Gastroenteritis in children: should we be sending them home with ondansetron?

This is a St Emlyn’s journal club post on paediatric emergency medicine. We have recently also covered new research on status epilepticus and abdominal trauma in children.

Autumn is here, and winter is on its way. As a registrar in the paediatric emergency department (PED) my busiest season is about to begin. The good children of Virchester will soon be flocking to our hospital with their coughs, sniffles, and wheezes. An unlucky few will also have diarrhea and vomiting. For some children, these gastrointestinal symptoms can be more distressing than pain.

Vanishingly few of the vomiting children I see at St Emlyn’s are truly dehydrated. Most can be discharged with safety netting and advice to use oral rehdyration solutions or — equally good! — dilute fruit juice. A minority will require treatment with an anti-emetic. For me, these are the children who are unwell, or have co-morbidities (such as chronic kidney disease) that would be exacerbated by dehydration. They are usually flagged up by our brilliant nurses at triage. If anti-emetics are unsuccessful, the next step is hydration via a nasogastric tube, with intravenous fluids as a “last resort” for the sickest cases.

Ondansetron is the drug of choice for vomiting children. Given in PED, ondansetron has been shown in multiple trials to improve oral intake, preventing the need for intravenous fluids and hospital admission. It appears to be more effective than other anti-emetics (e.g. domperidone) with a very low risk of cardiac side-effects, such as QT prolongation. It may also reduce diarrhea.

What is unclear, currently, is the role of ondansetron after emergency care. Should I be prescibing take-home doses for children who improve in the department? There is trial evidence to support this practice in out-of-hours primary care but I struggled to find much else in the literature. One recent-ish trial conducted in Turkey found some benefit from giving children (n=109) a second dose of ondansetron prior to leaving PED. But this does not really answer my question.

Fortunately, a large trial on this topic was published this year in the New England Journal of Medicine. The aim of this study was to see if a short course of oral ondansetron at home could improve symptoms and prevent “bounce-backs” to emergency care in children with gastroenteritis.

Abstract

Background & Objectives: Ondansetron improves outcomes when administered in emergency departments to children with acute gastroenteritis–associated vomiting. It is commonly prescribed at discharge to reduce symptoms, but evidence to support this practice is limited.
Methods: We conducted a double-blind, randomized superiority trial involving children 6 months to less than 18 years of age with acute gastroenteritis–associated vomiting in six pediatric emergency departments. Caregivers were provided with six doses of oral ondansetron or placebo to administer in response to ongoing vomiting during the first 48 hours after enrollment. The primary outcome was moderate-to-severe gastroenteritis, defined by a score of 9 or higher on the modified Vesikari scale (scores range from 0 to 20, with higher scores indicating greater severity), during the 7 days after enrollment. Secondary outcomes included the presence of vomiting, the duration of vomiting (defined as the time from enrollment to the last vomiting episode), the number of vomiting episodes within 48 hours after enrollment, unscheduled physician visits within 7 days after enrollment, and receipt of intravenous fluids.
Results: A total of 1030 children underwent randomization. Moderate-to-severe gastroenteritis occurred in 5.1% (23 of 452 participants for whom data were available) in the ondansetron group and 12.5% (55 of 441) in the placebo group (unadjusted risk difference, −7.4 percentage points; 95% confidence interval [CI], −11.2 to −3.7). After adjustment for site, weight, and missing data, ondansetron was associated with a lower risk of moderate-to-severe gastroenteritis than placebo (adjusted odds ratio, 0.50; 95% CI, 0.40 to 0.60). Although we did not observe any meaningful difference between the groups in the presence or median duration of vomiting, the total number of vomiting episodes within 48 hours after enrollment was lower with ondansetron than with placebo (adjusted rate ratio, 0.76; 95% CI, 0.67 to 0.87). The percentage of children who had unscheduled health care visits and the percentage who received intravenous fluids after enrollment did not differ substantially between the groups. The incidence of adverse events also did not differ meaningfully between the groups (odds ratio, 0.99; 95% CI, 0.61 to 1.61).
Conclusions: Among children with gastroenteritis-associated vomiting, the provision of ondansetron after an emergency department visit led to lower risk of moderate-to-severe gastroenteritis during the subsequent 7 days than the provision of placebo.
Freedman SB, […], Wallace A. Multidose ondansetron after emergency visits in children with gastroenteritis.
New England Journal of Medicine. 2025 Jul.

What was the study design?

This was a randomised controlled trial. Patients were recruited from six PEDs in Canada between 2019 and 2024. They were all tertiary centres.

Can you tell me about the patients?

The researchers aimed to recruit children (<18 years) with a clinical diagnosis of gastroenteritis. To be eligible for the trial, they needed to have received ondansetron in ED for treatment of vomiting.

The exclusion criteria included QT prolongation, ventricular arrythmia, and congenital cardiac disease. Patients were also excluded if they showed signs of bloody or bilious vomiting.

In total, 2431 children were screened and 1029 recruited into the trial. The majority were aged 3-4 years, which is fairly typical for gastroenteritis. The PED visit generally occurred in the early phase (first ~14 hours) of illness.

