Pharmacodynamic Profile of Telithromycin against Macrolide- and Fluoroquinolone-Resistant Streptococcus pneumoniae in a Neutropenic Mouse Thigh Model

Abstract

The new ketolide telithromycin has potent in vitro activity against Streptococcus pneumoniae, including strains resistant to penicillin, macrolides, and fluoroquinolones. The aim of the present study was to define the pharmacodynamic profile of telithromycin against S. pneumoniae strains with various resistance profiles in an in vivo system. Ten S. pneumoniae strains were studied; seven exhibited penicillin resistance, six demonstrated macrolide resistance, and two exhibited gatifloxacin resistance. The telithromycin MICs for all isolates were ≤0.5 μg/ml. Using the murine thigh infection model, CD-1/ICR mice were rendered neutropenic and were then inoculated with 10⁵ to 10⁶ CFU of S. pneumoniae per thigh. Telithromycin was administered orally at doses ranging from 25 to 800 mg/kg of body weight/day, with the doses administered one, two, three, or four times a day. The activity of telithromycin was assessed by determination of the change in the bacterial density in thigh tissue after 24 h of treatment for each treatment group and the untreated controls. Pharmacokinetic studies of telithromycin were conducted in infected, neutropenic animals. The levels of protein binding by telithromycin in mice ranged from 70 to 95% over the observed range of pharmacokinetic concentrations. By using either the total or the free concentrations of telithromycin, the area under the concentration-time curve (AUC)/MIC ratio was a strong determinant of the response against S. pneumoniae, regardless of the phenotypic resistance profile. The maximal efficacy (the 95% effective dose) against this cohort of S. pneumoniae strains and bacterial inhibition (stasis) of telithromycin were predicted by ratios of the AUC for the free drug concentration/MIC of approximately 1,000 and 200, respectively.

Telithromycin, the first of the new class of antimicrobials to be developed for clinical use, the ketolides, possesses potent activity against a variety of pathogenic gram-positive species, including penicillin- and macrolide-resistant Streptococcus pneumoniae strains (1, 2). Telithromycin is a semisynthetic derivative of the 14-membered-ring macrolide parent compound erythromycin A. As a consequence of chemical modifications made to the parent compound, telithromycin is highly acid stable, has an extended half-life, and lacks the ability to induce resistance to other macrolides (4). Additionally, these alterations allow telithromycin to retain activity against macrolide-resistant isolates due to the presence of the mefA and/or the ermB gene and against S. pneumoniae strains with 23S rRNA mutations (4, 14). Typical telithromycin MICs at which 90% of isolates are inhibited (MIC₉₀) for S. pneumoniae range from 0.125 to 0.5 μg/ml for both susceptible and penicillin- and/or macrolide-resistant S. pneumoniae isolates (1, 7, 11, 22). Telithromycin is also efficacious against fluoroquinolone-resistant S. pneumoniae strains; the MIC₉₀ for these isolates remain at equally low levels of 0.25 to 0.5 μg/ml (9, 15).

While telithromycin has activity against pneumococci in vitro, the pharmacodynamic profile of this agent against this important pathogen has not been fully described. The purpose of this present study was to generate data which describe the pharmacodynamic relationship between telithromycin exposure and treatment outcome as assessed by determination of the densities of S. pneumoniae isolates with various resistance patterns by using a well-recognized murine thigh infection model. These data were then used to delineate the optimal pharmacodynamic parameter and identify a parameter value associated with efficacy.

(This material was presented in part at the 41st Interscience Conference on Antimicrobial Agents and Chemotherapy, Chicago, Ill., 16 to 19 December 2001.)

MATERIALS AND METHODS

Antimicrobial test agents.

Telithromycin was supplied by Aventis (Antony, France). Penicillin G was obtained from Sigma Chemical Co. (St. Louis, Mo.), and azithromycin and gatifloxacin were supplied by Pfizer, Inc. (Groton, Conn.) and Bristol-Myers Squibb, Co. (Princeton, N.J.), respectively. Oral dosing solutions of telithromycin were prepared as 0.2-ml volumes by dissolving telithromycin powder in 10% (total volume) 95% ethanol and dilution of the solution with a 0.1 M phosphate buffer (pH 6.5). The maximal soluble dose of telithromycin was 200 mg/kg of body weight, which the mice tolerated at up to four daily doses.

Bacterial isolates and susceptibilities.

