qbd – understanding how excipient properties influence ... workshop qbd.pdf · qbd –...
TRANSCRIPT
QbD – Understanding How Excipient Properties
Influence Solid Oral Dosage Form Performance
Dr Amina Faham (Dow), Dr Liz Meehan (AstraZeneca)
ExcipientFest, Amsterdam NL
June 24, 2014
What do you understand by the term QbD,
in particular applied to excipients?
AstraZeneca – Do not share without permission
Traditional versus QbD approach
• In traditional approaches, industry focused on:
– Similar excipient lots are used during development and in
commercial manufacturing (avoiding variation)
– Optimized, fixed formulation and fixed process parameters
– Compliance with compendial specifications for excipients
• QbD approach encourages:
– Understanding variation of excipients properties as they relate
to critical process parameters and product quality attributes
– Building robustness and flexibility into manufacturing process
– Excipient specifications appropriate to ensure product quality
AstraZeneca – Do not share without permission
Product Quality Attributes –
Source and Effect/s
API
Variability Excipient
Variability
Process
Variability
Product
Variability
σσσσσ2
nsInteractio
2
Process
2
Excipients
2
API
2
Product
Ref: C. Moreton
Understanding
variability &
tolerating it
= Robustness
4 AstraZeneca – Do not share without permission
Excipient functionality and performance
• Quantitative performance requirements (i.e. critical material
attributes) of excipients
• Characterisation of excipients to determine their suitability for
intended use
• Must be evaluated and controlled to ensure consistent
performance throughout the product life-cycle (e.g. changes
in suppliers)
• Integral to the "Quality by Design" approach that should be
employed in drug product development
AstraZeneca – Do not share without permission
Quality by Design
API Excipients
Processing
Material
attributes
Material
attributes
Intermediate
attributes
Process
parameters
Drug
product
Product
attributes Safety and
efficacy
CMA CMA
CQA
CPP CMA
Spec range
MSA
AstraZeneca – Do not share without permission
Quality by Design
CQA=Critical quality attributes of the product
CMA=Critical material attributes of all input raw materials
CPP=Critical process parameters
MSA=measurement systems analysis
Target Drug Product Profile
CQA = f (CMA, CPP)
AstraZeneca – Do not share without permission
Why QbD for excipients?
• Excipient properties can affect CQAs of drug product
– Manufacturability (e.g. flow, compaction)
– Content uniformity (e.g. segregation)
– Bioavailability (e.g. disintegration, dissolution)
– Purity
– Stability (e.g. chemical and physical incompatibilities)
• It is important to understand and control the effects of
excipient variability
AstraZeneca – Do not share without permission
Challenges
• Excipients developed and manufactured specifically for
pharmaceutical use are often available in a range of special
grades (developed for specific formulation or process)
• There are multiple suppliers of nominally the same grade
– lot-to-lot/batch-to-batch/supplier inequivalence or variability
– variability in excipient properties should be anticipated and
appropriate controls must be in place to ensure consistent
performance
• Excipient applications for pharmaceutical development are
many and varied
AstraZeneca – Do not share without permission
Challenges
• Identification and control of critical material attributes may go
beyond monograph specifications and require a thorough
understanding of
– the formulation
– the process
– the physical and chemical properties of each ingredient
• Critical material attributes should be evaluated and
controlled to ensure that consistent product performance is
achieved throughout the product lifecycle
• Requires user/supplier collaboration
AstraZeneca – Do not share without permission
Challenges
• An excipient may have very different functions in the formulation
– e.g., diluent, lubricant, glidant
• It may require different performance characteristics
– e.g., particle size, size distribution, surface area depending on
its use in a formulation, manufacturing process, and dosage
form.
• The development, manufacture, and performance of
pharmaceutical dosage forms depend heavily upon the physical
and chemical properties of the excipients
– Physical • Particle morphology, powder property, polymorph, hygroscopicity, aqueous solubility,
pKa, and density
– Chemical • Identity, purity, incompatibility with drug substance or other excipients
– Mechanical • Flowability, compressibility
AstraZeneca – Do not share without permission
USP versus PhEur : different approach
USP Information Chapter <1059> Excipient Performance
• Overview of the key functional categories of excipients
identified in USP–NF.
