Résumé : This prospective, double-blind trial assessed whether the addition of a glycopeptide would be able to reduce the time to defervescence in neutropenic patients with cancer who had persistent fever 48-60 h after the initiation of empirical piperacillin-tazobactam monotherapy. Of 763 eligible patients, 165 with persistent fever were randomized to receive piperacillin-tazobactam therapy plus either vancomycin therapy or placebo. Defervescence was observed in 82 (95%) of 86 patients in the vancomycin group and in 73 (92%) of 79 patients in the placebo group (P=.52). The distributions of the time to defervescence were not statistically significant between the 2 groups (estimated hazard ratio, 1.03; 95% confidence interval, 0.75-1.43; P=.75). The number of additional episodes of gram-positive bacteremia and the percentage of patients for whom amphotericin B was empirically added to their therapy regimen were also similar in both groups. This study failed to demonstrate that the empirical addition of vancomycin therapy to the treatment regimen is of benefit to persistently febrile neutropenic patients with cancer.