Dr Kate Armon, Prof. Terence Stephenson, Dr Ursula Werneke, Miss Phillippa Eccleston and Dr Roderick MacFaul

Dr Kate Armon. BmedSci BMBS MRCP MRCPCH DCH DRCOG

Research Fellow, Academic Division of Child Health,

School of Human Development,

University of Nottingham, NG7 2UH.

Prof. Terence Stephenson, BSc BM DM BCh FRCP FRCPCH,

Professor of Child Health and Honorary Consultant Paediatrician,

Academic Division of Child Health

School of Human Development,

University of Nottingham

Dr Ursula Werneke, MD MSc MRCPsych

Specialist Registrar

Maudsley Hospital

London SE5 8AZ

Miss Philippa Eccleston, BSc (hons) RGN RSCN .

Nurse Researcher,

Academic Division of Child Health,

School of Human Development,

University of Nottingham.

Dr Roderick MacFaul MB ChB FRCP FRCPCH DCH

Consultant Paediatrician

Pinderfields General Hospital

Aberford Road

Wakefield

Dr kate Armon

Conflict of interest: none

Funding: Children Nationwide Medical Research Fund

Acknowledgements

Ackland F (Paed. Cons.), Arrowsmith W A (Paed. Cons.), Bailey R (A&E Cons), Barker R (Paed. Nurse), Bennett Britton S (Paed. Cons.), Boon A W (Paed. Cons.), Boyle R (Paed. SpR.), Cade A (Paed. SpR.), Carter E (Paed. Cons.), Charlton C P J (Paed. Cons. (gastro)), Cutting W A M (Paed. Cons.), Cutts J (Paed. Nurse), Devane S (Paed. Cons.), Edge J (Paed. Cons.), Ehrhardt P (Paed. Cons.), Gleeson E (A&E SpR), Green C (Paed. Cons.), Hewertson J (Paed. Cons.), Hodges S(Paed. Cons.), Huynh H (Paed. SpR.), Jefferson I (Paed. Cons.), Jenkins H (Paed. Cons. (gastro)), Kershaw C (Paed. Cons.), Laurent S (Paed. Cons.), Lewis H M (Paed. Cons.), Marcovitch H (Paed. Cons.), McGovern M C (Paed. SpR.), McGraw M E (Paed. Cons.), McLain B (Paed. Cons.), Mirfattahi M M B (Paed. Cons.), Puntis J (Paed. Cons. (gastro)), Rutter N (Paed. Cons.), Sajjanhar T (Paed. Cons. (A&E)), Smith R (Paed. Cons.), Smith S (Paed. Cons. (A&E)), Stephens S (Paed. SpR.), Sullivan C (Paed. Cons.), Thomas S (Paed. SpR.), Wells L (Paed. SpR.).

Key: Paed. = Paediatric, Cons. = Consultant, SpR = Specialist Registrar, A&E = Accident and Emergency, Gastro = Gastroenterology

• Children Nationwide for their generous funding of this research

• Jeanette Taylor-Meek for effective administration of the Delphi process

Diarrhoea, gastroenteritis, Delphi consensus, guideline

Table 1: Broad differential diagnosis of the child presenting with acute diarrhoea (+/- vomiting). The latter diagnoses are more likely to present chronically.

NB. The following features may be indicative of diagnoses other than acute viral gastroenteritis:

• Abdominal pain with tenderness/guarding and/or bilious vomiting (?surgical)

• Pallor, jaundice, oligoanuria, bloody stool (?HUS)

• Systemically unwell, out of proportion to the level of dehydration (other infections, surgical, CAH etc)
• Shock

Table 2: Assessment of severity of dehydration (if in doubt err by over-estimating % dehydration).

• Signs are ordered in each column by severity.
• If a pre-illness accurate weight is available calculate deficit from weight loss.
• Pinch test – Pinch skin of abdomen. Skin recoils instantly = normal, 1-2 sec = mild/moderate, >2sec = severe.

Table 3: Calculation of Oral Rehydration Solution (ORS) requirements in dehydration.

• Children who are dehydrated are thirsty and do not normally refuse ORS.
• Give fluid little and often. If the child is vomiting decrease volumes and increase frequency (every 5-10 minutes).
• Where carers are not willing/able to do this under supervision (or child is asleep) then rehydrate by NGT.
• Suitable ORS are Dioralyte, Diocalm Junior or Electrolade.

Table 4: Calculation of maintenance requirements.

E.g. A 22Kg child has maintenance requirements of: (10x100)+(10x50)+(2x20)=1,540mls/24hours

These requirements should be supplemented if the child has frequent or substantial watery stools or vomits.

Table 5: When to send a stool to the lab for microscopy, culture, sensitivity and virology.

• A history of blood +/- mucous in the stool
• Systemically unwell, severe or prolonged diarrhoea
• If the child is admitted to hospital (local policy)
• A history suggestive of food poisoning

• Recent travel abroad

Table 6: Management of feeding during gastroenteritis.

Table 7: Guide to drug treatment.

Abstract

Structured in the format recommended by Hayward et al for guideline reports.

