandially in
combination with 1 mg  kg1  day1 esomeprazole once daily
(PPI LLP group).
2. Group 2: Head of cot elevation (HE) to 20 degrees for 2 hours
postprandially in combination with esomeprazole (PPI HE
group).
3. Group 3: Left-side positioning in combination with antacid
(AA) therapy, which consisted of at 1.5, 3, or 5 mL of Mylanta
(McNeil Consumer Healthcare, Ft Washington, PA) once daily
in infants 0 to 2, 2 to 4, and 4 to 6 months of age, respectively
(AA LLP group).
4. Group 4: HE to 20 degrees in combination with antacid (AA
HE group).
The reasoning behind the decision to perform the trial sham
controlled, rather than placebo controlled, was 3-fold: First, there is
no evidence that HE and AA actually reduce GER frequency or acid
exposure. Second, being used widely in the community, parents
were familiar with HE and AA and this allowed for better concealment of which treatments were hypothesized to be efficacious.
Finally, by using sham therapies rather than a placebo, parents were

8 h study

Randomizaion

GERD symptoms
Informed consent

pH-impedance
Cardioresp
monitoring
Sx monitored
Video
GE breath test
I-GERQ-R

Positive symptom
assosiation (SAP > 95%)
to symptoms of crying,
coughing or
regurgitation

ALTE
Cigarette smoke
Surgery
Unsafe to enroll

Volume 59, Number 2, August 2014

Study Protocol

Eligibility

Exclusion

TX 2 wk
Gr
Med

Pos

PPI

LLP

PPI

HE

AA

LLP

AA

HE

8 h study
pH-impedance
Cardioresp
monitoring
Sx monitored
Video
GE breath test
I-GERQ-R

FIGURE 1. Study protocol. Overview of study procedures. AA antacid; ALTE acute life-threatening event; GE gastric emptying;
GERD gastroesophageal reflux disease; Gr group; HE head of cot elevation; I-GERQ-R reflux questionnaire; LLP left lateral position;
PPI proton-pump inhibitor; SAP symptom association probability; Sx symptoms.

238

www.jpgn.org

Copyright 2014 by ESPGHAN and NASPGHAN. Unauthorized reproduction of this article is prohibited.

JPGN

Volume 59, Number 2, August 2014

reassured that, if not the trial treatment, a possible treatment would


be administered nevertheless.
PPI and AA therapies were administered double blind. That
is, the medication was prepared by the hospital pharmacy, parents
were not advised as to what treatment had been given, and investigators charged with analysis of studies had no knowledge of
therapies given. Body positioning interventions were performed
single blind. That is, parents were not provided with specific
information regarding which positioning treatment was predicted
to be more effective.
After 14 days, combined pH-impedance, physiological
monitoring, video study, gastric emptying breath test, and the
I-GERQ-R questionnaire were repeated on therapy and in the
allocated body position. Adverse events were identified by asking
parents/caretakers/hospital staff an open question about any health
issues during the study.

Precautions to Mitigate Sudden Infant Death


Syndrome Risk
The literature shows that side positioning (side unspecified)
increases sudden infant death syndrome (SIDS) risk (23). It has
been suggested that the doubled risk of SIDS in the side-sleeping
position is probably because of the side position being relatively
unstable, resulting in infants turning to the prone position (24). In
addition to the exclusion of infants who were exposed to cigarette
smoke in the home, we incorporated several additional precautionary measures into this study. To reduce the chance of accidental
prone positioning, parents were instructed to side position their
infant with lower arm and shoulder forward using a purposedesigned pillow.
The following precautions were also taken for all of the
infants enrolled. All of the parents received instruction on SIDSsafe infant dressing and arrangement of bedding. Positioning protocols were limited to the first 2 hours after a feed and only allowed
when parents were able to observe their infant. Finally, O2 saturation was continuously monitored during this period.

