Efficacy of Telavancin in a Rabbit Model of Aortic Valve Endocarditis Due to Methicillin-Resistant Staphylococcus aureus or Vancomycin-Intermediate Staphylococcus aureus

The activities of telavancin and vancomycin were compared invitro and in the rabbit model of aortic valve endocarditis againsta methicillin-resistant Staphylococcus aureus strain, COL, anda vancomycin-intermediate S. aureus (VISA) strain, HIP 5836.Telavancin was bactericidal in time-kill studies at a concentrationof 5 µg/ml against both COL and HIP5836. Vancomycin wasbacteriostatic at 5 µg/ml and bactericidal at 10 µg/mlagainst COL and was bacteriostatic at 10 µg/ml againstVISA strain HIP 5836. Compared to untreated controls, a twice-dailyregimen of 30 mg/kg of telavancin reduced mean aortic valvevegetation titers of the COL strain by 4.7 log₁₀ CFU/g after4 days of therapy and sterilized 6/11 vegetations compared to3.4 log₁₀ CFU/g with 3/10 vegetations sterilized for a regimenof twice-daily vancomycin, 30 mg/kg; these differences werenot statistically significant. Telavancin was significantlymore effective than vancomycin in the VISA model, producinga 5.5 log₁₀ CFU/g reduction versus no reduction in CFU withvancomycin. In experiments comparing 2-day regimens of telavancinat 30 mg/kg and 50 mg/kg twice daily, organisms were rapidlyeliminated from vegetations, but the effect was not differentbetween the two doses. These results suggest that telavancinmay be an effective treatment for endocarditis and other seriousstaphylococcal infections accompanied by bacteremia, includinginfections caused by staphylococci not susceptible to vancomycin.

INTRODUCTION

Top

Introduction

Telavancin (formerly known as TD-6424) is an investigationallipoglycopeptide derivative of vancomycin (4, 5). It is morerapidly bactericidal than vancomycin in vitro, exhibits concentration-dependentkilling, and is active against strains with reduced susceptibilityto vancomycin (6, 7). This in vitro activity is likely due tothe novel, multiple mechanisms of action of telavancin, includingeffects on membrane permeability as well as inhibition of peptidoglycansynthesis (4). In animal models of soft-tissue infection, telavancinis efficacious and more potent than vancomycin or linezolid(3). Telavancin is in phase 3 clinical trials as a once dailyagent for treatment of serious infections caused by gram-positivepathogens. The purpose of the present study was to compare theefficacy of telavancin to vancomycin against antibiotic-resistantStaphylococcus aureus in a rabbit model of aortic valve endocarditis.

(This work was presented in part at the 43rd Interscience Conferenceon Antimicrobial Agents and Chemotherapy, Chicago, Illinois,14 to 17 September 2003.)

MATERIALS AND METHODS

Top

Materials and Methods

Bacterial strains. In separate experiments, endocarditis was established with eitherof two S. aureus isolates: the homogeneous, highly methicillin-resistantS. aureus (MRSA) strain COL or a vancomycin-intermediate S.aureus (VISA) strain, HIP 5836, which is a methicillin-resistant(nafcillin MIC, 16 µg/ml), ß-lactamase-producingclinical isolate from a patient in New Jersey (originally designatedstrain 992) (2, 9).

Susceptibility studies. MICs were determined by the standard microdilution method incation-supplemented Mueller-Hinton broth with an inoculum of2 x 10⁵ CFU/ml after a 24-h incubation at 35°C. Time-killstudies were performed using a 1:100 dilution of an overnightculture in cation-adjusted Mueller-Hinton broth to compare thebactericidal activities of telavancin and vancomycin; 50-µlsamples of culture were taken after 0, 4, and 24 h of incubationat 37°C, serially diluted 10-fold, and inoculated onto bloodagar. Colonies were counted after incubation for 24 h at 37°C.

Endocarditis model. The animal studies were approved by the Committee on AnimalResearch of the University of California, San Francisco. Endocarditisof the aortic valve was established in 2- to 3-kg New ZealandWhite rabbits by positioning a catheter across the aortic valveand securing it in place for the duration of the experiment.Animals were infected 48 h after insertion of the catheter byintravenous injection of 1 ml of bacterial suspension containing ${approx}$ 10⁶ CFU in 0.9% saline. Experiments were conducted to comparethe activities of telavancin and vancomycin in animals infectedwith either COL or the VISA strain HIP 5836 and to assess dose-responsewith telavancin.

