- Email: [email protected]
The pharmacodynamic effect of amoxycillin and danofloxacin againstActinobacillus pleuropneumoniaein an in-vitro pharmacodynamic model
Research in Veterinary Science 1999, 67, 93–97 Article No. rvsc.1998.0307, available online at http://idealibrary.com on
The pharmacodynamic effect of amoxycillin and danofloxacin against Actinobacillus pleuropneumoniae in an in-vitro pharmacodynamic model R. H. LINDECRONA*, C. FRIIS,* N. E. JENSEN† *Department of Pharmacology and Pathobiology. Royal Veterinary and Agricultural University and † The Danish Veterinary Laboratory, Copenhagen, Denmark. SUMMARY The pharmacodynamic effect of amoxycillin and danofloxacin against two strains of Actinobacillus pleuropneumoniae was evaluated in an in-vitro pharmacodynamic model. For amoxycillin peak concentrations of 0·5, 1, and 4 µg ml–1 and half-lives of 3 and 15 hours were examined. For danofloxacin peak concentrations of 0·125, 0·5, and 1·5 µg ml–1 and half-lives of 1·5 and 7 hours were evaluated. The initial bactericidal effect was measured as the reduction in colony count (log CFU ml–1) during the first three hours, and the overall pharmacodynamic effect as the area under the bacterial growth versus time curve (AUBC). The initial bactericidal effect of amoxycillin was maximal at peak concentrations of two to four times the MIC. Peak concentration and half-life only influenced the pharmacodynamic effect of amoxycillin if the antibiotic concentration fell below the MIC during the experiments, which is consistent with time > MIC as the most important parameter of pharmacodynamic effect of β-lactam drugs. For danofloxacin maximal bactericidal effect initially was observed at peak concentrations of at least eight times the MIC. The pharmacodynamic effect was dependent on the peak concentration. The half-life only influenced the pharmacodynamic effect of danofloxacin in experiments with a peak concentration MIC ratio of less than eight. This indicated that for danofloxacin the peak concentration was the major determinant of pharmacodynamic effect.
IN-VITRO pharmacodynamic infection models simulating antibiotic pharmacokinetics in vivo can provide useful information about the optimal dosing regimen of antimicrobials in clinical infections (Blaser and Zinner 1987). Many studies have been performed in in-vitro pharmacodynamic models with β-lactam drugs, aminoglycosides, fluoroquinolones and their combinations against human pathogens (Blaser et al 1987, Chambers et al 1991, Dudley et al 1991, Lister et al 1997, Palmer and Rybak 1996). So far only few studies have been carried out with veterinary pathogens in in-vitro pharmacodynamic models. Actinobacillus pleuropneumoniae is a major cause of swine pneumonia and causes great economic losses in the swine production worldwide. Treatment with antibiotics is an important tool in the control of this disease. However, little is known about the optimal dosing regimen. The purpose of this study was to evaluate the pharmacodynamic effects of amoxycillin and danofloxacin against two strains of Actinobacillus pleuropneumoniae in a one-compartment in-vitro pharmacodynamic model, simulating in-vivo drug concentrations of the drugs in bronchial secretions in the pig (Friis and Nielsen 1997, Agersø and Friis 1998a,b).
MATERIALS AND METHODS Bacterial strains Two strains of Actinobacillus pleuropneumoniae, biotype 1, serotype 2 (s4226 and BS9505611) were studied. Both strains were isolated from clinical outbreaks of pleuropneumoniae. Antibiotics Amoxycillin and danofloxacin were kindly supplied by the manufacturers (SmithKline Beecham, Tadworth, Surrey, UK and Pfizer Inc., New York, USA, respectively). Medium The medium used for susceptibility testing and model experiments was Tryptic Soy Broth (BBL, Becton Dickenson & Co., Cockeysville, MD, USA) supplemented with 10 g yeast extract per litre (Oxoid, Unipath Ltd. Basingstoke, UK). β-NAD (Sigma) was supplemented to a concentration of 0·02 per cent. In-vitro model
Corresponding author address: R. H. Lindecrona, The Danish Veterinary Laboratory, Bülowsvej 27, 1790 Copenhagen V, Denmark. 0034-5288/99/040093 + 5 $18.00/0
A schematic representation of the in-vitro pharmacodynamic model used in the experiments is presented in Fig 1. The in-vitro model consisted of a 300 ml magnetic stirred cell (Millipore) filled with 200 ml broth. The broth was © 1999 Harcourt Publishers Ltd
R.H. Lindecrona, C. Friis, N.E. Jensen
94
Flow rate F=Va *Ke
Diluent reservoir
Central reservoir
Elimination reservoir
FIG 1: Schematic representation of the one-compartment in-vitro model used in the experiments. A peristaltic pump is pumping antibiotic-free medium from the diluent reservoir into the central compartment. Simultaneously medium is pumped out of the central reservoir into the elimination reservoir at the same flow rate (F).
continuously stirred during the experiment. Elimination of the antimicrobial agents was achieved using a peristaltic pump (Masterflex) to supply fresh antibiotic-free medium to the compartment resulting in the simultaneous displacement of antibiotic-containing medium. The flow rate of the pump was set according to the half-life being simulated (Grasso et al 1978). Elimination of the bacteria was prevented by a filter (pore size 0·2 µm, Millipore) and a prefilter placed in the bottom of the compartment. Samples were taken through a silicone membrane. Atmospheric air diffused into the compartment through a filter (pore size 0·2 µm, Millipore). Antibiotics were delivered as a bolus into the compartment in doses resulting in initial drug concentrations from 0·5 to 4 µg ml–1 for amoxycillin and from 0·125 to 1·5 µg ml–1 for danofloxacin. Bacteria for inoculation were grown overnight at 37°C on meat and blood agar supplemented with 0·07 per cent β-NAD as previously described (Jacobsen and Nielsen 1995). Three colonies were suspended in 3 ml broth and incubated for one hour at 37°C. From this broth culture a McFarland standard of 0·5 in sterile saline was prepared. Two ml of this suspension were inoculated into the model resulting in a starting inoculum of 1 to 3 × 106 CFU ml.–1 The bacteria were incubated in the model for one hour before addition of the antibiotic. Experiments were run for 8 hours for strain s4226 and for 12 hours for strain BS9505611. All experiments were performed in triplicates.
administration of the antibiotic dose for the determination of of viable bacteria per millitre Serial tenfold dilutions were plated in triplicate on meat and blood agar supplemented with 0·07 per cent β-NAD. The plates were incubated for 24 to 36 hours at 37°C and the colonies counted. At times when the bacterial count was expected to be below the limits of detection or when antibiotic carry-over was expected to be a problem, 100-µl samples were diluted in 10 ml of 0·9 per cent saline and filtered using a Millipore system (0·45 µg pore size filters). Filters were placed aseptically on meat and blood agar and incubated for 24 to 36 hours, at which time the colonies were counted. The lower limit of detection was set to be 102 CFU/ml. The area under the bacterial growth versus time curve (AUBC) was calculated according to the trapezoidal rule up to the last sampling point (Gibaldi and Perrier, 1982). For the bacteria isolated at the beginning and at the end of the experiment MICS were determined in duplicate by a standard broth microdilution test (Sensititre®) as outlined by the NCCLS (1997). CFU
Statistical analysis The effect of peak concentration and half-life on the AUBC and the absolute change in log CFU/ml during the first 3 hours was evaluated by a two-way analysis of variance. P<0·05 was considered significant.
