Open access Original article

Antacid therapy for gastroesophageal

bmjpo: first published as 10.1136/bmjpo-2018-000287 on 9 July 2018. Downloaded from http://bmjpaedsopen.bmj.com/ on February 28, 2024 by guest. Protected by copyright.
reflux in preterm infants: a
systematic review
Elda Dermyshi, Charley Mackie, Phoebe Kigozi, Bernard Schoonakker, Jon Dorling

To cite: Dermyshi E, Mackie C, Abstract

Kigozi P, et al. Antacid therapy What is already known on this topic?
Background Gastro-oesophageal reflux is prevalent
for gastroesophageal reflux in in preterm infants. Despite widespread use in clinical
preterm infants: a systematic ►► Gastro-oesophageal reflux is a prominent condition
practice, there is still much controversy over the efficacy
review. BMJ Paediatrics Open among preterm infants.
and safety of drug interventions, particularly antacid
2018;2:e000287. doi:10.1136/ ►► Antacids are often used to treat gastro-oesophageal
bmjpo-2018-000287 therapy.
reflux disease (GORD), despite the lack of good qual-
Objective To systematically review the effects of antacid
ity evidence to support its use.
therapy on preterm infants with symptoms of gastro-
►► Studies have shown a significant positive correlation
Received 12 March 2018 oesophageal reflux, and to assess the safety of these
Revised 12 May 2018
between the use of histamine-2 receptor antago-
interventions.
Accepted 14 May 2018 nists and important complications.
Methods We carried out an electronic search of the
Cochrane central register of controlled trials (CENTRAL,
The Cochrane Library), MEDLINE (1966–present), EMBASE
(1980–present) and CINAHL (1982–present) as well as What this study hopes to add?
other online sources. Participants were preterm infants
(<37 weeks gestation) with gastro-oesophageal reflux ►► There is limited evidence supporting the use of ant-
disease who were receiving care on a neonatal unit. We acids in preterm infants.
assessed the effects of histamine-2 receptor antagonists, ►► Omeprazole reduced gastric and oesophageal pH,
proton pump inhibitors and alginates against placebo, but did not alter GORD symptoms. Esomeprazole and
primarily to see if they reduced the symptoms of reflux. lansoprazole had no significant effect on GORD signs
Results Six studies were included in this review. Meta- and symptoms.
analysis could not be carried out due to a lack of studies ►► Combined use of ranitidine and metoclopramide ap-
assessing the same intervention with the same outcomes. pears countereffective, with placebo periods giving
Omeprazole therapy significantly reduced the oesophageal less bradycardia episodes versus drug periods.
acid exposure percentage time with pH<4 (p<0.01)
and sodium alginate significantly decreased gastro-
oesophageal reflux episodes (p=0.024). Metoclopramide frequent vomiting, apnoea and bradycardia,
and ranitidine showed a significant increase in gastro- aspiration pneumonia, aversion to feeding
oesophageal reflux disease symptoms versus placebo and exacerbation of chronic lung disease
(p<0.04). No significant results were found for the use of
in term and preterm infants with associated
esomeprazole or lansoprazole versus placebo.
Conclusions There is insufficient evidence available to resource implications from longer hospital
conclude whether antacid therapy is effective or safe when stays.
treating gastro-oesophageal reflux disease in preterm Despite this, controversy remains over how
infants. Further research is needed into this topic and it should be treated. Currently, non-pharma-
caution should be taken when administering antacids to cological therapies are generally the first line
preterm infants. of treatment in GORD with pharmacolog-
Trial registration number CRD42017078778 ical therapies reserved for those who do not
© Author(s) (or their
employer(s)) 2018. Re-use respond.1 Antacids containing alginate, hista-
permitted under CC BY-NC. No mine-2 receptor antagonists (H2 RAs) and
commercial re-use. See rights Background proton pump inhibitors (PPI) are among the
and permissions. Published by Gastro-oesophageal reflux (GOR) is a prom- most common interventions used with 60%,
BMJ.
inent condition among preterm infants. 53% and 23% of UK neonatal units using
Neonatal Intensive Care Unit,
Symptoms such as apnoeas, desaturation, these products, respectively.2
Nottingham University Hospitals,
Nottingham, UK bradycardia, vomiting, poor weight gain and Studies have also shown a significant
irritability have been attributable to GOR, correlation between the use of H2 RAs
Correspondence to
which is called gastro-oesophageal reflux and important complications.3 4 Guillet et
Dr Elda Dermyshi; e​ lda.​ disease (GORD), when symptoms are severe. al showed H2-blocker use was associated
dermyshi@y​ ahoo.​com GORD has been reported to cause irritability, with an increased incidence of necrotising

Dermyshi E, et al. BMJ Paediatrics Open 2018;2:e000287. doi:10.1136/bmjpo-2018-000287 1

