Efficacy of colistin alone and in various combinations for the treatment of carbapenemase-producing Klebsiella pneumoniae experimental osteomyelitis

Background: Optimal treatment (Tx) of carbapenemase-producing Enterobacteriaceae infections is poorly defined. We evaluated the efficacy of colistin alone and in combinations in a new experimental model of carbapenemase-producing K. pneumoniae (CR-Kp) osteomyelitis.

Methods: KPC-99YC is a clinical strain intermediate to meropenem (MIC 4 mg/L), and susceptible to gentamicin (MIC 1 mg/L), colistin, and tigecyclin (MIC 0.5 mg/L). Plasma antibiotic concentrations were measured in rabbits, to select doses equivalent to those used in humans (H). Time-kill curves were performed at 4 x MIC. An osteomyelitis was induced in rabbits by tibial injection of a sclerosing agent, followed by 10⁹ CFU/mL (0.2ml) of KPC-99YC. Tx was started at day 14, for 7 days, in 6 groups (12 rabbits/group): (1) control, (2) group C: colistin 150 000 IU/kg tid (equivalent to 3 M IU tid in H), (3) group CG: colistin + gentamicin, 30 mg/kg im od (5 mg/kg od in H), (4) group CT: colistin + tigecyclin 14 mg/kg im bid (50 mg bid in H), and (5) group CM: colistin + meropenem 80 mg/kg sc tid (1 g tid in H), (6) group CMG. Bone samples were quantitatively cultured, and tested for colistin resistance.

Results: In vitro, colistin bactericidal effect was inhibited when combined with T, and enhanced when combined with M or G (>2 log₁₀ CFU). Bone samples were sterile in 0/11, 1/11, 1/12, 2/12, 1/12 and 7/10 rabbits in control, C, CG, CT, CM, and CMG groups, respectively. Median [interquartile range] bacterial concentrations (log₁₀ CFU/g) in bones were 5.4 [3.5, 6.1] (C), 4.9 [4.4, 5.6] (CG), 6.6 [6, 7.2] (CT), 4.5 [3.9, 5.2] (CM), 1.5 [1.5, 4], and 5.7 [4.6, 6.1] (controls). Bacterial concentrations were lower in CM and CMG groups as compared to controls (p=0.049 and 0.001), and higher in CT group as compared to C (p=0.01). C or CT were not better than controls (p>0.2). Colistin-resistant strains were detected in groups C, CT, CG but not in CM and CMG.

Conclusion: Addition of meropenem to colistin was the only effective therapy and prevented emergence of colistin-resistance, in agreement with the concept that i) colimycin should always be combined; ii) penems are the best companion drugs, whenever MIC<8 mg/L. The combination of colistin and tigecyclin was less effective than colistin alone, suggesting an antagonistic effect in vitro and in vivo.