ArticlePDF Available

Abstract and Figures

Limited clinical evidence plus much biochemical analysis suggests that carbonated drinks, such as cola (whether "flat" or otherwise) should not be recommended for oral rehydration in children with acute viral gastroenteritis. Parents and carers should be discouraged from using "flat" fizzy drinks for this purpose.
Content may be subject to copyright.
Towards evidence-based medicine for paediatricians
Edited by Bob Phillips
Archimedes seeks to assist practising clin-
icians by providing ‘‘evidence based’’
answers to common questions that are
not at the forefront of research but are at
the core of practice (format adapted from
BestBETS published in the Emergency
Medicine Journal). A full description of
the format is available online at http://
Readers wishing to submit their own
questions—with best evidence answers—
are encouraged to review those already
proposed at If your
question still hasn’t been answered, feel
free to submit your summary according to
the instructions for authors at http://adc.
You are the registrar on duty in the
children’s emergency department. A regis-
tered children’s nurse asks you about an
18-month-old child who came in with his
parents. He has been vomiting for the last
24 h and has today developed loose
watery diarrhoea. His mother is con-
cerned that his oral intake is poor and
his nappies are not as wet as normal. The
GP prescribed some oral rehydration solu-
tion yesterday, but the child is refusing to
drink it. A neighbour told the mother that
allowing the child to drink ‘‘flat’’ cola was
a good way to prevent him from getting
dehydrated. The nurse asks you if this is a
safe and acceptable treatment to recom-
mend for children. You have heard it
mentioned by parents of children with
gastroenteritis before but feel unsure
whether any evidence supports it.
In children with viral gastroenteritis
[subject] do ‘‘flat’’ fizzy drinks such as
Coca-Cola [intervention] compared with
oral rehydration solution [comparison]
offer adequate rehydration?
Primary searching used the dialog
DataStar interface Medline 1950 to date,
CINAHL(R) 1982 to date and EMBASE
1974 to date (search date: 18 April 2007).
Two search strategies were run, firstly
[gastroenteritis (textword) or gastroenter- (MeSH term)] AND [coca cola or or cola or fizzy
drink$] yielding 25 records and secondly
[coca cola or or
cola or fizzy drink$] AND [dehydrat$]
OR [rehydrat$] yielding nine records, all
of which were found by the first strategy.
From these, nine papers were relevant and
A further five relevant papers
not picked up by initial searching were
found from the articles reviewed.
A secondary search was performed on
The Cochrane Library (Issue 2, 2007)
using the following terms: [gastroenteritis
AND treatment], [gastroenteritis AND
carbonated drink$], [gastroenteritis AND
cola] and [gastroenteritis AND flat cola].
Three articles were found but none was
relevant to this question.
Although it is difficult to know what
terms to choose in order to find all
possible literature relating to such drinks,
the terms used were felt to be exhaustive.
Repeated searches using alternative terms
But at what cost?
It’s uncommon for us, as paediatricians, to be asked about how cost effective our treatments are.
Glancing at the media shows health stories about the new wonder drugs in adult cancer, or in
Alzheimer’s disease, and how they are being restricted by a heartless and miserly health system.
Where do these statements about ‘‘cost effectiveness’’ come from?
The basics of health economic analysis are simple. If you are comparing treatment Adrug and
Bepill, and they both work well, you want the one that is cheapest (a ‘‘cost minimisation’’ process; for
more on these descriptions, see Phillips from 2002
). But if Adrug works better than Bepill, then the
element of costs becomes more important. (Obviously, if the Adrug is cheaper, it’s a no-brainer!) The
question becomes: ‘‘How much is society/your insurance company/your patient prepared to pay for
the extra benefit?’’.
