Tobramycin, or Amikacin in an Experimental Model of Serratia marcescens Endocarditis: In Vitro-In Vivo Correlation

ANTIMICROBIAL AGENTS AND CHEMOTHERAPY, Jan. 1991, p. 111-116 0066-4804/91/010111-06$02.00/0 Copyright C 1991, American Society for Microbiology Vol. 35, No. 1 Impact of Dosage Schedule on the Efficacy of Gentamicin, Tobramycin, or Amikacin in an Experimental Model of Serratia marcescens Endocarditis: In Vitro-In Vivo Correlation GILLES POTEL,1* JOCELYNE CAILLON,1 BRUNO FANTIN,2 JOCELYNE RAZA,1 FLORENCE LE GALLOU,1 JEAN-YVES LEPAGE,' PHILIPPE LE CONTE,' DENIS BUGNON,1 DENIS BARON,' AND HENRI DRUGEON' Laboratoire d'antibiologie, Faculte de Medecine, 1 rue Gaston Veil, 44000 Nantes,' and Institut National de la Sante et de la Recherche Medicale U.13, H6pital Claude Bernard, 75019 Paris,2 France Received 21 May 1990/Accepted 16 October 1990 Aminoglycosides are usually considered to be concentration-dependent antibiotics and to have similar pharmacodynamic and pharmacokinetic properties. To verify the equivalent activity of the aminoglycosides on a susceptible strain, we tested the killing rate of three aminoglycosides (gentamicin, tobramycin, and amikacin) on one strain of Serratia marcescens both in vitro and in vivo by using a rabbit model of left-ventricle endocarditis. Despite similar MICs, the time-kill curve of gentamicin was consistently better than those of amikacin and tobramycin, whatever the concentration of each antibiotic used (1, 2, 4, 8, 16, or 32 mg/liter), after a 5-h incubation. The in vivo bacterial reduction in the vegetations was measured 24 h after administration of an intravenous 48-mg/kg bolus of each antibiotic or at the end of a 24-h continuous intravenous infusion of the same dose. Gentamicin was significantly more effective when administered as a bolus than when administered as a continuous infusion (2.8 0.2 versus 6.4 1.5 log1o CFU/g of vegetation, respectively; P < 0.01), whereas amikacin was more effective as a continuous infusion than as a bolus injection (3.6 2.0 versus 7.5 + 1.3 loglo CFU/g of vegetation, respectively; P < 0.01). Tobramycin was not very effective, whatever the dosage tested (approximately 6.5 to 7 loglo CFU/g). These results suggest that concentration-dependent bactericidal activities, both in vitro and in vivo, may vary greatly among aminoglycosides despite similar MICs. The bactericidal killing rates of aminoglycosides are usually considered to be concentration dependent. This property is defined by an increased bactericidal effect with increasing concentration for any given period of exposure. Consistent data from different animal models (8, 13) suggest that in cases of pseudomonal experimental infections, single, large daily doses of aminoglycosides are at least as effective as conventional intermittent dosing regimens. Despite infrequent in vivo comparative studies between aminoglycosides, Kapusnik et al. concluded that it is now possible to study the single-daily-dose regimen in humans and that any of the aminoglycosides may be used, because their pharmacokinetic and pharmacodynamic properties are comparable (12). A clinical multiple-dose study (including bacteremia and pneumonia) of humans has shown that a high peak-concentration/mic ratio was a factor of improved therapeutic outcome (16), suggesting that single daily dosing could further improve the prognosis of severely infected patients, allowing higher peak concentration of aminoglycosides. Furthermore, aminoglycosides are less toxic in single doses than in multiple doses (1, 21). Recent studies (4, 18) failed to show any evidence that aminoglycoside treatment once daily had greater ototoxicity or nephrotoxicity than the traditional three-times-daily regimen. Nevertheless, and before the routine clinical application of single daily doses, it should be pointed out that some bacterial species (like Serratia marcescens) have been excluded from experimental studies