model-based dose selection for phase iii rivaroxaban study ... · regular article model-based dose...

Regular Article

Model-based Dose Selection for Phase III Rivaroxaban Studyin Japanese Patients with Non-valvular Atrial Fibrillation

Takahiko TANIGAWA1,*, Masato KANEKO1, Kensei HASHIZUME1, Mariko KAJIKAWA2,Hitoshi UEDA3, Masahiro TAJIRI2, John F. PAOLINI4 and Wolfgang MUECK5

1Development Clinical Pharmacology Asia, Bayer Yakuhin Ltd., Osaka, Japan2Clinical Development Cardiovascular, Bayer Yakuhin Ltd., Osaka, Japan

3Medical Scientific Liaison, Bayer Yakuhin Ltd., Osaka, Japan4Bayer HealthCare Pharmaceuticals, Montville, NJ, USA

5Development Clinical Pharmacology, Bayer HealthCare AG, Wuppertal, Germany

Full text of this paper is available at http://www.jstage.jst.go.jp/browse/dmpk

Summary: The global ROCKET AF phase III trial evaluated rivaroxaban 20mg once daily (o.d.) for strokeprevention in atrial fibrillation (AF). Based on rivaroxaban pharmacokinetics in Japanese subjects andlower anticoagulation preferences in Japan, particularly in elderly patients, the optimal dose regimen forJapanese AF patients was considered. The aim of this analysis was dose selection for Japanese patientsfrom a pharmacokinetic aspect by comparison of simulated exposure in Japanese patients with those inCaucasian patients. As a result of population pharmacokinetics-pharmacodynamics analyses, a one-com-partment pharmacokinetic model with first-order absorption and direct link pharmacokinetic-pharmaco-dynamic models optimally described the plasma concentration and pharmacodynamic models (Factor Xaactivity, prothrombin time, activated partial thromboplastin time, and HepTest), which were also consistentwith previous works. Steady-state simulations indicated 15mg rivaroxaban o.d. doses in Japanese patientswith AF would yield exposures comparable to the 20mg o.d. dose in Caucasian patients with AF. Inconclusion, in the context of the lower anticoagulation targets in Japanese practice, the populationpharmacokinetic and pharmacodynamic modeling supports 15mg o.d. as the principal rivaroxaban dose inJ-ROCKET AF.

Keywords: anticoagulants; pharmacokinetics; atrial fibrillation; Factor Xa; Japan; rivaroxaban; stroke

Introduction

Atrial fibrillation (AF) is the most common sustained cardiacarrhythmia and it is predicted that AF will affect over 1 millionJapanese adults by 2030.1) AF is an independent risk factor forstroke that carries a 2.8- and 3.7-fold excess risk for Japanesewomen and men, respectively.2) The risk of stroke in patientswith AF increases with age, 1.5-fold per decade.3) Because Japanhas one of the most rapidly ageing populations in the world,4,5)

the economic burden imposed on Japanese healthcare systems bystroke is expected to become a critical issue.6) As AF is responsiblefor approximately 15% of all strokes,7) reducing AF-related strokewill clearly mitigate a significant portion of this burden.

Vitamin K antagonists (VKAs), such as warfarin, have beenthe mainstay of oral anticoagulant therapy.8) Japanese guidelines,like others,7,9,10) advocate the use of warfarin for stroke preven-tion in patients with AF and a moderate to high risk of stroke.11)

These guidelines reflect the benefit–risk analyses of Yasaka andothers,12,13) in which elderly patients aged over 70 years wererecommended to receive warfarin therapy to a lower targetinternational normalized ratio (INR) level. Consequently, theguidelines recommend that warfarin be dose adjusted to an INRrange of 2.0–3.0 in patients aged under 70 years and to an INRrange of 1.6–2.6 for patients aged 70 years or older.11) Registry datasuggest that Japanese clinicians indeed seek lower INR targets inpractice, and that lower targets are sought irrespective of age.14)

Warfarin use is complicated by multiple issues, including avariable dose response and numerous drug and food interactions,necessitating inconvenient regular clinic visits for INR monitoringand dose adjustment.15) For these reasons, and despite guidelinerecommendations for its use in patients with AF at a moderateto high risk of stroke,7,9–11) warfarin is often under-prescribed toeligible patients.16,17) Given the considerable limitations of VKAtherapy, a number of new oral anticoagulants that eliminate many

Received April 3, 2012; Accepted July 3, 2012J-STAGE Advance Published Date: July 17, 2012, doi:10.2133/dmpk.DMPK-12-RG-034*To whom correspondence should be addressed: Takahiko TANIGAWA, Ph.D., Bayer Yakuhin, Ltd., 2-4-9, Umeda, Kita-ku, Osaka 530-0001,Japan. Tel. ©81-6-6133-6392, Fax. ©81-6-6344-2259, E-mail: [email protected]

Drug Metab. Pharmacokinet. 28 (1): 59–70 (2013). Copyright © 2013 by the Japanese Society for the Study of Xenobiotics (JSSX)

59

http://www.jstage.jst.go.jp/browse/dmpk

http://dx.doi.org/10.2133/dmpk.DMPK-12-RG-034

of the drawbacks associated with VKAs15) have been developed.Many of these are either in,18,19) or have completed,20–23) phase IIItrials for stroke prevention in patients with AF.

Rivaroxaban is an oral, direct Factor Xa inhibitor currently indevelopment for the prevention of stroke in AF. Rivaroxaban has adual mode of elimination: approximately one-third of the drug iseliminated unchanged via the kidneys; two-thirds of the drugundergoes metabolic degradation in the liver, half of which isexcreted via the kidneys and half via the hepatobiliary route.24)

In previous phase I dose-ranging clinical studies, rivaroxaban hasbeen found to have predictable pharmacokinetics and pharmaco-dynamics.25,26) Maximum inhibition of Factor Xa activity andmaximum prolongation of prothrombin time (PT), activatedpartial thromboplastin time (aPTT) and HepTest, which has beendeveloped to monitor anticoagulant effect of heparin and especiallylow molecular weight heparins, occurred within 1–4 h of admin-istration of rivaroxaban. No accumulation of drug was foundbeyond steady state in a multiple-dose study.26)

A population pharmacokinetic and pharmacodynamic evalua-tion based on data obtained from phase IIb studies in patientsundergoing major orthopaedic surgery given rivaroxaban oncedaily (o.d.) or twice daily (b.i.d.) reiterated the predictablepharmacokinetics of rivaroxaban.27,28) It was previously determinedin phase II dose-ranging studies of rivaroxaban for deep veinthrombosis (DVT) treatment, conducted primarily in Caucasians,that 20mg o.d. was a suitable dose for the prevention of recurrenceof DVT29,30); this dataset informed the choice of a 20mg o.d. dosefor evaluation for stroke prevention in AF in the global ROCKETAF trial.

The global ROCKET AF trial tested a 20mg dose ofrivaroxaban (reduced to 15mg in patients with creatinine clearance[CrCl] 30–49mlmin¹1), in 45 countries, including countries inAsia other than Japan.23) A lower dose of rivaroxaban was exploredas possibly being more appropriate for use in Japan to accom-modate: (1) phase I studies with rivaroxaban in healthy Japanesesubjects (data on file), which showed a 20–40% increase in areaunder the plasma concentration–time curve (AUC) and a 20–30%increase in maximum plasma concentration (Cmax) compared withphase I studies in healthy Caucasian subjects26); (2) the uniqueaspects of Japanese clinical practice, i.e. the lower intensity ofwarfarin anticoagulation recommended in Japanese guidelines forolder patients; (3) the relatively lower body mass indices of theJapanese population compared with populations of other coun-tries,31) a demographic known to be associated with higher drugexposures (previously described for ximelagatran32,33)).

Although the 20mg rivaroxaban dose was well tolerated in theJapanese phase I trials, phase II studies with Japanese AF patientsconducted prior to the phase III study in Japanese AF patients wereperformed to see if a lower dose of rivaroxaban would be moresuitable to test in the Japanese phase III J-ROCKET AF trial inlight of expected higher exposure and the lower anticoagulationpreferences in Japan. These phase II studies focused on thepharmacokinetics-pharmacodynamics (PK-PD) and safety ofrivaroxaban in Japanese AF patients.