Measurement Methods
GER Monitoring
GER episodes were recorded using a single-use infant pHimpedance catheter with a pH sensor and 6 impedance channels.
The pH sensor was placed at the third vertebra above the diaphragm
and was confirmed by a thoracic x-ray. Depending on site preference, a ComforTec MII/pH probe was used in combination with a
Sleuth system recording device (Sandhill Scientific, Highlands
Ranch, CO), or the Unisensor pH-MII infant catheter was used
in combination with the Omega ambulatory recording system
(Medical Measurement Systems, Enschede, the Netherlands).
Studies were manually analyzed for the occurrence of liquid, mixed,
and gas GER according to accepted criteria (25). Following primary
analysis by 1 investigator, who was experienced with the use of both
software systems, each study was also checked for accuracy by a
second investigator. Acid exposure time in the distal esophagus and
total reflux index (% time pH < 4) >24 hours were determined
using automated software (GERD Check; Sandhill Scientific and
Omega analysis software [Medical Measurement Systems]).
For the purpose of accurate GER-symptom association,
symptoms suggestive of GER such as crying, irritability, fussing,
vomiting, burping, regurgitation, cough, sneeze, grimacing, and
back arching were continually observed by trained hospital staff and
recorded by direct input into the acquisition system. A 2-minute
time window was used for GER-symptoms association. To assess
www.jpgn.org

Treatment of Infant Gastroesophageal Reflux


GER-symptom association, the SAP was calculated. A positive
symptom association is defined by a SAP  95% for any symptom
type (26).

Physiological Monitoring
One of 2 cardiorespiratory and video data acquisition systems was used depending on site preference. In Adelaide and
Brisbane all cardiorespiratory and video data were acquired
simultaneously with GER monitoring on an Alice V sleep system
computer (Philips Respironics, Best, the Netherlands). The 8hr
physiological and video data were synchronized and merged with
GER data and analyzed in Bioview (Sandhill Scientific). In the
patients recruited in Amsterdam cardiorespiratory and video data
were acquired on the Embla polysomnography system with Remlogic analysis software (Broomfield, CO). pH-impedance tracings
were recorded on ambulatory Omega system and simultaneously on
the Remlogic system to ensure synchronization (GER analysis was
performed using the Omega software).
Cardiorespiratory tracings were analyzed for episodes of
bradycardia (heart frequency < 100/min), desaturation (transcutaneous SaO2 < 80%), and apnea (>20 seconds absence of nasal
flow and concurrent desaturation). Videos were also manually
reviewed to verify correct body positioning throughout the study
(ie, infant kept in randomized position for at least 50% of total time,
and at least 1 hour postprandially) and for the calculation of total
time of irritability, crying, and fussing (total crying time) as a posthoc measure of behavioral symptom severity.

Gastric Emptying Breath Test


Gastric half emptying time was measured with the 13C Naoctanoate (a stable isotope) breath test. 13C labeled Na-octanoate
(50 mg) was added to the infants feed, and breath samples were
taken before and for 4 hours after feeding. The 13CO2 excretion rate
was used to calculate gastric emptying (gastric emptying half time,
Tmax, and gastric emptying constant) using an established nonlinear
regression model (27).

Analysis and Outcome Measures


Investigators analyzed the pH/impedance data blinded to the
treatment of the infants. Primary outcome was defined as the change
in total aggregate number of symptoms and total crying time on
treatment. Secondary outcomes were the number of GER episodes
(acid and weakly acid, liquid, mixed, and gas), the number of crying,
coughing, and vomiting episodes (if present before treatment),
I-GERQ-R scores, gastric emptying, and cardiorespiratory events.
Additional post-hoc analysis was performed to assess the impact of
patient selection criteria based on automated pH-impedance analysis
(using autoscan) on symptomatic outcomes.

Statistical Analysis
Most variables were normally distributed and are presented
as mean (standard error of mean) unless otherwise stated. Primary
parameters at baseline were not significantly different (Student
t test). Consequently a paired t test was used to calculate the
difference between baseline and therapy per treatment group.
The change between baseline and therapy was compared among
the 4 treatment groups and for positioning intervention and medication separately using an analysis of covariance, with the number
of events after treatment as the dependent and the number of events
at baseline as a covariate. A Pearson correlation was used to

239

Copyright 2014 by ESPGHAN and NASPGHAN. Unauthorized reproduction of this article is prohibited.

