Diagnosis of Serious Conditions Delayed in Association with Ondansetron Treatment for Vomiting in the Pediatric Emergency Department

We evaluated the characteristics and sought risk factors for hospitalization in children who return to the emergency department within 7 days of discharge after oral or intravenous ondansetron treatment for vomiting. The secondary aim was to determine whether the diagnosis of any serious condition had been delayed as the result of discharge after ondansetron treatment. This retrospective analysis of the medical records of children who had been treated for vomiting with ondansetron in a tertiary care pediatric emergency department and revisited the emergency department within 7 days was performed between 2017 and 2019. We compared demographic and clinical features as well as management between hospitalized and discharged patients, focusing upon potentially delayed diagnoses of serious conditions. Fifty of the 89 ondansetron-treated children (56.2%) who revisited the emergency department were discharged home after their second emergency department visit and the remaining 39 (43.8%) were hospitalized. No parameter of the management of the first visit was predictive of the outcome of the revisit. Five revisit patients (5.6%) were newly diagnosed with a serious condition, with intussusception and ovarian torsion being the most substantial time-sensitive delays (the other diagnoses were pneumonia and aseptic meningitis). Physicians assessing patients who had been treated with ondansetron as supportive care for vomiting at an earlier visit to the pediatric emergency department should consider alternative diagnoses despite initial clinical improvement. No definitive risk factor for readmission was identified, but a high level of alertness to a possible meningeal or acute abdominal source is imperative.

No risk factor was identified among demographic and clinical parameters of the children treated with ondansetron on the first visitthat could predict the outcome of the revisit or the diagnosis of a serious etiology.
Clinicians should consider alternative serious diagnoses despite initial clinical improvement with ondansetron treatment.

Introduction
Vomiting is a very common complaint in infants and children who present to the emergency department (ED). It is often benign and can be managed solely with supportive measures. Physicians must be able to recognize life-threatening causes of vomiting and to avoid serious associated complications, such as electrolyte abnormalities, dehydration, or even bowel necrosis [1]. The differential diagnosis of vomiting is broad and includes pathologic and physiologic conditions affecting the gastrointestinal tract, the peritoneal cavity, the genitourinary system, and the central nervous system, as well as endocrine and metabolic functions [2].
The use of anti-emetics in children with vomiting is still controversial, but ondansetron, a 5-hydroxytryptamine-3-serotonin antagonist, is currently recommended in guidelines from international professional societies [3]. In a meta-analysis of randomized clinical trials evaluating the effectiveness and safety of antiemetics in children with acute gastroenteritis, ondansetron was the only medication that revealed a positive effect on the cessation of vomiting, on preventing hospitalizations, and on reducing the need for intravenous rehydration [4]. Ondansetron was also considered a safe intervention in the same meta-analysis; however, the topic of missed diagnoses was not evaluated [4]. One retrospective study suggested that children who receive ondansetron in the pediatric ED are more likely to have an unplanned return visit and to be admitted on that return visit than children who were discharged without antiemetic treatment, although the use of ondansetron itself did not appear to be associated with increased risks of masking serious diagnoses [5]. This suggests that ED physicians should consider the differential diagnosis of vomiting before initiating treatment with anti-emetics and carefully identify the children who are suitable for discharge after ondansetron treatment.
The primary goal of this study was to evaluate the characteristics and seek risk factors for hospitalization in children who return to the ED within 7 days of discharge after oral or intravenous ondansetron treatment for vomiting. The secondary aim was to detect whether the diagnosis of any serious condition had been delayed as a result of discharge from the ED of a child who showed clinical improvement after ondansetron treatment.

