Introduction

Acute gastroenteritis (AGE) is the main cause of acute vomiting in children under the age of 3 years and one of the most important reasons for admission to the pediatric emergency department (ED) and the hospital [1,2]. In the USA, 1.5 million children under 5 years are diagnosed with AGE annually and this condition accounts for 13% of all hospital admissions [1]. The most frequent complication is dehydration. In Europe, at least 230 deaths and over 87,000 hospitalizations of children under 5 years of age are reported every year.[3]

In the initial phase of AGE, vomiting is reported in 75% of children with rotavirus infection [4], and is distressing for both patients and their families. Vomiting is a direct cause of fluid loss and can also hamper successful treatment with oral rehydration solution (ORS). Symptomatic pharmacological treatment for vomiting is still a matter of debate and is not systematically included in current practice recommendations for pediatric AGE [5–7]. Physicians and parents in ED favor intravenous fluid therapy (IVT) for mild or moderate dehydration when vomiting is the major symptom [8, 9]. Thus, effective antiemetic treatment would lead to an important reduction in the use of IVT.

Various antiemetic agents are available and are often used off-label to prevent or reduce vomiting in children with AGE [10, 11]. In France, Spain, Italy and in other European countries, the dopamine receptor antagonist domperidone is the preferred antiemetic treatment [12]. Ondansetron is administered to only a small proportion of children and its use varies significantly among institutions [13, 14].

Literature evaluating the efficacy of symptomatic drugs in reducing acute vomiting for pediatric AGE focuses mainly on ondansetron [15–19]. Evidence exists that ondansetron compared with placebo increases the proportion of patients with cessation of vomiting, reduces the immediate hospital admission rate and the need for IVT. However, not all of these studies evaluate first-line oral rehydration therapy (ORT) during hospital stay before the administration of the antiemetic [18], and an adequate comparative evaluation between domperidone and ondansetron is missing [4, 20]. Concerning the use of domperidone, only few studies are available with small sample sizes, low methodological quality, and inconsistent results [4, 15, 17, 18–23].

The aim of the current trial was to assess whether the oral administration of ondansetron vs domperidone or placebo, after a first attempt with ORS, prevents IVT or nasogastric rehydration in children with vomiting during AGE.

Methods

Results

Participants were recruited between July 7, 2011 (first randomization) and November 3, 2013 (last follow-up). A total of 1438 children with AGE who had accessed the 15 EDs were assessed, 1313 were eligible for the first ORT attempt which was successful for 832 (63.4%). Of the remaining 481 children, 125 (25.9%) were excluded because parents or guardians did not give their consent while 356 were randomly assigned to the study groups: 119 to domperidone, 119 to ondansetron and 118 to placebo (Fig 1). The baseline characteristics of the groups were similar (Table 1).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Study flow chart: assessment, randomization, and follow-up.

https://doi.org/10.1371/journal.pone.0165441.g001

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Baseline Characteristics of the Study Patients.

https://doi.org/10.1371/journal.pone.0165441.t001

All the randomized subjects received the first dose of the study medication. For the majority of patients (315/356, 88.5%) vomiting did not interfere with the first administration, whereas 22 patients receiving domperidone (18.5%), 8 ondansetron (6.7%) and 11 placebo syrup (9.3%) needed a second dose within 15 minutes after the first one. Five children (four in the domperidone and one in the ondansetron group) vomited immediately after the second dose. Three of these received one dose of open-label ondansetron, and two were treated with intravenous rehydration. According to the intention to treat approach, these children were considered to belong to their respective randomization group (Fig 1). The average amount of syrup was 3.8 ml (SD 1.2) for ondansetron, 3.8 ml (SD 1.1) for domperidone, and 3.9 ml (SD 1.0) for placebo (p = 0.66).

