Intended for healthcare professionals

Rapid response to:

Practice Guidelines

Diarrhoea and vomiting caused by gastroenteritis in children under 5 years: summary of NICE guidance

BMJ 2009; 338 doi: (Published 22 April 2009) Cite this as: BMJ 2009;338:b1350

Rapid Response:

Diarrhoea and vomiting caused by gastroenteritis: clinical guideline--a critique

Diarrhoea and vomiting caused by gastroenteritis: clinical guideline for the NHS (April 2009)--a critique

Treatment of diarrhoea should rest on understanding of the underlying
pathophysiology: evidence-based medicine is essential to confirm the
validity of such an approach, not to replace it, especially when the
available evidence is fallible. O’Shaugnessy[1,2], who pioneered the use of
rationally based oral and parenteral fluid therapy for acute diarrhoea in
1832 appreciated the importance of treating what we now call acidosis; the
new Clinical Guideline on the treatment of diarrhoea commissioned by NICE
[RCOG Press, April 2009] seems not to and that is not its only defect.

The fundamental nature of diarrhoea
The Guideline includes several clinical definitions of diarrhoea. An
additional pathophysiological definition would illuminate the rationale
for oral rehydration therapy [ORT]: diarrhoea is fundamentally a failure
of enteric salt and water uptake sufficient to overwhelm the compensatory
capacity of the colon [3]. A well formulated oral rehydration solution
[ORS], has an appropriate sodium:glucose ratio for intestinal cotransport,
promoting intestinal sodium and water uptake, so attacking the underlying
problem; it is not simply symptomatic treatment.

Metabolic acidosis
With mild dehydration, restoration of extracellular fluid [ECF] volume,
through improved renal perfusion, may correct acidosis without provision
of bicarbonate precursor. In severe diarrhoea, bicarbonate precursor is
important in both oral and parenteral therapy- as already noted by
O’Shaugnessy by 1832. Indeed, the lethal potential of acidosis was
expounded 90 years ago by the American humorist H. L. Mencken ‘Life is a
struggle, not against sin, not against money power, but against hydrogen
ions…The dying man doesn’t struggle much and he isn’t much afraid. As his
alkalies give out he succumbs to a blest stupidity. His mind fogs. His
will power vanishes. He submits decently. He scarcely gives a damn.’ [3]

Neglecting acidosis in discussing the treatment of severe diarrhoea,
however, is unblest stupidity; the Guideline scarcely gives a damn; it
advocates saline for ECF volume replacement. Saline is appropriate
treatment for metabolic alkalosis; it can not also be correct for
acidosis. What evidence establishes that rapid i/v rehdyration with
saline, 40 ml/kg is appropriate even with shock [p72] ? With both severe
diarrhoea and shock, metabolic alkalosis is improbable: acidosis, with its
associated hyperkalaemia, is potentially dangerous. The suggested dose,
equivalent to 200ml of bicarbonate-free fluid per litre of ECF, would
dilute plasma bicarbonate eg from 12 mmol/l to 10 mmol/l, increasing base
deficit by 17%. Suggesting even 50-100 mmol/l could increase base deficit
by 50%. Although Table 5.4 [p68] includes 4 studies utilising lactated
Ringers, their efficacy in correcting acidosis is not discussed.

Sodium in ORSs
The sodium concentration of an ORS reflects the optimum for enteric
uptake, not faecal sodium loss or ‘normal’ intake. Clinical anxiety about
hypernatraemia rests substantially on the misapprehension that it is
caused by excess sodium intake: actually it reflects predominant water
loss or inadequate water intake. Every litre of ECF deficit inescapably
requires 140 mmol of sodium to correct it: the less per litre of ORS, the
more ORS will be needed unless, improbably, reduced sodium is outweighed
by enhanced absorbtion. The Guideline, commendably, acknowledges the
excessive preoccupation with hypernatraemia [p71]. Indeed, although the
differences are not statistically significant, the failure rate [need for
supplementation with i/v fluids] for low Na ORS [45mmol/l] was double that
with high Na [90 mmol/l] ORS [p62]. Nevertheless the verdict favours low
Na ORSs, essentially influenced by a Cochrane Review by Hahn et al [cited
as 2007, but analysed in detail by Michell in 2005[4];the essential data
remain unchanged, May 2009]. Of 13 studies providing this key evidence,
which has directed paediatric ORT towards hypotonic solutions, 9 aroused
serious reservations.

A glutamine-containing ORS has been used to treat neonatal diarrhoea in
calves for over 10 years, because it supports recovery of villus
architecture, whereas a conventional glutamine-free WHO-type ORS was
associated with deterioration of villus architecture[5-7].If children
responded similarly, glutamine would seem particularly important,
especially if they are already malnourished, before their diarrhoea.
Carneiro-Filhoe et al[8] refer to its ‘great potential in oral rehydration
for diarrhoea’ and a BMJ article in 2005[9] advocated research on this
solution for paediatric use. The Guideline has several recommendations
for research, but no mention of glutamine.