If you have two treatments which have the same outcome, this can be a difficult question to
answer (it’s a cost–effectiveness analysis). Imagine the difficulty that is added by asking the same
question but needing the answer to compare different conditions, for example asthma and
neurodevelopmental disability (a cost–utility analysis). What you need is a common metric to
compare the ‘‘benefit’’ across these conditions, which is what the quality adjusted life year (QALY) is
intended to be. There are lots of things to think about with QALYs (for an intelligent discussion see
Petrou and McIntosh
), but QALYs are a good starting point and may help avoid the divisive
judgement issues which could emerge when comparing surgery for chronic constipation,
chemotherapy for relapsed-refractory leukaemia and treatment for alcohol and drug dependence in
The advanced bits of health economics are not simple. They compare the presumptions about how
effective and costly things are, perform analysis on how sensitive the results are to variations in the
‘‘truths’’ that the model has supposed, and add factors that balance for the fact that we value present
health much higher than health in the future. (You know this is true – what happens when you’re
offered another piece of fruit cake?) These all lead to clouds of data, odd looking graphs, but usually a
relatively straightforward answer along the lines of ‘‘It probably costs £10 000 or so per QALY, and
the truth might be between £5000 and £42 000’’.
Which then puts you back to the beginning – it works, but at what cost?
1. Phillips B. Towards evidence based medicine for paediatricians. Arch Dis Child 2002;87:77.
2. Petrou S, McIntosh E. Measuring the benefits of growth hormone therapy in children: a role for
preference-based approaches? Arch Dis Child 2008;93:95–7.
Bob Phillips; Evidence-based On Call, Centre for Evidence-based Medicine, University Department of
Psychiatry, Warneford Hospital, Headington OX3 7JX, UK;
Clinical bottom line
cLimited clinical evidence plus much
biochemical analysis suggests that
carbonated drinks, such as cola (whether
‘‘flat’’ or otherwise) should not be
recommended for oral rehydration in
children with acute viral gastroenteritis.
cParents and carers should be discouraged
from using ‘‘flat’’ fizzy drinks for this
Arch Dis Child Educ Pract Ed August 2008 Vol 93 No 4 129 on March 18, 2013 - Published by ep.bmj.comDownloaded from
yielded no new papers and all the papers
found by the initial search.
Papers were mainly opinion or letters
offering personal experience. One rando-
mised controlled trial was found, but only
in a published abstract; the full study
could not be found.
Several biochemical
analyses-based papers and recommenda-
tions were found which alluded to the use
of carbonated drinks in children with
acute gastroenteritis. An analysis of this
data was included. Table 1 shows an
appraisal of the single randomised trial.
No good quality controlled randomised
trials were found. There were a small
number of relevant papers, but none were
published in recent years. Those relevant
to the clinical scenario were mainly
offering opinion.
However, importantly, papers relating
to the scientific analysis of the biochem-
ical constituents of various liquids, both
physiological (eg, oral rehydration solu-
tion) and non-physiological (eg, carbo-
nated drinks), were found. The issue of
biochemical composition of carbonated
drinks compared to oral rehydration solu-
tion is relevant in the management of a
mildly dehydrated child. The biochemical
analyses clearly show the low levels of
sodium (1.0–9.9 mmol/l) and potassium
(0–0.3 mmol/l) and the very high osmol-
ality (388–790 mOsm/kg of water) of
various carbonated drinks.
Table 2
summarises some of the data from these
studies. WHO recommendations, which
have changed since the publication of the
papers found in our searches, were for oral
rehydration solutions to contain 90 mmol/l
of sodium, 20 mmol/l of potassium and
333 mOsm/kg of water.
Also of note,
although not consistently reported in bio-
chemical data, is the amount of glucose in
such drinks. Branded cola contains approxi-
mately 550 mmol/l, while WHO oral rehy-
dration solution contained 111 mmol/l as
Currently, the WHO form-
ulation contains 75 mmol/l sodium and
the same amount of glucose.