Open access

enterocolitis (NEC) (OR 1.71; 95% CI 1.34 to 2.19; Types of interventions

bmjpo: first published as 10.1136/bmjpo-2018-000287 on 9 July 2018. Downloaded from http://bmjpaedsopen.bmj.com/ on February 28, 2024 by guest. Protected by copyright.
p<0.0001).3 We included all available RCTs evaluating antacid ther-
There continues to be a widespread use of antacid apies for GOR in preterm neonates. Antacid therapy
therapy in neonatal units today despite the evidence gaps. (administered by any method) should have been
This review was carried out to systematically evaluate the commenced after the diagnosis of GORD and continued
evidence of efficacy and safety of antacid treatment for for any duration.
GORD in preterm infants and to highlight potential The interventions considered were:
areas for future research. ►► H2 RAs versus a placebo or standard care/non-phar-
macological therapy.
►► PPIs versus a placebo or standard care/non-pharma-
cological therapy.
Objectives ►► Alginates versus a placebo or standard care/non-phar-
Primary objective macological therapy.
To assess the efficacy of antacid therapy in preterm Trials were not limited by dose, frequency or duration
infants diagnosed with GORD. of intervention.

Secondary objective Selection of studies

To assess the safety of antacid therapy in preterm infants Paired reviewers (ED, CM, BS, JD) independently
diagnosed with GORD. screened titles, abstracts and then full texts for eligibility,
assessed risk of bias and collected data from included
studies. Any disagreement between reviewers was resolved
through discussion or adjudication by a third reviewer
Materials and methods (BS, JD). In case of duplicate publications, the most
We used Preferred Reporting Items for Systematic Reviews recent and updated report of the study was included.
and Meta-analyses guidelines and the Cochrane Hand-
book of Systematic Reviews of Interventions approach Risk of bias and quality of evidence assessment
for conducting and reporting systematic reviews and The Cochrane Risk-of-Bias Tool was used to assess the
meta-analyses of randomised controlled trials (RCTs).5 6 risk of bias.7 The quality of the evidence of outcomes was
The methodology of this systematic review was published rated by the Grading of Recommendations Assessment,
in PROSPERO (www.​crd.​york.​ac.​uk/​PROSPERO; ref Development, and Evaluation (GRADE) approach.8
CRD42017078778).
Data extraction
Search methods for identification of studies From each eligible study the following information was
MEDLINE/PubMed, Embase, Wiley Online Library, collected: study characteristics (eg, author name, year of
Cochrane Library and Web of Science databases were publication, sample size, patient characteristics, antacid
searched to identify trials of antacid therapy in preterm type, duration of intervention, dosage and any of our
infants. Databases were screened for publications from preplanned clinical outcomes).
the earliest available date until 15 October 2017. No
Primary outcomes
language restrictions were applied. Ethical approval
A reduction in reflux symptoms assessed by a reflux
was not required because only published articles were
index score or bedside symptom charts.9 Clinical symp-
included in this review. A database search of ​clinicaltrials.​
toms include the following: total GOR episodes, vomit/
gov for ongoing and completed trials was also carried
regurgitations, choking/coughing, bradycardia attrib-
out, using the search terms infant or preterm and reflux
uted to GOR, behavioural/crying, feeding difficulties,
or gastroesophageal reflux. Trials reported as abstracts
irritability or pain, recurrent postprandial apnoeas and
or letters to the editor were included if sufficient data
oxygen desaturation within 2 hours postprandial period.
to fulfil the inclusion criteria were presented within the
report, or provided by authors. Full search strategy is Secondary outcomes
presented in online supplementary appendix 1. ►► Time taken to establish full enteral feeds
►► Length of hospital stay
Eligibility criteria ►► NEC (Bell’s stage 2 or greater)
All relevant randomised trials involving preterm infants ►► Suspected or proven sepsis
(<37 weeks gestation) with GORD (clinical diagnosis ►► Other adverse effects
and/or 24 hours intraoesophageal pH monitoring, or
impedance studies) receiving care on a neonatal unit. Statistical analysis
Crossover, randomised trials or quasi-randomised studies, We planned to analyse treatment effects in the indi-
described in some way as to suggest or imply that the vidual trials using Review Manager 5.3 software, with
study was randomised if the demographic detail of each risk ratio and risk difference for dichotomous data and
group was similar were included. mean difference for continuous data, with respective

