|UFDC Home||myUFDC Home | Help|
This item has the following downloads:
1 DOSE OPTIMIZATION AND BODY SIZE By RONG SHI A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF F LORIDA 2011
2 2011 Rong Shi
3 To my parents and my husband
4 ACKNOWLEDGMENTS I would like to first express my gratitude and great appreciation to my academic advisor, Dr. Hartmut Derendorf. It has been such a great honor to be trained by Dr Derendorf during my Ph.D. study I am deeply thankful to Dr Derendorf who accept ed me to his research group, guided, supported, inspired and encouraged me in my research in all respect. I truly believe the training with Dr. Derendorf will be a strong fo undation for my future career and will benefit me a lifetime. S pecial thanks also go to the membe rs of my supervisory committee, Dr. Anthony Palmieri, Dr. William Cary Mobley Dr Christoph N. Seubert, and Dr. M aria Grant for t heir valuable advice througho ut my doctoral research. I would like to specially thank Dr. Anthony Palmieri for his support during the stressful time in my Ph.D. program An extended special thanks go to the postdoctoral fellows Dr. Sreedharan N. Sabarinath Dr. Rajendra Pratap Singh and Dr. Michael Bewernitz and all my labmates I would also like to take this opportunity to express my thanks to Yufei Tang, for teaching and helping me with the analytical techniques, and for encouraging me over the years. I would like to thank Dr. Sihon g Song for his advice on teaching and my future career. I am also thankful for the technical and administrative support by Mrs. Patricia J. Khan, Ms. Robin Keirnan Sanchez and Mrs. Sarah L Scheckner Special thanks go to all my friends. My time during the internship in Clinical Pharmacology at Pfizer was very rewarding, and I would like to extend my appreciation to Dr. Diane Wang, Dr. Shuzhong Zhang, Dr. Chunze Li and Dr. Danny Chen for their suggestions, advice and support. I would like to thank all my C hinese friends for their help and support over the years.
5 Last but not least I deeply thank my parents, and my dear and loving husband Gary Nool whose love and unconditional support gave me the strength to complete this dissertation.
6 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ .. 4 LIST OF TABLES ................................ ................................ ................................ ............ 8 LIST OF FIGURES ................................ ................................ ................................ .......... 9 ABSTRACT ................................ ................................ ................................ ................... 10 CHAPTER 1 INTRODUCTION ................................ ................................ ................................ .... 12 Overview of Population Pharmacokinetics and Dose Optimization ......................... 12 Population Pharmacokinetic Approach ................................ ............................. 16 Population Pharmacokinetics in Dose Optimization during Drug Development ................................ ................................ ................................ 22 Body Size Measurement and Body Size Effect on Pharmacokinetics .................... 25 Measurements of Body Size ................................ ................................ ............. 27 Body Size E ffect on Pharmacokinetic parameters ................................ ............ 31 2 FIXED DOSING VERSUS BODY SIZE BASED DOSING OF THERAPEUTIC PEPTIDES AND PROTEINS IN ADULTS ................................ ............................... 40 Background ................................ ................................ ................................ ............. 40 Methods ................................ ................................ ................................ .................. 43 Data Collection ................................ ................................ ................................ 43 Population PK and PD Models ................................ ................................ ......... 43 PK and/or PD Simulation ................................ ................................ .................. 44 Calculation of Values ................................ ................................ ..................... 45 Performance Evalu ation ................................ ................................ ................... 46 Contribution of Body Size Effect to Overall Intersubject Variability of Relevant PK Parameters ................................ ................................ ............... 47 Results and Discussion ................................ ................................ ........................... 48 Data Collection ................................ ................................ ................................ 48 Dosing Approach Performance ................................ ................................ ........ 49 Performance evaluation b ased on AUC ................................ ..................... 49 Performance evaluation based on C max ................................ ..................... 50 Performance evaluation based on Pharmacodymanics ............................. 51 Contribution of Body Size Effect to Overall Intersubject Variability in PK Parameters ................................ ................................ ................................ .... 52 Relationship between the Type of Therapeutic Biologics and Body Size Eff ect on Pharmacokinetics ................................ ................................ ........... 53 Summary ................................ ................................ ................................ ................ 53
7 3 PEDIATRIC DOSING AND BODY SIZE IN BIOTHERAPEUTICS .......................... 63 Background ................................ ................................ ................................ ............. 63 General Pharmacokinetics in Pediatrics ................................ ........................... 64 Pharmacokinetics of Proteins and Peptides ................................ ..................... 66 Distribution ................................ ................................ ................................ 67 Elimination ................................ ................................ ................................ 68 Results and Discussion ................................ ................................ ........................... 69 Monoclonal Antibodies (mAbs) ................................ ................................ ......... 69 Growth Factors ................................ ................................ ................................ 76 Blood Factors ................................ ................................ ................................ ... 81 Hormone ................................ ................................ ................................ ........... 85 Other Proteins and Peptides ................................ ................................ ............ 88 Summary ................................ ................................ ................................ ................ 93 4 FIXED DOSING VERSUS BODY SIZE BASED DOSING OF THERAPEUTIC ANTICANCER DRUGS IN ADULTS ................................ ................................ ..... 100 Background ................................ ................................ ................................ ........... 100 Methods ................................ ................................ ................................ ................ 103 Data Collection ................................ ................................ ............................... 103 Population PK Models ................................ ................................ .................... 104 PK Simulation ................................ ................................ ................................ 104 Calculation of Values ................................ ................................ ................... 106 Performance Evaluation ................................ ................................ ................. 106 Contribution of Body Size Effect to Overall Intersubject Variability of Relevant PK Parameters ................................ ................................ ............. 107 Comparison of BW and BSA based Dosing at Population Level .................... 108 Results and Discussion ................................ ................................ ......................... 109 Data Collection ................................ ................................ ............................... 109 Dosing Approach Performance ................................ ................................ ...... 110 Performance evaluation based on AUC ................................ ................... 110 Performance evaluation based on C max ................................ ................... 113 Contribution of Body Size Effect to Overall Intersubject Variability in PK Parameters ................................ ................................ ................................ .. 114 Performance of BW and BSA based dosing ................................ ................... 114 Relationship between the Class of Anticancer Drugs and Body Size Effect on Pharmacokinetics ................................ ................................ ................... 115 Discussion and Summary ................................ ................................ ..................... 116 5 CONCLUSION ................................ ................................ ................................ ...... 129 LIST OF REFERENCES ................................ ................................ ............................. 132 BIOGRAPHICAL SKETCH ................................ ................................ .......................... 154
8 LIST OF TABLES Table page 2 1 Selected therapeutic peptides and proteins and their dosing a pproaches for adult patients ................................ ................................ ................................ ...... 55 2 2 Population pharmacokinetics/pharmacodynamics (PK/PD) models for the selected therapeutic peptides and proteins ................................ ........................ 56 2 3 Percentage contribution of body size measurements to the overall intersubject variability of PK parameters in selected proteins and peptides ....... 58 3 1 Total body water change by age ................................ ................................ ......... 96 3 2 Ti ssue distribution comparison of newborn and adults (Organ weight expressed as % of total body weight) ................................ ................................ 96 3 3 Renal functi on: glomerular filtration rate (GFR) and renal plasma flow (RPF) by age ................................ ................................ ................................ ................. 96 3 4 Pharmacokinetics of selected proteins and peptides in pediatrics. ..................... 97 4 1 BSA formulas ................................ ................................ ................................ .... 121 4 2 Selected anticaner drugs and their dosing approaches for adult patients. ....... 122 4 3 Population pha rmacokinetics/pharmacodynamics (PK/PD) models for the selected anticancer drugs. ................................ ................................ ................ 123 4 4 Percentage contribution of body size measurements to the overall intersubject variability of PK parameters in anticancer drugs ........................... 124
9 LIST OF FIGURES Figure page 1 1 Schematic diagram of relation between observed concentration, interindividual variability, int raindividual variability and errors. ............................ 38 1 2 Schematic diagram of relation between typical clearance, individual clearance, interindividual variability, intraindividual variability, and errors. ......... 39 2 1 The % difference of AUC for patients with extremely low and high body weight from those for patients with a median BW of 75 kg. ................................ 59 2 2 I nter subject variability of simulated AUC and Cmax of 1000 subjects after a single fixed dose or a body size based dose in proteins and peptides ............. 60 2 3 Comparison of the deviation (% dif ference) of AUC and Cmax for subjects with low and high extreme body size from the typical body size ......................... 61 2 4 The intersubject variability and d eviation of AUC and Cmax of the PD markers after fixe d and body size based dosing. ................................ ............................. 62 3 1 An example of fixed dosing. An example of body size based dosing. An example of fixed dosing by different age groups or different body size groups. 99 4 1 I nter subject variability of simulated AUC and Cmax of 1000 subjects after a single fixed dose or a body size anticancer drugs. ................................ ........... 125 4 2 D eviation (% difference) of AUC and Cmax for subjects with low and high extreme body size from the typical values ................................ ........................ 126 4 3 Visual comparison of the 5 th and 95 th percentile of cisplatin after a BS A based dose and a BW based dose for four BW or BSA value quartiles .......... 127 4 4 Comparison of BSA based dosing and BW based dosing at population and individual levels. ................................ ................................ ............................... 128
10 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy DOSE OPTIMIZATION AND BODY SIZE By Rong Shi May 2011 Chair : Hartmut Derendorf Major: Pharmaceutical Sciences Therapeutic biologics and anticancer small molecules are often administered based on body size. A previous study has found that fixed dosing performs similarly to body size based dosing in reducing inters ubject variability in drug exposure across the monoclonal antibodies ( mAbs ) studied. A few studies have questioned body surface area (BSA) based dosing for anticancer drugs. This dissertation extend s t he evaluation from mAbs to other therapeutic proteins a nd peptides. This dissertation also for the first time evaluates fixed dosing and BSA based dosing using population pharmacokinetic simulation. Eighteen therapeutic proteins and peptides as well as 28 anticancer drugs with published population pharmacokine tic (PK) models were selected for dosing approach evaluation. The relationships betwe en body size and drug exposure were evaluated, and simulation studies were conducted to compare the performance of the 2 dosing a pproaches. The results showed that fixed d osing performed better for more selected biologics and anticancer drugs than body size based dosing in terms of reducing intersubject variability in exposure at both population and individual levels. This result is consis tent with the findings for mAbs, an d previously published review for anticancer drugs. Therefore, fixed dosing is recommended for first in human studies of
11 proteins and peptides and small oncology drugs The final dosing approach for Phase III studies should be determined based on a full a ssessment of body size effect on PK/PD when data are available and the therapeutic window of the drug The pediatric dosing for protein and peptides was also discussed in this dissertation and selected proteins and peptides were reviewed for their body si ze based dosing. Pediatric dosing is more complicated than adult dosing. F or most of the evaluated pro teins and peptides, body size based dosing was appropriate. However, for certain biologics, a simplification can be made, such as one fixed dose for a ran ge of body weight/age and another fixed dose for a different range of body weight/age.
12 CHAPTER 1 INTRODUCTION Overview of Population Pharmacokinetics and Dose Optimization During drug development, the concept of drug response in pharmacokinetic and pharm acodynamic (PK / PD) investigations has been confirmed for many decades. However, drug response is influenced by substantial interindividual variability. Individual patients given the same dose and dose regimen may present widely varied responses in onset, m agnitude, and duration. Dose optimiza tion and dose individualization is, therefore, the common goal for scientists who are involved in drug development, both in the pharmaceutical industry and regulatory agencies. To achieve this goal does not seems to be an easy journey since the large between patient variability exists in pharmacokinetic, pharmacodynamic, physiological, physiochemical, and pathophysiological processes which ultimately effect the clinical outcomes of the medicines The task of the late st age of drug development is to select dose or doses to be proposed for therapeutic use, and the purpose is to optimize the dosage and reduce the risk of failing to meet the safety and efficacy criteria. The focus of this dissertation is on the interindividu al variability in pharmacokinetic s and pharmacodynamics, mainly pharmacokinetic s of the protein, peptides and anticancer drugs T he factor that can cont ribute to the interindividual variability in pharmacokinetic s and pharmacodynamic s can be intrinsic fac biological characteristics ( e.g. gender, race, age, weight, height, menopausal status, pregnancy) ; genetic differences ( e.g. polymorphisms in metabolizing enzymes and transports) ; disease related characteristics ( e.g. tumor type, cancer stages, surgery, liver function, renal function, albumin and alpha 1 acid glycoprotein levels) ; comedications ( e.g.
13 antibiotics, herbal supplements, over the counter medications); and environmental factors such as patient life style ( e.g. adh erence to medications, food, alcohol, smoking, coffee, exercise, stress) In order to optimize the dose and dosing regimen, study or studies may be characteristics that have a significant effect on PK and PD, and ultimately on the clinical outcome. The classic PK/PD analyses, in many cases, fail to identify the factors that are responsible for the interindividual variability. To individualize doses further for therapeutic agents with high inte rindiv idual PK or PD variability, the drug development team needs additional information of factors that might account for the intersubject variability. Population pharmacokinetics was originally proposed in the 1970s by Sheiner et al 1 In early 1990s, population pharmacokinetics dr ew the attention due to the increasing activities in the field and has expanded from a discipline mainly used for therapeutic drug monitoring to a critical tool in drug development In February 1999, the FDA issu Guidance for I ndustry : Population P which laid out the mechanisms and theory of population pharmacokinetics and highlighted its usage during the drug development process 2 Population pharmacokinetics w as defined in he study of the sources and correlates of variability in drug concentrations among indiv iduals who are the target patient population receiving clinically relevant doses of a drug of interest Population pharmacokinetics seeks to identify the measurable pathophysiologic factors that cause changes in the dose concentration relationship and the extent of these changes so that, if such changes are associated with clinically significant shifts in the therapeutic index, dosage can be
14 In 1997, a survey of 206 new drug applications showed that 47 of the applications contained population pharmacokinetics and/or population pharmacodynamic reports and early application of population PK provided helpful saf e ty, efficacy and dose optimization information in 83% of the 47 applications 3 In the traditional pharmacokinetic studies, the main focus has been the average behavior of the group, such as the mean plasma concentration time profile. The interindividual variabilit y in pharmacokinetics is often investigated through complex controlled studies. Additionally, concentra ting on one single variable in the traditional way of pharmacokinetic study make s it complicated when interactions among variables need to be understood. Compared to traditional pharmacokinetic studies that are mainly focused on collecting plasma concentration of the drug at a series of time point s population pharmacokinetic studies collect any further information that might explain the interindividual v ariability in the patient population ( e.g. gender age, body weight, body surface area, race, smoking status, drug response phenotypes etc.). Similar to traditional pharmacokinetics, population pharmacokinetics also estimates the average pharmacokinetic parameters ( e.g. typical value of clearance or volume of distribution ) The focus of the population pharmacokinetic analysis process is identifying the factors that do have an effect on PK parameters or exposures which is the m ajor purpose of the analysis ( e.g. body weight, gender, age, etc. which are called covariates in population pharmacokinetics or fixed effect s on pharmacokinetic parameters ) Population pharmacokinetic s also quantifies the magnitude of unexplained (random) variability whi ) The greater the
15 percentage of the interindividual variability can be explained by covariates, the more control we have in ensuring that the dosing of the therapeutic drug is both efficacious and safe. In other words, the less the intersubject variability accounts for the overall interindividual variability, the more information we have to adjust dosing according to the covariates for different group s of patient population. Concentrations of the drug may v ary between days or within a day due to the error of the analysis measurements. Samples taken in different clinic al study locations may have variation as well. This kind of variability is called residual intrasubject variability and interoccasion variabili ty usually expressed as e psilon Estimation of such variability is also important for therapeutic drug monitoring. Finding the optimum dose f or a population, subpopulation, or individual patient requires knowledge of the variability mentioned above. T he siginificance and usefulness of the population PK in drug development has been recognized more and more nowadays, it should be integrated as broad as it can be when appropriate. As early as in pre clinical studies, population PK can be applied to allom etric scaling and toxickinetic studies in small and large animals, as well as more importantly to guide the first in human (FIH) doses for clinical Phase I trials. In Phase I (usually healthy volunteers ) studies, population PK can be applied to understand the structural model with the mean PK parameters and potential covariates for healthy subjects, as well as the foundation to evaluate the difference between patient and healthy populations. Phase II studies data provides the best situation to uncover the d ose response of PK and PD for the product in patient population, and propose dose strategies that maxi mize the benefits and minimize the adverse effects. Phase III and IV
16 studies can provide opportunities to further explore the population PK model and find correlation of new appeared responses. Population Pharmacokinetic Approach In 1972, Sheiner et al. first proposed the theory of population pharmacokinetics, and suggested using computer program s to find the optimal dose for a number of drugs for individu al patients 4 In order to conduct population PK analysis, Lewis B. Sheiner and Stuart Beal developed NONMEM (non linear mixed effect model) software during the 1970s at Universi ty of California San Francisco. The difference between NONMEM and other statistics softwares that can also be used to conduct non linear mixed effect model is that there are numerous model libraries and specific terms that are easier for clinical pharmacologists and pharmacokineticists to ad apt as a tool. It has been used mainly in the academic research in the 1980s, and regulatory agencies began to take notice of population PK in the late 1980s. Since the FDA issued the guidance for population PK in 1999, NONMEM (Globomax) has been the gold standard for pharmaceutical companies and regulatory agents. The goal of this part of the chapter is introduce the concept of nonlinear mixed effects modeling, the mechanics of NONMEN and the utility of population PK in drug development. W e all understand that individuals vary in the pharmacokinetic and pharmacodynamic response s to the administration of certain dose s of drugs. Population PK is the tool to understand why there are variations and to what extent are the variations The s implest way to describ e population PK is the PK in a certain patient population. The goal is to e stimate mean PK parameters and between subject variability, e s timate individual PK parameters, and e stimate residual variability in order
17 to u nderstand measurable sources of variab ility in PK and describe their relationship to PK parameters The reason that industry and regulatory agencies conduct population PK is to understand factors that cause variability in PK. It is a highly efficient way to screen a large number of diverse ind ividuals from the certain target population. The advantage of population PK is to investigate multiple factors ( e.g. disease status, demographics (body weight, age, gender, etc.), drug interactions, food effects, etc. and combinations of all these factors at once instead of designing and conducting multiple clinical trials according. Nonlinear mixed effects model Nonlinear mixed effects model is a suit able tool to analyze repeated measurements data. The clinical PK/PD data is usually r epeated measurements over time. The function s to fit the PK/PD data are also very common to be nonlinear in the parameters. This is the reason why nonlinear mixed effects model became popular in PK / P D modeling. The n onlinear mixed effects model simultaneous ly estimate s param eters relating fixed effects and random effects to observed data. Fixed effects are observed or measured variables (e.g. dose, time, weight, age, gender, genetic difference, etc.), and random effects are unexplained random variability (e.g. inter individua l or residual errors). Based on the premise, the individual pharmacokinetic parameters of a patient population arise from a distribution which can be described by the population mean a nd the interindividual variance. Each pharmacokinetic parameter can be e xpressed as a population mean and a deviation for that individual Nonlinear mixed effects model is a one stage analysis that simultaneously estimates mean
18 parameters, fixed effect parameters, interindividual variability and residual random effects The c oncept of mixed effects model can be explained by a schematic approach as it applies to analysis of PK data. Let us take the example of concentration time profiles of a group of subjects 5 In Figure 1 1, e ach data point is identified using the subject ID in the concentration time profile The red line is the mean concentration time profile using a 1 compartment model that provides the best fit to all of the data in the figure This mean profile is a resul t of a mean c learance as shown in Figure 1 1C, the clearance CrCL (creatine clearance). T he horizont al line indicates that there is only one clearance value across the whole patient population and the volume of distribution has a similar case We all know that individuals vary. We will focus on s ubject 2 in the concentration time profile, Figure 1 1D. If only data from subject 2 were fit alone, a higher clearance will be predicted The blue line indicates the best fit for this particular subject due to a higher clearance. It is assumed that the individual pharmacokinetic parameter s arise from a distribu tion which can be described by the population mean and the interindividual variance. The deviation (the difference between the population mean and the individual parameter) is assumed to be a random variable with an expected mean of zero and variance 2 This variance describes biological population variability. Therefore, in order to fit each subject, we sample an ETA for each subject fro m this distribution shown in green (Figure 1 1, C and D ). However, we still have not predicted the concentrati ons for subject 2 accurately. This is where we should consider the second level of random variable. This random variable represents assay error, within
19 subject error (or the intraindividual variability) and model specification error. It is also assume d th at the distribution of the random variable epsilon is zero with a variance of sigma squared. In a wor d, there are two levels of random variability (one for inter individual variability and one for residual variability). The next step is to identify any possible covariates that can explain the interin dividual variability of the parameters For example suppose this compound is predominantly eliminated via the kidney, and assume there is a strong relationship between CL of the drug and CrCL. In Figure 1 1 D, since s ubject 2 ha ving lower concentrations, c learance is higher than the clearance value of the red line As a result a particular typical value can be assigned to this subject (instead of everyone has the same typical clearance value) ; here is the gre en line indicates the typical conc entration time profile of a subject with a certain CrCL (Figure 1 1 C and D) As can be see n the predicted concentrations are much compared to the mean profile in red However, there is still a difference between the typic al concentration time profile for subject 2 with a certain CrCL (in green) and the individual predicted profile (in blue) This difference is the interindividual variability in the new model that incorporates CrCL as a covariate for clearance. The reason f or this difference between the two profiles (also two models) is that CrCL has attributed some of the inter subject variability in CL but CrCL cannot explain all the interindividual variability in CL The g rey shaded area is now an area where we could expl ain the variability from subject to subject using CrCL The g oal would be to have this grey area as big as possible In order to achieve this goal, all the possible covariates ( e.g. body
20 weight, gender, age, genetic polymorphism, etc.) that could contribut e to the interindividual variability shou ld be evaluated carefully during the analysis The covariates, eta and epsilon can be better explained by Figure 1 2. In Figure 1 2 C, the clearance CrCL profile, instead of everyone h aving the same clearance, the typical value of clearance can be expressed by TVCL= 1+Cr CL* 2 and the typical clearance of the whole population depends on the value of CrCL. The individual clearance for a particular subject is expressed as CL i and CL i = TVCL The red profile is corresponding to TVCL and blue profile is corresponding to CL i Even after between the observed concentration points to the profile, and this is the random variable epsilon attributed by assay error, within sub ject error, and model specification error As mentioned before, fixed effects are observed or measured variables (e.g. dose, time, weight, age, gender, genetic difference, etc.), and incorporating them into equations can be shown for example: TVCL= 1+Cr CL* 2 or CL=TVCL*(CrCL/median CrCL)^ 2 To describe the population PK model using equations the two levels of random effects can be expressed by the following equations. In this example, I.V infusion administration is used Interindividual variability: Where CLi is individual clearance for the ith subject, TVCL is the typical value of clearance for the population, and CL,i is the interindividual variability on clearance for the ith subject.
21 Residual variability: Where y ij is the response, the concentration of ith subject at jth time point, CL i is individual clearance for the ith subject, k 0 is the infusion rate, V i is individual volume of distribution for the ith subject, T ij is the infusion duration of ith ij ij is error term of the intraindividual variability of ith subject at jth time point. Nested random effects : TVV is typical value of volume of dist ribution for the population, V ,i is the interindividual variability on volume of distribution for the ith subject. A simple case in population PK nonlinear mixed effect s model is explained above Th ere are different scenarios in PK that can be analyzed using nonlinear mixed effect s model if the variability is the interest. Population PD and population PK/PD are widely adapted in drug development as well. Since the same theory applies to population As lon g as the data is repeated measurements, and the goal is to understand not only the structure model but the variability nonlinear mixed effects model is the ideal candidate tool.
22 Population Pharmacokinetics in Dose Optimization during Drug Development The application of population PK can be involved in all phases of clinical drug develo pment: the transition from Phase I to Phase II/III, dose selection for Phase III d ose adjustment in special populations c onfirm drug interaction studies s upport for confi rmatory efficacy/safety p ediatrics studies and dosing, b ridging to different ethnic groups r es ults incorporated in product package i nsert and a lso applied in preclinical development Sheiner 6 first introduced the learning and confirming concept to the drug development process. In the first learning and confirming cycle (traditional Phase I and Phase IIa), the purpos e of Phase I is to find out what is the biggest short term dose that maybe given to human without introducing side effects ; and the purpose of Phase IIa is to confirm that the dose resulted from Phase I provide promising efficacy in selected group of pat ients. The decision after the first learning and confirming cycle is the go or no go call for continuation of the investment of the drug after reviewing whether there is strong enough indication of efficacy and absence of toxicity. In the second learning a nd confirming cycle, the purpose of the Phase IIb is to find a good, if not ideal, dosing and dosage regimen to provide a promising clinical outcome; and the purpose of the Phase III is to get the compound approval by confirming via a randomized clinical t rial in a large number of target patient population using the selected dosing and dosage regimen fro m Phase IIb to achieve reasonable benefit risk ratio. The application of population PK in the modeling and simulation in different phases and the learning and confirmation cycle is summarized as below. Phase I: PK/PD simulations in preclinical studies are very helpful in deciding the first in human dose. The PK data in Phase I is characterized by rich data at different
23 dose levels and relatively consisten t PK profile for each individual. The structural population PK model is usually best identifi ed due to the intense sampling in Phase I. In addition, related demographics for PK parameters as well as PD models can also be done in this ste p to be better prep ared for Phase II designs. Since the Phase I studies are usually carried out in healthy volunteers, the PKPD difference between healthy population and patient population can be studied later. Especially, for the PD models, if the difference is established between healthy and patient population for one of the compound s used previously for the same indication for the disease, then these models can be used for another compound via simulation to better guide the design of the Phase II studies. The importance of the population PK in t his p hase is no doubt, since this is the first time ever in human, for the new compound estimates of the basic structural model parameters and demographic covariates effects on the parameters could be evaluated. The estimation of int er and intra individual variability could provide valuable information for designing robust Phase II studies. Phase II a: It is the confirming step of the first learning and confirming cycle. Phase IIa studies are done in patients at different dose levels Phase II has more cohorts than Phase I and more dose levels than Phase III, which provides the greatest opportunity to apply population PK modeling and simulation knowledge for dosing strategies with promising benefit risk ratio. The population PK mod els give an overall idea of fixed effects and random effects in healthy population. In Phase IIa, the learning and confirming steps go along together The goal of this stage is the proof of concept. The concentration in this learning step is usually spars e sampling. After the proof of concept, the PK/PD steps in to analyze the relationship between dose, concentration and effects.
