Modification of Enrofloxacin Treatment Regimens for Poultry Experimentally Infected with Salmonella enterica Serovar Typhimurium DT104 To Minimize Selection of Resistance

Bacterial strains.Strains A (fully antibiotic sensitive) and B (isogenic multiple-antibiotic-resistant [MAR] mutant of A) of S. enterica serovar Typhimurium DT104 described previously were used throughout (30). Two nonisogenic gyrA mutants (strain C, VLA 5120/98, cyclohexane sensitive, ciprofloxacin MIC of 0.25 μg/ml, and strain D, VLA 2590/99, cyclohexane tolerant, ciprofloxacin MIC of 1 μg/ml) were also used for in vitro studies only. For enrofloxacin (ciprofloxacin), the MICs/MPCs for strains A, B, and C were 0.03/0.5 (0.03/0.25) μg/ml, 0.06/4 (0.06/2) μg/ml, and 0.5/4 (0.5/1) μg/ml, respectively (29).

Strains NCTC 10418, AG100, and AG102 were used as controls for determination of MICs (2) and cyclohexane tolerance (5). Cultures were grown overnight at 37°C in Luria-Bertani (LB) broth prior to infection of chicks, MIC determinations, and time-kill studies.

Antimicrobials, chemicals, and sera.Enrofloxacin (Baytril 10% oral solution) for dosing poultry was obtained from the National Veterinary Services, United Kingdom. Ciprofloxacin and enrofloxacin for MIC determinations and high-pressure liquid chromatography (HPLC) were kindly donated by Bayer HealthCare AG (Germany). Other antibiotics, organic solvents, and chicken sera for in vitro studies were obtained from Sigma-Aldrich (Poole, Dorset, United Kingdom).

Activity of enrofloxacin in vitro.The effect of increasing concentrations of enrofloxacin was determined in vitro against Salmonella in cecal contents, sera, and physiological (0.85%) saline. Cecal contents were collected from healthy untreated birds at necropsy. Serum (2 ml) was inoculated with ∼10⁵ CFU/ml (to correspond to the level of Salmonella in a severe systemic infection) of strain A, B, C, or D; cecal contents (2 g) and saline (2 ml) were inoculated with ∼10⁸ CFU/gram/ml (to correspond to the upper level of Salmonella in cecal contents of infected birds) of strain A, B, or D. For serum, enrofloxacin was added at 0 (control), 0.5, 1, 2, 4, and 8 μg/ml (to correspond to the determined concentration range in sera following recommended and modified dosage regimens). For cecal contents and saline sample, enrofloxacin was added at 0, 12.5, 25, 50, 100, and 200 μg/ml (to correspond to the determined concentration range in cecal contents following recommended and modified dosage regimens). The effect of enrofloxacin against strain A or B was determined only for cecal contents. Samples were incubated at 41.5°C (the normal body temperature of chickens), and viable counts of Salmonella were performed on selective agar (Rambach agar; Merck, Darmstadt, Germany) by the method of Miles et al. (27) at 0, 3, 6, and 24 h after addition of enrofloxacin.

Binding of enrofloxacin to cecal contents and sera.Sera and saline (as control) were spiked with 2 and 20 μg/ml and cecal contents with 20 and 200 μg/ml enrofloxacin and allowed to stand for 1 hour at 37°C (for binding equilibrium to occur). Complexes were then centrifuged for ∼30 min (until enough filtrate was collected) in a Vivaspin 15 (Vivascience AG, Hannover, Germany) column with a molecular mass cutoff of 10,000 Da. Original samples and filtrates were frozen at −80°C prior to quantification of enrofloxacin (and ciprofloxacin if present) by HPLC.

Dose calculations and routes of administration.Current data sheets for Baytril 10% oral solution recommend adding enrofloxacin to the drinking water, based on the calculated amount of water drunk in the house, to give 10 mg/kg of body weight of birds per day for 5 to 10 days for treatment of salmonellosis in poultry. However, previous data sheets (2003) recommended a dose of 50 ppm enrofloxacin added to the drinking water each day for 5 to 10 days; the two approaches are considered equivalent. Routes of administration were compared and included the data sheet-recommended method (continuously medicated water) compared to pulsed water treatment (“breakfast regimen”; dose given to birds before lights come on and before they start eating or drinking) and dosing by oral gavage.

Antibiotic dose was calculated as mg/kg per individual bird (gavage dosing) or mg/kg per group of birds given in one-third of the previously calculated amount of water drunk daily (pulsed dosing) or water at 50, 125, or 250 ppm enrofloxacin (continuous dosing). For the pulsed dose, once the water was drunk, it was replaced with fresh water. Birds received antibiotic at the beginning of each day before the lights in their accommodation came on.