No patient received nasogastric rehydration. Intravenous treatment was administered to 78 out of 356 children (21.9%): 14 in the ondansetron group (11.8%), 30 (25.2%) in the domperidone group, and 34 (28.8%) in placebo group. Ondansetron led to a relative risk reduction of IVT by over 50%, both vs domperidone (RR 0.47, 98.6% CI 0.23 to 0.97) and placebo (RR 0.41, 98.6% CI 0.20 to 0.83) (Table 2). The number needed to treat (NNT) for ondansetron vs domperidone was 8 (95% CI from 5 to 28) and 6 (95% CI from 4 to 15) vs placebo.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Outcome measures during Emergency Department stay.

https://doi.org/10.1371/journal.pone.0165441.t002

A stratified analysis showed no effect of the dehydration status (no vs mild to moderate dehydration) on the primary study outcome.

Ondansetron halved the risk of the observation stay exceeding 6 hours versus both domperidone and placebo while the number of patients admitted to hospital from the ED was similar in all three groups (Table 2). Ondansetron significantly reduced the number of subjects with episode of vomiting during ED stay and led to greater success of the second attempt with ORS vs domperidone and placebo; furthermore, ondansetron significantly reduced the need for laboratory tests compared with placebo (Table 2). No statistically significant difference was seen for the occurrence of diarrhea while the median number of diarrhea episodes was higher in the ondansetron group (2.0 episodes vs 1.0 in domperidone and 1.5 in placebo group).

After discharge, no statistically significant differences were seen among the three groups for subjects readmitted in ED for the same illness, subjects with episodes of vomiting and diarrhea in the 48 hours follow-up and number of episode of vomiting and diarrhea in the last 24 hours of follow-up (Table 3).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Outcome measures at telephone follow-up.

https://doi.org/10.1371/journal.pone.0165441.t003

No serious adverse event was observed. A total of 13 patients had one mild adverse effect: 6 after ondansetron, 5 after domperidone, and 2 after placebo. Episodes of drowsiness, asthenia, irritability, diarrhea or abdominal pain were common to all the three investigated groups (S3 Table). In one case (ondansetron group) the blinding was opened following an adverse event.

Discussion

Our study indicates a 63.4% success rate of the first attempt with ORS in over 1300 children with AGE without severe dehydration. This means that, in an ED setting, 6 out of 10 children aged 1–6 years with vomiting due to AGE and no or mild to moderate dehydration, can be successfully treated with oral rehydration solution alone, without the need for drugs. This finding is consistent with the estimates of the Cochrane review [17, 18].

In children who continue to vomit after the first ORT attempt, a single oral dose of ondansetron improves the chances of success of ORT. Ondansetron reduces by over 50% the number of patients requiring IVT vs both domperidone and placebo (Table 2), in agreement with the results of other RCTs [15–19, 25]. Our results provide clear evidence of benefit of ondansetron also with respect to the other study outcomes. Hospital admission rates are lower in the ondansetron group vs both domperidone and placebo. Differences among groups did not reach the statistical significance of the meta-analysis (RR 0.41; 95% IC 0.29 to 0.59) [18]. In the present study, the need for observation stay to last more than six hours is statistically significantly lower in the ondansetron group compared with the domperidone and placebo groups.

In agreement with the results of other RCTs, no difference in the percentage of patients readmitted to the ED within 48 hours of discharge was seen among the three groups. This percentage is lower than the one reported in Freedman's RCT and similar to that reported in another study [25, 28].

Only a few mild adverse events occurred after ondansetron administration, all of them resolved quickly without any consequence for the children. In particular, although the ondansetron group presented a higher median number of episodes of diarrhea (one additional episode on average), this increase is smaller than the one described in other RCTs [17, 18], and seems to have no clinical relevance, especially when weighed against the drug’s significant effect on reduction of vomiting. Unfortunately, our study does not have the statistical power to detect rare but serious adverse events, such as cardiac arrhythmias, and further studies, i.e. post-marketing surveillances, should be carried out to address this issue. Although outside the context of diarrhea, the FDA black box alert published in September 2011, recommends electrocardiogram monitoring in patients with potential “electrolyte abnormalities” receiving ondansetron, due to the risk of developing prolongation of the QT interval which can lead to an abnormal and potentially fatal heart rhythm, including Torsade de Pointes [4]. However, there is evidence that routine ECG and electrolyte screening before the administration of a single dose of oral ondansetron in children without known risk factors (i.e. history of arrhythmias, concomitant use of QT-prolonging drugs) are not necessary [29].