2 Dermyshi E, et al. BMJ Paediatrics Open 2018;2:e000287. doi:10.1136/bmjpo-2018-000287

 Open access

95% CIs. However, given the small number of included assessed for eligibility and six studies met our inclusion

bmjpo: first published as 10.1136/bmjpo-2018-000287 on 9 July 2018. Downloaded from http://bmjpaedsopen.bmj.com/ on February 28, 2024 by guest. Protected by copyright.
studies, their varying methodologies and interventions, criteria (figure 1).
we judged quantitative meta-analysis to be inappropriate All included studies were double-blind, randomised,
and instead report a narrative description of each study. placebo-controlled trials. Four of the six were crossover
Data are presented as reported in individual studies. We trials (Wheatley and Kennedy,10 Omari et al,11 Corvaglia
had also planned to conduct a number of subgroup anal-
et al,12 Corvaglia et al13), while the remaining two were
yses, which are detailed in the study protocol. The small
number of studies, with small sample sizes and variable parallel trials (Orenstein et al14 and Davidson et al15).
methods precluded subgroup analyses. The main characteristics of included RCTs are described
in table 1 and excluded studies are summarised in online
supplementary appendix 2.
Results
Description of studies Risks of bias assessments of trials are summarised in
A total of 20 139 records were identified by the initial figure 2 and online supplementary appendix 3. The eval-
search; 18 909 were excluded as they were duplicates, or uations of the level of evidence of outcomes according to
systematic reviews; 1230 titles and abstracts were screened the GRADE approach are summarised in online supple-
and 1202 were excluded. Twenty-eight full-text articles were mentary appendix 4.

Figure 1 Study flow diagram. RCT, randomised controlled trial.

Dermyshi E, et al. BMJ Paediatrics Open 2018;2:e000287. doi:10.1136/bmjpo-2018-000287 3

4
Table 1 Characteristics of included studies
Wheatley
Studies Corvagliaet al12 Corvaglia et al13 Davidson et al15 Omari et al11 Orenstein et al14 and Kennedy10
Open access

Methods Clinical trial— Clinical trial— Randomised, double-blind, Randomised, double- Multicentre, double-blind, Randomised,
crossover of treatment crossover of placebo controlled trial blind, placebo- randomised, placebo- controlled, blind
and placebo treatment and controlled, crossover controlled trial crossover study of
placebo design trial treatment and placebo
Participants 32 Preterm newborns 28 Preterm 52 Term infants or 10 Preterm infants with 162 Infants aged 16 weeks 18 Preterm<37 weeks
(gestational newborns with a gestational or a mean postmenstrual (median, range 4–51) and corrected
age≤33 weeks) (gestational postconceptional age of age of 36.1±0.7 (range, gestation at birth 35 weeks gestational age at
age≤33 weeks) 28–44 weeks 34–40 weeks) (median, range 25–39) enrolment<44 weeks
Diagnostic Frequent regurgitations Recurrent Two of the following clinical Infants with symptoms Infants with symptomatic Clinical diagnosis of
symptoms and/or postprandial postprandial findings: apnoea±bradycardia; of GORD, confirmed by GORD who remained GOR and bradycardia
desaturations) apnoeas ±oxygen desaturations, 24 hour pH monitoring symptomatic within attributed to GOR by
vomiting or gagging, and with significant reflux 1 hour after feeding clinicians
irritability or pain at least every index despite at least 1 week
second feed or at least twice of non-pharmacological
every 8 hours management
Interventions 0.25 mL/kg sodium 0.25 mL/kg sodium Esomeprazole 0.5 mg/kg or Omeprazole (0.7 mg/ Lansoprazole administered Metoclopramide
alginate was given alginate after one placebo once daily for up to kg) or placebo (days once daily at 0.2–0.3 mg/kg/ (0.2 mg/kg/dose every
four times at alternate single meal (DG 14 days 1–7) and then the day for infants age≤10 weeks 6 hours) and ranitidine
meals (DG meals), meal) or placebo (DF alternative treatment and at 1.0–1.5 mg/kg/day (2 mg/kg/dose) every
remaining four meals meal) regimen was given for for those age>10 weeks or 8 hours or placebo.
were placebo (DF the second week (days placebo Each infant was
meals) 8–14) (maximum 4 randomly assigned
weeks of study drug to one of two study
treatment) groups
Primary GOR features Apnoea episodes Change from baseline to Gastric acidity, Number and duration of Bradycardia episodes
outcomes (ie, number, acidity, and gastro- end of treatment in the oesophageal acid crying episodes during per day
duration and height of oesophageal total number of GORD- exposure and the or ≤1 hour after feeding and
GORs) features related symptoms and number and duration of frequency of various GORD
signs (vomiting, apnoea, acid reflux episodes symptoms quantified in daily
bradycardia, oxygen diaries
desaturation, gagging, back
arching, irritability, crying and
fussing)
Secondary NS NS Adverse events Number of vomiting, Adverse events NS
outcomes apnoea, bradycardia or
behavioural changes
DF, drug-free; DG, drug-given; GOR, gastro-oesophageal reflux; GORD, gastro-oesophageal reflux disease; NS, not specified.