Drug determination Samples were collected from the compartment immediately after injection and at 1, 2, 3, 4, 6, 8, and 12 hours during the experiment and were stored at –20°C until assayed. The concentration of amoxycillin and danofloxacin was determined by high-pressure liquid chromatography (HPLC) as described by Agersø and Friis (1998a,b) and Friis (1993) respectively. Pharmacodynamic analysis Samples were collected from the compartment before injection and at 0·5, 1, 2, 3, 4, 6, 8, and 12 hours after
RESULTS In-vitro susceptibility tests The MIC values for strain s4226 were 0·25 µg ml–1 for amoxycillin and 0·06 µg ml–1 for danofloxacin. For strain BS9505611, MIC values were 0·25 0·50 µg ml–1 for amoxycillin and 0·03 µg ml–1 for danofloxacin. For the experiments no change in MIC was observed except in one case. For strain s4226, the regimen with a peak concentration of danofloxacin of 0·125 µg ml–1 and a half-life of 7 hours, resulted in a four to eight times rise in MIC for the bacteria
Effect of amoxycillin and danofloxacin against A. pleuropneumoniae
The antibiotic concentrations during the experiment were fitted to a one-compartment pharmacokinetic model and selected pharmacokinetic parameters were calculated according to standard equations (Gibaldi and Perrier, 1982). For amoxycillin, mean initial concentrations (µg ml–1 ± SD) were 0·64±0·15, 1·15±0·39 and 3·96±0·74 and mean half-lives were (hours ±SD) 14·6±5·8 and 2·9±0·6. Mean initial danofloxacin concentrations (µg/ml ±SD) were 0·17±0·07, 0·57 ± 0·12 and 1·58±0·26 and mean half-lives were (hours ±SD) 6·6 ±1·1 and 1·2±0·3. Pharmacodynamics Amoxycillin. Mean growth curves for the amoxycillin experiments are presented in Fig 2. For strain s4226 no regrowth was seen within the 8-hour experiment period in any of the experiments. For strain BS9505611 less than a 1-log reduction of bacterial count was seen followed by regrowth of bacteria in the experiments with a peak concentration of 0·5 µg ml–1 and a half life of 3 hours. In these experiments the drug concentration was below the MIC throughout most of the study period. At the end of the 12-hour experiment period some regrowth was observed for all experiments except for the regimen with a peak concentration of 4 µg ml–1 and a half-life of 15 hours. The reductions in bacterial count after 3 hours are presented in Table 1. The effect of half-life and peak concentration on the initial reduction in bacterial count was significant for both strains. In some experiments a change in half life from 15 to 3 hours resulted in a reduced bactericidal effect. A maximal reduction in bacterial count was achieved with a peak concentration of 1 µg ml.–1 The AUBCS are listed in Table 2. For both strains the difference in the AUBC was significant (P <0·05) among the various peak concentrations. The main difference was seen in the experiments with the short half-life of 3 hours, where the lowest peak concentration of 0·5 µg ml–1 resulted in a larger AUBC than the higher peak concentrations. The influence of half-life on the AUBC was strongly significant (P<0·002) for strain BS9505611 but not for strain s4226. However, for the experiments with this strain the drug concentration exceeded the MIC throughout most of the study period. Danofloxacin. For the danofloxacin experiments mean growth curves are illustrated in Fig 3. The reductions in log CFUml from 0 to 3 hours are presented in Table 3. An effect of peak concentration was seen for both strains (P<0·05). For all experiments a rapid reduction in colony count within the first 3 hours was observed at a peak concentration of at least eight times the MIC. A change in half-life from 7 hours to 1·5 hours had no effect on the initial bactericidal activity, except for the
Log CFU ml–1
Pharmacokinetics
Strain s4226
8 6 4 2 0 2
0
4
6
8
Time (h) Strain BS9505611 8 Log CFU ml–1
isolated at the end of the experiment. However, the MIC fell to the original level upon passage on antibiotic-free meat and blood agar.
95
6 4 2 0 0
2
4
6 8 Time (h)
10
12
FIG 2: Amoxycillin versus two strains of Actinobacillus pleuropneumoniae in an in-vitro pharmacodynamic model. Results are means of triplicate experiments. Peak conc.: 0·5 µg ml–1 (●), 1 µg ml–1(■), 4 µg ml–1 (▲), t1/2=15 hours (full symbol), t1/2=3 hours (open symbol), control (+).
experiments with strain s4226 and a peak concentration of 0·5 µg ml–1, where a marked reduction in killing activity was observed. The AUBCS are listed in Table 4. The effect of peak concentration on the AUBC was strongly significant with the highest peak concentrations resulting in the smallest AUBC. The half-life only influenced the AUBC in two cases. For strain s4226, the experiments with a peak concentration of 0·5 µg/ml and for strain BS9505611, a peak concentration of 0·125 µg ml–1 a significant smaller AUBC was observed for the experiments with a longer half-life compared to the experiments with a shorter half-life.
DISCUSSION For amoxycillin, the peak concentrations ranging from 0·5 to 4·0 µg ml–1 did not influence the pharmacodynamic effect as long as the antibiotic concentration exceeded the MIC during the greater part of the experiment. These observations are in agreement with previous findings both in vivo and in vitro, where the time that the antibiotic concentration exceed the MIC (T > MIC) is recognised as the most important parameter for antibacterial efficiency of the β-lactam drugs (Drusano 1991, Craig and Ebert 1992). The bactericidal effect of amoxycillin reached a maximum at peak concentrations of two to four times the MIC of the actual strain. This observation has also been reported from in-vitro studies with penicillin against Streptococcus pneumoniae (Knudsen
R.H. Lindecrona, C. Friis, N.E. Jensen
96
TABLE 1: Reduction in log CFU/ml from 0 to 3 hours for experiments with amoxycillin against two strains of Actinobacillus pleuropneumoniae in an in-vitro pharmacodynamic model. Values are mean of three experiments ± SD.