Loots et al

JPGN

investigate treatment effects in relation to GER change or gastric


emptying change across all individual patients.
The additional post-hoc analyses of subgroups were performed using nonparametric tests because the groups were small
and not normally distributed. Wilcoxon signed rank test and MannWhitney U analyses were used to assess pre- and postdifferences
and differences between groups, respectively. Statistical analyses
were performed using SPSS Statistics 17.0 (SPSS, Chicago, IL).

RESULTS
Sixty-six infants were enrolled in the study between
September 2008 and May 2011 (16 patients in Amsterdam, 46
patients in Adelaide, and 4 patients in Brisbane). Fifty-one patients
were included in the study (27 boys, mean age 13.6 [range 226]
weeks). Eleven patients were born prematurely (median [range]
gestational age 32 [2837] weeks). Twelve patients received
treatment before enrolling in the study including PPIs, H2
antagonists, Mylanta, Gaviscon, Panadol, and Infacol (GlaxoSmithKline, Brentford, UK). No patient had undergone surgery.
Figure 2 shows the patient inclusion flowchart. After randomization, 6 patients were withdrawn, 3 patients were withdrawn before
the start of therapy by their parents without further specification of
motive, and 1 patient (randomized to AA therapy head elevation)
was withdrawn by the parents after 9 days on therapy. One patient
(randomized to PPI LLP) was withdrawn by the investigators
because of noncompliance. One patient (randomized to PPI HE)
was withdrawn after admission to hospital because of feeding
difficulties and a request by the treating pediatrician for the
medication to be changed. Baseline age and sex characteristics
per group are shown in Figure 2.
All of the patients were monitored for the risk of SIDS.
Physiological monitoring for apparent life-threatening events,
recurrent desaturations, brachio-/plagiocephaly, and neck muscle
dysfunction was unremarkable and gave no indication that the
interventions were unsafe to the infants when performed during
the short study period.

Treatment Effects on Gastroesophageal Reflux


Primary outcomes are presented in Table 1 for all of
the patients per treatment group. Compared with baseline, the

Volume 59, Number 2, August 2014

combination of PPI LLP was most effective in reducing the


number of GER episodes and esophageal acid exposure. Weakly
acid GER was reduced in the group treated with AA LLP, from
32 [4] to 18 [3], P 0.03 on treatment. Esophageal acid exposure
was reduced in all groups treated with PPI (Table 1).

Treatment Effects on Symptoms


Symptomatic changes are shown in Table 2. Of patients who
demonstrated vomiting before treatment, vomiting was reduced in
the AA LLP group (Table 2). Furthermore LLP, regardless of
medication, significantly reduced vomiting after treatment (from
7 [2] episodes at baseline to 1 [1] episode on treatment, P 0.042).
None of the treatment groups showed a symptomatic improvement
as measured by a decrease in total irritability, crying, and fussing
(total crying time) (Table 2). Symptoms of cough and crying were
not reduced in any of the treatment groups either (Table 3).
Parentally reported I-GERQ-R scores decreased in all of the groups,
reaching statistical significance for both the PPI LLP group (from
21 [1] to 18 [2], P 0.02) and AA HE group from (24 [2] to
20 [2], P 0.001).

Treatment Effects on Gastric Emptying


Gastric emptying results are shown in Table 3. Gastric
emptying halftime was significantly delayed compared with baseline in LLP patients (39 [19] minutes slower) vs HE patients (10 [8]
minutes faster, P 0.038) regardless of medication. Gastric emptying Tmax was significantly slower on therapy in the PPI LLP
group (55 [5] minutes vs 78 [8] minutes pre- and posttreatment,
respectively, P 0.015).

Treatment Effect of LLP and PPI Independent


of Other Therapies
LLP overall produced a greater reduction in total GER when
compared with HE overall (Table 4). PPI therapy overall produced a
greater reduction in reflux index compared with AA therapy overall
(Table 4).

66 patients eligible
9 patients not randomized
57 patients randomized
6 patients withdrawn*
51 patients included for
analysis

Group 1
PPI and LLP
N = 12 (4 boys)
Age: 12(3) wk

Group 2
PPI and HE
N = 14 (9 boys)
Age: 12(3) wk

Group 3
Antacid and LLP
N = 13 (7 boys)
Age: 14(2) wk

Group 4
Antacid and HE
N = 12 (7 boys)
Age: 17(2) wk

FIGURE 2. Patient flowchart. Withdrawn patients: before start therapy (N 3), no compliance (N 1, PPI LLP), parents did not wish to
proceed after 9 days on therapy (N 1, AA HE), reduced oral intake, and hospital admission (N 1, PPI HE). AA antacid; HE head of cot
elevation; LLP left lateral position; PPI proton-pump inhibitor.

