Untersuchungen zur Implementierung des Bouillon-Mikrodilutionsverfahrens zur Empfindlichkeitsprüfung von Bakterien gegenüber antimikrobiellen Wirkstoffen

The in-vitro susceptibility testing of bacterial pathogens is crucial for the therapeutic use of antimicrobial agents as well as for a meaningful resistance monitoring. The determination of quantitative data (minimal inhibitory concentrations [MIC values] using the broth microdilution method is considered the method of choice particularly since an internationally accepted guideline (Standard M31-A2 [2002]) has been presented by the National Committee for Clinical Laboratory Standards (NCCLS, now: Clinical and Laboratory Standards Institute, CLSI); this guideline contains breakpoints for veterinary use and, therefore, allows a classification of bacteria tested based on MIC values. Therefore, the implementation of this technique into routine diagnostics is of high interest particularly for accredited laboratories. For the inoculation of microtiter plates used in the broth microdilution method an inoculation density of 5x105 CFU/ml has to be achieved according to the CLSI standard. A published and commonly used method is the densitiy adjustment of the bacterial suspension using a densitometer. In this study, it was shown that the requested CFU/ml corresponded to the optical density of 0.00027 of a bacterial suspension determined in a spectrophotometer at 625 nm (OD625). The comparison of susceptibility data of clinical isolates determined by agar disk diffusion and, in parallel, by broth microdilution confirmed the comparability of both methods. For most antibiotics the frequency of “very major errors” was below 10%. However, a higher frequency of errors was observed for combinations of antimicrobial agents and bacteria for which no valid breakpoints are available. The comparative testing of cephalothin showed a frequency of “very major errors” of 20.3% for staphylococci and 18.4% for Enterobacteriaceae; for apramycin the frequency of “very major errors” was found to be 46% for Enterobacteriaceae. The comparison of microtitre plates from two manufacturers resulted in differences of two or more titration steps in an average of 5 % of the tests. These differences were generally independent from the density of the inoculum, but they varied with the antimicrobial substances. For 13 substances only minor differences (0 - 4.5%) were observed, whereas for tetracycline, tiamulin and erythromycin differences were 6.8 – 11.4%, and for ampicillin, penicillin and spectinomycin differences were 13.6 – 47.7%. Particularly for the combinations of ampicillin and penicillin with isolates Staphylococcus spp., but also spectinomycin with isolates from Pasteurella multocida and Mannheimia haemolytica a large number of divergent results occurred. For one of the manufacturers, the results of two test series also differed from one another and, most importantly, for spectinomycin the differences in MIC values also resulted in different classifications. The reference strains were used as recommended in the CLSI Standard M31-A2 for quality assurance and proved to be – with a single exception – to be within the accepted MIC ranges. The microtitre plates were coated for diagnostic purposes and, therefore, the concentrations tested for each antimicrobial agent do not allow the determination of unambiguous MIC values for the reference strains used. This, in turn, resulted in deficits in the quality assurance process of the coated plates performed by the manufacturers which became apparent when testing isolates from Pasteurella multocida, Mannheimia haemolytica, and Staphylococcus spp. . Therefore, in the future additional reference strains should be included in the quality assurance process. The comparison of two read-out systems, the visual method (accepted as reference method) and the automated spectrophotometrical analysis, showed a good agreement of both methods with most species tested (differences 0-3.4%). Only for Mannheimia haemolytica frequent differences were observed (17.3%); this is most likely due to the relatively weak growth of the organism which is not sufficient for spectrophotometrical detection. The classification of MIC values requires the existence of valid breakpoints. Not for all pathogens and indications such values are available. Therefore, in order to confirm the tentative breakpoints of apramycin for Escherichia coli suggested by the AVID (1999) the specific genotype (presence/absence of the apramycin resistance gene (aac[3’]-IV) was determined in clinical isolates that exhibited different MIC values for apramycin. The observation that 25 Escherichia coli isolates with MIC values of < 16µg/ml did not carry the resistance gene and six isolates with MIC values of > 32 µg/ml carried the resistance gene confirmed the proposed apramycin breakpoints of £ 16 µg/ml for susceptible isolates and ³ 32 µg/ml for resistant isolates.