• Guidance as to which properties might be important for a
particular material in a particular application.
• Cross-references to standard methods that can be used by
both manufacturers and users:
– Makes communication more straightforward
– Avoids an unnecessary plethora of test variations for a
particular parameter.
• Keeping the tests non-mandatory.
• Avoiding confusion with mandatory tests and labelling tests.
• Not imposing limits/specifications.
AstraZeneca – Do not share without permission
Extract from USP <1059>
“Not all critical material attributes of an excipient may be
identified or evaluated by tests, procedures, and acceptance
criteria in NF monographs. Excipient suppliers and users
therefore at times may wish to identify and control critical
excipient attributes that go beyond monograph
specifications.”
AstraZeneca – Do not share without permission
USP versus PhEur : different approach
PhEur
• Within each individual excipient monograph a section exists
for non-mandatory Functionality Related Characteristics
(FRCs) that should be considered
e.g. Croscarmellose sodium
Settling volume
Degree of substitution
Particle size distribution
Hausner ratio
e.g. Dibasic Calcium Phosphate
Particle size distribution
Bulk and tapped density
Powder flow
AstraZeneca – Do not share without permission
Excipient variability – how much do you
need to do?
• A risk based approach benefits both the patient and the business
– Not all excipients have an impact on product quality or safety
– Not all properties of an excipient are equally important
– In many cases normal excipient variation does not negatively impact
the quality and safety of the product
• The way forward
– Comprehensive studies of excipient properties are only needed when
the excipient properties are expected to impact the critical quality
attributes (CQAs) of the drug product
– The goal is to define control strategy for excipients
AstraZeneca – Do not share without permission
Case study to exemplify the approach
Microcrystalline
cellulose Degree of polymerisation
pH
Bulk density
Loss on drying
Residue on ignition
Conductivity
Ether soluble substances
Water soluble substances
Impurities
Particle size distribution
Mannitol Conductivity
Loss on drying
Reducing sugars
Assay
Particle size distribution
Porosity/Specific surface
area
Bulk density
Polymorphic form
Impurities
Sodium starch
glycolate pH
Loss on drying
Sodium chloride
Sodium glycolate
Assay (Na)
Bulk density
Rate/degree of
swelling
Magnesium
stearate Particle size
Specific surface area
Loss on drying
Stearic/palmitic acid
level
Assay (Mg)
•To explore every material attribute would require many
thousands of experiments
•Risk assessment is required to focus the experimental
programme
AstraZeneca – Do not share without permission
Assessing the risk of excipient variability
• Collect existing data/information on the raw materials
– Excipient monographs, literature examples, Handbook of
Pharmaceutical Excipients, supplier certificates of analysis, supplier
databases, etc
• Refer to target product profile
– target patient populations, geographical markets, etc
• For each excipient in the formulation, identify potential critical
material attributes (functionality) and potential risk factors (security
of supply, commercial and regulatory considerations)
• Score the potential risk for each material attribute and risk factor
AstraZeneca – Do not share without permission
Possible outcome after risk assessment
Microcrystalline
cellulose Degree of polymerisation
pH
Bulk density
Loss on drying
Residue on ignition
Conductivity
Ether soluble substances
Water soluble substances
Impurities
Particle size distribution
Mannitol Conductivity
Loss on drying
Reducing sugars
