Download Extract from Clinical Evaluation Report: Fentanyl citrate

AusPAR Attachment 2
Extract from the Clinical Evaluation
Report for fentanyl citrate
Proprietary Product Name: Abstral
Sponsor: A. Menarini Australia Pty Ltd
Date of CER: 24 July 2012
Therapeutic Goods Administration
About the Therapeutic Goods Administration (TGA)

The Therapeutic Goods Administration (TGA) is part of the Australian Government
Department of Health and is responsible for regulating medicines and medical devices.

The TGA administers the Therapeutic Goods Act 1989 (the Act), applying a risk
management approach designed to ensure therapeutic goods supplied in Australia
meet acceptable standards of quality, safety and efficacy (performance), when
necessary.

The work of the TGA is based on applying scientific and clinical expertise to decisionmaking, to ensure that the benefits to consumers outweigh any risks associated with
the use of medicines and medical devices.

The TGA relies on the public, healthcare professionals and industry to report problems
with medicines or medical devices. TGA investigates reports received by it to
determine any necessary regulatory action.

To report a problem with a medicine or medical device, please see the information on
the TGA website <http://www.tga.gov.au>.
About the Extract from the Clinical Evaluation Report

This document provides a more detailed evaluation of the clinical findings, extracted
from the Clinical Evaluation Report (CER) prepared by the TGA. This extract does not
include sections from the CER regarding product documentation or post market
activities.

The words [Information redacted], where they appear in this document, indicate that
confidential information has been deleted.

For the most recent Product Information (PI), please refer to the TGA website
<http://www.tga.gov.au/hp/information-medicines-pi.htm>.
Copyright
© Commonwealth of Australia 2013
This work is copyright. You may reproduce the whole or part of this work in unaltered form for your own personal
use or, if you are part of an organisation, for internal use within your organisation, but only if you or your
organisation do not use the reproduction for any commercial purpose and retain this copyright notice and all
disclaimer notices as part of that reproduction. Apart from rights to use as permitted by the Copyright Act 1968 or
allowed by this copyright notice, all other rights are reserved and you are not allowed to reproduce the whole or any
part of this work in any way (electronic or otherwise) without first being given specific written permission from the
Commonwealth to do so. Requests and inquiries concerning reproduction and rights are to be sent to the TGA
Copyright Officer, Therapeutic Goods Administration, PO Box 100, Woden ACT 2606 or emailed to
<[email protected]>.
Submission PM-2011-03151-3-1 Extract from the Clinical Evaluation Report for fentanyl citrate (Abstral)
Page 2 of 56
Therapeutic Goods Administration
Contents
List of abbreviations ________________________________________________________ 4
1.
Clinical rationale _______________________________________________________ 6
2.
Contents of the clinical dossier _______________________________________ 7
3.
4.
2.1.
Scope of the clinical dossier _________________________________________________ 7
2.2.
Paediatric data _______________________________________________________________ 7
2.3.
Good clinical practice ________________________________________________________ 7
Pharmacokinetics ______________________________________________________ 8
3.1.
Summary of pharmacokinetics _____________________________________________ 8
3.2.
Evaluator’s overall conclusions on pharmacokinetics __________________ 21
Pharmacodynamics ___________________________________________________ 21
4.1.
Studies providing pharmacodynamic data ______________________________ 21
4.2.
Summary of pharmacodynamics _________________________________________ 21
4.3.
Secondary pharmacodynamics ___________________________________________ 22
4.4.
Evaluator’s overall conclusions on pharmacodynamics ________________ 22
5.
Dosage selection for the pivotal studies ___________________________ 22
6.
Clinical efficacy ________________________________________________________ 23
6.1.
7.
Relief of breakthrough cancer pain ______________________________________ 23
Clinical safety __________________________________________________________ 40
7.1.
Studies providing evaluable safety data _________________________________ 40
7.2.
Pivotal studies that assessed safety as a primary outcome ____________ 41
7.3.
Patient exposure ___________________________________________________________ 41
7.4.
Adverse events _____________________________________________________________ 42
7.5.
Laboratory tests ___________________________________________________________ 52
7.6.
Post-marketing experience _______________________________________________ 53
7.7.
Evaluator’s overall conclusions on clinical safety _______________________ 53
8.
First round benefit-risk assessment ________________________________ 54
9.
First round recommendation regarding authorisation __________ 54
10.
Clinical questions ___________________________________________________ 54
11.
References ___________________________________________________________ 55
Submission PM-2011-03151-3-1 Extract from the Clinical Evaluation Report for fentanyl citrate (Abstral)
Page 3 of 56
Therapeutic Goods Administration
List of abbreviations
Abbreviation
Meaning
AE
adverse event
ALT
alanine transaminase
ANOVA
analysis of variance
AUC
area under the curve
AUC0-4h
area under the plasma concentration-time curve from 0 to 4 hours after the
dose
AUC0-6h
area under the plasma concentration-time curve from 0 to 6 hours after the
dose
AUC0-t
area under the plasma concentration-time curve from 0 to the last
quantifiable observation
AUC0-∞
area under the plasma concentration-time curve, extrapolated to infinity
AUCτ
area under the plasma concentration-time curve during the dose interval
BPI
Brief Pain Inventory
BTcP
breakthrough cancer pain
CI
confidence interval
CL
clearance
CL/F
clearance/fraction absorbed
Cmax
maximum plasma concentration
Ctrough
trough plasma concentration (at the end of the dosing interval)
CMI
Consumer Medicine Information
CNS
central nervous system
CYP3A4
cytochrome P-450, 3A4 isoenzyme
DAPOS
Depression, Anxiety, and Positive Outlook Scale
ECG
electrocardiograph
F
fraction absorbed
FDA
[United States] Food and Drug Administration
FEV1
forced expiratory volume over 1 second
Submission PM-2011-03151-3-1 Extract from the Clinical Evaluation Report for fentanyl citrate (Abstral)
Page 4 of 56
Therapeutic Goods Administration
Abbreviation
Meaning
ITT
intent-to-treat
IU
international unit(s)
IV
intravenous
LC-MS/MS
liquid chromatography - tandem mass spectrometry
LOQ
[lower] limit of quantification
LOCF
last observation carried forward
LS mean
least squares mean
MAO
monoamine oxidase
NNH
number-needed-to-harm
NNT
number-needed-to-treat
OR
odds ratio
PD
pharmacodynamic(s)
PI
Product Information
PID
pain intensity difference
PEFR
peak expiratory flow rate
PK
pharmacokinetic(s)
PP
per-protocol
PR
pain relief
QoL
quality of life
RMP
Risk Management Plan
Rsq
Residual sum of squares
SAE
serious adverse event
SD
standard deviation
SE
standard error (of the mean)
SOC
System Organ Class
SPID
sum of pain intensity difference
Submission PM-2011-03151-3-1 Extract from the Clinical Evaluation Report for fentanyl citrate (Abstral)
Page 5 of 56
Therapeutic Goods Administration
Abbreviation
Meaning
ss
(after a PK parameter)at steady state
t1/2
terminal elimination half-life
TEAE
treatment-emergent adverse event
Tmax
Time (after the dose) of the maximum plasma concentration
TOTPAR
total pain relief
V
volume of distribution
1. Clinical rationale
Among patients with cancer there is a high incidence of breakthrough cancer pain (BTcP), ie.
transient exacerbations of otherwise controlled chronic cancer pain. Treatment guidelines for
cancer pain recognise the need for an agent with appropriate potency, rapid onset and short
duration of effect, which can be tailored (by dose titration) to address the individual
requirements of patients suffering from BTcP.
Oral transmucosal fentanyl citrate has been identified as an agent that fulfils these criteria. One
product – in the form of a lozenge on a stick with the trade name Actiq – is currently registered
in Australia. The following PK information for Actiq is derived from the Australian PI (30
January 2012):

The absorption pharmacokinetics of fentanyl from Actiq are a combination of rapid
oromucosal absorption (bypassing first-pass metabolism) and slower gastrointestinal
absorption of swallowed fentanyl (subject to first-pass metabolism).

Approximately 25% of the total dose of Actiq is rapidly absorbed from the buccal mucosa.
The remaining 75% of the dose is swallowed and slowly absorbed from the gastrointestinal
tract. About 1/3 of this amount (25% of the total dose) escapes hepatic and intestinal firstpass elimination and becomes systemically available. The absolute bioavailability of Actiq is
thus about 50%.

The maximum plasma concentration of fentanyl after Actiq administration occurs after
about 20 to 40 minutes.

Once in the blood, fentanyl rapidly crosses the blood-brain barrier. Redistribution into
tissues and hepatic metabolism provide a duration of action that is suitable for the
treatment of BTcP.
One disadvantage of Actiq is that to maximise the speed and extent of absorption of fentanyl
across the oral mucosa, the patient must move the lozenge around the mouth, using the
attached stick, over a period about 15 minutes. Differences in the amount of movement between
successive uses of the product can lead to inconsistency in the speed and extent of fentanyl
absorption. In addition, moving the lozenge around the mouth would tend to encourage the
swallowing of saliva (containing fentanyl), diverting fentanyl from transmucosal to
gastrointestinal absorption. A rapidly-disintegrating sublingual tablet - such as Abstral - would
potentially be more convenient to use and also has the potential to provide a more consistent
dosage as well as improved bioavailability.
Submission PM-2011-03151-3-1 Extract from the Clinical Evaluation Report for fentanyl citrate (Abstral)
Page 6 of 56
Therapeutic Goods Administration
Comment: The potential advantages of Abstral are acknowledged. On the other hand, one
of the advantages of Actiq – particularly during initial dose titration – is that the product
can be removed from the mouth if signs of excessive opioid effects appear. This is less
feasible with a rapidly-disintegrating, mucosally-adherent, sublingual tablet such as
Abstral. It is therefore particularly important that patients using Abstral should be started
on the lowest available dose and that titration to a higher dose level should occur in
conservative steps.
In addition, the promise of improved bioavailability compared to Actiq does not appear to
have been realised (as discussed in greater detail later) and improved consistency of
absorption compared to Actiq was not demonstrated in the submitted data.
2. Contents of the clinical dossier
2.1.
Scope of the clinical dossier
The following clinical information was presented for evaluation:

14 pharmacokinetic studies (7 in opioid naïve healthy subjects, 1 in opioid tolerant patients
with cancer pain);

1 study of the effect of acidic and basic food/beverages on oral pH;

1 placebo controlled pharmacodynamic (dose exploration) study in opioid tolerant patients
with cancer pain;

1 placebo controlled efficacy/safety study in opioid tolerant patients with cancer pain;

1 uncontrolled, open label, long term safety study in opioid tolerant patients with cancer
pain;

Literature references.
2.2.
Paediatric data
The submission did not include paediatric data. The sponsor provided a justification for the
absence of a paediatric development program in children aged 0 to 5 years of age, based on the
factors such as the limited number of patients in this age group with cancer pain, questions
about “whether BTcP as an entity exists in the same way as we understand it in adults and older
children” and a lack of validated and accepted tools to identify and quantify BTcP in this age
group.
Comment: Sublingual administration of fentanyl without swallowing the product, be it a
tablet or liquid, would also be impractical in very young children. However, no justification
was provided for the lack of clinical data or a development program in children over the
age of 5 years.
2.3.
Good clinical practice
All of the clinical studies in the submission were stated to have been conducted in accordance
with Good Clinical Practice and in accordance with clinical trials legislation/regulations in the
countries in which they were performed.
Submission PM-2011-03151-3-1 Extract from the Clinical Evaluation Report for fentanyl citrate (Abstral)
Page 7 of 56
Therapeutic Goods Administration
3. Pharmacokinetics
3.1.
Summary of pharmacokinetics
Several PK studies and the pivotal efficacy study used the Abstral formulation proposed for
registration (Formulation A). A number of developmental formulations were also used in the
clinical pharmacology studies. Of these, Formulation 1 (used in 4 PK studies and 1 PD study)
was shown to be bioequivalent to Formulation A. As previously noted, Formulation 2 (used in 2
PK studies) was not directly compared to Formulation A, and cannot be regarded as
bioequivalent to the market formulation.
3.1.1.
3.1.1.1.
Summary PK parameters
Single dose
The pharmacokinetics of fentanyl after the administration of single doses of Abstral were
investigated for several formulations, for doses ranging from 50 μg to 800 μg, in healthy
subjects and in patients with cancer pain. Summary PK parameters for the market formulation
of Abstral (Formulation A) and the bioequivalent Formulation 1 are shown in Tables 1-7.
Table 1: Mean in vivo tablet dissolution time (minutes) after single doses of Abstral.
* Median; † 2 x 400 μg tablets; ‡ 4 x 200 μg tablets; # 2 x 800 μg tablets; all other doses given as single tablets.
Dose comp. = Type of comparison between different dose levels in that study (PG = parallel group, XO =
crossover).
Table 2: Median Tfirst (minutes) after single doses of Abstral.
Dose comp. = Type of comparison between different dose levels in that study (PG = parallel group, XO =
crossover).
† 800 μg dose given as 2 x 400 μg tablets; all other doses given as single tablets. * Data from Site 1 only.
Submission PM-2011-03151-3-1 Extract from the Clinical Evaluation Report for fentanyl citrate (Abstral)
Page 8 of 56
Therapeutic Goods Administration
Table 3: Mean Cmax (ng/mL) after single doses of Abstral.
Dose comp. = Type of comparison between different dose levels in that study (PG = parallel group, XO =
crossover).
† 2 x 400 μg tablets; ‡ 4 x 200 μg tablets; # 2 x 800 μg tablets; all other doses given as single tablets. § Swedishmanufactured tablets (data from this study for US-manufactured tablets [predating manufacturing process
change] have been disregarded).
* Data from Site 1 only.
Table 4: Median Tmax (minutes) after single doses of Abstral.
Dose comp. = Type of comparison between different dose levels in that study (PG = parallel group, XO =
crossover).
† 2 x 400 μg tablets; ‡ 4 x 200 μg tablets; # 2 x 800 μg tablets; all other doses given as single tablets. § Swedishmanufactured tablets (data from this study for US-manufactured tablets [predating manufacturing process
change] have been disregarded - see text). * Data from Site 1 only.
Submission PM-2011-03151-3-1 Extract from the Clinical Evaluation Report for fentanyl citrate (Abstral)
Page 9 of 56
Therapeutic Goods Administration
Table 5: Mean AUC0-t (ng.h/mL) after single doses of Abstral.
Dose comp. = Type of comparison between different dose levels in that study (PG = parallel group, XO =
crossover).
† 2 x 400 μg tablets; ‡ 4 x 200 μg tablets; # 2 x 800 μg tablets; all other doses given as single tablets. § Swedishmanufactured tablets (data from this study for US-manufactured tablets [predating manufacturing process
change] have been disregarded - see text).* Data from Site 1 only.
Table 6: Mean AUC0-∞ (ng.h/mL) after single doses of Abstral.
Note: The summary AUC0-∞ estimates were potentially unreliable in many studies due to inadequate
characterisation of the slope of the terminal portion of the log concentration-time curve and/or extrapolation
of more than 20% of AUC0-∞ in a significant proportion of subjects. See the individual study summaries for
details.
Dose comp. = Type of comparison between different dose levels in that study (PG = parallel group, XO =
crossover).
† 2 x 400 μg tablets; ‡ 4 x 200 μg tablets; # 2 x 800 μg tablets; all other doses given as single tablets. § Swedishmanufactured tablets (data from this study for US-manufactured tablets [predating manufacturing process
change] have been disregarded - see text). * Data from Site 1 only.
Submission PM-2011-03151-3-1 Extract from the Clinical Evaluation Report for fentanyl citrate (Abstral)
Page 10 of 56
Therapeutic Goods Administration
Table 7: Mean t1/2 (h) after single doses of Abstral.
Note: The summary t1/2 estimates were potentially unreliable in many studies due to inadequate
characterisation of the slope of the terminal portion of the log concentration-time curve in a significant
proportion of subjects. See the individual study summaries for details.
Dose comp. = Type of comparison between different dose levels in that study (PG = parallel group, XO =
crossover).
† 2 x 400 μg tablets; ‡ 4 x 200 μg tablets; # 2 x 800 μg tablets; all other doses given as single tablets. § Swedishmanufactured tablets (data from this study for US-manufactured tablets [predating manufacturing process
change] have been disregarded - see text). * Data from Site 1 only.
3.1.1.2.
Multiple dose
PK parameters after the administration of repeated doses of the market formulation of Abstral
(Formulation A) or the bioequivalent Formulation 1 in healthy subjects are summarised in
Tables 8-14, below (with single dose parameters from the same studies for comparison).
Table 8: Mean in vivo tablet dissolution time (minutes) after single and multiple doses of Abstral
in healthy subjects.
Dose comp. = Type of comparison between different dose levels in that study (PG = parallel group).
Table 9: Mean Cmax (ng/mL) after single and multiple doses of Abstral in healthy subjects.
Dose comp. = Type of comparison between different dose levels in that study (PG = parallel group).
Table 10: Median Tmax (min) after single and multiple doses of Abstral in healthy subjects.
Dose comp. = Type of comparison between different dose levels in that study (PG = parallel group).
Submission PM-2011-03151-3-1 Extract from the Clinical Evaluation Report for fentanyl citrate (Abstral)
Page 11 of 56
Therapeutic Goods Administration
Table 11: Mean Ctrough (ng/mL) after multiple doses of Abstral in healthy subjects.
Dose comp. = Type of comparison between different dose levels in that study (PG = parallel group).
Table 12: Mean AUC0-t (ng.h/mL) after single and multiple doses of Abstral in healthy subjects.
Dose comp. = Type of comparison between different dose levels in that study (PG = parallel group).
Table 13: Mean AUCτ (ng.h/mL) after single and multiple doses of Abstral in healthy subjects.
Dose comp. = Type of comparison between different dose levels in that study (PG = parallel group).
Table 14: Mean t1/2 (h) after single and multiple doses of Abstral in healthy subjects.
Dose comp. = Type of comparison between different dose levels in that study (PG = parallel group).
* Result may be unreliable due to exclusion of 5/10 subjects (see Study Summary).
No multiple dose PK studies were conducted in patients with cancer pain. There are no data for
Tfirst during repeated dosing, since regular q4h and q6h dosing meant that plasma fentanyl
concentrations remained detectable throughout the entire dose interval and therefore
immediately after each dose at steady state. There are also no data for AUC0-∞ at steady state as
this parameter was not assessed in Study 2246-EU005 and the results in Study-EU-002 were
unreliable due to exclusion of 8 of the 10 subjects.
3.1.2.
3.1.2.1.
3.1.2.1.1.
Pharmacokinetics in healthy subjects
Absorption
Sites of absorption
Abstral sublingual tablet are intended to be placed under the tongue where they adhere to the
buccal mucosa then dissolve, releasing fentanyl which is systemically absorbed via three
potential mechanisms:

Direct passage from the adherent tablet into the buccal mucosa.