Chick experiments—efficacy.All animal studies were conducted under the jurisdiction of the Animals Scientific Procedures Act (1986) and were reviewed by the local ethical review committee. For experiments I, II, III, and IV, 95, 150, 103. and 100 specific-pathogen-free (SPAFAS chicks specially reared and checked to be free of specific pathogens such as Salmonella) White Leghorn chicks (1 day old) were randomly separated into the appropriate number of groups for the experiment (Table 1). Each group of birds was housed together, but separately from other groups, with appropriate biosecurity between groups. Chicks received feed and water ad libitum and were monitored for condition twice daily throughout the experiments.

At 3 to 4 weeks of age birds (Table 1 shows bird numbers per group) were infected by gastric gavage (10) with a standardized inoculum of ∼10⁶ CFU per bird of S. enterica serovar Typhimurium strains A and B in a 1:1 mix in phosphate-buffered saline (0.1 ml of phosphate-buffered saline containing ∼10⁷ CFU/ml Salmonella), although in experiment III the birds received a lower (∼10⁴ CFU per bird) dose in order to investigate response to antibiotic treatment in birds with a lower colonization level. Strains A and B were given together to try to establish if antibiotic treatment selected strain B (the isogenic MAR mutant) in preference to strain A. In order to neutralize the crop acid, chicks were dosed orally by gavage with 0.2 ml of 10% sodium bicarbonate approximately 30 min prior to dosing by oral gavage with 0.1 ml of the Salmonella solution (7).

One week postinfection birds were treated with the different treatment regimens of enrofloxacin (Baytril 10% oral solution). In each experiment the 5-day data sheet-recommended treatment for salmonellosis (10 mg/kg or 50 ppm) was compared with 2 days at 125 ppm or 25 mg/kg. In experiments I, II, and III, comparisons were also made to a 1-day treatment at 250 ppm or 50 mg/kg. The shorter courses of antibiotic overall delivered the same amount of antibiotic as the 5-day treatment regimen. In experiment II half of the birds for each treated group were moved to clean accommodation 6 hours after the last antibiotic dose to determine if this affected recolonization of the birds and the extent of antibiotic resistance seen post-antibiotic treatment. The Salmonella organisms present before, during, and after antibiotic treatment were enumerated by both semiquantitative (cloacal swabs) and quantitative (enumeration of Salmonella organisms from 1 g of cecal contents) methods as described in our previous study (30).

Chick experiments—pharmacokinetics.To determine the antibiotic concentrations in cecal contents, liver, and sera following the different antibiotic gavage treatments in experiment II (Table 1), birds were killed at 3, 6, and 24 h (four birds per group) after antibiotic treatment.

To determine the pharmacokinetics of continuous versus pulsed water treatments, a separate experiment was performed. Eighty-day-old specific-pathogen-free White Leghorn chicks were randomly divided into four groups of 20 birds, and at 4 weeks of age the groups were treated with the continuous or pulsed water treatments at the recommended dose and 2.5× the recommended dose. Three birds were killed from each group at 2, 4, 6, 8, 10, and 24 h after the start of antibiotic treatment.

Samples of cecal contents, liver, lung, and sera taken were frozen at −80°C for subsequent analysis by HPLC. The concentrations of enrofloxacin and its metabolite ciprofloxacin in these samples were measured by HPLC (9) based on a method described previously (40).

Basic pharmacokinetic parameters such as AUC, half-life, and mean residence time were calculated using PK-Solutions software (Summit Research Services, Montrose, CO).

Replica plating, determination of MICs, cyclohexane tolerance, and gyrA mutations.To determine the emergence of strains with reduced susceptibility during and after antibiotic treatment, replica plating was performed as previously described (30) but only onto medium with 4× MICs (for strains A and B) of nalidixic acid or ciprofloxacin or medium overlaid with cyclohexane to detect the presence of MAR Salmonella (5, 28). Isolates obtained from postmortem counts, swabbing controls, and all treatment groups of birds in all experiments at 1, 7, 14, 21, and 28 days post-antibiotic treatment were replica plated. Additionally, some isolates from before treatment and during treatment (experiments I and II only) were replica plated.

MICs of ciprofloxacin, nalidixic acid, and tetracycline were determined by the method of the British Society for Antimicrobial Chemotherapy (2) against posttreatment isolates (n ∼ 40 per experiment) selected from all treatment groups and time periods. Isolates were deliberately selected to include fully sensitive strains and strains with reduced susceptibility.

Fifty strains from experiment II were tested for mutations in the quinolone resistance-determining region of gyrA as previously described (31).

Statistical analyses.For animal experiments, there is an ongoing requirement to use as few animals as possible. The numbers of chickens that were used in each experiment were based on numbers used in previous studies where it was possible to show statistically significant differences between Salmonella counts in enrofloxacin-treated and untreated birds (30) or on the numbers used in previous pharmacokinetic studies with enrofloxacin in animals (40).