Assay
Particle size distribution
Porosity/Specific
surface area
Bulk density
Polymorphic form
Impurities
Sodium starch
glycolate pH
Loss on drying
Sodium chloride
Sodium glycolate
Assay (Na)
Bulk density
Rate/degree of
swelling
Magnesium
stearate Particle size
Specific surface area
Loss on drying
Stearic/palmitic acid
level
Assay (Mg)
•Risk assessment reduces the number of potential CMAs to
consider for experimental work
•Some material attributes could be confounded providing further
simplification
AstraZeneca – Do not share without permission
Next steps
• Risk assessment scores identify the highest risks excipient
attributes
• Select/source excipient variants
– Batch select from a particular supplier and within grade (QbD
sample sets)
– From one supplier use different grades (more extreme
variation)
– From multiple suppliers (different ranges of variation)
• Perform risk mitigation work to study effect of excipient
variability (on process and/or product performance)
• Use outputs to define excipient control strategy
AstraZeneca – Do not share without permission
Excipient supplier-user collaboration
• Exchange of information between excipient supplier and user
is invaluable
• Provides benefits to both supplier and user
• IPEC QbD checklists developed to help facilitate this
• Available to IPEC Europe members as downloads from the
website
AstraZeneca – Do not share without permission
How HPMC Physicochemical Properties
Impact Matrix Tablet Performance
ExcipientFest, Amsterdam NL
June 24, 2014
DOW CONFIDENTIAL - Do not share without permission
Outline
• Background and HPMC materials
• HPMC physical properties and how they impact matrix tablet
performance
• HPMC chemical properties and how they impact matrix
tablet performance
23 DOW CONFIDENTIAL - Do not share without permission
Quality by Design (QbD) Means Design
the Product And The Process
• Design the product to meet patient requirements
• Design the process to consistently meet product critical quality attributes
• Understand the impact of starting materials and process parameters on product quality
• Identify and control the source of process variation
• Continually monitor and update the process to allow a consistent quality over time
DOW CONFIDENTIAL - Do not share without permission
Dr A.Faham
Quality by Design (QbD)
• The drug product must be safe and efficacious for the patient.
– I.e., Ensure the dosage form performs as expected.
• How robust is dosage form performance?
• How robust is the process to make the dosage form?
• How robust are the methods to characterize the dosage form?
• What is the impact of raw material variability? (API? Excipients?)
– Multiple suppliers?
– Lot-to-lot variability?
27 DOW CONFIDENTIAL - Do not share without permission
Dr A.Faham
Properties vs. Performance
• Raw material properties
– Physical
– Chemical
• Process
– Processability • E.g. Flowability
– Process steps and parameters which are critical to quality.
• Performance
– Dosage form physical properties
– Achieving desired performance • API release
– Is desired performance reproducible (e.g. from lot-to-lot, day-to-day)?
28 DOW CONFIDENTIAL - Do not share without permission
Dr A.Faham
HPMC Matrix Tablets for Modified-Release
• Hydrophilic matrix tablets are the most commonly utilized
MR dosage form.
– Simplest.
– Fastest to develop.
– Least expensive to manufacture.
• Hypromellose 2208 is the most
common rate-modifying excipient
used in hydrophilic matrices.
29
R = CH3 OH OCH3
OOHO
OCH3
OCH3
OO
HO
OH
O
OO
HOOH
O
O
HO OO
CH3OOCH3
OCH3OCH3
O
R
OH
DOW CONFIDENTIAL - Do not share without permission
Dr A.Faham
HPMC Sustained Release Matrix Tablets
Key Hypromellose Formulation Variables
• Level
• Molecular weight/viscosity