TABLE 2: Area under the bacterial growth versus time curve for experiments with amoxycillin against two strains of Actinobacillus pleuropneumoniae in an in-vitro pharmacodynamic model (experiments were run for 8 hours for strain s4226 and for 12 hours for strain BS9505611). Results are mean ±SD of three experiments.
Peak concentration (µg ml–1) Peak concentration (µg ml–1) Strain
t1/2 (hours)
0·5
1·0
4·0
S4226
15 3 15 3
1·9 (0·1) 1·9 (0·1) 2·2 (0·2) 0·3 (0·7)
2·9 (0·3) 2·0 (0·1) 2·4 (0·1) 1·9 (0·6)
2·7 (0·7) 2·5 (0·3) 2·9 (0·2) 1·8 (0·5)
BS9505611
TABLE 3: Reduction in log CFU/ml from 0 to 3 hours for experiments with danofloxacin against two strains of Actinobacillus pleuropneumoniae in an in-vitro pharmacodynamic model. Values are mean of three experiments ±SD.
Strain
t1/2 (hours)
0·5
1·0
4·0
s4226
15 3 15 3
32·2 (2·1) 35·1 (2·0) 45·1 (5·1) 68·3 (9·5)
31·9 (6·6) 31·8 (0·7) 46·7 (5·3) 52·7 (14·2)
27·2 (2·1) 28·4 (2·6) 36·0 (1·1) 50·1 (2·0)
BS9505611
TABLE 4: Area under the bacterial growth versus time curve for experiments with danofloxacin against two strains of Actinobacillus pleuropneumoniae in an in vitro pharmacodynamic model (experiments were run for 8 hours for strain s4226 and for 12 hours for strain BS9505611). Values are mean ± SD of three experiments.
Peak concentration (µg ml–1) Peak concentration (µg ml–1) Strain
t1/2 (hours)
0·125
0·50
1·50
s4226
7 1·5 7 1·5
2·7 (0·5) 2·8 (1·1) 3·7 (0·6) 2·9 (1·0)
4·4 (0·1) 2·5 (0·2) 4·3 (0·1) 3·9 (0·4)
4·0 (0·1) 3·8 (0·7) 4·2 (0·1) 4·2 (0·1)
BS9505611
Strain
t1/2 (hours)
0·125
0·50
1·50
s4226
7 1·5 7 1·5
35·1 (2·8) 31·0 (6·0) 35·9 (2·8) 52·9 (14·6)
17·9 (0·8) 30·7 (1·1) 29·5 (4·5) 33·0 (2·4)
19·0 (0·5) 21·3 (4·9) 27·2 (0·03) 32·2 (5·7)
BS9505611
Va volume of the central compartment; Ke elimination constant of the antibiotic
Strain s4226 Log CFU ml–1
8 6 4 2 0 0
2
4 Time (h)
6
8
Strain BS9505611
Log CFU ml–1
8 6 4 2 0 0
2
4
6 8 Time (h)
10
12
FIG 3: Danofloxacin versus two strains of Actinobacillus pleuropneumoniae in an in-vitro pharmacodynamic model. Results are means of triplicate experiments. Peak conc 0·125 µg ml–1 (●), 0·5 µg ml–1 (■), 1·5 µg ml–1 (▲), t1/2=7 hours (full symbol), t1/2=1·5 hours (open symbol), control (+).
et al 1995). A peak concentration of four times the MIC was needed for the efficacy of ceftazidime against Pseudomonas aeruginosa in an in-vitro pharmacodynamic model (Mouton and den Hollander 1994). Peak concentration as well as the area under the antibiotic concentration versus time curve (AUC) has been mentioned
as important factors of optimal pharmacodynamic effect of fluoroquinolones (Chambers et al 1991, Drusano et al 1993, Leggett et al 1991, Marchbanks et al 1993, Meinen et al 1995, Palmer and Rybak 1996). Several studies have indicated that a high peak concentration/MIC ratio is important to achieve clinical efficiency of fluoroquinolones and minimize problems with emergence of resistance (Chambers et al 1991, Marchbanks et al 1993, Drusano et al 1993). This is in agreement with the observations made in this study. A maximal bactericidal effect and no regrowth within the experimental period was seen at peak concentrations of at least eight times the MIC of the particular strain. Studying the pharmacodynamic effect of enoxacin against several bacterial species in an in-vitro model, Blaser et al (1987) also observed regrowth of bacteria with a MIC of four to 16 times the MIC of the original strain unless the peak concentration/MIC ratio exceeded eight. In our study, a shortening of the half-life (and thereby a reduction in the AUC) had no influence on the antibacterial activity as long as the peak concentration/MIC ratio exceeded eight. Only in experiments with a peak/MIC ratio of four to eight a shorter halflife resulted in a reduced bactericidal activity both initially and during the whole experimental period. This was also observed by Drusano et al (1993), where the AUC only influenced the outcome of a Pseudomonas infection in rats if the peak concentration/MIC ratio was lower than 10. Other studies have pointed to the AUC as the most significant determinant of clinical efficiency of fluoroquinolones. In a murine thigh-infection and pneumonitis model Leggett et al (1991) observed that the frequency of administration of the same daily dose of ciprofloxacin had little impact on the efficacy against Gram-negative bacilli. The same observation has been reported in a study with enrofloxacin against Escherichia coli and Staphylococcus aureus in neutropenic mice (Meinen et al 1995).
Effect of amoxycillin and danofloxacin against A. pleuropneumoniae
The change in MIC observed in this study in experiments with a peak/MIC ratio of two and a half-life of 7 hours might have resulted from a single step mutation in the gene of DNA gyrase or in genes regulating drug permeation (Wolfson and Hooper, 1989). However, these findings were not further investigated. The in-vitro model described in this study provides a useful tool for describing the antibacterial activity of different antibiotics and for predicting the optimal dosing regimen clinically. Moreover this is the first in-vitro study describing the pharmacodynamic effect of antibiotics against Actinobacillus pleuropneumoniae. Running the experiment for a longer period of time will further improve the in-vitro model. However in these experiments it was not possible to run the experiment period for more than twelve hours because of bacterial regrowth. This regrowth may be caused by bacteria adhering to the glass surface of the model as described by Haag et al (1986). In a study by Raemdonck et al. (1994), the in-vitro susceptibilities of Actinobacillus pleuropneumoniae strains isolated from clinical cases of pleuropneumoni were determined. The MIC90 for amoxycillin and danofloxacin were reported to be 0·5 µg ml–1 and 0·125 µg ml–1 respectively. Thus, for amoxycillin, a long half-life and a peak concentration of 1 µg ml–1 is needed at the site of infection to achieve an efficient treatment of Actinobacillus pleuropneumoniae infections. On the assumption that the amoxicillin concentration in the lungs is one third of the plasma concentration (Agersø and Friis, 1998a), a dose of 15 mg kg–1 intramuscularly would result in sufficient concentrations for the treatment of Actinobacillus pleuropneumoniae infections (Agersø & Friis, 1998b). For danofloxacin, a peak concentration of 1 µg ml–1 is the most important determinant of therapeutic efficiency, as well as for minimal emergence of resistance. A maximal danofloxacin concentration of 1 µg ml–1 could be achieved in all parts of the lung tissue following intramuscular administration of a dose of 2·5 mg kg–1 (Friis and Nielsen, 1997).
