UNDERSTANDING PELLET-FILLER INTERACTIONS IN …understanding pellet-filler interactions in...
187
UNDERSTANDING PELLET-FILLER INTERACTIONS IN COMPRESSION OF COATED PELLETS CHIN WUN CHYI NATIONAL UNIVERSITY OF SINGAPORE 2011
Transcript of UNDERSTANDING PELLET-FILLER INTERACTIONS IN …understanding pellet-filler interactions in...
UNDERSTANDING PELLET-FILLER INTERACTIONS IN
COMPRESSION OF COATED PELLETS
CHIN WUN CHYI
NATIONAL UNIVERSITY OF SINGAPORE
2011
UNDERSTANDING PELLET-FILLER INTERACTIONS IN
COMPRESSION OF COATED PELLETS
CHIN WUN CHYI
B.Sc. (Pharm.) Hons, NUS
A THESIS SUBMITTED
FOR THE DEGREE OF DOCTOR OF PHILOSOPHY
DEPARTMENT OF PHARMACY
NATIONAL UNIVERSITY OF SINGAPORE
ACKNOWLEDGEMENTS
I would like to express my heartfelt thanks to my supervisors, A/P Paul Heng
Wan Sia and A/P Chan Lai Wah for their patience, guidance, encouragements and
opportunities given by them throughout my candidature. I am also thankful for their
time and effort spent in correcting and improving this thesis. It has been an enriching
experience working with them. I would also like to thank Dr Celine Liew for her
suggestions to improve my work and for being so approachable.
I wish to thank National University of Singapore for providing the research
scholarship and resources for research.
My sincere appreciation goes to my colleagues and friends in GEANUS for
their support, help and company: Zhihui, Dawn, Sook Mun, Stephanie, Atul, Likun,
Christine, Kou Xiang, Srimanta, Asim, Bingxun, Poh Mun, Yien Ling, Teresa and
Mei Yin. They have made my graduate life more enjoyable and meaningful.
I am grateful and indebted to my parents and sister for their constant love,
support, patience and faith in me. Without them, I will never be able to come this far.
Thank you. I am also thankful to brothers and sisters in Christ who have been
supporting me through prayers.
Last but not the least, I wish to express my gratitude towards my Lord and
Saviour, Jesus Christ, who “causes all things to work together for good to those who
love God, to those who are called according to His purpose” (NASB, Romans 8:28).
Wun Chyi
July, 2011
i
TABLE OF CONTENTS
TABLE OF CONTENTS .............................................................................................i
SUMMARY ..................................................................................................................v
LIST OF TABLES .....................................................................................................vii
LIST OF FIGURES ..................................................................................................viii
LIST OF SYMBOLS .................................................................................................xii
1. INTRODUCTION....................................................................................................2
1.1. Background .......................................................................................................2
1.2. Coating of multiparticulates ............................................................................4
1.2.1. Polymers for coating .............................................................................4
1.2.2. Coat thickness .......................................................................................8
1.3. Coating core.......................................................................................................9
1.4. Compact fillers ................................................................................................12
1.5. Proportion of coated pellets ...........................................................................23
1.6. Compression pressure ....................................................................................26
1.7. Protective overcoat and undercoat................................................................26
1.8. Methods for assessing the coat integrity of compressed pellets..................30
2. HYPOTHESES AND OBJECTIVES ..................................................................41
3. EXPERIMENTAL.................................................................................................44
3.1. MATERIALS ..................................................................................................44
3.1.1. Pellet cores and coating materials......................................................44
3.1.2. Compact excipients .............................................................................44
3.2. METHODS ......................................................................................................46
3.2.1. Preparation of coating materials ........................................................46
3.2.2. Pellet coating .......................................................................................46
ii
3.2.3. Preparation and characterization of compacts ..................................48
3.2.4. Dissolution studies ..............................................................................49
3.2.5. Analysis of drug release data..............................................................50
3.2.6. Disintegration test ...............................................................................51
3.2.7. Evaluation of the densification behaviour of materials ....................51
3.2.8. Determination of particle true density................................................52
3.2.9. Determination of material bulk density, tapped density and Hausner
ratio......................................................................................................53
3.2.10. Particle size reduction of lactose ......................................................53
3.2.11 Particle size determination of compact fillers ...................................54
3.2.12. Particle size and coat thickness determination of pellets ................54
3.2.13. Method of retrieving pellets from compacts.....................................55
3.2.14. Scanning electron microscopy (SEM)..............................................56
3.2.15. Stereomicroscopy ..............................................................................56
3.2.16. Surface profilometry .........................................................................56
3.2.17. Determination of pellet core osmolality ...........................................56
3.2.18. Determination of pellet mechanical strength...................................57
3.2.19. Statistical analysis .............................................................................57
3.2.20. Multivariate data analysis.................................................................58
4. RESULTS AND DISCUSSION ............................................................................60
PART A. UTILIZATION OF LACTOSE FILLERS IN COMPACTS
CONTAINING COATED PELLETS......................................................................60
A.1. Influence of lactose filler particle size on the extent of coat damage during
compression .................................................................................................60
A.2. Influence of disintegration time on UC/C values ........................................62
iii
A.3. Other methods of assessing coat integrity of pellets compressed with
lactose filler of different particle size ........................................................65
A.3.1. Stereomicroscopy................................................................................65
A.3.2. Scanning electron microscopy (SEM) ...............................................68
A.3.3. Surface profilometry...........................................................................71
A.5. Utilization of micronized lactose in reducing the extent of coat damage
during compression.....................................................................................75
A.5.1. Relationship between disintegration time and UC/C ........................77
A.5.2. Effect of lactose blend particle size....................................................77
A.5.3. Effect of proportion of micronized lactose in the blend ...................79
A.5.4. Critical pellet volume fraction ...........................................................85
A.6. Conclusion ......................................................................................................86
PART B. INFLUENCE OF FILLER TYPE AND PELLET VOLUME
FRACTION IN COMPACTS CONTAINING COATED PELLETS ..................89
B.1. Physical characteristics of fillers and their influence on the extent of coat
damage during compression ......................................................................89
B.2. Relationship between the material compressibility and extent of pellet
coat damage .................................................................................................99
B.3. Comparison of the extent of coat damage in uncured and cured coated
pellets..........................................................................................................102
B.4. Influence of disintegration time on UC/C values.......................................103
B.5. Influence of compression pressure and pellet volume fraction................105
B.5.1. Influence of compression pressure and disintegration time on
different dissolution parameters .......................................................105
B.5.2. Relationship between compact characteristics and UC/C ..............111
iv
B.5.3. Influence of compression pressure on UC/C ..................................112
B.5.4. Influence of pellet volume fraction on UC/C ..................................120
B.6. Influence of compression speed on the extent of coat damage.................129
PART C. INFLUENCE OF PELLET SIZE AND PELLET TYPE IN
COMPACTS CONTAINING COATED PELLETS ............................................133
C.1. Dissolution profiles of uncompressed pellets .............................................133
C.1.1. Differences in drug release from coated MCC pellets and sugar
pellets .................................................................................................133
C.1.2. Differences in drug release from coated sugar pellets of different
particle size ........................................................................................138
C.2. Influence of pellet core material on the extent of coat damage ...............139
C.3. Influence of sugar pellet size on the extent of coat damage .....................144
5. CONCLUSION ....................................................................................................149
6. REFERENCES.....................................................................................................153
v
SUMMARY
Over the years, the interest in compression of coated multiparticulates, such as pellets,
has grown due to its numerous advantages. However, it is a challenge to preserve the
coat function after compression as the pressure could induce changes to the structure
of the coating. A non-disintegrating matrix could also form after compression. Due to
the multiple factors involved, it is not uncommon to find conflicting results in the
literature. It is thus important to further investigate and understand the factors that
could affect the pellet coat during compression.
When ethylcellulose coated pellets were compressed with different grades of lactose
fillers, there was a directly proportional linear relationship between d50 (median
particle size) and the extent of coat damage. The SEM micrographs showed larger and
deeper indentations on the surfaces of pellets compressed with coarser lactose filler
particles. The extent of coat damage was reduced with the addition of micronized
lactose to formulations containing coarse grade lactose filler. The maximum amount
of reduction in pellet coat damage was achieved when the lactose filler consisted of a
blend of 3:1 micronized lactose: coarse grade lactose. This, however, decreased the
flowability of the powder blend. Thus, micronized lactose could be used to mitigate
coat damage but its effect on the powder flow property should be carefully
considered. Unlike the lactose fillers, the extent of coat damage was not linearly
proportional to the particle size when MCC and crospovidone were used as fillers.
Irrespective of the filler type, compression of coated sugar pellets at various pressures
consistently showed three phases of pellet coat damage with two critical transition
points. There was minimal coat damage at the initial particle rearrangement phase,
followed by coat damage acceleration phase beyond the first critical transition point.
The pellet volume fraction at the first critical point was dependent on the filler type
vi
and was in the range of 0.33 to 0.39. After the second critical transition point which
corresponded to the percolation threshold for the coated pellets, the drug dissolution
rate decreased. The pellet volume fraction at the second critical point was
approximately 0.48 for all the filler types. It was found that pellet volume fraction was
an important factor affecting the extent of coat damage. In addition, the rate of
increase in pellet coat damage with pellet volume fraction was dependent on the filler
type. Among the four filler types used, the coat damage observed in pellets
compressed with micronized lactose was markedly lower. It was concluded that the
eventual extent of coat damage was dependent on the length of particle rearrangement
phase with respect to the compression pressure. It was also dependent on the rate of
increase in pellet coat damage with compression pressure and pellet volume fraction
in the acceleration phase.
In order to investigate the influence of pellet size and pellet core material on the
extent of pellet coat damage, sugar pellets of different size fractions and MCC pellets
were coated to approximately the same thickness. They were then compressed with
four filler types. It was found that the extent of coat damage for MCC pellets was not
dependent on the filler particle size. The smaller sugar pellets showed greater extent
of coat damage.
vii
LIST OF TABLES Table 1. Summary of the effects of parameters that were found to influence the extent
of coat damage during compression .............................................................29
Table 2. Release exponent (n) derived from the power law and the corresponding
release mechanism (Peppas and Sahlin, 1989). ............................................38
Table 3. Coating and drying conditions ......................................................................47
Table 4. Dispersing media of different materials ........................................................54
Table 5. Size distribution of lactose fillers..................................................................61
Table 6. Mean Ra and Rq values of uncompressed coated pellets and pellets
compressed with different grades of lactose. ................................................72
Table 7. Bulk density, tapped density and Hausner ratio of lactose fillers. ................74
Table 8. The UC/C and disintegration time of compacts produced with lactose blends
of different d50, d90 values..............................................................................76
Table 9. Py and W values of the different materials....................................................99
Table 10. Curve-fitting (R2) of dissolution profiles for compacts containing lactose
fillers formed at different pressures to different models.............................106
Table 11. Curve-fitting (R2) of dissolution profiles for compacts containing MCC
fillers formed at different pressures to different models.............................107
Table 12. Physical properties of pellets. ...................................................................134
Table 13. Curve-fitting of dissolution date and dissolution parameters of different
coated pellet cores.......................................................................................134
Table 14. Osmolality of solutions containing different concentrations of pellets. ...136
viii
LIST OF FIGURES
Figure 1: Factors governing coat function preservation during compression. .............5
Figure 2. A typical Heckel plot (Heckel, 1961b). .......................................................14
Figure 3. Plot of the compression stress against thickness of powder column (Asano
et al., 1997). .................................................................................................................21
Figure 4. Beer’s plot of chlorpheniramine maleate in purified water. ........................50
Figure 5. Dissolution profiles of coated pellets compressed with lactose 450M (□),
350M (▲), 200M (●), 150M (■), 100M (♦) and uncompressed pellets (○)................61
Figure 6. Relationship between UC/C and (A) d10, (B) d50 and (C) d90 of lactose filler.
......................................................................................................................................63
Figure 7. Relationship between UC/C and the average disintegration time of
compacts containing different grades of lactose filler. ................................................64
Figure 8. Relationship between UC/C and average disintegration time of compacts
containing different disintegrants. ...............................................................................65
Figure 9. Stereomicroscope images of uncompressed coated pellets and coated pellets
compressed with different grades of lactose. “I” denotes pellets retrieved from the
interior of the compacts; “S” denotes pellets from the surface of the compacts. ........66
Figure 10. SEM photomicrographs of (A) uncompressed pellets, as well as coated
pellets compressed with (B) lactose 100M and (C) lactose 450M (retrieved from
compact interior). .........................................................................................................69
Figure 11. SEM photomicrographs of coated pellets compressed with (A) lactose
100M and (B) lactose 450M (retrieved from compact surface). .................................70
Figure 12. Relationship between UC/C and Hausner ratio lactose fillers. .................75
Figure 13. Plot of UC/C values against disintegration time. ......................................77
Figure 14. Relationship between UC/C values and d50...............................................78
ix
Figure 15. Relationship between UC/C and d90 of lactose blends. .............................79
Figure 16. UC/C values of compacts containing blends with lactose 80M (◊) and
blends with lactose 125M (■). .....................................................................................80
Figure 17. Relationship between Hausner ratio and the proportion of micronized
lactose in lactose blend containing lactose 80M (◊) and lactose 125M (■).................81
Figure 18. Relationship between compact thickness and proportion of micronized
lactose in the lactose blend. .........................................................................................83
Figure 19. Plot of UC/C against average roughness, Ra, of compacts containing
different proportions of micronized lactose and lactose 80M formed using a
compression pressure of 11.3 MPa. .............................................................................84
Figure 20. Relationship between UC/C and pellet volume fraction. ..........................86
Figure 21. Relationship between the UC/C values of uncured (closed symbols) and
cured (opened symbols) pellets against d50 of lactose ( ), MCC (■), L-HPC (▲),
crospovidone ( ), DCP (×) and DCL 11 (+). ..............................................................90
Figure 22. SEM photomicrographs of (A-C) lactose and (D-F) MCC filler particles
(all at same magnification except micronized lactose). ...............................................93
Figure 23. SEM photomicrographs of (A-B) L-HPC, (C-D) crospovidone (Kollidon)
and (E) DCP filler particles (all at the same magnification)........................................94
Figure 24. Hausner ratio of different grades of (A) lactose ( ), (B) L-HPC (▲), (C)
MCC ( ) and (D) crospovidone ( ) against d50. .......................................................96
Figure 25. Relationship between UC/C and Hausner ratio of lactose ( ), L-HPC
(▲), MCC ( ) and crospovidone ( ). .......................................................................96
Figure 26. Plot of UC/C against (A) W coefficient and (B) yield pressure of lactose
(♦), MCC (■), L-HPC (▲) and crospovidone (◊)......................................................101
x
Figure 27. Plot of the UC/C values of coated pellets compressed with different
grades of MCC (▲), lactose ( ) and L-HPC (■) against the average disintegration
time. ...........................................................................................................................104
Figure 28. Loadings plot from PCA of dissolution parameters, compression pressure
(Press) and disintegration time (DT). KH: Higuchi dissolution rate constant; KP:
Power Law dissolution rate constant; n: release exponent from Power Law; LH: lag
time from Higuchi equation; LP: lag time from Power Law equation; T25, T50, T75:
time taken for 25 %, 50 % and 75 % of the drug to be released respectively; MDT:
mean dissolution time; VDT: variance of dissolution time; RD: relative dispersion of
concentration-time profiles; F2: similarity factor.......................................................110
Figure 29. Loadings plot from PCA of compact properties, compression pressure and
UC/C. App vol: compact apparent volume; PVF: pellets volume fraction. ..............112
Figure 30. Plot of UC/C against compression pressure for compacts containing
lactose 80M (■), PH200 (▲), PH 105 (Δ), and micronized lactose ( )....................113
Figure 31. Heckel plot (opened symbols, dotted line) and plot of UC/C (closed
symbols, solid line) against compression pressure for compacts containing (A)
lactose 80M, (B) micronized lactose. ........................................................................115
Figure 32. Heckel plot (opened symbol, dotted line) and plot of UC/C (closed
symbol, solid line) against compression pressure for compacts containing (C) PH200
and (D) PH105. ..........................................................................................................116
Figure 33. Plot of disintegration time (opened symbols) and UC/C (closed symbols)
against compression pressure for compacts containing (A) lactose 80M, or (B)
micronized lactose. ....................................................................................................121
Figure 34. Plot of disintegration time (opened symbols) and UC/C (closed symbols)
against compression pressure for compacts containing (A) PH200, or (B) PH105...122
xi
Figure 35. Relationship between UC/C and pellet volume fraction for lactose 80M
(♦), PH200 (▲), PH105 (Δ) and micronized lactose (■)...........................................123
Figure 36. Plot of UC/C against pellet volume fraction for compacts containing (A)
lactose 80M, (B) micronized lactose, (C) PH200 and (D) PH105.............................127
Figure 37. UC/C values of compacts containing lactose 80M (■), PH105 (Δ), PH200
(▲) and micronized lactose (□) formed at different compression speeds................130
Figure 38. Porosity of compacts containing PH200 (▲),PH105 (Δ), micronized
lactose (□) and lactose 80M (■) formed at different compression speeds.................132
Figure 39. Drug release profile of MCC pellets (▲), Group I sugar pellets (♦), Group
II sugar pellets (■) and Group III sugar pellets (Δ). ..................................................135
Figure 40. SEM photomicrographs of coated (A) Group I sugar pellets, (B) Group II
sugar pellets, (C) Group III sugar pellets and (D) MCC pellets at (1) 200 and (2) 1000
times magnification....................................................................................................137
Figure 41. (A)UC/C, (B) apparent volume, (C) porosity and (D) pellet volume
fraction of compacts containing group III sugar pellets (shaded) and MCC pellets
(clear) with different fillers. .......................................................................................141
Figure 42. A typical force-displacement curve of (A) a Group III sugar pellet and (B)
a MCC pellet. .............................................................................................................143
Figure 43. Plot of UC/C against average mean diameter of coated pellets compressed
with lactose 80M (♦), PH 200 (▲), PH 105 (Δ) and micronized lactose (■). ...........145
Figure 44. Plots of (A) apparent volume (B) porosity and (C) pellets volume fraction
of compacts containing lactose 80M (■), micronized lactose (□), PH200 (▲) and
PH105 (Δ) against the average mean pellet diameter of sugar pellets.......................146
xii
LIST OF SYMBOLS
ε: Porosity of the matrix
σo: Yield strength in psi
τ: Tortuosity factor of the capillary system
A: Y-intercept of extrapolated line from the linear portion of Heckel plot
AH: Total amount of drug present in the matrix per unit volume
bp: Burst effect
cH: Constant in Higuchi model
C1: Constant in Walker equation
Cs: Solubility of the drug per unit volume
CE: Compression energy
d: Time plasticity
dc: Average mean diameter of ethylcellulose coated pellets
duc: Average mean diameter of chlorpheniramine loaded pellets that were not coated
with ethylcellulose.
dx: Diameter of the material at x percentile of the cumulative percent undersize plot
D: Relative density at compression pressure, P
DO: Relative density of loose powder at zero pressure
DA: Relative density at y-intercept of extrapolated line from the linear portion of
Heckel plot
DH: Diffusion constant of the drug molecule in the external phase
e: Pressure plasticity
EE: Elastic energy
ER: Elastic recovery
f: Axial force on each particle in the compact
xiii
F1: Difference factor
F2: Similarity factor
Fp: Pellet volume fraction,
GP: General plasticity
H0: Minimal height of tablet under load
h: Compact thickness
H1: Height of tablet after 10 days
i: sample number
j: number of dissolution sample times
K: Gradient of linear portion of Heckel plot
K0: Zero order dissolution rate constant
KH: Higuchi dissolution rate constant
KP: Kinetic constant in Power Law
KV : Change in volume ratio as pressure increases logarithmically
m’: Proportionality constant
ΔMi: Fractional amount of drug released between ti and ti-1
Mt: Cumulative amount of drug released at time t
M∞: Cumulative amount of drug released at infinite time
MDT: Mean dissolution time
n: Release exponent in Power Law
P: Compression pressure
ρapp: Apparent compact density
ρT: True compact density
Py: Mean yield pressure
PE: Plastic energy
xiv
Q: Amount of drug released from the system at time t per unit area of exposure
Q0: Initial amount of drug in the solution
Qt: Amount of drug dissolved in time t
Rc: Radius of compact
R: Beginning of the compression cycle
Ri: Mean drug release for reference formulation at time point i
Ra: Represents the average roughness
Rq: Represents the root mean square of the roughness
RD: Relative dispersion of concentration-time profiles
S: Point where the compression stress reaches the maximum in a typical plot of
compression stress against the thickness of powder column
t: Time point
T: Point where the thickness of the powder column reaches the minimum in a typical
plot of compression stress against the thickness of powder column
t̂ : Time point between ti and ti-1
tL: Lag time
Tx%: Time taken for x % of drug to be released
Ti: Mean drug release for test formulation at time point i
T50%: Time taken for 50% of the drug to be released
Th: Ethylcellulose coat thickness
U: End point of hysteresis loop
UC/C: Ratio of the T50% value of uncompressed coated pellets to compressed coated
pellets
V: volume ratio or relative volume
V′: Volume at pressure P
xv
VO: True volume (at zero porosity)
VB: Bulk volume
Vap: Apparent volume of powder prior to compression
Vat: Apparent volume of tablets after compression
Vt: Tapped volume
VDT: Variance of dissolution time
ω: Fast elastic decompression
Ws: Weight of sample
W: Compressibility coefficient which expresses the percent change in volume ratio
when the pressure increased by a factor of 10
Wc: Weight of the compact
Wt: Weight of the material in cylinder for determining bulk and packed densities
1
PART 1: INTRODUCTION
2
1. INTRODUCTION
1.1. Background
Pharmaceutical solid dosage forms are commonly coated for the purpose of
modifying drug release rate, improving the stability of core substances, as well as for
separating incompatible substances. They are also coated to mask unpleasant taste or
odour of core substances and for aesthetic purposes (Cole, 1995). Besides tablets and
capsules, multiparticulates can also be coated. Multiparticulates are discrete particles
such as pellets, granules, powders or crystals, which may be filled into capsules or
compressed into tablets (Tang et al., 2005). Some coated tablets can cause irritation to
the mucosal lining of the gastrointestinal tract (GIT) when lodged at a particular site.
In addition, dose-dumping could occur if the coat of a sustained release tablet fails.
These disadvantages of coated tablets can be avoided by employing coated
multiparticulates which are well distributed along the GIT upon ingestion. As the dose
of the drug is divided into smaller subunits, coat failure of a few multiparticulates will
not contribute to failure of the whole dosage form. In addition, coated
multiparticulates offer additional benefits of physically separating incompatible drugs,
having a more predictable gastric emptying when administered together with a meal
and less likelihood of product failure (Abrahamsson et al., 1996). As a mixture of
pellets with different coat thickness can be combined in a single dosage form, a more
precise and controlled drug release can be achieved.
It is more desirable to compress the coated multiparticulates into tablets than to fill
them into capsules. This is because capsules can pose tampering concerns. In 1982,
Tylenol capsules were adulterated with cyanide, resulting in the deaths of seven
people in Chicago, United States. Foreign materials could be easily added into a
3
capsule by pulling apart the cap and the body of the capsule and replacing them back
without obvious signs of tampering. In addition, the use of animal gelatin makes
capsules less acceptable to vegetarians or certain religious groups. Furthermore, tablet
production is more efficient and incurs lower unit production costs in general. The
production rate of a capsule filling machine can vary from 5,000 to 150,000 per hour
(Jones, 2007) while output of over 600,000 tablets per hour can be achieved by a
rotary press (Alderborn, 2007). Besides, tablets allow for a more flexible dosing
regimen since they may be scored. A higher drug dose can also be delivered by
compressing coated multiparticulates instead of filling them into capsules (Béchard
and Leroux, 1992). The applications of compressed coated multiparticulates may also
be extended to orally disintegrating tablets containing coated particles. Orally
disintegrating tablets are specially designed solid dosage forms which disperse in the
mouth immediately upon contact with the saliva. They can thus be administered
without water (Hahm and Augsburger, 2008). Hence, these products can improve
compliance of patients especially those with difficulties in swallowing tablets or
capsules. Although there are certain attractive advantages to formulating coated
particles into tablets, coated particulates are more commonly filled into capsules. This
is mainly due to the challenges encountered in delivering undamaged compressed
coated particles. Firstly, tablet compression forces can cause structural damage to
outer coatings, hence affecting its sealant function. Secondly, coated particles might
fuse together during compression and remain as a single fused entity without
deaggregating upon tablet disintegration. It is essential for the coated particles to
disintegrate into individual coated particles in order to preserve the advantages of a
multiparticulate dosage form. Thirdly, segregation of the tablet constituents during
mixing and die filling may occur due to their differences in shape, size and
flowability. Thus, it is important to investigate the effects of compression forces on
4
coated multiparticulates and understand the factors governing coat integrity under
compression. The major factors governing coat integrity under compression are
shown in Figure 1 and will be discussed in detail in the following sections.