What was given in each trial arm?

All participants had received ondansetron in PED prior to discharge. This was given as an oral solution. The trial was unable to provide the orodispersible (“wafer”) form of the drug.

In the intervention group, caregivers were provided with further six doses of ondansetron (0.15mg / kg) to use as-required in the 48 hours after randomisation. For the control group, six doses of a volume-matched placebo (saline) solution were dispensed. Neither the caregivers nor the outcome assessors were aware of treatment allocation.

What outcome measures were used?

The primary end-point was the presence or absence of “moderate-to-severe” gastroenteritis. The researchers defined this as a score of 9 or more on a modified version of the Vesikari scale at seven days post-randomisation.

I had never heard of the Vesikari scale prior to reading this trial, and even after reading through the supplementary information, I am not sure that I fully grasp how it was used here. It seems that seven domains were involved:

Duration of diarrhea after ED discharge
Duration of vomiting after ED discharge
Number of watery stools
Number of vomiting episodes
Max recorded temperature
Unscheduled healthcare visits
Requirement for intravenous fluids or hospitalisation

Each of the above were rated 0-3 for a total score range of 0-21. Any combination of scores could bring participants to the threshold (>9) for the primary end-point.

Secondary outcomes included caregiver satisfaction and requirement for intravenous fluids or further healthcare usage.

What were the main results?

Patients who received ondansetron had a lower incidence of “moderate-to-severe” gastroenteritis: 5.1% had a Vesikari score of 9 or more, as opposed to 12.5% in the control group. The absolute risk reduction was 7.4% (95% CI -11.2 to -3.7) so for every thirteen children given ondansetron, one avoided the outcome.

Confusingly, no significant difference was seen between groups on the total Vesikari score: the mean score actually favoured the placebo group (-0.47) with a 95% confidence interval crossing zero (-0.99 to 0.06).

Ondansetron use was not associated with an actual improvement in the incidence or duration of vomiting. Caregiver satisfaction was unchanged, and there was no reduction in healthcare service use or intravenous fluid administration.

What should we take away from this study?

This was a large, multicentre, randomised trial. The researchers utilised a safe, inexpensive medication at a dose (0.15mg/kg) commonly prescribed to children. The participants were more medically complex than might be expected outside of academic centres, but otherwise, I would say that they were fairly typical of the patients I see on the shop floor. We are not going to get better data than this.

That said, I do disagree with the authors’ conclusion that ‘provision of ondansetron […] was beneficial’ for children with acute gastroenteritis. In fact, I think that they have published a negative trial. This is a great paper for critical appraisal, so if you have a journal club at your centre, make sure to discuss it — and let me know what you think! I am going to be quite critical, because there is a high chance that this study will influence future departmental (even national?) guidelines on gastroenteritis. It has already created a stir on social media, with some pretty sensational headlines knocking around.

To recap: a modest difference was found in the occurrence of ‘moderate-to-severe’ symptoms in the ondansetron group, which was not reflected in the mean Vesikari scores. Nor was caregiver satisfaction, which should, according to the authors’ protocol, have correlated with these ratings. Ondansetron had no effect on the incidence of vomiting post-discharge, or the amount of vomiting in children with persistent symptoms.

It is hard to make sense of these discrepancies. The simplest explanation to me is that the clustering of Vesikari scores in favour of ondansetron occurred items on the scale unrelated to vomiting. There are perhaps other reasons, including, of course, random chance. But even if we take the primary findings at face value, it is not clear to me that they demonstrate benefit. When I am working in PED, I do not prescribe ondansetron to stop children from vomiting. Rather, I prescribe it to prevent them from becoming dehydrated. In this trial, there were no differences found between groups on healthcare service use or intravenous fluid administration. This would suggest to me that ondansetron did not prevent the higher-risk children from deteriorating at home.

Ed — others may argue that simply preventing vomiting, irrespective of hydration, is an important outcome, and not being puked on, or puking, is a good thing — but that was not really the outcome of this study…

Something else that concerns me in this trial is the absence of specific child-oriented outcomes. Could the authors have measured the distress associated with vomiting? Or nausea, which can be equally as miserable? It might be that ondansetron is great for improving children’s symptoms between vomiting episodes, but we really do not know. This, to me, is an example of why it is so important to include the child’s voice in emergency care research.

Should this study change our practice?

No.

I rarely prescribe ondansetron as a take-home medication. I cannot see anything in this trial to change my mind. I appreciate that my experiences may not be shared by clinicians working in remote or low/middle-income countries, whose local population and pathogens are likely to be different.

A brilliant point made by Natalie May — that I failed to consider when I first read the trial! — is that in some ways, we want children to return to PED if their symptoms are not resolving. There is a diagnostic “test of time” for gastrointestinal symptoms, and discharging children with anti-emetics can interfere with this. If the vomiting is getting worse at home, it might be time for a re-assessment in PED. Has an appendicitis been missed, for example?

On the whole, I would urge caution in over-medicalising UK children with gastroenteritis. I do not believe that it helps. Our focus should be on empowering families to manage their child’s hydration for themselves. If you want to discharge your patient with something, give them an ice lolly.

Greg Yates