J Clin Pharmacol 2009;49:1071-1078 1071

Linezolid (Zyvox), belonging to oxazolidinone antibiotics, is commonly used for the treatment of patients infected with methicillin-resistant Staphylococcus aureus and vancomycin-resistant enterococci. Although linezolid has been approved worldwide, the Japanese pharmacokinetic (PK) profile has not been characterized in detail. The objective of this study is to develop a population PK model for linezolid that can be applied to a Japanese population. This population PK model was established based on the 1 Japanese phase III and 4 Caucasian phase II/III studies. A total of 2539 linezolid plasma concentration measure-ments from 455 patients, aged 18 to 98 years and body weight 30 to 190.5 kg, were used for the analysis. The data were analyzed using nonlinear mixed effects modeling. Body weight (BW), age, ethnicity, and gender were investi-gated as covariates. The final model was validated by the

bootstrap technique. The PK profiles of linezolid were described with a 1-compartment PK model with first-order absorption and first-order elimination. In the final popu-lation PK model, BW and age were influential covariates on clearance, and the distribution volume was affected by BW. The present population PK model of linezolid described well the PK profiles in Japanese patients who have lower BW and are relatively older compared with those in the United States/European Union.

Keywords: Population pharmacokinetics; linezolid; NONMEM; antibacterial agents; Japanese patients

Journal of Clinical Pharmacology, 2009;49:1071-1078© 2009 the American College of Clinical Pharmacology

Population Pharmacokinetic Analysis of Linezolid in Patients With Infectious

Disease: Application to Lower Body Weight and Elderly Patients

Sadahiro Abe, BSc, Koji Chiba, PhD, Brenda Cirincione, MSc, Thaddeus H. Grasela, PhD, Kaori Ito, PhD, and Toshio Suwa, PhD

From Clinical Pharmacology, Global Research & Development, Tokyo Laboratories, Pfizer, Tokyo, Japan (S. Abe); Department of Drug Develop­ment Science & Clinical Evaluation Faculty of Pharmacy, Keio University, Tokyo, Japan (S. Abe, K. Chiba, T. Suwa); Amylin Pharmaceuticals, Inc, San Diego, California (B. Cirincione); Cognigen Corporation, Buffalo, New York (T. H. Grasela); and Pharmacometrics, Global Research & Development, Pfizer, New London, Connecticut (K. Ito). Submitted for publication January 23, 2009; revised version accepted April 23, 2009. Address for correspondence: Sadahiro Abe, Pfizer Japan, Inc, Shinjuku Bunka Quint Building, 3­22­7, Yoyogi, Shibuya­ku, Tokyo, Japan 151­8589; e­mail: Sadahiro.abe@pfizer.com.DOI: 10.1177/0091270009337947

Linezolid is a member of the oxazolidinone class of synthetic antibacterial agents. The

drug has substantial antimicrobial activity against gram-positive bacteria, such as staphylococci, streptococci, and enterococci, including activity

against methicillin-resistant Staphylococcus aur-eus (MRSA) and vancomycin-resistant enterococci (VRE).1-5

The pharmacokinetics (PK) of linezolid following both oral and intravenous administration have been determined by clinical studies. Because of 100% absolute bioavailability in noncompartment analysis (NCA), no dosage adjustment is necessary when in travenous dosing is changed to the oral route.6 Under steady-state conditions, approximately 30% of the dose appears in the urine as the intact drug, whereas 40% is excreted as the major urine metabolite of hydroxyethyl glycine (PNU-142586), and 10% is a minor metabolite of aminoethoxyacetic acid (PNU-142300).7 For the mechanism of producing the predominant metabolite of PNU-142586, it is proposed that the oxidative reaction is mediated

1072 • J Clin Pharmacol 2009;49:1071-1078

ABE ET AL

nonenzymatically by a ubiquitous reactive oxygen species.8 The chemical oxidation of the predominant meta bolism is consistent with the results that the metabolism of linezolid was not inhibited by cyto-chrome P450 inhibitors.

