Effectiveness of Antibacterial Prophylaxis with Non-Absorbable Polymyxin B Compared to Levofloxacin after Allogeneic Hematopoietic Stem Cell Transplantation

Introduction:

Bacterial infection is a major cause of early mortality after hematopoietic stem cell transplantation (HSCT). Currently, fluoroquinolones are widely used as prophylaxis against bacterial infection during neutropenia in HSCT. This strategy was established based on the results of several meta-analyses which showed that fluoroquinolones improved survival outcomes in patients receiving chemotherapy including HSCT when compared to placebo or no treatment. However, there was no evidence that the sole use of fluoroquinolones decreased mortality rates compared to other antibiotics.

Intestinal flora has been investigated as a possible source of systemic infection during neutropenia through bacterial translocation from the intestinal tract. Some recent reports have shown that enterobacterial flora might be associated with bacteremia or graft-versus-host disease (GVHD) in HSCT. It was also shown in early trials that non-absorbable antibiotics which target intestinal flora improve the survival rate through decreased incidence of infection events in patients with neutropenia. However, it has been not sufficiently elucidated whether fluoroquinolones are able to reduce the infection rate more effectively than non-absorbable antibiotics, particularly in HSCT.

Here we retrospectively compared systemic levofloxacin (LVFX) with non-absorbable polymyxin B (PB) in allogeneic HSCT.

Patients and methods:

This was a retrospective cohort study. We reviewed the charts of patients who underwent allogeneic HSCT where prophylaxis against bacterial infection was given using LVFX or PB from the start of conditioning until neutrophil engraftment. The study was conducted at our institute between 2005 and 2013.

Results:

A total of 226 (86 patients with PB and 140 patients with LVFX) patients were analyzed. Median age was 46 years (range: 16-69) with a 52:48 female/male ratio. The percentages for disease diagnosis, disease risk, and conditioning intensity were statistically similar. In the latter part of the study, significantly more patients were given LVFX. With the recent developments in HSCT procedure, cord blood or an HLA-mismatched donor was used as a stem cell source and drug combinations other than the classical combination, i.e., calcineurin+ short-term MTX, were used as GVHD prophylaxis in the LVFX group more frequently than in PB group.

Median duration until neutrophil engraftment after HSCT was 16 days (range, 8-40) in the LVFX group and 15.5 days (range, 7-47) in the PB group (P= 0.74). The duration of prophylaxis was 11 days (range, 0-27) in the LVFX group and 12 days (range, 0-31) (P= 0.41) in the PB group. The type of antibiotic was changed in 96% of patients in the LVFX group and 94% in the PB group before engraftment, which was defined as failure of prophylaxis in this study (P=0.51). There was no significant difference in the incidence of clinically documented infection between the LVFX group (13%) and the PB group (17%) (P= 0.34). Microbiologically documented infection rates were 17% in the PB group and 11% in the LVFX group. The rate of resistance of etiological bacteria to LVFX was 100% in the LVFX group and 79% in the PB group (P=0.51).

No significant difference was observed between the two groups in the cumulative incidences of prophylaxis failure (P= 0.36), clinically documented infection (P= 0.26), GVHD (P=0.50) and non-relapse mortality within the first 100 days after HSCT (P= 0.62). Furthermore, there was also no significant difference in overall survival between the two groups (P=0.72). Similar results were obtained in multivariate analysis adjusted for following factors: sex, age, transplant source, disease risk, donor type, HLA disparity, intensity of conditioning, type of GVHD prophylaxis, use of TBI or ATG.

Conclusion:

Our results indicate that prophylaxis against bacterial infection even with PB, a topical antimicrobial agent, might be as effective as LVFX in the early phase after allogeneic HSCT, suggesting the importance of inhibiting pathogenic bacteria in the intestine to prevent systemic infection probably due to the bacterial translocation. Antibacterial prophylaxis with PB might be a feasible strategy to preserve quinolone as an option for later use.

Further optimization of oral prophylaxis against bacterial infection and/or a better understanding of influence of enterobacterial flora on after-transplant immunity are required.