Dissolution into saliva, followed by absorption across the buccal mucosa.

Dissolution into saliva, followed by swallowing and gastrointestinal absorption.
The sponsor asserts that oral absorption plays a minimal role and that absorption is
predominantly via the transmucosal route.
Submission PM-2011-03151-3-1 Extract from the Clinical Evaluation Report for fentanyl citrate (Abstral)
Page 12 of 56
Therapeutic Goods Administration
Comment: The first two absorption mechanisms avoid first-pass metabolism. First-pass
metabolism reduces the bioavailability of fentanyl by about two-thirds, so if oral
transmucosal absorption can be made predominant then increased bioavailability would
be anticipated compared to Actiq, from which only about 25% of the fentanyl is absorbed
via the oral transmucosal route and 75% is swallowed.
In the case of Abstral, the favouring of oral transmucosal absorption has been attempted
by a combination of an optimised tablet formulation (so that the tablet adheres well to the
mucosa but also does not dissolve too slowly) and careful adherence to the correct method
of administration (not moving the tablet around the mouth and not swallowing the drugcontaining saliva). In theory, this would lead to a higher bioavailability of fentanyl from
Abstral tablets, compared Actiq lozenges. As discussed elsewhere, however, this expectation
was not realised in practice.
3.1.2.1.2.
In vivo tablet dissolution
The time taken for Abstral sublingual tablets to dissolve under the tongue was assessed in a
number of the PK studies. Mean in vivo tablet dissolution times for the market formulation
(Formulation A) ranged from 5 to 16 minutes, with the mean around 7 minutes in most studies.
Comment: No attempt was made to correlate in vivo dissolution times with Tmax, Cmax or
AUC. On average, there was a delay between full dissolution of the Abstral tablets (around
5-16 minutes) and Tmax (around 30-60 minutes). This may represent a delay in the
distribution of fentanyl from the buccal mucosa to the systemic circulation (ie. the
formation of a transient depot of fentanyl in the buccal mucosa under the tablet
application site, from which fentanyl is then released to the systemic circulation) and/or a
delayed contribution to Cmax from swallowed fentanyl. In regard to the latter, the mean
fentanyl plasma concentration-time curves did not display the double peak that is
characteristic after Actiq administration (where the second peak represents intestinal
absorption of swallowed fentanyl). However when individual plasma-concentration time
curves were provided, a double peak was seen after Abstral administration in some of
them, showing that swallowing of the drug and subsequent intestinal absorption do occur
in some individuals.
As can be seen in the relevant tables in the EN3267-012 and-013 study summaries, Abstral
tablets dissolved more rapidly than Actiq lozenges. However, this did not lead to any
noticeable shortening of Tmax.
3.1.2.2.
3.1.2.2.1.
Bioavailability
Absolute bioavailability
The absolute bioavailability of the Abstral tablets proposed for marketing was 55% in Study
EN3267-012.
Comment: The absolute bioavailability of Abstral tablets was similar to that of Actiq
lozenges (50% in Study EN3267-012 and “about 50%” in the Actiq PI).
3.1.2.2.2.
Bioequivalence of clinical trial and market formulations
The pivotal efficacy study EN3267-005 used the formulation that is proposed for marketing
(Formulation A). Bioequivalence between Formulation A and a number of developmental
formulations was assessed as follows:

Formulation 1: A number of PK studies and the single PD study used a developmental
formulation (Formulation 1, also referred to as Formulation D) that was shown in a singledose study to be bioequivalent in respect of Cmax and AUC to Formulation A and to have a
similar Tmax to Formulation A (Study SuF-003).
Submission PM-2011-03151-3-1 Extract from the Clinical Evaluation Report for fentanyl citrate (Abstral)
Page 13 of 56
Therapeutic Goods Administration

Formulation 2: Another developmental formulation (Formulation 2) was used in two PK
studies but was not compared to Formulation A (or to Formulation 1 which is bioequivalent
to Formulation A). Only cross-study comparisons between Formulation 2 and Formulation A
are available.

Formulations B and C: Two further developmental formulations (B and C) were compared
to Formulation 1 in Study SuF-003, but were not used in any efficacy/safety studies and are
not proposed for registration, so the comparison is not relevant.
Comment: Formulation 2 has not been shown, either directly or indirectly, to be
bioequivalent to the market formulation (Formulation A). Accordingly, potentially
formulation-dependent PK data from the studies that used Formulation 2 have been
disregarded.
3.1.2.2.3.
Bioequivalence of different tablet strengths
In Study EN3267-003, 2 x 400 μg and 2 x 200 μg Abstral sublingual tablets of the formulation
proposed for marketing were bioequivalent to 1 x 800 μg tablet.
The dose-adjusted bioequivalence of different tablet strengths was also assessed in Study 2246EU-003, but that study used Formulation 2 and the results have been disregarded as discussed
above.
Comment: The 800 μg tablet is not proposed for registration in Australia and the
bioequivalence of 4 x 200 μg to 2 x 400 μg tablets was not statistically assessed. However,
the comparison to the 800 μg tablet provides indirect evidence that the 200 μg and 400 μg
tablets are likely to be bioequivalent on a dose-adjusted basis.
3.1.2.2.4.
Bioequivalence to relevant registered products
The bioavailability of Abstral Formulation A (proposed for registration) was compared to that of
(US-marketed) Actiq in three studies.
In the first study, EN3267-001, Abstral appeared to have double the bioavailability of Actiq.
However, the Actiq lozenges were not administered in accordance with the instructions in the
Actiq PI and a post-study audit revealed that many of the Actiq lozenges were not fully
consumed. Accordingly, the bioavailability comparison in Study EN3267-001 has been
disregarded.
Two further studies (EN3267-012 and EN3267-013) were performed. In those studies, the
Abstral formulation proposed for marketing (Formulation A) was found to be bioequivalent to
US-registered Actiq lozenges, in respect of Cmax, AUC0-t and AUC0-∞, using the standard
acceptance range of 80-125%. Tmax values were also similar for the two products. However,
although the Actiq lozenges were fully consumed in these studies, administration was still not in
strict accordance with the Actiq PI, which states that the lozenge should be manipulated in the
mouth so that it is consumed over a 15 minute period. Instead, the mean time to fully consume
the Actiq lozenges was 28 minutes in EN3267-012 and about 20 minutes in EN3267-013.
Comment: Consumption of the Actiq lozenges over a mean of 28 minutes instead of the PIrecommended 15 minutes in Study EN3267-012 did not appear to affect the fentanyl AUC,
given that the absolute bioavailability of Actiq in that study was the same as is reported in
the Actiq PI. By extension, the fentanyl AUC after Actiq administration in Study EN3267013 - where the consumption time was closer to the PI-recommended value - would also
not have been affected. However a small effect on Cmax could not be excluded.
3.1.2.2.5.
Influence of food and beverages
The influence of food or beverages on Abstral bioavailability was not directly assessed.
Comment: Intake of food or beverages while the Abstral tablet is dissolving would
presumably increase the swallowing of drug-containing saliva or could risk detaching the
Submission PM-2011-03151-3-1 Extract from the Clinical Evaluation Report for fentanyl citrate (Abstral)
Page 14 of 56
Therapeutic Goods Administration
tablet from the mucosa with subsequent swallowing of the tablet and reduced
bioavailability. The draft PI appropriately states that patients should be advised not to eat
or drink anything until the sublingual tablet is completely dissolved.
A different issue is whether the dissolution rate of Abstral tablets (with a potential effect on
bioavailability) is altered by changes in oral pH, for example due to prior consumption of
acidic or basic beverages. Oral pH did not appear to affect the dissolution of Abstral
Formulation A tablets in the multi-dose study 2246-EU-005, but the capacity to assess such
a relationship was very limited because there was little variability in oral pH values in the
study subjects, who were fasted for at least 1 hour before each dose.
Study EN3267-PH001 investigated the effect of acidic (orange juice, coffee) and basic
(milk) beverages on oral pH. When coffee and milk were held in the mouth for a period of 2
minutes, there was no meaningful effect on sublingual and buccal pH. A small effect was
seen after orange juice (paradoxically, a mean rise in pH of 0.53 pH units) but this was
transitory and the pH returned to baseline levels within approximately 10 minutes. The
data indicate that the effect of acid and basic beverages (and by extension foods) on the
absorption of Abstral is likely to be minimal, and no specific restrictions are needed in the
PI regarding beverage or food consumption prior to Abstral administration.
3.1.2.2.6.
Dose proportionality
Dose proportionality of bioavailability is potentially formulation dependent and was assessed
for the market formulation (Formulation A) or the bioequivalent Formulation 1 in 4 studies.

In Study EN3267-001, one group of healthy subjects received a single tablet of Abstral
Formulation A 100 μg and another group received a single tablet of Abstral Formulation A
800 μg. The study had a number of problems and only the PK data relating to Abstral from
Study Site 1 are considered to be reliable. Based on the data from that site, the dosenormalised ratios (low dose/high dose) and corresponding 95% CIs were 1.35 (0.93-1.97)
for Cmax and 1.26 (0.81-1.95) for AUC0-t.

In Study 2246-EU-005, multiple doses of Abstral Formulation A, 100 μg, 200 μg, 400 μg or
800 μg (as 2 x 400 μg tablets) were administered to four groups of healthy subjects (one
dose level per group). The pair wise ratios of AUCτ, AUC0-∞, C6h and Cmax were in proximity to
the ratio of the doses after single doses and at steady-state, consistent with dose
proportionality of these parameters. However, no formal statistical analysis of doseproportionality was performed.

In Study 2246-EU-001, single doses of Abstral Formulation 1, 50 μg, 100 μg, 150 μg or 200
μg were administered to four groups of healthy subjects (one dose level per group). Mixedeffect model analysis showed that none of the selected plasma PK parameters (AUC0-∞, AUC0t, Cmax, CLr and CL/F) met the criteria for dose proportionality, but Cmax did increase in a
linear manner with dose.

In Study EN3267-013, the authors concluded that dose-proportionality had been
demonstrated after single doses of Abstral 800 μg and 1600 μg on the grounds that “The
mean plasma fentanyl concentrations and the AUC and Cmax values for EN3267 [Abstral] at
single doses of 1 x 800 μg and 2 x 800 μg, and for Actiq at single doses of 1 x 800 μg and 1 x
1600 μg increased in a linear fashion (that is, at a rate and extent very close to a 2-fold
progression between the doses). Actiq is recognized as having dose proportional
pharmacokinetics and the equivalency of EN3267 and Actiq Cmax, AUC0-t, and AUC0-∞, also
demonstrates that EN3267 [Abstral] has dose proportional pharmacokinetics”. However,
the ratios were not actually presented in the study report and there was no statistical
assessment of the dose-proportionality of fentanyl pharmacokinetics. When the ratios were
calculated by the evaluator, they were found to be about 1.9 for AUC0-t and Cmax and about
1.7 for AUC0-∞, for both Abstral and Actiq. The ratios based on AUC0-∞ do not appear to be
Submission PM-2011-03151-3-1 Extract from the Clinical Evaluation Report for fentanyl citrate (Abstral)
Page 15 of 56
Therapeutic Goods Administration
consistent with dose-proportionality, but as discussed in the study summary, the results for
AUC0-∞ may be unreliable. The ratios for AUC0-t and Cmax were consistent with doseproportionality, although this was not confirmed by a formal statistical analysis.
Comment: The data indicate regarding dose-proportionality for Abstral Formulation A are
not conclusive:

In Study EN3267-001, the 95% CIs for the dose-normalised Cmax and AUC0-t ratios for
Abstral Formulation A include 1, so the data are consistent with dose-proportionality.
However, the point estimates are well away from 1 and the 95% CIs are wide, so the
data are also consistent with non-proportionality. It should also be remembered that
the dose-proportionality assessment is based on the comparison of two different
groups of subjects (rather than two different doses within the one set of subjects) and
may be confounded by differences between the two groups of subjects. Assignment to
the low or high dose group was randomised and this should have reduced the potential
for such confounding and bias, but with only 20 subjects in each group the power of
randomisation to provide comparable groups is limited.

In Study 2246-EU-005, the data for Abstral Formulation A were again consistent with
dose proportionality but were not statistically analysed.

In Study 2246-EU-001, inspection of the dose-normalised mean values provided in the
Study Summary shows that AUC0-t was clearly not dose proportional, which was
attributed in the Clinical Summary to underestimation of exposure at the 50 μg dose,
since it was impossible to follow plasma fentanyl concentrations for long enough after
that dose to depict the pharmacokinetic profile entirely. Cmax and AUC0-∞ values were
closer to dose-proportional but, as noted above, were not confirmed as such in the
statistical analysis.

In Study EN3267-013, the ratios for AUC0-t and Cmax were consistent with doseproportionality, although this was not confirmed by a formal statistical analysis and
dose-proportionality was not assessed within the dose range (up to 800 μg) that is
proposed for Abstral.
3.1.2.2.7.
Bioavailability during multiple-dosing
The bioavailability of Abstral during repeated dosing was investigated for the market
formulation (Formulation A) or the bioequivalent Formulation 1 in two studies:

In Study 2246-EU-005, four groups of healthy subjects took single doses of Abstral
Formulation A 100 μg, 200 μg, 400 μg or 800 μg (as 2 x 400 μg tablets), followed by a 72
hour washout period, then the same dose every 6 hours for 72 hours. Accumulation
parameters are summarised in Table 15.
Table 15: 2246-EU-005: Mean accumulation ratios after repeated administration of Abstral
Formulation A every 6 hours for 72 hours (13 doses).

In Study 2246-EU-002, healthy subjects took Abstral Formulation 1, 50 μg every 4 hours for
44 hours (12 doses). Accumulation parameters are summarised in Table 16.
Submission PM-2011-03151-3-1 Extract from the Clinical Evaluation Report for fentanyl citrate (Abstral)
Page 16 of 56
Therapeutic Goods Administration
Table 16: 2246-EU-002: Mean accumulation ratios after repeated administration of Abstral
Formulation 1 every 4 hours for 44 hours (12 doses).
Comment: During repeated 6-hourly dosing with Abstral Formulation A at doses of 100 to
800 μg, peak plasma fentanyl concentrations rose to about 1.5 to 2 times the peak
concentration after a single dose, while fentanyl exposure over the dose interval rose to
about 2 to 2.5 times the amount after a single dose. The mean accumulation ratios were
broadly similar across the range of doses (noting the limitations of the between-dose
comparison, which was based on data from different groups of patients, rather than data
for different doses from a single group of patients).
During repeated 4-hourly dosing with Abstral Formulation 1 at a dose of 50 μg, peak
plasma fentanyl concentrations rose to about 2.6 times the peak concentration after a
single dose, while fentanyl exposure over the dose interval rose to about 3 times the
amount after a single dose. The higher accumulation ratios for the 50 μg dose compared
the 100 to 800 μg doses presumably reflect the shorter dose interval at which the 50 μg
dose was administered (4 hours vs 6 hours for the other doses). Very similar accumulation
ratios for Cmax and AUCτ were seen with four-hourly dosing in Study 2246-EU-004, although
that study used a formulation of Abstral that has not been shown to be bioequivalent to the
one proposed for registration.
The anticipated accumulation during clinical usage of Abstral would generally be lower
than was seen in these studies, since the drug will be consumed at irregular intervals in
response to episodes of BTcP rather than regularly every 4 or 6 hours. For example, during
the Long-term Extension Phase of the pivotal efficacy study, patients took a mean of 2.9
Abstral doses per day (median 3 doses). Fentanyl accumulation might occasionally be
higher in some individuals, noting that the draft PI recommends a maximum of 4 doses per
day, but with only 30 minutes between doses when two doses are taken for the same
episode of BTcP, and no restriction on the interval between doses taken for separate BTcP
episodes. This could be avoided by specifying an appropriate minimum interval between
doses taken for separate BTcP episodes.
3.1.2.3.
Distribution, metabolism and elimination
The distribution, systemic clearance and excretion of fentanyl are not formulation-dependent
and are well described in the published literature. As described in the PIs for registered fentanyl
products:

Fentanyl is a highly lipophilic drug that is rapidly distributed to the brain, heart , lungs,
kidneys and spleen, followed by a slower redistribution to muscles and fat.

