Introduction to Population Pharmacokinetic / Pharmacodynamic Analysis with Nonlinear Mixed Effects Models E-Book

Joel S. Owen

• Herausgeber: John Wiley & Sons
• Kategorie: Wissenschaft und neue Technologien
• Sprache: Englisch

Seitenzahl: 586

Veröffentlichungsjahr: 2014

Beschreibung

This book provides a user-friendly, hands-on introduction to the Nonlinear Mixed Effects Modeling (NONMEM) system, the most powerful tool for pharmacokinetic / pharmacodynamic analysis. * Introduces requisite background to using Nonlinear Mixed Effects Modeling (NONMEM), covering data requirements, model building and evaluation, and quality control aspects * Provides examples of nonlinear modeling concepts and estimation basics with discussion on the model building process and applications of empirical Bayesian estimates in the drug development environment * Includes detailed chapters on data set structure, developing control streams for modeling and simulation, model applications, interpretation of NONMEM output and results, and quality control * Has datasets, programming code, and practice exercises with solutions, available on a supplementary website

Leseprobe

CONTENTS

Cover

Title page

Copyright page

Dedication

Preface

Acknowledgments

References

Chapter 1: The Practice of Pharmacometrics

1.1 Introduction

1.2 Applications of Sparse Data Analysis

1.3 Impact of Pharmacometrics

1.4 Clinical Example

References

Chapter 2: Population Model Concepts and Terminology

2.1 Introduction

2.2 Model Elements

2.3 Individual Subject Models

2.4 Population Models

2.5 Models of Random Between-Subject Variability (L1)

2.6 Models of Random Variability in Observations (L2)

2.7 Estimation Methods

2.8 Objective Function

2.9 Bayesian Estimation

References

Chapter 3: NONMEM Overview and Writing an NM-TRAN Control Stream

3.1 Introduction

3.2 Components of the NONMEM System

3.3 General Rules

3.4 Required Control Stream Components

3.5 Specifying the Model in NM-TRAN

3.6 Specifying Initial Estimates with $THETA, $OMEGA, and $SIGMA

3.7 Requesting Estimation and Related Options

3.8 Requesting Estimates of the Precision of Parameter Estimates

3.9 Controlling the Output

References

Chapter 4: Datasets

4.1 Introduction

4.2 Arrangement of the Dataset

4.3 Variables of the Dataset

4.4 Constructing Datasets with Flexibility to Apply Alternate Models

4.5 Examples of Event Records

4.6 Beyond Doses and Observations

References

Chapter 5: Model Building

5.1 Introduction

5.2 Analysis Planning

5.3 Analysis Dataset Creation

5.4 Dataset Quality Control

5.5 Exploratory Data Analysis

5.6 Base Model Development

5.7 Covariate Evaluation

5.8 Model Refinement

Appendix 5.1 Recommended Analysis Plan Content

References

Chapter 6: Interpreting the NONMEM Output

6.1 Introduction

6.2 Description of the Output Files

6.3 The NONMEM Report File

6.4 Error Messages: Interpretation and Resolution

6.5 General Suggestions for Diagnosing Problems

Appendix 6.1 Mixture Model Example Code

References

Chapter 7: Applications Using Parameter Estimates from the Individual

7.1 Introduction

7.2 Bayes Theorem and Individual Parameter Estimates

7.3 Obtaining Individual Parameter Estimates

7.4 Applications of Individual Parameter Estimates

References

Chapter 8: Introduction to Model Evaluation

8.1 Introduction

8.2 Internal Validation

8.3 External Validation

8.4 Predictive Performance Assessment

8.5 Objective Function Mapping

8.6 Leverage Analysis

8.7 Bootstrap Procedures

8.8 Visual and Numerical Predictive Check Procedures

8.9 Posterior Predictive Check Procedures

References

Chapter 9: User-Written Models

9.1 Introduction

9.2 $MODEL

9.3 $SUBROUTINES

9.4 A Series of Examples

References

Chapter 10: PK/PD Models

10.1 Introduction

10.2 Implementation of PD Models in NONMEM

10.3 $PRED

10.4 $PK

10.5 Odd-Type Data: Analysis of Noncontinuous Data

10.6 PD Model Complexity

10.7 Communication of Results

References

Chapter 11: Simulation Basics

11.1 Introduction

11.2 The Simulation Plan

11.3 Miscellaneous Other Simulation-Related Considerations

