lnternalional Journal of Clinical Pliannacology and Therapeutics, Vol. 48 — No. 5/2010 (309-318)
`
`Thorough QT study of the effect of
`fesoterodine on cardiac repolarization
`
`B. Malhotral, N. Wood2, R. Sachse3 and K. Gandelman‘
`
`’Pfizer lnc, New York, NY, USA, 2Pfizer Global Research and Development,
`Sandwich, Kent, UK and 3Schwarz Biosciences, ll/lonhelm, Germany
`
`
`
`0 rig in al
`@2010 Dustri-Verlag Dr. K. Feistle
`ISSN 0946-1965
`
`Key words
`OAB — fesoterodine —
`cardiovascular safety —
`antimuscarinic
`
`Abstract. Objective: Fesoterodine 4 mg
`and 8 mg once daily are indicated for the treat-
`ment of overactive bladder. A thorough QT
`study was conducted to investigate the effects
`of fesoterodine on cardiac repolarization.
`Materials and methods: In this parallel-group
`study, subjects were randomly assigned to re-
`ceive double-blind fesoterodine 4 mg, feso-
`terodine 28 mg, or placebo or open-label
`moxifloxacin 400 mg (positive control) for 3
`days. Electrocardiograms (ECGs) were ob-
`tained on Days -1 (baseline), 1, and 3. The
`primary analysis was
`the time-averaged
`changes from baseline for Fridericia’s-cor-
`rected QT interval (QTcF) on Day 3. Results:
`Among 261 subjects randomized to feso-
`terodine 4 mg (n = 64), fesoterodine 28 mg
`(n = 68), placebo (11 : 65), or moxifloxacin
`400 mg (n = 64), 256 completed the trial. The
`least squares mean changes in QTcF from
`baseline were 21.1, 20.5, 18.5, and 31.3 ms
`(maximum), and -5.1, -4.2, -5.2, and 7.6 ms
`(time-averaged at Day 3) for placebo, feso-
`terodine 4 mg,
`fesoterodine 28 mg, and
`moxifloxacin, respectively. The lower limit
`of the 95% confidence interval exceeded 5 ms
`for moxifloxacin. Conclusions: The results
`indicate that fesoterodine is not associated
`
`with QTC prolongation or other ECG abnor-
`malities at either therapeutic or suprathera-
`peutic doses.
`
`Introduction
`
`tion of fesoterodine 8 or 12 mg in a fasted
`state; these values are similar in extensive
`
`metabolizers and poor metabolizers [8, 14].
`
`Thorough QT studies have become part of
`the drug development process, as a standard
`regulatory requirement for assessing the ef-
`fect of a drug on cardiac repolarization and
`determining its proarrhythmic potential [3].
`The QT interval prolongation (the time for
`electrophysiologic depolarization and re-
`polarization of ventricles) may be a character-
`istic of torsade de pointes, which has been
`documented to occur beyond threshold limits
`of > 500 ms corrected QT interval (QTc) and
`> 60 ms QTc prolongation and, as such, is a
`risk factor for ventricular arrhythmia and sud-
`den death [1 1]. Cardiac safety concerns due to
`drug-related QT interval prolongation have
`been the reason some drugs have been with-
`drawn from the pharmaceutical market [1 1].
`For example, the antimuscarinic agent tero-
`diline (1991) [4, 7, 13] and the antihistamine
`terfenadine (1998) [5] were removed from
`the market by the United States Food and
`Drug Administration because of associated
`electrocardiographic (ECG) effects, includ-
`ing drug-induced proarrhythmia.
`
`No clinically relevant effects on the QTC
`interval have been previously reported in
`Phase I, II, and III clinical trials of feso-
`
`terodine in healthy volunteers or subjects
`with OAB [2, 12]. Similarly, no clinically rel-
`evant cardiac effects were observed at single
`doses of4 mg, 8 mg, and 12 mg in a pharma-
`cokinetic study in healthy volunteers that
`demonstrated a 2-fold increase in exposure to
`5-HMT in poor metabolizers compared with
`extensive metabolizers [8]. Because these
`Phase 1 studies of fesoterodine were not spe-
`cifically designed to dctcnninc the effect of
`fesoterodine on cardiac electrophysiology
`and QT interval, this thorough QT study was
`
`Received
`
`August 20, 2009;
`accepted
`January 10. 2010
`
`Correspondence to
`B. Malhotra, PhD
`Senior Director
`Clinical Sciences, Pfizer
`Inc, 655 3rd Avenue,
`New York. NY 10017,
`USA
`bima|.k.ma|hotra@
`pfizer.com
`
`Fesoterodine is a new antimuscarinic agent
`that has demonstrated safety and efficacy for
`the treatment of overactive bladder (OAB) [2,
`l0]. Fesoterodine is not detectable in plasma
`following oral administration; it is rapidly and
`extensively converted by nonspecific esterases
`to the active metabolite 5-hydroxymethyl tol-
`terodine (5-HMT) (Figure 1), which is also the
`active metabolite of tolterodine [15]. The
`mean plasma half-life of 5-HMT is approxi-
`mately 8 hours after administration of feso-
`terodine 4 mg and 9 hours after administra-
`
`Patent Owner, UCB Pharma GmbH — Exhibit 2083 - 0001
`
`

`
`Malhotra, Wood, Sachse and Gandelman
`
`310
`
`o
`reso/L?
