Page 81 - 15Diarrhoeaandvomiting

Diarrhoea and vomiting caused by gastroenteritis in children under 5 years

5.2.1 ORT versus IVT for children with all degrees and types of dehydration

Evidence overview
A high-quality Cochrane review compared the effectiveness of ORT with IVT for the treatment
83
of dehydration due to acute gastroenteritis in children. Altogether 17 trials were included
comparing an IVT arm with one or more ORT arms (oral or nasogastric). Nine of the trials were
conducted in high-income countries (six in the USA and one each in Canada, Australia and
Finland), one trial involved participants from both the USA and Panama, and the others were
conducted in relatively low-income countries. Most trials included children aged between
3 months and 5 years. One included children up to 17 years of age and three included newborn
babies younger than 14 days. All but two excluded children with hypovolaemic shock – in
one, children presenting with shock or severe dehydration were treated with initial IVT before
randomisation. Five trials excluded children with persistent vomiting, four included such cases,
and the remaining 11 did not provide any information on this matter. Overall, more children
were randomised to the ORT group (n = 1015) than to the IVT group (n = 796) because some
trials included more than one ORT arm.

All the included trials used ORS solution containing glucose or dextrose with sodium, potassium
and chloride, but the concentration of these constituents varied. In 14 trials, ORT was administered
by mouth but in four of these nasogastric tube administration was employed if necessary. In two
trials, ORS solution was given exclusively by nasogastric tube but in one of these the children had
previously failed to tolerate oral administration. In one trial, a combination of oral and nasogastric
administration was used. The primary outcome was failure to rehydrate but the definition of
failure varied between the studies. Secondary outcomes included weight gain at discharge,
incidence of hyponatraemia and hypernatraemia, duration of diarrhoea, total fluid intake and
total sodium intake at 6 and 24 hours. Safety outcomes included paralytic ileus, phlebitis, peri-
orbital oedema, abdominal distension and seizures. A meta-analysis was conducted using the
random effects model. [EL = 1++]

Randomisation was adequate in all but two of the trials. Most of the trials were small and of poor
quality. As double-blinding was not possible and arrangements for allocation concealment were
unclear in 16 trials, it is likely that the treatment effects could have been overestimated.
Children treated with ORT had a 4% higher risk of failure to rehydrate (using any definition)
compared with IVT, and this difference was statistically significant (18 trials; risk difference

(RD) 4%; 95% CI 1% to 7%) but with strong evidence of statistical heterogeneity (I² = 70%;
P < 0.001). When sensitivity analysis was performed using a homogeneous definition of ‘failure’
(limited to those with persistent vomiting, persistent dehydration and shock/seizures), the RD
was reduced to 2% with the lower limit of the 95% CI including the null value (RD 2%; 95% CI
0% to 4%). A subgroup meta-regression analysis was also performed using failure to rehydrate
as the dependent variable but no significant cause of heterogeneity was identified. Children
treated with ORT had a significantly shorter stay in hospital compared with those treated with
IVT (six trials; WMD −1.2 days; 95% CI −2.38 to −0.02 days) but again there was evidence
of significant heterogeneity. There were no statistically significant differences between the two
groups for the other outcomes – weight gain at discharge, mean duration of diarrhoea, incidence
of hyponatraemia or hypernatraemia, or the total fluid intake at 6 hours and 24 hours.
Regarding complications, the risk of phlebitis was significantly higher in the IVT group by 2%
(five trials; RD −2%; 95% CI −4% to −1%). More children in the ORT group developed paralytic
ileus although the difference was not statistically significant. There were no differences between
the two groups for the other complications and adverse effects – peri-orbital oedema, seizures or
abdominal distension.
A cumulative metagraph was developed (studies by ascending year) showing that the overall
estimate of failure was unlikely to change substantially with further trials. Additionally, the study
sample size (n = 1811) provided adequate power to support the observed results regarding failure
to rehydrate. However, the study lacked power to detect serious but rare adverse events in either
treatment group.

56