The study aimed to select an optimal dose regimen for JapaneseAF patients in terms of pharmacokinetics employing the result of acomparison of simulated steady-state area under the curve (AUCss)and maximum concentration (Cmax,ss) in Japanese patients withAF given rivaroxaban as 20mg or 15mg o.d. regimens (using apopulation pharmacokinetic model from the Japanese AF data inphase II trials) with a previously described virtual population ofCaucasian patients with AF given rivaroxaban 20mg o.d. (using apopulation pharmacokinetic model from Caucasian patients beingtreated for DVT).34) This study also evaluated the inter- andintra-individual variability in the derived pharmacokinetic andpharmacodynamic parameters of rivaroxaban.

The phase II studies used for this analysis are the first studieswith AF patients in order to investigate the pharmacokinetics,pharmacodynamic effects and safety of rivaroxaban in Japanesesubjects with atrial fibrillation, and this study is the first popula-tion PK-PD evaluation of rivaroxaban in patients with AF, whoseresults could be used for appropriate dose setting in the Japanesephase III trial.

Methods

Study design and patient selection: Three Japanese phase IItrials provided the data for this analysis. The number of subjectswith pharmacokinetic and pharmacodynamic data in the threestudies (denoted A, B and C) were A (36 subjects), B (72subjects) and C (74 subjects). These were all open-label, dose-ranging studies that assessed the safety, pharmacokinetics andpharmacodynamics of rivaroxaban in patients with non-valvularAF treated for 28 days. In these trials, the following rivaroxabandoses and dosing regimens were investigated: 2.5, 5, 10 and20mg (all b.i.d. [i.e. total daily dose of 5, 10, 20 and 40mg;studies A and B]); and 10, 15 and 20mg (all o.d. [study C]).The number of subjects, data availability and gender ratios foreach dosing regimen from the three trials are listed in Table 1.All three trials had similar durations, blood sampling times forpharmacokinetic and pharmacodynamic analyses, and inclusion

Table 1. Summary of clinical studies

Total daily dose(mg)

Study RegimenNo of subjects

(% male)No of pharmacokinetic

data pointsNo of PTdata points

No of aPTTdata points

No of Factor Xadata points

No of HepTestdata points

5.0 B 2.5mg b.i.d. 23 (69.6%) 102 125 125 125 125

10.0 B 5.0mg b.i.d. 26 (76.9%) 114 140 140 140 140

C 10.0mg o.d. 26 (80.8%) 124 150 150 149 150

15.0 C 15.0mg o.d. 25 (72.0%) 114 139 139 139 139

20.0 A 10.0mg b.i.d. 25 (96.0%) 128 128 127 128 128

B 10.0mg b.i.d. 23 (82.6%) 106 128 128 128 128

C 20.0mg o.d. 23 (87.0%) 109 132 132 132 132

40.0 A 20.0mg b.i.d. 11 (90.9%) 45 45 45 44 45

Total 182 (81.3%) 842 987 986 985 987

aPTT, activated partial thromboplastin time; b.i.d., twice daily; o.d., once daily; PT, prothrombin time.No of pharmacokinetic/pharmacodynamics data points shows the number of data points higher than LLOQ (Lower limit of quantification).

Takahiko TANIGAWA, et al.60

Copyright © 2013 by the Japanese Society for the Study of Xenobiotics (JSSX)

and exclusion criteria. Japanese male patients aged ²20 years andpostmenopausal female patients with electrocardiographicallydocumented non-valvular AF (¯30 days before randomization)were enrolled. Patients had one or more of the following riskfactors for thromboembolism: hypertension, diabetes mellitus,

coronary artery disease, heart failure or age ²60 years; thus mostpatients had a CHADS2 score of 1 or above, which was the scorefor risk stratification for stroke prevention in patients with atrialfibrillation,35) although patients with a CHADS2 score of 0 wereincluded in the three phase II studies (Table 2). Reasons forexclusion included a history or presence of stroke or transientischaemic attack, or planned cardioversion or active endocarditis.Active internal bleeding, or a prior or active condition associatedwith increased bleeding (e.g., thrombocytopenia or persistent,poorly controlled hypertension); treatment with anticoagulanttherapy within 1 week of randomization; liver cirrhosis/impairedliver function (transaminases or total bilirubin >2 © upper limitof normal [ULN]); or impaired renal function (serum creatinine>1.5 © ULN) were also exclusion criteria. Study participantdemographics are listed in Table 3. The mean age of the subjectswas 65.6 years (range: 30–92 years) and mean body weight was67.2 kg (range: 45–103 kg).

All subjects in these studies provided written, informed con-sent, and these studies were performed in accordance with theDeclaration of Helsinki and Good Clinical Practice guidelines.

Sampling and determination of plasma rivaroxaban con-centrations: Blood samples for pharmacokinetic and pharmaco-dynamic analyses were taken simultaneously except at baseline atthe time points described in Figure 1. Actual sampling time was

Table 2. Patient numbers from the safety analyses of the three phase IIstudies

Study CHADS2 scorea2.5mgb.i.d.

5mgb.i.d.

10mgo.d.

15mgo.d.

10mgb.i.d.

20mgo.d.

20mgb.i.d.

A

²2 — — — — 2 — 6

0 or 1+ additional risk factor

— — — — 9 — 4

0 — — — — 14 — 1

B

²2 8 12 — — 7 — —


16 14 — — 16 — —

0 0 0 — — 1 — —

C

²2 — — 12 8 — 6 —


— — 14 17 — 17 —

0 — — 0 0 — 1 —

aAdditional risk factors: age 65–74 years, female sex, coronary artery disease,cardiac myopathy, hyperthyroidism.b.i.d., twice daily; o.d., once daily.

Table 3. Patient demographics in this evaluation

Study A Study B Study C Overall

Mean (SD) (range) Mean (SD) (range) Mean (SD) (range) Mean (SD) (range)

Age (years) 59.3 (11.0) (34–81) 68.1 (10.1) (30–92) 66.2 (8.0) (45–85) 65.6 (10.0) (30–92)

Body weight (kg) 68.9 (8.5) (46–86) 66 (12.0) (45–103) 67.5 (9.4) (47–90) 67.2 (10.4) (45–103)

Height (cm) 167.9 (7.1) (153–183) 162.2 (8.2) (140–182) 164.7 (7.2) (145–178) 164.3 (7.8) (140–183)

Gender (male) 94.4% (n = 34/36) 76.4% (n = 55/72) 79.7% (n = 59/74) 81.3% (n = 148/182)

Body surface area (m2)a 1.8 (0.1) (1.5–2.0) 1.7 (0.2) (1.4–2.2) 1.7 (0.1) (1.4–2.1) 1.7 (0.2) (1.4–2.2)

Lean body mass (kg)b 53.7 (5.5) (38.7–64.4) 49.9 (7.9) (35.2–67.9) 51.6 (6.9) (34.7–65.1) 51.4 (7.2) (34.7–67.9)

Body fat (kg)c 15.2 (4.3) (5.6–25.7) 16 (5.9) (4.8–38.2) 15.9 (4.2) (8.9–26.5) 15.8 (4.9) (4.8–38.2)

Rivaroxaban dose (mg) 13.1 (4.7) (10–20) 5.8 (3.1) (2.5–10) 14.9 (4.1) (10–20) 10.9 (5.7) (2.5–20)

Values obtained at the time of screening were used

Albumin (g dl¹1) 4.5 (0.3) (3.9–5.1) 4.2 (0.3) (3.4–4.9) 4.3 (0.3) (3.4–4.9) 4.3 (0.3) (3.4–5.1)

Serum creatinine (mg dl¹1) 1.0 (0.2) (0.7–1.6) 0.9 (0.2) (0.4–1.5) 0.9 (0.2) (0.5–1.3) 0.9 (0.2) (0.4–1.6)

BUN (mgdl¹1) 16.6 (4.0) (9.0–25.0) 17.8 (4.9) (9.0–36.0) 15.7 (4.2) (7.0–29.0) 16.7 (4.6) (7.0–36.0)

SGOT (AST) (U l¹1) 28.0 (11.0) (13.0–55.0) 27.5 (9.4) (7.0–59.0) 30 (11.6) (12.0–73.0) 28.6 (10.7) (7.0–73.0)