REFERENCES AGERSØ H. & FRIIS C. (1998a) Penetration of amoxycillin into the respiratory tract tissues and secretions in pigs. Research in Veterinary Science, 64, 245–250. AGERSØ H. & FRIIS C. (1998b) Bioavailability of amoxycillin in pigs. Journal of Veterinary Pharmacology and Therapeutics, 21, 41–46. BLASER J., STONE B.B., GRONER M.C. & ZINNER S.H. (1987) Comparative study with enoxacin and netilmicin in a pharmacodynamic model to determine importance of ratio of antibiotic peak concentration to MIC for bactericidal activity and emergence of resistance. Antimicrobial Agents and Chemotherapy, 31, 1054–1060. BLASER J. & ZINNER S.H. (1987). In vitro models for the study of antibiotic activities. Progress in Drug Research, 31, 349–381. CHAMBERS S.T., PEDDIE B.A., ROBSON R.A., BEGG E.J. & BOSWELL D.R. (1991) Antimicrobial effects of lomefloxacin in vitro. Journal of Antimicrobial Chemotherapy, 27, 481–489.
97
CRAIG W.A. & EBERT S.C. (1992). Continuous infusion of beta-lactam antibiotics. Antimicrobial Agents and Chemotherapy, 36, 2577–2583. DRUSANO G.L. (1991) Human pharmacodynamics of beta-lactams, aminoglycosides and their combination. Scandinavian Journal of Infectious Disease, Suppl. 74, 235–248. DRUSANO G.L., JOHNSON D.E., ROSEN M. & STANDIFORD H.C. (1993) Pharmacodynamics of a fluoroquinolone antimicrobial agent in a neutropenic rat model of Pseudomonas sepsis. Antimicrobial Agents and Chemotherapy, 37, 483–490. DUDLEY M.N., BLASER J., GILBERT D., MAYER K.H. & ZINNER S.H. (1991) Combination therapy with ciprofloxacin plus azlocillin against Pseudomonas Aeruginosa: Effect of simultaneous versus staggered administration in an in vitro model of infection. Journal of Infectious Disease, 164, 499–506. FRIIS C. (1993) Penetration of danofloxacin into the respiratory tract tissue and secretions in calves. American Journal of Veterinary Research, 54, 1122–1127. FRIIS C. & NIELSEN J.P. (1997) Penetration of danofloxacin into the respiratory tract tissues and secretions in healthy and in Actinobacillus pleuropneumoniae infected pigs. Proceedings of the 7th EAVPT Congress, Madrid, Spain. Journal of Veterinary Pharmacology and Therapeutics, 20, suppl. 1, 108–109. GIBALDI M. & PERRRIER D. (1982) Pharmacokinetics, 2nd ed. Pp. 409–417. Marcel Dekker, Inc., New York. GRASSO S., MEINARDI G., DE CARNERI I. & TAMASSIA V. (1978) New in vitro model to study the effect of antibiotic concentration and rate of elimination on antibacterial activity. Antimicrobial Agents and Chemotherapy, 13, 570–576. HAAG R., LEXA P. & WERKHÄUSER I. (1986) Artifacts in dilution pharmacokinetic models caused by adherent bacteria. Antimicrobial Agents and Chemotherapy, 29, 765–768. JACOBSEN M.J. & NIELSEN J.P. (1995) Development and evaluation of a selective and indicative medium for isolation of Actinobacillus pleuropneumoniae from tonsils. Veterinary Microbiology, 47, 191–197. KNUDSEN J.D., FRIMODT-MØLLER N. & ESPERSEN F. (1995) Experimental Streptococcus pneumoniae infection in mice for studying correlation of in vitro and in vivo activities of penicillin against pneumococci with various susceptibilities to penicillin. Antimicrobial Agents and Chemotherapy, 39, 1253–1258. LEGGETT J.E., EBERT S., FANTIN B. & CRAIG W.A. (1991) Comparative doseeffect relations at several dosing intervals for beta-lactam, aminoglycoside and quinolone antibiotics against gram-negative bacilli in murine thigh-infection and pneumonitis. Scandinavian Journal of Infectious Disease, Suppl. 74, 179–184. LISTER P.D., PONG A., CHARTRAND S.A. & SANDERS C.C. (1997) Rationale behind high-dose amoxycillin therapy for acute otitis media due to penicillin-nonsusceptible pneumococci: Support from in vitro pharmacodynamic studies. Antimicrobial Agents and Chemotherapy, 41, 1926–1932. MARCHBANKS C.R., MCKIEL J.R., GILBERT D.H., ROBILLARD N.J., PAINTER B., ZINNER S.H. & DUDLEY M.N. (1993) Dose ranging and fractionation of intravenous ciprofloxacin against Pseudomonas aeruginosa and Staphylococcus aureus in an in vitro model of infection. Antimicrobial Agents and Chemotherapy, 37, 1756–1763. MEINEN J.B., MCCLURE J.T. & ROSIN E. (1995) Pharmacokinetics of enrofloxacin in clinically normal dogs and mice and drug pharmacodynamics in neutropenic mice with Escherichia coli and staphylococcal infections. American Journal of Veterinary Research, 56, 1219–1224. MOUTON J.W. & DEN HOLLANDER J.G. (1994) Killing of Pseudomonas aeruginosa during continuous and intermittent infusion of ceftazidime in an in vitro pharmacokinetic model. Antimicrobial Agents and Chemotherapy, 38, 931–936. NCCLS. (1997) Performance standards for antimicrobial disk and dilution susceptibility tests for bacteria isolated from animals. NCCLS Document M31-P Vol. 1, no. 1. PALMER S.M. & RYBAK M.J. (1996) Pharmacodynamics of once- or twice-daily levofloxacin versus vancomycin, with or without rifampin, against Staphylococcus Aureus in an in vitro model with infected platelet-fibrin clots. Antimicrobial Agents and Chemotherapy, 40, 701–705. RAEMDONCK D.L., TANNER A.C., TOLLING S.T. & MICHENER S.L. (1994) Antimicrobial susceptibility of Actinobacillus pleuropneumoniae, Pasteurelle multocida and Salmonella choleraesuis isolates from pigs. Veterinary Record, 134, 5–7. WOLFSON J.S. & HOOPER D.C. (1989) Bacterial resistance to quinolones: Mecanisms and clinical importance. Review of Infectious Disease, 11, 960–968. Accepted April 29, 1999
The pharmacodynamic effect of amoxycillin and danofloxacin against Actinobacillus pleuropneumoniae in an in-vitro pharmacodynamic model R. H. LINDECRONA*, C. FRIIS,* N. E. JENSEN† *Department of Pharmacology and Pathobiology. Royal Veterinary and Agricultural University and † The Danish Veterinary Laboratory, Copenhagen, Denmark. SUMMARY The pharmacodynamic effect of amoxycillin and danofloxacin against two strains of Actinobacillus pleuropneumoniae was evaluated in an in-vitro pharmacodynamic model. For amoxycillin peak concentrations of 0·5, 1, and 4 µg ml–1 and half-lives of 3 and 15 hours were examined. For danofloxacin peak concentrations of 0·125, 0·5, and 1·5 µg ml–1 and half-lives of 1·5 and 7 hours were evaluated. The initial bactericidal effect was measured as the reduction in colony count (log CFU ml–1) during the first three hours, and the overall pharmacodynamic effect as the area under the bacterial growth versus time curve (AUBC). The initial bactericidal effect of amoxycillin was maximal at peak concentrations of two to four times the MIC. Peak concentration and half-life only influenced the pharmacodynamic effect of amoxycillin if the antibiotic concentration fell below the MIC during the experiments, which is consistent with time > MIC as the most important parameter of pharmacodynamic effect of β-lactam drugs. For danofloxacin maximal bactericidal effect initially was observed at peak concentrations of at least eight times the MIC. The pharmacodynamic effect was dependent on the peak concentration. The half-life only influenced the pharmacodynamic effect of danofloxacin in experiments with a peak concentration MIC ratio of less than eight. This indicated that for danofloxacin the peak concentration was the major determinant of pharmacodynamic effect.
