Influence of Macrolide Susceptibility on Efficacies of Clarithromycin and Azithromycin against Streptococcus pneumoniae in a Murine Lung Infection Model

Abstract

We evaluated the activities of clarithromycin and azithromycin against 19 isolates of Streptococcus pneumoniae using a neutropenic lung infection model. The isolates included five susceptible isolates (clarithromycin and azithromycin MICs, ≤0.12 μg/ml), nine isolates exhibiting low-level, mefA-mediated resistance (clarithromycin and azithromycin MICs, 0.5 to 32 μg/ml), and five isolates expressing high-level, ermB-mediated macrolide resistance (clarithromycin and azithromycin MICs, ≥64 μg/ml). Infected mice were administered either saline (control), clarithromycin (4, 40, or 200 mg/kg of body weight twice daily or 200 mg/kg once daily), or azithromycin (4, 40, or 200 mg/kg once daily or 40 mg/kg twice daily) by oral gavage for 72 h. Mortality was assessed at regular intervals for 10 days, and survival in each group was compared to that of untreated controls. Animals infected with susceptible isolates demonstrated significant improvement in survival compared to the controls following treatment with either agent at doses of ≥40 mg/kg. In contrast, none of the regimens improved the survival of animals infected with isolates exhibiting high-level macrolide resistance. Among mice infected with strains expressing low-level resistance, significant improvement in survival compared to the controls was noted among isolates treated with clarithromycin at 40 (seven of nine isolates) and 200 (nine of nine isolates) mg/kg twice a day and with azithromycin at 40 (one of nine isolates) and 200 (three of nine isolates) mg/kg once a day. Animals infected with isolates of S. pneumoniae exhibiting low-level, mefA-mediated macrolide resistance responded to treatment with clarithromycin at rates similar to those observed among mice infected with fully susceptible isolates.

Streptococcus pneumoniae is a commonly encountered pathogen that is associated with a variety of community-acquired respiratory tract infections (6, 15, 19). Although prior to the 1980s S. pneumoniae was believed to be virtually universally susceptible to penicillin and a number of other commonly used antimicrobials, the decade of the 1990s witnessed the explosive emergence of resistance in the pneumococcus. Recent data from 34 U.S. medical centers revealed that ∼30% of S. pneumoniae isolates exhibited reduced susceptibility to penicillin, with 12% exhibiting high-level penicillin resistance (7). Additionally, 18 to 19% of pneumococcal isolates were resistant to macrolides. Interestingly, only 5.6% of the isolates were resistant to clindamycin, suggesting that the majority of the macrolide resistance was mediated by a mechanism other than target site alteration.

Macrolide resistance in S. pneumoniae is generally attributed to two primary mechanisms: target site alteration and active drug efflux. Methylation of the 23S rRNA moiety by a methyltransferase enzyme encoded by the ermB gene produces a conformational change at the macrolide ribosomal binding site. In addition to effectively blocking the macrolide target site on the ribosome, this conformational change also interferes with the ability of lincosamides and streptogramins to bind to their target sites. As a result, resistance arising from ermB-mediated ribosomal alterations is commonly referred to as macrolide, lincosamide, streptogramin B resistance. Phenotypes resulting from the expression of the ermB gene generally exhibit high levels of macrolide resistance (macrolide MICs, >64 μg/ml). Currently, ∼30% of the macrolide resistance among S. pneumoniae isolates in the United States is mediated by modification of the ermB gene (16). The second mechanism responsible for macrolide resistance in S. pneumoniae is the expression of an active drug efflux which is encoded by the mefA gene. Unlike ermB-mediated resistance, mefA-mediated resistance is relatively specific for 14- and 15-member macrolides and is thus referred to as the M phenotype. Additionally, the level of resistance resulting from efflux pump expression is generally less than that produced by target site modification (macrolide MICs of 1 to 32 μg/ml versus ≥64 μg/ml). Current data suggest that approximately two-thirds of the macrolide resistance observed among S. pneumoniae isolates in the United States is the result of expression of the mefA gene (16).