Dermyshi E, et al. BMJ Paediatrics Open 2018;2:e000287. doi:10.1136/bmjpo-2018-000287

bmjpo: first published as 10.1136/bmjpo-2018-000287 on 9 July 2018. Downloaded from http://bmjpaedsopen.bmj.com/ on February 28, 2024 by guest. Protected by copyright.
 Open access

of hospital stay, NEC and suspected or proven sepsis.

bmjpo: first published as 10.1136/bmjpo-2018-000287 on 9 July 2018. Downloaded from http://bmjpaedsopen.bmj.com/ on February 28, 2024 by guest. Protected by copyright.
Orenstein et al looked at treatment-emergent adverse
events (AEs) and serious adverse events (SAEs) including
upper respiratory tract infections, constipation, derma-
titis, ear infections, fever, lower respiratory tract infection,
respiratory tract congestion, rhinorrhoea, candidiasis,
diarrhoea (excluding infective), vomiting, alkaline phos-
phatase increase and others.

Effects of interventions
Sodium alginate (Gaviscon) versus placebo
Thirty-two patients with a median gestational age of
30 weeks were enrolled in Corvaglia et al.12 Participants
were fed eight times over a 24 hours period, with meals
alternatively given with drug (‘drug-given’ (DG)) and
without drug (‘drug-free’ (DF)). Twenty-eight patients
with a median gestational age of 30 weeks were enrolled
in Corvaglia et al.13 Participants were studied between the
hours of 09:00 and 18:00, when they were recorded twice,
for 3 hours each time, after one DG meal and one DF
meal, the order of which was randomly chosen.
There was significant decrease in total GOR episodes
detected only by pH monitoring, acid GOR episodes
detected by multichannel intraluminal impedance moni-
toring, reflux index detected only by pH monitoring and
proximal GORs.12 13
All other outcomes were not significant (liquid GORs,
mixed GORs, non-acid GOR episodes detected by multi-
channel intraluminal impedance monitoring, non-acid
MII-GOR-bolus exposure index and distal GORs).12 No
differences in the number of total apnoea episodes,
central apnoeas, obstructive apnoeas, mixed apnoeas,
desaturations, bradycardia, pathological apnoeas were
Figure 2 Risk of bias summary. found between DG and DF periods (p value was not
significant)13 (table 2).

A total of 302 participants were enrolled in the six Esomeprazole versus placebo
included trials, of which, four studies included only Fifty-two patients with a mean gestational age of 31 were
preterm infants. Omari et al11 included preterm infants enrolled in the study by Davidson et al. One did not have
between 34 and 40 weeks gestational age, Corvaglia et al12 valid efficacy data and was excluded from the reported
and Corvaglia et al13 included ≤33 weeks gestational age results. Participants were randomly selected to receive
and Wheatley and Kennedy10 included those with a gesta- either esomeprazole (n=25) or placebo (n=26), once
tional age of <37 weeks at birth and a corrected gesta- daily, for up to 14 days.15
tional age at enrolment of <44 weeks. Orenstein et al14 No significant results were obtained from this study,
and Davidson et al15 included both preterm infants and which was discontinued prematurely due to poor
full-term infants. The inclusion criteria for each study enrolment15 (table 3).
defined GORD differently (table 1).
Primary outcome: all six studies assessed various reflux Omeprazole versus placebo
symptoms. Four trials reported GOR episodes based Ten preterm infants with a mean postmenstrual age of
on 24-hour pH/impedance monitoring.11–13 15 Three 36.1±0.7 weeks and mean postnatal age of 50±9 days
trials reported bradycardia (Davidson et al,15 Omari et were enrolled in the study by Omari et al.11 Participants
al,11 Wheatley and Kennedy10) and three trials reported were given omeprazole for 7 days and placebo for 7 days
apnoea (Corvaglia et al,13 Davidson et al,15 Omari et al11). in randomised order. At the end of each week of inter-
The other reported outcomes included vomiting, apnoea, ventions, a 24-hour oesophageal and gastric pH moni-
bradycardia, oxygen desaturation, gagging, back arching toring study was performed. Analyses on the basis of pH
and irritability/crying/fussing.10 11 14 15 None of the recordings showed that omeprazole therapy significantly
studies reported on the prespecified secondary outcomes, reduced the oesophageal acid exposure % time pH<4 and
namely: time taken to establish full enteral feeds, length reduced gastric acidity % time pH<411 (table 3). There

Dermyshi E, et al. BMJ Paediatrics Open 2018;2:e000287. doi:10.1136/bmjpo-2018-000287 5