`
`"\|/
`
`"° 0
`
`TDl[EfDdIl'IE
`
`Because the terminal half—life of 5—HMT
`
`after oral administration is approximately
`8 — 9 hours [9], administering fesoterodine
`once daily for 3 clays is appropriate for evalu-
`ating the QTc effects of fesoterodine at steady
`state. Although both crossover and parallel-
`group study designs are suitable for thorough
`QT evaluations, we chose a parallel—group
`design for this 4—treatment study to minimize
`the anticipated subject drop-out rate and to al-
`low rapid completion of the trial [3]. Overall,
`the duration of dosing and the dosage of
`fesoterodine were suitable to allow character-
`
`ization of QTC effects at 5-HMT concentra-
`
`tions expected in the possible clinical situa-
`tion of a CYPZD6 poor metabolizer receiving
`the highest therapeutic dose of fesoterodine
`along with a potent CYP3A4 inhibitor.
`
`
`
`
`I CYP3A4
`
`5 -HMT
`
`5- CM
`Carboxy rnetabolutt
`
`I CVPJA4
`
`..
`- 3r——..
`O
`N-dnlkylated metabolite
`
`N»dz-allrylaled 5-HMT
`
`N»deIllu_.-lated 5~CM
`
`Figure 1. Fesoterodine (FESO) metabolic path-
`way. CYP = cytochrome P450; 5-HMT = 5-hydroxy-
`methyl tolterodine.
`
`Methods
`
`conducted to investigate the effects of feso-
`terodine on cardiac repolarization.
`
`The study was conducted in accordance
`with the International Conference on Hartno-
`
`nization (lCH)—E14 guidelines [3] for the
`evaluation of QT/QTc interval prolongation
`and proarrhythmic potential for non-anti-
`arrhythmic drugs. The guidelines dictate that
`a negative thorough QT study should demon-
`strate that the mean threshold for QT/QTc
`prolongation remains less than 5 ms and that
`the upper bound of the 95% 1-sided confi-
`dence interval for the largest time-matched
`mean effect on QT should exclude 10 ms.
`This placebo-controlled study was conducted
`at steady state after administration of a thera-
`peutic (4 mg/day) or supratherapeutic (28 mg/
`day) dosage of fesoterodine for 3 days in
`healthy cytochrome P450 (CYP) ZD6 exten-
`sive metabolizers and included moxifloxacin
`
`400 mg as a positive control. The 28—mg/day
`dosage was selected because it was previously
`defined as the maximum tolerated dose of
`
`fesoterodine. Furthermore, plasma exposures
`after a 28—mg/day dosage of fesoterodine are
`appropriate to cover a “worst case scenario” of
`exposure in a CYPZD6 poor metabolizer re-
`ceiving fesoterodine 8 mg/day plus a potent
`CYP3A4 inhibitor, based on the results of a
`previously conducted drug interaction study
`in healthy volunteers [9].
`
`Study design
`
`This was a single—site, randomized, pla-
`cebo- and positive-controlled, parallel-de-
`sign trial with multiple oral dose administra-
`tion of double—blind fesoterodine or placebo
`or open-label moxifloxacin. The treatment
`phase began 3 — 28 days after the eligibility
`assessment and consisted of 3 days of treat-
`ment as follows:
`
`fesoterodine 4-mg sustained release
`— l
`(SR) tablet + 6 placebo tablets matching
`fesoterodine
`
`— 7 fesoterodine 4-mg SR tablets
`— 7 placebo tablets matching fesoterodine
`— l moxifloxacin 400-mg tablet
`
`Moxifloxacin is often used in thorough
`QT studies as a positive control because of its
`well-defined QT/QTc effect (5 — 14 ms) [1 1]
`and, as such, provides context for the sensi-
`tivity of the study to detect small QTC changes
`of around 5 ms [3].
`
`Subjects
`
`The trial protocol, amendments, and sub-
`ject informed consent were reviewed by the
`appropriate Institutional Review Board (In-
`dependent
`Investigational Review Board
`Inc.). Informed consent was obtained from
`each subject and documented according to
`
`Patent Owner, UCB Pharma GmbH — Exhibit 2083 - 0002
`
`

`
`Fesoterodine and cardiac repolarization
`
`311
`
`the current version of the applicable regulatory
`and ICH Good Clinical Practice requirements.
`The study was conducted under in-house condi-
`tions at SFBC lntemational, a clinical research
`organization in Miami, FL, USA.