The MICs of telithromycin, penicillin, azithromycin, and gatifloxacin for all test organisms were determined in duplicate by the microdilution method according to NCCLS guidelines (17) with cation-adjusted Mueller-Hinton broth (20 to 25 mg of calcium per liter, 10 to 12.5 mg of magnesium per liter) with 5% lysed horse blood and incubation in ambient air. Trypticase soy agar with 5% sheep blood was used as the growth medium for all S. pneumoniae strains.

A total of 10 different S. pneumoniae isolates with various telithromycin susceptibilities were used for in vivo testing. This group of test isolates had the following phenotypic classifications: penicillin susceptible and resistant (both intermediate and resistant); macrolide susceptible (azithromycin MIC ≤ 0.5 μg/ml), low-level macrolide resistant (azithromycin MIC = 1.0 to 32 μg/ml), and high-level macrolide resistant (azithromycin MIC ≥ 64 μg/ml); and fluoroquinolone resistant (gatifloxacin MIC ≥ 4.0 μg/ml).

Thigh infection model.

Specific-pathogen-free, female CD-1/ICR mice (weight, approximately 25 g) were obtained from Harlan Sprague-Dawley, Inc., (Indianapolis, Ind.), and were used throughout the study. All procedures were carried out in accordance with the Hartford Hospital Institutional Animal Care and Use Committee. The animals were kept in accordance with National Research Council recommendations (19) and were allowed food and water ad libitum. Mice (three to four mice per treatment regimen) were rendered transiently neutropenic by intraperitoneal injection of cyclophosphamide at a dose of 150 mg/kg of body weight at 4 days and 1 day before inoculation. No mortality related to cyclophosphamide administration was observed before inoculation of any S. pneumoniae isolate. Colonies from a fresh growth of the S. pneumoniae isolate were transferred to cation-adjusted Mueller-Hinton broth with 5% lysed horse blood, incubated at 37°C to obtain logarithmic growth, and subsequently diluted in normal saline to an inoculum of 10⁶ to 10⁷ CFU/ml. The numbers of CFU per milliliter of the final inocula were confirmed by serial dilution and plating techniques. Thigh infection with each of the test isolates was produced by intramuscular injection of 0.1 ml of the inoculum into each thigh of the mice 2 h prior to the initiation of antimicrobial therapy. The animals were then randomized into control and treatment groups, as described below.

Pharmacokinetic studies.

The telithromycin concentrations in mouse serum were measured over 24 h. The animals were prepared as described above. Two hours after pneumococcal inoculation of the thighs, the mice were administered a single oral telithromycin dose of either 50 or 100 mg/kg. Blood was obtained by intracardiac puncture from six animals per dosage group at 0.08, 0.16, 0.25, 0.5, 1, 2, 4, 6, 8, 12, and 24 h following drug administration (a total of 66 animals) after the animals were euthanized. After centrifugation, serum was stored at −80°C until the telithromycin concentrations were determined by a validated high-performance liquid chromatography (HPLC) procedure.

Additional pharmacokinetic studies were performed to confirm the values of the parameters obtained with the doses described above. A 25-mg/kg dose was also administered to animals prepared exactly as described above for the first study. Blood was obtained by intracardiac puncture from four animals at 0.16, 1, 4, and 8 h following drug administration (total of 16 animals). The samples were handled as described above.

Telithromycin concentration determination.

Mouse serum samples were analyzed by HPLC to quantitate the telithromycin concentrations (16). Serum samples to which an internal standard (moxifloxacin; Bayer Corp., West Haven, Conn.) was added were deproteinized by the addition of acetonitrile and centrifugation. An aliquot of the supernatant was dried under a nitrogen stream and reconstituted with the mobile-phase buffer. The mobile-phase buffer consisted of triethylamine, phosphoric acid, and HPLC-grade water at a ratio of 0.37:0.30:99.3 (vol/vol); acetonitrile and methanol were added to the buffer at a ratio of 18:15:67 (vol/vol). The fluorescence of the sample extracts was monitored at 263 nm (filter cutoff, 418 nm) under a mobile-phase rate flow of 1.0 ml/min. Chromatograms were registered on an integrator (EZChrom Elite Chromatography Data System; Scientific Software, San Ramon, Calif.). The concentrations of telithromycin in the mouse serum samples used for pharmacokinetic analysis were calculated from a weighted (1/y) linear regression equation that comprised the peak-height ratios for standard samples prepared with known telithromycin concentrations in mouse serum. The range of detection of this assay was from 0.1 μg/ml (lower limit of detection) to 50 μg/ml, with an average r² value of 0.999. Assay variability was <4.0%, as established with both within-run and day-to-day-run quality control samples.