Open access

bmjpo: first published as 10.1136/bmjpo-2018-000287 on 9 July 2018. Downloaded from http://bmjpaedsopen.bmj.com/ on February 28, 2024 by guest. Protected by copyright.
Table 2 Effect of alginates (Gaviscon) use in preterm infants
Studies Antacids Control P values*
12
Corvaglia et al
 Total GORs 49.00 (28.50–67.00) 58.50 (33.50–75.75) 0.024
 Liquid GORs 21.50 (12.25–32.00) 21.50 (13.50–39.75) 0.432
 Gaseous GORs 2.00 (0.25–7.50) 3.00 (0.00–14.75) 0.040
 Mixed GORs 3.00 (2.00–5.75) 3.00 (1.00–5.00) 0.614
 pH-GORs 17.00 (6.00–29.75 29.00 (13.50–44.50) 0.002
 aMII-GORs 4.00 (2.00–8.25) 6.00 (2.25–11.75) 0.050
 NaMII-GORs 19.00 (10.00–32.75) 18.50 (8.50–33.75) 0.743
 RIpH 4.0 (1.8–13.1) 7.6 (3.3–17.0) 0.030
 aMII-GOR-BEI 0.2 (0.1–0.6) 0.4 (0.1–1.0) 0.036
 NaMII-GOR-BEI 1.2 (0.5–1.9) 0.9 (0.5–1.7) 0.822
 Distal GORs (no.) 18.00 (11.25–27.00 15.00 (9.25–26.00) 0.959
 Proximal GORs (no.) 5.50 (4.00–9.00) 7.50 (3.00–12.00) 0.030
13
Corvaglia et al
 Total GOR episodes 9 (0–33) 20.5 (1–42) 0.001
 pH-GOR 2 (0–26) 7.5 (0–23) 0.004
 aMII-GOR 1 (0–5) 3 (0–16) 0.001
 NaMII-GOR 4.5 (0–22) 6 (1–21) 0.145
 RIpH 0.9 (0–23.2) 8.4 (0–44.2) 0.001
 aMII-BEI 0.17 (0–2) 0.5 (0–8.1) 0.002
 NaMII-BEI 0.75 (0–5.7) 1.0 (0.1–9.2) 0.982
 Total apnoea episodes 9.5 (0–35) 9.5 (0–44) 0.99
 Central apnoeas 3.5 (0–25) 5 (0–34) 0.22
 Obstructive apnoeas 1 (0–8) 1 (0–10) 0.10
 Mixed apnoeas 3 (0–16) 4 (0–17) 0.98
 Desaturations 0.5 (0–10) 0 (0–12) 0.41
 Bradycardia 0 (0–3) 0 (0–3) 0.32
 Pathological apnoeas 0 (0–5) 0 (0–7) 0.69
Values are reported as median (IQR).
*P values as provided in the original publication. The level of significance was set at p≤0.05.
aMII-GOR, acid GOR episode detected by MII; aMII-GOR-BEI, acid MII-GOR-bolus exposure index; GOR, gastro-oesophageal reflux; MII,
multichannel intraluminal impedance monitoring; NaMII-GOR, non-acid GOR episode detected by MII; NaMII-GOR-BEI, non-acid MII-GOR-
bolus exposure index; pH-GOR, GOR episodes detected only by pH monitoring; RIpH, reflux index detected only by pH monitoring.

were no significant changes to symptom frequency (in minutes) of episodes averaged across feedings. No
(vomiting, apnoea, bradycardia, choking, behavioural significant results were obtained from this trial. SAEs,
changes) or blood results. particularly lower respiratory tract infections, occurred
more frequently with lansoprazole than with placebo
Lansoprazole versus placebo group (10 vs 2; p=0.032) (table 3).
One hundred sixty-two patients were enrolled in the
study by Orenstein et al, 44 of whom were premature
infants, with a median gestational age at birth of 35 (IQR Metoclopramide and ranitidine versus placebo
25–39) weeks.13 Participants were randomly selected to Eighteen patients were enrolled, and 17 completed the
take either lansoprazole (n=81) or a placebo (n=81) for study, with a gestational age of 29±3 weeks. There was
up to 4 weeks. There was a 35% loss of follow-up for partic- a significant decrease in the number of bradycardia
ipants receiving lansoprazole and 36% for participants episodes per day in the mean combined placebo time
receiving placebo. Lansoprazole and placebo produced periods compared with the mean combined drug time
identical responder numbers (54%). Responder status periods (3.6 (SD 2.7) vs 4.6 (SD 3.1)), p=0.04), and in
was defined as a ≥50% reduction from baseline in either bradycardia episodes over time (p<0.001), with fewer
percentage of feedings with crying episode(s) or duration episodes during placebo periods.10