SGPT (ALT) (U l¹1) 27.3 (13.2) (11.0–67.0) 24.1 (12.3) (11.0–82.0) 27.2 (14.4) (9.0–73.0) 26 (13.4) (9.0–82.0)

CrCl (mlmin¹1)d 84.2 (23.6) (31.8–130.8) 76.5 (29.8) (29.0–175.8) 80.6 (20.5) (42.4–138.9) 79.7 (25.2) (29.0–175.8)

Total bilirubin (mg0dl¹1) 0.9 (0.5) (0.3–2.4) 0.8 (0.3) (0.3–1.7) 0.8 (0.3) (0.3–1.8) 0.8 (0.3) (0.3–2.4)

Haematocrit (%) 45.3 (3.3) (37.8–51.7) 43.3 (4.6) (30.5–55.4) 43.2 (3.9) (34.2–52.9) 43.7 (4.2) (30.5–55.4)

Haemoglobin (g dl¹1) 15.2 (1.0) (12.2–17.4) 14.5 (1.6) (10.2–19.7) 14.4 (1.3) (11.3–18.3) 14.6 (1.4) (10.2–19.7)

Values obtained at the time after dosing on Day 28e

Albumin 1 (g dl¹1) 4.4 (0.3) (3.7–5.0) 4.2 (0.3) (3.5–5.0) 4.3 (0.3) (3.4–5.0) 4.3 (0.3) (3.4–5.0)

Serum creatinine 1 (mg dl¹1) 1.0 (0.2) (0.6–1.6) 0.9 (0.2) (0.5–1.5) 0.8 (0.2) (0.5–1.2) 0.9 (0.2) (0.5–1.6)

BUN 1 (mg dl¹1) 16.1 (4.7) (8.0–28.0) 17.1 (4.8) (9.0–37.0) 15.6 (3.6) (8.0–26.7) 16.3 (4.3) (8.0–37.0)

SGOT 1 (AST) (U l¹1) 26.2 (10.8) (10.0–60.0) 25.4 (8.4) (12.0–57.0) 32.2 (20.9) (12.0–146.0) 28.3 (15.4) (10.0–146.0)

SGPT 1 (ALT) (U l¹1) 25.3 (12.9) (10.0–66.0) 21.7 (12.2) (8.0–80.0) 35.6 (49.7) (9.0–328.0) 28.1 (33.6) (8.0–328.0)

CrCl 1 (mlmin¹1)d 84.5 (24.5) (31.8–138.0) 77.0 (31.9) (29–198.8) 82.3 (21.4) (42.7–139.5) 80.7 (26.6) (29.0–198.8)

Total bilirubin 1 (mg dl¹1) 0.9 (0.4) (0.5–2.1) 0.8 (0.3) (0.2–2.6) 0.8 (0.3) (0.3–1.6) 0.8 (0.3) (0.2–2.6)

Haematocrit 1 (%) 44.2 (3.1) (37.3–50.0) 42.7 (4.7) (27.6–56.1) 43 (3.5) (36.7–53.4) 43.1 (4.0) (27.6–56.1)

Haemoglobin 1 (g dl¹1) 14.9 (1.1) (12.2–17.0) 14.3 (1.7) (9.1–19.7) 14.3 (1.2) (12.3–18.5) 14.4 (1.4) (9.1–19.7)

All covariates excluding height were tested for inclusion in the pharmacokinetic and pharmacodynamic model.aCalculated body surface area (BSA) in [m2]. BSA = (body weight)0.425 © (height)0.725 © 0.007184. bCalculated lean body mass (LBM) in kg ¹ for males: LBM =

1.10 © (body weight) ¹ 128 © (body weight)2/(height)2; for females: LBM = 1.07 © (body weight) ¹ 148 © (body weight)2/(height)2. cCalculated body fat [FAT] inkg = (body weight) ¹ (LBM) (if FAT < 0, FAT = 0). dCreatinine clearance (CrCl) was calculated with the use of the Cockcroft–Gault formula. eThe covariates on day28 were coded as COV1 (e.g. ALB1) to distinguish from the value at baseline.BUN, blood urea nitrogen; SD, standard deviation; SGOT (AST), serum glutamic oxaloacetic transaminase (aspartate aminotransferase); SGPT (ALT), serum glutamicpyruvic transaminase (alanine aminotransferase).

Rivaroxaban Population PK-PD in Japanese Patients 61


employed for analysis. Plasma rivaroxaban was extracted and con-centrations were measured using a fully validated high perform-ance liquid chromatography/tandem mass spectrometer (HPLC/MS-MS) method (Agilent system 1100 coupled with an AppliedBiosystems MDS Sciex API 3000 tandem mass spectrometer)as described previously.36) Concentrations above the lower limitof quantification (LLOQ; 0.500 µg l¹1) were determined with aprecision of 2.00–11.50% and 3.41–7.90% (intra- and inter-assay,respectively) and an accuracy of 94.0–105.0% and 96.3–102.3%(intra- and inter-assay, respectively).

Pharmacodynamic assays: Factor Xa activity, PT, aPTTand HepTest were used as pharmacodynamic parameters. PT andaPTT examine the effects of drugs on coagulation stimulated bythe extrinsic and intrinsic clotting pathways, respectively. HepTestis used to monitor anticoagulation with LMWHs, and it measuresFactor Xa activity indirectly.

These pharmacodynamic parameters were determined usingmethods described previously.37)

Briefly, Factor Xa activity was determined using a two-stepphotometric assay: total Factor X in plasma was activated to FactorXa using Russell’s viper venom in the presence of calcium ions.Subsequently, a chromogenic substrate (S-2765; Chromogenix,Milan, Italy) was hydrolyzed by Factor Xa, releasing p-nitroani-line, which was quantified by spectrophotometry at 405 nm. PTwas determined using freeze-dried thromboplastin derived fromrabbit brain (Neoplastin Plusμ; Roche Diagnostics, Mannheim,Germany), with an International Sensitivity Index of 1.2. aPTTwas assessed using a kaolin-activated test (Roche Diagnostics). PT,aPTT and HepTest (Haemachem, St. Louis, MO, USA) weremeasured with a ball coagulometer KC 10 (Amelung, Germany) asper the manufacturer’s instructions.

Population pharmacokinetic and pharmacodynamic model-ing: Population pharmacokinetic and pharmacodynamic analyseswere performed separately, using non-linear mixed-effects model-ing (NONMEM) version V level 1.1 (GloboMax LLC, Hanover,MD, USA) running in a validated Linux server farm environment.Separate models were developed for each set of pharmacodynamicdata (Factor Xa, PT, aPTT and HepTest). The first-order condi-tional estimation with interaction method (FOCE-INTERACTION)was used for all analyses. All statistics and graphs were generatedusing S-Plus version 6 (Tibco Software Inc, Palo Alto, CA, USA)or SAS version 8.2 (SAS Institute Inc., Cary, NC, USA).

The covariates that had a significant influence on a pharmaco-

kinetic and pharmacodynamic parameter were determined usingthe output from the NONMEM post hoc estimation step.

Pharmacokinetic models: The population pharmacokineticmodel (including all relevant covariates for the population) wasestablished based on actual time of blood sampling relative tothe rivaroxaban dose and amount, plasma concentrations and thedemographic data of the study population. Both one- and two-compartment models were tested during model selection.

Pharmacodynamic models: The matched pharmacokineticand pharmacodynamic data (including baseline data, where phar-macokinetic data were regarded as 0) were employed for analysis.Variables for the population pharmacodynamic model includedmeasured plasma concentration, pharmacodynamic parametersand demographic data. Previously characterized models25,28,34) wereused as the starting model for all pharmacodynamic parameters.Prothrombin time (PT)

Introduction of the parameter FACT (the exponent factor ofplasma rivaroxaban concentration) improved the fit in the modelthat describes the relationship between plasma concentration andPT value.Activated partial thromboplastin time (aPTT)

Introduction of FACT improved the fit in the model that describedthe relationship between plasma concentration and aPTT value.HepTest

Introduction of the Hill coefficient, which is also an exponentfactor of plasma concentration, improved the fit in the model thatdescribed the relationship between plasma concentration andHepTest value.Factor Xa activity

As with the HepTest model, introduction of the Hill coefficientimproved the fit in the model to describe the relationship betweenplasma concentration and Factor Xa activity.

