Alimentary Pharmacology & Therapeutics

Meta-analysis: ondansetron for vomiting in acute gastroenteritis in children

H. SZAJEWSKA, D. GIERUSZCZAK-BIAEK & M. DYLA G

The Second Department of Pediatrics, The Medical University of Warsaw, Warsaw, Poland Correspondence to: Dr H. Szajewska, The Second Department of Pediatrics [II Katedra Pediatrii AM], The Medical University of Warsaw, 01-184 Warsaw, Dzialdowska 1, Poland. E-mail: hania@ipgate.pl

SUMMARY Background Vomiting is a common sympton in children with gastroenteritis, but its treatment remains controversial. Aim To investigate potential benecial effects of ondansetron, compared with placebo or no intervention, in treating vomiting during acute gastroenteritis in children. Methods The following electronic databases were searched through August 2006: MEDLINE, EMBASE, CINAHL and The Cochrane Library; additional references were obtained from reviewed articles. Only randomized-controlled trials (RCTs) were included. Results Four RCTs involving 490 patients with vomiting during acute gastroenteritis were included. Combined data from three RCTs (n 466) showed that ondansetron compared with the control signicantly increased the chance for vomiting cessation soon after drug administration [relative risk (RR): 1.3, 95% condence interval (CI): 1.21.5, number needed to treat (NNT): 5, 95% CI: 48], but this effect was not observed at 24 h (three RCTs, n 144, RR 1.2, 95% CI: 0.91.7). Ondansetron signicantly reduced the risk of intravenous rehydration (two RCTs, n 359, RR 0.4, 95% CI: 0.30.7, NNT 7, 95% CI: 514). Outcome measures not signicantly different after ondansetron treatment were the need for hospitalization and return emergency department visits. Conclusions Despite some clinical benets, there is insufcient evidence to recommend the routine use of ondansetron for vomiting during acute gastroenteritis in children.
Aliment Pharmacol Ther 25, 393400

Publication data Submitted 20 October 2006 First decision 20 November 2006 Resubmitted 23 November 2006 Resubmitted 10 December 2006 Accepted 12 December 2006

2007 The Authors Journal compilation 2007 Blackwell Publishing Ltd doi:10.1111/j.1365-2036.2006.03231.x

393

394 H . S Z A J E W S K A et al.

INTRODUCTION
Vomiting is a common symptom in children with gastroenteritis, but its treatment remains controversial. In position papers, respected scientic societies or expert groups recommend that the use of antiemetics be avoided in young children with vomiting associated with acute diarrhoea because of potential troublesome side effects and questionable benet.13 Recent Cochrane review found that there is some evidence, albeit weak and unreliable, that antiemetics such as ondansetron (5-HT3 serotonin antagonist) and metoclopramide (dopamine antagonist), compared with placebo, reduce the number of episodes of vomiting due to gastroenteritis in children.4 However, the authors provided only a descriptive summary of included trials and abstained from performing a meta-analysis. Given this, as well as an interest of both doctors and carers in interventions that will increase the likelihood that oral rehydration therapy will be successful, we thought it is important to undertake a meta-analysis of all relevant studies available to date to clarify the benets, if any, of ondansetron. We report the results of this systematic review and meta-analysis.

Searches
We searched MEDLINE (1966 to August 2006), EMBASE (1980 to August 2006), Cumulative Index to Nursing and Allied Health (CINAHL; 1982 to August 2006), The Cochrane Database of Systematic Reviews (Issue 3, 2006) and The Cochrane Controlled Trials Register (Issue 3, 2006) for randomized-controlled clinical trials using the following text word terms and MESH headings: ondansetron, diarrhea/diarrhoea, diarrh*, gastroenteritis, vomit*. Furthermore, the reference lists from the original studies and review articles were identied. No limit was imposed regarding the language of publication, but certain publication types (i.e. letters to the editor, abstracts, proceedings from scientic meetings) were excluded.

Data extraction
Each author independently assessed the titles and abstracts of potential papers identied according to the above-described search strategy. All potentially relevant articles were retained and the full text of these studies examined to determine which studies satised the inclusion criteria. Data extractions were carried out independently by all reviewers, using standard data extraction forms. We compared the extracted data to identify errors. One reviewer (HS) entered the data into The Cochrane Review Manager [REVMAN (Computer program), Version 4.2 for Windows. Oxford, England: The Cochrane Collaboration; 2003] for analysis. In case of disagreement, data were read again to reach a consensus. All discrepancies between the reviewers were resolved by this method. Only the consensus data were entered.

OBJECTIVE
The objective of this study was to systematically evaluate the effect of ondansetron in treating vomiting during acute gastroenteritis in children.

