<%BANNER%>

Pharmacokinetic-Pharmacodynamic Modeling of Armodafinil

Permanent Link: http://ufdc.ufl.edu/UFE0044918/00001

Material Information

Title: Pharmacokinetic-Pharmacodynamic Modeling of Armodafinil Effects on Electroencephalogram and Neurocognition of Sleep Deprived Adults
Physical Description: 1 online resource (129 p.)
Language: english
Creator: Conrado, Daniela Joice
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2012

Subjects

Subjects / Keywords: armodafinil -- electroencephalogram -- modeling -- pharmacodynamics -- pharmacokinetics
Pharmaceutics -- Dissertations, Academic -- UF
Genre: Pharmaceutical Sciences thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Subjective measures are extensively used to assess drug effects on the central nervous system. Electroencephalogram (EEG) offers the possibility of an objective and continuous measure which also does not require active cooperation of the subjects. The aim of this study was to evaluate EEG as a measure of wakefulness-promoting effect. We performed a double-blind, placebo-controlled, cross-over study consisting of three 48-hour sessions (placebo, 150 mg and 250 mg armodafinil oral tablets). In each session, the subjects (n=6) underwent 36-hour sleep deprivation period during which blood sampling, EEG recording and neurobehavioral assessments were simultaneously performed. The effects of sleep deprivation were significantly mitigated by armodafinil when compared to placebo. Shortly, (1) armodafinil increased the performance of the subjects during the behavioral tasks (i.e. psychomotor vigilance task, PVT, and go/no-go association task, GNAT, as measured by the mean reciprocal reaction time and error rates); (2) armodafinil increased the event-related brain activity in the central region of the brain during the execution of the both behavioral tasks, PVT and GNAT; (3) armodafinil mitigated the increase in the EEG delta power over the frontal, temporal and occipital region of the brain. Using a pharmacokinetic-pharmacodynamic modeling approach, the armodafinil effect on the behavioral measures and event-related brain activity was best described by an excitatory (Emax) drug effect model. On the other hand, armodafinil effect on the EEG delta power was best described by an inhibitory (Imax) drug effect model. In all the aforementioned models, drug effect was linked to the hypothetical drug concentrations in the site of action (i.e. effect compartment approach). A statistically significant correlation between the population predictions for the drug effect on the mean 1/RT of the PVT (i.e. a well-established measure of alertness) and EEG-based measures (brain activity and delta power) suggests EEG as a potential biomarker of armodafinil effect building foundation for further research on this topic.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Daniela Joice Conrado.
Thesis: Thesis (Ph.D.)--University of Florida, 2012.
Local: Adviser: Derendorf, Hartmut C.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2014-12-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2012
System ID: UFE0044918:00001

Permanent Link: http://ufdc.ufl.edu/UFE0044918/00001

Material Information

Title: Pharmacokinetic-Pharmacodynamic Modeling of Armodafinil Effects on Electroencephalogram and Neurocognition of Sleep Deprived Adults
Physical Description: 1 online resource (129 p.)
Language: english
Creator: Conrado, Daniela Joice
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2012

Subjects

Subjects / Keywords: armodafinil -- electroencephalogram -- modeling -- pharmacodynamics -- pharmacokinetics
Pharmaceutics -- Dissertations, Academic -- UF
Genre: Pharmaceutical Sciences thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Subjective measures are extensively used to assess drug effects on the central nervous system. Electroencephalogram (EEG) offers the possibility of an objective and continuous measure which also does not require active cooperation of the subjects. The aim of this study was to evaluate EEG as a measure of wakefulness-promoting effect. We performed a double-blind, placebo-controlled, cross-over study consisting of three 48-hour sessions (placebo, 150 mg and 250 mg armodafinil oral tablets). In each session, the subjects (n=6) underwent 36-hour sleep deprivation period during which blood sampling, EEG recording and neurobehavioral assessments were simultaneously performed. The effects of sleep deprivation were significantly mitigated by armodafinil when compared to placebo. Shortly, (1) armodafinil increased the performance of the subjects during the behavioral tasks (i.e. psychomotor vigilance task, PVT, and go/no-go association task, GNAT, as measured by the mean reciprocal reaction time and error rates); (2) armodafinil increased the event-related brain activity in the central region of the brain during the execution of the both behavioral tasks, PVT and GNAT; (3) armodafinil mitigated the increase in the EEG delta power over the frontal, temporal and occipital region of the brain. Using a pharmacokinetic-pharmacodynamic modeling approach, the armodafinil effect on the behavioral measures and event-related brain activity was best described by an excitatory (Emax) drug effect model. On the other hand, armodafinil effect on the EEG delta power was best described by an inhibitory (Imax) drug effect model. In all the aforementioned models, drug effect was linked to the hypothetical drug concentrations in the site of action (i.e. effect compartment approach). A statistically significant correlation between the population predictions for the drug effect on the mean 1/RT of the PVT (i.e. a well-established measure of alertness) and EEG-based measures (brain activity and delta power) suggests EEG as a potential biomarker of armodafinil effect building foundation for further research on this topic.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Daniela Joice Conrado.
Thesis: Thesis (Ph.D.)--University of Florida, 2012.
Local: Adviser: Derendorf, Hartmut C.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2014-12-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2012
System ID: UFE0044918:00001


This item has the following downloads:


Full Text

PAGE 1

1 PHARMACOKINETIC PHARMACODYNAMIC MODELING OF ARMODAFINIL: EFFECTS ON ELECTROENCEPHALOGRAM AND NEUROC OGNITION OF SLEEP DEPRIVED ADULTS By DANIELA JOICE CONRADO A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSIT Y OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2012

PAGE 2

2 2012 Daniela Joice Conrado

PAGE 3

3 To my Grandparents

PAGE 4

4 ACKNOWLEDGMENTS I would like to thank Dr. Hartmut Derendorf for giving me the wonderful opportunity of working on a project with a focus on quantitative neuropharmacology. His excellent mentorship will never be forgotten. I would also like to express my special gratitude to Drs. Micha el Bewernitz, Mingzhou Ding, C ristoph Seubert, Jean Cibula, Anthony Palmieri, Gunther Hochhaus and Stephan Eisenchenk for all their excellent guidance and contribution. I would like to acknowledge Dr. Julie Johnson who introduced me to the field of pharmac ogenomics and allowed me to enrich my expertise. I am also thankful to the office staff in the Department of Pharmaceutics, in special Patricia Khan and Sarah Foxx, and the personnel in the Clinical Research Center Shands Hospital for their hard work. I ex tend my gratitude to Drs. Issam Zineh, Michael Pacanowski and Ping Zhao for contributing to my training on clinical pharmacology. I would like to express my appreciation to our funding sources: Clinical Translational Science Institute at the University of Florida, Clinical Research Center at the University of Florida, Department of Pharmaceutics at the University of Florida, and CNPq (Brasilia Brazil; pre doctoral fellowship). Moreover, I would like to thank Dr. Teresa Dalla Costa who presented me to the f ield of pharmacokinetics and guided me during my master studies. I am also thankful to my labmates and friends In special, I would like to thank Angelo Zanotto, Bhargava Kandala, Cristian Cocconcelli, Danny Gonzalez Hari Ananthula, Joao Plinio Juchem Net o and Sherwin Sy Finally, I would like to thank my family for their support.

PAGE 5

5 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ .. 4 LIST OF T ABLES ................................ ................................ ................................ ............ 7 LIST OF FIGURES ................................ ................................ ................................ .......... 8 ABSTRACT ................................ ................................ ................................ ................... 12 CHAPTER 1 INTRODUCTION ................................ ................................ ................................ .... 14 Study Rationale ................................ ................................ ................................ ...... 14 Hypothesis and Specific Aims ................................ ................................ ................. 15 Background ................................ ................................ ................................ ............. 17 Excessive Sleepiness ................................ ................................ ....................... 17 Current scenario ................................ ................................ ........................ 17 Management strategi es and armodafinil ................................ .................... 18 Armodafinil pharmacokinetics ................................ ................................ .... 20 Pharmacodynamic Measures of Central Drug Effects ................................ ...... 21 Electroencephalogram based measures ................................ ................... 22 Significance and Innovation ................................ ................................ .................... 26 2 RESEARCH DESIG N AND METHODS ................................ ................................ .. 30 Clinical Study ................................ ................................ ................................ .......... 30 Bioanalytical Analysis ................................ ................................ ............................. 32 Sol vents and Chemicals ................................ ................................ ................... 32 Preparation of Standard Solutions and Quality Control Samples ..................... 32 Instrumentation ................................ ................................ ................................ 33 Sample Preparation ................................ ................................ .......................... 34 Chromatographic and Mass Spectrometer Conditions ................................ ..... 34 Validation Procedure ................................ ................................ ........................ 35 Stability Studies ................................ ................................ ................................ 36 Behavioral Based Measures ................................ ................................ ................... 36 E lectroencephalogram Based Measures ................................ ................................ 37 Pharmacokinetic Pharmacodynamic Modeling ................................ ....................... 39 Statistical Analysis ................................ ................................ ................................ .. 40 3 BIOANALYTICAL ANALYSIS ................................ ................................ ................. 46 Rationale ................................ ................................ ................................ ................. 46 Results ................................ ................................ ................................ .................... 46 Selectivity and Recovery ................................ ................................ .................. 46

PAGE 6

6 Linearity and Lower Limit of Quantification ................................ ....................... 46 Precision and Accuracy ................................ ................................ .................... 46 Stability ................................ ................................ ................................ ............. 47 4 BEHAVIORAL BASED MEASURES ................................ ................................ ...... 51 Rationale ................................ ................................ ................................ ................. 51 Results ................................ ................................ ................................ .................... 51 Psychomotor Vigilance Task ................................ ................................ ............ 51 Go/No Go Association Task ................................ ................................ ............. 52 5 ELECTROENCEPHALOGRAM BASED MEASURES ................................ ............ 60 Rationale ................................ ................................ ................................ ................. 60 Results ................................ ................................ ................................ .................... 60 Event Related Potentials ................................ ................................ .................. 60 Spectral Analysis ................................ ................................ .............................. 61 6 PHARMACOKI NETIC PHARMACODYNAMIC MODELING ................................ ... 83 Rationale ................................ ................................ ................................ ................. 83 Results ................................ ................................ ................................ .................... 83 Pharma cokinetics ................................ ................................ ............................. 83 Pharmacokinetic Pharmacodynamic Modeling ................................ ................. 84 Behavioral based measures ................................ ................................ ...... 85 Electroencephalogram based measures ................................ ................... 86 Correlation between behavioral alertness and electroencephalogram based measures ................................ ................................ ..................... 88 7 DISCUSSION AND CONCLUSIONS ................................ ................................ .... 106 Spectral Analysis ................................ ................................ ................................ .. 106 Psychomotor Vigilance Task ................................ ................................ ................. 109 Go/No Go Association Task ................................ ................................ ................. 111 Event Related Brain Activity ................................ ................................ ................. 113 Pharmacokinetics ................................ ................................ ................................ 114 Pharmacokinetic Pharmacodynamic Modeling ................................ ..................... 114 Study Limitations ................................ ................................ ................................ .. 117 S tudy Strengths ................................ ................................ ................................ .... 118 Conclusions ................................ ................................ ................................ .......... 118 LIST OF REFERENCES ................................ ................................ ............................. 120 BIOGRA PHICAL SKETCH ................................ ................................ .......................... 129

PAGE 7

7 LIST OF TABLES Table page 3 1 Intra and inter day variation of armodafinil determination in human plasma ..... 48 3 2 Intra and inter day variation of armodafinil determination in human plasma ..... 48 3 3 Bench top and autosampler stability of armodafinil in human plas ma ................ 48 3 4 Freeze thaw stability of armodafinil in human plasma ................................ ........ 49 6 1 Pharmacokinetic parameters of armodafinil. ................................ ...................... 89 6 2 Pharmacodynamic parameters of the armodafinil effect on the psychomotor vigilance task mean reciprocal reaction time. ................................ ..................... 89 6 3 Pharmacodyn amic parameters of the armodafinil effect on the go/no go association task mean reciprocal reaction time. ................................ ................. 89 6 4 Pharmacodynamic parameters of the armodafinil effect on the event related brain activity during the psychomotor vigilance task ................................ ........... 90 6 5 Pharmacodynamic parameters of the armodafinil effect on the event related brain activity during the go/no go ass ociation task (no go cond ition) ................ 90 6 6 Pharmacodynamic parameters of the armodafinil effect on the delta power of the occipital region of the brain. ................................ ................................ .......... 90

PAGE 8

8 LIST OF FIGURES Figure page 1 1 Schematic representation of the pharmacokinetic (PK) pharmac odynamic (PD) modeling approach ................................ ................................ ..................... 28 1 2 Schematic represen tation of different event related potential components ........ 29 2 1 Schematic representation of the study design ................................ .................... 41 2 2 Schematic repr esentation of the psychomotor vigilance task ............................ 42 2 3 Schematic representation of the go/no go task ................................ .................. 43 2 4 Schematic representati on of the 1 0 10 average virtual montage with 27 channels. ................................ ................................ ................................ ............ 44 2 5 Schematic representation of the electroencephalogram power spectral analysis. ................................ ................................ ................................ ............. 45 3 1 Representative chromatograms of calibration curve in human plasma.. ............ 50 4 1 Psychomotor vigilance task number of lapses plus false starts ......................... 54 4 2 Psychomotor vigilance task mean reciprocal reaction time ............................... 55 4 3 Go/no go association task mean reciprocal reaction time ................................ .. 56 4 4 Go/no go association task number of errors of omission ................................ ... 57 4 5 Go/no go association task number of errors of comission ................................ 58 4 6 and go/no go association task, GNAT ................................ ................................ 59 5 1 Grand average event related brain a ctivity during the execution of the psychomotor vigilance task ................................ ................................ ................. 63 5 2 Average event related brain activity, ERP, during the execution of the psychomotor vigilance task, PVT (Cz channel, positi ve peak at around 380 ms). ................................ ................................ ................................ .................... 64 5 3 related brain activity, ERP, during the execution of the psychomotor vigilance task, PVT, and the PVT mean re ciprocal reaction time ................................ ................................ ............. 65

PAGE 9

9 5 4 related brain activity, ERP, during the execution of the psychomotor vigilance task, PVT, and the PVT number of l apses plus false starts. ................................ ................................ ..... 66 5 5 Grand average event related brain activity during the execution of the go/no go association task (no go condition) ................................ ................................ 67 5 6 Average event related brain activity, ERP, during the execution of the go/no go association task, GNAT (go condition, Cz channel, positive peak at around 380 ms) ................................ ................................ ................................ .. 68 5 7 Average event related brain activity, ERP, during the execution of the go/no go association task, GNAT (no go condition, Cz channel, positive peak at around 380 ms) ................................ ................................ ................................ .. 69 5 8 ation analysis between the event related brain activity, ERP, during the go/no go association task, GNAT, and the GNAT mean reciprocal reaction time ................................ ................................ ................................ ....... 70 5 9 en the event related brain activity, ERP, during the go/no go association task, GNAT, and the GNAT number of errors of omission. ................................ ................................ ................................ ........ 71 5 10 relate d brain activity, ERP, during the go/no go association task, GNAT, and the GNAT number of errors of comission. ................................ ................................ ................................ ....... 72 5 11 related brain activity, ERP, during the go/no go association task, GNAT, and the psychomotor vigilance task, PVT, mean reciprocal reaction time ................................ ........................... 73 5 12 F rontal (F), temporal (T) and occipital (O) regions of the brain shown in the 1 0 10 average virtual montage ................................ ................................ ........... 74 5 13 Electroencephalogram delta band spectral power in the electrode F3 of the frontal area ................................ ................................ ................................ ......... 75 5 14 Electroencephalogram delta band spectral power in the electrode F4 of the frontal area ................................ ................................ ................................ ......... 76 5 15 Electroencephalogram delta band spectral power in the electrode T7 of the temp oral area. ................................ ................................ ................................ .... 77 5 16 Electroencephalogram delta band spectral power in the electrode T8 of the temporal area. ................................ ................................ ................................ .... 78 5 17 Electroencep halogram delta band spectral power in the electrode O1 of the occipital area. ................................ ................................ ................................ ..... 79

PAGE 10

10 5 18 Electroencephalogram delta band spectral power in the electrode Oz of the occipital area. ................................ ................................ ................................ ..... 80 5 19 Electroencephalogram delta band spectral power in the electrode O2 of the occipital area. ................................ ................................ ................................ ..... 81 5 20 the delta power the psychomotor vigilance task, PVT, mean reciprocal reaction time. ................................ ........... 82 6 1 Visual predictive check of the final armodafinil pharmacokinetic model.. ........... 91 6 2 Basic goodness of fit plots of the final armodafinil pharmacokinetic model ........ 92 6 3 Anticlockwise hysteresis loop between armodafinil pharmacokinetics and pharmacodynamics. ................................ ................................ ........................... 93 6 4 Schematic representation of the effect compartment approach ......................... 94 6 5 Visual predictive check of the final pharmacokinetic/pharmacodynamic model of the armodafinil effect on the psychomotor vigilance task mean reciprocal reaction time ................................ ................................ ................................ ....... 95 6 6 Basic goodness of fit plots of the final pharmacoki netic/pharmacodynamic model of the armodafinil effect on the psychomotor vigilance task mean reciprocal reaction time. ................................ ................................ ..................... 96 6 7 Visual predictive check of the final pharmacokinetic/pharmacodynam ic model of the armodafinil effect on the go/no go association task mean reciprocal reaction time ................................ ................................ ................................ ....... 97 6 8 Basic goodness of fit plots of the final pharmacokinetic/pharmacodynamic model of the a rmodafinil effect on the go/no go association task mean reciprocal reaction time ................................ ................................ ...................... 98 6 9 Visual predictive check of the final pharmacokinetic/pharmacodynamic model of the armodafinil effect on the event related brain activity, ERP, during the psychomotor vigilance task, PVT. ................................ ................................ ....... 99 6 10 Basic goodness of fit plots of the final pharmacokinetic/pharmacodynamic model of the armodafinil eff ect on the event related brain activity during the psychomotor vigilance task ................................ ................................ ............... 100 6 11 Visual predictive check of the final pharmacokinetic/pharmacodynamic model of the armodafinil effect on t he event related brain activity, ERP, during the go/no go association task, GNAT (no go condition). ................................ ........ 101