Model development: The pharmacokinetic and pharmaco-dynamic model structures were optimized with the goal ofreducing any unexplained variability. Covariates were incorporatedinto the models if the likelihood ratio test (LRT) showed areduction in objective function value (OFV) of >6.63 (in the caseof one degree of freedom [df]). After backwards removal from thefull model, components were retained if the LRT showed a changein OFV of >10.8 (in the case of one df). A decrease by >6.63 or>10.8 points between competing models corresponds to prespe-cified nominal P values of <0.01 and <0.001, respectively. Inaddition to goodness-of-fit (GOF) statistics using the LRT, GOF

Fig. 1. Sampling times for each of the three studies are described relative to rivaroxaban administrationStudies A and B: Baseline (B only) and on days 14 and 28 before dosing, 2 « 1 h after dosing and 4.5 « 1.5 h after dosing. Study C: Baseline and on days 14 and 281 « 1 h before dosing, 2 « 1 h after dosing and 5 « 2 h after dosing.



graphs were also inspected to judge the adequacy of the model. Tocover for the residual variability not reflected by the populationmodel, additional error, proportional error and combined errormodels were tested, whose error models were selected based onevaluation of objective function values and GOF plots.

Furthermore, the final pharmacokinetic and pharmacodynamicmodels were also validated by visual predictive check (VPC) anda non-parametric bootstrap method to check the distributions ofparameters.

Prediction and simulation: One thousand virtual Japanesepatients with AF were simulated using the data from these threephase II pharmacokinetic-pharmacodynamic studies conductedin patients residing in Japan with AF (Tables 1, 2 and 3) and theresults of an epidemiological investigation of the demographics ofJapanese patients with AF,38) which generated the ratio of male tofemale patients used in the simulation. Data regarding the covariateblood urea nitrogen (BUN) were generated for each gender byresampling the data from the three phase II studies. The resultantsimulated patient population was used in the final populationmodel to estimate individual exposure levels and pharmacody-namic parameters at steady state.

Simulation of virtual Caucasian patients with AF receiving20mg o.d. rivaroxaban has been described previously.34) In brief,the simulation was based on the data from two phase II dose-ranging trials of rivaroxaban for treatment of DVT,29,30) with thedemographics modified to reflect patients with AF enrolled in trialsof ximelagatran.39,40) US epidemiology and phase I rivaroxabandata were employed to describe the relationship between age andrenal function.41,42) Descriptive statistics of steady-state AUC0–24h

and Cmax were compared between the virtual Caucasian patientswith AF receiving 20mg rivaroxaban o.d. and the virtual Japanesepatients with AF receiving 20mg or 15mg rivaroxaban o.d.

Results

Number of patients and pharmacokinetic and pharmacody-namic data points: A total of 182 subjects resident in Japan (148men, 34 women) provided data for the population pharmacokineticand pharmacodynamic analyses giving 842 data points for phar-macokinetic evaluation. For pharmacodynamic evaluations (includ-ing the additional baseline data), 987 data points were obtained forboth PT and HepTest, 986 for aPTT and 985 for Factor Xa.

Pharmacokinetic model: Of the one- and two-compartmentpharmacokinetic models that were tested during the model-buildingprocess, a one-compartment model with first-order absorptionand proportional error model was found to optimally describethe plasma concentration versus time profile data taking intoaccount the sparse sampling approach used in these clinical studies(Fig. 2A), consistent with previous pharmacokinetic modeling inCaucasians.27,28,34) The final population estimates for pharmaco-kinetic parameters, such as the oral absorption rate constant (Ka),clearance (CL) and volume of distribution (V), are shown inTable 4; the residual (unexplained) variability of the model was40.2%. Relative bioavailability (F1) was introduced and estimatedat 24.4% to describe inter-individual variability from the populationF1 mean, which was defined as 1.0. GOF graph examples with bothpopulation predictions and individual predictions are shown inFigures 3A and 3B.

On the basis of empirical Bayes estimate plots, the followingcovariates in the full population pharmacokinetic model wereselected for evaluation: body surface area (BSA) and BUN forCL, lean body mass (LBM) for Ka, LBM for V, and serum glutamicpyruvic transaminase (SGPT) for F1. Each covariate was testedseparately during the backward elimination process and BUNwas the only covariate that remained in the model after this step.

Fig. 2. Final structure pharmacokinetic and pharmacodynamic models(A) Final structural pharmacokinetic model for rivaroxaban. (B) Final pharmacodynamic models, showing the correlation between rivaroxaban plasma concentrations(Cp) and the pharmacodynamic effects of Factor Xa activity, prothrombin time (PT), activated partial thromboplastin time (aPTT) and HepTest. BASE, baseline value;BUN, blood urea nitrogen; CL, apparent total clearance after oral administration; EC50, concentration resulting in half of the maximum effect; Emax, maximum effect;F1, relative bioavailability; ©, random effect parameters for inter-individual variability; FACT, exponent factor for plasma concentration; Ka, oral absorption rateconstant; V, apparent volume of distribution after oral administration.



CL was predicted to decrease/increase by 31.8% and 16.1%,respectively, when comparing a mean subject to the upper/lowerends of the demographic distribution in the dataset that wasused for this evaluation (BUN lower end: 7.0mg dl¹1, upper end:36.0mg dl¹1). According to the population pharmacokinetic model,the magnitude of moderate or severe renal impairment (representedby BUN) would be expected to have a demonstrable impact on thepharmacokinetics of rivaroxaban.

Pharmacodynamic models:Prothrombin time (PT)

The selected base model consisted of a direct link combinedpower plus a linear model with proportional error model. On thebasis of plots of empirical bayesian estimates, no covariates in thefull population PT model were selected. As final model, a directlink combined power plus linear model with proportional errormodel was found to best fit the study data (Fig. 2B). The baselinePT in the study population was 13.7 s and the slope of the

Table 4. Final population estimates for pharmacokinetic parameters

Model Parameter Population mean (CV %) SE/mean (%)

Bootstrap estimates95% confidence interval

Lower Upper

Pharmacokinetic

Ka (h¹1) 0.6 11.433 0.475 0.762

CL (l h¹1) 4.72 3.686 4.389 5.031

V (l) 42.9 6.224 37.391 48.092

BUN for CL ¹0.0165a ¹27.091 ¹0.0255 ¹0.00717

Inter-individual variability

Ka 0.463 (68%) 35.205 41.506% 91.372%

CL 0.0452 (21.3%) 27.655 14.200% 26.443%

F1 0.0596 (24.4%) 39.933 14.749% 34.824%

Residual variability (proportional error model) 0.162 (40.2%) 7.778 37.003% 43.152%

Prothrombin time (PT)

Baseline (s) 13.7 0.869 13.506 13.955

Slope (s0l µg¹1) 0.0227 16.652 0.017 0.031

FACT (¹) 1.1 2.645 1.046 1.154


Baseline 0.00633 (8.0%) 17.536 6.486% 9.238%

Slope 0.0731 (27.0%) 17.237 22.636% 31.187%


Activated partial thromboplastin time (aPTT)

Baseline (s) 32.6 0.92 32.054 33.190

Slope (s0l µg¹1) 0.0658 5.897 0.0583 0.0735

FACT (¹) 0.000156 13.141 0.000114 0.000200


Baseline 0.00914 (9.6%) 13.786 8.122% 10.666%

Slope 0.101 (31.8%) 31.386 21.004% 42.548%


HepTest

Baseline (s) 17.9 0.67 17.681 18.127

Emax (s) 43.2 4.977 38.926 47.308

EC50 (µg l¹1) 240 8.833 199.454 283.530

ALB1 on EC50 0.147b 31.497 0.0471 0.238

Hill coefficient (FACT) 1.18 3.034 1.125 1.268


Baseline 0.00228 (4.8%) 27.851 3.078% 5.287%

EC50 0.0113 (10.6%) 36.372 5.903% 13.775%


Factor Xa

Baseline (U ml¹1) 0.803 1.029 0.787 0.820

Age on baseline ¹0.00656c ¹19.512 ¹0.00912 ¹0.00409

Emax (¹) 0.928 2.608 0.875 0.986

EC50 (µg l¹1) 221 4.57 199.754 245.720

Hill coefficient (FACT) (¹) 1.16 2.983 1.103 1.249


Baseline 0.019 (13.8%) 12.158 11.969% 15.241%

EC50 0.0222 (14.9%) 19.64 11.354% 17.538%


aCalculated as CL © (1 – 0.0165 © [BUN – 16.73]); bCalculated as EC50 © (1 + 0.147 © [ALB1 – 4.28]); cCalculated as BASE* (1 – 0.00656 © [AGE – 65.59]).ALB1, albumin 1; BUN, blood urea nitrogen; CL, clearance; EC50, effective concentration generating 50% of effect; Emax, maximal effect; F1, bioavailability; FACT,exponent factor for plasma concentration; Ka, absorption rate constant; SE, standard error; V, volume of distribution.



