page 1 of 32 a new semi-physiological absorption model to assess

36
Page 1 of 32 A New Semi-Physiological Absorption Model to Assess the Pharmacodynamic Profile of 1 Cefuroxime Axetil using Nonparametric and Parametric Population Pharmacokinetics 2 3 J. B. Bulitta 1,# , C. B. Landersdorfer 1,# , M. Kinzig 1 , U. Holzgrabe 2 , F. Sorgel 1,3 4 5 6 1: IBMP – Institute for Biomedical and Pharmaceutical Research, Nürnberg-Heroldsberg, 7 Germany; 8 2: Institute of Pharmacy and Food Chemistry, University of Würzburg, Würzburg, Germany; 9 3: Department of Pharmacology, University of Duisburg – Essen, Essen, Germany; 10 11 Corresponding author: Fritz Sörgel, PhD, BSc, Professor, 12 IBMP – Institute for Biomedical and Pharmaceutical Research, 13 Paul-Ehrlich-Str. 19, D-90562 Nürnberg-Heroldsberg, Germany; 14 Phone: ++49-911-518290, Fax: ++49-911-5182920, e-mail: [email protected] 15 16 Running title: Population PK & PD of Cefuroxime Axetil 17 18 Key words: 19 population pharmacokinetics and pharmacodynamics, 20 nonparametric and parametric population pharmacokinetics, 21 Monte Carlo simulation, 22 semi-physiological Michaelis-Menten absorption model, 23 PK/PD MIC breakpoints 24 25 # Present address: Ordway Research Institute, Albany, NY 12208, USA. 26 27 This work was in part presented at the 46th Interscience Conference on Antimicrobial Agents and 28 Chemotherapy, 2006 (poster A-1119). 29 30 31 Dedication: This article is dedicated to Professor Ulrich Stephan, the Co-founder of IBMP who 32 passed away February 6th of 2009. Without his inspiration and support IBMP would not exist 33 neither would the present research have been performed. We keep him in our hearts. 34 Copyright © 2009, American Society for Microbiology and/or the Listed Authors/Institutions. All Rights Reserved. Antimicrob. Agents Chemother. doi:10.1128/AAC.00054-09 AAC Accepts, published online ahead of print on 15 June 2009 on April 1, 2018 by guest http://aac.asm.org/ Downloaded from

Upload: vuhanh

Post on 31-Jan-2017

217 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: Page 1 of 32 A New Semi-Physiological Absorption Model to Assess

Page 1 of 32

A New Semi-Physiological Absorption Model to Assess the Pharmacodynamic Profile of 1

Cefuroxime Axetil using Nonparametric and Parametric Population Pharmacokinetics 2

3 J. B. Bulitta1,#, C. B. Landersdorfer1,#, M. Kinzig1, U. Holzgrabe2, F. Sorgel1,3 4 5

6

1: IBMP – Institute for Biomedical and Pharmaceutical Research, Nürnberg-Heroldsberg, 7 Germany; 8 2: Institute of Pharmacy and Food Chemistry, University of Würzburg, Würzburg, Germany; 9 3: Department of Pharmacology, University of Duisburg – Essen, Essen, Germany; 10 11

Corresponding author: Fritz Sörgel, PhD, BSc, Professor, 12 IBMP – Institute for Biomedical and Pharmaceutical Research, 13 Paul-Ehrlich-Str. 19, D-90562 Nürnberg-Heroldsberg, Germany; 14 Phone: ++49-911-518290, Fax: ++49-911-5182920, e-mail: [email protected] 15 16

Running title: Population PK & PD of Cefuroxime Axetil 17

18

Key words: 19

population pharmacokinetics and pharmacodynamics, 20

nonparametric and parametric population pharmacokinetics, 21

Monte Carlo simulation, 22

semi-physiological Michaelis-Menten absorption model, 23

PK/PD MIC breakpoints 24

25

#Present address: Ordway Research Institute, Albany, NY 12208, USA. 26

27

This work was in part presented at the 46th Interscience Conference on Antimicrobial Agents and 28

Chemotherapy, 2006 (poster A-1119). 29

30

31

Dedication: This article is dedicated to Professor Ulrich Stephan, the Co-founder of IBMP who 32

passed away February 6th of 2009. Without his inspiration and support IBMP would not exist 33

neither would the present research have been performed. We keep him in our hearts. 34

Copyright © 2009, American Society for Microbiology and/or the Listed Authors/Institutions. All Rights Reserved.Antimicrob. Agents Chemother. doi:10.1128/AAC.00054-09 AAC Accepts, published online ahead of print on 15 June 2009

on April 1, 2018 by guest

http://aac.asm.org/

Dow

nloaded from

Page 2: Page 1 of 32 A New Semi-Physiological Absorption Model to Assess

Page 2 of 32

Abstract: 35

Cefuroxime axetil is widely used to treat respiratory tract infections. We are not aware of a 36

population pharmacokinetic (PK) model for cefuroxime axetil. Our objectives were to develop a 37

semi-physiological population PK model and evaluate the pharmacodynamic (PD) profile for 38

cefuroxime axetil. Twenty-four healthy volunteers received 250mg oral cefuroxime as suspension 39

after a standardized breakfast. LC-MS/MS was used for drug analysis, NONMEM and S-ADAPT 40

(results reported) for parametric population PK, and NPAG for nonparametric population PK 41

modeling. Monte Carlo simulations were used to predict the time of non-protein bound 42

concentration above the MIC (fT>MIC). A model with one disposition compartment, a saturable 43

and time-dependent drug release from stomach and fast drug absorption from intestine yielded 44

precise (r>0.992) and unbiased curve fits and an excellent predictive performance. Apparent 45

clearance was 21.7 L/h (19.8% CV) and volume of distribution 38.7 L (18.3%). Robust (≥90%) 46

probabilities of target attainment (PTA) were achieved by 250 mg Q12h (Q8h) cefuroxime for 47

MICs ≤0.375 mg/L (0.5 mg/L) for the bacteriostasis target fT>MIC≥40% and for MICs ≤0.094 48

mg/L (0.375 mg/L) for the near-maximal killing target fT>MIC≥65%. For the fT>MIC≥40% target, 49

PTAs for 250 mg cefuroxime Q12h were ≥97.8% against S. pyogenes and penicillin-susceptible 50

S. pneumoniae. Cefuroxime 250 mg Q12h (Q8h) achieved PTAs below 73% (92%) against H. 51

influenzae, M. catarrhalis, and penicillin-intermediate S. pneumoniae for susceptibility data from 52

various countries. Depending on the MIC distribution, 250 mg oral cefuroxime Q8h instead of 53

Q12h should be considered especially for more severe infections that require near-maximal 54

killing by cefuroxime. 55

on April 1, 2018 by guest

http://aac.asm.org/

Dow

nloaded from

Page 3: Page 1 of 32 A New Semi-Physiological Absorption Model to Assess

Page 3 of 32

Introduction 56

Cefuroxime axetil is the acetoxyethyl-ester prodrug of cefuroxime. Cefuroxime axetil is 57

reliably absorbed and can be taken with or without a meal, although its extent of bioavailability is 58

enhanced under the influence of food (21, 55). Cefuroxime has been successfully used in the 59

treatment of upper and lower respiratory tract infections as well as genitourinary tract infections 60

(46) and is active against H. influenzae, M. catarrhalis, S. pyogenes, K. pneumoniae, N. 61

gonorrhoeae, penicillin susceptible S. pneumoniae, and against some isolates of penicillin 62

intermediate S. pneumoniae (6, 7, 26-28, 35, 36, 38, 39, 42, 56). 63

A susceptibility breakpoint of ≤1 mg/L has been determined for cefuroxime by national 64

organizations in Britain (BSAC, (8)) and Germany (DIN, (17)). Susceptibility breakpoints from 65

the Clinical and Laboratory Standard Institute [CLSI, (11)] are ≤1 mg/L for S. pneumoniae and 66

≤4 mg/L for Haemophilus spp., Enterobacteriaceae, and Staphylococcus spp. 67

Several authors (32, 38, 41) determined the pharmacokinetic-pharmacodynamic (PKPD) 68

MIC breakpoint for cefuroxime axetil based on the average plasma concentration profiles, but did 69

not incorporate between subject variability (BSV) in their analysis. Ambrose et al. (2) determined 70

the PKPD MIC breakpoint for intravenous cefuroxime via Monte Carlo simulation (MCS) based 71

on literature data and Viberg et al. (52, 53) developed a population PK model for intravenous 72

cefuroxime. We are not aware of a population PK model or MCS for cefuroxime axetil. 73

Population PK and MCS methodology account for the BSV in PK parameters and for the 74

variability in the bacterial susceptibility. A PKPD target is used as surrogate measure to predict 75

successful microbiological or clinical outcome (13, 18, 23, 29, 30, 49). For beta-lactams the 76

duration that the non-protein bound plasma concentration exceeds the MIC (fT>MIC) best predicts 77

these outcomes (3, 14, 18). For cephalosporins, data from animal infection models showed that a 78

target time of 40% fT>MIC correlates with bacteriostasis at 24 h and 60-70% fT>MIC are required 79

on April 1, 2018 by guest

http://aac.asm.org/

Dow

nloaded from

Page 4: Page 1 of 32 A New Semi-Physiological Absorption Model to Assess

Page 4 of 32

for near-maximal bactericidal activity at 24 h (3, 12, 14, 18). Based on these PKPD targets, MCS 80

can predict the probability of target attainment (PTA) at various MICs. If the PTA vs. MIC 81

profile is combined with the expected MIC distribution of the pathogen(s) of interest in a local 82

hospital, the probability of successful microbiological or clinical outcome can be predicted. 83