despite a high mortality rate in clinical infections and a possible difference from other * Corresponding author. species in terms of optimal antimicrobial regimen. Juvin et al. (11) have shown for six clinical strains of S. marcescens that killing rates of aminoglycosides in vitro could be different from one another for the same strain, even when their MICs were similar, suggesting a possible difference in terms of pharmacodynamic properties of aminoglycosides. Furthermore, gentamicin is intrinsically the most active aminoglycoside against Serratia strains which are susceptible to all aminoglycosides (11). The aim of this work was to verify the in vivo relevance of these findings on a rabbit model of S. marcescens endocarditis, comparing the in vivo dynamics of bacterial killing of a single equivalent dose of gentamicin, tobramycin, or amikacin. To assess the in vivo impact of dosage schedule, we administered the same dosage of each drug as a pulse dosing or as a 24-h continuous intravenous (i.v.) infusion. MATERIALS AND METHODS Organism. The strain of S. marcescens used for inducing experimental endocarditis, strain HN229, was isolated from the urine of a hospital patient. This strain was found to be resistant to rabbit serum. Antibiotics. The three aminoglycosides tested were tobramycin (Eli-Lilly), gentamicin (Schering-Plough), and amikacin (Bristol). In vitro studies. (i) Antibiotic susceptibility tests. The MIC of each antibiotic was determined by a Mueller-Hinton broth dilution technique with Ca2' and Mg2+ supplementation (24) in 200-pdl wells, with an inoculum of 105 CFU/ml in the mid-exponential phase of growth. The MIC was defined as the lowest concentration of the drug producing no visible 111

112 POTEL ET AL. growth after an incubation of 18 h. After 24 h, a subculture was made on Mueller-Hinton agar (Difco Laboratories, Detroit, Mich.) The MBC corresponded to the lowest concentration of the drug permitting 0.1% of the bacteria to survive and was obtained by replicating 1 ml, using a Steers apparatus, onto agar plates with polyanethole sulfonic acid sodium salt (SPS; Sigma). (ii) Killing curves. Time-kill curves were drawn for each antibiotic at 11 concentrations in Mueller-Hinton broth: 0, 0.06, 0.12, 0.25, 0.5, 1, 2, 4, 8, 16, and 32 mg/liter. For each concentration the antibiotics were incubated in a microtube (1-ml tubes, Macrowell; Skatron, Lier, Norway) with an inoculum of 107 S. marcescens cells per ml in the stationary phase of growth. Surviving bacteria were counted in each tube after 1.5, 3, 5, and 24 h of incubation by a semiautomatic dilution micromethod involving an automatic 96-well dispenser (Skatron) and a Steers replicator distributing 2 ± 0.5 IlI of each dilution onto agar plates. A 3% solution of SPS was added, avoiding a carryover phenomenon. After a 24-h incubation, the first dilution with 5 to 30 colonies was read and the colony count was then multiplied by the dilution factor. The standard error of this count was 0.2 log1o CFU/ml. The sensitivity limit of detection is equal to 2.4 log1o CFU/ml. This method was detailed in a previous work (5) Ėxperimental endocarditis. In vivo studies were carried out on New Zealand White female rabbits (age range, 12 to 15 weeks; weight range, 2.5 to 3.5 kg). The animals were kept in individual cages and allowed free access to food and water throughout the experiment. Left-ventricular endocarditis was induced as described previously (19). At 24 h after introduction of a polyethylene catheter through the aortic valve, each rabbit received 1 ml of a suspension containing i07 organisms per ml, injected through the marginal ear vein. (i) Experimental design. At 48 h after inoculation, the animals were randomly assigned to one of the six following therapeutic regimens: 48 mg of gentamicin, tobramycin, or amikacin per kg administered as an i.v. bolus or as a 24-h i.v. continuous infusion. For the infusion, a catheter (22-gauge) was inserted into a marginal ear vein and connected to an electric syringe pump. The antibiotic was diluted iti sterile saline, and the infusion output chosen was equal to 2 ml/h. The animals were kept for 24 h in rabbit-restraining cages. Nine untreated rabbits made up the control group. As previously shown, the dose chosen (48 mg/kg) allowed for trough concentrations in vegetations averaging 2,ug/g of vegetation 24 h after an i.v. bolus (20), assuming that no bacterial regrowth could occur during the experiment; this permitted valuable comparisons between drugs. Moreover, three infected animals were assigned to each therapeutic regimen to determine the pharmacokinetics of each drug. A catheter was inserted into the left femoral artery to take serum samnples 5, 15, and 30 min and 1, 2, 4, 6, and 24 h after the bolus or 0.25, 0.5, 1, 2, 3, 6, and 24 h after the beginning of the infusion. (ii) Evaluation of therapy. The effect of each treatment was evaluated 24 h after the i.v. bolus of each antibiotic or at the end of the 24-h i.v. infusion. The animals were sacrificed with an i.v. bolus of thiopental. The heart was removed, and vegetations were excised and rapidly rinsed in sterile saline. Some of the vegetations were weighed and homogenized in a Thomas Teflon pestle tissue homogenizer with 0.5 ml of sterile saline. Serial dilutions of 50-pd aliquots were spread by using a Spiral System (Interscience) and quantitatively cultured on Trypcase-soy agar plates for 24 h at 370C. ANTIMICROB. AGENTS CHEMOTHER. Bacterial titers were expressed as log1o CFU/g of vegetation. We were able to detect quantities as small as 20 CFU/ml. Owing to this sensitivity limit, vegetations found to be sterile were considered to contain 20 CFU/ml of homogenate, and the value integrated for the calculation of the mean bacterial titer took into account the weight of vegetations. Part of each vegetation was frozen prior to antibiotic assays. (iii) Antibiotic assays. Concentrations of each antibiotic in serum were determined by using a microbiological assay with Bacillus subtilis ATCC 6633. The range of measurable concentrations with this strain was 0.06 to 1,ug/ml for all three antibiotics. After being weighed and homogenized with 0.3 ml of 0.1 M phosphate buffer, the vegetations were centrifuged and the supernatant fluid was sampled for microbiologal assay. The same strain of B. subtilis was used. (iv) Statistical evaluation. A Kruskall-Wallis test and then a Mann-Whitney test were performed to compare the bacterial titers measured in treated animals versus controls. A Fisher exact test was used to compare the number of sterile vegetations (no growth of the undiluted tissue homogenate) in each group versus the controls. RESULTS In vitro studies. (i) Antibiotic susceptibiity tests. The MICs (and the MBCs) for the S. marcescens strain studied were 0.5 -jg of gentamicin per ml and 1 jig of tobramycin and amikacin per ml. (ii) Killing curves. At concentrations of twice the MIC, the three aminoglycosides exhibited a slight bactericidal activity between 0 and 5 h, but regrowth was noted between 5 and 24 h only for tobramycin and amikacin; this regrowth reached the same level as the control (10 log CFU/ml). The time-kill curves of the three aminoglycosides at 4 times the MIC are represented in Fig. 1. Gentamicin exhibited the best killing, apparently sterilizing the culture after a 5-h incubation period. At a concentration 16 times the MIC (Fig. 2), gentamicin sterilized the culture after a 3-h incubation period, without regrowth at 24 h. Amikacin and tobramycin exhibited a concentration-dependent killing between 4 and 16 times the MIC, although their effect was less dramatic than that of gentamicin. The MICs for surviving bacteria after a 24-h exposure were unchanged compared with those for the parental strain. In vivo studies. (i) Concentrations in serum. The drug levels obtained in serum after a 48-mg/kg