Objective: An evidence and consensus based guideline for the management of the child who presents to hospital with diarrhoea (+/- vomiting). The guideline was developed for this common problem (18% of all paediatric medical attenders) where variation in practice occurs.

Options: Assessment, investigations (biochemistry and stool culture in particular), admission and treatment are addressed. The guideline aims to aid junior doctors in recognising children who need admission for observation and treatment and those who may safely go home.

Evidence: A comprehensive review of English language literature using the electronic databases Medline, Embase and Cochrane from 1966 to June 1998 was performed. Articles were selected if they addressed the specific clinical question and personal reviews were excluded. The literature was appraised, graded, and synthesised qualitatively. Statements of recommendation were made.

Consensus: An anonymous, postal Delphi consensus development was used. A National panel of 39 medical and nursing staff who frequently care for these children were asked to grade their agreement with the statements generated. They were sent the papers, appraisals, and literature review. On the second and third rounds they were asked to re-grade their agreement in the light of other panellists’ responses. Consensus was defined as 83% of panellists agreeing with the statement.

Recommendations: in brief. Clinical signs useful in assessment of level of dehydration are given. Admit to a paediatric facility mild-moderate and severely dehydrated children. For those with mild-moderate dehydration, estimate the deficit and replace over 4 hours with Oral Rehydration Solution (ORS, glucose based, 200-250mOsm/L) little and often. Use a naso-gastric tube where necessary. Recommence normal feeds following rehydration. Observe children at high risk of dehydration (<6 months, >8 stools/day, >4 vomits/day) and ensure at least maintenance fluid is given. Antidiarrhoeal medication is not indicated.

Validation: The guideline was piloted in the form of a care pathway in a paediatric accident and emergency department. The guideline reduced inappropriate investigations and intravenous infusions. The proportions of children admitted increased in keeping with the overall increases in medical admissions.

This policy is for the child presenting to an acute facility (accident and emergency or admissions / paediatric assessment unit) with acute diarrhoea (<7 days) with or without vomiting. Children presenting with vomiting alone or chronic diarrhoea (>7 days) are not considered.

The development group assumes that health care professionals will use general medical knowledge and clinical judgement in applying the recommendations in this document to the management of individual patients. These recommendations may not be appropriate for use in all circumstances.

Diarrhoea is defined as a change in bowel habit for the individual child resulting in substantially more frequent and/or looser stools. The UK incidence of diarrhoeal illness in children is not known, but it leads to high GP consultation rates (204/1000/year in the 0-4 age group) and hospital admissions (at least 7/1000/year <5 years old). It accounts for 16% of all paediatric medical presentations to A&E. In 1996 OPCS data for England and Wales states that there were 58 deaths from intestinal infections in children 0-15 years, accounting for 0.9% of all causes of death in this age group. This guideline was developed because gastroenteritis is common, management varies and junior doctors make many of the initial decisions.

The vast majority of children have an infective cause, of which at least 80% are viral. The commonest causative organism is rotavirus, with a peak rate of infection in infants 6-12 months, and marked seasonality, peaking in Jan to March each year. The remaining 14% are predominantly campylobacter, salmonella, shigella and ecoli, in decreasing frequency. Acute infective diarrhoea is however a diagnosis of exclusion and other less common causes of diarrhoea must be considered and excluded first (see Table 1).

Morbidity and mortality are caused primarily by water and electrolyte losses in the stool. Thus the key to management is the prevention of dehydration and promotion of rehydration in those already dehydrated (see annotations K and L). This relies on an accurate estimate of the level of dehydration (Table 2). Rehydration should be done orally whenever possible and the keys to success are small amounts of Oral Rehydration Solution (ORS) frequently (Table 3).

Three years from date of completed development (Jan 1999), due Jan 2002. Paediatric Accident and Emergency Research Group responsible for revisions.

A 'seed' algorithm was initially drawn up. This consisted of a flow chart detailing the decision path that the clinician might take when dealing with a child with diarrhoea. The decision points on this 'seed' algorithm were used to develop the key questions which required an answer. The questions were specified using the format described in evidence-based medicine texts , describing the patient population of interest, the problem, the intervention or test (with or without comparison) and the outcome. This set of key questions directed the literature search.

The areas considered were: incidence of diarrhoea in the paediatric population; differential diagnoses of acute diarrhoea and their incidence; validity of history and examination in refining the diagnosis and determining severity; diagnostic tests required; need for admission; use of oral and intravenous rehydration in acute infective gastroenteritis; types of oral rehydration solution; patient and / or parent preferences in management. The search strategies for each question type are shown in table 8 with the clinical question type in table 9.

The literature search for this guideline used the following databases: Medline (1966 to June 1998), Embase and Cochrane (to June 1998) confined to humans aged 0-16 and English language), the citations of the references found and references provided by colleagues. The following mesh headings and text words were used: diarrh*; diarrhea infantile; diarrhea to 14 years; diarrhoea; gastroenteritis; differential diagnosis; diagnos*; incidence; prevalence; aetiology; etiology; dehydration; patient admission; fluid therapy; intravenous; intravenous treatment; rehydration solution; administration, oral; enteral nutrition; faeces; feces; lactose intolerance; enteral; differential diagnosis.