In addition to increasing the extent of bioavailability (21, 55), administration after a meal 84

may cause a slower rate of cefuroxime absorption due to a prolonged gastric transit time. The rate 85

of gastric emptying after a high-fat meal is likely to be variable and may change over time. 86

Parameters describing the absorption phase may also not be normally or log-normally distributed. 87

As MCS based on parametric population PK models use parametric distributions to describe the 88

variability in PK parameters, we additionally applied nonparametric population PK modeling. 89

The latter offers the advantage that it does not assume any shape for the multivariate distribution 90

of PK parameters. However, sample sizes larger than 24 subjects may be required to adequately 91

describe the shape of the multivariate distribution by a nonparametric variability model. 92

Our first objective was to develop a semi-physiological population PK model for 93

cefuroxime axetil using parametric and nonparametric population PK methods. Secondly, we 94

sought to determine the PTA vs. MIC profiles and the probability of target attainment for specific 95

MIC distributions of various pathogens for Q12h and Q8h dosage regimens with oral cefuroxime. 96

on April 1, 2018 by guest

http://aac.asm.org/

Dow

nloaded from

Page 5: Page 1 of 32 A New Semi-Physiological Absorption Model to Assess

Page 5 of 32

Methods 97

Subjects: Twenty-four (24) healthy, male, Caucasian volunteers participated in the study 98

after they had given their written informed consent. Their demographic data were, average ± SD 99

[range]: age 24.5 ± 3.3 yrs [18-31 yrs], weight 73.8 ± 9.2 kg [58.2-93.6 kg], and height 179 ± 8.0 100

cm [166-193 cm]. The subjects’ health status was assessed by physical examination, 101

electrocardiography and laboratory tests including urinalysis and screening for drugs of abuse. 102

Intake of food and fluid was strictly standardized during the study days. Consumption of tobacco, 103

methylxanthines and alcohol in any form was prohibited from 12 h before administration of study 104

drug until the last sample. The volunteers were closely observed by physicians for the occurrence 105

of adverse events on the study days. The study protocol had been approved by the local ethics 106

committee and the study was conducted following the revised version of the Declaration of 107

Helsinki. 108

Study design and drug administration: The study was a single dose, single-center study. 109

All subjects received an oral suspension of 300.72 mg cefuroxime axetil (equivalent to 250 mg 110

cefuroxime) with 240 mL low-carbonated, calcium-poor mineral water at room temperature. The 111

study drug was administered directly after intake of a standardized breakfast with a significant 112

amount of fat. This breakfast contained 4 slices of crisp bread (50 g), 20 g margarine, 2 slices (40 113

g) of cheese (30% fat content), 25 g jam, 100 mL fruit tea, and 100 mL milk (3.5% fat content). 114

Blood sampling: All blood samples were drawn in heparinated tubes from a forearm vein 115

via an intravenous catheter. Blood samples were drawn immediately before administration and at 116

30, 60, and 90 min and at 2, 2.33, 2.67, 3, 3.33, 3.67, 4, 4.5, 5, 6, 8, 10, and 12 h after 117

administration of study drug. Samples were immediately centrifuged and immediately frozen and 118

stored at -70°C until analysis. 119

on April 1, 2018 by guest

http://aac.asm.org/

Dow

nloaded from

Page 6: Page 1 of 32 A New Semi-Physiological Absorption Model to Assess

Page 6 of 32

Drug analysis: Samples were analyzed by means of an LC-MS/MS method, validated for 120

0.1 mL samples of human plasma. Plasma samples (0.1 mL) were diluted with buffer containing 121

the internal standard and deproteinized by addition of 400 µL of acetonitrile. After thorough 122

mixing, the samples were centrifuged for 5 min at 3,600 rpm at approximately +4 °C, and 123

acetonitrile was removed by extraction with 1 mL dichloromethane. The mixture was centrifuged 124

again and 30 µL of the aqueous phase of each sample were then chromatographed on a reversed-125

phase column (Waters Symmetry® C8), eluted with an isocratic solvent system consisting of 126

ammonium acetate buffer and acetonitrile (70/30, v/v) and monitored by LC-MS/MS with a 127

multiple reaction monitoring method as follows: Precursor → product ion for cefuroxime m/z 128

423 → m/z 207 and internal standard m/z 426 → m/z 156, both analyses were in negative mode. 129

Under these conditions cefuroxime and the internal standard were eluted after approximately 1.4 130

and 1.5 minutes. The Mac Quan software (version 1.5, PE Sciex, Thornhill, Ontario, Canada, 131

1991 - 1997) was used for evaluation of chromatograms. 132

The linearity of the cefuroxime calibration curve was shown from 0.00900 mg/L to 133

10.2 mg/L. The coefficient of correlation for all measured sequences of cefuroxime was at least 134

0.999. The lowest calibration standard of 0.00900 mg/L was set as the lower limit of 135

quantification of the assay for cefuroxime in human plasma. There was no observation below this 136

quantification limit. For the spiked quality control standards of cefuroxime the inter-day precision 137

ranged from 3.2 to 5.0% with an inter-day accuracy (relative error) between -4.3 and 2.1%. The 138

intra-day precision and relative error of the cefuroxime assay ranged from 0.7 to 4.0% and from 139

-0.1 to 3.4%. 140

on April 1, 2018 by guest

http://aac.asm.org/

Dow

nloaded from

Page 7: Page 1 of 32 A New Semi-Physiological Absorption Model to Assess

Page 7 of 32

Population PK analysis 141

Computation: We applied the first order conditional estimation (FOCE) method with the 142

interaction estimation option in NONMEM version VI release 1.2 (NONMEM Project Group, 143

University of California, San Francisco, CA, USA) (5). Initial models were developed in 144

NONMEM V. Model development was primarily performed in NONMEM. The final population 145

PK model was additionally estimated in S-ADAPT (version 1.55, parallelized on a computer 146

cluster) using the importance sampling Parametric Monte-Carlo Expectation-Maximization 147

method (pmethod=8 in S-ADAPT) (4) and by the nonparametric adaptive grid (NPAG) algorithm 148

implemented in the USC*PACK (version 12.00) (34). WinNonlinTM Professional (version 4.0.1, 149

Pharsight Corp., Mountain View, CA, USA) was used for non-compartmental analysis and 150

statistics. 151

Parameter uncertainty was assessed by standard asymptotic formulas in S-ADAPT (4). As 152

NONMEM could not compute asymptotic standard errors in the $COV step for the final model, 153

nonparametric bootstrap methods with 100 replicates were used to calculate standard errors in 154

NONMEM as described previously (9). 155

Structural model: We considered one and two compartment disposition models with first-156

order, zero-order, or mixed-order (Michaelis-Menten) absorption with or without a lag-time. 157

Additionally, a semi-physiological model with a time-dependent release from stomach to 158

intestine and subsequent absorption into the central compartment was developed (Figure 1). The 159

differential equations for this model are (A1: amount of drug in stomach, A2: amount of drug in 160

intestine, A3: amount in central compartment, see Table 2 for parameter explanations): 161

A1Km

A1Vmax

dt

dA1

+

⋅−= (1) 162

A2kA1Km

A1Vmax

dt

dA2abs ⋅−

+

⋅= (2) 163

on April 1, 2018 by guest

http://aac.asm.org/

Dow

nloaded from

Page 8: Page 1 of 32 A New Semi-Physiological Absorption Model to Assess

Page 8 of 32

A3V

CLA2k

dt

dA3abs ⋅−⋅= (3) 164

All initial conditions of all three compartments were zero. The stomach compartment (A1) 165

received a bolus dose of 250 mg cefuroxime at 0 h. The model is simplified, as it did not contain 166

a specific compartment for the prodrug cefuroxime axetil. It was assumed that cefuroxime axetil 167

is converted to cefuroxime before the ester-prodrug reaches the peripheral sampling site. This 168

assumption is considered justifiable for an ester prodrug. The maximum rate of release (Vmax) of 169

drug from the stomach compartment was described by a time dependent function: 170

( )

+

⋅+⋅=

γγ

γ

TPMTC

TPMEmax1VmaxTPMVmax

50

0 (4) 171

Time is denoted as time past meal (TPM). Starting from a maximum rate of release at time zero 172

(Vmax0), the Hill-function modifies the maximum rate of release over time with TC50 denoting 173

the time of 50% of the maximal change and Emax the extent of maximal change. For 174

TPM>>TC50 the maximum rate of release approaches Vmax0 · (1 + Emax). Therefore, an Emax 175

of -1 represents complete inhibition of gastric release, an Emax of 0 an unchanged maximum rate 176

of gastric release, and an Emax of 1 a twice as fast maximum rate of gastric release. 177

Competing models were discriminated by their predictive performance assessed via visual 178

predictive checks (VPCs), their objective function (or log-likelihood), and standard diagnostic 179

plots. 180

Parameter variability and observation model: For parametric population PK modeling in 181

NONMEM and S-ADAPT, we estimated the BSV of PK parameters by assuming a log-normal 182

distribution. The maximum extent of inhibition or stimulation of gastric release for the ith subject 183