Pharmacokinetic analysis.

A one-compartment model (WinNonlin, version 3.0; Pharsight Corporation, Mountain View, Calif.) with first-order absorption and first-order elimination was used to characterize the dispositions of the 50- and 100-mg/kg telithromycin doses in infected mice. The resulting mean values of the volume of distribution (V), K_a, and k_el were used to simulate all dosing regimens. Confirmation of the mean concentrations resulting from the 25-mg/kg dose in the pharmacokinetic analysis were plotted with the corresponding dose-appropriate simulation curve to estimate the accurateness of the curve-fitting model.

Protein binding studies.

Freshly collected serum was used for the protein binding studies, as both the bioactive protein and the pH in serum collected and stored under refrigerated or frozen conditions may undergo substantial changes, which may lead to inaccurate determination of the level of protein binding in vivo. Whole blood was obtained from CD-1/ICR mice (noninfected, immunocompetent) by intracardiac puncture and was allowed to clot at room temperature. Serum was then collected so that binding studies could be conducted immediately after centrifugation.

Protein binding testing was conducted in triplicate with Amicon Centrifree Micropartition devices (molecular weight cutoff, 30,000; Millipore, Bedford, Mass.), according to the instructions provided in the manufacturer's package insert. An aqueous telithromycin stock solution was prepared; and dilutions were made in fresh mouse serum to yield final concentrations of 2, 5, 10, 15, and 25 μg/ml. Each of the serum solutions was heated at 37°C in a shaking water bath for 10 min. Exactly 0.9 ml of each serum solution was transferred into three ultrafiltration devices and centrifuged for 25 min at 1,000 × g at 10°C to achieve an ultrafiltrate volume of approximately 250 μl. The concentrations of telithromycin in both the initial serum solutions and the ultrafiltrates were determined in triplicate by a validated HPLC method. To determine the nonspecific binding of telithromycin to the filter apparatus, mouse serum ultrafiltrate was spiked with telithromycin and was handled as described above for the mouse serum.

The percent protein binding of telithromycin at each concentration of telithromycin prepared and the amount of nonspecific binding were calculated by using the following equation: percent protein binding = [(total concentration − filtered concentration)/total concentration] × 100, where the total concentration is the telithromycin concentration in the initial serum or ultrafiltrate solution, and the filtered concentration is the concentration in the serum ultrafiltrate or (refiltered) telithromycin-spiked ultrafiltrate.

Bacterial density assessment.

Once the animals were prepared and inoculated as described above, telithromycin treatment began at 2 h postinfection and was continued over a total of 24 h for each regimen. The telithromycin concentration and treatment regimen were varied to create a range of drug exposures for each S. pneumoniae isolate, with the total daily dosages increased from 2- to 20-fold. The range of doses of telithromycin administered orally was from 25 to 800 mg/kg/day, divided into doses of 25 to 200 mg/kg administered one, two three, or four times a day. The majority of the telithromycin regimens tested against all S. pneumoniae isolates were multiple-dose regimens (26 of 43 in total). Control animals received blank solution orally in the same volume and on the same schedule used for the telithromycin administrations. Half of the untreated control mice were enthanized at the initiation of therapy and the remaining half were euthanized after 24 h. The treated mice were killed after 24 h of therapy.

After the mice were euthanized, both thighs were removed and individually homogenized in normal saline. Serial dilutions were plated on Trypticase soy agar with 5% sheep blood. To limit antibiotic carryover, the cultures were incubated in a 5% CO₂ atmosphere (to enhance bacterial growth and decrease the activity of telithromycin) for 24 to 48 h, at which time the final CFU counts were determined (18). The limit of detection was 50 CFU per ml of thigh homogenate. The effect of telithromycin treatment was measured by determination of the change in bacterial density and was calculated by subtracting the mean log CFU per ml of thigh homogenate for the control mice obtained 2 h after bacterial inoculation from the log CFU per ml of thigh homogenate for the telithromycin-treated or untreated control mice at the end of therapy (24 h).

Pharmacodynamic and statistical assessments.