IN-VITRO pharmacodynamic infection models simulating antibiotic pharmacokinetics in vivo can provide useful information about the optimal dosing regimen of antimicrobials in clinical infections (Blaser and Zinner 1987). Many studies have been performed in in-vitro pharmacodynamic models with β-lactam drugs, aminoglycosides, fluoroquinolones and their combinations against human pathogens (Blaser et al 1987, Chambers et al 1991, Dudley et al 1991, Lister et al 1997, Palmer and Rybak 1996). So far only few studies have been carried out with veterinary pathogens in in-vitro pharmacodynamic models. Actinobacillus pleuropneumoniae is a major cause of swine pneumonia and causes great economic losses in the swine production worldwide. Treatment with antibiotics is an important tool in the control of this disease. However, little is known about the optimal dosing regimen. The purpose of this study was to evaluate the pharmacodynamic effects of amoxycillin and danofloxacin against two strains of Actinobacillus pleuropneumoniae in a one-compartment in-vitro pharmacodynamic model, simulating in-vivo drug concentrations of the drugs in bronchial secretions in the pig (Friis and Nielsen 1997, Agersø and Friis 1998a,b).
MATERIALS AND METHODS Bacterial strains Two strains of Actinobacillus pleuropneumoniae, biotype 1, serotype 2 (s4226 and BS9505611) were studied. Both strains were isolated from clinical outbreaks of pleuropneumoniae. Antibiotics Amoxycillin and danofloxacin were kindly supplied by the manufacturers (SmithKline Beecham, Tadworth, Surrey, UK and Pfizer Inc., New York, USA, respectively). Medium The medium used for susceptibility testing and model experiments was Tryptic Soy Broth (BBL, Becton Dickenson & Co., Cockeysville, MD, USA) supplemented with 10 g yeast extract per litre (Oxoid, Unipath Ltd. Basingstoke, UK). β-NAD (Sigma) was supplemented to a concentration of 0·02 per cent. In-vitro model
Corresponding author address: R. H. Lindecrona, The Danish Veterinary Laboratory, Bülowsvej 27, 1790 Copenhagen V, Denmark. 0034-5288/99/040093 + 5 $18.00/0
A schematic representation of the in-vitro pharmacodynamic model used in the experiments is presented in Fig 1. The in-vitro model consisted of a 300 ml magnetic stirred cell (Millipore) filled with 200 ml broth. The broth was © 1999 Harcourt Publishers Ltd
R.H. Lindecrona, C. Friis, N.E. Jensen
94
Flow rate F=Va *Ke
Diluent reservoir
Central reservoir
Elimination reservoir
FIG 1: Schematic representation of the one-compartment in-vitro model used in the experiments. A peristaltic pump is pumping antibiotic-free medium from the diluent reservoir into the central compartment. Simultaneously medium is pumped out of the central reservoir into the elimination reservoir at the same flow rate (F).
continuously stirred during the experiment. Elimination of the antimicrobial agents was achieved using a peristaltic pump (Masterflex) to supply fresh antibiotic-free medium to the compartment resulting in the simultaneous displacement of antibiotic-containing medium. The flow rate of the pump was set according to the half-life being simulated (Grasso et al 1978). Elimination of the bacteria was prevented by a filter (pore size 0·2 µm, Millipore) and a prefilter placed in the bottom of the compartment. Samples were taken through a silicone membrane. Atmospheric air diffused into the compartment through a filter (pore size 0·2 µm, Millipore). Antibiotics were delivered as a bolus into the compartment in doses resulting in initial drug concentrations from 0·5 to 4 µg ml–1 for amoxycillin and from 0·125 to 1·5 µg ml–1 for danofloxacin. Bacteria for inoculation were grown overnight at 37°C on meat and blood agar supplemented with 0·07 per cent β-NAD as previously described (Jacobsen and Nielsen 1995). Three colonies were suspended in 3 ml broth and incubated for one hour at 37°C. From this broth culture a McFarland standard of 0·5 in sterile saline was prepared. Two ml of this suspension were inoculated into the model resulting in a starting inoculum of 1 to 3 × 106 CFU ml.–1 The bacteria were incubated in the model for one hour before addition of the antibiotic. Experiments were run for 8 hours for strain s4226 and for 12 hours for strain BS9505611. All experiments were performed in triplicates.
administration of the antibiotic dose for the determination of of viable bacteria per millitre Serial tenfold dilutions were plated in triplicate on meat and blood agar supplemented with 0·07 per cent β-NAD. The plates were incubated for 24 to 36 hours at 37°C and the colonies counted. At times when the bacterial count was expected to be below the limits of detection or when antibiotic carry-over was expected to be a problem, 100-µl samples were diluted in 10 ml of 0·9 per cent saline and filtered using a Millipore system (0·45 µg pore size filters). Filters were placed aseptically on meat and blood agar and incubated for 24 to 36 hours, at which time the colonies were counted. The lower limit of detection was set to be 102 CFU/ml. The area under the bacterial growth versus time curve (AUBC) was calculated according to the trapezoidal rule up to the last sampling point (Gibaldi and Perrier, 1982). For the bacteria isolated at the beginning and at the end of the experiment MICS were determined in duplicate by a standard broth microdilution test (Sensititre®) as outlined by the NCCLS (1997). CFU
Statistical analysis The effect of peak concentration and half-life on the AUBC and the absolute change in log CFU/ml during the first 3 hours was evaluated by a two-way analysis of variance. P<0·05 was considered significant.