(Emaxi) was constrained to a lower bound of -1 by the following logit-transformation: 184

( )( )

++

+⋅+=

i

ii

BSVEmaxLg_Emaxexp1

BSVEmaxLg_Emaxexp101-Emax (5) 185

on April 1, 2018 by guest

http://aac.asm.org/

Dow

nloaded from

Page 9: Page 1 of 32 A New Semi-Physiological Absorption Model to Assess

Page 9 of 32

The Lg_Emax is the estimated population mean (arithmetic mean) on transformed scale and 186

BSVEmaxi the random deviate for the ith subject on transformed scale. This transformation 187

assures that all Emaxi range from -1 to 9. A sensitivity analysis showed that the upper limit of 9 188

did not affect the curve fits or predictive performance of this model. 189

Plasma concentration time profiles were simulated for at least 4,800 subjects for each 190

competing model to calculate the median and nonparametric 80% prediction interval (10% to 191

90% percentile) of the predicted concentrations. The same percentiles were calculated for the 192

observations to visually assess whether the simulated percentiles closely matched the percentiles 193

of the observations. For nonparametric population PK models in NPAG, this VPC was performed 194

either based on the nonparametric distribution of PK parameters characterized by the support 195

point matrix or based on a parametric, multivariate log-normal distribution of PK parameters. In 196

all three programs full variance-covariance matrices were estimated and used for MCS. 197

The residual unidentified variability was described by a combined additive and 198

proportional error model. We used the adaptive gamma option in NPAG to estimate the residual 199

error described by the assay error polynomial. 200

Monte Carlo simulation: A target of 60-70% fT>MIC has been identified for near-maximal 201

bactericidal activity of cephalosporins and a target of 40% is required for bacteriostasis (14, 18). 202

Therefore, we used a PKPD target of 65% fT>MIC for near-maximal bactericidal activity and 40% 203

fT>MIC for bacteriostasis. A range of MICs from 0.031 to 64 mg/L was considered. As the protein 204

binding for cefuroxime has been reported to range between 33 and 50% (22, 24, 46), we assumed 205

an average protein binding of 42% for cefuroxime. 206

We compared dosage regimens of 250 mg or 500 mg oral cefuroxime given every 12 h 207

(Q12h) or every 8 h (Q8h) at steady-state. For the final population PK models in NONMEM, S-208

ADAPT and NPAG, we simulated 10,000 subjects for each dosage regimen in absence of 209

on April 1, 2018 by guest

http://aac.asm.org/

Dow

nloaded from

Page 10: Page 1 of 32 A New Semi-Physiological Absorption Model to Assess

Page 10 of 32

residual error. NONMEM was used to simulate the full concentration time profiles at steady-state 210

with very frequent sampling based on the final population PK model and the estimated full 211

variance-covariance matrix. The fT>MIC values were calculated by linear interpolation between 212

simulated time points as previously described (9). 213

The PTA was estimated by calculating the fraction of subjects who attained the PKPD 214

target at each MIC. The highest MIC with a PTA of at least 90% was used as the PKPD MIC 215

breakpoint. 216

To put these PTAs into clinical perspective, we calculated the PTA expectation value (40) 217

for successful treatment against pathogens from specific MIC distributions as described 218

previously (9). The PTA expectation value is the PTA for treatment of infections caused by 219

bacteria from a specific MIC distribution (ideally the MIC distribution of each local hospital). 220

The PTA expectation value was calculated based on published MIC distributions. We 221

used susceptibility data from the UK (39) collected in 2002 and 2003 on H. influenzae (n=581), 222

M. catarrhalis (n=269), and S. pneumoniae (n=519), susceptibility data from Canada (56) 223

collected in 2001 and 2002 on H. influenzae (n=1350), and susceptibility data from Germany (7) 224

collected in 2002 on H. influenzae (n=300), M. catarrhalis (n=308), S. pneumoniae (n=331), and 225

S. pyogenes (n=340). Additionally, we used susceptibility data from a global surveillance study 226

(6) collected between 1997 and 2000 on penicillin susceptible S. pneumoniae (n=2102) and 227

penicillin intermediate S. pneumoniae (n=1024), susceptibility data from a European surveillance 228

study (27) collected between 1997 and 1999 on S. pneumoniae (n=2018) and S. pyogenes 229

(n=662), and susceptibility data from North America (26) collected between 1997 and 1999 on S. 230

pyogenes (n=119), S. pneumoniae (n=417), H. influenzae (n=300), and M. catarrhalis (n=231). 231

The PTA expectation values were also calculated for the MIC distributions for H. influenzae 232

(n=66,947), K. pneumoniae (n=34,629), M. catarrhalis (n=14,308), S. aureus (n=10,620), S. 233

on April 1, 2018 by guest

http://aac.asm.org/

Dow

nloaded from

Page 11: Page 1 of 32 A New Semi-Physiological Absorption Model to Assess

Page 11 of 32

pneumoniae (n=18,869), N. gonorrhoeae (n=655), and N. meningitidis (n=257) based on the 234

multinational database of the European Committee on Antimicrobial Susceptibility Testing 235

(EUCAST) (19). 236

on April 1, 2018 by guest

http://aac.asm.org/

Dow

nloaded from

Page 12: Page 1 of 32 A New Semi-Physiological Absorption Model to Assess

Page 12 of 32

Results 237

The PK parameters from non-compartmental analysis (Table 1) were in good agreement 238

with literature (42, 46). We found an average ± SD terminal half-life of 1.34 ± 0.13 h and peak 239

concentration of 2.64 ± 0.64 mg/L between 2 and 5 h post dose. The variability in terminal half-240

life (9.4% coefficient of variation) was lower than variability in apparent clearance (20%), peak 241

concentration (24%) and time of peak (24%). 242

A biphasic absorption pattern was found for 5 of 24 subjects and a “plateau-like” peak for 243

8 of 24 subjects (Figure 2). These shapes could not be described by standard first-order or zero-244

order absorption models that included a lag-time. Compared to the final semi-physiological 245

model, the objective function difference in NONMEM was 721 points for the first-order 246

absorption model with lag-time, 359 points for the zero-order absorption model with lag-time, 247

and 189 points for the Michaelis-Menten absorption model with lag-time (likelihood ratio test: 248

p<0.0001 for all comparisons). The semi-physiological absorption model with a two-249

compartment disposition model had a 25 point lower objective function compared to the same 250

absorption model with one disposition compartment. As the latter model showed precise curve 251

fits and an excellent predictive performance, we chose the simpler model as final model. 252

The individual curve fits for the semi-physiological model (Figure 1) were excellent in all 253

three programs (Figure 2). The model was flexible enough to fit profiles with “sharp” peaks, 254

“plateau-like” peaks, and dual peaks. No estimation algorithm (program) provided the best fit for 255

every subject. The linear regression plot of individual fitted vs. observed concentrations had a 256

slope of 1.007 and intercept of -0.014 mg/L in NONMEM (r = 0.9941), a slope of 1.011 and 257

intercept of -0.011 mg/L in S-ADAPT (r = 0.9928), and a slope of 1.003 and intercept of 258

+0.011 mg/L in NPAG (r = 0.9935). 259

on April 1, 2018 by guest

http://aac.asm.org/

Dow

nloaded from

Page 13: Page 1 of 32 A New Semi-Physiological Absorption Model to Assess

Page 13 of 32

The final estimates (Table 2 and Table 3) were precise. Relative standard errors were 31% 260

or less for all population means (except for Km, 70% in NONMEM and 39% in S-ADAPT) and 261

41% or less for all BSV estimates. Estimates for apparent clearance and volume of distribution 262

were similar between all three programs. For the absorption parameters, differences were more 263

apparent. The low Km/Dose (mean: 0.433%) from NONMEM indicated that the release from 264

stomach was estimated to be essentially a zero-order process and that Vmax was inhibited in 265

some subjects and stimulated in others as indicated by the negative or positive individual Emax. 266

In S-ADAPT, the release from stomach to intestine had partial first-order and zero-order 267

properties as indicated by the estimate of 42.6% for Km/dose. This rate of release was more 268

notably stimulated, since the median [90% percentile] of individual Emax were 1.82 [5.55]. 269

NPAG estimated the release from stomach to intestine primarily as a first-order process 270

(Km/Dose: 343%) and stimulation of gastric emptying was more pronounced compared to S-271

ADAPT. The mean time of 50% change in the maximum rate of gastric emptying (TC50) after a 272

standardized breakfast was 1.61 h in NONMEM and S-ADAPT and 2.08 h in NPAG. Individual 273

TC50 estimates were variable (Table 2). 274

The VPCs (Figure 3) indicated that the nonparametric simulation based on the support 275

points from NPAG yielded the closest match between predicted and observed concentrations. 276

This was expected, since this simulation is based on the nonparametric distribution of PK 277

parameters that yielded precise and unbiased fits for all 24 concentration-time profiles. The 278

parametric simulations based on the estimates from S-ADAPT and NPAG matched the median 279

and 10-90% percentile of the observations more closely between approximately 1 and 4 h 280

compared to NONMEM (Figure 3). For the three parametric simulations, S-ADAPT yielded the 281

best representation of the observations during the terminal phase followed by NONMEM. The 282

on April 1, 2018 by guest

http://aac.asm.org/

Dow

nloaded from

Page 14: Page 1 of 32 A New Semi-Physiological Absorption Model to Assess

Page 14 of 32

predicted variability was slightly too wide during the terminal phase for the parametric simulation 283