1.2. Coating of multiparticulates
1.2.1. Polymers for coating
The polymers most commonly used for pharmaceutical film coating can be broadly
classified as cellulose derivatives or acrylic polymers. For the purpose of sustained
drug delivery, ethylcellulose is an effective water-insoluble cellulose derivative. It is
marketed under the trade names, Surelease and Aquacoat. The acrylic polymers
available in the market for sustaining drug release include Eudragit (NE 30 D, RS 30
D, RL 30 D) and Kollicoat (SR 30 D) (Lehmann, 1997). All these polymers are
mostly water-insoluble and can be either dissolved in an organic solvent or dispersed
as fine pseudo-latex particles in an aqueous solvent. Due to the toxicity and
flammability of organic solvents, as well as stringent government regulations on their
use, water is generally the preferred liquid medium.
In several studies, the ethylcellulose coat was found to be ruptured after compression
of the coated multiparticulates (Bansal et al., 1993, Maganti and Celik, 1994, Sarisuta
and Punpreuk, 1994, Lundqvist et al., 1998, Dashevsky et al., 2004, Cantor et al.,
2009). This was probably due to the brittle nature of the ethylcellulose aqueous-based
film coat, which had tensile strength of less than 5 N/mm2, puncture strength of less
than 5 MPa and elongation at break of less than 12 % (Sarisuta and Punpreuk, 1994,
Bodmeier et al., 1997b, Heng et al., 2003). Due to the brittleness of the ethylcellulose
coat, it is likely that the coat would fracture or crack on compression. Thus, in order
5
Figure 1: Factors governing coat function preservation during compression.
Core properties: • Core size • Porosity • Deformability • Tensile
strength
Coating of multiparticulates: • Coat flexibility
and strength • Coat thickness • Protective
overcoat and undercoat
Proportion of coated multiparticulates
Tablet containing coated multiparticulates
Compression pressure Filler properties:
• Plasticity • Fast elastic
decompression • Percentage
reduction in volume
• Particle size • Deformability • Granule size
Powders
Granules
6
to preserve coat integrity, it is essential for the coat to be sufficiently plastic to
accommodate the core deformations during compression (Lehmann, 1997). However,
it should be kept in mind that the drug release properties of the coat could also be
altered when the coat is stretched and deformed, even though it is not ruptured (Tunon
et al., 2003b).
The compression characteristics of pellets coated with Surelease were studied by
Maganti and Celik (1994). It was found that Surelease imparted plasto-elastic
properties to the original brittle and elastic nature of uncoated pellets. Increase in rate
of compression also reduced the plastic flow and extent of consolidation, resulting in
weaker compacts. In addition, the sustained release properties were lost at low
compression pressures.
Due to the brittleness of ethylcellulose aqueous-based film coat, several methods have
been explored to recover or reduce the coat damage. In a study, pellets coated with
Aquacoat plasticized with 24 % dibutyl sebacate were tabletted with microcrystalline
cellulose and heated in an oven at 75 °C for 24 hours (Béchard and Leroux, 1992). It
was found that the drug release from heated tablets was slower than unheated tablets.
The partial recovery from coat damage after heating was attributed to sintering of
fissures after exposure to temperature above the film glass transition temperature. In
another study, micronized sodium chloride was used as a pore former of
ethylcellulose microcapsules (Tirkkonen and Paronen, 1993). Although the addition
of sodium chloride accelerated the drug release from uncompressed microcapsules, it
also made the microcapsule wall firmer and more resistant to damage during
compression. However, the microcapsules were not completely free from damage
after compression. Besides the methods discussed, undercoats and overcoats have also
been applied to ethylcellulose aqueous-based film coated pellets to reduce coat
7
damage during compression. This will be further elaborated in a later section (section
1.7) on protective overcoats and undercoats.
In contrast to the findings mentioned above, some other studies that used
ethylcellulose coating with ethanol as the solvent achieved preservation of coat
integrity after compression (Yao et al., 1998, Tunon et al., 2003a, b). This was
attributed to the higher tensile strength conferred by the organic solvent based
ethylcellulose film coats (Bodmeier et al., 1997b). It could also be due to the
differences in coating cores and cushioning additives used in the studies.
Most of the studies which used acrylic polymers for coating reported insignificant
changes in the drug release profiles of the coated multiparticulates after compression
(Vergote et al., 2002, Dashevsky et al., 2004, Debunne et al., 2004, Sawicki and
Lunio et al., 2005). In comparison with ethylcellulose films, acrylic films generally
have higher tensile strength and elongation at break, especially when plasticizers are
added (Lehmann et al., 1997). It has been reported that no or only small changes in
drug release were found in coatings with elongation values in excess of 75 %
(Bodmeier, 1997a). This indicates the importance of the tensile strength and
flexibility of the coat in determining its resistance to rupture during compression. The
use of flexible acrylic coating, such as Kollicoat SR with 10 %, w/w triethyl citrate as
the plasticizer, has also been shown to achieve similar drug release profiles for
compacts with 25-90 % pellet content, subjected to different compression forces (5-25
kN) (Dashevsky et al., 2004). Hence, such coatings enable the application of a larger
range of compression forces to produce tablets with desirable mechanical properties
and higher drug contents supplied by the pellets.
8
Although film coats with high tensile strength and elongation at break are less likely
to rupture upon compression, it should also be noted that excessive strength of the
pellets’ film coats may result in compacts possessing poor mechanical strength.
Aulton et al. (1994) reported that polymeric films exhibiting significantly greater
resilience than the uncoated cores were inappropriate if the coated cores were
intended to be compressed into tablets. This is due to the tendency of these film-
coated pellets to exhibit excessive elastic recovery on removal of compression load,
resulting in compacts with weaken mechanical strength. Furthermore, the addition of
excessive plasticizer in order to increase the coat flexibility can cause adhesion
problems during storage if the coated pellets were not tabletted immediately after
coating (Abbaspour et al. 2008).
1.2.2. Coat thickness
The thickness of a film coat applied onto a solid dosage form is typically between 20
to 100 μm (Hogan, 1995). The amount of coat applied is inversely proportional to the
drug release rate; the thicker the coat, the slower the drug release (Harris and Ghabre-
Sellassie, 1997, Wheatley and Steuernagel, 1997). This is due to the greater diffusion
pathway associated with thicker coats. The amount of coat to be applied would thus
be dependent on the desired drug diffusion rate.
Most of the studies that investigated the effect of coat thickness on its ability to resist
coat damage demonstrated that multiparticulates with thicker coats were more
resistant to coat damage (Beckert et al., 1996, Li et al., 1997, Lundqvist et al., 1998,)
However, in some cases where the film coats were very brittle, increasing the coat
thickness might not be able to reduce the damage to the coatings (Beckert et al., 1996,
Maganti and Celik, 1994).
9
Coat thickness had also been reported to be able to affect the deformation
characteristics of the pellets, as well as the mechanical strength and elastic recovery of
the compacts containing only Surelease coated pellets (Maganti and Celik, 1994). It
was concluded that coating pellets with Surelease changed the deformation
characteristics of the pellets from being elasto-brittle to elasto-plastic. With an
increase in the coat thickness, the total ability of the pellets to deform plastically and
elastically increased.
1.3. Coating core
Most of the reported studies on compression of coated multiparticulates had made use
of pellets as the coating cores (Beckert et al., 1996, Lundqvist et al., 1998, Salako et
al., 1998, Tunon et al., 2003a). The investigators studied on how the strength (Beckert
et al., 1996), deformability (Tunon et al., 2003a) and porosity (Tunon et al., 2003a) of
the core pellets affected their ability to stay intact during compression. Tunon et al.
(2003a) concluded that the use of more porous and deformable pellets was more
advantageous in terms of preserving the drug release profile after compression
compared with hard, non-porous pellets. It was suggested that cracking of pellet coat
and hence, increase in drug release, could had resulted from a built-up of local high
stresses during compression at pellet-pellet contact points. The porous pellets were
more deformable, enabling the pellet to pellet transmitted forces to spread out over a
larger area of contact, resulting in less coat damage. In addition, it was mentioned that
the ability of a coat to adapt to both shape and volume changes in the core indicated
that the coat was also compressible and thus rendered less permeable. In contrast to
the findings of Tunon et al. (2003a), Beckert et al. (1996) reported that harder pellets
could preserve the coat function to a greater extent compared to soft pellets. This was
attributed to the ability of harder pellets to withstand compression forces better than
10
softer pellets. These contradictory findings could be due to the brittleness of the film
coat (elongation at break < 5 %) used in that study. Unlike the elastic coating used in
the study done by Tunon et al. (2003a), the brittle coat used was unable to adapt to the
shape and volume changes in the soft pellets during compression. In addition, since
hard pellets have been reported to have a higher resistance to deform and fracture
compared to soft pellets (Salako et al. 1998), the greater degree of deformation by
softer pellets could have resulted in a larger extent of damage to the inflexible coat.
Although the extent of coat damage was less for harder pellets, the coat function was
not totally preserved as reflected from the liberation of more than 20 % of drug
through the enteric coat after two hours in 0.1 M HCl. The threshold pressure beyond
which the hard pellets were more easily and drastically damaged could have been
exceeded due to insufficient cushioning effects imparted by the external additives
(Salako et al. 1998). Though the ability of the core to deform could prevent the coat
from rupturing, it should be noted that coated cores should not be excessively porous
and/ or weak. In the study conducted by Tunon et al. (2003a), it was observed that the
uncompressed coated porous pellets showed an immediate burst release and faster
overall drug release compared to uncompressed pellets of lower porosity. This was
attributed to uneven distribution of coating polymer on the pellet surface due to
increased surface roughness of porous pellets. Furthermore, the use of cores and film
coats that are easily deformable could result in the densification of the core and coat
during compression. This process would reduce the water penetration rate into pellets,
causing the drug release to be prolonged and modified (Tunon et al., 2003b).
Therefore, cores should not be excessively deformable during compression in order to
maintain the original uncompressed pellets’ drug release profiles.
11
With regard to the size of pellets, smaller pellets coated to the same weight gain as
bigger pellets had faster drug release due to the larger specific surface area of smaller
pellets and thinner film coats applied. A greater extent of coat damage was also found
for smaller pellets (Béchard and Leroux, 1992), which could be explained by the
lower mechanical strength of the thinner coat formed. Hence, pellets of different sizes
have to be coated to the same coat thickness in order to investigate if pellet size has
any effect on the extent of coat damage when compressed. In a study conducted by
Haslam et al. (1998), pellets of different sizes were coated such that they had similar
dissolution profiles before compression. It was observed that the smallest pellets were
least affected by the compression process, and their drug release profiles were similar
to that of uncompressed millispheres. This correlated directly to the pellet strength. In
addition, Ragnarsson et al. (1987) also found that larger coated pellets showed more
damage after compression. However, the thickness of the coat and the mechanical
strength of the pellets were not reported.
Besides compressing coated pellets, some other investigators have compressed
microspheres (Torrado-Duran et al., 1991, Tirkkonen et al., 1993, Celik and Maganti,
1994), coated granules (Bansal et al., 1993, Shimizu et al., 2003) and coated drug
particles (Yao et al., 1997, 1998, Yuasa et al., 2001). It is not known if these cores are
superior to pellets in the preparation of coated multiparticulates for compression. If
the flowability of these coated particles is similar to that of the added excipients, any
likelihood of segregation may be avoided. The mechanical strength of the compact
may be improved when smaller coated particles are used, due to the increase in the
specific contact surface area with other particles for bonding. Nevertheless, it could be
a considerable challenge to coat smaller, non-spherical particles. Small particles tend
to move as aggregates or clusters when fluidized due to their lower mass and greater
12
specific surface area to attract one another. A lower coating spray rate is required to
reduce particle-particle agglomeration but dry coating condition usually generates
more static charges, resulting in greater particle agglomeration. In addition, the
surface roughness of smaller but less regular drug particles may result in the coating
solution not spreading uniformly over the particles’ surfaces upon coating application.
Thus, the surfaces of the irregularly shaped particles may not be uniformly coated,
depending on their movement and orientation during fluidization and coat application.
1.4. Compact fillers
In most of the reported studies, the coated multiparticulates were compressed with
other materials which act to protect the coats of the multiparticulates from the
compression forces. However, Tunon et al. (2003a) did not include any cushioning
additives in their tablet formulations besides 0.5 %, w/w magnesium stearate. The
drug release profiles were not affected by compression when porous, deformable
coated pellets were used. This demonstrated the possibility of preserving the
multiparticulates’ drug release profiles without the use of external cushioning
additives if the core and the coat are sufficiently deformable and flexible. However,
coated pellets in direct contact with the die and punches were likely to be damaged if
no cushioning material was added (Wagner et al., 2000a). The absence of change in
dissolution profile could be due to the incomplete disintegration of the coated pellets,
and the faster drug releasing pellets compensated by slower released, undamaged
pellets. In order to cushion the coated pellets, especially those in direct contact with
the die and punches, as well as to ensure disintegration of the tablet in the dissolution
medium, it is essential to include external cushioning additives and disintegrants into
the tablet formulation. The co-compressed external additives are also important in
improving tablet strength and friability (Türkoğlu et al., 2004).
13
The properties of the different additives in affecting their ability to cushion the coated
multiparticulates have been studied by several investigators. Torrado and Augsburger
(1994) concluded that materials with low yield pressure had more superior cushioning
properties. They suggested that cushioning materials with lower yield pressure than
the pellets and the pellet coats were able to absorb the energy of compression as they
were preferentially deformed. Yield pressure values, obtained from the reciprocal of
the gradient of the Heckel plot, is a measure of the plasticity of a compressed
material; the smaller the yield pressure, the greater the plasticity of the material
(Paronen and Ilkka, 1996). Heckel (1961a, b) described the density-pressure
relationship in powder compression according to the following equation:
ln AKPD
+=⎟⎠⎞
⎜⎝⎛−11 Equation 1
where D is the relative density of the compact at pressure P. D is calculated by
dividing the apparent density of the compact by the particle true density of the
material. Constant K is the gradient of the linear portion, and constant A is the y-
intercept of the extrapolated line from the linear portion of the plot of ln ⎟⎠⎞
⎜⎝⎛− D11
versus P (Figure 2). The Heckel equation is based on the assumption that the rate of
change in density with pressure is directly proportional to the pore fraction.
14
P
A
ln ⎟⎟⎠
⎞⎜⎜⎝
⎛− OD11
ln AKPD
+=⎟⎠⎞
⎜⎝⎛
−11
DA
DO
DB = DA - DO
Figure 2. A typical Heckel plot (Heckel, 1961b).
The density-pressure relationship is linear and valid except at lower pressure where
the relationship is non-linear due to the predominance of particle movement and
rearrangement at the initial stage of compression (Heckel, 1961a). The densification
of powder during compression has been described as a three-stage process (Heckel,
1961b):
a) densification by filling the die,
b) densification by individual particle movement and rearrangement processes,
and
c) densification by particle deformation after interparticulate bonding has
become appreciable.
15
During densification, the three stages overlap with one another and do not occur
individually.
Constant A represents the degree of packing achieved at low pressure as a result of
rearrangement processes before appreciable amounts of interparticulate bonding takes
place (Heckel, 1961a). It can also be expressed as relative density, DA, using the
following equation:
DA = 1- e-A Equation 2
DO is the relative density of the loose powder at zero pressure. The difference between
DA and DO, DB, therefore represents the extent of particle movement and
rearrangement before consolidation. The variations of DO, DA and DB were
determined to be primarily a function of geometry i.e. particle size and shape (Heckel,
1961b). Constant K is a measure of the ability of material to deform and a correlation
was made between K and yield strength (Heckel, 1961b):
O31Kσ
≅ Equation 3
where σo is the yield strength in psi. The mean yield pressure, Py has been defined by
Hersey and Rees (1971) to be the reciprocal of constant K. Py has been shown to
correlate with the nominal single particle fraction strength (Patel et al., 2007).
The density-pressure relationship can be obtained by the “at-pressure” or “zero-
pressure” method. In the “at-pressure” method, the movement of the punches and the
change in pressure are monitored by an instrumented tablet press. In this case, the
accuracy of the monitoring system is of utmost importance. In the “zero-pressure”
method, compacts are made at different pressure and their densities are determined
16
after removal from die. A few disadvantages of this method have been discussed by
Heckel (1961a). Firstly, this method is relatively slower than the “at-pressure”
method. Secondly, data can only be obtained at pressure greater than that required to
form coherent compacts. On the other hand, the “at-pressure” method also has a few
disadvantages over the “zero-pressure” method. Firstly, the density value at pressure
contains an elastic component which can give a false low value to the yield pressure
(Fell and Newton, 1971). Secondly, tablet press instrumentation can be quite costly
and requires technical knowledge and skills (Armstrong, 2008).
Although the Heckel plot has been commonly used in the last few decades to
investigate the compression behaviour of materials, it has received several criticisms.
For some materials, the linear portion of the Heckel plot could not be found (Rue and
Rees, 1978). In addition, inconsistent yield strength values of the same materials have
been reported over the years. This could be due to the pressure dependency of Py
values. It has been shown that regardless of the densification and deformation
mechanism (elastic, plastic or brittle fracture behaviour), different Py values were
obtained when the same material was subjected to different compression pressures,
especially at high pressure (Patel, 2010). The study demonstrated that besides plastic
deformation, Py value was influenced by elastic deformation and strain hardening of
materials. Other than the method of determining the density-pressure relationship (at-
pressure or zero-pressure), the speed of compression could also affect the Py value
(Fell and Newton, 1971). It has also been shown that other experimental conditions
such as mode of die filling, state and type of lubrication, as well as the dimensions of
the die can influence the Py value (York, 1979). Hence, the values of the parameters
derived from the Heckel plot are not universally applicable. For comparing the
compression behaviour of different materials, it is also essential to standardize the
17
experimental conditions. The Heckel equation was critically evaluated and compared
with the Walker equation by Sonnergaard (1999). It was mentioned that due to the
normalization of volume through multiple transformations, an error of 1 % in true
density leads to an error of more than 10 % in the Py values. A small change in the
gradient derived from the linear portion of Heckel plots also results in a great change
in the Py value. Furthermore, the pressure range that corresponds to the linear portion
of Heckel plot frequently exceeds the relevant pressure range in the production of
tablets. In comparison with the Walker equation, Sonnergaard (1999) also showed
that the data fitted the Walker equation well at lower pressure range. This pressure
range might be of closer relevance to the pressure range used for tablet production
compared to the pressure range in the linear portion of Heckel plots. It was also
concluded that the Heckel model was less reproducible and had less discriminative
power as a general compression constant. Patel (2007) also concluded that the Walker
parameters had less pressure dependency while the Heckel plot had less
discriminative power. In order to eliminate the elastic component in the determination
of Py from the Heckel plot, the “zero-pressure” method might be considered a better
method. In addition, the compression behaviour of the materials could also be
investigated using the Walker equation since it has been considered to be superior to
the Heckel equation by some investigators. The compression behaviour of the
materials in the more relevant compression pressure range used for producing the
compacts could also be better understood using the Walker equation. Therefore, some
caution had to be taken when interpreting parameters derived using the Heckel
equation.
18
The Walker equation assumes that the rate of change of volume is proportional to the
pressure (Walker, 1923):
0V'V = - KV log P + C1 Equation 4
where V′ is the volume at pressure P, V0 is the true volume (at zero porosity), and the
volume ratio or relative volume, 0V'V , is represented by V. KV is the change in volume
ratio as the pressure increases logarithmically and C1 is a constant. In order to express
the change in volume ratio in percentage, V is multiplied by 100 in the equation:
100 V = -W log P + C1 Equation 5
where W is the compressibility coefficient which expresses the percent change in
volume ratio when the pressure increased by a factor of 10.
Besides the Heckel equation, a new three-dimensional model developed by Picker
(2000) was used to evaluate the compression behaviour of materials. It allows the
simultaneous evaluation of the three important variables in tabletting; pressure,
displacement (presented as ln ⎟⎠⎞
⎜⎝⎛− D11 ) and time. The data was fitted into a twisted
plane. A few parameters describing the compression behaviour of the material could
be derived. These include parameter d, which quantifies the time plasticity, e, which
quantifies the pressure plasticity, and ω, which indicates the amount of fast elastic
decompression (Picker, 2000, 2004, Schmid and Picker-Freyer, 2009). A material
which densifies quickly has a high d value, while a material which densifies easily
under low pressure has a high e value. In addition, a material which exhibits
relaxation (elasticity) during decompression has a low ω value (Picker, 2000, 2004).
19
This model was subsequently used to study the compression behaviour of external
additives for tabletting pressure-sensitive materials (Picker, 2004, Schmid and Picker-
Freyer, 2009). In both studies, the materials were characterized by a term called
general plasticity (GP). GP is a combination of the normalized parameters from the
model and the elastic recovery (ER) after tabletting:
GP = dnorm + enorm + ωnorm - ERnorm Equation 6
0
01
HHH100(%)ER −
= Equation 7
where H1 is the height of the tablet after 10 days, and H0 is the minimal height of the
tablet under load. The parameters were normalized using the maximum values of the
parameters determined in the studies. This eliminated the influence of the actual
values of the parameters.
According to the studies (Picker, 2004, Schmid and Picker-Freyer, 2009), the extent
of coat damage correlated well with the GP, ω and e values of the materials. It was
concluded that materials with lower GP, ω and e values induced less coat damage.
This indicated that materials which exhibited high fast elastic decompression for
instantaneous release of pressure from the pellets and low plasticity were more
suitable for preventing coat damages. However, as mentioned earlier, Torrado and
Augsburger (1994) found that materials with lower Py values had better pellet coat
protective effect. As low Py values from Heckel plots were often interpreted as high
plasticity (Paronen and Ilkka, 1996, Sonnergaard, 1999), the findings of Torrado and
Augsburger (1994) appears to be contradictory to the findings of Picker (2004), as
well as Schmid and Picker-Freyer (2009). This could be because the elastic
component of the materials was included in the Py values derived from Heckel plots,
20
giving a false low value to the yield pressure (Fell and Newton, 1971). Clearly, Py
values have to be interpreted with caution, particularly when the “at-pressure” method
was used to obtain the Heckel plots.
Another method used to investigate the compression behaviour of external additives
in compressed coated multiparticulate system is the stress-powder column thickness
compression profile (Yuasa et al., 2001). Figure 3 shows a typical plot of compression
stress against the thickness of powder column. ‘R’ denotes the beginning of the
compression cycle, ‘S’ is the point where the compression stress reaches the
maximum, and ‘T’ is the point where the thickness of the powder column reaches the
minimum. The area of the hysteresis loop (area RSU), the area of the decompression
process (area STU), and the total area of compression (RSU + STU) corresponds to
the plastic energy (PE), elastic energy (EE), and compression energy (CE)
respectively (Asano et al., 1997). The percentage of plastic energy is calculated by
expressing the PE/CE ratio in percentage. It has been reported by Yuasa et al. (2001)
that materials with a high percentage of plastic energy and high percentage reduction
in volume have high stress dispersibility, which prevented the film fracture of coated
particles.
The percentage reduction in volume was calculated using the following equation:
Dp = (Vap – Vat) / Vap x 100 Equation 8
where Vap is the apparent volume (cm3) of the powder prior to compression and Vat is
the apparent volume (cm3) of the tablet after compression (Yuasa et al., 2001).
21
Thickness of powder column
EE
Stress
PE
U T R
S
Figure 3. Plot of the compression stress against thickness of powder column (Asano et al., 1997).
Other than correlating the compression behaviour of the external additives to the
cushioning capability of the external additives, the influence of the particle size of
external additives on the protection of the film coat from mechanical damage during
compression has also been studied (Yao et al., 1998). It was concluded that additives
with smaller particle size resulted in superior protection of the film coat from
compression damage. It was reported that 20 μm appeared to be the upper limit of the
particle size for additive to confer good protective effect. However, the mechanism by
which micronized additives could have contributed to the superior protective effect
was not studied. In contrast to the results of this study, another study reported that
improvement in coat protection could be achieved by compressing the coated pellets
with external additives of larger particle size (Avicel PH 200, Meggle lactose EP,
grade D 10) (Haslam et al., 1998). It was explained that the usage of external
22
additives of larger particle sizes resulted in increased excipient-excipient interaction,
producing an environment where less compression forces would be impacted on the
coated pellets.