Drug determination Samples were collected from the compartment immediately after injection and at 1, 2, 3, 4, 6, 8, and 12 hours during the experiment and were stored at –20°C until assayed. The concentration of amoxycillin and danofloxacin was determined by high-pressure liquid chromatography (HPLC) as described by Agersø and Friis (1998a,b) and Friis (1993) respectively. Pharmacodynamic analysis Samples were collected from the compartment before injection and at 0·5, 1, 2, 3, 4, 6, 8, and 12 hours after
RESULTS In-vitro susceptibility tests The MIC values for strain s4226 were 0·25 µg ml–1 for amoxycillin and 0·06 µg ml–1 for danofloxacin. For strain BS9505611, MIC values were 0·25 0·50 µg ml–1 for amoxycillin and 0·03 µg ml–1 for danofloxacin. For the experiments no change in MIC was observed except in one case. For strain s4226, the regimen with a peak concentration of danofloxacin of 0·125 µg ml–1 and a half-life of 7 hours, resulted in a four to eight times rise in MIC for the bacteria
Effect of amoxycillin and danofloxacin against A. pleuropneumoniae
The antibiotic concentrations during the experiment were fitted to a one-compartment pharmacokinetic model and selected pharmacokinetic parameters were calculated according to standard equations (Gibaldi and Perrier, 1982). For amoxycillin, mean initial concentrations (µg ml–1 ± SD) were 0·64±0·15, 1·15±0·39 and 3·96±0·74 and mean half-lives were (hours ±SD) 14·6±5·8 and 2·9±0·6. Mean initial danofloxacin concentrations (µg/ml ±SD) were 0·17±0·07, 0·57 ± 0·12 and 1·58±0·26 and mean half-lives were (hours ±SD) 6·6 ±1·1 and 1·2±0·3. Pharmacodynamics Amoxycillin. Mean growth curves for the amoxycillin experiments are presented in Fig 2. For strain s4226 no regrowth was seen within the 8-hour experiment period in any of the experiments. For strain BS9505611 less than a 1-log reduction of bacterial count was seen followed by regrowth of bacteria in the experiments with a peak concentration of 0·5 µg ml–1 and a half life of 3 hours. In these experiments the drug concentration was below the MIC throughout most of the study period. At the end of the 12-hour experiment period some regrowth was observed for all experiments except for the regimen with a peak concentration of 4 µg ml–1 and a half-life of 15 hours. The reductions in bacterial count after 3 hours are presented in Table 1. The effect of half-life and peak concentration on the initial reduction in bacterial count was significant for both strains. In some experiments a change in half life from 15 to 3 hours resulted in a reduced bactericidal effect. A maximal reduction in bacterial count was achieved with a peak concentration of 1 µg ml.–1 The AUBCS are listed in Table 2. For both strains the difference in the AUBC was significant (P <0·05) among the various peak concentrations. The main difference was seen in the experiments with the short half-life of 3 hours, where the lowest peak concentration of 0·5 µg ml–1 resulted in a larger AUBC than the higher peak concentrations. The influence of half-life on the AUBC was strongly significant (P<0·002) for strain BS9505611 but not for strain s4226. However, for the experiments with this strain the drug concentration exceeded the MIC throughout most of the study period. Danofloxacin. For the danofloxacin experiments mean growth curves are illustrated in Fig 3. The reductions in log CFUml from 0 to 3 hours are presented in Table 3. An effect of peak concentration was seen for both strains (P<0·05). For all experiments a rapid reduction in colony count within the first 3 hours was observed at a peak concentration of at least eight times the MIC. A change in half-life from 7 hours to 1·5 hours had no effect on the initial bactericidal activity, except for the
Log CFU ml–1
Pharmacokinetics
Strain s4226
8 6 4 2 0 2
0
4
6
8
Time (h) Strain BS9505611 8 Log CFU ml–1
isolated at the end of the experiment. However, the MIC fell to the original level upon passage on antibiotic-free meat and blood agar.
95
6 4 2 0 0
2
4
6 8 Time (h)
10
12
FIG 2: Amoxycillin versus two strains of Actinobacillus pleuropneumoniae in an in-vitro pharmacodynamic model. Results are means of triplicate experiments. Peak conc.: 0·5 µg ml–1 (●), 1 µg ml–1(■), 4 µg ml–1 (▲), t1/2=15 hours (full symbol), t1/2=3 hours (open symbol), control (+).
experiments with strain s4226 and a peak concentration of 0·5 µg ml–1, where a marked reduction in killing activity was observed. The AUBCS are listed in Table 4. The effect of peak concentration on the AUBC was strongly significant with the highest peak concentrations resulting in the smallest AUBC. The half-life only influenced the AUBC in two cases. For strain s4226, the experiments with a peak concentration of 0·5 µg/ml and for strain BS9505611, a peak concentration of 0·125 µg ml–1 a significant smaller AUBC was observed for the experiments with a longer half-life compared to the experiments with a shorter half-life.
DISCUSSION For amoxycillin, the peak concentrations ranging from 0·5 to 4·0 µg ml–1 did not influence the pharmacodynamic effect as long as the antibiotic concentration exceeded the MIC during the greater part of the experiment. These observations are in agreement with previous findings both in vivo and in vitro, where the time that the antibiotic concentration exceed the MIC (T > MIC) is recognised as the most important parameter for antibacterial efficiency of the β-lactam drugs (Drusano 1991, Craig and Ebert 1992). The bactericidal effect of amoxycillin reached a maximum at peak concentrations of two to four times the MIC of the actual strain. This observation has also been reported from in-vitro studies with penicillin against Streptococcus pneumoniae (Knudsen
R.H. Lindecrona, C. Friis, N.E. Jensen
96
TABLE 1: Reduction in log CFU/ml from 0 to 3 hours for experiments with amoxycillin against two strains of Actinobacillus pleuropneumoniae in an in-vitro pharmacodynamic model. Values are mean of three experiments ± SD.
TABLE 2: Area under the bacterial growth versus time curve for experiments with amoxycillin against two strains of Actinobacillus pleuropneumoniae in an in-vitro pharmacodynamic model (experiments were run for 8 hours for strain s4226 and for 12 hours for strain BS9505611). Results are mean ±SD of three experiments.