Appendix 11.1 Example Simulation Output

Appendix 11.2 Example Simulation Output with a Different Seed Value

Appendix 11.3 Simulation with Parameter Uncertainty

References

Chapter 12: Quality Control

12.1 Introduction

12.2 QC of the Data Analysis Plan

12.3 Analysis Dataset Creation

12.4 QC of Model Development

12.5 Documentation of QC Efforts

12.6 Summary

References

Index

End User License Agreement

List of Tables

Chapter 02

Table 2.1 Commonly used Greek characters in population PK

Table 2.2 Individual parameters of a population PK model for

n

subjects

Table 2.3 Presentation of residual error (%CV) for an additive plus CCV model

Chapter 03

Table 3.1 Desired inclusions and exclusions of data, based on study number and disease severity

Chapter 04

Table 4.1 Example of a data record

Table 4.2 Example of data records in unlike units that lead to an error in analysis

Table 4.3 Example of incorrect expression of simultaneous events in the analysis dataset

Table 4.4 Example of correct expression of simultaneous events in the analysis dataset

Table 4.5 Date formats for input of calendar dates

Table 4.6 Example of correct use of TIME and calendar date

Table 4.7 Example of incorrect use of TIME and calendar date

Table 4.8 Default dose and observation compartment assignments for the specific ADVAN routines

Table 4.9 General requirements for DV and AMT variable values on the various types of EVID records

Table 4.10 Example of a single dosing record describing multiple additional doses

Table 4.11 Expanded table of the individual dosing records specified by the ADDL example given before

Table 4.12 Example of dosing records for repeated, irregular dosing patterns

Table 4.13 Example records demonstrating alternative expressions for specifying the value of time

Table 4.14 Example records demonstrating alternative methods to define a zero-order input dosing record

Table 4.15 Example of dosing and observation records in a NONMEM dataset

Table 4.16 Example of an SS dosing scenario

Table 4.17 Example of an SS dosing scenario with observations made before and after achieving SS conditions

Table 4.18 Example of an SS dosing scenario with an additional dose

Table 4.19 Example of SS dosing with an irregular dosing interval

Table 4.20 Example of dosing from multiple routes of administration

Table 4.21 Example of records describing a dose followed by simultaneous PK and PD observations

Table 4.22 Example data records for a $PRED model

Table 4.23 Example records demonstrating a changing covariate over time

Chapter 05

Table 5.1 Summary statistics of patient characteristics by study

Table 5.2 Summary statistics of patient characteristics by dose group

Table 5.3 Summary of select NONMEM dose-related data items

Table 5.4 Summary of BLQ data by dose and sampling timepoint

Table 5.5 Planned covariates to be evaluated during covariate assessment, and on which parameter each is to be evaluated

Table 5.6 Summary of a given round of the forward selection process

Table 5.7 Summary of the first round of backward elimination

Chapter 07

Table 7.1 Example dataset records for one subject

Table 7.2 Example table file output records, including AUC, for one subject

Table 7.3 Example table file output records with individual parameters, subset to one record per subject

Table 7.4 Example of template records for simulation of the administration of five doses and prediction of concentration values from 96 to 120 hours

Table 7.5 Example of template records following the

one-to-many

merge to include individual parameters and sufficient records for prediction of concentration values from 96 to 120 hours

Chapter 08

Table 8.1 Example table comparing bootstrap results to the final parameter estimates

Chapter 09

Table 9.1 Example dosing records for simultaneous bolus and zero-order input: fixed amounts through each route

Table 9.2 Example dosing records for simultaneous bolus and zero-order input: estimated amounts through each route

Table 9.3 Example dosing records for parallel absorption processes, with estimated amounts absorbed through each route