All calculated changes in the values of log CFU per ml of thigh homogenate were compared to each of the following pharmacodynamic parameters: the percentage of time that the telithromycin concentrations were above the MIC within the 24-h treatment period (T > MIC), the peak concentration (C_max)/MIC, and the area under the concentration-time curve (AUC)/MIC. These comparisons were evaluated for significant correlations by using Spearman's rank correlation coefficient test. A P value of <0.05 was considered significant. The data were then characterized by use of an inhibitory effect sigmoidal maximum effective concentration (E_max) model (WinNonlin), and the static and 95% maximum log reduction (95% effective dose [ED₉₅]) values were established. The values of the pharmacodynamic parameters which represent stasis were determined from the line of the E_max curve corresponding to the telithromycin treatments that inhibited log-phase growth (i.e., that resulted in no net change in growth) of the S. pneumoniae isolates at 24 h.

RESULTS

Of the 10 strains studied, 7 were nonsusceptible to penicillin, 6 were macrolide resistant, and 2 were gatifloxacin resistance (Table 1). Six strains were resistant to two classes of antimicrobials; the remaining isolates were resistant to one of the four classes tested (n = 2) or were sensitive to all classes of antimicrobials tested (n = 2). Despite the phenotypic profile of resistance to the β-lactam, macrolide, and fluoroquinolone classes, telithromycin displayed potent in vitro activities against the pneumococci used in this study. All S. pneumoniae isolates were susceptible to telithromycin, with MICs ranging from 0.006 to 0.5 μg/ml.

TABLE 1.

Median MICs of telithromycin, penicillin, gatifloxacin, and azithromycin for the S. pneumoniae isolates tested^a

Strain no.b	Characteristicc	TEL MIC (μg/ml)d	PEN		GAT		AZM
			MIC (μg/ml)	NCCLSe	MIC (μg/ml)	NCCLSe	MIC (μg/ml)	NCCLSe
93	gyrA parC parE	0.006	0.5	I	32.0	R	≤0.03	S
21	S, serotype 1	0.008	0.03	S	1.0	Sf	0.06	S
44	S, serotype 3	0.008	0.06	S	1.0	S	0.03	S
128	ETP resistant	0.008	8.0	R	0.5	S	0.03	S
126	MDR (ETP resistant)	0.015	8.0	R	0.5	S	>16.0	R
95	ermB	0.03	0.03	S	0.5	S	>128	R
92	gyrA parC	0.03	0.25	I	8.0	R	0.5	S
109	mefA, serotype 9V	0.125	4.0	R	0.5	S	16.0	R
102	mefA	0.25	4.0	R	0.5	S	8.0	R
71	MDR (ermB), serotype 14	0.5	8.0	R	0.375	S	>128	R

^aTEL, telithromycin; PEN, penicillin; GAT, gatifloxacin; AZM, azithromycin.

^bA total of eight isolates were nonsusceptible to the drugs tested: seven were nonsusceptible to penicillin, two were nonsusceptible to gatifloxacin, and five were nonsusceptible to azithromycin. None of the isolates was nonsusceptible to telithromycin.

^cKnown genotype, phenotype, and/or serotype of strain. S = susceptible to β-lactams; fluoroquinolones, macrolides, ketolides, and carbapenems; R, resistant; ETP, ertapenem; MDR, multidrug resistant.

^dThe NCCLS breakpoint has been established at a value of 1.0 μg/ml.

^eNCCLS, NCCLS breakpoint interpretations. R, resistant; I, intermediate; S, susceptible.

^fIntermediate to levofloxacin at 4.0 μg/ml.

The pharmacokinetic parameters calculated from dosing regimen simulations were a V of 0.289 liter, K_a of 3.28 h⁻¹, and k_el of 0.078 h⁻¹. The resultant elimination half-life (t_1/2β) for all regimens was 8.9 h, and the time to C_max was 1.17 h. Figure 1 displays the observed mean and predicted telithromycin concentrations after administration of a single 25-, 50-, or 100-mg/kg dose to infected mice. The concentrations observed for the 25-mg/kg dose validated the appropriateness of the concentration-time profile, as delineated by the preceding data. With the exception of the mean concentration at 8 h, when the concentrations were overestimated by model, the mean concentrations fell within 25% of the predicted values. Total AUCs for the range of telithromycin doses administered (25 to 200 mg/kg) were 29.8 to 953.2 μg · h/ml. Total C_maxs ranged from 2.12 to 39.2 μg/ml.