• Substitution type
• Particle size distribution
Actives and other excipients can cause the formulation to be
more sensitive to HPMC properties
DOW CONFIDENTIAL - Do not share without permission
Dr A.Faham
How HPMC Physical Properties
Impact Matrix Tablet Performance
31 DOW CONFIDENTIAL - Do not share without permission
Dr A.Faham
32
For a selected hypromellose
product, polymer level is usually
the major drug release rate
controlling factor
– Ford et al. 1985. IJP, 24:327-
338 and 339-350
Drug release may be more sensitive
to variations in hypromellose
properties at low hypromellose
levels
(< 30%)
10% propranolol HCl, METHOCEL™ K4M balance lactose, 0.5% mag stearate
Hypromellose Level
DOW CONFIDENTIAL - Do not share without permission Dr A.Faham
30
40
50
60
70
80
90
100
0 10 20 30 40 50 60 70 80 90 100
HPMC Particle Size (% thru 230 mesh)
Dru
g R
ele
ased
(%
)
caffeine (50%), K15M (30%) - 6 hr
metoprolol tartrate (20%), K4M (30%) - 3 hr
theophylline (50%), K4M (30%) - 6 hr
Particle Size
34 DOW CONFIDENTIAL - Do not share without permission
Dr A.Faham
30
40
50
60
70
80
90
100
0 10 20 30 40 50 60 70 80 90 100
HPMC Particle Size (% thru 230 mesh)
Dru
g R
ele
as
ed
(%
)
acetaminophen (50%), K100M (30%) - 6 hr
hydrochlorothiazide (50%), K100 LV (30%) - 3 hr
ketoprofen (20%), K4M (30%) - 12 hr
Particle Size
35
DOW CONFIDENTIAL - Do not share without permission
Dr A.Faham
METHOCEL™ K15M Premium CR
0
20
40
60
80
100
0 120 240 360 480 600 720
Time (min)
% P
P d
isso
lved
High % thru 230 mesh/ Low Level High % thru 230 mesh/ High LevelLow % thru 230 mesh/ Low Level Low % thru 230 mesh/ High LevelCenter Point/ Low Level Center Point/ High Level
Propranolol HCl release: effect of particle size
f2 = 48.23
f2 = 94.14
• Higher polymer level slower drug release
• Higher polymer level lower variability
• Drug release were significantly affected by coarser P/S for lower polymer level
36
DOW CONFIDENTIAL - Do not share without permission
Dr A.Faham
How HPMC Chemical Properties
Impact Matrix Tablet Performance
37 DOW CONFIDENTIAL - Do not share without permission
Dr A.Faham
38
Selection of Hypromellose substitution grade
Hypromellose grade
has a significant effect
on dissolution
Methylcellulose and
Hypromellose 2906 (A
and F Chemistry)
typically are not used for
CR applications
DOW CONFIDENTIAL - Do not share without permission Dr A.Faham
METHOCEL™ K15M Premium CR
0
20
40
60
80
100
0 120 240 360 480 600 720Time (min)
% P
P d
iss
olv
ed
High Viscosity/ Low Level High Viscosity/ High Level
Low Viscosity/ Low Level Low Viscosity/ High LevelCenter Point/ Low Level Center Point/ High Level
• Higher polymer level slower drug release
• Higher polymer level lower variability
• Drug release were consistent across viscosity range
Propranolol HCl release: effect of viscosity
f2 = 66.90
f2 = 74.21
The similarity factor (f2) was calculated by comparing high vs. low end of the selected physicochemical property
39 DOW CONFIDENTIAL - Do not share without permission
Dr A.Faham
40
50% diclofenac sodium, 40% METHOCEL™ K15M
9.5% lactose, 0.5% mag stearate
Hypromellose Substitution
40% salicylic acid, 30% METHOCEL™ K15M
29% lactose, 1% mag stearate
DOW CONFIDENTIAL - Do not share without permission
Dr A.Faham
Paracetamol Model Example
Ingredient % w/w Weight per tablet (mg)
Paracetamol* 50 250
METHOCEL™ K4M or Pilot Plant HPMC 30 150
Lactose 18 90
Magnesium stearate 1 5
Talc 1 5
Total 100 500
Actual tablet weight: 502 ± 3 mg
Hardness: 94 ± 8 N
* Paracetamol:
Analgesic
Aqueous solubility: 14 mg/mL
DOW CONFIDENTIAL - Do not share without permission
Dr A.Faham
Batch-to-Batch Consistency
• Batch-to-batch consistency with commercial
METHOCEL™:
• Reproducible modified-release
performance.
