RESUMEN
BACKGROUND: Recurrent FUO (fever of unknown origin) is a rare subtype of FUO for which diagnostic procedures are ill-defined and outcome data are lacking. METHODS: We performed a retrospective multicentre study of patients with recurrent FUO between 1995 and 2018. By multivariate analysis, we identified epidemiological, clinical and prognostic variables independently associated with final diagnosis and mortality. RESULTS: Of 170 patients, 74 (44%) had a final diagnosis. Being ≥ 65 years of age (OR = 5.2; p < 0.001), contributory history (OR = 10.4; p < 0.001), and abnormal clinical examination (OR = 4.0; p = 0.015) independently increased the likelihood of reaching a diagnosis, whereas lymph node and/or spleen enlargement decreased it (OR = 0.2; p = 0.004). The overall prognosis was good; 58% of patients recovered (70% of those with a diagnosis). Twelve (7%) patients died; patients without a diagnosis had a fatality rate of 2%. Being ≥ 65 years of age (OR = 41.3; p < 0.001) and presence of skin signs (OR = 9.5; p = 0.005) significantly increased the risk of death. CONCLUSION: This study extends the known yield of recurrent FUO and highlights the importance of repeated complete clinical examinations to discover potential diagnostic clues during follow-up. Moreover, their overall prognosis is excellent.