PAGE 11

11 6 12 Basic goodness of fit plots of the final pharmacokinetic/pharmacodynamic model of t he armodafinil effect on the event related brain activity during the go/no go association task (no go condition). ................................ .................... 102 6 13 Visual predictive check of the final pharmacokinetic/pharmacodynamic model of the armodafinil effect on the delta power of the occipital region of the brain 103 6 14 Basic goodness of fit plots of the final pharmacokinetic/pharmacodynamic model of the armodafin il effect on the delta power of the occipital region of the brain ................................ ................................ ................................ ........... 104 6 15 Correlation between behavioral alertness and electroencephalogram based measures ................................ ................................ ................................ .......... 105

PAGE 12

12 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 PHARMACOKINETIC PHARMACODYNAMIC MODELING OF ARMODAFINIL: EFFECTS ON E LECTROENCEPHALOGRAM AND NEUROCOGNITION OF SLEEP DEPRIVED ADULTS By Daniela Joice Conrado December 2012 Chair: Hartmut Derendorf Major: Pharmaceutical Science s Subjective measures are extensively used to assess drug effects on the central nervous system. Electroencephalogram (EEG) offers the possibility of an objective and continuous measure which also does not require active cooperation of the subjects. The aim of this study was to evaluate EEG as a measure of wakefulness promoting effect. We performed a double blind, placebo controlled, cross over study consisting of three 48 hour sessions (placebo, 150 mg and 250 mg armodafinil oral tablets). In each session, the subjects (n=6) underwent 36 hour sleep deprivation period during which blood sampling, EEG recording and neurobehavioral assessments were simultaneously performed. The effects of sleep deprivation were significantly mitigated by armoda finil when compared to placebo. Shortly, (1) armodafinil increased the performance of the subjects during the be havioral tasks ( i.e., psychomotor vigilance task, PVT, and go/no go association task, GNAT, as measured by the mean reciprocal reaction time and error rates ) ; (2) armodafinil increase d the event related brain activity in the central region of the brain dur ing the execution of the both behavioral tasks, PVT and GNAT; (3) armodafinil mitigated the increase in the EEG delta power over the frontal, temporal and

PAGE 13

13 occipital region of the brain. Using a pharmacokinetic pharmacodynamic modeling approach, the armodaf inil effect on the behavioral measures and event related brain activity was best described by an excitatory (Emax) drug effect model. On the other hand, armodafinil effect on the EEG delta power was best described by an inhibitory (Imax) drug effect model. In all the aforementioned models, drug effect was linked to the hypothetical drug concentrations in the site of action ( i.e., effect compartment approach). A statistically significant correlation between the population predictions for the drug effect on t he mean 1/RT of the PVT ( i.e., a well established measure of alertness) and EEG based me as ures (brain activity and delta power) suggest s EEG as a potential biomarker of armodafinil effect building foundation for further research on this topic.

PAGE 14

14 CHAPTER 1 INTRODUCTION Study Rationale The CTSI funding opportunity has led to the consolidation of a multi disciplinary research team involving investigators from the College of Pharmacy, College of Engineering and College of Medicine. As a result of this collabora tion, the present research addresses a genuine need in the pharmaceutical area through integration of clinical pharmacology and biomedical engineering related concepts and approaches. A therapeutic effect is obtained when the minimum effective concentratio n of a drug reaches the site of action and binds to its target. While the drug target interaction constitutes the pharmacodynamics (PD), the amount of drug available in the biophase will be determined by its pharmacokinetic (PK) properties. Although drug c oncentrations are easily measured and sometimes used to guide treatment, they are not always directly correlated with drug effect. The concentration effect relationship may vary between subjects, and even within subjects under the influence of pathophysiol ogical variables. Defining how concentration and effect are correlated and which parameters affect this relation ship is the objective of the PK PD modeling (Figure 1 1) 1 2 It allows an accurate assessment of the clinical significance of PK and/or PD changes due to drug drug interactions, disease states, age, gender, and other genetic and environmental factors 3 Moreover, a reproducible, precise and accurate pharmacodynamic measure has the ability to enhance drug development 4 by optimization of study design, it can reduce cost and time of development. While assessing the pharmacokinetic component is relatively straightforward, measuring pharmacological effect may be cumbersome. The latter is particularly true for

PAGE 15

15 central nervous system (CNS) acting drugs 3 Most methods for measuring central drug effects rely on subjective assessments 5 7 Self rated scales require individuals to rate their perceived intensity of some outcome measure, such as sleep quality after treatment with a sedative drug 8 This limitation presents an opportunity to investigate more objective pharmacodynamics measures to describe drug response. Electroencephalogram (EEG) brings the possibility of an objective, nearly continuous and non invasive measure of central effec ts 3 9 11 There are a few drug classes for which EEG has been systematically investigated with the prime exa mples being benzodiazepines 12 anesthetics an d opioids 9 The goal of the present research was to evaluate EEG based approaches as measures of stimulant effect. Armodafinil (Nu vigil ), a novel non amphetamine stimulant medication, was selected due to its low potential for abuse 13 In addition, its closely related drug modafinil has demonstrated effects detectable with EEG 11 14 The main hypothesis is that armodafinil related changes on EEG are correlated with a well established neurocognitive measure of alertness ( i.e., psychomotor vigilance task) and drug concentrations. Such a wide range of measures may provide insight of the clinical meaning of changes on EEG. This hypothesis can be stratified and inve stigated as below. Hypothesis and Specific Aims Hypothesis 1 Armodafinil effect on the neurocognitive performance of sleep deprived healthy adults is correlated with its concentrations. Specific aim 1 To describe the relationship between neurocognitive p erformance and armodafinil concentrations. We (a) develop ed and validate an analytical method to measure armodafinil concentrations in human plasma; (b) determine d the pharmacokinetics after single oral dose administration of armodafinil

PAGE 16

16 150 mg and 250 mg in sleep deprived healthy adults; (c) determine d and compare d the individual performance of the subjects by two neurocognitive tests (psychomotor vigilance task, PVT, and go/no go association task, GNAT) after armodafinil and placebo administration; (d) es tablish ed a non linear mixed effects PK PD model to describe the relationship between armodafinil effect on neurocognition and its concentration time profile. Hypothesis 2 Armodafinil effect on the EEG power spectrum and event related brain activity of sl eep deprived healthy adults are correlated with its concentrations. Specific aim 2 To describe the relationships between ERP s and EEG frequency specific power and armodafinil concentrations. We will (a) analyze and compare the ERP produced during the exec ution of two neurocognitive tests (PVT and GNAT) after armodafinil and placebo administration; (b) establish a non linear mixed effects PK PD model to describe the relationship between armodafinil effect on event related brain activity and its concentratio n over time; (c) perform a power spectral analysis of eyes closed and eyes open EEG recording after armodafinil and placebo administration; (d) establish a non linear mixed effects PK PD model to describe the relationship between armodafinil effect on spec ific frequency ranges and its concentration time profile. Hypothesis 3 Armodafinil related EEG changes are correlated with its effect on alertness of sleep deprived healthy adults. Specific aim 3 To describe the relationship between drug related EEG chan ges and behavioral alertness. We will (a) describe the relationship between armodafinil

PAGE 17

17 effect on EEG power spectrum and alertness ( i.e., PVT) ; (b) describe the relationship between armodafinil effect on event related brain activity and alertness ( i.e., PV T) Background E xcessive Sleepiness Current s cenario Excessive daytime sleepiness (EDS) is a common problem among patients visiting sleep clinics. It is defined as sleepiness occurring in situations when an individual would typically be expected to be aler t 15 It has been estimated by epidemiologic studies that EDS affects up to 13% of the population 16 The four major causes of EDS are: (1) CNS pathologic disorders such as narcolepsy and idiopathic hypersomnia; (2) qualitative or quantitative sleep deficiencies such a s sleep apnea and sleep deprivation; (3) circadian rhythm misalignments such as jet lag or shift work; and (4) medication side effects 17 Chronic excessive sleepiness is related to substantial morbidity such as reduce d quality of life, impaired work or academic performance and psychosocial distress. It has also been associated to an increased risk of accidents with a consequent impact on public health 16 The National Highway Traffic Safety Administration reported that sleepiness is the main leading factor in about 100,000 of the annual police reported crashes in the United States 18 Ap proximately 4% of all fatal motor vehicle accidents (i.e., 1,500 deaths) occurring every year has been associated to drowsy driving. Moreover, it has been suggested that sleep deprivation played an important role in some of the major disasters of the last decades. Those include the nuclear accident at the Three Mile Island (1979), the nuclear meltdown at Chernobyl (1986), the explosion of the space shuttle Challenger (1986) 19 and the grounding of the Exxon Valdez oil tanker (1989) 20

PAGE 18

18 Management strategies and armodafinil Management strategies for EDS depend on the etiology of the disorder. In respect to the pharmacotherapy, ED S patients are normally treated with amphetamine like stimulants. The most frequently utilized ones are dexam ph etamine, methamphetamine, methylphenidate, mazindol and pemoline 21 Even though approved to alleviate EDS, these agents are associated with potential for abuse, adverse cardiovascular and central effects 22 Modafinil (Provigil ), a more recent non amphetamine wakefulness promoting agent, has demonstrated less potential for abuse 23 It is a racemic mixture of R modafinil and S modafinil approved to promote wakefulness in individuals with excessive sleepiness related to obstru ctive sleep apnea/hypopnea syndrome (OSAHS), narcolepsy and shift work sleep disorder (SWSD). Armodafinil (Nuvigil ), approved by the U.S. Food and Drug Administration agency ( FDA ) in 2007, is the R enantiomer of modafinil. While approved for the same ind ications, armodafinil has an elimination half life time that is three to four times longer than that of the S enantiomer 24 Armodafinil, hence, can sustain higher concentrations late in the day when compared to the racemate 25 The consequence of this has been demonst rated in studies with OSAHS 26 and narcoleptic 27 patients where armodafinil sustained wakefulness throughout the day. Armodafinil exact mechanism of action remains unknown 28 Although armodafinil is a rela tively new drug, its analogous compound, modafinil, has received extensive attention in the literatur e. Studies on modafinil are considere d to shed light on armodafinil pharmacology given that both enantiomers demonstrated

PAGE 19

19 similar pharmacological profile i n pre clinical and in vitro studies 28 Modafinil potential mechanisms of action have been extensively presented in a systematic review : 29 its effect has been related with en hanced adrenergic, glutamatergic histaminergic and hypocretin activity and reduced aminobutyric acid (GABA) activity in certain areas of the brain. There is evidence supporting that modafinil has a distinct mechanism of action when compared to t he conventio nal amphetamine like stimulants I nterestingly, modafinil produces neuronal activation that is more restricted to wakefulness areas ( e.g., h ypothalamus) as opposed to amphetamine like compounds which have a more widespread neuronal activation 29 30 U nlike conventional stimulants, modafinil has been shown to exclusively reduce the release o f GABA and only weakly e nhance the release of dopamine in the nucleus accumbens of rats 31 Moreover, it has been suggested that modafinil modulates the hypocretin system by causing activation of hypocretin secreting neurons 32 Hypocretin or orexin is a hypothalamic peptide that plays a role in the regulation of sleep and wakefulness 33 This pep tide seems to be able to stimulate glutaminergic and histaminergic systems leading to a rousal 29 Therefore modafinil effect on monoamine systems seems to be less important than modulation of GABA, glutamate, hypocretin and histamine 29 Indeed, compared to amphetamines like compounds, modafinil does not produce feelings associated to drug abuse and does not affect the sleep style 14 Several p otential neurological and psychiatric indications have been investigated for modafinil and are likely to be extended for armodafinil A systematic review points out that compelling evidence exists for the use of modafinil in attention deficit and

PAGE 20

20 hyperactive disorder, post anesthetic sedation, cocaine withdrawal and as an adjunct in major depressive disorder 29 Armodafinil pharmacokinetics Armodafinil pharmacokinetics has been previously described in a pooled analysis of three randomized clinical trials 34 Armodafinil exhibit ed linear pharmacokinetics in the dose range of 50 to 400 mg. The rate but not the extent of absorption was affected by food intake. Maximum plasma concentrations were reached at approximately 2 .3 and 6.0 hours post dose in fasted and fed state (after consumption of a fatty meal) respectively. After the maximum plasma concentration, armodafinil concentrations declined in a monoexponential manner havi ng a half life time of approximately 15 hours Armodafinil lipophilic nature precluded the intravenous administration of the drug and, hence, the determination of its absolute oral bioavailability 13 After the multiple dose administration, s teady state seem ed to be reached by the 7 th day post dosing and the systemic exposure was 1.8 times when compared to the single dose 34 Armodafinil mass balance data is not available. Modafinil is metabolized by the liver with less than 10% of the parent compound being excreted in the urine 28 Armodafinil suffers hydrolytic deamidation, S oxidation and aro matic ring hydroxylation with subsequent glucuronide conjugation of the hydroxylated products 28 Amide hydrolysis is the most noticeable metabolic pathway followed by sulfone formation through cytochrome P450 (CYP) 3A4/5 28 In a dedicated drug drug interaction study in healthy subjects, armodafinil did not induce CYP1A2 but was a moderate CYP3A4 inducer and a moderate CYP2C19 inhibitor 35 Although armodafinil was well tolerated when co administered with midazolam and omeprazole 35 one cannot exclude the need

PAGE 21

21 to perform dose adjustment of other co administered CYP3A4 and CYP2C19 substrates. Pha rmacodynamic Measures of Central Drug Effects CNS acting drugs are an important part of the therapeutic arsenal constituting more than 14% of the dispensed prescriptions in the U.S. market 36 While therapeutic drug monitoring is a common practice for many drugs in this class, assessment solely of concentration le vels may be of limited importance. Drug concentration and effect may be correlated in a complex and/or indirect manner. The concentration effect relationship may vary between subjects, and within subjects due to the influence of pathophysiological or exter nal factors. As a consequence, it becomes important to describe the PK PD relationship to account for the pharmacodynamic component of the variability in drug response. A pharmacodynamic measure can be classified as biomarker, surrogate endpoint or clinica l outcome. Although clinical outcomes are considered the primary assessment of effectiveness and safety 2 biomarkers normally occur earlier and can be determined in a more robust manner 37 Moreover, a reproducible, precise and accurate pharmacodynamic measure has the ability to enhance PK PD modeling and consequently expedite drug development. Psychometric tests Measuring pharmacol ogical effect is still cumbersome for many CNS acting drugs. Central drug effects are commonly measured using psychometric tests. Although they might focus on assessment of sensory, cognitive or motor functions, it is the integration of sensory and motor s ystems through the cognitive processing what is measured 38 Psychometric tests can be more or less subjective and robust. They normally require active cooperation of the individuals. In addition, they will likely produce correlated measures when learning effects occur with repeated testing 3

PAGE 22

22 Electroencephalogram based measures Unlike psychometric tests, EEG is considered to meet almost all criteria for an ideal pharmacodynamic measure. Learning effects in EEG based measures have not been reported; they yield an objective, (nearly )continuous, sensitive and non invasive 3 9 39 Source of electroencephalogram activity The electroencephalogram converts small electrical signals arising from neuronal synaptic activity into waveforms 40 Neuronal communi cation occur through synapses: (a) an action potential leads the presynaptic terminal to release neurotransmitter; (b) the neurotransmitter binds to the membrane post synaptic terminal leading to a change in the membrane potential; (c) a sequence of change in the membrane potential are summed in the body of the target neuron resulting in a post synaptic potential; (d) as a consequence, the ion channels open resulting in an action potential which, in turn, will lead to the same sequence of events previously described 40 The post synaptic potentials from these cells are summed together in the extracellular fluid around them and conducted in the following sequence: (a) the volume of th e brain; (b) cerebrospinal fluid; (c) blood; (d) bone; (e) muscle; (f) skin comprising the head; (g) scalp electrodes; (h) EEG amplifier. Electroencephalogram frequencies The clinical utility of a typical scalp EEG resides in four major frequency ranges: (a) b eta ( over 12 or 13 /sec or Hz), usually anterior ; (b) a lpha (8 12 or 13 Hz ), usually posterior ; (c) t heta (4 8 Hz ), usually widespread ; (d) d elta ( under 4 Hz ), associate to drowsiness/sleep 40 41 Another less specific nomenclature is calling waves under 8 Hz as slow waves and waves over 13 Hz as fast waves 41

