To compare the incidence, characteristics, treatment course, and clinical outcome of children with culture-negative versus culture-positive septic arthritis, we reviewed all 105 children treated for septic arthritis at our institution from 1990 to 1997. Seventy-six children had a clinical presentation consistent with an isolated joint infection. All underwent a joint aspiration with fluid analysis including culture. All were followed up until resolution of their symptoms. Culture of the synovial aspirates identified an etiologic organism in only 30% of cases. No significant differences existed between the culture-positive and culture-negative groups in most clinical and laboratory criteria. No other diagnoses were demonstrated. All patients underwent joint drainage, received comparable antibiotic therapy, and had complete resolution of their infections. The current literature reports deceptively low rates of 18-48% for culture-negative septic arthritis. Seventy percent of children with clinical findings of septic arthritis had negative synovial fluid cultures. As the two culture groups were comparable and no other diagnoses were demonstrated, the culture-negative cases were likely infections. Thus we recommend the same aggressive treatment in those cases with and without identification of a causative organism.