24 The population PK/PD models then could estimate for patients certain demographic characteristics will/will not achieve the targ eted exposure that corresponds to final clinical outcome. Phase IIb: Phase IIb is the dose finding step. The knowled ge about variation in PK and PD in patient population will be improved. The previous knowledge about PK and PD will help to design the dose finding studies via simulations. The data observed will further improve the population PK/PD models. An optimized dose for Phase III will be proposed using simulations with latest u pdated population PK/PD models in order to achieve the ultimate goal of ma ximizing the beneficial effects while minimizing adverse effects. Phase III: The purpose of the Phase I II is to confirm efficacy in a larger patient population as well as learning side effects about this population. It is usually better to use sparse samp ling design in a small group with exposure data, because sometimes confirmation of efficacy fails. Using the data from this small group, we can validate the population PK model. The demographic information in Phase III ( e.g. body size, gender, disease stat us, age, etc.) is richer than previous phase s. This in formation will provide more improvement for the population PK model. Simulations from the updated models could provide good recommendation for new Phase III to confirm the efficacy. The simulations can also help dose adjustment with special populations to be indicated in the drug label. Overall, in the past decade, population PK/PD has been applied more and more in drug development by pharmaceutical industry. T he theories, methods and application s of p opulation PK/PD modeling and simulation has been actively published in clinical
25 pharmacology journals and international conferences, such as American College of Clinical Pharmacology American Association of Pharmaceutical Scientists American Society for Clinical Pharmacology and Therapeutics Population Approach Group in Europe which is more focused on population PK and PD started in Europe and the new emerging American conference that focused only on PK/PD modeling the American Conference on Pharmacome trics (ACOP, the first ACOP meeting was hold in 2008) The guidance by FDA in 1999 speeded up the growth of population PK/PD in drug development. It is and has been shown to be a great tool that needs to be integrated in the modern drug dosing finding stu dies to build more efficient, robust, informative, and cost effective clinical studies at different stages. It is more encouraged to be implemented throughout the entire drug development process including the preclinical animal studies T he early the imple mentation is the more information and benefits it can provide. Body Size Measurement and Body Size E ffect on Pharmacokinetics Although with not sufficient scientific evidence b ody size based dose is usually used for biotherapeutic molecules and small oncology drugs. Flat fixed dosing is most commonly used for small chemical molecules in adults. Whether body size based dosing should be used in patients, mainly depends on whether body size has a significant effect on PK and PD of the drug, as well as i ts therapeutic index. If the drug has a wide therapeutic window, a flat fixed dose for every adult patient is cho sen, regardless of the effect of body size on PK and PD. The reason of this decision is because flat fixed dosing provide a serie s of advantage s than body size based dosing: 1) it is easier for pharmacists to prepare and easier for nurse to administer; 2) patients usually have better compliance; 3) there is less risk for medical errors, since no
26 calculation is needed for dosing each patient; 4) i t cos t s less, since there is no drug wasted. On the other hand, if the drug has a narrow therapeutic index, not only the body size effect, but all the other possible factors that can affect PK/PD of the drug should be systemically evaluated for an appropri ate dosing strategy that provides the optimal clinical outcomes. The goal of evaluat ing the factors that could affect the PK/PD of the drug is to optimize a dosing strategy to achieve low interindividual variability in the patient population by adjusting the dose when necessary according to the factors ( e.g. body size, gender, pharmacogenomics differences), and most importantly achieve the optimal risk/benefit ratio. If the interindividual variability cannot be explained by any obvi ous factor, then flat f ixed dosing should be chosen during drug development and clinical practice. If the re is enough evidence showed that the drug then the drug development team needs to choose a body size measurement for the dosi ng in this case. Pinkel has suggested using body surface area (BSA) instead of body weight as the dosing criterion for anticancer chemotherapy 7 Body wei ght has long been used in the dosing criterion, and it is a straight forward measurement that is easier to observe than body surface area. However, BSA was believed to be better correlated with basal metabolic rate as well as blood volume than body weight across difference species ( e.g. rabbits, guinea pigs, and mice) and human s 7 BSA based dosing has been ex tensively used in chemotherapy. T he history of BS A and BSA formulas will be discussed in detail in Chapter 4. Besides body weight and BSA, there are other body size measurement s such as body mass index (BMI), ideal bodyweight (IBW), percent IBW, adjusted bodyweight,
27 lean bodyweight (LBW). The b ody size measurements and body size effect on pharmacokinetics as well as obesity pharmacokinetics are reviewed in this section. Measurement s of Body Size Since it is hard to measure body fat and fat free mass directly, many methods have been used to indirectly me asure body composition. These indirect measures of body composition are often derived from readily measurable body size, thus body weight, height, and gender. B ody mass index proposed by Quetelet i n 1869 8 It was intuitively suggested that body size should t first Quetelet proposed that body volume would be ideally explained by height to the third power (HT 3 ) This idea was introduced by Keys et al. to describe the incidence of coronary heart disease in men 9 Keys then reported that the ratio of total body weight to HT 2 best to describe the result renamed as body mass index (BMI) BMI is currently the and recommen ded measure to classify obesity It is used to separate people into four major groups: underweight BMI < 18.5 kg/m 2 ; normal weight BMI18.5 24.99 kg/m 2 ; overweight BMI 25 30 kg/m 2 and obesity BMI > 25 kg/m 2 Obesity is then further classified as modera te, BMI30 34.99 kg/m 2 ; severe, BMI 35 39.99 kg/m 2 ; and morbid BMI > 40 kg/m 2 Although BMI is widely used by medical profess ionals, it is not that helpful in pharmacokinetics and dosage calculations. The biggest limitation for BMI is that it does not diff erentiate between the fat tissue mass and lean muscle mass. If BMI is used, then the patient with a large muscle mass would be given the same dose as a patient with a large fat
28 mass. As a result, BMI is unlikely to be come a good body size measurement for dosing. Ideal body weight. Ideal body weight (IBW) was a size measurement first derived from insurance data collected by the Metropolitan Life Insurance Compan y of New York. The company re ported evidence that relates size to mortality in woman and man, re spectively in 1942 and 1943 10 11 This finding was updated in 1959 12 and 1960 13 using the data obtained during the Build and Blood Pressure Study 14 T he study recruited more than 4.5 million people; however, ideal body weights for height were derived based on only 360 000 people This version of IBW from was reported unrelated to total body weight (TBW) and is an estimate of weight corrected for height gender and body frame size. Equation s for estimation was report ed by Blackburn in 1977 15 However, the emp irical estimate of IBW that is often used in pharmacokinetics was derived by Devine in 1974 in a case study of gentamicin 16 The Devine formula was not related to the Metropolitan Life Insurance Company data, even though they are similar The Devine formula is expressed as: IBW (kg) = 45.4 kg (49.9 kg ifmale) + 0.89 (heigh t in cm 152.4). Unlike BSA and BMI mentioned before, IBW introduced gender into the equations. The D evine equation is the most common citation in the PK studies IBW also seems to not be an ideal metric for dosing calculation, since it basically suggests that patients with the same gender and height should be given the same dose regardless of body fat and muscle composition. Adjusted body weight In order to improve IBW as dosing calculation for size, the concept of adjusted body weight (ABW) came into the picture in the pharmacokinetics. ABW was the first body size measurement that was derived for the purpose of
29 pharmacokinetics studies. ABW was introduced to the field as part of a noncompartmental analysis for aminoglycoside dosing in 1983 by Bauer 17 The mean correct ion factor (C F) was estimated to be 0.45 for gentamicin, 0.37 for tobramycin and 0.42 for admikacin, and later and average value of 0.4 were used for aminoglycoside generally. The equation of ABW (kg) = ideal body weight + [ 0.4 (actual body weight ideal body weight )] was incorporated in to basic pharmacokinetics text books 18 19 ABW ov ercome the drawback by introducing total body weight into the equation and utilizes the difference between total body weight and IBW for dosing adjustment. In clinic practice, ABW is often used for aminoglycoside dosing calculation 20 Since the CF is determined on case by case base, so far ABW is used mainly for aminoglycosides. Fa t free mass. Fat free mass (FFM) was first reported by Rathbun and Pace in 1945 to validate the relationship between weight and fat mass 21 FFM was derived using live weight, eviscerated wet weight and eviscerated dry weight of guinea pigs in order to find out the total fat mass. Since direct measure of human body fat is difficult, sever al indirect methods were proposed: derivation from height and total body weight 22 skinfold thi c kness measurement 23 underwater weighing ( density test) 24 and total body potassium test 24 Bioelectrical impedance anal ysis (BIA) has been combined with height, total body weight and gender to determine FFM 25 27 Garrow and Webster presented the regressed FFM equations to evaluate the effect of obesity on the pharmacokinetics of glibenclamide 28 The study used the fat mass estimated by skinfold thickness, underwater weighing and total body potassium tests, an d derived equations to evaluate how BMI correlates with fat mass.
30 Lean body weight. Fractional fat mass (FM frac ) has been used by James for obesity report for the Department of Health and Social Security Medical Research Council in the UK 29 It was then used to estimate lean body weight (LBW), where LBW is the difference between total body weight and fat mass weight 30 LBW concept relates closely to the concept of FFM, and it consists of bone, muscle, extracellular fluid and vital organs 31 The often used equations by James to estimate LBW ar e : for males, LBW (kg) = 1.10 x TBW 0.0128 x BMI x TBW; females, LBW (kg) = 1.07 x TBW 0.0148 x BMI x TBW 29 Green and Duffull have reported that the formulae may not be physiologically accurate when estimating the extremes of height and weight 32 A semi mechanistically derived equation for estimating LBW was reported in 2005, using bioelectrical impedance data 31 The equations for this LBW estimation are: for male: LBW (kg) = (9270 x TBW) / (6680 + 216 x BM I ); for female: LBW (kg) = (9270 x TBW)/( 8780 + 244 x BMI ) Th ese LBW estimates do not decrease with increasing total body weight; i t was suggested to be a more appropriate method of calculating LBW in obese and normal weight individuals. Predicted Normal Weight. Predicted normal weight (PNW) was derived by Duffull to predict the normal weight estimate for dosing obese patients 33 This weight descriptor was developed to better describe the pharmacokinetics properties of the drugs. PNW was developed to describe the expected normal body weight of obese patients ight and their predicted normal fat mass (excluding the excess portion of the fat mass) and height For males, PNWT (kg) = 1.57 x TBW 0.0183 x BMI x TBW 10.5. For female, PNWT (kg) = 1.75 x TBW 0.0242 x BMI x TBW 12. 6. However, because that
31 PNWT is derived using the previous equations for LBW, its accuracy was also questioned at the extreme values of height and weight 33 The estimation of body composition in elderly patients is more complex because of the fact that the ratio of adi pose tissue and lean tissues increa se s with increasing age 33 34 The semimechanistic LBW was suggested to account for changes related to age due to the fact that it was derived using bioimpedance data 31 The changing of pediatric body composition is even more complicated than in elderly individuals, and the effect of body composition on pharma cokinetics will be discussed in Chapter 3 in detail. Body Size Effect on Pharmacokinetic parameters The prevalence of obesity has been rising dramatically in the last decade. Due to the fact that obese patients often have other diseases such as type 2 di ab etes mellitus and heart disease more and more physicians are facing the dosing adjustment challenges for obese patients. Since there is not much pharmacokinetic information of obese patients available in the literature for most of the drugs, finding the right dose for a new obese patient could be difficult. Green and Duffull have summarized the impact of different size descriptors on the pharmacokinetic parameters clearance (CL) and volume of distribution (Vd) in obese patients 35 The size descriptors included TBW, height, IBW, LBW, ABW, BMI, BSA, FFM, PNWT. on CL and Vd were evaluated using either regression of the parameter against a size descriptor or as part of p opulation PK model. Green and Duffull concluded that none of the single size descriptor is dramatically better than others in terms of explaining the variability of pharmacokinetic CL and Vd in obese patients. However, they suggested that since there is a strong empirical and mechanistic theory that support s the u se of LBW for CL and TBW for Vd, i f the drug does not degrade in fat tissue, the clearance
32 was believed to be related more to LBW than TBW. Vd was concluded to c onsistently increase with excess a dipose tissue and TBW is more related to the physicochemical properties of drugs, especially for lipophilic drugs. Volume of distribution. Volume of distribution (Vd) is a theoretical term that relates total drug amount in the body and plasma or tissue dr ug concentration. Vd presents roughly to what extent a drug distributes into extravascular spaces. As a result, drugs with larger tissue uptake usually have larger volumes of distribution. However, with only the information of Vd it is not sufficient to de termine the sites of distribution. The way to find out the site of distribution is to measure directly the tissue concentration at certain sites which is usually not readily observed in the clinical studies. The emerging techno logy microdialysis made dire ct measurement of tissue concentration possible 36 Though the brain and lung microdialysis is not often conducted in human, skin microdialysis including adipose tissue and skeletal muscle is readily performed in patients to measure endogenous compound and free drug concentration at the sites. Hollenstein, Brunner et al. has reported that obese and non obese subjects presented very different drug plasma concentrat ion but very similar free tissue concentration measurable by microdialsys when given weight adjusted dose of ciprofloxacin 37 Barbour Derendorf et al. reported the micro dialysis measured c oncentrations in the interstitial space fluid of soft tissues following a s t andard 1.5 g cefuroxime dose may be high enough to prevent infections with Gram positive organisms in morbidly obese patients undergoing abdominal surgery 38 In general, increase of body weight leads to increasing of body water, muscle mass, adipose tissue and organ blood flow, which all may contribute to the increase of
33 volume of distribution. Followed by dru g administration in a patient, the Vd of a drug will depend on both the physicochemistry of the drug and physiology of the patient. The physical and chemical properties of the drug determine the rate and extent of the drug distribution to the body, such as the molecular size, ionization degree, lipophilicity and the permeability to the biological membranes. The physiological conditions that affect the distribution of the drug includes degree of tissue perfusion, tissue size, permeability of the tissue and very importantly protein binding and tissue binding The lipophilicity can be assessed by the partition (octanol/water) coefficient of the drug. It makes sense that Vd of lipophilic drugs is usually larger in obese patients than normal weight subjects, si nce obese patients have an exce ss amount of adipose tissue compared with non obese patients. However, the proportion and amount of fat tissue is not the single factor for even lipophilic drugs that affect Vd, Vd varies due to the affinity of the drugs to the tissue as well. Phy siological changes in obesity can significantly change the drug distribution in obese patients. These changes include increased cardiac output, increased blood volume, larger organ mass, larger lean body mass, significant absolute a mount and proportion of adipose tissue mass, as well as altered plasma or tissue binding. The adipose tissue has smaller proportion of water than lean muscle which leads to a smaller proportion of water per total body weight in obese subjects. Hydrophilic drugs distribute poorly to the adipose tissue, the Vd changes little in obese patients compared to non obese patients. In this case, obese patients might be overdosed if water soluble drugs are given per body weight.
34 Tissue perfusion and plasma protei n binding could potential ly change the volume of distribution. Tiss ue perfusion was reported to decrease in obese subjects b y Summers et al. after evaluating the s ubcutaneous abdominal tissue blood flow 39 Abel et al. summarized the obesity influence on cardiac structure and function 40 In terms of plasma protein binding, it does not seem that obesity has a big impact on drug binding to albumin for alprazolam, triazolam and phenytoin 41 42 Contradictory conclusion was 1 acid glycoprotein in obese subjects 43 46 Vd plays an important role in do sing obese patients, since it is an essential parameter for determining load dose for a lot of drugs. A good understanding of obesity impact on Vd is therefore the key information. In many pharmacokinetic analyses, Vd can be expressed as an absolute value with a unit of liter or a weight/BSA normalized value with a unit of liter per kilogram (L/kg) or liter per square meter (L/m 2 ). For hydrophilic drugs, there is little change in absolute Vd (L) in obese compared with non obese subjects, but decreased bo dy weight normalized Vd (L/kg). For extremely lipophilic drugs, both the in absolute Vd (L) and body weight normalized Vd (L/kg) are increased in obese patients. There are some advantages body weight normalized Vd(L/kg) to absolute Vd (L). By comparing th e body weight normalized Vd (L/kg) values in obese and non obese subjects, one can get some information about the drug distribution to the excess adipose tissue. For some drug s body weight adjusted dose is often used for dosing obese patients assuming t he body weight normalized Vd (L/kg) is constant in both obese and non obese subjects. But in the reality, bo dy weight normalized Vd (L/kg) is not always the same for both obese and non obese
35 populations. Especially for hydrophilic drugs the absolute Vd increased in the obese but not much body weight normalized Vd (L/kg) is much lower in obese patients than in non obese subjects. This indicates that the drug does not distribute much to the fat tissue. Therefore IBW or LBW might be a better method than T BW when dosing obese patients. If the absolute Vd does not increase much, then a flat dose could be given to both obese and non obese subjects for the loading dose. Additionally, if the adipose tissue is not the site of target or the site of adverse events then distribution might not be the interest of dosing strategy D ifferent dosing methods might have different effect on distribution but the one that benefit s other important pharmacokinetic properties should be chosen in this case Clearance. Clearan ce (CL) is often considered to be the most important pharmacokinetic parameter when finding the right dose regimen. Maintenance dose regimen needs the clearance information to be determine d CL mainly depends on the physiological properties of the patien ts when compared with Vd which is affected by both the physiochemical properties of the drugs and physiological properties of the patients For a given organ, CL can be seen as the volume of blood being removed in a certain time, or the rate of blood volum e removed from the organ. One of the main organs for drug elimination is kidneys. Renal clearance depends on glomerular filtration rate (GFR), tubular reabsorption, and tubular secretion. The influence of obesity on renal functions is still not well known 47 According to the Cockcroft Gault formula developed in 1976 if control ling for the gender and age of the patients, then creantinine clearance is increased with increasing body weight 48 However, in the published studies when obesity effect on GFR is investigated, i t has
36 been shown GFR increased, decreased, or remain unchanged when comparing obese and non obese subjects. In 2007, Pai et al. investigated the influence of morbid obesity on daptomycin and reported that GFR values were found to be 60% higher in the morb idly obese patients than non obese patients 49 However, there was no difference found for the two groups for Vd, total clearance, renal clearance or protein binding. For many drugs, hepatic elimi nation in liver plays an important role in the total drug clearance. Obese individuals have been reported to have higher risk of developing nonalcoholic fatty liver disease 50 Impairmen t of hepatic microcirculation has been found in fatt y liver 51 Thus, these changes in obese individuals might change their hepatic blood flow and hepatic function, and as a result change d the hepatic clearance of the d rugs. Emery et al. has reported increased cytochrome P450 2E1 activity in morbidly obese subjects with nonalcoholic fatty liver disease and recovery after weight loss 52 Green and Duffull have summarized different size descriptor s for clearance for various drugs. They concluded that there is no single best size descriptor for determining clearance and reported that 35% of the studies in which LBW was considered. It was suggested from a physiological standpoint that LBW should be used generally, due to the more close relationship with liver/kidneys to LBW than TBW 35 In a more recent publication, Han, Duffull et al. emphasized further the correlation between LBW and clearance by suggesting a simple solution when dosing obese patients 53 It was proposed that obese subjects have higher absolute drug cl earance than non obe se subjects; clearance does not linearly increase with increasing total body weight; and clearance and LBW are linearly correlated. With the first two points
37 generally accepted, the third point has been challenged by Mathijssen and Spar reboom who claimed that BSA is a good size describer for the summarized anticancer drugs 54 Soon after the report, Han et al respon ded to the publication by suggesting LBW and BSA calculated for patients of 40 120 kg and 150 190 cm are highly correlated 55 However the weight range seem not be able to represent obe se and non obese populations. Overall, currently, there is no single, well proven size descriptor to present drug CL in obe se and non obese individuals. descriptor requires the consideration of all the factors that are height, weight and gender. When CL is incorporated in the population pharmacokinetics analysis, gender could be included as a covariate, and as a result, the size descriptor does not necessarily describe the gender difference; the same method can be appl ied to age effect on size It is true that no single size descriptor can be the tool to provide constant exposure to the whole patient population. We should not expect the size measurement to be able to explain all the interindividual variability as well. As the population PK analysis is raising rapidly in the pharmaceutical industry and regulatory agencies the overall picture (all factors that could affect PK/PD including size, genetic polymorphisms, gender, age, disease) needs our attention when findi ng the optimal dose for patients instead of only focus ing on size effect. With the blooming of pharmacogenomics combined with population PK/PD modeling and simulation, pharmacokinetics and pharmacodynamics stand more and more critical roles in dose optimi zation
38 Figure 1 1. Schematic diagram of relation between observed concentration, interindividual variability, intraindividual variability and errors
39 Figure 1 2. Schematic diagram of relation between typical clearance, individual clearance, i nterindividual variability, intraindividual variability, and errors
40 CHAPTER 2 FIXED DOSING VERSUS BODY SIZE BASED DOSING OF THERAPEUTIC PEPTIDES AND PROTEINS IN ADULTS 1 Background In contrast to small molecule drugs, which are commonly dosed with fixed doses, therapeutic biologics such as body weight (BW) and body surface area (BSA). It has been generally perceived that dosing biologics variability in pharmacokinetics (PK) and/or pharmacodynamics (PD) and thereby optimize their therapeutic outcomes, based on the theory that patients with larger body size would have a larger volume of distribution and a higher elimination capacity. The validity of th e above perception, however, has recently been challenged for dosing of mAb therapeutics 56 Through simulation studies of 12 approved mAbs with published population PK and/ or PD models, Wang et al. 56 compared the performance of fixed and body size based dosing with regard to reducing intersubject PK and/or PD variability in adult patients. In contrast to the expectation that body size based dosing should produce less intersub ject variability, their r esults showed that fixed dosing p erformed similarly to body size based dosing across the 12 mAbs evaluated, with fixed dosing being better for some mAbs (7 of 12) and body size based dosing being better for the others (5 o f 12). The study indicated that de termination of body size as a significant covariate on clearance (CL) parameter does no t necessarily justify body size based dosing as a better approach than fixed dosing even when the only consideration is to reduce PK and/or PD variability, because simpl e body size correction as mg/kg or 1 Reprinted with permission. Zhang S, Shi R, Li C, Parivar K, Wang DD. Fixed Dosing Versus Body Size Based Dosing of Therapeutic Peptides and Proteins in Adults. J Clin Pharmacol 2011
41 mg/m 2 could overcorrect the effect of body size on exposure. The determinant factor is of body size on CL is described by a power function, as shown in equation below: (2 1) In terms of reducing intersubject variability in the area under the concentration time curve (AUC), body size based dosing would work the best when CL is body size 1), and fixed dosing would work the best when CL is not affected by value is less th (Figure 2 1). The contribution of body size effect to overall PK and/or PD intersubject variability was also evaluated, as many other factors, besides body size, could also contribute to the intersubject variability in PK and/or PD of a given drug. These factors include intrinsic factors (eg, age, gender, disease states, and genetic polymorphism) and environmental factors (eg, concomitant medication and smoking). Depending on the contribut ion of body size effect to the overall inte rsubject variability, body size based dosing may or may not be necessary even when it significantly reduces intersubject variability in certain PK and/or PD parameters. On the basis of the findings from this anal ysis, the authors recommended (1) using fixed dosing in first in human (FIH) adult studies for mAbs because of its convenience, better compliance, less risk for medical errors, and cost effectiveness and (2) selecting
42 the final dosing approach for Phase II I trials based on a full evaluation of the effect of body size and other influential factors on PK and PD of the study mAb along with its therapeutic window after sufficient data become available during drug development 56 As a continuation of the mAb study 56 the present study was set out to evaluate th e clinical benefit of body size based dosing for other therapeutic biologics namely, therapeutic peptides and proteins. These biologics should be exa mined separately considering that the sources of the variability in exposure are likely to be different between mAbs and therapeutic peptides and proteins since they do not necessarily share the same distribution and elimination mechanisms. In contrast to mAbs, which usually share the same IgG structure with a molecular weight of ~150 kDa, therapeutic peptides and proteins comprise a much more diverse group of molecules. Large therapeutic proteins may share similar distribution and elimination mechanisms to mAbs. They are generally distributed via convection and eliminated via intracellular catabolism following fluid phase or receptormediated endocytosis. However, mAbs differ from most therapeutic proteins in that a significant fraction of mAbs is protected from protein catabolism by FcRnmediated recycling, whereas most of the therapeutic peptides and proteins do not have such protective mechanism with the exception of fusion proteins containing the Fc region of IgGs 57 58 For smaller therapeutic peptides and proteins, depending on the molecular size and physicochemical properties (eg, charge and lipophilicity), renal excret ion and diffusion into tissues may play an important role in their overall elimination and distribution mechanisms in addition to catabolism and convection 59 With increased amount of PK and PD informati on of this class of therapeutics accumulated in recent years, it is time to compare the clinical benefits of
43 the 2 dosing approaches for therapeutic peptides and proteins. The objectives of the present study are therefore to 1) systemically evaluate the pe rformanc e of fixed dosing and body size based dosing for therapeutic pepti des and proteins in adults and 2) recommend a dosing strategy for clinical trials conducted in adults at different stages of the development of therapeutic biologics other than mAbs. Methods Data Collection Data used in this simulation study were collected from the population PK/PD studies of therapeutic peptides and proteins published in peer reviewed journals. The selection criteria included the availability of population PK and/or PD models for adult patients or healthy volunteers and adequate assessment of the effect of body size on the PK (and/or PD) parameters. Population PK and PD Models The population PK and/or PD models of the selected peptides and proteins (see Table 2 1 ) we re obtained from published reports. General properties of these population PK and PD models and the effect of body size, such as BW or BSA, on the PK and/or PD parameters are summarized in Table 2 2 Mixed effect models were used to describe the PK and/or PD of all the selected therapeutic peptides and proteins. The jth observation for the ith individual was given by i 1ij 2ij are the residual errors followin g normal distributions with a mean of zero 1 2 and 2 2 i can be further described by
44 where z i population values of PK and/ o i is the intersubject variability 2 Simulation analysis was conducted using NONMEM (version VI; GloboMax, Hanover, Maryland). PK and/or PD Simulation Population per formance. Simulations to evaluate population performance of a dosing approach were conducted in the same way as previously described 56 Briefly, Monte Carlo simulation was conducted using the final PK and/or PD model reported f or each peptide or protein to obtain the concentration time profiles following both fixed dosing and body size based dosing approaches. The dose used for simulation was the dose recommended in the labeling for marketed products or the dose used in the repo rted clinical trials for biologics under development. The median value of body size (BW or BSA) was used as the conversion factor for dose determination so that the dose used in the fixed dosing approach is the same as the dose for the participants with me dian (typi cal) body size in the body size based dosing approach. For all simulation studies, 1000 participants were simulated per dosing approach. The sampling points were chosen based on the PK or PD properties of therapeutic peptides and proteins, and th e same sampling schedule was used for both dosing approaches. For each simulation study, with the exception of BW, values of influential covariates were r andomly generated using S PLUS 7.0 (TIBCO, Palo Alto, California) assuming normal or lognormal distri bution. The values of parameters used for generating these covariates were selected by trial and error, with a goal of reproducing
45 the patient population by matching the median, standard deviation, or the range of covariates to those reported in the corres ponding population PK/PD study. BSA was generated assuming normal distribution with a median of 1.82 m 2 and a range of 1.2 to 2.4 m 2 56 BW values were generated as previously described 56 Historic al data have shown that BW does not follow normal distribution, whereas transformation of BW by a power and can adequately describe the right tail of natural BW distributio n 60 Therefore, the BW values were generated assuming a normal distribution of Z randomly generated BW (1000 participants) has a median of 75.7 kg and a range of 38.8 to 187.2 kg, which in general covers the range of those reported in the population PK/PD studies for selected peptides and proteins (Table 2 2 ). Indi vidual performance. To evaluate the individual performance, we simulated the PK profiles for the participants with typical, low extreme, and high extreme body size. The typical, low extreme, and high extreme BW/BSA used in this study were 75.7 kg/1.8 m 2 4 0 kg/1.3 m 2 and 140 kg/2.3 m 2 respectively. For covariates other than BW/BSA, typical values were used. The intersubject variability and residual errors were all set to zero for the simulations conducted for both dosing approaches. Calculation of Values A simple way to assess wh ether fixed dosing or body size based dosing may be ze effect on CL, as defined in E quation ( 2 1), obtained from covariate analysis However, the covariate models used to characterize the effect of body size on CL for the selected molecules are not all in the form of E quation ( 2 1) (Table 2 2 ). Therefore, an effort was
46 evaluated. For those not reported, the following steps were used to obtain the generate a series of CL values over the range of body size reported based on the reported original covariate model and (2) fit the generated CL ve rsus body size data using equation (Pharsight, Mountain View, California). AUC calculation. The AUC for each participant was calculated as dose/CL if the molecule exhibited a linear PK. When the molecule exhibi ted a nonlinear PK, the AUC was calculated by integration of the concentration time curve by the trapezoidal method using S PLUS. PK C max (PD C max or C min ) determination. The maximum concentration ( C max ) for each participant was determined as the maximal c oncentration from the simulated concentration time profile of the participant. For PD measurements, C max or the minimum concentration ( C min ) of the PD marker, whichever reflects the maximal drug effect, was determined from the PD marker time profile of the participant. Performance Evaluation The performance of a dosing approach in terms of reducing intersubject variability e relationship as described by E quation (2 1) based on the follow ing criteria: fixed dose is better. fixed and body size based dosing are similar. body size based dosing is better. These criteria apply to both population and individual performances. To be consistent with the mAb paper1 and to co nfirm the predictive performance of
47 value, results from simulation studies for the comparison of the 2 dosing approaches were also presented as described below. Population performance was assessed by comparing the intersubject variability (expressed as % coefficient of variation [CV % ]) in the exposure (AUC and C max ) of 1000 participants simulated following the 2 dosing approaches. The dosing approach that produced less intersubject variability provides a better population performance. Individual pe rformance was evaluated by comparing the percentage difference in the exposure between participants with extreme body size and typical body size following the 2 dosing approaches. The dosing approach that resulted in a smaller difference in PK or PD exposu re between participants with extreme body size and those with typical body size has a better individual performance. Contribution of Body Size Effect to Overall Intersub ject Variability of Relevant PK Parameters A simulated PK data set of 100 participants was generated using the published covariate values were randomly generated as described above. This simulated data set body size as covariate(s) for any PK parameters. The percentage change in the the contribution of body size effect to the overall intersubject variability of the relevant PK parameters.
