Fever of unknown origin-predictors of outcome. A prospective multicenter study on 164 patients.

BACKGROUND: To date, the studies that have been done on fever of unknown origin have mostly been descriptive. Therefore, we know the etiogical spectrum and how it has changed since 1966 for many regions of the world. However, we do not know if there are clinical or laboratory predictors of severe outcome. Being able to estimate the severity of the disease early on would allow one to determine how intensive the diagnostic work-up should be. METHODS: A multicenter cohort study was carried out on 164 consecutive patients who met the classic, modified criteria of fever of unknown origin. The study lasted 2 years (1997-1998) and included a follow-up period of another 2 years. The main outcome measured was the final diagnosis established at the end of follow-up. RESULTS: When the white cell count was abnormal, the relative risk for a serious disease was 1.49 (CI: 1.15-1.94; p=0.004), when anemia was present, the relative risk was 1.55 (CI: 1.21-1.98; p=0.003), and for high alanine aminotransferase (ALAT), bilirubin, or lactate dehydrogenase (LDH), the relative risks were 1.57 (CI: 1.21-2.02; p=0.010), 1.57 (CI: 1.18-2.08; p=0.007), and 3.43 (CI: 1.81-6.48; p=0.0002), respectively. In multivariate analysis, the odds ratios for serious diseases were 2.7 (CI: 1.17-6.4; p=0.02) for abnormal white cell count, 2.8 (CI: 1.14-7.16; p=0.02) for anemia, 4.3 (CI: 1.6-11.5; p=0.003) for high serum bilirubin, and 5.3 (1.5-18.6; p=0.009) for high serum ALAT. CONCLUSIONS: In patients having a fever of unknown origin, anemia, abnormal white cell count, and high ALAT and bilirubin are independent predictors of severe outcome.