Materials and Methods
This retrospective study was based upon data from the medical records of patients who presented twice within 7 days to the pediatric ED at the Dana-Dwek Children's Hospital, a tertiary care pediatric hospital, between 1 July, 2017 and 31 December, 2019. We included patients if they were under 18 years of age, had been discharged after satisfactory treatment with ondansetron because of vomiting on the first pediatric ED visit, and who returned to the pediatric ED within 7 days because of vomiting. Patients with previous neurological or neurosurgical conditions, with medical conditions that involve impaired gut motility, with a history of abdominal surgery, as well as oncology patients were excluded from the study.
Each patient's demographic information and disease variables were extracted from the medical records, including age, sex, history of present illness, past medical history, assigned triage score using the Pediatric Canadian Triage Acuity Scale [6], management and diagnosis at the first pediatric ED visit, and disposition and diagnosis at the return visit. Triage score was categorically defined as urgent in cases documented by the triage nurse as Pediatric Canadian Triage Acuity Scale 1-3 and not urgent in Pediatric Canadian Triage Acuity Scale 4-5.

Statistical Analysis
The statistical analyses were performed with the IBM SPSS statistics version 24 program (IBM, New York, NY, USA). Differences in demographic and clinical variables were compared between patients who were admitted and those who were again discharged after the second pediatric ED visit. The distributions of categorical and continuous variables were described by a univariate analysis. Categorical variables were analyzed by Chi-square or Fisher exact tests. Continuous variables that were not normally distributed were analyzed by a Mann-Whitney test. All tests were twotailed, and p values equal to or below 0.05 were considered significant. For the secondary outcome, we defined a new diagnosis made at the revisit as being serious if an urgent surgical or other procedure or treatment (e.g., antibiotic treatment) was required. Logistic regression modeling of all the variables from the first visit was performed to identify risk factors for admission or for a new serious diagnosis on the revisit.
The study was approved by the Tel Aviv Sourasky Medical Center Institutional Review Board. Patient/parental consent was not required for this retrospective medical chart review. The study is reported according to the STROBE guidelines.

Results
During the 2.5-year study period, 3838 patients presented twice within 1 week to the pediatric ED at the Dana-Dwek Children's Hospital, representing a 5.5% revisit rate. Of them, 1095 patients (28.5%) required hospitalization. Among these revisit patients, 96 had been treated with ondansetron for vomiting during the first visit. After exclusion of seven patients (three neurosurgical patients, three with previous abdominal disease/surgery, and one oncology patient), the study cohort included 89 patients (Fig. 1). Fifty patients (56.2%) were discharged home after their second pediatric ED visit, while 39 patients (43.8%) were admitted to the hospital.
The mean (± standard deviation) age of the study population was 3.79 ± 4.3 years, with a slight male predominance (55%). The youngest patient was 6 months old, and the oldest was 17 years and 3 months. The discharge diagnosis at the first pediatric ED visit was "vomiting" in 52 patients (58.4%). Table 1 summarizes the demographic and clinical characteristics of the admitted and non-admitted study participants. No significant difference was found in the management of the first pediatric ED visit between those two groups, including the pharmacological (ondansetron given orally or intravenously) and intravenous rehydration protocol. No demographic, clinical, or ED management parameter was associated with admission on the revisit in logistic regression modeling.
A new serious diagnosis was established in five of the 89 patients (5.6%) who returned to the pediatric ED within 2.5-25 h after the first visit (Table 2). Two surgical cases had a delay in management, including a 10.5-month-old infant with persistent vomiting who was diagnosed as having intussusception, and a 7-year-old girl who was diagnosed as having ovarian torsion based upon the results of peritoneal signs on a physical examination. One case of aseptic meningitis was diagnosed by the detection of meningeal signs on a physical examination, and lobar pneumonia was suspected because of fever and respiratory symptoms and confirmed in two patients during the return visit. The non-serious diagnoses that led to hospitalization included acute gastroenteritis (n = 20), viral infection (n = 5), ketotic hypoglycemia (n = 3), ileo-ileal intussusception (n = 1), and vomiting that resolved without an identified diagnosis (n = 6). There were no cases of bacterial meningitis among the study cohort. In logistic regression modeling, no risk factor was identified as predicting a serious diagnosis. Revisit cases included in the study n = 89 All revisit cases n = 3838 New serious diagnosis n = 4 New serious diagnosis n = 1