Linezolid has been approved for use in more than 60 countries worldwide. Recently in Japan, linezolid has also been granted an expanded indication for MRSA in addition to VRE. For the approval, the Ministry of Health, Labor and Welfare (MHLW) had requested to compare the efficacy and safety between Japanese and Caucasians. In the ICH-E5 guideline, various factors are listed as an example of intrinsic factors such as genetic polymorphism, age, gender, height, body weight (BW), lean body mass, body composition, and organ dysfunction. In fact, the influence of BW difference in PK between Japanese and Caucasians is demonstrated in several population PK studies.9,10 The average BW of the Japanese population is ~55 kg, whereas that of the Caucasian population is >70 kg.11 Age should also be considered an important intrinsic factor for the Japanese population because MHLW reported that the average age of Japanese MRSA patients is older comparative-ly.12

The population PK analysis of linezolid had been investigated in previous studies using the 1- compartment model with parallel linear and Michaelis- Menten (MM) elimination,13 2-compartment model with linear elimination,14 2-compartment model with MM elimination,15 2-compartment model with paral-lel linear and MM elimination,16,17 and 2-compartment model with inhibition compartment model.18 How-ever, these population PK studies have not inves-tigated ethnic factors with sufficient numbers of Asian patients. We developed a new population PK model through the evaluation of various covariates related to ethnic factors.

METHODS

Participants and Trial Design

The data obtained from 1 Japanese phase III study and 4 Caucasian phase II/III studies were used for the present population PK analysis. All of Japanese MRSA infectious patients were administered 600 mg of linezolid twice daily (bid) by intravenous and/or oral route, whereas non-Japanese infectious disease patients were primarily administered 600 or 625 mg. The study protocol was approved by the institutional review board, and written informed consent was

obtained from all patients. The study was performed in accordance with the International Con ference on Harmonization guidelines and good clinical practice standards.

All plasma concentrations of linezolid were determined by a validated high-performance liquid chromatography–ultraviolet (HPLC-UV) assay.19 The limit of quantification was 0.01 µg/mL.

Population Pharmacokinetics Analysis

A 1-compartment open PK model with first-order absorption and first-order elimination was chosen as the structural model. The data were analyzed by a nonlinear mixed effects modeling approach using the NONMEM software system (Version V, Level 1.1; GloboMax LLC, Ellicott City, Maryland) with its library subroutines ADVAN2, TRANS2.20 S-Plus (Version 7; Insightful Corporation, Seattle, Washington) was used for data analysis. The first-order conditional estimation (FOCE) method was used for all analyses. The PK parameters obtained from the model were total body clearance (CL), vol-ume of distribution (V), and absorption rate constant (Ka). The Ka was estimated for patients receiving linezolid orally. The interindividual variability was modeled using an exponential error model as given in equation (1).

Pj = TVP ⋅ exp (ηj), (1)

where Pj describes the individual PK parameter, TVP represents the mean typical value of the PK parameter in the population, and ηj represents the interindividual variability, which is assumed to be normally distributed with a mean of 0 and vari-ance of ω2. During stepwise covariate model building, BW, age, gender, and ethnicity were considered in the analysis.

The influence of BW on CL was fitted to equations (2) and (3) by the ordinary least squares method using NONMEM with breakpoint.

CL = θ1 + θ2 ⋅ AGE (AGE ≥ P1), (2)

CL = θ1 (AGE < P1), (3)

where P1 represents the upper limit of the age, which does not influence the CL. As described in the next section, the limit was decided to be 58 years.

For each model, NONMEM computed the mini mum value of the objective function (OFV), a stat istic that

POPULATION PK ANALYSIS OF LINEZOLID

SPECIAL POPULATIONS 1073

was proportional to minus twice the log likelihood of the data. The covariate model was built in a stepwise fashion. Each covariate was added to the base model one at a time during forward selection. A decrease in the OFV of at least 3.84 (χ2, P ≤ .05, degrees of freedom [df] = 1) was considered significant for adding a single covariate into the model. The full model was developed by incorporating all significant covariates.