6 Dermyshi E, et al. BMJ Paediatrics Open 2018;2:e000287. doi:10.1136/bmjpo-2018-000287

 Open access

bmjpo: first published as 10.1136/bmjpo-2018-000287 on 9 July 2018. Downloaded from http://bmjpaedsopen.bmj.com/ on February 28, 2024 by guest. Protected by copyright.
Table 3 Effect of proton pump inhibitors use in preterm infants
Studies Antacids Control
Esomeprazole vs placebo 25 infants 26 infants P values*
15
Davidson et al
 Total number of GORD-related signs and symptoms, percentage of −14.7% 14.1% 0.92
change from baseline after 14 days of treatment
 Gastrointestinal events, percentage of change from baseline −8.39% 10.16% 0.42
 Neurobehavioural events, percentage of change from baseline −3.54% −3.98% 0.94
 Cardiorespiratory events, percentage of change from baseline −38.94% −41.17% 0.89
Omeprazole vs placebo 10 infants 10 infants
11
Omari et al
 Gastric acidity (%time pH<4), mean±SEM 13.9±5.1 53.8±6.8 <0.0005
 Oesophageal acid exposure (%time pH<4), mean±SEM 4.9±3.4 19.0±4.5 <0.01
 No. of acid GOR episodes, mean±SEM 119.4±20 59.6±26.7 <0.05
 No. of oesophageal acid GOR>5 min, mean±SEM 8.0±2.1 3.0±2.0 <0.01
Lansoprazole vs placebo 81 infants 81 infants
14
Orenstein et al
 Primary efficacy: responder rate, n (%) 44 (54%) 44 (54%) NS
 AEs, n (%) 50 (62%) 37 (46%) NS
 SAEs, n (%) 10 (12%) 2 (2%) 0. 032
*P values as provided in the original publication.
AEs, adverse events; GOR, gastro-oesophageal reflux; GORD, gastro-oesophageal reflux disease; NS, not significant; SAEs, serious adverse
events.

Discussion natural improvement of the outcome overtime, due to

This systematic review reveals that there is insufficient the infants’ growth.10
evidence to support the efficacy and safety of antacid
therapies in preterm infants. Four out of the six studies Alginates
included in the review were crossover trials, where the Sodium alginate significantly reduced acid GOR episodes,
patient receives both interventions at different time inter- although had no effect on the reduction of apnoeas.
vals. The carry-over effect, where the intervention taken Corvaglia et al12 reports that participants were observed
in the first period is still effective when the second inter- over a 24-hour period, and data were collected after eight
vention is being taken, is a major limitation in crossover meals. There may have been selective reporting in the
designs. These effects cannot be estimated separately. To study by Corvaglia et al,13 with authors only reporting data
from a 6-hour period of observation instead of the full
minimise the risk of a carry-over effect, it can be effec-
24 hours data, thus presenting more significant results.
tive to allow a ‘washout’ period between interventions.
This discrepancy diminishes the validity of the reports
Wheatley and Kennedy allowed a 24-hour washout period
and applicability to clinical practice.
at the beginning of the second and third time periods.10
Omari et al,11 Corvaglia et al12 and Corvaglia et al13 did Proton pump inhibitors
not appear to have a washout period. When deciding Omari et al showed that 0.7 mg/kg omeprazole given
whether to use a crossover design, it is important to once daily was effective in reducing the frequency of
consider whether the outcome that is being treated will acid reflux episodes and the overall degree of oesoph-
change naturally over time. This does not seem like it ageal acid exposure in premature infants.11 Despite the
would be an issue in the studies by Corvaglia et al12 and normalisation of acid reflux in most patients, the number
Corvaglia et al,13 due to the overall treatment and obser- of symptomatic events was not significantly changed. The
vation period being short at just 24 hours. It may be prob- drug-dosing regimen used appeared safe based on AE
lematic, however, for the study by Omari et al,11 with each reporting and blood screening. However, due to the small
infant receiving a week of each intervention. Wheatley number of participants enrolled in the study (n=10), it
and Kennedy studied infants over a 2-week period and would be difficult to state whether this evidence is appli-
splits one intervention so it is given at the start and end of cable in everyday practice and more trials must be carried
the 2-week period. This study shows a significant decrease out into the efficacy of omeprazole. There were no signif-
in bradycardias over time, which may be evidential of a icant differences in the number of GORD-related signs

Dermyshi E, et al. BMJ Paediatrics Open 2018;2:e000287. doi:10.1136/bmjpo-2018-000287 7

Open access

and symptoms between neonates receiving esomeprazole gestation were to be included in this review; however,