It is
concluded that there is little direct clinical
data to answer our question but many
scientific analyses demonstrating the inap-
propriateness of non-physiological liquids
for the treatment of acute gastroenteritis
in children compared with physiologically-
based oral rehydration solutions. No study
reported the degree of carbonation (ie, how
‘‘flat’’ or ‘‘fizzy’’ the carbonated drinks
were) in relation to electrolyte content or
osmolality. Therefore carbonated drinks,
‘‘flat’’ or otherwise, including cola, provide
Table 2 Results from biochemical analysis of various carbonated drinks from selected studies compared to oral rehydration solution
WHO formula:
oral rehydration
Dibley et al
Weitzman et a
Head et al
Jacobs and
this article (UK)
Sodium (mmol/l)
Oral rehydration solution* 90
Coca-Cola 4.0 2.5 6.1 5.6 3.0 ,5.0
Pepsi 3.0 1.4 – 1.0 –
Lemonade (Schweppes) 5.0 – – – – –
Sprite – 3.0 9.9 – – –
7UP – 4.5 – – 4.0
Potassium (mmol/l)
Oral rehydration solution 20
Coca-Cola 0.2 0 0 0 0.1 ,5.0
Pepsi 0.1 0 0.1 –
Lemonade (Schweppes) 0.1 – – – – –
Sprite – 0 0.1 – – –
7UP – 0.15 – – 0 –
Osmolality (mmol/kg water)
Oral rehydration solution 333
Coca-Cola 630 605 656 Not reported 469 410
Pepsi 583 624 – 576 –
Lemonade (Schweppes) 420 – – – – –
Sprite – 526 428 – – –
7UP – 485 – – 388
Information from several sources, including this article.
*WHO formula is not commonly used in the UK; UK formulations vary in sodium concentration (50–60 mmol/l) and WHO formula composition is now lower than that cited from
Dibley et al above.
Table 1 Does drinking ‘‘flat’’ cola prevent dehydration in children with acute gastroenteritis?
Citation Study group Study type Outcome Key results Comments
Feldman et al,
110 infants
Oral electrolyte
solution (OES)
group mean age:
8.7 months;
unspecified clear
fluid group mean
age: 10.5 months
Randomised controlled trial
comparing oral electrolyte
solutions with unspecified
clear fluids (including fruit
juices, carbonated
beverages, liquid gelatine,
weak tea and water).
OES group
ingested a larger
volume of fluids
and had fewer
treatment failures
No difference in purge
rate, time to onset of
solids, time to
discharge, 36-h serum
electrolyte levels or
causative pathogens
between groups
More detail on
methodology not
available; only abstract
reported, full study not
Limitations: although there
were a reasonable number
of subjects, the control
group includes a mixture
of different fluids and not
specifically or only ‘‘flat’’
carbonated fluid
Level of evidence 1b
(details of randomisation
not available)
130 Arch Dis Child Educ Pract Ed August 2008 Vol 93 No 4 on March 18, 2013 - Published by ep.bmj.comDownloaded from
inadequate fluid and electrolyte replace-
ment and cannot be recommended.
M Jacobs, Children’s Emergency Department, Watford
General Hospital, Watford, Hertfordshire WD18 0HB, UK
AReece, Children’s Emergency Department, Watford
General Hospital, Vicarage Road, Watford, Hertfordshire
WD18 0HB, UK;
Acknowledgements: Dr Bharat Patel, Consultant
Biochemist, Watford General Hospital for analysing a sample
of Coca-Cola and Janet Reynolds, Librarian, Watford General
Hospital for advice regarding literature searches.
Competing interests: None.
Arch Dis Child Educ Pract Ed 2008;93:129–131.
1. Anon. Acute diarrhoea in infants: oral rehydration is
crucial. Prescrire Int 2000;9(49):146–53.
2. Hefelfinger DC. More on cola drinks and rehydration
in acute diarrhea. N Engl J Med 1987;316(5):280.
3. Weizman ZN. Cola drinks and rehydration in acute
diarrhea. N Engl J Med 1986;315(12):768.
4. Dibley M, Phillips F, Mahoney TJ, et al. Oral
rehydration fluids used in the treatment of
diarrhoea. Analysis of the osmolalities, and
sodium, potassium and sugar contents of
commercial and home-made products. Med J Aust
5. Guignard JP. Potassium in coca-cola. Lancet
6. Wicklund S, Devroye M, Archbald CK, et al. Nurses’
drug alert: cola drinks not recommended for diarrhea.
Am J Nurs 1987;11(1):4.