(Figure 3) using the variance-covariance matrix derived from the support point matrix in NPAG. 284

As the VPCs showed that the nonparametric simulation based on NPAG and the 285

parametric simulation based on S-ADAPT had the best predictive performance, breakpoints from 286

MCS are reported for these two models. Breakpoints of the other two simulation models were 287

within one 1.5-fold dilution. The PTA vs. MIC profiles were similar for these two models (Figure 288

4). For the bacteriostasis target fT>MIC ≥ 40%, the PKPD MIC breakpoint in NPAG / S-ADAPT 289

was 0.5 / 0.375 mg/L for 250 mg Q12h and 0.5 / 0.5 mg/L for 250 mg Q8h. As these PK models 290

are linear with dose, 500 mg Q12h achieved breakpoints of 1 / 0.75 mg/L and 500 mg Q8h 291

achieved 1 / 1 mg/L for the fT>MIC ≥ 40% target. For the near-maximal killing target fT>MIC ≥ 292

65%, PKPD MIC breakpoints were identical between S-ADAPT and NPAG and were 293

0.094 mg/L for 250 mg Q12h, 0.375 mg/L for 250 mg Q8h, 0.188 mg/L for 500 mg Q12h, and 294

0.75 mg/L for 500 mg Q8h. Additional simulations with a hypothetical faster rate of absorption 295

showed that the PKPD MIC breakpoints were lower, if rate of absorption was faster. The 296

decrease in breakpoints was most pronounced for the 65% fT>MIC target and Q12h dosing. 297

High PTA expectations values (≥97.8% for the 40% fT>MIC target) were achieved by all 298

three dosage regimens shown in Table 4 against S. pyogenes, penicillin susceptible S. 299

pneumoniae, N. gonorrhoeae, and N. meningitidis (results not shown for the latter two 300

pathogens). High (>90%) PTA expectation values were achieved against some but not all MIC 301

distributions for S. pneumoniae. The PTA expectation values were notably lower for penicillin 302

intermediate S. pneumoniae, H. influenzae, M. catarrhalis, S. aureus, and K. pneumoniae (results 303

not shown for the latter two pathogens). 304

on April 1, 2018 by guest

http://aac.asm.org/

Dow

nloaded from

Page 15: Page 1 of 32 A New Semi-Physiological Absorption Model to Assess

Page 15 of 32

Discussion 305

Cefuroxime axetil has been used widely for treatment of community-acquired upper and 306

lower respiratory tract infections (including community-acquired pneumonia) (46) that are often 307

caused by S. pneumoniae, H. influenzae, M. catarrhalis, and S. pyogenes. The reported MIC90’s 308

of cefuroxime against H. influenzae and M. catarrhalis are often 2 or 4 mg/L (42, 46). Whereas 309

most isolates would be deemed susceptible by the CLSI breakpoint of ≤4 mg/L for those 310

pathogens, a significant number of isolates would be considered resistant according to the BSAC 311

and DIN susceptibility breakpoint of ≤1 mg/L. To evaluate which breakpoint is in better 312

agreement with the predicted PKPD MIC breakpoint, we applied parametric and nonparametric 313

population PK modeling and MCS for cefuroxime axetil. 314

The fT>MIC best predicts the clinical and microbiological success for cephalosporins. As 315

the prolonged absorption phase of cefuroxime axetil has a notable influence on the fT>MIC values, 316

it was critical to develop a population PK model that adequately captures the rate of absorption 317

and its between subject variability that was observed in our and in literature studies. We 318

intensively qualified the predictive performance of our population PK model to assure that the 319

model predicted PKPD MIC breakpoints are sound. An 8-fold increase (from 125 to 1000 mg) in 320

the oral cefuroxime axetil dose after a meal causes a 7.5-fold increase in the area under the curve 321

(AUC) and 6.5-fold increase in peak concentrations (Cmax) (21) and causes no systematically 322

altered time of peak concentration (Tmax) (21). Data from rats suggest a saturable component for 323

the rate of absorption (43-45). 324

We found a range of complex absorption patterns in healthy volunteers (Figure 2). Models 325

with first-order or zero-order absorption with or without a lag-time can describe the dose-326

proportionality in AUC and Cmax (21), but cannot describe a mixed-order rate of absorption and 327

the complex absorption profiles observed in our study. A mixed-order absorption model with a 328

on April 1, 2018 by guest

http://aac.asm.org/

Dow

nloaded from

Page 16: Page 1 of 32 A New Semi-Physiological Absorption Model to Assess

Page 16 of 32

lag-time could describe the plasma concentration time profiles of cefuroxime axetil at one dose 329

level (results not shown). However, such a mixed-order absorption model would predict a notable 330

increase in Tmax with cefuroxime doses. This is in disagreement with the data by Finn et al. (21). 331

A mixed-order absorption model cannot describe profiles with a dual peak. 332

Food increases the extent of bioavailability of cefuroxime axetil from 36% in fasting 333

subjects to 52% after a meal (21). Similar results were found by Williams & Harding (55). In 334

both studies (21, 55), Tmax is prolonged by approximately 0.6 to 0.7 h for administration with 335

food. Potential reasons for the increased bioavailability and slightly longer Tmax under fed 336

conditions include a more complete dissolution of cefuroxime axetil due to a longer residence 337

time in the stomach and due to bile acid secretion stimulated by the presence of lipids in the 338

intestine (37, 51). 339

The proposed absorption model (Figure 1) is in agreement with the observations of 340

literature studies at various dose levels of cefuroxime axetil (21, 55). One limitation of our study 341

is that we only had data at 250 mg oral cefuroxime. Therefore, our simulation results for 500 mg 342

oral cefuroxime q12h should be interpreted conservatively. The saturable rate of absorption is 343

described by a mixed-order release of drug from stomach to intestine that is primarily saturated 344

due to the presence of food and not due to the cefuroxime axetil dose. In our model, Vmax and 345

Km are expressed as fractions of dose and this causes Tmax to be independent of dose. The 346

second peak in some profiles was described by an increase in rate of gastric release over time. 347

This semi-physiological model proved to be robust (Table 2) and to yield excellent 348

individual curve fits (Figure 2) for all three population PK algorithms and programs. This 349

absorption model was able to capture relevant features of complex oral absorption profiles (25, 350

33, 54) which showed that the rate of gastric emptying is important for the absorption of 351

amoxicillin and clavulanic acid (54). 352

on April 1, 2018 by guest

http://aac.asm.org/

Dow

nloaded from

Page 17: Page 1 of 32 A New Semi-Physiological Absorption Model to Assess

Page 17 of 32

Estimation of a full variance-covariance matrix and its use during simulations yielded the 353

best predictive performance, caused no instability during estimation and did not prolong runtimes 354

in S-ADAPT and NPAG. This saves modeling time, since there are fewer decisions about the 355

choice of the parameter variability model in S-ADAPT and NPAG compared to NONMEM. 356

NPAG does not directly estimate the variance-covariance matrix rather than always derives this 357

full matrix from the estimated support points. Estimating a full variance-covariance matrix in 358

NONMEM tended to cause model instability (i.e. unsuccessful termination messages and 359

inability of NONMEM to obtain asymptotic standard errors) and notably increased estimation 360

times in NONMEM. 361

The most important difference between the parametric and nonparametric approach is that 362

the former describes BSV by a parametric, multivariate distribution (often a multivariate log-363

normal distribution). In contrast, nonparametric methods use a discrete set of support points to 364

exactly store the BSV and correlation structure of all estimated PK parameters in the studied 365

patient population. In the simplest case, each support point essentially represents a complete set 366

of PK parameters for one patient and has a probability of 1 / number of subjects. 367

As the variability of individual PK parameter estimates in the studied subject population 368

is “exactly” represented by the set of support points, it was expected that the nonparametric VPC 369

had the best predictive performance (Figure 3). The parametric VPC using estimates from S-370