The plasma protein binding of fentanyl is 80-85%. The main binding protein is alpha-1-acid
glycoprotein, but both albumin and lipoproteins contribute to some extent. The free fraction
of fentanyl increases with acidosis.

The mean volume of distribution at steady state (Vss) is about 4 L/kg.

Fentanyl is metabolised by CYP3A4 in the liver and intestinal mucosa where it is converted
to a number of pharmacologically inactive metabolites, the major one being norfentanyl.
Fentanyl has a high hepatic extraction ratio (0.8-1.0), so the hepatic clearance of fentanyl
approaches hepatic blood flow.
Submission PM-2011-03151-3-1 Extract from the Clinical Evaluation Report for fentanyl citrate (Abstral)
Page 17 of 56
Therapeutic Goods Administration

Within 72 hours aft the intravenous administration of fentanyl, about 75% of the dose is
excreted in the urine, mostly as metabolites, with less than 10% as unchanged drug. About
1% is excreted unchanged in the faeces. The metabolites are mainly excreted in the urine,
while faecal excretion is less important.

The total plasma clearance of fentanyl is about 0.5 L/h/kg.

The terminal elimination half-life of fentanyl after intravenous administration is about 8
hours.
Results for t1/2 from PK studies of Abstral are summarised in Table 7 (single dose studies) and in
Table 14 (multiple dose studies).
Comment: In the single dose studies, t1/2 was generally consistent with published data.
The half-life of fentanyl tended to increase with dose and was unusually low in Study SuF003 and for the 100 μg dose in Study 2246-EU-001, but these are probably artefacts caused
by the limitations of the fentanyl assay and/or the collection of blood samples over too
short a time period after the dose. It is probable that t1/2 was sometimes underestimated in
the single dose studies, particularly after low doses when fentanyl concentrations dropped
below the LOQ relatively soon after the dose. In the multiple dose studies, much higher
plasma fentanyl concentrations were achieved and they remained quantifiable for a
considerably longer period. The terminal half-life then appeared to be much longer (18-24
hours). The sponsor considers that the true terminal elimination phase was not captured in
the single dose studies and was only seen in the multiple dose studies. It is also possible that
the long terminal half-life in the multiple dose studies (measured after the final dose of
Abstral) could be due to the slow redistribution of accumulated fentanyl from adipose
tissue and muscle back into the systemic circulation, maintaining low plasma fentanyl
levels over a prolonged period and prolonging the terminal half-life. The finding that
steady-state was achieved within 2 to 3 days after starting Abstral is further evidence that
the effective half-life is in the order of 8 to 10 hours, rather than 20 hours.
3.1.3.
Pharmacokinetics in the target population
The pharmacokinetics of fentanyl after the administration of single doses of Abstral
Formulation 1, 100, 200 and 400 μg, were analysed in 8 patients with cancer pain in Study SuF001.
Comment: Only a small number of patients were studied and no formal statistical analysis
was conducted, but AUC0-∞ and Cmax appeared to be dose-proportional in cancer pain
patients, based on the similarity of dose-normalised mean values across the three dose
levels. As in healthy subjects, Tmax appeared to lengthen as the dose increased.
A comparison of PK parameters in cancer pain patients versus healthy subjects relies on
cross-study comparisons and very limited PK data in patients with cancer pain, and may
not be reliable. With this caveat, there did not appear to be any large PK differences
between cancer pain patients and healthy subjects, although Cmax and AUC0-∞ were perhaps
a little higher in the cancer patients for a given dose of Abstral.
3.1.4.
3.1.4.1.
Pharmacokinetics in other special populations
Pharmacokinetics in subjects with impaired hepatic function
The pharmacokinetics of fentanyl after Abstral administration were not specifically studied in
patients with hepatic impairment. However, fentanyl is known to be predominantly eliminated
by hepatic metabolism to norfentanyl and other inactive metabolites, with only about 10% of a
dose excreted as unchanged drug in the urine. Impaired liver function would increase the
bioavailability of any swallowed fentanyl and decrease the systemic clearance of fentanyl, which
could lead to increased and prolonged opioid effects.
Submission PM-2011-03151-3-1 Extract from the Clinical Evaluation Report for fentanyl citrate (Abstral)
Page 18 of 56
Therapeutic Goods Administration
Comment: Impaired liver function can occur in patients with cancer pain for various
reasons, including the effects of concomitant treatment and metastatic liver disease. It
would not be clinically appropriate to completely exclude such patients from using Abstral
by making hepatic impairment a contraindication, but caution should be observed when
the drug is used in patients with hepatic impairment. The draft Abstral PI includes a
Precaution to this effect, which is similar to the one in the Actiq PI.
3.1.4.2.
Pharmacokinetics in subjects with impaired renal function
The pharmacokinetics of fentanyl after Abstral administration were not specifically studied in
patients with renal impairment and limited data are available regarding the effects in general of
renal impairment on fentanyl pharmacokinetics. Renal impairment would not be expected to
have a direct impact on the elimination of fentanyl, which is predominantly due to hepatic
metabolism. However renal impairment produces changes (eg. in plasma proteins) that would
affect the volume of distribution of fentanyl and unbound fentanyl concentrations.
Comment: For various reasons, including age-related deterioration in renal function and
the side effects of some antineoplastic agents, renal impairment is not uncommon in
patients with cancer pain. It would not be clinically appropriate to completely exclude such
patients from using Abstral by making renal impairment a contraindication, but caution
should be observed when the drug is used in patients with renal impairment. The draft
Abstral PI includes a Precaution to this effect, which is similar to the one in the Actiq PI.
3.1.4.3.
Pharmacokinetics according to age
The effect of age on the pharmacokinetics of fentanyl was not specifically studied after Abstral
administration. However, published data for other dose forms show that the elimination of
fentanyl is slower and the terminal elimination half-life is longer in the elderly. In addition, the
elderly are known to be more sensitive to the pharmacological effects of opioids, including
fentanyl.
Comment: Abstral should be used with caution in elderly patients, especially during
titration. The draft Abstral PI includes statements relating to use in the Elderly under
Precautions and Dosage and Administration.
3.1.4.4.
3.1.4.4.1.
Pharmacokinetics related to genetic factors
Ethnicity
The pharmacokinetics of fentanyl after the administration of single 50 to 200 μg doses of
Abstral Formulation 1 were comparable in healthy Japanese and Caucasian subjects (Study
2246-EU-001).
3.1.4.4.2.
Gender
The pharmacokinetics of fentanyl after the administration of single and repeated doses of
Abstral were compared between women and men in two studies:

In Study 2246-EU-005, four groups of healthy male and female subjects took single doses of
Abstral Formulation A 100 μg, 200 μg, 400 μg or 800 μg (as 2 x 400 μg tablets), followed by
a 72 hour washout period, then the same dose every 6 hours for 72 hours. Summary PK
parameters are compared in women and men in Table 17.
Submission PM-2011-03151-3-1 Extract from the Clinical Evaluation Report for fentanyl citrate (Abstral)
Page 19 of 56
Therapeutic Goods Administration
Table 17: 2246-EU-005. Gender comparison (females versus males) of summary PK parameters
after administration of Abstral fentanyl citrate sublingual tablets (market Formulation A) as a
single dose or q6h for 72 hours.
Comparisons are based on arithmetic means except for Tmax (based on medians). Negative values indicate that
the mean (or median for Tmax) was lower in females than in males

In Study 2246-EU-002, healthy male and female subjects took Abstral Formulation 1, 50 μg
every 4 hours for 44 hours (12 doses). After a single dose, the mean Cmax in women was 25%
higher than in men, the median Tmax was 25% shorter, and the mean AUC0-4h was 7% higher.
At steady state, however, all of these parameters were comparable in women and men (<3%
between-gender difference in mean values).
Comment: In Study 2246-EU-005, there were no consistent differences between males and
females in respect of Cmax or Tmax after a single dose or at steady-state. Systemic exposure
(AUC) was similar in males and females after a single dose and also at steady state. Halflife was consistently longer in females compared to males after a single dose and also at
steady state, but this did not appear to lead to an increase in Cmax or AUC during repeated
6-hourly dosing.
The report of study 2246-EU-002 attributed the higher Cmax and slightly higher AUC0-4h in
women after a single dose (compared to men) to the lower body weight of the female
subjects. However, this does not explain the shorter Tmax in women after the first dose. The
number of subjects of each gender was small (5 each), and the data may not be
representative.
Overall, the pharmacokinetics of fentanyl after Abstral administration do not appear to
differ between men and women to a clinically meaningful extent, after allowing for
differences in weight. The half-life of fentanyl was longer in women than men, but this did
not lead to higher plasma concentrations of fentanyl during repeated 4 or 6 hourly dosing.
It is thus reasonable to recommend the same starting dose, titration steps, maximum
Abstral dose and maximum number of doses per day for men and women.
3.1.5.
Pharmacokinetic interactions
Pharmacokinetic interactions were not specifically examined in the submitted studies. However,
fentanyl is known to be metabolised by CYP3A4 in the liver and intestinal wall. Grapefruit juice
and drugs that inhibit CYP3A4 - such as macrolide antibiotics (eg. erythromycin), azole
antifungal agents (for example, ketoconazole and itraconazole), certain protease inhibitors (eg.
ritonavir) - would be expected to increase the bioavailability of any fentanyl that is swallowed
during Abstral administration (eg. from swallowed drug-containing saliva), and would also slow
the elimination of fentanyl, with resultant increased or prolonged opioid effects. Conversely,
inducers of CYP3A4 would be expected to increase both the intestinal and hepatic metabolism
of fentanyl, leading to reduced efficacy.
Opioid antagonists (for example, naloxone and naltrexone), and partial agonists or drugs with
mixed agonist/antagonist activity (eg. pentazocine, butorphanol, buprenorphine, nalbuphine)
would be expected to reduce or eliminate the efficacy of Abstral, as well as precipitating opioid
Submission PM-2011-03151-3-1 Extract from the Clinical Evaluation Report for fentanyl citrate (Abstral)
Page 20 of 56
Therapeutic Goods Administration
withdrawal symptoms (given that Abstral is only intended to be used in opioid-tolerant
patients).
The use of opioids concurrently with monoamine oxidase (MAO) inhibitors or within 2 weeks of
the cessation of MAO inhibitors is usually contraindicated, on the basis that serious and
sometimes fatal reactions have been reported when the opioid analgesic pethidine has been
administered to patients receiving MAO inhibitors.
Comment: The draft Abstral PI includes Precautions regarding these interactions that are
similar to the ones in the Actiq PI, with the exception of the statement regarding MAO
inhibitors. The Actiq PI contraindicates the use of the drug concurrently with MAO
inhibitors, or within 2 weeks after the cessation of the use of MAO inhibitors. In the draft
Abstral PI, this is listed only as a Precaution rather than a Contraindication.
3.2.
Evaluator’s overall conclusions on pharmacokinetics
The pharmacokinetics of fentanyl after the administration of Abstral have been satisfactorily
elucidated in the submitted studies or by reference to the published literature.
4. Pharmacodynamics
4.1.
Studies providing pharmacodynamic data
The submission included one pharmacodynamic study, SuF-002.
4.2.
Summary of pharmacodynamics
4.2.1.
Primary pharmacodynamics
Study SuF-002 examined the pharmacodynamic effect of Abstral on BTcP. Twenty-three
patients on background slow-release opioid therapy for cancer pain took single doses of Abstral
100 μg, 200 μg, 400 μg and placebo, in a randomised crossover fashion, for consecutive episodes
of BTcP (but with a washout of at least 1 day between doses).
Pain intensity was assessed using a 100 mm visual analogue scale (VAS), before each dose of
study drug and 5, 10, 15, 20 and 30 minutes after each dose. Due to an error in the preparation
of the patient diaries, efficacy was not assessed for the planned 60 minute post-dose period. The
primary PD (efficacy) variable was the pain intensity difference (PID), ie. the change in pain
intensity from the pre-dose value at each time point. A single dose of Abstral 400 μg was
statistically superior overall to placebo for the relief of BTcP. Onset of action was noted at the
first post-dose assessment (5 minutes), with the difference between Abstral and placebo
becoming statistically significant at 15 minutes and persisting until the final assessment at 30
minutes. The mean PID reached 20 mm, which is regarded as a clinically meaningful pain
reduction, less than 15 minutes after a 400 μg dose of Abstral, compared to 30 minutes after a
dose of placebo.
Lower Abstral doses (100 μg and 200μg) were not significantly superior to placebo during the
first 30 minutes after the dose, but did provide satisfactory relief in some patients.
No unexpected safety concerns were raised. Most of the AEs were related to the underlying
cancer and/or background opioid therapy; 2 patients had typical opioid-type TEAEs soon after
the administration of Abstral 400 μg.
Submission PM-2011-03151-3-1 Extract from the Clinical Evaluation Report for fentanyl citrate (Abstral)
Page 21 of 56
Therapeutic Goods Administration
4.3.
Secondary pharmacodynamics
Fentanyl displays typical μ-opioid pharmacodynamic actions other than analgesia. The most
relevant of these in relation to the potential adverse effects of Abstral are: sedation; respiratory
depression; reduced blood pressure and heart rate; decreased intestinal motility (leading to
constipation); and variable effects on urinary tract smooth muscle that can lead to urinary
urgency or difficulty in urination. Miosis, cough suppression and hyporeflexia are other μ-opioid
pharmacological effects that are listed in the Actiq PI.
4.3.1.
Pharmacodynamic interactions
Pharmacodynamic interactions are expected between fentanyl and a range of other drugs. Such
interactions may be beneficial or adverse. The most important beneficial interaction is the
increased analgesic effect that occurs when fentanyl is used in combination with other analgesic
agents, including other opioids, and which forms that basis for the use of Abstral for BTcP. The
most adverse pharmacological interaction relate to the central nervous system and
gastrointestinal effects of fentanyl described above. Sedation, respiratory depression and
hypotension are expected to be increased when Abstral is used in combination with other CNS
depressants such as opioids, general anaesthetics, skeletal muscle relaxants, sedative
antidepressants, sedative antihistamines, barbiturates, benzodiazepines, hypnotics,
antipsychotics, clonidine and related substances, or alcohol. Abstral would also be expected to
increase the risk of gastrointestinal adverse effects such as nausea, vomiting and constipation
when used in combination with other opioids.
Comment: The potential for increased gastrointestinal side effects cannot be avoided given
that Abstral is intended to be used to supplement chronic opioid therapy. Careful dose
titration and the use of appropriate measures to combat constipation should minimise the
impact of this interaction. The complete avoidance of CNS depressant drugs is not feasible,
nor is it necessary. However caution should be observed when Abstral is administered
together with other CNS depressants and the draft PI includes suitable precautionary
statements regarding this interactions.
4.4.
Evaluator’s overall conclusions on pharmacodynamics
The submitted PD study (SuF-002) provides evidence that Abstral Formulation 1 (and thus the
bioequivalent Formulation A that is proposed for registration) should be capable of significantly
relieving acute BTcP, with a sufficiently rapid onset of action, provided that an adequate dose is
used for the individual patient.
SuF-002 also indicated that 100 μg would be a reasonable starting dose of Abstral in subsequent
efficacy studies (on the basis that it did not appear to be associated with excessive adverse
effects and may provide adequate pain relief in some individuals), and that subsequent efficacy
studies should provide for titration of the Abstral dose up to at least 400 μg.
SuF-002 only measured pain severity for 30 minutes after the dose, and did not by itself show
that the duration of action of Abstral is sufficient to treat BTcP. Being a single dose study, the PD
study also did not examine whether the therapeutic and adverse effects of Abstral remained
consistent across successive episodes of BTcP.
5. Dosage selection for the pivotal studies
The report of the pivotal study (EN3267-005) stated: “The study medication, EN3267 [Abstral],
was examined in doses of 100, 200, 300, 400, 600 (two 300 μg tablets), and 800 μg (two 400 μg
tablets) during this study. These doses were provided to allow the flexibility for each patient to
achieve a personally effective dose that avoided intolerable adverse effects. The dose range
chosen for this study was based on pharmacokinetic studies, which examined doses up to 800
Submission PM-2011-03151-3-1 Extract from the Clinical Evaluation Report for fentanyl citrate (Abstral)
Page 22 of 56
Therapeutic Goods Administration
μg and on studies in opioid-tolerant cancer population using doses up to 400 μg, which
demonstrated that the study medication was well tolerated.”
Comment: What is not made explicit in the report of EN3267-005, but is evident from the
data package as a whole, is that the maximum dose of Abstral in the pivotal study was
limited to 800 μg on the basis that this is half the maximum approved dose of Actiq,
combined with a belief that the bioavailability of Abstral is double that of Actiq. As is
discussed elsewhere, this belief was subsequently shown to be mistaken.
6. Clinical efficacy
6.1.
Relief of breakthrough cancer pain
6.1.1.
Pivotal efficacy study: EN3267-005
The submission contained one adequate, well-controlled efficacy study, EN3267-005, which
used a randomised, double-blind, multiple crossover design to compare Abstral and placebo.
The relevant TGA-adopted guideline, CPMP/EWP/612/00 Note for Guidance on Clinical
Investigation of Medicinal Products for Treatment of Nociceptive Pain, states that crossover
designs are useful as exploratory studies of treatments for nociceptive pain, but parallel group
designs are recommended for confirmatory studies. One reason for this, as stated in the
Guideline, is the risk of carryover effects between the two study treatments. However, issues of
practicality and ethical considerations mean that it would be difficult, if not impossible, to
conduct a successfully blinded parallel-group study involving a placebo arm when confirmation
of efficacy requires the treatment of multiple pain episodes and where the dose of the active
agent needs to be titrated over a period of several days. Under these circumstances, reliance on
a crossover design is reasonable.
6.1.1.1.
Study design, location and dates
EN3267-005 was a double-blind, randomised, placebo-controlled, multiple crossover,
multicentre study to evaluate the efficacy and safety of Abstral (Formulation A) for the
treatment of BTcP in opioid tolerant cancer patients. The study was conducted in the USA from
January 2006 to December 2008.
EN3267-005 enrolled patients aged ≥17 with cancer-related pain. The most important entry
criteria were:

Receiving a stable, fixed-schedule oral opioid regimen equivalent to 60 to 1000 mg of oral
morphine per day or transdermal fentanyl equivalent to 50 to 300 μg/h. The fixed-dose
opioid regimen must have been taken for at least 14 days before screening, and must have
been expected to remain unchanged for the duration of the Open-label Titration and
Double-blind Treatment Phases of the study (ie. up to approximately 4 weeks);

On a stable dose of opioid for relief of BTcP. This drug/dose combination was to be used, if
required, as rescue medication during the course of the study.