i.v. bolus of gentamicin, tobramycin, and amikacin were quite similar each time, with a peak level (5 min after the injection) around 300,ug/ml and a trough level (24 h after the injection) around 1,ug/ml. The mean half-life at 1B phase for the three aminoglycosides, calculated between 30 min and 6 h, was 54 min (r = 0.97; P < 0.001). The steady state of the 24-h i.v. infusion was reached between 3 and 6 h. The mean level in serum was slightly lower for gentamicin (7.9 + 2.1,ug/ml) than for tobramycin (9.2 + 1.9,ug/ml) or amikacin (8.0 ± 1.9,ug/ml), but this difference was not significant. (ii) Experimental endocarditis. The in vivo effect of a 48-mg/kg dose of each antibiotic administered as a bolus is shown in Table 1. Gentamicin was the most effective antibiotic and had a significant antibacterial effect on vegetations, whereas amikacin and tobramycin did not. The mean trough concentration in vegetations was similar in each group (around 2,ug/g of vegetation). When the same dose was administered as a continuous 24-h i.v. infusion (Table 2), amikacin proved the most effective, more so than gentamicin (P < 0.05), whereas tobramycin had no significant antibac-

VOL. 35, 1991 DOSAGE SCHEDULE OF AMINOGLYCOSIDES 113 L06 CFU/fIL 4 TIME (HOURS) FIG. 1. Time-kill curves of gentamicin (*), tobramycin (O), and amikacin (K) at a concentration equal to four times the MIC versus control (El). Surviving bacteria were counted after 1.5, 3, 5, and 24 h of incubation. terial effect, despite concentrations in vegetations far above the MICs of the three drugs. Furthermore, the MICs for the surviving bacteria in the vegetations were identical to those for the parental strain. These results show clearly that gentamicin was more effective when administered as a bolus, whereas the same dose of amikacin was more effective on S. marcescens when administered as a 24-h continuous infusion rather than as a LOG CFU/IL 10 bolus. The antibacterial efficacy of tobramycin was not affected by the dosage schedule. DISCUSSION In our in vivo experiments, we chose to study the antibacterial effect of a single i.v. injection of each antibiotic, comparing the in vivo efficacy of two therapeutic regimens: 24 6 FIG. 2. Surviving bacteria were counted after 1.5, 3, 5, and 24 h of incubation. (El). 3 TIME (HOURS) Time-kill curves of gentamicin (*), tobramycin (U), and amikacin (O) at a concentration equal to 16 times the MIC versus control

114 POTEL ET AL. ANTIMICROB. AGENTS CHEMOTHER. TABLE 1. In vivo mean antibacterial effect of gentamicin, amikacin, or tobramycin in vegetations 24 h after an i.v. 48-mg/kg dose administered as a bolus Treatment group No. of rabbits Mean vegetation titer ± SD No. of sterile Trough concn in vegetations sacrificed (log CFU/g of vegetation) vegetations/total no." (glg/g of vegetation) (mean ± SD) Control 11 7.8 ± 0.4 0/11 Gentamicin 9 2.8 ± 0.2b 6/9c 2.2 ± 1.2 Amikacin 10 7.5 ± 1.3 0/10 1.9 ± 1.8 Tobramycin 7 6.8 ± 1.9 0/7 2.3 ± 1.5 a Number of culture-negative vegetations/total number of animals sacrificed. b P < 0.01 versus control (Mann-Whitney U test). c P = 0.004 versus control (Fisher's exact test). bolus or continuous infusion. S. marcescens endocarditis, although uncommon in humans, provided a useful experimental model of acute gram-negative bacterial infection and was also used to assess the relationship between the bacterial titer and the antibiotic level in infected sites (6). Moreover, a single i.v. injection allowed us to study the in vivo intrinsic activity of each drug more accurately, avoiding complex interactions of dosing intervals or postantibiotic effects. Our in vitro results showed clearly that, despite similar MICs and MBCs, the killing rate of gentamicin was markedly higher than those of tobramycin and amikacin, whatever the concentration used. In vivo, gentamicin exhibited the best antibacterial effect when administered as a bolus, whereas amikacin proved most