Ceftiofur, whether administered as hydrochloride or sodium salt, was metabolized rapidly to desfuroylceftiofur. The plasma half-life of ceftiofur sodium after intravenous dosing in swine was approximately 10 min [16], but the in vitro activity of desfuroylceftiofur can be predicted by the content of ceftiofur according to a previous study [17]. In our study, we determined the effect of ceftiofur, instead of desfuroylceftiofur, against 135 Actinobacillus pleuropneumoniae strains. The MIC values ranged from 0.0075 µg/mL to 4 µg/mL, and the MIC50 and MIC90 values were 0.015 µg/mL and 0.5 µg/mL, respectively, which were slightly different from the results of ceftiofur against 50 Actinobacillus pleuropneumoniae strains (MIC ranged from 0.0039 − 0.0015 µg/mL) in Salmon’s study.
The epidemiologic cutoff was established based on the pretest of drug susceptibility and then analyzed by statistical methods. At present, CLSI and EUCAST guidelines mainly involve analysis by nonlinear regression, which was proposed by Turnidge [18]. EUCAST based on the nonlinear regression complied the ECOFFinder software[19] the principle of which is the same as Turnidge’s method. In this study, the epidemiologic cutoff was simulated by ECOFFinder. When the MIC value was 0.125 µg/mL, the difference between the actual value and predicted value was smallest, and the confidence interval was 99.9%. The cutoff value would encompass at least 95% of the wild-type isolates [20], Therefore, the epidemiologic cutoff was 0.125 µg/L according to the ECOFFinder results.
There are currently 16 serotypes of Actinobacillus pleuropneumoniae [2, 21]. The prevalent serotypes are different all over the world. The main prevalent serotypes in Canada are serotypes 1, 3 and 5, those in Europe are serotypes 1, 2, 5, 7 and 9 [22], and those in the United States are serotypes 1, 5 and 7 [23]. However, serotypes 1, 3 and 7 were the main prevalent serotypes in China. Moreover, serotype 1 was more virulent than other serotypes[24, 25] Therefore, the strain App BW39 of serotype 1 was selected in this experiment, and the MIC90 was calculated. According to the MIC and MBC values in vitro and ex vivo, the MBC/MIC ratio was 2 and 1, respectively. Therefore, ceftiofur is a bactericidal drug with an MBC/MIC ratio < 4 [26]. According based on the killing curve in vitro and ex vivo, ceftiofur achieved a maximum bactericidal effect at a concentration of 4–8 MIC. The time-dependent antimicrobial activity was defined as achieving the maximum bactericidal effect at a concentration of 3–4 MIC [27]. The results suggest that ceftiofur has a time-dependent action against Actinobacillus pleuropneumoniae both in vitro and ex vivo. According to the killing curve in vitro, when the drug concentration was 1/2 MIC-1 MIC, bacterial growth was not inhibited. Concentrations between 1 MIC and 2 MIC showed a slight inhibitory effect on the bacterial growth. Ceftiofur achieved a maximum bactericidal effect from 4 MIC-8 MIC. The ex vivo killing curve showed that the plasma collected between 0.33 and 2 h achieved the maximum bactericidal effect at the highest ceftiofur concentration. When the concentration was below the MIC at 96 h, ceftiofur no longer inhibited the growth of bacteria. The in vitro and ex vivo killing curves showed the same situation: the concentration from 4 MIC-8 MIC achieved the maximum bactericidal effect. When the concentration increased, the bactericidal effect no longer increased. These findings suggest that ceftiofur has a time-dependent action against Actinobacillus pleuropneumoniae both in vitro and ex vivo.
To establish a pharmacodynamic cutoff, we needed to obtain the PK-PD parameters, MIC distribution and pharmacodynamic targets. The most appropriate PK/PD index to determine the antibacterial efficacy and predict the therapeutic efficacy for β-lactam antibiotics is the percentage of duration that the concentration of free drug is greater than the MIC (%T > MIC), especially for multiple modalities of drug administration [28]. The AUC/MIC ratio is the most appropriate index when the terminal half-life is relatively long relative to the dosage interval [29, 30]. In addition, the different parameters based on the values of T > MIC and bacterial count (log10 CFU/mL) cannot be acquired using ex vivo PK/PD integration, and the index of the AUC/MIC ratio was also used to describe the character of the antibacterial activity of time-dependent killing with prolonged persistent efficacy [31]. For these reasons, the AUC/MIC parameter was regarded as the most appropriate PK/PD index to describe the antibacterial activity of ceftiofur in serum. An AUC/MIC ratio > 125 h is generally considered the best activity indicator [15, 32]. However, the target parameters might be different for different drugs, target organisms or bacteria(Ahmad, 2015). For example, the AUC24 h/MIC values for three levels of preventive, therapeutic, and bacterial eradication were 20.9, 45.2 and 71.7 h, respectively, in previous study on the effects of florfenicol against Actinobacillus pleuropneumoniae [33]. In another study, the AUC24 h/MIC values for three levels of preventive, therapeutic, and bacterial eradication were 24.6, 43.8 and 58.4 h, respectively [34]. However, prior to our study, no PK-PD integration modeling analyses were performed for ceftiofur against Actinobacillus pleuropneumoniae. In our study, the inhibitory sigmoidal Emax model was employed for the simulation of the PK/PD integration model to obtain the AUC24 h/MIC value. The ex vivo values of AUC24 h/MIC of ceftiofur against bacteriostatic, bactericidal and eradication were 45.73, 63.83 and 69.04 h, respectively. Through Monte Carlo simulations, when the MIC values were 2 µg/mL and 4 µg/mL, the probability of target attainment was 100% and 88.94%, respectively. Therefore, the COPD was calculated as 2 µg/mL.
PD-PD integration modeling can be used to select rational dose regimes in veterinary medicine[15]. The ex vivo AUC24 h/MIC ratios of ceftiofur achieving bacterial and eradication effects against BW39 were 63.83 and 69.84, respectively, which were much lower than the in vivo AUC24 h/MIC value (358.84 ± 57.42 h) after intramuscular injection administration (5 mg/kg). This demonstrated that the recommended dosage of 5 mg/kg could guarantee clinical efficacy against A. pleuropneumoniae with an MIC90 of 1 µg/ml. Based on the dose estimation equation and Monte Carlo simulation, the ‘fu’ in the equation was the free fraction of the drug in plasma. Ceftiofur had a high protein binding rate, which was 90% in cow and adult cattle [35, 36],The EMA[37] reported that the protein rate of ceftiofur in swine was 0.7, which was applied in this research. The predicted daily doses for the 50% and 90% targets of ceftiofur with bactericidal activity against Actinobacillus pleuropneumoniae were 2.97 and 3.21 mg/kg body weight, respectively. The Monte Carlo simulation combined with the dosage equation and taking into consideration the PK/PD integration could set the target percentage to simulate models for all data in relation to incidence in swine [38, 39].