48 Results and Discussion Data Collection A total of 18 therapeutic biologics were identified based on the data collection criteria specified in the Me thods section (Tables 2 1 and 2 1 ), including 7 therapeutic peptides and 11 therapeutic proteins. The effect of body size on various PK parameters, including CL, intercompartment clearance (Q), volume of the central compartment (V or V1), and volume of the peripheral compartment (V2), is summarized in Table 2 1 Body size has been found to be covariate(s) of 1 or m ore of the PK parameters for 11 t herapeutic biologics (Table 2 2 ). The effect of body size on PK parameters of these 11 biologics was generally w ell estimated with a couple of exceptions. The relative standard error (RSE) or 95% confidence intervals of the parameters that describe the body size effect on CL and/or volume of distribution (V/V1 or V2) are summarized below to help the interpretation o f these simulation studies. The RSE ranged from 16% to 26% for CL and 12% to 33% for volumes of distribution except for octreotide (89.5% for V). In the case of darbepoetin alfa, where standard error of the estimate was not reported, 95% confidence interva ls It should also be pointed out that for 4 of the selected biologics namely, degarelix, hematide, recombinant factor VIIa, and rhGH although body size had been found not to be covariates of their PK param eters, the population PK models were developed based on data from fewer than 30 patients (Table 2 2). Therefore, some caution needs to be exercised in interpretation of the following results. Among these 18 selected biologics, 12 are administered based on their body size in adult patients (Table 2 1). Interestingly, for some products that a re administered using body size based dosing, such as hematide and onercept (Table 2 1), body size
49 measures (BW) had been shown not to be a covariate of their PK paramet ers (Table 2 2). Dosing Approach Performance Performance evaluation b ased on AUC As discussed in the mAbs work 56 the comparative performance of the 2 dosing approaches bas the power function as defined in equation ( 2 all the 18 sele cted therapeutic biologics and listed in Table 2 either directly obtained from the published reports or obtained in such a way as described in the Methods section. For products with body size found not to be a covariate of CL, a zero 2, 12 of 18 would perform better for 12 mol based dosing would perform The results of the simulation studies for comparing the performance of these 2 dosing app roaches at the population level are presented in Figure 2 2A. Consistent with intersubject va riability in AUC when body size based dosing was adopted, whereas the oth er 12 biologics fixed dosing was (enfuvirtide), the variability was similar for the 2 dosing approaches. Similar results were also obtained for individual performance
50 ( Figure 2 3A). For all therapeutic biologics investigated, the dosing approach that had better population performance also had better individual performance. It should be noted that the zero difference in AUC between patients with extreme body size and typi cal body size following a fixed dose for daptomycin, degarelix, emfilermin, lanreotide autogel, octreotide acetate, onercept, recombinant factor VIIa, and rhGH is a result of the lack of BW effect on their CL (Figure 2 3A). The consistent conclusions obtai the optimal dosing approach if AUC is the exposure parameter of the main concern. This approach can also be used for comparative performance evaluation based on C max if body size is only a covariate for the V/V1 but not for any other parameters, as is the case for emfilermin and octreotide. However, if the body size variable is a covariate on more than one PK parameter that could af fect C max simulation needs to be conducted to evaluate the performance of the 2 dosing approaches in terms of reducing C max variability. Performance e valuation b ased on C max The population and individual performances of the 2 dosing approaches based on Cm ax were evaluated by simulation studies and shown in Figures 2 2B and 2 3B, respectively. At the population level, body size based dosing resulted in less intersubject variability in Cmax for 7 of 18 biologics, whereas fixed dosing produced less variabili ty in Cmax for the other 11 biologics (Figure 2 2 B). At the individual level, body size based dosing produced a smaller percentage difference in C max between participants with extreme and typical body sizes for 6 of 18 biologics, whereas fixed dosing produ ced a smaller percentage difference for the other 12 biologics (Figure 2
51 3B). The results from both population and individual level evaluations are again very consistent with the only exception of enfuvirtide, for which bod y size based dosing was shown to have better individual performance but worse population performance. It was noted that body size based dosing tends to overdose patients with large body size and underdose patients with small body size. The opposite is true for fixed dosing, that is, overd ose patients with small body size but underdose patients with large body size (Figure 2 3A, B). Performance evaluation based on Pharmacodymanics Among the 18 selected biologics, PD models, in addition to the PK models, have been reported for 3 products (ab atacept, darbepoetin alfa, and etanercept) 61 63 Therefore, the 2 dosing approaches were also evaluated for their performance in reducing PD variability. The PD re sponse of abatacept namely, IL 6 levels was described by an indirect response model, in which the IL 6 degradation rate was stimulated by abatacept according to an Emax model 63 The PD response of darbepoetin namely, the hemoglobin levels was described by a modified indirect response model, wherein serum darbepoetin stimulated the production of hemoglobin through an Emax model after weekly administration of darbepoeti n 61 A logistic regression model was adopted to describe the exposure response relationship for etanercept 62 The cumulative AUC of etanercept was used as the exposure variable, and the American College of Rheumatology response criterion of 20% improvement (ACR20) was used as the binominal clinical outcome. Body size measures were not identified as covariates for any PD parameters in any of these PD models. The results of PD performa nce for all 3 biologics are shown in Figure 2 4. Although the comparative performance of 2 dosing approaches in reducing PD variability is in the same order as
52 that in reducing drug exposure variability, the difference between the performances of the 2 dos ing approaches based on PD is smaller than that based on the PK for all these 3 therapeutic agents (Figures 2 2, 2 3, and 2 4). For example, the intersubject variability in drug exposure of etanercept following the 2 dosing approaches was shown to be 47.5% (fixed dosing) versus 45.7% (body sized based dosing) for AUC and 37.4% (fixed dosing) versus 31.4% (body size based dosing) for C max at the population level. However, the intersubject variability in its PD measures was 82.2% (fixed dosing) versus 81.9% ( body size based dosing) for AUC and 70.8% (fixed dosing) versus 70.3% (body size based dosing) for C max of the PD effect. As the ultimate goal of a clinical trial is to achieve its efficacy and safety endpoints, efficacy data, safety data, and data of sur rogate PD markers, when available, are more important than PK data alone. The smaller difference in intersubject PD variability between the 2 dosing approaches suggested that the clinical benefit, if any, of bo dy size measurements, as shown in this study for abatacept, darbepoetin alfa, and etanercept. Contribution of Body Size Effect to Overall Intersubject Variability in PK Parameters The contribution of the effect of body size to the overall intersubject variability of relevant PK parameters was evaluated for 8 therapeutic biologics, and the res ults are summarized in Table 2 3 It was observed that the effect of body size had a small and, in some cases, moderate contribution to the ov erall intersubject variability of major PK parameters, ranging from 1.8% to 18.4% for CL and from 0.42% to 26.9% for V/V1 (Table 2 3 parameters appear to correlate with the relative cont ribution of body size to the overall
53 variability, although the rank orders of the two are not exactly the same. This discrepancy may be due to the difference in the extent of the contribution of other identified and unidentified factors, such as demographi c characteristics and disease conditions. When body size only explains a very small percentage of the intersubject variability for example, in the case of darbepoetin alfa and PEG interferon alpha 2b adjusting the dose based on body size would lead to a m inimal reduction in the variability in AUC. On the other hand, if body size is a major source for inte rsubject variability, body size based dosing may provide a clinical benefit when supported by other factors, such as a narrow therapeutic window. Relation ship between the Type of Therapeutic Biologics and Body Size Effect on Pharmacokinetics The therapeutic biologics investigated for this analysis included 7 therapeutic peptides and 11 therapeutic proteins. Among the 11 therapeutic proteins, including 3 fus ion proteins and 1 pegylated protein, no apparent correlation betwe en either the type or the size of therapeutic proteins and the body size effect on their PK was observed. esult, fixed dosing would work better for all the peptides evaluated. Whether this can be generalized to other peptides remains a topic for further investigation when more data become available. Summary As a continuation of the mAb work 56 body size based dosing and fixed dosing were evaluated for 18 nonmAb therapeutic biologics in terms of their population and
54 individual performances in reducing intersubject PK and/or PD variability in adult patients. The resu lts demonstrated th at body size based dosing did not always result in less intersubject variability in drug exposure and PD measurements. In fact, fixed dosing showed better performance for 12 of 18 evaluated biologics based on both AUC and C max assessments. Even if the 4 bi ologics whose population PK models were developed based on data from fewer than 30 patients were excluded from the analysis, fixed dosing would still show better performance for 8 of 14 evaluated biologics. Therefore, the recommendations made for mAbs dosi ng 1 also apply to non mAb biologics. For adult FIH studies, fixed dosing is recommended because it offers advantages in ease of dosing preparation, reduced cost, and reduced chance of dosing errors. When sufficient data become available, a full assessment of body size effect on PK and/or PD should be conducted. The final dosing approach for P hase III trials in adults should be selected based on the established body size effect on the PK and PD, the therapeutic window of the therapeutic products, and other f actors that may affect the outcome of the study.
55 Table 2 1. Selected therapeutic peptides and proteins and their dosing approaches for a d ult p atients Generic name Brand name Approval date MW (Da) Type Target Dosing approach Abatacept Orencia 2005 92,3 00 fusion protein CD80/CD86 mg/kg Daptomycin Cubicin 2004 1,620 peptide LTA synthesis mg/kg Darbepoetin alfa Aranesp 2001 37,100 protein EpoR g/kg Degarelix Firmagon 2008 1,632 peptide GnRHR mg Emfilermin Discontinued 22,007 protein LIFR g/kg E nfuvirtide Fuzeon 2003 4,492 peptide gp41 mg Erythropoietin alpha EPOGEN 1989 30,400 protein EpoR Units/kg Erythropoietin beta NeoRecormon 1993 30,000 protein EpoR g/kg Etanercept Enbrel 1999 150,000 fusion protein TNF mg Hematide In development NR 1 pegylated peptide EpoR mg/kg Lanreotide autogel Somatuline 2007 1,096 peptide IGF 1 mg Octreotide acetate Sandostatin 1988 1,019 peptide SSTR2/5 g Onercept Discontinued 18,000 fusion protein TNFR mg/kg PEGinterferon alpha 2b PEG Intron A 2001 19,271 protein IFNAR1/2 g/kg Plitidepsin Aplidin 2004 1,110 peptide EGFR mg/m 2 Recombinant Factor VIIa NovoSeven 1999 50,000 protein TF g/kg rhGH 1 Norditropin 1987 22,000 protein GH receptor mg/kg u hFSH 1 Metrodin HP Discontinued 30,000 prot ein FSH receptor IU 1 Abbreviation: NR: not reported; rhGH: recombinant human growth hormone; u hFSH: Urinary human follicle stimulating hormone.
56 Ta ble 2 2. Population pharmacokinetics/pharmacodynamics (PK/PD) models for the selected therapeutic peptides and p roteins Generic name Type MW (Da) Dosing Structure Model Covariates, Mean (SD or range) Covariate model N CL Ref Abatacept fusion protein 92,300 mg/kg 2 CMT, linear BW: 78.3 (21.0) CL=CL0+CL1x(BW/78.3) 388 0.4 63 Daptomycin peptide 1,620 mg/kg 2 CMT linea r BW : 75.1 (48.2 152.8) CL=[CLr+0.14(Temp(C) 37.2)]*y Q=3.46+0.0593x(BW 75.1) V2=[3.13+0.0458x(BW 75.1)*z 282 0 64 Darbepoetin alfa protein 37,100 g/k g 2 CMT, linear BW: 70.8 (36,123) CL=TVCLx(0.737) a x(BW/70) 0.623 V=TVVx(BW/70) 0.639 140 0.623 61 Degarelix peptide 1,632 mg 3 CMT linear no info WT is not a covariate 24 0 65 Emfilermin protein 22,007 g/kg 1 CMT linear BW: 62 (48 83) TV=V+6.7x(WT 62) 64 0 66 Enfuviride peptide 4,492 mg 1 CMT, linear BW: 72.2 (12.7) CL/F=CL0+CL1x(BW/70) 534 0.5 67 Erythropoietin alfa protein 30,400 Units/kg 1 CMT, linear BW: 72.2 (18.96) CL=TVCLx(BW/70) 0.75 V=TVVx(BW/70) 1.37 48 0.75 68 Erythropoietin beta protein 30,000 g/kg 1 CMT, linear BW: 62.0 (5 1.0 79.0) ka=TVkaxBW 1.92 ke=TVkex(CrCL) 0.542 x AGE 1.13 V=TVVx(BW) 0.776 48 0.776 69 Etanercept fusion protein 150,000 mg 1 CMT, linear BW:73.6 (19.2) CL=TVCLx(BW/70) 0.75 V=TVVx(BW/70) 182 0.75 70 Hematide PEGylated peptide 76,000 mg/kg 1 CMT linear & nonlinear BW: 76.9 (59 96) WT is not a covariate 28 0 71 Lanreotide autogel peptide 1,096 mg 3 CMT linear BW:67.1 (12.4) WT is not a covariate 50 0 72 Octreotide acetate peptide 1,019 g 1 CMT linear BW: 77.1 (51 103) V=TVVx(BW/81) 0.362 59 0 73 Onercept fusion protein 18,000 mg/kg 2 CMT linear, BW: 73.1 (11.2) WT is not a covariate 48 0 74
57 Table 2 2. Continued Generic name Type MW (Da) Dosing Structure Model Covariates, Mean (SD or range) Covariate model N CL Ref PEG interferon alpha 2b protein 19,271 g/kg 1 CMT, linear BW: 80 (41 149) CL=TVCLx(BW/70) 0.455 817 0.455 75 Plitidepsin peptide 1,110 mg/m 2 3 CM T nonlinear BSA:1.78 (1.29 3.32) WT is not a covariate 283 0 76 Recombinant Factor VIIa protein 50,000 g/kg 2 CMT linear no info WT is not a covariate 28 0 77 rhGH protein 22,000 mg/kg 1 CMT nonlinear BW:77.9 WT is not a covaria te 21 0 78 u hFSH protein 30,000 IU 1 CMT, linear BW: 59 (7.4) CL/F=TVCL/Fx(1+0.017x(BW 58.5)) 62 0.99 79 1 Abbreviation: rhGH: recombinant human growth hormone; u hFSH: Urinary human follicle stimulating hormone.
58 Table 2 3. Percen tage contribution of body size m easurements to the overall intersubject variability of pharmacokinetics (PK) parameters in selected proteins and peptides Biologics % contribution of BW to the intersubject variability CL V1 CL V1 u hFSH 13.6 NA 0.99 0 Etanercept 18.4 26.9 0.75 1 Erythropoietin alfa 8.48 11 0.75 1.37 Darbepoetin alfa 1.80 0.42 0.623 0.639 Enfuviride 3.17 NA 0.5 0 PEG interferon alpha 2b 2.12 NA 0.455 0 Abatacept 3.46 NA 0.4 0 Emfiler min NA 4.16 0 1.66
59 Figure 2 1 The % difference of AUC for patients with extremely low body weight (BW) (40 kg, colored broken lines) and extremely high BW (140 kg, colored solid lines) from those for patients with a median BW of 75 kg as a function values following a fixed (red) and a BW based (green) dose, assuming The shaded area represents AUC values within100 20% o f typical AUC. Reprinted from 56 with permission
60 Figure 2 2. Compa rison of the inter subject variability of simulated AUC (A) and C max (B) of 1000 subjects after receiving a single fixed (solid bar) dose or a body size (BW/BSA) based dose(open bar) for selected proteins and peptides
61 Figure 2 3. Comparison of the de viation (% difference) of AUC (A) and C max (B) for subjects with low (open bar) and high (solid bar) extreme body size (BW/BSA) measurements from the typical values (AUC and C max for subjects with median body size measurements after a fixed dose (red) or a body size (BW/BSA) based dose (black) for selected proteins and peptides
62 Figure 2 4. (A) The intersubject variability of AUC and C max (or C min ) of the PD markers for abatacept, darbepoetin alfa and etanercept across 1000 subjects after fixed (solid bar) and body size based (open bar) dosing. (B) Deviation of AUC and C max (or C min ) of the PD markers for subjects with low (open bar) and high (solid bar) extreme BW from the typical value after a fixed (red) and body size based (black) dose of abatacept, darbepoetin alfa and etanercept
63 C HAPTER 3 PEDIATRIC DOSING AND BODY SIZE IN BIOTHER APEUTICS 1 Background Due to the complexity and costs of pediatric safety and efficacy studies, pharmaceutical companies are somewhat reluctant to study drugs and biolo gical products in children. Without safety and efficacy studies in children, physicians are often forced to make empirical assumptions to treat children on a trial and error basis 80 The clinical outcomes of such treatments in children can be promising, marginal or harmful. Physiological development during childhood can produc e significant effects on drug absorption, distribution, metabolism and excretion. After birth, the changes in gastrointestinal absorption, secretion, motility, metabolism and transport, as well as first pass effects will affect the absorption of the drug; the changes in body composition, tissue perfusion and plasma protein binding will affect the distribution of the drug; maturation in cytochrome P450 enzyme mediated metabolism and Phase II metabolism will affect hepatic clearance; and maturation of glomeru lar filtration and renal tubular function will affect renal clearance 81 In general, all the effects of maturation of the pharmacokinetics of a given drug are not well understood. Usu ally, drugs are given with two types of dosing strategies: flat fixed dosing and body size based dosing (Figure 3 1 a and b). The most common dosing approach for pediatrics is body surface area (BSA) / body weight adjusted dosing. Small children are rarely given the same dose as adults (Figure 3 1a). However, a convenient dosing approach that sometimes provides accurate dosing and less intersubject variability is often overlooked by body size based 1 Reprinted with Permission. Sh i R, Derendorf H. Pediatric Dosing and Body Size in Biotherapeutics. Pharmaceutics 2010; 2 : 389 418
64 dosing. This dosing approach provides a fixed dose for a ce rtain age or certain body size group (Figure 3 1c). Fixed dosing for a patient group provides quite a few advantages compared to body size based dosing: ease of preparation and administration, less risk of medical errors, better patient compliance, and cos t effectiveness. When body weight or body surface area adjusted pharmacokinetic parameters can explain the difference between pediatrics and adults, body weight or BSA dose adjustment can provide comparable exposure in pediatrics as in adults. However, thi s is not always the situation. More often, the trend and extent of the pharmacokinetic difference between pediatrics and adults across different age groups are not predictable. Clearance and volume of distribution of drugs can be higher but also possibly l ower in younger children, compared with older children or adults 82 Therefore, simply adjusting the pediatric dose according to the body weight/BSA may not be a n accurate dosing approach. Age should also be accounted for the maturation in pediatrics. Sometimes, even when taking age into consideration for dose determination, it might still not accurately account for all variables related to the different stages of maturation as well as the physiological differences between pediatrics and adults. More importantly, any dose adjustment should decrease the variability in the resulting exposure, which would be the proof of if it makes sense to apply the dose adjustment. General Pharmacokinetics in Pediatrics The definitions of pharmacokinetics in children are as follows: Premature: gestational age < 36 weeks Full Neonates: 0 1 months Infants: 1 12 months
65 Children: 1 12 years Adolescent: 12 16 years Unlike adults, the pharmacokinetics in pediatrics is remarkably affected by the growth and development of children. Body composition an d organ function change over the development of childhood. The total body water, extracellular fluid and intracellular fluid constitution of the body weight in fetuses, premature or full term neonates, infants, adolescent and adults are shown in Table 3 1 83 84 The total body water and extracellular fluid decrease dramatically among premature, full term and infan ts. Additionally, fat contributes to 3% of the total body weight in premature neonates, and 12% of the total body weight in full term neonates; and it is more than 20% by the age of 4 5 months. Protein mass in infants before they start walking is around 20 % and increases to 40% in adults. Lean muscle tissue contains about 75% water by weight. Therefore, total body water, fat and muscle change at different ages may produce significant changes in volume of distribution and systemic concentration of the drug. Different organs such as the heart, liver, and kidney account for more body weight than adults in percentage (Table 3 2 83 ). This can explain the cases when infants or children have a faster body weight normalized clearance than adults, since infants or children have relatively large r liver or kidney per body size compared with adults. The glomerular filtration rate (GFR) is an important factor in clearance. Table 3 3 lists the GFR and renal plasma flow (RPF) change by age. Infants to children from 1 10 days, 1 month and 6 months rou ghly doubl e the GFR in the three stages 83 GRF reaches maturation at age 1 and stays almost constant from the age of 1 to 70 years. GRF/vECF(extracellular fluid volume) ratio and GRF/BSA ratio was studied in 130 patients (age range 1 80 years; 40 patients < 12 years). Neither GFR mea surement
66 showed a significant correlation with age in children. In adults, GFR/vECF significantly decreased with age; however, no significant association was shown between age and GFR/BSA 85 Besides GFR and RPF, the cardiac output (Q, the volume of blood being pumped by the heart in the time interval of one minute) may change by body weight (per kg) or by body surface (cardiac index= Q/BSA) in children at different ages. The cytochrome P450 enzymes constitute the major system for Phase I metabolism, and some CYP enzymes appear to be switched on by birth, while in others onset was later than b irth 86 87 However, for proteins and peptides, endopeptidase or receptor mediated transport processes are involved in hepatic metabol ism instead of CYP enzyme 88 89 The guidance for general industry considerations for pediatric pharmacokinetic studies for drugs and biological products by the FDA has summarized the effect of age and growth on ADME and p rotein binding in pediatrics 90 The guidance has also pointed out ediatric population, growth and developmental changes in factors influencing ADME also lead to changes in pharmacokinetic measures and/or parameters. To achieve AUC and C max values in children similar to values associated with effectiveness and safety in a dults, it may be important to evaluate the pharmacokinetics of a drug over the entire pediatric age range in which the drug will be used. Where growth and development are rapid, adjustment in dose within a single patient over time may be important to maint Pharmacokinetics of Proteins and Peptides Generally, pharmacokinetic principles are equally applied to the large molecule proteins and peptides and small conventional molecules. The underlying mechanisms for absorption, di stribution, metabolism and excretion (ADME) of biologics are usually quite different from that of small molecule s 91 92 Therefore, in order to interpret and
67 apply the pharmacokinetics of biologics, a thorough understating of ADME in these proteins and peptides is required. With few exceptions, almost all proteins and peptides are administered with intravenous, subcutane ous, or intramuscular dosage forms. D istribution The volume of distribution of a molecule is affected by its physiochemical properties (such as lipophilicity and charge), protein binding, and possibly active transporters. Due to the large size of protein s and peptides, they usually exhibit small volume of distribution, limited by the volume of extracellular space because of their mobility and inability to pass through membrane 93 94 However, binding to the intravascular/extravascular proteins or active tissue uptake can significantly increase the volume distribution of biologics 95 The pharmacokinetics of proteins and peptides are usually described by a two compartment model, and volume of distribution of the central compartment as well as the steady state volume of distribution (Vss) are usually reported 96 The typical volume of distribution of the central compartment is equal or slightly bigger than the plasma volume of 3 8 L, and the Vss usually falls in the range of 14 20 L. It should be noted that the assumption to obta in Vss is not suitable for many biologics, since proteolysis in peripheral tissues may contribute a significant portion of overall elimination process for these drugs 92 Therefore, caution should be taken when interpreting Vss for proteins and peptides. Take antibodies for example. The Vss values reported in the literature for antibody pharmacokinetics studie s are usually based on noncompartmental analysis which assumes that the site of elimination reaches rapid equilibrium with plasma and that elimination is only through the central compartment. These assumptions do not necessarily apply to antibodies since a ntibodies also degrade in the tissues 92 Protein binding has been reported to affect the transport and
68 re gulation of some proteins and peptides such as growth ho rmone recombinant human growth factor, cytokines and fusion proteins (e.g. enfuvirtide) 97 100 Eliminatio n Peptides usua lly have short elimination half lives, which is desirable for drugs like hormones, while large proteins like antibodies have an elimination half life of around 21 days 92 B iotechnological p eptides and proteins are almost exclusively metabolized through the same catabolic pathways as dietary proteins and endogenous biologics. With few exceptions renal or biliary excretions are generally negligible for most peptides and proteins. Proteases and peptidase are widely available throughout the body. Therefore, besides metabolism in liver and kidneys blood and other tissues are also the sites of exten sive metabolism for proteins and peptides Renal elimination was reported in small proteins and peptides through predominantly three routes. Glomerular filtration of interleukin 11, growth hormone, and insulin was described previously 101 103 Some of the small linear peptides are eliminated through hydrolysis by brush border enzymes on the luminal membrane, such as angoitensin I/II, glucagons and luteinizing hormo ne releasing hormone 104 106 Peritubular extraction of immunoreactive growth hormone and insulin has also been reported 101 107 Several proteins and peptides were reported to be the substrates for hepatic metabolism including insulin, glucagon, epidermal growth factor, and antibodies 89 108 Endopeptidase or receptor mediated transport processes were observed in the liver as w ell 88 89 Many therapeutic proteins and peptides are endogenous molecules; receptor mediated uptake followed by intracellular metabolism can take place in the organs that express receptors for these molecules. Since the number of receptors is limited, saturation can happen within the therapeutic concentration range This saturation of the receptor mediated elimin ation