Table 9.4 Example dosing and observation records for simultaneous modeling of parent and metabolite concentrations

Chapter 10

Table 10.1 Three options for estimation of the PD models

Table 10.2 Example of data for use with a model in $PRED

Table 10.3 Examples of noncontinuous data types

List of Illustrations

Chapter 01

Figure 1.1 Propofol clearance (L/min) versus TBW (kg) in morbidly obese children and adolescents (

r

 = 0.721) (Adapted from Diepstraten, et al. ADIS. Clin Pharmacokinet 2012; 51(8):547, Figure 1. Adis 2012 Springer International Publishing AG. All rights reserved. Used with kind permission from Springer Science + Business Media B.V.).

Chapter 02

Figure 2.1 One-compartment model with first-order absorption and first-order elimination.

Figure 2.2 Example of a random-variable distribution.

Figure 2.3 Concentration independence of the additive residual error model. The error lines are equidistant from the median at all values of time, though the curvature makes them appear to be smaller at higher concentrations.

Figure 2.4 Concentration dependence of the CCV residual error model.

Figure 2.5 Residual error for the additive plus CCV model.

Figure 2.6 Relationship of sigma, the SD of the residual variability, with predicted concentration for various residual error models.

Figure 2.7 Relationship of residual variance (%CV) with predicted concentration for various residual error models.

Figure 2.8 Illustration of the impact of the magnitude of residual variability.

Chapter 03

Figure 3.1 Schematic illustrating the components of the NONMEM system.

Figure 3.2 Representative eta biplot: scatterplot of the ETA on clearance (CL) versus the ETA on volume (

V

2).

Chapter 05

Figure 5.1 Typical pharmacometric model building process.

Figure 5.2 Frequency distributions of patient characteristics. Panel (a) illustrates the distribution of weight and Panel (b) provides the distribution of age in the population; the number below each bar represents the midpoint of the values in that bar.

Figure 5.3 Boxplots of continuous patient characteristics versus categorical characteristics. Panel (a) illustrates the distribution of weight by gender. Panel (b) provides the distribution of weight by race group. Panel (c) illustrates the distribution of age by gender. Panel (d) provides the distribution of age by race group in the population; the box spans the interquartile range for the values in the subpopulation, the whiskers extend to the 5th and 95th percentiles of the distributions, with the asterisks representing values outside this range, and the boxes are joined at the median values.

Figure 5.4 Scatterplot of the relationship between patient age and weight.

Figure 5.5 Frequency distribution of dose amount.

Figure 5.6 Frequency distribution of TSLD (time of sampling relative to the most recent previous dose administered). Summary statistics are also provided to quantify the frequency of sampling in various time bins or ranges of TSLD.

Figure 5.7 Scatterplot of drug concentration versus TSLD. A semi-log scale is used for the

y

axis and time of sampling is normalized to time relative to the most recent dose administration.

Figure 5.8 Scatterplot of trough concentrations versus time since the start of treatment.

Figure 5.9 Scatterplot of trough concentrations versus time since the start of treatment with a smoothing spline through the data. A smoothing spline is added to the scatterplot to illustrate the central tendency of the trough concentrations over time.

Figure 5.10 Scatterplot of drug concentrations versus TSLD with different symbols used to represent the different dose groups.

Figure 5.11 Scatterplot of drug concentrations versus TSLD with different panels used to compare the data following different dose levels. Solid triangles represent data following the administration of the 50-mg dose, asterisks represent data following the administration of the 75-mg dose, and solid dots represent data following the administration of the 100-mg dose.

Figure 5.12 Scatterplot of dose-normalized drug concentrations versus TSLD with different symbols used to represent the different dose groups.

Figure 5.13 Scatterplot of dose-normalized drug concentrations versus TSLD with different symbols used to represent the different dose groups and smoothing splines added for each dose group.

Figure 5.14 Concentration versus time profiles for a sample of patients with full-profile sampling. A semi-log scale is used for the

y

axis.

Figure 5.15 Scatterplot of trough concentrations versus time since the start of treatment, joined by subject.

