Phenotypical and genotypical resistance testing of Pasteurella multocida isolated from different animal species in Austria

Leonard Barta¹Anna Stöger²Daniel Polzer¹Werner Ruppitsch³Friedrich Schmoll¹Tatjana Sattler^1,4*

• ¹Austrian Agency for Helath and Food Safety, Mödling, Austria
• ²Austrian Agency for Health and Food Safety, Vienna, Austria
• ³Medizinische Universitat Innsbruck, Innsbruck, Austria
• ⁴Leipzig University, Leipzig, Germany

Pasteurella multocida is an economically important pathogen in veterinary medicine. Data on its antimicrobial resistance vary widely across regions. Furthermore, most of the found literature focuses on phenotypic resistance testing. To date, no study has examined P. multocida resistance in Austria, and no national surveillance program exists. In this study, we tested 276 isolates of P. multocida from different hosts including farm animals, pets, wildlife and humans. Susceptibility testing was performed using three different variants of the broth microdilution method against 16 antibiotics, applying veterinary specific breakpoints referenced from CLSI: the CAMHB method using cation adjusted Mueller Hinton Broth, the LHB method supplemented with laked horse blood and the LHB+CO2 method, which additionally included an enriched CO2 atmosphere. Whole genome sequencing was then performed to identify resistance genes. Genomic data and the results from the phenotypical resistance testing were compared to determine the most suitable method for the detection of resistance. About 20% of bovine isolates and 9% of pig isolates carried at least one resistance gene. No resistance genes were detected in isolates from other hosts. The most commonly detected resistance genes were against tetracyclines, aminoglycosides and sulphonamides. Resistance against florfenicol and macrolides was scarce and only present in bovines. Three or more different resistance genes were found in 3% of porcine strains and 10% of cattle strains. In pig isolates, the comparison of phenotype and genotype revealed a good concordance rate using both the CAMHB and LHB methods. Method LHB+CO2 yielded major discrepancies in macrolide susceptibility results. In cattle, CAMHB method showed a high concordance, however, it failed to identify resistant isolates. While the LHB and LHB+CO2 methods demonstrated effective detection of resistance genes, they were associated with a higher rate of false-positive results for ampicillin resistance. We recommend performing antimicrobial resistance testing of P. multocida with the supplementation of LHB. Despite the occurrence of false positive results in resistance, it is still the most suitable method to detect resistance genes. Our results suggest good efficacy of antibiotics against P. multocida in Austria, however, the risk posed by strains carrying multiple resistance genes should not be overlooked.