Article Text
Abstract
Recent guidelines focus on a non-interventionist approach to management of gastro-oesophageal reflux in infancy and emphasise the importance of explanation, reassurance and simple measures such as attention to feeding. Relying on clinical history alone leads to over diagnosis of disease, and widely used medications are often ineffective for symptom relief and carry significant risk of harm. The association between vomiting in infancy and other problems such as crying and poor feeding should not be interpreted as implying causality. When there are strong pointers to underlying gastro-oesophageal reflux disease, invasive investigations are required in order to formulate appropriate intervention.
- Gastroenterology
- General Paediatrics
Statistics from Altmetric.com
Introduction
Gastro-oesophageal reflux (GOR) in young babies is a physiological phenomenon that in the vast majority can be expected to resolve spontaneously. Management involves repeated explanation and reassurance, together with simple advice regarding feeding and positioning. Gastro-oesophageal reflux disease (GORD) occurs when this normal event results in the occurrence of symptoms or complications.1 GOR can result simply from overfeeding, a primary disorder of function of the upper gastrointestinal tract, or a systemic disorder affecting motility such as neurological impairment in cerebral palsy. While uncomplicated reflux does not require treatment, the key issue for clinicians is to differentiate between simple, ‘physiological’ GOR and pathological GORD once less common causes of vomiting have been discounted (box 1).
Some possible causes of persistent vomiting in an infant, and the signs and symptoms suggestive of underlying pathology
Gastrointestinal obstruction from:
pyloric stenosis—frequent forceful (projectile) vomiting; needs surgical referral
malrotation with intermittent volvulus—yellow or green vomit; needs immediate referral to surgical unit
Hirschsprung's disease
Other gastrointestinal disorders:
gastro-oesophageal reflux (consider overfeeding)
gastro-oesophageal reflux disease
eosinophilic oesophagitis/gastritis
gastroparesis/pseudo-obstruction
constipation
Neurological disease:
hydrocephalus—raised intracranial pressure, bulging fontanelle
subdural haematoma
space-occupying lesion
Infection: often systemically unwell, with fever
gastroenteritis
septicaemia
meningitis
urinary tract infection
pneumonia
otitis media
hepatitis
Metabolic/endocrine disease:
congenital adrenal hyperplasia
urea cycle defects
amino and organic acidaemias
Renal disease:
obstructive uropathy
renal failure
Cardiac disease:
congestive heart failure
Fabricated or induced illness
Features indicative of pathology:
weight faltering
aspiration pneumonia
projectile vomiting
bilious (yellow or green) vomit
haematemesis
blood in stools
diarrhoea
constipation
fever
abdominal distension/mass/tenderness
hepatosplenomegaly
developmental delay
lethargy
increased head circumference
hypertonia/hypotonia
seizures
poor feeding/pain on feeding
The latter can often be identified by a careful history and examination including growth assessment, followed by targeted investigations. Growth faltering is always a red flag sign indicating likely pathology and the need for investigation. Due to the association between physiological GOR in early life and common behavioural problems such as feeding difficulties, crying and fussing, a diagnosis of ‘GORD’ is often made through causality being wrongly inferred. This in turn leads to both unnecessary treatment and iatrogenic disease (box 2).
Medications used for gastro-oesophageal reflux disease: actions and potential side effects2
Alginates (eg, Gaviscon): Reduce the symptoms of indigestion, oesophagitis and vomiting, but little effect on reflux overall. Use in conjunction with thickening agents or antiregurgitation milks poses risk of obstruction from agglutinated feed.
Cisapride: Increases motility in the upper gastrointestinal tract acting directly as a serotonin 5-HT4 receptor agonist, increasing acetylcholine release in the enteric nervous system. It was withdrawn from the market in 2000 because of concerns regarding long QT syndrome and risk of fatal arrhythmia. In one survey of 46 hospitals, >19% of preterm infants cared for in the neonatal intensive care unit had been treated with this agent,3 sometimes referred to as ‘vitamin C’.4 A Cochrane review concluded that there was no evidence that cisapride was effective for treating gastro-oesophageal reflux disease (GORD) in children.5
Domperidone: A peripheral D2 receptor antagonist that increases motility and gastric emptying. Widely prescribed since the unavailability of cisapride, there is little clinical trial evidence of efficacy in GORD. Potential serious cardiac side effects mean it should not be used for long-term treatment.6
Erythromycin: A macrolide antibiotic that increases gastrointestinal motility by acting directly on gut motilin receptors. There is no evidence of efficacy in GORD7; potential side effects include nausea and vomiting, hepatic damage, anaphylaxis, arrhythmias and pyloric stenosis.8
Metoclopramide: A dopamine antagonist that stimulates motility, accelerates gastric emptying and increases lower oesophageal sphincter tone. Ineffective, its use was complicated by adverse effects including extrapyramidal reactions such as dystonia and tardive dyskinesia.