`Key inclusion criteria were healthy sub-
`jects aged 45 — 65 years with a body mass in-
`dex between 19 and 32 kg/m2 (inclusive). Eli-
`gible subjects also had no clinically relevant
`abnormal findings on the physical examina-
`tion, ECG, blood pressure, pulse rate, medical
`history, or clinical laboratory results at the eli-
`gibility assessment visit and were character-
`ized as extensive metabolizers for CYPZD6.
`
`Key exclusion criteria were medical history
`of any serious disease ofthe internal organs or
`of the central nervous system; a history or
`presence of urinary retention, obstructive dis-
`turbance of bladder emptying, micturition
`disturbance, nocturia, or pollakiuria, for ex-
`ample, prostatic hyperplasia, or urethral stric-
`ture; a history of ischemic heart disease or a
`positive diagnostic cardiac stress test within
`12 weeks before the start of the trial; a supine
`pulse rate of< 50 bpm or > 100 bpm, supine
`systolic blood pressure of < 100 mmHg or
`> 160 mmHg, or a supine diastolic blood
`pressure of > 95 mmHg; and any clinically
`relevant changes in ECG, such as second- or
`third-degree AV block, or prolongation of the
`QRS interval to > 110 ms, the PR interval to
`> 240 ms, or QTc (Bazett’s correction, ma-
`chine read) to > 480 ms.
`
`ECG assessments
`
`Electrocardiograms were obtained digi-
`tally using a Mortara Instrument H—l2 ECG
`continuous recorder. The ECGs were stored
`
`on a flashcard approximately every 10 sec-
`onds and were not available for review until
`
`the card was received by the central ECG lab-
`oratory and analyzed. Electrocardiographic
`interval and morphology changes were based
`on change from baseline, where baseline was
`the mean of the 36 recordings obtained on
`Day -1. Triplicate 12-lead ECGS (within 1
`minute) were obtained and averaged for each
`ofthe following time points: at 1, 2, 3, 4, 6, 8,
`10, 12, 14, 16, 18, and 23.5 hours postdose on
`Days 1 and 3 and at matching time points for
`baseline on Day -1, resulting in a total of 108
`ECGs per subject. For practical purposes, it
`
`was necessary to have the last ECG collected
`at 23.5 rather than 24 hours postdose.
`
`0 Tc interval analysis of ECGs
`
`Centrally read ECG data were used to de-
`termine: (l) central tendency of QTC changes
`using maximum,
`time-averaged, and time-
`matched changes from baseline; and (2) outlier
`analyses. In addition, correlations of Frideri-
`ca’s-corrected (QTcF), Bazett’s-corrected QT
`interval (QTCB), and individually dCt6I1111I16Cl
`QT correction (QTCI) values with 5-HMT
`plasma concentrations were assessed.
`The primary correction formula was
`QTCF, the length of the QT interval corrected
`for heart rate by Fridericia’s formula: QTcF =
`QT/(RR)"3. Additional correction formulas
`that were included but considered secondary
`were QTCB and QTcl. The additional cor-
`rected QT interval QTcB is defined as the
`length of the QT interval corrected for heart
`rate by Bazett’s formula: QTcB = QT/(RR)"-7-.
`The corrected QT interval (QTCI) was deter-
`mined for each subject by iterating the QT-RR
`relationship using the 36 baseline ECGs
`(Day -1) to find an estimate for the exponent
`such that the slope of this relationship was
`closest to 0 or other appropriate method. The
`QTcI is the individually determined QT cor-
`rection, and the goal was to find [3 such that
`QTcI is a constant, where QTcI = QT/(RR)l5.
`This implies log(QTcl) = log(QT) — [3 X
`log(RR). Because log(QTcl) is a constant,
`this equation can be rewritten as log(QT) = oi
`+ [3 X log(RR). Therefore, the exponent esti-
`mate can be obtained by numerical iteration
`such that slope for QT-RR relationship is
`closest to 0 or using regression analysis on
`log—transformed data based on the least
`squares (LS) approach.
`
`Pharmacokinetic assessments
`
`Venous blood samples were drawn from
`all subjects at the following time points on
`Days 1 and 3: predose (Day 1) and 1, 2, 3, 4, 6,
`8, 12, and 23.5 hours postdose. Plasma was
`separated from the blood samples collected
`after
`fesoterodine treatments and stored
`
`the samples were sent to the
`frozen until
`bioanalytical unit of Schwarz BioSciences,
`Monheim, Germany. Plasma samples were
`
`Patent Owner, UCB Pharma GmbH — Exhibit 2083 - 0003
`
`

`
`Malhotra, Wood, Sachse and Gandelman
`
`312
`
`assayed for 5—HMT using a validated liquid
`chromatography tandem mass spectrometry
`method with a lower limit of quantification of
`0.04 ng/ml.