METHODS Inclusion and exclusion criteria

We included randomized-controlled trials (RCT) in which the intervention was ondansetron compared with a placebo or no intervention for children with acute gastroenteritis. In the case of trials with more than one experimental arm, we included data on ondansetron only. The outcome measures were cessation of vomiting, need for intravenous rehydration, need for hospitalization, return visit to the emergency department, amount of oral rehydration solution (ORS) intake, length of stay in emergency department and adverse effects. In addition to these outcome measures, a priori we decided to extract other data reported by the investigators if we found that they addressed issues relevant to the current review. Only studies with >80% follow-up were included.

Study quality
The reviewers independently, but without blinding to the authorship or journal, assessed the included trials for: (i) allocation concealment, (ii) blinding of investigators, participants, outcome assessors and data analysts, (iii) intention-to-treat analysis and (iv) comprehensive follow-up.5

Statistical methods
The data was analysed using The Cochrane Review Manager. The binary measure for individual studies and pooled statistics was reported as the risk ratio (RR)
2007 The Authors, Aliment Pharmacol Ther 25, 393400 Journal compilation 2007 Blackwell Publishing Ltd

M E T A - A N A L Y S I S : O N D A N S E T R O N T R E A T M E N T O F V O M I T I N G I N C H I L D R E N 395

or risk benet (RB) between the experimental and control groups with 95% condence intervals (CI). Number needed to treat (NNT) was calculated as the inverse of the pooled absolute risk differences and 95% CI. The weighted mean difference between the treatment and control groups was selected to represent the difference in continuous outcomes. The weights given to each study were based on the inverse of the variance. Heterogeneity was quantied by v2 and I2, which can be interpreted as the percentage of the total variation between studies that is attributable to heterogeneity rather than to chance. A value of 0% indicates no observed heterogeneity, and larger values show increasing heterogeneity. For simplicity, we present results of only the xed effects (FE) model for all analyses, unless there was a heterogeneity in which case we present results of both random effects (RE) and FEs models. We took no formal steps to look for publication bias, such as by plotting effect sizes or by calculating test statistics. In most cases, there are few studies on any given effect, and any formal method would have had little power.

intravenous rehydration, but heterogeneity was signicant for the other outcomes (i.e. cessation of vomiting in 24 h, I2 66%, hospitalization, I2 52%). A summary of the outcomes is presented in Table 2.

Cessation of vomiting
Four RCTs69 reported data on cessation of vomiting at certain time intervals (Figure 1). Based on the results of three RCTs68 (n 466), the relative chance of cessation of vomiting in the emergency department in the ondansetron group compared with the control group was increased by 30% (RB: 1.3, 95% CI: 1.21.5). The NNT was 5 (95% CI: 48). However, there was no difference in the cessation of vomiting within 24 h (two RCTs,8, 9 n 144, RR: 1.2, 95% CI: 0.91.7, FE, and 1.6, 95% CI: 0.55, RE).

Intravenous rehydration
Pooled results of two trials6, 8 (n 359) showed a reduction in the risk of intravenous rehydration in those treated with ondansetron compared with control (RR 0.4, 95% CI: 0.30.7), NNT 7 (95% CI: 514; Figure 2).

RESULTS
Table 1 summarizes the characteristics of the included trials. Four RCTs69 involving 490 participants (247 in the experimental group and 243 in the control group) met our predened inclusion criteria. All were placebo-controlled and full peer-reviewed publications. Three studies were based in countries with a high Human Development Index (HDI;10 USA, Canada), and one study was based in a country with a medium HDI (Venezuela). The age of the participants varied from 1 month to 22 years. Moreover, the treatment protocol varied between studies. In two studies, ondansetron was given in a single intravenous dose (0.15 mg/kg7 or 0.3 mg/kg9), and in two trials, it was administered orally (28 mg depending on body weight, single dose,6 or 1.64 mg/kg depending on age, six doses).8 The aetiology of the acute gastroenteritis was provided in only one RCT.9 Allocation concealment was adequate in three studies.68 Double blinding was used in all studies. In addition, all investigators reported the proportion of patients who were randomized but lost to follow-up, which was <20% in all studies. No signicant heterogeneity was found for cessation of vomiting in the emergency department and
2007 The Authors, Aliment Pharmacol Ther 25, 393400 Journal compilation 2007 Blackwell Publishing Ltd

Hospitalization
Based on the pooled results of three studies68 (n 466), there was no signicantly reduced risk of hospitalization in the ondansetron group compared with the control group (RR: 0.6, 95% CI: 0.41.01, FE, and 0.6, 95% CI: 0.21.4, RE).