240

www.jpgn.org

Copyright 2014 by ESPGHAN and NASPGHAN. Unauthorized reproduction of this article is prohibited.

JPGN

Volume 59, Number 2, August 2014

Treatment of Infant Gastroesophageal Reflux

TABLE 1. Effect of different treatments upon GER variables


GER variables

Pretreatment

Total no. GER


PPI LLP
PPI HE
AA LLP
AA HE
Reflux index, %
PPI LLP
PPI HE
AA LLP
AA HE

Posttreatment

Difference



69
50
47
55

(13)
(6)
(6)
(9)

46
41
29
44

(10)

(6)

(3)
(7)

8.9
6.9
2.9
5.8

(3.118.1)
(2.121.8)
(0.66.9)
(2.812.6)

1.1
0.5
4.8
1.1

(04.4)

(0.12.2)
(0.89.0)
(0.29.9) (0.066)

23
9
18
11

(5)
(3)
(6)
(8)

4
5.1
0.9
2.2

(12.9 to 0.2)
(21.2 to 0.1)
(2.0 to 2.8)
(8.9 to 0.2)

Mean (SEM) or median (IQR) and mean difference shown for each treatment group. Pretreatment was significantly different from posttreatment using



P < 0.001). P values for ANOVA across the 4 treatment
paired t test or signed rank test (P values 0.050.10 shown in parentheses, P < 0.05, P < 0.01,
groups for pre-/posttreatment and difference were not significant. AA antacid; ANOVA analysis of variance; GER gastroesophageal reflux; HE head
of cot elevation; IQR interquartile range; LLP left lateral position; PPI proton-pump inhibitor; SEM standard error of mean.

Relation Between GER Change, Gastric


Emptying Change, and Symptom Change

P < 0.001). Slower gastric emptying did not correlate with any
specific symptom subtype.

Among all of the patients, an improvement in the number of


symptoms overall did not correlate with the reduction in GER (total
number of GER r 0.13, P 0.4, acid GER r 0.18, P 0.26,
weakly acid GER r 0.005, P 0.96 and reflux index r 0.006,
P 0.972). Reduced GER did not correlate with any specific
symptom subtype.
A relation between symptom improvement overall and a
slowing of the gastric emptying rate was observed (r 0.54,

Post-Hoc Stratification Based on Automated


Impedance Analysis

TABLE 2. Effect of different treatments upon symptom assessments


Symptom assessments
Total crying time, min
PPI LLP
PPI HE
AA LLP
AA HE
No. cough Sx
PPI LLP
PPI HE
AA LLP
AA HE
No. vomit Sx
PPI LLP
PPI HE
AA LLP
AA HE
No. cry Sx
PPI LLP
PPI HE
AA LLP
AA HE

Pretreatment

Posttreatment

Difference

1
9
17
8

92
71
106
74

(10)
(11)
(19)
(20)

92
81
88
66

(10)
(10)
(10)
(13)

22
23
30
32

(7)
(6)
(12)
(9)

25
25
31
42

(7)
(6)
(14)
(12) (0.1)

4
2
2
11

(9)
(6)
(5)
(6)

6
6
2
2

(2)
(2)

(0)
(1)

1
2
3
0

(1)
(3)
(1)
(0)

48
49
54
35

(8)

(7)
(9)
(7)

5
17
7
5

(5)
(10)
(13)
(10)

7(2)
8 (3)
7 (2)
3 (1)
48
30
60
38

(9)
(7)
(12)
(10)

(7)
(9)
(18)
(13)

Mean (SEM) or median [IQR] and mean difference shown for each

treatment group. Pretreatment was significantly different from posttreatment using paired t test or signed rank test (P values 0.050.10 shown