Ranitidine: Inhibits H2 receptors of gastric parietal cells and is effective at suppressing acid secretion. Side effects include abdominal pain, nausea vomiting and diarrhoea; tolerance develops quickly. To prevent rebound hypersecretion of acid, stepwise withdrawal is advised; acid suppression by raising stomach pH will negate the effect of antiregurgitation milk (both also apply to proton pump inhibitors (PPI)).
Lansoprazole and omeprazole: PPI that inactivate the H+/K+ -ATPase pump in parietal cells preventing gastric acid secretion. Side effects include headache, nausea, vomiting, abdominal pain, diarrhoea or constipation. Gastric polyps,9 increased risk of community acquired pneumonia10 and vitamin B12 deficiency11 have also been reported. No evidence of additional benefit if combined with an H2 receptor inhibitor. PPI given via feeding tubes may cause blockage; although omeprazole is available as an unlicensed solution, it is extremely expensive.
Guidelines and definitions
To assist clinicians in the diagnosis and management of GOR/GORD, a number of successive evidence-based guidelines have been published, both in full12–14 and in summary form,15 ,16 most recently from the National Institute for Health and Care Excellence.17 These commonly start with definitions:
GOR: The passage of gastric contents into the oesophagus, with or without regurgitation/vomiting; it is considered physiological in infants when symptoms are absent or not troublesome,
GORD: The presence of troublesome symptoms (ie, those that adversely affect the well-being of the child) and/or complications (eg, tissue damage or inflammation, as in oesophagitis, reactive airways disease or pulmonary aspiration).
Regurgitation (posseting, or spitting up): The effortless return of previously swallowed food or secretions into or out of the mouth. In children <1 year of age, regurgitation may be regarded as normal.
Vomiting: The ejection of ingested stomach contents through the mouth.
While important, these definitions immediately raise difficulties such as pinning down ‘troublesome’ symptoms, the key discriminator between GOR and GORD. Additionally, a ‘symptom’ is defined as “a departure from normal function indicating the presence of disease or abnormality”,18 such that health professionals will be more inclined to interpret vomiting as a sign of disease rather than as a physiological variant. Also, vomiting might not be troubling to the child, but deeply troublesome to parents because other aspects of behaviour such as crying are misinterpreted as being directly related. Indeed, doctors too commonly fall into this same trap, for example, attributing Apparent Life-Threatening Events to reflux seen on a contrast swallow.19–21 While GOR or GORD and feed refusal frequently coexist, a relation between them has not been demonstrated.13 In some situations, parental factors themselves (maternal psychopathology involving fear of inadequate feed intake) may lead to the development of GORD (ie, the child with GOR who then develops feeding problems).22 ,23 To establish a causal relationship between reflux and symptoms/tissue damage requires invasive investigations (all of which have their limitations),13 as well as long-term follow-up, to evaluate the effectiveness or otherwise of interventions. Rather than entities that can be neatly separated, GOR and GORD are best seen as a part of a continuum of disease.13
Have GOR guidelines proved helpful?
Fifty-page-long evidence-based guidelines13 are useful for background reading but not for quick reference at the point of care.24 Since the objective of a clinical guideline is to improve patient management, the ideal measure of a good guideline is improvement in clinical outcome. This is difficult to assess and the practical measure of a good guideline may have to be: it has evidence-based recommendations, has been tested among the target group for feasibility and acceptance, and is used in practice.24 A good guideline will encourage consistent practice and is amenable to audit and critique, leading to a cycle of continuous improvement.24 Grol et al25 investigated which attributes of clinical practice guidelines influenced their use in a general practice setting. Recommendations that were non-controversial, clear, did not demand a change in existing practice routine and were based on research evidence were more likely to be followed.