Discussion
The use of ondansetron has become a useful adjunct in the treatment of acute gastroenteritis in the pediatric ED. Its potential to reduce episodes of vomiting and thereby reduce the need for intravenous fluid rehydration has been well established [4,7]. The focus of this study was to evaluate the characteristics of those patients who had been discharged from the pediatric ED after treatment with ondansetron for vomiting and ultimately required admission to hospital at a second ED visit within 7 days. We compared them to the patients who similarly revisited the ED and were discharged a second time with no significant health risk. The findings of this study emphasize the need to maintain a high level of suspicion of alternative differential diagnoses, notwithstanding a satisfactory clinical response to the antiemetic in a child who presents to the ED with vomiting. Our rate of hospitalizations of revisit patients treated with ondansetron was 43.8% (39/89), while the rate of hospitalizations of all revisit patients to the pediatric ED during the study period was significantly lower (1095/3838, 28.5%, p = 0.0016). This reinforces the finding of others who observed that children who receive ondansetron in the pediatric ED appear more likely to be admitted on a return visit than those who did not receive ondansetron [5]. In our cohort, there were no significant differences in the demographic and clinical data between those who were admitted to the hospital on the revisit to the ED and those who were discharged home again, precluding our ability to identify any risk factors for hospitalization. Our rate of a serious new diagnosis was 5.6% (5/89) among revisit patients, with intussusception and ovarian torsion being the most critically time-sensitive delays. No risk factor that could predict a new serious diagnosis among our study cohort was identified. In one large retrospective study of pediatric ED revisit patients within 72 h because of vomiting who had been treated with ondansetron and sent home, 25.9% of patients (309/1192) were admitted on the revisit and 3.86% (46/1192) had an alternative diagnosis, the most common of which were appendicitis, intussusception, and pyelonephritis. However, as a similar proportion of each diagnosis was missed in the revisit group that had not received ondansetron, those authors concluded that ondansetron use was not associated with an increased risk of masking a serious diagnosis [5].
There are several potential limitations to our study. First, owing to its retrospective design, it relies upon documentation during the ED visits. That said, its retrospective nature enabled us to collect data of unbiased physician practice and admission decisions. Second, the study was conducted in a single center with access to revisit patients in our institution alone, and may have missed the patients who revisited a different ED. However, being a large tertiary care urban medical center it lowers the possibility of that having occurred. Finally, the rare occurrence of the alternative diagnoses that present with vomiting in children (e.g., bacterial meningitis, intestinal obstruction) resulted in a small group of delayed serious diagnoses, precluding the possibility of evaluating significant risk factors. In order to better understand these risk factors, and compare them to patients with vomiting who were not treated with ondansetron, a prospective study is needed.

Conclusions
In this study of revisit pediatric patients after treatment with ondansetron, we observed that almost one-half of the children who had been discharged home after the first visit will be admitted, but we could not identify specific risk factors that influence the decision to admit. Physicians who recommend the use of ondansetron as supportive care for vomiting in the pediatric ED should anticipate subsequent improvement in the child's condition but maintain a high degree of alertness to the possibility of life-threatening and time-sensitive causes of the vomiting. Our experience was that the diagnoses of intussusception, ovarian torsion, and aseptic meningitis were missed, and we expect that other serious time-sensitive diagnoses could fail to be spotted as well. We urge ED physicians to maintain a high level of suspicion for their presence.

Authors' Contributions
All authors contributed to the study conception and design. GPS, ZS, and DM prepared the material and performed the data collection and analysis. GPS and AR conducted the data analysis. GPS wrote the first draft of the manuscript, AR revised it critically for important intellectual content, and all authors commented on previous versions of the manuscript. All authors read and approved the final manuscript.

Declarations
Funding The authors declare that no funds, grants, or other support were received during the preparation of this article.

Conflicts of interest/competing interests Gili Palnizky Soffer, Zeev
Schnapp, Dana Miroluz, and Ayelet Rimon declare that they have no relevant financial or non-financial interests to disclose.
Ethics approval This is an observational study. The Tel Aviv Sourasky Medical Center Institutional Review Board approved the study with a waiver of informed consent.

Consent for publication Not applicable.
Availability of data and material The datasets generated and analysed during the current study are available from the corresponding author on reasonable request.