69 contributes a major source for nonlinear pharmacokinetics of many proteins and peptides 1 09 Nonlinear pharmacokinetics due to receptor mediated drug disposition has been often reported for monoclonal antibodies 110 111 In t he current review, we summarize the effect of body size (body weight or body surface area) or age on pharmacokinetic parameters of selected biological products in pediatric patients such as clearance, volume of distributio n, area under the curve (AUC) maximum concentration ( C max ), half life etc. Comparison of these parameters and the relationship of the parameters with body weight and age between pediatrics and adults are also included. The current review selectively incl udes the original clinical pharmacokinetic safety studies in pediatrics that have recorded body weight/BSA or age and/or incorporate body size and age in pharmacokinetic analysis. The dosing strategy of biologics in pediatrics is discussed accordingly. Res ults and Discussion An overview of pharmacokinetics of selected FDA approved proteins and peptides is presented in Table 3 4. Comparisons of the pharmacokinetic parameters between pediatrics and adults, as well as the effect of body size and/or age on thes e parameters are discussed in the following section. Monoclonal Antibodies (mAbs) Basiliximab In a pharmacokinetic and dosing rational study, 39 pediatric renal trans plant patients were enrolled 112 In part 1 of the study, pediatric patients were given 12 mg/m 2 of basiliximab; in study part 2, infants and children received two 10 mg doses and adolescents received two 20 mg doses. Basiliximab clearance in infant s and children (n = 25, 17 6 mL/h) was reported approximately half that of adults (n = 169, 37 15 mL/h) from a previous study 113 and was independent of age ( 1 11 years), body
70 weight (9 37 kg), and body surface area (0.44 1.20 m 2 ). Clearance in adolescents (12 16 years, n =14, 31 19 mL/h) was comparable to the adult values. A similar designed study was done in liver transplant pediatri c patients 114 and together; these data support a simple dosing algorithm for basiliximab in pediatric transplant patients. An adjusted fixed dose of two 10 mg is recommended for pediatr ic patients less than 35 kg, and a dose of two 20 mg is recommended for pediatric patients weighting 35 kg or more just like for adults. In another study of basiliximab in pediatric renal recipients on comedication with mycophenolate mofetil, patients were classif ied by age as 16 children (3 11 years) and 27 adolescents (12 18 years) 115 This study confirmed that the dosing strategy mentioned by the studies above p rovides consistent exposure for children and adolescents. Body surface area adjusted basiliximab clearance was reported to be significantly higher in children. However, children were given higher dose than adolescents (0.54 0.18 vs. 0.42 0.08 mg/kg). Si milar total AUC were observed in the two groups (101 68 g d/mL in children vs. 102 42 g d/mL in adolescents) which resembled those of adults (107 44 g d/ mL) from a previous study 116 Significantly larger central and steady state volumes of distribution were reported in children and adolescents than in adults, whereas half lives were similar, 10.1 7.6, 12.1 5.0 and 11.5 4.3 days respectively in child ren, adolescents and adults. All the data implies that the body surface area adjusted dosing approach does not offer any apparent advantage over the simpler fixed dose approach. Daclizumab In a study of daclizumab, pediatric renal transplant recipients we re divi ded into different age groups ( 12 years (n = 18), and 13 17
71 years (n = 25), and the analysis indicates that bodyweight and race (black vs. nonblack) were found to be significant influences on the pharmacokinetics of daclizum ab in pediatric patients 117 A 4.2 fold range in clearance (CL) (4.50 19.0 mL/h) and a 7.4 fold range in central volume of distribution (V1: 0.64 4.71 L) were less proportional than a 12 fold range of bodyweight (7.5 89.5 kg). As a result, body weight adjusted dosing leads to lower exposure in the younger patient group (<5 years), and higher exposure in patients with larger body weight. The pharmadynamic results showed that the difference in exposur e did not affect the safety and extent of daclizumab saturation in different age groups. Palivizumab Intramuscular humanized monoclonal antibody palivizumab in premature infants and infants with bronchopulmonary dysplasia was studied using body weight adj usted dosing 118 Sixty five infants (ages 4.6 to 7.6 months) were enrolled of whom 11 (17%) received 5 mg/kg, 6 (9%) received 10 mg/kg and 48 (74%) received 15 mg/kg palivizumab. Mean serum palivizumab concentrations (ranges) measured at 2 days and were 28.4 (13.0 41.1) and 91.1 (52.3 174.0) g/mL for the 5 mg/kg and 15 mg/kg dose groups, respectively, and at 30 days the palivizumab levels were 12.5 (4.2 to 2 6.2) and 49.2 (13.5 to 132.0) g/mL. The study concluded that monthly injections of 15 mg/kg were able to maintain mean serum concentrations above 40g/mL. The safety and pharmacokinetics o respirato ry syncytial virus infection 119 Mean palivizumab levels were 61.2 and 303.4 g /mL at 60 m in after infusion and 11.2 and 38.4 g/mL at 30 days, after the infusion of 5 and 15 mg/kg palivizumab respectively. The mean half life was 22.6 and 16.8 days after the infusion of 5 and 15 mg/kg palivizumab, respectively. The mean area under the
72 curve was 487g/mL after 5 mg/kg and 2386 g/mL after 15 mg/kg. No significant differences in clinical outcomes between placebo and palivizumab 5 or 15 mg/kg were observed. Infliximab Infliximab has been studied for the first time in a clinical trial in patients y ounger than 12 months (6 infants and 10 children) 120 The pharmacokinetics of infliximab (5 mg/kg) did not differ as age increases. Standard body weight adjusted d osing provided peak concentrations similar to those reported previously, regardless of subject age. The peak concentrations were similar to those observed in the study with peak and trough levels reported after a dose of 6 mg/kg in 62 children (ages 4 to 1 7 years) with pauciarticular j uvenile rheumatoid arthritis 121 The single dose of 5 mg/kg used in the study with infants and children exhibited comparable system ic infliximab exposure to that reported previously for therapeutic drug monitoring of inflixim ab in adolescents and adults 122 The estimated pharmacokinetic para meters median (CV) in the 5 pediatric patients were volume of distribution (V) 3.0L (13%), clearance (CL) 0.0083 L/h (40%), and half life (t1/2) 10.9 days (20%) 43 The parameters are consistent with a study that reported a median t1/2 of 9.5 days and a median C L of 0.0098 L/h by Cornillie 123 and anot her study reported a t1/2 of 8 to 12 days (n=108, 1,5,10,20mg/kg) 45 In yet another study, 21 pediatrics, ages 8 17 ye ars, were given an inflixi mab dose of 1, 5, and 10mg/kg, serum infliximab concentrations were reported to be proportional to dose, and the pharmacokinetic profile in pediatric patients w as similar to that in adults 124 Gemtuzumab Gemtuzumab is derived from the murine anti CD33 antibody hP67.6. In a pediatric pharmacokinetic study of gentuzumab, twenty nine patients were
73 grouped into three age categories: infant (0 2 years), children (3 11 years), an d adolescents (12 16 years) 125 Dosages of 6, 7.5, and 9 mg/m2 gentuzumab were given to the pediatric patients. Pharmac okinetic parameters of hP67.6 antibody for the first dose are consistent and statistically different from that of the second dose. Increases in AUC and decreases in CL and Vss from the first dose to the second dose in pediatrics agree with those of the adu lts. Reported mean pharmacokinetic parameters in pediatrics are similar to t he values reported in adults 126 Children given the dose of 9mg/m 2 had the hP67.6 para meters of: C max CL, 0.12 0.15 L/h/m 2 ; Vss, 6.5 5.5L; and t1/2, 64 44h after the first dose. Concentration vs. time profiles of hP67.6 was similar for the first dose among age. The mean C max for infants was a bit lower than the children, and the C max for children was 22.8% higher than the adolescents. Infants and children AUC are 2.3% and 33.5% higher than adolescents AUC. CL in infants and children was 80.9% lower and 72.0% lower than in the adolescents. The C L (L/h) values after administration of 9 mg/m 2 gentuzumab in infants, children, adolescents and adults were 0.03 0.02, 0.06 0.03, 0.26 0.30 and 0.27 0.23, respectively. The body surface area adjusted CL (L/h/m 2 ) values in infants, children, adolesc ents and adults were 0.05 0.02, 0.05 0.05, 0.17 0.21 and 0.15 0.13, respectively. Therefore, both absolute CL (L/h) and body surface area adjusted CL (L/h/m 2 ) increase from infants to adults. Volume of distribution showed the same trend: lower in i nfants and children than in adolescents. Body weight adjusted Vss (L/kg) was larger in adults and infants than children and adolescents. There was no statistically significant correlation observed between hP67.6 CL and body weight or CL and age. Intersubje ct variability within age groups was relatively large for the
74 pharmacokinetic parameters. Overall, the body surface area adjusted dose provides comparable exposure for pediatric patients. Alemtuzumab (Campath 1H) In a Phase II study, Campath 1H 0.6 mg/kg (max 30mg) was administered in 13 (8 male) pediatric patients, media n (range) age 8 (3 20) years 127 The study concluded that Campath 1H exposure in pediatric s with acute lymphoblastic leukemia (ALL) tends to be lower than that in adults with chroni c lymphocytic leukemia (CLL) 128 and this observation may be due to the m ore rapid clearance in children. That indicated that children may have higher body weight normalized clearance than adults. Mould et al. reported that adult patients with a Campath 1H trough concentrations >13.2 mg/mL had a 50% chance of achieving either c omplete re mission or partial remission 128 while Montillo et al. reported that all patients with a Campath 1H AUC 0 12 >5 mg hr/mL achieved a complete remission 129 In a study of 30 CLL patients, mean peak and trough plasma concentration was 10.7 mg/mL (2.8 26.4 mg/mL) and 5.4 mg/mL (0.5 to 18.3 mg/mL). It was found that no t all patients showed beneficial clinical response, and higher blood peak concentrations correlate with better clinical outcome s. 130 Cetuximab In a Phase I study, twenty seven children (ages 1 12 years) and 19 adolescents (ages 13 18 years) received escalating weekly doses of cetuximab (75,150,250 mg/m 2 ) 131 In the dose range studied, cetuximab exhibited nonlinear pharmacokinetics, since the AUC does not increase proportionally as dose increases. The clearance after non compartmental analysis decreased with increasing dose in both children and adolescents. In children, c 2 as cetuximab dose increased from 75 to 250 mg/m 2 Similar results were reported in
75 adolescents. The receptor mediated clearance might explain this dose dependent elimination of cetuximab, and the receptors are likely to be saturated at higher doses. The mean steady state volume of distribution across all doses and age groups was around 2 L/m 2 indicating limited distribution of cetuximab into the extracellular space. Overall, cetuximab exhibits nonlinear pharmac okinetics and similar profiles among age groups. Estimates of the pharmacokinetic parameters (clearance, area under the curve, and volume of distribution) at steady state in both the children and adolescent subgroups were comparable to those previously rep or ted in adults 132 The body surface area adjusted dosing seems to provide consistent exposure in children and adolescents compared to that of adults and the pharmacokinetics does not seem to correlate with age in pediatric patients. Bevacizumab In a Phase I study of 20 pediatric cancer patients, ages 1 to 20 years (median 13 years), 10 females and 10 males, bevacizumab exposure was proportiona l to dose (5, 10, 15 mg/kg) 133 The study showed a large degree of interpatient variability in children, and it was simila r to that observed in adults 134 Bevacizumab exhibits linear pharmacokinetics at the dose range of 1 to 20 mg/kg in adults . Median clearance and mean residence time in children and adults are 4.1 vs. 3.9 mL/d/kg and 16.3 vs. 12.4 days, resp e ctively 133 134 In a population pharmacokinetic study of bevacizumab, gender difference was found in ad ult patients 135 but with a limited number of pediatric patients, the gender analysis was not performed in the study 128 Natalizumab In a pediatric study, 38 adolescent patients (ages 12 17 years) with active pediatric Crohn Disease received 3 intravenous infusions of natalizumab ( 3
76 mg/kg) at 0, 4 and 8 weeks 136 The natalizumab peak level and half life after the first and third infusions are 61.0 vs. 66.3 mg/mL and 92.3 vs. 96.3 h. Natalizumab showed time invariant pharmacokinetics and no accumulation on repeated monthly dosing. The C max and half life of natalizumab (3 mg/kg) in the adolescents were reported to be lower and shorter compared with those in adults after a fixed dose of 300 mg. The study showed that the dose of 3 mg/kg in a dolescent patients may reduce the symptoms of severe or moderate Crohn Disease. Overall, the study concluded that the magnitude of the clinical benefit to adolescent patients is u nknown, because the body weight based dosing 3 mg/kg did not provide adequate receptor saturation in adolescents. This study is an example where simple body weight adjustment for dose in adolescents has the potential of underdosing the population. Growth Factors Epoetin alfa and delta In a pharmacokinetic and pharmacodynamic stud y, twelve children with cancer were enrolled, and six (median age 15.2 years; range 9.3 18.6 years) were randomi zed to receive erythropoietin (EPO) 137 In this s tudy, children were randomized to receive i.v. EPO 600 IU/kg (max dose 40,000 IU) or placebo weekly for 16 weeks. Doses for all children were increased to 900 IU/kg (max dose 60,000 IU) due to not observing a 1 g/dl increase in hemoglobin by study week 3 o r 4. EPO clearance after the first dose showed relatively big intersubject variability (0.19 1.08 L/h/m 2 ), but the clearance after the 10th and 11th dose showed much less intersubject variability (0.15 to 0.25 L/h/m 2 ). Additionally, the AUC0 24 of EPO incr eased proportionally with EPO dose in these children. In a previous study in adults, the mean half life and clearance after the first EPO dose were 7.7 h (range 3.5 12.6 h) and 0.4 L/h (rang e 0.3 0.7 L/h), respectively 138 If
77 adjusting for BSA, the study exhibits similar pharmacokinetic parameters as the above study in children with cancer. For example, an adult with a typical BSA of 1.73 m 2 has clearance of 0.4 L/h which equates to 0.2 L/h/m 2 Another intravenous EPO (40 IU/kg) study in children (ages 9 16 years) reported mean half life and clearance of 5.6 h (range 4.4 6.7 h) and 10.1 mL/h /kg (range 7.1 14.9 mL/h/kg) 139 The study also concluded that, after i.v. administration, clearance in pediatrics was two fold of that in adults, and after s.c. dosing, bioavailability was two fold of that in adults. This study showed similar clearance to the cancer children study when adjusting the clearance in cancer children for body weight (12.4 mL/h/kg). Unlike children, premature infants (birth weight <1.25kg) showed greater serum erythropoietin cleara nce and larger volume of distribution than adults 140 Two more studies have reported greater clearance in preterm infants than adults after continuous intravenous or multiple subcutaneous EPO administration, and larger bioavailability was reported in preterm infants than adult s given subcutaneous EPO 140 141 A population pharmacokinetics of intravenous and subcutaneous epoetin delta in pediatric patients wit h chronic kidney disease discussed the covariate effects on epoetin delta and epoetin alf a pharmacokinetic parameters 142 Sixty patients, 47 of them received i.v or s.c epoetin delta and 13 of them received i.v. or s.c. epoetin alfa. In the population pharmacokinetic modeling building, V and CL were allometrically scaled by body weight by fixing the power exponents to 0.75 for CL and 1 for V. Age was included in the final model by a power function, normalized by the reference age of 10 years for children order than 10; sex, dialysis type, and drug type were also included in the model. The typical pharmacokinetic estimates were CL (0.268 L/h), V (1.03 L), Ka
78 (0.055 4 h 1), and bioavailability (0.708 epoetin delta and on predialysis. The epoetin delta pharmacokinetic parameters were similar in children as compared with those in adul ts when normalized by weight 143 The subcutaneous epoetin alfa reported lower bioavailability than subcutaneous epoetin delta. Darbepoetin a lfa This is a randomized, open label, crossover study in pediatric patients with chronic kidney disease (CKD) me an age 11 (range 3 16 years) 144 s.c darbepoetin alfa. The mean clearance and half life of darbepoetin alfa was 2.3 mL/h/ kg and 22.1 h after i.v. administration. Absorption was shown to be the rate lim iting step after s.c. darbepoetin alfa; the mean half life was 42.8 h and mean bioavailability was 54%. Beside slightly faster absorption for s.c. administration, darbepoetin alfa disposition in pediatrics were shown to be simila r to that in adults patient s 145 Darbepoetin alfa exhibited roughly two to four fold longer terminal half life than previously reported in epoetin in pediatric patients 139 Previous studies in adult patients with CKD showed darbepoetin alfa half life was approximately 3 fold longer than that of i.v. epoetin ( 25.3 h v. s. 8.5 h) and around 2 fold longer than s.c. ep oetin ( 48.4 h v.s. 24 h) 145 147 In another study in pediatric patients with chemotherapy induced anemia (CIA) sixteen patients (mean age 12 years, range 5 18 years) were given darbepoetin al 148 After a single dose of s.c. darbepoetin alfa, the mean (SD) terminal half life of 49.4 (32) h was found to be similar to the 48.2 h in pediatric CKD patients 144 T he lack of dose proportionality in the C max between the 0.5 y due to population
79 differences rather than nonlinear pharmacokinetics. Darbepoetin alfa showed linear pharmacokinetics in adults patients. In a study in neonates, a single i.v. dose (4 mg/kg) of darbepoetin was given to 10 10.5 g/dl. The birth weight of the neonates was 1128g (median, ranged from 704 to 3025 g), and were 26.0 40.0 weeks old (median, 29.2 weeks). The mean (range) half life, V and CL in the preterm neonates were 10.1 h (range 9.0 22.7 h), 0.77 l/kg (range 0.1 8 3.05 l/kg), and 52.8 mL/h/kg (range 22.4 158.0 mL/h/kg) respectively. In preterm neonates, there was no significant correlation between age and darbepoetin pharmacokinetic parameters. V was found to be correlated with both age and gestational age in the term and near term neonates. Darbepoetin i.v. pharmacokinetics in neonates was compared with children, and neonates had a shorter half life, a larger V and larger CL than children 149 Filgrastim A different dosage adjustment besides body weight or age based dose was used for granulocyte colony stimulating factor (G CSF) in pediatric patients, ages 2 to 17 years 150 Because G CSF clearance increases with increasing absolute neutrophil count (ANC), the dose optimizing study of G CSF was conducted by giving 8 patients filgrastim at a single dose of 10 mg/kg/day subcutaneo usly for peripheral blood progenitor cell (PBPC) mobilization. This preliminary pharmacokinetics of G CSF seems to indicate that an ANC adjusted G CSF dosing adjustment might improve PBPC mobilization in pediatric patients. Interferon In a Phase I pharmac okinetic study of interferon children, ages 3 15 years, with relapsed acute lymphocyte leukemia (ALL) were given IFN amuscularly for over 25 days 151 Single doses of 2.5 to 15
80 MU/m 2 (total doses of 60 to 200 MU/m 2 ) were given to the subjects. The serum levels of IFN and the AUC are similar to those reported in adult cancer patients, but slightly lower 152 The study did not discuss the body surface area, body weight or age effect on pharmacokinetic parameters. The individual AUC was reported, but due to the unknown information of the total dose the relationship of age or body size with dose adjusted AUC could not be evaluated. A safety and pharmacokinetic study of PEG interferon alpha2a was done in 14 children 2 to 8 years (mean age 4.4 years) with chronic hepa titis virus infection (HCV) 153 Mean (range) weight was 20.1 kg (13.3 45.3 kg). The drug dose was 2 )/ (1.73 m 2 ) x180 body weight was found to be a linear covariate for apparent volume of distribution of the central compartment. The study showed wide intersubject variability with the apparent clearance ra nge of (6.6 35.5 mL/h) in the 14 pediatric patients, which suggested the necessity of individualized dosing. When compared with data from a Phase III 48 weeks adult study, the mean Ctrough in children was comparable to that in adults 154 The mean AUC0 168h was 25% higher in pediatrics than in adult. Standard interferon (INF) has shown better efficacy in pediatric patients; additionally, children seem to tolerate pe gylated INF better than adults. As a result the study concluded that the higher drug exposure in pediatric patients may have potentially good efficacy outcomes. Among the 56 pediatric patients (ages 3 16 years) who participated in the multiple dose pharmac okinetics of interferon alfa 2b study, 20, 19, and 17 subjects received ribavirin 8, 12, and 15 mg/kg/d, respectively 155 Median (range) body weight of the
81 subjects was 40.4 kg (10 95). The pharmacokinetics of interferon alfa 2b in children was approximately twice that of adults on a body surface area basis. The dose normalized AUC0 12h and C max are similar to the multiple dose pharmacokinetics in adults. Blood Factors Factor VII Pharmacokinetics of activated recombinant coagulation factor VII (NovoSeven) was compared in children v s. adults with haemophilia A 156 Twelve children (2 12 ye ars) received rFVIIa at one single dose of 90 and 180 g/kg. In children, the plasma FVII concentration is dose proportional in the dose range of 90 180g/kg. Direct comparison of the results for adults (ages 18 55 years) and children (2 12 years) reflects that plasma clearance was significantly higher in pediatrics than in adults for both the FVII: C and FVIIa clot activity assays. The total body weight normalized clearance was significantly faster in children than in adults with both assays (rFVII:C, 58 v s. 39 mL/kg/ h and rFVIIa, 78 vs. 53 mL/kg/ h). This difference suggests a higher metabolic activity per kg body weight in children than in adults and is likely correlated with age related differences in body composition, such as different liver volume per kg body wei ght, as previously described 157 This difference also suggested a higher dose of rFVII might be needed for children to achieve the comparable levels as in adults. The relationship between clearance and weight was illustrated by a linear regression in a review as CL (mL/kg/h) =76.8 0.488 x (Weight 43.6kg) (p<0.002) 158 Volume of distribution at steady state tends to be larger in children than in adults, but not sig nificantly (196 vs. 159 mL/kg). The dose normalized AUC0 12 was 30% lower in children than in adults. This study is important for pediatric dosing of FVII as it provides predictable pharmacokinetics in children.
82 Factor VIII A study for the first time anal yzed the effect of age and BMI on pharmacokinetic parameters in young children for p rediction of dosage regimen 159 Pediatric patients (52 boys, one girl, mea n SD age 3.1 1.5 years) were given an intravenous bolus dose of rAHF PFM (recombinant anti haemophilic factor -protein free method) 50 IU/ kg. Body mass index (BMI) was a significant predictor of Factor VIII distribution. Vss decreased linearly as BMI increases, and age was a significant covariate for half life and MRT. In another study of rAHF PFM with 111 subjects, median age of 18, rAHF PFM mean (SD) half life was 12.0 4.3 h 160 A study in a premature infant showed the half life of Factor VIII was 6.43 h, and 6 20 h in children from other studies 161 Twenty one patients (ages 8 42 years), including 12 pediatric patients, received single doses of 24 51 U/ kg 162 The pharmacokinetic parameters were CL: 81 606 mL/h, V: 1.6 9.7 L, half life: 7.8 to 18.3 h. Weight wa s found to correlate with clearance and Vss, and a positive correlation of age and half life was reported. It was shown that when body weight increased from 40 to 80 kg, this 100% increase in body weight corresponds to an increase of 42% in clearance and a n increase of 60% in Vss. Clearly this increase of clearance and Vss is not proportional to the weight change. Normalization of clearance and Vss for total body weight will therefore not correctly explain the interindividual differences but rather over cor rect them. Additionally, body weight adjusted clearance in mL/h/ kg and Vss in L/ kg seems to decrease with age. The half life of FVIII tended to be shorter in pediatrics than in adults. In a retrospective study, patients 7 to 77 years old (one child, 16 t eenagers, and 44 adults, body weight 21 to 120 kg) were given FVIII to determine t he
83 pharmacokinetics of FVIII 163 The body weight normalized pharmacokineti c parameters of pediatrics were comparable with those observed in adults. Covariate analysis showed that V1 is significantly related to body weight and BSA. Including BSA in the model decreases substantially the unexplained V1 variability (from 34.1% to 21 .1%). In a study of factor VIII (FVIII), 34 patients (ages 7 74 years), including 16 children, were used for model building 164 Body weight and age were found to be significant covariates. FVIII was administered at around 60 U/kg in the small children, decreasing to 10 U/kg or less in middle age patients. The dose requirements after obtaining individual PK data showed a much greater variation than the dose range used. Weight normalized clearance (CL/kg) of FVIII has been reported to decrease with age and/or body weight during growth from infancy to adulthood, and half life showe d the opposite trend 162 165 166 Pharmacokinetics of FVIII was well described by a two compartment model. In the model building process, the exponent on clearance and volume of distribution was set to 0.75 for the clearance parameters (CL and Q) and to 1 for the volume (V1 and V2) terms. In addition to the influence of body weight on clearance, age showed a significant effect only on weigh t adjusted CL, which decreased by 1.5 mL/h per year of age with a reference age of 24. Age had showed no significant correlation with weight adjusted V1, which was in line with a previous observation that age was not correlated with in vivo rec overy ( C max divided by dose) 167 The study concluded that the right dosage of FVIII cannot be only calculated from body weight and/or age, and suggested that starting doses for most patients to be 1,000 U every other day. Individual FVIII concentrations sho uld then be checked for further dose adjustment.
