Evaluation of the Etest method for fosfomycin susceptibility of ESBL-producing Klebsiella pneumoniae

Sir, Escherichia coli and Klebsiella pneumoniae are the most commonly isolated microorganisms in cases of uncomplicated lower urinary tract infection (UTI). Because of the increased isolation of extended-spectrum β-lactamase (ESBL)-producing E. coli and K. pneumoniae from the urine samples of women with uncomplicated UTI,1 it is important to find alternative oral drugs for the treatment of such infections. Fosfomycin may be a useful alternative for treatment because of its pharmacokinetic and microbiological properties.2 In previous reports it has been demonstrated that fosfomycin activity against ESBL-producing isolates is similar to that observed against non-ESBL-producing isolates. Moreover, cross-resistance with other classes of antimicrobial agents has not been observed.3 In the USA, fosfomycin is approved only for the treatment of uncomplicated UTIs caused by E. coli, including ESBL producers.4 Further clinical studies are required to assess the clinical efficacy of fosfomycin for treatment of UTIs caused by ESBL-producing K. pneumoniae. An important issue that hampers the clinical assessment of this antibiotic is the lack of reliable and easily performed methods of evaluating in vitro activity against K. pneumoniae. So far, the only approved susceptibility testing method for fosfomycin is agar dilution.4 However, in a recent report where broth microdilution and disc diffusion were compared with the reference method, complete agreement was observed for E. coli with both methods while marked discrepancies between them were found for K. pneumoniae.5 Because the agar dilution method is laborious and time consuming, the present study was designed to evaluate the accuracy of the Etest method against 138 ESBL-producing K. pneumoniae isolates in comparison with the agar dilution method. K. pneumoniae strains were collected in two Spanish studies.5 In both studies ESBL production was confirmed by broth microdilution, according to CLSI guidelines.4 The MICs of fosfomycin (Zambon, Milan, Italy) were determined by the agar dilution method in Mueller–Hinton medium (Oxoid Ltd, Basingstoke, UK), supplemented with 25 mg/L G-6-P (Sigma Chemical Co., St Louis, USA). A 2-fold dilution across a range of 0.25–512 mg of fosfomycin per litre was used. The inocula were prepared to achieve 1 × 104 cfu per spot.4 At the same time, strips of fosfomycin containing 25 mg/L G-6-P (AB Biodisk, Solna, Sweden) were tested in Mueller–Hinton agar, following the manufacturer's recommendations. The tests were performed in duplicate and the strains with errors were re-tested. In the absence of accepted fosfomycin breakpoints for K. pneumoniae, those proposed by the CLSI for E. coli have been used in the present report.4 Susceptibility results of the Etest method were compared with those obtained from the agar dilution method. Agreement and discrepancies were classified in terms of: in complete agreement, very major errors, major errors and minor errors. Etest results were compared with agar dilution values and discrepancies of more than 4-fold were repeated. These repeated test values became the definitive result. Differences in proportion between paired categorical data were calculated by the McNemar test. The Wilcoxon test was used to compare MIC values. Statistical significance was established at P < 0.05. For evaluation of the results, all strains with intermediate susceptibility to fosfomycin were considered as resistant strains. Using the agar dilution method, 128 (92.7%) strains were susceptible to fosfomycin with MIC values of ≤64 mg/L. MIC50 and MIC90 of fosfomycin were 16 and 64 mg/L, respectively (range 0.5–512 mg/L). Only four strains were resistant to fosfomycin (MIC ≥ 256 mg/L) and another six isolates showed intermediate susceptibility to this agent (MIC 128 mg/L). Using the Etest method, 124 (89.8%) of all 138 strains tested were susceptible to fosfomycin (P = 0.138, McNemar test). Rates of very major, major and minor errors were 2.1%, 5.8% and 5.8%, respectively. Complete categorical agreement (±1 dilution) was 82%, and non-significant differences were found between the MIC values of both methods (P = 0.609, Wilcoxon test). The recommended results for an acceptable performance of susceptibility testing included an overall category error rate of <10% and an essential agreement with the reference method (±1 dilution) of ≥90%.6 In our study, essential agreement was low (82%) and the overall error rate was 13.7%. As previously described, our results suggest excellent in vitro activity for fosfomycin against ESBL-producing K. pneumoniae strains. Fosfomycin is only recommended for E. coli UTI treatment, and the CLSI and other committees have subsequently omitted K. pneumoniae from their recommendations because of inadequate clinical data. Further clinical studies are required to assess the efficacy of fosfomycin for the treatment of UTIs caused by K. pneumoniae and this kind of study needs accurate MIC tests. Our data confirm that the MIC data for fosfomycin against K. pneumoniae is markedly dependent on the method used,5 and the Etest is not a reliable alternative for the agar dilution method. This study was performed under the auspices of the Spanish Network for Research of Infectious Diseases (REIPI, Instituto de Salud Carlos III, Spain). This study was partially supported by Zambon Laboratories (Spain). None to declare.