The present rate of macrolide resistance among S. pneumoniae isolates in the United States is roughly 18 to 22% (7, 20). Despite this appreciable level of resistance, concordance between in vitro susceptibility data and clinical outcomes among patients with respiratory tract infections is lacking (10). A hypothesis generated to explain these discordant observations is that although the macrolide MICs for pneumococci expressing the mefA gene are frequently interpreted as resistant according to approved breakpoints, the relatively low-level resistance that is expressed may be overcome by the by drug concentrations that are achievable following commonly used macrolide dosing regimens. The purpose of this study was to describe the activities of clarithromycin and azithromycin against pneumococci for which the study agents had a range of MICs and which expressed no, mefA, or ermB resistance determinants. Specifically, we desired to determine whether isolates expressing the mefA gene responded more like susceptible isolates or like isolates exhibiting high-level resistance (ermB mediated).

MATERIALS AND METHODS

Animals.

Six- to 8-week-old Swiss-Webster mice were obtained from Charles Rivers (Portage, Mich.). The mice used were virus free and weighed between 24 and 29 g. The mice were housed with three to four mice per cage and were allowed food and water ad libitum following University of Iowa guidelines on the use of laboratory animals. The mice were allowed to acclimate for 10 days prior to the initiation of experimental procedures. Study procedures were approved by the University of Iowa Animal Care and Use Committee.

Antibacterial agents.

Clarithromycin standard powder (Abbott Laboratories, Abbott Park, Ill.) was dissolved in ethanol and diluted to a final concentration of 30 mg/ml with sterile water. Azithromycin 200-mg/5-ml powder for oral use (Pfizer, Inc., New York, N.Y.) was obtained from McKesson Corp. (LaCrosse, Wis.). Azithromycin was prepared by adding 9 ml of sterile saline, in accordance with the manufacturer's instructions, to the commercially available 200-mg/5-ml powder, resulting in a final stock solution of 40 mg/ml. Macrolide stock solutions were diluted with sterile water immediately prior to use to yield the appropriate drug concentration in the final administration volume of 0.3 ml. Drug solutions were stored at 2 to 8°C for no more than 72 h.

Test isolates.

Nineteen clinical isolates of S. pneumoniae for which the macrolide MICs ranged from ≤0.03 to >64 μg/ml were obtained from the Department of Pathology, University of Iowa College of Medicine, and used for testing. Additionally, molecular characterization was performed with isolates for which the MICs were ≥0.5 μg/ml and in which the presence of the mefA and/or ermB gene(s) was noted (Abbott Laboratories) (17, 18). The organisms were stored at −70°C in a Microbank storage system (Pro-lab Diagnostics, Ontario, Canada). Each isolate was subcultured twice on blood agar plates (Remel, Lenexa, Kans.) prior to use.

Serotype determination.

Isolates were serotyped via the Quellung reaction, using specific pneumococcal antiserum (Statens Serum Institute, Copenhagen, Denmark). Briefly, a 0.5-McFarland standard suspension of each isolate was prepared in phosphate-buffered saline (pH 7.2). Equal volumes (3 μl) of methylene blue, antiserum, and the cell suspension were mixed and placed on a glass slide. Samples were then examined microscopically for binding of the pneumococcal capsular polysaccharide with the specific antibody contained in the antiserum.

In vitro susceptibility testing.

The MICs of clarithromycin and azithromycin were determined in ambient air against each of the study isolates according to broth microdilution techniques using Mueller-Hinton broth with 3% lysed horse blood (Becton Dickinson Microbiology Systems, Cockeysville, Md.) as described by the NCCLS (11).

Lung infection model.

Mice were rendered neutropenic with cyclophosphamide (150 mg/kg) administered by intraperitoneal injection 4 and 1 day prior to infection (1). The animals were infected by intratracheal inoculation with a 3-McFarland standard suspension of S. pneumoniae (1 × 10⁸ to 5 × 10⁸ CFU/ml) in sterile 0.9% sodium chloride and 50% sucrose. Prior to infection, the mice were anesthetized with halothane. Each animal was held vertically, and 50 μl (∼5 × 10⁶ CFU) of the bacterial suspension was instilled in the oropharynx. The nares of the animal were blocked, forcing inhalation of the suspension into the lungs. Immediately postinfection, the mice were briefly exposed to 100% oxygen to facilitate recovery. These procedures result in a starting inoculum of ∼1 × 10⁶ CFU/g of lung tissue at 12 h postinfection.