P < 0.001). P values for ANOVA
in parentheses, P < 0.05, P < 0.01,
across the 4 treatment groups for pre/post treatment and difference were not
significant. AA antacid; ANOVA analysis of variance; GER
gastroesophageal reflux; HE head of cot elevation; IQR interquartile
range; LLP left lateral position; PPI proton-pump inhibitor; SEM
standard error of mean; Sx symptoms.

www.jpgn.org

Impedance baseline recordings were reanalyzed by automated analysis software. Differences in GER parameters based
on automated analysis were statistically similar to manual analysis;
however, 7 patients who were SAP positive based on manual
analysis were found not to be SAP positive on automated analysis.
When these patients were removed from the database, we were able
to observe some treatment effects in the remaining patients. In the
PPI LLPtreated group (n 7), crying in the (autoscan-determined) SAP-positive infants reduced after treatment from 99
(65103) to 62 (3296) episodes (P 0.018, Wilcoxon signed
rank test) and in the PPI HE group (n 5) vomiting reduced from
17 (633) to 10 (1.520) (P 0.042, Wilcoxon signed rank test).
TABLE 3. Effect of different treatments upon gastric emptying variables
GE variables
GEt1=2
PPI
PPI
AA
AA
GEtmax
PPI
PPI
AA
AA

Pretreatment

Posttreatment

Difference

39
16.8
39
8

(22)
(12)
(31)
(11)

22
13
8
12

(7)
(6)
(12)
(9)

LLP
HE
LLP
HE

53
46
57
62

(6)
(12)
(10)
(11)

92
29
96
61

(22)
(7)
(39)
(14)

LLP
HE
LLP
HE

55
65
71
66

(5)
(7)
(4)
(9)

78
52
79
77

(8)
(3) (0.051)
(13)
(7)

Mean (SEM) or median (IQR) and mean difference shown for each
treatment group. GEt1=2 : gastric emptying half time. GEtmax : gastric emptying time to maximum 13C concentration. GEGEC: gastric emptying constant.

Pretreatment was significantly different from posttreatment using paired
t test or signed rank test (P values 0.050.10 shown in parentheses,



P < 0.05, P < 0.01,
P < 0.001). P values for ANOVA across the
4 treatment for the difference in GEtmax was 0.061. AA antacid;
ANOVA analysis of variance; GER gastroesophageal reflux; HE head
head of cot elevation; IQR interquartile range; LLP left lateral position;
PPI proton-pump inhibitor; SEM standard error of mean.

241

Copyright 2014 by ESPGHAN and NASPGHAN. Unauthorized reproduction of this article is prohibited.

Loots et al

JPGN

Volume 59, Number 2, August 2014

TABLE 4. Differences in GER, GE, and symptom variables in relation to body positioning or acid-suppression therapy
LLP vs HE
Difference

PPI vs AA

LLP

GER variables
Total no. GER
No. acid GER
No. WA GER
Reflux index (%)
GE variables
GE 1/2 time
Sx assessments
Total ICF, min
No. cough Sx
No. vomit Sx
No. cry Sx

21
10
11
3.4

(4)
(3)
(3)
(2.1)

39 (19)
8
3
2
1

(9)
(5)
(1)
(7)

HE

10
9
0
4.5

(4) (0.056)
(3)
(4) (0.052)
(1.7)

10 (8)
1
6
1
5

PPI

(8)
(4)
(1)
(7)

16
12
4
6.8

AA

15
6
7
0.9

(3)
(3)
(3)
(2.1)

8 (13)
4
3
2
11

(5)
(3)
(5)

(1.4)

24 (17)
12
6
2
6

(6)
(6)
(1)
(5)


(11)
(4)
(1)
(8) (0.09)



Data presented as mean (SEM). Treatments were significantly different using t test (P values 0.050.10 shown in parentheses, P < 0.05, P < 0.01,
P < 0.001). AA antacid; GER gastroesophageal reflux; HE head of cot elevation; IQR interquartile range; LLP left lateral position;
PPI proton-pump inhibitor; SEM standard error of mean; Sx symptoms; WA Weakly acidic.


Adverse Events
During the study 5 patients experienced adverse events, of
which 2 were serious. The adverse events were urinary tract
infection, constipation, diarrhea, and vomiting (following immunization). Of the 2 patients with serious adverse events, 1 patient
(randomized to PPI HE) was admitted to hospital after 5 days of
treatment and successfully treated for rotavirus infection. Another
patient (randomized to PPI HE) was admitted because of reduced
oral intake and weight loss. None of the adverse events were judged
to be related to the study therapies by the treating pediatricians.