Return visit to emergency department

A meta-analyses of three trials68 (n 462) showed no reduction in return visits to the emergency department (RR: 1.3, 95% CI: 0.82.2).

Amount of ORS intake

The only RCT6 (n 214) that reported this outcome found increased intake of ORS in the ondansetron compared with the placebo group [mean difference 43 mL (95% CI: 15.570.5)]. The same study reported no signicant difference in the length of stay in the emergency department (mean difference )14 min, 95% CI: )29.6 to 1.6).

Table 1. Characteristics of included trials

Allocation concealment Blinding ITT FU Age 6 months 0.3 mg/kg, 8 years single dose, intravenous Dose Unclear Yes Yes 24/24 12/12 AGE with emesis (vomited twice within 1 h) N (experiment/ Inclusion control) criteria Control group Study denitions Aetiology of AGE

Author (country)

Generation of allocation sequence

396 H . S Z A J E W S K A et al.

Cubeddu et al.9 (Venezuela)

Unclear

Freedman et al.6 Adequate (Canada) Adequate Yes Yes 210/214 107/107 (on day 3) 206/214 (on day 7)

6 months AGE, 1 episode 10 years of non-bilious, non-bloody vomiting within 4 h preceding triage, 1 episode of diarrhoea, mild-to-moderate dehydration Vomited at least 6 months 5 times during the 12 years proceeding 24 h

2007 The Authors, Aliment Pharmacol Ther 25, 393400 Journal compilation 2007 Blackwell Publishing Ltd Yes No* 145 74/71

Ramsook et al.8 Adequate (random Adequate (USA) number allocation tables)

50% HRV, Placebo Emetic episode: 17% AV, a single vomit 25% bacteria or retch or any number of continuous vomits and/or retches Episode of vomiting: an expulsion of stomach contents Treatment failure: 2 emetic episodes within any 90-min period over 18 h following study drug administration or if they experiences 3 emetic episodes during the hour following the end of administration of the study treatment 28 mg, Placebo Proportion of children No data depending who vomited while on body receiving ORT weight, single Vomiting episode dose, orally forceful expulsion of stomach contents. Non-productive retching, spilling of oral contents, and drooling were not considered vomiting No data 1.64 mg/kg Placebo Frequency of emesis depending during the 48 h after on age, the enrolment 6 doses, Rates of intravenous orally administration Treatment failure any patient requiring admission

M E T A - A N A L Y S I S : O N D A N S E T R O N T R E A T M E N T O F V O M I T I N G I N C H I L D R E N 397
Aetiology of AGE

No data AGE, identied as 1 month 0.15 mg/kg Placebo Vomiting episode requiring 22 years (max. 8 mg), any episode of single dose, forceful expulsion intravenous uids, 3 episodes intravenous of stomach contents. Non-productive retching, of vomiting in the previous 24 h spilling of oral contents, drooling were not considered as vomiting episodes

Diarrhoea
Three studies (Table 3) provided data regarding the number of episodes of diarrhoea.6, 8, 9 As either mean values only (without the standard deviation of the outcomes) or no data were given, pooling of the results was not feasible. Therefore, we report the results in a narrative format. The evidence about the number of episodes of diarrhoea while in the emergency department is inconclusive; one RCT showed a signicant difference in the number of diarrhoeal stools,6 while another did not show a signicant difference.8 In a 24-h follow-up, two trials8, 9 showed signicantly more episodes of diarrhoea in the ondansetron group compared with the placebo group. Based on the results of one trial, the same effect was observed at 48 h.8
ITT, intention-to-treat analysis; FU, completeness to follow-up; AGE, acute gastroenteritis; HRV, human rotavirus; AV, adenovirus. * Available case analysis (assessed by the reviewers).

Dose

Control group

Study denitions

Adverse events
Of the four trials included in the review, three reported that therapy was well tolerated and that adverse events were similar in both groups.6, 7, 9

Age

N (experiment/ Inclusion control) criteria

DISCUSSION Principal ndings

Our review of the results on using ondansetron for vomiting associated with acute gastroenteritis in children shows that such treatment was effective in increasing the chance for cessation of vomiting during the rst hours after drug administration. However, there was no such effect at 24 h. In our opinion, this can be explained by the fact that vomiting is usually self-limited, and there is a spontaneous remission of vomiting with time. Ondansetron, compared with placebo, was effective in reducing the need for intravenous rehydration. Approximately seven children would need to be treated with ondansetron to prevent one case of intravenous rehydration. Such ndings may be useful in promoting oral rehydration therapy, which currently remains underused. We did not nd any signicant reduction in hospital admissions, a nding one would consider to be clinically relevant. The observed heterogeneity in relation to this outcome seems to be accounted for largely by the study of Ramsook et al.8 In this trial, the total dose of the drug was higher compared with two other trials.6, 7 Other factors may be patients characteristics and/or