PAGE 23

23 The alpha rhythm is not only characterized by its frequency, but also by its distribution and reactivity; as a normal consequence, alpha rhythm and alpha frequency are not synonymous 41 This rhythm occurs during wakefulness in the posterior part of the brain showing high voltage in the occipital area ; although the amplitude can vary it should be mostly below 50 V 42 A lpha rhythm is best seen with eyes closed and in situations of physical relaxation and fairly mental inactivi ty; indeed, it is characteristically blocked by attention, in particular, visual and mental efforts 42 It has been also pointed out that alpha rhythm disappears with drowsiness 42 Conversely, beta actitivity is more prominent in the frontal area when the subject is allowed to fall sleep; in addition, drugs such as barbiturates and benzodiazepines has been shown to increase beta activity 42 The theta and delta rhythm play an important role in the childhood, but are also present in the adult life, and are evident in states of drowsiness and sleep 42 Quantitative electroencephalogram analysis Quantitat ive EEG can defined 41 The quantification of EEG signal can be done in the time domain ( i.e., aperiodic analysis) or in the frequency domain ( i.e., fast Fourier analysis) 9 41 43 In the aperiodic analysis the quantitation is performed directly from the analysis of the signal represented in the time domain ; a wave is defined as an oscillation in voltage between two local minima 9 Therefore, the frequency and the amplitude of an EEG signal are calculated wave by wave generating two main variables: (a) the total number of waves per second and (b) total voltage per second for a determined frequency range 43

PAGE 24

24 Fast Fourier transform assumes that the raw EEG waveforms are a summation of waves; as a consequence, this approach decomposes the raw EEG sign al into multiple sine waves with different phases, frequencies and amplitudes 41 In this sense, the squaring of the Fourier coefficien ts ( i.e., numerical coefficients representing these sine waves ) within a specific frequency range generates the power spectrum in a certain time period. The power spectrum contain considerably less data than the raw EEG making the storage of data more effi cient 41 A drawback is that the original EEG signal cann ot be reconstructed from the power spectrum since it does not comprise phase information 41 Event related brain activity Event related brain activity or event related potential ( ERP) is another type of EEG the brain activity due to a specific event which can be an 44 ERP is proposed to designate the general class of 44 ERPs offer the possibility of a continuous assessment of brain processing between a stimulus and a response, allowing one to determine which stage of the information processing is altered when applying an experimental paradigm ( e.g., sensory, cognitive, motor process) 44 In other words, ERPs are not composed of an overlap of individual cognitive processes 44 Conversely the output of a behavioral measure comes from innumerous individual cognitive processes, and the variation in reaction time and accuracy are difficult to associate with a specific variation of a cognitive process 44 Another advantage is that ERPs can be measured in the ab sence of a behavioral response allowing one to

PAGE 25

25 continuously monitor brain processing and compare between conditions of attended with unattended stimulus 44 ERPs are obtained by recording EEG during the time where the subjects are being exposed to multiple repetitions of a certain stimulus ( e.g., visual, somatosensory or a uditory stimulus) 45 In order to filter out background noise, one has to record the exact time in which the stimulus is presented, align the trials based on st imulus appearance and averag e the trials altogether ( i.e., stimulus locked average) 44 The assumption is that the ERP waveforms (related to the stimulus appearance) are identical in each trial repetition but the noise occurs randomly and, hence, cancels out during the average 44 The result of trials average is a sequence of positive and negative voltage deflections called ERP peaks, wave s or components representing the flow of information through the brain (Figure 1 3). 44 The early peak (P1) is a mandatory sensory response provoked by visual stimulus and highly influenced by physical attributes of the stimulus ( i.e., exogenous component) ; conversely, the late peak (P3) depends completely on the task perf ormed by the subject not being directly influenced by the physical attributes of the stimulus ( i.e., endogenous component) 44 The P300 latency is related to the time required to categorize a stimulus; P300 amplitude, in turn, is related to attention and perception and gets larger as the target probability gets smaller 44 46 Finally, as early components, such as P1, are related to sensory processing they are more noticeable in the o ccipital region of the brain ; l ater components such as P3, reflect complex cognitive process ( e.g., categorization of stimulus) being more noticeable in the parietal/central/frontal regions of the brain 44

PAGE 26

26 Pharmaco electroencephalogram Pharmaco the study of drug effects on the electroencephalogr ams of animals, healthy volunteers or patients with respect to the possibility of applying the information obtained to drug 47 EEG has been systematically studied in a few drug classes. Changes in EEG parameter not only were able to differentiate benzodiazepines with a wide range of intrinsic efficacy and affinity by the GABA benzodiazepine receptor complex 48 but also reflected the depth of anesthesia after administration of intravenou s anesthetics 9 Likewise, remifentanil development was expedited based on a well established EEG parameter 49 50 which showed to be correlated with its analgesic efficacy 51 The greatest challenging is that the correlation between EEG parameters and drug effect is not clear yet in many situations 52 It has been claimed that a better understanding of the clinica l meaning of EEG parameters would be obtained if an alternative measure of drug effect were performed in combination with EEG recording 3 Although it has been noticed more than two decades ago, studies investigating correlations between changes in EEG parameters and well established measures of specific drug effects are still limited. Significance and Innovation In light of the previous observations, this research consists of a thorough evaluation of the armodafinil effect in sleep deprived healthy adults. In order to minimize the likelihood of empirical correlations, we conducted EEG recording with simultaneous assessment of armodafinil concentrations as well as alternate measures of armodafinil effect. Amo ng the utilized pharmacodynamic measures, PVT is a well established measure of armodafinil effect on alertness. Investigating the presence of correlation between a specific EEG parameter and alertness will provide insight of the clinical

PAGE 27

27 meaning of changes on EEG. Indeed, a PK PD model integrating both pharmacodynamic measures would allow us to suggest E EG as a potential biomarker of armodafinil effect. Finally, we believe that a PK PD modeling approach incorporating an objective CNS effect measure such as EEG can provide a means to optimize the dose regimen under various clinical scenarios ( e.g., disease states and concomitant medications). It will be of great value to guide appropriate dose selection in clinical trials pursuing new indications for armodafinil and expedite development of novel compounds with an analogous mechanism of action. This research project demonstrates innovation given that no published studies have applied such a wide range of CNS based measures in the context of a PK PD analysis of armodafinil or its analogous compound ( i.e., modafinil)

PAGE 28

28 Figure 1 1. Schematic representation o f the pharmacokinetic (PK) pharmacodynamic (PD) modeling approach. Adapted from Derendorf H, Meibohm B. Modeling of pharmacokinetic/pharmacodynamic (PK/PD) relationships: Concepts and perspectives. Pharmaceutical Research 1999;16:176 85 (Page 177, Figure 1 ) 1

PAGE 29

29 Figure 1 2 Schematic representation of different event related potential components Negative voltage is plotted upward. Adapted from Luck, S. J. What is an ERP. 2012. (Acc essed 10/17, 2012, at http://erpinfo.org/what is an erp Figure 2) 53

PAGE 30

30 CHAPTER 2 RESEARCH DESIGN AND METHODS Clinical Study A double blind, placebo controlled, crossover clinical study was performed at the UF Clinical Research Center, Shands Hospital, University of Florida. T he study protocol was in accordance with the Declaration of Helsinki and was approved by the local Institutional Review Board. The study consisted of three 48 hour visits separated by washout periods of at least 7 days. In each visit, the subjects underwen t a 24 hour sleep deprivation period. After that a single dose of the study drug was administered (placebo, 150 mg, or 250 mg armodafinil oral tablets). The subjects then underwent additional 12 hours of sleep deprivation during which pharmacodynamics and pharmacokinetics assessments were simultaneously carried out. Six subjects were enrolled following a set of inclusion and exclusion criteria. They were healthy adults, between 18 and 35 years of age, with a normal body weight (BMI between 18.5 and 29.9), a nd without clinical sleep abnormalities. At the screening, the study physician conducted an Epworth Sleepiness Scale evaluation and discussed with the subject about his or her sleep history. Eligible subjects agreed not to consume alcohol three days prior to and during any study session. In addition, they did not use any other medication one week prior to and during any study session. They were recommended to avoid caffeine for one week prior to a study session, but required to avoid caffeine for 24 hours p rior to and during a study session. Females had a negative pregnancy test at screening and admission and were abstinent, sterile, postmenopausal, or practicing an effective method of birth control except steroidal contraceptives. Smokers or subjects who us ed a nicotine containing product within 12

PAGE 31

31 months of screening were not eligible. Likewise, subjects with evidence of use of drugs of abuse (including but not limited to barbiturates, opiates, cocaine, cannabinoids, amphetamines, and benzodiazepines) were excluded. On the day of any study session, a set of pre specified procedures were performed. On admission, subjects were interviewed about duration of sleep as well as inclusion and exclusion criteria. Urine drug screen (for all subjects) and pregnancy tes t (for female subjects) were also performed. Subjects had a training session on each consisted of simultaneous EEG recording and execution of the neurobehavioral assays. baseline session was performed. The procedures were identical to the resting baseline session. Subsequently, the subjects received the study drug and 10 experimental sessions were performed along the period of 12 hours. The experimental sessions were identical to the baseline session but blood sampling was also performed. Nature and order of the procedures are provided in Figure 2 1 Blood collection was performed through an i ntravenous catheter placed in the arm vein of the subjects. Blood sampling was conducted prior to drug dosing and at 0.5, 1, 1.5, 2, 3, 4, 6, 8, 10, and 12 hours after dosing. Blood samples were approximately 4 mL in volume and were collected into EDTA con taining collection tubes and centrifuged at 1,300 g for 10 min at 4 C. Plasma was stored at 80 C until analysis. Subjects received a standardized and caffeine free diet during each study visit. In the morning of the drug dosing, subjects were encouraged to have breakfast prior to the administration of the study drug because of the full schedule of data collection.

PAGE 32

32 Protein binding determination was conducted at 2 hours and 12 hours after drug dosing. Once the blood samples were obtained from the subjects, those were processed as previously described. An aliquot of plasma was taken to determine the armodafinil total concentration. The remaining volume was placed in a 37C water bath for 30 minutes. After this time, 1 mL of the sample was added to a filtrate device (Centrifee; Millipore Corporation, Billerica, MA) which was centrifuged at 1,800 g for 25 minutes at 37C. The filtrate device underwent previous acclimation in the centrifuge at 37C. Upon completion of centrifugation, the filter was removed from the device and the ultrafiltrate (<11% of the total plasma volume) was submitted to visual inspection for membrane damage. The ultrafiltrate yielded armodafinil unbound concentration. Simultaneously, samples of armodafinil in aqueous solution (50 4200 ng m L 1 ) were run to assess armodafinil binding to the membrane of the filtration device ( i.e., non specific binding). The difference of armodafinil total and unbound concentrations corrected by non specific binding was assumed to be the armodafinil plasma pro tein binding. Bioanalytical Analysis Solvents and Chemicals P rednisolone acetate and a nalytical grade ammonium acetate were purchased from Sigma Aldrich (St. Louis, MO). High performance liquid chromatography grade methanol was purchased from Fisher Scient ific (Fair Lawn, NJ). Analytical grade acetic acid was obtained from Fisher Scientific (Fair Lawn, NJ). Double distilled water was obtained in house (Department of Pharmaceutics, University of Florida). Preparation of Standard Solutions and Quality Control Samples Two individual weighing of armodafinil was performed to prepare standard stock solutions of 480 g mL 1 in methanol. One stock solution was used to generate the

PAGE 33

33 working solutions for the calibration curve standards and the other, the quality contr ol (QC) samples. The calibration cu rve samples were prepared by a sequence of dilution steps: (a) 100 L of the 480 g mL 1 stock solution was diluted with 100 L of methanol d ouble distilled water (1:1) to yield a 240 g mL 1 solution of armodafinil; (b) 20 L of the 24 0 g mL 1 solution was diluted with 980 L of human plasma to yield a final concentration of 4 8 g mL 1 ; (c) 250 L of the 4.8 g mL 1 armodafinil solution was diluted with 250 L of human plasma to yield a final concentration of 2.4 g mL 1 This last step was repeated 7 times in order to obtain armodafinil final concentrations of 1.2, 0.6, 0.3, 0.15, 0.075, 0.0375, 0.01875 g mL 1 QCs were obtained by three dilution steps: (a) 400 L of the 4.8 g mL 1 armodafinil solution was diluted wit h 80 L of human plasma to yield a final concentration of 4 g mL 1 ; (b) 96 L of the 4.0 g mL 1 armodafinil solution was diluted with 384 L of human plasma to yield a final concentration of 0.8 g mL 1 ; (c) 30 L of the 0 .8 g mL 1 armodafinil solution was diluted with 450 L of human plasma to yield a final concentration of 0. 05 g mL 1 A s tandard stock solution of prednisolone (IS) was obtained by dissolving 25 mg of the standard i n methanol acetonitrile (1:1) to a final volume of 25 mL to yield a fin al concentration of 0.2 g mL 1 Instrumentation Analyses were performed on a high performance liquid chromatography system (Perkin Elmer; Waltham, MA). Detection was performed on a triple quadrupole mass spectrometer (API4000; Applied Biosystems, Carlsbad CA) equipped with an electrospray ionization interface.

PAGE 34

34 Sample Preparation Prior to chromatographic analysis, p lasma samples (100 L each) were deproteinized with 400 L methanol:acetonitrile (1:1) containing 0.2 g mL 1 of IS, vortexed for 25 s and cent rifuged at 10,000 rpm for 10 minutes. The respective supernatant was separated and 5 L were injected into the system. IS final concentration into the samples was 0.16 g mL 1 Calibration curv e, QC, and experimental plasma samples were processed in the s ame manner. Chromatographic and Mass Spectrometer Conditions The mobile phase used for the chromatographic separation consisted of methanol water (90:10, v/v) containing 10 mM ammonium acetate buffer The apparent pH was adjusted to 4 with an aqueous solut ion of acetic acid. T he mobile phase was filtered and isocratically delivered at a flow rate of 1.0 mL min 1 (split ratio 3:4). The analysis was performed utilizing a reversed phase analytical column (Supelco C18, 5 m, 25 cm x 4.6 mm; Sigma Aldrich, St. L ouis, MO). Detection was performed on a triple quadrupole mass spectrometer (API4000; Applied Biosystems, Carlsbad, CA) equipped with an electrospray ionization interface operating in a positive mode (ESI + ) The spectrometer was programmed in the multiple reaction monitoring mode to allow for the transition of the precursor ion to its respective fragment. The decay of the mass to charge ratio (m/z) 274.2 precursor ion to the m/z 167.2 product ion for armodafinil and the decay of the m/z 403.3 precursor ion to the m/z 307.1 product ion for the IS were monitored.

PAGE 35

35 Validation Procedure Selectivity, linearity, intra and inter day precision, accuracy and lower limit of quantification were evaluated through standard calibration curves and quality control sets 54 E ach concentration level of the standard calibration curve and QC samples were analyzed in triplicate in three different days. Different stock solution s were used for each curve and QC samples set Prednisolone acetate was utilized as an IS to correc t for potential loss of the analyte during sample preparation and analysis; in additional it helped correct potential variability in the detection stage. The linearity of the c alibration curves based on the peak area ratio (armodafinil/IS) as a function of nominal concentration was investigated in the concentration range of 18.75 4800 ng mL 1 (weighted linear regression, concentration 1 ). Intra and inter day accuracy and precision of the bioanalytical method were determined by analyzing nine replicates ( th ree /per day during three days) of i ndividually prepared QC samples for each concentration level ( 0.05 0.8 and 4.0 g mL 1 ) Precision (%) was determined by calculating the relative standard deviation (RSD) of the experimental concentrations; accuracy (%) was calculated as the percentage of the experimental and the nominal sample concentration ratio The utilized criteria for acceptability of the validation procedure included accuracy from 85 to 115% and precision within 15%. In particular, the lower limit of quantification (LLOQ) should not exceed 20% of precision and not extrapolate the range of 80 120% for accuracy. The extraction efficiency was determined by comparing extracted calibration curve and QC samples with unextracted standards ( i.e., recovery o f 100%).

PAGE 36

36 Stability Studies Studies were carried out to evaluate a rmodafinil stability under the (potential) experimental conditions First, b ench top stability was determined in order to mimic plasma sample s handling at room temperature for 4 h ours Second s tability of the processed plasma samples was determined at room temperature for 6 h in the autosampler Third, armodafinil stability in human plasma was also determined after three freeze and thaw cycles. Stability studies were performed using three rep licates of the low and high concentration level of the QC samples. Behavioral Based Measures The subjects executed two neurobehavioral assays: a psychomotor vigilance task (PVT) and a go/no go association task (GNAT). Both tasks were based on published lit erature 55 57 In the 10 minute PVT, the subjects were requested to focus on a cross symbol located at the cente r of a computer screen and respond to the appearance of a visual intervals (Figure 2 2) The subjects were instructed to respond as quickly as possible by pressing a r esponse button on an external keyboard logic (EXKEY; BeriSoft Cooperation, Frankfurt, Germany). Responses with reaction time (RT) less than 100 milliseconds post stimulus were not regarded as valid ( i.e., false starts or errors of commission). The primary PVT output variables were: (a) the number of false starts (errors of commission) plus number of performance lapses (errors of omission: no response within 500 milliseconds post stimulus), and ( b ) the mean reciprocal reaction time ( 1/ RT) 57 In order to calculate the mean 1/RT, individua l RTs were divided by 1,000 before the reciprocal transformation was applied.

