Minimum bactericidal concentration (MBC)

After determining the MIC, 20 μL of the cell suspension in the microtiter plate was transferred to a new 96-well round-bottom microtiter plate filled with 180 μL DE broth for 5 min. Subsequently, 12.5 μL of each well was spotted on PCA-plates. Plates were incubated at 37 °C for 24 h and the MBC was determined. The MBC was defined as the lowest concentration where no visible growth on the agar plate was observed (~ 5 log CFU reduction).

For both animal categories the antibiotic resistance prevalence and the accompanying 95% confidence interval was calculated for each antibiotic based on the standard error of the binomial distribution in Microsoft Excel (Microsoft, 2016). The association between active components used (absent = 0, present = 1) during disinfection and antibiotic resistance at each farm was tested by means of binary logistic regression analysis taking the corresponding antibiotic use (TI100) into account as co-variable. First, the independent variables (‘use of QACs-GA’, ‘use of F’, ‘use of PA-H₂O₂’ and ‘use of other active components’) were tested univariable for all antibiotics (n = 13 by combining sulfamethoxazole and trimethoprim resistance). Those variables with univariable P-values of < 0.20 were retained for further analysis in a multivariable model. Subsequently, with the retained variables, a multivariable logistic regression model was constructed using the stepwise backward elimination procedure starting with the global model and gradually excluding all non-significant factors. Multivariate binary logistic regression models were used for each antibiotic. As multiple models were tested to evaluate the effect of the different active components on the different types of antibiotic resistance a bonferroni correction for multiple testing was performed. P-values ≤0.0038 (after Bonferroni correction) were considered as significant. All statistics were performed using SPSS Statistics 25.0 (IBM Corporation, Armonk, NY).