For comparison of telavancin and vancomycin, rabbits were infectedwith one of the two strains and then were randomly assignedto one of three groups: untreated controls; a 4-day treatmentregimen of telavancin administered intravenously at 30 mg/kgtwice a day; or a 4-day regimen of vancomycin administered intravenouslyat 30 mg/kg twice a day. Therapy was begun 16 to 18 h afterbacterial challenge.

Experiments performed to assess the dose-response of telavancinwere conducted with rabbits infected with the COL strain. Rabbitswere randomly assigned to one of the following three groups:untreated controls; a 2-day treatment with telavancin administeredintravenously at a dose of 30 mg/kg twice daily; or a 2-daytreatment with telavancin administered twice daily at a doseof 50 mg/kg. Therapy was begun 16 to 18 h after bacterial challenge.

Controls were sacrificed approximately 18 h after infectionto determine the number of organisms present prior to therapy.Rabbits assigned to treatment groups were sacrificed 18 to 24h after the last dose of telavancin or vancomycin. Aortic valvevegetations were harvested and weighed. Samples were homogenizedin 0.5 ml of 0.9% saline and quantitatively cultured onto bloodagar. After incubation at 37°C for 24 to 48 h, colonieswere counted and the result was expressed as log₁₀ CFU per gramof vegetation. The lower limit of detection by this method isapproximately 1 log₁₀ CFU/g. Vegetations in which no organismswere detected were scored as sterile and assigned a value of0 log₁₀ CFU/g for purposes of data analysis. Differences betweenmean vegetation titers of the several groups were tested forstatistical significance by analysis of variance followed byStudent's t test adjusted for multiple comparisons by Bonferronicorrection.

Serum drug concentrations of telavancin were determined fromblood samples obtained from the marginal ear vein at 1 h and12 h after dosing. Serum was collected and frozen at –70°C.Telavancin was assayed by the manufacturer using a validatedliquid chromatography/mass spectrometry/mass spectroscopy method(3). Telavancin has a longer elimination half-life in humans(7 to 9 h) (8) than in rabbits (1 to 2 h) (data not shown).Telavancin administered to rabbits at 30 mg/kg twice a day wasestimated to achieve similar 24-h area-under-the-curve (AUC)and time above the MIC as a clinical dose of 7.5 mg/kg (30-mininfusion) in humans (8). Pharmacokinetic data for vancomycinhave been determined previously (1). Vancomycin has a half-lifeof 80 min in rabbits, and a regimen of 30 mg/kg administeredintravenously twice a day produces serum concentrations (totaldrug) 1 h after dosing of 29 µg/ml, which is similar tothat seen in humans (1).

RESULTS

Top

Results

Susceptibility studies. The MICs of telavancin were 1 µg/ml against the MRSA strainCOL and 4 µg/ml against the VISA strain HIP 5836. VancomycinMICs were 2 and 8 µg/ml, respectively. Telavancin wasbactericidal (>3 log₁₀ decrease in CFU at 24 h of drug exposure)in time-kill studies at a concentration of 5 µg/ml againstboth COL and HIP 5836 (Fig. 1). Vancomycin was bacteriostaticat 5 µg/ml and bactericidal at 10 µg/ml againstCOL. It did not inhibit growth of the VISA strain, HIP 5836,at 5 µg/ml and was bacteriostatic at 10 µg/ml.

View larger version (19K):
[in this window]
[in a new window]
FIG. 1. Representative time-kill curves for methicillin-resistant S. aureus strain COL (A) and vancomycin-intermediate S. aureus strain HIP 5836 (B) for vancomycin at 5 µg/ml and 10 µg/ml (VAN 5 and VAN 10, respectively) and telavancin at 5 µg/ml and 10 µg/ml (TLV 5 and TLV 10, respectively).

Endocarditis model. The mean ± standard deviation values for serum concentrationin infected rabbits at 1 h after a 30-mg/kg dose and a 50-mg/kgdose of telavancin were 102 ± 11 (n = 6) and 133 ±24 µg/ml (n = 6) (P < 0.025), respectively. The meanconcentration 12 h after the 30-mg/kg dose was 0.4 ±0.6 (n = 11). Telavancin concentrations at 12 h following the50-mg/kg dose were not assayed.

Telavancin reduced vegetation titers of COL by a mean of 4.7log₁₀ CFU/g compared to untreated rabbits (Table 1), a statisticallysignificant decrease. Vancomycin produced a mean reduction of3.4 log₁₀ CFU/g, a decrease that was not significantly differentfrom either controls or telavancin-treated rabbits. Six of 11telavancin-treated rabbits had sterile vegetations, comparedto 3 of 10 vancomycin-treated rabbits (P = 0.38, Fisher exacttest).