Pharmacokinetics studies.

Mice were rendered neutropenic as described above. Antibiotics were administered as a single dose via oral gavage. Clarithromycin and azithromycin single-dose pharmacokinetics were evaluated at doses of 4, 40, and 200 mg/kg. Plasma samples were collected from three mice at 0.5, 1, 1.5, 2, 3, 4, 5, 8, 12, and 24 h following the dose administration. Blood was collected in heparinized tubes and centrifuged at 5,000 × g for 10 min. The plasma was stored at −80°C until it was analyzed. Data from each time point were averaged, and concentration-versus-time profiles were constructed for each drug at each dosing level.

Drug analysis.

Samples were analyzed by a validated high-performance liquid chromatography method. Clarithromycin, azithromycin, and an internal standard were selectively removed from the plasma sample by liquid-liquid extraction. Briefly, the method combined a plasma aliquot (a sample or spiked standard) with internal standard, 0.5 M Na₂CO₃, and ethyl acetate-hexane (1:1 by volume). The samples were vortexed vigorously, followed by centrifugation. The organic layer was transferred and evaporated to dryness with a gentle stream of nitrogen at room temperature. The samples were reconstituted by vortexing them with the mobile phase. The parent drug and internal standard were separated from coextracted contaminants on a 50- by 3-mm 5-μm Kromasil C₁₈ column (Keystone) with an acetonitrile-trifluoroacetic acid (40:60 by volume) mobile phase at a flow rate of 0.25 to 0.3 ml/min with a 10- to 25-μl injection volume. The compounds of interest were quantified utilizing selective ion monitoring detection (m/z, 748.4 and 749.6 for clarithromycin and azithromycin, respectively), using a turbo ion spray source on an API 2000 mass spectrometer (PE Sciex). Triplicate standards, prepared in mouse plasma, at each of five separate concentrations were characterized by good accuracy (82 to 117% of theory) and reproducibility (percent coefficient of variation, 2.4 to 10.7%), with an estimated limit of quantitation of ∼3 ng/ml for both compounds. The plasma samples were analyzed as a single batch.

Treatment regimens.

Treatment was initiated 12 h postinfection and continued for 72 h. The regimens included the control (saline); clarithromycin at 4, 40, and 200 mg/kg twice daily; and azithromycin at 4, 40, and 200 mg/kg once daily. Additional regimens, clarithromycin at 200 mg/kg once a day and azithromycin at 40 and 200 mg/kg twice daily, were also subsequently tested. Ten mice were included in each treatment group. The animals were evaluated every 8 h for a total of 10 days postinfection. Mice surviving beyond 10 days were humanely euthanized with CO₂ gas.

Data analysis.

Pharmacokinetic parameters were calculated by noncompartmental methods using WinNonlin (Scientific Consulting Inc., Piscataway, N.J.). Median survival time was calculated using life tables (SPSS Inc., Chicago, Ill.). Pairwise comparisons of rates of survival between each group and the control were conducted using Gehan's generalized Wilcoxon test. Statistical significance was set at a P value of ≤0.05. The net change in the percentage of animals surviving 84 h after initiation of therapy (end of treatment) versus the control was graphed for all isolates (grouped by serotype) for the control, clarithromycin at 200 mg/kg twice a day, and azithromycin at 40 mg/kg once a day (the regimens determined to simulate human exposure). Additionally, end-of-treatment survival data were plotted for clarithromycin and azithromycin regimens simulating human exposure as a function of the MIC.

RESULTS

In vitro susceptibility.