Though many studies have shown that the physical properties of the external additives
could influence their cushioning capability, there was a study (Dashevsky et al. 2004)
which reported insignificant differences in the protective effects of different additives
(Avicel PH 200, Flowlac and Ludipress). All the three additives demonstrated good
cushioning properties and drug release from the tablets containing these additives was
almost identical to the original uncompressed coated pellets. This observation could
be due to the high flexibility (elongation at break up to 137 %) of the film coat used.
This suggests that the different capability of the external additives in protecting the
film coat during compression has little significance if the pellets have highly resilient
coats.
The studies that have been discussed so far had used additives in their powdered form
for compression. Segregation of pellets and powdered additives could had occurred,
resulting in poor component mix in the compacts. Segregation occurs when the
components of a formulation have different size, shape and density (Twitchell, 2007).
In addition, the poor flowability of fine powdered additives could affect the overall
flowability of the mixture, producing compacts with poor drug and weight uniformity.
In order to resolve this problem, many investigators used drug containing pellets,
disintegrant granules and cushioning granules of similar size, shape and density to
make the compacts. In addition, it was reported that granules could reduce the
deformation of pellets at the compact surface to a greater extent compared to additives
in the powder form (Wagner et al., 2000a). Hence, because of the potential
advantages that cushioning granules could offer, many studies were geared towards
23
finding the properties that cushioning granules should possess in order to preserve the
coat integrity of pellets during compression. Most of the studies concluded that
granules that were softer and more easily deformable acted better as cushioning
agents (Tunon et al., 2003b, Vergote et al., 2002). This was due to the ability of soft
granules to deform preferentially and form a matrix encasing the pellets (Salako et al.,
1998). Thus, several studies were conducted to find ways of improving the
deformability of granules. Methods that have been explored include the incorporation
of soft materials such as polyethylene glycol (Nicklasson and Alderborn, 1999) and
paraffin wax (Vergote et al., 2002) in the formulation of cushioning granules.
Cushioning granules with different porosity and deformability, were also prepared by
varying the proportion of water to ethanol in the granulation liquid (Tunon et al.,
2003b), varying microcrystalline cellulose / lactose ratio and adding different types of
superdisintegrants into granules made by freeze-drying (Habib et al., 2002). It should
be noted that because of the ability of these cushioning granules to deform and
consolidate, they often resulted in matrices with poor disintegration properties
(Vergote et al., 2002). Therefore, it is essential to include sufficient amounts of
suitable disintegrants into the compact formulations in order to preserve the
advantages of coated drug containing pellets, as well as to avoid modification to the
drug release profile as a result of a non-disintegrating compact. It was demonstrated
that with the use of sufficiently deformable cushioning granules and suitable
disintegrants, it was possible to produce compressed coated pellets that were able to
disintegrate and deliver individual coated pellets undamaged (Vergote et al. 2002).
1.5. Proportion of coated pellets
In order to prevent the coated pellets in a compact from damage, sufficient amount of
external additives has to be included to form a network of deformable materials which
24
is able to cushion the coated pellets from the compression forces. There has to be an
optimal amount of coated pellets that could be packed into the volume of a compact
without being deformed or at least, not being excessively deformed, if a certain
amount of particle deformation is tolerated. Random close packing is used to
characterize the maximum volume fraction of solid objects, which is obtained by
packing these objects in a random manner via shaking or tapping. The volume
fraction of monodisperse spheres in a random close pack is 0.64 (Wikipedia, 2008).
The maximum volume fraction of pellets that could be packed into the restrictive
dimension of a tablet without being deformed will be smaller than 0.64, especially
when the size of the pellets is relatively big when compared to the tablet size. When
considering the quantity of coated spherical pellets to be added into the tablet
formulation, the volume fraction of the coated pellets relative to additives should be
considered.
It was reported that tablets with homogeneous distribution of pellets, as well as good
mass and content uniformity could be obtained when 70 %, w/w (35 %, v/v) of pellets
were compressed with Avicel PH 101 (Wagner, 1999, 2000b). However, by
incorporating such a high pellet content into the tablet formula, extensive coat damage
was observed and tablet disintegration was slow (Wagner, 2000b). Hence, the pellet
content was reduced to 60 %, w/w (29 %, v/v) in order to comply with requirements
of USP 23 for disintegration time and drug release. However, the flowability of the
mixture was compromised with the reduction of pellet content. Thus, a glidant,
Aerosil 200, was added to achieve good mass and content uniformity. This study
demonstrated the importance and influence of pellet content on the extent of coat
damage, tablet disintegration, as well as the flowability of the mixture which affects
the mass and content uniformity.
25
In other studies, the influence of the proportion of coated pellets on the friability of
tablets was also investigated. In contrast to the findings of Wager (2000b), an increase
in the proportion of coated pellets was found to increase compact friability, decrease
disintegration time and mechanical strength (Lundqvist et al., 1997, Debunne et al.,
2004). This was attributed to the disruption of the continuous matrix formed by
readily deformable cushioning additives with the incorporation of a larger proportion
of less deformable coated pellets (Debunne et al., 2004). Hence, the bonds formed
between the coated pellets were weaker than that between the cushioning additives,
forming weaker compacts as the proportion of coated pellets increased. On the other
hand, the bonds between the coated pellets used in the study conducted by Wagner
(2000b) were probably stronger than the bonds between the filler particles. This
resulted in the formation of stronger compacts with longer disintegration time as the
proportion of pellets in the compacts increased.
With an increase in the proportion of coated pellets, the extent of coat damage either
increased or remained almost similar (Dashevsky et al., 2004, Vergote et al., 2002,
Lundqvist et al, 1998, Beckert et al., 1996, Torrado and Augsburger., 1994). In a
study conducted by Dashevsky et al. (2004), the extent of coat damage was found to
increase with an increase in the proportion of coated pellets when the pellet coats
(Aquacoat ECD 30 with 25 %, w/w triethyl citrate) were brittle. On the other hand,
when the pellet coats (Kollicoat SR 30 D with 10 %, w/w triethyl citrate) were more
flexible, the extent of coat damage was independent of the proportion of coated pellets
used. This indicates that the extent of coat damage with an increase in the proportion
of coated pellets was dependent on the type of polymer coat used. The greater extent
of coat damage with higher proportion of coated pellets was mainly attributed to a
reduction in the cushioning effect of the additives. It was also partially attributed to
26
decreasing ability of formulations to deform plastically at lower concentrations of
cushioning additives as revealed by the Heckel plot (Vergote et al., 2002).
1.6. Compression pressure
The influence of compression pressure on coat integrity of compressed pellets varied
and was found to depend on the other factors that affected coat integrity. In one study
(Dashevsky et al., 2004), coat integrity either remained unchanged or decreased with
an increase in the compression pressure, depending on the type of pellet coat. With a
flexible coat (Kollicoat SR 30 D with triethyl citrate), drug release profile remained
similar to uncompressed coated pellets regardless of the compression pressure used.
However, when a less flexible coat (Aquacoat ECD with triethyl citrate) was used,
increase in compression pressure resulted in an increase in the extent of coat rupture
as reflected by the faster dissolution profiles. In some other studies, an increase in the
compression force resulted in decreased rate of drug dissolution (Lundqvist et al.,
1998, Torrado and Augsburger, 1994, Bansal et al., 1993). This was probably due to
increased disintegration time rather than decreased pellet coat damage with an
increase in compression pressure. Furthermore, an increase in the compression
pressure also could had resulted in densification of the compressed pellets, producing
a decrease in the rate of water penetration and thus slower drug release (Tunon et al.,
2003b).
1.7. Protective overcoat and undercoat
Overcoats and undercoats have been used with the intention to protect the functional
coat layer during compression. Overcoats are layered over the functional coat while
undercoats are layered underneath the functional coat. The overcoats and undercoats
usually contain polymeric materials (Haslam et al., 1998, Altaf et al., 1998, 1999, El-
27
Gazayerly et al., 2004, Chambin et al., 2005). In some cases, in addition to the
polymeric overcoat, a final outermost layer of excipients was sprayed onto the coated
pellets for faster disintegration and to provide additional cushioning effect (Altaf et
al., 1998, 1999, El-Gazayerly et al., 2004). The excipients that have been used include
mannitol (Altaf et al., 1998), microcrystalline cellulose and/or polyethylene glycol
with sodium starch glycolate (Altaf et al., 1999), as well as lactose with sodium starch
glycolate (El-Gazayerly et al., 2004). Pellets layered with these excipients were
compressed directly without additional filler materials. This prevented the prevalent
problem of segregation between coated pellets and filler materials. However, the
outermost cushioning layer did not protect the functional coat completely in these
studies. In addition, there were difficulties compressing pellets multi-layered with the
cushioning excipients. This was due to the increase in pellet size after layering and
decrease in the surface area-to-volume ratio for bonding (Altaf et al., 1998).
Hypromellose was employed for overcoating the coated pellets in two studies
(Haslam et al., 1998, Chambin et al., 2005). It was chosen due to its water solubility
and thus, lower possibility to affect the drug release profile. Furthermore, the
hypromellose film coat is soft and pliable and therefore suitable as a cushioning layer
around the coated pellet. The overcoat did not appear to protect the sustained release
coat in one of the reported studies (Chambin et al., 2005). In addition, the
hypromellose overcoat was found to increase the drug release rate. Haslam et al.
(1998) reported that hypromellose overcoat was able to protect the underlying coat of
small pellets (0.5 mm) when compressed with filler materials of larger particle size.
The soft and pliable overcoat appeared to have the capability to protect the underlying
coat through viscous deformation during compression. Hence, hypromellose
overcoats can be useful for protecting the functional coat during compression.
28
In another study (El-Gazayerly et al., 2004), an undercoat consisting of polyethylene
oxide and hypromellose was spray coated on drug-layered pellets before applying the
Surelease coat. The final outer disintegrant layer consisting of lactose and sodium
starch glycolate was then sprayed into the pellets. The coated pellets were compressed
into tablets without additional filler materials. It was found that tablets containing
pellets spray coated to sufficient weight gain of undercoat had similar drug release
profile as the uncompressed coated pellets. This was attributed to ability of
polyethylene oxide to hydrate and retain water within its structure, providing
sufficient viscosity to seal the cracks formed in the Surelease coat during
compression. Hence, the undercoat was able to provide a sealing effect to coat
ruptures. Polyethylene oxide was also used in another study as an overcoat (Altaf et
al. 1999). However, the drug release was still faster than the uncompressed pellets at
the later phase of the dissolution study.
29
Table 1. Summary of the effects of parameters that were found to influence the extent of coat damage during compression
Parameters affecting the extent
of coat damage during compression
Extent of coat damage
References
Coat flexibility / strength
Bansal et al., 1993, Magnati and Celik, 1994, Sarisuta and Punpreuk, 1994, Bodmeier, 1997a,
Lehmann, 1997, Lundqvist et al., 1998, Vergote et al., 2002, Dashevsky et al., 2004, Debunne et al.,
2004, Cantor et al., 2009
Coat thickness Beckert et al., 1996, Lundqvist et al., 1998, Salako et al., 1998, Tunon et al., 2003a
Protective overcoat / undercoat Haslam et al., 1998, Altaf et al., 1998, 1999, El-
Gazayerly et al., 2004, Chambin et al., 2005
Pellet core size / Ragnarsson et al., 1987, Béchard and Leroux, 1992, Haslam et al., 1998
Pellet core porosity Tunon et al., 2003a
Pellet core deformability Tunon et al., 2003a
Pellet core strength Beckert et al., 1996
Compression pressure / /- Lundqvist et al., 1998, Torrado and Augsburger,
1994, Bansal et al., 1993, Dashevsky et al., 2004
Plasticity of fillers / Torrado and Augsburger, 1994, Picker, 2004, Schmid and Picker-Freyer, 2009
Fast decompression elasticity of fillers Picker, 2004, Schmid and Picker-Freyer, 2009
% reduction in volume of fillers Yuasa et al., 2001
Filler particle size / Yao et al., 1998 Haslam et al., 1998
Proportion of coated pellets / /-
Wagner, 1999, 2000b Dashevsky et al., 2004, Vergote et al., 2002,
Lundqvist et al, 1998, Beckert et al., 1996, Torrado et al., 1994
: Denotes decrease in level of coat damage when factor parameter increased
: Denotes increase in level of coat damage when factor parameter increased
-: Denotes insignificant change in level of coat damage when factor parameter increased
30
1.8. Methods for assessing the coat integrity of compressed pellets
The most commonly reported method employed to assess the coat integrity of
compressed pellets involves the use of dissolution rates. The drug release profiles of
compressed pellets are usually compared with those of uncompressed pellets. Any
change in the coating layer is expected to be reflected by the dissolution profile due to
changes in coat functionality. Several methods can be employed to compare the drug
dissolution profiles. As a first step in obtaining an understanding of the dissolution
data, the dissolution data points of the different formulations can be plotted on the
same graph, allowing the profiles to be compared visually. This method is also known
as the exploratory data analysis method (O’Hara et al., 1998). The error bars
corresponding to the standard deviation of each dissolution data point are entered in
order to determine if the dissolution profiles are significantly different. If the error
bars of the different dissolution curves do not overlap, the dissolution profiles may be
considered to be significantly different (O’Hara et al., 1998). However, the
comparisons made are not necessarily quantitative and may not be feasible when more
formulations with differing dissolution profile shapes are to be compared. Other
methods of comparing dissolution profiles can be divided into model-independent and
model-dependent methods. Through these methods, parameters which characterize the
dissolution profile may be determined and compared by one-way analysis of variance
(ANOVA). In addition, the parameters obtained from the dissolution profile of one
formulation can also be expressed as a ratio of that obtained from the other
formulation.
One major advantage of model-independent methods is that it does not require the
dissolution profile to be fitted into a mathematical dissolution model. Hence, these
methods are useful for comparing dissolution profiles with complicated or different
31
release mechanisms. The model-independent methods include the comparison of time
points and mean dissolution time (MDT). In the former method, the time taken for a
certain percentage of drug to be released is determined from the dissolution curve
(tx%). Although it is relatively simpler to determine Tx%, drug dissolution profile
comparisons made using this parameter are specific to a single point on the
dissolution profile and not the entire dissolution profile. For the latter method, MDT
is calculated using the following equation:
MDT = ∑
∑
=
=
Δ
Δj
1ii
j
1iii
M
Mt̂ Equation 9
where i is the sample number, j is the number of dissolution sample times, t̂ is the
time point between ti and ti-1 ( 2ttt̂ 1ii −+
= ) and ΔMi is the fractional amount of drug
released between ti and ti-1 (Costa and Sousa Lobo, 2001, Costa et al., 2003). In
contrast to the time point method, MDT is a parameter which characterizes the
dissolution profile more completely since all the dissolution data points are used in its
determination. However, differences in the dissolution rates may not be detected for
formulations with zero order release mechanism, in particular dissolution profiles
with incomplete release of drug (Podczeck, 1993, Costa et al., 2003). The MDT value
is also dependent on the dose-solubility ratio (Lánsky and Weiss, 1999, Rinaki et al.,
2003a). Besides MDT, the variance of the dissolution time (VDT) and the relative
dispersion (RD) of the concentration-time profiles can be determined by the following
equations (Costa et al., 2003):
32
∑
∑
=
=
Δ
Δ−= n
1ii
n
1ii
2i
M
M)MDTt̂(VDT , and Equation 10
2MDTVDTRD = Equation 11
The RD value has been correlated to the dissolution mathematical models and drug
release mechanisms (Pinto et al., 1997). Thus, a change in the RD value could
indicate a compression-induced change to the structure of the pellet coating
(Lundqvist et al., 1998, Tunon et al., 2003b). The MDT, VDT and RD values have
been used to compare the dissolution profiles of coated and uncoated pellets (Chopra
et al., 2002, Sousa et al., 2002, Santos et al., 2004).
The difference factor (F1) and the similarity factor (F2) introduced by Moore and
Flanner (1996) have also been used to quantify the differences between two
dissolution profiles. These parameters are also classified as model-independent and
can be determined using the following equations:
)(100
2/TR
TRF n
1iii
n
1iii
1 ×+
−=∑
∑
=
= , and Equation 12
∑=
− ×−+×=n
1i
5.02
ii2 }100]TR)n/1(1log{[50F Equation 13
where n is the total number of sampling time points, Ri and Ti are the mean drug
release for the reference and test formulations at time point i respectively. The F1 and
F2 values range from 0 to 100. The U.S. Food and Drug Administration (FDA, Center
for Drug Evaluation and Research, 1995, 1997), as well as the European Agency for
33
the Evaluation of Medicinal Products (EMEA, 1999) consider two dissolution profiles
to be similar if the F2 value falls between 50 to 100. Though F2 has been adopted by
the FDA (Center for Drug Evaluation and Research, 1995, 1997) and European
Agency for the Evaluation of Medicinal Products (EMEA, 1999) to assess the
similarity between two dissolution profiles, it has several limitations. Firstly, F2 does
not take into account of the dissolution curve shape and the variation in spacing
between sampling times (Costa, 2001). Simulation results also showed that F2 was too
liberal in concluding similarity between dissolution profiles (Liu and Chow, 1996, Liu
et al., 1997). This could potentially lead to false conclusions. Furthermore, F2 is a
function of the mean differences and does not take into account of the intra-batch
variation of test and reference batches (Shah et al., 1998). The selection and
determination of the dissolution data points for caluculation also play a critical role in
affecting the F2 value (Polli et al., 1997). Despite its limitations, F2 has been
concluded to be superior to Tx% and MDT when comparing dissolution profiles with
crossover (Pillay and Fassihi, 1998).
The model-dependent methods involve fitting the dissolution data points to a
mathematical model. A disadvantage of these methods is that a good fit to a
mathematical model is required. In some cases, it could be difficult to find a model
which describes the dissolution profile correctly due to the multiple drug release
mechanisms involved. In addition, parameters derived from the models can only be
used for comparison when all the dissolution profiles fit considerably well to the same
dissolution model. Nevertheless, this method provides an additional advantage of
providing insights into the drug release mechanisms. In addition, a change in the drug
release mechanism as a result of modifications to product formulation or processing
condition can be detected. In this study, some of the commonly used mathematical
34
models were employed for curve-fitting. They include zero order model, Higuchi
model and power law (Korsmeyer-Peppas model). The Higuchi and zero order models
represent two limited cases in the release mechanisms and the power law can be used
as a decision parameter between these two models (Costa and Sousa Lobo, 2001).
The zero order model describes a constant rate of drug release with time according to
the equation:
tKQQ 00t += Equation 14
where Qt is the amount of drug dissolved in time t, Q0 is the initial amount of drug in
the solution and K0 is the zero order rate constant. Zero order drug release can be
achieved in osmosis driven systems which works by pumping of saturated drug
solution through the coating membrane with drilled orifices (Liu et al., 2000). The
saturated drug solution can also be transported through pores formed by the
dissolution of water-soluble component in the membrane or membrane expansion
with the influx of water (Schltz and Kleinebudde, 1997). Besides coated matrices,
zero order drug release could also be achieved in polymeric matrices through the
synchronization of swelling and surface erosion (Yang and Fassihi, 1996).
An equation which expresses the drug release rate from a planar system having a
homogeneous matrix in the form of ointment bases in perfect sink condition was
derived by Higuchi (1961, 1963):
ssHH C)CA2(tDQ −= Equation 15
where Q is the amount of drug released from the system at time t per unit area of
exposure, DH is the diffusion constant of the drug molecule in the external phase, AH
is the total amount of drug present in the matrix per unit volume, and Cs is the
35
solubility of the drug per unit volume. The equation was derived from Fick’s first law
and a one-dimensional theoretical drug concentration profile of an ointment in contact
with a perfect sink release media (Higuchi, 1961). Fick’s law states that the permeant
flux is a product of the concentration gradient in the membrane and the diffusion
coefficient of the permeant molecules through the membrane (Baker, 1987). Another
expression for describing the drug release from a planar system having a granular
matrix was also introduced (Higuchi, 1963):
tC)CA2(DQ ssHH ε−τε
= Equation 16
where Q is the amount of drug released from the system at time t per unit area of
exposure, DH is the diffusion constant of the drug molecule in the external phase, ε is
the porosity of the matrix, τ is the tortuosity factor of the capillary system, AH is the
total amount of drug present in the matrix per unit volume, and Cs is the solubility of
the drug per unit volume. It is apparent from these expressions that the drug release is
proportional to the square root of time. The Higuchi model is commonly simplified
and expressed as:
tKQ Ht = Equation 17
where Qt is the amount of drug dissolved in time t, and KH is the Higuchi dissolution
rate constant. As the Higuchi model assumes the fitted graph to pass through the
origin, it has been modified to include the dissolution lag time observed in some
matrices (Ford et al., 1991):
ctKQ Ht += Equation 18
36
where Qt is the amount of drug dissolved in time t, KH is the Higuchi dissolution rate
constant and cH is a constant. The lag time period is defined as –cH/AH.
Besides ointments, the Higuchi model has been used to characterize the drug release
from polymeric tablet matrices (Ford et al., 1985, Campos-Aldrete and Villafuerte-
Robles 1997) and uncoated pellets (O’Connor and Schwartz, 1993).
Due to the need for a general mathematical expression to describe deviations from the
Fickian diffusional release of drugs, the following equation, also known as the power
law was developed by Korsmeyer et al. (1983):
nP
t tKMM
=∞
Equation 19
where Mt and M∞ are the cumulative amount of drug released at time t and infinite
time respectively, KP is a kinetic constant characteristic of the drug/polymer system,
and n is the release exponent, indicative of the mechanism of drug release. When n=0,
the expression describes a drug release profile that follows the zero order kinetics,
also known as Case-II transport. On the other hand, an n value of 0.5 is indicative of
Fickian diffusion as the predominant drug release mechanism, corresponding to the
Higuchi model. An n value which falls between 0.5 and 1 indicates an anomalous
behaviour, with non-Fickian kinetic corresponding to coupled fusion/polymer
relaxation. The use of this equation for porous systems will probably lead to n values
less than 0.5 due to the combined mechanisms of diffusion through a swollen matrix
and water-filled pores (Peppas, 1985). Depending on the matrix geometry of the
dosage form, the n value indicative of each release mechanism may shift slightly.
These shifted values for cylinders and spheres are shown in Table 2 (Peppas and
Sahlin, 1989). Nevertheless, regardless of the matrix geometry, the n value indicative
37
of zero order release kinetics is double of that for Fickian release mechanism. In order
to take into account of dissolution lag time, the equation was modified (Ford et al.,
1991, Kim and Fassihi, 1997):
nLP
t )tt(KMM
−=∞
Equation 20
where Mt and M∞ are the cumulative amount of drug released at time t and infinite
time respectively, tL is the lag time, KP is a kinetic constant characteristic of the
drug/polymer system, and n is the release exponent, indicative of the mechanism of
drug release. For matrices which exhibit an initial burst release, the following
equation can be used (Kim and Fassihi, 1997):
p
n
Pt btk
MM
+=∞
Equation 21
where Mt and M∞ are the cumulative amount of drug released at time t and infinite
time respectively, kP is a kinetic constant characteristic of the drug/polymer system, n
is the release exponent, indicative of the mechanism of drug release, and bp is the
burst effect. The tL and bp values will be zero in the absence of lag time and burst
effect. Although the power law equation was recommended to be applied only to the
initial 60 % of drug release profile when it was first introduced (Peppas, 1985), it has
been shown to be applicable to the entire drug release profile (Rinaki et al., 2003b).
38
Table 2. Release exponent (n) derived from the power law and the corresponding release mechanism (Peppas and Sahlin, 1989).
Mechanism Release exponent (n)
Film Cylinder Sphere
Fickian diffusion 0.50 0.45 0.43
Anomalous transport (non-
Fickian) 0.50 < n < 1.00 0.45 < n < 0.89 0.43 < n < 0.85
Zero order release or Case II transport
1.00 0.89 0.85
As can be seen, dissolution profiles can be compared by different methods and each
method has its own advantages and limitations. Some of the methods mentioned have
been used in assessing the coat damage in compacts containing coated
multiparticulates. The model-independent methods are more commonly used. These
include the single point comparison, exploratory data analysis method and
comparison of MDT values. Since dissolution studies remain as the main tool to
assess the extent of pellet coat damage, it is important to ensure that the chosen
method of data analysis yields informative and accurate conclusions. In addition, one
major disadvantage of dissolution studies is that drug dissolution profile is also
affected by the disintegration time of the compact. Compacts with longer
disintegration time are expected to exhibit correspondingly slower drug release rates.