Domperidone did not appear to be superior in any of the primary and secondary outcomes (Table 2). The available evidence of the efficacy of domperidone consist of few studies with small sample sizes, low methodological quality and inconsistent results [4, 15, 17–23]. The only randomized trial comparing oral domperidone with placebo showed that the drug, used in combination with ORT, does not reduce vomiting in the early stages of AGE [23].

We chose to include domperidone in our RCT since it is commonly prescribed to children with gastroenteritis in several countries, including Italy [10–12, 23, 30], and is licensed in Europe for the “treatment of nausea and vomiting”, also in the pediatric population, despite the lack of evidence on efficacy. Recently the authorization for the use of domperidone for these clinical conditions has been subjected to restrictions because of the possible risks of severe arrhythmias, particularly in the case of drug overdose [4, 31]. Furthermore, case reports and post marketing surveillances have reported the occurrence of extrapyramidal reactions associated with the use of domperidone [32, 33].

The main limitation of our study is the premature closure of the enrollment, and the consequent failure to reach the initially estimated sample size. This has already been described for publicly-funded trials in the UK [34]. In our study this was due to the success of the first ORT which made it difficult to have patients to randomize and to the non-occurrence of the annual rotavirus diarrhea outbreak during the second year of the study. Furthermore, in the absence of literature evidence, the sample calculation of the efficacy of domperidone was estimated to be intermediate between ondansetron and placebo, but the study findings did not confirm this hypothesis. However, our RCT is the largest carried out on this topic so far.

The use of placebo in our RCT could be questioned. When the study started, available evidence, albeit weak and unreliable [17], suggested the efficacy of ondansetron for symptomatic treatment of vomiting during AGE in children but a proper evaluation of domperidone, largely used in clinical practice, was lacking. Furthermore, at the time, no formal indications for the use of ondansetron in the treatment of AGE was given by clinical practice guidelines. Consequently, and in accordance with the document on ethical considerations for clinical trials in the pediatric population [35], we felt that the use of placebo was legitimate. This matter was also discussed by the Bioethic Committee of the coordinating centre before the approval of the study protocol.

Our trial included a large number of patients with a limited age range (1–6 years), with no or mild to moderate dehydration. This reflects the actual clinical setting in Italy. The selected age range allowed us to enroll most of the children with community acquired AGE, which is prevalent in younger children, and to ensure an adequate safety profile, since ondansetron is approved for vomiting (after chemotherapy) in children over 6 months. This study contributes to provide evidence for better management of this young population characterized by high incidence of AGE associated to dehydration, and scanty evidence of effective therapeutic approaches.

Our study presents several strengths. First, the complete independence of the study, which was made possible by the public funding received from AIFA. Second, the initial use of ORS before administering the drugs and the inclusion in the trial only of children who had failed this first attempt. This made it possible to adequately evaluate the role of the drugs under study and to confirm in a field study the role and applicability of ORT in children with AGE without severe dehydration.

The results of our study have relevant implications for practice. Our trial fully confirms the results of the most recent systematic review on the use of ondansetron which suggests that the clinical practice guidelines for the treatment of children with AGE should be revised to include the use of a single dose of oral ondansetron in the case of continued vomiting after a first attempt with ORS [18]. This simple intervention reduces the need for IVT and laboratory tests, the discomfort of vomiting and the time spent in the ED. Currently, some guidelines, in particular those from North America [7], already suggest the use of ondansetron in pediatric emergency departments in infants and children between six months and 12 years of age, while others, including ESPGHAN and NICE [4, 6], are more conservative as a result of the FDA warning on the potentially severe adverse effects of ondansetron.