Peak concentration (µg ml–1) Peak concentration (µg ml–1) Strain
t1/2 (hours)
0·5
1·0
4·0
S4226
15 3 15 3
1·9 (0·1) 1·9 (0·1) 2·2 (0·2) 0·3 (0·7)
2·9 (0·3) 2·0 (0·1) 2·4 (0·1) 1·9 (0·6)
2·7 (0·7) 2·5 (0·3) 2·9 (0·2) 1·8 (0·5)
BS9505611
TABLE 3: Reduction in log CFU/ml from 0 to 3 hours for experiments with danofloxacin against two strains of Actinobacillus pleuropneumoniae in an in-vitro pharmacodynamic model. Values are mean of three experiments ±SD.
Strain
t1/2 (hours)
0·5
1·0
4·0
s4226
15 3 15 3
32·2 (2·1) 35·1 (2·0) 45·1 (5·1) 68·3 (9·5)
31·9 (6·6) 31·8 (0·7) 46·7 (5·3) 52·7 (14·2)
27·2 (2·1) 28·4 (2·6) 36·0 (1·1) 50·1 (2·0)
BS9505611
TABLE 4: Area under the bacterial growth versus time curve for experiments with danofloxacin against two strains of Actinobacillus pleuropneumoniae in an in vitro pharmacodynamic model (experiments were run for 8 hours for strain s4226 and for 12 hours for strain BS9505611). Values are mean ± SD of three experiments.
Peak concentration (µg ml–1) Peak concentration (µg ml–1) Strain
t1/2 (hours)
0·125
0·50
1·50
s4226
7 1·5 7 1·5
2·7 (0·5) 2·8 (1·1) 3·7 (0·6) 2·9 (1·0)
4·4 (0·1) 2·5 (0·2) 4·3 (0·1) 3·9 (0·4)
4·0 (0·1) 3·8 (0·7) 4·2 (0·1) 4·2 (0·1)
BS9505611
Strain
t1/2 (hours)
0·125
0·50
1·50
s4226
7 1·5 7 1·5
35·1 (2·8) 31·0 (6·0) 35·9 (2·8) 52·9 (14·6)
17·9 (0·8) 30·7 (1·1) 29·5 (4·5) 33·0 (2·4)
19·0 (0·5) 21·3 (4·9) 27·2 (0·03) 32·2 (5·7)
BS9505611
Va volume of the central compartment; Ke elimination constant of the antibiotic
Strain s4226 Log CFU ml–1
8 6 4 2 0 0
2
4 Time (h)
6
8
Strain BS9505611
Log CFU ml–1
8 6 4 2 0 0
2
4
6 8 Time (h)
10
12
FIG 3: Danofloxacin versus two strains of Actinobacillus pleuropneumoniae in an in-vitro pharmacodynamic model. Results are means of triplicate experiments. Peak conc 0·125 µg ml–1 (●), 0·5 µg ml–1 (■), 1·5 µg ml–1 (▲), t1/2=7 hours (full symbol), t1/2=1·5 hours (open symbol), control (+).
et al 1995). A peak concentration of four times the MIC was needed for the efficacy of ceftazidime against Pseudomonas aeruginosa in an in-vitro pharmacodynamic model (Mouton and den Hollander 1994). Peak concentration as well as the area under the antibiotic concentration versus time curve (AUC) has been mentioned
as important factors of optimal pharmacodynamic effect of fluoroquinolones (Chambers et al 1991, Drusano et al 1993, Leggett et al 1991, Marchbanks et al 1993, Meinen et al 1995, Palmer and Rybak 1996). Several studies have indicated that a high peak concentration/MIC ratio is important to achieve clinical efficiency of fluoroquinolones and minimize problems with emergence of resistance (Chambers et al 1991, Marchbanks et al 1993, Drusano et al 1993). This is in agreement with the observations made in this study. A maximal bactericidal effect and no regrowth within the experimental period was seen at peak concentrations of at least eight times the MIC of the particular strain. Studying the pharmacodynamic effect of enoxacin against several bacterial species in an in-vitro model, Blaser et al (1987) also observed regrowth of bacteria with a MIC of four to 16 times the MIC of the original strain unless the peak concentration/MIC ratio exceeded eight. In our study, a shortening of the half-life (and thereby a reduction in the AUC) had no influence on the antibacterial activity as long as the peak concentration/MIC ratio exceeded eight. Only in experiments with a peak/MIC ratio of four to eight a shorter halflife resulted in a reduced bactericidal activity both initially and during the whole experimental period. This was also observed by Drusano et al (1993), where the AUC only influenced the outcome of a Pseudomonas infection in rats if the peak concentration/MIC ratio was lower than 10. Other studies have pointed to the AUC as the most significant determinant of clinical efficiency of fluoroquinolones. In a murine thigh-infection and pneumonitis model Leggett et al (1991) observed that the frequency of administration of the same daily dose of ciprofloxacin had little impact on the efficacy against Gram-negative bacilli. The same observation has been reported in a study with enrofloxacin against Escherichia coli and Staphylococcus aureus in neutropenic mice (Meinen et al 1995).
Effect of amoxycillin and danofloxacin against A. pleuropneumoniae
The change in MIC observed in this study in experiments with a peak/MIC ratio of two and a half-life of 7 hours might have resulted from a single step mutation in the gene of DNA gyrase or in genes regulating drug permeation (Wolfson and Hooper, 1989). However, these findings were not further investigated. The in-vitro model described in this study provides a useful tool for describing the antibacterial activity of different antibiotics and for predicting the optimal dosing regimen clinically. Moreover this is the first in-vitro study describing the pharmacodynamic effect of antibiotics against Actinobacillus pleuropneumoniae. Running the experiment for a longer period of time will further improve the in-vitro model. However in these experiments it was not possible to run the experiment period for more than twelve hours because of bacterial regrowth. This regrowth may be caused by bacteria adhering to the glass surface of the model as described by Haag et al (1986). In a study by Raemdonck et al. (1994), the in-vitro susceptibilities of Actinobacillus pleuropneumoniae strains isolated from clinical cases of pleuropneumoni were determined. The MIC90 for amoxycillin and danofloxacin were reported to be 0·5 µg ml–1 and 0·125 µg ml–1 respectively. Thus, for amoxycillin, a long half-life and a peak concentration of 1 µg ml–1 is needed at the site of infection to achieve an efficient treatment of Actinobacillus pleuropneumoniae infections. On the assumption that the amoxicillin concentration in the lungs is one third of the plasma concentration (Agersø and Friis, 1998a), a dose of 15 mg kg–1 intramuscularly would result in sufficient concentrations for the treatment of Actinobacillus pleuropneumoniae infections (Agersø & Friis, 1998b). For danofloxacin, a peak concentration of 1 µg ml–1 is the most important determinant of therapeutic efficiency, as well as for minimal emergence of resistance. A maximal danofloxacin concentration of 1 µg ml–1 could be achieved in all parts of the lung tissue following intramuscular administration of a dose of 2·5 mg kg–1 (Friis and Nielsen, 1997).