Hence, the disintegration time itself could play a dominant role in affecting the
perceived rate of drug release and obscure the real extent of coat damage. It is thus
essential to carry out disintegration studies for the compacts in order to ensure that the
observed drug release is largely caused by changes in the coat integrity rather than a
consequence of the disintegration time.
39
As a result of this disadvantage of dissolution studies, other methods to detect coat
damage should be used to substantiate the conclusions derived from the dissolution
profiles. These methods include capturing images of the compressed pellet surface
using scanning electron microscopy (SEM) and light microscopy. As these methods
provide qualitative results, they are often used only to substantiate any observed
changes in dissolution studies and validate the conclusions made from the dissolution
data.
40
PART 2: HYPOTHESES AND
OBJECTIVES
41
2. HYPOTHESES AND OBJECTIVES
The compression of coated multiparticulates without affecting the coat function
remains a challenge. Due to the multiple factors involved in influencing the pellet coat
damage during compression, conflicting results were often reported in different
studies. A review of the literature reveals a lack of understanding of the pellet-filler
interaction and mechanism of pellet coat damage during compression. Very often,
studies were more product-driven and geared only towards finding the optimal
formulation in such dosage forms without an in-depth understanding of the cause of
pellet coat damage or protection. Hence, this study aimed to understand the pellet-
filler interactions during compression so as to mitigate the extent of pellet coat
damage in such systems.
Physical properties of fillers that have been correlated to pellet coat damage include
the compression behaviour, particle size as well as porosity. Contradictory results
have been reported for the former two properties (Torrado and Augsburger, 1994,
Haslam et al., 1998, Yao et al., 1998, Picker, 2004). It is unknown how the properties
affected the pellet-filler interactions, hence affecting the extent of pellet coat damage.
Furthermore, no studies to date have investigated the two properties simultaneously.
In addition, pellet-filler interactions were also expected to change with compression
pressure and pellet volume fraction (Torrado et al., 1994, Beckert et al., 1996,
Lundqvist et al, 1998, Wagner, 2000b, Vergote et al., 2002). It was thus hypothesized
that pellet coat damage could be mitigated by compressing the coated pellets with
fillers of appropriate particle size and compression behaviour to a suitable pellet
volume fraction.
42
Besides the properties of fillers, the pellet core properties have been reported to affect
the extent of coat damage during compression. The pellets which are commercially
available include sugar pellets and MCC pellets of different size fractions. However,
comparative studies of the two pellet types and pellet sizes have not been carried out
so far. It was hypothesized that the two pellet type interact differently with the filler
particles and pellet size is a factor affecting the extent of coat damage.
To test the aforementioned hypotheses, studies were carried out with the following
objectives:
1) Investigate the influence of particle size of lactose fillers on the extent of pellet
coat damage, as well as the pellet-filler interactions.
2) Investigate the influence of particle size and compressibility of various fillers, as
well as the pellet volume fraction on the extent of pellet coat damage.
3) Investigate the influence of pellet core size and pellet type on the extent of pellet
coat damage.
Coat damage was assessed by the determination of drug release properties. Purified
water was used as the dissolution media as the coating polymer, ethylcellulose is
insoluble in water while the model drug used, chlorpheniramine maleate is freely
soluble in water. Damages to coats that could potentially impact drug release
properties would be detected by changes in the dissolution profiles.
43
PART 3: EXPERIMENTAL
44
3. EXPERIMENTAL
3.1. MATERIALS
3.1.1. Pellet cores and coating materials
Sugar pellets (Nu-pareil, Hanns G, Werner GmbH + Co., Germany) of size fraction
500 - 600 µm were used as the coating cores for all the studies. Two additional size
fractions of sugar pellets, 250 - 355 µm and 355 - 425 µm, as well as microcrystalline
cellulose (MCC) pellets (Celphere CP-507, Asahi Kasei, Japan) of size fraction 500 -
600 µm were used to investigate the influence of coating cores. The coating materials
for the drug deposition layer consisted of 10 %, w/w hypromellose (HPMC;
Methocel-E3, Dow Chemical, Midland, MI) as the film former, 1 %, w/w polyvinyl
pyrrolidone (PVP; Plasdone C-15, ISP Technologies, Wayne NJ) as the co-film
former and 5 %, w/w chlorpheniramine maleate (CPM; BP grade, China) as the model
drug. Aqueous ethylcellulose dispersion (Surelease, Colorcon, US) was used to coat
the drug loaded pellets in order to achieve sustained drug release. Purified water was
used as the vehicle for all coating materials.
3.1.2. Compact excipients
The coated pellets were compressed with several types of lactose fillers. These
included a few grades of crystalline α-lactose monohydrate (Pharmatose 80M, 100M,
125M, 200M, 350M, and 450M, DMV, Netherlands), micronized lactose (see section
3.2.10 for method of size reduction) and spray-dried lactose (Pharmatose DCL11,
DMV, Netherlands). Other fillers employed were low-substituted hydroxypropyl
cellulose (L-HPC; LH-11, LH-20, LH-22, LH-31, Shin-Etsu Chemical, Japan),
microcrystalline cellulose (MCC; Avicel PH101, Avicel PH102, Avicel PH301,
45
Avicel PH302, Ceolus KG801, Ceolus KG802), crospovidone (Kollidon CL, CL-F,
CL-SF, CL-M, BASF, Germany), as well as dibasic calcium phosphate (DCP;
Emcompress, Edward Mendell, USA). Crospovidone (Polyplasdone XL, ISP, US) and
magnesium stearate (Riedel-de Haën, Germany) were used as disintegrant and
lubricant respectively. For the studies on the effect of compression pressure,
compression speed and coating cores, the fillers employed were micronized lactose,
Pharmatose 80M, Avicel PH200 and Avicel PH105.
In order to elucidate the effect of disintegration time on the dissolution profile,
another set of compacts were prepared using different disintegrants to vary the
disintegration time. The disintegrants used were sodium starch glycolate (Primojel,
AVEBE, Netherlands), microcrystalline cellulose (MCC; Avicel PH101, Asahi Kasei
Chemical, Japan), pregelatinized corn starch (Starch 1500, Colorcon, US) and low-
substituted hydroxypropyl cellulose (LHPC-11, Shin-Etsu Chemical, Japan).
Pharmatose 200M and magnesium stearate were employed as the diluent and
lubricant respectively.
46
3.2. METHODS
3.2.1. Preparation of coating materials
Hypromellose solution was prepared by first dispersing the hypromellose powder in
purified water heated up to 90 °C with a magnetic stirrer. Cold water was added while
stirring was continued for another 30 min. The mixture was then left in the
refrigerator to hydrate overnight at approximately 5 °C. The other materials (CPM
and PVP) were added into the hypromellose solution the next day and topped up to
the required weight.
The Surelease aqueous dispersion was diluted from 25 %, w/w solids to 15 %, w/w
solids and stirred for approximately 30 min before use.
3.2.2. Pellet coating
The pellets were coated with two consecutive functional layers, namely the drug
deposition layer followed by the sustained release layer using the Precision coater
module of the Multiprocessor (MP1 GEA-Aeromatic Fielder, UK). The coating and
drying conditions after coating are described in Table 3. In method A, the pellets
coated with ethylcellulose were dried at 60 °C for 24 h so as to produce cured coated
pellets. However, in method B, the pellets coated with ethylcellulose were dried at 35
°C for 24 h so as to produce uncured coated pellets. For the study in which the coated
pellets were compressed with a variety of fillers, both cured and uncured pellets were
used to investigate the effect of coat quality on the extent of coat damage induced
during compression. For other studies, coating method A was used. The fines and
agglomerates were removed using appropriate sieves after coating.
47
Table 3. Coating and drying conditions
Coating Method Process parameters Drug deposition layer Sustained release layer
Pellet load (kg) 0.5 0.5
80* 80*
70# 60# Airflow rate (m3/h)
50@ 50@ Atomizing air pressure (bars) 2 2
5.3* 5.6*
5.5# 5.4# Spray rate (g/min)
3.9@ 4.4@
Inlet air temperature (°C) 70 60 Accelerator insert inlet air diameter (mm) 30 30
18* 18*
10# 10# Partition gap (mm)
6@ 6@ Hot air oven drying temperature (°C) 60 60
A
Drying duration (h) 24 24
Pellet load (kg) 1 1
Airflow rate (m3/h) 80 100 Atomizing air pressure (bars) 2 2
Spray rate (g/min) 11.8 14
Inlet air temperature (°C) 70 60 Accelerator insert inlet air diameter (mm) 24 24
Partition gap (mm) 18 18
Drying temperature (°C) 60 35
B
Drying duration (h) 24 24
*500-600 µm sugar and MCC pellets; #355-425 µm sugar pellets; @250-355 µm sugar pellets
48
3.2.3. Preparation and characterization of compacts
All the compacts weighed 1 g each and contained 50 %, w/w coated pellets, 44 %,
w/w filler, 5 %, w/w disintegrant and 1 %, w/w lubricant. The ingredients for each
compact were weighed out individually and mixed in a 25 mL beaker for 1 min with a
spatula. The lubricant, magnesium stearate, was then added into the beaker and the
contents mixed for 30 s. The compacts were formed using a universal testing machine
(Autograph, AG-100kNE, Shimadzu, Japan) with a 100 kN load cell and 15 mm flat
face punches. Compacts were prepared at constant compression pressure of 28.3 MPa
(5kN) and varying compression speed of 1 mm/min, 10 mm/min, 50 mm/min, 100
mm/min and 300 mm/min to investigate the effect of compression speed. In the
investigation of the effect of compression pressure, compacts were prepared at
constant compression speed of 10 mm/min while the compression pressure was varied
in the range of 1.1 to 424.4 MPa (0.2 to 75 kN). Compacts were stored in a desiccator
for at least three days before further analysis.
Compact porosity, ε, was determined using the following equation:
T
app1ρ
ρ−=ε Equation 22
where ρapp is the apparent compact density and ρT is the true compact density. The
apparent compact density ρapp was calculated by:
hrW
2
capp π=ρ Equation 23
where Wc is the weight of the compact, r is the radius of the compact and h is the
compact thickness.
49
The true density, ρT, was calculated by:
i321T z....cba ρ+ρ+ρ+ρ=ρ Equation 24
where a, b c and z are the proportions of components 1, 2, 3 and i in the compact with
true densities ρ1, ρ2, ρ3, and ρn respectively.
The pellet volume fraction, Fp, was calculated according to the following equation:
app
pp V
VF = Equation 25
where Vp and Vapp are the true volume of the pellets and the apparent volume of the
compact respectively.
3.2.4. Dissolution studies
The dissolution studies were carried out using paddles rotated at 100 rpm in 500 mL
of deaerated purified water maintained at 37 °C ± 0.5 °C (Optimal DT-1 dissolution
tester, Optimal Instruments, US). At 5, 10, 15, 20, 30, 45, 60, 90, 120, 180 and 240
minutes dissolution time points, 5 mL samples were withdrawn and replaced with
purified water maintained at 37 °C ± 0.5 °C. At the end of the dissolution run, the
pellets were crushed with a glass rod and stirred at 300 rpm for 30 min. The final
sample was then withdrawn for the determination of total drug content. All samples
were filtered though membrane filters with 0.45 µm pore size (Regenerated cellulose
filters, Sartorius Stedim Biotech, Germany). The filtered samples were assayed
spectrophotometrically (UV-1201, Shimadzu, Japan) at 262 nm using the Beer’s plot
(Figure 4). Dissolution tests were carried out in triplicates and the average results
calculated.
50
Drug concentration (%w/v) x 10-3
y = 0.131x R2 = 1.000
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
0 1 2 3 4 5 6 7
Abs
orba
nce
Figure 4. Beer’s plot of chlorpheniramine maleate in purified water.
For ultraviolet (uv) spectroscopic studies, the components of the compacted coated
pellet systems were either insoluble (ethylcellulose, magnesium stearate and
microcrystalline cellulose), or soluble but non-uv active (lactose and sucrose). Thus,
the only uv-active ingredient was the drug, chlorpheniramine maleate.
3.2.5. Analysis of drug release data
Dissolution profiles of 1 g compacts and equivalent amount of uncompressed pellets
(500 mg) were compared. For most of the studies, the ratio of the T50% value (time
taken for 50 % of the drug to be released) of uncompressed coated pellets to
compressed pellets (UC/C) was used as an index to indicate the extent of coat
damage. The greater the value of UC/C, the greater was the extent of coat damage.
For the study on the effect of compression pressure, model-independent dissolution
parameters, T25%, T50%, T75%, MDT, VDT, RD and F2, were also determined. In
51
addition, the dissolution profiles were fitted into the zero order model, Higuchi model
and the power law model. Curve-fitting of the dissolution data was carried out using
KaleidaGraph 4.0 (Synergy software, US). The goodness-of-fit of the dissolution data
to the models was compared using R2 values. The equations for the different models
can be found in the section 1.8. As the uncompressed coated pellets exhibited a lag
phase at the beginning of dissolution, the equations which take into account of the lag
phase were used. The dissolution parameters were compared using ANOVA.
3.2.6. Disintegration test
The disintegration tests were conducted according to the procedure of the British
Pharmacopoeia disintegration test for tablets and capsules (Test A, Appendix XII A,
2008) using a disintegration tester (DT2-, Sotax, US). Purified water maintained at 37
°C ± 1 °C was used as the liquid media. Six compacts of each formulation were
subjected to the disintegration test and the results averaged.
3.2.7. Evaluation of the densification behaviour of materials
Compacts of the materials studied, weighing 1 g each, were formed using the
Autograph with pressures ranging from 5.7 to 67.9 MPa. The die wall and the
punches were pre-lubricated with magnesium stearate suspended in isopropyl alcohol
and dried before each compression cycle. A total of three compacts were formed at
each pressure. The compression behaviour of the materials was studied using the
Heckel equation (“zero-pressure” method):
ln AKPD
+=⎟⎠⎞
⎜⎝⎛−11 Equation 1
52
where D is the relative density of the compact at pressure P. D was calculated by
dividing the apparent density of the compact by the particle true density of the
material. The yield pressure was determined from the reciprocal of the slope at the
linear portion of the plot (R2 > 0.97).
Besides Heckel analysis, the densification behaviour of the materials were also
evaluated using the Walker equation:
100 V = -W log P + C1 Equation 5
where V is the ratio of volume at pressure P to the true volume (at zero porosity), W
is the compressibility coefficient which expresses the percent change in volume ratio
when the pressure increased by a factor of 10 and C1 is a constant.
3.2.8. Determination of particle true density
A helium pycnometer (Penta-pycnometer, Quantachrome, USA) was used to measure
the true volume of the material according to the USP method. The materials were
dried in a hot air oven and stored in a desiccator overnight before measurements. True
volume analysis was carried out until the percent deviation for three consecutive runs
was equal to or less than 0.005 % or until a maximum of five runs had been reached.
The true density of the material, ρt, was calculated by dividing the weight of the
sample, Ws, by its true volume V0 according to the following equation:
0
t Vwρ = Equation 26
53
3.2.9. Determination of material bulk density, tapped density and Hausner ratio
An excess quantity of material was passed through a 1.00 mm screen into a 50 mL
graduated cut cylinder. The excess material was scraped from the top of the cylinder
using a spatula. The weight of the material in the cylinder (Wt) was determined and
the bulk density, in g per mL, calculated using the formula:
Bulk density = B
tt
VW Equation 27
where VB is the volume of the cut cylinder.
The cut cylinder filled with the material was mechanically tapped using the
Stampvolumeter, Stav 2003 (Jel, Germany) until there was no further change in the
volume of the material (Vt). The tapped density in g per mL was calculated using the
formula:
Tapped density = t
t
VW Equation 28
Hausner ratio was calculated by the formula:
Hausner ratio = tV
50 Equation 29
3.2.10. Particle size reduction of lactose
Communition of lactose particles (Pharmatose 100M) was carried out by jet milling
(100 AFG, Hosokawa Alpine AG, Germany) using a classifier wheel rotational speed
of 18000 rpm. Prior to the milling process, lactose powder was dried in the hot air
54
oven at 60 °C for at least 6 h and stored in a desiccator overnight. The lactose powder
was then milled in batches of 300 g load.
3.2.11 Particle size determination of compact fillers
The particle size distributions of the compact fillers were determined by laser
diffractometry (LS 230, Coulter Corporation, USA) using the wet module. The
dispersing media used for suspending the materials during sizing are shown in Table
4. The materials were suspended in the dispersing media immediately before particle
size determination. A total of three determinations were carried out for each material.
The particle size distribution of the materials was expressed as d10, d50 (median size)
and d90, which were the diameters of the material at 10, 50 and 90 percentiles of the
cumulative percent undersize plot respectively.
Table 4. Dispersing media of different materials
Material Dispersing medium*
Lactose Isopropyl alcohol (IPA)
MCC Water
L-HPC IPA
Crospovidone Water
Corn starch IPA
DCP Water
*all materials are very poorly soluble or insoluble in the dispersing media
3.2.12. Particle size and coat thickness determination of pellets
Images of uncoated pellets and coated pellets were captured at 22.5 x magnification
using the stereomicroscope (SZ 61, Olympus, Japan) connected to a colour video
55
camera (DP 71, Olympus, Japan). The captured images were analyzed using an image
analysis software (Image-Pro Plus Version 6.3, Media Cybernetics, USA). The mean
diameter of the pellets obtained from the image analysis was used to calculate the
ethylcellulose coat thickness (Th) according to the following equation:
2d-dTh ucc= Equation 30
where dc is the average mean diameter of ethylcellulose coated pellets and duc is the
average mean diameter of chlorpheniramine-loaded pellets that were not coated with
ethylcellulose. The mean diameter is defined as the average length of diameters
measured at two degree intervals and passing through the object’s centroid. The
average mean diameter of sugar pellets was determined from the mean diameter of
200 pellets.
3.2.13. Method of retrieving pellets from compacts
The pellets on the surface of the compacts were removed and collected by gently
brushing the surface of the compact using a small brush (Microbrush™, Microbrush
Corporation, US). After removing the pellets from the surface, the compact was
further brushed to collect the pellets from the core. All the pellets collected were
mounted on a slide with a double-sided tape. With the help of a stereomicroscope (BX
61, Olympus, Japan) and the Microbrush, the excipient particles adhered to the
surface of the pellets were brushed away. Using the above method, pellets were
collected from three compacts for each formulation for assessment of pellet coat
damage using the stereomicroscope, scanning electron microscope and non-contact
optical profiler.
56
3.2.14. Scanning electron microscopy (SEM)
Samples were pre-treated by gold sputtering (JFC-1100, Jeol, Japan) before
photomicrographs of the samples were taken by SEM (Phenom, FEI company, U.S.,
JSM-5200, Jeol, Japan).
3.2.15. Stereomicroscopy
Images of the pellets collected were taken at 40 x magnification using the
stereomicroscope (BX 61, Olympus, Japan) connected to a colour video camera
(DXC-390P 3CCP, Sony, Japan). The coat damage of pellets was assessed based on
the presence of indentations, cracks and fractures.
3.2.16. Surface profilometry
The vertical scanning interferometry (VSI) mode of the surface profiler (Wyko
NT1100, Veeco, U.S.) was used for surface analysis of the pellets at 42.9 x
magnification with a scan area of 109 μm x 144 μm. The coat integrity of pellets was
assessed based on two parameters, Ra and Rq. Ra represents the average roughness. It
is the arithmetic mean of the absolute values of the surface departures from the mean
plane (Veeco Metrology Group, 1999). Rq represents the root mean square of the
roughness, obtained by squaring each height value in the dataset, then taking the
square root of the mean of the squares of the height values. It represents the standard
deviation of the surface heights (Veeco Metrology Group, 1999).
3.2.17. Determination of pellet core osmolality
A vapour pressure osmometer (Vapro, Wescor, U.S.) was used to determine the
osmolality of pellet cores by measuring the vapour pressure depression. Different
57
amount of pellets were dissolved in purified water (0.1, 0.05 and 0.01 g/mL) and
filtered through membrane filters with 0.45 µm pore size before analysis. The
osmolality for each pellet concentration was measured in triplicates and the results
averaged.
3.2.18. Determination of pellet mechanical strength
The mechanical strength of coated sugar pellets of size fractions 500-600 µm, 355-
425 µm and 250-355 µm, as well as coated MCC pellets of size fraction 500-600 µm
were determined using a tensile tester (EZ Tester-100N, Shimadzu, Japan). Before
analysis, the pellets with different sizes (500-600 µm, 355-425 µm and 250-355 µm)
were sieved (sieves with aperture size 600 µm, 425 µm and 355 µm respectively).
Pellets lodged within the apertures of the sieves were gently brushed off and used for
further analysis. The compression speed was set at 0.5 mm/min. A total of 15
determinations were carried out for each type of pellet. The load stress at break of the
pellet was calculated from the first peak observed on the force-displacement graph.
The inverse of the slope of the force-displacement curve was regarded as a measure of
the pellet deformability (Bashaiwoldu et al., 2004).
3.2.19. Statistical analysis
Where comparison of two sample means was required, independent samples t-test was
performed. However, if more than two sample means were compared, one-way
analysis of variance (ANOVA) was performed with Tukey’s post-hoc testing. The
correlation between two sets of data was assessed using the square of the Pearson
product moment correlation coefficient (R2). The statistical tests were performed
using Minitab Release 14 (Minitab Inc., USA).
58
3.2.20. Multivariate data analysis
When multiple variables and a large amount of data were required to be analysed,
Principal Component Analysis (PCA) served as a useful tool to identify important
variables and inter-variable relationships. The PCA was performed using
Unscrambler® X version 10.0.1 (CAMO software AS, Norway). Compacts containing
different fillers were prepared using a range of pressures and their model independent
and dependent dissolution parameters were compared using PCA. In addition, PCA of
the compact properties and compression pressure was also performed. The loadings
plot generated from PCA of the variables served as a “map” for identifying important
trends and variables, so that further in-depth analysis of the data could be further
performed for better understanding.
59
PART 4: RESULTS AND
DISCUSSION
60
4. RESULTS AND DISCUSSION
PART A. UTILIZATION OF LACTOSE FILLERS IN COMPACTS
CONTAINING COATED PELLETS
A.1. Influence of lactose filler particle size on the extent of coat damage during
compression
For the preparation of compressed multiparticulates, an important component of the
compact was the filler material selected. Lactose is one of the most commonly used
tablet filler due to its stability, safety and relatively lower cost. Various grades of
lactose with different physical properties such as particle size and flowability are
available commercially. Depending on the crystallization and drying conditions,
various isomeric forms of lactose can be manufactured. These include α-lactose
monohydrate, β-lactose anhydrous and α-lactose anhydrous (Edge et al., 2006). Since
crystalline α-lactose monohydrate is available in numerous grades, it was chosen to be
used as filler in this part of the study. Coated sugar pellets were compressed with
several grades of crystalline α-lactose monohydrate with different particle sizes (d50)
at the same compression pressure. Five percent, w/w sodium starch glycolate as
disintegrant and 1 %, w/w magnesium stearate as lubricant were also included.
A low compression pressure was used (28.3 MPa), resulting in compacts that were
quite friable but coherent. This allowed the pellets to be retrieved easily from the
compacts for subsequent examination. The size parameters of the different lactose
grades are shown in Table 5. Figure 5 shows that the compressed pellets exhibited a
faster drug release regardless of the lactose grade. This indicates that the functional
coats of the pellets were damaged in the process of compression.
61
Time (min)
0
10
20
30
40
50
60
70
80
90
100
0 100 200 300 400
% d
rug
Table 5. Size distribution of lactose fillers.
Lactose grade d10 (μm) d50 (μm) d90 (μm)
100M 52.6 169.3 288.4
150M 8.3 56.4 193.6
200M 6.8 35.9 110.2
350M 5.4 31.1 82.1
450M 3.3 22.2 54.0 Figure 5. Dissolution profiles of coated pellets compressed with lactose 100M (□), 150M (▲), 200M (●), 350M (■), 450M (♦) and uncompressed pellets (○).
62
The UC/C value is a ratio of the T50% value of uncompressed coated pellets to
compressed pellets and it is indicative of the extent of coat damage. The UC/C values
were plotted against d10, d50, and d90 of the different lactose fillers in Figure 6.