The MICs for the 19 study isolates were determined (Table 1). According to current clarithromycin and azithromycin breakpoints, five isolates (R1, R11, R20, 65, and R101) were macrolide susceptible, nine isolates (R24, 134, 136, 710, 787, 963, 964, 1437, and 1633) exhibited low-level macrolide resistance (these isolates were found to possess the mefA gene), and five isolates exhibited high-level macrolide resistance. Of the isolates exhibiting high-level resistance, four (140, 144, 159, and 659) were ermB positive and one (430) possessed both the mefA and ermB genes. The clarithromycin and azithromycin MICs were ≤0.12 μg/ml for susceptible isolates, ranged from 0.5 to 32 μg/ml for mefA-positive isolates, and were ≥64 μg/ml for ermB- and mefA-ermB-positive isolates.

TABLE 1.

Test isolate serotypes, susceptibilities, and comparative drug activities versus control based on Kaplan-Meier analysis

Isolate	Serotype (penicillin MIC [μg/ml])a	Control % Survival/median survival timeb	Clarithromycin					Azithromycin
			MIC (μg/ml)	Survivalc				MIC (μg/kg)	Survivalc
				4 mg/kg BID	40 mg/kg BID	200 mg/kg BID	200 mg/kg QD		4 mg/kg QD	40 mg/kg QD	200 mg/kg QD	40 mg/kg BID
Susceptible
R1	23F (2.0)	52.4/99	0.06			∗	∗	0.12			∗	∗
R11	23F (0.015)	38.9/72	0.06		∗	∗		0.12		∗
R20	NT (0.008)	8.7/62	≤0.03		∗	∗		0.06		∗	∗	∗
65	NT (0.015)	0/50	≤0.03		∗	∗	∗	≤0.03		∗	∗	∗
R101	23F (0.015)	26.7/69	≤0.12		∗	∗		≤0.12	∗	∗	∗	∗
mefA positive
R24	6B (NA)	11.8/56	1		∗	∗		2
134	6A (0.25)	28.0/52	1	∗	∗	∗		2
136	19F (1.0)	16.7/47	1	∗	∗	∗		2		∗
710	9V (1.0)	0.0/51	8			∗	∗	32			∗
787	6B (0.06)	26.1/70	16		∗	∗		16
963	19F (4.0)	0/33	8			∗		8
964	19F (2.0)	16.7/70	0.5		∗	∗	∗	1
1437	14 (0.03)	10.5/54	16		∗	∗		16			∗
1633	19F (2.0)	22.7/72	0.5		∗	∗	∗	2		∗	∗	∗
ermB positive
140	6B (1.0)	0/52	>64					>64			∗
144	14 (2.0)	11/63	>64					>64
159	23F (2.0)	0/561	>64					>64
659	23F (1.0)	15.8/661	>64					>64
mefA and ermB positive
430	19F (2.0)	0.0/41	>64					>64

^aNT, nontypeable; NA, not available.

^bValues represent the percentage of mice surviving at 84 h following the initiation of therapy/median survival time (in hours).

^c∗, statistically significant improvement in survival of mice up to 84 h following initiation of therapy compared to control based on life table calculations (P < 0.05). BID, twice daily; QD, daily.

Pharmacokinetics.

Pharmacokinetic data for clarithromycin and azithromycin following each of the dosing regimens are presented in Table 2. Based on these data, it was determined that the 200-mg/kg dose of clarithromycin administered every 12 h produced a level of drug exposure in the mice that was similar to the human exposure following the administration of a 500-mg dose every 12 h (3). Likewise, azithromycin administered to mice as a single daily dose of 40 mg/kg closely approximated human exposure following ingestion of a 250-mg dose administered once daily (azithromycin package insert, Pfizer, Inc., 1996).

TABLE 2.

Observed clarithromycin and azithromycin pharmacokinetics in mice and targeted human parameters

Parametera	Value
	Clarithromycin				Azithromycin
	Murine			Human (500 mg) (4)	Murine			Human (250 mg) (1)
	4 mg/kgb	40 mg/kg	200 mg/kg		4 mg/kg	40 mg/kg	200 mg/kg
Cmax (μg/ml)	0.03 (0.01)	0.25 (0.11)	3.74 (1.68)	2-3 (0.9-1.35)	0.02 (0.02)	0.23 (0.21)	2.1 (1.89)	0.24 (0.20)
Cmin (μg/ml)	0c	0	0.26 (0.12)	0.42-1.5 (0.13-0.45)	0	0.01	0.39 (0.35)	0.05 (0.04)
Half-life (h)	0.5	3.0	13.3	4.0-14.5	22.5	11.3	16.0	>40
AUC0-τ (μg · h/ml)	0.03 (0.01)	0.37 (0.17)	16.83 (7.57)	19-30 (5.7-9)	0.15 (0.14)	0.82 (0.74)	15.94 (14.35)	2.1 (1.72)