effective when administered as a continuous infusion, more so than gentamicin and tobramycin. The differences observed in vivo cannot apparently be explained by pharmacokinetic parameters. The dose chosen (48 mg/kg) proved capable of achieving similar mean trough concentrations for each drug 24 h after a bolus (2.2, 1.9, and 2.3,ug/g of vegetation for gentamicin, amikacin, and tobramycin, respectively). Furthermore, the half-lives in serum were very similar from one drug to another. Nevertheless, despite similar pharmacokinetics in serum, it is not clear that the areas under the concentration-time curves are similar in the vegetations themselves. Therefore, Carbon et al. (2) showed that of the aminoglycosides injected in 1.5-mg/kg doses, gentamicin induced higher interstitial levels than did tobramycin. In this work, we did not determine the early concentrations of aminoglycosides after administration of the bolus in the vegetations themselves, and a possible difference (in terms of local concentrations between drugs) cannot be excluded. Finally, it must be stressed that the tobramycin bolus has a mediocre in vivo antibacterial effect, whereas its distribution is rapid and homogeneous, as has been recently shown in quantitative autohistoradiography (3) Ṫhe differences observed may be due to the antibiotic itself: the activity of gentamicin in vitro (in time-kill curve terms) is significantly more concentration dependent than that of amikacin or tobramycin, which may explain the better in vivo activity of a gentamicin bolus. It has been shown in vitro that the antibacterial effect of aminoglycosides on gram-negative bacteria necessitates an active uptake, which consumes energy (9, 23). In recent research, it has been shown with Escherichia coli and Salmonella typhi that the level of this uptake is genetically determined (15). The hypothesis can thus be put forward that this level of uptake may be different from one aminoglycoside to another and that, for the same bacterium, the higher the uptake, the more efficient a high concentration (gentamicin) would be. On the contrary, a lower level for tobramycin and amikacin would be responsible for the poor result observed in vivo with the bolus. In continuous administration, only amikacin enables five of eight animals to clear the bacteria, confirming its time-dependent activity on S. marcescens. As the steady state is similar for the three antibiotics tested, it is hard to understand why tobramycin does not cause the same results as amikacin with this regimen, since their in vitro behavior is similar. It is possible that local physicochemical conditions (ph, anaerobiosis) are less favorable to its activity on S. marcescens, although this fact, to our knowledge, has never been demonstrated. Therefore, the active concentrations in vivo are probably higher than the MICs in vitro and may be different from one drug to another. Finally, it is possible that our strain of S. marcescens produces a 6'-aminoglycoside acetyltransferase (AAC 6') enzyme, which preferentially inactivates tobramycin and amikacin (10, 17, 22), even though the MICs for surviving bacteria (both in vitro and in vivo) were apparently identical to those for the parental TABLE 2. In vivo mean antibacterial effect of gentamicin, amikacin, or tobramycin in vegetations after an i.v. 24-h continuous infusion of a 48-mg/kg dose No. of rabbits Mean vegetation titer ± SD No. of sterile Concn in vegetation at steady state Treatment group sacrificed (log CFU/g of vegetation) vegetations/total no." (p.g/g of vegetation) (mean ± SD) Control 11 7.8 ± 0.4 0/11 Gentamicin 7 6.4 1.5b 0/7 10.5 ± 2.5 Amikacin 8 3.6 ± 2.0C 5/8d 8.2 ± 3.7 Tobramycin 7 6.5 ± 2.2 0/7 11.3 ± 3.6 a Number of culture-negative vegetations/total number of animals sacrificed. b p < 0.05 versus control (Mann-Whitney U test). C P < 0.01 versus control (Mann-Whitney U test). d p = 0.004 versus control (Fisher's exact test).

