PURPOSE: To evaluate the frequency of etiologies of Horner syndrome in children and suggest an imaging and laboratory diagnostic protocol to evaluate for neuroblastoma and other lesions in a child presenting with Horner syndrome and no known cause. DESIGN: Retrospective chart and data review. METHODS: A retrospective review of all children seen at a large pediatric neuro-ophthalmology referral center with a diagnosis of Horner syndrome between 1993 and 2005 with particular attention to underlying etiologies and the results of imaging and urine catecholamine studies. RESULTS: Fifty-six children met criteria for Horner syndrome and further review. Twenty-eight children (50%) had no previously identified cause for Horner syndrome. Of these children, 24 (85.7%) had urine catecholamine metabolite studies, and all had negative results. Twenty (71.4%) had complete modern imaging of the brain, neck, and chest. Of the 18 children who had complete imaging and urine studies, responsible mass lesions were found in six (33%). Four had neuroblastoma, one had Ewing sarcoma, and the other had juvenile xanthogranuloma. Of all patients (diagnosis known and unknown), neoplasm was the etiology in 13 of 56 (23%) of patients. CONCLUSIONS: We confirm that Horner syndrome in a child of any age without a surgical history requires a complete examination to exclude a mass lesion. In such patients, we recommend brain, neck, and chest magnetic resonance imaging (MRI) with and without contrast as well as urinary catecholamine metabolite testing. However, imaging is more sensitive than urine testing in this setting.
PURPOSE: To evaluate the frequency of etiologies of Horner syndrome in children and suggest an imaging and laboratory diagnostic protocol to evaluate for neuroblastoma and other lesions in a child presenting with Horner syndrome and no known cause. DESIGN: Retrospective chart and data review. METHODS: A retrospective review of all children seen at a large pediatric neuro-ophthalmology referral center with a diagnosis of Horner syndrome between 1993 and 2005 with particular attention to underlying etiologies and the results of imaging and urine catecholamine studies. RESULTS: Fifty-six children met criteria for Horner syndrome and further review. Twenty-eight children (50%) had no previously identified cause for Horner syndrome. Of these children, 24 (85.7%) had urine catecholamine metabolite studies, and all had negative results. Twenty (71.4%) had complete modern imaging of the brain, neck, and chest. Of the 18 children who had complete imaging and urine studies, responsible mass lesions were found in six (33%). Four had neuroblastoma, one had Ewing sarcoma, and the other had juvenile xanthogranuloma. Of all patients (diagnosis known and unknown), neoplasm was the etiology in 13 of 56 (23%) of patients. CONCLUSIONS: We confirm that Horner syndrome in a child of any age without a surgical history requires a complete examination to exclude a mass lesion. In such patients, we recommend brain, neck, and chest magnetic resonance imaging (MRI) with and without contrast as well as urinary catecholamine metabolite testing. However, imaging is more sensitive than urine testing in this setting.