84 Factor IX A 6 year follow up study was done for coagulation factor, Factor VIII and Factor XI, in children and adults with haemophilia 168 The median CL of FVIII: C was 3.0 (range, 1.1 9.9) mL/h/ kg, the Vss was 0.050 (0.028 0.129) L/ kg, and the half life was 11 (5.1 33) h. Clearance increases with increasing body weight in this patient population. A 100% increase in weight, from 40 to 80 kg, corresponds with a 39% increase in CL for FIX. The median CL of FIX: C was 3.9 (2.9 4.5) mL/h/ kg, the Vss was 0.14 (0.08 0.20) L/kg, and the half life was 32 (26 49) h. The prophylactic dose of coagulation factor, in U/kg, was higher for children, especially small children, because of the higher weight adjusted CL in children than adults. Pharmaco kinetics of Factor IX was studied in 56 patients, ages 4 56 years 169 FIX: C clearance and volume of distribution at steady state increased linearly with bo dy weight, with a faster increase in children and adolescents but remaining relatively constant during adulthood. The body weight adjusted CL and Vss, shown as functions of age, indicated a decrease of 0.68% of CL/body weight per year, and CL/lean body mas s decreased by 0.40% per year. The slope between the two regressions was not statistically different, which indicates that dose adjustment of rFIX (recombinant FIX) to lean body mass did not reduce this variability compared to body weight dose adjustment. Vss/body weight decreased by 0.68% per year, while Vss/lean body mass decreased by 0.38% per year, and they were not statistically different. The terminal half life of FIX: C exhibited no correlation with age, nor MRT. The high intersubject variation in di sposition and required doses of rFIX suggests the need for individual dose titration. Drotrecogin alfa In the first study reporting the use of drotrecogin alfa (activated) in pediatric patients, the overall mean weight adjusted clearance was 0.53 L/h/kg across
85 all infusion rates and age groups (n =63 ). 170 No correlation was found between infusion rate and age group. Weight normalized clearance decreases significa ntly with age in patients <18 years old, although combined pediatric and adult weight normalized clearance was not found to depend significantly on age or body weight. The mean weight normalized CL in patients <3 months (n=11) (0.608 L/h/kg) was 22% higher than that in all patients 3 months or older (0.497 L/h/kg) and 19% higher than that in adult patients (18 years or older). The higher CL in the small children was expected to have slightly lower steady state concentration than in the older patients. Hormo ne Insulin h ormones A trial enrolled 32 children and adolescents (19 girls and 13 b oys; ages 132.5 years, range 6 17 years) to compare the pharmacokine tics of detemir and glargine 171 BMI was 15 24 kg/m 2 for children ages 6 12 years and 18 29 kg/m 2 for adolescents ages 13 17 years, but the study did not mention weight or age effect on variability of the two drugs. Pediatric patients were randomized to receive a se quence of 0.4 U/kg of detemir and glargine. The study concluded that the intersubject variability in pharmacokinetic was significantly lower for detemir than for glargine in type 1 diabetes mellitus (T1DM) children and adolescents. The smaller pharmacokine tic variability was most likely due to the smaller variability in absorption with detemir, which is also likely to be associated with a more predictable therapeutic range. An insulin comparison study in pediatrics reported that insulin aspart had a quicke r onset and shorter duration of action compared with human insulin, meaning aspart is more appropriate to be injected immediately before a meal, which makes it a more practical product 172 In this study, postprandial plasma glucose increments did not differ between the human insulin and insulin aspart. Slightly higher blood glucose
86 concentration observed after breakfast and dinner with insulin aspart administration. In another study, subcutaneous insulin aspart or human insulin (0.15 IU/kg body weight) was given 5 min before breakfast in 9 children (ages 6 12 years) and nine adolescents (ages 13 17 years) with T1DM 173 Insulin aspart exhibits significantly higher C max SD than human insulin (881321pmol/L vs. 422193 pmol/L, p < 0.001). C max and AUC of insulin were found to be related with age in the study. The change of glucose AUC and C max are smaller for insulin aspart than human insulin in children. It was surprising for the investigators to find higher levels of both insulin aspart and human insulin in the adolescents than in children. Additionally, the insulin dosage in this study does not reflect the usual dosage of insulin in adolescents (1.01.5 U/kg per 24 h) and smaller children or adults (0.51.0 U/kg per 24 h). There is not much comparison of pharmacokinetic study in children and adults in the literature. Pharmacokinetic study was done to compare insulin glulisine and regular human insulin analogous in childr en and adolescents with T1DM 174 Ten children (ages 5 11 years) and 10 adolescents (ages 12 17 years) were enrolled. The concentration time profile for insulin glulisine is similar for children and adolescents, whereas human insulin exhibits 64% hig her concentration in adolescents. The higher concentration in adolescents of human insulin is in line with the previous study 173 The difference is suggested to be caused by disparities i n residual endogenous insulin secretion in adolescents and children or simply the fact that adolescents were given a larger meal than the children. adults with T2DM. Thirteen adolescent patients (ages 10 16 years, 7 females, 6 males, body mass index, 32.5 5.0 kg/m 2
87 placebo followed by a standa rdized meal 15 minutes later 175 There is no demographic effect, such as age, sex, race, or degree of obesity, found on exenatide pharmacokinetics in adults during clinical development. The exenatide AUC was found to be dose proportional in these adolescent patients. Postprandial plasma glucose levels were significantly decreased with both doses of exenatide compared with the placebo from 1 to 3 hours after administra tion. The geometric meanSE exenatide AUC0 C max exenatide. Not all exenatide levels were detectable i AUC0 C max (85.111.5 pg/mL) are comparable to those in adults with T2DM (n = 39) (AUC0 360m an d C max 113.012.2 pg/mL) 142 176 1 77 With this finding, the study suggested that the should be explored in adolescent patients. Growth hormone Somatropin inhalation powder and subcutaneous humatrope pharmacokineti cs were compared in pediatrics with growth hormone deficiency, ages 6 16 years, weighted 18.0 5 2.0 kg, and mean BMI 17.5 kg/m 2 178 Participants were randomize d to one of three dose levels: 1) 8.4 mg/d somatropin or 0.5 mg/d humatrope; 2) 16.8 mg/d somatropin or 1.0 mg/d humatrope; 3) 33.6 mg/d somatropin or 2.0 mg/d humatrope. At least two subjects were assigned to each dose level within each of the weight rang es: 18.0 29.9, 30.0 39.9, and 40.0 52.0 kg. The mean serum growth hormone area under the curve of somatropin was dose proportional. There was no
88 significant effect of weight and age on somatropin and humatrope pharmacokinetic parameters. Height was found t o be a significant covariate for somatropin AUC, somatropin C max and humatrope AUC, respectively, which indicates taller subjects tended to have higher AUC and C max A novel sustained release recombinant human growth hormone, LB03002 once a week s.c. inje ction was studied in 37 children (24 boys, 13 girls, ages 6.52.1 years), at dos es of 0.2, 0.5 or 0.7 mg/kg 179 C max and AUC was dose proportional in the dose rang e of 0.2 0.7 mg/kg, and was comparabl e with the levels in adults 180 This study shows body weight adjusted dosing of LB03002 gives comparable exposure in pe diatrics as in adults. Nutropin Depot was administered subcutaneously in 138 pediatrics, and the C max and total growth hormone (AUC0 28 d) were approximately proportional to the dose administered (0.75 mg/kg twice a month and 1.5 mg/kg once a month) 181 Zomacton 2 IU/m 2 jet injected and needle injecte d was studied in 18 pediatric patients, and the AUC, C max and T max are s imilar in both groups 182 The study reported the individual BMI, age and sex information for the subjects, but due to the limited number of patients, no correlation was demonstrated with its pharmacokinetic parameters. Other Proteins and Peptides Interleukin In a dose escalation study in children, adolescents and young adults of recombinant human interleukin 11 (rhIL 11), C max and AUC are dose proportional, with mean C max level (range, 7.6 25.5 ng/mL) and A UC (range, 56.7 dose range o f 25 183 The pharmacokinetics of intravenous and subcutaneous rhIL 11 at a dose range of 3 tudie d in 30 healthy male
89 adults 184 The adult mean C max and AUC was reported to be dose proportional, which is similar to the pediatric study. At their overlappi children and adults, it did not seem to show a difference between the C max and Tmax in adults and pediatrics, but the half life and AUC were significantly shorter and lower in children, indicating higher rhIL 11 clea rance in pediatrics than adults. The AUC in (MTD) of rhIL 11 in children and adolescents Etanercept In a population pharmacokinetic study, 69 patients with juvenile rheumatoid arthritis(JRA), aged 4 to 17 years, received twi ce weekly subcu taneous etanercept 0.4mg/kg 185 Sex was a covariate for CL/F, and power exponent of body surface area was found to be 1.41 when normalizing BSA by the typical BSA of 1.071 m 2 Body weight was found to be a significant covariate for V/F with typical body weight of 30.8 kg. This analysis justified the body weight based dose adjustment for etanercept in JRA patients. Age (<17 years) was identified as one o f the most important covariates on CL in the population pharmacokinetic analysis of pooled data obtain ed from 10 clinical studies 186 The correlation between age an d CL is no longer apparent when age is 17 years and older. Body weight was also found as a significant covariate for both apparent clearance and volume of distribution in rh eumatoid arthritis patients 70 Gender difference was found in apparent clearance in these adults with a mean level of 0.117 L/h in female and 0.138 L/h in male, but the difference was not statistically significant. A similar trend was found in JRA p atients with the population mean CL/F of 0.0576 L/h
90 (95%CI: 0.0525 0.0657 L/h) in females and 0.0772 L/h (95% CI: 0.066 0.0870 L/h) in males 185 Elimination mechanis m of etanercept is not known much, and there was no appropriate explanation for the gender difference reported both in children and adults. In this JRA patients study, simulation was conducted to find out whether body surface area (BSA) or body weight adju stment would be a better dosing regimen 185 To calculate the dose for the BSA based regimen, it was assumed that a patient with the weight of the population median ( i.e., 30.8 kg) and a patient with the BSA of the population median (i.e., 1.071 m 2 ) received the same total dose of etanercept. Therefore, for example, 11.5mg/m 2 (= 0.4mg/kg 30.8 kg/1.071 m 2 ) was chosen to be the dose per unit BSA for the BSA based dosag e in the simulation. In the middle 2 quartiles, the body surface area and body weight dosing adjustment yielded similar PK profiles. Interestingly, the s imulated PK profiles of the BSA based dosing were slightly higher t han body weight based dosing, and th e opposite was observed in the higher quartile. The study also conclud ed that the current body weight based dosing in patients weighing equal to or less than 23 kg may have less drug exposure compared to patients weighing more than 23 kg. But the pharmaco kinetic difference of etanercept was not known to lead to clinical difference in JRA patients. Enfuvirtide Enfuvirtide is approved for HIV treatment in adults and dosage recommendations exist for children ages 6 years or older. The safety and efficacy stu dy of 2.0 mg/kg (maximum 90 mg) subcutaneous enfuvirtide twice daily for 48 weeks was conducted in 52 treatment experienced, HIV 1 infected pediatric patients (ages 3 to 16 years) 187 There was no significant difference observed in enfuvirtide meanSD pharmacokinetic param eters in children (n=12, ages 5 to 11 years) and adolescents
91 (n=13, ages 12 to 16 years): steady state C max was no meaningful difference in the pharmacokinetic values between children and adolescents. In treatment experienced HIV 1 infected chil dren (3 12 years), 60 mg/m 2 subcutaneous enfuvirtide twice daily reported mean single dose AUC0 12h of 56.4 to AUC0 BID 90 mg enfu virtide 188 189 The pediatric study showed that the body weight adjusted dosing in children was independent of age, body weight, body surface area, and sexual maturity. In a population pharmacokinetic analysis study by Zhang et al., 43 patients (20 adolescents and 23 children) were included, mean age was 11 years, and mean body weight was 35.7 kg. 113 Body weight was a covariate for CL/F but not V/F. The population parameters CL/F, V/F, and Ka for a 33 kg patient were 1.31 L/h, 2.31 L, and 0.105 h 1, respectively. Age did not seem to affect the enfuvirtide exposure. This analysis approves the body weight based enfuvirtide dosing in pediatrics. In HIV 1 infected adults, enfuvirtide reported a small volume of distribution (5.48 L), low clearance (1.4 L/h), and high plasma pro t ein binding (92%). Body weight based dosing (2 mg/kg BID) provides similar pharmacokinetic profiles to th ose observed with 90 mg BID 190 Pharmacokinetic parameter CL/F (1.31 L/h for a 33 kg patient) from the pediatric study is comparable to the reported value from a previous study in HIV 1 infected pediatric patients (CL = 1.42 L/h and F = 0.90 for a 21.3 kg patient) 20 and also comparable to the adult population analysis with CL/F of 1.82 L/h for a 70 kg male patient an d 1.45 L/h for a 70 kg female patients 191 The mentioned pediatric enfuvirtide study by Soy et al
92 involved 26 children (mean age 8.2 yea rs and range 4.0 12.1 years) 99 Patient weight was found to have an effect on CL and V, but the effect was not statistically significant. ed in adults by Zhang et al 190 Additionally, in the plot of CL (L/h) vs. weigh, even the data covers a large weight range, but it does not seem to necessarily capture all differences between childr en and adults. However, if plotting the Soy et al data as CL/kg (L/h/kg) vs .weight, the trends seem to be decreasing, but it is unknown whether the trend is statistically significant. L Asparaginase In a study of pediatrics with acute lymphoblastic leuk emia, 271 patients were given 500, 750, 1000, and 250 0 IU/m 2 PEG L Asparaginase 192 After (1 to 17 years) nor the body surface area had any influence on the distribution of Asparaginase activity. The study concluded that normalization of dose based on body surface area was appropriate in the pediatric patients studied. A statistical analysis u sing linear regression was done to compare chemotherapy dose modifications in obese and non obese pediatric patients with acute l ymphoblastic leukemia (ALL) 193 Obese ALL children were reported to have a 7% decrease in the mean relative modification of L asparaginase compared with non obese children. The result was statistically significant even after taking into consideration gender, age, race, and study cent er. It is found that the difference of dose modifications was gre ater among older children ( 10 18 y ears) than small children ( 2 9 years). It is pointed out that the obesity driven dose modification among older children is possibly because of higher BSAs an d the chemotherapy doses.
93 Summary Most of the studies in the current review showed that body weight or BSA dose adjustment produced comparable exposure for proteins and peptides. However, not all pharmacokinetic studies result in promoting dosing adjustme nt. For basiliximab, a fixed dose of two 10 mg doses for patients less than 35 kg and a fixed doses of two 20 mg dose for patients more than 35 kg wa s recommended for pediatrics 112 Children should pediatric patients cannot always be explained by changes in body size. Simply adjusting dose linearly according to the body weight/BSA c annot always achieve desirable exposure in pediatrics. Anderson and Holford have proposed that growth and development can be evaluated using readily observable demographic information such as weight and age 194 196 Weight was suggested to be an essential covariate for determining dose in pediatrics. The range of body weights in children is much greater than that in adults and can range 200 fold (0.5 100 kg). An established framework was believed to support the allometry used in pediatric pharmacokinetics. The coefficient exponent of body weight/typical body weight was suggested to be 0.75 for clearance and 1 for volume. Fat free mass may be better than total b ody weight when variations in fat affect body composition. A sigmoid Emax model was used to describe gradual maturation of clearance from small children to adults. Future issues were suggested in pediatric pharmacokin etics and pharmacodynamics 191 1) Determination of in vivo maturation of clearance enzyme pathways; 2) Analysis of the placenta concentration to total clearance; 3) Investigation of elimination pathway triggered by birth; 4) Understanding the impact of hormonal cha nges on clearance pathway in adolescents; 5) Refining
94 PBPK models for children; 6) Further understanding of pharmacodynamic difference between children and adults. Overall, the difference of pharmacokinetics of proteins and peptides in pediatric patients is due to catabolic enzymes, changes in body composition, elimination organs, and receptor mediated endocytosis. The differences lead to the changes in volume of distribution, clearance and absorption of proteins and peptides. The above factors can be dram atically affected by body weight, BSA, height, age, and these covariates may be highly correlated and not mutually exclusive. Due to the complexity of the contributors evolved, the direction and extent of the difference are not always readily predictable. Clearance and volume of distribution of proteins and peptides can be higher but also lower when the comparisons are done in children and adults or younger children and older children. In the current review, most of the proteins and peptides show a more rap id body size adjusted clearance (e.g., L/h/kg) in children than in adults, such as alemtuzumab, epoetin, factor VII, factor VIII, and factor XI, while both absolute CL (L/h) and body surface area adjusted CL (L/h/m2) of gemtuzumab are smaller in infants in adults. Enfuvirtide does not have consistent conclusions from different studies. One pediatric study showed that the pharmacokinetics of body weight adjusted enfuvirtide in children was independent of age, body weight, body surfac e area, and sexual maturi ty 187 but from the figure of CL (L/h) vs. weight reported by Soy et al, the decreasing CL/body weight does not seems to support the body weight adjusted dose of e nfuvirtide. Some of the studies showed that body size adjusted dose for certain proteins and peptides produce comparable exposure in children and adults, and the
95 pharmacokinetics of these products are not affected by age, for example, infliximab, cetuximab drotrecogin alfa, L Asparaginase. Though there is not much obvious similarities for drugs that should not follow simple body size linear adjustment, quit a few monoclonal antibodies are among them. This may due of the fact that monoclonal antibodies o ften are reported to have nonlinear pharmacokinetics. Basiliximab has less PK variability if use 10mg (weight<35kg) and 20mg (weight>35kg) in pediatrics. Daclizumab tends to underdose younger patient and over dose larger children. Alemtuzumab and natalizum ab underdose children and followed by not desired clinical outcomes. Eating disorders such as anorexia and bulimia are rising in adolescent girls in the United States. On the other side, the rate of obesity in adolescents is also increasing. Anorexia rel ated hospitalizations in children younger than 12 surged 119 per cent between 1999 and 2006 197 As the pharmacokinetic parameters may be even more complex, simple body weight adjusted dose might not be suitable for such particular pop ulation. The finding from this review suggests the need to continue the study of proteins and peptides in this particular population, and mechanism based population pharmacokinetic and pharmacodynamic models with consideration of body size and maturity mig ht be helpful in explaining and extrapolating the pharmacokinetics and pharmacodynamics of the studies. Dose adjustment in pediatrics should lead to not only consistent exposure compared with adults, but also decreased intersubject variability in the expo sure; only then does it make sense to apply the adjustment.
96 Table 3 1. Total body water change by age Age Total body water (%) Extracellular fluid (%) Intracellular fluid (%) Fetus (<3 months) 90 65 25 Neonate (Premature) 85 50 35 Neonate (Full term) 75 40 35 Infant (4 6 months) 60 23 37 Adolescent 60 20 40 Adult 60 20 40 Table 3 2. Tissue distribution comparison of newborn and adults (Organ weight expressed as % of total body weight) Organ Newborn Adults Muscle 25 40 Skin 4 6 Heart 0.5 0.4 L iver 5 2 Kidney 1 0.5 Brain 2 2 Table 3 3. Renal function: glomerular filtration rate (GFR) and renal plasma flow ( RPF ) by age Age GFR (mL/min) RPF (mL/min) 1 10 days 15 45 20 125 1 month 30 60 100 400 6 months 50 100 400 500 1 years 80 120 500 600 1 70 years 80 140 500 700 70 80 years 70 110 250 450 80 90 years 45 85 200 400
97 Table 3 4. Pharmacokinetics of selected proteins and peptides in pediatrics Generic Name Class Route Pharmacokinetics Alemtuzumab mAbs i.v. More rapid clearance in chi ldren than in adults. Basiliximab mAbs i.v. CL (ml/h) in infants and children is about half that of adults. Use 35kg as a cut off weight for 10 or 20mg in pediatrics. Bevacizumab mAbs i.v. BW based dose exhibits similar PK parameters in children and adu lts, and large variability in both populations. Cetuximab mAbs i.v. Dose dependent nonlinear elimination. BSA based dose provides similar exposure in children and adults, and age has no effect on PK. Daclizumab mAbs i.v. The 4.2 fold range in CL, 7.4 fol d range in V are less proportional than a 12 fold range in body weight Darbepoetin Alfa Growth factor i.v., s.c. The lack of dose proportionality is likely due to pediatric population rather than nonlinear PK; neonates have a shorter half life, larger V a nd CL than children. Drotrecogin alfa Blood factor i.v. Weight normalized clearance decreases significantly with age in patients <18 years old. Enfuvirtide Peptide s.c. One study justified body weight (BW) based pediatric dosing. Epoetin Alfa Growth factor i.v., s.c. CL (mL/h/kg) and bioavailability in pediatrics were two fold of that in adults. Epoetin Delta Growth factor i.v., s.c. BW adjusted PK parameters are similar in children and in adults. Etanercept Fusion protein s.c. The analysis justifi ed the body weight based dose adjustment for etanercept in JRA patients; gender difference was reported both in children and adults. Exenatide Incretin s.c. The max recommended adult dose instead of half of the max dose was suggested to be explored in ad olescent patients. Factor VII Blood factor i.v. Total body weight normalized clearance was significantly faster in children than in adults. Factor VIII Blood factor i.v. BW adjusted clearance in mL/h/ kg and Vss in L/ kg seems to decrease with age. Fact or VIX Blood factor i.v. Higher weight adjusted CL in children than adults. Filgrastim Growth factor s.c. ANC adjusted G CSF dosing adjustment might improve PBPC mobilization in pediatric patients. Gemtuzumab mAbs i.v. Both faster CL (L/h) and CL (L/h/m2 ) in adults than children and infants.
98 Table 3 4. Continued Generic Name Class Route Pharmacokinetics Humatrope Growth hormone s.c. No significant effect of weight and age on humatrope pharmacokinetic parameters. Infliximab mAbs i.v. BW based dose p rovides similar exposure in children and adults; PK of infliximab does not differ as age increases. Insulin aspart Insulin s.c. In pediatrics, insulin aspart had a quicker onset than human insulin; aspart has a higher exposure in adolescents than in chil dren. Insulin detemir Insulin i.v. Less PK variability in insulin determir than glargine. Insulin glulisin Insulin i.v. The profile of insulin glulisine is similar for children and adolescents, whereas human insulin exhibits higher level in adolescents. Interferon Interferon s.c. Higher drug exposure in pediatrics; wide intersubject variability suggests further individualized dosing. Interferon Interferon i.v. BSA based PK parameters in pediatrics is about twice that in adults. Interferon Interferon i.v., i.m. No BW/BSA or age effected was discussed. Slightly lower exposure in pediatrics than in adults. Interleukin Cytokines i.v., s.c. Higher rhIL 11 clearance in pediatrics than adults Asparaginase Enzyme i.m, i.p. After adjusting dose by BSA, neither age nor the BSA had any influence on the distribution. LB03002 Growth hormone s.c. Body weight adjusted dosing gives comparable exposure in pediatrics as in adults. Natalizumab mAbs i.v. BW base dose tends to underdose adolescents. Nutropi n Growth hormone s.c. Drug exposure was approximately proportional to the dose. Palivizumab mAbs i.v., i.m. BW based dose for palivizumab, but body weight effect not discussed; no significant clinical outcome between placebo, 5 and 15 mg/kg were observed. Somatropin Growth hormone Inhaled No significant effect of weight and age on somatropin pharmacokinetic parameters. Zomacton Growth hormone s.c. No BW/BSA or age correlation was analyzed for its pharmacokinetic parameters. A bbreviations: i.v.: intrave nous, s.c.: subcutaneous, mAbs: monoclonal antibodies, CL: clearance, PK: pharmacokinetics, BSA: body surface area, V: Volume of distribution, JRA: juvenile rheumatoid arthritis, ANC: absolute neutrophil count, G CSF: granulocytecolony simulating factor, P BPC: peripheral blood progenitor cell, rhIL 11: interleukin 11.
99 Figure 3 1. (a) An example of fixed dosi ng (b) An example of body size based dosing (c) An example of fixed dosing by different age groups or different body size groups
100 CHAPTER 4 FIX ED DOSING VERSUS BODY SIZE BASED DOSING OF THERAPEUTIC ANTICANCER DRUGS IN ADULTS Background The classic cytotoxic anticancer molecules are widely known for their narrow therapeutic window. It was perceived that since patients with larger body size have a bigger metabolizing capacity, faster clearance, and a larger volume of distribution, a higher dose should be given to larger patients. As a result, anticancer drugs are usually and it was believe d that BSA based dosing provide s precise and accurate dosing even t h ough there was no solid scientific support BSA was used mainly for the purpose of allometric scaling at the beginning. In the 1950s, BSA was introduced to dose oncology d rugs in pediatric patients 198 In 1958, Pinkel first propose d that the maximum tolerated dose (MTD) when adjusted by BSA (mg/m2) was similar in different animals and human s 7 The anticancer agents Pinkel reviewed were mechlorethamine, methotrexate, 6 mercaptopurine, actinomycin D and triethylenethiophosphoramide (thiotepa). In 1966, Freireich et al compared the observed human MTD with the predicted MTD by preclinical animal data for 18 anticancer drugs, and conclu ded that the MTD expressed as mg/m2 in animals accurately predicted the MTD dose in human. Although Pinkel and Fr eireich suggested BSA based dosing be used to determine the MTD, nei ther publication recommended BSA based d osing in dose escalation in Phase I studies. Without soli d scientific investigation, BSA based dosing was then ad opted in adults Phase I trials, and then carried through Phase II and III studies, and ultimately used in the approved labeling for most anticancer drugs
101 Body surface area is d if ficult to measure directly; therefore it needs to be estimated using formula s that incorporate body weight and height in the calculation. The original formula to calculate BSA was first developed by DuBois and DuBois in 1916 from a study that enrolled on ly nine subjects 199 : BSA=0.20247 x Height (m ) 0.725 x Weight (kg) 0.425 The nine subj ects weigh ed 25 to 90 kg. A mold was made of their body, and was cut into small flat pieces, and then the surface area was calculated. The study included only one child, so the prediction of pediatric BSA was beyond the range of this formula. One commonly used formul a was published by Mosteller 200 The major formulas to calculate BSA are summarized in Table 4 1 In 1970, Gehan and George derived the formula BSA=0.02350 x Height (cm) 0.422246 x Weight (kg) 0.51456 which work s well both in adults and children. Gehan and George also validated the original Dubois formula in more than 401 subjects, and the Dubois for mula was reported to overestimate the BSA by more than 15% in about one fifth of the people, while underestim ating only 1% of the people 201 This finding indicat ed the lack of accuracy of the DuBois formula to determine the BSA of cancer patients and to further individualize dose for cancer patients. This fin ding however, surprisingly did not result in the Gehan and George BSA formulas to replace the Dubois furmul a as the medical standard. It was not until 1987, when Mosteller has modified and simplified the original formula, was it commonly used by medical professionals : BSA= Height (cm) x Weight (kg) ]/3600) 200 Moste ller formula has been extensively used in oncology to calculate dosage for many chemotherapy drugs However, the use of BSA formula in obese patients is still uncertain. In the current clinical use, many patients are assigned a BSA
102 of 2 m 2 when patients have actual BSA larger than this value 202 Though we have listed several formulas, the c orrelations among the formulas were reported high (r> 0.97) indicating very small diff erences 203 However, significant difference s of BSA values was observed for overweight and obese adults between the DuBois formula and other formulas 203 This stud y compared different formulas, and reported that for overweight adults, the DuBois formula underestimated the BSA by 3% for male and 5% for fem ale when compared to Mosteller formula. N owadays, a 3D scan can be used to determine BSA using high technology tools 204 In 1958, Pinkel proposed that BSA correlates to the pharmacokinetics of the anticancer drugs better than body weight 7 BSA is believed to provide better individualized medication in daily practice, but many other factors can affect systemic exposure of anticancer drug concentration in patients. These factors includes the e.g. gender, race, age, weight, height, menopausal status, pregnancy), genetic diffe rences ( e.g. polymorphisms in metabolizing enzymes and transports), disease related characteristics ( e.g. tumor type, cancer stages, surgery, liver function, renal function, albumin and alpha 1 acid glycoprotein levels), comedications ( e.g. antibiotics herbal supplements, over the counter medications), and patient life style ( e.g. adherence to medications, food, alcohol, smoking, coffee, exercise, stress). Body size ( body weight or BSA) is only one of those factors that affect s the interindividual vari ability from patient to patient, and simply just linearly adjust ing dose accordingly to BSA may lead to reduced, unchanged, or even increased intersubject variability in cancer patients. For anticancer cytotoxic drugs, after the BSA dose adjustment, the in tersubject variability, if expressed as coefficient of variation
103 (standard deviation divided by the mean and multiply by 100), is often still in the range of 25% to 70% if not more 205 As ea rly as 1990, Crochow et al. has questioned the dose adjustment by body surface area 198 Baker et al. reported in a retrospective study that for the 33 investigational anticancer agents, BSA based dosing significantly reduce d the interpatient variability in clearance for only 5 drugs, such as docosa hexaenoic acid paclitaxel, 5 fluorouracil/eniluracil, paclitaxel, temozolomide, and troxacitabine. It was reported that for some anticancer agents, BSA based dosing reduced intersubject variability between the ranges of 15% to 35%; as a result, BSA can onl y explain up to one third of the total interindividual variability. The objectives of this study are 1) for the first time utilize population pharmacokinetics approach to systemically evaluate intersubject variability for anticancer drugs 2) provide a mode l based analysis method on dosing strategies for small oncology drugs under development and 3) recommend a dosing strategy for clinical trials conducted in adults at different stages of the development of anticancer drugs. M ethods Data Collection Data use d in t he current simulation were collected from the population PK/PD studies of anticancer drugs published. The selection criteria included the availability of population PK models for adult patients and assessment of the body size effect on the PK paramet ers.
104 Population PK Models The population PK and/or PD models of the selected anticancer drug (see Table 4 1) were obtained from published journal articles General properties of these population PK models and the body size effect such as BW or BSA, on the PK parameters are summarized in Table 4 2. Mixed effect models were used to describe the PK of all the selected small oncology drugs The jth observation for the ith individual was given by i is a set of PK parameters for the 1ij 2ij are the residual 1 2 2 2 i can be further described by where z i is a set of fixed effects on the PK par i is the intersubject variability following a normal 2 Simulation analysis was investigated using NONMEM (version VI; GloboMax, Hanov er, Maryland). PK Simulation Population performance. Simulations to evaluate population performance of two dosing approaches were investigated the same way as previously described 56 Briefly, Monte Carlo simulation was conduc te d using the final PK model reported for each anticancer agent to generate the concentration time profiles following both fixed dosing and body size based dosing approaches. The dose used for simulation was one of the dose s recommended in the labeling for m arketed products. The median value of body
105 size (BW or BSA) was used as the assumed standard for dose determination so that the dose used in the fixed dosing dataset is the same as the dose for the participants with median (typi cal) body size in the body s ize based dosing dataset For all simulation studies, 1000 subject s were simulated for each dosing approach. The sampling points were chosen based on the PK properties of the anticancer drugs, often the reported sampling points in the selected PK model art icle is used The exact same sampling schedule was used for both fixed and body size based dosing approaches. For each simulation study, values of the covariates were randomly generated using S PLUS 8 .0 (TIBCO, Palo Alto, California) assuming normal, log normal or binomial distribution. The values of parameters used for generating these covariates were selected by trial and error, with a goal of reproducing the patient population by matching the median, standard deviation, and/ or the range of covariates to those reported in the corresponding population PK/PD study. Individual performance. To evaluate the individual performance, the PK profiles were simulated for the participants with typical, low extreme, and high extreme body size. The typical, low extreme and high extreme BW/BSA used in this study were 75.7 kg/1.8 m 2 40 kg/1.3 m 2 and 140 kg/2.3 m 2 respectively. For covariates other than BW/BSA, typical values were used. The intersubject variability and residual errors were all set to zero for the simul ations conducted for both dosing approaches. The reason for fixing other covariates values to be the same, and inter and intra subject variabilities to be zero, is that the exposure difference among the typical, low extreme and high extreme body size pat ients is only attributed to the difference of dosing approaches.
106 Calculation of Values A simple way to assess wh ether fixed dosing or body size based dosing may be better in reducing intersubject variability in AUC is to evaluate ( the powe r function of the equation CL=TVCL x ( individual BW/ median BW ) of the body size effect on CL. However, the covariate models used to characterize the effect of body size on CL for the selected anticancer drugs are not all in the form of this equation (Ta ble 4 2 ). Therefore, a translation all anticancer drugs proces s were used to obtain calculate a series of CL va lues over the range of BW or BSA reported based on the reported original paper with PK model reported and (2) fit the generated CL versus b ody size data using equation CL=TVCL x (individual BW/median BW) M ountain View, California). AUC calculation The AUC for each subject was calculated as dose/CL if the molecule presented line ar PK. When the molecule present ed nonlinear PK, the AUC was calculated by integration of the concentration time curve by the trape zoidal rule using S PLUS 8.0 (TIBCO, Palo Alto, California) PK C max determination. The maximum concentration ( C max ) for each individual was calculate d as the maximal concentration from the simulated concentration time profile of the subject Performance Evaluation The performance of the two approach es in terms of reducing intersubject variability in AUC was relationship based on the following criteria:
107 fixed dose is better. fixed and body size based dosing are similar. body size based dosing is better. These criteria apply to both population and individual performances. To be consistent with chapter 2 the comparison of simulation studies of the 2 dosing approaches were presented as described below. Population perfor mance was evaluate d by comparing the intersubject variability (expressed as % coefficient of variation [CV %] ) in the exposure (AUC and C max ) of 1000 subject s simulated following both fixed and body size based dosing approaches Coefficient of variation of the 1000 subjects for each anticancer drug is calculated by ratio of the mean o f the 1000 subjects and the standard deviation of the 1000 subjects. A good dosing approach should provide consistent exposure for the whole population. The dosing ap proach that produced less intersubject variability should be a better dosing approach for the patient population Individual performance was evaluated by comparing the percentage difference in the C max or AUC between subject s with extreme low/high body siz e and typical body size following the 2 dosing approaches. The dosing approach that provides a smaller percentage difference in PK exposure between individual s with extreme body size and those with typical body size has a better individual performance. Con tribution of Body Size Effect to Overall Intersub ject Variability of Relevant PK Parameters A simulated PK data set of 100 participants was generated using the published levant covariate values were randomly generated as described above. This simulated data set
108 body size as covariate(s) for any PK parameters. The percentage change in the i the overall intersubject variability of the relevant PK parameters. Comp arison o f BW and BSA based D osing at Population Level Once we confirmed that body size based dosing is suitable for a particular anticancer drug, how do we decide which body size to be used for dosing calculation? An example simulation is given to compare bo dy wei ght and body surface area based dosing, since these two are the most common dosing approaches for anticancer drugs. M onte Carlo simulation was conducted using the final PK model reported for each anticancer agent to generate the concentration time profiles following both body weight based dosing and body surface area based dosing approaches. The dose used for simulation was one of the doses recommended in the labeling for marketed products. Each simulation is done with one dataset. Body weight based d osing and body surface area based dosing methods have separate datasets. The only difference between the two datasets is the dosing information. The median value body weight subject in the body weight data set is assumed to be given the same dose as the median v alue body surface area in the body surface area dataset. For all simulation studies, 1000 subjects were simulated for each dosing approach. The sampling points were chosen based on the PK properties of the anticancer drugs, often the reported sampling poin ts in the selected PK model article is used. The exact same sampling schedule was used for both fixed and body size based dosing approaches.