Each covariate from the full model was deleted one at a time to obtain the final model using the backward elimination procedure. An increase of OFV from the full model of at least 10.83 (χ2, P ≤ .001, df = 1) was used as the chosen criterion for retaining the covariate in the model.

Model Evaluation

The bootstrap resampling method was used to evalu-ate the stability of the final model. The final popula-tion PK model was fitted repeatedly to the 500 additional bootstrap data sets. The obtained mean of the parameter estimates was compared with the final parameter estimates from the original data set. Empirical 95% confidence intervals were constructed by observ-ing the 2.5 and 97.5 quartiles of the resulting parameter distributions for those bootstrap runs.

RESULTS

Model Building

A total of 2539 sets of plasma concentrations from 455 patients were available for the analysis. Demographic background for the population partici-pating in the present population PK analysis is summarized in Table I. The pooled study popula-tion included approximately 42% women and 20% Japanese. The age of these patients ranged from 18 to 98 years with the mean (SD) of 58.6 (18.3) years. The mean BW was 73.1 (24.4) kg with a range of 30 to 190.5 kg. A scatterplot of linezolid concentrations versus time since last dose is shown in Figure 1.

To determine the structural PK model, we compared using OFV among 1-compartment (OFV: 11302) and 2-compartment (11314) models with linear elimina-tion and 1-compartment (14606) and 2-compartment (11841) models with parallel linear and MM elimina-tions from the central compartment. As a result, a 1-compartment model with linear elimination was the most appropriate model for analysis of the present data set, which was provided for all further model building. Bioavailability was fixed at 1 in the structural model because it was reported as approximately 100% previously in NCA6 and population PK analysis.15

Table I Summary of Patient Characteristics

Study 1 Study 2a Study 3a Study 4 Study 5 Total

Dose, mg 625 600, 625 600 600 600 —Number of patients 55 94 155 60 91 455Treatment, % IV + PO dose 58.2 68.1 54.2 1.7 18.7 43.5 IV dose 41.8 31.9 45.8 70.0 81.3 52.7 PO dose 0 0 0 28.3 0 3.7Ethnicity, % Japanese 0 0 0 0 100 20 Non-Japanese 100 100 100 100 0 80Gender, % Male 61.8 67.0 53.5 38.3 61.8 58.2 Female 38.2 33.0 46.5 61.7 31.9 41.8Age, y [Min-Max] 55.4 ± 15.5 51.5 ± 15.9 57.7 ± 18.9 59.2 ± 18.5 68.9 ± 16.8 58.6 ± 18.3 [31-85] [20-85] [18-98] [23-88] [22-95] [18-98]BW, kg [Min-Max] 75.6 ± 20.6 90.2 ± 26.1 72.9 ± 20.7 78.6 ± 21.4 51.0 ± 13.1 73.1 ± 24.4 [40.8-149.7] [47.2-190.5] [34.5-158.8] [38.3-135.9] [30-110] [30-190.5]Target population CAP SSTI Bacteremia UTI, SSTI, Pneumonia, peritonitis, SSTI, bacteremia sepsisBacterial species Streptococcus Gram- Gram- VRE MRSA pneumoniae positive positive

BW, body weight; CAP, community-acquired pneumonia; IV, intravenous; MRSA, methicillin-resistant Staphylococcus aureus; SSTI, soft issue infections; UTI, urinary tract infection; VRE, vancomycin-resistant enterococci.a. One sample of 1 patient in study 2 and 16 samples of 9 patients in study 3 were after administered 400 mg and 200 to 580 mg, respectively. Those data were included in the population pharmacokinetic (PK) analyses because no significant deviation from linear PK was observed at doses up to 625 mg.23