Generally, UC/C correlated positively with d10, d50 and d90 values of the lactose fillers
(R2 > 0.9). Hence, the particle size of lactose filler can be regarded as a strong factor
influencing the UC/C value. The increasing trend of UC/C with mean particle size
suggests that the extent of coat damage during compression could be reduced when
filler materials of smaller particle size were used. This trend is in agreement with the
findings of Yao et al., 1998. The contributory factors of the observed trend, as well as
the interactions of lactose particles with the coated pellets during compression were
investigated and presented in the following sections.
A.2. Influence of disintegration time on UC/C values
In order to assess if disintegration time could have contributed to the increasing trend
of the UC/C value with increasing lactose particle size, the UC/C values of the
compacts containing different grades of lactose were plotted against their
disintegration time (Figure 7). From the plot, formulations with longer disintegration
time showed lower UC/C values, which tend to level off. To further investigate the
extent of influence of disintegration time on UC/C value, compacts of the coated
pellets containing the same filler (lactose 200 M) but different disintegrants were
prepared and tested. A few types of disintegrants were employed so as to vary the
disintegration time. Assuming that the small amount of disintegrant (5 %) in each
formulation did not affect the extent of coat damage under compression pressure
significantly, any changes in the values of UC/C would be largely due to the
63
(A)
(B)
(C)
Figure 6. Relationship between UC/C and (A) d10, (B) d50 and (C) d90 of lactose filler.
y = 0.15x + 3.11 R2 = 0.96
0
2
4
6
8
10
12
0 10 20 30 40 50 60d10 (μm)
UC
/C
d50 (μm)
y = 0.05x + 2.12 R2 = 0.99
0
2
4
6
8
10
12
0 50 100 150 200
UC
/C
y = 0.03x + 0.78 R2 = 0.91
0
2
4
6
8
10
12
0 100 200 300 400
d90 (μm)
UC
/C
64
Figure 7. Relationship between UC/C and the average disintegration time of compacts containing different grades of lactose filler.
differences in disintegration time. The UC/C values of compacts containing the
different disintegrants were plotted against their average disintegration time (Figure
8).
Figure 8 shows a linear decrease in UC/C with an increase in the average
disintegration time. The correlation between the UC/C values and the average
disintegration time is statistically significant according to the Pearson’s correlation.
However, as the disintegration time varied from 35 to 100 s, UC/C decreased only
from 6 to 4. In comparison with the plot for UC/C values of coated pellets compressed
with different grades of lactose against disintegration time (Figure 7), a smaller
variation of the disintegration time from 12 to 50 s resulted in a much greater
reduction in UC/C (9 to 2). This shows that the decrease in UC/C value with
y = -0.19x + 11.16 R2 = 0.81
y = 76.66x-0.84 R 2= 0.96
0
2
4
6
8
10
12
0 20 40 60 Average disintegration time (s)
UC
/C
65
y = -0.03x + 6.95R2 = 0.93
3.0
3.5
4.0
4.5
5.0
5.5
6.0
6.5
0 20 40 60 80 100 120Average disintegration time (s)
UC
/C
Figure 8. Relationship between UC/C and average disintegration time of compacts containing different disintegrants.
lactose particle size was not solely due to the disintegration time, but mainly due to
the decrease in the extent of pellet coat damage.
A.3. Other methods of assessing coat integrity of pellets compressed with lactose
filler of different particle size
A.3.1. Stereomicroscopy
The images of uncompressed pellets and pellets compressed with different grades of
lactose (100M, 200M, 350M and 450M) were taken using a stereomicroscope
connected to a colour video camera. The images are shown in Figure 9.
For the coated pellets retrieved from the interior of the compacts, the surfaces of the
pellets compressed with lactose 100M and 200M appeared slightly rougher compared
to the pellets compressed with 350M and 450M lactose grades. These findings
66
I 100M I 200M I 350M
I 450M S 100M S 200M
S 350M S 450M Uncompressed
Figure 9. Stereomicroscope images of uncompressed coated pellets and coated pellets compressed with different grades of lactose. “I” denotes pellets retrieved from the interior of the compacts; “S” denotes pellets from the surface of the compacts.
67
indicate that the pellet coats were less well protected by lactose of larger particle size
during compression. The level of damage observed was in agreement with the results
from the dissolution studies. The rougher pellet surfaces caused by contacts with the
coarser lactose particles during the application of compression pressure had resulted
in more severe level of coat damage. The uncompressed pellets and the pellets
compressed with the finer grades of lactose (350M and 450M) did not differ
appreciably in appearance and no coat damage was observed visually. These findings
contrasted with the results of the dissolution studies, which had suggested some coat
damage for pellets compressed with finer grades of lactose powder.
For the pellets retrieved from the surface layers of the compacts, coat damage and
cracks were observed in pellets compressed with lactose 100M and 200M. The pellets
appeared more severely damaged than those retrieved from the interior of the
compact. In addition, these compressed pellets appeared slightly distorted in shape
compared to the uncompressed coated pellets. However, it was generally difficult to
detect visual coat damage for pellets compressed with finer grades of lactose.
The stereomicroscope images of pellets could only show major defects on the coats.
Minor defects such as hairline cracks on the coats of pellets could not be easily
identified from the stereomicroscope images. The detection of coat rupture or damage
might be better if coloured coatings were used. With the addition of colorants, the
coat could be more easily distinguished from the pellet core and excipient. However,
the addition of colorants into the coating formulation could affect the tensile strength
of the coat and thus, possibly the extent of coat damage during compression (Hogan,
1995). This may complicate the interpretation of results for determining the influence
of filler particle size on the extent of damage in sustained release ethylcellulose
coatings.
68
A.3.2. Scanning electron microscopy (SEM)
Images of uncompressed coated pellets and coated pellets compressed with lactose
100M and 450M were taken using a scanning electron microscope. In comparison
with the images taken using white light stereomicroscopy, the SEM images were
clearer and were better defined.
As can be seen from the SEM images, all the compressed pellets appeared “damaged”
compared to the uncompressed pellets (Figures 10 and 11). The surfaces of the pellets
compressed with lactose 100M had more indentations which were larger and deeper
compared to pellets compressed with lactose 450M. The surfaces of pellets
compressed with finer lactose particles (450M) appeared smoother.
Due to the angularity of the lactose particles, pressure was likely to concentrate on the
edges or tips of the lactose particles. When the lactose particles were in direct contact
with the coated pellets, the pressure-concentrated edges could cut into the coat of the
pellets as the die volume decreased during compression. A lactose particle could be
clearly seen anchored on the surface of a pellet as shown by the arrow in Figure
11A(i). As longer lactose particles were able to penetrate deeper into the pellets,
larger crevices can be seen on the surfaces of pellets compressed with the coarser
lactose grade (Figures 10B and 11A). In contrast, the depth of indentation brought
about by finer lactose particles was limited by its diameter. Particles with diameter
less than the coat thickness would not be able to traverse through the entire depth of
the coat. In addition, for smaller lactose particles, the compression force was
dispersed over a greater area of contact among the particles due to greater specific
69
(A)(i) (A)(ii)
(B)(i) (B)(ii)
(C)(i) (C)(ii)
Figure 10. SEM photomicrographs of (A) uncompressed pellets, as well as coated pellets compressed with (B) lactose 100M and (C) lactose 450M (retrieved from compact interior).
70
(A)(i) (A)(ii)
(B)(i) (B)(ii)
Figure 11. SEM photomicrographs of coated pellets compressed with (A) lactose 100M and (B) lactose 450M (retrieved from compact surface).
surface area. This resulted in a reduction in pressure at the points of contact between
the coated pellets and lactose particles. This again reduced the extent and depth of
surface indentations.
Due to the brittle nature of sugar pellets, the indentations on the pellets could also act
as points of weakness for crack propagation with further application of force,
eventually leading to pellet fragmentation. In addition to indentations caused by
lactose particles, other forms of damage to the surfaces of coated pellets include
71
peeled coats (Figure 10B(i)) and cracks on coats (Figures 10B(ii) and 11B(i)). Besides
surface damage, pellet core deformation or fragmentation could also be seen. Figures
10C(i) and 10C(ii) show fractured pellet cores, while Figures 11A(ii) and 11B(ii)
show pellets with flattened surfaces. The flattened surfaces could be brought about by
compression against the surface of punches or neighbouring pellets.
A.3.3. Surface profilometry
From the images taken by stereomicroscopy and SEM, the roughness of the pellet
coat surface was found to vary after compression with lactose powders of different
mean particle sizes. Since quantitative measurements of the coat roughness could not
be achieved by stereomicroscopy, non-contact surface profiler was used in an attempt
to evaluate the roughness and investigate if coat roughness correlated with the level of
coat damage.
The average roughness values, Ra of the pellets retrieved from the surfaces and
interior of the compacts are shown in Table 6. The disadvantage of using Ra as a
roughness parameter is that it is not a direct measure for the presence or absence of
high peaks and deep valleys commonly associated with cracks. Hence, the
corresponding Rq values, which represent the standard deviation of the surface heights
(Veeco Metrology Group, 1999) were determined. If there are no large deviations
from the mean surface, Rq will be close to the Ra in value. On the other hand, Rq will
be larger than Ra if an appreciable number of large cracks and holes are present
(Veeco Metrology Group, 1999). Therefore, the ratio of Rq to Ra is a more useful
indicator for the likelihood of large protuberances or surface roughness.
From Table 6, the pellets obtained from either the compact interior or surface showed
a general decrease, albeit statistically insignificant, in the Ra and Rq values as the
72
Table 6. Mean Ra and Rq values of uncompressed coated pellets and pellets compressed with different grades of lactose.
Lactose grade
Median particle
size (μm)
Location of pellets
in compact
Ra (SD) (nm)
Rq (SD) (nm) Rq /Ra
Nil (uncompressed) N.A. - 584 (384) 729 (446) 1.25
100 M 169.3 Surface 2407 (1634) 2991 (1803) 1.24 200 M 35.9 Surface 2504 (1745) 3152 (2065) 1.26 350 M 31.1 Surface 1661 (1232) 2177 (1587) 1.31 450 M 22.2 Surface 1250 (1073) 1611 (1303) 1.29 100 M 169.3 Interior 3018 (1940) 3943 (2517) 1.31 200 M 35.9 Interior 1629 (1009) 2213 (1422) 1.36 350 M 31.1 Interior 861 (519) 1162 (625) 1.35 450 M 22.2 Interior 1381 (923) 1932 (1613) 1.40
lactose particle size decreased. The results from the Tukey’s post-hoc comparison
showed that Ra and Rq values of coated pellets compressed with lactose 100M and
200M were significantly greater than the uncompressed coated pellets. Hence, this
shows that the coarser lactose grades induced pellet coat damage through surface
indentations. The extent of surface indentations caused by the coarser lactose grades
were shown to be statistically significant. However, the Ra and Rq values of coated
pellets compressed with lactose 350M and 450M were not significantly different from
those of the uncompressed coated pellets. According to the results of the t-test, the Ra
values of the pellets collected from the surface are not significantly different from
pellets collected from the compact interior, except for pellets compressed with lactose
350M. Interestingly, the overall Rq/Ra values of the pellets collected from the compact
interior were generally larger than the pellets collected from the surface. This
indicates that there were protuberances found on the pellets from the compact core but
not on the pellets at the surface of the compact. This could be attributed to pellets
73
impacting against the flat surfaces of the punches, resulting in brittle fracture of
pellets. The brittle fracture of pellets yielded fragments, which filled the voids and
rebonded on the surfaces in contact with the punches. This has also been observed in
heavy magnesium carbonate compacts by Train (1956).
A.4. Bulk density, tapped density and Hausner ratio of lactose fillers
The bulk density, tapped density and Hausner ratio of the various grades of lactose
with different d50 values are shown in Table 7. For the lactose grades used in this
study, the bulk density varied directly with the tapped density but inversely with
Hausner ratio.
The coarser lactose particles generally had greater bulk and tapped densities, as well
as lower Hausner ratio. A high bulk density indicates a close packing structure of
particles even before application of pressure. This could exacerbate the coat-piercing
effects of the coarse lactose particles at the filler-pellet contact points as they were
brought closer to one another. Hausner ratio is defined as the ratio of the tapped
density to the bulk density. It has also been shown to be indicative of the
interparticulate friction of the powders under low pressure (Hausner, 1967). Since the
coarser lactose grades had lower Hausner ratio, the interparticulate friction would be
lower. Hence, the coarser lactose particles were expected to undergo particle
rearrangement over a shorter range of low compression pressures before extensive
particle fragmentation. As a result, extensive particle deformation would likely to
have occurred at a lower compression pressure for the coarser grades of lactose
compared to finer grades. This could have contributed to the greater extent of pellet
coat damage observed in compacts containing coarser grades of lactose.
74
Table 7. Bulk density, tapped density and Hausner ratio of lactose fillers.
Lactose grade
d50 (μm)
Bulk density (g/cm3)
Tapped density (g/cm3)
Hausner ratio
100M 169.3 0.69 0.89 1.29
150M 56.4 0.49 0.91 1.85
200M 35.9 0.41 0.83 2.03
350M 31.1 0.40 0.78 1.97
450M 22.2 0.33 0.72 2.19
On the other hand, due to the lower initial density and Hausner ratio of the finer
lactose grades, coupled to their greater resistance to fragment, (Roberts and Rowe,
1986; York, 1978), particle rearrangement was the predominant mechanism of
densification (Fell and Newton, 1971; Nordstrom et al., 2009). The transition in the
stage of compression from particle rearrangement to particle deformation occurred at
a higher compression pressure compared to coarser grades of the same material
(Huffine and Bonilla, 1962; York, 1978). This could reduce the extent of coat damage
during compression. During compression, instead of piercing into the pellet coat, the
finer lactose particles could reposition and slip into any voids available due to their
smaller dimensions and lower packing density. Furthermore, unlike the coarser
lactose particles, the finer lactose particles could arrange themselves to conform to the
shape of the coated pellets. This allowed the compression force to be distributed over
a greater contact area.
From Figure 12, as the Hausner ratio increased, the UC/C value decreased linearly.
This further showed that interparticulate friction and particle rearrangement before
fragmentation played a role in affecting the extent of pellet coat damage.
75
Figure 12. Relationship between UC/C and Hausner ratio of lactose fillers.
A.5. Utilization of micronized lactose in reducing the extent of coat damage
during compression
Due to the lower level of coat damage observed when finer lactose grades were used,
lactose 100M was micronized (section 3.2.10) by a jet mill and used as a co-filler in
the compact. As can be seen in Table 8, d50 of the micronized lactose was only 2.3
µm. The micronized lactose was also mixed with two other coarse grades of directly
compressible lactose (80M and 125M) at different concentrations to assess any
identical protective effect of micronized lactose in the presence of coarse grade
lactose. All the compacts were made using the same compression pressure.
y = -9.03x + 22.19
R2 = 0.98
0
2
4
6
8
10
12
1.0 1.5 2.0 2.5
Hausner ratio
UC
/C
76
Table 8. The UC/C and disintegration time values of compacts containing lactose blends with different d50 and d90 values.
Filler proportions (%, w/w)
Lactose 80M
Lactose 125M
Micronized lactose
d50 (µm)
d90 (µm) UC/C Disintegration
time (s)
100 0 0 204.2 (2.1)
412.5 (9.3)
10.27 (1.20) 103 (28)
0 100 0 69.2 (1.2)
127.2 (1.6)
7.05 (0.46) 116 (65)
0 0 100 2.3 (0.0)
4.4 (0.1)
3.35 (0.21) 80 (42)
50 50 0 91.8 (5.9)
341.2 (13.4)
6.66 (0.44) 116 (24)
75 25 0 122.1 (1.6)
362.7 (9.9)
7.98 (0.08) 141 (35)
25 75 0 77.7 (2.8)
250.1 (21.8)
6.59 (0.61) 170 (55)
50 0 50 9.9 (1.2)
237.7 (18.2)
5.13 (0.39) 66 (36)
75 0 25 92.9 (10.8)
351.2 (25.3)
6.45 (0.58) 67 (44)
25 0 75 5.7 (0.3)
154.0 (12.4)
3.31 (0.12) 88 (27)
0 50 50 9.5 (2.4)
120.7 (0.8)
3.66 (0.04) 78 (25)
0 75 25 55.3 (0.6)
143.9 (0.7)
3.69 (0.20) 71 (37)
0 25 75 6.1 (0.4)
139.7 (10.9)
3.24 (0.29) 94 (32)
50 25 25 74.9 (4.3)
332.3 (18.9)
5.44 (0.18) 146 (39)
25 50 25 62.8 (1.7)
265.4 (11.6)
4.97 (0.18) 60 (41)
25 25 50 7.7 (0.8)
130.5 (13.8)
4.37 (0.10) 121 (60)
Standard deviation shown in parentheses.
77
A.5.1. Relationship between disintegration time and UC/C
The UC/C values of all the formulations were plotted against the disintegration time
in Figure 13. Unlike the previous finding in section A.2, Figure 7, the UC/C value did
not have any clear correlation with disintegration time. This was probably due to the
minimal effect disintegration time had on UC/C value as explained earlier. The extent
of coat damage superseded the disintegration time in influencing the UC/C value.
Figure 13. Plot of UC/C values against disintegration time. A.5.2. Effect of lactose blend particle size
Figure 14 shows that there was a positive, linear relationship between UC/C and d50
of the lactose blends (R2 = 0.86), although the relationship was not as strong as that
for compacts containing only one type of lactose filler (R2 = 0.99, section 4.A.1,
Figure 6B). Nevertheless, the general correlation between d50 and UC/C shows that
0
2
4
6
8
10
12
50 70 90 110 130 150 170 190
Disintegration time (s)
UC
/C
78
Figure 14. Relationship between UC/C values and d50.
the median particle size of the lactose blends played an important role in affecting the
extent of coat damage. As mentioned in the earlier section (4.A.1), pellet coat damage
was probably extensively induced by coats being pierced by protrusions of the lactose
particles. Larger lactose particles with longer dimensions were capable of creating
larger and deeper indentations on the pellets coats.
From the plot showing the relationship between UC/C and d90 (Figure 15), it is
observed that the UC/C value increased gradually as d90 increased up to
approximately 330 µm. Beyond this d90 value, a much sharper increase in UC/C value
is noted. This finding shows that the pellet coat damage increased drastically when the
lactose filler consisted of only 10 %, w/w of particles larger than 330 µm.
y = 0.03x + 3.48R2 = 0.86
0.0
2.0
4.0
6.0
8.0
10.0
12.0
0 50 100 150 200 250
d50
UC
/C
79
Figure 15. Relationship between UC/C and d90 of lactose blends.
A.5.3. Effect of proportion of micronized lactose in the blend
The compacts containing binary lactose blends showed a non-linear decrease in UC/C
as the proportion of micronized lactose in the blend increased (Figure 16). This shows
that the addition of micronized lactose to coarse lactose reduced pellet coat damage
during compression. It is interesting to note that the reduction in UC/C was not
linearly proportional to the percentage of micronized lactose in the filler blend. In
addition, when the proportion of micronized lactose was 75 %, w/w, the UC/C values
for both lactose 80M and 125M were not statistically different (p < 0.05). Hence,
particle size of the coarser lactose grade in the lactose blend did not play such an
y = 0.06x - 13.43 R2 = 0.95
y = 0.01x + 2.57R2 = 0.59
0
2
4
6
8
10
12
14
0 100 200 300 400 500
d90 (μm)
UC
/C
80
Proportion of micronized lactose (%, w/w)
2
3
4
5
6
7
8
9
10
11
0 20 40 60 80 100
UC
/C
Figure 16. UC/C values of compacts containing blends with lactose 80M (◊) and blends with lactose 125M (■). important role when the proportion of micronized lactose in the filler blend was 75 %,
w/w or higher. However, the flowability of the lactose blend consisting of this
proportion of micronized lactose was as poor as with only micronized lactose (Figure
17). Overall, powder flowability was compromised when the protective effect of
micronized lactose reached its optimal level.
In one of the studies conducted by Riepma et al. (1991), the compression of binary
mixtures consisting of different particle size fractions of lactose was investigated. It
was reported that percolation of the finer lactose fraction in the matrix of coarse
lactose particles could decrease the fragmentation propensity of the coarser lactose
particles. When two materials with a large difference in particle size were mixed to
form compacts, the material with lower particle size would adhere and coat the larger
particles, forming a percolating network. As a result, the compressibility of the binary
81
Figure 17. Relationship between Hausner ratio and the proportion of micronized lactose in lactose blend containing lactose 80M (◊) and lactose 125M (■).
mixture was closer to the compressibility of the adhering material (Barra et al., 1999,
Busignies et la., 2006). Hence, in the present study, the micronized lactose adhered
onto the coarser lactose particles, transforming the compression behaviour of the
binary mixture to that of micronized lactose when it formed a percolating network.
As a result, the pellet-filler interactions were similar to that of micronized lactose
particles, concealing the coat damaging effects of the coarse lactose particles in the
binary mixture. The point at which the percolation network occurs can be derived by
observing the change in Hausner ratio and compact thickness with the proportion of
micronized lactose (Figures 17 and 18). This is because Hausner ratio and compact
thickness of compacts formed at the same pressure are indicative of the
interparticulate friction and compressibility of the powder mixture.
Hau
sner
ratio
1.2
1.4
1.6
1.8
2.0
2.2
2.4
0.0 20.0 40.0 60.0 80.0 100.0
Proportion of micronized lactose (%, w/w)
82
As the proportion of micronized lactose increased from 0 to 50 %, w/w, the Hausner
ratio increased from approximately 1.5 to 2.1. However, the Hausner ratio of the
lactose blends was similar to that of micronized lactose when the proportion of
micronized lactose was more than 50 %, w/w. This suggests that the micronized
lactose formed a percolating network at a proportion of 50 %, w/w. The higher
Hausner ratio with the addition of micronized lactose also meant higher
interparticulate friction in the blend. By increasing the interparticulate friction of the
lactose blend, particle rearrangement was likely to occur over a greater compression
pressure range, reducing strong pellet-filler interactions and pellet coat damage.
Coupled to the greater resistance of smaller lactose particles to fragment (Roberts and
Rowe, 1986; York, 1978), the higher interparticulate friction resulted in thicker
compacts with the addition of micronized lactose (Figure 18). The compact thickness
increased only slightly when the proportion of micronized lactose increased from 0 to
50 %, w/w. In comparison, the increase in compact thickness was greater when the
proportion of micronized lactose increased from 50 to 100 %, w/w (Figure 18). This
trend again suggests that the micronized lactose formed a percolating network when
the proportion of micronized lactose was more than 50 %, w/w. Above 50 %, w/w of
micronized lactose, the compact thickness was dependent on the compressibility of
micronized lactose and the amount of micronized lactose. Small particles were
reported to be more cohesive, which restricted densification (Roberts and Rowe,
1986). Therefore, there was a greater increase in compact thickness as the proportion
of micronized lactose increased from 50 to 100 %, w/w. An increase in the compact
thickness would result in a corresponding decrease in pellet volume fraction which
could reduce the extent of pellet coat damage. This is further discussed in the next
section.
83
4.75
4.85
4.95
5.05
5.15
5.25
5.35
0 20 40 60 80 100
Proportion of micronized lactose (%, w/w)
Com
pact
thic
knes
s (m
m)
Figure 18. Relationship between compact thickness and proportion of micronized lactose in the lactose blend. The percolation condition of micronized lactose particles in the blend also meant that
a greater proportion of the pellets were in contact with the micronized lactose
particles rather than the coarser lactose particles. This was reflected by the decreasing
surface roughness (Ra values) of compacts containing increasing proportions of
micronized lactose with lactose 80M (Figure 19). In this experiment, compacts of the
binary lactose blends were made using a lower compression pressure in order to
preserve the surface integrity of the filler particles. The roughness values of the
compacts were thus representative of the roughness of the particle surfaces in contact
with the coated pellets. From Figure 19, it can be seen that as the roughness
decreased, the UC/C value also decreased, indicating a reduction in the extent of coat
damage. Compacts containing coarse lactose 80M only as filler had the roughest
surface whereas compacts containing only micronized lactose were the smoothest. For
84
Figure 19. Plot of UC/C against average roughness, Ra, of compacts containing different proportions of micronized lactose and lactose 80M formed using a compression pressure of 11.3 MPa. Proportion of micronized latose shown in parentheses.
the compacts containing various proportions of micronized lactose, the greater the
proportion of micronized lactose, the smoother would be the compact surface.