`The pharmacokinetic variables assessed
`were area under the plasma concentration
`versus time curve from 0 to the last measured
`
`concentration greater than the lower limit of
`quantification during a dose interval at
`steady-state conditions (AUC04); maximum
`drug concentration in plasma during a dose
`interval at steady state (Cmax); time of ob-
`served Cmax (tmax); and apparent total body
`clearance of drug.
`
`Safety
`
`Safety was monitored by assessing ad-
`verse events, vital signs, ECG recordings, and
`laboratory tests.
`
`Statistical methods
`
`Analyses were conducted in the pharma-
`codynamics population set, which included
`all randomized subjects who received at least
`one dose of trial medication and had a suffi-
`cient number of ECG assessments to calcu-
`
`the pharmaco-
`
`late reliable estimates for
`dynamic parameters.
`For analyses of central tendency of QTcF
`changes (the primary variable), three meth-
`ods were explored: maximum,
`time-aver-
`aged, and time—matched QTcF changes from
`baseline. The primary method was a time-av-
`eraged analysis on Day 3. The secondary
`methods were placebo-subtracted maximum
`and time—matched changes from baseline for
`QTCF on Day 3. For time-averaged analyses,
`the baseline corresponded to the mean of the
`36 recordings obtained on Day -1, and for
`time—matched analyses the baseline corre-
`sponded to the mean of the three recordings
`obtained at the same time point on Day -1.
`In the time-averaged analysis, all values
`within each sampling day were averaged to
`obtain a single value. This analysis averaged
`over any possible circadian rhythms. The
`mean time-averaged value was summarized
`by day and treatment group with descriptive
`statistics. For Days 1 and 3, change from
`baseline was summarized in a similar manner.
`
`For each treatment group, a 2-sided 95% con-
`
`fidence interval for the change from time-av-
`eraged baseline value on Days 1 and 3 was
`presented.
`The second method for analysis was a
`time—matched analysis for which the mean
`value at each sampling time point (mean of
`three ECGS at that time point; there were 12
`time points per day) was summarized by time
`point for baseline, Day 1, and Day 3 by treat-
`ment group with descriptive statistics. For
`Days 1 and 3, the time-matched change from
`baseline values was summarized by treatment
`group with descriptive statistics. In the time-
`matched change from baseline, each post-
`baseline value was compared with the corre-
`sponding value at baseline for that specific
`time point during the day.
`The final method involved examining the
`maximum change from baseline. The maxi-
`mum change from baseline was defined as the
`maximum time—matched change from baseline
`observed for each subject across Days 1 and 3.
`The maximum change from baseline in QTcF
`was summarized by treatment group with de-
`scriptive statistics. A 2-sided 95% confidence
`interval for maximum change from baseline
`was presented for each treatment group.
`Time-averaged change from baseline and
`maximum change from baseline were ana-
`lyzed in analysis of covariance (ANCOVA)
`models with gender and treatment group as
`factors, and the corresponding mean baseline
`value of ECG as a covariate. These analyses
`were exploratory, and each pairwise compar-
`ison of factor levels was performed at the 5%
`level of significance without multiplicity ad-
`justment. 95% confidence intervals were de-
`rived for all pair wise mean differences be-
`tween treatments (i.e., comparing each active
`treatment vs. placebo, the two fesoterodine
`dose groups vs. moxifloxacin, and the higher
`dose vs. lower dose for fesoterodine) using
`the LS means and corresponding residual er-
`ror from the ANCOVA model. 90% confi-
`
`dence intervals (equivalent to 95% l-sided
`confidence intervals) were also presented for
`the pair wise differences between treatments.
`For the outlier analysis, the number and
`percentage of subjects who had an absolute
`QTcF of< 450, 450 to < 480, 480 to < 500,
`and 2 500 ms and the number and percentage
`of subjects who had an increase from baseline
`QTcF of< 30, 30 to < 60, and Z 60 ms were
`summarized by day, time point, and treatment
`
`Patent Owner, UCB Pharma GmbH — Exhibit 2083 - 0004
`
`

`
`Fesoterodine and cardiac repolarization
`
`313
`
`group. This analysis was repeated by gender.
`The proportion of subjects meeting outlier
`criteria for each treatment group was com-
`puted using the longest of the triplicate ECG
`intervals at a given time point after treatment,
`compared with baseline values.
`Although QTcF was the primary correc-
`tion method for this trial, all analyses ofQTcF
`were repeated for the secondary variables
`(QTcl, QTcB, uncorrected QT, heart rate, PR
`interval, QRS interval). All p values associ-
`ated with QTcl are considered exploratory.
`Because the lCH—El4 guidelines state that
`QTCB overcorrects at elevated heart rates and
`undercorrects at heart rates below 60 bpm, it
`is not an ideal correction for assessing the ef-
`fects of fesoterodine on cardiac repolariza—
`tion. Therefore, because Fridericia’s correc-
`tion is more accurate than Bazett’s correction
`
`in subjects with such altered heart rates, only
`QTcF and QTcl results are presented.