The references generated were limited to human, English language and child (aged 0-16). The remaining references were sifted for relevance to the clinical questions according to their titles and abstracts. They were then further sifted for inclusion and exclusion criteria. Articles with abstracts that demonstrated the following were examined: Those which addressed the specific clinical question; thorough scientific reviews of the literature; review or clinical guideline written by a national body; large, well designed clinical trial; Exclusions: Abstracts were excluded if the study was non-experimental, descriptive or opinion based.

Relevant articles were appraised and abstracted using a pre-designed proforma. This was developed from epidemiological texts . The articles were assessed as to whether they met the primary and secondary guides as suggested by the evidence-based medicine group in their series of articles in the Journal of the American Medical Association. Sixty articles were of sufficient quality and relevance to be included in the literature review, based on the key clinical questions and the evidence found. Draft management statements were constructed for each step. The initial 'seed' algorithm was revised to incorporate the evidence found.

Members – as listed in the acknowledgments.

The panellists selected were drawn from the United Kingdom, represented practice in both urban and rural settings and were clinicians who would be involved in management of a child after presentation at hospital. We did not include general practitioners, parents or patients. Ninety-six medical (consultant, registrar and SHO) and nursing staff from mixed adult/paediatric A&E departments, paediatric A&E departments, general paediatric departments (both teaching hospital and district general hospital) and specialist paediatric gastroenterology services were invited of whom 54 agreed to be included.

Delphi process

The definition of consensus is crucial to the ‘consensus development’ process, and should be decided before the process starts. For ‘nominal group consensus development’ rules have been developed to assess agreement when statements have been ranked on a 9-point scale . We chose to apply this to the Delphi method since the same scale was used. One sixth of the ratings furthest from the median were removed. This is done so that outliers (who may not have understood the question, or are unique in their views) do not overly influence the results. Consensus within the panel (known as ‘relaxed’ agreement for a nominal group) is defined as all remaining panellists’ responses falling within 3 boxes of each other on the Likert scale. Consensus agreement with the statement as presented to the panel is defined as all remaining responses falling in boxes 7-9 (thus agreement both between the panel members and with the statement as given, known as ‘strict agreement’ for a nominal group).

All statements that achieved ‘strict’ consensus were removed from subsequent rounds and used for guideline construction. For statements that did not gain consensus, modified and new statements were used in the second round. After the extreme one sixth of responses were removed, the range, inter-quartile range and median of the remaining responses were reported back to the panellists, along with any comments made. Panellists were asked to re-consider the statements in the light of the responses and comments of the rest of the panel. A third round consisted of statements that had not yet achieved consensus.

The guideline

All statements, which had achieved ‘strict’ consensus, were used to generate a guideline in algorithm format. Each box is numbered and annotated, referring the user to the evidence on which decisions are based. The literature review, level of evidence, grade of recommendation and consensus base of each decision point is given. The algorithm formed the basis of an integrated care pathway, which was used to pilot the guideline (described later) and to study its impact.

Throughout, the word ‘admit’ is defined as follows: ‘any admission to a paediatric facility with paediatric trained staff for observation, further investigation and management regardless of the expected length of stay’.

The algorithm is shown at the beginning of this document and the annotations for each box follow.

Corresponding with letters A-S on the algorithm

Summary table of the levels of evidence and grade of recommendation – fully reported in appendix 1

For each of the alphabetised boxes on the algorithm the literature is discussed. A statement is then made in Italics, which was put to the Delphi panel. The level of evidence for the statement and whether it achieved consensus with the panel follows.

2. Definition of diarrhoea

Diarrhoea is present when there is an increase in the frequency, volume or liquidity of the stool relative to the usual habit of the individual. There is a great variability in stool patterns amongst normal infants. Typically children excrete 5-10 g/Kg per day , but this can vary tremendously. Table 10 shows the range of stool pattern in normal infants. Most papers accept a working definition of diarrhoea as follows:

RECOMMENDATION

Diarrhoea is defined as a change in bowel habit for the individual child resulting in substantially more frequent and/or looser stools.

(Based on Level Vb evidence and Delphi consensus, grade D recommendation)

3. Differential diagnosis

Once a child has attended Accident and Emergency with a presenting complaint of diarrhoea with or without vomiting we need to know the possible differential diagnoses and the likelihood of these. Unfortunately there is very little data available to help with this clinical question. Conway et al performed a prospective hospital cohort study of patients initially thought to have acute gastro-enteritis and subsequently given other diagnoses. They included patients with vomiting alone. 1,148 were enrolled of which 59(5%) were found to have other diagnoses, which included infections other than in the GI tract, pyloric stenosis, feeding problems and cows milk protein intolerance.

Fleischer in his textbook of paediatric emergency medicine (Vb,D) gives a differential diagnostic list for children presenting with diarrhoea. In the absence of published evidence a modified list of differentials was sent to the Delphi panel and consensus agreement was achieved on table 1of the algorithm.