ADAPT had the best predictive performance among the parametric VPCs. We reported the 371

results from a parametric MCS in S-ADAPT with 10,000 virtual patients and from a 372

nonparametric MCS in NPAG. As every support point had the same probability for this study 373

with frequent sampling, the latter MCS was identical to simulating from the individual PK 374

parameter estimates of the 24 subjects. 375

on April 1, 2018 by guest

http://aac.asm.org/

Dow

nloaded from

Page 18: Page 1 of 32 A New Semi-Physiological Absorption Model to Assess

Page 18 of 32

The PTA vs. MIC profiles from S-ADAPT and NPAG were similar (Figure 4). For 250 376

mg oral cefuroxime q12h or q8h, PKPD MIC breakpoints fell between 0.375 and 0.5 mg/L for 377

the bacteriostasis target 40% fT>MIC, but were approximately 4-fold lower (0.094 mg/L) for q12h 378

dosing of 250 mg and the near-maximal killing target 65% fT>MIC. Dosing 250 mg (500 mg) 379

every 8h increased the latter breakpoint to 0.375 mg/L (0.75 mg/L). Koeth et al. (32) determined 380

a PKPD MIC breakpoint of ≤1 mg/L for susceptibility for standard cefuroxime dosage regimens 381

based on the fT>MIC ≥ 40-50%. This higher breakpoint is expected, since Koeth et al. (32) used 382

average PK parameters for simulation and did not include BSV. 383

We did not manually increase the BSV in clearance and volume of distribution to mirror 384

the higher variability in critically ill patients, as the relatively low PKPD MIC breakpoints for 385

oral cefuroxime do not support treatment of critically ill patients. A higher variability in PK 386

parameters and lower average extent of bioavailability after intake of cefuroxime in the fasting 387

state (21, 55) are expected to result in lower PKPD MIC breakpoints than reported here. As the 388

sample size of 24 subjects probably did not allow us to obtain precise estimates of the between 389

subject variability in the whole patient population, the results of our MCS should be interpreted 390

conservatively. The probability of clinical success of the simulated cefuroxime axetil dosage 391

regimens will ultimately depend on the MIC distribution of the pathogen(s) of interest in the local 392

hospital. Dosing 250 mg cefuroxime Q8h instead of Q12h had only a small benefit for S. 393

pyogenes, penicillin susceptible S. pneumoniae, since 250 mg Q12h achieved high PTA 394

expectation values, especially for the bacteriostasis target (Table 4). 395

Administering cefuroxime axetil every 8 h yielded notably higher PTA expectation values 396

for some but not all MIC distributions of H. influenzae and M. catarrhalis. Although 500 mg oral 397

cefuroxime q8h are above the typically recommended oral cefuroxime dose, parenteral doses of 398

up to 6,000 mg split into four daily doses are recommended for severe infections. 399

on April 1, 2018 by guest

http://aac.asm.org/

Dow

nloaded from

Page 19: Page 1 of 32 A New Semi-Physiological Absorption Model to Assess

Page 19 of 32

To put the results of our MCS into a clinical perspective, we compared our PTA 400

expectation values to the microbiological and clinical outcomes in clinical studies. Clinical data 401

in children with pneumococcal acute otitis media suggest a breakpoint of about 0.5 mg/L for oral 402

cefuroxime (15). Shah et al. (48) studied hospitalized patients and outpatients in 14 countries in 403

Europe, Africa, and South America with acute exacerbation of chronic bronchitis (AECB) and 404

find a 60% bacteriological overall cure rate for 250 mg oral cefuroxime twice daily. This cure 405

rate is comparable to the placebo response rate for AECB (31, 47), depending on the severity of 406

disease. Interestingly 56% (22 of 39) of the H. influenzae isolates were eradicated. The authors 407

report a clinical cure rate of 66% / 61% (per-protocol / intention-to-treat) at their clinical endpoint 408

(5-14 days post treatment) and of 53% / 39% at the follow-up 3 to 4 weeks post treatment. 409

Although the authors did not report the MICs in those patients, the failures for the treatment of 410

H. influenzae show a sub-optimal effectiveness of oral cefuroxime against this pathogen. 411

Chodosh et al. (10) find a significantly lower microbiological eradication rate for 500 mg 412

oral cefuroxime bid (82%) vs. 500 mg oral ciprofloxacin bid (96%) in an outpatient trial with 413

AECB patients. Cefuroxime eradicated S. pneumoniae in 100% of the cases (13/13), but had only 414

an eradication rate of 76% (19/25) for M. catarrhalis and of 86% (19/22) for H. influenzae. In an 415

outpatient trial with AECB patients, de Abate et al. (16) found a clinical cure rate of 77% for 416

250 mg oral cefuroxime bid which was significantly lower than the clinical cure rate of 89% for 417

400 mg gatifloxacin once daily. The microbiological eradication rate was 77% for cefuroxime 418

and 90% for gatifloxacin. 419

A trial in patients with community-acquired pneumonia (20) showed a significantly lower 420

microbiological eradication rate of H. influenzae for combinations of intravenous ceftriaxone (1 421

to 2 g once daily or bid) and/or oral cefuroxime axetil (500 mg bid) compared to intravenous 422

and/or oral levofloxacin (500 mg once daily). The former regimen had an eradication rate of 79% 423

on April 1, 2018 by guest

http://aac.asm.org/

Dow

nloaded from

Page 20: Page 1 of 32 A New Semi-Physiological Absorption Model to Assess

Page 20 of 32

and the latter of 100%. Upchurch et al. (50) found a clinical cure rate of 74.5% for treatment of 424

acute bacterial sinusitis with 250 mg oral cefuroxime for 10 days but did not document the 425

bacterial etiology. Alvarez-Sala et al. (1) found a clinical success rate of approximately 82% for 426

patients with H. influenzae and S. pneumoniae for 250 mg oral cefuroxime Q12h. 427

In conclusion, we developed a semi-physiological population PK model for oral 428

cefuroxime which provided precise and unbiased individual curve fits for complex absorption 429

profiles and had an excellent predictive performance. The nonparametric VPC based on NPAG 430

showed a better predictive performance than the best parametric VPC in S-ADAPT. The PK/PD 431

MIC breakpoint was 0.375 to 0.5 mg/L for 250 mg oral cefuroxime Q12h and 0.5 mg/L for 250 432

mg oral cefuroxime Q8h for the bacteriostasis target fT>MIC ≥ 40%. Dosing 250 mg cefuroxime 433

Q8h instead of Q12h increased the breakpoint for the near-maximal killing target 65% fT>MIC 434

from 0.094 mg/L to 0.375 mg/L. These breakpoints were (slightly) lower than the susceptibility 435

breakpoint of ≤1 mg/L provided by the BSAC and DIN, whereas the CLSI breakpoint of ≤4 mg/L 436

is higher for most pathogens. Oral cefuroxime (250 mg Q12h or Q8h) achieved high PTA 437

expectation values against S. pyogenes (≥96.7%) and penicillin-susceptible S. pneumoniae 438

(≥91.4%), but notably lower PTA expectation values against M. catarrhalis, penicillin-439

intermediate S. pneumoniae, and H. influenzae for most studied MIC distributions from various 440

countries. Administering 250 mg oral cefuroxime Q8h instead of Q12h was most beneficial for 441

the near-maximal killing target for MICs between 0.094 and 0.375 mg/L. Future clinical studies 442

that assess the MIC of the causative pathogen are warranted to validate these predictions for the 443

clinical and microbiological success for Q12h and Q8h cefuroxime axetil dosage regimens. 444

445

Acknowledgment 446

We thank Dr George Drusano for fruitful discussions about this project.447

on April 1, 2018 by guest

http://aac.asm.org/

Dow

nloaded from

Page 21: Page 1 of 32 A New Semi-Physiological Absorption Model to Assess

Page 21 of 32

448

Table 1: PK parameters from non-compartmental analysis for 250 mg oral cefuroxime 449

given as cefuroxime axetil 450

451

Parameter Unit Average ± SD Median [Min - Max]

Area under the curve from time zero to infinity

mg h L-1 11.9 ± 2.49 11.6 [8.49 - 18.1]

Peak concentration mg L-1 2.64 ± 0.64 2.54 [1.65 - 3.90]

Time of peak concentration h 2.98 ± 0.73 2.83 [2.00 - 5.00]

Terminal half-life h 1.34 ± 0.13 1.35 [1.08 - 1.54]

Apparent total clearance L h-1 21.8 ± 4.29 21.5 [13.8 - 29.4]

Apparent volume of distribution at steady-state a

L 54.1 ± 15.7 51.2 [34.7 - 102]

Apparent volume of distribution during the terminal phase

L 41.7 ± 7.64 44.4 [27.4 - 54.9]

Time of total concentration above 1 mg/L

h 4.97 ± 0.79 4.90 [3.43 - 7.08]

Time of total concentration above 0.5 mg/L

h 6.90 ± 0.77 6.75 [5.13 - 9.02]

452

453

a: To calculate the mean residence time after iv bolus administration, mean input time was 454

estimated by 0.5 times time of peak concentration assuming approximately zero-order 455

kinetics of drug input. 456

on April 1, 2018 by guest

http://aac.asm.org/

Dow

nloaded from

Page 22: Page 1 of 32 A New Semi-Physiological Absorption Model to Assess

Page 22 of 32

457

Table 2: Population PK parameters for 250 mg oral cefuroxime given as cefuroxime axetil 458

after a breakfast with a significant amount of fat 459

460

Parameter Symbol Unit Population mean (Standard error)

i

Median [10-90% percentile] of individual estimates

NONMEM S-ADAPT NPAG

Fixed effects

Apparent clearance CL/F L h-1 23.0 (8.1%)

21.7 [16.8-27.9]

21.7 (4.1%)

21.5 [17.1-27.8]

21.8 c

21.8 [17.0-28.3]

Apparent volume of distribution

V/F L 40.7 (6.3%)

39.7 [28.8-47.6]

38.7 (4.1%)

40.2 [30.4-45.8]

34.8

35.6 [25.5-48.0]

Maximum rate of release from stomach to intestine at time zero divided by dose

Vmax0 / Dose

1/h 0.381 (12%) a

0.308 [0.200-0.471]

0.505 (22%)a

0.418 [0.164-1.48]

1.53 a

1.35 [0.500-2.66]

Fraction of dose associated with 50% of Vmax0

Km / Dose 0.433% (70%) a

0.488% [0.180-3.09]

42.6% (39%) a

38.1% [4.70-478]

343% a

363% [193-498]

Fastest half-life of gastric release, if fraction of dose in stomach << Km/Dose

Ln(2) / Vmax0 ·

Km min 0.534 [0.270-6.36] 30.4 [13.0-120] 104 [49.3-155]

Absorption half-life from intestine to central comp.

