Fever of unknown origin.

FUO is an uncommon problem in pediatric patients if one uses a strict definition of at least 2 weeks of fever. Perhaps the best definition, as suggested by Lorin and Feigin, is more than 1 week of fever and a negative initial evaluation, including examination and preliminary laboratory results. Most patients, especially younger ones, will spontaneously improve or have common disorders, usually respiratory-related infections. Most patients will have uncommon manifestations of more commonly recognized disorders. Mortality rates, even in those series that include a week in the hospital without a diagnosis, are 15% to 20% at maximum--half that seen in most series of adult patients. Common mistakes are failure to document fever, failure to perform a complete history or physical examination, a shotgun approach that uses the laboratory to make diagnoses, and overuse or inappropriate use of newer imaging techniques in an undirected fashion. Patience, persistence, repeated histories and physical examinations, and continued observation offer the best chance of making a diagnosis in difficult cases. In the current decade, we continue to learn about new manifestations of old disorders--cat-scratch disease, Kawasaki disease, neonatal syphilis, and Epstein-Barr virus infection are examples. The use of older and well-established methods (history and physical examination) and the addition of newer techniques (ultrasound, CT, MRI, etc.) to pursue suggested diagnoses offer the best current approaches to the patient with prolonged and unexplained fever.