PAGE 37

37 In the 10 minute GNAT, the subjects were requested to respond to one form of visual stimuli and hold a response to another form. The visual stimulus consisted of four squares arranged as a (a) right slanted line, (b) left slanted line, (c) right slanted diamond, or (d) left slanted diamond (Figure 2 3) 56 One of the four stimuli was displayed on the screen after a random time delay. The subjects initiated each trial by pressing and holding a button on the keyboard. They were instructed to release the button (go condition) for line, and keep holding it (no go condition) for diamond disregarding orientation of the stimulus. Go and no go trials occurred in a random fashion with equal probability in every session. The output variables were: (a) the mean 1/RT for the go trials; (b) the number of failures to respond for the go trials (errors of omission); (c) the number of failures to inhibit a response for the no go trials (error s of commission); (d) the total number of errors (errors of omission plus errors of commission). The mean 1/RT was determined as aforementioned for the PVT task. All output variables were recorded by a dedicated computer (ERTS; BeriSoft Cooperation, Frankfurt, Germany) and then extracted for analysis (M ATLAB; MathWorks Inc, Natick, MA). E le ctroencephalogram Based Measures Scalp EEG recordings were acquired using a battery operated 128 channel amplifier with a 1024 Hz sampling rate and a 24 bit converter (Biosemi; Amsterdam, The Netherlands). 128 scalp el ectrodes were placed using a spandex electrode cap. Four electro oculogram (EOG) electrodes below and above the left eye (vertical EOG) and in the left and right side of the eyes (horizontal EOG) were attache d to control for eye artifacts.

PAGE 38

38 EEG recordin g was quantitatively analyzed by two different approaches: (1) the event related potential (ERP) analysis was applied to the EEG recording performed during the execution of the behavioral assays ( i.e., PVT and GNAT) ; (2) the EEG power spectral analysis was applied to the 2 minute EEG recording alone (Figure 2 1 ). In the ERP analysis, the data were pre processed and stimulus locked average was performed (BESA 5.3; BESA GmbH, Grfelfing, Germany). A low cutoff filter of 0.1 Hz (6 db/oct, forward) and a high c utoff filter of 40 Hz (24 db/oct, zero phase) were applied to the data. An adaptive artifact correction was utilized in order to correct for vertical EOG. The maximum amplitude value of accepted trials was set to 120 V. The data were epoched from 500 to 1000 ms with zero millisecond corresponding to the stimulus onset. Baseline was defined in the time window of 200 to 0 ms. The original recording was re referenced and interpolated to the 10 10 average virtual montage with 27 channels (Figure 2 4). The ou tput variables were amplitude, mean amplitude and area of the peaks. In the EEG power spectral analysis, fast Fourier transform was applied and power estimates were obtained (BESA 5.3; BESA GmbH, Grfelfing, Germany). The data were epoched in 4 second inte rvals (0.25 Hz of frequency resolution). The maximum amplitude value of accepted epochs was 120 V. Likewise, the original recording was re referenced and interpolated to the 10 10 average virtual montage with 27 channels (Figure 2 4). Power estimates were averaged within the following frequency ranges: (1 4 Hz), (4 8Hz), (8 12 Hz) and (12 25 Hz) (Figure 2 5 )

PAGE 39

39 Pharmacokinetic Pharmacodynamic Modeling The PK PD modeling was conducted using a non linear mixed effects model approach (NONMEM 7.2; ICON D evelopment Solutions, Ellicot City, MD). Based on preliminary studies 34 we started by utilizing a one compartment model with first order elimination and first o rder absorption to described the armodafinil pharmacokinetics Inclusion of a lag time was tested given the reported effect of meal in the rate of absorption 34 Using the outcome of the non linear mixed effects pharmacokinetic model, we performed a sequential fit of the PK PD data. Specifically, we fixed the population pharmacokinetic parameter estimates and respective variabilities including the individual pharm acokinetic data, and fit the pharmacodynamic data. The observed effect was defined by two components: time varying baseline and drug effect. Two options were tested: additive and proportional drug effect relative to baseline. The time varying baseline func tion was determined by fitting of the placebo group data. Different functions ( e.g., constant, linear, exponential, Weibull, inverse bateman function, polynomial, cosine) were tested in order to best describe the data over time. Likewise, several functions to describe the drug effect were tested ( e.g., linear, Emax, sigmoidal Emax, ) Diagnostics for time delay of the drug effect with respect to the concentration were performed. These included data based ( e.g., hysteresis plot) and PK PD mode l based ( e.g., basic goodness of fit and conditional weighted residuals, CWRES, vs. derivative of concentration, CDER) approaches. Overall model selection was guided by objective function values (NONMEM 7.2; ICON Development Solutions, Ellicot City, MD). In addition, visual predictive check and

PAGE 40

40 goodness of fit plots were generated using the Perl coded PsN Toolkit 58 and the R based package Xpose 59 Statistical Analysis Output variables among the three treatment groups were compared using a gene ralized linear mixed model (SAS 9.3; SAS Institute Inc., Cary, NC). Data that were not normally distributed underwent transformation ( i.e., log or reciprocal transformation). The fixed effects of treatment, time, fatigue baseline, prior treatment (carry ov er) and treatment*time interaction were investigated. Subject and visit were set as random effects. Different covariance structures were investigated to account for the repeated hoc multipl e comparisons.

PAGE 41

41 Figure 2 1. Schematic representation of the study design

PAGE 42

42 Figure 2 2 Schematic representation of the psychomoto r vigilance task The f igure shows a regular computer screen with the visual stimulus corresponding to a re.

PAGE 43

43 Figure 2 3 Schematic representation of the go/no go task The figure shows a regular computer screen with the visual stimulus consisting of four squares arranged as a right slanted line, a left slanted line, a right slanted diamond, or a left sla nted diamond Adapted from reference Zhang Y, Chen Y, Bressler SL, Ding M. Response preparation and inhibition: the role of the cortical sensorimotor beta rhythm. Neuroscience 2008;156:238 46 (Page 239, Figure 1) 56

PAGE 44

44 Figure 2 4 Schematic representation of the 1 0 10 average virtual montage with 27 channels The original recording was re referenced and interpolated using BESA 5.3 ( BESA GmbH, Grfelfing, Germany )

PAGE 45

45 Figure 2 5. Schematic representation of the electroencephalogram power spectral analysis. Preprocessing and analysis were performed in BESA 5.3 ( BESA GmbH, Grfelfing, Germany )

PAGE 46

46 CHAPTER 3 BIOANALYTICAL ANALYS IS Rationale Developing and validating an analytica l method to measure armodafinil concentrations in human plasma. Results Selectivity and Recovery The selectivity of the analytical method was demonstrated by comparing representative chromatograms: blank plasma, blank plasma spiked with armodafini l (18.75 ng mL 1 ) and processed with IS solution (160 ng mL 1 ); blank plasma spiked with armodafinil (4800 ng mL 1 ) and processed with IS solution (160 ng mL 1 ) (Figure 3 1). The extraction recovery of armodafinil was higher than 90%. Linearity and Lower L imit of Quantification Calibration curves of peak area ratio (armodafinil/IS) as a function of nominal concentration were linear (weighted linear regression, concentration 1 ) in the investigated concentration range (18.75 4800 ng mL 1 ). The lower limit of quantification (LLOQ) was 18.75 ng mL 1 corresponding to the lowest concentration that could be determined with precision of 20% and accuracy from 80 to 120% under the experimental condition s (Tables 3 1 and 3 2). Precision and Accuracy The intra and inte r day relative standard deviation (R.S.D.) values for armodafinil determination in human plasma are shown in Table 3 1. The intra and inter assay precision in the quantification of QC samples were less than or equal to 13.3 and

PAGE 47

47 6.6%, respectively. The acc uracy in their determination was from 84.9 to 110% (Table 3 2). Stability The accuracy and precision in the quantification of QC samples (low and high level of concentration) for bench top (four hours) and autosampler (six hours) are shown in Table 3 3. A fter the three freeze thaw cycles the precision in the quantification of QC samples were less than or equal to 8.6% ; t he accuracy in their determination was from 94.6 to 10 1 % (Table 3 4 ). A long term stability study on both R and S modafinil in human plas ma showed that both enantiomers were stable for at least 2 months at 20 C 60 Moreover, an even longer stability study showed that modafinil was stable for approxim ately 6 months at 20 C in human plasma 61

PAGE 48

48 Table 3 1. Intra and inter day variation of armodafinil determination in human plasma Nominal concentration (ng mL 1 ) Day Experimental c oncentrations a Mean (ng mL 1 ) S D RS D (%) Intra day variation 18.75 (LLOQ) 1 19.05 1.30 6.8 2 18.62 3.23 17.3 3 18.39 2.90 15.7 50 1 49 3 6.2 2 52 3 5.3 3 50 7 13.3 800 1 782 12 1.5 2 864 16 1.9 3 765 28 3.6 4000 1 4031 117 2.9 2 3769 61 1.6 3 3953 206 5.2 Inter day variat ion 18.75 (LLOQ) 18.69 0.33 1.8 50 50 2 3.6 800 804 53 6.6 4000 3918 134 3.4 a n=3 observations. SD, standard deviation; RSD, relative standard deviation ; LLOQ, lower limit of quantification Table 3 2. Intra and inter day variation of arm odafinil determination in human plasma Concentration (ng mL 1 ) Range (ng mL 1 ) Accuracy (%) a 18.75 (LLOQ) 15.06 21.56 80.3 115 50 42 55 84.9 110 800 733 880 91.6 110 4000 3700 4160 92.5 104 a n=9 observations. LLOQ, lower limit of quant ification. Table 3 3. Bench top and autosampler stability of armodafinil in human plasma Nominal concentration (ng mL 1 ) a Bench top stability Autosampler stability 50 4000 50 4000 Mean measured concentrations 46 4093 52 3812 Precision (RSD) 2.0 1.5 13 2.4 Accuracy (%) 92.4 102 103 95.3 a n=3 observations per concentration level in each condition.

PAGE 49

49 Table 3 4. Freeze thaw stability of armodafinil in human plasma Nominal concentration (ng mL 1 ) a Freeze thaw Cycle 1 Freeze thaw Cycle 2 Fr eeze thaw Cycle 3 50 4000 50 4000 50 4000 Mean measured concentrations 50 3787 47 3828 50 3809 Precision (RSD) 1.9 3.5 8.6 2.5 6.9 1.5 Accuracy (%) 101 94.7 94.6 95.7 100 95.2 a n=3 observations per concentration level in each cycle.

PAGE 50

50 F igure 3 1. Representative chromatograms of calibration curve in human plasma. Multiple reaction mode with positive electrospray ionization interface is shown using m/z ratios of 274.2 > 167.2 and 403.3 > 307.1 for armodafinil and internal standard (predni solone) respectively. Blank plasma A) MRM 274.2 > 167.2 B ) MRM 403.3 > 307.1. B lank plasma spiked with armodafinil (18.75 ng mL 1 ) and processed with IS solution (160 ng mL 1 ) C) MRM 274.2 > 167.2 D ) MRM 403.3 > 307.1. B lank plasma spiked with armodafinil (4800 ng mL 1 ) and processed with IS solution (160 ng mL 1 ) E ) MRM 274.2 > 167.2 F ) MRM 403.3 > 307.1. A B D F C E

PAGE 51

51 CHAPTER 4 BEHAVIORAL BASED MEASURES Rationale In an attempt to assess the armodafinil effect on the subjects, the present study comprised two different approaches: behavioral and EEG based measures. This chapter addresses the first approach by involving two neurocognitive tests (PVT and GNAT). PVT has been selected for three main reasons: i) it meets the criteria to an accurate assessment of neurobehavio ral degradation due to sleep loss ; 55 ii) it has been recently pointed out as the most widely used measure of behavioral alertness ; 57 and iii) it has been used as an efficacy endpoint in clinical trials of armodafinil 62 In this sense, a PK PD model involving a reliable measure of armodafinil effect on alertness will give insight about the drug effect and help to apply the knowledge to the development of compounds with an analogous mechanism of action. GNAT, in turn, brings foundation to the ERP analysis, since it has been previously applied to investigate the effect of response inhibition and production o n ERP studies 63 65 In additional, it is a well known measure of impulsive behavior being applied to study the effect of stimulant drugs such as methylphenidate in children with ADHD ; 66 67 it becomes interesting due to the fact that modafinil, the analogous compound of armodafinil, has showed to alleviate the symptoms of ADHD 68 Results Psychomotor Vigilance Task Armodafinil increased the PVT performance as measured by number of lapses plus false starts and mean reciprocal response time (1/RT) when compa red to placebo ( a rmodafinil 150 mg vs. placebo: lapses plus false starts, P = .0081 mean 1/RT, P <

PAGE 52

52 .0001 ; a rmodafinil 250 mg vs. placebo: lapses plus false starts, P = .0218 mean 1/RT, P = .0254 ) A significant treatment versus time interaction was obser ved in both PVT metrics (lapses plus false starts : P < .0001 ; mean 1/RT: P = .0012 ), indicating that the change in alertness over time varies with treatment. Fatigue baseline adjustment was performed in order to account for the significantly different base line values across the different treatments (lapses plus false alarms: P < .0001 ; mean 1/RT: P < .0001 ). Armodafinil produced a significant decrease of the number of lapses plus false starts from 1.6 to 11.6 hours post dose (Figure 4 1 ) and a significant i ncrease of the mean 1/RT from 2.6 to 9.6 hours post dose (Figure 4 2 ). The two doses of armodafinil were not significantly different from each other in respect to both PVT metrics (lapses plus false starts, P = .9644 mean 1/RT, P = .5095 ) Go/No Go Assoc iat ion Task Armodafinil increased the GNAT performance as measured by mean 1/RT of go trials, number of errors of omission (go trials) and errors of commission (no go trials) ( a rmodafinil 150 mg vs. placebo: mean 1/RT, P < .0001 errors of omission, P = .0 038, errors of commission, P < .0001 ; a rmodafinil 250 mg vs. placebo: mean 1/RT, P < .0001 errors of omission, P = .0048, errors of commission, P < .0001 ) A significant treatment versus time interaction was observed in all three GNAT metrics ( mean 1/RT: P = .0257 ; errors of omission: P < .0001 ; errors of commission: P < .0001 ). Fatigue baseline adjustment was performed for the three metrics given their significant difference across the treatments (1/RT: P = .0001 ; errors of omission: P < .0021 ; errors of commission: P < .0001 ). Armodafinil improved the mean 1/RT of go trials from 1.8 to 9.8 hours post dose ( Figure 4 3 ). In respect to accuracy, armodafinil significantly reduced errors of omission from 1.8 to 11.8 hours post dose (Figure 4 4 ); errors of

PAGE 53

53 comm ission were also significantly reduced from 3.8 to 9.8 hours post dose (Figure 4 5 ). The two doses of armodafinil were not significantly different from each other in respect to the three GNAT metrics (mean 1/RT, P = .9362 errors of omission, P = .9979, er rors of commission, P < .5417 ). Finally, a statistically significant correlation was observed between the mean 1/RT during GNAT and PVT (Figure 4 6)

PAGE 54

54 Figure 4 1. Psychomotor vigilance task numbe r of lapses plus false starts. A ) Mean observed data. B ) Least square mean fatigue baseline adjusted data. Simple effect comparisons of treatment*time by time. Specific time points represented by different letters are statistically significant different (Adj. Tukey p<0.05). rBL, resting baseline; fBL, fatigu e baseline. A B

PAGE 55

55 Figure 4 2. Psychomotor vigilance task mean reciprocal reaction time A ) Mean observed data. B ) Least square mean fatigue baseline adjusted data. Simple effect comparisons of treatment*time by time. Specific time points represented by different letters are statistically significant different (Adj. Tukey p<0.05). rBL, resting baseline; fBL, fatigue baseline. A B

PAGE 56

56 Figure 4 3. Go/no go association task m ean reciprocal reaction time. A ) Mean observed data. B ) Least square mean fatigue b aseline adjusted data. Simple effect comparisons of treatment*time by time. Specific time points represented by different letters are statistically significant different (Adj. Tukey p<0.05). rBL, resting baseline; fBL, fatigue baseline. A B

PAGE 57

57 Figure 4 4. Go/no go association task number of errors of omission A ) Mean observed data. B ) Least square mean fatigue baseline adjusted data. Simple effect comparisons of treatment*time by time. Specific time points represented by different letters are statistical ly significant different (Adj. Tukey p<0.05). rBL, resting baseline; fBL, fatigue baseline. A B

PAGE 58

58 Figure 4 5. Go/no go association task number of errors of com ission. A ) Mean observed data. B ) Least square mean fatigue baseline adjusted data. Simple eff ect comparisons of treatment*time by time. Specific time points represented by different letters are statistically significant different (Adj. Tukey p<0.05). rBL, resting baseline; fBL, fatigue baseline. A B

PAGE 59

59 Figure 4 6. een the psychomotor vigilance task PVT and go/no go association task GNAT. O bserved mean reciprocal reaction time after administration of placebo, 150 mg and 250 mg armodafinil.