42
0
20
40
60
80
100
0 200 400 600 800 1000 1200 1400
Para
ceta
mo
l Re
leas
ed
(%)
Time(min)
Batch no. 1 Batch no. 2
Batch no. 3 Batch no. 4
Batch no. 5 Batch no. 6
Batch no. 7 Batch no. 8
Batch no. 9 Batch no. 10
Batch no. 11 Batch no. 12
Batch no. 13 Batch no. 14
Batch no. 15 Batch no. 16
Batch no. 17 Batch no. 18
Batch no. 19 Batch no. 20
Commercial
Batch No. %Me %HP
50% Cumulative
Volume Particle
Size (µm) %NaCl
2% Viscosity
(mPa·s)
1 22.8 8.3 93.8 0.2 3711
2 23.1 8.7 91.9 0.3 4514
3 22.2 9.1 84.3 0.3 3638
4 22.6 8.4 88.7 0.1 4953
5 22.7 8.2 94.1 0.2 4015
6 23.0 8.5 97.8 0.2 4444
7 23.3 8.7 102.1 0.3 3506
8 23.2 8.8 110.8 0.3 3897
9 23.1 8.6 109.1 0.3 3615
10 23.1 8.6 103.7 0.3 3615
11 22.2 8.6 96.7 0.6 3756
12 23.0 8.8 107.9 0.3 3810
13 23.0 8.7 103.1 0.4 4325
14 23.3 8.7 99.3 0.3 3775
15 23.4 8.7 99.3 0.3 3849
16 22.9 8.5 98.8 0.4 4364
17 22.8 7.9 101.9 0.3 4562
18 23.6 8.4 104.3 0.3 4322
19 23.1 8.7 101.2 0.4 4057
20 23.0 8.7 100.8 0.4 3839
Average 23.0 8.6 99.2 0.3 3996
Std Deviation 0.4 0.3 6.6 0.1 414Rogers TL, Petermann O, Adden R, and Knarr M (2011). Investigation and rank -ordering of
hypromellose 2208 properties impacting modified release performance of a hydrophilic matrix tablet,
Twenty-Sixth Annual Meeting, Proceedings of the American Association of Pharmaceutical Scientists,
Washington DC, Poster no. R6168.
900 mL pH 5.7 phosphate buffer at 37 °C
50 rpm paddle speed
Tablets placed in sinkers
n=6 standard deviation was never more than 2%
DOW CONFIDENTIAL - Do not share without permission Dr A.Faham
METHOCEL™ FRCs Impacting Performance
• Based on this model, rank-order of METHOCEL FRC impact is as follows: %HP (p <
0.05) > 2% viscosity (p = 0.06) > particle size (p = 0.13) > %Me (p = 0.75).
• Correlations between paracetamol release and HP substitution vs. 2% viscosity
reflect findings from the model.
• Paracetamol release increases with increasing HP content .
– Trend occurs over a narrow range of 79-86% paracetamol released at 22 hr, reflecting
reproducible batch-to-batch modified-release performance.
43
Rogers TL, Petermann O, Adden R, and Knarr M (2011). Investigation and rank -ordering of
hypromellose 2208 properties impacting modified release performance of a hydrophilic matrix tablet,
Twenty-Sixth Annual Meeting, Proceedings of the American Association of Pharmaceutical Scientists,
Washington DC, Poster no. R6168.
ESTABLISHING THE PERFORMANCE DESIGN SPACE
DOW CONFIDENTIAL - Do not share without permission Dr A.Faham
Pilot Plant HPMC vs. Commercial
METHOCEL™
– Expanded design space boundaries with pilot plant HPMC.
• HP substitution was purposefully varied.
– Premise:
• There is ‘insufficient’ batch-to-batch variability in commercial METHOCEL to investigate
performance design space proactively.
• We cannot explore the allowable pharmacopeial design space.
– Where are the boundaries of robustness?
– What if we miss optimal performance ‘sweet spots’?
44
Rogers TL, Knarr M, Petermann O, and Adden R (2011). Expanding design space boundaries within
pharmacopeial limits: Impact of atypical hydroxypropoxyl substitution on drug release from HPMC
matrices, Twenty-Sixth Annual Meeting, Proceedings of the American Association of Pharmaceutical
Scientists, Washington DC, Poster no. R6167.