Micronized lactose filled the asperities of the coarser lactose particles, creating a
smoother surface. Thus, when these particles were in contact with the coated pellets,
less coat damage was encountered during compression. Nevertheless, it should also be
borne in mind that coarse filler particles may also fragment into smaller particles and
present sharp edges, which could contribute to pellet coat damage. In addition, the
micronized lactose also acted as a lubricating layer, allowing ‘ball bearing assisted
slippage’ or spatial adjustments to the pellets during compression, and thus,
mitigating compression related damage.
UC
/C
0
1
2
3
4
5
6
7
8
0 2000 4000 6000 8000 10000
Ra
(0 %)
(25 %)
(50 %)
(75 %)
(100 %)
85
To conclude, these observations show that a percolating network of micronized
lactose was formed when the proportion of micronized lactose was more than 50 %,
w/w. The formation of the percolation network resulted in pellet-filler interactions
that were similar to that of micronized lactose. This brought about the non-linear
reduction in UC/C as the proportion of micronized lactose increased (Figure 16).
A.5.4. Critical pellet volume fraction
The pellets and powder filler can be regarded as a two compartment model (Kühl et
al., 2002). The pellets formed the non-deforming compartment with a nearly constant
volume during compression while the powder filler was the deforming compartment
of the matrix. This model assumed that the powder filler was much more deformable
than the pellets and would have deformed sacrificially in place of the pellets under
pressure. For the pellets to remain as a non-deforming compartment, the die volume
must at least be sufficient to accommodate the pellets in a cubic arrangement with the
deforming compartment filling up the inter-pellet spaces. The theoretical cubic
packing of spheres corresponds to a volume fraction of 0.3116 (Stauffer and Aharony,
1992). From Figure 20, it can be seen that the UC/C value increased steeply when the
packing volume fraction of pellets in the compact increased from 0.39, which
approximates to that for theoretical cubic packing of spheres. This indicates that there
was a critical pellet volume fraction at which the pellet coat damage increased steeply
as the pellet volume fraction increased. Similarly, Kühl et al. (2002) reported that at a
pellet volume fraction of 0.42, the work of compression increased steeply, indicating
the beginning of stronger mechanical interactions between the pellets. In the present
study, when the pellet volume fraction was below 0.39, the UC/C value remained
relatively constant as the pellet volume fraction decreased (Figure 20). However, the
86
Figure 20. Relationship between UC/C and pellet volume fraction.
UC/C values were more than one even when the pellet volume fraction was less than
0.39. This could be due to some extent of coat damage on the surfaces of pellet cores
brought about by compressive interaction and contact with the filler particles, as well
as other coated pellets during compression
A.6. Conclusion
As the particle size of lactose filler particles increased, the extent of pellet coat
damage also increased. This was shown by SEM photomicrographs to be largely due
to the deeper and larger indentations created by the pressure concentrated edges of the
larger lactose particles in comparison with the finer lactose particles. Besides surface
damage, other forms of damage to the coated pellets include compression-induced
pellet core deformation and fragmentation. The finer grades of lactose fillers
0
2
4
6
8
10
12
14
0.365 0.37 0.375 0.38 0.385 0.39 0.395 0.4 0.405 0.41
Pellet volume fraction
UC
/C
0.393
87
decreased the extent of pellet coat damage through several mechanisms. Compression
force could be dissipated over a larger specific surface area of smaller particles,
resulting in a reduction in the amount of transmitted pressure at the inter-particle
contact points. The higher Hausner ratio of finer grades of lactose indicates greater
interparticulate friction and hence particle rearrangement over a larger range of
compression pressure. This allowed the pellets to undergo spatial adjustments and less
force induced interactions over a larger range of pressure during compression.
Furthermore, due to the greater resistance for smaller lactose particles to fracture,
their lower bulk density and greater interpaticulate friction, they formed compacts
with lower pellet volume fraction, resulting in lower extent of pellet coat damage.
It was shown that the addition of micronized lactose to coarser lactose fillers reduced
the extent of pellet coat damage. However, the overall material flow was greatly
compromised as the proportion of micronized lactose was increased. Therefore,
micronized lactose can be employed to mitigate the extent of coat damage caused by
coarser lactose particles with careful consideration of its effect on the overall flow
property. The pellet coat damage decreased in a non-linear fashion as the proportion
of micronized lactose was increased. The non-linear relationship was attributed to the
percolating network of micronized lactose formed when the proportion of micronized
lactose was more than 50 %, w/w. This resulted in similar pellet-filler interactions as
that of the micronized lactose, concealing the damaging effects of the coarse lactose
particles in the binary mixture fillers. The formation of the percolation network of
micronized lactose also enabled the pellets’ contacts with micronized lactose to be
considerably greater than the contacts with the coarse lactose particles. The outcome
was reduced number of large and deep indentations induced by the coarse lactose
particles. The extent of coat damage increased steeply when the pellet volume fraction
88
was more than a critical value of 0.39, demonstrating the importance of pellet volume
fraction in affecting the pellet coat damage.
89
PART B. INFLUENCE OF FILLER TYPE AND PELLET VOLUME
FRACTION IN COMPACTS CONTAINING COATED PELLETS
In Part 4A, it was shown that lactose particle size played an important role in
determining the extent of coat damage. The pellet-filler interactions during
compression with lactose fillers were also investigated. However, there are many
other excipients used in pharmaceutical dosage forms which are of different physical
characteristics and could behave differently from lactose during compression. Hence,
the purpose of this part of the study was to examine how the physical characteristics
and compression behaviour of various filler types influence the extent of pellet coat
damage. The study was also aimed at investigating the influence of pellet volume
fraction and compression rates on the extent of coat damage when different filler
types were used.
B.1. Physical characteristics of fillers and their influence on the extent of coat
damage during compression
Several particle size fractions of fillers were compressed with uncured (dried at 35
°C) and cured coated pellets (dried at 60 °C). The fillers used include lactose, MCC,
L-HPC, crospovidone and DCP.
As mentioned in Part 4A, the UC/C value increased linearly with lactose particle size
(Figures 21). This trend was attributed to the formation of deeper and larger
indentations created by the larger lactose particles when they were in contact with the
surfaces of the coated pellets during compression. The closely packed powder
structure of coarser lactose fillers exacerbated their damaging effects. In addition, due
to the larger specific surface area of fine particles, the amount of pressure exerted
onto the pellet surface by each particle was reduced when finer grades of lactose
90
Figure 21. Relationship between the UC/C values of uncured (closed symbols) and cured (opened symbols) pellets against d50 of lactose ( ), MCC (■), L-HPC (▲), crospovidone ( ), DCP (×) and DCL 11 (+).
+
d50 (μm)
×
y = 0.02x + 1.39R2 = 0.98
y = 0.05x + 2.12R2 = 0.99
y = 0.10x + 2.55 R2 = 1.00
2
y = 0.02x + 1.10 R2 = 1.00
1
2
3
4
5
6
7
8
9
10
11
12
0 50 100 150 200
UC
/C
91
fillers were used. There was greater allowance for spatial adjustments to the pellets
during compression, thereby reducing the force induced interactions when the pellets
were compressed with finer lactose grades.
Interestingly, particle size did not have the same influence on the extent of coat
damage when MCC was used as the filler. For MCC, the UC/C value increased non-
linearly as the d50 increased from 27.6 to 109.8 µm. However, when the d50 of the
MCC filler was 168.4 µm, the UC/C value decreased (Figure 21). Hence, in
comparison with lactose, the extent of coat damage was not the greatest when the
coated pellets were compressed with the coarsest grade of MCC. These findings
indicate that in addition to particle size, some other material properties had also
contributed to the protective effects imparted by the fillers.
MCC is manufactured by acid hydrolysis of α-cellulose of fibrous plant pulp. After
hydrolysis, mechanical shearing of the slurry is carried out, followed by
neutralization, washing, filtering and spray-drying to produce the MCC particles.
MCC particles consist of cellulose microcrystals (with diameters ranging from 1 to 10
µm) bound together randomly as porous aggregates. The porous structure of the MCC
particles plays an important role by conferring superior compression behaviour to
MCC. It had been reported that the total pore volume, mean and median pore size
values of compacts decreased with increasing compression pressure (Westermarck,
1999). The decrease in the values was attributed to the collapse of the internal pores
of MCC particles and the reduction in interparticle voids during compression. The
plastic deformation of MCC particles was brought about by the presence of slip
planes and dislocations on a microscale, and the deformation of the spray-dried
agglomerates on a macroscale (Shangraw, 1990, Bolhuis and Chowhan, 1996).
92
During compression, the internal pores of the microcrystal aggregates would collapse
and result in the plastic deformation of the MCC particles, which then conformed to
the shapes of the surrounding coated pellets. As a result, unlike the coarse lactose
particles, the edges of the coarse MCC particles were less likely to create large and
deep indentations. This effect was expected to be more pronounced in coarser grades
of MCC since the modal pore size MCC powders was reported to increase as the
particle size increased (Doelker, 1993). It was also observed by Tunon et al. (2003b)
that the deformation and surface indentations of coated pellets could be reduced by
using filler pellets that were more porous. Hence, this explains the reduction in UC/C
value for the coarsest grade of MCC (d50 = 168.4 µm), as well as the non-linear
relationship between UC/C and MCC particle size (Figure 21).
Similarly for crospovidone, the extent of coat damage did not increase linearly with
particle size. The extent of coat damage was greater when compressed with Kollidon
CL-SF (d50 = 42.2 µm) compared to Kollidon CL-M (d50 = 5.2 µm). However, the
coat damage decreased as the d50 of crospovidone filler increased from 42.2 to 94.8
µm. This could again be explained by the increase in total porosity and median pore
diameter as the particle size of crospovidone increased (Shah and Augsburger, 2001).
From the SEM photomicrographs (Figures 22 and 23), MCC and crospovidone
particles appeared as composites of smaller primary particles. Since the agglomerates
were larger and more porous, they were likely to be more deformable and hence
conferred protective properties when compressed with coated pellets.
Surprisingly, despite the deformability of MCC and crospovidone particles, the UC/C
values were higher when the coated pellets were compressed with finer grades of
MCC and crospovidone fillers (d50 < 80 µm) compared to lactose fillers of the same
93
(A) Lactose 100M (B) Lactose 450M
(C) Micronized lactose (D) Avicel PH 200
(E) Avicel PH102 (F) Avicel PH101
Figure 22. SEM photomicrographs of (A-C) lactose and (D-F) MCC filler particles (all at same magnification except micronized lactose).
94
(A) LH-11 (B) LH-31 (C) Kollidon CL (D) Kollidon CL-M
(E) DCP
Figure 23. SEM photomicrographs of (A-B) L-HPC, (C-D) crospovidone (Kollidon) and (E) DCP filler particles (all at the same magnification).
95
size range. This was probably due to the comparatively lower Hausner ratio of MCC
and crospovidone (Figure 24). Generally, as the Hausner ratio of the fillers increased,
the UC/C value decreased (Figure 25). Hausner ratio is indicative of the
interparticulate friction at low pressures. Materials with higher Hausner ratio had
higher interparticulate friction, and particle rearrangement was likely to occur over a
wider range of compression pressure. As a result, compression pressure could be
dissipated partly through particle rearrangement even at higher compression
pressures. In addition, this allowed the pellets to undergo spatial adjustments at higher
compression pressures, reducing the forceful pellet-filler interactions. Since the MCC
and crospovidone fillers used in this study had comparatively lower Hausner ratios,
extensive particle deformation was likely to have occurred at a lower compression
pressure, resulting in higher UC/C values compared to lactose fillers. In addition, the
compressibility of MCC and crospovidone fillers also played a role in the greater
pellet coat damage than lactose fillers in this size range. This will be further discussed
in the next section.
The effect of L-HPC particle size on the extent of coat damage was similar to
crystalline α-lactose monohydrate fillers in the d50 range of 20 to 60 μm. There was a
linear increase in coat damage with particle size (R2 = 1.00). Unlike MCC and
crospovidone, the mean pore size and the micropore volume of all the different grades
of L-HPC were comparable except for the finest grade of L-HPC, LH-31, which had a
considerably higher micropore volume or intraparticle porosity (Alvarez-Loreanzo et
al., 2000). A reduction in the intraparticle porosity was observed when LH-31 was
compressed (Alvarez-Loreanzo et al., 2000). In addition, LH-31 compacts had a
smaller macropore volume than the other L-HPC compacts. These observations
indicate that the LH-31 particles deformed and collapsed internally during
96
C) y = -0.0033x + 1.84R2 = 0.24
A) y = -0.0055x + 2.22 R2 = 0.96
B) y = -0.0057x + 2.04 R2 = 0.81
D) y = -0.0071x + 2.08 R2 = 0.96
1
1.2
1.4
1.6
1.8
2
2.2
2.4
0 50 100 150 200
d50 (μm)
Hau
sner
ratio
A
B
C
D
Figure 24. Hausner ratio of different grades of (A) lactose ( ), (B) L-HPC (▲), (C) MCC ( ) and (D) crospovidone ( ) against d50.
Figure 25. Relationship between UC/C and Hausner ratio of lactose ( ), L-HPC (▲), MCC ( ) and crospovidone ( ).
y = -4.80x + 14.11R2 = 0.40
0
2
4
6
8
10
12
1.2 1.4 1.6 1.8 2 2.2 2.4
Hausner ratio
UC
/C
97
compression. Coupled with its small particle size compared to other grades, the LH-
31 particles were able to arrange and deform to conform to the shape of the
surrounding coated pellets during compression. This reduced the pressure at the
contact points between the filler particles and the coated pellets, thus reducing surface
damage (Tunon 2003b). Hence, it was not unexpected to observe the lowest extent of
coat damage in pellets compressed with LH-31 compared to other grades of L-HPC.
Unlike LH-31, the coarser grades of L-HPC appeared as rod-shaped extrudates
(Figure 23A and B). Due to their lower intraparticle porosity, particle shape
deformation and internal collapse might not be as extensive (Alvarez-Loreanzo et al.,
2000). Therefore, it was more likely for the pellet surfaces to be indented and
damaged by the surrounding coarse and rod-shaped filler particles during
compression.
In order to investigate if the non-linear relationship between the particle size and
UC/C also appeared in porous particles which deformed by fragmentation, coated
pellets were compressed with DCP (Emcompress) and DCL 11. These two materials
are known to be porous and deformed predominantly by fragmentation (Landin et al.,
1994, Cal et al., 1996). The UC/C value for compacts containing either DCP or DCL
11 was indeed relatively lower than the other fillers despite their comparatively larger
particle size (Figure 21). As can be seen from the SEM photomicrograph of DCP
particles, they were aggregates of small primary particles (Figure 23E). DCL 11 is
manufactured by spray drying lactose suspension, forming spherical, porous particles
made up of microcrystals ‘glued’ together with amorphous lactose (Bolhuis and
Chowhan, 1996). Hence, due to the presence of pores in their collapsible structures,
both DCP and DCL 11 could undergo structural collapse upon application of pressure.
98
An increase in the number of small pores after compression of DCP had also been
observed through mercury intrusion porosimetry in a study, indicating particle
structural collapse (Vromans et al. 1985). Since both DCP and DCL 11 consolidated
by fragmentation (Landin et al., 1994), it could be concluded that internal collapse of
structure at relatively low compression pressure had reduced pellet coat damage,
regardless of the consolidation mechanism.
In conclusion, particle size was an important factor affecting the extent of pellet coat
damage during compression for both fragmenting fillers (e.g. lactose) and plastically
deforming fillers (e.g. L-HPC). It was found that large particles created bigger and
deeper indentations on pellet surfaces during compression, resulting in a greater
extent of coat damage than similar but smaller particles. However, if pores were
present as in large agglomerates, the structure of the particle could collapse internally
during compression. This would dissipate the compression pressure, hence reducing
the impacting surface indentations created at the pellet-filler contact points, resulting
in reduced extent of coat damage. The latter effect was exhibited by both fragmenting
(e.g. DCP and DCL11) and plastically deforming materials (e.g. MCC and
crospovidone). Hence, the coat damaging effects of coarse filler particles could be
mitigated through particle deformation and structural collapse. On the other hand,
when the filler particles were smaller in size and less likely to create deep indentations
on pellet coat, fillers with higher Hausner ratio were found to induce a lower extent of
coat damage. This was because fillers with greater Hausner ratio were likely to
undergo particle rearrangement over a higher range of compression pressure, thus
dissipating the pressure exerted and reducing the extent of particle deformation at
higher pressures.
99
B.2. Relationship between the material compressibility and extent of pellet coat
damage
Heckel and Walker equations were used to derive the compression parameters that
characterize material compressibility. These parameters were used to correlate with
the UC/C values in order to examine for any relationship between material
compressibility and coat damage. The yield pressure (Py) of the fillers was determined
from the linear region (34 to 68 MPa) of the Heckel plot (R2 > 0.97). In addition, the
compressibility coefficient, W was determined from the linear region (22.6 to 68
MPa) of the Walker plot (R2 > 0.98). The Py and W values of the different materials
are shown in Table 9.
Table 9. Py and W values of the different materials
Material Material grade
Py (MPa) W d50
(µm)
80M 208.3 21.5 204.2
100M 357.1 26.2 169.3
200M 256.4 18.3 35.9 Lactose
Micronized 204.1 44.5 2.3
PH 200 76.3 97.5 168.4
PH 102 70.9 84.3 109.8
PH 101 74.1 85.9 72.5 MCC
PH 105 77.5 80.7 27.6
LH-11 89.3 79.0 126.7
LH-20 73.5 64.1 55.6 L-HPC
LH-31 88.5 56.2 19.6
CL 107.5 75.9 94.8
CL-SF 104.2 82.7 42.2 Crospovidone (Kollidon)
CL-M 107.5 87.1 5.2
100
The Py value is a measure of the plasticity of a material. A smaller Py value indicates
greater plasticity (Paronen and Ilkka, 1996). On the other hand, the W value is a
measure of the irreversible compressibility of a system of particles (Sonnergaard,
1999). A larger W value indicates greater compressibility. The variation in particle
size within each type of material did not have a strong influence on both constants
(Table 9). The Py value for the different materials decreased in the following order:
lactose > crospovidone (Kollidon) > L-HPC ≥ MCC. By comparison, the W value for
the different materials increased in the following order: lactose < L-HPC <
crospovidone ≤ MCC. It is not unexpected that lactose of different grades had much
greater Py values than the other types of material since it is known to consolidate
mainly by fragmentation.
The relationship between UC/C and Py or W is shown in Figure 26. There was a
general increase in UC/C as the W value increased within each material type. This
observation suggests that the coated multiparticulate system was sensitive to volume
changes. When the coated pellets were compressed with a material grade which was
more compressible compared to other grades, the system underwent a greater volume
change, resulting in a higher level of coat damage with volume reduction. In the
previous section, it was found that when the coated pellets were compressed with
finer grades of lactose fillers (< 80 μm), the extent of coat damage was lower
compared to other fillers of comparable particle size (Figure 21). In addition to the
longer phase of particle rearrangement, the lower compressibility of lactose fillers
could also have contributed to the lower extent of coat damage.
From the plot of UC/C against Py, the relationship followed a U-shaped trend, with
minima at about 150 MPa. Materials with higher plasticity could impart cushioning
101
(A)
(B)
Figure 26. Plot of UC/C against (A) W coefficient and (B) yield pressure of lactose (♦), MCC (■), L-HPC (▲) and crospovidone (◊).
0
2
4
6
8
10
12
14
16
0 20 40 60 80 100W
UC
/C
Yield pressure (MPa)
0
2
4
6
8
10
12
14
16
50 100 150 200 250 300 350 400
UC
/C
102
properties by deforming and ‘absorbing’ the compression energy (Torrado and
Augsburger, 1994, Yuasa et al., 2001). Hence, as the yield pressure decreased from
350 to 200 MPa, there was a general decrease in UC/C values. However, this trend
was absent when the yield pressure was below 150 MPa. This could be attributed to a
greater extent of volume reduction when the pellets were compressed with materials
of greater plasticity or compressibility. Although plastic deformation of materials
could cushion the coated pellets from the compression force, at the same time, the
amount of space available to accommodate the pellets and materials was
compromised. The U-shaped relationship demonstrates that the amount of protective
effect was an interplay between coat damage caused by material volume reduction
and force dissipation through filler particle deformation.
B.3. Comparison of the extent of coat damage in uncured and cured coated
pellets
Cured coated pellets were produced by drying the coated pellets at 60°C while the
uncured coated pellets were produced by drying the pellets at 35°C (Part 3.2, Table
3). The glass transition temperature of Surelease is about 35°C (O’Donnell and
McGinity, 1996). Below this temperature, ethylcellulose exists in a glassy state
characterized by minimal chain movement. Above this temperature, ethylcellulose
exists in a rubbery state characterized by increased polymer chain movement and
elasticity (O’Donnell and McGinity, 1996). Polymer movement is required for
coalescence and the formation of continuous film. Hence, drying the coated pellets at
35 and 60°C produced uncured and cured coated pellets respectively.
Similar trends were observed in both cured and uncured coated pellets (Figure 21).
However, in general, the UC/C values for the uncured pellets were higher (2.8 to
103
10.7) than those for cured pellets (1.1 to 4.9) for approximately the same size range of
fillers investigated (2 to 170 µm). In addition, the uncured pellets were more sensitive
to the particle size of fillers. This was due to incomplete coalescence of the
ethylcellulose particles for pellet coats which were not completely cured. As a result,
a discrete, continuous, strong film was not formed. Points of weakness existed at
regions where the ethylcellulose particles were not completely fused to one another.
Consequently, the coat was relatively weaker and more easily damaged to a greater
extent during compression. The ratio of tensile strength to elastic modulus (low values
correlated to more coating defects) of Surelease coat has been shown to increase as
the coalescence temperature increased from 30 to 50°C (O’Donnell and McGinity,
1996). In addition, Bodmeier and Paeratakul (1994) reported that films made from
Aquacoat (ethylcellulose pseudolatex) plasticized with triethyl citrate had slightly
higher puncture strength and % elongation when dried at 60°C instead of 40°C.
B.4. Influence of disintegration time on UC/C values
In compacts containing coated multiparticulates, the relationship between UC/C and
disintegration time was found to be rather complex. If the coated pellets were intact
and undamaged after compression and the disintegration time was reasonably short,
the rate of drug diffusion through the functional coat would be the rate-limiting step in
drug release. The UC/C value would thus be minimally affected by the disintegration
time. However, if the coat barrier around the pellet cores was greatly damaged during
compression, disintegration time of the tablet matrix might become the rate-limiting
step of the drug release process instead. As a result, the compact disintegration time
might influence the UC/C value to a larger extent than the pellet coat. In this present
study, coated pellets were compressed with a variety of fillers that produced different
104
level of coat damage. The compact disintegration time for the different formulations
was determined to ascertain if compact disintegration time was a major determinant in
regulating drug release, especially during the early phases. From Figure 27, the
relationship between UC/C and disintegration time was unclear except for compacts
containing lactose fillers. It was already shown in Part 4A that although there was a
relationship between UC/C and disintegration time for compacts containing lactose,
the dominant influencing factor of UC/C value was the extent of coat damage.
Furthermore, with the exception of one formula (compacts containing LH-31), all
compacts disintegrated in less than one minute. Thus, the extent of coat damage was
more likely to be the dominating factor influencing the UC/C value in this study than
compact disintegration time.
Figure 27. Plot of the UC/C values of coated pellets compressed with different grades of MCC (▲), lactose ( ) and L-HPC (■) against the average disintegration time.
0
2
4
6
8
10
12
14
0 50 100 150 200
Average disintegration time (s)
UC
/C
105
B.5. Influence of compression pressure and pellet volume fraction
It was earlier shown that the interaction between the coated pellets and lactose filler
was different from MCC fillers during compression. The size of filler particles was a
factor affecting the extent of coat damage in compressed pellets. Hence, coarse and
fine grades of both lactose and MCC fillers were chosen to study the influence of
compression pressure and pellet volume fraction. The fillers employed include lactose
80M, micronized lactose, PH200 and PH105. According to the Heckel plots for the
fillers, the yield pressures of the fillers were 208.3, 204.1, 76.3 and 77.5 MPa,
respectively. The compression pressure studied ranged from 1.13 to 424 MPa.
B.5.1. Influence of compression pressure and disintegration time on different
dissolution parameters
The UC/C value, which was the ratio of T50% values of compressed and uncompressed
pellets, was used as an index to indicate the extent of pellet coat damage from
dissolution studies. The dissolution profiles can also be described by other model
independent and dependent parameters. Different dissolution parameters had been
used by different investigators to assess the extent of pellet coat damage for
compressed multiparticulates. However, there was a lack of information on the
suitability of the different dissolution parameters for use to assess pellet coat damage.