7. Anon. Nurses’ drug alert: Coca-Cola is not for
diarrhea... should not be used for potassium
replacement in infant diarrhea. Am J Nurs
8. Weitzman Z, Mozes S. Electrolyte content and
osmolality of Israel soft drinks and their unsuitability
for oral rehydration in infantile diarrhea. Harefuah
9. Anon. Is coca-cola suitable for rehydration of very
young children with diarrhea? Ned Tijdschr Geneeskd
10. Subcommittee on Acute Gastroenteritis and
Provisional Committee on Quality Improvement.
Practice parameter: the management of acute
gastroenteritis in young children. Pediatrics
11. Feldman W, Pennie R, Ritter H. Oral electrolyte
solutions vs unspecified clear fluids in the
management of mild gastroenteritis in infants
(abstract). Am J Dis Child 1986;140:303.
12. Klish WJ. Use of oral fluids in treatment of diarrhea.
Pediatr Rev 1985;1:27–30.
13. Listernick R, Zieseri E, Davis AT. Outpatient oral
rehydration in the United States. Am J Dis Child
14. Head J, Hogarth M, Parsloe J, et al. Soft drinks,
electrolytes and sick children. Lancet 1983;1:1450.
15. RCPCH. In: BNF for children. London, Royal College of
Paediatrics and Child Health, 2007:517.
A 13-year-old boy with a diagnosis of
attention deficit hyperactivity disorder
(ADHD) comes to the clinic with his
mother for a review. He was started on
atomoxetine 6 weeks prior to this visit for
hyperactive/impulsive symptoms and
poor concentration. The boy was
admitted to the hospital 1 week ago for
changed behaviour, disorientation, irrele-
vant speech and self-harming behaviour.
He was reported to be very aggressive and
hostile towards other children and adults.
Investigation for organic causes was
undertaken and no clear explanation was
found. The patient was discharged on day 3
upon improvement of symptoms.
Although the diagnosis of ADHD had been
made some time ago, stimulant medica-
tions were not used because of the risk of
abuse and drug diversion. His mother
ascribes this hospitalisation to a side effect
of atomoxetine as he had not shown such
behaviour before starting on treatment and
there is no recent history of head injury,
illness or change in family dynamics. She
asks your opinion about increased aggres-
sion and hostility related to atomoxetine.
In a school age child with ADHD
[patient] can atomoxetine [intervention]
cause increased aggression and hostility
Search terms: keywords ‘‘ADHD’’,
‘‘Atomoxetine’’; textwords ‘‘Aggression’’,
Primary search
EMBASE 1996–July week 2 2007 and
Medline 1996–July week 2 2007.
Secondary search
The Cochrane Library.
In Embase the keywords ‘‘atomoxetine’’
(all subheadings) and ‘‘attention deficit
hyperactivity disorder’’ (all subheadings)
were exploded and combined with ‘‘and’’.
The references retrieved were combined
with the truncated textwords ‘‘Agressi$’’
or ‘‘hostil$’’. The results were exploded,
combined with ‘‘Evidence Based
Medicine’’. The textwords were combined
with ‘‘OR’’ with the keyword. In Medline
the exploded MeSH term ‘‘attention deficit
hyperactivity disorder’’ with all subhead-
ings was used. The results were combined
with the textwords ‘‘Atomoxetine’’ or
‘‘Strattera’’ using ‘‘AND’’. The results were
combined with the publication types meta-
analysis or randomised clinical trials.
The search was carried out on 13 July
2007 and a total of eight articles were
obtained in Embase combining the above
searches. After searching abstracts, one
relevant article was retrieved (table 1).
The other seven articles did not study
aggression or hostility in association with
atomoxetine. Medline did not identify any
study. The search of a secondary database
in the Cochrane Library showed no results.
ADHD is a pervasive disorder charac-
terised by levels of inattention, impulsive-
ness and hyperactivity inappropriate to
developmental stage. The Diagnostic and
Statistical Manual of Mental Disorders-
fourth edition (DSM-IV) of the American
Psychiatric Association identifies three
subtypes of ADHD: predominantly inat-
tentive, hyperactive-impulsive and a com-
bined type.