PAGE 60

6 0 CHAPTER 5 ELECTRO ENCEPHALOGRAM BASED MEASURES Rationale We aim to propose EEG based approaches as a potential measure of armodafinil effect given their numerous advantages over behavioral measures. Differences in reaction time and error rates obtained from behavioral measures are not easily attributed to a particular cognitive process. ERP, a measure of brain activity due to a specific event, adds to our research in the sense that it yields a continuous monitoring of neural sub processes ( e.g., sensory, cognitive and motor processes) with a high time resolution 44 Another advantage of ERP is that it allows to measure brain activity even in the absence of behavioral response. EEG spectral power, in turn, offers the possibility of assessing alertness during a resting waking state. Indeed, it presents the advant age aforementioned advantages of an EEG based measure. Results Event Related Potentials ERP analyses were performed during execution of the PVT and GNAT. During the PVT an a rmodafinil effect was evident in the Cz channel where the amplitude of the positive ERP peak at around 380 ms was increased (Figure 5 1 ; a rmodafinil 150 mg vs. placebo: P = .0526 ; a rmodafinil 250 mg vs. placebo: P = .0001 ) No significant treatment versus time interaction ( P = .2827 ) and fatigue baseline difference ( P = .7199 ) were observed. The average amplitude over time for the three treatment groups is presented in the Figure 5 2. The two doses of armodafinil were not significantly different from each other in respect t o the ERP peak at around 380 ms ( P =

PAGE 61

61 .2006 ). S tatistically significant correlation s were observed between the average ERP amplitude and the both PVT metrics (Figure s 5 3 and 5 4 ). Likewise during the GNAT an armodafinil effect was pronounced in the Cz ch annel where the amplitude of the positive ERP peak at around 380 ms was increased ( Figure 5 5 ; a rmodafinil 150 mg vs. placebo: go condition P < .0001 no go condition P <.0001 ; a rmodafinil 250 mg vs. placebo: go condition P < .0001 no go condition P < .0001 ) No significant treatment versus time interaction was observed for the go ( P = .8377) and no go conditions ( P = .7242 ) Fatigue baseline adjustment was performed in order to account for the significantly different baseline values across the differen t treatments during the no go condition ( P < .0001 ) ; for consistency, it was kept in the model even when no significant difference was observed during the go condition ( P = .2625 ). The average amplitude over time for the three treatment groups during the g o and no go conditions is presented in the Figure s 5 6 and 5 7 respectively. The two doses of armodafinil were not significantly different from each other in respect to the ERP peak at around 380 ms for both conditions (go condition: P = .9943 ; no go cond ition: P = .8583 ). S tatistically significant correlation s were observed between the average ERP amplitude for both conditions and the GNAT metrics (Figure s 5 8 5 9, 5 10 ) as well as the PVT mean reciprocal reaction time ( Figures 5 1 1 ) Spectral Analysis A rmodafinil decreased the EEG power in the delta frequency range ( 1 4 Hz, delta power ) over the frontal, temporal and occipital region of the brain (Figure 5 1 2 ) when compared to placebo ( a rmodafinil 150 mg vs. placebo: F3, P = .0018 F4, P = .0008 T7, P < .0001 T8, P < .0001 O1, P = .0081 Oz: P = 0018 O2: P < 0001 ; a rmodafinil 2 50 mg vs. placebo: F3, P = .0348 F4, P = .0032 T7, P =.0040 T8, P = .0173 O1, P

PAGE 62

62 < .0001 Oz: P < 0001 O2: P < 0001 ) No significant treatment versus time interactio n was observed in any of the aforementioned region s ( F3: P = .9343; F4: P = .7949 ; T7: P = 0.1750 ; T8: P = .1426 ; O1: P = .7100 ; Oz: P = .7016 ; O2: P = .6342 ) Fatigue baseline adjustment was performed in order to account for the significantly different ba seline values across the different treatments ( T7 : P = 0216 ; O1: P < .0001 ; Oz: P < .0002 ; O2: P = .0026 ) ; for consistency, it was kept in the model even when no significant difference was observed ( F3: P = .6507; F4: P = .2744; T8: P = .6074 ) The averag e delta power over time for the three treatment groups in the F3, F4, T7, T8, O1, Oz and O2 electrodes are presented in the Figures 5 13 through 5 19 respectively. The two doses of armodafinil were not significantly different from each other in respect to delta power at any tested region ( F3: P = .5233; F4: P = .9200 ; T7: P = .5120 ; T8: P = .2841 ; O1: P = .2173 ; Oz: P = .3264 O2: P < .8441 ) A statistically significant correlation was observed between the delta power and the mean PVT 1/RT (Figur e 5 20 ).

PAGE 63

63 Figure 5 1. Grand average event related brain activity during the execution of th e psychomotor vigilance task. A ) The vertical bar (time=0) corresponds to the appearance of the visual stimulus (time window: 200 to 500 ms; positive pote ntials are pl otted downward). B ) Topographic map at approximately 6 hours after the treatment administration (380 ms after appearance of visual stimulus). To be included B A

PAGE 64

64 Figure 5 2 Average event related brain activity, ERP, during the execution of the psychomotor vigilance task PVT (Cz channel, positive peak at around 380 ms). A) Mean observed data. B ) Least square mean fatigue baseline adjusted data Simple effect comparisons of treatment*time by time. Specific time points represented by different letters are statistically significantly different (Adj. Tukey p<0.05). rBL, resting baseline; fBL, fatigue baseline. B A

PAGE 65

65 Figure 5 orrelation analysis between the event related brain activity, ERP, during the execution of the psychomotor vigilance task, PVT, and the PVT mean reciprocal reaction time. Observed data after administration of placebo, 150 mg and 250 mg armodafinil.

PAGE 66

66 Figure 5 related brain activity, ERP, during the execution of the psychomotor vigilance task, PVT, and the PVT number of lapses plus false starts. Observed data after administration of placebo, 150 mg and 250 mg armodafinil.

PAGE 67

67 Figure 5 5 Grand average event related brain activity during the execution of the go/no go association task ( no go condition) A ) The vertical bar (time=0) corresponds to the appearance of the visual stimulus (time window: 200 to 7 00 ms; positive potentials are plotted downward). B ) Topographic map at approximately 6 hours after the treatment administration (380 ms after appearance of visual stimulus). To be included B A

PAGE 68

68 Figure 5 6. Average event related brain activity, ERP, during the execution of the go/no go association task, GNAT ( go condition, Cz channel, pos itive peak at around 380 ms). A) Mean observed data. B ) Leas t square mean fatigue baseline adjusted data. Simple effect comparisons of treatment*time by time. Specific time points represented by different letters are statistically significantly different (Adj. Tukey p<0.05). rBL, resting baseline; fBL, fatigue base line. B A

PAGE 69

69 Figure 5 7. Average event related brain activity, ERP, during the execution of the go/no go association task, GNAT ( no go condition, Cz channel, pos itive peak at around 380 ms). A) Mean observed data. B ) Least square mean fatigue baseline ad justed data Simple effect comparisons of treatment*time by time. Specific time points represented by different letters are statistically significantly different (Adj. Tukey p<0.05). rBL, resting baseline; fBL, fatigue baseline. B A

PAGE 70

70 Figure 5 8. Pear related brain activity, ERP, during the go/no go association task, GNAT and the GNAT mean reciprocal reaction time Observed data for ERP during the go A ) and no go B ) condition after administration of placebo, 150 mg and 250 mg armodafinil. ( b ) (a) B A

PAGE 71

71 Figure 5 related brain activity, ERP, during the go/no go association task, GNAT, and the GNAT number of errors of omission Observed data for ERP during the no go con dition after administration of placebo, 150 mg and 250 mg armodafinil.

PAGE 72

72 Figure 5 related brain activity, ERP, during the go/no go association task, GNAT, and the GNAT number of errors of comission. Ob served data for ERP during the go condition after administration of placebo, 150 mg and 250 mg armodafinil.

PAGE 73

73 Figure 5 1 1 related brain activity, ERP, during the go/no go association task, GNAT and t he psychomotor vigilance task PVT mean reciprocal reaction time. Observed data for ERP during the go A) and no go B ) condition after administration of placebo, 150 mg and 250 mg armodafinil. A B

PAGE 74

74 Figure 5 12. F rontal (F) temporal (T) and occipital (O) r egion s of the brain shown in the 1 0 10 average virtual montage. Electrodes represented by filled blue circles evidenced a significant decrease in the EEG power in the delta frequency range in the armodafinil groups when compared to placebo.

PAGE 75

75 Figure 5 13. Electroencephalogram delta band spectral power in the electrode F3 of the frontal area. A ) Mean ( S.E.M.) observed data. B ) Least square mean fatigue baseline adjusted data. Simple effect comparisons of treatment*time by time. Specific time points represented by different letters are statistically significantly different (Adj. Tukey p<0.05). rBL, resting baseline; fBL, fatigue baseline. B A

PAGE 76

76 Figure 5 14. Electroencephalogram delta band spectral power in the elec trode F4 of the frontal area. A ) M ean ( S.E.M.) observed data. B ) Least square mean fatigue baseline adjusted data. Simple effect comparisons of treatment*time by time. Specific time points represented by different letters are statistically significantly different (Adj. Tukey p<0.05). rBL resting baseline; fBL, fatigue baseline. B A

PAGE 77

77 Figure 5 15. Electroencephalogram delta band spectral power in the elect rode T7 of the temporal area. A ) M ean ( S.E.M.) observed data. B ) Least square mean fatigue baseline adjusted data. Simple effect c omparisons of treatment*time by time. Specific time points represented by different letters are statistically significantly different (Adj. Tukey p<0.05). rBL, resting baseline; fBL, fatigue baseline. B A

PAGE 78

78 Figure 5 16. Electroencephalogram delta band s pectral power in the elect rode T8 of the temporal area. A ) M ean ( S.E.M.) observed data. B ) Least square mean fatigue baseline adjusted data. Simple effect comparisons of treatment*time by time. Specific time points represented by different letters are st atistically significantly different (Adj. Tukey p<0.05). rBL, resting baseline; fBL, fatigue baseline. B A

PAGE 79

79 Figure 5 17. Electroencephalogram delta band spectral power in the electr ode O1 of the occipital area. A ) M ean ( S.E.M.) observed data. B ) Leas t square mean fatigue baseline adjusted data. Simple effect comparisons of treatment*time by time. Specific time points represented by different letters are statistically significantly different (Adj. Tukey p<0.05). rBL, resting baseline; fBL, fatigue base line. B A

PAGE 80

80 Figure 5 18. Electroencephalogram delta band spectral power in the electr ode Oz of the occipital area. A ) M ean ( S.E.M.) observed data. B ) Least square mean fatigue baseline adjusted data. Simple effect comparisons of treatment*time by time Specific time points represented by different letters are statistically significantly different (Adj. Tukey p<0.05). rBL, resting baseline; fBL, fatigue baseline. B A

PAGE 81

81 Figure 5 19. Electroencephalogram delta band spectral power in the electr ode O2 of the occipital area. A ) M ean ( S.E.M.) observed data. B ) Least square mean fatigue baseline adjusted data. Simple effect comparisons of treatment*time by time. Specific time points represented by different letters are statistically significantly different (Adj. Tukey p<0.05). rBL, resting baseline; fBL, fatigue baseline. B A

PAGE 82

82 Figure 5 delta power the psychomotor vigilance task, PVT, mean reciprocal reaction time. Observed data after administration of placebo, 150 mg and 250 mg armodafinil.

PAGE 83

83 CHAPTER 6 PHARMACOKINETIC PHARMACODYNAMIC MODE LING Rationale We believe in the value of establishing a model to correlate armodafinil concentrations with neurocognitive and EEG based measures of drug effect. Furthermore we went further by investigating the clinical meaning of EEG parameters. In other words, we investigated the correlation between a well established measured of the armodafinil effect on alertness which is PVT, and EEG parameters bringing insight of the clin ical meaning of changes on EEG. Results Pharmacokinetics Armodafinil pharmacokinetics was determined in order to establish the pharmacokinetic pharmacodynamic (PK PD) relationship of the drug. The plasma concentration profile of armodafinil was best descri bed by a one compartment model with first order elimination and absorption with lag time as follows: ( 6 1 ) where C is the total plasma concentration over time t ; F is the oral bioavailability of the drug; k a is the first order ab sorption rate constant; k e is the first order elimination rate constant; Vd is the apparent volume of distribution; t lag is the delay between the dosing time and the appearance of concentration in the sampling compartment. The final parameter estimates and variability are presented in Table 6 1. The visual predictive check plots (Figure 6 1) together with the goodness of fit plots (Figure

PAGE 84

84 6 2) for armodafinil plasma concentrations demonstrated the conformity between the model predictions and the observed da ta. Pharmacokinetic P harmacodynamic Modeling In this study, be havioral and EEG based assessments were investigated as measures of drug effect For the purpose of the PK PD modeling we focused on a speci fic metric for each PD measure: 1) PVT, mean 1/RT ; 2) GNAT, mean 1/RT; 3) ERP during PVT, amplitude of the positive peak at around 380 ms in the Cz electrode; 4) ERP during GNAT, amplitude of the positive peak at around 380 ms in the Cz electrode; 5) Spectral analysis delta power of the occipital region of t he brain (O2 electrode). A delay in effect (counterclockwise h ysteresis) was observed for all PD measures suggesting disequilibrium between plasma and effect site concentration after oral administratio n of armodafinil ( Figure 6 3 ). An effect compartment ap proach was used to account for this delay: i n the one compartment pharmacokinetic model the effect compartment is linked to the plasma concentrations (central compartment) by the rate constant k 1e ; k e0 in turn, represents the rate constant for the drug e limination from the effect compartment (Figure 6 4) 69 The rate of change of the drug amount in the effect compartment can then be described as follow : 69 ( 6 2 ) where X e represents the hypothetical drug amount in the effect compartment, X represents the drug am ount in the central compartment. In the steady state, we have that or:

PAGE 85

85 ( 6 3 ) S ince and : ( 6 4 ) where C is the drug concentration in the centra l compartmental, V d is the volume of the central compartment C e is the drug concentration in the effect compartment and V e is the volume of the effect compartment. As in the steady state we can define: ( 6 5 ) Now, dividing Equation 6 2 by V e we have: ( 6 6 ) Behavioral based measures The observed effect was defined by two components: time varying baseline and drug effect The time varying baseline function was determined by fitting the data corresponding to the placebo group. Placebo model or baseline The following models describe d the placebo effect on the mean 1/RT for the PVT and GNAT, respectively: ( 6 7 ) ( 6 8 ) where S(t) is the placebo effect over time, S 0 is the mean baseline, is the amplitude, is the peak related parameter and t is time.

PAGE 86

86 Drug effect model For the groups receiving armodafinil treatment, the m ean 1/ RT increased as a function of the drug concentration in the effect compartment ( i.e., hypothetically the site of action) Therefore, an excitatory model was used to describe the drug effect on the mean 1/RT for both behavioral measures, PVT and GNAT : ( 6 9 ) ( 6 10 ) where E max is the maximum excitatory drug effect and EC e 50 is the apparent drug concentration at the effect site producing 50% of the E max The final parameter estimates and variability f or the placebo and drug effect model corresponding to both behavioral measures are presented in Table 6 2 and 6 3. The visual predictive check plots together with the goodness of fit plots for armodafinil effect on both behavioral measures demonstrated the conformity between the model predictions and the observed data (Figures 6 5 through 6 8). Electroencephalogram based m easures Likewise, the observed effect was defined by two components: time varying baseline and drug effect. The time varying baseline fun ction was determined by fitting the data corresponding to the placebo group. Placebo model or baseline The following models were used to describe the placebo effect on the amplitude of the positive peak at around 380 ms in the Cz electrode ( i.e., event re lated brain activity) during the execution of both behavioral measures, PVT and GNAT:

PAGE 87

87 ( 6 11 ) ( 6 12 ) To describe the placebo effect on the delta power of the occipital region (O2 electrode), the followin g equation was used: ( 6 13 ) Drug effect model For the groups receiving armodafinil treatment, the event related brain activity increased as a function of the drug concentration in the effect compartment ( i.e., hypothetically the site of action). Therefore, an excitatory model was used to describe the drug effect on the brain activity during the execution of both behavioral measures, PVT and GNAT: ( 6 14 ) ( 6 15 ) The final paramete r estimates and variability for the placebo and drug effect model corresponding to the brain activity during both behavioral measures are presented in Table 6 4 and 6 5. The visual predictive check plots together with the goodness of fit plots for armodafi nil effect on brain activity demonstrated the conformity between the model predictions and the observed data (Figures 6 9 through 6 12).

PAGE 88

88 To describe the drug effect on the delta power of the occipital region (O2 electrode), an inhibitory model was used giv en that armodafinil mitigated the increase in delta power due to the sleep loss : ( 6 16 ) where I max is the maximum inhibitory drug effect and IC 50 is the drug concentration producing 50% of the I max The final parameter estimates a nd variability for the placebo and drug effect model corresponding to the delta power is presented in Table 6 6. The visual predictive check plots together with the goodness of fit plots for armodafinil effect on delta power demonstrated the conformity bet ween the model predictions and the observed data (Figure 6 13 and 6 14 ). Correlation between behavioral alertness and electroencephalogram based measures The correlation between behavioral alertness and EEG based measures was investigated in order to bring insight into the clinical significance of the drug related changes on EEG. PVT represents the most widely used measure of behavioral alertness with the mean 1/RT being one of the primary metrics 57 A statistically significant correlation between the population predictions for the drug effect on the mean 1/RT of the PVT and EEG based measures (brain activity and delta power) is shown in Figure 6 15.