FIG. 1.

Pharmacokinetic profile of telithromycin after administration of single doses to infected mice. ▴, observed profile with 100 mg/kg; •, observed profile with 50 mg/kg, •, observed profile with 25 mg/kg; dashed line, predicted profile with 100 mg/kg; heavy solid line, predicted profile with 50 mg/kg; solid line, predicted profile with 25 mg/kg.

The percentage of protein-bound telithromycin in CD-1/ICR mice was measured at concentrations ranging from 2 to 25 μg/ml and ranged from 94.6 to 69.8% (Table 2). The free versus the bound telithromycin concentrations followed Michaelis-Menten kinetics, as shown in Fig. 2. Unbound telithromycin concentrations were determined on the basis of this plot of the bound versus the unbound telithromycin concentrations and were used for the pharmacodynamic assessments. Any overestimation of the telithromycin concentration by the predictive model used would therefore be overcome, particularly at lower concentrations. Thus, the range of AUCs of free telithromycin was 1.79 to 235.9; free C_maxs ranged from 0.13 to 21.9 μg/ml. The level of nonspecific binding of telithromycin in the filter device was 0.98% ± 6.45%.

TABLE 2.

Protein binding of telithromycin in CD-1/ICR mouse serum

Nominal concn (μg/ml)	Measured concn (μg/ml)	% Bound	% SD
2	1.76	94.6	0.003
5	5.25	93.7	0.009
10	10.9	89.4	0.019
15	16.2	79.2	0.015
25	26.6	69.8	0.011

FIG. 2.

Free versus bound telithromycin (TEL) concentrations in CD-1/ICR mouse serum (r² = 0.981).

Previously reported data on protein binding in mice support these findings, with 89% binding seen at concentrations less than 1.0 μg/ml and 66 to 79% binding observed at concentrations of 5 and 10 μg/ml (O. Vesga, C. Bonnat, and W. A. Craig, Abstr. 37th Intersci. Conf. Antimicrob. Agents Chemother., abstr. F-255, 1997). Overall, these data compare favorably with the serum protein binding levels determined in humans (approximately 60 to 70%) over the range of clinically achievable exposures (20).

Infection was initiated in the murine model by injection of 0.1 ml of 10⁶ to 10⁷ CFU of S. pneumoniae into each thigh. An assessment of the inoculation technique over the course of the study (quantitative culture prior to the initiation of therapy, i.e., 2 h after inoculation of the thigh) revealed adequate recovery of pneumococci (5.75 ± 0.43 log CFU; range, 5.11 to 6.48 log CFU; n = 6 to 8 thighs per isolate) from the thighs for all isolates. After 24 h, the expected increase in bacterial density was seen, with an average increase of 2.21 ± 0.99 log CFU for the untreated control animals (n = 6 to 8 thighs per isolate) inoculated with all isolates except S. pneumoniae 93. At 24 h, the number of CFU of isolate 93 declined 0.5 log in this model. The corresponding changes in the numbers of CFU were complementary to data acquired for the other isolates, and the pharmacodynamic assessment in this study was not significantly altered by the inclusion or the exclusion of the values for S. pneumoniae 93; thus, these data were retained in the data set. Antibiotic carryover was minimized by incubation in CO₂ and a lengthened incubation time. A value of 50 CFU for the thigh cultures was used as the lower limit of detection.

A significant (P ≤ 0.001) correlation coefficient was observed by Spearman's rank test for the relationships between the change in CFU and AUC/MIC, the change in CFU and percent T > MIC, and the change in CFU and C_max/MIC for telithromycin (free and total) after 24 h of treatment. While the pharmacodynamic profile of telithromycin was well predicted by both the total and the free concentrations of the compound, the bioactive portion or the free concentration-versus-time profile was integrated with the quantitative culture results for the purposes of parameterization in the present study. Figure 3 demonstrates the pharmacodynamic profile of telithromycin obtained with free drug concentrations, as determined with the sigmoid E_max model.

FIG. 3.

Sigmoidal effect model for pharmacodynamic parameters of free telithromycin concentration versus change in log₁₀ CFU in mouse thighs after 24 h of treatment. Correlation coefficients: r² = 0.86 for AUC/MIC (a); r² = 0.84 for C_max/MIC (b); r² = 0.61 for percent T > MIC (c). The different symbols and numerical values correspond to the different pneumococcal isolates noted in Table 1.