REFERENCES AGERSØ H. & FRIIS C. (1998a) Penetration of amoxycillin into the respiratory tract tissues and secretions in pigs. Research in Veterinary Science, 64, 245–250. AGERSØ H. & FRIIS C. (1998b) Bioavailability of amoxycillin in pigs. Journal of Veterinary Pharmacology and Therapeutics, 21, 41–46. BLASER J., STONE B.B., GRONER M.C. & ZINNER S.H. (1987) Comparative study with enoxacin and netilmicin in a pharmacodynamic model to determine importance of ratio of antibiotic peak concentration to MIC for bactericidal activity and emergence of resistance. Antimicrobial Agents and Chemotherapy, 31, 1054–1060. BLASER J. & ZINNER S.H. (1987). In vitro models for the study of antibiotic activities. Progress in Drug Research, 31, 349–381. CHAMBERS S.T., PEDDIE B.A., ROBSON R.A., BEGG E.J. & BOSWELL D.R. (1991) Antimicrobial effects of lomefloxacin in vitro. Journal of Antimicrobial Chemotherapy, 27, 481–489.
97
CRAIG W.A. & EBERT S.C. (1992). Continuous infusion of beta-lactam antibiotics. Antimicrobial Agents and Chemotherapy, 36, 2577–2583. DRUSANO G.L. (1991) Human pharmacodynamics of beta-lactams, aminoglycosides and their combination. Scandinavian Journal of Infectious Disease, Suppl. 74, 235–248. DRUSANO G.L., JOHNSON D.E., ROSEN M. & STANDIFORD H.C. (1993) Pharmacodynamics of a fluoroquinolone antimicrobial agent in a neutropenic rat model of Pseudomonas sepsis. Antimicrobial Agents and Chemotherapy, 37, 483–490. DUDLEY M.N., BLASER J., GILBERT D., MAYER K.H. & ZINNER S.H. (1991) Combination therapy with ciprofloxacin plus azlocillin against Pseudomonas Aeruginosa: Effect of simultaneous versus staggered administration in an in vitro model of infection. Journal of Infectious Disease, 164, 499–506. FRIIS C. (1993) Penetration of danofloxacin into the respiratory tract tissue and secretions in calves. American Journal of Veterinary Research, 54, 1122–1127. FRIIS C. & NIELSEN J.P. (1997) Penetration of danofloxacin into the respiratory tract tissues and secretions in healthy and in Actinobacillus pleuropneumoniae infected pigs. Proceedings of the 7th EAVPT Congress, Madrid, Spain. Journal of Veterinary Pharmacology and Therapeutics, 20, suppl. 1, 108–109. GIBALDI M. & PERRRIER D. (1982) Pharmacokinetics, 2nd ed. Pp. 409–417. Marcel Dekker, Inc., New York. GRASSO S., MEINARDI G., DE CARNERI I. & TAMASSIA V. (1978) New in vitro model to study the effect of antibiotic concentration and rate of elimination on antibacterial activity. Antimicrobial Agents and Chemotherapy, 13, 570–576. HAAG R., LEXA P. & WERKHÄUSER I. (1986) Artifacts in dilution pharmacokinetic models caused by adherent bacteria. Antimicrobial Agents and Chemotherapy, 29, 765–768. JACOBSEN M.J. & NIELSEN J.P. (1995) Development and evaluation of a selective and indicative medium for isolation of Actinobacillus pleuropneumoniae from tonsils. Veterinary Microbiology, 47, 191–197. KNUDSEN J.D., FRIMODT-MØLLER N. & ESPERSEN F. (1995) Experimental Streptococcus pneumoniae infection in mice for studying correlation of in vitro and in vivo activities of penicillin against pneumococci with various susceptibilities to penicillin. Antimicrobial Agents and Chemotherapy, 39, 1253–1258. LEGGETT J.E., EBERT S., FANTIN B. & CRAIG W.A. (1991) Comparative doseeffect relations at several dosing intervals for beta-lactam, aminoglycoside and quinolone antibiotics against gram-negative bacilli in murine thigh-infection and pneumonitis. Scandinavian Journal of Infectious Disease, Suppl. 74, 179–184. LISTER P.D., PONG A., CHARTRAND S.A. & SANDERS C.C. (1997) Rationale behind high-dose amoxycillin therapy for acute otitis media due to penicillin-nonsusceptible pneumococci: Support from in vitro pharmacodynamic studies. Antimicrobial Agents and Chemotherapy, 41, 1926–1932. MARCHBANKS C.R., MCKIEL J.R., GILBERT D.H., ROBILLARD N.J., PAINTER B., ZINNER S.H. & DUDLEY M.N. (1993) Dose ranging and fractionation of intravenous ciprofloxacin against Pseudomonas aeruginosa and Staphylococcus aureus in an in vitro model of infection. Antimicrobial Agents and Chemotherapy, 37, 1756–1763. MEINEN J.B., MCCLURE J.T. & ROSIN E. (1995) Pharmacokinetics of enrofloxacin in clinically normal dogs and mice and drug pharmacodynamics in neutropenic mice with Escherichia coli and staphylococcal infections. American Journal of Veterinary Research, 56, 1219–1224. MOUTON J.W. & DEN HOLLANDER J.G. (1994) Killing of Pseudomonas aeruginosa during continuous and intermittent infusion of ceftazidime in an in vitro pharmacokinetic model. Antimicrobial Agents and Chemotherapy, 38, 931–936. NCCLS. (1997) Performance standards for antimicrobial disk and dilution susceptibility tests for bacteria isolated from animals. NCCLS Document M31-P Vol. 1, no. 1. PALMER S.M. & RYBAK M.J. (1996) Pharmacodynamics of once- or twice-daily levofloxacin versus vancomycin, with or without rifampin, against Staphylococcus Aureus in an in vitro model with infected platelet-fibrin clots. Antimicrobial Agents and Chemotherapy, 40, 701–705. RAEMDONCK D.L., TANNER A.C., TOLLING S.T. & MICHENER S.L. (1994) Antimicrobial susceptibility of Actinobacillus pleuropneumoniae, Pasteurelle multocida and Salmonella choleraesuis isolates from pigs. Veterinary Record, 134, 5–7. WOLFSON J.S. & HOOPER D.C. (1989) Bacterial resistance to quinolones: Mecanisms and clinical importance. Review of Infectious Disease, 11, 960–968. Accepted April 29, 1999