^aC_max, maximum concentration of drug in serum; C_min, minimum concentration of drug in serum; AUC_0-τ, area under the concentration time curve from time zero to the end of the dosing interval. Drug concentrations are expressed as total (calculated free). The protein binding values used to determine the free-drug concentrations were as follows: clarithromycin, 55% bound in mice (personal communication from C. Olsen, [Abbott Laboratories]), 70% bound in humans (4); azithromycin, 10% bound in mice (1), 18% bound in humans (1).

^bDose.

^c0, value that is below the limit of assay detection.

Survival data.

Survival data are summarized and presented in Table 1. Statistically significant improvement in survival compared to the control was noted among mice infected with macrolide-susceptible isolates and treated with clarithromycin at doses of 40 (four of five) and 200 (five of five) mg/kg twice daily and 200 mg/kg once daily (two of five). Similar improvements were noted among mice treated with azithromycin at doses of 40 (four of five) and 200 (four of five) mg/kg once daily and 40 mg/kg twice daily (four of five). Among animals infected with isolates exhibiting low-level macrolide resistance, treatment with clarithromycin at 40 and 200 mg/kg twice daily resulted in improved rates of survival for seven of nine and nine of nine isolates, respectively. Improved rates of survival were noted among azithromycin-treated mice for two of nine, four of nine, and one of nine isolates with 40 and 200 mg/kg once daily and 40 mg/kg twice daily, respectively. Treatment-limiting toxicity was observed following the administration of azithromycin at 200 mg/kg twice daily. As a result, use of this regimen was discontinued and the data gathered were excluded from analysis. Kaplan-Meier curves generated with a representative isolate exhibiting low-level macrolide resistance are displayed in Fig. 1. Survival at the end of treatment following the administration of regimens simulating human exposure plotted as a function of the MIC are presented in Fig. 2. Among clarithromycin-treated mice, the survival differences between mice infected with susceptible or mefA-containing isolates and controls ranged between 21 and 74% and between 20 and 73%, respectively. In contrast, for azithromycin-treated animals, the survival differences between mice infected with susceptible or mefA-containing isolates and controls ranged from 18 to 81% and from −17 to 57%, respectively. The difference in survival between azithromycin-treated mice and controls infected with ermB-containing isolates ranged between −16 and 25%.

FIG. 1.

Representative Kaplan-Meier curves for a mefA-positive isolate of S. pneumoniae (S. pneumoniae 963). (A) •, control; ▿, clarithromycin at 4 mg/kg twice a day; ▪, clarithromycin at 40 mg/kg twice a day; ⋄, clarithromycin at 200 mg/kg twice a day; ▴, clarithromycin at 200 mg/kg once a day. (B) •, control; ▿, azithromycin at 4 mg/kg once a day; ▪, azithromycin at 40 mg/kg once a day; ⋄, azithromycin at 200 mg/kg once a day; ▴, azithromycin at 40 mg/kg twice a day.

FIG. 2.

Correlation between MIC and end-of-treatment survival with simulated human exposures of clarithromycin (A) and azithromycin (B) for mice infected with test isolates.

Neither clarithromycin nor azithromycin administration at any dosing level appreciably affected the survival of mice infected with isolates exhibiting high-level macrolide resistance (Table 1 and Fig. 3). A statistically significant difference was noted between controls and mice infected with isolate 140 that were treated with azithromycin at 200 mg/kg once daily; however, the median difference in survival was only 10 h (53 versus 63 h).

FIG. 3.

