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

Abstract Objectives In a new experimental model of carbapenemase-producing Klebsiella pneumoniae osteomyelitis we evaluated the efficacy of colistin alone and in various combinations and examined the emergence of colistin-resistant strains and cross-resistance to host defence peptides (HDPs). Methods KPC-99YC is a clinical strain with intermediate susceptibility to meropenem (MIC = 4 mg/L) and full susceptibility to gentamicin, colistin and tigecycline (MICs = 1 mg/L) and fosfomycin (MIC = 32 mg/L). Time–kill curves were performed at 4× MIC. Osteomyelitis was induced in rabbits by tibial injection of 2 × 108 cfu. Treatment started 14 days later for 7 days in seven groups: (i) control; (ii) colistin; (iii) colistin + gentamicin; (iv) colistin + tigecycline; (v) colistin + meropenem; (vi) colistin + meropenem + gentamicin; and (vii) colistin + fosfomycin. Results In vitro, colistin was rapidly bactericidal, but regrowth occurred after 9 h. Combinations of colistin with meropenem or fosfomycin were synergistic, whereas combination with tigecycline was antagonistic. In vivo, colistin alone was not effective. Combinations of colistin with meropenem or fosfomycin were bactericidal (P < 0.001) and the addition of gentamicin enhanced the efficacy of colistin + meropenem (P = 0.025). Tigecycline reduced the efficacy of colistin (P = 0.007). Colistin-resistant strains emerged in all groups except colistin + fosfomycin and two strains showed cross-resistance to HDP LL-37. Conclusions In this model, combinations of colistin plus meropenem, with or without gentamicin, or colistin plus fosfomycin were the only effective therapies. The combination of colistin and tigecycline should be administered with caution, as it may be antagonistic in vitro and in vivo.