Allocation concealment Blinding ITT

2007 The Authors, Aliment Pharmacol Ther 25, 393400 Journal compilation 2007 Blackwell Publishing Ltd

Author (country)

Reeves et al.7 Adequate (computer Adequate (USA) randomization code)

Table 1. (Continued)

Generation of allocation sequence

Yes

Yes 107 54/53

FU

398 H . S Z A J E W S K A et al.

Figure 1. Ondansetron vs. control. Cessation of vomiting during the rst hours and 24 h after administration of treatment.

Table 2. Summary of the results on the effectiveness of ondansetron vs. control Comparison or outcome Cessation of vomiting (while in ED or during ORT) Cessation of vomiting in 24 h Intravenous rehydration Hospitalization Return visit to emergency department RCT(s) 3 2 2 3 3 Experiment/ control 235/231 76/68 181/178 235/231 235/227 Statistical method RR RR RR RR RR Effect size (95% CI) 1.3 1.2 0.4 0.6 1.3 (1.21.5) (0.91.7) (0.30.7) (0.41.01) (0.82.2) NNT (95% CI) 5 (48) N.S. 7 (514) N.S. N.S.

RCT, randomized-controlled trial; RR, relative risk; RB, relative benet; WMD, weighted mean difference; negative values indicate that the outcome was shorter (or reduced) in the ondansetron group than in the control group; NNT, number needed to treat; CI, condence interval.

Figure 2. Ondansetron vs. control. Intravenous rehydration.

different criteria for hospital admission (no generally agreed upon criteria). Moreover, we did not nd any signicant reduction in visits to the emergency department between the group treated with ondansetron and the placebo group.

The 5-HT3 receptor antagonists, including ondansetron, are generally well tolerated. The most common adverse events experienced by patients are headache, diarrhoea, constipation and fever.11 Our review showed that the use of ondansetron compared with placebo
2007 The Authors, Aliment Pharmacol Ther 25, 393400 Journal compilation 2007 Blackwell Publishing Ltd

M E T A - A N A L Y S I S : O N D A N S E T R O N T R E A T M E N T O F V O M I T I N G I N C H I L D R E N 399

Table 3. Ondansetron compare with placebo effect on diarrhoeal stools Outcome Number of episodes of diarrhoea while undergoing ORS,6 or during ED stay8 Number of episodes of diarrhoea in 24 h Number of episodes of diarrhoea in 48 h Reference Freedman et al.6 Ramsook et al.8 Ramsook et al.8 Cubeddu et al.9 Ramsook et al.8 Experiment/control 107/107 74/71 64/54 12/12 62/51 Results 1.4 vs. 0.5 0.7 vs. 0.6 4.7 vs. 1.4 No data 2.9 vs. 0.96 P-value <0.001 N.S. 0.002 0.013 0.015

was not associated with an increased risk of adverse effects, except for signicantly more episodes of diarrhoea during a 24- to 48-h follow-up period. However, this outcome measure may not be optimal, particularly when evaluating the effect of ondansetron in treating vomiting during acute gastroenteritis in children. Instead, stool output, as recommended by the World Health Organization,12 may represent a more adequate outcome measure and should be considered in future trials. Lack of data did not allow for analysis based on factors that may inuence the outcome (i.e. aetiology of gastroenteritis, age of participants). Thus, based on the results, it is not possible to identify children who will benet most from the treatment with ondansetron. For example, it would be interesting to know whether the drug will reduce vomiting in children with rotavirus gastroenteritis, for which severe vomiting is typical.

Previous studies
Our ndings are similar to those of the recent Cochrane Review.4 However, our review focuses on different outcome measures (e.g. cessation of vomiting), which are of practical importance to paediatricians. Secondly, we provided not only a descriptive summary of included trials, but we also conducted a meta-analysis of the extracted data. In our opinion, the latter was feasible, at least for the outcomes for which there was no statistical heterogeneity between studies (e.g. cessation of vomiting).

ively small number of patients; in total, only 247 children received ondansetron. Study limitations also included a wide age range of participants and no widely agreed-on denition of outcome measures (e.g. hospital admission, need for intravenous rehydration), and inclusion of patients with different degrees of dehydration. Possible issues with these studies were that all had pharmaceutical sponsorship. Given the small number of studies, statistical conclusions on determinants of heterogeneity might be awed. Of more importance is the fact that we cannot fully exclude publication bias, which is an important threat to the validity of systematic reviews and is difcult to combat except though the registration of all RCTs. We did not perform a statistical test for the detection of publication bias, as we are aware that these tests have very low power in the meta-analysis of only few trials.13 However, to limit the risk of publication bias, we did not impose restrictions by language or year of publication and made attempts to identify unpublished trials which strengthens our meta-analysis. Nevertheless, we are aware that, it is likely that, publication bias interfere with the analysis.