`Descriptive statistics were conducted for
`secondary safety endpoints,
`including ad-
`verse events, changes in vital signs, physical
`examination,
`l2—lead ECG parameters, he-
`matology, and serum chemistry parameters.
`
`Sample size
`
`The sample size chosen for this trial was
`based on precedents set by similar ECG
`safety studies. From published trials,
`it
`is
`known that moxifloxacin induces a QTc pro-
`longation of approximately 5 — 10 ms; the de-
`tectable difference between active treatment
`
`and placebo is usually defined as 5 ms. The
`standard deviation in the published moxi-
`floxacin trials varied substantially and was
`approximately l2 — I3 ms on average. Repli-
`cate ECG measurements were planned for
`this trial, and this was expected to decrease
`the standard deviation to approximately 10
`ms. Setting the significance level to or = 5%
`(2-sided), and requiring a power of 80%, it
`was estimated that 64 evaluable subjects were
`needed per treatment group, resulting in a
`total of 256 subjects.
`
`Pharmacokinetic analyses
`
`analyses were per-
`Pharmacokinetic
`formed on the pharmacokinetic set, which in-
`cluded all randomized subjects who had re-
`
`ceived at least one dose of trial medication,
`had at least one pharmacokinetic measure-
`ment, and had no major protocol deviation
`that would have rendered the pharmaco-
`kinetic data unreliable. Analysis of variance
`with confidence intervals was perfonned for
`all pharmacokinetic endpoints except t,,,,,,,, for
`which a nonparametric test was performed
`among subjects in the two fesoterodine treat-
`ment groups.
`
`Results
`
`Subjects
`
`Subjects’ baseline demographics were simi-
`lar across groups (Table 1). Most subjects in
`each group were Hispanic, and the average age
`was approximately 53 years. Of the 261 sub-
`jects randomized to fesoterodine 4 mg (n = 64),
`fesoterodine 28 mg (n : 68), placebo (11 : 65) or
`moxifloxacin ( n = 64), 256 (98%) completed
`the trial. Two subjects in the fesoterodine
`28-mg group discontinued because of ad-
`verse events; 2 subjects in the fesoterodine
`28-mg group withdrew consent; and 1 subject
`in the placebo group discontinued, stating
`“other reasons."
`
`Electrocardiographic parameters
`
`At all time points during the trial, most
`subjects had normal findings for ECG param-
`eters, and none of the abnormal findings were
`considered clinically relevant. Time—aver—
`aged by day and maximum changes from
`baseline in QTcF are shown in Table 2. There
`were no significant differences in the model-
`based adjusted mean change from baseline in
`QTcF on Days 1 or 3 between the feso-
`terodine 4-mg (4.6 and4.2, respectively) or
`fesoterodine 28—mg
`(-7.3 and -5.2,
`re-
`spectively) and placebo (-4.6 and -5.1, re-
`spectively) groups. The mean changes from
`baseline in QTcF after treatment with moxi-
`floxacin (4.0 on Day 1 and 7.6 on Day 3) were
`significantly greater than those in the 3 other
`treatment groups on Days 1 and 3 (p < 0.001).
`The mean placebo-subtracted change from
`baseline in QTcF on Day 3 was approxi-
`mately l0 ms (maximum) and 13 ms (time-
`averaged), and the lower limit of the 95%
`
`Patent Owner, UCB Pharma GmbH — Exhibit 2083 - 0005
`
`

`
`Malhotra, Wood, Sachse and Gandelman
`
`314
`
`Table 1. Baseline demographics.
`
`(n = 55)
`
`(n = 54)
`
`(n = 53)
`
`(n = 54)
`
`T3
`
`27.0 1 3.24
`
`Black
`
`Hispanic
`
`Latino
`
`1 (1.5)
`
`53 (95.9)
`
`0
`
`W5)
`4 (6.3)
`
`59 (92.2)
`
`0
`
`45-9)
`1 (1.5)
`
`52 (91.2)
`
`1 (1.5)
`
`2 (3.1)
`
`59 (92.2)
`
`1 (1.5)
`
`BMI, kglmz (mean 1 so)
`
`25.5 1 2.55
`
`25.7 1 3.2
`
`25.2 1 3.05
`
`BMI = body mass index, FESO = fesoterodine, SD = standard deviation.
`
`Table 2. Summary of time-averaged QTcF by day and maximum changes from baseline in QTcF.
`
`LS mean
`
`difference
`
`Day 1
`
`3-2“
`FESO 4 mg
`4 mg vs. PBO n 0.975
`-2.3, 2.4
`7.3
`FESO 25 mgn 25 mg vs. PBO
`0.022
`-5.0, -0.4
`MOXI
`64
`4.0
`MOXI vs. PBO
`< 0.001
`6.3, 11.0
`
`Day 3
`
`-2“
`
`K 4
`
`7.6
`
`MOXI vs. PBO
`
`12.7
`
`< 0.001
`
`10.6. 14.8
`
`IVIOXI
`
`6
`
`-2“
`FESO4mg
`4mg vs.PBO
`0.749
`-3.5, 2.8
`FESO2B mgn 25 mg vs.PBO
`0.124
`—5.3,0.7
`
`CI = confidence interval. FESO = fesoterodine, L8 = least squares. MOXI = moxifloxacin, PBO = placebo.