Tabs f min 9.00 (29%)

14.2 [3.48-30.9]

9.34 (15%)

9.20 [4.70-16.2]

16.9 17.5 [4.00-32.2]

Time past meal at which Vmax changed by 50%

TC50 h 1.61 (14%)

2.06 [1.09-3.26]

1.61 (20%)

1.53 [0.619-5.05]

2.08

2.33 [1.14-5.31]

Maximum fractional change on transformed scale

Lg_Emax - -3.59 (8.9%)

-3.14 [-4.15- -1.79]

-0.762 (31%)

-0.937 [-1.96-0.641]

0.515 c

-0.50 [-3.44-4.95]

Maximum fractional change Emax - -0.582 [-0.845-0.426] 1.82 [0.242-5.55] 2.78 [-0.582-8.93]

Hill coefficient γ - 10 (fixed) g 10 (fixed) g 10 (fixed) g

Random effects b

BSV(CL/F) 0.202 b (18%) h 0.198 b (31%) h 0.193 e

BSV(V/F) 0.201 (16%) 0.183 (32%) 0.258

BSV(Vmax0 / Dose) 0.401 (22%) 0.900 (30%) 0.518

BSV(Km / Dose) 1.23 (37%) 1.80 (27%) 0.403

BSV(Tabs) 0.867 (23%) 0.600 (41%) 0.634

BSV(TC50) 0.536 (25%) 0.946 (31%) 0.580

BSV(Lg_Emax) 1.09 d (30%) 0.990 d (29%) 3.03 d

Proportional residual error 0.0754 (15%) 0.0851 (5.5%) 0.0620

Additive residual error mg/L 0.0065 (24%) 0.0029 (48%) 0.0062

461 462

on April 1, 2018 by guest

http://aac.asm.org/

Dow

nloaded from

Page 23: Page 1 of 32 A New Semi-Physiological Absorption Model to Assess

Page 23 of 32

a: Vmax0 and Km were estimated and are reported as a fraction of the cefuroxime dose. This assumes 463

that the rate of drug release from stomach is primarily determined by the meal and not by the amount 464

of cefuroxime axetil. This assumption yields a linear increase in peak concentrations and no change 465

of time to peak with dose which is consistent with literature data. 466 b: These estimates for random effects represent the apparent coefficient of variation for the between 467

subject variability. All variability estimates are reported as square roots of the estimated variance, 468

since this square root is an approximation of the coefficient of variation of a normal distribution on 469

natural logarithmic scale. 470 c: For NPAG, the medians of the support point matrix are provided. For Lg_Emax, the arithmetic mean 471

is reported, since the distribution of Lg_Emax was more symmetric and since this choice for 472

Lg_Emax yielded a better predictive performance in the visual predictive check. 473 d: The variability estimate for Lg_Emax is reported as the standard deviation on the transformed scale. 474 e: Estimates are coefficients of variation calculated based on the standard deviation and mean from the 475

support point matrix. 476 f: The corresponding rate constant kabs (unit 1/h) is calculated as ln(2) / (Tabs/60). 477 g: The Hill coefficient was initially estimated and estimates ranged between 10 and 15. To improve 478

model stability, the Hill coefficient was subsequently fixed to 10. This had no notable effect on the 479

curve fits and the objective function. 480 h: The value in brackets represents the uncertainty of the respective between subject variability 481

estimate. The value in brackets is the relative standard error of the estimated variance. 482 i: Standard errors are derived from a nonparametric bootstrap for NONMEM and are asymptotic 483

standard errors for S-ADAPT. 484

485

on April 1, 2018 by guest

http://aac.asm.org/

Dow

nloaded from

Page 24: Page 1 of 32 A New Semi-Physiological Absorption Model to Assess

Page 24 of 32

486

Table 3: Variance-covariance matrix for the final population PK model in S-ADAPT (see 487

Table 2 for parameter explanations) 488

489

CL/F V/F Vmax0/Dose Km/Dose Tabs TC50 Lg_Emax

CL/F 0.0393

V1/F 0.0326 0.0333

Vmax0/Dose 0.1143 0.0877 0.8100

Km/Dose 0.2693 0.2066 1.5925 3.2560

Tabs 0.0083 0.0142 -0.1036 -0.1640 0.3604

TC50 0.0931 0.0389 0.5324 1.0200 0.0459 0.8945

Lg_Emax 0.1530 0.1131 0.5702 1.1824 0.0414 0.8197 0.9793

490

491

on April 1, 2018 by guest

http://aac.asm.org/

Dow

nloaded from

Page 25: Page 1 of 32 A New Semi-Physiological Absorption Model to Assess

Page 25 of 32

492

Table 4: PTA expectation values for various cefuroxime axetil dosage regimens and PKPD 493

targets for the parametric Monte Carlo simulation based on S-ADAPT 494 495

Pathogen Region and year (no. of

isolates) PTA expectation value

PKPD Target fT>MIC ≥40% fT>MIC ≥65%

Dosage regimen 250 mg 250 mg 500 mg 250 mg 250 mg 500 mg

Q12h Q8h Q8h Q12h Q8h Q8h

S. pyogenes US & Canada 1997-99 (n=119) (26)

99.2% 99.2% 99.2% 98.9% 99.2% 99.2%

Germany 2002 (n=340) (7) 99.6% 99.8% 100% 99.3% 99.6% 99.9%

Europe 1997-1999 (n=662) (27) 98.9% 99.3% 99.6% 96.7% 98.9% 99.4%

S. pneumoniae US & Canada 1997-99 (n=417) (26)

65.9% 67.6% 70.2% 56.4% 65.6% 68.1%

UK 2002-2003 (n=519) (39) 94.4% 95.0% 96.2% 90.6% 94.2% 95.2%

Germany 2002 (n=331) (7) 97.3% 98.0% 98.2% 93.4% 97.0% 98.0%

Europe 1997-1999 (n=2018) (27) 71.5% 74.1% 77.8% 63.2% 70.8% 74.7%

EUCAST (n=18,869) (19) 100% 100% 100% 98.0% 100% 100%

PSSP US & Canada 1997-99 (n=249) (26)

97.8% 98.5% 99.3% 91.4% 97.6% 98.6%

Global 1997-2000 (n=2102) (6) 98.2% 99.0% 99.6% 92.7% 98.0% 99.2%

Europe 1997-1999 (n=1274) (27) 98.4% 99.1% 99.6% 93.5% 98.2% 99.2%

PISP US & Canada 1997-99 (n=70) (26)

44.5% 52.5% 65.1% 10.9% 43.4% 55.0%

Global 1997-2000 (n=1024) (6) 46.0% 53.9% 64.4% 19.9% 44.0% 55.7%

Europe 1997-1999 (n=458) (27) 39.9% 48.8% 60.8% 18.0% 37.8% 50.9%

H. influenzae US & Canada 1997-99 (n=300) (26)

18.5% 45.3% 83.8% 1.7% 16.0% 53.4%

Canada 2001-2002 (n=1350) (56) 19.3% 44.4% 81.6% 4.8% 17.3% 52.2%

UK 2002-2003 (n=581) (39) 41.1% 67.6% 88.8% 3.9% 33.7% 72.4%

Germany 2002 (n=300) (7) 72.4% 91.6% 98.1% 13.2% 64.7% 93.6%

EUCAST (n=66,947) (19) 41.5% 72.7% 99.1% 4.9% 34.9% 79.3%

M. catarrhalis US & Canada 1997-99 (n=231) (26)

26.2% 49.5% 82.5% 3.6% 22.7% 56.0%

UK 2002-2003 (n=269) (39) 35.1% 63.0% 93.0% 3.6% 30.0% 69.7%

Germany 2002 (n=308) (7) 11.8% 35.5% 77.4% 1.5% 10.5% 43.6% 496 PSSP: Penicillin susceptible S. pneumoniae, PISP: Penicillin intermediate S. pneumoniae. 497

on April 1, 2018 by guest

http://aac.asm.org/

Dow

nloaded from

Page 26: Page 1 of 32 A New Semi-Physiological Absorption Model to Assess

Page 26 of 32

Figure legends: 498

499

Figure 1: Structure of the final PK model. For the simulation of Vmax/Vmax0 profiles, 500

Emax values of 3 (long dashed line), 1 (continuous line) or -0.5 (dotted line), a 501

time of 50% change (TC50) of 2.5 h and a Hill coefficient of 10 were used. 502

503

Figure 2: Individual curve fits from NONMEM (dashed line), S-ADAPT (dotted line), and 504

NPAG (continuous line) overlaid on observations (markers) 505

506

Figure 3: Visual predictive check for a single oral dose of 250 mg cefuroxime as cefuroxime 507

axetil after a breakfast with a significant amount of fat for the final model in each 508

program. The continuous lines are the 10, 50, and 90% percentiles of the simulated 509

concentrations, the broken lines are the same percentiles of the observations, and 510

the markers show the observations. Ideally, the continuous and corresponding 511

broken lines should fall onto each other. The insets show the terminal phase on 512

semi-logarithmic scale. 513

514

Figure 4: Probability of target attainment vs. MIC profiles for the PKPD targets fT>MIC ≥ 515