Figure 5.16 Scatterplot of drug concentrations (on a linear scale) versus TSLD with different panels used to compare the data following different dose levels. Solid triangles represent data following the administration of the 50-mg dose, asterisks represent data following the administration of the 75-mg dose, and solid dots represent data following the administration of the 100-mg dose.

Figure 5.17 Concentration (on a linear scale) versus time profiles for a sample of patients with full-profile sampling.

Figure 5.18 Representative concentration versus time profiles for various disposition patterns.

Figure 5.19 Scatterplot of drug concentrations versus TSLD with different symbols used to represent the different renal function categories.

Figure 5.20 Scatterplot of drug concentrations versus TSLD with different symbols used to represent the different renal function categories and smoothing splines added for each renal function category.

Figure 5.21 Scatterplot of drug concentrations versus TSLD with different panels used to compare the data from the different renal function categories. Solid triangles represent data from subjects with moderate renal impairment, asterisks represent data from subjects with mild renal impairment, and solid dots represent data from subjects with normal renal function.

Figure 5.22 Scatterplot of an endpoint versus a predictor variable for a large dataset using open symbols.

Figure 5.23 Scatterplot of an endpoint versus a predictor variable for a large dataset using open symbols and a smoothing spline added.

Figure 5.24 Scatterplot of an endpoint versus a predictor variable, plotted separately by study with smoothing splines. Panel (a) represents the endpoint versus predictor data for the subjects from Study A. Panel (b) is for Study B. Panel (c) is the data from Study C. Panel (d) is the data from Study D.

Figure 5.25 Scatterplot of observed concentrations versus population predicted values illustrating a reasonably good fit. Note that the axes are scaled to the range of the data.

Figure 5.26 Scatterplot of observed concentrations versus population predicted values with equal axes.

Figure 5.27 Scatterplot of weighted residuals versus population predicted illustrating a good fit. Panel (a) includes no apparent outlier points, while Panel (b) includes two apparent outliers associated with high weighted residuals.

Figure 5.28 Scatterplot of weighted residuals versus population predicted values illustrating a poorly fitting model.

Figure 5.29 Scatterplot of weighted residuals versus population predictions illustrating a U-shaped pattern, indicative of the need for an additional compartment in the model.

Figure 5.30 Scatterplot of weighted residuals versus population predictions illustrating the need for log-transformation.

Figure 5.31 Scatterplot of residuals versus population predictions illustrating the fan-shaped pattern indicative of the proportional residual variability model.

Figure 5.32 Scatterplot of residuals versus population predictions illustrating a constant spread across the range of predictions indicative of an additive error model.

Figure 5.33 Scatterplot of weighted residuals versus time since the previous dose illustrating a good fit.

Figure 5.34 Scatterplot of weighted residuals versus TSFD illustrating patterns indicative of a biased fit over the time course of treatment.

Figure 5.35 Scatterplot of conditional weighted residuals versus population predictions illustrating a good fit.

Figure 5.36 Scatterplot of conditional weighted residuals versus TSFD illustrating a good fit.

Figure 5.37 Typical value concentration versus time profile overlaid on the raw concentration versus time since previous dose data.

Figure 5.38 Typical value concentration versus time profiles for a one- and two-compartment model fit overlaid on the raw concentration versus time since previous dose data. The solid line indicates the typical profile associated with the one-compartment model and the dashed line indicates the typical profile associated with the two-compartment model.

Figure 5.39 Typical value concentration versus time profiles for hypothetical subjects representing two different patient types overlaid on the raw concentration versus time since previous dose data.

Figure 5.40 Scatterplot of observed concentrations versus population predicted values illustrating a limited number of predicted values based on the design and the model.

Figure 5.41 Scatterplot of observed concentrations versus individual predicted values illustrating a good fit.

Figure 5.42 Scatterplot of the absolute value of the individual weighted residuals versus the individual predictions, indicating no lack of fit.

