Linezolid is an alternative to vancomycin for the long-term treatment of gram-positive bacterial orthopedic infections because of its antibacterial spectrum and oral bioavailability, but duration-related myelosuppression could offset its advantages. To evaluate the hematologic effects of these agents, we prospectively studied 65 consecutive adults with gram-positive bacterial orthopedic infections requiring > or =2 weeks of vancomycin therapy (n=52) or linezolid therapy (n=20). Trends suggesting higher incidence of hematologic effects among the patients receiving vancomycin were not significant, regardless of whether the end point was lowest cell count during therapy or change from baseline. The only difference was a higher incidence of thrombocytopenia (<150x10(9) platelets/L) in the subset of the linezolid recipients who had received vancomycin within 2 weeks before starting linezolid therapy than in the linezolid recipients who had not received vancomycin (5 [71%] of 7 patients vs. 2 [15%] of 13; P=.02). All hematologic effects were reversible. In conclusion, hematologic effects were detectable through weekly monitoring and were reversible; therefore, concern about myelosuppression need not preclude linezolid use for orthopedic infections requiring long-term therapy.