A total of 799 GNB isolates, including Enterobacteriaceae (n-500) and GNB non-fermented (NFGNB, n-299), were recovered from various clinical samples collected from November 2013 to December 2014 at Taiwan University National Hospital, Taipeh, Taiwan. The agar dilution method, the disk scattering method and two automated sensitivity systems (Phoenix and Vitek 2) were used to test the sensitivity of isolates to CFP-SUL. The vulnerability categories (sensitivity, medium or resistance) to the PSC-SUL, produced by each method, were interpreted according to the CFP-SUL interpretation points proposed above. The results of the categorical agreement and errors between the agar dilution method and the other three methods were analyzed. Among automated systems, MicroScan showed in this study the best compliance of the overall category with the results of micro-dilution of broth. This is consistent with the qi et al. ratio, which showed a strong correlation between the presence or absence of mutations in the RNA 23S r gene and the phenotype reported by the MicroScan system (18). VITEK 2 also had a high category agreement, provided that the advanced expert system, which reported all non-receptive linezolid staph as vulnerable, was not used. However, neither system gave a categorical result to strains classified as non-sensitive, leaving this gap in the report. This is potentially very confusing for prescribers. If the repetitive tests correct the error, the repeat result should be maintained as definitive, as shown in example 1 in Table S4. If MIC values are in log2 dilution ±1, but there is a category error, the results and error status must be indicated. This scenario is more common when there is no intermediate point and/or resistance mechanism that gives MIC values near the stopping point by the BMD method.
These errors should be considered less critical, as the test works as intended, given the restrictions imposed on STAs. If the two repeated MIC results are different, but one is equal or in ±1 dilution of the initial result, this value must be selected, as shown in example 2 in Table S4. The variability observed in Table S4 is particularly problematic because the exact outcome of the MIC is not clearly defined and may be a characteristic of the isolate to be evaluated when tested against an antimicrobial or a particular class. Such variability is, in most years, a characteristic of a combination between the active substance and the isolate, and these performance problems may not constitute a systematic problem with ASAT, but rather a problem with the isolate studied. Three tests can be used; in this case, the MIC is the dominant MIC (2 out of 3 values) or the mode. For STAs, the errors were within the FDA`s indicated range, namely that the main margin of error must be less than 3% of all sensitive organisms tested and a very large margin of error of 1.5% or less .