correlation between PT and rivaroxaban concentration was0.0227 s0l µg¹1 (Table 4). No covariates exhibited a statisticallysignificant effect on PT prolongation. Examples of GOF graphsare shown in Figures 3C and 3D.Activated partial thromboplastin time (aPTT)

The selected base model consisted of a direct link combinedpower plus a linear model with proportional error model. On thebasis of plots of empirical bayesian estimates, no covariates in fullpopulation aPPT model were selected. As final model, a direct linkcombined power plus linear model with proportional error modelwas found to fit the study data best (Fig. 2B). The baseline aPTTin the study population was 32.6 s and the slope of the correlationbetween aPTT and rivaroxaban concentration was 0.0658 s0l µg¹1

(Table 4). No covariates exhibited a statistically significant effecton aPTT prolonging properties. Examples of GOF graphs areshown in Figures 3E and 3F.HepTest

The selected base model consisted of a direct link Emax modelwith Hill coefficient with a proportional error model. On the basisof plots of empirical bayesian estimates, the following covariates infull population HepTest model were selected for evaluation: studynumber (STUD) and albumin at screening (ALBU) for baselineand STUD and albumin at day 28 (ALB1) for the effectiveconcentration generating 50% of effect (EC50). As final model, adirect link Emax pharmacodynamic model including the Hill

coefficient with a proportional error model was found to fit thestudy data best (Fig. 2B). Emax was 43.2 s, the baseline HepTestvalue in the study population was 17.9 s and EC50 was 240µg l¹1

(Table 4). Plasma albumin level was selected as a covariate as itwas found to influence EC50 in the analysis. EC50 was predicted toincrease by 10.6%, or decrease by 12.9% when comparing a meansubject to the upper or lower ends of the albumin concentrationrange (albumin range: 3.4–5.0 g dl¹1), respectively, of the demo-graphic distribution in the data set that was used for this evaluation.Examples of GOF graphs are shown in Figures 3G and 3H.Factor Xa activity

The selected base model consisted of a direct link inhibitoryEmax model with Hill coefficient, with a proportional error model.On the basis of plots of empirical bayesian estimates, the followingcovariates in the full population model were selected forevaluation: age (AGE) and serum glutamic oxaloacetic trans-aminase at screening time (SGOT) for baseline and study number(STUD) for Emax. As the final model, a direct link Emax phar-macodynamic model including the Hill coefficient with a propor-tional error model was found to fit the study data best (Fig. 2B).Emax was 0.928, the baseline Factor Xa activity value in the studypopulation was 0.803 unitsml¹1 and the EC50 for rivaroxaban was221 µg l¹1 (Table 4). Age was found to influence baseline levels.Baseline levels were predicted to decrease/increase by 17.32% and23.35% when comparing a mean subject to the upper/lower ends of

Fig. 3. Goodness-of-fit graphs with lines of identity(A) Population predictions compared with observed values for rivaroxaban plasma concentrations. (B) Individual predictions compared with observed values forrivaroxaban plasma concentrations. (C) Population predictions compared with observed values for PT. (D) Individual predictions compared with observed values for PT.(E) Population predictions compared with observed values for aPTT. (F) Individual predictions compared with observed values for aPTT. (G) Population predictionscompared with observed values for HepTest. (H) Individual predictions compared with observed values for HepTest. (I) Population predictions compared with observedvalues for Factor Xa activity. (J) Individual predictions compared with observed values for Factor Xa activity.



the demographic age distribution in the data used in this study (agerange: 30–92 years). Examples of GOF graphs are shown inFigures 3I and 3J.

Simulation results: The demographics of the virtual patientswith AF are reported in Table 5. The simulations revealed thatthe distribution of both Cmax,ss and AUC0–24,ss in Japanese patientswith AF receiving a 15mg o.d. dose of rivaroxaban would alsoapproximate Cmax,ss and AUC0–24,ss in Caucasian patients withAF receiving a 20mg o.d. dose, albeit slightly lower in the caseof Cmax,ss. The simulations for the 20mg o.d. dose of rivaroxabanshowed that the AUC0–24,ss and Cmax,ss distribution for Japanesepatients with AF would be slightly higher than AUC0–24,ss

and Cmax,ss distribution for Caucasian patients with AF (Fig. 4).This might lead to a larger proportion of patients with higher Cmax,ss

and to a lesser extent AUC0–24,ss in Japanese patients with AFtreated with 20mg o.d. rivaroxaban compared with Caucasianpatients with AF receiving 20mg o.d. doses of rivaroxaban(Fig. 4).

Model validation: All final pharmacokinetic and pharmaco-dynamic models were validated by visual inspection of GOF

plots, VPC and bootstrap simulation comprehensively. GOF graphswith both population predictions and individual predictions for allmodels are shown in Figure 2. While the population predictionsversus dependent variable plots were equally distributed aroundthe line of identity (Fig. 3A), the individual predictions versusdependent variable plots showed a small bias, especially forpharmacokinetics (Fig. 3B). The model seemed to slightly under-estimate and overestimate the plasma rivaroxaban concentrationat lower and higher concentrations, respectively. Because of thesparse sampling approach, it proved necessary to simplify thepharmacokinetic model in the patient population pharmacokineticanalyses to an oral, one-compartment model, resulting in a trendto underestimate Cmax—as theoretically expected. In addition,1,000 non-parametric bootstrap estimations were performed on theoriginal dataset, and all pharmacokinetic parameters were esti-mated based on the final model, confirming the robustness ofthe model and the good precision in all estimated parameters(Table 4). VPCs were conducted to further investigate the stabilityand robustness of the final models. Individual empirical bayesianestimates for pharmacokinetics were estimated using the popula-tion means and individual patient information by simulating200 sub-problems. Ninety percent prediction intervals were thencalculated and compared with the actual observations. The resultsof the VPCs are shown in Figures 5 and 6. The observations wererandomly distributed within the calculated 90% prediction inter-vals, demonstrating no obvious bias in the estimated populationpharmacokinetic or pharmacodynamic parameters of the finalmodel. Only a few observations were outside the 90% predictioninterval, demonstrating that the variability of the parameters wasadequately captured.

Discussion

Previously, phase I results with rivaroxaban had shown thathealthy Japanese subjects (data on file) experienced a 20–40%

Table 5. Demographics for virtual patients with atrial fibrillation

Parameter Ethnicity Mean « SD (range)

Lean body mass (kg) Japanese Male: 54 « 6 (40–68)

Female: 40 « 4 (35–49)

Caucasian Male: 61 « 7 (41–80)

Female: 47 « 5 (33–59)

Age (years) Japanese 67 « 11 (30–92)

Caucasian 71 « 8 (51–92)

Serum creatinine (mg dl¹1) Japanese 0.88 « 0.22 (0.40–1.60)

Caucasian 1.20 « 0.36 (0.43–2.51)

Male/female Japanese 62%/38%

Caucasian 69%/31%

SD, standard deviation.

