Comparative Assessment of Tedizolid Pharmacokinetics and Tissue Penetration between Diabetic Patients with Wound Infections and Healthy Volunteers via In Vivo Microdialysis

ABSTRACT

Herein, we present pharmacokinetic and tissue penetration data for oral tedizolid in hospitalized patients with diabetic foot infections (DFI) compared with healthy volunteers. Participants received oral tedizolid phosphate 200 mg every 24 h for 3 doses to achieve steady state. A microdialysis catheter was inserted into the subcutaneous tissue near the margin of the wound for patients or into thigh tissue of volunteers. Following the third dose, 12 blood and 14 dialysate fluid samples were collected over 24 h to characterize tedizolid concentrations in plasma and interstitial extracellular fluid of soft tissue. Mean ± standard deviation (SD) tedizolid pharmacokinetic parameters in plasma for patients compared with volunteers, respectively, were as follows: maximum concentration (C_max), 1.5 ± 0.5 versus 2.7 ± 1.1 mg/liter (P = 0.005); time to C_max (T_max) (median [range]), 5.9 (1.2 to 8.0) versus 2.5 (2.0 to 3.0 h) (P = 0.003); half-life (t_1/2), 9.1 ± 3.6 versus 8.9 ± 2.2 h (P = 0.932); and plasma area under the concentration-time curve for the dosing interval (AUC_p), 18.5 ± 9.7 versus 28.7 ± 9.6 mg · h/liter (P = 0.004). The tissue area under the concentration-time curve (AUC_t) for the dosing interval was 3.4 ± 1.5 versus 5.2 ± 1.6 mg · h/liter (P = 0.075). Tissue penetration median (range) was 1.1 (0.3 to 1.6) versus 0.8 (0.7 to 1.0) (P = 0.351). Despite lower plasma C_max and delayed T_max values for patients with DFI relative to healthy volunteers, the penetration into and exposure to tissue were similar. Based on available pharmacodynamic thresholds for tedizolid, the plasma and tissue exposures using the oral 200 mg once-daily regimen are suitable for further study in treatment of DFI.

INTRODUCTION

The lifetime risk of infection leading to a foot ulcer among diabetic patients is estimated to be as high as 25% (1). These infections are frequently associated with poor clinical outcomes. Two-thirds of all nontraumatic amputations are attributed to diabetic foot infections (DFI) (2). Additionally, the burden to the health care system associated with treatment of DFI continues to rise, and cost of treatment is estimated to exceed 1 billion dollars annually in the United States (3). As antibiotic resistance rates have also risen, methicillin-resistant Staphylococcus aureus (MRSA) has become more frequent in acute bacterial skin and skin structure infections (ABSSSI) (4). As such, there exists a clear and increasing need among clinicians, regulatory agencies, and the pharmaceutical industry to innovate with an aim to provide new treatment options for this vulnerable patient population.

Tedizolid (Sivextro, Merck & Co., Inc., Kenilworth, NJ), the active moiety of the prodrug tedizolid phosphate, is an oxazolidinone antibiotic currently approved for the treatment of ABSSSI caused by Staphylococcus aureus (including MRSA), Streptococcus pyogenes, Streptococcus agalactiae, and Enterococcus faecalis (5). In addition to these Gram-positive pathogens, tedizolid also displays in vitro activity against anaerobes typically implicated in DFI, such as Bacteroides fragilis, including isolates resistant to imipenem, piperacillin-tazobactam, and metronidazole (6). Given its spectrum of activity and the option for use of the oral formulation against the backdrop of an ever-increasing rise in bacterial resistance rates, tedizolid presents an attractive opportunity for potential use in the treatment of DFI.

Efficacy of any antimicrobial agent is in part tied to its ability to penetrate into the affected tissues and achieve a therapeutic concentration (7–10). Although the pharmacokinetic profile of tedizolid has been described for numerous populations, penetration into tissue during an active infection, as in the case of DFI, has yet to be elucidated (7). Such data would lend necessary support for utilization of this agent in the treatment of these infections.

