Objective: To determine the relative proportions of the diagnostic categories in patients with fever of unknown origin who were examined in the 1980s.
Study design: Prospective case series.
Setting: General Internal Medicine Service based at University Hospital, Leuven, Belgium.
Patients: One hundred ninety-nine consecutive patients meeting the classic criteria of fever of unknown origin who were treated in the 1980s.
Main outcome measurement: The final diagnosis established at discharge or during follow-up.
Results: Infections were found in 45 patients (22.6%), tumors were found in 14 (7%), multisystem diseases were found in 42 (21.5%), drug-related fever was found in six (3%), factitious fever was found in seven (3.5%), habitual hyperthermia was found in five (2.5%), miscellaneous diseases were found in 29 (14.5%), and no diagnosis was reached in 51 (25.6%).
Conclusions: Tumors were a less important cause of fever of unknown origin in the 1980s. The same holds true for some infectious diseases, such as abscesses and hepatobiliary disorders. Multisystem diseases were more frequently found, and the number of undiagnosed cases increased. Although these shifts in the disease spectrum in fever of unknown origin most probably resulted from a constellation of factors, we suspect that these changes are mainly due to easy and early diagnosis by new diagnostic modalities, such as ultrasonography and computed tomography, of previously common causes of fever of unknown origin.