Fig. 4. Comparison of estimated levels of exposureMaximum drug concentration in plasma at steady state [Cmax,ss] and area under the curve from 0 to 24 h at steady state [AUC0–24,ss] for once-daily (o.d.) regimen of 15mgand 20mg rivaroxaban for Japanese patients and 20mg for Caucasian patients. From the top: 10% point, 25% point, median, 75% point, 90% point, “:”: 5% point and95% point.



increase in AUC and a 20–30% increase in Cmax compared withhealthy Caucasian subjects.26) This phenomenon was also observedin Japanese patients receiving ximelagatran.32,33) The primaryexplanation given for the increased exposure to ximelagatran inJapanese patients was weight- and age-related decreases in renalelimination of the active metabolite melagatran. Because theJapanese population typically has a greater proportion of elderlypeople than most other countries,5) and its residents generallyhave lower body mass indices than countries with predominantlyCaucasian populations,31) it would appear that for a given dose of arenally eliminated drug, a population of Japanese patients mightbe expected to experience increased exposure compared with anaverage population of Caucasian patients. Indeed, results fromclinical pharmacology studies have revealed that renal function(i.e., renal CL), and to a lesser extent age and body weight arefactors that influence the exposure level of rivaroxaban.34,42–44)

This study used data pooled from three phase II studies ofJapanese patients with AF. Study A included 10mg and 20mg

b.i.d. rivaroxaban doses. As a total daily dose of 40mg wasconsidered too high for Japanese patients with AF, studies B and Cwere carried out to evaluate 2.5, 5 and 10mg b.i.d and 10, 15 and20mg o.d. doses, respectively. These studies included patients withCHADS2 scores of 1 and 0 as well as scores ²2, thereby providinga broad representation of patients with AF. Japanese guidelinesrecommend that patients with AF and a CHADS2 score ²2 shouldreceive oral anticoagulation therapy, and this should also beconsidered in patients with a CHADS2 score of 1.11) Moreover,these guidelines also state that oral anticoagulant therapy should beconsidered for patients with a CHADS2 score of 0 in the presenceof at least one the following risk factors for stroke: cardiomyop-athy, age 65–74 years, female gender, coronary artery disease orthyrotoxicosis.

Analysis of data pooled from the three phase II studies ofJapanese patients with AF showed that the pharmacokinetics ofrivaroxaban were well described by a one-compartment model withfirst-order absorption and elimination with a covariate related to

Fig. 5. Visual predictive check plots representing plasma rivaroxaban concentration versus time stratified by regimen (2.5mg twice daily [b.i.d.], 5mg b.i.d.,10mg once daily [o.d.] and b.i.d., 15mg o.d. and 20mg o.d. and b.i.d.)Each panel shows observed rivaroxaban concentrations (closed circles) and the predicted (n = 200) 5, 95 percentile points (dotted lines). Pharmacokinetics of:(A) 2.5mg b.i.d. dose; (B) 5mg b.i.d. dose; (C) 10mg o.d. dose; (D) 10mg b.i.d. dose; (E) 15mg o.d. dose; (F) 20mg o.d. dose; (G) 20mg b.i.d. dose.



renal function against CL (represented by BUN). This observationis consistent with previous pharmacokinetic and pharmacodynamicstudies of rivaroxaban therapy for DVT and for the prevention ofvenous thromboembolism after orthopaedic surgery.27,28,34)

The relationships between plasma concentration and pharmaco-dynamic parameters were best described by an Emax model forFactor Xa activity and HepTest, and by linear models for PT andaPTT (by including a baseline value) using direct link models,yielding results consistent with previous pharmacokinetic andpharmacodynamic evaluations of rivaroxaban in subjects andpatients from both Japanese studies (data on file) and globalphase I and II venous thromboembolic prevention and DVTtreatment studies.25–28,34,42) Hence the relationships betweenrivaroxaban plasma concentrations and pharmacodynamic parame-ters in patients with AF in Japan reported in this study areconsistent with these previous findings.

Data available to date have consistently shown a dose-propor-tional relationship and no significant ethnicity in the relationshipbetween rivaroxaban plasma concentrations and pharmacodynamicparameters across all ethnicities,25–28,34,42,45) which was also shownin this study.

This study utilized the data from three phase II studies in Japa-nese patients with AF, which included a range of doses bracketingthe 20mg o.d. dose of rivaroxaban selected for the global ROCKETAF trial. However, in practice, Japanese clinicians prefer to uselower doses of warfarin, even in patients under 70 years of age,primarily driven by concerns about bleeding. Recent registry datafor Japanese patients with AF showed that 66% of patientsreceiving warfarin had an INR within the range 1.6–2.6 (mean forall patients 1.9), irrespective of age.14) Results from the currentmodelling study indicate that the simulated steady-state exposure of

Japanese patients with AF treated with rivaroxaban 15mg o.d.approximates both the Cmax,ss and AUC0–24,ss achieved in simulatedCaucasian patients with AF treated with rivaroxaban 20mg o.d.34)

The finding that the simulation of the 20mg o.d. dose ofrivaroxaban in Japanese patients with AF achieved slightly higherpharmacokinetic parameters than in Caucasian patients with AFtreated with the same dose is the same tendency of the mini-mal effect of age and body weight seen previously.34,42,43) Thisdifference in exposure levels between Japanese and Caucasianmight include some ethnic factors which cannot be explained fromonly age, body weight and renal function modelled in Japanese orCaucasian population pharmacokinetic models.34) This could bebecause the information obtained from sparse sampling data inPhase II trials might be limited.

In addition, the 20mg o.d. dose for Japanese patients wouldcarry the risk of more patients including “fragile” patients whoare very old and/or have a much lower body weight, thoughexperiencing higher Cmax,ss and to a lesser extent AUC0–24,ss

compared with Caucasian patients with AF receiving a 20mgo.d. dose in these simulation results. Considering these factorstogether, the 15mg o.d. dose for Japanese patients with AFappeared to provide a greater margin to allow for potentialindividual variability, particularly in the context of preferred loweranticoagulant doses in Japanese clinical practice.11,14) These resultsprovide important support for the selection of 15mg o.d. as thedose of rivaroxaban to be evaluated for the prevention of strokein Japanese patients with AF, integrating both ethnic pharmaco-kinetic effects using population pharmacokinetic and pharmaco-dynamic modeling (likely explained by a lower distribution ofbody mass indices) as well as the clinical practice differences inJapan, where the preference is to utilize lower levels of anti-

Fig. 6. Visual predictive check plots of pharmacodynamic parameters versus plasma rivaroxaban concentrationEach panel shows observed rivaroxaban concentrations and pharmacodynamic parameters (closed circles) and the predicted (n = 200) 5, 95 percentile points (dottedlines). (A) Prothrombin time. (B) Activated partial thromboplastin time. (C) Factor Xa activity. (D) HepTest.



coagulation. The rivaroxaban dose was lowered to 10mg o.d. inpatients with renal impairment in J-ROCKET AF, in a similarmanner to the global ROCKET AF trial.23)

Further PK-PD sampling and also population PK-PD evaluationwere planned in the ROCKET AF and J-ROCKET AF trial toconfirm whether this dose setting is appropriate from a PK-PDpoint of view.

Acknowledgments: The authors thank Mark Hillen and JoanneMcGrail, who provided medical writing services with fundingfrom Bayer HealthCare Pharmaceuticals and Johnson & JohnsonPharmaceutical Research & Development, LLC. The authorswould also like to thank Dr. Martin Blunck, Dr. Partha Nandy,Dr. Ihab Girgis and Dr. Steven Xu for their insightful commentsabout the manuscript.

References

1) Ohsawa, M., Okayama, A., Sakata, K., Kato, K., Itai, K., Onoda, T. andUeshima, H.: Rapid increase in estimated number of persons with atrialfibrillation in Japan: an analysis from national surveys on cardiovasculardiseases in 1980, 1990 and 2000. J. Epidemiol., 15: 194–196 (2005).

2) Tanizaki, Y., Kiyohara, Y., Kato, I., Iwamoto, H., Nakayama, K.,Shinohara, N., Arima, H., Tanaka, K., Ibayashi, S. and Fujishima, M.:Incidence and risk factors for subtypes of cerebral infarction in a generalpopulation: the Hisayama study. Stroke, 31: 2616–2622 (2000).

3) Stroke Risk in Atrial Fibrillation Working Group: Independent predictorsof stroke in patients with atrial fibrillation: a systematic review.Neurology, 69: 546–554 (2007).

4) Lutz, W., Sanderson, W. and Scherbov, S.: The coming acceleration ofglobal population ageing. Nature, 451: 716–719 (2008).

5) World Health Organization: World Population Ageing 1950–2050;Chapter IV. Demographic Profile of the Older Population. http://www.un.org/esa/population/publications/worldageing19502050/pdf/90chapteriv.pdf(updated 2009; last accessed 10 January 2011).

6) Turin, T. C., Kokubo, Y., Murakami, Y., Higashiyama, A., Rumana, N.,Watanabe, M. and Okamura, T.: Lifetime risk of stroke in Japan. Stroke,41: 1552–1554 (2010).