Sample
identification %Me %HP
50% cumulative
volume particle
size (µm) %NaCl
2% viscosity
(mPa-s)
Prototype No. 1 24.2 8.6 78.5 0.1 4466
Prototype No. 2 23.0 11.4 72.0 0.1 4346
Prototype No. 3 24.0 9.1 64.6 0.1 2730
Prototype No. 4 24.4 6.0 84.8 < 0.1 5292
Prototype No. 5 23.1 11.2 70.3 0.1 3356
Prototype No. 6 24.4 6.6 66.8 < 0.1 5476
Prototype No. 7 23.3 7.8 70.5 < 0.1 5092
Prototype No. 8 23.4 9.5 66.1 < 0.1 4999
Prototype No. 9 23.7 10.2 52.4 < 0.1 5009
See previous section for FRCs of commercial batches investigated
4
5
6
7
8
9
10
11
12
HP
Co
nte
nt (
%)
Commercial Batches 1 through 21 Pilot Plant Batches 1 through 9
Breadth of minimum and maximum HP content (4–12%) according to the harmonized
pharmacopeia (USP, PhEur, and JP).
DOW CONFIDENTIAL - Do not share without permission Dr A.Faham
Modified-Release Performance
• Pilot plant HPMC data
“brackets” commercial
METHOCEL data for HP
substitution and paracetamol
release.
• Paracetamol release increases
with increasing HP substitution.
• Efficiently determined that
formulation is robust.
Rogers TL, Knarr M, Petermann O, and Adden R (2011). Expanding design space boundaries within
pharmacopeial limits: Impact of atypical hydroxypropoxyl substitution on drug release from HPMC
matrices, Twenty-Sixth Annual Meeting, Proceedings of the American Association of Pharmaceutical
Scientists, Washington DC, Poster no. R6167.
DOW CONFIDENTIAL - Do not share without permission Dr A.Faham
Indapamide Example
Ingredient % w/w Weight per tablet (mg)
Indapamide* 2.5 5
Pilot Plant HPMC 40 80
Lactose 40 80
Microcrystalline cellulose 16.5 33
Magnesium stearate 0.5 1
Talc 0.5 1
Total 100 200
Actual tablet weight: 200 ± 3 mg
Hardness: 83 ± 8 N
Friability: Weight loss ≤ 0.16%
* Indapamide:
Antihypertensive
Aqueous solubility: 75 µg/mL
DOW CONFIDENTIAL - Do not share without permission
Dr A.Faham
Modified-Release Performance
• Only variable was the HPMC batch used.
– Same formulation composition.
– Tried to hold everything constant except HPMC batch.
47
Proactively determined that API and formulation are very sensitive to variation in %HP
substitution.
High risk of batch failure.
0.1% SLS in 900 mL water at 37˚C
50 rpm paddle speed
Tablets placed in hanging baskets
n=6 standard deviation was never more
than 5%
0
20
40
60
80
100
0 200 400 600 800 1000 1200 1400
Ind
apam
ide
Re
leas
ed
(%)
Time (min)
% indapamide released at 17 hr
ranged from 60 to 90%
Breaking point in
modified release performance
Step-change increase in API
release
DOW CONFIDENTIAL - Do not share without permission Dr A.Faham
Performance Design Space
Breaking point
in modified
release
performance
Above HP content of 7.8%
Step-change increase in
API release
Modulation of API
release spans
∆ of ~35%
Potential extent of
variation unacceptable
Proactive exploration of
design space identified
highly responsive API
HPMC specification
recommended
DOW CONFIDENTIAL - Do not share without permission
Dr A.Faham
Summary
• Modified release performance is most significantly impacted by HP
substitution of METHOCEL™
– HP substitution is the primary factor modulating modified release
• Forced-variation prototypes enabled expansion of the design
space boundaries of our model formulation
– APIs highly ‘responsive’ to METHOCEL™ FRCs
49 DOW CONFIDENTIAL - Do not share without permission Dr A.Faham