The Salmonella counts were transformed to their logarithm to base 10. One-way or two-way analyses of variance according to the design were performed to test for treatment effects. In some of the experiments small counts were recorded as an upper limit; these were regarded as left-censored observations, and the appropriate censored regression model was fitted instead. Following a significant F test (P < 0.05) for the treatment effects or interactions in the analyses of variance, treatment contrasts were tested using t tests. Likelihood ratio tests were used for the corresponding tests in the censored regression models. Where the majority of pairwise treatment differences were of interest, Bonferroni adjustments were made to the P values. In one experiment where all of the observations for a treatment were censored, the Kruskal-Wallis nonparametric analysis of variance was followed by Dunnett's test for comparisons with a control treatment.

In vitro activity of enrofloxacin.In sera at all concentrations tested, enrofloxacin rapidly killed the fully sensitive strain A (reduction of >10⁴ CFU/ml in 3 h), although at 24 h there was some regrowth in the presence of 0.5 and 1 μg/ml (Fig. 1A). The activity of enrofloxacin was progressively less for strains B, C, and D (Fig. 1B, 1C, and 1D, respectively).

• Open in new tab
• Download powerpoint

FIG. 1.

Effects of various concentrations of enrofloxacin on survival of Salmonella serovar Typhimurium strains in chicken sera at 41°C. Counts are means of two experiments, and error bars are standard errors of the means. The dotted line is the limit of detection. A. Parent strain A (fully sensitive; ciprofloxacin MIC, 0.03 μg/ml). B. MAR strain B (isogenic mutant of strain A; ciprofloxacin MIC, 0.6 μg/ml). C. VLA 512/98 (strain C, nonisogenic gyrA mutant, cyclohexane sensitive; ciprofloxacin MIC, 0.25 μg/ml). D. VLA 2590/99 (strain D, nonisogenic gyrA mutant, cyclohexane tolerant; ciprofloxacin MIC, 1 μg/ml). Symbols: filled diamonds, control; open squares, 0.5 μg/ml enrofloxacin; open triangles, 1.0 μg/ml enrofloxacin; crosses, 2.0 μg/ml enrofloxacin; open diamonds, 4.0 μg/ml enrofloxacin; open circles, 8.0 μg/ml enrofloxacin.

In saline, enrofloxacin at 50 μg/ml at 6/24 h reduced the counts for strains A, B, and D to 0.025%/0.001%, 0.23%/0.023%, and 6.55%/1.04%, respectively, of the control counts at that time (results not shown).

Enrofloxacin at 50 μg/ml in cecal contents at 6 h did not reduce the counts for strains B and D (compared to controls) and reduced the counts for strain A only to 80% of control (results not shown). At 24 h there were no viable Salmonella organisms in control cecal contents in vitro. The pH of the cecal contents was in the range of 4.6 to 5.2.

When enrofloxacin was tested at 0, 12.5, 25, 50, 100, and 200 μg/ml against strain A or B only in cecal contents, there was no increased enrofloxacin activity with increasing concentrations of enrofloxacin as seen with sera (results not shown).

Binding of enrofloxacin to cecal contents and sera.Enrofloxacin (mean ± standard deviation) was bound 45.3% (±9.7%) and 54.8% (not determined) to cecal contents fortified with 20 and 200 μg/ml enrofloxacin, respectively, and 56.4% (±31.8%) and 50.1% (±1.6%) to sera fortified with 2 and 20 μg/ml enrofloxacin, respectively.

Calculated dose and dose uptake time.For birds that received the dose continuously in the drinking water at 50 and 125 ppm, the mean dose (± standard deviation) that they actually received when calculated for two separate days of dosing was 14.8 (±0.71) mg/kg and 32.3 (±0.42) mg/kg, respectively. On the basis of these results, the birds that received the 50-ppm dose received 1.48× the dose of those that received the 10-mg/kg dose and the birds that received the 125-ppm dose received 1.29× the 25-mg/kg dose. For the birds that received the dose pulsed in the water, the dose was drunk in ca. 4 h.

Efficacy of the different enrofloxacin treatment regimens.Infecting the birds with Salmonella did not make them clinically ill, and so efficacy was related to the ability to reduce Salmonella counts in cecal contents. The different 1-day treatments at 5× the recommended treatment concentration did not reduce the Salmonella counts to the same extent as the 2- or 5-day treatments did but in some instances still significantly reduced the Salmonella counts compared to bacterial counts in control birds (Table 1). The 2-day treatments significantly reduced the Salmonella counts compared to counts in control birds (P < 0.01), and in all but one treatment group, the 2-day treatment showed reductions similar to those of the 5-day treatment.