109 For each simulation study, values of the covariates were randomly generated using S PLUS 8.0 (TIBCO, Palo Alto, C alifornia) assuming normal, lognormal or binomial distribution as described in the population performance section If there are two covariates related to each other such as body weight or creatinine clearance, both of the covariates are ranked from low to high. Individual simulation for BW and BSA is done the same way as described above for fixed dosing and body size based dosing approaches. Performance e valuation of BW and BSA based dosing Cisplatin was chosen as an example to demonstrate the compariso n of BW and BSA based dosing. BW and BSA population data set were both divided in to four groups according to the BW or BSA value ( and ) Take BW for example, if the BW range is 40 140kg, then BW could be divided into four groups: and 75kg 75kg > and 100 100kg In order to compare BW and BSA based dosing, it was assume d that the person with median BSA and the person with median BW receive the same amount of total dose from the two datasets. The cut off values for each group depends on the range and distribution of the body weight For each group 5 th and 95 th quartile o f each group were plotted for both BW and BSA. Visual comparison of the two profiles can be made for BW and BSA based dosing. The similar analysis at population level and individual level as mentioned above for all the selected drugs are also made to comp are BW and BSA based dosing. R esults and Discussion Data Collection A total of 2 8 anticancer drugs were selected based on the data collection criteria specified in the Methods section (Tables 4 2 ) The effect of body size on various PK
110 parameters, includi ng CL, intercompartment clearance (Q), volume of the central compartment (V or V1), and volume of the peripheral compartment (V2), is summarized in Table 4 2 Body size has been found to be covariate(s) of 1 or m ore of the PK parameters for 20 anticancer d rugs (Table 4 3 ). It should also be noted that for i fos famide, although body size had been found not to be covariates of its PK parameters, the population PK models were developed based on data from only 24 patients (Table 4 3 ). One should be cautious wit h the interpretation of the results. Among these 28 selected anticancer drug s, nineteen are administered based on body size in adult patients (Table 4 2 ). Interestin gly, for some products that are administered using body weight based dosing, such as cladri bine, cyclophosphamide and Thiot epa (Table 4 2 ), body size measures (BW or BSA ) had not been shown to be a covariate of their PK parameters (Table 4 3 ). Similarly, etoposide, gemcitabine, ifosfamide, and melphalan are given per BSA, body size was not foun d significant to be a covariate for their PK models as well. Dosing Approach Performance Performance evaluation based on AUC As discussed in the chapter 2 the performance of AUC for the two dosing function as defined in E quation ( 2 cutoff point. T he exponent of power function n CL were obtai ned for all the 2 8 selected anticanc er drugs and listed in Table 4 3 The were directly recorded from the published models if they are available or obtained in such a way as described in the Methods For
111 anticance r drugs where body size was not found to be significant covariate of CL a zero 4 3 1 5 of 2 than 0.5, and 1 3 han 0.5. These results suggest that fixed dosing would perform better for 1 4 based dosing would perform better for 1 4 The only exception is Melphalan. Overall, t he two dosing approaches would perform not too different ly across all the molecule s The results of the simulation studies for comparing the performance of two dosing approaches at the population level are presented in Figure 4 1 A. Consistent with the 14 presente d less i ntersubject variability in AUC when body size based dosing was adopted, whereas the other 1 4 AUC when fixed dosing was used, w ith only one exception, melphalan For melphalan, etoposide and m ethotrexate body size was not, but creatinine clearance was a covariate of CL or V For melphalan, etoposide creatinine CL was a covariate for CL, and for m ethotrexate creatinine CL was a covariate for V. Both e toposid and m ethotrexate the population AUC variabil ity is presents less variability. However for melphalan, body size based dosing presents less va riability AUC at population level The reason might be that creatinine CL is highly related to the body weight of in the melphalan study. For the BSA simulation dataset, both BSA and creatinine CL were ranked from low to high for all the subjects. The study did not mention the relation bet ween BSA and creatinine CL. If there is high relation between the two, then BSA is likely to provide less interindividual variability to the AUC
112 across the population. In the other hand, for e toposid and m ethotrexate body size is likely not highly related to creatinine CL in these two study; th us fixed dosing provides less interindividual variability for AUC. Similar results were also obtained for individual performance (Figure 4 2 A). For most of the anticancer drugs investigated, the dosing approach that had better population performance also had better individual performance. The only exception is melphalan. The difference between population and individual AUC for melphalan might be because of the creatinine CL values. In the population dataset, creatinine CL ranges from 30 to 195 ml/min Howe ver, for the individual dataset, creatinine CL was randomly assigned in order for low extreme, high extreme and typical bo dy size patients, and the value range of creatinine CL is not as wide as the body size range. This mig ht be the reason for melphalan t o have different population and individual variability results. It should be noted that the zero difference in AUC between patients with extreme body size and typical body size following a fixed dose for capecitabine, cladribine, cyclophosphamide, dexameth asone, erlotinib, etoposide, everolimus, gemcitabine, ifosfamide, imatinib, irinotecan, methotrexate, pemetrexed, and Thiotepa is a result of the lack of BW /BSA effect on their CL (Figure 2 2 A). from the the optimal dosing approach if AUC is the exposure parameter of the main concern. However, when other body size related measurements are the covariates for CL such as creatinine CL, one should be more cautious
113 Performance e valuation b ased on C max The population and individual performances of the two dosing approaches based on C max were evaluated by simulation studies and shown in Figures 4 1 B and 4 2 B, respec tively. At the population level, body size based dosing resulted in less intersubject variability in C max for 14 of 2 8 anticancer drugs whereas fixed dosing produced less variability in C max for the other 1 4 compounds (Figure 4 1 B). At the individual leve l, body size based dosing produced a smaller percentage difference in C max between participants with extreme and typical body sizes for 1 3 of 2 8 anticancer drugs whereas fixed dosing produced a smaller percentage difference for the other 1 5 compounds (Fig ure 4 2 B). The results from both population and individual level evaluations are again very consistent with the only exception of c apecitabine for which fixed based dosing was shown to have better individual performance but a slightly worse population per formance. The difference between the population and individual CV% for C max of c apecitabine is very little, less than one percent. on V 3 (volume of distribution of the deep peripheral compartment). The reaso n for this inconsistent resu lt might be because that the body size effect on volume of distribution of the deep peripheral compartment is dil uted, so the effect is not as obvious as for the central volume of distribution. As shown for the population performance, there is almost no di fference between fixed dosing and body size based dosing. It was noted that body size based dosing tends to overdose patients with large body size and underdose patients with small body size. The opposite is true for fixed dosing, that is, overdose patient s with small body size but underdose patients with large body size (Figure 4 2 A, B).
114 Contribution of Body Size Effect to Overall Intersubject Variability in PK Parameters The contribution of the effect of body size to the overall intersubject variability o f relevant PK parameters was evaluated for 19 anticancer drugs with body size as a covariate and the results are summarized in Table 4 4. It was observed that the effect of body size had a small and, in some cases, moderate contribution in a few cases, b igger contribution to the overall intersubject variability of major PK parameters, ranging from 3.32 % to 60.9 % for CL and from 0. 55 % to 75 5 9% for V/V1 (Table 4 4 ). It was observed that BSA or body weight can explain a relatively big ger portion of the inte rsubject variability for busulfan, cisplatin and thalidomide. BSA/BW contributes to 39.18%, 59.55%, and 60.90% intersubject variability of CL for busulfan, cisplatin and thalidomide, respectively. The contributions to volume of distribution for these three compounds are 75.59%, 45.24% and 48.44. BSA/BW also explains 38.28% and 35.67% intersubject variability of CL for imatinib and paclitaxel. For these compounds, body size seems to be a major source for intersubject variability, body size based dosing may p rovide a clinical benefit when supported by other factors, such as a narrow therapeutic window. When body size only explains a very small percentage of the intersubject variability for example, in the case of 5 Fluorouracil, doxorubicin, and oxaliplatin a djusting the dose based on body size would lead to a minimal reduction in the variability in AUC Performance of BW and BSA based dosing After body size based dosing is confirmed to be better than fixed dosin g, the next step is to find out whether body weig ht or BSA is appropriate for dosing adjustment. The evaluation method to compare BW and BSA based dosing was introduced in Methods
115 section. Figure 4 3 shows the example of cisplatin, the 5 th and 95 th percentile of the concentration time profiles were plott ed for four groups according to the four quartiles of BW or BSA value ( It is very obvious that the middle two quartiles, there is tiny difference between the profiles after BSA based dosing and BW based dos ing (Figure 4 3 B and C) For the fourth quartile, BSA based dosing seems slightly underdose the patients compared to BW based dosing (Figure 4 3 D) In the other hand, there is a little larger difference between the two dosing approaches in the first quar tile body size group but t he different is not dramatic either (Figure 4 3 A). A comparison of CV% of AUC and C max at population level between BW and BSA based dosing is shown in Figure 4 4. BW based dosing showed less CV% for both AUC and C max at popula tion level (Figure 4 4A) The similar observation was found at individual level, BW showed less % difference for both AUC and C max (Figure 4 4B) In summary, the result show s little difference between BW and BSA based dosing when visually comparing the 5 th and 95 th concentration time profiles. BW based dosing was showing to have less intersubject variability at both population and individual level than BSA based dosing. Relationship between the Class of Anticancer Drugs and Body Size Effect on Pharmacokinet ics The 28 anticancer drugs selected in this study fall in a variety of drug classes. There are 8 alkylating antineoplastic agents, 4 topoisomerase inhibitors, 3 antimcrotubule agents, 2 tyrosine kinase inhibitors, 2 mTOR ( mammalian target of rapamycin ) in hibitors, 1 nucleoside analog, 1 DNA crossliker. The rest of the anticancer agents are either inhibitors of certain enzymatic pathway or involving in multiple
116 pathways. There is no apparent cor relation between the drug class and t body size effect on CL or V There is also no apparent correlation between the drug class and the performance of AUC and C max at either population level or individual level. Among the drugs that body size contributes to relatively larger per cent age of intersubject variability there is no trend observed in terms of drug classes. Overall, there is no observed correlation between the drug class of an anticancer drug and the body size effect on the pharmacokinetic parameters. Discussion and Sum mary Baker et al. has conducted analysis evaluating the role of body surface area in dosing anticancer drugs 206 They defined the criteria as 1) a linear regressi on coefficient between BSA and CL (R) .50; 2) P<.01; and 3) a relative reduction in the variability of clearance .15%. They have concluded that b ody surface area based dosing statistically significantly reduced interpatient variability in drug clearance for only 5 of the 33 agents: docosahexaenoic acid (DHA) paclitaxel, 5 fluorouracil / eniluracil, paclitaxel, temozolomide, and troxacitabine. Our results presents a lower intersubject variability in AUC at both population and individual level s for 5 fluor ouracil for BSA based dosing Our result of p aclitaxel and temozolomide is also in consistent with showed less intersubject variability for AUC and C max at both population and individual level s for BSA ba sed dosing. Interestingly, temozolomide was approved with a fixed dosing on its labeling and more than one study have shown that CL increased with BSA for both gender, and BSA should be used for dosing temozolomide 207 208 Surprisingly, in our study, body size is only showed to be able to explain 2.24% intersubject variability of temozolomide CL. Our results show that BSA contributes to 35.67% intersubject variability of paclitaxel CL.
117 BSA normalized dose showed bene fits for paclitaxel. It was reported that the distribution of paclitaxel in the blood strongly depends on its formulation vehicle (Cremophor EL dehydra ted ethanol USP; Bristol Myers Squibb, Wallingford, CT) 209 due to the fact that paclitax el has strong affinity to Cremophor EL in the blood. Blood volume has long been demonstrated to correlate with BSA in 1986 210 Sparreboom evaluated the disposition of Cremophor EL and reported that Cremophor EL has a volume of distribu tion that is similar to the blood volume and body surface area is a significant covariat e for Cremophor EL clearance 211 Therefore, BSA effect on PK parameter s of pac litaxel is likely related to the affinity of paclitaxel to its vehicle Cremophor EL 212 and the distribution of paclitaxel depends on the distribution of Cremophor EL which is linked to blood volume. Previous studies showed that there is no correlation between BSA and busulfan CL when busulfan is given by i.v. administration 213 but BSA has significant influence on busulfan CL when busulfan is given orally 214 Our result of the s imulation shows that after busul fan administration, body weight based dosing provides less intersubject variability for AUC and C max at both population and individual level. The PK model for busulfan contain s body weight as a covariate for both V and CL Body weight contributes to 39.18% of intersubject variability of CL and 75.59% intersubject variability of V. Overall, our results supports body size based dosing for busulfan. Cisplatin was found to have low er intersubject variability for AUC and C max in our study at population level, but difference is minimized. In our study c isplatin was also found to decrease intersubject variability for AUC and C max at individual level. Additionally, for cisplatin, BSA co ntributes to 59.55 % of intersubject variab ility of CL and
118 49.25 % intersubject variability of V. A previous study by de Jongh concluded that body surface area based dosing does not increase accuracy of cisplatin exposure 215 I t is debatable for docetaxel whether BSA has a clinically meaningful effect on its clearance. value sma ller than 0.001, and our result s s how that less intersubject variability for AUC and Cmax at both population and individual level with BSA based dosing 206 However, the intersuject variability that attributes to BSA among patients is only 3.5% 206 and the impact of transaminases and alkaline phosphatase levels on CL has be en shown to be more clinical ly releva nt 216 Our finding for c apecitabine aligns with previous study that BSA has no influence on CL of c apecitabine 217 Both our finding and another study supports the conclusion for c yclophosphamide that neither BSA nor body weight has correlation with the clearance of the c yclophosphamide 218 A study has shown that BSA has no correlat ion with CL for methotrexate 219 which is in agreement with our results. The impact of body size on irinotecan clearance has been studied by Mathijssen 220 This study examined 82 patients, and BSA normalized clearance exhibited higher intersubje ct variability. The metabolite of irinotecan was also studied, and BSA normalized clearance for the metabolite also showed higher intersubject variability. Mathijssen recommended that alternative dosing strategies should be studied for irinotecan. Our resu lt for irinotecan indicates that fixed dosing provides less intersubject variability for AUC and C max at both population and individual levels.
119 In our study, besides temozolomide, other three anticancer drugs (mitomycin, temsirolimus, thalidomide) that ar e approved for fixed dosing administration present less intersubject variability for AUC and C max when body size based dosing is used in simulation. Especially for thalidomide, body weight was tested to contribute to 60.90% and 48.44% intersubject variabil ity in CL and V, respectively. As a result, for these drugs, further analysis and evaluation needs to be done to find a better dosing strategy. Overall, finding a good dosing strategy is challenging for anticancer drugs. Some lessons were reported in the liter ature: drugs with approved body size based dosing shows no correlation of BSA with their PK parameters, and drugs with approved fixed dosing exhibits a correlation of body size with their PK parameters. The reason for this situation might be a lack of assessment of body size effect on PK or PD of these anticancer drugs. At the end, we recommend fixed dosing to be used for first in human studies for anticancer drugs under development Once data is available, body size effect should be evaluated on PK a nd PD of anticancer drugs. Since most of the small oncology drugs are c ytotoxic agents with narrow therapeutic windows often times, adjustment of dosage is made according to the body size. However, a large intersubject variability can still exist This co uld lead to failure of the treatment or harmful toxicity in the patient population T he goal is to detect any possible factors that could significantly impact the PK and PD of the drugs, and further dosing strategies should be made based on these factors b ut not restricted only to body size. The analysis of covariates that affect PK and PD is extremely important information for the Phase III dose selection. Fortunately, the challeng e of BSA based dosing for anticancer drugs started two decades ago, and
120 nowa days the importance of the analysis has been accepted by more and more professionals.
121 Table 4 1. BSA formulas Year Authors Number of Subjects Formula 1916 DuBois, DuBois 9 (1 Child) BSA (m) = 0.20247 x height (m) 0.725 x weight (kg) 0.425 or BSA (m ) = 0.007184 x height (cm) 0.725 x weight (kg) 0.425 1935 Boyd 231 BSA (m) = 0.0003207 x weight(g) 0.7285 0.0188 log(Weight(g)) x height (cm) 0.3 1970 Gehan, George 401 BSA (m) = 0.0235 x height (cm) 0.42246 x weight(kg) 0.51456 1978 Haycock et al 81 BSA (m) = 0.024265 x height (cm) 0.3964 x weight(kg) 0.5378 1987 Mosteller Unknown or
122 Table 4 2 Selected anticaner drugs and their d osing approaches for adult p atients Generic Name Brand Names Dosing Drug Class 5 Fluorouracil Adrucil, Efudex, Fluo roplex, Carac mg/kg, mg/m 2 T hymidylate synthase inhibitor Busulfan Busulfex, Myleran mg/kg A lkylating antineoplastic agent Capecitabine xeloda mg/m 2 Thymidylate synthase inhibitor Cisplatin Platinol, PlatinolAQ mg/m 2 A lkylating antineoplas tic agent Cladribine Leustatin mg/kg A denosine deaminase inhibitor Cyclophosphamide Cytoxan, Neosar mg/kg A lkylating antineoplastic agent Dexamethasone Decadron, Dexasone, Diodex, Hexadrol, Maxidex mg A nti inflammatory Docetaxel Taxotere m g/m 2 Antimicrotubule agent Doxorubicin Adriamycin PFS, Adriamycin RDF, Rubex,Doxil mg/m 2 T opoisomerase II inhibitor Erlotinib Tarceva mg T yrosine kinase inhibitor Etoposide Etopophos, Toposar, VePesid mg/m T opoisomerase II inhibitor Everolimu s Afinitor mg mTOR inhibitor Gemcitabine Gemzar mg/m 2 N ucleoside analog Ifosfamide Ifex mg/m 2 A lkylating antineoplastic agent Imatinib Gleevec mg T yrosine kinase inhibitor Irinotecan Camptosar mg/m 2 Topoisomerase I inhibitor Melphalan Al keran mg/m 2 A lkylating antineoplastic agent Methotrexate Amethopterin, Rheumatrex, Trexall mg Dihydrofolate reductase inhibitor Mitomycin Mutamycin mg DNA crosslinker Oxaliplatin Eloxatin mg/m 2 A lkylating antineoplastic agent Paclitaxel Onx al, Taxol mg/m 2 Antimicrotubule agent Pemetrexed Alimta mg/m 2 T hymidylate synthase, dihydr ofolate reductase, and glycinamide ribonucleotide formyltransferase inhitor Temozolomide Temodar mg A lkylating antineoplastic agent Temsirolimus Torisel m g mTOR inhibitor Thalidomide Thalidomid mg Multiple pathways Thiotepa Thioplex mg/kg A lkylating antineoplastic agent Topotecan Hycamtin mg/m 2 Topoisomerase I inhibitor Vinorelbine Navelbine mg/m 2 Antimicrotubule agent mTOR : mammalian target of rapamycin.
123 Table 4 3 Population pharmacokinetics/pharmacodynamics (PK/PD) models for the s elected anticancer drugs Genetic Name (CL) (V) Covariate models Ref N 5 Fluorouracil 1 1 CL=TVCL*BW V=TVV*BW 221 44 Busulfan 0.83 0.89 CL=TVCL*(BW/60)^0.833 V=TVV*(BW/60)^0.889 222 30 Capecitabine 0 0.81 V3=TVV3*(BSA/1.8)^0.812 223 481 Cisplatin 1.85 1 .38 CL=TVCL*(BSA/1.74)^1.85 V=TVV*(BSA/1.74)^1.38 224 32 Cladribine 0 0 BSA was not found significant to be a covariate 225 161 Cyclophosphamide 0 0 BSA was not found significant to be a covariate 226 46 Dexamethasone 0 0.63 V1=TVV1*(WT/57.5)^0.626 227 897 Docetaxel 1 .11 0 CL=TVCL*(BSA/1.53)^1.11*(ALB/3.7)^0.251*(97/AAG)^ 0.776 228 200 Doxorubicin 1.4 0 CL=TVCL*(BSA/1.8)^1.4)*(AST/21)^ 0.24*(AGE/56)^ 0.54 226 46 Erlotinib 0 0.73 V=TVV*(BW/74)^0.73 229 42 Etoposide 0 0 BSA was not found signifi cant to be a covariate,but CRCL is on CL 230 52 Everolimus 0.4 0.32 CL=CL0+CL1* (WT 71)+CL2*(AGE 44))*1.2^RACE, V=V0+V1*(WT 71) 231 673 Gemcitabine 0 0 BSA was not found significant to be a covariate 232 94 Ifosfamide 0 0 BSA was not found significant to be a covariate 233 24 Imatinib 0.30 0.4 1 CL=TVCL*(WT/80)^0.301*(HB/13)^0.897*(WBC/16)^ 0.105, V=TVV*(WT/80)^0.405*(HB/13)^0.676*(WBC/16)^ 0.0 70 234 371 Irinotecan 0 0.2 V=TVV*(1+0.004*(WT 80)) 235 78 Melphalan 0 0 BSA was not found significant to be a covariate, but CRCL is on CL 236 64 Methotrexate 0 0 BSA was not found significant to be a covariate,but CRCL is on V 237 51 Mitomycin 1.63 0 CL=TVCL*(BSA/1.53)^1.63 238 47 Oxaliplatin 1.1 1.29 CL=TVCL*(WT/71)^1.1*(CRCL/87)^ 0.57*(0.6*GEN(F)),V1=T VV1*(WT/71)^1.29,Q=TVQ*(W T/71)^1.01 239 56 Paclitaxel Non linea r 1.17 V3=TVV3*(BSA/1.8)^1.17, VMT=VMT0*1.2^GEN*(BSA/1.8)^0.911 240 45 Pemetrexed 0 1.32 V=TVV*(BSA)^1.32 241 287 Temozolomide 1.05 0 CL=TVCL*(BSA)^1.32 208 445 Temsirolimus 1.28 0 CL=TVCL*(BSA)^1.28 242 50 Thalidomide 1 V=TVV*WT, ke was modeled instead of CL 243 65 Thiotepa 0 0 BSA was not found significant to be a covariate 244 65 Topotecan 0.75 1 CL=TVCL*(WT/70)^0.75,V=TVV*(WT/70) Q=TVQ*(WT/70)^0.75,V2=TVV2*(WT/70) 245 245 Vinorelbine 1.25 0 CL=TVCL*(BSA/1.61)^1.25 246 30 Ref: Reference number in the reference lis t; N is the number of patients enrolled in the study.
124 Table 4 4. Percentage contribution of body size measurements t o the overall intersubject variability of pharmacokinetics (PK) parameters in anticancer drugs Anticancer drugs % contribution of BW /BSA to the intersubject variability CL V1 5 Fluorouracil 3.3 2 0.55 Busulfan 39.1 8 75.59 Capecitabine 14.4 8 17.0 8 C isplatin 59.5 5 45.2 4 Dexamethasone 9.6 0 NA Docetaxel 11.2 0 NA Doxorubicin 5.61 NA Erlotinib NA 23.81 Everolimus 9.69 14.49 Imatinib 38.28 NA Irinotecan NA 10.85 Mitomycin 19.82 NA Oxaliplatin 4.58 6.7 5 Paclitaxel 35.6 7 NA Pemetrexed NA 17 .02 Temozolomide 2.24 NA Thalidomide 60.90 48.4 4 Topotecan 12.5 0 NA Vinorelbine 11.54 NA
125 Figure 4 1 Comparison of the inter subject variability of simulated AUC (A) and C max (B) of 1000 subjects after receiving a single fixed (solid bar) do se or a body size (BW/BSA) based dose(open bar) for selected anticancer drugs
126 Figure 4 2 Comparison of the deviation (% difference) of AUC (A) and C max (B) for subjects with low (open bar) and high (solid bar) extreme body size (BW/BSA) measurements from the typical values (AUC and C max for subjects with median body size measurements after a fixed dose (red) or a body size (BW/BSA) based dose (black)
127 Figure 4 3 Visual c omparison of the 5 th and 95 th percentile of cisplatin concentration time prof iles after a BSA based dos e and a BW based dos e for four BW or BSA value quartiles groups ( A. B. C. 50% > and 75%, and D.
128 Figure 4 4. Comparison of BSA based dosing and BW based dosing at population and individual levels. A. Comparison of the inter subject variability of simulated AUC and C max of 1000 subjects afte r receiving a single BSA based (solid bar) dose or BW based dose(open bar) for cisplatin B. Comparison of the deviation (% difference) of AUC and C max for subjects with low (open bar) and high (solid bar) extreme body size (BW/BSA) measur ements from the t ypical values AUC and C max for subjects with median body size measurements after a BW based dose (red) or BSA based dose (black)
129 CHAPTER 5 CONCLUSION Small molecule drugs, besides anticancer agents, are usually administered using a flat fixed dosing for the target patient population, w hile most small oncology drugs and biotherapeutic proteins and peptides are administered per body size which was and may be still believed to provide less interindividual variability and optimized risk/benefit ratio. A prev ious study has evaluated body size based dosing and fixed dosing were evaluated for 12 mAb s in terms of their population and individual performances in reducing intersubject PK and/or PD variability in adult patients 56 This di ssertation systemically evaluated therapeutic proteins and peptides besides mAbs as well as anticancer drugs via population PK/PD simulation studies in terms of inter individual PK and or/PD variability in adults. The object ives are : 1) systemically evaluat e the performance of fixed dosing and body size based dosing for therapeutic peptides and proteins in adults using population pharmacokinetics approach; 2) for the first time, utilize population pharmacokinetics approach to systemically evaluate intersubje ct variability for anticancer dru gs ; 3 ) provide a model based analysis method on dosing strategies for biotherapeutic large molecules and oncology drugs under development 4) recommend a dosing strategy for clinical trials conducted in adults at different s tages of the development of therapeutic biologics and anticancer drugs The simulation analysis results demonstrated that body size based dosing did not always result in less intersubject variability in drug exposure and fixed dosing does not always prese nts higher intersub ject variability than body size based dosing. In fact, fixed dosing showed better performance for 12 of 18 evaluated biologics based on both AUC and C max assessments. Similarly, fixed dosing showed better performance for half
130 of the eval uated anticancer drugs based on both AUC and C max assessments Therefore, the recommendations f or adult FIH studies, fixed dosing is recommended because it offers advantages in ease of dosing preparation, reduced cost, and reduced chance of dosing errors. When sufficient data become available, a full assessment of body size effect on PK and/or PD should be conducted. The final dosing approach for Phase III trials in adults should be selected based on the established body size effect on the PK and PD, the th erapeutic window of the therapeutic products, and other imprant factors that may affect the clinical outcomes Body size based dosing of biologics is also discussed in pediatrics. Most of the studies in the current study showed that body weight or BSA dose adjustment produced comparable exposure for proteins and peptides in pediatrics and adults However, not all pharmacokinetic studies result in promoting dosing adjustment. For basiliximab, a fixed dose of two 10 mg doses for patients l ess than 35 kg and a fixed dose of two 20 mg dose for patients more than 35 kg was recommended for pediatrics 112 Some studies suggested dividing children into different groups as a dosing strategy. Children in pediatric patients cannot always be explained by changes in body size. Simply adjusting dose linearly according to the body weight/BSA does not necessarily provide desirable exposure in pediatrics. Overall, individualized dosing is not equivalent to body size based dosing. Dose optimization using population PK/PD modeling and simulation can help to identify factors that contribute to the ove rall intersubject variability. A good dosing approach
131 should be made bas ed on those factors and provide consistent exposure to the patient population and overall beneficial clinical outcomes in the target patient groups.