Previous studies designed to address questions of this nature have employed a technique known as in vivo microdialysis, which allows for the quantitative assessment of interstitial extracellular fluid concentration (7–10). This method provides additional benefits relative to traditional methods of drug quantification in tissue (i.e., direct tissue sampling), as interstitial extracellular fluid may be continuously sampled and contamination via blood during tissue processing can be avoided (11). Here, we sought to describe the pharmacokinetics of oral tedizolid in the plasma and interstitial extracellular fluid obtained from infected tissue of hospitalized patients admitted with DFI compared with those from healthy volunteers.

RESULTS

Participants.

Eighteen participants were enrolled, of which 16 completed the study (10 patient participants and 6 healthy volunteers). One patient withdrew due to an adverse event after the first tedizolid dose; a second patient was withdrawn due to lack of required inclusion/exclusion criteria before any doses were received. Baseline characteristics for both cohorts are provided in Table 1. The patient participant group was poorly controlled with respect to glucose management (HbA1c, 9.8 ± 2.0) and 9 of 10 patients possessed PEDIS grade 3 (i.e., moderate) wound infections. Patients were also significantly older than healthy volunteers. Both plasma and interstitial extracellular tissue fluid (i.e., tissue) data were available for all participants except for one healthy volunteer, whose microdialysis catheter dislodged during sample collection.

TABLE 1.

Characteristics of 10 patients with DFI and 6 healthy volunteers participating in the tedizolid in vivo microdialysis study^a

Characteristic	Value for:		P
	Patients (n = 10)	Healthy volunteers (n = 6)
Age (yrs)	51 ± 17	33 ± 9	0.017
Male gender	5 (50)	3 (50)	0.102
Height (inches)	66.6 ± 5.0	65.7 ± 4.5	0.713
Weight (kg)	89.4 ± 21.0	79.5 ± 17.9	0.350
Body mass index (kg/m2)	31.5 ± 6.3	28.5 ± 5.4	0.342
ClCr (ml/min)	90 ± 32 (19–127)	102 ± 27 (71–135)	0.487
HgbA1Cb (%)	9.8 ± 2.0	NA
Albumin (g/dl)	3.1 ± 0.4 (2.7–3.6)	4.6 ± 0.3 (4.2–5.0)	<0.001
PEDIS grade
2 (mild)	1 (11.1)	NAc
3 (moderate)	9 (88.8)	NA

^aData are reported as mean ± standard deviation (range) or number (percentage).

^bValues reflect laboratory results obtained at enrollment prior to receipt of the study drug.

^cNA, not applicable.

Safety and tolerability.

Among DFI patients, one withdrew after a single dose, following an episode of epigastric pain with nausea and vomiting. A second patient experienced a transient period of hypothermia, which resolved within approximately 16 h. No medical intervention was required. Tedizolid was otherwise well-tolerated among remaining patients. Two of the participants in the healthy volunteer group experienced abnormalities in urinalysis, including the presence of uric acid crystals and asymptomatic bacteriuria, both of which resolved within 72 h. Three participants in the volunteer group also experienced transient headaches (one of which was associated with nausea and some dizziness) which resolved within 12 h. No chemistry, blood count, or liver function test changes occurred between the start of therapy and completion of study interventions.

Plasma pharmacokinetics.

Plasma pharmacokinetic parameters for both patients with DFI and healthy volunteers are summarized in Table 2. The C_max achieved among patients was significantly lower than that among healthy volunteers (1.5 ± 0.5 versus 2.7 ± 1.0 mg/liter; P = 0.005). The T_max was also delayed in patients (5.9 [range, 1.2 to 8.0] versus 2.5 [range, 2.0 to 3.0] h; P = 0.003). The area under the plasma concentration-time curve over the dosing interval (AUC_0–τ) for patients was 36% lower than that of healthy volunteers (18.5 ± 9.7 versus 28.7 ± 9.6 mg · h/liter; P = 0.011). Protein binding was similar between patients and participants (81% ± 3% versus 79% ± 4%, respectively). Total and unbound plasma concentration-time plots for patients with DFI compared with healthy volunteers are provided in Fig. 1.