40% (top) and fT>MIC ≥ 65% (bottom) for the nonparametric Monte Carlo 516

simulation based on NPAG (continuous lines) and the parametric Monte Carlo 517

simulation based on S-ADAPT (broken lines) 518

519

on April 1, 2018 by guest

http://aac.asm.org/

Dow

nloaded from

Page 27: Page 1 of 32 A New Semi-Physiological Absorption Model to Assess

Page 27 of 32

520 References 521

1. Alvarez-Sala, J. L., P. Kardos, J. Martinez-Beltran, P. Coronel, and L. Aguilar. 522

2006. Clinical and bacteriological efficacy in treatment of acute exacerbations of chronic 523

bronchitis with cefditoren-pivoxil versus cefuroxime-axetil. Antimicrob Agents 524

Chemother 50:1762-1767. 525

2. Ambrose, P. G., S. M. Bhavnani, R. N. Jones, W. A. Craig, and M. N. Dudley. 2004. 526

Presented at the 44th Interscience Conference on Antimicrobial Agents and 527

Chemotherapy, Washington, DC, October 30 to November 2, 2004. Use of 528

pharmacokinetic-pharmacodynamic and Monte-Carlo simulation as decision support for 529

the re-evaluation of NCCLS cephem susceptibility breakpoints for Enterobacteriaceae. 530

[abstract #A-138] 531

3. Ambrose, P. G., S. M. Bhavnani, C. M. Rubino, A. Louie, T. Gumbo, A. Forrest, and 532

G. L. Drusano. 2007. Pharmacokinetics-pharmacodynamics of antimicrobial therapy: it's 533

not just for mice anymore. Clin Infect Dis 44:79-86. 534

4. Bauer, R. J. 2007 S-ADAPT/MCPEM User’s Guide (Version 1.55). Software for 535

Pharmacokinetic, Pharmacodynamic and Population Data Analysis, Berkeley, CA. 536

5. Beal, S. L., L. B. Sheiner, and A. J. Boeckmann. 2006. NONMEM Users Guides 537

(1989-2006). Icon Development Solutions, Ellicott City, Maryland, USA. 538

6. Bouchillon, S. K., D. J. Hoban, J. L. Johnson, B. M. Johnson, D. L. Butler, K. A. 539

Saunders, L. A. Miller, and J. A. Poupard. 2004. In vitro activity of gemifloxacin and 540

contemporary oral antimicrobial agents against 27247 Gram-positive and Gram-negative 541

aerobic isolates: a global surveillance study. Int J Antimicrob Agents 23:181-196. 542

7. Brauers, J., S. Bagel, and M. Kresken. 2005. Aktuelle Resistanzsituation bei 543

bakteriellen Erregern von ambulant erworbenen Atemwegsinfektionen. Chemotherapie 544

Journal 14:45-53. 545

8. British Society for Antimicrobial Chemotherapy. 2008. BSAC Methods for 546

Antimicrobial Susceptibility Testing. Version 7.1, February 2008 547

9. Bulitta, J. B., S. B. Duffull, M. Kinzig-Schippers, U. Holzgrabe, U. Stephan, G. L. 548

Drusano, and F. Sorgel. 2007. Systematic comparison of the population 549

pharmacokinetics and pharmacodynamics of piperacillin in cystic fibrosis patients and 550

healthy volunteers. Antimicrob Agents Chemother 51:2497-2507. 551

on April 1, 2018 by guest

http://aac.asm.org/

Dow

nloaded from

Page 28: Page 1 of 32 A New Semi-Physiological Absorption Model to Assess

Page 28 of 32

10. Chodosh, S., J. McCarty, S. Farkas, M. Drehobl, R. Tosiello, M. Shan, L. Aneiro, 552

and S. Kowalsky. 1998. Randomized, double-blind study of ciprofloxacin and 553

cefuroxime axetil for treatment of acute bacterial exacerbations of chronic bronchitis. The 554

Bronchitis Study Group. Clin Infect Dis 27:722-729. 555

11. Clinical and Laboratory Standards Institute. 2007. Performance Standards for 556

Antimicrobial Susceptibility Testing: Seventeenth Informational Supplement, vol. 27. 557

Clinical and Laboratory Standards Institute, Wayne, PA, USA. 558

12. Craig, W. A. 2003. Basic pharmacodynamics of antibacterials with clinical applications 559

to the use of beta-lactams, glycopeptides, and linezolid. Infect Dis Clin North Am 17:479-560

501. 561

13. Craig, W. A. 1998. Choosing an antibiotic on the basis of pharmacodynamics. Ear Nose 562

Throat J 77:7-11; discussion 11-12. 563

14. Craig, W. A. 1998. Pharmacokinetic/pharmacodynamic parameters: rationale for 564

antibacterial dosing of mice and men. Clin Infect Dis 26:1-12. 565

15. Dagan, R., O. Abramson, E. Leibovitz, D. Greenberg, R. Lang, S. Goshen, P. 566

Yagupsky, A. Leiberman, and D. M. Fliss. 1997. Bacteriologic response to oral 567

cephalosporins: are established susceptibility breakpoints appropriate in the case of acute 568

otitis media? J Infect Dis 176:1253-1259. 569

16. De Abate, C. A., R. A. McIvor, P. McElvaine, and e. al. 1999. Smokers treated with 570

gatifloxacin had a high clinical cure rate: gatifloxacin vs cefuroxime axetil in patients 571

with acute exacerbations of chronic bronchitis. J Respir Dis 20 Suppl.:S23-29. 572

17. Deutsches Institut für Normung e. V. 2004. Medizinische Mikrobiologie - 573

Empfindlichkeitsprüfung von mikrobiellen Krankheitserregern gegen Chemotherapeutika 574

- Teil 4: Bewertungsstufen für die minimale Hemmkonzentration - MHK-Grenzwerte von 575

antibakteriellen Wirkstoffen. DIN 58940-4. Beuth Verlag, Berlin. 576

18. Drusano, G. L. 2004. Antimicrobial pharmacodynamics: critical interactions of 'bug and 577

drug'. Nat Rev Microbiol 2:289-300. 578

19. European Committee on Antimicrobial Susceptibility Testing (EUCAST). 2008. 579

Antimicrobial wild type distributions of microorganisms 580

(http://www.escmid.org/research_projects/eucast/, last accessed on December 20, 2008). 581

20. File, T. M., Jr., J. Segreti, L. Dunbar, R. Player, R. Kohler, R. R. Williams, C. 582

Kojak, and A. Rubin. 1997. A multicenter, randomized study comparing the efficacy 583

on April 1, 2018 by guest

http://aac.asm.org/

Dow

nloaded from

Page 29: Page 1 of 32 A New Semi-Physiological Absorption Model to Assess

Page 29 of 32

and safety of intravenous and/or oral levofloxacin versus ceftriaxone and/or cefuroxime 584

axetil in treatment of adults with community-acquired pneumonia. Antimicrob Agents 585

Chemother 41:1965-1972. 586

21. Finn, A., A. Straughn, M. Meyer, and J. Chubb. 1987. Effect of dose and food on the 587

bioavailability of cefuroxime axetil. Biopharm Drug Dispos 8:519-526. 588

22. Foord, R. D. 1976. Cefuroxime: human pharmacokinetics. Antimicrob Agents 589

Chemother 9:741-747. 590

23. Garraffo, R., H. B. Drugeon, and D. Chiche. 1997. Pharmacokinetics and 591

pharmacodynamics of two oral forms of cefuroxime axetil. Fundam Clin Pharmacol 592

11:90-95. 593

24. Hara, T., H. Saeki, A. Yamamoto, K. Ohta, T. Suzuki, M. Ara, M. Sugawara, Y. 594

Sano, H. Sakuma, and K. Satoh. 1988. Concentrations of cefuroxime in the skin. Jpn J 595

Antibiot 41:10-17. 596

25. Higaki, K., S. Y. Choe, R. Lobenberg, L. S. Welage, and G. L. Amidon. 2008. 597

Mechanistic understanding of time-dependent oral absorption based on gastric motor 598

activity in humans. Eur J Pharm Biopharm 70:313-325. 599

26. Hoban, D. J., S. K. Bouchillon, J. L. Johnson, G. G. Zhanel, D. L. Butler, K. A. 600

Saunders, L. A. Miller, and J. A. Poupard. 2003. Comparative in vitro potency of 601

amoxycillin-clavulanic acid and four oral agents against recent North American clinical 602

isolates from a global surveillance study. Int J Antimicrob Agents 21:425-433. 603

27. Hoban, D. J., S. K. Bouchillon, J. L. Johnson, G. G. Zhanel, D. L. Butler, K. A. 604

Saunders, L. A. Miller, and J. A. Poupard. 2003. Comparative in vitro surveillance of 605

amoxicillin-clavulanic acid and four oral comparators against 21232 clinical isolates from 606

europe. Eur J Clin Microbiol Infect Dis 22:261-267. 607

28. Ison, C. A., J. W. Mouton, K. Jones, K. A. Fenton, and D. M. Livermore. 2004. 608

Which cephalosporin for gonorrhoea? Sex Transm Infect 80:386-388. 609

29. Jacobs, M. R. 2003. How can we predict bacterial eradication? Int J Infect Dis 7 Suppl 610

1:S13-20. 611

30. Jacobs, M. R., S. Bajaksouzian, A. Zilles, G. Lin, G. A. Pankuch, and P. C. 612

Appelbaum. 1999. Susceptibilities of Streptococcus pneumoniae and Haemophilus 613

influenzae to 10 oral antimicrobial agents based on pharmacodynamic parameters: 1997 614

