674 - Variations in Agreement and Epidemiological Cutoff Value (ECV) between Fosfomycin (FOF) Agar Dilution and Broth Microdilution Using Standard- and High-Inoculum Protocols for Klebsiella pneumoniae (KP)

Background: FOF has been used in the treatment of multidrug-resistant (MDR) KP infections despite established susceptibility breakpoints. At present, agar dilution (AD) is considered the reference method for FOF while broth microdilution (BMD) is specifically recommended against despite its convenience over AD. We therefore sought to assess FOF activity against KP, along with essential and categorical agreement between AD and BMD methods to determine if BMD could be used as a reliable testing method.

Methods: Minimal inhibitory concentration (MIC) values were determined for a convenience collection of 69 KP isolates (59.4% MDR) from three US institutions. MIC testing was conducted in duplicate on separate days using AD and BMD methods; essential and categorical agreement were calculated using AD as the reference method. Fourteen isolates were also analyzed using high-inoculum AD (10^5.3-5.9 CFU/mL) similar to the BMD method. MIC values were categorized using Clinical and Laboratory Standards Institute (CLSI) interpretive criteria for Escherichia coli (≤ 64 mg/L, susceptible). ECVs were determined according to CLSI methodology.

Results: MIC values varied between methods, withMIC₅₀/MIC₉₀ values being 32/256 mg/L for AD and 128/256 mg/L for BMD. Using E. coli criteria, susceptible/intermediate/resistant rates were 82.6/2.9/14.5% (AD) and 44.9/21.7/33.3% (BMD). Essential agreement was 44.9% and categorical agreement was 60.8%. When using high-inoculum AD, MIC values were on average three-fold higher compared to standard-inoculum AD, with 10 of the 14 (71.4%) isolates brought into essential agreement with BMD. Calculated ECVs were 128 mg/L for standard-inoculum AD and 1024 mg/L for BMD.

Conclusion: Our collection of KP displayed high MIC values to FOF, in addition to substantial discrepancies between AD and BMD methods. Essential agreement increased with the use of high-inoculum AD testing, which better correlated with BMD results. ECV for BMD was three dilutions higher than that for standard-AD ECV. Based on these results, we recommend further investigation of BMD for FOF testing using a larger isolate collection, along with optimization of currently recommended testing methods. In light of these results, KP-specific breakpoints should also be examined.