Although considered ‘excellent’ by some,23 a recent study evaluated whether the 2009 North American Society for Paediatric Gastroenterology, Hepatology, and Nutrition (NASPGHAN)–European Society for Pediatric Gastroenterology, Hepatology, and Nutrition (ESPGHAN) clinical practice guideline for GOR13 was in fact being implemented by clinicians.26 A structured questionnaire including a series of case scenarios was sent to a random sample of general paediatricians across 11 European countries. From 567 replies (a return of 42%), <2% were adhering to the guideline in all respects. Common departures from the guideline included 46% diagnosing GORD from clinical symptoms alone, 39% prescribing proton pump inhibitors (PPI) in infants simply because of unexplained crying and/or distressed behaviour, and 36% prescribing PPI in infants with GOR (ie, uncomplicated recurrent regurgitation and vomiting). The overall rate of paediatricians over prescribing PPI was 82%. These findings published in 2014 can be interpreted as evidence that paediatricians are unaware of, or in disagreement with, the 2009 guideline13 and were greatly overdiagnosing GORD, commonly prescribing medication despite the lack of efficacy for the symptoms being treated.
McCracken27 reflected upon a number of possible explanations for non-implementation of guidelines: they may not have been well publicised, have changed radically in a short period from an earlier guideline and are having to compete against other guidelines for attention. She highlighted the significant changes in 200913 compared with the 2001 NASPGHAN guideline,12 published in the same specialist gastroenterology journal. The former stated “a trial of time-limited medical therapy for GOR is useful for determining if GOR is causing a specific symptom”; for the infant with recurrent crying and irritability, “… expert opinion suggest two diagnostic and treatment strategies. Empiric treatment with either a sequential or simultaneous two-week trial of a hypoallergenic formula and acid suppression…” In contrast, the 2009 recommendations opined: “there is no evidence to support an empiric trial of acid suppression as a diagnostic test in infants and young children where symptoms of GORD are less specific”, and in “infants with unexplained crying and/or distressed behaviour. … there is no evidence to support empiric use of acid suppression; … a time-limited trial of anti-secretory therapy may be considered, but there is potential risk of adverse effects”. So instead of a simple treatment trial in 2001 the 2009 guidance advised the need for invasive testing before the use of PPI in infants. While it is considered that cow milk protein allergy manifests in some infants in a way indistinguishable from GORD, evidence from the literature supporting a trial of extensively hydrolysed protein formula or an amino acid-based feed is acknowledged to be limited and based on retrospective cohort or case–control studies.13 There are significant cost implications here since even when a trial of hypoallergenic formula has little effect parents may be motivated to continue a milk provided on prescription.
Many infants are not seriously ill but have complaints that are compatible with GORD; 24 h pH or impedance studies and endoscopy are largely the domain of paediatric gastroenterologists. When in addition to minor symptoms seeming not to warrant specialty referral there are also long waiting times, and parents do not want invasive tests, it is likely that the guideline may be seen as impractical. However, an appropriate standpoint may well be that of ‘watchful waiting’ before either investigation, empirical treatment or referral. The principle of primum non nocere should prevent or delay trials of medical therapy with drugs that commonly cause unwanted side effects and occasionally pose serious risk (box 2). This should also be the standpoint in secondary care, where there may be more inclination to opt for empirical treatment or investigations in order to be seen to be ‘doing something’.
The study by Quitadamo et al26 did not explore the reasons for failure to comply with the guideline recommendations13 and therefore could not suggest specific interventions to improve matters. It seemed likely to one commentator28 that traditional practice patterns and national healthcare systems, including defensive medicine related to threat of litigation, would influence decisions regarding treatment and investigation. To this list one might add the promotion to community and hospital staff of products by drug manufacturers and makers of hypoallergenic and antiregurgitation infant formula. Additional studies in other parts of the world, it was suggested, would throw light on these matters.28 In fact, previous studies already indicate that there might be similar findings outside Europe. In 2002, a questionnaire was sent to 6000 randomly selected members of the American Academy of Pediatrics.29 Of the 1245 who responded, 54% started to test for GORD inappropriately early (in newborns) and 19% wrongly believed acid suppression was best achieved by H2 blockers. If acid suppression was indicated, only 36% followed guideline recommendations for therapy duration and 52% for dosing. These findings lead the authors to conclude that paediatric care providers did not let limited knowledge of evidence-based GORD management stand in their way of ordering diagnostic tests or prescribing treatment. The nub of the matter appears to be that when presented with a vomiting infant by anxious parents doctors feel more inclined to intervene than give explanation, reassurance and simple advice on feeding. To change this situation will require an understanding of what is normal, an appreciation of the limitations of therapy and wide dissemination of clear and concise guidance.17
The natural history of GOR
Information on the natural history of GOR is important so that appropriate explanation and reassurance can be given both to parents and professionals. In a cross-sectional survey in primary care in the USA, almost a thousand parents of healthy infants up to 13 months of age were given a questionnaire about symptoms of GOR.30 Regurgitation at least once a day was reported in half of infants up to 3 months old but in only 5% of children 10–12 months of age. The proportion of children with regurgitation four or more times a day was 23% at 5 months, but only 7% at 7 months of age. A quarter of parents regarded regurgitation as a problem at 6 months of age, but by 7 months this decreased to 14%. Parental perception of regurgitation as a problem was associated with the frequency and volume of regurgitation, increased crying or fussing, reported discomfort with spitting up and frequent back arching. Reported treatment included a change in formula in 8.1%, thickened feedings in 2.2%, termination of breast feeding in 1.1% and medication in 0.2%.