`
`confidence interval exceeded 5 ms, conf'1rm— time-averaged ECG parameters showed no
`ing the sensitivity of the study to detect a
`notable differencesbetween treatment groups
`mean change of approximately 5 ms.
`in the absolute values and model—based ad-
`The m0del—based adjusted mean placebo~
`justed mean changes from baseline on Day 1
`subtracted change from baseline in QTCF was
`Of Day 3 in the PR iml-31”V-‘=11 01” the QRS dura-
`close to 0 ms, and the upper limit of95% con-
`lion.
`fidence interval was well below 10 ms for
`Overall, there wasanegative model-based
`both the therapeutic and suprathcrapcutic
`adjusted mean time—matchcd change from
`doses of fesoterodine. Similar results were
`baseline QTCF following dosing with either
`observed for QTCI (Table 3). Results of the
`fesoterodine dose on Days 1 and 3 (Figure 2).
`
`Patent Owner, UCB Pharma GmbH — Exhibit 2083 - 0006
`
`

`
`Fesoterodine and cardiac repolarization
`
`315
`
`Table 3. Summary oftime-averaged QTcl by day and maximum change from baseline.
`
`Treatment
`
`Day 1
`
`Day 3
`
`FESO 4 mg
`
`FESO 23 mg
`
`Endpoint
`LS mean
`
`Comparison
`
`Treatment
`difference
`
`p value
`
`95% CI
`
`4
`
`4 mg vs. PBO
`E 23 mg vs. PBO
`MOXI vs. PBO
`
`-1.1
`-5.0
`8.1
`
`0.435
`0.002
`< 0.001
`
`-4.4, 2.1
`-8.2, -1.3
`4.9, 11.4
`
`4 mg vs. PBO
`
`28 mg vs. PBO
`
`64
`
`MOX|vs.PBO
`
`12.5
`
`<0.001 M
`
`PBO
`
`65 --2
`54
`4mg vs.PBO
`-0.5
`0.309
`—4.9,3.s
`
`CI = confidence interval, FESO = fesoterodine, L8 = least squares, MOXI = moxifloxacin, PBO = placebo.
`
`with placebo treatment. Assay sensitivity was
`shown by an increase from baseline in QTcF
`following treatment with moxifloxacin.
`
`Results of the subjects with a new onset
`QT or QTC outlier value were consistent with
`the absence of any QT prolongation effect as-
`sociated with fesoterodine treatment (Table
`4). No notable differences were seen in the
`number of QTCF outliers between placebo
`and either of the fesoterodine treatment
`
`groups. In contrast, there was a higher inci-
`dence of QTcF outliers following treatment
`with moxifloxacin. No subject had a new on-
`set QTcF value > 480 ms. The percentages of
`subjects with new onset values for QTCF of
`> 450 ms were 4.6% (placebo), 3.1% (feso-
`terodine 4 mg), 0% (fesoterodine 28 mg), and
`10.9% (moxifloxacin). The new onset values
`in QTcF that were > 450 ms in the feso-
`terodine 4-mg group represented one occur-
`rence each in 2 subjects. One subject in the
`moxifloxacin treatment group had a change
`from baseline of 2 60 ms in the QTcF interval
`at Day 3; no other subject had a change from
`baseline in QTcF that was 2 60 ms at any time
`point during the trial. The percentage of sub-
`jccts with increases in QTCF that were 30 ~ 60
`ms was higher in the rnoxifloxacin group
`(50%) compared with the placebo (15.4%),
`
`a 9IDII12l)ll1Sl6l?l8l9202I221324
`TlmI,lI
`—D— hwlrnu in-Ml
`- Q - Madllaumtn-M:
`
`—l.— Ha:uho|n—o5l
`-0- rnonrnq In-ail:
`
`D 5
`
`Time-Matched,ChangeFromBaselineInQTcF,ms
`
`
`ChanglFromBanlincinQTCF,ms
`Time-Matchod
`
`8c
`
`uv-
`
`u'nau-S533
`
`S
`
` 7
` II9IDH111!lII5l6'l?|5l9N2l2Z1!2-I
`
`2:J:
`
`—O-Hxzhnln-fill
`“-0- iuollmgln-61)
`
`Tlm|,h
`—u—mn4mq1n—a-u
`- 9 -Ilnlillouolrtln-Ml
`
`Figure 2. Time-matched change from baseline in
`QTCF on (A) Day 1 and (B) Day 3. Feso = fesotero-
`dine.
