Abstract
Background: Urinary tract infections are among the most common human infections. Due to the progressive increase in ESBL-producing bacteria and the unavailability of new antibiotics, re-evaluation of 'old' antibiotics is needed. However, the pharmacodynamics of nitrofurantoin under variable pH conditions are poorly understood. We determined the pharmacodynamic properties of nitrofurantoin at different pH levels using time-kill assays. Methods: Time-kill assays were performed at four pH levels (5.5, 6.5, 7.5 and 8.5), exposing the bacteria to 2-fold increasing concentrations from 0.125 to 32 times the MIC. Seven ESBL-positive and two ESBL-negative strains (MICs 8-32 mg/L) were used. The Deltalog10 cfu/mL values at 6 and 24 h were plotted against each log10-transformed concentration and analysed with non-linear regression analysis using the sigmoid maximum effect model with variable slope. Geometric means normalized by the MIC of the EC50, stasis and 1 and 3 log10 cfu/mL kill were calculated. Results: Minimum bactericidal effects differed significantly by species and pH level. At pH 5.5-6.5 bactericidal effects were observed at >/=0.5 x MIC for Escherichia coli and Enterobacter cloacae. At pH 8.5 only the two highest concentrations were considered bactericidal. Strong pH-dependent pharmacodynamic output parameters were observed in 6 h and especially 24 h modelling. At 24 h, pH 5.5-6.5 for E. coli and Klebsiella pneumoniae required significantly lower nitrofurantoin concentrations compared with pH 7.5 or 8.5. Although for E. cloacae similar strong decreasing trends were visible with decreasing pH, none of the tested pharmacodynamic parameters was significant. Conclusions: Nitrofurantoin bactericidal activity against Enterobacteriaceae significantly increases at lower pH levels. Bactericidal activity of nitrofurantoin may be overestimated or underestimated, which may have implications for therapy and the interpretation of clinical breakpoints.