A predictive relationship was found to exist between the change in CFU and the AUC for free drug concentrations (AUC_free)/MIC (r² = 0.86). A bacteriostatic effect appeared to occur, with an AUC_free/MIC of approximately 200, while the highest bactericidal effects (ED₉₅s) were observed when the AUC_free/MIC ratios were >1,000. Additionally, the C_max of the free drug (C_{max free})/MIC was also predictive of reductions in the numbers of CFU (r² = 0.84), with bacterial stasis appeared at a C_{max free}/MIC ratio of 11, and maximal bacterial killing was evident at a C_{max free}/MIC ratio of approximately 90. The goodness of fit for percent T > MIC of the free drug (MIC_free) was poor (r² = 0.61) compared to those for the other parameters. This is illustrated in Fig. 3, in which, despite the achievement of a 100% T > MIC_free, quantitative cultures revealed a wide range of in vivo activities (greater than 2 logs of bacterial growth to 4.5 logs of bacterial killing) in the face of maximum exposure for this parameter.

In addition to the potent in vitro activity of telithromycin in spite of the pneumococcal phenotypic resistance profile noted in Table 1, this compound was also potent in vivo. This is supported by the observations in Fig. 3, which shows that similar exposures of the isolates to telithromycin according to the various pharmacodynamic parameters studied (e.g., AUC_free/MIC) resulted in comparative in vivo antimicrobial effects.

DISCUSSION

The ketolides are a new class of antimicrobials with potent activities against the pathogens commonly associated with community-acquired respiratory tract infections (CARTIs), including S. pneumoniae. Telithromycin is the first of the ketolide compounds to be introduced into clinical practice for the treatment of CARTIs. Moreover, numerous in vitro studies have established the activity of this antimicrobial against isolates of S. pneumoniae which are resistant to those antibiotics traditionally used for the treatment of CARTIs, such as the β-lactams, erythromycin, as well as the newer fluoroquinolone compounds (2, 6, 7, 9). These observations are also supported by the present in vitro portion of the study, as telithromycin maintained potent activity (MICs ≤ 0.5 μg/ml) regardless of the phenotypic profile of susceptibility of our pneumococci to penicillin, azithromycin, and gatifloxacin.

While telithromycin has been demonstrated to have in vivo activity in several animal models of infection (3, 21, 23), limited data which describe the pharmacodynamic profile of this compound against S. pneumoniae are available (Vesga et al., 37th ICAAC). In turn, we evaluated the relationship between telithromycin exposure and outcome by assessing the changes in bacterial density in the pneumococcal thigh infection model. The results of the present study revealed that both AUC/MIC and C_max/MIC were correlated with the in vivo bactericidal activity of telithromycin. Despite the inclusion of some dose fractionation studies, we were unable to identify the pharmacodynamic parameter which is the most predictive of outcome, since the parameters noted above are highly interrelated. Furthermore, our analysis revealed that both of these parameters were the greatest predictors of success for the outcome studied, regardless of whether the total or the free concentration-versus-time profile of telithromycin was used for this assessment. By taking into consideration the fact that both parameters are predictive of the bactericidal activity of telithromycin, it can be concluded that the bactericidal nature of telithromycin is concentration dependent.

Our results are in accordance with the previous findings of an in vivo study reported by Vesga and colleagues (Vesga et al., 37th ICAAC) regarding the high predictive value of the AUC/MIC ratio for telithromycin, followed by C_max/MIC, in a similar neutropenic infection model. Furthermore, while the 10 pneumococcal strains used in the present study were different from the single isolate used in the previous study noted above, the 24-h AUC_free/MIC exposures required to produce both static and bactericidal effects in the murine models were similar. In both studies, telithromycin AUC_free/MIC ratios of approximately 200 appeared to produce stasis, whereas ratios ≥1,000 were needed to maximize the dynamic response (ED₉₅), as judged by the bactericidal activity of the compound in infected thighs.

While the extent of protein binding may substantially alter the magnitude of the pharmacodynamic parameter of interest, the degree of protein binding of telithromycin in mice is comparable to that in humans, as noted above. Therefore, since the binding of telithromycin to serum proteins is comparable for the two species, the data generated from the murine model should be predictive of activity in humans without mathematical transformation. Moreover, the similarities in the correlation of our pharmacodynamic data, irrespective of the use of total or free telithromycin concentrations, and data on the levels of protein binding among species of interest allow the use of pharmacodynamic parameterization with either the total or the free concentrations of telithromycin in humans.