Some of these infants (62 cases who had been identified as ‘regurgitators’ and 92 controls who were not) were then followed up in a further study to explore any adverse outcomes after a year31: no parent of children within either group considered regurgitation still to be a problem. Only one control and no case subject was reported to be regurgitating one or more times a day. Parents of infants who had troublesome regurgitation were more likely to report feeding problems ("my child takes more than an hour to eat meals": 8% vs 0%; "I get upset when I think about our meals": 14% vs 4%). However, feed refusal was reported in 9% of controls, highlighting that feeding problems in young children are common (and no doubt multifactorial in origin). Infants with daily or problematic regurgitation at 6–12 months outgrew this within the following year, but were more likely to be reported to have feeding problems. What these two studies tell us is that GOR is very common in well infants, peaking at 4 months, with most better at 7 months of age. Parents perceived vomiting to be a problem much more frequently than treatment was given. Although the data reported cannot tell us whether later feeding problems could have been prevented by treatment interventions, nor what the risks of undertreating or overtreating may be, it is clear that even when symptoms were troublesome to parents many infants did well without treatment.
Empirical treatment with acid suppression
Changes in guideline recommendations relate to increased knowledge about the risks and benefits of empirical treatment. Orenstein et al10 conducted a double-blind, randomised, placebo-controlled trial of acid suppression using a PPI (lansoprazole) given to infants aged 1–12 months with symptoms that are commonly (mis)interpreted as indicative of GORD. Subjects had to have symptoms that persisted despite 1 week of non-pharmacological therapy (ie, advice to reduce tobacco smoke exposure; undertake both frequent winding and give small feeds; use hypoallergenic or thickened formula or dairy avoidance by a breastfeeding mother; minimise seated or awake supine positioning; avoid vigorous handling after feeds). Eighty-one children were entered into each limb of the study, their symptoms including regurgitation (100%), crying (100%), stopping feeding during a feed (80%), refusing feed (64%), back arching (89%), coughing (74%), wheezing (42%) and hoarseness (25%). In all, 54% of subjects in each group responded; significantly more lower respiratory tract infections were noted in the lansoprazole group. This study strongly suggests that symptoms commonly interpreted as signifying GORD were no more likely to improve with acid suppression than with placebo and that acid reflux was therefore not the underlying cause. In fact, many of these infants probably had physiological reflux and other reasons for crying. A recent systematic review has added further weight to these findings.32
Crying
Crying mediates nurturing and safety, which in turn facilitates attachment.33 Community studies show that around a fifth of parents consider their otherwise healthy baby to have crying/fussing problems at 2 months of age. In the first six weeks of life, median cry–fuss time is 2 h (with considerable inter-individual variation, up to 6 h), reducing to 72 min by 10–12 weeks;34 5% continue to be unsettled at 5 months. There is no evidence that GOR underlies this problem and there is no basis from available evidence to support an empirical trial of acid suppression in infants with unexplained crying, irritability or sleep disturbance.10 ,13 ,32
Conclusions
GOR is common in young infants and is often associated with problems such as feed refusal, crying and back arching while not being the cause of these symptoms. Guidelines now emphasise the importance of expectant management and reassurance rather than empirical trials of treatment. Medical professionals need to be aware of the natural history of physiological reflux and of crying in young infants, and avoid treatment with drugs that pose a risk of significant side effects. When there are strong grounds to suspect underlying pathology, investigations including contrast study, endoscopy and 24 h pH and/or impedance are warranted to establish the relationship between acid reflux and symptoms, and the presence of tissue damage. Existing guidelines provide a detailed overview of the available literature and evidence-based advice regarding management.17
References
Footnotes
Competing interests None.
Provenance and peer review Commissioned; internally peer reviewed.