Abstract
BACKGROUND: Burkholderia cepacia complex (Bcc) is a collection of intrinsically drug-resistant Gram-negative bacteria that cause life-threatening disease in people with cystic fibrosis (CF). Standard antimicrobial susceptibility testing methods have poor predictive value for clinical outcomes in Bcc infections, probably due in part to differences between in vitro testing conditions and the environment in which Bcc grow in the lungs of people with CF. OBJECTIVES: To compare the activity of commonly used antibiotics under standard in vitro testing conditions with activity in conditions mimicking those found in vivo. METHODS: Two Bcc strains were grown alone and with six different antibiotics (minocycline, ceftazidime, meropenem, tobramycin, levofloxacin, trimethoprim-sulfamethoxazole) in two different media: standard cation-adjusted Mueller-Hinton broth and an artificial sputum medium designed to simulate the environment in the lungs of people with CF through addition of components including mucin, free DNA and amino acids. Two different starting conditions were used for time-kill assays: a standard ∼5 × 106 cfu/mL inoculum, and a high-density inoculum in which bacteria were grown for 72 hours before addition of antibiotics. Growth detection was performed by colony enumeration and by detection of resazurin reduction. RESULTS: There were major discrepancies between standard susceptibility results and activity in our models. Some antibiotics, including ceftazidime, showed minimal activity in all time-kill assays despite low minimal inhibitory concentrations, while others, notably tobramycin, were more active in high-density growth conditions than in standard time-kill assays. CONCLUSIONS: This work underscores the urgent need to develop more clinically relevant susceptibility testing approaches for Bcc.