Issues for research
From what has already been said, these should include

• Would diarrhoeic children, especially those who are malnourished,
benefit from a glutamine-containing ORS ?

• Is saline, as opposed to lactated Ringers solution, appropriate i/v
therapy in the face of metabolic acidosis; if so, how severe can the
dehydration and the acidosis be for replacement of ECF volume with
solutions lacking a bicarbonate precursor to be effective in correcting
the acidosis ?

• Stool output is a notoriously fallible guide to the efficacy of
rehdration1 yet it is widely used as a criterion in judging ORSs; should
such papers be excluded from systematic reviews ?

Meta-analysis of fallible criteria
Meta-analysis of unreliable data can not create reliable conclusions. The
feasible measurements in paediatric cases of severe diarrhoea, especially
in epidemics in isolated communities, are very restricted. The fallibility
of clinical criteria in assessing severity of dehydration and success of
rehydration are well known[4] and reaffirmed in the Guidance [17, 42-48].
Even loss of bodyweight has pitfalls as an index of severity of
dehydration. Death can occur with relatively little loss of weight if
there is a huge accumulation of fluid within the gut, causing lethal
hypovolaemia before the output of fluid faeces can increase[10].

In contrast, ORT in calves rests on a line of research which defined
the therapeutic targets which demarcate likely death from likely
survival10, and then evaluated the beneficial effects of ORSs of different
composition on plasma and ECF volume, plasma Na, K and pH, GFR, PCV (not
just an index of hypovolaemia; in shock a raised PCV reminds us that blood
viscosity rises exponentially with PCV, further impeding tissue perfusion)
and villus architecture. It also corroborated the fallibility of stool
output as a guide to effective rehydration. This is not surprising; if
most of the ORS restores ECF volume, loss of the remainder as increased
faecal output is inefficient but clinically unimportant. Moreover faecal
output becomes normal before the small intestine, because of the
compensatory capacity of the colon. Calves are not children, but unlike
adult cattle, they are non-herbivorous, non-ruminant, functionally simple
stomached; prima facie they offer a valid model for the evaluation of ORT
in acute diarrhoea. Effectiveness of ORSs in improving plasma
electrolytes, pH, plasma volume and GFR in diarrhoeic calves reflected
adequate content of Na and bicarbonate precursor; improvement of pH and
GFR were also potentiated by glutamine.

This debate is not simply academic: for millions of children
throughout the world it is a matter of life and death. Ten thousand die of
this eminently treatable disease daily, mostly in poor countries: the
media are missing the real medical headlines. The burden of proof is not
whether calves are like children; since more reliable data are obtainable
from calves the burden of proof is to establish, with due precautions,
whether similar principles could save children’s’ lives[7].

A.R. Michell, Dept Biochemical Pharmacology, Harvey Institute, Barts
& the London School of Medicine & Dentistry, London EC1M 6BQ.

1 O’Shaugnessy. WB Lancet 1831; 1: 490.

2 O’Shaugnessy. WB Report on the chemical pathology of malignant cholera.
London: S.Wiley, 1832

3 Michell AR. The clinical biology of sodium Oxford, Elsevier. Chapter 3.

4 Michell AR. Why has oral rehydration for calves and children diverged:
direct vs indirect criteria of efficacy. Research in Veterinary Science
2005; 79:177-81.

5 Brooks HW, Hall, GA, Wagstaff AJ, MichellAR. Detrimental effects on
villus form and function during conventional oral rehydration for
diarrhoea in calves: . alleviation by a nutrient
oral rehydration solution containing glutamine. Veterinary Journal 1998; 155: 263-274.

6 Michell AR. Oral rehydration for diarrhoea: symptomatic treatment or
fundamental therapy Journal of Comparative Pathology 1998; 118: 175-193.

7 Michell AR. How may advances in oral rehydration therapy for animals
benefit children ? Pharmaceutical Journal 2004; 272: 580-581.

8 Carneiro-Filho BA, Bushen OY, Brito GA, Lima AA, Guerrant RL. Glutamine
analogues as adjunctive therapy for infectious diarrhoea. Current
Infectious Disease Reports 2003; 5: 114-119

9 Michell AR. Why shouldn’t children benefit from oral rehydration
solutions for calves ? British Medical Journal 2005; 331: 1267

10 Groutides CP, Michell AR. Changes in plasma composition in calves
surviving or dying from diarrhoea. British Veterinary Journal 1990; 146:

Competing interests:
Consultant on oral rehydration to veterinary pharmaceutical companies

Competing interests: No competing interests

08 May 2009
Alastair R Michell
Professor of Comparative Medicine [ Univ. of London]
London EC1 6BQ
Dept Biochemical Pharmacology, Harvey Institute, Barts Hospital