TABLE 2.

Summary of plasma pharmacokinetic parameters for oral tedizolid among patients with DFI compared with healthy volunteers^a

Parameterb	Valuec for:		P
	Patients with DFI (n = 10)	Healthy volunteers (n = 6)
Cmax (mg/liter)	1.5 ± 0.5	2.7 ± 1.1	0.005
Tmax (h)	5.9 (1.2–8.0)	2.5 (2.0–3.0)	0.003
t1/2 (h)	9.1 ± 3.6	8.9 ± 2.2	0.932
AUC0-τ (mg · h/liter)	18.5 ± 9.7	28.7 ± 9.6	0.004
ƒAUCp (mg · h/liter)	3.9 ± 2.9	6.0 ± 1.8	0.011
CL/F (liter/h)	15.0 ± 6.8	11.4 ± 3.3	0.481
V/F (liter)	177.3 ± 53.7	143.4 ± 50.4	0.143
Protein binding (%)	81 ± 3	79 ± 4	0.549
AUCt (mg · h/liter)	3.4 ± 1.5	5.2 ± 1.6	0.075
Tissue penetration	1.1 (0.3–1.6)	0.8 (0.7–1.0)	0.351

^aDosed to steady state indicates 3 doses of 200 mg q24h.

^bC_max, peak concentration; T_max, time to reach peak concentration; t_1/2, half-life; AUC_0–τ, area under the plasma concentration-time curve over the dosing interval; ƒAUC_p, free area under plasma concentration-time curve corrected for protein binding; CL/F, clearance at steady state corrected for bioavailability; V/F, volume of distribution at steady state corrected for bioavailability; AUC_t, area under tissue concentration-time curve.

^cData are reported as mean ± standard deviation or median (range).

FIG 1.

Average tedizolid plasma concentration-time profiles (48 to 72 h) from diabetic patients and healthy volunteers following oral tedizolid phosphate 200 mg q24h.

Tissue exposure and penetration.

Average in vivo recovery for the microdialysis catheters was 63% ± 16% and 75% ± 12% in patients and healthy volunteers, respectively. Figure 2 displays the tissue concentration-time profiles for patients compared with those of healthy volunteers. The area under the concentration-time curve in tissue (AUC_t) observed in patients was 3.4 ± 1.5, compared with 5.2 ± 1.6 mg · h/liter (P = 0.075) in the volunteer group. The tissue penetration ratio (median [range]) was calculated to be 1.1 (0.3 to 1.6) and 0.8 (0.7 to 1.0) (P = 0.351) for patients with DFI and healthy volunteers, respectively.

FIG 2.

Average tedizolid tissue concentration-time profiles (48 to 72 h) from diabetic patients and healthy volunteers following oral tedizolid phosphate 200 mg q24h.

Pharmacodynamic analyses.

Using the previously defined pharmacodynamic threshold (area under the concentration-time curve over 24 hours in the steady state for the free, unbound fraction of a drug divided by the MIC [fAUC/MIC]) of 3 and an MIC₉₀ for Staphylococcus aureus of 0.5 mg/liter, all 10 patient participants (100%) and all 6 volunteers (100%) were determined to reach this target in plasma (12–14). Assuming the same threshold applied in tissue, 7 of 10 patient participants (70%) and all 5 volunteers (100%) with tissue concentration data available achieved this threshold in the interstitial extracellular fluid.

DISCUSSION

Given its in vitro spectrum of activity covering many bacterial pathogens implicated in ABSSSI, including MRSA, tedizolid is an attractive candidate for further studies in the treatment of DFI. In this study, the plasma pharmacokinetics and penetration into interstitial extracellular fluid of infected wound tissue from patients with DFI were compared with those of tissue from healthy volunteer participants.