View this table:
[in this window]
[in a new window]
TABLE 1. Comparative treatment outcomes between telavancin (TLV) and vancomycin (VAN) in the rabbit model of aortic valve endocarditis caused by either of two strains of Staphylococcus aureus

Vancomycin had no effect on the vegetation titers of rabbitsinfected with the VISA strain HIP 5836, with the organism burdenin vegetations being slightly higher than in the controls. Noneof the rabbits treated with vancomycin had sterile vegetations,and four died during treatment. Telavancin produced an approximately5.5 log₁₀ reduction in CFU compared to the controls. Four rabbitshad sterile vegetations, and one died during the treatment period.

In the 2-day dose escalation studies comparing telavancin administeredintravenously at a dose of 30 mg/kg and 50 mg/kg twice a day,both regimens significantly reduced the overall mean vegetationtiters compared to untreated controls (Table 1). The two dosageregimens had similar efficacy. Notably, the number of residualorganisms in vegetations after 2 days and 4 days of exposureto telavancin were similar, suggesting that most of the killingoccurs early during therapy.

DISCUSSION

Top

Discussion

Telavancin is an investigational lipoglycopeptide with multiplemechanisms of action (4, 5). In vitro it is more active thanvancomycin, and its spectrum of activity includes vancomycin-intermediateand -resistant staphylococci and enterococci. This expandedspectrum is thought to be due to the presence of a hydrophobicside chain on the vancosamine sugar of the molecule that interfereswith critical membrane functions (10). The antibacterial activitiesof telavancin and vancomycin in vivo in the rabbit model ofaortic valve endocarditis resembled their relative in vitroactivities. The MICs of telavancin were half those of vancomycinagainst the MRSA strain COL and the VISA strain HIP 5836. Intime-kill studies telavancin was bactericidal against both strains,whereas vancomycin was only bactericidal at the 10 µg/mlconcentration against the COL strain. Telavancin was more activethan vancomycin for endocarditis caused by either strain, althoughthe difference was statistically significant only for infectioncaused by the VISA strain, for which vancomycin was ineffective.In contrast, telavancin sterilized vegetations in four of sixrabbits infected with the VISA strain and reduced mean vegetationtiters by more than 5 log₁₀ CFU/g.

There was an appreciable early bactericidal effect present invivo (as occurred in vitro) with mean reductions of log₁₀ CFU/gof 3.3 to 4.2 after 2 days compared to 4.7 after 4 days of treatment.The mean serum concentrations of 102 µg/ml and 133 µg/mlthat were achieved 1 hour after dosing with 30 mg/kg and 50mg/kg, respectively, are similar to that of approximately 90to 110 µg/ml observed in humans receiving 7.5 to 12.5mg/kg (8). While these serum concentrations are statisticallysignificantly different for the 30-mg/kg versus the 50-mg/kgdose, it is unlikely that only a 33% increase in serum concentrationwould be therapeutically significant, as the results indicate.It was not possible, therefore, to assess the role of the concentrationdependence of telavancin in these experiments.

Serum binding of telavancin in the rabbit was not measured inthis study. However, serum binding of telavancin has been measuredin both mouse and human plasma to be 94 to 96% (3) and 93% (8),respectively. Therefore, if one assumes that serum binding inrabbits is in the same range, the free drug levels in the rabbitsin this study are likely to be comparable to the free drug levelsat the clinical doses in humans.

In summary, telavancin was bactericidal in vitro and in vivoagainst two MRSA strains, including a VISA strain. It was atleast as active as vancomycin for the strain susceptible toboth telavancin and vancomycin and more active than vancomycinagainst the VISA strain. Rapid sterilization of vegetationsoccurred within 2 days of treatment. These results suggest thattelavancin may be an effective treatment for endocarditis aswell as other serious staphylococcal infections, including thosecaused by staphylococci not susceptible to vancomycin.

ACKNOWLEDGMENTS

Financial support for these studies was provided by Theravance,Inc.

FOOTNOTES

* Corresponding author. Mailing address: University of California at San Francisco, Division of Infectious Diseases at SFGH, Box 0811, San Francisco, CA 94143. Phone: (415) 206-5437. Fax: (415) 648-8425. E-mail: chipc{at}itsa.ucsf.edu.

REFERENCES

Top