Experiencing 1 to 4 episodes of BTcP per day.

Patients with moderate to severe ulcerative mucositis (≥grade 2 according to WHO criteria)
were excluded from the study.
The study initially consisted of Screening, followed by an Open-label Titration Phase, then a
Double-blind Treatment Phase, then an Open-label Long-term Extension Phase. The protocol
was amended 3 times during the course of the study:
Submission PM-2011-03151-3-1 Extract from the Clinical Evaluation Report for fentanyl citrate (Abstral)
Page 23 of 56
Therapeutic Goods Administration

Protocol Amendment 1 added an observed test dose of Abstral 100 μg to the screening
phase and added Quality of Life (QoL) assessments.

Protocol Amendment 2 added an interim analysis and associated stopping rules.

Protocol Amendment 3 (made after the interim analysis demonstrated positive efficacy
according to the criteria set in Protocol Amendment 2) removed the Double-blind
Treatment Phase, so that patients who completed the Open-label Titration Phase then
proceeded directly to the Open-label, Long-term Extension Phase. In October 2008, it was
determined that the study had collected a suitable quantity of safety and efficacy data and
the study was terminated early on 31 December 2008. Any patients that were still enrolled
(eg. still in the Long-term Extension Phase) were to undergo an end-of-study visit by this
date.
The study phases are described in further detail below.


Screening
–
Patients were screened to determine if they satisfied the entry criteria.
–
Each patient identified a particular type or location of BTcP as their “target pain,” which
was the only pain treated with study medication throughout the study
–
Prior to Protocol amendment 1, patients who satisfied the entry criteria proceeded
directly to the Open-label Titration Phase.
–
After Protocol Amendment 1, patients who satisfied the entry criteria were given an
open-label test dose of Abstral 100 μg while under observation. Patients who could not
tolerate the test dose were withdrawn from the study and excluded from the efficacy
analyses. Patients who tolerated the test dose were eligible to proceed to the Open-label
Titration Phase.
Open-label Titration Phase. During the Open-label Titration Phase, patients had up to 2
weeks to self-determine a single appropriate dose of Abstral for adequate treatment of
BTcP.
–
Patients took a single dose of Abstral 100 μg for the first BTcP episode.
–
For each subsequent BTcP episode, patients titrated the dose up (if pain relief had been
inadequate at the previous dose) or down (if adverse effects were intolerable at the
previous dose), within the range of 100 to 800 μg, until they found a dose that provided
adequate pain relief without intolerable adverse effects and that remained stable for 2
consecutive days. The doses available to each patient were 100 μg, 200 μg, 300 μg, 400
μg (each as single tablets), 600 μg (2 x 300 μg) and 800 μg (2 x 400 μg).
–
For each dose of study drug, patients recorded efficacy data (described below) in a
patient diary.
–
If pain relief was unsatisfactory 30 minutes after taking a dose of Abstral, patients were
to take rescue medication (ie. their previously prescribed drug/dose of BTcP
medication).
–
Patients were instructed to wait at least 2 hours after the treatment of one BTcP episode
before treating the next episode.
–
Patients who were unable to titrate to a satisfactory and stable dose (eg. because of lack
of efficacy, adverse effects or unstable dose requirements), or who did not continue to
Submission PM-2011-03151-3-1 Extract from the Clinical Evaluation Report for fentanyl citrate (Abstral)
Page 24 of 56
Therapeutic Goods Administration
have 1-4 episodes of BTcP per day,1 were withdrawn from the study and excluded from
the efficacy analyses.
–


Patients who achieved a satisfactory, stable dose of Abstral and continued to have 1-4
episodes of BTcP per day were eligible to proceed to the Double-blind Treatment Phase
(or, after the interim analysis and protocol amendment, directly to the Open-label Longterm Extension Phase).
Double-blind Treatment Phase. Patients who entered the Double-blind Treatment Phase
were issued with a blister card of 10 sequentially numbered doses of study medication that
contained 7 doses of Abstral (at the dose level determined for that patient by the Open-label
Titration Phase) and 3 doses of matching placebo, organised in a randomised fashion.
–
Patients were instructed to take one tablet, in the order presented in the blister card, for
consecutive episodes of BTcP over a period of up to 2 weeks.
–
For each dose of study drug, patients recorded efficacy data (described below) in a
patient diary.
–
If pain relief was unsatisfactory 30 minutes after taking a dose of study medication,
patients were to take rescue medication (ie. their previously prescribed drug/dose of
BTcP medication).
–
Patients were instructed to wait at least 2 hours after the treatment of one BTcP episode
before treating the next episode.
–
Patients who completed the Double-blind Treatment Phase were eligible to enter the
Open-label Long-term Extension Period.
Open-label Long-term Extension Phase. During the Open-label Long-term Extension
Phase, patients used Abstral to treat BTcP episodes over a period of up to 12 months.
–
Investigators could adjust the dose of Abstral as required (eg. due to lack of efficacy or
AEs) within the range of 100 to 800 μg.
–
Patients returned to the study site monthly (± 2 weeks) and site personnel contacted
patients via telephone monthly (ie. approximately 2 weeks after the monthly visits).
The Abstral tablets used in this study were Formulation A, which is proposed for registration
in Australia. Patients were instructed to place the study medication under the tongue into the
deepest part of the oral cavity and allow it to dissolve. Patients were told not to move or touch
the medication once it had fastened to the mucous membrane, not to eat or drink until the tablet
had completely dissolved, not to swallow the medication, and to allow it to dissolve completely
in the mouth without chewing or sucking. Patients were told that the medication should have
dissolved completely in approximately one minute. The proper method for using the study
medication was described to the patient and reviewed regularly throughout the course of the
study.
Although it was expected that episodes of BTcP would be treated with study medication,
patients could choose, for the purpose of convenience, not to treat an episode of BTcP with
study medication if it occurred during sleeping hours (which would mean having to remain
awake to perform the required efficacy assessments). In these instances, patients could use
their previously prescribed drug/dose of BTcP medication to treat the episode.
Comment: Since study patients were suffering up to 4 BTcP episodes per day, the time
between successive doses of study medication would have been short enough to allow
Patients could have a day without an episode of BTcP during the 2-week Open-Label Titration Phase, as long as at
least one episode was recorded either on the previous day or on the subsequent day (ie. patients could not have 2
consecutive days without at least one episode of BTcP).
1
Submission PM-2011-03151-3-1 Extract from the Clinical Evaluation Report for fentanyl citrate (Abstral)
Page 25 of 56
Therapeutic Goods Administration
carryover of effect between some of the doses. For those placebo doses that followed a dose
of Abstral, this carryover would tend to increase the apparent efficacy of placebo, thereby
reducing the difference between Abstral and placebo. On the other hand, when two or more
Abstral doses were taken in succession, any carryover would tend to increase the apparent
efficacy (and also the apparent adverse effects) of the second and subsequent Abstral
doses, thereby increasing the apparent difference between Abstral and placebo. Depending
on the randomisation sequence, 3 or 4 out of the 7 Abstral doses received by each patient
directly followed a preceding Abstral dose. Either 2 or 3 of the placebo doses received by
each patient directly followed an Abstral dose (patients never received 2 placebo doses in
succession, but some received placebo as the first dose of the Double-blind Treatment
Phase). Finally, most of the efficacy endpoints were based on the change from the predose
pain value, rather than absolute pain measurements. Considering these factors, one would
expect the potential carryover effects favouring Abstral and placebo to approximately
balance out.
6.1.1.2.
Efficacy assessments
During the Open-label Titration and Double-blind Treatment Phases, the following efficacy
data were recorded using electronic patient diaries:

Dose of study treatment. During the Open-label Titration Phase, patients recorded the
strength, date and time of each dose of study treatment. During the Double-blind Treatment
Phase, patients recorded the identification number, date and time of each dose of study
treatment.

Pain intensity. Pain intensity was rated on an 11-point scale, where 0 indicated “no pain”
and 10 indicated “pain as bad as you can imagine”, in response to the question: “Please rate
your pain by indicating the one number that tells how much pain you have right now.” Pain
intensity was rated immediately before treating a BTcP episode with study medication (ie.
at 0 minutes), at 10, 15, 30, and 60 minutes after treating the episode, and at the time of
treatment with rescue medication, if applicable.

Pain relief. Pain relief (PR) was rated on a 5-point scale in response to the question: “How
much pain relief do you have now compared to immediately prior to taking the study
medication?”. The rating scale was 0 = None, 1 = Slight, 2 = Moderate, 3 = A lot, 4 =
Complete. Pain relief was rated at 10, 15, 30, and 60 minutes after treatment of a BTcP
episode with study medication and at the time of treatment with rescue medication, if
applicable.

Rescue Medication Use. Patients recorded the date and time at which any rescue
medication was used.

Patient Global Evaluation of Medication. Patients recorded a global evaluation of study
medication 60 minutes after each dose of study drug or upon the use of rescue medication,
in response to the following prompt: “Please rate your overall satisfaction with the pain
medication you took for this episode of breakthrough pain. The rating scale is: 1 = excellent,
2 = very good, 3 = good, 4 = fair, 5 = poor”. Patients were also asked at each study visit
regarding their overall degree of satisfaction with their usual BTcP medication (at the
screening visit) or with the study medication (at subsequent visits), categorised as “very
satisfied”, “satisfied”, “no preference”, “dissatisfied” or “very dissatisfied”.
During the Open-label Long-term Extension Phase, patients were asked to complete a daily
paper diary to record the following efficacy data:

Occurrence or non-occurrence of BTcP.

The dose of study medication used to treat BTcP episodes (with provision for up to 6
episodes per day).
Submission PM-2011-03151-3-1 Extract from the Clinical Evaluation Report for fentanyl citrate (Abstral)
Page 26 of 56
Therapeutic Goods Administration

The number of BTcP episodes treated with other (ie. non-study) medication.

Patient satisfaction with study medication was recorded as described above at each monthly
visit.
Quality of life. After Protocol amendment 1, Quality of Life (QoL) was assessed at screening, at
the end of the Double-blind treatment phase, and at each monthly visit during the Open-label,
Long-term Extension Phase. QoL was assessed using the Brief Pain Inventory (BPI) and
Depression, Anxiety, and Positive Outlook Scale (DAPOS) questionnaires.
6.1.1.3.
Efficacy endpoints
6.1.1.3.1.
Primary endpoint
The primary efficacy endpoint was the sum of pain intensity difference (SPID) from baseline to
30 minutes after dosing. Baseline for each episode was defined as the pain score recorded
before taking study medication for that episode. The PID at each time point was the score at
Baseline minus the score at that time point, so pain relief resulted in positive PID values. The
SPID for an episode was the area under the PID versus time curve up to the cutoff time point (30
minutes for the primary efficacy endpoint), calculated using the trapezoidal rule.2 The mean
SPID across episodes for each treatment for each patient then was calculated and used in the
statistical analysis. Thus, for each patient the 7 episodes treated with Abstral were averaged
into one value and the 3 episodes treated with placebo were averaged into one value and then
used in the analysis.
Comment: The SPID is analogous to the AUC of a drug in plasma, but based on a measure
of drug effect rather than drug concentration. It is a common efficacy endpoint in
analgesic trials. A higher SPID corresponds to greater “total” pain relief over the
assessment period. Other clinically relevant aspects related to the SPID include:

The duration over which the SPID is determined. In the case of BTcP, an assessment
period of 30 minutes is arguably too short to assess whether the analgesic effect is
adequately maintained throughout an episode of BTcP. The SPID over 60 minutes
would provide a more clinically relevant assessment, but is affected by differential use
of rescue medication after 30 minutes in a manner that would tend to exaggerate the
difference between Abstral and placebo (because of the combination of the LOCF
analysis and the natural resolution over time of BTcP episodes).3

The shape of the PID curve. This shows whether pain relief is concentrated around a
particular time after the dose or is adequate throughout the assessment period.
6.1.1.3.2.
Secondary endpoints
Secondary efficacy endpoints were:

Pain intensity difference (PID) at each post-baseline time point (ie. 10, 15, 30, and 60
minutes).
The EN3267-005 study report stated that when calculating the SPID and TOTPAR, the base of the trapezoid (ie. the
time between successive PID or PR values) would be expressed in hours (ie. 5 minutes would be expressed as 1/12,
30 minutes as 1/2). Examination of the reported SPID values (for which no units were given) revealed that this was
not actually done and the time in minutes rather than hours was used in calculating the SPID and TOTPAR. This
increased the magnitude of the SPID and TOTPAR for both treatments (and the between-treatment difference in SPID
and TOTPAR) by a factor of 60 compared to the planned analysis. However it does not affect the between-treatment
statistical comparison of these endpoints.
3 The LOCF analysis appropriately disregarded PID values after the use of rescue medication (not to do so would have
led to an underestimation of the true difference between Abstral and placebo). However, given the higher use of
rescue medication in the placebo group, the LOCF analysis meant that a higher proportion of placebo SPIDs at 60
minutes were based on data where the PID had been locked at the value observed when rescue medication was taken,
thereby disregarding the subsequent contribution to the SPID that would otherwise have flowed from natural
resolution of the BTcP episode after that time point.
2
Submission PM-2011-03151-3-1 Extract from the Clinical Evaluation Report for fentanyl citrate (Abstral)
Page 27 of 56
Therapeutic Goods Administration

Pain relief (PR) scores at each post-baseline time point (ie. 10, 15, 30, and 60 minutes). PR
was scored as: 0 = None, 1 = Slight, 2 = Moderate, 3 = A lot, 4 = Complete.

Total pain relief (TOTPAR) at 30 and 60 minutes. The TOTPAR score for a pain episode was
calculated as the area under the patient’s PR score vs time (in hours after the dose) curve for
that episode, using the trapezoidal rule.

SPID at 60 minutes (as per the primary efficacy endpoint but incorporating data to 60
minutes after the dose).

Percentage of treated BTcP episodes that required use of rescue medication.

Percentage of responders (defined as patients who have at least a 30% decrease in pain
intensity from Baseline to 30 minutes).

Percentage of the 30% and 50% responder rates at 10 and 15 minutes.