CONCLUSIONS AND FUTURE RESEARCH

The use of ondansetron was associated with some clinical benets in the treatment of vomiting in children with acute gastroenteritis. However, in our opinion, if a child is at low risk for dehydration, treatment with ondansetron is not required. In other cases, ondansetron may be considered, but there is insufcient evidence to recommend the routine use of the drug. Further well-conducted clinical studies using validated outcomes are recommended to address the cost-effectiveness of using ondansetron to treat children with acute diarrhoea, its effect on diarrhoea, and to further delineate the groups (e.g. viral vs. other aetiology of diarrhoea) deriving the greatest clinical benet from

Limitations
Any meta-analysis is only as good as the constituent studies. Except one, all trials included in our analysis had some methodological limitations, including unclear or inadequate allocation concealment and no intention-to-treat analysis. Data are based on a relat 2007 The Authors, Aliment Pharmacol Ther 25, 393400 Journal compilation 2007 Blackwell Publishing Ltd

400 H . S Z A J E W S K A et al.

ondansetron therapy. Before the results of such studies are available, clinicians may wish to consider the Centers for Disease Control and Prevention recommendation that the reliance on pharmacological agents shifts the therapeutic focus away from appropriate uid, electrolyte and nutritional therapy.1

ACKNOWLEDGEMENT
Declaration of funding interests: this study was funded in full by The Medical University of Warsaw.

REFERENCES
1 King CK, Glass R, Bresee JS, Duggan C. (Centers for Disease Control and Prevention). Managing acute gastroenteritis among children: oral rehydration, maintenance, and nutritional therapy. MMWR Recomm Rep 2003; 52 (RR-16): 116. 2 Walker-Smith J, Sandhu B, Isolauri E, et al. Recommendations for feeding on childhood gastroenteritis. Guidelines prepared by the ESPGHAN Working Group on Acute Diarrhoea. J Pediatr Gastroenterol Nutr 1997; 24: 61920. 3 Armon K, Stephenson T, MacFaul R, Eccleston P, Werneke V. An evidence and consensus based guideline for acute diarrhea management. Arch Dis Child 2001; 85: 13242. 4 Alhashimi D, Alhashimi H, Fedorowicz Z. Antiemetics for reducing vomiting related to acute gastroenteritis in chil5

dren and adolescents. Cochrane Database Syst Rev 2006; 3: CD005506. Higgins JPT, Green S. (eds) Cochrane Handbook for Systematic Reviews of Interventions 4.2.5 (updated May 2005). In: The Cochrane Library, Issue 3. Chichester, UK: John Wiley & Sons, Ltd, 2005. Freedman SB, Adler M, Seshadri R, Powell E. Oral ondansetron for gastroenteritis in a pediatric emergency department. N Engl J Med 2006; 354: 1698705. Reeves JJ, Shannon MW, Fleisher GR. Ondansetron decreases vomiting associated with acute gastroenteritis: a randomized controlled trial. Pediatrics 2002; 109: e62. Ramsook C, Sahagun-Carreon I, Kozinetz CA, et al. A randomized clinical trial comparing oral ondansetron with placebo in children with vomiting from

10

11

12

13

acute gastroenteritis. Ann Emerg Med 2002; 39: 397403. Cubeddu LX, Trujillo LM, Talmaciu I, et al. Antiemetic activity of ondansetron in acute gastroenteritis. Aliment Pharmacol Ther 1997; 11: 18596. Human Development Report 2005. Available at: http://hdr.undp.org; Last accessed October 2006. Finn AL. Toxicity and side effects of ondansetron. Semin Oncol 1992; 19: 5360. Word Heath Organization. The Rational Use of Drugs in the Management of Acute Diarrhoea in Children. Geneva, Switzerland: World Health Organization, 1990. Higgins JPT, Green S. (eds) Cochrane handbook for systematic reviews of interventions 4.2.4 (updated March 2005). In: The Cochrane Library, Issue 2. Chichester, UK: John Wiley & Sons, Ltd, 2005.

2007 The Authors, Aliment Pharmacol Ther 25, 393400 Journal compilation 2007 Blackwell Publishing Ltd