In Vitro Activity of Human-Simulated Epithelial Lining Fluid Exposures of Ceftaroline, Ceftriaxone, and Vancomycin against Methicillin-Susceptible and -Resistant Staphylococcus aureus

Abstract

Staphylococcus aureus, including methicillin-susceptible (MSSA) and -resistant (MRSA) strains, is an important pathogen of bacterial pneumonia. As antibiotic concentrations at the site of infection are responsible for killing, we investigated the activity of human-simulated epithelial lining fluid (ELF) exposures of three antibiotics (ceftaroline, ceftriaxone, and vancomycin) commonly used for treatment of S. aureus pneumonia. An in vitro pharmacodynamic model was used to simulate ELF exposures of vancomycin (1 g every 12 h [q12h]), ceftaroline (600 mg q12h and q8h), and ceftriaxone (2 g q24h and q12h). Four S. aureus isolates (2 MSSA and 2 MRSA) were evaluated over 72 h with a starting inoculum of ∼10⁶ CFU/ml. Time-kill curves were constructed, and microbiological response (change in log₁₀ CFU/ml from 0 h and the area under the bacterial killing and regrowth curve [AUBC]) was assessed in duplicate. The change in 72-h log₁₀ CFU/ml was largest for ceftaroline q8h (reductions of >3 log₁₀ CFU/ml against all strains). This regimen also achieved the lowest AUBC against all organisms (P < 0.05). Vancomycin produced reliable bacterial reductions of 0.9 to 3.3 log₁₀ CFU/ml, while the activity of ceftaroline q12h was more variable (reductions of 0.2 to 2.3 log₁₀ CFU/ml against 3 of 4 strains). Both regimens of ceftriaxone were poorly active against MSSA tested (0.1 reduction to a 1.8-log₁₀ CFU/ml increase). Against these S. aureus isolates, ELF exposures of ceftaroline 600 mg q8h exhibited improved antibacterial activity compared with ceftaroline 600 mg q12h and vancomycin, and therefore, this q8h regimen deserves further evaluation for the treatment of bacterial pneumonia. These data also suggest that ceftriaxone should be avoided for S. aureus pneumonia.

INTRODUCTION

Pneumonia remains a common cause of morbidity and mortality, particularly among patients who require hospitalization (1). Staphylococcus aureus has emerged as an important causative pathogen of both community-acquired (CA) and hospital-acquired (HA) bacterial pneumonia (2). Historically, β-lactams have been deemed first-line empirical therapy due to their safety and spectrum of activity against common etiologies of pneumonia, including S. aureus (3, 4). Largely due to a convenient administration schedule, ceftriaxone is one of the most commonly used empirical agents for community-acquired pneumonia in the inpatient setting. However, when methicillin-resistant S. aureus (MRSA) is of concern in these patients, vancomycin is often the empirical choice (with linezolid offered as an alternative first-line agent in the Infectious Diseases Society of America [IDSA] MRSA guidelines) (5).

Ceftaroline (CPT) is a novel cephalosporin with potent in vitro anti-MRSA activity (6). Ceftaroline fosamil, the prodrug form of ceftaroline, is approved for the treatment of community-acquired bacterial pneumonia (CABP) caused by methicillin-susceptible S. aureus (MSSA); however, the potential utility in the setting of MRSA pneumonia remains unknown. Several in vitro models of infection have demonstrated reliable bactericidal activity of human-simulated plasma concentrations of ceftaroline against MRSA isolates (7, 8, 9). Furthermore, ceftaroline therapy reduced the bacterial burden of MRSA in murine and rabbit pneumonia models compared with that in control animals (10, 11). However, it is the achievable concentration in epithelial lining fluid (ELF) at the bronchopulmonary segment that is responsible for activity against S. aureus in the treatment of pneumonia (12). No pharmacodynamic (PD) studies (in vitro or in vivo) have been conducted to compare the efficacy of ceftaroline against other antistaphylococcal agents at humanized pharmacokinetic exposures observed in the ELF.

Until recently, ELF penetration and achievable exposure of ceftaroline and vancomycin were not available (13, 14). These data are particularly important for most β-lactams and glycopeptides (vancomycin), as these antibiotic classes generally have ELF-to-plasma concentration ratios of ≤1 (12). In this study, we investigated the activity of human-simulated ELF exposures of ceftaroline versus ceftriaxone against MSSA and ceftaroline versus vancomycin against MSSA and MRSA.