Patient global evaluation of study medication.
Comment: As will be discussed later, some of these secondary efficacy endpoints are
potentially informative as to whether Abstral provides analgesia for BTcP that is clinically
meaningful and not merely statistically significant.
6.1.1.4.
6.1.1.4.1.
Statistical methods
Determination of sample size
Data from SuF-002 were re-analysed to determine the SPID up to 30 minutes in that study. The
mean ± SD treatment difference between Abstral 400 μg and placebo in that study was found to
be 3.61 ± 7.1. Based on this, a sample size of 83 was required to provide 90% power using a 2sided test at the 5% significance level. Assuming a 40% dropout rate, a total of 140 patients
were to be enrolled into the Open-label Titration Phase of the study. In addition, at least 10
patients were required to be randomly assigned to each of the Abstral low dose (100 to 400 μg)
and high dose (600 to 800 μg) subgroups. Therefore, enrolment was planned to continue until
there were 10 patients randomly assigned to each dose subgroup and there were at least 83
patients randomly assigned to treatment in the Double-blind Treatment Phase of the study.
Comment: The sample size was not based on the ability to detect a predefined clinically
meaningful analgesic effect, but rather to detect an effect of the magnitude that was
statistically significant (but not necessarily clinically meaningful) in an earlier study. As a
consequence, statistically significant differences between Abstral and placebo in Study
EN3267-005 need to be interpreted carefully with regard to their magnitude and clinical
relevance.
6.1.1.4.2.
Analysis of the primary efficacy endpoint
The primary efficacy endpoint was analysed using an analysis of variance (ANOVA) model with
fixed effects for treatment, centre, and sequence, and random effect for patient. The observed
margins option was used in estimating the least squares (LS) means for treatment groups to
weight each centre according to the number of patients treated in that centre. Least squares
means, p-values, and 95% CIs of the between-treatment difference were calculated. Missing
data were imputed as follows:

If the baseline value was missing for the first episode in the Double-blind Treatment Phase,
the baseline value from the last episode in the Open-label Titration Phase was used.

For all other episodes in the Double-blind Treatment Phase where the baseline value was
missing, the baseline value from the previous episode was used.

Missing post-baseline values for an episode were imputed using the last observation carried
forward (LOCF) method.
Submission PM-2011-03151-3-1 Extract from the Clinical Evaluation Report for fentanyl citrate (Abstral)
Page 28 of 56
Therapeutic Goods Administration
6.1.1.4.3.
Analysis of secondary efficacy endpoints

PID and PR scores at each post-baseline time point, SPID at 60 minutes after dosing, and
TOTPAR at 30 and 60 minutes after dosing were analysed in the same manner as the
primary endpoint.

The patient global evaluation after each dose of study medication was analysed in the same
manner as the primary endpoint. Results for the patient global evaluation of medication at
the clinic visits were summarised by visit.

The percentage of treated BTcP episodes that required the use of rescue medication was
summarised for each of the 2 treatments and analysed using mixed effects logistic
regression with repeated measures using generalised estimating equations with logit link
and the following covariates: treatment, episode, and centre. Odds ratios and 95% CIs were
also calculated.

The percentage of responders was summarised for each of the 2 treatments and analysed in
the same manner as the use of rescue medication.
6.1.1.4.4.
Interim and end-of-study analyses
An Interim analysis was added at approximately 75% of the planned enrolment. In order to
preserve the overall type I error of 0.05, Pocock’s group sequential procedure was applied to
the analysis of the primary efficacy endpoint. The resulting critical p-values were 0.0414 for the
interim analysis performed at 75% of data, and 0.0239 for the End-of-Study analysis performed
at 100% of the data. The decision rule for the interim analysis based on the primary efficacy
endpoint was:

If the p-value was <0.0414, the efficacy portion of the study would be considered positive,
the efficacy portion of the study would be stopped (ie. the protocol would be amended to
remove the Double-blind Treatment Phase), but patients would continue to be enrolled for
the safety portion of the study (Open-label Titration followed directly by Open-label Longterm Extension).

If the p-value was ≥0.0414 and the effect size was 0.356 or greater, then the study would
continue to its full enrolment of 83 randomised patients. (Note: an effect size of 0.356
corresponded to the study, if it were continued, retaining 50% power to demonstrate a
statistically significant result at the End-of-Study analysis using the adjusted critical p-value
of 0.0239.)

If the p-value was 0.0414 or greater and the effect size was less than 0.356, then the study
would be stopped due to futility.
As it turned out, the first element of the decision rule was satisfied and the Double-blind
Treatment Phase was stopped due to positive efficacy (according to the definition in the
stopping rules). The interim analysis therefore became the primary efficacy analysis. The Endof-Study analysis, which is regarded as secondary in respect of efficacy, included 3 additional
patients who completed the Double-blind Treatment Phase between closure of the database for
the interim analysis and the discontinuation of the Double-blind Treatment Phase. The End-ofStudy analysis also included 12 patients who went directly from the Open-label Titration Phase
to the Long-term Extension Phase.
The statistical analyses of secondary endpoints were not adjusted for the interim analysis or for
multiplicity.
6.1.1.4.5.
Analysis populations
The following populations were used in the efficacy and/or safety analyses:
Submission PM-2011-03151-3-1 Extract from the Clinical Evaluation Report for fentanyl citrate (Abstral)
Page 29 of 56
Therapeutic Goods Administration
6.1.1.4.5.1.
Interim analysis

All Treated Patients (Open-label Titration Phase): Included all patients who received at least
one dose of the open-label titration medication, including the test dose.

All Treated Patients (Double-blind Treatment Phase): Included all patients who received at
least one dose of double-blind medication.

Intent-to-Treat (ITT) Population: This was the primary efficacy analysis population. It
included all randomly assigned patients who received at least one dose of double-blind
study medication and provided baseline and at least one post-baseline pain intensity score
during the Double-blind Treatment Phase.

Per-Protocol (PP) Population: Included all ITT patients with evaluable episodes who were
also compliant with the protocol. A patient must have had at least one evaluable episode to
be included in the PP population. All unevaluable episodes were excluded from the analyses.
Unevaluable episodes were excluded for reasons such as: the pain treated was not target
BTcP, changed the fixed dose of pain medication, incomplete episodes, etc. An incomplete
episode was defined as an episode without baseline and at least one post-baseline pain
intensity assessment before 60 minutes. Evaluable episodes were defined by the following
criteria:
–
Use of no more than one dose of study medication for each BTcP episode.
–
Use of rescue medication no sooner than 30 minutes after study medication use.
–
Allowance of at least 2 hours between BTcP episodes treated with study medication.
–
Use of study medication to treat only target BTcP.
6.1.1.4.5.2.
End-of-study analysis

All Treated Patients (Open-label Titration Phase): Defined as per the Interim analysis.
Demographic data for the study and safety data for the Open-label Titration Phase were
summarised using this population.

All Treated Patients (Double-blind Treatment Phase): Defined as per the Interim analysis.
Safety data for the Double-blind Treatment Phase were summarised using this population.

All Treated Patients (Long-term Extension Phase): Included all patients who received at
least one dose of the open-label medication in the Long-term Extension Phase. Safety data
for the Long-term Extension Phase were summarised using this population.

All Treated Patients (Overall): Included all patients who received at least one dose of Abstral
during the entire study. Safety data during the entire study were summarised using this
population.

Intent-to-Treat Population: Defined as per the Interim analysis.

Per-Protocol Population: Defined as per the Interim analysis.
6.1.1.4.6.
Study patients
Patient disposition proceeded as follows:

136 patients underwent initial eligibility assessment; 5 were excluded from further
participation.

131 patients entered the Open-label Titration Phase; 53 patients discontinued during the
Open-label Titration Phase and 78 completed the Open-label Titration Phase.

Of the 78 patients who completed the Open-label Titration Phase, 66 entered the Doubleblind Treatment Phase, and 12 proceeded directly to the Long-term Extension Phase (after
Protocol amendment 3).
Submission PM-2011-03151-3-1 Extract from the Clinical Evaluation Report for fentanyl citrate (Abstral)
Page 30 of 56
Therapeutic Goods Administration

Of the 66 patients who entered the Double-blind Treatment Phase, 3 were still engaged in
the Double-blind Treatment Phase at the time the Interim (primary) analysis; 63 patients
were therefore included in the Interim (primary) analysis, but 2 of these did not have any
data for the Double-blind Treatment Phase, so the ITT population for the Interim analysis
included 61 patients. The End-of-Study (secondary) analysis of efficacy included all 66
patients. 6 patients discontinued during the Double-blind Treatment Phase and 60 patients
completed the Double-blind Treatment Phase.

72 patients entered the Long-term Extension Phase (60 after completing the Double-blind
Treatment Phase and 12 directly from the Open-label Titration Phase; 47 patients
discontinued during the Long-term Extension Phase and 25 completed the Long-term
Extension Phase (either after 12 months or when the study was discontinued early as
previously described).
A flow chart summarising patient disposition, together with the reasons for discontinuation at
each stage of the study, is presented below in Figure 1.
Figure 1: EN3267-005: Patient flow diagram.
Comment: Of the 131 patients who commenced titration with Abstral:

11 (8.3%) withdrew during titration due to lack of efficacy (this potentially includes
patients who were unable to titrate to a stable Abstral dose)

11 patients (8.3%) withdrew during titration due to treatment-related AEs.

It is also feasible that at least some of the 10 patients who withdrew consent during the
titration period did so because they were unsatisfied with Abstral.
Submission PM-2011-03151-3-1 Extract from the Clinical Evaluation Report for fentanyl citrate (Abstral)
Page 31 of 56
Therapeutic Goods Administration
None of these patients - who in clinical practice would be regarded as treatment failures were included in the efficacy analyses. This means that the so-called ITT population was
not really a true ITT population but one that had been enriched with Abstral responders.
This has implications for the interpretation of the results as will be discussed later.
Long term safety data were available from this study, but only for a very small number of
patients compared to the expected clinical usage if Abstral is marketed.
Briefly, baseline demographic characteristics of the study participants are:

The overall mean age was 55 years, with an age range of 21 to 80 years.

Most of the patients (81%) were in the 18-64 years age group. Only 5 patients aged ≥75
years were enrolled, of whom 2 discontinued during titration and only 1 received doubleblind medication and was included in the efficacy analysis.

46% of the participants were male and 54% were female.

81% were Caucasian, with the remainder being mainly Hispanic or African American.
Data regarding the medications that patients were currently taking for their background cancer
pain and for BTcP at the time of screening were not provided, although what medications had
been used during the 6 months prior to screening were noted.
The median time to reach a stable Abstral dose (amongst those who were able reach a stable
dose) was 7 days. Data regarding the open-label stabilisation dose in the additional 12 patients
who proceeded directly from the Open-label Titration Phase to the Long-term Extension Phase
were not provided, but the distribution of Abstral doses and time to stabilisation would
presumably have been similar.
Comment: Information was not provided regarding the type/location of cancer or the
type/location of the “target” BTcP. This information is relevant to assessing the
generalisability of the study results.
Information was also not provided regarding the baseline severity of pain episodes, either
overall or according to study treatment. This meant that the mean PID and the betweentreatment difference in mean PID at the various observation times could not be assessed in
terms of their magnitudes relative to the baseline pain intensity.
Patients who entered the study had been using opioids (most commonly fentanyl, morphine
or oxycodone) for the treatment of background cancer pain and an opioid or combination
analgesic for BTcP (most commonly hydrocodone/paracetamol, oxycodone/paracetamol,
morphine or oxycodone).
About half of the patients who were successfully titrated required an Abstral dose above
400 μg. Thus, one would expect that in Australian clinical practice about half of the
patients using Abstral would be taking 2 tablets for each episode of BTcP, given the
Sponsor's proposal to not register in Australia the 600 and 800 μg tablet strengths that are
available overseas. The proportion of patients requiring 2 tablets per dose would be
expected to increase over time due to increasing opioid tolerance. Data regarding the
percentage of patients taking each Abstral dose level during the Long-term Extension
Phase were not presented, but the median dose of Abstral during that phase was 600 μg,
compared to 400 μg in the Double-blind Treatment Phase.
6.1.1.5.
6.1.1.5.1.
Efficacy results
Evaluable BTcP episodes
During the Open-label Titration Phase only 16 of 1001 episodes (1.6%) were considered
unevaluable. During the Double-blind Treatment Phase, 14 of 393 episodes (3.6%) treated with
Abstral were unevaluable and 12 of 168 episodes (7.1%) treated with placebo were
Submission PM-2011-03151-3-1 Extract from the Clinical Evaluation Report for fentanyl citrate (Abstral)
Page 32 of 56
Therapeutic Goods Administration
unevaluable. In both phases, the main reason episodes were considered not evaluable was the
use of rescue medication sooner than 30 minutes after the dose of study medication. The
remaining unevaluable episodes were due to patients waiting less than 2 hours between doses
of study medication.
As previously noted, 61 patients provided efficacy data for the Interim (primary) analysis.
However, the related study tables all show “N=57” for placebo.
Comment: The reason for the lower N for placebo is not given in the study report. However,
it can be deduced - from other information in the report - that 4 of the subjects in the ITT
population at the Interim analysis must have provided data for at least one double-blind
dose of Abstral, but not for any doses of placebo (either because they discontinued before
taking any placebo doses or because all of their placebo doses were not evaluable as
described above).
6.1.1.5.2.
6.1.1.5.2.1.
Primary endpoint - double-blind treatment phase
Interim (primary) analysis
Figure 2 shows the LS Mean PID-time curves for Abstral and placebo in the ITT population
during the Double-blind Treatment Phase (Interim analysis).
Figure 2: EN3267-005: LS Mean PID vs time after dose during the Double-blind Treatment
Phase. ITT population, Interim (primary) analysis.
In the ITT population, SPID at 30 minutes after treatment of BTcP with Abstral was significantly
higher (better) than with placebo treatment (p=0.0004; LS mean difference 14.1; 95% CI for LS
mean difference 6.52-21.64 [Table 18]). Note that the adjusted critical p-value for this
comparison was 0.0414 rather than the usual 0.05, due to the added interim analysis.
Table 18: EN3267-005: SPID (units*min) at 30 minutes during the Double-blind Treatment Phase.
ITT population, Interim (primary) analysis.
Submission PM-2011-03151-3-1 Extract from the Clinical Evaluation Report for fentanyl citrate (Abstral)
Page 33 of 56
Therapeutic Goods Administration
The difference between Abstral and placebo remained statistically significant in the perprotocol (PP) population (p=0.0077; LS mean difference 13.76; 95% CI for LS mean difference
3.85-23.67).
6.1.1.5.2.2.
End-of-study (secondary) analysis
In the ITT population, SPID at 30 minutes after treatment of BTcP with Abstral was significantly
higher (better) than with placebo treatment (p=0.0004; LS mean difference 15.0; 95%CI for LS
mean difference 7.00-22.96). Note that the adjusted critical p-value for this comparison was
0.0239 rather than the usual 0.05, due to the added interim analysis.
6.1.1.5.2.3.
Subgroup analyses of SPID
The SPID at 30 and 60 minutes in the ITT and PP populations were examined in several
subgroups, but without statistical testing of differences between Abstral and placebo within or
between the subgroups. The results for the Interim (primary) analysis were as follows:

Gender
–


The mean SPIDs at 30 and 60 minutes favoured Abstral in both males and females but
the between-treatment difference was somewhat higher in women than in men. In the
ITT population, the mean SPID in women was 50.8 at 30 minutes and 153.9 at 60
minutes for Abstral and 36.6 and 109.0, respectively, for placebo. In men, the mean SPID
was 48.1 at 30 minutes and 131.7 at 60 minutes for Abstral and 36.6 and 99.5,
respectively, for placebo. These correspond to (unadjusted) between-treatment
differences of 14.2 in women compared to 11.5 in men for SPID at 30 minutes and 44.9
in women compared to 32.2 in men for SPID at 60 minutes. The same pattern of results
and gender difference was seen in the PP population.
Age group (18-64, 65-74, >74 years)
–
The majority of patients were in the 18 to 64 years age group. In the ITT population,
patients in this age group reported higher (better) mean SPIDs after Abstral than after
placebo at both 30 minutes (51.2 vs 36.6, difference = 14.6) and 60 minutes (145.7 vs
100.6, difference = 45.1). The same pattern was seen in the PP population: 57.4 vs 43.6
(difference = 13.8) at 30 minutes and 160.8 vs 118.5 (difference = 42.3) at 60 minutes.
–
The 65-74 years age group of the ITT population included only 8 patients for Abstral and
7 patients for placebo. In this age group, the difference between Abstral and placebo was
not consistently favourable: The mean SPIDs after Abstral and placebo were 37.3 vs 35.2
at 30 minutes (difference = 2.3), and 121.3 vs 125.0 at 60 minutes (difference = -3.7). In
the PP population (N=6 for Abstral, 5 for placebo) the mean SPIDs were higher (better)
for Abstral than placebo at both time points: 27.7 vs 17.3 (difference = 10.4) at 30
minutes; 98.5 vs 73.1 (difference = 25.4) at 60 minutes.
–
The >74 years age group had only 1 patient in the ITT population, in whom SPID at 30
and 60 minutes was higher (better) after Abstral than after placebo. The PP population
had no patients in this age group.
Route of background opioid medication (oral, transdermal, other)
–
The route of administration of background fixed-schedule opioid medication was
categorised as oral, transdermal or “other”. Some patients were using medication by
both the oral and transdermal routes. In both the ITT and PP populations, for SPID at 30
and 60 minutes, the route of administration of background opioid medication did not
appear to affect the difference between Abstral and placebo.
–
In the ITT population, the mean SPIDs at 30 and 60 minutes were higher (better) after
Abstral than after placebo amongst the 60 patients who were using oral background
Submission PM-2011-03151-3-1 Extract from the Clinical Evaluation Report for fentanyl citrate (Abstral)
Page 34 of 56
Therapeutic Goods Administration
opioid medication: 48.6 vs 35.1 (difference = 13.5) at 30 minutes; 140.5 vs 99.8
(difference = 40.7) at 60 minutes.