1074 • J Clin Pharmacol 2009;49:1071-1078

ABE ET AL

In the forward selection, the effect of each covariate of age, BW, gender, and ethnicity on CL and age, BW, gender, and ethnicity on V was tested for statistical significance. Because the most sig nificant covariate from the forward selection was the influence of BW on CL, BW effect was added to the base model. In the next step, the most significant covariate was the influence of BW on V, and the effect was also added to the model. Then, the remaining trend was only observed in age and gender on CL. Age, gender, and BW were found as significant covariates in stepwise forward selection, whereas evident correlation between age, gender, and BW was not observed (data not shown). Figure 2 shows the relationship between the CL and age. The CL was decreased with age increasing from about 60 years, which was visually inspected. Then, the breakpoint was computed in the NONMEM analysis. The breakpoint analysis in equations (2) and (3) provided the upper limit P1 as 58 years. Thus, as the results of the forward selection procedure, BW, age, and gender were found to be significant predictors of CL, and BW was found for V.

The final model was determined by a stepwise backward elimination of each parameter. During each step of backward elimination, because the gender on CL resulted in the smallest nonsignificant increase in the objective function, it was removed from the model.

The final population PK model and parameter estimates are presented in Table II. Diagnostic plots of the final population PK model of linezolid are shown in Figure 3. The final model showed an improvement of fit compared with the base model.

Model Evaluation

The reliability of the final population PK model was confirmed by reestimating the model parame-ter estimates and their 95% confidence interval using nonparametric bootstrap approaches.21,22 The average parameter values obtained from the boot-strap analyses and the final estimates from the original data set are compared in Table III. The geo-metric means of these 500 parameter estimates were within a 10% difference from the final popu-lation PK parameters obtained with the original data set.

Age (year)

Cle

aran

ce (

L/h

)

20 40 60 80 100

24

68

1012

14

Figure 2. Relationship between the clearance and age for all patients.

Co

nce

ntr

atio

n (

µg/m

L)

0 20 40 60

0.01

0.10

1.00

10.0

010

0.00

0 20 40 60

0.01

0.10

1.00

10.0

010

0.00

(A) Japanese(B) Non-Japanese

Time since last dose (h) Time since last dose (h)

Co

nce

ntr

atio

n (

µg/m

L)

Figure 1. Plasma concentrations of linezolid (µg/mL) versus time after last dose.

POPULATION PK ANALYSIS OF LINEZOLID

SPECIAL POPULATIONS 1075

DISCUSSION

The use of linezolid is rapidly increasing following approval for VRE and MRSA indications in Japan. The PK information of linezolid is, however, limited

for the Japanese population. This is the first study to determine the population PK of linezolid including Japanese patients.

The patient characteristics of Japanese MRSA in -fectious disease are different from the United States/

Table II Final Population Pharmacokinetic Parameters of Linezolid

Population mean parameterCL L/h θ1 × (WTKGj/69.5)θ3 + θ4×(110 – AGE) × (1 – AGECj)+ θ4 × 52 × AGECjV L θ2 × (WTKGj/69.5)θ5

Interindividual variability (coefficient of variation [CV%])ωCL 46.6%ωV 25.9%Residual variability (CV%)σ2 8.14%

WTKGj = the body weight (BW) (kg) of the jth patient (centered on a median BW of 69.5 kg). AGECj = indicator variable in the jth patient with a value of 1 for total age ≥58 years and 0 otherwise. In (110 – AGE), the value of 110 was chosen because it is larger than the oldest age in value in the database to prevent the minus estimate of TVCL during the minimization process.

Population predicted concentration (µg/mL)

Wei

gh

ted

res

idu

al

0 20 40 60

–15

–10

–50

510

15

Time (h)

Wei

gh

ted

res

idu

al

0 200 400 600 800

–10

–50

510

Individual Predicted Concentration (µg/mL)

Ob

serv

ed C

on

cen

trat

ion

(µg

/mL

)

0 10 20 30 40 50 60

020

4060

Individual Predicted Concentration (µg/mL)

Ob

serv

ed C

on

cen

trat

ion

(µg

/mL

)

0 10 20 30 40 50 60

020

4060

Figure 3. Goodness-of-fit plots for the final model. Plot of population and individual predicted versus observed linezolid concentrations and plot of conditional weighted residuals versus population predicted concentrations and time.