Thus, several dissolution parameters were derived from the dissolution profiles in this
study for a comparative evaluation. The model independent dissolution parameters
derived include T25%, T50%, T75%, MDT, VDT, RD and F2. The dissolution profiles
were also fitted to zero order, Higuchi and power law models. The goodness-of-fit
values of the dissolution profiles to the different models were compared using R2
(Tables 10 and 11). Regardless of the compression pressure and filler type, the
106
Table 10. Curve-fitting (R2) of dissolution profiles for compacts containing lactose fillers formed at different pressures to different models.
R2 value Compression pressure (MPa)
Filler type Zero order Higuchi Power law
0 - 0.906 0.969 0.984 1.13 Lactose 80M 0.931 0.979 0.991 2.83 Lactose 80M 0.945 0.987 0.994 5.66 Lactose 80M 0.932 0.983 0.992 11.3 Lactose 80M 0.959 0.993 0.996 22.6 Lactose 80M 0.940 0.984 0.995 34.0 Lactose 80M 0.952 0.987 0.992 45.3 Lactose 80M 0.924 0.982 0.988 67.9 Lactose 80M 0.907 0.964 0.976 90.5 Lactose 80M 0.941 0.982 0.986 113 Lactose 80M 0.980 0.996 0.997 147 Lactose 80M 0.973 0.993 0.995 181 Lactose 80M 0.980 0.996 0.997 226 Lactose 80M 0.972 0.994 0.997 283 Lactose 80M 0.962 0.990 0.997 340 Lactose 80M 0.965 0.990 0.994 424 Lactose 80M 0.962 0.989 0.996 1.13 Mic Lac 0.911 0.972 0.986 2.83 Mic Lac 0.917 0.975 0.987 5.66 Mic Lac 0.919 0.975 0.986 11.3 Mic Lac 0.933 0.984 0.992 22.6 Mic Lac 0.933 0.986 0.990 34.0 Mic Lac 0.967 0.998 0.999 45.3 Mic Lac 0.951 0.992 0.997 67.9 Mic Lac 0.934 0.987 0.996 90.5 Mic Lac 0.941 0.987 0.997 113 Mic Lac 0.934 0.984 0.996 147 Mic Lac 0.892 0.962 0.982 181 Mic Lac 0.940 0.984 0.989 226 Mic Lac 0.909 0.967 0.992 283 Mic Lac 0.927 0.979 0.993 340 Mic Lac 0.914 0.972 0.992 424 Mic Lac 0.925 0.976 0.993
107
Table 11. Curve-fitting (R2) of dissolution profiles for compacts containing MCC fillers formed at different pressures to different models.
R2 value Compression pressure (MPa)
Filler type Zero order Higuchi Power law
1.13 PH200 0.918 0.979 0.991 2.83 PH200 0.921 0.973 0.989 5.66 PH200 0.933 0.980 0.992 11.3 PH200 0.923 0.979 0.991 22.6 PH200 0.913 0.977 0.994 34.0 PH200 0.920 0.980 0.995 45.3 PH200 0.962 0.993 0.999 67.9 PH200 0.954 0.989 0.989 90.5 PH200 0.911 0.975 0.975 113 PH200 0.869 0.951 0.984 147 PH200 0.982 0.997 0.997 181 PH200 0.979 0.997 0.998 226 PH200 0.961 0.989 0.997 283 PH200 0.980 0.997 0.998 340 PH200 0.970 0.994 0.998 424 PH200 0.966 0.992 0.997 1.13 PH105 0.915 0.974 0.987 2.83 PH105 0.917 0.971 0.989 5.66 PH105 0.942 0.987 0.991 11.3 PH105 0.926 0.981 0.991 22.6 PH105 0.921 0.983 0.995 34.0 PH105 0.933 0.984 0.994 45.3 PH105 0.905 0.972 0.994 67.9 PH105 0.919 0.970 0.997 90.5 PH105 0.899 0.967 0.982 113 PH105 0.929 0.981 0.997 147 PH105 0.927 0.978 0.990 181 PH105 0.883 0.954 0.990 226 PH105 0.882 0.952 0.985 283 PH105 0.911 0.970 0.990 340 PH105 0.906 0.967 0.988 424 PH105 0.901 0.964 0.986
108
goodness-of-fit for zero order model was always the worst. The goodness-of-fit for
Higuchi model was relatively better, indicating that drug release followed the Fick’s
law and was concentration gradient dependent. The power law model fitted the best to
all dissolution profiles as a result of its variable exponent, n. Due to the poor fit for
zero order model, the rate constant derived from this model was not considered for
subsequent analysis. As more than three different dissolution parameters were derived
from each dissolution profile, the principal component analysis (PCA) was performed
(The Unscrambler® X 10.0.1, CAMO software AS, Norway) to determine the inter-
variable relationships simultaneously. The PCA is a technique used in multivariate
data analysis to identify the most prominent variables and uncover the trends present
in a set of data. It involves the plotting of n objects (the different formulations formed
at different compression pressures in this case) in a p-dimensional variable space (the
variables describing the object e.g. compression pressure, Tx%, MDT etc.). The line
that has the best simultaneous fit to all the objects in the variable space is identified
through the use of least-squares optimization principle. This central axis is known as
the First Principle Component (PC1) and it explains the main variance in the data set.
Thereafter, a second principle component (PC2) lying orthogonal to PC1 in the
direction of the next largest variance is plotted. Subsequently, more orthogonal PCs,
each lying along a maximum variance direction in decreasing order can be plotted
until the system describes the objects considerably. A common origin for the PCs
acting as the “center-of-gravity” of the data points is identified by the software. This
origin is the average point of all the data points. The different variables describing the
objects are plotted on the loading plot according to how much each variable
contributes to each PC. The inter-variable relationships can thus be derived from the
109
loadings plot. More details on PCA can be found elsewhere in reported literature
(Esbensen et al., 2001).
The loadings plot from PCA of compression pressure and dissolution parameters is
shown in Figure 28. PC1 explained 67 % of the total data variation while PC2
explained only 13 % of the total data variation. Together, the two PCs explained 80 %
of the variation in data. Variables which lie near to the origin are considered to be of
little importance. The variables contributed differently to the two PCs, depending on
how close they are to each PC axis and the periphery of the plot. Variables which lied
on the periphery, close to a particular PC axis contributed strongly to that PC. Thus,
the variables in groups A (MDT, VDT, F2, lag time, T25%, T50% and T75%) and B
(dissolution rate constant from the models and compression pressure) contributed to
PC1 significantly. They did not contribute much to PC2. Since the two groups lied on
the opposite sides of the origin along the PC1 axis, the variables in the two groups
were negatively correlated to each other. The variables within the same group were
positively correlated to one another. As the extent of coat damage was expected to
increase with compression pressure, the dissolution parameters that were positively or
negatively correlated to the compression pressure could be used for assessing the
extent of pellet coat damage. In addition, these parameters might show similar trends
and give similar information as the compression pressure increased. RD was a major
contributor of PC2 only and was independent of the variables in groups A and B. This
indicates that RD was not correlated to compression pressure and was therefore not a
suitable parameter to assess the pellet coat damage during compression. As the release
exponent, n, derived from the power law lied near to the origin, it had very little
contribution to the explained data variance. This suggests that n was not a suitable
variable to monitor the pellet coat damage during compression. Compact
110
Figure 28. Loadings plot from PCA of dissolution parameters, compression pressure (Press) and disintegration time (DT). KH: Higuchi dissolution rate constant; KP: Power Law dissolution rate constant; n: release exponent from Power Law; LH: lag time from Higuchi equation; LP: lag time from Power Law equation; T25, T50, T75: time taken for 25 %, 50 % and 75 % of the drug to be released respectively; MDT: mean dissolution time; VDT: variance of dissolution time; RD: relative dispersion of concentration-time profiles; F2: similarity factor.
disintegration time had a weak contribution to PC1 since it was relatively nearer to the
origin. Thus, it had weak positive and negative correlations with the variables in
groups B and A respectively. This probably indicates that the influence of
disintegration time on the dissolution parameters was not consistent throughout the
RD
DTPress
KP
KH
n
LP
T25
VDT T50 F2 LH MDTT75
B
A
111
whole range of compression pressures, and was generally weak. The PCA values of
the compression pressure and dissolution parameters had provided an insight to the
relationships amongst the important variables. In addition, variables which did not
show any significant relationship with the compression pressure and other dissolution
parameters were identified from the loadings plot. However, PCA did not provide any
information on the actual variation of the dissolution parameters with the compression
pressure for each type of filler. Since T50% was strongly correlated to most of the other
variables, the use of UC/C for the assessment of pellet coat damage was warranted.
B.5.2. Relationship between compact characteristics and UC/C
Several properties of the compacts containing different fillers formed at different
compression pressure were determined. These properties included compact apparent
volume, porosity, pellet volume fraction, disintegration time and UC/C. The loadings
plot from PCA (Figure 29) shows that the data variation was largely explained by PC1
(77 %). PC2 explained only 14 % of the data variation. Most of the variables
contributed more to PC1 than PC2 except for disintegration time. The loadings plot
also shows that compression pressure, UC/C and pellet volume fraction (group A)
were positively correlated to one another, albeit not perfectly correlated. On the other
hand, compact porosity and apparent volume were negatively correlated to the
variables in group A. Though the disintegration time was in the same quadrant of the
loadings plot as the variables in group A, it lied nearer to PC2. This indicates that
disintegration time was not as strongly correlated to the variables in group A. As
mentioned earlier, disintegration time did not have any consistent correlation with the
variables in the range of compression pressure used. As PCA only provided rough
indications to possible important relationships between various variables, the
112
Figure 29. Loadings plot from PCA of compact properties, compression pressure and UC/C. App vol: compact apparent volume; PVF: pellets volume fraction.
following sections will examine more closely into the relationships of these variables
as the compression force increased.
B.5.3. Influence of compression pressure on UC/C
The changes in UC/C with compression pressure for the four types of fillers are
shown in Figure 30. At almost all the compression pressures tested, UC/C values for
compacts containing micronized lactose was the lowest, followed by PH105, PH200
and lactose 80M. The difference between the coarse grade filler and the fine grade
filler for lactose was greater than that for MCC. In addition, there was a stark
difference between the UC/C values of compacts containing micronized lactose and
A
113
Figure 30. Plot of UC/C against compression pressure for compacts containing lactose 80M (■), PH200 (▲), PH 105 (Δ), and micronized lactose ( ). the other types of fillers. This difference became even more pronounced at higher
compression pressures. This suggests that a different mechanism of interparticulate
interaction occurred during compression for micronized lactose. Due to the small
particle size of the micronized lactose (d50 = 2.3 µm), the particles did not create deep
indentations on the pellet coat. The greatest depth of indentation that could be created
was possibly not more than the diameter of the filler particle. As the average coat
thickness of the coated pellets was approximately 15 µm (refer to Part C, Table 12,),
indentations caused by micronized lactose particles would not traverse through the
0
2
4
6
8
10
12
14
16
18
0 100 200 300 400 500Compression pressure (MPa)
UC
/C
114
entire thickness of the coat. Moreover, the diameters of any indentation produced by
micronized lactose particles were unlikely to be much wider than their diameters. Due
to a greater resistance for smaller lactose particles to fragment (York, 1978, Roberts
and Rowe, 1986), particle rearrangement was probably the predominant mechanism
for densification (Fell and Newton, 1971). It was found that the transition in the stage
of compression from particle rearrangement to particle deformation occurred at a
much higher compression pressure compared to coarser grades of the same material
(Huffine and Bonilla, 1962, York, 1978). This prolonged stage of particle
rearrangement could help to reduce the extent of coat damage during compression.
During compression, instead of piercing into the pellet coat, the micronized lactose
particles probably repositioned and slipped into any adjacent small voids available
due to their small dimensions. Furthermore, unlike the coarser lactose particles, the
micronized lactose particles arranged themselves to conform to the outer contours of
the coated pellets. This allowed the compression force to be distributed over a greater
number of contact points and therefore larger area. The greater specific surface area of
the micronized particles also reduced the axial force, f, on each particle in the
compact according to an equation derived by Huffine and Bonilla, 1962:
app
2Pd'mfρ
= Equation 31
where m′ is the proportionality constant, P is the pressure exerted on the mass of
particles, d is the diameter of particles, and ρapp is the apparent density.
The relationships between UC/C and compression pressure for the fillers followed
sigmoidal patterns (Figure 30). In Figures 31 and 32, UC/C and ln ⎟⎠⎞
⎜⎝⎛− D11 were
115
A)
B)
Figure 31. Heckel plot (opened symbols, dotted line) and plot of UC/C (closed symbols, solid line) against compression pressure for compacts containing (A) lactose 80M, (B) micronized lactose.
0
0.5
1
1.5
2
2.5
3
3.5
0 50 100 150 200 250 300 350 400 450 Compression pressure (MPa)
Ln (1
/1-D
) (op
ened
sym
bols
)
0
2
4
6
8
10
12
14
16
18
UC
/C (c
lose
d sy
mbo
ls)
(b) y = 0.012x+1.2 R2 = 0.98
(a) y = 0.17x-0.37 R2 = 0.99
(c) y = 0.0020x+2.2 R2 = 1.0
(a)
(b)
(c)
0
0.5
1
1.5
2
2.5
3
3.5
0 50 100 150 200 250 300 350 400 450 Compression pressure (MPa)
Ln (1
/1-D
) (op
ened
sym
bols
)
0
0.5
1
1.5
2
2.5
3
3.5
4
4.5
5
UC
/C (c
lose
d sy
mbo
ls)
(c) y = 0.0040x+1.6R2 = 1.0
(a) y = 0.016x+0.625 R2 = 0.98
(b) y = 0.0075x+1.2 R2 = 0.99 (a)
(b)
(c)
116
C)
D)
Figure 32. Heckel plot (opened symbol, dotted line) and plot of UC/C (closed symbol, solid line) against compression pressure for compacts containing (C) PH200 and (D) PH105.
Ln (1
/1-D
) (op
ened
sym
bols
)
0
0.5
1
1.5
2
2.5
3
3.5
4
0 50 100 150 200 250 300 350 400 450 Compression pressure (MPa)
0
2
4
6
8
10
12
14
UC
/C (c
lose
d sy
mbo
ls)
(b) y = 0.013x+0.90 R2 = 1.0
(a) y = 0.090x-0.19 R2 = 0.99
(c) y = 0.0042x+2.0 R2 = 0.99(a)
(b)
(c)
0
0.5
1
1.5
2
2.5
3
3.5
0 50 100 150 200 250 300 350 400 450 Compression pressure (MPa)
Ln (1
/1-D
) (op
ened
sym
bols
)
0
2
4
6
8
10
12
14
16
UC
/C (c
lose
d sy
mbo
ls)
b) y = 0.015x+0.80R2 = 1.0
a) y = 0.11x+0.18R2 = 1.0
c) y = 0.0014x+2.7R2 = 0.97
(a)
(b)
(c)
117
plotted against the compression pressure on the same graph for individual fillers. The
plot of ln ⎟⎠⎞
⎜⎝⎛− D11 against the compression pressure is also known as the Heckel plot
(Heckel, 1961a, b). The changes in UC/C can be related to the stages of densification,
as well as the rate of change in compact density with pressure. The expression,
ln ⎟⎠⎞
⎜⎝⎛− D11 , is a mathematical transformation of the relative density, which is a ratio of
compact apparent density to true density. A larger value is indicative of higher
compact relative density.
At low compression pressures of approximately 1.13 to 11.3 MPa, the UC/C values
for all the materials were close to unity (lag phase). This indicates that the rates of
drug release for the compressed coated pellets and the uncompressed pellets were
rather similar. During this lag phase before appreciable pellet coat damage occurred,
there were adequate void spaces in the immediate vicinity for particle rearrangement
without causing any pronounced coat damage. From the plots in Figures 31 and 32, it
can be seen that the lag phase for UC/C coincides with the initial non-linear portion of
the Heckel plot. This portion has been known to correlate with the particle
rearrangement stage of powder densification (Heckel, 1961b). Hence, this observation
supports the supposition that there was minimal or no coat damage when the particles
were loosely packed and underwent mainly rearrangement during compression.
As the compression pressure increased further to between approximately 11 to 113
MPa, the UC/C values rose linearly and rapidly for all the fillers except for
micronized lactose. The rates of increase in UC/C values for the different fillers
decreased in the following order: lactose 80M (0.17) > PH200 (0.11) > PH105 (0.090)
> micronized lactose (0.016) (Figures 31 and 32). This order is identical to the
118
decreasing order of UC/C values for compacts compressed with the four filler types.
The acceleration phase for UC/C values coincided with the stage when relative
density increased most rapidly. At this relatively linear stage of the Heckel plot, the
consolidation of particles occurred. As the compact density increased, the coated
pellets and fillers were brought to closer contact. The void spaces suitable for particle
rearrangement, as well as the volume capable of accommodating the coated pellets
and fillers were also reduced, increasing the number of interparticulate contact points.
Coat damage occurred at the pellet-filler or pellet-pellet contact points if the stress
was not completely alleviated through filler particle deformation or rearrangement
into void spaces. As mentioned earlier, the ability to relieve stress was dependent on
the packing structure of the different fillers. In addition, coarser particles which did
not deform or collapse internally during compression had pellet coats damaged to a
greater extent through lancing coat surfaces. It is interesting to note that the rate of
increase in UC/C for the different fillers in this attritive phase did not correlate with
the rate of increase in compact density (Figures 31 and 32). The rate of increase in
UC/C for the different fillers decreased in the following order: lactose 80M (0.17) >
PH200 (0.11) > PH105 (0.090) > micronized lactose (0.016). However, the rate of
increase in density for the different fillers decreased in the order: PH200 (0.015) >
PH105 (0.013) > lactose 80M (0.012) > micronized lactose (0.0075) (Figures 31 and
32). Those findings indicate that the extent of coat damage was not solely affected by
the compressibility of material. The lacerating effects of the coarse lactose particles
during compression had also possibly contributed to the extensive rate of coat damage
despite a relatively lower rate of volume reduction with compression pressure.
The critical compression pressure (CCP) at the transition between the UC/C lag phase
and attritive phase was calculated from the linear equation of the UC/C acceleration
119
phase. The CCP is the compression pressure at which the UC/C value is one (Figures
31 and 32). It was determined that the CCP of the different fillers increased in the
following order: PH200 (7.45) < lactose 80M (8.06) < PH105 (13.22) < micronized
lactose (23.4). The CCP for the finer grades of both lactose and MCC were greater
than the coarser grades. This showed that the finer grades of lactose and MCC
underwent particle rearrangement over a wider range of compression pressures prior
to extensive particle deformation and coat damage. Thus, the earlier postulation that
fillers with greater interparticulate friction (higher Hausner ratio) had a longer phase
of particle rearrangement was valid. The longer phase of particle rearrangement was
detected as one of the mechanisms which resulted in lower extent of coat damage
when finer grades of fillers were used.
At higher compression pressures, the rate of increase in UC/C decelerated and started
to level off. A sudden increase in the UC/C value (Figures 31 and 32, circled) was
observed during this phase for all fillers. This sharp increase in UC/C coincided with
the beginning of the phase on the Heckel plot where the rate of compact densification
decreased. The decrease in rate of compact densification indicates a stronger
resistance to densification and a reduction in compressibility. This suggests the
beginning of a compression phase when the particles in the compact were in even
closer contact with one another after the earlier phases of rearrangement and particle
deformation. Consequently, pressure exerted accumulated at the interparticulate
contact points, resulting in forceful interparticulate interactions as particle
rearrangement ceased. After the sharp increase in UC/C, further increase in UC/C
value was not observed. Instead, the UC/C value even decreased appreciably. This
could be due to several reasons. Firstly, due to the decreased rate of densification
within this range of compression pressure, further increase in UC/C from the peak
120
value should not be too great. Secondly, as mentioned earlier, the pellets were in
closer proximity in this stage. Some of the pellets were fused or agglomerated. The
pellets remained agglomerated even upon compact disintegration, hence decreasing
the total surface area for drug release, resulting in a decrease in UC/C value. This was
reflected by the increasing trend in disintegration time after the peak of UC/C value
(Figures 33 and 34).
B.5.4. Influence of pellet volume fraction on UC/C
In order to investigate the influence of pellet volume fraction on the extent of pellet
coat damage, the pellet volume fraction of the compacts made at different
compression pressures were calculated. The UC/C values were then plotted against
the pellet volume fraction in Figure 35 so as to correlate the extent of pellet coat
damage to pellet volume fraction for the different fillers. The pellet volume fraction
and the rate of change in UC/C values for the individual fillers are shown on the plots
in Figure 36. The differences in UC/C value at each pellet volume fraction for all the
four types of fillers were less pronounced or large compared to when plotted against
the compression pressure (Figure 21). Nevertheless, the UC/C values for micronized
lactose were still distinctly lower than those for the other fillers. The following rank
order was noted: lactose 80M > PH200 > PH105 > micronized lactose (Figure 35).
The above observations suggest that the different protective effects of the fillers were
partially due to their compressibility or volume change under pressure. However, in
particular for micronized lactose, other mechanisms of pellet coat protection, which
has been discussed previously, had also contributed to the preservation of coat
121
A)
B)
Figure 33. Plot of disintegration time (opened symbols) and UC/C (closed symbols) against compression pressure for compacts containing (A) lactose 80M, or (B) micronized lactose.
Compression pressure (MPa)
0
10
20
30
40
50
60
70
80
90
100
0 50 100 150 200 250 300 350 400 450
Dis
inte
grat
ion
time
(s)
0
0.5
1
1.5
2
2.5
3
3.5
4
4.5
5
0
2
4
6
8
10
12
14
0 50 100 150 200 250 300 350 400
Dis
inte
grat
ion
time
(s)
0
2
4
6
8
10
12
14
16
18
UC
/C
Compression pressure (MPa)
122
0
20
40
60
80
100
120
140
0 100 200 300 400 500
Compression pressure (MPa)
Dis
inte
grat
ion
time
(s)
0
2
4
6
8
10
12
14
16
UC
/C
0
50
100
150
200
250
300
0 100 200 300 400 500
Compression pressure (MPa)
Dis
inte
grat
ion
time
(s)
0
2
4
6
8
10
12
14
UC
/C
A)
B)
Figure 34. Plot of disintegration time (opened symbols) and UC/C (closed symbols) against compression pressure for compacts containing (A) PH200, or (B) PH105.
123
Figure 35. Relationship between UC/C and pellet volume fraction for lactose 80M (♦), PH200 (▲), PH105 (Δ) and micronized lactose (■).
Pellet volume fraction
0
2
4
6
8
10
12
14
16
18
0.24 0.29 0.34 0.39 0.44 0.49
UC
/C
124
function during compression. Generally, it can be seen from Figure 35 that the UC/C
values remained close to unity initially. After a certain pellet volume fraction, the
UC/C value increased sharply to a peak value. This corresponded to the initial lag
phase followed by an acceleration phase respectively. Thereafter, the UC/C values
started to decrease. Regardless of the type of filler, two critical pellet volume fractions
(CPVF) were observed. The first CPVF corresponded to start of the acceleration
phase. The second CPVF corresponded to the peak of UC/C value.
The transitions and critical points observed could possibly be explained by the
percolation theory. A percolation system consists of sites in an infinitely large lattice.
Each site in the lattice is occupied randomly, independent of its neighbours, forming
loose clusters. A percolating or continuous cluster is formed when the occupied sites
span through the length, width and height of the system by being in contact with one
another continuously (Stauffer and Aharony, 1992). The condition or concentration at
which a percolating or continuous cluster is first formed is known as the percolation
threshold. Some property of the system may be altered abruptly at the percolation
threshold (Leuenberger et al., 1992, 1996). The percolation theory has been applied to
tablet compression and dissolution kinetics (Leuenberger et al., 1992, 1996, Leu and
Leuenberger, 1993, Beckert et al., 1998, Kuny and Leuenberger, 2003). Tablet
properties which had been observed to change at the percolation thresholds included
disintegration time, tensile strength, deformation hardness (Leuenberger et al., 1992,
1996, Leu and Leuenberger, 1993), content uniformity (Beckert et al., 1998) and
activity of pressure sensitive enzyme (Kuny and Leuenberger, 2003). Several
percolation thresholds were observed in the process of compact formation (Holman,
1991, Leuenberger et al., 1992). Generally, the first percolation threshold occurred
when a component formed a continuous network of particles bonded to one another.