While similarities in protein binding between mice and humans allow comparisons of the pharmacodynamic profile of telithromycin without concern for marked changes in the bioactive (i.e., free drug) concentration-versus-time profile, it must be noted that the present studies with mice result in conservative estimates of drug efficacy due to the neutropenic status of the host. As a result, while the pharmacodynamic parameter most predictive of outcome (i.e., AUC/MIC) remains constant in the absence or the presence of neutrophils, the magnitude of drug exposure required to produce any given reduction in bacterial density is substantially reduced in the immunocompetent host. To this end, Andes and colleagues (D. Andes, O. Vesga, and W. A. Craig, Abstr. 4th Int. Conf. Macrolides, Azalides, Streptogramins, Ketolides Oxazolidinones [ICMAS-KO4], abstr. 1-19, 1998) evaluated the impacts of neutrophils on the bacteriostatic activity of telithromycin and a structurally related ketolide against S. pneumoniae strains with various macrolide susceptibilities in the murine thigh infection model and reported that the potencies of the ketolides are enhanced (1.8- to 24-fold) in the presence of neutrophils. Moreover, although neutropenic conditions were used in the present study, the enhanced in vivo potency of another ketolide, ABT-773, against pneumococci in immunocompetent hosts has also recently been demonstrated (5).

Moreover, while the use of the murine thigh infection model has been paramount in the pharmacodynamic profiling of various antimicrobial classes, this model may not accurately portray the resulting drug accumulation at the site of pulmonary infection (i.e., the concentrations in epithelial lining fluid) and may therefore overestimate the magnitude of drug exposure required for microbiologic eradication. This issue has been raised with the macrolide clarithromycin in a study previously conducted by our group (24). Moreover, an additional study was conducted with clarithromycin to evaluate the differences in the dispositions and the resulting changes in bacterial density in murine pneumonia and thigh infection models (12). In that study the activity of the macrolide was evaluated against seven S. pneumoniae isolates with efflux-mediated resistance in both models. In addition, the intrapulmonary disposition of clarithromycin was also determined in the pneumonia model. Consistent bacterial killing was observed in the pneumonia model, whereas no drug effect was observed in the thigh infection model. This difference in the bacterial killing profile between the models was supported by the higher macrolide concentrations observed in epithelial lining fluid compared to those observed in serum. While this apparent difference between the pneumonia and thigh infection models was not detected in a previous study (W. A. Craig and D. R. Andes, Abstr. 41st Intersci. Conf. Antimicrob. Agents Chemother., abstr. A-1097, 2001) conducted with clarithromycin and three S. pneumoniae strains, the doses of the antimicrobial used in that study (0.38 to 48 mg/kg administered every 6 h) were considerably different from the regimen of 200 mg/kg administered every 12 h used in our previous study (12). As a result of these murine studies, we must interpret the data that have been generated with the thigh infection model cautiously when assessing the magnitude of exposure required to ensure microbiologic eradication by treatment with telithromycin in humans, since this compound also displays the ability to accumulate in epithelial lining fluid, as the concentrations at this pulmonary site exceed the corresponding concentrations in plasma (10, 13). Lastly, it must be noted that the murine model evaluates the effectiveness of telithromycin over a relatively short (24-h) treatment period rather than the 5- to 10-day regimen used in the clinical setting. Because of these factors, one might anticipate that a substantially lower AUC/MIC ratio in immunocompetent humans would translate into high microbiologic eradication rates. While differences among the pharmacodynamic parameter estimates derived through evaluations with various animal models of infection and infected patients do exist, animal models provide important insights into the parameter most closely linked to efficacy in humans and also provide conservative estimates of the drug exposure required to ensure optimal efficacy prior to the initiation of clinical trials with humans.

In conclusion, our data demonstrate the in vitro and in vivo potencies of telithromycin against S. pneumoniae strains exhibiting a wide range of phenotypic profiles of resistance to the anti-infective compounds commonly used for the management of CARTIs. Moreover, the findings of the present study confirm the appropriateness of the AUC/MIC ratio as a dynamic predictor of response for telithromycin. With its enhanced pharmacokinetic and pharmacodynamic profile and the ability to overcome S. pneumoniae resistance (8, 20), telithromycin should be a useful addition to the antimicrobial arsenal for the treatment of patients with CARTIs.