–
The mean SPIDs at 30 and 60 minutes were also higher (better) after Abstral than after
placebo amongst the 20 patients who were using transdermal background opioid: 40.1
vs 27.2 (difference = 12.9) at 30 minutes; 127.5 vs 85.2 (difference = 42.3) at 60
minutes.
–
The ITT population only had 1 patient in the “other” route category, in whom SPIDs at
30 and 60 minutes were better after Abstral than after placebo.
Dose group (low and high)
–
Abstral dosage was characterised as low (100 to 400 μg) or high (600-800 μg). About
half of the patients end up in each dose group after titration. In both the low and high
dose groups of the ITT population, the mean SPIDs at 30 and 60 minutes were higher
(better) after Abstral than placebo: 53.1 vs 40.6 at 30 minutes and 152.2 vs 114.0 in the
low dose group; 45.2 vs 31.9 at 30 minutes and 132.1 vs 93.1 at 60 minutes.
Comment: The placebo-subtracted effect of Abstral on SPID was only 11% greater in women
than in men at 30 minutes but about 40% greater in women than in men at 60 minutes. The
corollary of this is that the effect of Abstral on SPID at 60 minutes in men was not only lower
than in women but also lower than in the overall study population. It is unclear whether the
effect of gender was statistically significant, since gender was not included as a term in the
ANOVA models used to analyse the SPID data. It is also unclear whether the lower effect of
Abstral on SPID at 60 minutes in men would remain statistically significant, since there was
no statistical analysis of the efficacy of Abstral within the subgroup of males. As previously
noted, although SPID to 30 minutes was the primary efficacy endpoint in this study, SPID to
60 minutes is arguably more clinically relevant, given the typical duration of an episode of
BTcP. As discussed later, the SPID at 60 minutes has also been shown to bear a relationship
to a clinically meaningful treatment effect. Accordingly, the apparent gender difference in
the placebo-subtracted effect of Abstral on SPID at 60 minutes is a matter that requires
explanation.
The number of elderly patients in the Study EN3267-005 was too small to make any
meaningful conclusions about whether the efficacy of Abstral differed according to age. The
between-treatment differences in mean SPID were noticeably lower in the 65-74 years age
group than in the 18-64 age group in the ITT population, but less so in the PP population.
This suggests that the apparently lower efficacy of Abstral amongst elderly patients in the
study might be due to reduced medication adherence (eg. due to AEs and other factors)
rather than inherently lower efficacy. Whatever the explanation, however, it is reasonable to
conclude that efficacy has not been convincingly demonstrated in the elderly and this should
be reflected in the PI.
The efficacy of Abstral did not appear to be affected by the route of administration (oral or
transdermal) of background opioid medication.
The effect of Abstral was greater than that of placebo in both the high and low dose groups.
The finding in the high dose group provides some reassurance that Abstral doses can be
satisfactorily administered as two tablets, since both dose levels in that group (600 and
800 μg) were administered as two tablets.
6.1.1.5.3.
Secondary endpoints - double-blind treatment phase
As previously noted, the statistical analyses of secondary endpoints were not adjusted for
multiplicity or for the addition of the interim analysis.
6.1.1.5.3.1.
Interim (primary) analysis
In the ITT population:
Submission PM-2011-03151-3-1 Extract from the Clinical Evaluation Report for fentanyl citrate (Abstral)
Page 35 of 56
Therapeutic Goods Administration

SPID at 60 minutes was significantly higher (better) after Abstral doses than after placebo
doses (p=0.0002; LS mean difference 41.6; 95%CI for LS mean difference 20.44-62.79).

PID was significantly higher (better) after Abstral doses than after placebo doses at all time
points. The LS mean difference between treatments was 0.28, 0.55, 0.87 and 0.98 at 10, 15,
30 and 60 minutes, respectively, after the dose).

PR was significantly higher (better) after Abstral doses than after placebo doses at all time
points. The LS mean difference between treatments was 0.15, 0.36, 0.54 and 0.53 at 10, 15,
30 and 60 minutes, respectively, after the dose).

TOTPAR at 30 minutes and TOTPAR at 60 minutes were significantly higher (better) after
Abstral doses than after placebo doses.

BTcP intensity reduction of ≥30% from baseline was seen after a significantly higher
proportion of Abstral doses compared to placebo doses (63.5% vs 41.8%; p<0.0001; Odds
ratio [OR] 2.79; 95% CI for OR 1.74-4.49).

Rescue medication was used after 11.2% of Abstral doses compared to 27.4% of placebo
doses (p-value and OR “not evaluable due to sparse data” using the planned statistical test).

Patient Global Evaluation of Medication was not assessed in the Interim (primary) analysis.
Results in the PP population were consistent with those seen in the ITT population, with
statistically significant differences between Abstral and placebo for: SPID at 60 minutes; PID at
all time points; PR at 15, 30 and 60 minutes (not at 10 minutes); TOTPAR at 30 and 60 minutes;
and the percentage of doses that were followed by a ≥30% reduction in BTcP intensity. Rescue
medication use was numerically less common after Abstral than placebo in the PP population,
but not statistically tested.
6.1.1.5.3.2.
End-of-study (secondary) analysis - double-blind treatment phase
The End-of-Study efficacy analysis of the Double-blind Treatment Phase included the 3
additional patients who were engaged in the Double-blind Treatment Phase at the time of the
Interim analysis and gave similar results to the Interim analysis for all of the secondary efficacy
endpoints that had been examined in the Interim analysis. In relation to other endpoints that
were examined only in the End-of-Study analysis:

Patient Global Evaluation of Medication. The LS mean score was significantly lower
(better) after Abstral compared to placebo (LS mean 3.06 vs 3.64; LS mean difference -0.58;
p=0.0006).

QoL. The timing of the QoL assessments in relation to the Double-blind Treatment Phase - at
Visit 1 (Screening) and Visit 5 (end of the Double-blind Treatment Phase) - combined with
the crossover design, meant that the study could not distinguish between QoL effects related
to Abstral and QoL effects related to placebo.
6.1.1.5.4.
Efficacy analysis - open-label long-term extension phase
The Open-label Long-term Extension Phase of the study did not include a placebo comparison.
Data on the Patient Global Evaluation of Medication over the course of the study were provided
for the 64 ITT patients who completed the Double-blind Treatment Phase. Amongst these 64
patients:

Only 50% had been satisfied or very satisfied with their prestudy (non-Abstral) BTcP
medication; 88% were satisfied or very satisfied with open-label Abstral at the end of
titration.

80% remained satisfied or very satisfied with study treatment (a mix of Abstral and
placebo) at the end of the Double-blind Treatment Period.
Submission PM-2011-03151-3-1 Extract from the Clinical Evaluation Report for fentanyl citrate (Abstral)
Page 36 of 56
Therapeutic Goods Administration

The number of patients remaining in the study at each visit during the Long-term Extension
decreased over time, but amongst those who remained in the study, the proportion who
were satisfied or very satisfied with Abstral remained high at around 90%.

Data were not provided regarding the percentage of patients who were satisfied or very
satisfied with Abstral at their last available assessment.
Comment: The data regarding satisfaction with pre-study BTcP medication compared to
Abstral are very likely to be biased, since patients who were satisfied with their existing
BTcP medication would be unlikely to volunteer for the study. However, the findings do
indicate that Abstral can be a useful alternative in patients who are unsatisfied with their
current BTcP medication.
The information regarding satisfaction during the Long-term Extension is clearly biased in
favour of Abstral since it excludes any patients who discontinued early because they were
unsatisfied with the product.
6.1.1.5.5.
Evaluator’s comments on efficacy in the pivotal study
Statistically significant differences between Abstral and placebo were demonstrated for the
primary efficacy endpoint and most secondary efficacy endpoints.
In clinical practice, treatment with Abstral would be regarded as having failed in patients who
cannot obtain adequate relief of BTcP despite titrating to the maximum dose, who cannot settle
on a stable dose of Abstral for their BTcP, or who are unable to continue taking Abstral because
of treatment-related AEs. In Study EN3267-005, however, these patients were counted as
“screen failures”, removed from the study and excluded entirely from the efficacy analyses. The
population contributing to the efficacy analyses was thus enriched with responders to Abstral,
does not represent a true ITT population, and the resulting comparisons between Abstral and
placebo were biased in favour of Abstral.
The study demonstrates that amongst patients who have been selected because they respond to
and can tolerate Abstral, the short-term effect of Abstral in those patients is greater than that of
placebo. This provides no information about how much better than placebo Abstral would be in
an unselected population or after a prolonged period of use.
The study report, Clinical Overview and Summary of Efficacy asserted that the statistically
significant differences between Abstral and placebo represented a clinically meaningful
treatment effect, but did not provide any evidence to support that assertion. As previously
noted, the sample size was not determined with reference to the demonstration of a predefined
clinically meaningful difference, so statistical superiority over Abstral over placebo cannot be
assumed to represent a clinically meaningful treatment effect.
The sponsor did provide two published papers4 which presented the reasonable argument that
if a patient in a clinical trial of a treatment for BTcP does not require rescue medication in
addition to the study medication for an episode of BTcP, then this of itself represents a clinically
meaningful treatment effect of the study medication in that patient for that episode. Based on
the analysis of published data from placebo- and active-controlled studies of Actiq, the
published papers then showed that this clinically meaningful effect corresponded to:

A maximum %PID of ≥33% during the 60 minutes after a dose, where the %PID is calculated
as the percentage change in PI from the baseline value. This was not analysed in Study
EN3267-003, but a near-equivalent (10% lower) effect is represented by the secondary
endpoint in Study EN3267-005 of a ≥30% decrease in pain intensity from baseline. A ≥30%
reduction in pain intensity is also commonly accepted as a clinically worthwhile treatment
Farrar JT, et al. (2000) Defining the clinically important difference in pain outcome measures. Pain 88: 287-294;
Farrar JT, et al. (2003) Clinically important changes in acute pain outcome measures: a validation study. J Pain
Symptom Manage. 25: 406-11.
4
Submission PM-2011-03151-3-1 Extract from the Clinical Evaluation Report for fentanyl citrate (Abstral)
Page 37 of 56
Therapeutic Goods Administration
effect in studies supporting the registration of analgesic medications. A substantially higher
clinical effect is represented by the secondary endpoint in Study EN3267-005 of a ≥50%
decrease in pain intensity from baseline.

A ≥33% increase in “%Max TOTPAR” over the 60 minutes after a dose. The definition of
“%Max TOTPAR” was not clearly expressed, but the endpoint does not appear to be the
same as TOTPAR in Study EN3267-005.

A PR score of ≥2 (moderate), on a 0-4 scale, during the 60 minutes after a dose. The PR score
at any given time point was determined using the same method as Study EN3267-005, but it
was not clear whether a clinically meaningful effect was represented by a PR score of ≥2 at
any single time point during the 60 minutes after treatment, or a mean PR score of ≥2 over
all time points during the 60 minutes, or an “overall” PR assigned by the patient at the end of
the 60 minutes on the basis of pain relief over the entire period. In any case, information
regarding the frequency with which Abstral and placebo use led to a PR ≥2 was not
presented for Study EN3267-005.

A Global Medication Performance score of ≥2 (good) on a 0-4 scale where 0 = poor and 4 =
excellent. This corresponds to a Patient Global Evaluation of ≤3 (good) on the scale used in
Study EN3267-005 (which ran in the reverse direction from 1 = excellent to 5 = poor).

A SPID of ≥2 over 60 minutes. The SPID was stated to be “the sum of 4 PID measurements
over 60 minutes, divided by 4” (that is, a simple average of the 4 PID measurements).
Examination of the studies on which the analysis was based5 shows that the SPID was
actually calculated as an AUC in a comparable manner to Study EN3267-005, but with the
time component expressed in hours instead of minutes. A SPID ≥2 (units*h) over 60 minutes
using that definition thus corresponds to a SPID ≥120 (units*min) at 60 minutes in EN3267005.
The inherently meaningful efficacy endpoint to which all of the above refer, namely the
avoidance of rescue medication, is also available as a secondary endpoint in EN3267-005.
Having defined a clinically meaningful analgesic effect for an individual dosing episode, one can
then examine whether an effect of that size was significantly more common with Abstral than
with placebo, and if it was, then it follows that the difference between Abstral and placebo must
be clinically meaningful (albeit within the inherently biased population that made it through to
the Double-blind Treatment Phase).
Considering in turn the endpoints from the above list that are available in Study EN3267-005:

SPID at 60 minutes:
–

The presentation of the SPID data in Study EN3267-005 did not include the frequency
with which Abstral and placebo use led to a SPID ≥120 at 60 minutes.
30% or 50% decrease in pain intensity
–
A 30% decrease in pain intensity from baseline was seen after 63.5% of Abstral doses
compared to 41.8% of placebo doses in the Interim (primary) analysis. The odds ratio
for Abstral compared to placebo was 2.79, with a lower 95%CI boundary of 1.73 and pvalue <0.0001.
–
Data regarding a 50% reduction in pain intensity were not assessed in the Interim
analysis but at the End-of-Study analysis this level of response was achieved after 24.7%
of Abstral doses compared to 19.9% of placebo doses. The result favoured Abstral but
Farrar JT, et al. (1998) Oral transmucosal fentanyl citrate: randomized, double-blinded, placebo- controlled trial for
treatment of breakthrough pain in cancer patients. J Natl Cancer Inst. 90: 611-616; Coluzzi PH, et al. (2001)
Breakthrough cancer pain: a randomized trial comparing oral transmucosal fentanyl citrate (OTFC) and morphine
sulfate immediate release (MSIR). Pain 91: 123-130
5
Submission PM-2011-03151-3-1 Extract from the Clinical Evaluation Report for fentanyl citrate (Abstral)
Page 38 of 56
Therapeutic Goods Administration
could not be assessed for statistical significance using the planned method and was not
tested using an alternative method.

Pain relief:
–

Patient Global Evaluation of Medication:
–

The mean difference in PR between Abstral and placebo at each time point was always
less than 2, but this does not necessary mean that Abstral did not produce a clinically
meaningful effect (PR ≥2) at a significantly higher frequency than placebo. However, as
noted above, data regarding the frequency with which Abstral and placebo use led to a
PR ≥2 were not presented.
Data for the Patient Global Evaluation of Medication were only presented for the End-ofstudy analysis. As with PR, the presentation of the data for the Patient Global Evaluation
of Medication did not include the frequency with which Abstral and placebo use led to a
clinically meaningful vale (Patient Global Evaluation ≤3).
Rescue medication was required less frequently after Abstral than placebo, but the
difference could not be assessed for statistical significance using the planned method and
was not tested using an alternative method.
In summary, the data relating to a 30% reduction in pain intensity indicate that the statistically
significant difference between Abstral and placebo was indeed clinically meaningful within the
biased population that was studied. However, the data relating to other endpoints which could
have confirmed this (most importantly SPID at 60 minutes, the use of rescue medication and the
Patient Global Evaluation of Medication) were presented and analysed in a manner that
prevented an assessment of whether Abstral had a clinically meaningful effect on those
endpoints.
It should also be remembered that an unbiased comparison between a candidate medication
and placebo is required if the true population benefit of the drug is to be assessed. As noted
above, the double-blind assessment in EN3267-005 was enriched with proven Abstral
responders and does not provide an unbiased comparison. This means that the true (placebosubtracted) population benefit of Abstral could not be satisfactorily quantified, e.g. in terms of
the “number-needed-to-treat” or NNT (that is, the number of patients who need to be treated
with the drug to produce a clinically worthwhile treatment effect in one additional patient
compared to placebo). In turn, this means that the benefits of Abstral could not be quantitatively
compared with its risks (for example in terms of the “number-needed-to-harm” or NNH).
Finally, the effect of Abstral on the clinically important endpoint of SPID at 60 minutes appeared
to be reduced in men compared to women and this finding has not been adequately explored or
explained.
6.1.2.
Other efficacy studies
SuF-002 was listed as an efficacy study in the submission Table of Contents, but patients only
received a single dose of Actiq at each dose level and a single dose of placebo. It is more
appropriately regarded as a PD/dose-finding study and has therefore been covered elsewhere
in this report.
The open-label study EN3267-007 included baseline and post-baseline assessments of the
Patient Global Evaluation of Medication and QoL, but the lack of a control group means that
these assessments do not provide meaningful evidence of efficacy.
6.1.3.
Analyses performed across trials (pooled & meta analyses)
Not applicable.
Submission PM-2011-03151-3-1 Extract from the Clinical Evaluation Report for fentanyl citrate (Abstral)
Page 39 of 56
Therapeutic Goods Administration
6.1.4.
Evaluator’s conclusions on clinical efficacy for BTcP
The design and analysis of Study EN3267-005 mean that it will have exaggerated the population
benefit of Abstral compared to placebo, and did not permit an unbiased assessment of the NNT.
Nevertheless, the study demonstrates that there is a large subpopulation of BTcP sufferers who
can be identified - via a process of dose titration - as responders to Abstral, and that the
response to Abstral in those individuals is greater than the “placebo effect”.
Study SuF-002 provided only limited supporting efficacy data that involved the treatment of
only a single BTcP episode with placebo and each dose level of Abstral, and did not cover the full
dose range of Abstral as proposed in the PI.
Study EN3267-007 did not provide meaningful efficacy data due to the lack of a control group.
The Abstral dose range that was examined in the submitted studies, and that is now proposed
for approval in Australia, appears to reflect a mistaken belief that the bioavailability of Abstral is
double that of Actiq and therefore that the starting dose, titration steps, and maximum doses of
Abstral should be half those that are approved for Actiq. The recommended Abstral starting
does, titration steps and maximum dose may therefore be unduly conservative. However, it
would be unwise to change the Abstral starting dose, titration steps and maximum dose to
higher ones that have not been studied in any clinical trial. The more conservative titration
steps are also advisable for another reason: unlike an Actiq lozenge which can be easily
removed from the mouth, it is less feasible to terminate absorption from a rapidlydisintegrating and mucosally adherent Abstral sublingual tablet should excessive opioid side
effects start to develop.
7. Clinical safety
7.1.
Studies providing evaluable safety data
7.1.1.
Pivotal efficacy study
In Study EN3267-005, the following safety data were collected:

Adverse events (AEs) were elicited by open-ended questioning and by observation of
patients at each study visit. An AE was any unfavourable or unintended change in signs,
symptoms, laboratory test (if performed), or worsening of a pre-existing condition
associated temporally with the use of the study medication, whether or not it was
considered related to the study medication. All AEs, including observed and volunteered
problems, complaints, signs, or symptoms, were recorded on the electronic case report
form, regardless of whether the AE was associated with the use of study medication. This
included AEs resulting from concurrent illness, reactions to concurrent medication use, or
progression of disease states. Recurrent symptoms of a chronic pre-existing condition were
not considered AEs unless they occurred in a worse or unexpected pattern during study
medication. A treatment-emergent AE (TEAE) was any condition that was not present
before treatment with the study medication but appeared after treatment, was present at
treatment initiation but worsened during treatment, or was present at treatment initiation
but resolved and then reappeared while the individual was on treatment

The oral mucosa was examined at each clinic visit for the duration of the study, preferably
by the same examiner for each patient. Oral erythema and ulcers were recorded as AEs. The
presence, severity, and location (site of sublingual study medication application or other
oral site) of mucositis were recorded. After Protocol Amendment 1, erythematous mucositis
was rated as: 0 = none; 1 = mild (areas of tissue just noticeably red); 2 = moderate (red); 3 =
severe (deep beefy red colour).
Submission PM-2011-03151-3-1 Extract from the Clinical Evaluation Report for fentanyl citrate (Abstral)
Page 40 of 56
Therapeutic Goods Administration

Laboratory tests were specified only at Screening. Additional laboratory tests were
performed during the study in some patients as part of their clinical management, but the
results were generally not reported except in some cases where laboratory abnormalities
were reported as AEs.

Vital signs were assessed at Screening and during the first hour after the test dose of
Abstral. Physical examination and 12-lead ECG were assessed at Screening.
Comment: Throughout Study EN3267-005, patients received background opioid therapy
for their cancer pain as well as a range of concomitant medications. The reported TEAEs
therefore represent a combination of the effects of Abstral, of background and concomitant
medications, and of the patients underlying cancer and other medical conditions. However
a causal relationship to Abstral could reasonably be deduced for some TEAEs, particularly
those that occurred during the titration period and that represented the new onset of a
typical opioid-type side effect or hypersensitivity reaction.
The multiple crossover design of the Double-blind Treatment Phase meant that TEAEs
during that phase could not be reliably assigned to one or the other of Abstral and placebo.
This means that the pivotal study does not provide placebo-controlled safety data, does not
allow a determination of the placebo-subtracted TEAE rate for Abstral and does not permit
calculation of the NNH.
7.1.2.
7.1.2.1.
Other studies
EN3267-007
Study EN3267-007 was a long-term, open-label, uncontrolled study that enrolled opioidtolerant cancer pain patients. After screening, eligible patients entered a Titration Period during
which participants had up to 2 weeks to determine a single effective and tolerable dose of
Abstral in the range 100 to 800 μg. The titration scheme was essentially the same as in the
pivotal Study EN3267-005. Successfully-titrated patients then entered an open-label
Maintenance Period of up to 12 months during which they used Abstral (maximum dose 800 μg)
to treat episodes of BTcP. Data were collected regarding AEs (including specific assessments of
oral mucosal effects), clinical chemistry, haematology, urinalysis, ECG, vital signs and physical
examination.
7.1.2.2.
Clinical pharmacology studies
Data relating to AEs, laboratory tests, physical examination, vital signs and ECG were collected
during the clinical pharmacology studies. Data from the clinical pharmacology studies in the
original submission were aggregated by the sponsor in the Summary of Clinical Safety.
Corresponding safety information from the clinical pharmacology studies submitted as
supplementary data were not included in the Summary of Clinical Safety but were later provided
separately for each study in the evaluation.
7.2.
Pivotal studies that assessed safety as a primary outcome
Not applicable. Study EN3267-007 primarily examined the safety of Abstral, but was
uncontrolled and therefore not regarded as pivotal.
7.3.
Patient exposure
7.3.1.
Pivotal study
In Study EN3267-005:

131 cancer pain patients received at least one dose of Abstral.

52 patients were exposed to Abstral for ≥3 months.
Submission PM-2011-03151-3-1 Extract from the Clinical Evaluation Report for fentanyl citrate (Abstral)
Page 41 of 56
Therapeutic Goods Administration

The mean exposure duration was 123.7 days (giving a cumulative exposure during the study
of 131×123.7 or 16,205 patient-days). The median exposure duration was 51 days.

Patients took a cumulative total of 41,596 doses of Abstral during the study (3,167 during
the Open-label Titration Phase, 414 during the Double-blind Treatment Phase and 38,015
during the Long-Term Extension Phase).

About half of the patients who were successfully stabilised on Abstral were taking a dose of
100-400 μg and half were on 600-800 μg.

Amongst the 72 patients who entered the Long-term Extension Phase, the mean dose of
Abstral per BTcP episode over the course of the Long-term Extension Phase was 551 μg,
with a median dose of 600 μg and a range of 100-800 μg.
7.3.2.
7.3.2.1.
Other studies
EN3267-007
In the long-term open-label safety study EN3267-007:

139 cancer pain patients received at least one dose of Abstral.

62 patients were exposed to Abstral for ≥3 months.

The mean exposure duration was 138.8 days (giving a cumulative exposure during the study
of 139 ×138.8 or 19,293 patient-days). The median exposure duration was 79 days.

About half of the 96 patients who were successfully stabilised on Abstral were taking a dose
of 100-400 μg and half were on 600-800 μg.
7.3.2.2.
Clinical pharmacology studies
In the clinical pharmacology studies in the original submission (including Suf-002), a total of
221 healthy subjects and 41 cancer pain patients received at least one dose of Abstral. Healthy
subjects received from 1 to 20 doses of Abstral, at dose levels of 50 to 800 μg. Cancer pain
patients received from 1 to 3 doses of Abstral at dose levels of 100 to 400 μg.
In the supplementary clinical pharmacology studies, a total of 162 healthy subjects received at
least one dose of Abstral, at dose levels of 400 or 800 μg.
7.4.
Adverse events
7.4.1.
All adverse events (irrespective of relationship to study treatment)
7.4.1.1.
7.4.1.1.1.
Pivotal study
Adverse events in general
In Study EN3267-005, 96 patients (73.3%) reported at least 1 TEAE. TEAEs reported in ≥5% of
patients overall are summarised in Table 19. In descending order of incidence, they were
nausea (22.1% of patients); vomiting (11.5%); fatigue (8.4%); anaemia, diarrhoea, stomatitis
(6.9% each); dizziness, headache, somnolence, weight decreased, urinary tract infection (6.1%
each), asthenia, pain in extremity, pneumonia, upper respiratory tract infection (5.3% each).
Submission PM-2011-03151-3-1 Extract from the Clinical Evaluation Report for fentanyl citrate (Abstral)
Page 42 of 56
Therapeutic Goods Administration
Table 19: EN3267-005: Summary of TEAEs reported in ≥5% of patients overall, according to SOC,
Preferred Term and study phase. Final analysis, All-treated population.
Comment: The most common TEAEs were a combination of typical opioid side effects and
AEs that are expected given the patients’ underlying cancers and the treatment of those
cancers. Constipation does not feature in the above list, but patients would have been using
preventative treatment for constipation due to their background opioid therapy. A wide
range of TEAEs were reported at frequencies below 5%, which is again not surprising
given the patient’s background medication and underlying medical conditions.
7.4.1.1.2.
Adverse events related to the application site
Two patients had evidence of oral mucositis at screening. During the Open-label Titration Phase
(Visit 2 and Visit 3), 6/131 patients (4.9%) had evidence of oral mucositis. During the Doubleblind Treatment Phase (Visit 4 and Visit 5), 3/131 patients (2.3%) had evidence of oral
Submission PM-2011-03151-3-1 Extract from the Clinical Evaluation Report for fentanyl citrate (Abstral)
Page 43 of 56
Therapeutic Goods Administration
mucositis. The number of patients with evidence of oral mucositis did not increase during the
Long-term Extension Phase. Throughout the study, there were fewer patients reported to have
both ulcers and erythema than patients with erythema alone. In all cases, the erythema was
reported as mild or moderate.
One patient experienced oral mucosal blistering that was reported as a TEAE. The oral mucosal
blistering in this patient was present during the Open-label Titration Phase and the Doubleblind Treatment Phase, and resolved after the patient had entered the Long-term Extension
Phase despite continuing to take Abstral.
TEAEs that could potentially represent local adverse effects of Abstral at or near the application
site included: stomatitis (9 patients [7%]); pharyngolaryngeal pain (4 patients [3.1%]); lip
ulceration, dry mouth (3 patients [2.3%] each); dry mouth (3 aphthous stomatitis, oral pain,
tongue ulceration, gingival swelling, gingival ulcer, gingivitis, dysgeusia, throat tightness (1
patient [0.8%] each). Of these, 2 cases of dry mouth, and 1 case each of stomatitis, aphthous
stomatitis, gingival ulceration, lip ulceration, pharyngolaryngeal pain, dysgeusia and throat
tightness were classified as treatment-related.
7.4.1.2.
7.4.1.2.1.
Other studies
EN3267-007
In the long-term open-label safety study EN3267-007, 116 patients (83.5%) reported at least
one TEAE. TEAEs reported in ≥5% of patients overall are summarised in Table 20. In
descending order of incidence, they were nausea (23.0% of patients); fatigue (15.1%); vomiting
(12.9%); dehydration (11.5%); peripheral oedema (10.8%); arthralgia (10.1%); back pain,
insomnia (9.4% each); asthenia, constipation (8.6% each), anorexia, cancer pain, diarrhoea,
somnolence, stomatitis (7.9% each); anaemia, dizziness (7.2% each); headache (6.5%);
abdominal pain, anxiety, depression, dry mouth, hypotension, weight decreased (5.8% each);
rash, thrombocytopenia (5.0% each). Overall, a similar percentage of men and women
experienced TEAEs (82.5% and 84.2%, respectively. The most commonly reported TEAEs in
women were nausea (28.9%), arthralgia, back pain, and vomiting (13.2% each), whereas the
most commonly reported TEAEs in men were dehydration (19.0%), fatigue (15.9%), and nausea
(15.9%).
Submission PM-2011-03151-3-1 Extract from the Clinical Evaluation Report for fentanyl citrate (Abstral)
Page 44 of 56
Therapeutic Goods Administration
Table 20: EN3267-007: Summary of TEAEs reported in ≥5% of patients overall, according to SOC,
Preferred Term and study phase. All-treated population (N=139).
Submission PM-2011-03151-3-1 Extract from the Clinical Evaluation Report for fentanyl citrate (Abstral)
Page 45 of 56
Therapeutic Goods Administration
Table 20 (continued): EN3267-007: Summary of TEAEs reported in ≥5% of patients overall,
according to SOC, Preferred Term and study phase. All-treated population (N=139).
A total of 18/139 patients (12.9%) showed evidence of oral mucositis during the course of the
study. The number of patients with evidence of oral mucositis did not increase during the
Maintenance Period with only one patient reporting oral mucositis at End-of-Study. TEAEs that
might represent local adverse effects of Abstral but which did not necessarily meet the criteria
for oral mucositis included stomatitis (11 patients); dry mouth (8 patients); mouth ulceration (4
patients); gingival bleeding, gingivitis, oral hypaesthesia, oral discomfort, oral soft tissue
disorder, tongue disorder, tongue dry, tongue ulceration, application site irritation, application
site reaction, and mucosal inflammation, (1 patient each).
7.4.1.2.2.
Clinical pharmacology studies
TEAEs reported in ≥1% of healthy volunteers or cancer patients in the clinical pharmacology
studies in the original submission are summarised in Table 21. In the originally-submitted
clinical pharmacology studies, TEAEs were reported in 71% of the healthy subjects after Abstral
administration and 49% of the cancer patients. In the supplementary clinical pharmacology
studies, TEAEs were reported in approximately one-third of the Abstral recipients, all of whom
were healthy volunteers. The lower incidence of TEAEs in the supplementary clinical
pharmacology studies reflects the fact that naltrexone was used in all of those studies to reduce
the risk of opioid-related AEs. TEAEs in the supplementary clinical pharmacology studies were
broadly similar in nature to those summarised in Table 21.
Submission PM-2011-03151-3-1 Extract from the Clinical Evaluation Report for fentanyl citrate (Abstral)
Page 46 of 56
Therapeutic Goods Administration
Table 21: Number (%) of participants reporting TEAEs in the clinical pharmacology studies in the
original submission (AEs reported in ≥1% of healthy subjects or cancer patients).
† All 3 cases of tremor were reported in healthy subjects in StudyEN3267-001, where they were correctly
included in the ‘Nervous system disorders’ SOC. These events were incorrectly included in the ‘Eye disorders’
SOC. They were also incorrectly included in the totals for the ‘Eye disorders’ SOC in Clinical Summary. The
evaluator has moved them to the correct SOC and adjusted the number of healthy subjects and the total
number of participants with AEs in the ‘Eye disorders’ SOC. The number of healthy subjects and the total
number of participants with AEs in the ‘Nervous system disorders’ SOC has not been adjusted because:
(a) The evaluator was unable to determine whether or not the patients who experienced tremor are already
included in the total for the ‘Nervous system disorders’ SOC (due to also having some other nervous system
AE). The listing that would have clarified this was not included in the submission (Study EN3267-001).
Submission PM-2011-03151-3-1 Extract from the Clinical Evaluation Report for fentanyl citrate (Abstral)
Page 47 of 56
Therapeutic Goods Administration
(b) The maximum possible number of additional healthy volunteers with AEs in the ‘Nervous system disorders’
SOC is 3, which is very small compared to the number already appearing in that SOC (133).
* Examination of the original terms used by the investigators show that the dermatitis/irritation occurred at
ECG recording tab sites, pCO2 probe sites and BP cuff sites, not at the drug application site.
7.4.2.
7.4.2.1.
Treatment-related adverse events (adverse drug reactions)
Pivotal study
In Study EN3267-005, TEAEs judged to be at least possibly related to study treatment were
reported in 41/131 (31.3%) of the patients. Treatment-related TEAEs reported in more than
2% of patients were nausea (12.2% of patients), vomiting (5.3%), somnolence (4.6%) and
headache (3.8%). Treatment-related rash, allergic pruritus and drug hypersensitivity were
reported in 1 patient each. As previously noted, treatment-related TEAEs of dry mouth (2
cases), stomatitis, aphthous stomatitis, gingival ulceration, lip ulceration, pharyngolaryngeal
pain, dysgeusia and throat tightness (1 case each) may represent adverse effects of Abstral at or
near the application site.
7.4.2.2.
7.4.2.2.1.
Other studies
EN3267-007
In the long-term open-label safety study EN3267-007, treatment-related TEAEs were reported
in 49/139 (35.3%) of the patients. Treatment-related TEAEs reported in more than 2% of
patients were nausea (8.6% of patients); constipation, somnolence (5.8% each); fatigue (4.3%);
dry mouth, headache (3.6%); dizziness (2.9%) and vomiting (2.2%).
7.4.2.2.2.
Clinical pharmacology studies
Treatment-related TEAEs in the clinical pharmacology studies in the original submission are
summarised in Table 22. In the supplementary clinical pharmacology studies, 37/162 (22.8%)
of the Abstral-treated subjects experienced treatment-related TEAEs. These are summarised in
the individual study summaries and were broadly similar in nature to those summarised in
Table 22 (but generally occurred at lower frequencies than in the originally-submitted clinical
pharmacology studies due to the concomitant administration of naltrexone).
Submission PM-2011-03151-3-1 Extract from the Clinical Evaluation Report for fentanyl citrate (Abstral)
Page 48 of 56
Therapeutic Goods Administration
Table 22: Number (%) of participants reporting TEAEs that were categorised as being at least
possibly related to study treatment in the clinical pharmacology studies in the original submission
(AEs reported in ≥1% of healthy subjects or cancer patients).
*Data for application site dermatitis and application site irritation were transposed in the source table in the
Clinical Summary (corrected in the above table). In addition, examination of the original terms used by the
investigators show that the dermatitis/irritation occurred at ECG recording tab sites, pCO2 probe sites and BP
cuff sites, not at the drug application site.
The total number and percentage of study participants with treatment-related AEs was not reported, nor were
the number and percentage of patients with treatment-related AEs in each SOC.
7.4.3.
7.4.3.1.
Deaths and other serious adverse events
Pivotal study
A total of 10 (7.6%) patients died in Study EN3267-005: 2 during the Open-label Titration Phase
and 8 during the Long-term Extension (Table 23). Six deaths (6/131 patients [4.6%]) were
considered to be a result of progression of the underlying disease (metastatic cancer). Other
causes of death included pneumonia, septic shock, myocardial infarction, and completed suicide
(one of each).
Submission PM-2011-03151-3-1 Extract from the Clinical Evaluation Report for fentanyl citrate (Abstral)
Page 49 of 56
Therapeutic Goods Administration
Table 23: EN3267-005: Number (%) of patients who died, with the fatal TEAEs listed according to
SOC, Preferred Term and study phase. End-of-study analysis, All-treated population.
None of the deaths were considered to be related to study medication. The event of completed
suicide did not involve overdosing with study medication.
A total of 24 patients (18.3%) experienced at least one treatment-emergent SAE: 6 (4.6%)
patients in the Open-label Titration Phase and 18 (25%) patients in the Long-term Extension
Phase). Most SAEs were considered to be related to the patients’ underlying medical conditions
rather than study treatment. SAEs that were classified as at least possibly related to Abstral
were:

1 case of mild affect lability during Abstral titration. Abstral was discontinued and the event
resolved.

