Fever of unknown origin in elderly patients.

OBJECTIVE To describe the spectrum of diseases that may give rise to fever of unknown origin in elderly patients and to delineate the diagnostic approach in these patients. DESIGN Subgroup analysis of a prospectively collected case series followed more than 2 years. SETTING General Internal Medicine Service based at University hospital, Leuven, Belgium. PATIENTS Forty-seven consecutive patients, older than 65 years, meeting the classic criteria of fever of unknown origin. MEASUREMENTS The final diagnosis established and the clinical value of diagnostic procedures. RESULTS Infections, tumors and multisystem diseases (encompassing rheumatic diseases, connective tissue disorders, vasculitis including temporal arteritis, polymyalgia rheumatica, and sarcoidosis) were found in 12 (25%), six (12%) and 15 patients (31%), respectively. Drug-related fever was the cause in three patients (6%), miscellaneous conditions were found in five patients (10%), and six patients (12%) remained undiagnosed. Microbiologic investigations were diagnostic in eight cases (16%), serologic tests yielded one diagnosis, immunologic investigations had a diagnostic value in four cases, standard X-rays yielded a diagnostic contribution in 10 cases, ultrasonography and computed tomography were diagnostic in 11 cases, Gallium scintigraphy had a diagnostic contribution in 17 cases, and biopsies yielded the final diagnosis in 18 cases. CONCLUSIONS Multisystem diseases emerged as the most frequent cause of fever of unknown origin in the elderly, and temporal arteritis was the most frequent specific diagnosis. Infections, particularly tuberculosis, remain an important group. The percentage of tumors was higher in our elderly patients than in the younger ones but still clearly lower than in other recent series of FUO in adults. The number of undiagnosed cases was significantly lower in elderly patients than in younger individuals (P < or = 0.01). The investigation of elderly patients with FUO should encompass routine temporal artery biopsy and extensive search for tuberculosis if the classic tests such as blood count, chemistry, urinalysis, cultures, chest X-rays, and abdominal ultrasonography do not yield any clue. Gallium scintigraphy should be considered as the next step and not as a last-resort procedure.