Pharmacokinetics.The higher doses of antibiotic resulted in a proportionate increase in antibiotic concentration in sera and tissues (Tables 2 and 3; Fig. 2). The mean peak antibiotic concentrations obtained following different dosage routes were gavage > pulsed water > continuous water treatment, and this corresponded to a shorter time to reach the peak antibiotic concentrations and less time that antibiotic was present in sera or tissues for gavage or pulsed treatments compared to antibiotic given continuously in the water (Table 2; Fig. 2).

• Open in new tab
• Download powerpoint

FIG. 2.

Graphs showing mean (n = 3) fluoroquinolone (enrofloxacin and ciprofloxacin) concentrations in chicken sera and tissues at 0, 2, 4, 6, 8, 10, and 24 h after dosing with Baytril 10% oral solution. (As enrofloxacin can be metabolized to ciprofloxacin in vivo, a percentage of the antibiotic detected was ciprofloxacin.) (A) Cecal contents; (B) liver; (C) lung; (D) sera. Values are for total antibiotic. Symbols: dashed lines and squares, continuous dosing; solid lines and circles, pulsed dosing; open squares or circles, recommended dose; filled squares or circles, 2.5× standard dose.

Some basic pharmacokinetic parameters are shown in Table 3 for pulsed and continuous water treatments at recommended and 2.5× the recommended treatment dose only. As would be expected, the pulsed treatment led to higher C_max/MIC ratios but to slightly lower AUC/MIC ratios.

Recolonization of chicken gut post-antibiotic treatment.Overall, there were no statistical differences in the extent of Salmonella recolonization after the end of antibiotic treatment for the different treatment groups within an experiment, including for birds that were moved to clean accommodation 6 h after their last antibiotic dose in experiment II (results not shown). Moving the birds to clean accommodation also did not affect whether the birds were shedding reduced-susceptibility Salmonella organisms or not (results not shown). However, there was considerable variation in the extent of recolonization between different experiments; in particular in experiments I and III, at 4 weeks after the end of antibiotic treatment, mean counts for different treatment groups were all ≤log 1.5, whereas in experiments II and IV mean counts for different treatment groups ranged from log 3.6 to 6.8 (results not shown).

Selection of Salmonella isolates with reduced susceptibility.The number of colonies replica plated and the percentages of strains with reduced susceptibility to ciprofloxacin for each treatment group and experiment are shown in Table 1. None of the control (Table 1) or pretreatment (results not shown) birds gave rise to strains with reduced susceptibility. The cyclohexane-tolerant MAR strain B was present in high numbers during the first few days of antibiotic therapy, but then the numbers rapidly decreased (results not shown).

In experiment III, where the lower dose of Salmonella was given to birds, there were no strains with reduced susceptibility isolated (Table 1).

Excluding isolates from experiment III, the numbers of colonies replica plated from all experiments and treatment routes after the end of antibiotic treatment were 5,025 for the 5-day treatments, 7,756 for the 2-day treatments, and 5,088 for the 1-day treatments. The percentages of these colonies/isolates that showed reduced susceptibility (i.e., ability to grow on agar with 4× MICs for the respective strains) to ciprofloxacin were 64% for the 5-day treatment, 52% for the 2-day treatment, and 14% for the 1-day treatment (results not shown).

MICs of ciprofloxacin against Salmonella isolates recovered from different treatment groups.Based on the MICs against selected isolates, the 5-day treatments tended to select for strains with a higher level of reduced susceptibility than the 2- or 1-day treatment regimens. The ciprofloxacin MIC₁₀₀ was 0.03, 0.5, 0.5, and 2 μg/ml for isolates (n = 12, 35, 57, and 66) from control groups and from all of the 1-, 2-, and 5-day treatment regimens, respectively (results not shown). Compared with the pulsed or continuous regimens, the gavage regimen appeared to select for strains with a lower level of reduced susceptibility with MIC₁₀₀s of 0.25, 0.06, and 0.5 μg/ml for isolates (n = 16, 13, and 24) from the gavage 1-, 2-, and 5-day treatment regimens, respectively (results not shown).

Determination of mutations within the quinolone resistance-determining region of gyrA.Fifty isolates from experiment II were examined for mutations in gyrA by denaturing HPLC. Isolates were taken from all treatment groups and the control group and from birds during (n = 21 isolates) and after (n = 29 isolates) antibiotic treatment. Isolates were wild type (n = 24) or had the following mutations: Asp87-Asn (n = 7), Asp87-Tyr (n = 8), and Ser83-Phe (n = 11). There was no evidence that specific mutations were associated with specific treatment regimens. Most of the strains which had mutations in gyrA and were also cyclohexane resistant had an Asp87-Asn mutation (seven/eight strains). The presence of a mutation in gyrA was associated with reduced susceptibility to ciprofloxacin and resistance to nalidixic acid. No isolates with mutations in gyrA were obtained from control birds.