RECOMMENDATION

Table 1 to aid the differential diagnosis of the child presenting with diarrhoea

With respect to life threatening causes of diarrhoea, three studies yield helpful clinical information:

Macdonald and Beattie , (Vb,D) carried out a retrospective review of children with intussusception over a 10 year period. Population Incidence was found to be about 1 per 1000 in the first year of life. There were 110 children in whom 32% had diarrhoea at first presentation. 26% were shocked or dehydrated, 83% were vomiting, and 32% had bloody stool. The peak age of diagnosis was 5 months with 80% under 1. Only 42% were diagnosed correctly within 3 hours of admission.

Milford et al , (III,C) reported on the clinical and epidemiological aspects of HUS in the British Isles (2987-1989), finding cases through the British Paediatric Surveillance Unit and other sources. The overall incidence in children aged 0-15 years was 0.91/100,000. The peak incidence was in the age group 1-2 years at 3.3/100,000/year. 298 children were reported over the three-year surveillance. A prodrome of diarrhoea was present in 273 (95%) of cases and in 199 it was bloody. Diagnostic features on presentation were pallor in 92%, jaundice in 35% and oligo/anuria in 38%.

Reynolds , (Vb,D) looked retrospectively at children presenting with abdominal pain to the A&E department. 371 children were identified over 4 seasonally diverse months. The final diagnoses were medical in 64.4%, surgical in 6.5% and nonspecific in 29.1%. Guarding and abdominal tenderness were the two signs most strongly associated with a surgical diagnosis.

RECOMMENDATION

1. Estimation of severity of dehydration

Weight loss

The severity of dehydration is most accurately assessed in terms of weight loss as a percentage of total body weight (prior to the dehydrating episode). An accurate weight immediately pre-illness is rarely available in the clinical situation, but if it is (for example a recent weight in the parent held record) dehydration can be estimated with some accuracy .

Clinical signs

In a prospective cohort study of children between 3 and 18 months of age in Egypt, Duggan (III,C) found that ‘prolonged skinfold’, dry oral mucosa, sunken eyes and ‘altered neurological status’ were the best clinical signs correlating with dehydration as determined by post rehydration weight gain. In a similarly designed study, with children <4 years old, Mackenzie etal (III, C) found ‘decreased skin turgor’, decreased peripheral perfusion and deep (acidotic) breathing to be the best clinical signs. A urea of >6.5mmol/L on serum blood sample and pH<7.35 on blood gas were positive investigations associated with dehydration. However the sensitivity and specificity of all these signs were very low.

In both studies clinical estimates of percentage dehydration greatly overestimated actual dehydration. The textbook estimation of dehydration came originally from the Medical Research Council descriptions in 1952, and was modified by Ironside in 1970, and more recently by Santosham in 1987 . The percentage weight loses on which they were based were not subject to confirmation in a clinical study. More recently two authors have looked at this. Duggan (III,C) found that those thought to be mildly dehydrated by Santosham’s scale showed weight gains of 3.6-3.9%, those moderately dehydrated showed gains of 4.9-5.3% and those severely dehydrated 9.5-9.8%.

Mackenzie etal (III,C) looked only at children who were thought to be moderately dehydrated (7-10% estimated, with one of; diminished skin turgor, sunken eyes, dry mucous membranes, oliguria, recent weight loss). They found weight gains of 3.4-4.0%. These studies need repeating on larger numbers with rigorously defined clinical signs in order to confirm the findings. Nevertheless it is clear that estimation of dehydration may not be accurate.

Capillary refill time >2 seconds has been proposed as a useful indicator of dehydration . This sign lacks sensitivity and specificity . Gorelick etal (II,B) showed that in healthy children the capillary refill time was abnormally prolonged following 15 minutes in a cool (air conditioned) room. However, given it’s limitations it is important to note that a child who is severely dehydrated is very unlikely to have a normal capillary refill time. Likewise a prolonged capillary refill in a child with diarrhoea should be taken as a sign of dehydration until proven otherwise .

On the basis of the above studies and WHO guidelines on assessment of the dehydrated child table 2 was developed. Thus children are divided into three groups only, namely no dehydration, mild to moderate dehydration and severe dehydration. The number and severity of signs present will indicate the degree. The table was refined by the Delphi process and achieved consensus.

RECOMMENDATION

Table 2 for estimating level of dehydration.

(Level III evidence and Delphi consensus, grade C recommendation)

 

2. Investigations (plasma)

No studies have addressed this issue directly. Most episodes of dehydration caused by diarrhoea in developed countries are isonatraemic . Even when there is derangement of electrolytes in the serum, this is due to relative losses of salt and water. There will still be a total body depletion of sodium in hypernatraemic patients . It is clear from several hospital cohort studies that derangement of electrolytes in acute gastroenteritis in the UK is now rare. Table 11 summarizes three recent UK papers looking at hospital cohorts of children with GE. Approximately 1% of these admissions had hypernatraemia. None of these studies reported hypokalaemia or hyponatraemia, which are commonly found in patients dehydrated with cholera.

Holliday (II, B) eloquently argues from published evidence that Oral Rehydration Solution with appropriate amounts of solutes and given in the correct quantity is sufficient in itself to correct electrolyte abnormalities (see also Meyers ). It is thus unnecessary to measure electrolytes in those children who will be rehydrated with Oral Rehydration Solution (ORS).