1076 • J Clin Pharmacol 2009;49:1071-1078

ABE ET AL

European Union. For example, it is well known that Japanese BW is lower than that of Caucasians.11 In addition, there are larger patient populations with older age in Japanese patients for MRSA infectious disease.12 In fact, in our data, BW and age in Japanese patients were lower and older than those in Caucasian patients.

The PK of linezolid has been investigated by NCA or population PK in several studies. One previous NCA concluded that linezolid had linear PK charac-teristics.23 By population PK analysis, Antal et al13 proposed 1 compartment with parallel linear and MM elimination using concentration data from patients at steady state following intravenous and oral administrations. Meagher et al16 developed a 2- compartment model with parallel linear and MM elimination from the central compartment using steady-state concentrations after intravenous and oral administrations in a compassionate use study. Beringer et al15 compared linear and MM eliminations from the central compartment in 2-compartment models using single-dose concentrations and con-cluded that both models performed equally. Thus, the linear elimination pathway is pivotal to describe linezolid PK. Moreover, Whitehouse et al14 success-fully described intravenous concentration data by a 2-compartment model with only linear elimination. These various results indicate that the existence of saturable elimination pathway may be controversial. Actually, results of the structural model selection in the present study showed that the simple 1-compart-ment model was the best-fitted model. Possible

explanations of the 1-compartment model selected rather than 2-compartment model lie in the follow-ing different conditions: first, the present data were obtained at steady state. It was shown in a previous study18 that the distribution phase was shorter after multiple dosing than single dosing. Second, the present study had a longer infusion interval (≥1 hour in 81% of patients) than those in previous studies (≤1 hour).14,18 Linezolid is usually used in multiple dosing regimens, and the package inserts recommend an infusion period over 30 to 120 minutes,24,25 which is consistent with conditions in the present studies and presumably reflect the practical conditions.

CLs after multiple dosing were compared bet-ween previous and the present models. Antal et al13 reported that the CLs were 10.99, 6.96, and 4.51 L/h for the concentrations of 1.02, 5.86, and 13.96 µg/mL, respectively, at 84.4 kg and 52 years, whereas CL by the present model resulted in 6.0 L/h. Meagher et al16 estimated the mean average CL as 6.85 L/h, whereas the CL for their age (mean: 55-56 years) and BW (65 kg) was 5.2 L/h by our model.

In the present study, BW was found to be a significant covariate on V. The estimated V was similar to that in previous NCA23 following intravenous administra-tion at the mean BW of 75.3 to 79.4 kg (45-46 vs 51-53 L). The influence of BW to V is also demonstrated by previous PK analysis.13-15 V was calculated by the method of Antal et al13 as 52.6 to 59.1 L at 84.4 kg and 52 years, which was corresponding to the present esti-mation (56.0 L) at the same BW. The value by Meagher et al16 was 65.8 L at 65 kg (mean, 55-56 years), whereas our estimation (44.2 L) was smaller. Because of variability of values among studies, the value in the present analysis is reasonably within the range of previous estimates.

Previous population PK studies provided a Ka of 0.75 to 0.80 h–1 using final models without lag time.13,15 To investigate the influence of Ka on other estimated parameters in the present model, we per-formed a sensitivity analysis of the Ka with the range of 0.53 to 0.80 h–1. As a result, there were insignificant increases in the OFV and negligible changes in the remaining parameters (data not shown), suggesting that the present Ka was estimated adequately.