Alteration in a drug's pharmacokinetic profile is multifactorial and such alterations may potentially affect treatment optimization. In the current investigation, differences in tedizolid plasma C_max and T_max were observed between these patients with DFI and the healthy volunteers. Patients with DFI also had significantly lower plasma AUC (area under the concentration-time curve) compared with that of healthy volunteers. Differences in the pharmacokinetic profile for infected patients are not unreasonable to expect, given the acuity of illness and the complexity of the overall clinical picture. Patient participants also tended to have higher body mass index (BMI) values than their counterparts, which may have contributed to the observed differences, although previous literature suggests that administration of tedizolid in obese patients did not result in clinically significant changes in AUC (15, 16). Alternatively, the observed differences may be attributed to a nonfasted state among the inpatient group, as enforcement of daily fasting conditions was challenging in this hospitalized diabetic population. The presence of food has been shown to affect both C_max and T_max for tedizolid. Flanagan and colleagues administered a single 600-mg dose of oral tedizolid to 11 volunteers under both fasted and nonfasted conditions (17). A 26.3% reduction in C_max during nonfasted conditions was observed. T_max was also delayed among fed volunteers (median [range], 8.0 [4.0 to 12.0]) compared with fasting individuals (2.0 [1.5 to 3.0]). Overall exposure in plasma AUC (AUC_p), however, was unaffected, which is in contrast to our observations, where plasma AUC was 36% lower in the patient cohort. This is an important observation, given that the pharmacodynamic driver for tedizolid is the AUC/MIC ratio (the area under the concentration-time curve over 24 h in the steady state divided by the MIC) (12, 13). Collectively, these studies suggest that the observed reduction in AUC_p here is unlikely to be the result of a nonfasted state among patient participants or their greater BMI values.

Notably, baseline albumin levels were significantly different among groups. Tedizolid is highly protein bound and as such, changes in volume of distribution might be expected. In the current investigation, however, no differences in protein binding were observed. Furthermore, although no statistical difference in volume of distribution (V/F) was observed, patient participants did have numerically lower V/F compared with volunteers. Such observations may also be attributed to acuity of illness, as diabetes has been shown to affect protein metabolism (18). When corrected for protein binding, fAUC_p (area under the concentration-time curve for the free, unbound fraction of a drug) in plasma remained significantly lower for the patients with DFI (Table 2).

The observed delayed absorption may be attributed to pathogenesis of disease (19). Gastroparesis among diabetics has been linked to prolonged oral absorption times (19, 20). In the current study, although no patient participant had a formal diagnosis of gastroparesis, such an effect may have further contributed to the observed plasma differences relative to the comparator group (i.e., T_max and C_max).

Data regarding a drug's ability to penetrate into tissue of otherwise healthy volunteers provides a necessary benchmark for comparison and future investigation into potential clinical application. Our own healthy volunteer cohort displayed a high degree of penetration relative to plasma with little variability, 0.8 (range, 0.7 to 1.0), which echoes results previously reported. A study conducted by Sahre and colleagues sought to characterize distribution among 12 healthy adult participants of tedizolid into interstitial fluid concentrations following a single oral dose of 600 mg using microdialysis in both skeletal muscle and adipose tissue (7). As with the current study, results showed similar unbound drug concentrations in plasma and tissue. The penetration ratio (AUC_t/fAUC_p) for adipose and skeletal muscle was 1.1 ± 0.2 and 1.2 ± 0.2, respectively. Compared with the current investigation, Sahre and colleagues reported higher rates of drug recovery in vivo. The observed difference may be explained by the utilization of a slower perfusion rate, which allows for a greater degree of diffusion of tedizolid across the dialysis membrane.