bmjpo: first published as 10.1136/bmjpo-2018-000287 on 9 July 2018. Downloaded from http://bmjpaedsopen.bmj.com/ on February 28, 2024 by guest. Protected by copyright.
or lansoprazole versus placebo.14 15 SAEs occurred more both Davidson et al and Orenstein et al included data
frequently with lansoprazole than with placebo group. It for full-term infants as well as preterm, some of whom
is unclear whether loss to follow-up caused a significant were >37 weeks gestation. The authors were contacted
imbalance in characteristics between lansoprazole and to obtain exclusively preterm data, however, replies were
placebo group. Therefore, applicability into everyday not received. We included these studies in this review due
practice is low because loss to follow-up can severely to the high percentage of preterm infants enrolled in the
compromise validity as those lost to follow-up could trials. The methods stated that the only interventions that
have a different prognosis than those who complete the were to be considered were H2 RAs, PPIs and alginates;
study. The number of AEs was similar between neonates however, Wheatley and Kennedy assessed the combined
receiving esomeprazole versus placebo. Both Orenstein et effects of both metoclopramide (dopamine receptor
al and Davidson et al had notable conflicts of interest that antagonist) and ranitidine (H2 RA).10 We still decided
were reported in the study, as shown in the online supple- to report this outcome as the inclusion of the H2 RA as
mentary appendix 3. The trials of both studies were spon- it is of interest to the reader in general who must bear in
sored by drug companies, which may have affected the mind this was a combined intervention. Studies included
design and outcomes of the trial as well as the reporting in the review were heterogeneous in terms of design,
of results. However, no significant results were found in study characteristics such as age of participants and inter-
either trial, and so no results were reported in favour of ventions considered for the treatment of GORD. Studies
the drug under trial.14 15 also had small sample sizes. This limits the conclusions
that can be drawn from this review; however, it highlights
Histamine-2 -receptor antagonists the gaps in the evidence.
A retrospective cohort study conducted by Romaine et
al16 in the USA concluded that H2 blocker use was asso- Agreements or disagreements with other studies or reviews
ciated with increased risk of the combined outcome of To our knowledge, this review is the first to look into the
death, NEC or sepsis in hospitalised very low birthweight effects of antacids in preterm infants.
infants (VLBW). Another recent retrospective cohort Terrin et al in a retrospective study of 274 VLBW infants
study showed that ranitidine use was associated with an reported that the risk of NEC, nosocomial infection and
increased risk of infections and mortality in preterm mortality were significantly higher in the infants exposed
infants, but not with NEC.17 Wheatley and Kennedy to ranitidine.4 However, non-prospective, non-controlled
showed that ranitidine did not reduce, and may have and unblinded design features limited its significance.
increased, bradycardia episodes in preterm infants with A Cochrane review by Tighe et al looking at the effects
bradycardia attributed to GOR.10 Wheatley and Kennedy of pharmacological treatment for the management of
compared the combination of two interventions together GORD in children concluded that although there is
against a placebo, ranitidine, a H2 RA and metoclopr- evidence to support pharmacological use in older chil-
amide, a dopamine receptor antagonist. With regard to dren, use in infants is unsupported due to lack of robust
applicability, the data derived from this study actually RCT evidence.21
suggest that combining ranitidine and metoclopramide Cohen et al in a recent review suggested that the use of
may be detrimental to patients and should therefore be GORD medications should only be used after non-phar-
avoided in clinical practice, as it showed a significant macological measures have been taken with incomplete
increase in bradycardia episodes during drug periods. success as acid suppression may place immune-deficient
This may be caused by significant interactions between infants and children at risk for the development of lower
the two drugs that could either decrease the efficacy of respiratory tract infections and nosocomial sepsis.22
either or both of drugs or perhaps cause AEs. Leucuta et
al found pharmacokinetic changes, such as an increased
half-life, in metoclopramide, when taken with raniti-
dine.18 It is also quite likely that this is a chance finding, Authors' conclusions
given the small number of participants enrolled in the There is insufficient evidence on the efficacy and safety
study (n=18). Previous studies into the combined efficacy of antacids in preterm infants.
of ranitidine and metoclopramide suggest that this treat- The lack of research in this area of medicine is a
ment is effective at increasing gastric pH and reducing problem that must be addressed in this population of
the side effects of GORD, and do not mention any signif- patients. Adequately powered, RCTs in preterm infants
icant drug induced side effects or drug interactions.19 20 are needed to determine the safety and efficacy of these
commonly used medications.

Limitations Contributors ED had full access to all study data and takes responsibility for the
integrity and accuracy of the data. Study concept and design: JD and BD conceived
A number of limitations are worth noting, not all studies and designed the study. Acquisition of data: CM and ED. Analysis and interpretation
met the inclusion criteria outlined in the methods. of data: ED, CM, PK, BD and JD. Drafting of the manuscript: ED, PK and JD. Critical
We initially stated that only preterm infants<37 weeks revision of the manuscript for important intellectual content: JD. All authors

8 Dermyshi E, et al. BMJ Paediatrics Open 2018;2:e000287. doi:10.1136/bmjpo-2018-000287

 Open access

approved the final manuscript as submitted and agree to be accountable for all 9. Orenstein SR, Cohn JF, Shalaby TM, et al. Reliability and validity