PAGE 89

89 Table 6 1 P ha rmacokinetic parameters of armodafinil Parameter Estimate (RSE%) BSV% (RSE%) CL/F (L h 1 ) 2.2 (7.7) 9.4 (87) V/F (L) 41 (9.8) 21 (27) k a (h 1 ) 1.3 (18) 27 (52) t lag (h) 1.1 (18) 54 (25) A proportional between subject variability (BSV) was used for all parameters. RSE%, estimate of standard error divided by parameter estimate; CL/F, total clearance divided by oral bioavailability; V/F, apparent volume of distribution divided by oral bioavailability; k a absorption rate constant; t lag time delay between drug dosing and its appearance in the sampling compartment. Table 6 2. Pharmacodynamic parameters of the armodafinil effect on the psychomotor vigilance task mean reciprocal reaction time. Parameter Estimate (RSE%) BSV% (RSE%) S 0 (ms 1 ) 3.6 (6.4) 14 ( 32) 0.066 (27) NE 6.9 (16) 17 (65) k e0 (h 1 ) 0.78 (41) NE E max (ms 1 ) 1.3 (8.2) NE EC e 50 ( g mL 1 ) 4.1 (33) 138 (58) A proportional between subject variability (BSV) was used in the case where it was estimated RSE%, estimate of standard error di vided by parameter estimate; S 0 me an baseline; amplitude; peak; k e0 elimination rate constant from the effect compartment; E max maximum excitatory drug effect ; EC e 50 apparent drug concentration at the effect site producing 50% of E max ; NE, not es timated. Table 6 3 Pharmacodynamic parameters of the armodafinil effect on the go/no go association task mean reciprocal reaction time. Parameter Estimate (RSE%) BSV% (RSE%) S 0 (ms 1 ) 2.2 (6.4) 14 (34) 0.044 (65) NE 4.7 (9.4) NE k e0 (h 1 ) 0.69 (38) NE E max (ms 1 ) 1.2 (21) NE EC e50 ( g mL 1 ) 7.6 (21) 123 (75) A proportional between subject variability (BSV) was used in the case where it was estimated. RSE%, estimate of standard error divided by parameter estimate; S 0 mean baseline; amplitu de; peak; k e0 elimination rate constant from the effect compartment; E max maximum excitatory drug effect; EC e50 apparent drug concentration at the effect site producing 50% of E max ; NE, not estimated.

PAGE 90

90 Table 6 4 Pharmacodynamic parameters of the ar modafinil effect on the event related brain activity during the psychomotor vigilance task. A proportional between subject error was used unless stated otherwise. Parameter Estimate (RSE%) BSV% (RSE%) S 0 ( V) 1.2 (36) 213 (22)* k e0 (h 1 ) 0.88 (28) 1.9 (3 4) E max ( V) 3.5 (9.6) NE EC e50 ( g mL 1 ) 5.3 (18) 143 (51) *Additive between subject variability (BSV) RSE%, estimate of standard error divided by parameter estimate; S 0 mean baseline; k e0 elimination rate constant from the effect compartment; E max maximum excitatory drug effect; EC e50 apparent drug concentration at the effect site producing 50% of E max ; NE, not estimated. Table 6 5 Pharmacodynamic parameters of the armodafinil effect on the event related brain activity during the go/no go asso ciation task (no go condition). A proportional between subject error was used unless stated otherwise. Parameter Estimate (RSE%) BSV% (RSE%) S 0 ( V) 5.7 (34) 494 (30)* 0.14 (25) NE 2.6 (24) NE k e0 (h 1 ) 1.1 (45) 88 (84) E max ( V) 8.0 (15) NE EC e5 0 ( g mL 1 ) 5.7 (19) 121 (70) *Additive between subject variability (BSV) RSE%, estimate of standard error divided by parameter estimate; S 0 mean baseline; amplitude; peak; k e0 elimination rate constant from the effect compartment; E max maximum excitatory drug effect; EC e50 apparent drug concentration at the effect site producing 50% of E max ; NE, not estimated. Table 6 6 Pharmacodynamic parameters of the armodafinil effect on the delta power of the occipital region of the brain. Parameter Es timate (RSE%) BSV% (RSE%) S 0 ( V 2 ) 14 (21) 44 (24) 0.23 (33) NE 7.0 (5.5) NE k e0 (h 1 ) 0.50 (38) NE I max ( V 2 ) 10 (24) NE I C e 50 ( g mL 1 ) 2.2 (12) NE A proportional between subject variability (BSV) was used in the case where it was estimated. R SE%, estimate of standard error divided by parameter estimate; S 0 mean baseline; amplitude; peak; k e0 elimination rate constant from the effect compartment; I max maximum inhibitory drug effect; IC e 50 apparent drug concentration at the effect site producing 50% of I max ; NE, not estimated.

PAGE 91

91 Figure 6 1. Visual predictive check of the final armodafinil pharmacokinetic model Observed data (blue circles) are shown with their median (red circles). The median (solid black line) and the 5 95th percent ile prediction interval (black broken lines) from the model are also presented.

PAGE 92

92 Figure 6 2. Basic goodness of fit plots of the final armodafinil pharmacokinetic model A) Population predictions. B) Individual predictions. C ) I ndividual weighted re siduals. D ) Conditional weighted residuals. A B C D

PAGE 93

93 Figure 6 3. Anticlockwise hysteresis loop between armodafinil pharmacokinetics and pharmacodynamics. The plot shows that the peak of armodafinil effect on the psychomotor vigilance task, PVT, reaction time i s delayed with respect to its peak plasma concentration (subject 6 after administration of armodafinil 250 mg).

PAGE 94

94 Figure 6 4. Schematic representation of the effect compartment approach. Adapted from Sheiner LB, Stanski DR, Vozeh S, Miller RD, Ham J. Si multaneous modeling of pharmacokinetics and pharmacodynamics: application to d tubocurarine. Clin Pharmacol Ther 1979;25:358 71 (Page 360, Figure 1). 69

PAGE 95

95 Figure 6 5. Visual predictive check of the final pharmacokinetic/pharmacodynamic model of the armodafinil effect on the psychomotor vigilance task mean reciprocal reaction time. Observed data (blue circles) are shown with their median (red circles). The median (solid black line) and the 5 95th percentile prediction interval (black broken lines) fr om the model are also presented.

PAGE 96

96 Figure 6 6. Basic goodness of fit plots of the final pharmacokinetic/pharmacodynamic model of the armodafinil effect on the psychomotor vigilance task mean reciprocal reaction time. A) Population predictions. B) In dividual predictions. C ) I ndividual weighted residuals. D ) Conditional weighted residuals. A B C D

PAGE 97

97 Figure 6 7. Visual predictive check of the final pharmacokinetic/pharmacodynamic model of the armodafinil effect on the go/no go association task mean reciproca l reaction time. Observed data (blue circles) are shown with their median (red circles). The median (solid black line) and the 5 95th percentile prediction interval (black broken lines) from the model are also presented.

PAGE 98

98 Figure 6 8. Basic goodness of fit plots of the final pharmacokinetic/pharmacodynamic model of the armodafinil effect on the go/no go association task mean reciprocal reaction time. A) Population predictions. B) Individual predictions. C ) I ndividual weighted residuals. D ) Conditiona l weighted residuals. A B C D

PAGE 99

99 Figure 6 9. Visual predictive check of the final pharmacokinetic/pharmacodynamic model of the armodafinil effect on the event related brain activity ERP, during the psychomotor vigilance task PVT Observed data (blue circles) a re shown with their median (red circles). The median (solid black line) and the 5 95th percentile prediction interval (black broken lines) from the model are also presented.

PAGE 100

100 Figure 6 10. Basic goodness of fit plots of the final pharmacokinetic/pha rmacodynamic model of the armodafinil effect on the event related brain activity during the psychomotor vigilance task. A) Population predictions. B) Individual predictions. C ) I ndividual weighted residuals. D ) Conditional weighted residuals. A B C D

PAGE 101

101 Figure 6 11. Visual predictive check of the final pharmacokinetic/pharmacodynamic model of the armodafinil effect on the event related brain activity, ERP, during the go/no go association task, GNAT (no go condition) Observed data (blue circles) are shown with th eir median (red circles). The median (solid black line) and the 5 95th percentile prediction interval (black broken lines) from the model are also presented.

PAGE 102

102 Figure 6 12. Basic goodness of fit plots of the final pharmacokinetic/pharmacodynamic mod el of the armodafinil effect on the event related brain activity during the go/no go association task (no go condition). A) Population predictions. B) Individual predictions. C ) I ndividual weighted residuals. D ) Conditional weighted residuals. A B C D

PAGE 103

103 Figure 6 13. Visual predictive check of the final pharmacokinetic/pharmacodynamic model of the armodafinil effect on the delta power of the occipital region of the brain. Observed data (blue circles) are shown with their median (red circles). The median (solid bl ack line) and the 5 95th percentile prediction interval (black broken lines) from the model are also presented.

PAGE 104

104 Figure 6 14. Basic goodness of fit plots of the final pharmacokinetic/pharmacodynamic model of the armodafinil effect on the delta powe r of the occipital region of the brain. A) Population predictions. B) Individual predictions. C ) I ndividual weighted residuals. D ) Conditional weighted residuals. A B C D

PAGE 105

105 Figure 6 15. Correlation between behavioral alertness and electroencephalogram bas ed measures. A ) Correlation between event related brain activity, ERP, during the psychomotor vigilance task, PVT, and PVT mean reciprocal reaction time (alertness) B ) Correlation between EEG delta power and PVT mean reciprocal reaction time (alertness) Blue circles are p opulation predictions after armodafinil administration. (a) (b ) B A

PAGE 106

106 CHAPTER 7 DISCUSSION AND CONCL USIONS To our knowledge, the present study is the first one correlating armodafinil (or its analogous compound modafinil) effect on quantitative EEG wi th a well established measure of alertness and drug concentrations through a PK PD modeling approach. We developed and compared PK PD models describing the armodafinil effect on EEG and behavioral based measures in sleep deprived healthy adults We observ ed that the two be havioral tasks ( i.e., PVT and GNAT), increased the event related brain activity and decreased the EEG power in the delta frequency range when compared to placebo; overall, th e drug groups started to differentiate from the placebo group after two hours post dose with the difference becoming less evident toward the last experimental time point ( i.e., approximately 12 hours post dose). All pharmacodynamics measures were correlated with the ap parent armodafinil concentrations at the site of action ( i.e., effect compartment approach) PVT in particular, is a widely used measure of alertness ; t herefore, establishing the correlation between a specific EEG parameter and alertness provide s insight of the clinical meaning of changes on EEG in the context of a wakefulness promoting drug Spectral A nalysis EEG has been systematically investigated on a few drug classes such as benzodiazepines 12 anesthetics and opioids 9 PK PD models using the EEG beta frequency band amplitudes as a measure of effect have been successfully established for diazepam, flunitrazepam, midazolam, clobazam, oxazepam, and bretazenil 70 Alfentanil, fentanyl, sufentanil and remifentanil have their analgesic effects reflected on the EEG spectral edge 49 71 In particular, a PK PD model utilizing EEG spectral edge as

PAGE 107

107 an effect measure identified the importance of taking into account age and lean body mass when determining the dose regimen of remifentanil 72 Systematic PK PD studies of the effect of CNS stimulant drugs on quantitative EEG along with a clinically relevant correlation still lacks in the literature. In a previous study quantitative EEG was evaluated as a potential pharmacodynamic measure of the CNS stimulant effect of dextroamphetamine in non sleep deprived healthy male subjects 73 Although the EEG alpha power was compared with alternative measures of CNS stimulation ( e.g., continuous performance, self rated mood scales and neuroendocrine tests), a well established measure of a lertness and assessment of drug profile is not identical to that of the amphetamine like stimulants 29 31 62 and any generalization among these different therapeutic classes would be premature. It becomes important to highlight that m odafinil effect on the EEG power s pectrum has been reported in the literature 11 14 74 75 but a systematic correlation of the drug related EEG changes with simultaneous plasma concentrations and alertness was not conducted In the present study, we observed that armodafinil mitigated the slowing of the brain acti vity due to sleep deprivation. In other words, it decreased the EEG power in the delta frequency range (1 4 Hz, i.e., EEG delta power) over the frontal, temporal and occipital region of the brain when compared to placebo. This effect was evident during the eyes open in opposed to the eyes closed EEG recording sessions Drug related effects on either theta, alpha or beta frequency ranges were not observed in both eyes open and closed conditions. Our findings are virtually consistent with a helicopter

PAGE 108

108 simula tion study where modafinil reduced EEG activity in the delta and theta but had no effect in the alpha and beta frequency ranges of sleep deprived healthy pilots 74 The absence of armodafinil effect on the theta activity in the pres ent study may be related to the different study design In our study a single dose of armodafinil 150 mg or 250 mg was administered after 24 hours of sleep deprivation; on the other hand, three repeated doses of modafinil 200 mg after 16, 20 and 24 hours o f sleep deprivation were administered to the sleep deprived healthy pilots 74 Another study in healthy subjects showed that modafinil 300 mg not only decreased the EEG power in the delta and theta frequencies but also increased the EEG power in the alpha frequency range 14 In that study, the subjects underwent 60 hours of sleep deprivation and received three doses of modafinil: the first do se aimed to prevent decrease in vigilance ( at 11:30 pm previously sleep deprivation) ; the second dose aimed to restore vigilance ( 5:30 am, after approximately 45 hours of sleep deprivation) and (3) investigate the effect of drug on recovery sleep ( 3:30 pm ten hours after administration of the second dose ) 14 Importantly, the authors noted that between the second and third dose administrations drug and placebo di d not differ much in respect to the EEG power in the alpha frequency range This period between the third and second dose would more closely represent our study design since the second drug dose was administered after a period of sleep deprivation in order to restore vigilance. Indeed, the authors suggested that when the alpha activity is largely reduced ( i.e., after a period of sleep deprivation), modafinil may not be effective enough to increase the EEG alpha power 14 In addition, the authors performed exclusively eyes closed EEG recording and stated that only the waking portion s of the EEG recordings were considered. In our study, we made efforts

PAGE 109

109 to include epoch s that well represent ed the respective recording sessions to prevent any potential selection bias; naturally, it was not always possible to avoid transition periods between wakefulness and sleep during eyes closed sessions. Psychomotor V igilance Task PVT, a measure of vigilant attention, meets the criteria to an accurate assessment of neurobehavioral degradation due to sleep loss 55 It has been recently pointed o ut as the most widely used measure of behavioral alertness 57 Indeed, d egradation in attention, particularly vigilant attention, is one of the most reliable effect s of sleep deprivation 76 77 PVT fulfills the requirements for an accurate measurement of the neurocognitive effec t of sleep loss in that : (a) it reflects an aspect of waking cognitive function ( e.g., ability to use attention) ; (b) it is easily and relatively quickly performed; (c) it has minor learning effects; (d) it is sensitive; (e ) its output variab les are of eas y interpretation; (f ) it has been suggest to reflect and drowsy related impaired driving 55 In the present study, the number of lapses plus false starts and mean 1/RT were utilized as primary PVT metrics. In a previous study, these both metrics showed the highest sensitivity in measuring sleep deprivation when considering both acute and chronic sleep deprivation. The mean 1/RT, in particular, scored very high effect sizes in measuring sleep loss after acute total sleep deprivation ( i.e., a 33 hour period of total sleep deprivation) and chronic sleep deprivation ( i.e., four hours of sleep per night during five consecutive nights) 55 The number of lapses combined with the number of false starts, in turn, presented a comparable high effect size in the condition of acute total sleep depriva tion but its effect size was slightly lower in the condition of chronic sleep deprivation 55

PAGE 110

110 The PVT has advantages over the maintenance of wakefulness test (MW T) and multiple sleep latency test (MSLT) which were used as primary efficacy endpoint in clinical trials for modafinil and armodafinil In the MWT, OSAHS and narcoleptic patient s w ere requested to stay a wake in a dark room, and four 20 minute sessions we re performed every 2 hours. The metric for each visit was the mean time for the patient to fall sleep over the four sessions. In the MSLT, SWSD patients basically underwent the same procedure, except that now the patients were requested not to resist falli ng sleep 62 Both MWT and MSLT are normally performed in a laboratory setting and, hence, provide artificial conditions that may stimulate sleep onset in children 78 and inhibit sleep onset in adults. Moreover, the PVT allows for a repeated use and, co nsequently concomitant assessment of alertness in a rich pharmacokinetic study. The result is a better description of the time course of effect over time, an important feature of an exploratory study ( i.e., biomarker investigation) such as ours A rmodafin il increased the PVT performance as measured by our primary metrics when compared to placebo. Both armodafinil doses produced a significant cumulative increase in the level of alertness during the 12 hour experimental period; subjects not only responded fa ster to the appearance of the stimulus but also improved their response accuracy ( i.e., fewer errors of commission and omission). A significant treatment versus time interaction was observed for both PVT metrics reflecting the change in alertness over time with treatment. We performed a fatigue baseline adjustment to account for the fact that the subjects did not respond to the sleep deprivation period in an identical manner previous each treatment visit. This adjustment along with a placebo controlled cros sover study helped us to eliminate potential

PAGE 111

111 confounder fa ctors PVT has been used as a secondary pharmacodynamic endpoint in a large phase one clinical trial of armodafinil in healthy subjects undergoing acute sleep deprivation 62 In that trial, the PVT was applied every 2 hours starting at approximately 1.5 hours post dose. The investigators started to evidence a significant armodafinil effect (as measured by the RT and lapses of performance) at the second time point what is consistent with our results. Likewise, modafinil improved the r eaction time of OSAHS 79 and SWSD 80 patients as well as sleep deprive d healthy subjects 81 Moreover, we observed that two doses of armodafinil were not significantly different from each other in respect to both PVT metrics. Consis tently, the FDA concluded that no consistent increase in armodafinil effect was observed between the different doses tested in OSAHS, narcolepsy and SWSD patient populations 62 The A gency further stated that the high dose of 250 mg once a day was accepted since there was no safety concern Go/No Go Association Task The GNAT is a well known measure of impulsive behavior 56 63 67 The capacity to in hibit an irrelevant response is a fundamental component of cognitive tasks 82 Deficit in response inhibition have been associate d with several conditions such as sleep deprivation and attention deficit and hyperactivity disorder (ADHD) 82 85 Indeed, a study comparing adult patients with n arcolepsy, idiopathic hypersomnia and ADHD demonstrated a high percentage of overlapping symptoms which may lead to misdiagnosing ; 86 in particular, the inatten tion score, as measured by the ADHD Rating Scale, was significantly correlated with excessive daytime sleepiness, as measured by the Epworth Sleepiness Scale. A systematic review suggested that children with ADHD had higher daytime sleepiness and more compared with controls 78 In