Figure 5.43 Scatterplot of the absolute value of the individual weighted residuals versus the individual predictions, indicating an increasing spread with increasing individual predictions. This pattern may result from the inappropriate use of an additive RV model when a CCV model may be warranted.

Figure 5.44 Scatterplot of the absolute value of the individual weighted residuals versus the individual predictions, indicating a decreasing spread with increasing individual predictions. This pattern may result from the use of a proportional or CCV RV model when an additive model would be more appropriate.

Figure 5.45 Scatterplots of the EBEs of clearance versus continuous covariates. Panel (a) shows the plot of clearance versus weight, Panel (b) shows the plot of clearance versus age, and Panel (c) shows the plot of clearance versus creatinine clearance.

Figure 5.46 Scatterplots of the EBEs of volume versus continuous covariates. Panel (a) shows the plot of volume versus weight, Panel (b) shows the plot of volume versus age, and Panel (c) shows the plot of volume versus creatinine clearance.

Figure 5.47 Boxplots of the EBEs of clearance versus categorical covariates. Panel (a) shows the plot of clearance versus gender and Panel (b) shows the plot of clearance versus race. The box represents the 25th to 75th percentiles of the data, with the whiskers extending to the 5th and 95th percentiles of the data and the asterisks representing values outside this range; the boxes are joined at the median values.

Figure 5.48 Boxplots of the EBEs of volume versus categorical covariates. Panel (a) shows the plot of volume versus gender and Panel (b) shows the plot of volume versus race. The box represents the 25th to 75th percentiles of the data, with the whiskers extending to the 5th and 95th percentiles of the data and the asterisks representing values outside this range; the boxes are joined at the median values.

Figure 5.49 Lineplot illustrating representative linear models for the relationship between apparent oral clearance and a covariate.

Figure 5.50 Lineplots illustrating representative power models with varying exponents for the relationship between apparent oral clearance and a covariate.

Figure 5.51 Lineplots illustrating the natural log transformation of representative power models for the relationship between apparent oral clearance and a covariate.

Figure 5.52 Lineplots illustrating representative exponential models for the relationship between apparent oral clearance and a covariate.

Figure 5.53 Lineplots illustrating the natural log transformation of representative exponential models for the relationship between apparent oral clearance and a covariate.

Figure 5.54 Lineplots illustrating representative piece-wise linear models for the relationship between apparent oral clearance and a covariate.

Figure 5.55 Scatterplots illustrating the relationship between apparent oral clearance and creatinine clearance. Panel (a) shows the plot of drug clearance versus (raw) creatinine clearance and Panel (b) shows the plot of drug clearance versus centered creatinine clearance.

Figure 5.56 Scatterplot illustrating the relationship between apparent oral clearance and creatinine clearance with the fitted line and the confidence band about the fitted line.

Figure 5.57 Scatterplots of delta-clearance versus continuous covariates. Panel (a) shows the plot of delta-clearance versus weight, Panel (b) shows the plot of delta-clearance versus age, and Panel (c) shows the plot of delta-clearance versus creatinine clearance.

Figure 5.58 Scatterplots of delta-volume versus continuous covariates. Panel (a) shows the plot of delta-volume versus weight, Panel (b) shows the plot of delta-volume versus age, and Panel (c) shows the plot of delta-volume versus creatinine clearance.

Figure 5.59 Boxplots of delta-clearance versus categorical covariates. Panel (a) shows the plot of delta-clearance versus gender and Panel (b) shows the plot of delta-clearance versus race. The box represents the 25th to 75th percentiles of the data, with the whiskers extending to the 5th and 95th percentiles of the data and the asterisks representing values outside this range; the boxes are joined at the median values.

Figure 5.60 Boxplots of delta-volume versus categorical covariates. Panel (a) shows the plot of delta-volume versus gender and Panel (b) shows the plot of delta-volume versus race. The box represents the 25th to 75th percentiles of the data, with the whiskers extending to the 5th and 95th percentiles of the data and the asterisks representing values outside this range; the boxes are joined at the median values.