DISCUSSION
In this randomized sham-controlled study we assessed the
influence of body positioning and medical therapies on the frequency of GER and typical symptoms in infants with positive SAP
findings on pH-impedance monitoring. This is the first study to only
include patients with a positive SAP and to assess the influence of
therapies designed to reduce reflux triggering and reduce gastric
acidity. LLP therapy reduced the number of impedance-detected
GER episodes, and PPI therapy reduced acid GER. The use of both
therapies in combination was the most effective way of reducing
total GER frequency and esophageal acid exposure; however,
despite this potent effect on GER, there was no concomitant
reduction in symptoms other than vomiting.
The purpose of this study was to determine the effect of
known reflux-reducing interventions on GER and typical symptoms
that would, in present clinical practice, lead to the prescription of
antireflux therapies. With overwhelming clinical trial evidence that
PPIs are ineffective for reducing typical GERD symptoms in the
infant group (13,15,16,28), the present trial addresses the possibility
that PPIs may lack efficacy because acid suppression does not
reduce the total reflux burden of combined acid, weakly acidic, and
nonacidic episodes (15,17). LLP was chosen based on previously
published observations that LLP significantly reduces total reflux
and because pharmacological reflux inhibitor treatments, such as
baclofen, have adverse effects, which preclude use in infants 0 to
6 months (18,19).
Despite use of therapies that clearly ameliorated the total
reflux burden, no objective symptomatic improvement was demonstrated in our patients. Parent reported I-GERQ-R scores improved
in infants randomized to AA HE (sham sham group) as well as

242

PPILLP (active active group). The improvement in the


I-GERQ-R score in the AA HE group, that did not show a
concomitant reduction in GER, appears counterintuitive. This
finding suggests that AA HE treatment was having a therapeutic
effect unrelated to GER or, alternatively, that parental familiarity
with this widely used conservative therapy could have influenced
the subjective I-GERQ-R scores.
Possible explanations for the lack of a symptom response
overall include the relatively short treatment period and the fact that
objective symptom monitoring only occurred for 8 daytime hours.
Patient selection required a positive SAP, but this may also have
been suboptimal. The SAP is recognized to be the strongest
statistical approach to determine whether GER episodes and symptoms are causally related (8,26,29); however, SAP is problematic.
SAP was originally conceived, designed, and validated (26) for the
assessment of symptoms of (adult) heartburn. The number of
symptoms typically recorded per infant study is much higher
(120 [7] symptoms per patient) than the typical number of symptoms of (adult) heartburn. The frequency of GER is also higher in
infants (30), and the combination of high numbers of symptoms and
reflux episodes increases the chance that a GER episode and a
symptom event are associated in time. Poor inter- and intrarater
reliability of manual scoring of pH-impedance recordings may be
problematic (31). Interestingly, a post-hoc analysis of patient
recordings using automated analysis software yielded fewer
patients who were SAP positive and a significant improvement
in crying symptoms in infants randomized to the PPI LLP group.
These findings in relation to automated analysis clearly warrant
further exploration.
Across all of the patients randomized to the 4 different
treatment arms, an inverse relation was demonstrated between
gastric emptying rate and symptom change. Although our observation that slower gastric emptying may improve symptoms
seems counterintuitive, because delayed GE is thought to exacerbate vomiting, it may also contribute to satiety. There is also
evidence in the neonatal literature showing that recommended
doses of the prokinetic drug cisapride may slow or leave GE
unchanged, rather than accelerate it (22,32). We have published
several articles on the influence of lateral positioning on TLESRs
and GER showing in LLP that GER is reduced while gastric
emptying is paradoxically delayed (1820). Clearly the relation
between gastric emptying, GER, and typical symptoms requires
further investigation.
www.jpgn.org

Copyright 2014 by ESPGHAN and NASPGHAN. Unauthorized reproduction of this article is prohibited.