VOL. 35, 1991 strains. New in vitro experiments are now in progress in an attempt to elucidate this important point. Experimental research comparing the dose-efficacy or concentration-efficacy relation of several aminoglycosides on the same type of bacterium are rare. Recently, Legett et al. (14) studied the effect of different doses and the interval of administration of netilmicin and gentamicin on a murine model of Klebsiella pneumoniae thigh infection or pneumonia. The authors found no difference in activity between the two antibiotics, but there were no data available in this research concerning the in vitro time-kill curves. On the other hand, much experimental research has shown the concentration-dependent activity of aminoglycosides (8, 12, 14), but none of this research has been carried out for S. marcescens infections. Moore et al. (16) do not identify the results obtained on S. marcescens in terms of the aminoglycoside used. Interestingly, Garraffo et al. studied the serum antibacterial activity of an i.v. infusion of amikacin on the same strain of S. marcescens in healthy volunteers. Serum killing curves showed a time-dependent activity against this strain, but a concentration-dependent activity against Escherichia coli and Enterobacter cloacae (7). Finally, our results show that the in vivo activities of three aminoglycosides against S. marcescens can be different, despite similar MICs. Their pharmacodynamics (in time-kill curve terms) are different, both in vitro and in vivo. Moreover, the in vivo effect of a bolus can indeed be predicted by the in vitro time-kill curves, which demonstrate the superior activity of gentamicin. Despite an initial dose with concentrations close to those observed in humans, amikacin is less efficient as a bolus than in continuous administration. Thus, concentration-dependent killing can differ markedly among aminoglycosides, and concentration-dependent killing may be relatively less important than time-dependent killing. The relative contributions of time- and concentration-dependent killing to overall in vivo antibacterial effect are difficult to predict from in vitro studies. Our results do not agree with those of Kapusnik et al. (12), and optimum dosing regimens determined with one aminoglycoside and species should not be generalized to others. Other in vitro studies are now necessary to explain the mechanism(s) which would enable these differences to be elucidated. In short, the type of antibacterial activity (time or concentration dependent) of aminoglycosides depends both on the molecule chosen and on the bacterium under treatment. The weak activity of amikacin in bolus form on S. marcescens (compared with that of gentamicin) could explain in part the absence of agreement in the literature concerning the optimum therapeutic dosage regimen of aminoglycosides (13). ACKNOWLEDGMENTS This work was supported in part by grants from Facultd de Mddecine de Nantes. We thank M. 0. Hervy for her helpful secretarial assistance. REFERENCES 1. Bennett, W. M., C. E. Plamp, D. N. Gilbert, R. A. Parker, and G. A. Porter. 1979. The influence of dosage regimen on experimental gentamicin nephrotoxicity: dissociation of peak serum levels from renal failure. J. Infect. Dis. 140:576-580. 2. Carbon, C., A. Contrepois, and S. Lamotte-Barrillon. 1978. Comparative distribution of gentamicin, tobramycin, sisomicin, netilmicin, and amikacin in interstitial fluid in rabbits. Antimicrob. Agents Chemother. 13:368-372. 3. Cremieux, A. C., B. Maziere, J. M. Vallois, M. Ottaviani, A. Azancot, H. Raffoul, A. Bouvet, J. J. Pocidalo, and C. Carbon. 1989. Evaluation of antibiotic diffusion into cardiac vegetations DOSAGE SCHEDULE OF AMINOGLYCOSIDES 115 by quantitative autoradiography. J. Infect. Dis. 159:938-944. 4. de Vries, P. J., R. P. Verkooyen, P. Lequit, and H. A. Verbrugh. 1990. Prospective randomized of once-daily versus thrice-daily netilmicin regimens in patients with intraabdominal infections. Eur. J. Clin. Microbiol. Infect. Dis. 9:161-168. 5. Drugeon, H. B., J. Caillon, M. E. Juvin, and J. L. Pirault. 1987. Dynamics of ceftazidime-pefloxacin interaction shown by a new killing curve-chequer board method. J. Antimicrob. Chemother. 19:197-203. 