Comorbidity is the norm
rather than the exception and is more
frequently diagnosed in children with
A systematic review of five
studies indicated pooled prevalence esti-
mates of 35% for oppositional defiant
disorder (95% CI 27.2 to 43.8), 25% for
conduct disorder (95% CI 12.8 to 41.3),
18% for depressive disorders (95% CI 11.1
to 26.6) and 25% for anxiety disorders (95%
CI 17.6 to 35.3).
Over 28% of children have
more than one co-existing disorder (95% CI
7.6 to 56.3).
Due to the impulsivity,
children with ADHD have increased diffi-
culty controlling anger or aggressive out-
Because of comorbid conditions,
management of ADHD may require more
than one treatment modality.
Atomoxetine is a selective inhibitor of
norepinephrine re-uptake in the brain,
licensed for the treatment of ADHD.
four double-blind, placebo-controlled clin-
ical trials leading to Food and Drug
Administration approval in the USA,
,70% of children and adolescents
responded to atomoxetine by reductions
in scores on a variety of measures for
ADHD symptomatology.
No major risks were identified in these
trials, but a small percentage of subjects
seem to have experienced mood destabili-
Four subjects discontinued the
drug because of irritability or aggression.
There is a paucity of studies addressing the
use of atomoxetine and aggression/hosti-
lity as an adverse effect. Randomised
controlled trials have only limited ability
to detect rare side effects. The review by
Polzer et al
has only about 40% power to
detect a true difference in rates of aggres-
sion/hostility between atomoxetine and
placebo, and only about 15% power to
detect a difference between atomoxetine
and methylphenidate use. As such, the
present data cannot exclude the potential
side effect of aggression with atomoxetine.
Arch Dis Child Educ Pract Ed August 2008 Vol 93 No 4 131 on March 18, 2013 - Published by ep.bmj.comDownloaded from
doi: 10.1136/adc.2008.142521
2008 93: 129-131Arch Dis Child Educ Pract Ed
M Jacobs and A Reece
Updated information and services can be found at:
These include:
This article cites 14 articles, 1 of which can be accessed free at:
Email alerting the box at the top right corner of the online article.
Receive free email alerts when new articles cite this article. Sign up in
(22 articles)Diarrhoea (41 articles)Clinical trials (epidemiology) (204 articles)Child health (7 articles)Infection (gastroenterology) (10 articles)Foodborne infections
Articles on similar topics can be found in the following collections
To request permissions go to:
To order reprints go to:
To subscribe to BMJ go to: on March 18, 2013 - Published by ep.bmj.comDownloaded from
ResearchGate has not been able to resolve any citations for this publication.
This practice parameter formulates recommendations for health care providers about the management of acute diarrhea in children ages 1 month to 5 years. It was developed through a comprehensive search and analysis of the medical literature. Expert consensus opinion was used to enhance or formulate recommendations where data were insufficient. The Provisional Committee on Quality Improvement of the American Academy of Pediatrics (AAP) selected a subcommittee composed of pediatricians with expertise in the fields of gastroenterology, infectious diseases, pediatric practice, and epidemiology to develop the parameter. The subcommittee, the Provisional Committee on Quality Improvement, a review panel of practitioners, and other groups of experts within and outside the AAP reviewed and revised the parameter. Three specific management issues were considered: (1) methods of rehydration, (2) refeeding after rehydration, and (3) the use of antidiarrheal agents. Main outcomes considered were success or failure of rehydration, resolution of diarrhea, and adverse effects from various treatment options. A comprehensive bibliography of literature on gastroenteritis and diarrhea was compiled and reduced to articles amenable to analysis. Oral rehydration therapy was studied in depth; inconsistency in the outcomes measured in the studies interfered with meta-analysis but allowed for formulation of strong conclusions. Oral rehydration was found to be as effective as intravenous therapy in rehydrating children with mild to moderate dehydration and is the therapy of first choice in these patients. Refeeding was supported by enough comparable studies to permit a valid metaanalysis. Early refeeding with milk or food after rehydration does not prolong diarrhea; there is evidence that it may reduce the duration of diarrhea by approximately half a day and is recommended to restore nutritional balance as soon as possible. Data on antidiarrheal agents were not sufficient to demonstrate efficacy; therefore, the routine use of antidiarrheal agents is not recommended, because many of these agents have potentially serious adverse effects in infants and young children. This practice parameter is not intended as a sole source of guidance in the treatment of acute gastroenteritis in children. It is designed to assist pediatricians by providing an analytic framework for the evaluation and treatment of this condition. It is not intended to replace clinical judgment or to establish a protocol for all patients with this condition. It rarely will provide the only appropriate approach to the problem. A technical report describing the analyses used to prepare this parameter and a patient education brochure are available through the Publications Department of the AAP.