Figure 5.61 Bivariate scatterplots of random-effect terms (eta biplots). Panel (a) shows the plot of eta

CL

versus , with no apprent trend, Panel (b) shows the plot of eta

V

2

versus , with no apparent trend, and Panel (c) shows the plot of eta

V

2

versus eta

CL

, with a positive relationship evident.

Figure 5.62 Histograms of random-effect terms. Panel (a) shows the frequency distribution of eta

CL

and Panel (b) shows the frequency distribution of eta

V

with a normal kernel density estimate overlaid on each plot.

Figure 5.63 Distribution of weighted residuals from the proposed final model. Panel (a) shows the histogram of conditional weighted residuals and Panel (b) shows the corresponding Q–Q plot of weighted residuals.

Chapter 07

Figure 7.1 Overview of Bayesian parameter estimation in a typical PK project.

Chapter 08

Figure 8.1 Example plot illustrating objective function mapping for the apparent volume of distribution.

Figure 8.2 Example plot illustrating objective function mapping for the absorption rate constant, where an apparent local minimum was obtained.

Figure 8.3 Example cross-validation scatterplots for the apparent volume of distribution and the absorption rate constant. Panel (a) illustrates a successful cross-validation result for volume where all estimates fall within the CI and Panel (b) illustrates a cross-validation exercise, which resulted in two subsets whose parameter estimates fall outside the CI.

Figure 8.4 Example VPC plot illustrating the simulation-based prediction interval overlaid on the raw data, the corresponding percentiles based on the observed data, and the CI about the median and 90% prediction interval bounds as shaded regions. Medians and percentiles are plotted at the mean time since dose of the data observed within each time since dose interval.

Figure 8.5 Example VPC plot illustrating the CI about the median and 90% prediction interval bounds as shaded regions with the corresponding percentiles based on the observed data overlaid. Medians and percentiles are plotted at the mean time since dose of the data observed within each time since dose interval.

Figure 8.6 Example posterior predictive check plot illustrating the distribution of simulated estimates of the (dose-normalized) maximum concentration with the mean observed maximum concentration (vertical line) overlaid.

Chapter 09

Figure 9.1 Simultaneous absorption model.

Figure 9.2 Illustration of the partitioning of the administered dose.

Figure 9.3 Two-compartment model with parallel zero-order and first-order drug absorption.

Figure 9.4 Catenary compartment model with first-order absorption.

Figure 9.5 One-compartment model with first-order and zero-order absorption, dose fractions known and complete bioavailability.

Figure 9.6 Parent and metabolite model.

Chapter 10

Figure 10.1 A general depiction of the PK/PD indirect response model. The black rectangles indicate the inhibition of either input or output rates of the response and the white rectangles indicate the stimulation of the rate of change in response.

Guide

Cover

Table of Contents

Begin Reading

Pages

iii

iv

v

xiii

xiv

xv

xvi

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

32

33

34

35

36

38

39

40

41

42

43

44

45

47

48

49

50

51

52

53

54

55

56

57

58

59

60

61

62

63

64

65

66

67

68

69

70

71

72

73

74

75

76

77

78

79

80

81

82

83

84

85

86

87

88

89

90

91

92

93

94

95

96

97

98

99

100

101

102

103

104

105

106

107

108

109

110

111

112

113

114

115

116

117

118

119

120

121

122

123

124

125

126

127

128

129

130

131

132

133

134

135

136

137

138

139

140

141

142

143

144

145

146

147

148

149

150

151

152

153

154

155

156

157

158

159

160

161

162

163

164

165

166

167

168

169

170

171

172

173

174

175

176

177

178

179

180

181

182

183

184

185

186

187

188

189

190

191

192

193

194

195

196

197

198

199

200

201

202

203

204

205

206

207

208

209

210

211

212

213

214

215

216

217

218

219

220

221

222

223

224

225

226

227

228

229

230

231

232

233

234

235

236

237

238

239

240

241

242

243

244

245

246

247

248

249

250

251

252

253

254

255

256

257

258

259

260

261

262

263

264

265

266

267

268

269

270

271

272

273

274

275

276

277

278

279

280

281

282

283

284

285

286

287

288

289

290

291

292

293

294

295

296

297

298

299

300

301

Introduction to Population Pharmacokinetic / Pharmacodynamic Analysis with Nonlinear Mixed Effects Models