If a child is severely dehydrated with circulatory compromise, they will need rapid Intravenous (IV) infusions to restore circulation and renal perfusion (20ml/kg boluses until circulation restored). Suitable fluids are normal saline or ringers lactate . All children having IV rehydration should have a U&E measured, as hypernatraemia will alter the rate at which IV rehydration fluids are given (discussed below). Further measurements of U&E should be made as rehydration progresses.

The American Academy of Pediatrics (AAP) suggest in their practice parameter that Electrolyte levels should be measured in moderately dehydrated children whose histories or physical findings are inconsistent with straightforward diarrhoeal episodes, and in all severely dehydrated children. This is based on a consensus view (Va, D). They also state that clinicians should be aware of the clinical features of hypernatraemic dehydration, namely a doughy feel to the skin +/- irritability and fever. There is no quoted evidence on which this statement is based.

RECOMMENDATION

1. IV Rehydration fluid in severe gastroenteritis

If a child is severely dehydrated with circulatory compromise, they will need rapid IV infusions to restore circulation and renal perfusion (20ml/kg boluses until circulation restored). Suitable fluids are normal saline (Vb,D) or ringers lactate the compositions of which are given in Table 12. Once their circulation is stable replacement of the remaining estimated deficit can commence (details below).

At the time of Delphi consensus development for this guideline there was controversy over the use of crystalloid versus colloid, which many of the Delphi panelists brought up in their comments. There was no literature on the particular issue of crystalloid versus colloid in the resuscitation of infants and children with diarrhoea. In studies in adults crystalloid is know to be as effective for rapid restoration of circulating fluid volume. Until we have more evidence for children (and this is likely to have to come from a developing country as the numbers of children presenting in shock with diarrhoea are so small in the UK) we will have to use the current literature which does not include a randomised controlled trial on crystalloid versus colloid. Sharifi , Mackenzie and Jenkins all report the use of crystalloid in severe dehydration. No studies report the use of colloid in diarrhoea.

The child with dehydrating diarrhoea is different from a child with shock secondary to trauma or sepsis. The dehydrated child has lost water and salts from all body compartments. In severe dehydration the final compartment to decompensate is the intravascular one. The child will have a high haematocrit and will not have lost any plasma proteins. It thus seems reasonable from a theoretical point of view to restore what has been lost, namely water and salts.

It is argued that if crystalloids are used they diffuse more readily into the interstitial and intracellular compartments. As these compartments are depleted in dehydrating diarrhoea, this seems a theoretically good thing, as long as further fluid is given to maintain intravascular volume.

RECOMMENDATION

Children who have severe dehydration with circulatory compromise secondary to acute gastroenteritis should have their circulation restored by rapid IV infusion of normal saline or ringers lactate with a 20ml/kg bolus over one hour (faster if necessary). An experienced paediatrician should be involved early.

(Based on Level I evidence and Delphi consensus, grade A recommendation)

2. Further boluses

In severe dehydration one bolus may not be sufficient.

RECOMMENDATION

A further bolus of 20ml/kg should be given if the circulation is still compromised. If further boluses are required (>40ml/Kg) involve an anaesthetist early as intubation and ventilation should be considered

(Based on Level Vb evidence and Delphi consensus, grade D recommendation)

3. Hypernatraemic dehydration

It is acknowledged following the evidence of several randomised controlled trials in USA, Europe and developing countries that ORS is quicker in the correction of dehydration and acidosis and safer than IV therapy. Moreover the use of Oral Rehydration Therapy (ORT) appears to reduce the risk of seizure during correction of hypernatraemic dehydration, Pizarro et al(Vb,D) reported no seizures among 35 infants with hypernatraemic dehydration whose deficit was repaired with WHO-ORT over 12 hours.

In the largest Randomised Controlled Trial (RCT) of IV versus ORT Sharifi et al(II,B) randomly assigned 470 children aged 1 to 18 months (without malnutrition) admitted to hospital in Tehran with severe acute GE to receive either ORS (administered initially by Naso-gastric tube (NGT)) or IV fluid. Of the 34 hypernatraemic patients in the ORT group, 2(6%) developed generalised seizures compared with 6 of 24 (25%) in the intravenous group. Similar results were found by Mackenzie (II,B) in a study in Australia, the ORS group fairing at least as well as the IV group.

RECOMMENDATION

The child with with hypernatraemic dehydration (Na>150mmol/L) secondary to acute gastro-enteritis should be given slow ORS, aiming to give the estimated deficit over 12 hours. Monitor electrolytes closely

(Based on Level I evidence and Delphi consensus, grade A recommendation)

4. Oral versus IV rehydration in the severely dehydrated child following restoration of circulating fluid volume.

The above literature suggests that further rehydration should be done with ORS .

RECOMMENDATION

Further rehydration can be done orally with ORS (given little and often) if the patient is stable and their mental state allows it.

(Based on Level I evidence and Delphi consensus, grade A recommendation)

 

5. Ward management of rehydration

The overriding principles of the management of gastro-enteritis are rehydration and prevention of dehydration. Several excellent systematic reviews have been published on this subject. The American Paediatric Association (I,A) and the European Society of Paediatric Gastroenterology and Nutrition (I,A) have both produced practice guidelines concerning management of children with gastro-enteritis, and their main recommendations are based on meta-analysis of good randomised controlled trials. In addition to these there is also a good systematic review published by Murphy (I,A).