JPGN

Volume 59, Number 2, August 2014

It is important to recognize that LLP was used in this clinical


trial as a concept test for pharmacological reflux inhibition,
which may emerge from future drug development. We present
evidence that LLP may reduce vomiting symptoms; however,
caution needs to be exercised because side-positioning contravenes
SIDS-safe guidelines and may contribute to plagiocephaly and neck
muscle dysfunction. We undertook several measures to eliminate
any potential for these outcomes, including the use of a brace pillow
to prevent rolling to the prone position and careful control and
monitoring of the interventions used. The requirement for such
measures, and the widespread availability of safe treatments for
regurgitation and vomiting, precludes lateral positioning from being
recommended for use in the community under any circumstances.
In conclusion, symptomatic GERD implies a disease
causation for distressing infant symptoms that provokes a targeted
therapeutic response. We present the first sham-controlled trial in
GER-symptompositive infants, which used therapies designed to
reduce total GER triggering and acid exposure. A reduction in total
GER and/or esophageal acid exposure did not result in a significant
improvement in symptoms other than vomiting; however, automated analysis appeared to identify infants who responded to
positioning therapy. This is an important new insight for future
research owing to the inherent advantages of automated analysis in
reducing the time and costs associated with administration of pHMII tests.
Acknowledgments: The authors thank Jan James, Ros Lontis,
Louise Goodchild, Amanda Warden, Rosa Katsikeros, and Julia van
de Pol for their help and support with conducting this study, and the
University of Pittsburgh for access to the I-GERQ-R.

REFERENCES
1. Van Howe RS, Storms MR. Gastroesophageal reflux symptoms in
infants in a rural population: longitudinal data over the first six months.
BMC Pediatr 2010;10:7.
2. Omari TI, Barnett C, Snel A, et al. Mechanisms of gastroesophageal
reflux in healthy premature infants. J Pediatr 1998;133:6504.
3. Omari TI, Barnett CP, Benninga MA, et al. Mechanisms of gastrooesophageal reflux in preterm and term infants with reflux disease. Gut
2002;51:4759.
4. Holloway RH, Penagini R, Ireland AC. Criteria for objective definition
of transient lower esophageal sphincter relaxation. Am J Physiol
1995;268 (1 pt 1):G12833.
5. Boeckxstaens GE, Hirsch DP, Fakhry N, et al. Involvement of cholecystokinin A receptors in transient lower esophageal sphincter relaxations triggered by gastric distension. Am J Gastroenterol 1998;93:
18238.
6. Mittal RK, Holloway RH, Penagini R, et al. Transient lower esophageal
sphincter relaxation. Gastroenterology 1995;109:60110.
7. Zerbib F, Bruley D, Scarpignato V, et al. Endogenous cholecystokinin in
postprandial lower esophageal sphincter function and fundic tone in
humans. Am J Physiol 1998;275 (6 pt 1):G126673.
8. Loots CM, Benninga MA, Davidson GP, et al. Addition of pHimpedance monitoring to standard pH monitoring increases the yield
of symptom association analysis in infants and children with gastroesophageal reflux. J Pediatr 2009;154:24852.
9. Rosen R, Lord C, Nurko S. The sensitivity of multichannel intraluminal
impedance and the pH probe in the evaluation of gastroesophageal
reflux in children. Clin Gastroenterol Hepatol 2006;4:16772.
10. Salvatore S, Arrigo S, Luini C, et al. Esophageal impedance in children:
symptom-based results. J Pediatr 2010;157:94954.
11. Barron JJ, Tan H, Spalding J, et al. Proton pump inhibitor utilization
patterns in infants. J Pediatr Gastroenterol Nutr 2007;45:4217.

