Mezlocillin, ceftizoxime, and amikacin alone and in combination against six Enterobacteriaceae in a neutropenic site in rabbits.

Closed-space, locally neutropaenic infection sites were simulated in rabbits by implanting subcutaneous semipermeable chambers inoculated with 5 X 10(4) cfu/ml of Escherichia coli (one strain), Citrobacter diversus (one strain), Klebsiella pneumoniae (two strains) or Serratia marcescens (two strains). Four hours after inoculation, treatment was begun with amikacin (15 mg/kg), mezlocillin (100 mg/kg), or ceftizoxime (50 mg/kg) alone or in two-drug combinations. Antibiotics were given intramuscularly every 6 h for 16 doses. Chambers were sampled for bacterial counts at the start of therapy, and 20, 44, and 92 h thereafter. Blood and chamber antibiotic levels were determined during the treatment course. In-vivo results were compared to in-vitro tests of inhibition, killing and synergism. Ceftizoxime alone was successful in vivo (greater than 6 log reduction in count) against the three strains of klebsiella and citrobacter, mezlocillin against one serratia strain, and amikacin alone against none of the strains. The best in-vitro correlation was with the minimum inhibitory concentration (MIC) at a high (10(8] inoculum for the beta-lactams and with the anaerobic MIC for amikacin. Among combinations, amikacin plus mezlocillin alone was successful against the three strains of klebsiella and serratia, but neither amikacin nor mezlocillin added to ceftizoxime were more successful than ceftizoxime alone. In-vitro chequerboard synergism was not predictive of in-vivo success. Mezlocillin alone was inactivated in vivo by all the strains except Ser. marcescens, and the E. coli strain inactivated both mezlocillin and ceftizoxime alone and in combination.