MATERIALS AND METHODS

Bacterial strains and susceptibility testing.

Four S. aureus isolates were studied (1 community-acquired MSSA, 1 community-acquired MRSA [USA300], 1 nosocomial-acquired MSSA, and 1 nosocomial-acquired MRSA). Isolates were selected from the organism bank at the Center for Anti-Infective Research and Development (Hartford, CT), based on the goal of testing a range of antibiotic MICs. Ceftaroline, ceftriaxone, and vancomycin MICs were determined by broth microdilution in accordance with Clinical and Laboratory Standards Institute (CLSI) recommendations (15). MICs were performed in triplicate, and the modal MIC was reported.

Antibiotic test agents.

Ceftriaxone (Sandoz Inc.; lot DP9493) and vancomycin (Hospira Inc.; lot 27400DD) powders for injection were obtained from the Department of Pharmacy at Hartford Hospital, Hartford, CT, USA. Analytical-grade ceftaroline (lots 01FOR01-03-29 and 01FOR01-04-29), the active component of the prodrug ceftaroline fosamil, was provided by Forest Laboratories, Inc. (Jersey City, NJ, USA).

Simulated drug exposures.

For all agents, we assumed the observed concentration in the ELF was that of free drug only (i.e., no further correction for protein binding was done for ELF). ELF drug exposures of ceftaroline 600 mg every 12 h (q12h) and q8h were simulated from median concentrations obtained from bronchoalveolar lavage (BAL) samples of 25 healthy subjects (13). For vancomycin, population pharmacokinetic estimates obtained from a BAL study of 10 healthy volunteers were employed to simulate the ELF concentration profile of a dosing regimen of 1 g q12h (14). The ELF penetration for ceftriaxone has not been reported; therefore, 100% penetration was assumed as a best-case scenario for a β-lactam. This exposure was simulated based on 100% penetration of the free-drug (90% protein binding) plasma concentration using a previous population pharmacokinetic study of critically ill patients (16). Steady-state concentration profiles for vancomycin 1 g q12h and ceftriaxone 2 g q12h and q24h were simulated in ADAPT 5 (17) using the median population pharmacokinetic parameters derived from the respective studies for similar comparisons to median concentrations of ceftaroline. The target peak, trough, and half-life (t_1/2) values for each regimen are shown in Table 1, and the target free area under the curve-to-MIC ratio (fAUC/MIC) or free time above the MIC (fT>MIC) exposure is shown in Table 2.

TABLE 1.

Target and observed pharmacokinetic ELF exposures of all isolates against each regimen

Regimen	ELF exposure (mg/liter)				t1/2 (h)
	Peak		Trough
	Target	Observed	Target	Observed	Target	Observed
Ceftriaxone 2 g q24ha	8	8.2 ± 0.5	0.6	0.7 ± 0.1	6.3	6.6 ± 0.2
Ceftriaxone 2 g q12ha	11	10.2 ± 0.7	3.1	3.1 ± 0.3	6.1	6.5 ± 0.3
Ceftaroline 600 mg q12hb	3.4	3.7 ± 0.4	0.0	0.0	1.3	1.4 ± 0.1
Ceftaroline 600 mg q8hb	3.6	3.7 ± 0.8	0.09	0.09 ± 0.04	1.3	1.3 ± 0.1
Vancomycin 1 g q12hc	29.2	30.6 ± 2.9	3.3	4.1 ± 0.5	5.3	5.9 ± 0.4

^aCeftriaxone concentrations in ELF have not been determined. As such, it was assumed that penetration into ELF was 100% and that ceftriaxone ELF exposure was synonymous with the plasma concentration profile as described in reference 16.

^bCeftaroline ELF exposure derived from reference 13.

^cVancomycin ELF exposure derived from reference 14.

TABLE 2.

