[A prospective, controlled, non-randomized study of three cohorts to test the effectiveness of two antibiotics levofloxacin and cotrimoxazol on antibacterial prophylaxis in neutropenic patients].

BACKGROUND AND

OBJECTIVE

At present, there is strong concern about the efficacy of current antimicrobial prophylaxis for the management of neutropenic patients. The purpose of this study was to test the effectiveness of levofloxacin, a new quinolone with expanded activity against grampositive bacteria, versus cotrimoxazol as a prophylactic treatment for granulocytopenic patients.

PATIENTS AND METHOD: In this prospective and controlled study, we included 249 consecutive episodes of neutropenia, such as those resulting from lymphoma and leukemia treatment, during 28 months (from November 1999 to February 2002). These episodes were divided into 3 cohorts: the first was treated with levofloxacin, the second with cotrimoxazol and the third was a subgroup without antibiotic prophylaxis (control group). The incidence of infection, rate of mortality, and reduction of hospitalization rate for treatment with parenteral antibiotics were tested.

RESULTS

There was a reduction in documented infections (clinically or microbiologically) when comparing the levofloxacin cohort with the control cohort (p < 0.0001) and the levofloxacin cohort with the cotrimoxazol group (p < 0.01). The reduction in the hospitalization rate for treatment with parenteral antibiotics reached statistical significance when comparing the levofloxacin group with the control cohort (p < 0.001) and levofloxacin group with the cotrimoxazol group (p < 0.05). Although the rate of global mortality was lower in the levofloxacin group than in the other two groups, no statistical significance was observed.

CONCLUSIONS

Our results show that levofloxacin effectively reduces the incidence of infection, the rate of hospitalization and the requirement for parenteral antibiotics. Although we found a reduction in the overall mortality and in the infection-related mortality, the corresponding data did not reach statistical significance.