In the 1970’s the WHO adopted a glucose-electrolyte solution for the treatment of diarrhoea that contained 90mmol/l of sodium. Since then there have been many controlled trials looking at the ideal concentration of electrolytes and carbohydrate in ORS. In developing countries rapid loses of sodium and potassium are documented, particularly with cholera, nevertheless a recent multicentre trial in 4 developing countries found that reduced osmolarity ORS (224mmol/l) had advantages over standard ORS (311mmol/l) in the treatment of non cholera diarrhoea (II,B). In developed countries diarrhoea tends to be isotonic, and therefore replacement of large quantities of sodium is not so imperative, and indeed may be harmful. Studies from Finland have confirmed that reduced osmolarity ORS is preferable in European children. European Society of Paediatric Gastroenterology And Nutritian (ESPGAN) have published guidelines on the ideal composition of ORS for children of Europe which have taken into account the results of these trials. See table 12 for the composition of ORS recommended and those commercially available.

Many papers have been published looking at the different types of carbohydrate to be used in ORS. A recent meta-analysis of 13 clinical trials examining the effect of rice based ORS on total stool output and duration of diarrhoea showed that there appeared to be some benefit in those with cholera, but in those with non-cholera diarrhoea it was uncertain (I, A).

RECOMMENDATION

RECOMMENDATION

*An attempt was made to define little and often further. The literature discusses the correct administration of ORS and recommends that it be given in 5ml aliquots every 1-2 minutes. Only if this is well tolerated with no vomiting the size of the aliquots may be increases with decreasing frequency. However this regime was thought to be too labor intensive for the UK by the Delphi panelists and did not achieve consensus.

1. Replacement of losses in the child at risk of dehydration.

There are no trials concerning this issue, but the AAP practice parameter and Murphy’s review recommend that ongoing losses for the infant at high risk of dehydration should be replaced with 10ml/Kg of ORS for each loose stool and substantial vomit (I,A). However, the Delphi panel did not agree with this point. An alternative of replacing the estimated volume lost was put to the panel but this also did not achieve consensus. In view of the lack of consensus and no evidence in the literature either way, this issue will have to be decided at a local level.

2. Criteria for admission of children with gastroenteritis

See annotation P

3. Literature concerning the need for stool culture

Once a diagnosis of acute gastro-enteritis has been made clinically, the question of the aetiology of the infection arises. For the individual, it would be important to know what is causing the symptoms if treatment of the infection could eliminate them. As we shall discuss later on, treatment is rarely necessary and therefore stool culture for this reason alone is not productive.

Some might argue that we would have a clearer idea of the prognosis if we knew the aetiology. With respect to acute risk of dehydration this does not seem to be the case. The risk of dehydration was the same for all aetiologic agents except cholera in both Faruque (III,C) and Bhattacharya’s studies (III,C). Fortunately in the UK cholera is only seen rarely in children who have travelled abroad. With respect to predicting which infections are likely to become chronic, it may be useful to know the pathogen. However when a child presents acutely it is unnecessary to make this distinction. If they present with diarrhoea lasting 5 or more days, a stool sample can be taken at this point.

In the UK a history of travelling abroad must be taken seriously. There have been two case series reported in the literature of children with malnutrition and severe chronic diarrhoea treated in UK hospitals following an extended trip abroad (Vb,D). A history of foreign travel therefore is a good reason to check a stool culture.

From a public health point of view it is clearly important to know which organisms in the community are causing infections, and more specifically whether there is any evidence of outbreaks of disease. With respect to food poisoning (Shigella, salmonella, campylobacter) it is important that the source of any outbreak is traced and dealt with.

Thus it is clear from the public health point of view that some stool samples should be sent for culture. However if all patients with a short spell of diarrhoea had a stool sample sent to the laboratory for culture the lab would be overwhelmed. It therefore seems reasonable to try to limit stool specimens sent to those likely to have important (bacterial or parasitic) infections.

DeWitt etal (III,C) looked at the value of various features of history and examination and stool screening tests in predicting whether diarrhoea was caused by a bacterial agent. They studied 200 children less than 4 years old presenting to a primary care centre in the USA with diarrhoea of less than 10 days. The best predictor on clinical grounds alone was a cluster of 3 historical variables- abrupt onset of diarrhoea, more than 4 stools per day and no vomiting before the onset of diarrhoea. This cluster had a sensitivity of 86% a specificity of 60%, PPV of 27% and NPV of 96%.

Diarrhoea which is frankly bloody is more likely to be caused by invasive bacteria than viruses. Finkelstein etal (III,C) found that in 1,035 infants under 1 year of life with diarrhoea (of which 108 (10.4%) had a bacterial cause), a history of blood in the stool was the best individual predictor of bacterial infection (sensitivity 39%, specificity 88%, PPV 30%, NPV 92%). Temperature >39° C and >10 stools per day were also useful indicators. Conway (Vb, D) looked at 1148 children <16 years with diarrhoea, in whom 153 (13%) had bacterial, protozoal or mixed pathogen aetiology. They found that the bacterial group had a statistically significant higher stool frequency of >7 per day, but the difference was of little use to the clinician (36% in bacterial group and 26% in rotavirus group, no figures given to calculate sensitivity etc). They also found that the stool more frequently contained blood or mucus (25% in the bacterial group compared with 2.8% in the viral group). In Milford’s study of HUS (III,C), 199 children (73%) had a prodrome of bloody diarrhoea, with 178 of these growing colliforms in the stool.