Organism MIC and targeted and observed pharmacodynamic ELF exposures of each regimen against individual isolates^a

Regimen	MRSA 412			MRSA 146			MSSA F2-21			MSSA F1-9
	MIC (mg/liter)	ELF exposure		MIC (mg/liter)	ELF exposure		MIC (mg/liter)	ELF exposure		MIC (mg/liter)	ELF exposure
		Target	Observed		Target	Observed		Target	Observed		Target	Observed
Ceftriaxone 2 g q24h	ND	ND	ND	ND	ND	ND	2	55.8	59.6	4	28.5	30.4
Ceftriaxone 2 g q12h	ND	ND	ND	ND	ND	ND	2	100	100	4	79.2	80.3
Ceftaroline 600 mg q12h	1	25.3	28.0	0.5	37.6	41.9	0.25	49.3	51.5	0.25	49.3	50.2
Ceftaroline 600 mg q8h	1	40.1	40.0	0.5	58.6	59.6	0.25	76.1	77.4	0.25	76.1	74.5
Vancomycin 1 g q12h	2	54.2	63.3	1	108.4	120.1	2	54.2	63.2	1	108.4	121.2

^aELF exposures are shown as %ƒT>MIC for ceftriaxone and ceftaroline and the fAUC/MIC of the dosing interval for vancomycin. ND, not determined.

In vitro pharmacodynamic model.

A one-compartment in vitro chemostat model was used for all experiments, as previously described (18, 19). Briefly, each experiment consisted of two treatment models and one growth control model running simultaneously for each isolate and treatment regimen. The models were placed in a 37°C water bath for optimal temperature control. Magnetic stir bars were utilized to ensure adequate mixing of the contents of each model. Experiments were performed with a starting (0-h) inoculum of ∼10⁶ CFU/ml. Inoculum preparation included making a bacterial suspension of ∼10⁸ CFU/ml from two subsequent subcultures of the test isolate (incubated for 18 to 24 h overnight at 37°C) with 0.9% normal saline for injection. Each model was filled with cation-adjusted Mueller-Hinton Broth (CAMHB) (Becton, Dickinson and Company, Sparks, MD) and inoculated. Bacteria were allowed to enter into log growth phase for 30 min, and then antibiotics were administered at simulated exposures.

Antibiotic administration was initiated 0.5 h after inoculation, corresponding to 0 h of the experimental phase. For antibiotic regimens with steady-state trough concentrations (all regimens except ceftaroline 600 mg q12h), a bolus dose was administered into the model at 0 h to achieve the desired trough. Following the initial bolus, each antibiotic was infused with drug-containing broth to establish a peak concentration over a 1-h period to replicate distribution into the ELF as observed in humans. After the 1-h infusion, fresh drug-free broth was supplied via a peristaltic pump (Masterflex L/S model 7524-40; Cole-Palmer Instrument Company, USA) which was set to achieve the human simulated half-life in ELF of the antibiotic being tested. For vancomycin, from 1 h to 3 h, broth was infused at a rate that simulated the rapid alpha elimination of vancomycin, followed by a reduced terminal elimination rate from 3 h to 12 h to achieve target AUC exposure. Each experiment was conducted over 72 h and performed at a minimum of two replicates to ensure reproducibility.

Samples were obtained from each of the models at various time points (16 to 24 samples per model depending on the antibiotic regimen) throughout the 72-h experiment and serially diluted in normal saline to assess changes in bacterial density over time. Aliquots of each diluted sample were plated on Trypticase soy agar plates with 5% sheep blood and incubated at 37°C for 18 to 24 h for quantitative culture. Time-kill curves were constructed by plotting log₁₀ CFU/ml against time. The area under the bacterial killing and regrowth curve (AUBC) was calculated for each model, as a measure of the antibacterial activity over the 72-h study duration. Differences in AUBC and changes in log₁₀ CFU/ml after 72 h were compared by analysis of variance (ANOVA) with the Student-Newman-Keuls method for multiple comparisons using SigmaPlot 12.0 (Systat Software Inc., San Jose, CA, USA). An a priori P value of <0.05 was considered statistically significant. The lower limit of detection for bacterial density was 1.7 log₁₀ CFU/ml.

Antibiotic concentration determination.