bmjpo: first published as 10.1136/bmjpo-2018-000287 on 9 July 2018. Downloaded from http://bmjpaedsopen.bmj.com/ on February 28, 2024 by guest. Protected by copyright.
aspects of the work. of an infant gastroesophageal reflux questionnaire. Clin Pediatr
1993;32:472–84.
Funding This research received no specific grant from any funding agency in 10. Wheatley E, Kennedy KA. Cross-over trial of treatment for
thepublic, commercial or not-for-profit sectors. bradycardia attributed to gastroesophageal reflux in preterm infants.
J Pediatr 2009;155:516–21.
Competing interests None declared.
11. Omari TI, Haslam RR, Lundborg P, et al. Effect of omeprazole
Patient consent Not required. on acid gastroesophageal reflux and gastric acidity in preterm
infants with pathological acid reflux. J Pediatr Gastroenterol Nutr
Provenance and peer review Not commissioned; externally peer reviewed. 2007;44:41–4.
Open access This is an open access article distributed in accordance with the 12. Corvaglia L, Aceti A, Mariani E, et al. The efficacy of sodium alginate
(Gaviscon) for the treatment of gastro-oesophageal reflux in preterm
Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which
infants. Aliment Pharmacol Ther 2011a;33:466–70.
permits others to distribute, remix, adapt, build upon this work non-commercially, 13. Corvaglia L, Spizzichino M, Zama D, et al. Sodium Alginate
and license their derivative works on different terms, provided the original work is (Gaviscon®) does not reduce apnoeas related to gastro-
properly cited, appropriate credit is given, any changes made indicated, and the oesophageal reflux in preterm infants. Early Hum Dev
use is non-commercial. See:©http://​creativecommons.​org/​licenses/​by-​nc/​4.​0/. 2011b;87:775–8.
14. 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
References with symptoms of gastroesophageal reflux disease. J Pediatr
1. Corvaglia L, Monari C, Martini S, et al. Pharmacological therapy of 2009;154:514–20.
gastroesophageal reflux in preterm infants. Gastroenterol Res Pract 15. Davidson G, Wenzl TG, Thomson M, et al. Efficacy and safety of
2013;2013:1–12. once-daily esomeprazole for the treatment of gastroesophageal
2 Rossor T, Andradi G, Bhat R, et al. Investigation and management reflux disease in neonatal patients. J Pediatr 2013;163:692–8.
of gastro-oesophageal reflux in United Kingdom neonatal intensive 16. Romaine A, Ye D, Ao Z, et al. Safety of histamine-2 receptor blockers
care units. Acta Paediatr 2018;107:48–51. in hospitalized VLBW infants. Early Hum Dev 2016;99:27–30.
3. Guillet R, Stoll BJ, Cotten CM, et al. Association of H2-blocker 17. Santana RNS, Santos VS, Ribeiro-Júnior RF, et al. Use of ranitidine
therapy and higher incidence of necrotizing enterocolitis in very low is associated with infections in newborns hospitalized in a neonatal
birth weight infants. Pediatrics 2006;117:e137–e142. intensive care unit: a cohort study. BMC Infect Dis 2017;17:375.
4. Terrin G, Passariello A, De Curtis M, et al. Ranitidine is associated 18. Leucuţa A, Vlase L, Farcău D, et al. Pharmacokinetic interaction
with infections, necrotizing enterocolitis, and fatal outcome in study between ranitidine and metoclopramide. Rom J Gastroenterol
newborns. Pediatrics 2012;129:e40–e45. 2004;13:211–4.
5. Liberati A, Altman DG, Tetzlaff J, et al. The PRISMA statement for 19. Hong JY. Effects of metoclopramide and ranitidine on preoperative
reporting systematic reviews and meta-analyses of studies that gastric contents in day-case surgery. Yonsei Med J 2006;47:315–8.
evaluate health care interventions: explanation and elaboration. J 20. Bala I, Prasad K, Bhukal I, et al. Effect of preoperative
Clin Epidemiol 2009;62:e1–e34. oral erythromycin, erythromycin-ranitidine, and ranitidine-
6. Higgins J, Green S. Cochrane Handbook for Systematic Reviews of metoclopramide on gastric fluid pH and volume. J Clin Anesth
Interventions, version 5.1.0 (updated March 2011). The Cochrane 2008;20:30–4.
Collaboration. 2011. www.​cochrane-​handbook.​org 21. Tighe M, Afzal NA, Bevan A, et al. Pharmacological treatment of
7. Higgins JP, Altman DG, Gøtzsche PC, et al. The Cochrane children with gastrooesophageal reflux (Review). Cochrane Database
Collaboration's tool for assessing risk of bias in randomised trials. Syst Rev 2014;11:CD008550.
BMJ 2011;343:d5928. 22. Cohen S, Bueno de Mesquita M, Mimouni FB, et al. Adverse effects
8. Guyatt G, Oxman AD, Akl EA, et al. GRADE guidelines: 1. reported in the use of gastroesophageal reflux disease treatments
Introduction-GRADE evidence profiles and summary of findings in children: a 10 years literature review. Br J Clin Pharmacol
tables. J Clin Epidemiol 2011;64:383–94. 2015;80:200–8.

Dermyshi E, et al. BMJ Paediatrics Open 2018;2:e000287. doi:10.1136/bmjpo-2018-000287 9