Based on these studies Table 5 was developed.

RECOMMENDATION concerning requirement for stool culture

• A history of blood +/- mucous in the stool (Level III and Delphi consensus, grade C)
• Systemically unwell, severe or prolonged diarrhoea (no literature, Delphi consensus)
• If the child is admitted to hospital (no literature AND no Delphi consensus, this must be decided at a local level)
• A history suggestive of food poisoning (no literature, Delphi consensus)
• Recent travel abroad (Level Vb and Delphi consensus, grade D)
• Abrupt onset of diarrhoea with more than 4 stools per day and no vomiting pre diarrhoea (Level III, BUT no Delphi consensus. To be decided at a local level).

1. Role of medication in gastroenteritis

Several trials looking at the use of these agents have been reported. They are thoroughly reviewed in the AAP practice parameter and Murphy’s systematic review.

RECOMMENDATION

There are several trials reported in the literature investigating the treatment of Campylobacter jejuni. Murphy reports that one randomised controlled trial indicated that if erythromycin was started at first presentation before the results of the stool culture were available, the clinical course of the illness was shortened. Other randomised trials in which erythromycin was started after isolation of the organism showed a shortened period of bacterial excretion, but no effect on the clinical course of the illness (I,A). Trials investigating the antibiotic treatment of Yersinia enterocolitica and E coli. have not found any benefit.

Invasive Salmonella typhi (i.e. a systemic infection featuring malaise, meningismus, and fever) clearly needs treatment and sensitivities are required. For other salmonella infections there is no evidence that anti-microbials help (III, C), although they should be considered in young infants, immunocompromised children, and those who are systemically unwell. Anti-microbial treatment is however worthwhile in Shigella, Amoebiasis and Giardiasis .

RECOMMENDATION

• Salmonella typhi – ciprofloxacin until sensitivities available
• Shigella – trimethoprim (or ciprofloxacin for trimethoprim-resistent strains)
• Giardia – metronidazole
• Amoebiasis – metronidazole and Diloxanide furoate.

Table 9 Clinical questions and question type for the diarrhoea guideline.

Table 11: Frequency of deranged electrolytes in acute gastroenteritis in developed countries.

NR= Not reported.

Table 12: Composition of fluids for intravenous and oral rehydration.

*Glucose given with fructose 1mmol/L and sucrose 94mmol/L

A study was undertaken to evaluate the implementation of this guideline in a paediatric A&E on the management of the child presenting with diarrhoea, with or without vomiting.

Method: A care pathway was developed with nursing and medical staff, based on the guideline algorithm (appendix 3). This was used as the documentation for children presenting with diarrhoea, and followed the child to the ward if admitted.

The key elements of the assessment and management of the child with diarrhoea, using this guideline were used to develop a data collection form. These data were collected from the notes of children attending A&E (both GP referrals and self referrals) during a four month period in 1997 and compared with those attending during a four month period in 1999, following implementation of the care pathway. Data were compared using SPSSâ , Chi-square and Man-Whitney-U tests.

Results: 292 children attended with diarrhoea pre care pathway and 239 post. There was no difference in age, sex, or time of arrival. Numbers admitted increased from 27% to 34%. During the same period there was a 14% increase in admissions of children presenting with all other medical problems. There was no change in the numbers of children returning to A&E having been discharged.

The time taken from seeing the doctor to discharge was reduced by 15 minutes from a median of 55 minutes to 40 minutes, and the total time in the department reduced by 24 minutes from median 102 to 78 minutes.

The number of children investigated for FBC and U&E fell (17 to 6, c ² p = 0.02 and 18 to 7, p=0.02 respectively), and intravenous infusions fell (13 to 2, c ² p=0.002). Other investigations remained the same. Documentation of symptoms, signs and management plan was improved.

Conclusion: The implementation of a care pathway for diarrhoea reduced the numbers of unnecessary investigations and unnecessary IV canulations. It also reduced the time spent in the A&E department. The proportion of attenders admitted increased in keeping with the overall increase in medical admissions. A study on the appropriateness of admission of both samples is planned.

Audit recommendations

The following data are valuable for monitoring compliance with guideline recommendations and auditing the impact of the guideline on clinical practice:

• Proportion of children admitted within the three levels of dehydration (none, with or without risk factors for dehydration, mild/moderate and severe) pre and post guideline implementation.

• Proportion of children returning to hospital (within 7 days) with the same presenting problem before and after guideline implementation.
• Proportion of children investigated by clinical chemistry and microbiology (include data on criteria for stool samples) pre and post guideline implementation. Record frequency of abnormal results.

• Proportion of children within each category of dehydration level who have a canula sited with or without commencement of IV rehydration, pre and post guideline implementation.