Broth samples were simultaneously obtained with bacterial density and assayed for ceftaroline, ceftriaxone, and vancomycin concentrations to ensure target exposures were achieved. Peak, trough, and a minimum of three time points for each dosing interval were assayed for the determination of each antibiotic's half-life. All pharmacokinetic samples were immediately frozen and stored at −80°C until analyses. Ceftriaxone and vancomycin concentrations were determined by a high-performance liquid chromatography (HPLC) method at the Center for Anti-Infective Research and Development (Hartford, CT, USA) (20, 21). The vancomycin and ceftriaxone assays were linear over a range of 2 mg/liter to 40 mg/liter (r² ≥ 0.998) and 1 mg/liter to 40 mg/liter (r² ≥ 0.999), respectively. The mean interday coefficient of variation (%CV) for high (40-mg/liter)- and low (2-mg/liter)-check samples were 0.57% and 6.09% for vancomycin, respectively. The mean intraday coefficients of variation were 5.94% and 6.13%, respectively. For ceftriaxone, the mean interday coefficients of variation for high- and low-check samples were 5.54% and 5.09%. The mean intraday coefficients of variation for ceftriaxone were 3.79% and 4.38%, respectively. Ceftaroline samples were analyzed by Keystone Bioanalytical, Inc. (North Wales, PA, USA) using a validated liquid chromatography-tandem mass spectrometry (LC-MS/MS) assay. A quadratic regression (weighted 1/concentration²) gave the best fit for the calibration curve over a concentration range of 50 to 50,000 ng/ml for ceftaroline (r² ≥ 0.999). The precision values (%CV) for high- and low-check samples were within 3%.

Resistance.

Screening for emergence of subpopulations with reduced susceptibilities was assessed throughout the 72-h experiment. The presence of resistant subpopulations was identified by plating the direct bacterial suspension from each model, including control models, on antibiotic-containing plates at drug concentrations 3 times the baseline MIC at 24, 48, and 72 h and incubating them for 48 h at 37°C. Any growth observed on the antibiotic-containing plates after incubation was considered emergence of resistance.

RESULTS

Pharmacokinetic analysis.

The observed pharmacokinetic parameters and associated pharmacodynamic exposures during all experiments are presented in Table 1 and Table 2, respectively. Targeted concentrations and exposures were achieved in all models.

Antibacterial results.

The average bacterial densities of the starting inoculum for MRSA 412, MRSA 146, MSSA F2-21, and MSSA F1-9 isolates were 6.4 ± 0.1, 6.4 ± 0.2, 6.3 ± 0.1, and 6.4 ± 0.1 log₁₀ CFU/ml, respectively. Control models grew to 8.8 ± 0.3, 9.1 ± 0.1, 8.9 ± 0.1, and 9.3 ± 0.0 log₁₀ CFU/ml for MRSA 412, MRSA 146, MSSA F2-21, and MSSA F1-9 isolates, respectively. Growth control models that ran with each experiment were similar throughout the study, regardless of pump flow rate. Time-kill curves for all treatment regimens against each isolate are shown in Fig. 1A to D. The mean changes in bacterial density after 72 h and AUBCs over 72 h of the tested isolates are provided in Table 3.

FIG 1.

Mean bacterial densities over 72 h by isolate. (A) MRSA 412; (B) MRSA 146; (C) MSSA F1-9; (D) MSSA F2-21. Closed diamonds, growth control; open circles, ceftaroline 600 mg q12h; closed circles, ceftaroline 600 mg q8h; open triangles, vancomycin 1 g q12h; closed squares, ceftriaxone 2 g q24h; open squares, ceftriaxone 2 g q12h.

TABLE 3.

Corresponding summary activity of test drug regimens

Isolate	Endpoint	Value by antibiotic regimenb
		CPT q12h	CPT q8h	CRO q24h	CRO q12h	VAN q12h
MRSA 412	Δlog10 CFU/ml at 72 h	0.2 ± 0.6	−4.1 ± 0.8	ND	ND	−2.0 ± 0.2
	AUBCc	313 ± 26	176 ± 6	ND	ND	220 ± 19
MRSA 146a	Δlog10 CFU/ml at 72 h	−2.3 ± 1.4	−4.3 ± 0.4	ND	ND	−3.3 ± 1.6
	AUBC	232 ± 47	154 ± 1	ND	ND	222 ± 43
MSSA F2-21	Δlog10 CFU/ml at 72 h	−0.2 ± 0.7	−3.5 ± 1.6	1.7 ± 0.3	−0.1 ± 0.0	−1.9 ± 0.5
	AUBC	268 ± 44	172 ± 28	476 ± 2	308 ± 13	225 ± 31
MSSA F1-9	Δlog10 CFU/ml at 72 h	−1.9 ± 2.8	−4.3 ± 0.3	1.8 ± 0.3	0.1 ± 0.2	−0.9 ± 0.3
	AUBC	240 ± 78	164 ± 4	472 ± 27	406 ± 7	328 ± 38

^aUSA300.