PAGE 112

112 another study including 2,463 children 6 15 years old sleep related problems ( e.g., dyssomnia, parasomnia, daytime inadvertent napping) were found to be correlated with ADHD like symptoms ( e.g., inattention, hyperactivity and impulsivity ) 87 Despite of the overlapping symptoms of sleep disorders and ADHD, further studies are required to establish their pathophysiological relationship. Armodafinil increased the GNAT performance as measured by mean 1/RT of go trials, number of errors of omission (go trials) and errors of commission (no go trials). S ubjects not only responded faster to the appearance of the go stimuli but also improved their response accuracy ( i.e., fewer errors of commission and omission). A significant treatment versus time interaction was observed for all GNAT metrics reflecting the important change in performance with tr eatment. A fatigue baseline adjustment was also performed to account for the fact that the subjects did not respond to the sleep deprivation period in an identical manner previous each treatment visit. Likewise evidenced in the PVT, we observed that two do ses of armodafinil were not significantly different from each other in respect to GNAT metrics. Response inhibition tasks have been applied to study the effect of stimulant drugs in children with ADHD 66 67 88 M ethylphenidate a classical stimulant drug was able to improve the GNAT performance by reducing th e number of failures to inhibit a response in the no go trials (errors of commission) 67 In anot her response inhibit ion task ADHD patients demonstrated a greater ability to inhibit response when compared to placebo 88 Indeed, several studies proposed modafinil a s an effective therapy in the treatment of ADHD Children and adults with ADHD demonstrated reduced inattention, hyperactivity and impulsivity when treated with modafinil 68 88 92

PAGE 113

113 GNAT also brings foundation to the ERP analysis, since it has been previously applied to investigate the effect of response inhibition and production on ERP studies 63 65 This topic will be further discussed in the next section. Event Related Brain Activity ERP is an increase in the brain activity due to a specific event which can be an internal or external stimulus 44 In the present study, the event corresponds to the external visual stimuli presented during the execution of the neurocognitive assay s (PVT and GNAT). A nalyse s of the ERP during execution of the PVT and GNAT have been performed. Armodafinil effect has been more evident in the Cz channel where the amplitude of the positive ERP peak at around 380 ms was increased when compared to the placebo group. Overall, no significant treatment versus time interaction and fatigue baseline difference were observed; the exception was seen in the no go condition of the GNAT where significantly different fatigue baseline values of ERP amplitude across the different treatment s were observed. The overall non significance of fatigue baseline values of ERP amplitude suggests that the effect of sleep deprivation on this me a sure has a lower within subject in respect to behavioral measures where fatigue baseline was always significa ntly different Importantly, we conducted selective averaging in that only co rrect response trials were averaged using a s timulus locked average approach; hence, error related interference was prevented The GNAT has been applied to study the effect of res ponse inhibition and production on ERP with most of the studies using visual stimuli 63 65 By using visual letter and symbol stimuli, it has been found a P3 with m aximum at Pz in go trials, and similar amplitude at Cz and Pz in no go trials 93 A response inhibition task was used to investigate the effect of methylphenidate on inhibitory control in children with attention

PAGE 114

114 deficit/hyperactivity disorder (ADHD) 66 The no go P3 mean amplitude was increased by methylphenidate when compared to placebo This results is consistent with our findings and with the fact that modafinil showed to improve the symptoms of ADHD in children and adults 68 88 92 In an ERP study using a visual odd ball paradigm in patients with narcolepsy, the latency of the positive peak aro und 300 ms (P3) was able to differentiate between responders and non responders to modafinil therapy 94 Pharmacokinetics The observed monophasic decline of the ar modafinil concentrations is consistent with a pooled pharmacokinetic analysis from three randomized studies on armodafinil 34 The estimates of armodafinil cleara nce and volume of distribution are similar to the mean values reported in the literature (2.32 L/h and 42.4 L, respectively) 34 The delayed appearance of armodaf inil in the systemic circulation ( i.e., lag time) can be attributed to the fact that the subjects received the study drug after breakfast; indeed, it has been reported that food affects the rate of absorption of armodafinil 34 Considering armodafinil neutral character (negative logarithm of the acid dissociation constant pK a 19.25) 95 this decrease in the rate of armodafinil absorption with food could be attributed to the altered g astric emptying rate instead of a pH dependent dissolution. It is coherent with the fact that the dissolution profiles of armodafinil tablets were basically superimposed in media of different pH ( i.e., 2.0, 6.4 and 7.4); in those, a complete release of the active content was evidenced in approximately 15 minutes 62 Pharmacokinetic Pharmacodynamic Modeling In the present study behavioral and EEG based measures of armodafinil effect were correlated with the apparent drug concentrations at the site of action by an effect compartment approach The rate constant for the drug elimination from the hypothetical

PAGE 115

115 effect compartment, k e0 characterizes the temporal component of the equilibration between drug concentration and effect 69 Based on the established PK PD models, the population estim ates for k e0 ranged from 0.5 to 1.1 h 1 implying an average half time of equilibration between armodafinil plasma concentration and effect of around 1 hour. The limitation of the effect compartment approach though, resides in its inability to identify whi ch of the possible factors would be the rate limiting step in the equilibration process between drug concentration and effect ; namely, tissue perfusion, drug diffusion from capillaries to the tissue tissue :blood partition coefficient for the drug, or post receptor events ; 69 in this case, we are referring to the tissue where the site of action belongs to. At least two mechanisms could be proposed to justify the disequilibrium between armodafinil plasma concentration and effect First, a rmodafinil was demonstrated to be a substrate of the P glycoprotein through a permeability in vitro assay (Madin Darby canine kidney cells transfected with the human MDR1 gene, MDR MDCK) 62 The P glycoprotein, a well known efflux transporter, plays an important role in limiting the cellular uptake of d rugs not only from the intestinal lumen into the epithelial cells, but also from the systemic circulation into the brain 96 Interestingly the MDR MDCK permeability assay has been proposed a s a blood brain barrier permeability model 97 98 Second, despite of the fact that armodafinil mechanism of action is not well understood, modafinil studies have suggested that modafinil modu lates the hypocretin system by causing activation of hypocretin secreting neurons, which, in turn, stimulate glutaminergic and histaminergic systems leading to arousal 29 32 Therefore, the hypothesis that such a sequence of events would constitute the rate limiting step for armodafinil onset of effect cannot be ruled out.

PAGE 116

116 To our knowledge, there is only one PK PD study of armodafinil/modafinil in the literature. That study was conducted by armodafinil sponsor (Cephalon ) and consisted of a pooled analysis of two randomized, double bind placebo controlled, multiple dose trials in 463 SWSD patients treated for 12 weeks with modafinil or armodafinil 99 The PD measure, however, was a 20 minute MSLT which corresponded to the primary clinical endpoint in the SWSD clinical trial s Under the assumption that modafinil and armodafinil are equipotent, the authors estimated an EC 50 of 4.6 g mL 1 for both d rugs. Interestingly, i n the present study, we estimated a similar EC e50 of 4.1 g mL 1 using PVT as our healthy subjects receiving a single dose of armodafinil This observation could help substantiate (a) the utility of our model of total acute sleep deprivation in mimicking a chronic condition; (b) the sensitivity of the PVT to sleep loss; (c) our single dose study as a good predict or of a multiple dose condition. In this study, we developed PK PD mod els for five different pharmacodynamic measures. Based on their expected clinical relevance, we can categorize them in (a) alertness and (b) impulsive behavior related models. The first category includes the models for PVT ( i.e., a behavioral measure), ev ent related brain activity during PVT, and EEG spectral analyses ( i.e., the EEG based measures). The second category includes the models for GNAT ( i.e., a behavioral measure) and event related brain activity during GNAT ( i.e., the EEG based measures). Ther e is no clear trend in the EC e50 values when comparing behavioral and EEG based measures. Focusing on the alertness related models, the EC e50 estimates were relatively close between the model for PVT and brain activity ( EC e50 4.1 vs. 5.3 g mL 1 respectiv ely ) in opposed to the one

PAGE 117

117 for spectral analysis (EC e50 2.2 g mL 1 ). Focusing on the impulsive behavior related models, the EC e50 estimates was higher in the model for GNAT than brain activity (7.6 vs. 5.7 g mL 1 respectively). In this scenario where we have used an empirical model to describe armoda finil effect, we would like to raise a hypothesis to explain these difference s in the EC e50 estimates. The concept of a translation mechanism between drug receptor interaction ( i.e., occupancy) and drug respo nse, the so called stimulus response relationship, has been originally proposed to explain why two agonists with similar a receptor occupancy profile could lead to significantly different responses 100 101 This conce pt was later applied to characterize the stimulus response relationship (or transducer function) for GABA A receptor agonists 70 102 In this context, we could not rule out the presence of different tran sducer functions translating the stimulus produced by the drug receptor interaction to the different measure of resp onse. Study Limitations The present study has some limitations that need to be considered. F irst, we conducted the study in a relatively small sample size. We made efforts to increase the the power by designing a placebo controlled cross over study Second, our subjects underwent a total of 36 hours sleep deprivation in the context of a single do se study. In respect to the EEG spectral analysis, it could explain the absence of a significant armodafinil effect in the theta frequency range, a phenomenon observed in modafinil studies where the subjects underwent longer sleep deprivation periods and m ultiple armodafinil administrations Third, it was not always possible to avoid transition periods between wakefulness and sleep during the eyes closed sessions; however, we sought to include epochs that well represented each recording session. Finally, th e effect

PAGE 118

118 compartment approach consists of an empirical model and does not allow us to identify which of the possible factors is the rate limiting step in the equilibration process between drug concentration and effect. Nevertheless, we believe that the use of a mechanistic model to describe the armodafinil effect is still premature given that its mechanism of action is not well understood. Study Strengths Besides the innovative and significant aspects of this study, we can highlight some other following str engths. First, we conducted a placebo controlled cross over study which increased the power and allowed us to separate the placebo effect or time varying baseline. Second, a rich data collection was performed with simultaneous assessment of pharmacokinetic s and pharmacodynamics yielding a more accurate description of armodafinil effect over time Third, EEG metrics were correlated not only with drug concentrations but also with behavioral measures. T he latter including a well es tablished measure of alertne ss, suggested the clinical relevance of EEG as a measure of alertness. Finally, the similarity between the EC 50 estimates for alertness (PVT and MSLT models, respectively) obtained in our study and in a pooled analysis of two phase 3 clinical trials studie s on modafinil/armodafinil could advocate for the clinical utility of our model : (a) our model of total acute sleep deprivation s eemed to mimic a chronic condition reasonably well ; (b) t he single dose study seemed to be a reasonable predictor of a multiple dose condition. Conclusions A rmodafinil mitigated the neurobehavioral de gradation due to sleep loss. The correlation between armodafinil related changes on EEG and alertness suggests EEG as a potential biomarker of armodafinil effect. Ultimately, it may g uide dosage selection

PAGE 119

119 in phase 3 clinical trials pursuing new indications for armodafinil and expedite the development of novel compounds with an analogous mechanism of action.

PAGE 120

120 LIST OF REFERENCES 1. Derendorf H, Meibohm B. Modeling of pharmacokinetic/pharmacodynamic (PK/PD) relationships: Concepts and perspectives. Pharmaceutical Research 1999;16:176 85. 2. Derendorf H, Lesko LJ, Chaikin P, et al. Pharmacokinetic/pharmacodynamic modeling in drug research and development. Journal of Cli nical Pharmacology 2000;40:1399 418. 3. Dingemanse J, Danhof M, Breimer DD. Pharmacokinetic Pharmacodynamic Modeling of Cns Drug Effects an Overview. Pharmacology & Therapeutics 1988;38:1 52. 4. Keizer RJ, Schellens JH, Beijnen JH, Huitema AD. Pharmacody namic biomarkers in model based drug development in oncology. Curr Clin Pharmacol 2011;6:30 40. 5. Norris H. Action of Sedatives on Brain Stem Oculomotor Systems in Man. Neuropharmacology 1971;10:181 &. 6. Vansteveninck AL, Schoemaker HC, Pieters MSM, Kroo n R, Breimer DD, Cohen AF. A Comparison of the Sensitivities of Adaptive Tracking, Eye Movement Analysis, and Visual Analog Lines to the Effects of Incremental Doses of Temazepam in Healthy Volunteers. Clinical Pharmacology & Therapeutics 1991;50:172 80. 7 dos Santos FM, Goncalves JCS, Caminha R, et al. Pharmacokinetic/Pharmacodynamic Modeling of Psychomotor Impairment Induced by Oral Clonazepam in Healthy Volunteers. Therapeutic Drug Monitoring 2009;31:566 74. 8. Mathias S, Zihl J, Steiger A, Lancel M. Ef fect of repeated gaboxadol administration on night sleep and next day performance in healthy elderly subjects. Neuropsychopharmacology 2005;30:833 41. 9. Mandema JW, Danhof M. Electroencephalogram Effect Measures and Relationships between Pharmacokinetics and Pharmacodynamics of Centrally Acting Drugs. Clinical Pharmacokinetics 1992;23:191 215. 10. Danhof M. Electroencephalography parameters as biomarkers: Extrapolation from laboratory animals to humans. Methods and Findings in Experimental and Clinical Pha rmacology 2002;24:63 4. 11. Saletu M, Anderer P, Saletu Zyhlarz G, et al. EEG tomographic studies with LORETA on vigilance differences between narcolepsy patients and controls and subsequent double blind, placebo controlled studies with modafinil. J Neurol 2004;251:1354 63.

PAGE 121

121 12. Laurijssens BE, Greenblatt DJ. Pharmacokinetic pharmacodynamic relationships for benzodiazepines. Clin Pharmacokinet 1996;30:52 76. 13. Garnock Jones KP, Dhillon S, Scott LJ. Armodafinil. Cns Drugs 2009;23:793 803. 14. Chapotot F, Pi geau R, Canini F, Bourdon L, Buguet A. Distinctive effects of modafinil and d amphetamine on the homeostatic and circadian modulation of the human waking EEG. Psychopharmacology 2003;166:127 38. 15. Littner MR, Kushida C, Wise M, et al. Practice Parameters for clinical use of the multiple sleep latency test and the maintenance of wakefulness test An American Academy of Sleep Medicine Report Standards of practice committee of the American Academy of Sleep Medicine. Sleep 2005;28:113 21. 16. Schwartz JR, Roth T, Hirshkowitz M, Wright KP. Recognition and management of excessive sleepiness in the primary care setting. Prim Care Companion J Clin Psychiatry 2009;11:197 204. 17. Roth T, Roehrs TA. Etiologies and sequelae of excessive daytime sleepiness. Clin Th er 1996;18:562 76. 18. Lyznicki JM, Doege TC, Davis RM, Williams MA. Sleepiness, driving, and motor vehicle crashes. Council on Scientific Affairs, American Medical Association. JAMA 1998;279:1908 13. 19. Mitler MM, Carskadon MA, Czeisler CA, Dement WC, Di nges DF, Graeber RC. Catastrophes, sleep, and public policy: consensus report. Sleep 1988;11:100 9. 20. Division of Sleep Medicine at Harvard Medical School. Sleep, Performance, and Public Safety. WGBH Educational Foundation, 2007. (Accessed 09/17, 2011, a t http://healthysleep.med.harvard.edu/healthy/matters/consequences/sleep performance and public safety ) 21. Nishino S, Mignot E. Drug treatm ent of patients with insomnia and excessive daytime sleepiness Pharmacokinetic considerations. Clinical Pharmacokinetics 1999;37:305 30. 22. Berman SM, Kuczenski R, McCracken JT, London ED. Potential adverse effects of amphetamine treatment on brain and behavior: a review. Mol Psychiatry 2009;14:123 42. 23. Kumar R. Approved and investigational uses of modafinil : an evidence based review. Drugs 2008;68:1803 39.

