Antibacterial prophylaxis for patients with neutropenia after HCT may decrease morbidity and mortality. We sought to determine whether levofloxacin prophylaxis is as effective as ceftazidime (which has been utilized at our center for the past decade) for this indication. Consecutive cohorts of allogeneic HCT recipients received ceftazidime (2 g IV q8h, N=221) or levofloxacin (750 mg PO or IV, N=103) as antibacterial prophylaxis during neutropenia (absolute neutrophil count < 500 cells/μL). The two groups were comparable with regards to age, gender, underlying disease, conditioning regimen, and donor match. Compared to the levofloxacin cohort, significantly more patients in the ceftazidime cohort received bone marrow rather than peripheral blood stem cells (p=0.007). Overall, 71 (32%) of patients in the ceftazidime cohort had at least one bacterial infection as compared to 31 (30%) in the levofloxacin cohort; time to first infection was similar as well. Eleven percent of ceftazidime patients and 8% of levofloxacin patients experienced 2 or more infectious episodes. In total, the ceftazidime cohort had 104 infections (mean 0.47 infections per patient) and the levofloxacin cohort had 44 infections (mean 0.43 infections per patient). Breakthrough infections with skin-colonizing gram-positive cocci were modestly more prevalent in the levofloxacin cohort than in the ceftazidime cohort (38.6% of infections vs. 34.6%), as were enteric gram-negative rods (6.8% vs. 5.8%). Oral and gastrointestinal gram-positive cocci (22.7% vs. 30.8%), and environmental gram negative rods (11.4% vs. 12.5%), were more prevalent in the ceftazidime cohort. Of the seven viridans streptococci infections isolated in the ceftazidime cohort, only one was resistant to levofloxacin, yet all of the five isolates in the levofloxacin cohort were resistant. Patients in the levofloxacin cohort had a significantly higher probability of having a fever, as compared to patients in the ceftazidime cohort (p=0.004). Changes in antibiotic regimen were more likely to occur in the levofloxacin cohort (87 of 103 patients) as compared to the ceftazidime cohort (154 of 220 patients), p=0.0002. Relapse-free survival in the first 200 days after HCT, however, did not differ by prophylaxis regimen. Our data demonstrates that levofloxacin is an appropriate alternative to ceftazidime for use as antibacterial prophylaxis in allogeneic HCT recipients, and may be more cost effective given its oral bioavailability. Prospective, randomized trials are needed to elucidate the ideal prophylaxis regimen.