^bValues for each isolate are the means from two to four independent time-kill experiments. Data are presented as means + SD of all time-kill curves (two to four per isolate). Negative numbers indicate reductions in CFU from 0 h. ND, not determined; CPT, ceftaroline; CRO, ceftriaxone; VAN, vancomycin. AUBC statistics for MRSA: CPT q8h > CPT q12h = VAN. AUBC statistics for MSSA: CPT q8h > CPT q12h = VAN > CRO q12h > CRO q24h. Δlog₁₀ CFU/ml at 72 h statistics for MRSA: CPT q8h > CPT q12h = VAN. Δlog₁₀ CFU/ml at 72 h statistics for MSSA: CPT q8h superior to all other regimens. CPT q12h and VAN superior to CRO q24h. No differences between (P > 0.05) all other regimens were noted against MSSA.

^cAUBC, area under the bacterial killing and regrowth curve (expressed in log₁₀ CFU · h/ml over 72 h).

For MRSA 412, the largest reduction in CFU at 72 h was observed with ceftaroline q8h, then vancomycin, and then ceftaroline q12h (P < 0.05). There was no difference in CFU at 72 h for MRSA 146 among the three regimens. Against MSSA F2-21, ceftaroline 600 mg q8h displayed significantly larger reductions in 72-h CFU counts than all regimens except vancomycin (P = 0.067). Vancomycin 1 g q12h was numerically superior to ceftriaxone 2 g q24h (P = 0.009) for this isolate; no other statistical differences were observed. There was no difference in CFU at 72 h for MSSA F1-9 among the five regimens.

When data from all isolates were combined, ceftaroline 600 mg every 8 h achieved the lowest AUBC (P < 0.05 against all regimens) and the greatest reduction in log₁₀ CFU/ml after 72 h (>3-log reductions). Vancomycin 1 g q12h and ceftaroline 600 mg q12h displayed rapid initial killing against both MRSA and MSSA isolates achieving ≥3-log reductions within the first 24 h. However, by 72 h, regrowth was observed against both regimens, unlike the ceftaroline 600 mg q8h regimen, which maintained >3 log₁₀ CFU of kill for the 72-h duration. Regrowth was most notable for ceftaroline 600 mg q12h, occurring in the final portion of each dosing interval throughout all experiments. As a result, minimal activity at 72 h was observed for ceftaroline 600 mg q12h against 2 of the 4 isolates, MRSA 412 (CPT MIC = 1 mg/liter) and MSSA F2-21 (CPT MIC = 0.25 mg/liter). Notably, ceftriaxone regimens, particularly 2 g q24h, demonstrated the least activity against the MSSA isolates, although due to some variability among ceftaroline q12h at the 72-h sampling, statistical significance was observed only for the AUBC endpoint (Table 3).

No growth was observed on any of the drug-containing plates for any of the isolates, signifying there was no emergence of resistance observed during these experiments.

