Use of Ondansetron for Vomiting After Head Trauma: Does It Mask Clinically Significant Traumatic Brain Injury?

Abstract

We describe ondansetron use in children with head injury evaluated in pediatric emergency departments and its association with return visits and late diagnoses of intracranial injuries requiring intervention. Children ages 6 months to 18 years discharged without neuroimaging from 35 pediatric emergency departments with a diagnosis of head injury from 2009 to 2013 were identified retrospectively from the Pediatric Health Information System. We evaluated the rates of ondansetron use during the study period and of the association of ondansetron treatment with the diagnosis of intracranial injury, skull fracture, and return visits within 72 hours requiring admission or operative intervention. We identified 218,904 encounters during the study period. Of these, 5894 patients (2.8%) were given ondansetron. There was significant variation in the use of ondansetron during the index visit between hospitals (0.1%-5.7%), and ondansetron use significantly increased over the study period. Return visits within 72 hours were more likely for patients treated with ondansetron during the index visit (3.7% vs 1.9%; adjusted odds ratio, 1.99; 95% confidence interval, 1.7-2.4). These patients were more likely to be admitted than those not treated initially with ondansetron (7% vs 4%; adjusted odds ratio, 1.97; 95% confidence interval, 1.09-3.55). There were no significant differences in rates of skull fractures, intracranial injury, intensive care unit admission, or operative intervention between groups. Ondansetron use during an initial emergency department visit for head trauma in children not requiring neuroimaging is associated with a higher likelihood of return within 72 hours and subsequent admission. There were no differences in rates of missed skull fractures, intracranial injury, intensive care admission, or operative intervention for groups who were and were not treated with ondansetron; however, this study was underpowered to detect significant differences in these categories. Future investigations with greater numbers would be required to confidently assess these critical differences.