Representative Kaplan-Meier curves for an ermB-positive isolate of S. pneumoniae (S. pneumoniae 430). (A) •, control; ▿, clarithromycin at 4 mg/kg twice a day; ▪, clarithromycin at 40 mg/kg twice a day; ⋄, clarithromycin at 200 mg/kg twice a day; ▴, clarithromycin at 200 mg/kg once a day. (B) •, control; ▿, azithromycin at 4 mg/kg once a day; ▪, azithromycin at 40 mg/kg once a day; ⋄, azithromycin at 200 mg/kg once a day; ▴, azithromycin at 40 mg/kg twice a day.

Serotypes.

Of the 19 S. pneumoniae isolates used in this study, two were nontypeable. The remaining 17 isolates belonged to one of six serotypes, 23F, 6B, 6A, 14, 19F, and 9V (Table 1). Among isolates expressing low-level, mefA-mediated resistance, the worst end-of-treatment survival, regardless of treatment regimen, was noted for isolates 710 and 1433, which belonged to serotypes 9V and 14, respectively (Fig. 4). The end-of-treatment survival rates for these isolates were ∼0 to 10% for controls, 0 to 20% for the azithromycin-treated (40-mg/kg daily) group, and 30 to 40% for the clarithromycin-treated (200-mg/kg twice-daily) group. In contrast, survival at the end of treatment among isolates of serotypes 6B, 6A, and 19F ranged from 10 to 50% among controls, from 10 to 70% for the azithromycin-treated (40-mg/kg daily) group, and from 60 to 100% for the clarithromycin-treated (200-mg/kg twice-daily) group.

FIG. 4.

Net change in percent survival at the end of treatment versus control for mice infected with susceptible (A), mefA-positive (B), and ermB-positive (C) isolates grouped by serotype. Shaded bars, clarithromycin at 200 mg/kg twice a day; open bars, azithromycin at 40 mg/kg once a day.

DISCUSSION

Discordance between the S. pneumoniae susceptibility profile and the clinical outcome of respiratory tract infections has been well documented (2, 10, 12). Data such as these have prompted clinicians and professional organizations to question the relevance of current pneumococcal breakpoints in the management of respiratory tract infections (8). With respect to the macrolides, intrapulmonary pharmacokinetic data suggest that concentrations of clarithromycin and perhaps azithromycin in epithelial lining fluid and alveolar macrophages exceed the MICs for some pneumococcal isolates currently categorized as resistant (4, 5, 13, 14). These data, along with the discordant findings of clinical data, prompted us to evaluate the efficacies of clarithromycin and azithromycin against isolates of S. pneumoniae for which they exhibited a range of MICs in a lung infection model.

Upon examining the human intrapulmonary pharmacokinetic profile of clarithromycin, one would predict that a good microbiologic response would be noted with isolates exhibiting low-level macrolide resistance (for which the clarithromycin MICs were 1 to 16 μg/ml). Although quantitative data regarding the intrapulmonary accumulation of clarithromycin in mice has not been published, Kohno and colleagues have reported qualitative data demonstrating a relative accumulation of clarithromycin in murine lungs (9). In the present study, we did note a response to clarithromycin among mice infected with mefA-containing isolates. This finding may be explained by accumulation of clarithromycin in the lungs. Slightly surprising was the fact that similar improvement in survival was not evident among mice infected with these low-level-resistant isolates following treatment with azithromycin. However, inspection of the published human pharmacokinetic data reveals that intrapulmonary, epithelial-lining fluid, and alveolar-macrophage clarithromycin concentrations are routinely 10-fold higher than those noted with azithromycin (4, 13, 14). Rodvold and colleagues reported that, following routine doses of clarithromycin and azithromycin, the peak concentrations of these agents in the epithelial lining fluid were ∼34.5 and 2.2 μg/ml, respectively. Although corresponding intrapulmonary pharmacokinetic data are not available for azithromycin in mice, it may be reasonable to hypothesize that differences in the absolute concentrations of azithromycin and clarithromycin in murine lungs follow a pattern similar to that noted in humans. Therefore, one might infer that the current breakpoints for clarithromycin may underestimate the activity of the drug in the management of respiratory tract infections. It is important to note that we did utilize a neutropenic infection model. Since it is widely accepted that clarithromycin and azithromycin both accumulate in polymorphonuclear cells, it is important to consider the relative effect of neutropenia on the pharmacokinetics of these agents. The absence of polymorphonuclear cells would likely negatively affect the pulmonary concentrations of both agents; however, the impact on azithromycin concentrations may be more evident.