U.S. Surveillance study. Antimicrob Agents Chemother 43:1901-1908. 615

on April 1, 2018 by guest

http://aac.asm.org/

Dow

nloaded from

Page 30: Page 1 of 32 A New Semi-Physiological Absorption Model to Assess

Page 30 of 32

31. Jorgensen, A. F., J. Coolidge, P. A. Pedersen, K. P. Petersen, S. Waldorff, and E. 616

Widding. 1992. Amoxicillin in treatment of acute uncomplicated exacerbations of 617

chronic bronchitis. A double-blind, placebo-controlled multicentre study in general 618

practice. Scand J Prim Health Care 10:7-11. 619

32. Koeth, L. M., D. Felmingham, M. R. Jacobs, and F. Rossi. 2004. Antimicrobial 620

resistance of Streptococcus pneumoniae and Haemophilus influenzae in Sao Paulo, Brazil 621

from 1996 to 2000. Int J Antimicrob Agents 23:356-361. 622

33. Kong, F., and R. P. Singh. 2008. A model stomach system to investigate disintegration 623

kinetics of solid foods during gastric digestion. J Food Sci 73:E202-210. 624

34. Leary, R., R. W. Jelliffe, A. Schumitzky, and M. Van Guilder. 2001. An adaptive grid 625

non-parametric approach to pharmacokinetic and dynamic (PK/PD) models. IEEE 626

Computer Society., Bethesda, MD. 627

35. Lesmana, M., C. I. Lebron, D. Taslim, P. Tjaniadi, D. Subekti, M. O. Wasfy, J. R. 628

Campbell, and B. A. Oyofo. 2001. In vitro antibiotic susceptibility of Neisseria 629

gonorrhoeae in Jakarta, Indonesia. Antimicrob Agents Chemother 45:359-362. 630

36. Llanes, R., J. Sosa, D. Guzman, A. Llop, E. A. Valdes, I. Martinez, S. Palma, and M. 631

I. Lantero. 2003. Antimicrobial susceptibility of Neisseria gonorrhoeae in Cuba (1995-632

1999): implications for treatment of gonorrhea. Sex Transm Dis 30:10-14. 633

37. Mackay, J., A. E. Mackie, J. L. Palmer, A. Moult, and N. S. Baber. 1992. 634

Investigations into the mechanism for the improved oral systemic bioavailability of 635

cefuroxime from cefuroxime axetil when taken after food (Proceedings of the BPS, Sept 636

1991). British Journal of Clinical Pharmacology 33:226P-227P. 637

38. Mason, E. O., Jr., L. B. Lamberth, N. L. Kershaw, B. L. Prosser, A. Zoe, and P. G. 638

Ambrose. 2000. Streptococcus pneumoniae in the USA: in vitro susceptibility and 639

pharmacodynamic analysis. J Antimicrob Chemother 45:623-631. 640

39. Morrissey, I., M. Robbins, L. Viljoen, and D. F. Brown. 2005. Antimicrobial 641

susceptibility of community-acquired respiratory tract pathogens in the UK during 2002/3 642

determined locally and centrally by BSAC methods. J Antimicrob Chemother 55:200-643

208. 644

40. Mouton, J. W., M. N. Dudley, O. Cars, H. Derendorf, and G. L. Drusano. 2005. 645

Standardization of pharmacokinetic/pharmacodynamic (PK/PD) terminology for anti-646

infective drugs: an update. J Antimicrob Chemother 55:601-607. 647

on April 1, 2018 by guest

http://aac.asm.org/

Dow

nloaded from

Page 31: Page 1 of 32 A New Semi-Physiological Absorption Model to Assess

Page 31 of 32

41. Peric, M., F. A. Browne, M. R. Jacobs, and P. C. Appelbaum. 2003. Activity of nine 648

oral agents against gram-positive and gram-negative bacteria encountered in community-649

acquired infections: use of pharmacokinetic/pharmacodynamic breakpoints in the 650

comparative assessment of beta-lactam and macrolide antimicrobial agents. Clin Ther 651

25:169-177. 652

42. Perry, C. M., and R. N. Brogden. 1996. Cefuroxime axetil. A review of its antibacterial 653

activity, pharmacokinetic properties and therapeutic efficacy. Drugs 52:125-158. 654

43. Ruiz-Balaguer, N., A. Nacher, V. G. Casabo, and M. Merino. 1997. Nonlinear 655

intestinal absorption kinetics of cefuroxime axetil in rats. Antimicrob Agents Chemother 656

41:445-448. 657

44. Ruiz-Balaguer, N., A. Nacher, V. G. Casabo, and M. Merino Sanjuan. 2002. 658

Intestinal transport of cefuroxime axetil in rats: absorption and hydrolysis processes. Int J 659

Pharm 234:101-111. 660

45. Ruiz-Carretero, P., M. Merino-Sanjuan, A. Nacher, and V. G. Casabo. 2004. 661

Pharmacokinetic models for the saturable absorption of cefuroxime axetil and saturable 662

elimination of cefuroxime. Eur J Pharm Sci 21:217-223. 663

46. Scott, L. J., D. Ormrod, and K. L. Goa. 2001. Cefuroxime axetil: an updated review of 664

its use in the management of bacterial infections. Drugs 61:1455-1500. 665

47. Sethi, S., and T. F. Murphy. 2004. Acute exacerbations of chronic bronchitis: new 666

developments concerning microbiology and pathophysiology--impact on approaches to 667

risk stratification and therapy. Infect Dis Clin North Am 18:861-882, ix. 668

48. Shah, P. M., F. P. Maesen, A. Dolmann, N. Vetter, E. Fiss, and R. Wesch. 1999. 669

Levofloxacin versus cefuroxime axetil in the treatment of acute exacerbation of chronic 670

bronchitis: results of a randomized, double-blind study. J Antimicrob Chemother 43:529-671

539. 672

49. Stoeckel, K., W. L. Hayton, and D. J. Edwards. 1995. Clinical pharmacokinetics of 673

oral cephalosporins. Antibiot Chemother 47:34-71. 674

50. Upchurch, J., M. Rosemore, R. Tosiello, S. Kowalsky, and R. Echols. 2006. 675

Randomized double-blind study comparing 7- and 10-day regimens of faropenem 676

medoxomil with a 10-day cefuroxime axetil regimen for treatment of acute bacterial 677

sinusitis. Otolaryngol Head Neck Surg 135:511-517. 678

on April 1, 2018 by guest

http://aac.asm.org/

Dow

nloaded from

Page 32: Page 1 of 32 A New Semi-Physiological Absorption Model to Assess

Page 32 of 32

51. Vasu, S., C. Adithan, C. H. Shashindran, M. Asad, K. Koumaravelou, and I. Topno. 679

2000. Effect of two types of Indian breakfast on bioavailability of cefuroxime axetil. 680

Indian J Med Res 112:104-108. 681

52. Viberg, A., O. Cars, M. O. Karlsson, and S. Jonsson. 2008. Estimation of cefuroxime 682

dosage using pharmacodynamic targets, MIC distributions, and minimization of a risk 683

function. J Clin Pharmacol 48:1270-1281. 684

53. Viberg, A., A. Lannergard, A. Larsson, O. Cars, M. O. Karlsson, and M. Sandstrom. 685

2006. A population pharmacokinetic model for cefuroxime using cystatin C as a marker 686

of renal function. Br J Clin Pharmacol 62:297-303. 687

54. Weitschies, W., C. Friedrich, R. S. Wedemeyer, M. Schmidtmann, O. Kosch, M. 688

Kinzig, L. Trahms, F. Sorgel, W. Siegmund, S. Horkovics-Kovats, F. Schwarz, J. 689

Raneburger, and H. Monnikes. 2008. Bioavailability of amoxicillin and clavulanic acid 690

from extended release tablets depends on intragastric tablet deposition and gastric 691

emptying. Eur J Pharm Biopharm 70:641-648. 692

55. Williams, P. E., and S. M. Harding. 1984. The absolute bioavailability of oral 693

cefuroxime axetil in male and female volunteers after fasting and after food. J Antimicrob 694

Chemother 13:191-196. 695

56. Zhanel, G. G., L. Palatnick, K. A. Nichol, D. E. Low, and D. J. Hoban. 2003. 696

Antimicrobial resistance in Haemophilus influenzae and Moraxella catarrhalis respiratory 697

tract isolates: results of the Canadian Respiratory Organism Susceptibility Study, 1997 to 698

2002. Antimicrob Agents Chemother 47:1875-1881. 699

700

701

on April 1, 2018 by guest

http://aac.asm.org/

Dow

nloaded from

Page 33: Page 1 of 32 A New Semi-Physiological Absorption Model to Assess

on April 1, 2018 by guest

http://aac.asm.org/

Dow

nloaded from

Page 34: Page 1 of 32 A New Semi-Physiological Absorption Model to Assess

on April 1, 2018 by guest

http://aac.asm.org/

Dow

nloaded from

Page 35: Page 1 of 32 A New Semi-Physiological Absorption Model to Assess

on April 1, 2018 by guest

http://aac.asm.org/

Dow

nloaded from

Page 36: Page 1 of 32 A New Semi-Physiological Absorption Model to Assess

on April 1, 2018 by guest

http://aac.asm.org/

Dow

nloaded from