6. Fantin, B., B. Pangon, G. Potel, J. M. Vallois, F. Caron, A. Bure, and C. Carbon. 1989. Ceftriaxone-netilmicin combination in single-daily-dose treatment of experimental Escherichia coli endocarditis. Antimicrob. Agents Chemother. 33:767-770. 7. Garraffo, R., H. B. Drugeon, P. Dellamonica, E. Bernard, and P. Lapalus. 1990. Determination of optimal dosage regimen for amikacin in healthy volunteers by study of pharmacokinetics and bactericidal activity. Antimicrob. Agents Chemother. 34: 614-621. 8. Gerber, A. U., W. A. Craig, H. P. Brugger, C. Feller, A. P. Vastola, and J. Brandel. 1983. Impact of dosing intervals on activity of gentamicin and ticarcillin against Pseudomonas aeruginosa in granulocytopenic mice. J. Infect. Dis. 147:910-917. 9. Hancock, R. E. W., and A. Bell. 1988. Antibiotic uptake into Gram-negative bacteria. Eur. J. Clin. Microbiol. Infect. Dis. 7:713-720. 10. John, J. F., W. F. McNeill, K. E. Price, and P. A. Kresel. 1982. Evidence for chromosomal site specifying amikacin resistance in multiresistant Serratia marcescens. Antimicrob. Agents Chemother. 21:587-591. 11. Juvin, M. E., H. B. Drugeon, J. Caillon, and J. L. Pirault. 1987. Comparaison de l'activitd bactdricide de trois aminosides: gentamicine, tobramycine, amikacine. Pathol. Biol. 35:461-465. 12. Kapusnik, J. E., C. J. Hachbarth, H. F. Chambers, T. Carpenter, and M. A. Sande. 1988. Single, large, daily dosing versus intermittent dosing of tobramycin for treating experimental pseudomonas pneumonia. J. Infect. Dis. 158:7-12. 13. Le Bel, M., and M. Spino. 1988. Pulse dosing versus continuous infusion of antibiotics. Pharmacokinetic-pharmacodynamic considerations. Clin. Pharmacokinet. 14:71-95. 14. Legett, J. E., B. Fantin, S. Ebert, K. Totsuka, B. Vogelman, W. Calame, H. Mattie, and W. A. Craig. 1989. Comparative antibiotic dose-effect relations at several dosing intervals in murine pneumonitis and thigh-infection models. J. Infect. Dis. 159:281-292. 15. Merlin, T. L., G. E. Davis, W. L. Anderson, R. K. Moyzis, and J. K. Griffith. 1989. Aminoglycoside uptake increased by tet gene expression. Antimicrob. Agents Chemother. 33:1549-1552. 16. Moore, R. D., P. S. Lietman, and C. R. Smith. 1987. Clinical response to aminoglycoside therapy: importance of the ratio of peak concentration to minimal inhibitory concentration. J. Infect. Dis. 155:93-99. 17. Morohoshi, T., M. Toriya, S. Yokoiyama, K. Fujimoto, K. Hayano, S. Goto, and A. Tsuji. 1984. The acetylation of 6'-amino group of amikacin by a new enzyme prepared from Serratia sp. J. Antibiot. 37:1687-1691. 18. Nordstrom L., H. Ringberg, S. Cronberg, 0. Tjernstrom, and M. Walder. 1990. Does administration of an aminoglycoside in a single daily dose affect its efficacy and toxicity? J. Antimicrob. Chemother. 25:159-173. 19. Perlman, B. B., and L. R. Freedman. 1971. Experimental endocarditis. II. Staphylococcal infection of the aortic valve following placement of a polyethylene catheter in the left side of the heart. Yale J. Biol. Med. 44:206-213. 20. Potel, G., B. Pangon, B. Fantin, J. M. Vallois, F. Faurisson, and C. Carbon. 1986. Program Abstr. 26th Intersci. Conf. Antimicrob. Agents Chemother., abstr. 575. 21. Powell, S. H., W. B. Thompson, M. A. Luthe, R. C. Stern, D. A. Grossniklaus, D. D. Bloxham, D. L. Groden, M. R. Jacobs, A. 0. Discenna, H. A. Cash, and J. D. Klinger. 1983. Once daily vs continuous aminoglycoside dosing: efficacy and toxicity in ani-

116 POTEL ET AL. ANTIMICROB. AGENTS CHEMOTHER. mal and clinical studies of gentamicin, netilmicin and tobramycin. J. Infect. Dis. 147:918-932. 22. Sanders, C. C., and C. Watanakunakorn. 1986. Emergence of resistance to f-lactams, aminoglycosides and quinolones during combination therapy for infection due to Serratia marcescens. J. Infect. Dis. 153:617-619. 23. Taber, H. W., J. P. Mueller, P. F. Miller, and A. S. Arrow. 1987. Bacterial uptake of aminoglycoside antibiotics. Microbiol. Rev. 51:439-457. 24. Washington, J. A., II, and V. L. Sutter. 1980. Dilution susceptibility test: agar and macro-broth dilution procedures, p. 453-458. In E. H. Lennette, A. Balows, W. J. Hausler, Jr., and J. P. Truant (ed.), Manual of clinical microbiology, 3rd ed. American Society for Microbiology, Washington, D.C.