7) Singer, D. E., Albers, G. W., Dalen, J. E., Fang, M. C., Go, A. S.,Halperin, J. L., Lip, G. Y. and Manning, W. J.: American College ofChest Physicians. Antithrombotic therapy in atrial fibrillation: AmericanCollege of Chest Physicians evidence-based clinical practice guidelines(8th Edition). Chest, 133: 546S–92S (2008).

8) Hart, R. G., Pearce, L. A. and Aguilar, M. I.: Meta-analysis:antithrombotic therapy to prevent stroke in patients who have nonvalvularatrial fibrillation. Ann. Intern. Med., 146: 857–867 (2007).

9) Camm, A. J., Kirchhof, P., Lip, G. Y., Schotten, U., Savelieva, I., Ernst,S., Van Gelder, I. C., Al Attar, N., Hindricks, G., Prendergast, B.,Heidbuchel, H., Alfieri, O., Angelini, A., Atar, D., Colonna, P.,De Caterina, R., De Sutter, J., Goette, A., Gorenek, B., Heldal, M.,Hohloser, S. H., Kolh, P., Le Heuzey, J. Y., Ponikowski, P., Rutten, F. H.,Vahanian, A., Auricchio, A., Bax, J., Ceconi, C., Dean, V., Filippatos, G.,Funck-Brentano, C., Hobbs, R., Kearney, P., McDonagh, T., Popescu,B. A., Reiner, Z., Sechtem, U., Sirnes, P. A., Tendera, M., Vardas, P. E.,Widimsky, P., Vardas, P. E., Agladze, V., Aliot, E., Balabanski, T.,Blomstrom-Lundqvist, C., Capucci, A., Crijns, H., Dahlof, B., Folliguet,T., Glikson, M., Goethals, M., Gulba, D. C., Ho, S. Y., Klautz, R. J.,Kose, S., McMurray, J., Perrone, F. P., Raatikainen, P., Salvador, M. J.,Schalij, M. J., Shpektor, A., Sousa, J., Stepinska, J., Uuetoa, H.,Zamorano, J. L. and Zupan, I.: Guidelines for the management of atrialfibrillation: The Task Force for the Management of Atrial Fibrillation ofthe European Society of Cardiology (ESC). Eur. Heart J., 31: 2369–2429(2010).

10) Fuster, V., Ryden, L. E., Cannom, D. S., Crijns, H. J., Curtis, A. B.,Ellenbogen, K. A., Halperin, J. L., Kay, G. N., Le Huezey, J. Y., Lowe,J. E., Olsson, S. B., Prystowsky, E. N., Tamargo, J. L. and Wann, L. S.:2011 ACCF/AHA/HRS focused updates incorporated into the ACC/AHA/ESC 2006 guidelines for the management of patients with atrialfibrillation: a report of the American College of Cardiology Foundation/American Heart Association Task Force on practice guidelines.Circulation, 123: e269–e367 (2011).

11) Japanese Circulation Society Joint Working Group: Guidelines for

pharmacotherapy of atrial fibrillation (JCS 2008): digest version. Circ.J., 74: 2479–2500 (2010).

12) Yasaka, M. and Yamaguchi, T.: Secondary prevention of stroke inpatients with nonvalvular atrial fibrillation: optimal intensity of anti-coagulation. CNS Drugs, 15: 623–631 (2001).

13) Yasaka, M., Minematsu, K. and Yamaguchi, T.: Optimal intensity ofinternational normalized ratio in warfarin therapy for secondaryprevention of stroke in patients with non-valvular atrial fibrillation.Intern. Med., 40: 1183–1188 (2001).

14) Atarashi, H., Inoue, H., Okumura, K., Yamashita, T., Kumagai, N. andOrigasa, H.: Present status of anticoagulation treatment in Japanesepatients with atrial fibrillation. Circ. J., 75: 1328–1333 (2011).

15) Hylek, E. M.: The need for new oral anticoagulants in clinical practice.J. Cardiovasc. Med. (Hagerstown), 10: 605–609 (2009).

16) Nieuwlaat, R., Capucci, A., Camm, A. J., Olsson, S. B., Andresen, D.,Davies, D. W., Cobbe, S., Breithardt, G., Le Heuzey, J. Y., Prins, M. H.,Lévy, S. and Crijns, H. J.: European Heart Survey Investigators: Atrialfibrillation management: a prospective survey in ESC member countries:the Euro Heart Survey on Atrial Fibrillation. Eur. Heart J., 26: 2422–2434 (2005).

17) Waldo, A. L., Becker, R. C., Tapson, V. F. and Colgan, K. J.: NABORSteering Committee: Hospitalized patients with atrial fibrillation and ahigh risk of stroke are not being provided with adequate anticoagulation.J. Am. Coll. Cardiol., 46: 1729–1736 (2005).

18) Ruff, C. T., Giugliano, R. P., Antman, E. M., Crugnale, S. E., Bocanegra,T., Mercuri, M., Hanyok, J., Patel, I., Shi, M., Salazar, D., McCabe,C. H. and Braunwald, E.: Evaluation of the novel factor Xa inhibitoredoxaban compared with warfarin in patients with atrial fibrillation:design and rationale for the Effective aNticoaGulation with factor xAnext GEneration in Atrial Fibrillation-Thrombolysis In MyocardialInfarction study 48 (ENGAGE AF-TIMI 48). Am. Heart J., 160: 635–641 (2010).

19) Lopes, R. D., Alexander, J. H., Al-Khatib, S. M., Ansell, J., Diaz, R.,Easton, J. D., Gersh, B. J., Granger, C. B., Hanna, M., Horowitz, J.,Hylek, E. M., McMurray, J. J., Verheugt, F. W. and Wallentin, L.:ARISTOTLE Investigators: Apixaban for Reduction In Stroke andOther ThromboemboLic Events in Atrial Fibrillation (ARISTOTLE)trial: design and rationale. Am. Heart J., 159: 331–339 (2010).

20) Connolly, S. J., Ezekowitz, M. D., Yusuf, S., Eikelboom, J., Oldgren, J.,Parekh, A., Pogue, J., Reilly, P. A., Themeles, E., Varrone, J., Wang, S.,Alings, M., Xavier, D., Zhu, J., Diaz, R., Lewis, B. S., Darius, H., Diener,H. C., Joyner, C. D. and Wallentin, L.: RE-LY Steering Committeeand Investigators: Dabigatran versus warfarin in patients with atrialfibrillation. N. Engl. J. Med., 361: 1139–1151 (2009).

21) Connolly, S. J., Ezekowitz, M. D., Yusuf, S., Reilly, P. A. and Wallentin,L.: Newly identified events in the RE-LY trial. N. Engl. J. Med., 363:1875–1876 (2010).

22) Connolly, S. J., Eikelboom, J., Joyner, C., Diener, H. C., Hart, R.,Golitsyn, S., Flaker, G., Avezum, A., Hohnloser, S. H., Diaz, R., Talajic,M., Zhu, J., Pais, P., Budaj, A., Parkhomenko, A., Jansky, P.,Commerford, P., Tan, R. S., Sim, K. H., Lewis, B. S., Van Mieghem,W., Lip, G. Y. H., Kim, J. H., Lanas-Zanetti, F., Gonzalez-Hermosillo,A., Dans, A. L., Munawar, M., O’Donnell, M., Lawrence, J., Lewis, G.,Afzal, R. and Yusuf, S.: AVERROES Steering Committee andInvetigators: Apixaban in patients with atrial fibrillation. N. Engl. J.Med., 364: 806–817 (2011).

23) Patel, M. R., Mahaffey, K. W., Garg, J., Pan, G., Singer, D. E., Hacke, W.,Breithardt, G., Halperin, J. L., Hankey, G. J., Piccini, J. P., Becker, R. C.,Nessel, C. C., Paolini, J. F., Berkowitz, S. D., Fox, K. A. A. and Califf,R. M.: ROCKET AF Investigators: Rivaroxaban versus warfarin innonvalvular atrial fibrillation. N. Engl. J. Med., 365: 883–891 (2011).

24) Weinz, C., Schwarz, T., Kubitza, D., Mueck, W. and Lang, D.:Metabolism and excretion of rivaroxaban, an oral, direct Factor Xainhibitor, in rats, dogs and humans. Drug Metab. Dispos., 37: 1056–1064(2009).