132 LIST OF REFERENCES 1. Sheiner LB, Rosenberg B, Marathe VV. Estimation of population characteristics of pharmacokinetic parameters from routine clinical data. J Pharmacokinet Biopharm 1977; 5: 445 479. 2. United States Food and Drug Administration. Guidance for Industr y. Population pharmacokinetics. 1999. 3. Williams PJ, Ette EI. The role of population pharmacokinetics in drug development in light of the Food and Drug Administration's 'Guidance for Industry: population pharmacokinetics'. Clin Pharmacokinet 2000; 39: 38 5 395. 4. Sheiner LB, Rosenberg B, Melmon KL. Modelling of individual pharmacokinetics for computer aided drug dosage. Comput Biomed Res 1972; 5: 411 459. 5. Sheiner LB, Grasela TH. An Introduction to Mixed Effect Modeling Concepts, Definitions, and Ju stification. J Pharmacokinet Biopharm 1991; 19: S11 S24. 6. Sheiner LB. Learning versus confirming in clinical drug development. Clin Pharmacol Ther 1997; 61: 275 291. 7. Pinkel D. The use of body surface area as a criterion of drug dosage in cancer chem otherapy. Cancer Res 1958; 18: 853 856. 8. Quetelet LAJ. Physicale sociale C. Muquardt; etc.: Bruxelles,, 1869. 9. Keys A, Fidanza F, Karvonen MJ, Kimura N, Taylor HL. Indices of relative weight and obesity. J Chronic Dis 1972; 25: 329 343. 10. Ideal Weights for Women. Stat Bull Metropolitan Life Insurance Co 1942. 11. Ideal Weights for Men. Stat Bull Metropolitan Life Insurance Co 1943. 12. New Weight Standards for Men and Women. Stat Bull Metropolitan Life Insurance Co 1959; 41: 6. 13. Mortality a mong overweight men and women. Stat Bull Metropolitan Life Insurance Co 1960; 41: 6. 14. Mallon P. Build and Blood Pressure Study Society of Actuaries: New York, 1959. 15. Blackburn GL, Bistrian BR, Maini BS, Schlamm HT, Smith MF. Nutritional and metab olic assessment of the hospitalized patient. JPEN J Parenter Enteral Nutr 1977; 1: 11 22.
133 16. Devine BJ. Case study number 25 gentamicin therapy. Drug Intell Clin Pharm 1974; 8: 650 655. 17. Bauer LA, Edwards WA, Dellinger EP, Simonowitz DA. Influence of weight on aminoglycoside pharmacokinetics in normal weight and morbidly obese patients. Eur J Clin Pharmacol 1983; 24: 643 647. 18. Bauer LA. Applied clinical pharmacokinetics McGraw Hill: New York, 2001. 19. Winter ME. Basic pharmacokinetics Lippin cott Williams and Williams: London, 2004. 20. Bearden DT, Rodvold KA. Dosage adjustments for antibacterials in obese patients: applying clinical pharmacokinetics. Clin Pharmacokinet 2000; 38: 415 426. 21. Rathbun EN, Pace N. Studies on Body Composition 1. The Determination of Total Body Fat by Means of the Body Specific Gravity. J Biol Chem 1945; 158: 667 676. 22. Hume R. Relationship between Total Body Water and Surface Area in Normal and Obese Subjects. J Clin Pathol 1971; 24: 234 &. 23. Durnin JVG, Rahaman MM. Assessment of Amount of Fat in Human Body from Measurements of Skinfold Thickness. Br J Nutr 1967; 21: 681 &. 24. Womersley J, Boddy K, King PC, Durnin JV. Estimation of the fat free mass of twenty subjects from measurements of total body pota ssium, body density, skinfold thickness, and height and weight. Proc Nutr Soc 1972; 31: 35A. 25. Gray DS, Bray GA, Gemayel N, Kaplan K. Effect of obesity on bioelectrical impedance. Am J Clin Nutr 1989; 50: 255 260. 26. Lukaski HC, Bolonchuk WW, Hall CB, Siders WA. Validation of tetrapolar bioelectrical impedance method to assess human body composition. J Appl Physiol 1986; 60: 1327 1332. 27. Segal KR, Van Loan M, Fitzgerald PI, Hodgdon JA, Van Itallie TB. Lean body mass estimation by bioelectrical imped ance analysis: a four site cross validation study. Am J Clin Nutr 1988; 47: 7 14. 28. Jaber LA, Antal EJ, Slaughter RL, Welshman IR. The pharmacokinetics and pharmacodynamics of 12 weeks of glyburide therapy in obese diabetics. Eur J Clin Pharmacol 1993; 45: 459 463. 29.
134 30. Morgan DJ, Bray KM. Lean Body Mass as a Predictor of Drug Dosage Implications for Drug Therapy. Clin Pharmacokinet 1994; 26: 292 307. 31. Janmahasatian S Duffull SB, Ash S. Quantification of lean body weight. Clin Pharmacokinet 2005; 44: 1051 1065. 32. Green B, Duffull S. Caution when lean body weight is used as a size descriptor for obese subjects. Clin Pharmacol Ther 2002; 72: 743 744. 33. Duffull SB, Dooley MJ, Green B, Poole SG, Kirkpatrick CM. A standard weight descriptor for dose adjustment in the obese patient. Clin Pharmacokinet 2004; 43: 1167 1178. 34. Greenblatt DJ, Abernethy DR, Divoll M. Is Volume of Distribution at Steady State a Meaningful Kinetic Variable. J Clin Pharmacol 1983; 23: 391 400. 35. Green B, Duffull SB. What is the best size descriptor to use for pharmacokinetic studies in the obese? Br J Clin Pharmacol 2004; 58: 119 133. 36. Chaurasia CS, Muller M, Bashaw ED, Benfeldt E, Bo linder J, Bullock R et al. AAPS FDA Workshop White Paper: microdialysis principles, application, and regulatory perspectives. J Clin Pharmacol 2007; 47: 589 603. 37. Hollenstein UM, Brunner M, Schmid R, Muller M. Soft tissue concentrations of ciprofloxaci n in obese and lean subjects following weight adjusted dosing. Int J Obes Relat Metab Disord 2001; 25: 354 358. 38. Barbour A, Schmidt S, Rout WR, Ben David K, Burkhardt O, Derendorf H. Soft tissue penetration of cefuroxime determined by clinical microdia lysis in morbidly obese patients undergoing abdominal surgery. Int J Antimicrob Agents 2009; 34: 231 235. 39. Summers LKM, Samra JS, Humphreys SM, Morris RJ, Frayn KN. Subcutaneous abdominal adipose tissue blood flow: Variation within and between subjects and relationship to obesity. Clinical Science 1996; 91: 679 683. 40. Abel ED, Litwin SE, Sweeney G. Cardiac remodeling in obesity. Physiol Rev 2008; 88: 389 419. 41. Abernethy DR, Greenblatt DJ, Divoll M, Smith RB, Shader RI. The Influence of Obesity on the Pharmacokinetics of Oral Alprazolam and Triazolam. Clin Pharmacokinet 1984; 9: 177 183. 42. Abernethy DR, Greenblatt DJ. Phenytoin Disposition in Obesity Determination of Loading Dose. Arch Neurol 1985; 42: 468 471.
135 43. Benedek IH, Fiske WD, 3rd, Griffen WO, Bell RM, Blouin RA, McNamara PJ. Serum alpha 1 acid glycoprotein and the binding of drugs in obesity. Br J Clin Pharmacol 1983; 16: 751 754. 44. Benedek IH, Blouin RA, McNamara PJ. Serum protein binding and the role of increased alpha 1 acid g lycoprotein in moderately obese male subjects. Br J Clin Pharmacol 1984; 18: 941 946. 45. Cheymol G, Poirier JM, Barre J, Pradalier A, Dry J. Comparative pharmacokinetics of intravenous propranolol in obese and normal volunteers. J Clin Pharmacol 1987; 27 : 874 879. 46. Derry CL, Kroboth PD, Pittenger AL, Kroboth FJ, Corey SE, Smith RB. Pharmacokinetics and pharmacodynamics of triazolam after two intermittent doses in obese and normal weight men. J Clin Psychopharmacol 1995; 15: 197 205. 47. Cheymol G. Ef fects of obesity on pharmacokinetics implications for drug therapy. Clin Pharmacokinet 2000; 39: 215 231. 48. Cockcroft DW, Gault MH. Prediction of creatinine clearance from serum creatinine. Nephron 1976; 16: 31 41. 49. Pai MP, Norenberg JP, Anderson T, Goade DW, Rodvold KA, Telepak RA et al. Influence of morbid obesity on the single dose pharmacokinetics of daptomycin. Antimicrob Agents Chemother 2007; 51: 2741 2747. 50. Saadeh S. Nonalcoholic Fatty liver disease and obesity. Nutr Clin Pract 2007; 22: 1 10. 51. Ijaz S, Yang WX, Winslet MC, Seifalian AM. Impairment of hepatic microcirculation in fatty liver. Microcirculation 2003; 10: 447 456. 52. Emery MG, Fisher JM, Chien JY, Kharasch ED, Dellinger EP, Kowdley KV et al. CYP2E1 activity before and aft er weight loss in morbidly obese subjects with nonalcoholic fatty liver disease. Hepatology 2003; 38: 428 435. 53. Han PY, Duffull SB, Kirkpatrick CM, Green B. Dosing in obesity: a simple solution to a big problem. Clin Pharmacol Ther 2007; 82: 505 508. 54. Mathijssen RHJ, Sparreboom A. Influence of Lean Body Weight on Anticancer Drug Clearance. Clin Pharmacol Ther 2009; 85: 23 24.
136 55. Han PY, Duffull SB, Kirkpatrick CMJ, Green B. Response to "Influence of Lean Body Weight on Anticancer Drug Clearance". Clin Pharmacol Ther 2009; 85: 24 24. 56. Wang DD, Zhang S, Zhao H, Men AY, Parivar K. Fixed dosing versus body size based dosing of monoclonal antibodies in adult clinical trials. J Clin Pharmacol 2009; 49: 1012 1024. 57. Lin JH. Pharmacokinetics of Biot ech Drugs: Peptides, Proteins and Monoclonal Antibodies. Curr Drug Metab 2009. 58. Wang W, Wang EQ, Balthasar JP. Monoclonal antibody pharmacokinetics and pharmacodynamics. Clin Pharmacol Ther 2008; 84: 548 558. 59. Lin JH. Pharmacokinetics of biotech dr ugs: peptides, proteins and monoclonal antibodies. Curr Drug Metab 2009; 10: 661 691. 60. Hermanussen M, Danker Hopfe H, Weber GW. Body weight and the shape of the natural distribution of weight, in very large samples of German, Austrian and Norwegian con scripts. Int J Obes 2001; 25: 1550 1553. 61. Agoram B, Heatherington AC, Gastonguay MR. Development and evaluation of a population pharmacokinetic pharmacodynamic model of darbepoetin alfa in patients with nonmyeloid malignancies undergoing multicycle che motherapy. AAPS J 2006; 8: E552 563. 62. Lee H, Kimko HC, Rogge M, Wang D, Nestorov I, Peck CC. Population pharmacokinetic and pharmacodynamic modeling of etanercept using logistic regression analysis. Clin Pharmacol Ther 2003; 73: 348 365. 63. Roy A, Mo uld DR, Wang XF, Tay L, Raymond R, Pfister M. Modeling and simulation of abatacept exposure and interleukin 6 response in support of recommended doses for rheumatoid arthritis. J Clin Pharmacol 2007; 47: 1408 1420. 64. Dvorchik B, Arbeit RD, Chung J, Liu S, Knebel W, Kastrissios H. Population pharmacokinetics of daptomycin. Antimicrob Agents Chemother 2004; 48: 2799 2807. 65. Tornoe CW, Agerso H, Nielsen HA, Madsen H, Jonsson EN. Pharmacokinetic/pharmacodynamic modelling of GnRH antagonist degarelix: a co mparison of the non linear mixed effects programs NONMEM and NLME. J Pharmacokinet Pharmacodyn 2004; 31: 441 461. 66. Goggin T, Nguyen QT, Munafo A. Population pharmacokinetic modelling of Emfilermin (recombinant human leukaemia inhibitory factor, r hLIF) in healthy
137 postmenopausal women and in infertile patients undergoing in vitro fertilization and embryo transfer. Br J Clin Pharmacol 2004; 57: 576 585. 67. Mould DR, Zhang X, Nieforth K, Salgo M, Buss N, Patel IH. Population pharmacokinetics and exposure response relationship of enfuvirtide in treatment experienced human immunodeficiency virus type 1 infected patients. Clin Pharmacol Ther 2005; 77: 515 528. 68. Chakraborty A, Natarajan J, Guilfoyle M, Morgan N, Vercammen E, Cheung W. Population pharmacok inetics of erythropoietin in critically ill subjects. J Clin Pharmacol 2005; 45: 193 202. 69. Hayashi N, Kinoshita H, Yukawa E, Higuchi S. Pharmacokinetic analysis of subcutaneous erythropoietin administration with nonlinear mixed effect model including e ndogenous production. Br J Clin Pharmacol 1998; 46: 11 19. 70. Lee H, Kimko HC, Rogge M, Wang D, Nestorov I, Peck CC. Population pharmacokinetic and pharmacodynamic modeling of etanercept using logistic regression analysis. Clin Pharmacol Ther 2003; 73: 3 48 365. 71. Woo S, Krzyzanski W, Duliege AM, Stead RB, Jusko WJ. Population pharmacokinetics and pharmacodynamics of peptidic erythropoiesis receptor agonist (ERA) in healthy volunteers. J Clin Pharmacol 2008; 48: 43 52. 72. Troconiz IF, Cendros JM, Pera ire C, Ramis J, Garrido MJ, Boscani PF et al. Population pharmacokinetic analysis of lanreotide Autogel in healthy subjects : evidence for injection interval of up to 2 months. Clin Pharmacokinet 2009; 48: 51 62. 73. Zhou H, Chen TL, Marino M, Lau H, Mill er T, Kalafsky G et al. Population PK and PK/PD modelling of microencapsulated octreotide acetate in healthy subjects. Br J Clin Pharmacol 2000; 50: 543 552. 74. Glatt S, Fuseau E, Buraglio M, Nguyen QT. Population pharmacokinetics of onercept in healthy subjects. Clin Pharmacokinet 2005; 44: 1295 1304. 75. Jen JF, Glue P, Ezzet F, Chung C, Gupta SK, Jacobs S et al. Population pharmacokinetic analysis of pegylated interferon alfa 2b and interferon alfa 2b in patients with chronic hepatitis C. Clin Pharmac ol Ther 2001; 69: 407 421. 76. Nalda Molina R, Valenzuela B, Ramon Lopez A, Miguel Lillo B, Soto Matos A, Perez Ruixo JJ. Population pharmacokinetics meta analysis of plitidepsin (Aplidin) in cancer subjects. Cancer Chemother Pharmacol 2009; 64: 97 108.
138 77. Girard P, Nony P, Erhardtsen E, Delair S, Ffrench P, Dechavanne M et al. Population pharmacokinetics of recombinant factor VIIa in volunteers anticoagulated with acenocoumarol. Thromb Haemost 1998; 80: 109 113. 78. Klitgaard T, Nielsen JN, Skettrup MP Harper A, Lange M. Population pharmacokinetic model for human growth hormone in adult patients in chronic dialysis compared with healthy subjects. Growth Horm IGF Res 2009; 19: 463 470. 79. Karlsson MO, Wade JR, Loumaye E, Munafo A. The population pharm acokinetics of recombinant and urinary human follicle stimulating hormone in women. Br J Clin Pharmacol 1998; 45: 13 20. 80. Roberts R, Rodriguez W, Murphy D, Crescenzi T. Pediatric drug labeling: improving the safety and efficacy of pediatric therapies. Jama 2003; 290: 905 911. 81. Alcorn J, McNamara PJ. Pharmacokinetics in the newborn. Adv Drug Deliv Rev 2003; 55: 667 686. 82. Rodriguez W, Selen A, Avant D, Chaurasia C, Crescenzi T, Gieser G et al. Improving pediatric dosing through pediatric initiati ves: what we have learned. Pediatrics 2008; 121: 530 539. 83. Derendorf H. Pediatrics. http://www.cop.ufl.edu/education/graduat e programs/pharmaceutics/pha 5128 basic principles of dose optimizaton ii/ (accessed on 30 Sep 2010). 84. Schoeller DA. Changes in total body water with age. Am J Clin Nutr 1989; 50: 1176 1181; discussion 1231 1175. 85. Peters AM, Henderson BL, Lui D. I ndexed glomerular filtration rate as a function of age and body size. Clin Sci (Lond) 2000; 98: 439 444. 86. Hines RN. Ontogeny of human hepatic cytochromes P450. J Biochem Mol Toxicol 2007; 21: 169 175. 87. Koukouritaki SB, Manro JR, Marsh SA, Stevens J C, Rettie AE, McCarver DG et al. Developmental expression of human hepatic CYP2C9 and CYP2C19. J Pharmacol Exp Ther 2004; 308: 965 974. 88. Authier F, Danielsen GM, Kouach M, Briand G, Chauvet G. Identification of insulin domains important for binding to and degradation by endosomal acidic insulinase. Endocrinology 2001; 142: 276 289.
139 89. Smedsrod B, Einarsson M. Clearance of tissue plasminogen activator by mannose and galactose receptors in the liver. Thromb Haemost 1990; 63: 60 66. 90. FDA. Guidance fo r Industry General Considerations for Pediatric Pharmacokinetic Studies for Drugs and Biological Products. http://www.fda.gov/downloads/Drugs/Gu idanceComplianceRegulatoryInformation /Guidances/ucm072114.pdf (accessed on 1 Dec 2010). 1998. 91. Lobo ED, Hansen RJ, Balthasar JP. Antibody pharmacokinetics and pharmacodynamics. J Pharm Sci 2004; 93: 2645 2668. 92. Tang L, Persky AM, Hochhaus G, Meibo hm B. Pharmacokinetic aspects of biotechnology products. J Pharm Sci 2004; 93: 2184 2204. 93. Reilly RM, Sandhu J, Alvarez Diez TM, Gallinger S, Kirsh J, Stern H. Problems of delivery of monoclonal antibodies. Pharmaceutical and pharmacokinetic solutions. Clin Pharmacokinet 1995; 28: 126 142. 94. Zito SW. Pharmaceutical biotechnology: A programmed text. Technomic Pub.Co.: Lancaster, PA, 1997. 95. Tan AC, Russel FG, Thien T, Benraad TJ. Atrial natriuretic peptide. An overview of clinical pharmacology an d pharmacokinetics. Clin Pharmacokinet 1993; 24: 28 45. 96. Colburn W. Peptide, peptoid, and protein pharmacokinetics/pharmacodynamics. vol. 3. Harvey Whitney Books: Cincinnati,OH, 1991. 97. Eppler SM, Combs DL, Henry TD, Lopez JJ, Ellis SG, Yi JH et al A target mediated model to describe the pharmacokinetics and hemodynamic effects of recombinant human vascular endothelial growth factor in humans. Clin Pharmacol Ther 2002; 72: 20 32. 98. Piscitelli SC, Reiss WG, Figg WD, Petros WP. Pharmacokinetic stu dies with recombinant cytokines. Scientific issues and practical considerations. Clin Pharmacokinet 1997; 32: 368 381. 99. Soy D, Aweeka FT, Church JA, Cunningham CK, Palumbo P, Kosel BW et al. Population pharmacokinetics of enfuvirtide in pediatric patie nts with human immunodeficiency virus: searching for exposure response relationships. Clin Pharmacol Ther 2003; 74: 569 580. 100. Toon S. The relevance of pharmacokinetics in the development of biotechnology products. Eur J Drug Metab Pharmacokinet 1996; 21: 93 103.
140 101. Johnson V, Maack T. Renal extraction, filtration, absorption, and catabolism of growth hormone. Am J Physiol 1977; 233: F185 196. 102. Rabkin R, Ryan MP, Duckworth WC. The renal metabolism of insulin. Diabetologia 1984; 27: 351 357. 103 Takagi A, Masuda H, Takakura Y, Hashida M. Disposition characteristics of recombinant human interleukin 11 after a bolus intravenous administration in mice. J Pharmacol Exp Ther 1995; 275: 537 543. 104. Carone FA, Peterson DR. Hydrolysis and transport o f small peptides by the proximal tubule. Am J Physiol 1980; 238: F151 158. 105. Carone FA, Peterson DR, Flouret G. Renal tubular processing of small peptide hormones. J Lab Clin Med 1982; 100: 1 14. 106. Maack T, Johnson V, Kau ST, Figueiredo J, Sigulem D. Renal filtration, transport, and metabolism of low molecular weight proteins: a review. Kidney Int 1979; 16: 251 270. 107. Nielsen S, Nielsen JT, Christensen EI. Luminal and basolateral uptake of insulin in isolated, perfused, proximal tubules. Am J Ph ysiol 1987; 253: F857 867. 108. Authier F, Posner BI, Bergeron JJ. Endosomal proteolysis of internalized proteins. FEBS Lett 1996; 389: 55 60. 109. Meibohm B, Derendorf H. Pharmacokinetics and pharmacodynamics of biotech drugs. Wiley VCH: Weinheim, 200 3. 110. Racine Poon A, Botta L, Chang TW, Davis FM, Gygax D, Liou RS et al. Efficacy, pharmacodynamics, and pharmacokinetics of CGP 51901, an anti immunoglobulin E chimeric monoclonal antibody, in patients with seasonal allergic rhinitis. Clin Pharmacol T her 1997; 62: 675 690. 111. Schulman ES. Development of a monoclonal anti immunoglobulin E antibody (omalizumab) for the treatment of allergic respiratory disorders. Am J Respir Crit Care Med 2001; 164: S6 11. 112. Kovarik JM, Offner G, Broyer M, Niaudet P, Loirat C, Mentser M et al. A rational dosing algorithm for basiliximab (Simulect) in pediatric renal transplantation based on pharmacokinetic dynamic evaluations. Transplantation 2002; 74: 966 971. 113. Kovarik JM, Kahan BD, Rajagopalan PR, Bennett W, Mulloy LL, Gerbeau C et al. Population pharmacokinetics and exposure response relationships for
141 basiliximab in kidney transplantation. The U.S. Simulect Renal Transplant Study Group. Transplantation 1999; 68: 1288 1294. 114. Kovarik JM, Gridelli BG, Mart in S, Rodeck B, Melter M, Dunn SP et al. Basiliximab in pediatric liver transplantation: a pharmacokinetic derived dosing algorithm. Pediatr Transplant 2002; 6: 224 230. 115. Hocker B, Kovarik JM, Daniel V, Opelz G, Fehrenbach H, Holder M et al. Pharmacok inetics and immunodynamics of basiliximab in pediatric renal transplant recipients on mycophenolate mofetil comedication. Transplantation 2008; 86: 1234 1240. 116. Kovarik JM, Pescovitz MD, Sollinger HW, Kaplan B, Legendre C, Salmela K et al. Differential influence of azathioprine and mycophenolate mofetil on the disposition of basiliximab in renal transplant patients. Clin Transplant 2001; 15: 123 130. 117. Pescovitz MD, Knechtle S, Alexander SR, Colombani P, Nevins T, Nieforth K et al. Safety and pharma cokinetics of daclizumab in pediatric renal transplant recipients. Pediatr Transplant 2008; 12: 447 455. 118. Saez Llorens X, Castano E, Null D, Steichen J, Sanchez PJ, Ramilo O et al. Safety and pharmacokinetics of an intramuscular humanized monoclonal a ntibody to respiratory syncytial virus in premature infants and infants with bronchopulmonary dysplasia. The MEDI 493 Study Group. Pediatr Infect Dis J 1998; 17: 787 791. 119. Saez Llorens X, Moreno MT, Ramilo O, Sanchez PJ, Top FH, Jr., Connor EM. Safety and pharmacokinetics of palivizumab therapy in children hospitalized with respiratory syncytial virus infection. Pediatr Infect Dis J 2004; 23: 707 712. 120. Burns JC, Best BM, Mejias A, Mahony L, Fixler DE, Jafri HS et al. Infliximab treatment of intravenous immunoglobulin resistant Kawasaki disease. J Pediatr 2008; 153: 833 838. 121. Ruperto N, Lovell DJ, Cuttica R, Wilkinson N, Woo P, Espada G et al. A randomized, placebo controlled tr ial of infliximab plus methotrexate for the treatment of polyarticular course juvenile rheumatoid arthritis. Arthritis Rheum 2007; 56: 3096 3106. 122. Ternant D, Mulleman D, Degenne D, Willot S, Guillaumin JM, Watier H et al. An enzyme linked immunosorben t assay for therapeutic drug monitoring of infliximab. Ther Drug Monit 2006; 28: 169 174. 123. Cornillie F, Shealy D, D'Haens G, Geboes K, Van Assche G, Ceuppens J et al. Infliximab induces potent anti inflammatory and local immunomodulatory activity
142 but no systemic immune suppression in patients with Crohn's disease. Aliment Pharmacol Ther 2001; 15: 463 473. 124. St Clair EW, Wagner CL, Fasanmade AA, Wang B, Schaible T, Kavanaugh A et al. The relationship of serum infliximab concentrations to clinical im provement in rheumatoid arthritis: results from ATTRACT, a multicenter, randomized, double blind, placebo controlled trial. Arthritis Rheum 2002; 46: 1451 1459. 125. Buckwalter M, Dowell JA, Korth Bradley J, Gorovits B, Mayer PR. Pharmacokinetics of gemtu zumab ozogamicin as a single agent treatment of pediatric patients with refractory or relapsed acute myeloid leukemia. J Clin Pharmacol 2004; 44: 873 880. 126. Dowell JA, Korth Bradley J, Liu H, King SP, Berger MS. Pharmacokinetics of gemtuzumab ozogamici n, an antibody targeted chemotherapy agent for the treatment of patients with acute myeloid leukemia in first relapse. J Clin Pharmacol 2001; 41: 1206 1214. 127. Angiolillo AL, Yu AL, Reaman G, Ingle AM, Secola R, Adamson PC. A phase II study of Campath 1 H in children with relapsed or refractory acute lymphoblastic leukemia: a Children's Oncology Group report. Pediatr Blood Cancer 2009; 53: 978 983. 128. Mould DR, Baumann A, Kuhlmann J, Keating MJ, Weitman S, Hillmen P et al. Population pharmacokinetics p harmacodynamics of alemtuzumab (Campath) in patients with chronic lymphocytic leukaemia and its link to treatment response. Br J Clin Pharmacol 2007; 64: 278 291. 129. Montillo M, Tedeschi A, Miqueleiz S, Veronese S, Cairoli R, Intropido L et al. Alemtuzu mab as consolidation after a response to fludarabine is effective in purging residual disease in patients with chronic lymphocytic leukemia. J Clin Oncol 2006; 24: 2337 2342. 130. Hale G, Rebello P, Brettman LR, Fegan C, Kennedy B, Kimby E et al. Blood co ncentrations of alemtuzumab and antiglobulin responses in patients with chronic lymphocytic leukemia following intravenous or subcutaneous routes of administration. Blood 2004; 104: 948 955. 131. Trippett TM, Herzog C, Whitlock JA, Wolff J, Kuttesch J, Ba gatell R et al. Phase I and pharmacokinetic study of cetuximab and irinotecan in children with refractory solid tumors: a study of the pediatric oncology experimental therapeutic investigators' consortium. J Clin Oncol 2009; 27: 5102 5108. 132. Tan AR, Mo ore DF, Hidalgo M, Doroshow JH, Poplin EA, Goodin S et al. Pharmacokinetics of cetuximab after administration of escalating single dosing
143 and weekly fixed dosing in patients with solid tumors. Clin Cancer Res 2006; 12: 6517 6522. 133. Glade Bender JL, Ada mson PC, Reid JM, Xu L, Baruchel S, Shaked Y et al. Phase I trial and pharmacokinetic study of bevacizumab in pediatric patients with refractory solid tumors: a Children's Oncology Group Study. J Clin Oncol 2008; 26: 399 405. 134. Gordon MS, Margolin K, T alpaz M, Sledge GW, Jr., Holmgren E, Benjamin R et al. Phase I safety and pharmacokinetic study of recombinant human anti vascular endothelial growth factor in patients with advanced cancer. J Clin Oncol 2001; 19: 843 850. 135. Lu JF, Bruno R, Eppler S, N ovotny W, Lum B, Gaudreault J. Clinical pharmacokinetics of bevacizumab in patients with solid tumors. Cancer Chemother Pharmacol 2008; 62: 779 786. 136. Hyams JS, Wilson DC, Thomas A, Heuschkel R, Mitton S, Mitchell B et al. Natalizumab therapy for moder ate to severe Crohn disease in adolescents. J Pediatr Gastroenterol Nutr 2007; 44: 185 191. 137. Freeman BB, 3rd, Hinds P, Iacono LC, Razzouk BI, Burghen E, Stewart CF. Pharmacokinetics and pharmacodynamics of intravenous epoetin alfa in children with can cer. Pediatr Blood Cancer 2006; 47: 572 579. 138. Lim VS, DeGowin RL, Zavala D, Kirchner PT, Abels R, Perry P et al. Recombinant human erythropoietin treatment in pre dialysis patients. A double blind placebo controlled trial. Ann Intern Med 1989; 110: 10 8 114. 139. Evans JH, Brocklebank JT, Bowmer CJ, Ng PC. Pharmacokinetics of recombinant human erythropoietin in children with renal failure. Nephrol Dial Transplant 1991; 6: 709 714. 140. Widness JA, Veng Pedersen P, Peters C, Pereira LM, Schmidt RL, Low e LS. Erythropoietin pharmacokinetics in premature infants: developmental, nonlinearity, and treatment effects. J Appl Physiol 1996; 80: 140 148. 141. Ohls RK, Veerman MW, Christensen RD. Pharmacokinetics and effectiveness of recombinant erythropoietin ad ministered to preterm infants by continuous infusion in total parenteral nutrition solution. J Pediatr 1996; 128: 518 523. 142. Kendall DM, Riddle MC, Rosenstock J, Zhuang D, Kim DD, Fineman MS et al. Effects of exenatide (exendin 4) on glycemic control o ver 30 weeks in patients with type 2 diabetes treated with metformin and a sulfonylurea. Diabetes Care 2005; 28: 1083 1091.