During covariate analysis, we evaluated relation-ships between CL and age first. The profiles of CL appeared to be flat for patients up to 58 years and afterward decreased linearly with increasing age. Therefore, we estimated the breakpoint and set it in our population PK analysis. To understand the influ-ence of age and the significant covariate of BW on

Table III Comparison of Parameter Estimates for Final Model With Estimates

From 500 Bootstrap Samples

Results of 500 Bootstrap Simulation

Final 95% Bootstrap Estimates Confidence Mean/Final of the Model Interval Estimate Parameters Mean Lower-Upper Ratio, %

θ1, L/h 1.28 1.27 0.541-2.02 99θ2, L 47.0 47.0 44.7-49.7 100Ka, h–1 0.583 0.532 0.223-0.947 91θ3 1.91 1.97 1.302-2.849 103θ4 0.0788 0.0791 0.0621-0.0953 100θ5 0.903 0.904 0.805-1.021 100ωCL 0.211 0.209 0.169-0.253 99ωV 0.0642 0.0607 0.037-0.087 95ωKa 3.27 2.97 1.28-5.12 91

POPULATION PK ANALYSIS OF LINEZOLID

SPECIAL POPULATIONS 1077

linezolid exposure (AUC0-24), we conducted a simu-lation with various ranges and combinations of BW and age (Figure 4). The plots of AUC0-24 estimated from individual CL were overlapped with the simu-lated surface in the final model analysis. The AUC0-24 showed a tendency to increase as BW decreased and/or as age increased. We actually estimated the mean AUC0-24 for patients who had a combination of older age (≥80 years) and lower BW (<50 kg) or younger age (<60 years) and higher BW (≥50 kg) as follows: the patient group with age ≥80 years and BW <50 kg had 730 µg⋅h/mL of AUC0-24, and the patient group with age <60 years and BW ≥50 kg had 207 µg⋅h/mL, showing a 3.5-fold difference. In the present data set, the exposure of Japanese patients was approximately 1.9 times larger than that of Caucasian patients. On the other hand, the number of patients with age ≥65 years and BW <50 kg accounted for 40% of Japanese patients but only 4% of Caucasian patients. This can explain the larger exposure in Japanese patients in the present analyses.

The safety of linezolid with larger exposure for the Japanese MRSA infectious disease patient has been reported recently,26 which indicated that hemato-logical adverse events were observed in linezolid-treated patients. However, all of these adverse events were mild and reversible. They suggested that the higher AUC0-24 by the lower BW and/or older age does not cause severe adverse events in Japanese. These results suggest that dosing adjustment based on age and BW is not necessary.

As for the influence of BW to the linezolid expo-sure, Stein et al23,27,28 reported that the steady-state mean AUC0-12 of obese patients after oral dosing at 600 mg was 92 µg⋅h/mL, which is lower than that of nonobese patients (138-141 µg⋅h/mL). Population

PK analysis including obese patients16 also demon-strated that higher BW patients received lower expo-sure. One possible explanation for the BW influence may lie in the metabolism of linezolid. Wienkers et al8 proposed that linezolid metabolism was medi-ated by ubiquitous reactive oxygen species (ROS) in vivo. They showed that the morpholine moiety of linezolid was oxidized by chemical oxidation of the Fenton reaction, which gave different products from enzymatic oxidation but corresponded to in vivo human metabolites. Furukawa et al29 reported that ROS production was enhanced by endogenous fat accumulation. Thus, the ubiquitous ROS caused by obesity may play a role in the increased CL.

The influence of age was also discussed previ-ously. Sisson et al30 showed that linezolid PK were not influenced by age. Although the result is different from ours, it can be explained by the following: the elderly BW was 10% higher than younger patients in their population. So, effect of age might be blinded by higher BW of the elderly. Meagher et al16 reported that the relationship between total CL and age was likely to be an inverse association. In many drugs, the plasma concentration increases due to the decrease in kidney function as a result of aging.31 However, the PK of linezolid were not different between patients with renal impairment and healthy volunteers,32 and the fraction of urinary excretion was only 35%. It is difficult to explain the increase in AUC by change in renal clearance. Previous population studies15,16 also discussed that CLcr would not influence CL. We speculate that the ROS production might be related to aging. The mechanism(s) of the increase of exposure by aging remains unexplained, and further investiga-tion will be required.

In the present analysis, the gender effect was removed in the backward elimination. Sisson et al30 reported that women had about 20% lower V and CL than men in the NCA. One population PK study detected a gender effect on V,13 but 2 other studies showed no gender effect.14,15 Thus, gender effect is presumably ignorable.