Joel S. Owen

Professor of PharmaceuticsUnion University, School of PharmacyDepartment of Pharmaceutical SciencesJackson, Tennessee

Jill Fiedler-Kelly

Vice President, Pharmacometric Services & CSOCognigen Corporation

Adjunct Associate ProfessorUniversity at Buffalo, Department of Pharmaceutical SciencesBuffalo, New York

Copyright © 2014 by John Wiley & Sons, Inc. All rights reserved.

Published by John Wiley & Sons, Inc., Hoboken, New Jersey.Published simultaneously in Canada.

No part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, scanning, or otherwise, except as permitted under Section 107 or 108 of the 1976 United States Copyright Act, without either the prior written permission of the Publisher, or authorization through payment of the appropriate per-copy fee to the Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, [phone], fax [phone], or on the web at www.copyright.com. Requests to the Publisher for permission should be addressed to the Permissions Department, John Wiley & Sons, Inc., 111 River Street, Hoboken, NJ 07030, [phone], fax [phone], or online at http://www.wiley.com/go/permission.

Limit of Liability/Disclaimer of Warranty: While the publisher and author have used their best efforts in preparing this book, they make no representations or warranties with respect to the accuracy or completeness of the contents of this book and specifically disclaim any implied warranties of merchantability or fitness for a particular purpose. No warranty may be created or extended by sales representatives or written sales materials. The advice and strategies contained herein may not be suitable for your situation. You should consult with a professional where appropriate. Neither the publisher nor author shall be liable for any loss of profit or any other commercial damages, including but not limited to special, incidental, consequential, or other damages.

For general information on our other products and services or for technical support, please contact our Customer Care Department within the United States at [phone], outside the United States at [phone] or fax [phone].

Wiley also publishes its books in a variety of electronic formats and by print-on-demand. Not all content that is available in standard print versions of this book may appear or be packaged in all book formats. If you have purchased a version of this book that did not include media that is referenced by or accompanies a standard print version, you may request this media by visiting http://booksupport.wiley.com. For more information about Wiley products, visit us at www.wiley.com.

Library of Congress Cataloging-in-Publication Data:

Owen, Joel S., author. Introduction to population pharmacokinetic/pharmacodynamic analysis with nonlinear mixed effects models / Joel S. Owen, Jill Fiedler-Kelly.  p. ; cm. Includes bibliographical references and index. ISBN 978-0-470-58229-9 (cloth)I. Fiedler-Kelly, Jill, author. II. Title.  [DNLM: 1. Pharmacokinetics. 2. Nonlinear Dynamics. 3. Software. QV 38] RM301.5 615.7–dc23

         2014002869

To Melanie, Janie, Matthew, Annette, and Mark for their love and patience. To numerous friends and colleagues at Union University and First Baptist Church, Jackson, Tennessee, for their interest and encouragement.

To Tad, Brynna, and Finn for their love and understanding and without whose support this would not have been possible.

Preface

Nonlinear mixed effects modeling is an analytical approach capable of the efficient use of data to support informed decision making. In drug development and in clinical practice, the culmination of the work of the pharmacometrician is making informed decisions regarding the treatment of patients, either during the development of new treatments or in the clinical care of specific patients. The application of nonlinear mixed effects models to sparse clinical data was a concept brilliantly conceived and implemented via the NONMEM system by Dr. Lewis B. Sheiner and Dr. Stuart L. Beal, supported over the years by Alison J. Boeckmann. The impetus for development of the analytical approach and supporting software was the need to study drugs in the population in which they were to be used, and realization of the direct correlation between the patients who are most vulnerable to pharmacokinetic differences and those who are most difficult to study (Sheiner and Benet 1985). Of course, there have been many additional contributors to the development of this field over the past few decades; their contributions have created an environment where pharmacometric analysis approaches now flourish and contribute significantly to informed decision making.