1 case of non-fatal accidental overdose with opioids and lorazepam during the Long-Term
Extension in a woman who was taking multiple medications, including 3 different opioids
(oxycodone 180 mg daily; morphine 30 mg as needed, total daily amount not confirmed; and
Abstral 100μg per dose, total daily amount not confirmed).
7.4.3.2.
7.4.3.2.1.
Other studies
EN3267-007
A total of 19 (13.7%) patients died in the long-term open-label safety study EN3267-007. None
of the deaths were considered to be related to Abstral. A total of 46 patients (33.1%)
experienced at least one treatment-emergent SAE: 7 (5.0%) patients in the Titration Period and
40 (41.7%) patients in the Long-term Maintenance Period. All of the SAEs were considered to be
related to the patients’ underlying medical conditions or concomitant therapy, and none were
considered to be related to Abstral.
Submission PM-2011-03151-3-1 Extract from the Clinical Evaluation Report for fentanyl citrate (Abstral)
Page 50 of 56
Therapeutic Goods Administration
7.4.3.2.2.
Clinical pharmacology studies
There were no deaths in the originally-submitted or supplementary clinical pharmacology
studies. SAEs occurred only in the originally submitted clinical pharmacology studies, with no
SAEs in the supplementary clinical pharmacology studies.
One SAE in the clinical pharmacology studies was considered to be related to Abstral. A healthy
male subject who had not received pre-emptive naltrexone experienced severe vomiting after
the fourth dose of Abstral 200 μg q4h. This was considered an SAE due to the need for
intervention (IV fluids, naloxone and metoclopramide) and the severity of the event. The SAE,
classified as probably related to study treatment, resolved the next day. Further investigation
revealed the presence of non-typhoid salmonella in the subject’s stool sample taken at the time
of discharge from the study. Nonetheless, the SAE was kept as ‘probably’ related to study drug.
7.4.4.
Discontinuation due to adverse events
7.4.4.1.
Pivotal study
A total of 30 of 131 patients (22.9%) experienced at least one TEAE that led to discontinuation:
15 patients in the Open-label Titration Phase, 2 patients in the Double-blind Treatment Phase,
and 13 patients in the Long-term Extension Phase). The most common SOCs for TEAEs that led
to discontinuation during the Open-label Titration Phase were gastrointestinal disorders
(4/131 patients [3.1%]) and psychiatric disorders (3/131 patients [2.3%]). The most common
SOC for TEAEs that led to discontinuation during the Long-term Extension Phase was neoplasms
benign, malignant, and unspecified (4/72 patients [5.6%]).
A total of 15/131 (11.5%) patients discontinued due to a total of 17 treatment-related TEAEs.
The events were: dyspnoea, nausea (2 each); abdominal pain, affect lability, allergic pruritus,
anxiety, diarrhoea, disorientation, drug hypersensitivity, headache, mental status changes,
paranoia, rash, throat tightness, and vomiting (1 each).
Comment: The data show the expected pattern of early discontinuations due to treatmentrelated TEAEs (particularly during titration), with later discontinuations that were mostly
related to progression of the underlying cancer. However some patients did discontinue
during the Long-term Extension Phase due to treatment-related TEAEs (abdominal pain,
allergic pruritus and dyspnoea).
7.4.4.2.
7.4.4.2.1.
Other studies
EN3267-007
Overall, 37 patients (26.6%) experienced TEAEs leading to discontinuation during EN3267-007:
14/139 (10.1%) in the Titration Period and 23/96 (24.0%) in the Maintenance Period. Overall,
metastatic prostate cancer and nausea (each in 5 patients) were the most frequently cited
TEAEs leading to discontinuation. The most common TEAEs resulting in discontinuation during
the Titration Period were nausea (3 patients), somnolence (3 patients), metastatic prostate
cancer (2 patients), and fatigue (2 patients). In the Maintenance Period, the most frequently
reported events causing discontinuation were metastatic prostate cancer (3 patients),
metastatic lung cancer (2 patients), and nausea (2 patients). A total of 11/139 (7.9%) patients
discontinued due to a total of 12 treatment-related TEAEs. The events were: nausea (3
patients); nausea and depression (1 patient); somnolence (3 patients); fatigue, lethargy,
headache , pruritus (1 patient each). Of the treatment-related withdrawals, 10 occurred during
the Titration Period and the remaining one was during the first month of the Maintenance
Period.
7.4.4.2.2.

Clinical pharmacology studies
A total of 14 volunteers discontinued due to AEs (4 from Study 2246-EU-004, 1 from Study
2246-EU-005, 5 from EN3267-001, 3 from EN3267-003 and 1 from EN3267-013). The AE
leading to the discontinuation of the largest number of volunteers was respiratory
Submission PM-2011-03151-3-1 Extract from the Clinical Evaluation Report for fentanyl citrate (Abstral)
Page 51 of 56
Therapeutic Goods Administration
depression, in 4 volunteers (all opioid naïve, from a study where pre-emptive opioid
antagonist was not administered).

No cancer patients in the clinical pharmacology studies discontinued treatment due to
TEAEs.
7.5.
Laboratory tests
7.5.1.
Pivotal study
Post-baseline laboratory data were not routinely collected in Study EN3267-005. Post-baseline
laboratory tests were performed in some patients as part of their clinical management, but the
results were not presented in the study report. Clinically significant abnormalities of these tests
could be reported as TEAEs.

Clinical chemistry abnormalities reported in this manner were: hypokalaemia (3 patients
[2.3%]); abnormal liver function test (2 patients [1.5%]); increased alkaline phosphatase,
increased creatinine, increased carcinoembryonic antigen, increased testosterone
hypoglycaemia, hyponatraemia (1 patient [0.8%] each). Hepatitis and jaundice were each
reported in 1 patient, as were proteinuria and renal failure.

Haematological abnormalities reported in this manner were anaemia (9 patients [6.9%]),
neutropenia (5 patients [3.8%]), pancytopenia (3 patients [2.3%]), thrombocytopenia (2
patients [1.5%] and leukopenia (1 patient [0.8%]).
None of these abnormalities were classified as treatment-related.
7.5.1.1.
7.5.1.1.1.
Other studies
EN3267-007
Laboratory parameters were examined at intervals during the long-term open-label safety study
EN3267-007. Screening and subsequent laboratory parameters reflected a population in a
diseased state, and many patients had abnormal baseline and post-baseline values. The lack of a
control group also meant that any effects of Abstral could not be separated from changes due to
the underlying cancer, other concomitant illnesses, or concomitant medications.
7.5.1.1.2.
Clinical pharmacology studies
Laboratory parameters were examined before and after Abstral administration in the clinical
pharmacology studies.

One subject developed a clinically-significantly raised alanine transaminase (ALT) level of
80 IU/L (normal range: 8-45 IU/L), one day after his fourth dose of Abstral (200 μg, after
previous doses of 50, 100 and 150 μg). ALT was still raised (70 IU/L) one day later and had
returned to normal (16 IU/L) 12 days later. The abnormality was classified as possibly
treatment-related.

No other clinically significant treatment-related clinical chemistry or haematology changes
were observed in the clinical pharmacology studies.
7.5.2.
7.5.2.1.
Other clinical trial safety assessments
ECG
In the pivotal Study EN3267-005, ECGs were routinely performed only at screening.
In the long-term open-label safety study EN3267-007, ECGs were recorded at baseline and at
intervals during the study. There were no clinically significant ECG changes during the study.
Submission PM-2011-03151-3-1 Extract from the Clinical Evaluation Report for fentanyl citrate (Abstral)
Page 52 of 56
Therapeutic Goods Administration
In the clinical pharmacology studies in the original submission, ECGs were performed before
and after Abstral administration. No clinically relevant abnormalities were seen. ECGs were
routinely performed only at screening in the supplementary clinical pharmacology studies.
7.5.2.2.
Vital signs
In the pivotal Study EN3267-005, vital signs were routinely recorded only at screening.
In the long-term open-label safety study EN3267-007, vital signs were recorded at baseline and
at interval during the study. There were some individual abnormalities, as expected given the
study population, but mean values were similar at screening and End-of-Study.
In the clinical pharmacology studies, vital signs were recorded before and after Abstral
administration. The administration of higher doses of Abstral to opioid-naïve subjects was
sometimes followed by decreases in respiratory rate, oxygen saturation and blood pressure. No
other clinically relevant abnormalities were seen.
7.6.
Post-marketing experience
From first international launch to 30 April 2010, the sponsor received a total of 30 case reports
with 52 suspected adverse reactions. The majority (71%) of the adverse reactions were
considered non-serious by the reporters.
One fatal case was received. This case was a poorly documented regulatory authority report of a
patient who had died (cause of death not reported) in whom “high levels” (not quantified) of
fentanyl were found at post-mortem. Other relevant information was not available, for example:
the specific fentanyl product(s) involved; route of administration; indication for the patient’s
use of fentanyl; whether the patient was opioid-tolerant or opioid-naïve and whether the “high
levels” occurred as a result of usage at the recommended dose, or deliberate or accidental
overdose).
A number of the spontaneously-reported post-market AEs represent possible local effects of
Abstral at or near the application site.
Comment: The suspected post-market adverse reactions comprise a mixture of systemic
opioid effects and administration site adverse effects. The data are now 2 years out of date
and an updated summary of post-marketing adverse event reports should be sought from
the Sponsor.
7.7.
Evaluator’s overall conclusions on clinical safety
In the submitted studies, Abstral was associated with typical opioid-type adverse effects, the
most common being gastrointestinal effects such as nausea and vomiting, and effects reflecting
the CNS depressant action of fentanyl, such as somnolence, fatigue and asthenia. A small
percentage of patients had AEs that could represent local effects of Abstral at or near the
application site. Hypersensitivity/allergic reactions and rashes were also seen in a small
percentage of patients. Overall, the types of AEs are what one would expect for a sublingually
administered fentanyl product.
Long term safety data were relatively limited, being available from only 170 patients 72 in
EN3267-005 and 98 in EN3267-007). However, no unexpected safety issues were identified and
the systemic safety profile of Abstral is expected to be broadly similar to that of Actiq (based on
the similarity of fentanyl plasma concentrations during use of the two products).
The Abstral studies did not reveal any major problems in relation to AEs at or near the
application site, but the database is limited. Given the underlying frequency with which oral AEs
occur in the general population and in patients with cancer, it is unlikely that preregistration
studies will be able to either detect or exclude, with a satisfactory degree of confidence, an
Submission PM-2011-03151-3-1 Extract from the Clinical Evaluation Report for fentanyl citrate (Abstral)
Page 53 of 56
Therapeutic Goods Administration
increased risk of such events in users of Abstral. Rather, if Abstral use is associated with oral
adverse effects, then this would only become apparent after marketing and consequent usage in
a much larger number of patients (as was the case, for example, with Actiq).
The Sponsor’s proposal to not register the 300 μg, 600 μg and 800 μg tablet strengths in
Australia means that some patients will be titrated to doses that can only be achieved by using
two different strength tablets, with a consequent risk of administration error that could be
avoided if these tablet strengths were available. In addition, the need to take two tablets instead
of one (for titration and also to achieve doses above 400 μg) may have an adverse impact on
compliance, that could be avoided if the sponsor were to market in Australia all of the strengths
that are available overseas.
As expected, the incidence of opioid-type AEs, including respiratory depression, was relatively
high amongst non-opioid-tolerant healthy subjects in the clinical pharmacology studies. This
serves to underline the recommendation that Abstral should only be used in patients who are
opioid-tolerant.
8. First round benefit-risk assessment
The design and analysis of Study EN3267-005 mean that it will have exaggerated the population
benefit of Abstral compared to placebo, and did not permit an unbiased assessment of the NNT.
Nevertheless, the study demonstrates that there is a large subpopulation of BTcP sufferers who
can be identified - via a process of dose titration - as responders to Abstral, and that the
response to Abstral in those individuals is greater than the “placebo effect”. The dose-titration
process by which responders are identified is associated with AEs, so that some individuals who
start Abstral will therefore experience AEs without eventually gaining any benefit from the
drug. However, treatment-related AEs during the titration period were generally minor and
transient, and overall the benefit-risk balance is considered to be favourable.
9. First round recommendation regarding authorisation
Authorisation of Abstral is recommended, provided that:

The PI and Consumer Medicine Information (CMI) are revised to more accurately reflect the
available information.

The sponsor provides satisfactory responses to the clinical questions.
10. Clinical questions
The Periodic Safety Update Reports (PSURs) included in the submission are now 2 years old.
The sponsor should provide up to date PSURs for review before authorisation.
The sponsor should be asked to comment on the gender difference in Sum of Pain Intensity
Differences (SPID), particularly at 60 minutes after the dose, in the pivotal Study EN3267-005.
As noted in the Clinical Pharmacology Study Summaries, a few minor issues could be addressed
if the Sponsor were to make available certain appendices to the study reports that were omitted
from the material provided to the TGA. However, these minor issues would not affect the
registration decision, so it is not necessary to obtain or evaluate the data.
Submission PM-2011-03151-3-1 Extract from the Clinical Evaluation Report for fentanyl citrate (Abstral)
Page 54 of 56
Therapeutic Goods Administration
11. References
1.
Center For Drug Evaluation and Research. Application number 022510Orig1s000.
Summary Review
<http://www.accessdata.fda.gov/drugsatfda_docs/nda/2011/022510Orig1s000SumR.pdf
>
Accessed 2 March 2012.
2.
Center For Drug Evaluation and Research. Application number 022510Orig1s000. Clinical
Pharmacology and Biopharmaceutics Review(s).
<http://www.accessdata.fda.gov/drugsatfda_docs/nda/2011/022510Orig1s000ClinPharm
R.pdf>
Accessed 2 March 2012.
3.
Australian Product Information for Actiq (30 Jan 2012).
4.
Stengel B. (2010) Chronic kidney disease and cancer: a troubling connection. J Nephrol. 23:
253-262.
5.
Farrar JT, et al. (2000) Defining the clinically important difference in pain outcome
measures. Pain 88: 287-294.
6.
Farrar JT, et al. (2003) Clinically important changes in acute pain outcome measures: a
validation study. J Pain Symptom Manage. 25: 406-411.
7.
Farrar JT, et al. (1998) Oral transmucosal fentanyl citrate: randomized, double-blinded,
placebo- controlled trial for treatment of breakthrough pain in cancer patients J Natl Cancer
Inst. 90: 611-616.
8.
Coluzzi PH, et al. (2001) Breakthrough cancer pain: a randomized trial comparing oral
transmucosal fentanyl citrate (OTFC) and morphine sulfate immediate release (MSIR). Pain
91: 123-130.
Submission PM-2011-03151-3-1 Extract from the Clinical Evaluation Report for fentanyl citrate (Abstral)
Page 55 of 56
Therapeutic Goods Administration
PO Box 100 Woden ACT 2606 Australia
Email: [email protected] Phone: 1800 020 653 Fax: 02 6232 8605
http://www.tga.gov.au