Among the inpatient group, we observed tissue exposures for tedizolid to be similar to that of free plasma and not statistically different from those in healthy comparators, albeit more variable (1.1 [range, 0.3 to 1.6]). These data are promising, given that treatment of DFI presents unique pharmacokinetic and pharmacodynamic challenges. Reduced peripheral blood flow to the extremities may mean that adequate antimicrobial concentrations can be difficult to achieve; however, this does not appear to be observed with currently available oxazolidinones. A study of similar design comparing plasma and tissue concentrations of linezolid among nine diabetic patients with lower limb wound infections was conducted by our group (9). Patients received 600 mg of linezolid every 12 h for 3 to 4 doses (until reaching steady state). Two microdialysis catheters were placed in each patient, one at the margin of the wound and one into healthy thigh tissue. The mean penetration ratio was found to be 1.42 (range, 1.08 to 2.23; n = 8) in thigh and 1.27 (range, 0.86 to 2.26; n = 7) in wound. No statistical difference was observed between insertion sites (P = 0.648).

Although the perceived clinical value of a compound is often assessed by its degree of tissue penetration, ultimately it is the absolute free drug concentration at the site of infection relative to the target pathogen MIC that correlates to efficacy (12). While pharmacodynamic targets are not currently available for the interstitial extracellular space of wound tissue, profiling of plasma has provided some insight regarding the required exposures to result in substantive bacterial kill. Flanagan and colleagues conducted a population plasma pharmacokinetic analysis designed to determine the probability of target attainment for a 200-mg dose of tedizolid (12). Using the pharmacodynamic target fAUC_p/MIC, a ratio of 3 was achieved with a 98.3% probability based on an MIC₉₀ for Staphylococcus aureus of 0.5 mg/liter (12–14). In the current investigation, using the aforementioned target, 10 of 10 patients achieved the threshold in plasma and 7 of 10 achieved the threshold in tissue. Although the true pharmacodynamic target in tissue for tedizolid is unknown and in the current analysis is assumed to be equivalent to that of plasma, given the sufficiently high free drug exposures in plasma and the tissue penetration (i.e., roughly equal to that of plasma) observed in the current patient population, we believe the exposure achieved is suitable for killing Gram-positive pathogens. More studies need to be conducted to determine the optimal duration of therapy with tedizolid for DFI.

Tedizolid was largely well tolerated in this study. Noting the single episode of epigastric pain with accompanying nausea previously mentioned, data from clinical trials suggests nausea is in fact the most common treatment-emergent event with tedizolid, occurring in about 8% of cases, and thus it is not unexpected (21). The most commonly associated side effect among the volunteer group was headache, also frequently occurring in clinical trials. No hematologic abnormalities (thrombocytopenia, leukopenia, or anemia) were observed in either group over the 4-day study period.

We acknowledge that the comparative design of the study may be limited by the fact that volunteer samples were collected under fasted conditions and compared with samples from patient participants who were nonfasted. However, our results among volunteers demonstrated consistency compared with previous data, which provides assurance of the robustness of the experimental model. Furthermore, the clinical reality of treating acutely ill diabetic patients is such that fasted conditions are often not reasonable or advisable. As such, these data provide practical insight into the exposure of tedizolid in the DFI population as it would be intended for use.

In conclusion, pharmacokinetic differences were observed between these patients with DFI and healthy volunteers receiving oral tedizolid phosphate. Although plasma AUC_0–τ was 36% lower in these diabetic patients, penetration into the interstitial extracellular fluid of lower limb soft tissue was not significantly different between groups, and tissue concentrations were similar to free plasma for both. Based on the free plasma exposures achieved in patients with DFI and the pharmacodynamic targets associated with a high probability of bacterial kill, these data lend support to further study of tedizolid phosphate 200 mg once daily for the treatment of patients with DFI.

MATERIALS AND METHODS

Participants.

This was an open-label pharmacokinetic study approved by Hartford Hospital's institutional review board (approval number HHC-2015-0268) in 10 patients admitted to Hartford Hospital with DFI and six healthy volunteers. Eligible patients and healthy volunteers provided written informed consent before participating in the study. Patient participants were included if they displayed a documented medical history of type 1 or type 2 diabetes mellitus, receipt of insulin or oral antihyperglycemic agents, and were hospitalized with a chronic ongoing DFI, categorized by the Infectious Disease Society of America's PEDIS (perfusion, extent/size, depth/tissue loss, infection, and sensation) rating scale as either grade 2 or 3 (i.e., mild to moderate) (22). Healthy volunteers >18 years of age were identified via hospital and local advertisement in the Greater Hartford area.