It is hoped that the results of the present study are used for investigating the relationships between exposure and safety/efficacy, especially in Asia, where people have lower BW, and in aging countries such as Japan.

In conclusion, the present population PK analysis has successfully characterized the linezolid PK not only for Caucasian patients but also Japanese patients, including older and lower BW populations.

Financial disclosure: None declared.

40

Weight (kg) 80

120

160

18

3654

7290

Age (year)

050

010

0015

00

AU

C (

µg·h

/mL

)

Figure 4. Surface and observed value plots of AUC as a function of body weight (BW) and age.

1078 • J Clin Pharmacol 2009;49:1071-1078

ABE ET AL

REFERENCES

1. Mutnick AH, Biedenbach DJ, Tumidge JD, Jones RN. Spectrum and potency evaluation of a new oxazolidinone, linezolid: report from the SENTRY Antimicrobial Surveillance Program, 1998-2000. Diagn Microbiol Infect Dis. 2002;43:65-73.2. Ballow CH, Jones RN, Biodenbach DJ; North American ZAPS Research Group. A multicenter evaluation of linezolid antimicro-bial activity in North America. Diagn Microbiol Infect Dis. 2002; 43:75-83.3. Noskin GA, Siddiqui F, Stosor V, Hacek D, Peterson LR. In vitro activities of linezolid against important gram-positive bacterial pathogens including vancomycin-resistant enterococci. Antimicrob Agents Chemother. 1999;43:2059-2062.4. von Eiff C, Peters G. Comparative in-vitro activities of moxi-floxacin, trovafloxacin, quinupuristin/dalfopristin and line-zolid against staphylococci. J Antimicrob Chemother. 1999;43: 569-573.5. Zurenko GE, Yagi BH, Schaadt RD, et al. In vitro activities of U-100592 and U-100766, novel oxazolidinone antibacterial agents. Antimicrob Agents Chemother. 1996;40:839-845.6. Welshman IR, Sisson TA, Jungbluth GL, Stalker DJ, Hopkins NK. Linezolid absolute bioavailability and the effect of food on oral bioavailability. Biopharm Drug Dispos. 2001;22:91-97.7. Slatter JG, Stalker DJ, Feenstra KL, et al. Pharmacokinetics, metabolism, and excretion of linezolid following an oral dose of [14C] linezolid to healthy human subjects. Drug Metab Dispos. 2001;29:1136-1145.8. Wienkers LC, Wynalda MA, Feenstra KL, Gao P, Slatter JG. In vitro metabolism of linezolid (PNU-100766): lack of induction or inhibition of cytochrome P450 enzymes and studies on the mechanism of formation of the major human metabolite, PNU-142586. Abstr Intersci Conf Antimicrob Agents Chemother Intersci Conf Antimicrob Agents Chemother. 1999;39:3. Abstract no. 11.9. Tatami S, Yamamura N, Sarashina A, Young CL, Igarashi T, Tanigawara Y. Pharmacokinetic comparison of an angiotensin II receptor antagonist, telmisartan, in Japanese and Western hyper-tensive patients using population pharmacokinetic method. Drug Metab Pharmacokinet. 2004;19:15-23.10. Tanigawa T, Heining R, Kuroki Y, Higuchi S. Evaluation of interethnic differences in repinotan pharmacokinetics by using population approach. Drug Metab Pharmacokinet. 2006;21:61-69.11. Tamura K. Initial empirical antimicrobial therapy: duration and subsequent modifications. Clin Infect Dis. 2004;15(suppl 1): S59-S64.12. Annual report for the in-hospital infection 8-1 2000 [in Japanese]. http://www.mhlw.go.jp/shingi/2002/10/dl/s1010-3h1.pdf. Accessed April 20, 2009.13. Antal EJ, Grasela T, Bergstrom T, Bruss JB, Wong E. The role of population PK/PD analysis during the implementation of a bridging strategy for linezolid. Presented at: 36th Drug Information Association Meeting; November 16-19, 2000; Hong Kong.14. Whitehouse T, Cepeda JA, Shulman R, et al. Pharmacokinetic studies of linezolid and teicoplan in the critically ill. J Antimicrob Chemother. 2005;55:333-340.