DISCUSSION

Staphylococcus aureus has increasingly become a significant pathogen of bacterial pneumonia over the past 20 years (2, 22). Methicillin resistance presents substantial challenges to the diagnosis and therapeutic management of S. aureus, for which vancomycin is recommended as a first-line treatment option (4, 5). However, cure rates with vancomycin for S. aureus pneumonia (MRSA and MSSA) have been less than desirable, with mortality rates reported as high as 50% (2, 23, 24). Although the poor performance of vancomycin in pneumonia is likely multifactorial, a contributing factor may be the relatively lower exposures achieved in the ELF than in plasma (14). Given the implications of therapeutic concentrations at the site of infection on outcome, combined with the absence of clinical data with S. aureus pneumonia, in vitro pharmacodynamic studies can play a pivotal role in bridging the gap between antimicrobial activity and clinical utility (25). In the current study, including two MSSA and two MRSA strains, ELF exposures of ceftaroline 600 mg every 8 h were bactericidal and achieved significantly greater kill than other tested regimens. ELF exposures to vancomycin 1 g every 12 h produced reliable reductions (0.9 to 3.3 log₁₀ CFU/ml) after 72 h against MRSA and MSSA isolates, while the activity of ceftaroline 600 mg every 12 h was more variable, with 0.2- to 2.3-log reductions at 72 h against 3 of 4 strains. Most surprising, ceftriaxone proved to be inferior against MSSA, with minimal activity against susceptible isolates even when dosed aggressively at 2 g every 12 h and assuming 100% penetration into ELF.

While vancomycin resistance (MICs ≥ 16 μg/ml) is rare among S. aureus strains, poor outcomes and high rates of mortality have been reported when using vancomycin for treatment of MRSA pneumonia against isolates with vancomycin MICs at the higher end of the susceptible range (1.5 to 2 μg/ml) (26). Based on this report and others in bacteremia, a consensus guideline from the IDSA, Society of Infectious Diseases Pharmacists, and American Society of Health-Systems Pharmacists suggests that alternative treatment options should be carefully considered when the vancomycin MIC exceeds 1 μg/ml (27). This is because for vancomycin, a total AUC/MIC ratio of approximately 400, or fAUC/MIC of 200 assuming 50% protein binding, is considered the pharmacodynamic driver of efficacy and successful patient outcomes, an endpoint that is challenging to attain with conventional vancomycin dosing against isolates with MICs of >1 mg/liter (28). As a result, we anticipated vancomycin would produce reliable activity against MRSA 146 and MSSA F1-9 with MICs of 1 μg/ml (mean ELF fAUC/MIC_0–24 ≅ 240) and poor activity against MRSA 412 and MSSA F2-21 with MICs of 2 μg/ml (mean ELF fAUC/MIC_0–24 ≅ 125). In contrast, human-simulated ELF exposures of vancomycin 1 g every 12 h in this experiment effectively reduced the bacterial burden over 72 h against all four S. aureus isolates (Table 3). It should be noted that reductions in vancomycin efficacy against strains with MICs of 2 μg/ml have been reported primarily in patients with severe and/or complicated infections, which may have contributed to poor clinical outcomes. Similar to our findings, previous in vitro pharmacodynamic studies of vancomycin have observed appreciable reductions in bacterial density against MRSA with vancomycin MICs of 2 μg/ml, despite suboptimal AUC/MIC exposures (9, 19, 29, 30). This highlights the need for additional studies to better understand the factors and mechanisms determining outcome of S. aureus infection. Notably, vancomycin ELF exposure is approximately 100% of unbound plasma concentrations according to the data produced in the healthy-volunteer study, and thus the achievable exposure at the infection site is quite similar to free-drug exposure in plasma, which may explain the activity observed (14).

Ceftaroline fosamil is the first approved beta-lactam with activity against MRSA (6). Large international surveillance programs have demonstrated potent in vitro activity of ceftaroline against S. aureus, with MIC₅₀ and MIC₉₀ values of 0.5 and 1 μg/ml, respectively (31, 32). However, as very few patients in the CABP clinical trials for ceftaroline had MRSA isolated, little is known concerning the antibacterial activity of ceftaroline in patients with MRSA pneumonia (33, 34). The clinical utility of ceftaroline fosamil compared to vancomycin and ceftriaxone in MRSA pneumonia was being explored in a phase IV clinical study; however, the study was closed prematurely due to slow enrollment (35). Instead, clinical data supporting the use of ceftaroline fosamil for the treatment of MRSA pneumonia have to date been limited to retrospective case series, primarily in the setting of salvage therapy due to elevated vancomycin MICs or prior anti-MRSA therapy failure (36, 37, 38). The largest collection of patient data on the use of ceftaroline for MRSA pneumonia comes from the Clinical Assessment Program and Teflaro Utilization Registry (CAPTURE), a multicenter cohort study of 398 patients with CABP. Of the 64 patients with confirmed MRSA CABP, 42 (66%) patients were determined to be clinical successes (37). In another multicenter, retrospective observational cohort of patients treated with ceftaroline for lower respiratory tract infections, Casapao and colleagues observed clinical success rates of 91% and 69% for S. aureus CABP and hospital-acquired pneumonia (HAP), respectively (39). Of note, the dosage regimen employed in these studies was not provided. Similar success rates were observed by Pasquale and colleagues with ceftaroline 600 mg q12h in the treatment of 10 patients with MRSA nosocomial pneumonia, with vancomycin MICs ranging from 0.75 to 2 μg/ml (36). Clinical success (improvement or cure) was achieved in six patients, while three patients expired due to a noninfectious cause, and one patient had a relapse of infection.