The primary purpose of this study was to evaluate the efficacies of clarithromycin and azithromycin, at doses in accordance with regimens that provided exposure similar to that noted in humans following approved dosing regimens, against S. pneumoniae isolates exhibiting no, low-level, and high-level macrolide resistances. To achieve our target parameters (Table 2) and to ensure adequate exposure, we used three dosing levels (4, 40, and 200 mg/kg) for each drug. Based on pharmacokinetic data, it was determined that twice-daily dosing of clarithromycin and once-daily dosing of azithromycin most closely approximated human exposure. Although the overall exposure resulting from these regimens achieved our targets, we were concerned that the elimination of azithromycin might have been more rapid than desired. Therefore, we elected to also test azithromycin at 40 and 200 mg/kg administered twice daily in an effort to prolong the duration of drug exposure. Despite doubling the azithromycin daily dose, no appreciable difference with respect to survival was noted between the 40-mg/kg once- and twice-daily regimens. Additionally, it was noted that the 200-mg/kg twice-daily regimen resulted in excessive mortality.

Using this model of infection, 90 to 100% of the control mice died within 5 days following infection. However, it was noted that the rates at which specific isolates resulted in death varied considerably. This isolate-specific variability in the rate of survival was most likely related to the virulence of the isolates. We determined the serotype of each isolate and used this to allow strains to be grouped by capsular-antigen similarities. Although the virulence of isolates did not alter the ranked efficacies of the antimicrobials with respect to the survival benefit among animals infected with mefA-positive strains, the median end-of-treatment survival rates did vary considerably among treated mice. Specifically, examining the mefA-positive isolates, it was noted that serotypes 9V and 14 were associated with the most rapid mortality. Accounting for organism virulence and rates of killing may significantly impact attempts to correlate MICs with efficacies. More virulent isolates produce greater mortality even among treated animals regardless of the MIC. Therefore, strain virulence needs to be considered prior to selecting test organisms, treatment duration, and observation periods. Failure to do so may hinder the ability to adequately assess the impact of the MIC on treatment success. However, antimicrobials may lessen the virulence of an isolate via a subinhibitory effect on virulence or by immunodulation. Therefore, evaluation of isolates with various degrees of virulence may help to provide a more realistic assessment of the overall impact of drug therapy on outcome.

The clinical presentation and histology of pneumonia is quite different from the infection induced using this lung infection model. Because of the acute nature of this experimental infection, many of the histological changes noted in pneumonia are absent. Additionally, following our procedures, 100% of the mice become bacteremic within 72 h following infection. Therefore, it may be argued that the mortality resulting from these procedures is secondary to bacteremia rather than pneumonia. This experimental nuance is important to discuss because of pharmacokinetic differences between clarithromycin and azithromycin. Unlike clarithromycin, the concentrations of azithromycin in serum decline to subtherapeutic levels quickly following dose administration. As a result, one might suspect that azithromycin is minimally effective against a bacteremic infection in a neutropenic model. Although this may contribute to the poor activity noted for azithromycin in this model, it should be pointed out that both agents were equally effective against macrolide-susceptible pneumococcal isolates. Therefore, enough azithromycin was present in either the serum or lung tissue to kill the susceptible microbes.

Efflux-mediated macrolide resistance among S. pneumoniae isolates results in the expression of a relatively low level of resistance. Additionally, since this mechanism of resistance results in a reduction in the intracellular accumulation of the drug rather than an alteration of the target site, it has been theorized that the resistance may be overcome if enough drug is able to reach the target site. In this study, we demonstrated that following simulated human exposure to clarithromycin, animals infected with pneumococci possessing low-level macrolide resistance responded to therapy in a manner similar to that of animals infected with fully susceptible isolates. These data suggest that further studies are warranted to better define the relationships among drug susceptibility, resistance determinants, and outcome for clarithromycin against S. pneumoniae exhibiting mefA-mediated resistance.