DOUBLE BETA-LACTAM REGIMEN COMPARED TO AN AMINOGLYCOSIDE BETA-LACTAM REGIMEN AS EMPIRIC ANTIBIOTIC-THERAPY FOR FEBRILE GRANULOCYTOPENIC CANCER-PATIENTS

In a prospective, randomized trial, 205 febrile episodes in granulocyt openic cancer patients were treated with ceftazidime with or without t obramycin (C +/- T), both agents being administered only if the initia l granulocyte count was below 200/mul, or ceftazidime plus piperacilli n (C + P). The overall response rate was 71% (39 of 60 for C +/- T and 45 of 58 for C + P). Logistic regression analyses documented no evide nce of a significant difference between the two regimens in overall tr eatment effect after accounting for the linear effects of potentially important variables, such as infection type and granulocyte count. Alt hough the response rates for the subgroup of patients with bacteremias was better with the C + P regimen (P = 0.06), there was no difference in response for patients with bacteremia and profound (< 100/mul) sus tained granulocytopenia. The double beta-lactam combination demonstrat ed in vitro synergism in 73%; antagonism was not seen. Both regimens p roduced excellent serum bactericidal levels (C +/- T geometric mean pe ak 1:170; C + P peak 1:137) against gram-negative but not gram-positiv e pathogens (1:4; 1:7 respectively) that had caused bacteremia. Emerge nce of resistance and significant coagulopathy and/or bleeding did not occur during therapy. Antibiotic-related nephrotoxicity was noted in 7 of 95 trials in the C + P and in 6 of 89 trials in the C +/- T group (P = 0.19). The incidence of secondary infections in patients with pr ofound (< 100/mul) sustained granulocytopenia was lower in the C +/- T group (P = 0.04). Alimentary canal anaerobic flora preservation with C +/- T, and suppression with C + P, was demonstrated. These results s uggest that these regimens are of similar effectiveness and neither is associated with major toxicity.