Consistent with these observations of clinical utility, we observed initial precipitous reductions in organism burden in excess of 3-log reductions in CFU within 24 h for both ceftaroline 600 mg q12h and q8h regimens against MRSA isolates. Notably, ceftaroline 600 mg q8h maintained bacterial reductions throughout the study duration and was statistically superior to the q12h dosing regimen, a potential advantage not seen in previous studies that simulated human serum exposures (40). Of note, previous studies have simulated plasma concentrations, while this study focused on ELF exposures, which were approximately 23% of free-drug concentrations in plasma (13).

Ceftriaxone was included in this experiment against the MSSA isolates, as it is a first-line empirical treatment of CABP and is routinely continued once MSSA is confirmed as the causative pathogen, as the antibiotic is approved for the treatment of S. aureus infections at a dose of 2 to 4 g per day (41). However, in a 24-h in vitro pharmacodynamic model of 5 clinical MSSA isolates with ceftriaxone MICs of 2 to 8 μg/ml, Iacovides and colleagues found ceftriaxone 2 g every 24 h to be poorly active and statistically inferior to cefazolin and vancomycin regimens (42). Only the single strain with a ceftriaxone MIC of 2 μg/ml produced appreciable bacterial killing (∼2 logs) after 24 h. Our data support that ceftriaxone 2 g every 24 h displays poor activity over 72 h against susceptible MSSA isolates with MICs of ≥2 μg/ml. Furthermore, while increasing the frequency of administration to every 12 h provided some additional activity, ceftriaxone was still significantly less active than comparator agents over the course of the 72-h study duration. Importantly, this was assuming the best-case scenario of 100% penetration into ELF. Given the fact that no other β-lactam, with the exception of cefepime (43), has ever reported 100% ELF penetration, it is likely that ceftriaxone ELF exposure is lower than we simulated, which would only further reduce activity.

While in vitro modeling assesses the antibacterial effect of the agent at concentrations observed at the site of infection, interpretation and application of these results should be made cautiously and in context of the inherent limitations of the in vitro system (i.e., absence of host immune system, shorter treatment window of 72 h, and a simplification of a complex in vivo disease). This is evident in the ceftaroline and ceftriaxone experiments, where minimal activity was observed against some isolates despite fT>MIC exposures (∼40 to 50%) associated with efficacy for cephalosporins (44). Furthermore, appreciable regrowth was observed in the final ∼4 h of each dosing interval when ceftaroline ELF concentrations fell below the MIC and approached 0 μg/ml for the regimen of 600 mg q12h, which is consistent with the minimal in vitro postantibiotic and postantibiotic sub-MIC effect of ceftaroline against S. aureus (45). Lastly, this study used a starting inoculum of ∼10⁶ CFU/ml, which allowed isolates to grow ∼10⁹ CFU/ml over the course of a few hours. While this starting inoculum is well above the threshold for onset of pneumonia in animal infection models (46), it is well established that vancomycin and beta-lactam antibiotics can portray reduced bactericidal activity in the presence of higher inoculums (29); therefore, the observed results may not be applicable to scenarios that begin with a higher inoculum than ∼10⁶ CFU/ml.

In this in vitro experiment, ceftaroline 600 mg q8h achieved the greatest killing against these S. aureus isolates and warrants further investigation for the treatment of pneumonia. Additionally, ceftaroline 600 mg q12h achieved similar activity to vancomycin 1 g q12h in these experiments. Lastly, our data suggest that ceftriaxone be replaced with an alternative antistaphylococcal agent when MSSA is the cause of bacterial pneumonia.