`
`Similar results were observed for QTcI. The
`magnitude of the decrease seen following
`treatment with fesoterodine was comparable
`
`Patent Owner, UCB Pharma GmbH — Exhibit 2083 - 0007
`
`

`
`Malhotra, Wood, Sachse and Gandelman
`
`316
`
`Table 4. Summary of subjects with new onset QTc outlier values.
`
`Number of subjects (%J
`
`Placebo
`
`(n = 65)
`
`3 (4.6)
`
`FESO 4 mg
`(n = 54)
`
`2(3.1)
`
`FESO 23 mg
`(n=68)
`
`0
`
`MOXI
`
`(n = 64)
`
`7(10.9)
`
`Increase of 30 to < 60 ms
`Increase of 2 60 ms
`
`10 (15.4)
`0
`
`12(18.B)
`0
`
`12(17.6)
`0
`
`QTc|
`
`> 450 ms
`
`> 480 ms
`
`4 (6.2)
`
`2 (3.1)
`
`5 (7.3)
`
`1 (1.6)
`
`2 (2.9)
`
`0
`
` 3 (12.5)
`
`32 (50.0)
`1 (1.6)
`
`3(12.5)
`
`
`Increase of 30 to < 60 ms
`7 (10.8)
`11 (16.2)
`Increase of 2 60 ms
`0
`1 (1.5)
`
`1 (1.6)
`
`31 (43.4)
`1 (1.6)
`
`FESO = fesoterodine, MOXI = moxifloxacin.
`
`Table 5. Summary statistics of pharmacokinetic parameters of 5-HMT.
`
`Fesoterndine 4 mg
`
`Fesoterodina 28 mg
`
`Day 1
`(n = 64)
`
`Day 3
`(n = 64)
`
`Day 1
`(n = 67)
`
`Day 3
`(0 = 64)
`
`Mean (CV)
`
`132.2 (202.6)
`
`Auc0_.
`
`ng x hfrnl
`
`24.4 (41.3)
`
`23.6 (33.0)
`
`212.4 (40.3)
`
`242.5 (44.6)
`
`2.33 (33.1)
`E 3.4 (53.1)
`212.3 (74.7)
`
`2.66 (33.3)
`3.4 (40.2)
`177.1 (34.1)
`
`13.1 (40.9)
`4.1 (36.1)
`155.2 (44.4)
`
`20.7 (40.0)
`4.1 (36.7)
`
`AUC0_, = area underthe concentration vs. time curve from time 0 to time of last measurable concentration.
`CLIF = apparent oral clearance, Cmax = maximum plasma concentration, CV = coefficient of variation. tmax
`= time to reach Cmax.
`
`fesoterodine 4—mg (18.8%), and fesoterodine
`28—mg (17.6%) groups.
`
`Pharmacokinetic parameters
`
`The values for the pharmacokinetic pa-
`rameters Cmax and AUC0_, show the expected
`increase when comparing Day 1 and Day 3
`data for each dose level (Table 5). Based on
`visual examination of the plots of 5-HMT
`trough concentrations on Days 1, 2 and 3, and
`consistent with the terminal half life of about
`
`8 hours [8], it was apparent that steady state was
`achieved on Day 3. A 7-fold increase in the dose
`(from 4 to 28 mg) resulted in a similar 7-fold in-
`
`crease in the pharmacokinetic parameters Cm‘
`and AUC.(-H, which is consistent with dose-pro-
`portional phatmacokinetics of fesoterodine.
`
`The relationship between model-based
`adjusted mean change in QTcF and mean
`plasma concentration of 5-HMT on Day 3 is
`shown for fesoterodine 4 mg (Figure 3A) and
`28 mg (Figure 3B). The plasma concentration
`of 5-HMT shows the expected pharmaco-
`kinetic profile change over time, whereas the
`change in QTCF over time is relatively stable,
`suggesting that there is no correlation be-
`tween the QTcF interval and plasma concen-
`tration of 5-HMT. Similar findings were ob-
`served between QTCI values and 5-HMT
`plasma concentrations.
`
`Patent Owner, UCB Pharma GmbH — Exhibit 2083 - 0008
`
`

`
`Fesoterodine and cardiac repolarization
`
`317
`
`CInngoi'IQl'cF,I-ns)>
`
`ossssesas
`
`
`DD
`
`0.!
`
`ID
`
`I5
`
`1D
`
`1.5
`
`In
`
`5-5
`
`ID
`
`4.5
`
`SD
`
`5.5
`
`6D
`
`related to trial medication, and pharyngitis
`was judged by the investigator to be not re-
`lated to trial medication. Both events resolved
`
`in 5 days. A 59-year-old man withdrew from
`the trial because of urinary retention requir-
`ing catheterization reported on Day 1. The
`event was severe and judged by the investiga-
`tor to be probably related to trial medication.
`The event resolved in 3 days.
`There were no abnormal clinical labora-
`
`tory findings or changes in vital signs that
`were determined to be clinically significant in
`this subject population. No clinically impor-
`tant changes from baseline were apparent at
`any time point for systolic or diastolic blood
`pressure. The mean placebo-corrected in-
`crease in heart rate associated with a dosage
`of 4 mg/day and 28 mg/day of fesoterodine
`was 3 bpm and 11 bpm, respectively.