PAGE 122

122 24. Darwish M, Kirby M, Hellriegel ET, Robertson P. Armodafinil and Modafinil Have Substantial ly Different Pharmacokinetic Profiles Despite Having the Same Terminal Half Lives Analysis of Data from Three Randomized, Single Dose, Pharmacokinetic Studies. Clin Drug Invest 2009;29:613 23. 25. Darwish M, Kirby M, Hellriegel ET. Comparison of steady sta te plasma concentrations of armodafinil and modafinil late in the day following morning administration: post hoc analysis of two randomized, double blind, placebo controlled, multiple dose studies in healthy male subjects. Clin Drug Investig 2009;29:601 12 26. Hirshkowitz M, Lankford A, Roth T, Yang R, Rippon GA. Adjunctive Armodafinil Improves Wakefulness Throughout the Day in Cpap Treated Patients with Excessive Sleepiness Associated with Obstructive Sleep Apnea. Sleep 2009;32:A194 A5. 27. Rosenberg R, H arsh JR, Walsh JK, Yang RH, Rippon GA, Roth T. Armodafinil Sustains Wakefulness throughout the Day in Patients with Excessive Sleepiness Associated with Narcolepsy. Neurology 2009;72:A430 A. 28. Cephalon Inc Nuvigil (armodafinil) Tablets: U.S. Labeling In formation. 2010. (Accessed 09/19, 2011, at http://www.accessdata.fda.gov/drugsatfda_docs/label/2010/021875s005s008s01 5lbl.pdf ) 29. Ballon JS, Feifel D. A syst ematic review of modafinil: Potential clinical uses and mechanisms of action. J Clin Psychiat 2006;67:554 66. 30. Engber TM, Koury EJ, Dennis SA, Miller MS, Contreras PC, Bhat RV. Differential patterns of regional c Fos induction in the rat brain by amphet amine and the novel wakefulness promoting agent modafinil. Neuroscience letters 1998;241:95 8. 31. Ferraro L, Antonelli T, O'Connor WT, Tanganelli S, Rambert FA, Fuxe K. Modafinil: an antinarcoleptic drug with a different neurochemical profile to d ampheta mine and dopamine uptake blockers. Biol Psychiatry 1997;42:1181 3. 32. Scammell TE, Estabrooke IV, McCarthy MT, et al. Hypothalamic arousal regions are activated during modafinil induced wakefulness. The Journal of neuroscience : the official journal of th e Society for Neuroscience 2000;20:8620 8. 33. Sakurai T. The neural circuit of orexin (hypocretin): maintaining sleep and wakefulness. Nature reviews Neuroscience 2007;8:171 81. 34. Darwish M, Kirby M, Hellriegel ET, Yang R, Robertson P, Jr. Pharmacokinet ic profile of armodafinil in healthy subjects: pooled analysis of data from three randomized studies. Clin Drug Investig 2009;29:87 100.

PAGE 123

123 35. Darwish M, Kirby M, Robertson P, Jr., Hellriegel ET. Interaction profile of armodafinil with medications metabolize d by cytochrome P450 enzymes 1A2, 3A4 and 2C19 in healthy subjects. Clin Pharmacokinet 2008;47:61 74. 36. IMS. Institute for Healthcare Informatics. The Use of Medicines in the United States: Review of 2010. In. Parsippany, NJ; 2011:1 36. 37. Colburn WA, L ee JW. Biomarkers, validation and pharmacokinetic pharmacodynamic modelling. Clin Pharmacokinet 2003;42:997 1022. 38. Stolerman IP, ed. Encyclopedia of psychopharmacology. Berlin: Springer; 2010. 39. Herrmann WM, Irrgang U. An Absolute Must in Clinico Phar macological Research Pharmaco Electroencephalography, Its Possibilities and Limitations. Pharmacopsychiatria 1983;16:134 42. 40. Isley MR, Krauss G, Levin K, Litt B, Shields R, Wilbourn AJ. Electromyography/electroencephalography Redmond, Washington: Spa ceLabs Medical, Inc.; 1993. 41. Fisch BJ. EEG Primer: Basic Principles of Digital and Analog EEG. 3rd ed. Kidlington, Oxford, UK: Elsevier; 1999. 42. Nierdermeyer E, Silva FL. Electroencephalography. 4th ed. Baltimore: Williams & Wilkins; 1999. 43. Buhrer M, Maitre PO, Hung O, Stanski DR. Electroencephalographic effects of benzodiazepines. I. Choosing an electroencephalographic parameter to measure the effect of midazolam on the central nervous system. Clin Pharmacol Ther 1990;48:544 54. 44. Luck SJ. An int roduction to the event related potential technique. Cambridge: Massachusetts Institute of Technology; 2005. 45. Bewernitz M, Derendorf H. Electroencephalogram based pharmacodynamic measures: a review. Int J Clin Pharm Th 2012;50:162 84. 46. Duncan CC, Barr y RJ, Connolly JF, et al. Event related potentials in clinical research: guidelines for eliciting, recording, and quantifying mismatch negativity, P300, and N400. Clin Neurophysiol 2009;120:1883 908. 47. Barbanoj MJ, Riba J, Morte A, Antonijoan RM, Jane F. Basics of PK PD using QEEG: Acute/repetitive administration, interactions. Focus on anxiolytics with different neurochemical mechanisms as examples. Methods and Findings in Experimental and Clinical Pharmacology 2002;24:67 83.

PAGE 124

124 48. Mandema JW, Kuck MT, Dan hof M. Differences in Intrinsic Efficacy of Benzodiazepines Are Reflected in Their Concentration Eeg Effect Relationship. Brit J Pharmacol 1992;105:164 70. 49. Minto CF, Schnider TW, Egan TD, et al. Influence of age and gender on the pharmacokinetics and p harmacodynamics of remifentanil. I. Model development. Anesthesiology 1997;86:10 23. 50. Minto CF, Schnider TW, Shafer SL. Pharmacokinetics and pharmacodynamics of remifentanil. II. Model application. Anesthesiology 1997;86:24 33. 51. Glass PSA, Hardman D, Kamiyama Y, et al. Preliminary Pharmacokinetics and Pharmacodynamics of an Ultra Short Acting Opioid Remifentanil (Gi87084b). Anesth Analg 1993;77:1031 40. 52. Tuk B, Oberye JJL, Pieters MSM, et al. Pharmacodynamics of temazepam in primary insomnia: Ass essment of the value of quantitative electroencephalography and saccadic eye movements in predicting improvement of sleep. Clinical Pharmacology & Therapeutics 1997;62:444 52. 53. Luck, S. J. What is an ERP. 2012. (Accessed 10/17, 2012, at http://erpinfo.org/what is an erp ) 54. U.S. Food and Drug and Administration.Center for Drug Evaluation and Research. Bioanalytical Method Validation. 2001. (Accessed 01/17, 2010, at http://www.fda.gov/downloads/Drugs/GuidanceComplianceRegulatoryInformation /Guidances/ucm070107.pdf ) 55. Dorrian J, Rogers NL, Dinges DF. Psychomotor Vigilance Performance: Neurocogni tive Assay Sensitive to Sleep Loss. In: Kushida C, ed. Sleep Deprivation: Clinical Issues, Pharmacology and Sleep Loss Effects. New York: Marcel Dekker; 2005:39 70. 56. Zhang Y, Chen Y, Bressler SL, Ding M. Response preparation and inhibition: the role of the cortical sensorimotor beta rhythm. Neuroscience 2008;156:238 46. 57. Basner M, Dinges DF. Maximizing Sensitivity of the Psychomotor Vigilance Test (PVT) to Sleep Loss. Sleep 2011;34:581 91. 58. Lindbom L, Pihlgren P, Jonsson N. PsN Toolkit A collecti on of computer intensive statistical methods for non linear mixed effect modeling using NONMEM. Comput Meth Prog Bio 2005;79:241 57. 59. Jonsson EN, Karlsson MO. Xpose an S PLUS based population pharmacokinetic/pharmacodynamic model building aid for NONM EM. Comput Meth Prog Bio 1999;58:51 64.

PAGE 125

125 60. Cass QB, Ferreira Galatti T. A method for determination of the plasma levels of modafinil enantiomers, (+/ ) modafinic acid and modafinil sulphone by direct human plasma injection and bidimensional achiral chiral chromatography. Journal of pharmaceutical and biomedical analysis 2008;46:937 44. 61. Gorman SH. Determination of the D and L enantiomers of modafinil in human plasma utilizing liquid liquid extraction and high performance liquid chromatography. Journal of chromatography B, Biomedical sciences and applications 1999;730:1 7. 62. Nuvigil (armodafinil) Tablets. U.S. Food and Drug and Administration Drug Approval Package. 2007. (Accessed 09/19, 2011, at http://www.accessdata.fda.gov/drugsatfda_docs/nda/2007/021875_nuvigil_toc.cf m ) 63. Kiefer M, Marzinzik F, Weisbrod M, Scherg M, Spitzer M. The time course of brain activations during response inhibition: evidence from event related potentials in a go no go task. Neuroreport 1998;9:765 70. 64. Falkenstein M, Hoormann J, Hohnsbein J. ERP components in Go Nogo tasks and their relation to inhibition. Acta Psychol 1999;101:267 91. 65. Bokura H, Yamaguchi S, Kobayashi S. Elec trophysiological correlates for response inhibition in a Go/NoGo task. Clin Neurophysiol 2001;112:2224 32. 66. Pliszka SR, Liotti M, Bailey BY, Perez R, 3rd, Glahn D, Semrud Clikeman M. Electrophysiological effects of stimulant treatment on inhibitory cont rol in children with attention deficit/hyperactivity disorder. Journal of child and adolescent psychopharmacology 2007;17:356 66. 67. Trommer BL, Hoeppner JAB, Zecker SG. The Go No Go Test in Attention Deficit Disorder Is Sensitive to Methylphenidate. J Ch ild Neurol 1991;6:S128 S31. 68. Biederman J, Pliszka SR. Modafinil improves symptoms of attention deficit/hyperactivity disorder across subtypes in children and adolescents. J Pediatr 2008;152:394 9. 69. Sheiner LB, Stanski DR, Vozeh S, Miller RD, Ham J. S imultaneous modeling of pharmacokinetics and pharmacodynamics: application to d tubocurarine. Clin Pharmacol Ther 1979;25:358 71. 70. Visser SA, Wolters FL, Gubbens Stibbe JM, et al. Mechanism based pharmacokinetic/pharmacodynamic modeling of the electroen cephalogram effects of GABAA receptor modulators: in vitro in vivo correlations. The Journal of pharmacology and experimental therapeutics 2003;304:88 101.

PAGE 126

126 71. Scott JC, Cooke JE, Stanski DR. Electroencephalographic quantitation of opioid effect: comparati ve pharmacodynamics of fentanyl and sufentanil. Anesthesiology 1991;74:34 42. 72. Minto CF, Schnider TW, Egan TD, et al. Influence of age and gender on the pharmacokinetics and pharmacodynamics of remifentanil .1. Model development. Anesthesiology 1997;86: 10 23. 73. Slattum PW, Venitz J, Barr WH. Comparison of methods for the assessment of central nervous system stimulant response after dextroamphetamine administration to healthy male volunteers. J Clin Pharmacol 1996;36:1039 50. 74. Caldwell JA, Caldwell J L, Smythe NK, Hall KK. A double blind, placebo controlled investigation of the efficacy of modafinil for sustaining the alertness and performance of aviators: a helicopter simulator study. Psychopharmacology 2000;150:272 82. 75. Saletu M, Anderer P, Semlit sch HV, et al. Low resolution brain electromagnetic tomography (LORETA) identifies brain regions linked to psychometric performance under modafinil in narcolepsy. Psychiat Res Neuroim 2007;154:69 84. 76. Goel N, Rao H, Durmer JS, Dinges DF. Neurocognitive consequences of sleep deprivation. Seminars in neurology 2009;29:320 39. 77. Lim J, Dinges DF. Sleep deprivation and vigilant attention. Annals of the New York Academy of Sciences 2008;1129:305 22. 78. Cortese S, Konofal E, Yateman N, Mouren MC, Lecendreux M. Sleep and alertness in children with attention deficit/hyperactivity disorder: a systematic review of the literature. Sleep 2006;29:504 11. 79. Dinges DF, Weaver TE. Effects of modafinil on sustained attention performance and quality of life in OSA pat ients with residual sleepiness while being treated with nCPAP. Sleep Med 2003;4:393 402. 80. Czeisler CA, Walsh JK, Roth T, et al. Modafinil for excessive sleepiness associated with shift work sleep disorder. New Engl J Med 2005;353:476 86. 81. Killgore WD S, Kahn Greene ET, Grugle NL, Killgore DB, Balkin TJ. Sustaining Executive Functions During Sleep Deprivation: A Comparison of Caffeine, Dextroamphetamine, and Modafinil. Sleep 2009;32:205 16. 82. Garavan H, Ross TJ, Stein EA. Right hemispheric dominance o f inhibitory control: an event related functional MRI study. Proc Natl Acad Sci U S A 1999;96:8301 6.

PAGE 127

127 83. Mander BA, Reid KJ, Baron KG, et al. EEG measures index neural and cognitive recovery from sleep deprivation. The Journal of neuroscience : the offici al journal of the Society for Neuroscience 2010;30:2686 93. 84. Harrison Y, Horne JA, Rothwell A. Prefrontal neuropsychological effects of sleep deprivation in young adults -a model for healthy aging? Sleep 2000;23:1067 73. 85. Chuah YM, Venkatraman V, Din ges DF, Chee MW. The neural basis of interindividual variability in inhibitory efficiency after sleep deprivation. The Journal of neuroscience : the official journal of the Society for Neuroscience 2006;26:7156 62. 86. Oosterloo M, Lammers GJ, Overeem S, d e Noord I, Kooij JJS. Possible confusion between primary hypersomnia and adult attention deficit/hyperactivity disorder. Psychiat Res 2006;143:293 7. 87. Gau SSF. Prevalence of sleep problems and their association with inattention/hyperactivity among child ren aged 6 15 in Taiwan. Journal of sleep research 2006;15:403 14. 88. Turner DC, Clark L, Dowson J, Robbins TW, Sahakian BJ. Modafinil improves cognition and response inhibition in adult attention deficit/hyperactivity disorder. Biol Psychiatry 2004;55:10 31 40. 89. Rugino TA, Copley TC. Effects of modafinil in children with attention deficit/hyperactivity disorder: an open label study. Journal of the American Academy of Child and Adolescent Psychiatry 2001;40:230 5. 90. Rugino TA, Samsock TC. Modafinil in children with attention deficit hyperactivity disorder. Pediatric neurology 2003;29:136 42. 91. Swanson JM, Greenhill LL, Lopez FA, et al. Modafinil film coated tablets in children and adolescents with attention deficit/hyperactivity disorder: results of a randomized, double blind, placebo controlled, fixed dose study followed by abrupt discontinuation. J Clin Psychiatry 2006;67:137 47. 92. Taylor FB, Russo J. Efficacy of modafinil compared to dextroamphetamine for the treatment of attention deficit hyperac tivity disorder in adults. Journal of child and adolescent psychopharmacology 2000;10:311 20. 93. Pfefferbaum A, Ford JM, Weller BJ, Kopell BS. ERPs to response production and inhibition. Electroencephalography and clinical neurophysiology 1985;60:423 34. 94. Sangal RB, Sangal JM, Belisle C. Visual P300 latency predicts treatment response to modafinil in patients with narcolepsy. Clin Neurophysiol 1999;110:1041 7.

PAGE 128

128 95. Wang W, Xiang SY, Zhou XJ, Ji YB, Xiang BR. Enantiomeric Separation and Determination of t he Enantiomeric Impurity of Armodafinil by Capillary Electrophoresis with Sulfobutyl Ether beta cyclodextrin as Chiral Selector. Molecules 2012;17:303 14. 96. Lin JH, Yamazaki M. Role of P glycoprotein in pharmacokinetics: clinical implications. Clin Pharm acokinet 2003;42:59 98. 97. Gumbleton M, Audus KL. Progress and limitations in the use of in vitro cell cultures to serve as a permeability screen for the blood brain barrier. J Pharm Sci Us 2001;90:1681 98. 98. Wang Q, Rager JD, Weinstein K, et al. Evalua tion of the MDR MDCK cell line as a permeability screen for the blood brain barrier. Int J Pharmaceut 2005;288:349 59. 99. Darwish M, Bond M, Ezzet F. Armodafinil and modafinil in patients with excessive sleepiness associated with shift work disorder: a ph armacokinetic/pharmacodynamic model for predicting and comparing their concentration effect relationships. J Clin Pharmacol 2012;52:1328 42. 100. Kenakin T. Pharmacologic Analysis of Drug Receptor Interaction. 3rd ed. Philadelphia: Philadelphia; 1997. 101. Kenakin T. Molecular Pharmacology. Cambridge: Blackwell Science; 1997. 102. Visser SA, Smulders CJ, Reijers BP, Van der Graaf PH, Peletier LA, Danhof M. Mechanism based pharmacokinetic pharmacodynamic modeling of concentration dependent hysteresis and bip hasic electroencephalogram effects of alphaxalone in rats. The Journal of pharmacology and experimental therapeutics 2002;302:1158 67.

PAGE 129

129 BIOGRAPHICAL SKETCH Daniela Joice Conrado was born in Vacaria ( Rio Grande do Sul Brazil). She received her h igh s cho ol diploma in 1997 focusing on a nalytical c hemistry. In 1999, she got into the College of Pharmacy Federal University of Rio Grande do Sul (Porto Alegre, R io Grande do Sul, Brazil) receiving her Bachelor in Science degree in 2003. Her final undergraduate p roject was a scientific review on the issues involving pharmacokinetic of antimicrobial agents in burn patients. In 2004, she joined a Master of Science program in the same aforementioned u Preclinical Pharmacokineti c Evaluation of LASSBio 579: An N phenylpiperazine Antipsychotic Prototype In 2006, she received a Master of Science degree as well as a diploma of s pecialist in i ndustrial p harmacy by the same u niversity. Then, she worked as a lecturer in the Federal Un iversity of Rio Grande do Sul (2006 2007) and Integrated Regional University of Alto Uruguai and Misses ( Erechim, Rio Grande do Su l Brazil; 2007 2008); t he main courses were p harmacokinetics, p harmacodynamics and p harmaceutical p ractice and c osmetology. In 2009, Daniela joined Professor Hartmut research group at the Department of Pharmaceutics, College of Pharmacy University of Florida (Gainesville, F lorida ) She focused on quantitative electroencephalography as a means to assess a wakefulness promoting effect.