RESUMO
OBJECTIVE: Pediatric headaches are a common presentation to emergency departments accounting for almost half a million annual visits. Providers are left with the difficult task of deciding who has a secondary headache etiology that warrants neuroimaging. METHODS: We conducted a retrospective case-control study. Patients from a pediatric neuro-oncology clinic database with brain cancer and a headache at presentation were identified as cases. Controls were patients from 2 local pediatric tertiary care emergency departments with a final diagnosis of headache after negative neuroimaging. Clinical factors were decided a priori, and logistic regression was used to determine which clinical factors were related to case/control status. RESULTS: A total of 334 patients (203 controls and 131 cases) were included. Patients with a history of headaches had 0.5 (95% confidence interval [CI]: 0.3-0.9; P = 0.03) times the odds of being a case. Patients with vomiting had increased odds of being a case compared with controls regardless of the time of day (early morning 1.8 [95% CI: 1.0-3.2; P = 0.04] and non-early morning 6.6 [95% CI: 2.0-21.7; P < 0.01]). Patients with neurological signs had 10.3 (95% CI: 5.4-19.4; P < 0.01) times the odds of being a case, and patients with an associated seizure had 10.9 (95% CI: 3.8-30.7; P < 0.01) times the odds of being a case. CONCLUSIONS: This study identified clinical factors associated with pediatric brain neoplasms that may guide acute neuroimaging decisions. This study also provides insight into potential clinical factors to be studied prospectively to derive a clinical decision rule.