Participants were excluded if they were <18 years of age; pregnant or breast-feeding; displayed a history of hypersensitivity to linezolid or a history of hypersensitivity to lidocaine or lidocaine derivatives; or were anemic, thrombocytopenic, or leukopenic (defined respectively as a hematocrit, platelet, or white blood cell count <75% of the lower limit of normal). Those in the patient group were also excluded if they displayed no palpable pedal pulses on physical exam or were likely to require multiple surgeries that could affect the placement of the microdialysis catheter. Healthy volunteers were additionally excluded if they tested positive on a urine drug screen for cocaine, tetrahydrocannabinol, opiates, benzodiazepines, and amphetamines within 28 days of study initiation; had a history of regular alcohol consumption exceeding 7 drinks/week for women or 14 drinks/week for men within 6 months of screening; used tobacco- or nicotine-containing products in excess of the equivalence of 5 cigarettes per day; used prescription or nonprescription drugs, vitamins, or dietary supplements within 7 days or 5 half-lives (whichever is longer) prior to the first dose of study medication (with the exception of acetaminophen at doses of ≤1 g/day); or used herbal supplements or hormonal methods of contraception within 14 days of first dose of study medication. Healthy volunteer studies were conducted in the Clinical Research Center at Hartford Hospital.

Baseline laboratory values, including complete blood count with differential liver enzymes, serum electrolyte, serum creatinine, urinalysis, and glycosylated hemoglobin (patient group only), as well as a physical examination were conducted prior to initiation of study medication. These tests were repeated upon completion. Any adverse events or lab abnormalities were recorded.

Study medication.

Tedizolid phosphate 200 mg tablets were provided by Merck and Co. (Lots PFGWA, expiring 28 February 2017, and WHBG, expiring 28 February 2018). Participants received oral tedizolid phosphate 200 mg every 24 h for 3 doses to achieve steady state. Other anti-infective agents besides the study drug, with the exception of linezolid, were permitted for the purpose of treatment of hospitalized patients. Doses were administered without regard to food among the patient participants. Healthy volunteers were required to fast for 8 h before and 4 h after the administration of each dose in the Clinical Research Center.

Microdialysis procedure.

After preparation of the insertion site and local anesthesia with 0.5% lidocaine solution, a microdialysis catheter (63 microdialysis catheter; M Dialysis, Inc., North Chelmsford, MA) with a membrane length of 30 mm and a molecular cutoff of 20 kDa was inserted under sterile conditions into the interstitial extracellular tissue near the margin (<10 cm) of the lower extremity wound for patients and into thigh tissue for volunteers. Probes were continuously perfused with normal saline at a flow rate of 2 μl/min using microinfusion pumps (106 microdialysis pump; M Dialysis, Inc., North Chelmsford, MA) for at least 20 min prior to the beginning of sample collection.

Sample collection.

Venous blood was obtained via peripheral intravenous catheter at 48 h from the start of the first dose (i.e., immediately before administration of the third dose), and at 49, 50, 50.5, 51, 51.5, 52, 54, 56, 60, 64, and 72 h. Blood samples were collected using a 10-ml BD Vacutainer (Becton, Dickinson and Company, Franklin Lakes, NJ) containing sodium heparin, immediately centrifuged (2,000 × g for 10 min) to separate plasma, and then stored at −80°C until analysis. Dialysate samples of 120 μl were collected simultaneously with plasma in 200-μl microvials (M Dialysis, Inc., North Chelmsford, MA) at 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 64, 68, and 72 h following administration of the first dose. Dialysate samples were stored in amber polypropylene tubes and immediately frozen at −80°C until analysis.

Microdialysis probe recovery by in vivo retrodialysis.