In view of the ineffectiveness demonstrated in the present trial, the inconclusive results of previous studies and the possible side effects reported in literature, domperidone should not be used for the symptomatic treatment of vomiting as a consequence of AGE in children.

The aims of our work did not include a pharmacoeconomic analysis, but a previous study showed that the administration of oral ondansetron to children with dehydration and vomiting secondary to AGE led to significant monetary savings compared to a no-ondansetron policy [36]. Appropriate strategies are needed to successfully incorporate oral ondansetron into clinical practice in order to maximize its potential benefits [37]. Further studies are needed to understand if, in the real context of ED care, the use of ondansetron in children at high risk of dehydration can effectively reduce the number of cases receiving IVT. Indeed, despite the increasing use of ondansetron over the years in the United States and Canada, the percentage of children requiring IVT doesn’t seem to have decreased [38].

In conclusion, our trial showed that, in a context of emergency care, 6 out of 10 children aged 1–6 years with vomiting due to AGE and without severe dehydration can be managed effectively with the administration of ORS alone. In children who continue to vomit or refuse ORT, a single oral dose of ondansetron reduces the need for IVT, the percentage of children who vomit and the number of episodes of vomiting, thereby facilitating the success of ORT. A small, not clinically relevant, increase in the number of episodes of diarrhea was observed in the ondansetron group. Domperidone was not effective for the symptomatic treatment of vomiting during AGE.

Supporting Information

Acknowledgments

The SONDO (Study ONdansetron vs DOmperidone) Investigators were:

Institute for Maternal and Child Health—IRCCS "Burlo Garofolo”, Trieste, Italy: Alessia Fratte, Alessandra Knowles, Marta Massaro, Gianni Messi, Rossella Paparazzo, Elisa Zanelli;

Pediatric Emergency Department, Presidio Ospedale dei Bambini, A.O. Spedali Civili, Brescia, Italy: Chiara Donati, Annamaria Sorlini;

Pediatric Emergency Department, Azienda Ospedaliera Universitaria Integrata, Verona, Italy: Linda Balanzoni, Davide Silvagni;

Department of Pediatrics, Treviso Hospital, Treviso, Italy: Paola Berlese, Mariaelena Cavicchiolo;

Department of Pediatrics, Ospedale Castelli, Verbania, Italy: Paola Angellotti, Luigina Boscardini;

Department of Pediatrics, Ospedale Maggiore, Bologna, Italy: Andrea Lambertini, Chiara Landini;

Emergency Department, Pediatric Hospital A. Meyer, Firenze, Italy: Francesco Mannelli, Simona Montano;

Department of Pediatrics, Ospedale di Macerata, Macerata, Italy: Roberta Piccinini, Luciana Maria Taccari;

Emergency Department, Ospedale Pediatrico Bambino Gesú, IRCCS, Roma, Italy: Francesca Faa;

Emergency Room and Emergency Medicine Division, G. Gaslini Institute, Genova, Italy: Laura Bergamino, Pasquale Di Pietro;

Emergency Department, Ospedale Infantile Regina Margherita, Torino, Italy: Francesca Bosetti, Emanuele Castagno;

Department of Pediatrics, Ospedale G.B. Morgagni—L. Pierantoni, Forlì, Italy: Martina Fornaro, Benedetta Mainetti;

Department of Pediatrics and Pediatric Emergency, "San Giuseppe Moscati" National Hospital (AORN), Avellino, Italy: Maria Grazia Scalone, Giovanna Roberta Vega;

Pediatric Emergency Department, Azienda Ospedaliera—University of Padova, Padova, Italy: Susanna Masiero, Stefania Scanferla.

We acknowledge the statistical support provided by Dr. Marcella Montico for the interim analysis.