`
`Discussion
`
`This trial demonstrated that fesoterodine
`
`does not affect the QTc interval in healthy
`subjects aged 45 — 65 years after a therapeutic
`(4 mg) or a supratherapeutic dose (28 mg) of
`fesoterodine. The ICH-E14 guidelines dictate
`that negative thorough QT studies should
`demonstrate that
`the mean threshold for
`
`QT/QTc prolongation remain less than 5 ms
`and the upper bound of the 95% 1-sided con-
`fidence interval for the largest time—matched
`mean effect on QT exclude 10 ms. This was
`the case for fesoterodine. The time-matched
`
`results and time-averaged changes in QTcF
`interval from baseline, the primary endpoint,
`did not show an increase following feso-
`terodine 4 and 28 mg doses relative to pla-
`cebo. This was also observed with the QTcl
`values; there were no significant differences
`between the fesoterodine treatment groups
`and placebo. No correlation was evident be-
`tween 5-HMT plasma levels and the cor-
`rected QTCF interval as a marker for myo-
`cardial repolarization. The results observed
`with moxifloxacin were consistent with those
`
`in previous studies using moxifloxacin as a
`positive control [1 l] and confirm the sensitiv-
`ity of this study to determine QT effects.
`According to the ICH-E14 regulatory guid-
`ance [3], applying the most accurate correction
`factor available is important in a thorough QT
`study. Considering that QTcB overcorrects at
`
`ChangeI1OTCF.msWbééééééaasusassaa
`
`
`
`Plum: Concentration ol 5|-|M‘|'. ngiml
`
`Figure 3. Correlation between change in QTCF
`and plasma concentration of 5-hydroxymethyl
`tolterodine (5-HMT) with (A) fesoterodine 4 mg and
`(B) fesoterodine 28 mg on Day 3. Data representthe
`pharmacodynamic and pharmacokinetic sets.
`
`Safety
`
`Fesoterodine was well tolerated. Subjects
`receiving fesoterodine 28 mg recorded a
`greater
`frequency of
`treatment-emergent
`treatment-related adverse events than did
`
`subjects receiving fesoterodine 4 mg, pla-
`cebo, or moxifloxacin, respectively, includ-
`ing abdominal pain (7.4% vs. 0% vs. 0% vs.
`0%). constipation (5.9% vs. 0% vs. 0% vs.
`0%). dry mouth (4.4% vs. 0% vs. 0% vs. 0%),
`vomiting (4.4% vs. 0% vs. 0% vs. 1.6%), uri-
`nary retention (4.4% vs. 0% vs. 0% vs. 0%),
`and pharyngitis (1.5% vs. 0% vs. 0% vs. 0%)
`but not headache (5.9% vs. 6.3% vs. 6.2% vs.
`4.7%). Adverse events experienced by sub-
`jects in the fesoterodine treatment groups, ex-
`cept for one case ofconj unctival hemorrhage.
`were consistent with those seen in other trials
`
`and were expected for an antimuscarinic
`drug.
`Two subjects. both in the fesoterodine
`28-mg group. withdrew from the trial because
`of treatment-emergent adverse events. A 65-
`year-old woman discontinued because of dry
`mouth and pharyngitis reported on Day 1.
`Both events were mild in intensity; dry mouth
`was judged by the investigator to be possibly
`
`Patent Owner, UCB Pharma GmbH — Exhibit 2083 - 0009
`
`

`
`Malhotra, Wood, Sachse and Gandelman
`
`318
`
`USA. Nolan Wood was an employee ofPfizer
`Inc, Sandwich, Kent, UK, at the time this
`
`study was conducted. Richard Sachse is an
`employee of Schwarz BioSciences, Mon-
`heim, Germany.
`
`References
`
`elevated heart rates and undercorrects at heart
`
`rates < 60 bpm, it is not an ideal correction
`factor [6]. The QTCI is considered suitable for
`a thorough QT/QTc study. In this trial, the pri-
`mary analysis was based on the QTcF because
`it is more accurate than QTCB in subjects with
`altered heart rate following treatment with
`antimuscarinic drugs [3].
`The safety profile in this population of
`healthy volunteers was consistent with other
`Phase I, II, and III trials [1, 2, 10]. Adverse
`events (dry mouth and headache) were ob-
`served more frequently after administration
`of fesoterodine 28 mg. In this population of
`healthy adult volunteers, changes in vital
`signs were not considered clinically relevant.
`As expected, there was a modest dose—related
`increase in heart rate on Days 1 and 3 for both
`the fesoterodine 4-mg and 28-mg treatment
`groups, and the effect was more pronounced
`with fesoterodine 28 mg; the ability to in-
`crease heart rate is a known effect of anti-
`
`muscarinic drugs.
`In concl