At the flow rates used during in vivo microdialysis, not all drug molecules are expected to cross the dialysis membrane, so concentrations must be corrected by calibrating each catheter. This is known as in vivo retrodialysis and was conducted with a known supraphysiologic tedizolid concentration (300 mg/liter) after all sampling was completed (7–10). The selected concentration was significantly greater than expected concentrations in tissue so that loss of tedizolid could be estimated across each catheter. Recovery of tedizolid via retrodialysis was calculated as follows: % in vivo recovery = 100 − (concentration_dialysate/concentration_perfusate × 100). All dialysate concentrations were then corrected for recovery before pharmacokinetic analysis as follows: concentration_tissue = 100 × (concentration_sample/% in vivo recovery).

Protein-binding studies.

Protein-binding was assessed for each participant using Centrifree Ultrafiltration devices (Millipore Corporation; Billerica, MA) with 30-kDa molecular cutoff filters. Plasma samples collected 3 h after the last dose were transferred into 3 ultrafiltration devices, and centrifuged for 40 min at 25°C at 2,000 × g to generate ultrafiltrate samples (C_{ultrafiltrate}). Aliquots of plasma at 3 h were also retained for determination of the total drug concentration (C_plasma). Protein binding was calculated as % protein binding = 100 − (100 × C_{ultrafiltrate}/C_plasma).

Analytical procedures.

All plasma, dialysate, and ultrafiltrate samples were analyzed by validated high-performance liquid chromatography (HPLC) assay at the Center for Anti-Infective Research & Development in Hartford, CT. The plasma assay was linear over a range of 0.2 to 5 mg/liter (r² ≥ 0.994). The mean interday coefficient of variation values for low- and high-quality check samples were 4.60% and 4.59%, respectively (n = 19). Accuracy values for interday low- and high-quality check samples were 103.1% and 102.2%, respectively. The mean intraday coefficient of variation values for low and high check samples were 2.75% and 1.43%, respectively (n = 10). Accuracy values for intraday low and high quality check samples were 105.6% and 107.9%, respectively. The dialysate and ultrafiltrate samples were run on a saline standard curve that was linear over a range of 0.05 to 5 mg/liter (r² ≥ 0.994). The mean interday coefficient of variation values for low and high check samples were 6.04% and 4.22%, respectively. The mean intraday coefficient of variation values for low and high check samples were 4.61% and 3.75%, respectively. The lower limits of detection for the plasma and saline assays were 0.2 and 0.05 mg/liter, respectively.

Pharmacokinetic analysis.

Noncompartmental pharmacokinetic analyses for tedizolid were conducted using Phoenix WinNonlin v6.3 (Certara, Princeton, NJ) to determine pharmacokinetic parameters for plasma concentrations, including plasma area under the concentration-time curve from 0 to 24 h, (AUC_0–24) (assessed by the linear-log trapezoidal method), elimination rate constant (k_el),terminal elimination rate constant (λ_z), half-life (t_½) (calculated as ln[2]/λ_z), total plasma clearance (CL/F), and apparent volume of distribution during terminal phase after nonintravenous administration (V/F). The maximum concentration (C_max) and time to C_max (T_max) were calculated by observation of the concentration-time profiles. Tissue penetration was calculated as the ratio of the area under the concentration-time curve in tissue (AUC_t) to free area under plasma concentration-time curve corrected for protein binding (fAUC_p).

Pharmacodynamic analyses.

A previously defined pharmacodynamic threshold ratio of 3 (fAUC/MIC) derived in the translational murine infection model was employed for both tissue and plasma. Target values obtained from both volunteers and patient participants were calculated using the MIC₉₀ for Staphylococcus aureus of 0.5 mg/liter (12–14).

Statistical analysis.

Pharmacokinetic parameters and tissue penetration ratios were compared between groups using a paired t test or Mann-Whitney U test for nonnormally distributed data (Sigma Plot, version 12.5; Systat Software, Inc. San Jose, CA). A P value of <0.05 was considered statistically significant.