The practicing physician cannot help but be somewhat surprised by the current interest and publicity given to oral rehydration therapy for diarrheal disease. Indeed, oral rehydration therapy has been used to some extent by all physicians who deal with diarrhea, and the history of its use as a folk remedy is probably as long as the history of diarrheal illness. Why, then, has interest in this rather mundane therapy reemerged? Only recently have we begun to understand how oral fluids are absorbed, and this has resulted in changes in the composition and indications for use of these fluids. Even though the need for fluid intake during an episode of diarrhea has appeared always to have been a part of folk medicine, the medical profession did not consider this practice until the early 19th century. In 1832, after William O'Shoughnessy, an Irish physician, described the chemical composition of the stools in cholera, Thomas Latta of Scotland attempted to treat cholera by the intravenous infusion of water and salts. Of the 15 cases he reported in The Lancet, five patients survived. Latta was criticized severely for this therapy, but it was pointed out (in discussion in The Lancet) that these five patients were saved from almost certain death.
Twenty-nine dehydrated, well-nourished infants, who were 3 to 24 months of age and had acute gastroenteritis, were enrolled in a prospective randomized study that compared the safety, efficacy, and costs of oral vs intravenous rehydration. The study was designed to assess the use of a holding room in the emergency room for the outpatient rehydration of dehydrated infants. The oral solution that was used contained 60 mEq/L of sodium, 20 mEq/L of potassium, 50 mEq/L of chloride, 30 mEq/L of citrate, 20 g/L of glucose, and 5 g/L of fructose. Thirteen of 15 patients were successfully rehydrated orally as outpatients; two patients, who were subsequently discovered to have urinary tract infections, required hospitalization due to persistent vomiting. Orally rehydrated outpatients spent a mean of 10.7 hours in the holding room, as compared with intravenously rehydrated inpatients, who were hospitalized for a mean of 103.2 hours. Outpatient oral rehydration therapy was significantly less costly than inpatient intravenous therapy (+272.78 vs +2,299.50). Our results indicate that oral rehydration is a safe and cost-effective means of treating dehydrated children in an outpatient setting in the United States. The use of a holding room for observation in the emergency room can markedly decrease health care costs and unnecessary hospitalizations.
Analysis of 91 commercial dietary "clear" fluids, including carbonated beverages, cordials, powdered drinks, jellies, fruit juices, fruit juice drinks, syrups, and soups showed a range of zero to 175 mmol/L of sodium, zero to 52.1 mmol/L of potassium, zero to 839 mmol/L of reducing sugars and an osmolality of 50 to 914 mmol/kg water. Home-made oral rehydration solutions prepared by a group of mothers and medical staff members also showed unacceptably wide variability in their composition. The composition of products specifically indicated for the treatment of diarrhoea was also remarkably diverse, with a range of 24 to 100 mmol/L of sodium, 14.2 to 21.0 mmol/L of potassium, 45 to 326 mmol/L of glucose, and 170 to 460 mmol/kg water osmolality. On the basis of our current knowledge of water and electrolyte absorption, neither the commercial dietary "clear" fluids nor the home-made solutions can be recommended for the treatment of diarrhoea. Of the therapeutic products, only "Unicef Oral Rehydration Salts" has been subjected to clinical trails and found safe and effective. However, the formulation of Diolyte appears appropriate for the treatment of mild diarrhoeal dehydration.