25) Mueck, W., Becka, M., Kubitza, D., Voith, B. and Zuehlsdorf, M.:Population model of the pharmacokinetics and pharmacodynamics ofrivaroxaban–an oral, direct Factor Xa inhibitor–in healthy subjects. Int. J.Clin. Pharmacol. Ther., 45: 335–344 (2007).

26) Kubitza, D., Becka, M., Wensing, G., Voith, B. and Zuehlsdorf, M.:Safety, pharmacodynamics, and pharmacokinetics of BAY 59-7939—anoral, direct Factor Xa inhibitor—after multiple dosing in healthy malesubjects. Eur. J. Clin. Pharmacol., 61: 873–880 (2005).

27) Mueck, W., Borris, L. C., Dahl, O. E., Haas, S., Huisman, M. V., Kakkar,A. K., Kälebo, P., Muelhofer, E., Misselwitz, F. and Eriksson, B. I.:Population pharmacokinetics and pharmacodynamics of once- and twice-daily rivaroxaban for the prevention of venous thromboembolism in



http://www.ncbi.nlm.nih.gov/sites/entrez?cmd=Retrieve&db=PubMed&list_uids=16195640&dopt=Abstract




http://www.un.org/esa/population/publications/worldageing19502050/pdf/90chapteriv.pdf

http://www.un.org/esa/population/publications/worldageing19502050/pdf/90chapteriv.pdf































patients undergoing total hip replacement. Thromb. Haemost., 100: 453–461 (2008).

28) Mueck, W., Eriksson, B. I., Bauer, K. A., Borris, L., Dahl, O. E., Fisher,W. D., Gent, M., Haas, S., Huisman, M. V., Kakkar, A. K., Kälebo,P., Kwong, L. M., Misselwitz, F. and Turpie, A. G. G.: Populationpharmacokinetics and pharmacodynamics of rivaroxaban—an oral, directFactor Xa inhibitor—in patients undergoing major orthopaedic surgery.Clin. Pharmacokinet., 47: 203–216 (2008).

29) Agnelli, G., Gallus, A., Goldhaber, S. Z., Haas, S., Huisman, M. V., Hull,R. D., Kakkar, A. K., Misselwitz, F. and Schellong, S.: ODIXa-DVTStudy Investigators: Treatment of proximal deep-vein thrombosis withthe oral direct Factor Xa inhibitor rivaroxaban (BAY 59-7939): theODIXa-DVT (oral direct Factor Xa inhibitor BAY 59-7939 in patientswith acute symptomatic deep-vein thrombosis) study. Circulation, 116:180–187 (2007).

30) Büller, H. R., Lensing, A. W., Prins, M. H., Agnelli, G., Cohen, A.,Gallus, A. S., Misselwitz, F., Raskob, G., Schellong, S. and Segers, A.:Einstein-DVT Dose-Ranging Study investigators: A dose-ranging studyevaluating once-daily oral administration of the Factor Xa inhibitorrivaroxaban in the treatment of patients with acute symptomatic deep veinthrombosis. The EINSTEIN-DVT Dose-Ranging Study. Blood, 112:2242–2247 (2008).

31) World Health Organization: Global database on body mass index. http://apps.who.int/bmi/index.jsp (updated 2010; last accessed 10 January2011).

32) Johansson, L. C., Andersson, M., Fager, G., Gustafsson, D. and Eriksson,U. G.: No influence of ethnic origin on the pharmacokinetics andpharmacodynamics of melagatran following oral administration ofximelagatran, a novel oral direct thrombin inhibitor, to healthy malevolunteers. Clin. Pharmacokinet., 42: 475–484 (2003).

33) Wernevik, L. C., Nystrom, P., Andersson, M., Johnsson, G., Bylock, A.,Nakanishi, T. and Eriksson, U. G.: Comparable pharmacokinetics andpharmacodynamics of melagatran in Japanese and Caucasian volunteersafter oral administration of the direct thrombin inhibitor ximelagatran.Clin. Pharmacokinet., 45: 85–94 (2006).

34) Mueck, W., Lensing, A. W., Agnelli, G., Decousus, H., Prandoni, P. andMisselwitz, F.: Rivaroxaban: population pharmacokinetic analyses inpatients treated for acute deep-vein thrombosis and exposure simulationsin patients with atrial fibrillation treated for stroke prevention. Clin.Pharmacokinet., 50: 675–686 (2011).

35) Gage, B. F., Waterman, A. D., Shannon, W., Boechler, M., Rich, M. W.and Radford, M. J.: Validation of clinical classification schemes forpredicting stroke: results from the National Registry of Atrial Fibrillation.JAMA, 285: 2864–2870 (2001).

36) Rohde, G.: Determination of rivaroxaban—a novel, oral, direct Factor Xainhibitor—in human plasma by high-performance liquid chromatog-raphy-tandem mass spectrometry. J. Chromatogr. B Analyt. Technol.Biomed. Life Sci., 872: 43–50 (2008).

37) Kubitza, D., Becka, M., Voith, B., Zuehlsdorf, M. and Wensing, G.:Safety, pharmacodynamics, and pharmacokinetics of single doses of BAY59-7939, an oral, direct Factor Xa inhibitor. Clin. Pharmacol. Ther., 78:412–421 (2005).

38) Tomita, F., Kohya, T., Sakurai, M., Kaji, T., Yokoshiki, H., Sato, M.,Sasaki, K., Itoh, Y., Konno, M. and Kitabatake, A.: Prevalence andclinical characteristics of patients with atrial fibrillation: analysis of20,000 cases in Japan. Jpn. Circ. J., 64: 653–658 (2000).

39) Halperin, J. L.: Executive Steering Committee SPORTIF III and V StudyInvestigators. Ximelagatran compared with warfarin for prevention ofthromboembolism in patients with nonvalvular atrial fibrillation: Ration-ale, objectives, and design of a pair of clinical studies and baseline patientcharacteristics (SPORTIF III and V). Am. Heart J., 146: 431–438 (2003).

40) Olsson, S. B.: Executive Steering Committee of the SPORTIF IIIInvestigators: Stroke prevention with the oral direct thrombin inhibitorximelagatran compared with warfarin in patients with non-valvular atrialfibrillation (SPORTIF III): randomised controlled trial. Lancet, 362:1691–1698 (2003).

41) Jones, C. A., McQuillan, G. M., Kusek, J. W., Eberhardt, M. S., Herman,W. H., Coresh, J., Salive, M., Jones, C. P. and Agodoa, L. Y.: Serumcreatinine levels in the US population: third National Health andNutrition Examination Survey. Am. J. Kidney Dis., 32: 992–999 (1998).

42) Kubitza, D., Becka, M., Roth, A. and Mueck, W.: Dose-escalation studyof the pharmacokinetics and pharmacodynamics of rivaroxaban inhealthy elderly subjects. Curr. Med. Res. Opin., 24: 2757–2765 (2008).

43) Kubitza, D., Becka, M., Zuehlsdorf, M. and Mueck, W.: Body weighthas limited influence on the safety, tolerability, pharmacokinetics, orpharmacodynamics of rivaroxaban (BAY 59-7939) in healthy subjects.J. Clin. Pharmacol., 47: 218–226 (2007).

44) Kubitza, D., Becka, M., Mueck, W., Halabi, A., Maatouk, H., Klause, N.,Lufft, V., Wand, D. D., Philipp, T. and Bruck, H.: Effects of renalimpairment on the pharmacokinetics, pharmacodynamics and safety ofrivaroxaban—an oral, direct Factor Xa inhibitor. Br. J. Clin. Pharmacol.,70: 703–712 (2010).

45) Zhao, X., Sun, P., Zhou, Y., Liu, Y., Zhang, H., Mueck, W., Kubitza, D.,Bauer, R. J., Zhang, H. and Cui, Y.: Safety, pharmacokinetics andpharmacodynamics of single/multiple doses of the oral, direct Factor Xainhibitor rivaroxaban in healthy Chinese subjects. Br. J. Clin. Pharma-col., 68(1): 77–88 (2009).










http://apps.who.int/bmi/index.jsp

http://apps.who.int/bmi/index.jsp





















model-based dose selection for phase iii rivaroxaban study ... · regular article model-based dose...

Documents