144 143. Smith WB, Dowell JA, Pratt RD. Pharmacokinetics and pharmacodynamics of epoetin delta in two studies in healthy volunteers and two studies in patients with chronic kidney disease. Clin Ther 2007; 29: 1368 1380. 144. Lerner G, Kale AS, Warady BA, Jabs K, Bunchman TE, Heatherington A et al. Pharmacokinetics of darbepoetin alfa in pediatric patients with chronic kidney disease. Ped iatr Nephrol 2002; 17: 933 937. 145. Macdougall IC, Gray SJ, Elston O, Breen C, Jenkins B, Browne J et al. Pharmacokinetics of novel erythropoiesis stimulating protein compared with epoetin alfa in dialysis patients. J Am Soc Nephrol 1999; 10: 2392 2395. 146. Ateshkadi A, Johnson CA, Oxton LL, Hammond TG, Bohenek WS, Zimmerman SW. Pharmacokinetics of intraperitoneal, intravenous, and subcutaneous recombinant human erythropoietin in patients on continuous ambulatory peritoneal dialysis. Am J Kidney Dis 199 3; 21: 635 642. 147. Macdougall IC, Roberts DE, Coles GA, Williams JD. Clinical pharmacokinetics of epoetin (recombinant human erythropoietin). Clin Pharmacokinet 1991; 20: 99 113. 148. Blumer J, Berg S, Adamson PC, Loew T, Rossi G, Hastings C. Pharmacok inetic evaluation of darbepoetin alfa for the treatment of pediatric patients with chemotherapy induced anemia. Pediatr Blood Cancer 2007; 49: 687 693. 149. De Palo T, Giordano M, Palumbo F, Bellantuono R, Messina G, Colella V et al. Clinical experience w ith darbepoietin alfa (NESP) in children undergoing hemodialysis. Pediatr Nephrol 2004; 19: 337 340. 150. Faulkner LB, Tucci F, Tamburini A, Tintori V, Lippi AA, Bambi F et al. G CSF serum pharmacokinetics during peripheral blood progenitor cell mobilizat ion: neutrophil count adjusted dosage might potentially improve mobilization and be more cost effective. Bone Marrow Transplant 1998; 21: 1091 1095. 151. Wells RJ, Weck PK, Baehner RL, Krivit W, Raney RB, Ortega JA et al. Interferon alpha n1 in children w ith recurrent acute lymphocytic leukemia: a phase I study of pharmacokinetics and tolerance. J Interferon Res 1988; 8: 309 318. 152. Knost JA, Sherwin SA, Abrams PG, Ochs JJ, Foon KA, Williams R et al. The treatment of cancer patients with human lymphobla stoid interferon. A comparison of two routes of administration. Cancer Immunol Immunother 1983; 15: 144 148. 153. Schwarz KB, Mohan P, Narkewicz MR, Molleston JP, Nash SR, Hu S et al. Safety, efficacy and pharmacokinetics of peginterferon alpha2a (40 kd) in children with chronic hepatitis C. J Pediatr Gastroenterol Nutr 2006; 43: 499 505.
145 154. Zeuzem S, Feinman SV, Rasenack J, Heathcote EJ, Lai MY, Gane E et al. Peginterferon alfa 2a in patients with chronic hepatitis C. N Engl J Med 2000; 343: 1666 1672. 155. Gonzalez Peralta RP, Kelly DA, Haber B, Molleston J, Murray KF, Jonas MM et al. Interferon alfa 2b in combination with ribavirin for the treatment of chronic hepatitis C in children: efficacy, safety, and pharmacokinetics. Hepatology 2005; 42: 1010 1018. 156. Villar A, Aronis S, Morfini M, Santagostino E, Auerswald G, Thomsen HF et al. Pharmacokinetics of activated recombinant coagulation factor VII (NovoSeven) in children vs. adults with haemophilia A. Haemophilia 2004; 10: 352 359. 157. Murry DJ, Crom WR, Reddick WE, Bhargava R, Evans WE. Liver volume as a determinant of drug clearance in children and adolescents. Drug Metab Dispos 1995; 23: 1110 1116. 158. Klitgaard T, Nielsen TG. Overview of the human pharmacokinetics of recombinant activated f actor VII. Br J Clin Pharmacol 2008; 65: 3 11. 159. Blanchette VS, Shapiro AD, Liesner RJ, Hernandez Navarro F, Warrier I, Schroth PC et al. Plasma and albumin free recombinant factor VIII: pharmacokinetics, efficacy and safety in previously treated pedia tric patients. J Thromb Haemost 2008; 6: 1319 1326. 160. Tarantino MD, Collins PW, Hay CR, Shapiro AD, Gruppo RA, Berntorp E et al. Clinical evaluation of an advanced category antihaemophilic factor prepared using a plasma/albumin free method: pharmacokin etics, efficacy, and safety in previously treated patients with haemophilia A. Haemophilia 2004; 10: 428 437. 161. Gale RF, Hird MF, Colvin BT. Management of a premature infant with moderate haemophilia A using recombinant factor VIII. Haemophilia 1998; 4 : 850 853. 162. Carlsson M, Berntorp E, Bjrkman S, Lethagen S, Ljung R. Improved cost effectiveness by pharmacokinetic dosing of factor VIII in prophylactic treatment of haemophilia A. Haemophilia 2007; 3: 96 101. 163. Bolon Larger M, Chamouard V, Bress olle F, Boulieu R. A limited sampling strategy for estimating individual pharmacokinetic parameters of coagulation factor VIII in patients with hemophilia A. Ther Drug Monit 2007; 29: 20 26. 164. Bjorkman S, Folkesson A, Jonsson S. Pharmacokinetics and do se requirements of factor VIII over the age range 3 74 years: a population analysis based on 50 patients with long term prophylactic treatment for haemophilia A. Eur J Clin Pharmacol 2009; 65: 989 998.
146 165. Matucci M, Messori A, Donati Cori G, Longo G, Va nnini S, Morfini M et al. Kinetic evaluation of four Factor VIII concentrates by model independent methods. Scand J Haematol 1985; 34: 22 28. 166. van Dijk K, van der Bom JG, Lenting PJ, de Groot PG, Mauser Bunschoten EP, Roosendaal G et al. Factor VIII h alf life and clinical phenotype of severe hemophilia A. Haematologica 2005; 90: 494 498. 167. Bjorkman S, Folkesson A, Berntorp E. In vivo recovery of factor VIII and factor IX: intra and interindividual variance in a clinical setting. Haemophilia 2007; 13: 2 8. 168. Ahnstrom J, Berntorp E, Lindvall K, Bjorkman S. A 6 year follow up of dosing, coagulation factor levels and bleedings in relation to joint status in the prophylactic treatment of haemophilia. Haemophilia 2004; 10: 689 697. 169. Bjorkman S, Shapiro AD, Berntorp E. Pharmacokinetics of recombinant factor IX in relation to age of the patient: implications for dosing in prophylaxis. Haemophilia 2001; 7: 133 139. 170. Barton P, Kalil AC, Nadel S, Goldstein B, Okhuysen Cawley R, Brilli RJ et al. S afety, pharmacokinetics, and pharmacodynamics of drotrecogin alfa (activated) in children with severe sepsis. Pediatrics 2004; 113: 7 17. 171. Danne T, Datz N, Endahl L, Haahr H, Nestoris C, Westergaard L et al. Insulin detemir is characterized by a more reproducible pharmacokinetic profile than insulin glargine in children and adolescents with type 1 diabetes: results from a randomized, double blind, controlled trial. Pediatr Diabetes 2008; 9: 554 560. 172. Danne T. Flexibility of rapid acting insulin an alogues in children and adolescents with diabetes mellitus. Clin Ther 2007; 29 Suppl D: S145 152. 173. Mortensen HB, Lindholm A, Olsen BS, Hylleberg B. Rapid appearance and onset of action of insulin aspart in paediatric subjects with type 1 diabetes. Eur J Pediatr 2000; 159: 483 488. 174. Danne T, Becker RH, Heise T, Bittner C, Frick AD, Rave K. Pharmacokinetics, prandial glucose control, and safety of insulin glulisine in children and adolescents with type 1 diabetes. Diabetes Care 2005; 28: 2100 2105. 175. Malloy J, Capparelli E, Gottschalk M, Guan X, Kothare P, Fineman M. Pharmacology and tolerability of a single dose of exenatide in adolescent patients with type 2 diabetes mellitus being treated with metformin: a randomized, placebo controlled, singl e blind, dose escalation, crossover study. Clin Ther 2009; 31: 806 815.
147 176. Buse JB, Henry RR, Han J, Kim DD, Fineman MS, Baron AD. Effects of exenatide (exendin 4) on glycemic control over 30 weeks in sulfonylurea treated patients with type 2 diabetes. Diabetes Care 2004; 27: 2628 2635. 177. DeFronzo RA, Ratner RE, Han J, Kim DD, Fineman MS, Baron AD. Effects of exenatide (exendin 4) on glycemic control and weight over 30 weeks in metformin treated patients with type 2 diabetes. Diabetes Care 2005; 28: 1092 1100. 178. Walvoord EC, de la Pena A, Park S, Silverman B, Cuttler L, Rose SR et al. Inhaled growth hormone (GH) compared with subcutaneous GH in children with GH deficiency: pharmacokinetics, pharmacodynamics, and safety. J Clin Endocrinol Metab 200 9; 94: 2052 2059. 179. Peter F, Savoy C, Ji HJ, Juhasz M, Bidlingmaier M, Saenger P. Pharmacokinetic and pharmacodynamic profile of a new sustained release GH formulation, LB03002, in children with GH deficiency. Eur J Endocrinol 2009; 160: 349 355. 180. Bidlingmaier M, Kim J, Savoy C, Kim MJ, Ebrecht N, de la Motte S et al. Comparative pharmacokinetics and pharmacodynamics of a new sustained release growth hormone (GH), LB03002, versus daily GH in adults with GH deficiency. J Clin Endocrinol Metab 2006; 91: 2926 2930. 181. Kemp SF, Fielder PJ, Attie KM, Blethen SL, Reiter EO, Ford KM et al. Pharmacokinetic and pharmacodynamic characteristics of a long acting growth hormone (GH) preparation (nutropin depot) in GH deficient children. J Clin Endocrinol Meta b 2004; 89: 3234 3240. 182. Houdijk EC, Herdes E, Delemarre Van de Waal HA. Pharmacokinetics and pharmacodynamics of recombinant human growth hormone by subcutaneous jet or needle injection in patients with growth hormone deficiency. Acta Paediatr 1997; 86: 1301 1307. 183. Cairo MS, Davenport V, Bessmertny O, Goldman SC, Berg SL, Kreissman SG et al. Phase I/II dose escalation study of recombinant human interleukin 11 following ifosfamide, carboplatin and etoposide in children, adolescents and young adult s with solid tumours or lymphoma: a clinical, haematological and biological study. Br J Haematol 2005; 128: 49 58. 184. Aoyama K, Uchida T, Takanuki F, Usui T, Watanabe T, Higuchi S et al. Pharmacokinetics of recombinant human interleukin 11 (rhIL 11) in healthy male subjects. Br J Clin Pharmacol 1997; 43: 571 578. 185. Yim DS, Zhou H, Buckwalter M, Nestorov I, Peck CC, Lee H. Population pharmacokinetic analysis and simulation of the time concentration profile of
148 etanercept in pediatric patients with juve nile rheumatoid arthritis. J Clin Pharmacol 2005; 45: 246 256. 186. Zhou H, Buckwalter M, Boni J, Mayer P, Raible D, Wajdula J et al. Population based pharmacokinetics of the soluble TNFr etanercept: a clinical study in 43 patients with ankylosing spondyl itis compared with post hoc data from patients with rheumatoid arthritis. Int J Clin Pharmacol Ther 2004; 42: 267 276. 187. Wiznia A, Church J, Emmanuel P, Eppes S, Rowell L, Evans C et al. Safety and efficacy of enfuvirtide for 48 weeks as part of an opt imized antiretroviral regimen in pediatric human immunodeficiency virus 1 infected patients. Pediatr Infect Dis J 2007; 26: 799 805. 188. Church JA, Cunningham C, Hughes M, Palumbo P, Mofenson LM, Delora P et al. Safety and antiretroviral activity of chro nic subcutaneous administration of T 20 in human immunodeficiency virus 1 infected children. Pediatr Infect Dis J 2002; 21: 653 659. 189. Church JA, Hughes M, Chen J, Palumbo P, Mofenson LM, Delora P et al. Long term tolerability and safety of enfuvirtide for human immunodeficiency virus 1 infected children. Pediatr Infect Dis J 2004; 23: 713 718. 190. Zhang X, Lin T, Bertasso A, Evans C, Dorr A, Kolis SJ et al. Population pharmacokinetics of enfuvirtide in HIV 1 infected pediatric patients over 48 weeks of treatment. J Clin Pharmacol 2007; 47: 510 517. 191. Patel IH, Zhang X, Nieforth K, Salgo M, Buss N. Pharmacokinetics, pharmacodynamics and drug interaction potential of enfuvirtide. Clin Pharmacokinet 2005; 44: 175 186. 192. Vieira Pinheiro JP, Lanver sa C, Wurthwein G, Beier R, Casimiro da Palma J, von Stackelberg A et al. Drug monitoring of PEG asparaginase treatment in childhood acute lymphoblastic leukemia and non Hodgkin's lymphoma. Leuk Lymphoma 2002; 43: 1911 1920. 193. Baillargeon J, Langevin A M, Lewis M, Thomas PJ, Mullins J, Dugan J et al. L asparaginase as a marker of chemotherapy dose modification in children with acute lymphoblastic leukemia. Cancer 2005; 104: 2858 2861. 194. Anderson BJ, Allegaert K, Holford NH. Population clinical pharma cology of children: modelling covariate effects. Eur J Pediatr 2006; 165: 819 829. 195. Anderson BJ, Holford NH. Mechanism based concepts of size and maturity in pharmacokinetics. Annu Rev Pharmacol Toxicol 2008; 48: 303 332.
149 196. Anderson BJ, Holford NH Mechanistic basis of using body size and maturation to predict clearance in humans. Drug Metab Pharmacokinet 2009; 24: 25 36. 197. Sehgal N. Hospitalization rate surges 119% in pre teens due to eating disorders -study. (accessed on 3 Dec 2010). 2010 http://www.themoneyt imes.com/featured/20101129/hospitalization rate surges 119 preteens due eating disordersstudy id 10143305.htm?utm_source=feedburner&utm_medium=feed&utm_campaign=Fe ed%3A+TheMoneyTimes+%28The+Money+Times%29 198. Crawford JD, Terry ME, Rourke GM. Simplifica tion of drug dosage calculation by application of the surface area principle. Pediatrics 1950; 5: 783 790. 199. Du Bois D, Du Bois EF. A formula to estimate the approximate surface area if height and weight be known. 1916. Nutrition 1989; 5: 303 311; disc ussion 312 303. 200. Mosteller RD. Simplified calculation of body surface area. N Engl J Med 1987; 317: 1098. 201. Gehan EA, George SL. Estimation of human body surface area from height and weight. Cancer Chemother Rep 1970; 54: 225 235. 202. Field KM, Kosmider S, Jefford M, Michael M, Jennens R, Green M et al. Chemotherapy dosing strategies in the obese, elderly, and thin patient: results of a nationwide survey. J Oncol Pract 2008; 4: 108 113. 203. Verbraecken J, Van de Heyning P, De Backer W, Van Gaal L. Body surface area in normal weight, overweight, and obese adults. A comparison study. Metabolism 2006; 55: 515 524. 204. Yu CY, Lo YH, Chiou WK. The 3D scanner for measuring body surface area: a simplified calculation in the Chinese adult. Appl Ergon 2003; 34: 273 278. 205. Mathijssen RH, de Jong FA, Loos WJ, van der Bol JM, Verweij J, Sparreboom A. Flat fixed dosing versus body surface area based dosing of anticancer drugs in adults: does it make a difference? Oncologist 2007; 12: 913 923. 206. Bake r SD, Verweij J, Rowinsky EK, Donehower RC, Schellens JH, Grochow LB et al. Role of body surface area in dosing of investigational anticancer agents in adults, 1991 2001. J Natl Cancer Inst 2002; 94: 1883 1888. 207. Hammond LA, Eckardt JR, Baker SD, Eckha rdt SG, Dugan M, Forral K et al. Phase I and pharmacokinetic study of temozolomide on a daily for 5 days schedule in patients with advanced solid malignancies. J Clin Oncol 1999; 17: 2604 2613.
150 208. Jen JF, Cutler DL, Pai SM, Batra VK, Affrime MB, Zambas DN et al. Population pharmacokinetics of temozolomide in cancer patients. Pharm Res 2000; 17: 1284 1289. 209. Sparreboom A, van Zuylen L, Brouwer E, Loos WJ, de Bruijn P, Gelderblom H et al. Cremophor EL mediated alteration of paclitaxel distribution in h uman blood: clinical pharmacokinetic implications. Cancer Res 1999; 59: 1454 1457. 210. Wolner E, Domanig E, Jr., Elkadi A, Helmer F, Romer P. [On a simple relation between blood volume and body surface area]. Z Kreislaufforsch 1968; 57: 79 84. 211. Spar reboom A, Verweij J, van der Burg ME, Loos WJ, Brouwer E, Vigano L et al. Disposition of Cremophor EL in humans limits the potential for modulation of the multidrug resistance phenotype in vivo. Clin Cancer Res 1998; 4: 1937 1942. 212. van Zuylen L, Karls son MO, Verweij J, Brouwer E, de Bruijn P, Nooter K et al. Pharmacokinetic modeling of paclitaxel encapsulation in Cremophor EL micelles. Cancer Chemother Pharmacol 2001; 47: 309 318. 213. Grochow LB, Baraldi C, Noe D. Is Dose Normalization to Weight or B ody Surface Area Useful in Adults. Journal of the National Cancer Institute 1990; 82: 323 325. 214. Gibbs JP, Gooley T, Corneau B, Murray G, Stewart P, Appelbaum FR et al. The impact of obesity and disease on busulfan oral clearance in adults. Blood 1999; 93: 4436 4440. 215. de Jongh FE, Verweij J, Loos WJ, de Wit R, de Jonge MJ, Planting AS et al. Body surface area based dosing does not increase accuracy of predicting cisplatin exposure. J Clin Oncol 2001; 19: 3733 3739. 216. Bruno R, Vivier N, Veyrat F ollet C, Montay G, Rhodes GR. Population pharmacokinetics and pharmacokinetic pharmacodynamic relationships for docetaxel. Investigational New Drugs 2001; 19: 163 169. 217. Cassidy J, Twelves C, Cameron D, Steward W, O'Byrne K, Jodrell D et al. Bioequival ence of two tablet formulations of capecitabine and exploration of age, gender, body surface area, and creatinine clearance as factors influencing systemic exposure in cancer patients. Cancer Chemother Pharmacol 1999; 44: 453 460. 218. Powis G, Reece P, A hmann DL, Ingle JN. Effect of body weight on the pharmacokinetics of cyclophosphamide in breast cancer patients. Cancer Chemother Pharmacol 1987; 20: 219 222.
151 219. Teresi ME, Riggs CE, Webster PM, Adams MJ, Noonan PK, O'Donnell JP. Bioequivalence of two m ethotrexate formulations in psoriatic and cancer patients. Ann Pharmacother 1993; 27: 1434 1438. 220. Mathijssen RH, Verweij J, de Jonge MJ, Nooter K, Stoter G, Sparreboom A. Impact of body size measures on irinotecan clearance: alternative dosing recomme ndations. J Clin Oncol 2002; 20: 81 87. 221. Climente Marti M, Merino Sanjuan M, Almenar Cubells D, Jimenez Torres NV. A Bayesian method for predicting 5 fluorouracil pharmacokinetic parameters following short term infusion in patients with colorectal can cer. J Pharm Sci 2003; 92: 1155 1165. 222. Takama H, Tanaka H, Nakashima D, Ueda R, Takaue Y. Population pharmacokinetics of intravenous busulfan in patients undergoing hematopoietic stem cell transplantation. Bone Marrow Transplant 2006; 37: 345 351. 22 3. Gieschke R, Burger HU, Reigner B, Blesch KS, Steimer JL. Population pharmacokinetics and concentration effect relationships of capecitabine metabolites in colorectal cancer patients. Br J Clin Pharmacol 2003; 55: 252 263. 224. Urien S, Brain E, Bugat R Pivot X, Lochon I, Van ML et al. Pharmacokinetics of platinum after oral or intravenous cisplatin: a phase 1 study in 32 adult patients. Cancer Chemother Pharmacol 2005; 55: 55 60. 225. Lindemalm S, Savic RM, Karlsson MO, Juliusson G, Liliemark J, Alber tioni F. Application of population pharmacokinetics to cladribine. BMC Pharmacol 2005; 5: 4. 226. Joerger M, Huitema AD, Richel DJ, Dittrich C, Pavlidis N, Briasoulis E et al. Population pharmacokinetics and pharmacodynamics of doxorubicin and cyclophosph amide in breast cancer patients: a study by the EORTC PAMM NDDG. Clin Pharmacokinet 2007; 46: 1051 1068. 227. Nakade S, Ohno T, Kitagawa J, Hashimoto Y, Katayama M, Awata H et al. Population pharmacokinetics of aprepitant and dexamethasone in the preventi on of chemotherapy induced nausea and vomiting. Cancer Chemother Pharmacol 2008; 63: 75 83. 228. Minami H, Kawada K, Sasaki Y, Tahara M, Igarashi T, Itoh K et al. Population pharmacokinetics of docetaxel in patients with hepatic dysfunction treated in an oncology practice. Cancer Sci 2009; 100: 144 149. 229. Thomas F, Rochaix P, White Koning M, Hennebelle I, Sarini J, Benlyazid A et al. Population pharmacokinetics of erlotinib and its
152 pharmacokinetic/pharmacodynamic relationships in head and neck squamous cell carcinoma. Eur J Cancer 2009; 45: 2316 2323. 230. You B, Tranchand B, Girard P, Falandry C, Ribba B, Chabaud S et al. Etoposide pharmacokinetics and survival in patients with small cell lung cancer: a multicentre study. Lung Cancer 2008; 62: 261 272 231. Kovarik JM, Hsu CH, McMahon L, Berthier S, Rordorf C. Population pharmacokinetics of everolimus in de novo renal transplant patients: impact of ethnicity and comedications. Clin Pharmacol Ther 2001; 70: 247 254. 232. Jiang X, Galettis P, Links M, Mitchell PL, McLachlan AJ. Population pharmacokinetics of gemcitabine and its metabolite in patients with cancer: effect of oxaliplatin and infusion rate. Br J Clin Pharmacol 2008; 65: 326 333. 233. Freyer G, Tranchand B, Ligneau B, Ardiet C, Souquet PJ, Court Fortune I et al. Population pharmacokinetics of doxorubicin, etoposide and ifosfamide in small cell lung cancer patients: results of a multicentre study. Br J Clin Pharmacol 2000; 50: 315 324. 234. Schmidli H, Peng B, Riviere GJ, Capdeville R, Hensl ey M, Gathmann I et al. Population pharmacokinetics of imatinib mesylate in patients with chronic phase chronic myeloid leukaemia: results of a phase III study. Br J Clin Pharmacol 2005; 60: 35 44. 235. Klein CE, Gupta E, Reid JM, Atherton PJ, Sloan JA, P itot HC et al. Population pharmacokinetic model for irinotecan and two of its metabolites, SN 38 and SN 38 glucuronide. Clin Pharmacol Ther 2002; 72: 638 647. 236. Mougenot P, Pinguet F, Fabbro M, Culine S, Poujol S, Astre C et al. Population pharmacokine tics of melphalan, infused over a 24 hour period, in patients with advanced malignancies. Cancer Chemother Pharmacol 2004; 53: 503 512. 237. Faltaos DW, Hulot JS, Urien S, Morel V, Kaloshi G, Fernandez C et al. Population pharmacokinetic study of methotre xate in patients with lymphoid malignancy. Cancer Chemother Pharmacol 2006; 58: 626 633. 238. van Ruth S, Mathot RA, Sparidans RW, Beijnen JH, Verwaal VJ, Zoetmulder FA. Population pharmacokinetics and pharmacodynamics of mitomycin during intraoperative h yperthermic intraperitoneal chemotherapy. Clin Pharmacokinet 2004; 43: 131 143. 239. Bastian G, Barrail A, Urien S. Population pharmacokinetics of oxaliplatin in patients with metastatic cancer. Anticancer Drugs 2003; 14: 817 824.
153 240. Henningsson A, Spa rreboom A, Sandstrom M, Freijs A, Larsson R, Bergh J et al. Population pharmacokinetic modelling of unbound and total plasma concentrations of paclitaxel in cancer patients. Eur J Cancer 2003; 39: 1105 1114. 241. Latz JE, Chaudhary A, Ghosh A, Johnson RD. Population pharmacokinetic analysis of ten phase II clinical trials of pemetrexed in cancer patients. Cancer Chemother Pharmacol 2006; 57: 401 411. 242. Boni JP, Leister C, Bender G, Fitzpatrick V, Twine N, Stover J et al. Population pharmacokinetics of CCI 779: correlations to safety and pharmacogenomic responses in patients with advanced renal cancer. Clin Pharmacol Ther 2005; 77: 76 89. 243. Eriksson T, Hoglund P, Turesson I, Waage A, Don BR, Vu J et al. Pharmacokinetics of thalidomide in patients wit h impaired renal function and while on and off dialysis. J Pharm Pharmacol 2003; 55: 1701 1706. 244. Huitema ADR, Mathot RAA, Tibben MM, Schellens JHM, Rodenhuis S, Beijnen JH. Population pharmacokinetics of thioTEPA and its active metabolite TEPA in pati ents undergoing high dose chemotherapy. Br J Clin Pharmacol 2001; 51: 61 70. 245. Mould DR, Holford NHG, Schellens JHM, Beijnen JH, Hutson PR, Rosing H et al. Population pharmacokinetic and adverse event analysis of topotecan in patients with solid tumors Clin Pharmacol Ther 2002; 71: 334 348. 246. Deporte Fety R, Simon N, Fumoleau P, Campone M, Kerbrat P, Bonneterre J et al. Population pharmacokinetics of short intravenous vinorelbine infusions in patients with metastatic breast cancer. Cancer Chemother Pharmacol 2004; 53: 233 238.
154 BIOGRAPHICAL SKETCH Rong Shi was born in Oct ober 1981, in P.R. China. In 2004, she received her Unive rsity of Technology, Hangzhou, P R. China. The same year, she joined the graduate prog ram at University of Missouri Rolla (now Missouri University of Science and Technology), and received her Master of Science degree in Chemistry two years later In 2006, Rong Shi enrolled in the graduate program in Pharmaceutic s at University of Florida, under the supervision of Dr. Hartmut Derendorf to work on her Doctor of Philosophy. Rong has finished two internships during her Ph.D. study in the department of Clinical P harmac ology at Pfizer La Jolla, CA an Pfizer Groton, CT. Rong completed her Ph.D. in May 2011.