• Monitoring length of time for rehydration in dehydrated children, and duration of ‘starvation’ prior to recommencing feeds.

• Monitoring of length of time taken to manage a child presenting with diarrhoea from consultation to admission or discharge.

Levels of evidence, as suggested by Muir Gray1

Suggested grading based on Cook2

Grade A, supported by level I evidence and therefore highly recommended.

Grade B, supported by level II evidence, and therefore recommended.

Grade C, supported by level III, evidence. Several potential clinical actions might be considered appropriate.

Grade D, supported by level IV and V evidence. The consensus route would have to be adopted.

2. Cook D, Guyatt G, Laupacis A and Sackett D. Rules of Evidence and Clinical Recommendations on the Use of Antithrombotic Agents. Chest 1992: 102: 305S-311S.

What is gastroenteritis?

Gastroenteritis is an infection in the gut, which leads to diarrhoea and/or vomiting (sickness). Diarrhoea is frequent watery poo. The infection may also give your child a temperature and tummy pain. It is usually caused by a virus, which the body clears on it’s own without treatment. The diarrhoea and vomiting may lead to dehydration (too much water lost from the body).

What do I do?

Your doctor has carefully looked for signs of dehydration and has not found any. They are therefore happy for you to take your child home. You must encourage your child to drink.

What is enough fluid?

• Your child’s weight today is

• He/She needs to take in at least of fluid over a 24 hour period.

• A teaspoon is 5mls. 1oz is 30mls. A typical beaker holds about 200 mls.

What kind of drinks should I give?

You can give any drink that your child usually has including milk. However try not to give very concentrated or sugary drinks like real fruit juices or fizzy drinks unless they are well diluted with water (4 times as much water as drink).

What if my child is vomiting?

• Stop all solid food until the vomiting has settled
• If your baby is breast fed continue to feed on demand
• If your baby is formula fed, give feeds in very small amounts (approximately 1oz(30ml)) often (every 20 minutes or so). If they continue to vomit, stop milk feeds for 4 hours and give cooled boiled water instead, little and often.

• Any other fluids that your baby or child has should be given as above (about 1oz(30ml) every 20 minutes) by bottle, spoon or cup. Do not offer a full bottle or cup as large amounts may make your child vomit again.

• As the vomiting settles you can start to offer larger amounts of fluid less often and the child’s usual solid food.
• Try rice, pasta, potatoes, toast, plain biscuits. Don’t worry if they are not hungry.
• AVOID fatty foods and sugary foods.

What about the diarrhoea?

Diarrhoea usually continues for 6-7 days. As long as your child is drinking and is improving in themselves this does not matter.

What about the temperature?

If your child has a temperature or appears to have tummy pains then give Paracetamol according to the instructions on the bottle.

What is Dioralyte/ Diocalm junior/ Electrolade?

• These are all names for salt and sugar solutions that are made up with water, to replace what is being lost. You will only be given these by your doctor if your child is dehydrated or at risk of becoming dehydrated.
• When a child is not dehydrated they may be used to supplement the child’s normal fluid.

• Try to give each time your child has a very loose poo, or large vomit. Give small amounts often. If your child does not like the taste try adding a drop of juice or sugar-free squash.

When should I ask for help or advice?

Seek advice if:

• The diarrhoea has blood in it
• Your child becomes more sleepy, lethargic or irritable than usual
• Your child has 5 or more vomits in 24 hours
• Your child has 9 or more loose poos in 24 hours
• The diarrhoea continues for more than 7 days

You could call your Health Visitor or General Practitioner

You could call the short stay unit at the Queen’s Medical Centre up to 48 hours after being on the ward

Telephone No

You could call NHS Direct

Telephone No

You could call Children’s A&E at the Queen’s Medical Centre up to 48 hours after being seen there

Telephone No

When can my child return to school or nursery?

When the diarrhoea has settled to 2 or 3 formed poos a day they are safe to return.

What about my baby’s sore bottom?

Frequent diarrhoea can make your baby’s bottom sore.

• Try to change the nappy as soon as it is dirty.
• Clean carefully with cotton wool and water or baby lotion (some wipes are alcohol based which can be sore on a red bottom).

• Apply barrier cream or Vaseline liberally.

How can I stop it happening again?

• Gastroenteritis is an infection that can be passed on from person to person or in contaminated food.
• Always wash hands before preparing any foods or eating and after nappy changes or going to the toilet.

• It is very important to wash and sterilise all baby bottles, teats and feeding equipment.

The following 3 pages show the care pathway exactly as used in the Nottingham paediatric A&E

during the implementation study

C. Signs of circulatory compromise go to BOX A

Complete Nursing, History and Examination for all. Then complete Box A or B or C as applies. DATE………………………

MEDICAL assessment: Doctor name:………………………………………………Date/Time………………………………………

Appraisal of the Paediatric A&E Research Group guideline for the management of the child presenting to hospital with diarrhoea with or without vomiting.

Appraisal performed by Dr Kate Armon

(in compliance with the RCPCH Quality of Practice Committee Report)

eLetters:

Read all eLetters

Duration of treatment of moderately dehydrated cases with glucose oral rehydration solution

Elif Ozmert

ADC Online, 30 Nov 2001 [Full text]