www.jpgn.org

Treatment of Infant Gastroesophageal Reflux

12. Higginbotham TW. Effectiveness and safety of proton pump inhibitors


in infantile gastroesophageal reflux disease. Ann Pharmacother 2010;
44:5726.
13. Orenstein SR, Hassall E, Furmaga-Jablonska W, et al. Multicenter,
double-blind, randomized, placebo-controlled trial assessing the efficacy and safety of proton pump inhibitor lansoprazole in infants with
symptoms of gastroesophageal reflux disease. J Pediatr 2009;154:514
20.
14. van der Pol RJ, Smits MJ, van Wijk MP, et al. Efficacy of proton-pump
inhibitors in children with gastroesophageal reflux disease: a systematic
review. Pediatrics 2011;127:92535.
15. Omari TI, Haslam RR, Lundborg P, et al. Effect of omeprazole on acid
gastroesophageal reflux and gastric acidity in preterm infants with
pathological acid reflux. J Pediatr Gastroenterol Nutr 2007;44:414.
16. Moore DJ, Tao BS, Lines DR, et al. Double-blind placebo-controlled
trial of omeprazole in irritable infants with gastroesophageal reflux.
J Pediatr 2003;143:21923.
17. Omari TI, Lundborg P, Sandstrom M, et al. Pharmacodynamics and
systemic exposure of esomeprazole in preterm infants and term neonates
with gastroesophageal reflux disease. J Pediatr 2009;155:2228.
18. Omari TI, Rommel N, Staunton E, et al. Paradoxical impact of body
positioning on gastroesophageal reflux and gastric emptying in the
premature neonate. J Pediatr 2004;145:194200.
19. van Wijk MP, Benninga MA, Dent J, et al. Effect of body position
changes on postprandial gastroesophageal reflux and gastric emptying
in the healthy premature neonate. J Pediatr 2007;151:58590.
20. van Wijk MP, Benninga MA, Davidson GP, et al. Small volumes of feed
can trigger transient lower esophageal sphincter relaxation and gastroesophageal reflux in the right lateral position in infants. J Pediatr 2010;
156:7448.
21. Omari T. Gastroesophageal reflux in infants: can a simple left side
positioning strategy help this diagnostic and therapeutic conundrum?
Minerva Pediatr 2008;60:193200.
22. McClure RJ, Kristensen JH, Grauaug A. Randomised controlled trial of
cisapride in preterm infants. Arch Dis Child Fetal Neonatal Ed
1999;80:F1747.
23. Blair PS, Platt MW, Smith IJ, et al. Sudden infant death syndrome and
sleeping position in pre-term and low birth weight infants: an opportunity for targeted intervention. Arch Dis Child 2006;91:1016.
24. Mitchell EA. Recommendations for sudden infant death syndrome
prevention: a discussion document. Arch Dis Child 2007;92:1559.
25. Sifrim D, Holloway R, Silny J, et al. Acid, nonacid, and gas reflux
in patients with gastroesophageal reflux disease during ambulatory
24-hour pH-impedance recordings. Gastroenterology 2001;120:1588
98.
26. Weusten BL, Roelofs JM, Akkermans LM, et al. The symptom-association probability: an improved method for symptom analysis of 24-hour
esophageal pH data. Gastroenterology 1994;107:17415.
27. Omari TI, Symonds E, Davidson GP, et al. Is the correction factor used
in the breath test assessment of gastric emptying appropriate for use in
infants? J Pediatr Gastroenterol Nutr 2005;41:3324.
28. Winter H, Kum-Nji P, Mahomedy SH, et al. Efficacy and safety of
pantoprazole delayed-release granules for oral suspension in a placebocontrolled treatment-withdrawal study in infants 111 months old with
symptomatic GERD. J Pediatr Gastroenterol Nutr 2010;50:60918.
29. Omari TI, Schwarzer A, vanWijk MP, et al. Optimisation of the refluxsymptom association statistics for use in infants being investigated by
24-hour pH impedance. J Pediatr Gastroenterol Nutr 2011;52:40813.
30. Lopez-Alonso M, Moya MJ, Cabo JA, et al. Twenty-four-hour esophageal impedance-ph monitoring in healthy preterm neonates: rate and
characteristics of acid, weakly acidic, and weakly alkaline gastroesophageal reflux. Pediatrics 2006;118:e299308.
31. Loots CM, van Wijk MP, Blondeau K, et al. Interobserver and intraobserver variability in pH-impedance analysis between 10 experts
and automated analysis. J Pediatr 2012;160:441.
32. Barnett CP, Omari T, Davidson GP, et al. Effect of cisapride on gastric
emptying in premature infants with feed intolerance. J Paediatr Child
Health 2001;37:55963.

243

Copyright 2014 by ESPGHAN and NASPGHAN. Unauthorized reproduction of this article is prohibited.