15. Beringer P, Nguyen M, Hoem N, Louis S, Gill M, Gurevitch M, Wong-Beringer A. Absolute bioavailability and pharmacokinetics of linezolid in hospitalized patients given enteral feedings. Antimicrob Agents Chemother. 2005;49:3676-3681.16. Meagher AK, Forrest A, Rayner CR, Birmingham MC, Schentag JJ. Population pharmacokinetics of linezolid in patients treated in a compassionate-use program. Antimicrob Agents Chemother. 2003;47:548-553.17. Turnak MR, Forrest A, Hyatt JM, et al. Multiple-dose pharma-cokinetics of linezolid—200, 400, and 600 mg PO Q.12 H. In Abstracts of the 38th Interscience Conference on Antimicrobial Agents and Chemotherapy, San Francisco, CA, 1999. Washington, DC: American Society for Microbiology; 1999:16. Abstract A51.18. Plock N, Buerger C, Joukhadar C, Kljucar S, Kloft C. Does line-zolid inhibit its own metabolism? Population pharmacokinetics as a tool to explain the observed nonlinearity in both healthy volun-teers and septic patients. Drug Metab Dispos. 2007;35:1816-1823.19. Peng GW, Stryd RP, Murata S, et al. Determination of linezolid in plasma by reversed-phase high-performance liquid chromatog-raphy. J Pharm Biomed Anal. 1999;20:65-73.20. Boeckman AJ, Sheiner LB, Beal SL. NONMEM Users Guides. San Francisco: NONMEM Project Group, University of California at San Francisco; 1998.21. Ette EI. Stability and performance of a population pharma-cokinetic model. J Clin Pharmacol. 1997;37:486-495.22. Ette EI, Onyiah LC. Estimating inestimable standard errors in population pharmacokinetic studies: the bootstrap with Win-sorization. Eur J Drug Metab Pharmacokinet. 2002;27:213-224.23. Stalker DJ, Jungbluth GL, Hopkins NK, Batts DH. Pharmacokinetics and tolerance of single- and multi-dose oral or intravenous linezolid, an oxazolidinone antibiotic, in healthy volunteers. J Antimicrob Chemother. 2003;51:1239-1246.24. Zyvox (Linezolid) [package insert]. New York: Pfizer, Inc; 2004.25. Zyvox (Linezolid) [package insert] (in Japanese). Tokyo: Pfizer Japan, Inc; 2007.26. Kohno S, Yamaguchi K, Aikawa N, et al. Linezolid versus vancomycin for the treatment of infections caused by methicillin-resistant Staphylococcus aureus in Japan. J Antimicrob Chemother. 2007;60:1361-1369.27. Stein GE, Schooley SL, Peloquin CA, et al. Pharmacokinetics and pharmacodynamics of linezolid in obese patients with cel-lulitis. Ann Pharmacother. 2005;39:427-432.28. Perry CM, Jarvis B. Linezolid: a review of its use in the manage-ment of serious gram-positive infections. Drugs. 2001;61:525-551.29. Furukawa S, Fujita T, Shimabukuro M, et al. Increased oxida-tive stress in obesity and its impact on metabolic syndrome. J Clin Invest. 2004;114:1752-1761.30. Sisson TL, Jungbluth GL, Hopkins NK. Age and sex effects on the pharmacokinetics of linezolid. Eur J Clin Pharmacol. 2002; 57:793-797.31. ElDesoky ES. Pharmacokinetic-pharmacodynamic crisis in the elderly. Am J Ther. 2007;14:488-498.32. Brier ME, Stalker DJ, Aronoff GR, et al. Pharmacokinetics of linezolid in subjects with renal dysfunction. Antimicrob Agents Chemother. 2003;47:2775-2780.