125
This was also the point of transition from powder to a coherent compact. In a binary
mixture, the threshold at which the second component percolated was also found. The
final threshold point occurred when the percolating pore network became
disconnected pore clusters distributed randomly in the percolating network of solid
matrix. This was also the point of transition from a particulate structure to a
continuum solid body (Holman, 1991).
In the current compressed multiparticulate system, two CPVFs were observed
(Figures 35). The first CPVF was likely to correlate with the threshold for the
formation of a percolating network of particles bonded to one another. Before this
critical threshold, particles underwent rearrangement with the application of
compression pressure. In addition, transmission of forces did not occur
homogeneously throughout the system and forces were not transmitted to some
particles (Holman, 1991). The system consisted of isolated clusters of bonded
particles. Bonds were primarily formed between the surfaces of the particles. Hence,
pellet coat damage was minimal. Beyond this bond percolation threshold for the
particles, particle deformation became more extensive as particles could no longer be
easily displaced in the lateral direction (Leuenberger et al., 1992). Particle
deformation included deformation of filler particles through plastic flow or brittle
fracture, pellet surface indentations, as well as pellet core deformation or fracture,
depending on the type of filler. It was thus not surprising for the extent of pellet coat
damage to rise sharply in this attritive phase. The linear increase in UC/C values with
pellet volume fraction implied that the amount of pellet coat damage was directly
dependent on the compact density in this attritive phase. The second CPVF observed
was likely to correlate with the percolation threshold of the coated pellets. With the
progressive deformation of filler particles and coated pellets before this point, the
126
coated pellets were brought closer to one another. Eventually this led to the formation
of a percolating network of coated pellets in the system. Surelease coated pellets were
found to form compacts with higher tensile strength and longer disintegration time
compared to compacts formed from uncoated pellets (Maganti and Celik, 1994). This
was attributed to the binding or adhesive properties of Surelease. Hence, the
formation of a network of coated pellets in the current study brought about the
increase in compact disintegration time at or after the percolation threshold (Figures
33 and 34). This contributed to the observed decreasing trend in UC/C values with
compression pressure beyond the second CPVF.
The first CPVF was calculated from the linear equation of the attritive phase shown in
Figure 36 by assuming the UC/C value (y value) to be one. The rank order of the
fillers with increasing first CPVF was PH200 (0.33) < PH105 (0.35) < lactose 80M
(0.37) < micronized lactose (0.39). This showed that the first CPVF varied for
different fillers. This was attributed to the differences in filler particle shape, size
distribution and deformation mechanism (Holman, 1991). Coincidentally, this range
of pellet volume fraction was close to the pellet volume fraction (0.35) at which
extensive coat damage was also observed by Wagner (2000b). Therefore, pellet coat
damage could be avoided by producing compacts with pellet volume fraction below
approximately 0.3, provided that the compact formed is of a reasonable mechanical
strength. Besides, it was also interesting to note that the first CPVF for lactose fillers
was higher than the MCC fillers. This meant that the compacts containing MCC
underwent the attritive phase at lower pellet volume fractions. This finding further
supported the postulation that MCC fillers underwent a shorter phase of particle
rearrangement due to the lower interparticulate friction (lower Hausner ratio)
127
(A) (B)
(C) (D)
Figure 36. Plot of UC/C against pellet volume fraction for compacts containing (A) lactose 80M, (B) micronized lactose, (C) PH200 and (D) PH105.
y = 108.94x - 38.77R2 = 1.00
0
2
4
6
8
10
12
0.35 0.4 0.45 0.5
Pellet volume fraction
UC
/C
y = 38.80x - 14.24 R2 = 0.93
0
2
4
6
8
10
12
0.3 0.35 0.4 0.45
Pellet volume fraction
UC
/C
0.386
0.5
y = 59.93x - 18.69R2 = 0.99
0
2
4
6
8
10
12
0.2 0.3 0.4
Pellet volume fraction
UC
/C
0.325
y = 63.58x - 21.15R2 = 0.99
0
2
4
6
8
10
12
0.25 0.3 0.35 0.4 0.45
Pellet volume fraction
UC
/C
128
compared to lactose fillers. This was also the reason given for the observed higher
UC/C values observed when coated pellets were compressed with finer grades of
MCC fillers (< 80 μm) compared to lactose filler of similar particle size. However,
despite the longer phase of particle rearrangement for lactose fillers, the lactose filler
of coarser grade (lactose 80M) induced a greater extent of coat damage compared to
the MCC fillers. This was due to the significantly higher rate of increase in UC/C
value with pellet volume fraction for lactose 80M compared to MCC fillers (Figure
36). The sharp edges of coarse lactose particles could pierce the pellet coat, leading to
the greater increase in UC/C value with pellet volume fraction in the attritive phase.
From the rank order of the fillers, it was also evident that for each filler type, the
smaller grade had a lower CPVF and thus a longer rearrangement phase with respect
to pellet volume fraction. Consequently, the fillers with smaller particle size had a
greater potential to preserve pellet coat function during compression with an attritive
phase at higher pellet volume fractions. Therefore, a longer phase of particle
rearrangement with respect to pellet volume fraction (higher CPVF value) delayed the
attritive phase and hence lowered the potential for pellet coat damage. However, the
eventual extent of pellet coat damage was also dependent on the rate of increase in
coat damage with pellet volume fraction besides the value of the first CPVF.
The second critical pellet volume fraction was approximately 0.48. The filler type had
little influence on the value of this second critical pellet fraction. This was not
unexpected since this point correlated to the percolation threshold of the coated pellets
and all the compacts contained the same type of coated pellets. For coating materials
which were able to accommodate core deformations during compression up to the
second critical point, it was essential not to exceed this point. This was because the
coated pellets would fuse together beyond this point, resulting in slower- or non-
129
disintegrating matrices, thus affecting the dissolution profile even if the coat remains
undamaged.
During the attritive phase, the UC/C values of compacts containing lactose 80M
increased the most rapidly as the pellet volume fraction increased. This was followed
by PH105, PH200 and micronized lactose (Figure 36). The slightly slower change in
UC/C with pellet volume fraction for PH200 compared to PH105 was probably due to
the greater porosity and deformability of PH200 as discussed in the earlier sections.
Hence, PH200 had a greater ability to disperse stress through structural deformation
compared to PH105. However, the UC/C values of compacts containing PH200 as the
filler were generally higher than that containing PH105 as filler (Figure 35). This was
probably due to the higher value of the first CPVF for PH105, indicating the start of
the attritive phase at a higher pellet volume fraction. This again illustrated the
influence of the first CPVF value on the extent of pellet coat damage.
B.6. Influence of compression speed on the extent of coat damage
In this part of the study, coated pellets were compressed with four filler types at
different compression speeds (rate of upper punch displacement) to the same
compression pressure (28.3MPa). With the exception of the fastest compression speed
(300 mm/min), the UC/C values for the different fillers decreased in the following
order: lactose 80M > PH200 > PH105 > micronized lactose (Figure 37). At the fastest
compression speed, the UC/C values for PH105 were slightly higher than PH200.
Generally, there was a decrease in the UC/C value when the compression speed
increased from 1 mm/min to 10 mm/min. From 10 mm/min to 100 mm/min, the
differences in the UC/C value were generally insignificant statistically (Figure 37).
However, when the compression speed increased from 100 mm/min to 300 mm/min,
130
Figure 37. UC/C values of compacts containing lactose 80M (■), PH105 (Δ), PH200 (▲) and micronized lactose (□) formed at different compression speeds. *One-way ANOVA shows insignificant difference in means of the two consecutive points (p >0.05).
there was a sharp increase in UC/C values for lactose 80M and a much smaller
increase in UC/C values for the other fillers. The discontinuity in the decreasing trend
of UC/C values with compression speeds suggested that the effect of compression
speed on the compression behaviour of compact materials was not straightforward.
It was reported that as the compression speed increased and dwell time decreased, the
extent of plastic deformation of Surelease coat was reduced (Maganti and Celik,
1994). In addition, the elastic expansion during decompression was also found to
increase at higher compression speeds. Hence, the reduction in plastic deformation
and greater elasticity of the coat material could increase its resistance to deformation
Log Compaction speed (mm/min)
0
1
2
3
4
5
6
7
8
0.1 1 10 100 1000
UC
/C
* *
** **
* * *
131
and coat damage. However, besides the Surelease coat, the yield pressure and
elasticity of the fillers also increased with higher compression speeds, especially for
materials which deformed plastically (Roberts and Rowe, 1986, Maarschalk et al.,
1996, Maarschalk et al., 1997). These observations were due to insufficient time for
the materials to react and deform when the compression speed was increased
(Armstrong and Palfrey, 1989). Since the extent of changes in yield pressure and
elasticity with compression speed varied for different materials, the differences in the
elasticity and yield pressure between the coated pellets and fillers could change as the
compression speed increased. This could modify the consolidation behaviour of the
filler particles and coated pellets as they interacted with each other during
compression (Nokhodchi and Rubinstein, 1998). For example, if the relative increase
in yield pressure for a coarse grade filler was greater than the coated pellets, the yield
pressure of the filler could be much higher than that of the coated pellets at a certain
compression speed. This could lead to a greater extent of pellet damage and
indentations on the pellet surfaces during compression. Therefore, the effect of
compression speed on the extent of coat damage varies, depending on the relative
amount of changes in the elasticity and yield pressure of the different compact
components.
It was shown in several previous reported studies that as the elasticity and yield
pressure of materials increased with compression speed, the porosity of the compacts
formed also increased (Maarschalk et al., 1996, Tye et al., 2005). This was not
surprising as consolidation and densification were expected to be reduced as the
material exhibits greater elasticity and resistance to yield. However, from Figure 38, it
can be seen that for all the fillers, the porosity of the compacts decreased significantly
(p > 0.05) when the compression speed increased from 100 to 300 mm/min. This
132
Figure 38. Porosity of compacts containing PH200 (▲),PH105 (Δ), micronized lactose (□) and lactose 80M (■) formed at different compression speeds. *One-way ANOVA shows insignificant difference in means of two consecutive points (p >0.05).
suggested that there was a modification in the consolidation behaviour of the compact
components. It also indicated a greater extent of pellet or filler densification. Since
there was a corresponding increase in UC/C value as the compression speed was
increased from 100 to 300 mm/min, the reduction in porosity was likely to be a
reflection of a greater extent of pellet densification.
1000 0.15
0.20
0.25
0.30
0.35
1 10 100
Log [Compaction speed (mm/min)]
Poro
sity
*
* *
*
*
*
* *
133
PART C. INFLUENCE OF PELLET SIZE AND PELLET TYPE IN
COMPACTS CONTAINING COATED PELLETS
C.1. Dissolution profiles of uncompressed pellets
C.1.1. Differences in drug release from coated MCC pellets and sugar pellets
The sugar pellets were classified into Group I, II or III according to their size range.
The average mean diameter of the pellets is shown in Table 12. Pellets in Group I had
the smallest mean diameter while those in Group III had the largest mean diameter.
The dissolution parameters and dissolution profiles of the uncompressed sugar pellets,
as well as that of MCC pellets are shown in Table 13 and Figure 39 respectively. The
dissolution profiles of both the sugar and MCC pellets followed the Higuchi model
more closely (Table 13). Hence, the predominant mechanism of drug release from the
different types of core was similar. This is in agreement with what has been reported
by Muschert et al. (2009). Since Higuchi model was derived from Fick’s law, the drug
release from MCC and sugar pellets was dependent on the concentration gradient
across the insoluble semi-permeable ethylcellulose coat.
Compared to the sugar pellets, the Higuchi dissolution rate constant of MCC pellets
was higher, indicating that the overall drug release rate from MCC pellets was faster
than the sugar pellets. This contradicted the observation of faster drug release rate
from soluble sugar pellets compared to insoluble pellets in some studies (Tang et al.,
2000, Kállai et al., 2010). However, another study showed faster drug release from
MCC pellets compared to sugar pellets (Muschert et al. 2009). In order to elucidate
the cause for the higher drug release rate from MCC pellets, the osmolality of
solutions containing different concentrations of dissolved pellets (without drug and
ethylcellulose coat) was determined (Table 14). In addition, SEM photomicrographs
134
Table 12. Physical properties of pellets.
Pellet core Group I
sugar pellets
Group II sugar pellets
Group III sugar pellets
MCC pellets
Average mean diameter of drug
loaded pellets (µm) 358.0 453.5 655.3 664.5
Average mean diameter of
ethylcellulose coated pellets (µm)
388.2 484.0 688.2 696.9
Coat thickness (µm) 15.1 15.3 16.5 16.2
True density of coated pellets
(g/cm3) 1.36 1.37 1.43 1.42
Surface roughness (Ra) of coated pellets (nm)
2787.1 (752.6)
2406.4 (618.8)
2068.7 (507.5)
2357.1 (633.2)
Load stress at break (N/mm2)
12.4 (2.2)
11.4 (1.9)
14.7 (1.9)
43.6 (5.7)
Table 13. Curve-fitting of dissolution date and dissolution parameters of different coated pellet cores.
Zero order Higuchi order
Pellet core R2 R2 KH tL
Group I sugar pellets 0.906 0.969 5.89 (0.11) 2.37 (0.38)
Group II sugar pellets 0.890 0.981 5.75 (0.03) 2.10 (0.14)
Group III sugar pellets 0.850 0.967 5.58 (0.08) 1.27 (0.11)
MCC pellets 0.928 0.969 11.89 (0.36) 2.72 (0.03)
135
Figure 39. Drug release profile of MCC pellets (▲), Group I sugar pellets (♦), Group II sugar pellets (■) and Group III sugar pellets (Δ).
0
10
20
30
40
50
60
70
80
90
100
0 50 100 150 200 250 300 350 Time (min)
% d
rug
rele
ased
136
Table 14. Osmolality of solutions containing different concentrations of pellets.
Pellet concentration (g/mL) 0.01 0.05 0.1
Osmolality of sugar pellet solution (mmol/kg) 48 118 248.3
Osmolality of MCC pellet solution (mmol/kg) 33.3 32.0 33.0
of the coated pellets were taken (Figure 40) and surface roughness of the coated
pellets determined (Table 12).
Although the dissolution rate constant of coated MCC pellets was greater than that of
coated sugar pellets, the lag period (tL) of MCC pellets was significantly longer than
that of the sugar pellets (group III). From Table 14, the osmolality of the solutions
containing dissolved sugar pellets was higher than that of the MCC pellets at all pellet
concentrations. This was expected since sucrose was soluble and MCC was insoluble
in water. Hence, the sugar pellets were more osmotically active than the MCC pellets.
Consequently, the coated sugar pellets drew in water at a faster rate, resulting in a
more rapid initial drug release. The length of the lag phase was predominantly
affected by the osmotic activity of the core.
Cracks were observed on the surface of the coated MCC pellets (Figure 40). Thus,
besides the membrane barrier, diffusion of drug occurred through the cracks, resulting
in higher drug release rate from MCC pellets. Cracks formed could be due to the
difference between the expansion coefficients of the coat and the substrate (Rowe,
1997, 1980, Hancock and Rowe, 1998) during coating. A greater volume expansion of
the core compared to the coat during coating could have caused the formation of
cracks. According to the two-tailed Student’s t test, the surface roughness of the
137
Figure 40. SEM photomicrographs of coated (A) Group I sugar pellets, (B) Group II sugar pellets, (C) Group III sugar pellets and (D) MCC pellets at (1) 200 and (2) 1000 times magnification.
A1 A2
B1 B2
C1 C2
D1 B2
Crack
138
coated MCC pellets was significantly higher than the coated sugar pellets (p-value =
0.014, Table 12). Hence, the faster drug release could also have been brought about
by the rougher surface of coated MCC pellets (Porter, 1997).
C.1.2. Differences in drug release from coated sugar pellets of different particle
size
Using optical microscopy, the sugar pellets of different particle size were found to
have similar ethylcellulose coat thickness (Table 12). The theoretical amount of
coating material deposited on each pellet was also calculated by dividing the
theoretical weight gain of pellet after coating by the surface area of each pellet. It was
found that the sugar pellets of different particle size were coated to comparable
amount of coating material per unit surface area. This observation agrees with the
ethylcellulose coat thickness determined for the pellets of different particle size.
Comparing the dissolution rate constant for the sugar pellets of different size
fractions, Group I pellets (smallest pellets) exhibited the fastest drug release. This
could be attributed to the higher specific surface area of smaller pellets. However, the
drug release became slower than that of the Group III pellets (largest pellets) after 75
% of the drug was released (Figure 39). For sugar pellets in Group II (medium sized
pellets), the drug release profile was similar to that of Group III sugar pellets initially.
After about 35 % of the drug was released, the drug release was slower than that of
Group III pellets. Nevertheless, the dissolution rate constant derived from the Higuchi
model decreased as the pellet size increased (Table 13). Hence, the overall drug
dissolution rate was still the fastest for Group I pellets (smallest), followed by Group
II pellets and Group III pellets (largest pellets). This shows that the drug dissolution
rate was still mainly dependent on the specific surface area of coated pellets. The
139
slower drug dissolution in the later stages of drug dissolution exhibited by Group I
and II pellets could be due to an earlier depletion of drug in the pellets compared to
Group III pellets. The earlier depletion of drug in Group I and II pellets could be
brought about by the faster initial dissolution rate from these pellets. From Table 13,
the lag time of the Group I pellets (smallest pellets) was the longest, followed by
Group II pellets and Group III pellets (largest pellets). This could be due to the greater
specific surface area of the smallest pellets which required a longer time to be wetted
before drug dissolution could take place.
C.2. Influence of pellet core material on the extent of coat damage
As discussed in Parts 4A and B, the filler particle size had an influence on the extent
of pellet coat damage during compression. Thus, sugar and MCC pellets of the same
size fraction were compressed with coarse and fine grades of two filler types. Lactose
and MCC were chosen as the fillers as they consolidated by different deformation
mechanisms. Furthermore, their particle sizes affected the UC/C value to different
extents as observed in previous studies. Lactose 80M (d50 = 204.2 µm) and
micronized lactose (d50 = 2.3 µm) were used as the coarse and fine grades of lactose
respectively. On the other hand, PH200 (d50 = 168.4 µm) and PH105 (d50 = 27.6 µm)
were used as the coarse and fine grades of MCC respectively.
The UC/C values, apparent volume, porosity and the pellet volume fraction of the
compacts containing Group III sugar pellets and MCC pellets with different fillers are
shown in Figure 41. From Figure 41A, it can be seen that for both lactose and MCC
fillers, the extent of coat damage of sugar pellets was significantly greater (p < 0.05)
when compressed with the coarse grade filler. The UC/C value was the lowest when
micronized lactose was used as the filler. In contrast, compression of MCC pellets
140
with the fine grade fillers did not result in any significant reduction (p > 0.05) in the
UC/C value. Interestingly, the UC/C value of these compacts containing micronized
lactose was not the lowest among the four fillers. A significantly lower (p < 0.05)
UC/C value was observed for MCC pellets compared to sugar pellets for all the fillers
except micronized lactose. When micronized lactose was compressed with MCC
pellets, the UC/C value was approximately double (UC/C = 2.34) of that for sugar
141
0.00
1.00
2.00
3.00
4.00
UC
/C
0.80
0.90
1.00
1.10
App
aren
t vol
ume
(cm
3 )
*
0.00
0.10
0.20
0.30
0.40
Poro
sity
**
*
0.300.320.340.360.380.40
Lactose 80M Micronizedlactose
PH200 PH105
Filler type
Pelle
t vol
ume
frac
tion
*
*
(A)
(B)
(C)
(D)
Figure 41. (A)UC/C, (B) apparent volume, (C) porosity and (D) pellet volume fraction of compacts containing group III sugar pellets (shaded) and MCC pellets (clear) with different fillers. *Two-sample t-test shows insignificant difference in means between compacts containing sugar pellets and MCC pellets (p >0.05).
142
pellets (UC/C = 1.10). The lower UC/C values of MCC pellets compressed with other
fillers could be due to the higher load stress at break for MCC cores compared to
sugar cores (Table 12). A higher load stress at break suggests a greater tensile
strength. Load stress at break was determined from the first peak observed on the
force displacement curve. The drop in the force detected at the first peak was much
smaller for the MCC pellets compared to sugar pellets (Figure 42). This suggests that
the MCC pellets only underwent surface tensile failure, resulting in cracks on the
surfaces of the pellets at the first peak (Salako et al., 1998). Further application of
stress did not result in total core failure or core fragmentation. Instead, further core
shape deformation and densification were observed. The dominating mechanism of
compression for MCC pellets had also been reported to be deformation and
densification (Johansson, et al., 1995, 1998). Unlike the MCC pellets, the tensile
stress for the sugar pellets dropped markedly at the first peak, suggesting major flaws
and failure to have occurred inside the sugar pellet cores (Salako et al., 1998).
Visually, the sugar pellets were observed to have fragmented after tensile tests.
Hence, the sugar pellet cores were considered to be more brittle compared to MCC
pellet cores. Brittle fracture of cores could result in coat damage and loss of coat
function. In contrast, loss of coat function for cores that undergo densification and
shape deformation were more gradual. Coat rupture occurred when the coat was
unable to accommodate the changes in core deformation (Tunon et al., 2003a).
Compacts containing MCC pellets had a greater apparent volume than equivalent
compacts containing sugar pellets when compressed with different fillers (Figure
41B). Furthermore, the average deformability of the MCC pellets was significantly
smaller than the sugar pellets (0.0125 mm/N versus 0.0187 mm/N respectively). This
further indicates a smaller extent of densification in MCC pellets during compression
143
Figure 42. A typical force-displacement curve of (A) a Group III sugar pellet and (B) a MCC pellet.
compared to sugar pellets, resulting in a lower level of coat damage seen. The harder
MCC cores were also more resistant to the surface indentations brought about by
sharp edges of coarser filler particles. Moreover, the effects of surface indentations on
coat damage for deformable cores such as MCC pellets were not as severe as for
brittle cores. This was because surface indentations on brittle cores had acted as
additional points of weakness from which cracks could propagate, leading to pellet
core fragmentation. This was less likely to occur for MCC pellets since they were
mechanically stronger and comparatively less brittle than sugar pellets. Hence, unlike
the sugar pellets, coat damage for MCC pellets was largely due to the extent of core
deformation rather than surface indentation or pellet fragmentation. The extent of
MCC core densification during compression depended mainly on the deformability
Forc
e (N
)
0
2
4
6
8
10
12
14
0 0.05 0.10 0.15 0.20Displacement (mm)
(A)
(B)
144
and yield pressure of the filler particles. The yield pressure of the filler decreased in
the following order: lactose 80M (208.3 MPa) > micronized lactose (204.1 MPa) >
PH105 (77.5 MPa) > PH200 (76.3 MPa) This order corresponds to the increasing
order of apparent volume and porosity of compacts containing MCC pellets (Figure
41B and C). The order also corresponds to the decreasing order of pellet volume
fraction (Figures 41D). Therefore, compacts containing the filler with the highest
yield pressure had the lowest apparent volume, porosity and highest pellet volume
fraction. As a result, MCC pellets compressed with both grades of lactose fillers were
damaged to a greater extent than when compressed with both grades of MCC fillers.
C.3. Influence of sugar pellet size on the extent of coat damage
Regardless of the coated sugar pellet size, the UC/C values of compacts containing
lactose 80M was always the highest, followed by PH 200, PH 105 and micronized
lactose (Figure 43). In agreement with the earlier experimental work, the rank order
of the UC/C values (lactose 80M > PH200 > PH105 > micronized lactose) of
compacts containing different fillers did not correlate with the rank order of compact
porosities (lactose 80M < micronized lactose < PH105 < PH200) and pellet volume
fractions (lactose 80M > micronized lactose > PH105 > PH200) (Figures 43, 44B and
C). This was true for sugar pellets of different sizes. Nevertheless, it had been shown
in the earlier studies carried out that when the compacts contained the same filler
type, the pellet coat damage increased as the compact porosity decreased.
Except for Group I pellets (smallest pellets) compressed with PH 105, the UC/C
values of Groups I and II pellets were higher than that of Group III pellets (largest
pellets) (Figure 43). This was unexpected as smaller pellets had been reported to
deform to a lesser degree than coarser pellets during compression (Johansson et al.,
145
Figure 43. Plot of UC/C against average mean diameter of coated pellets compressed with lactose 80M (♦), PH 200 (▲), PH 105 (Δ) and micronized lactose (■).
0
1
2
3
4
5
6
7
350 400 450 500 550 600 650 700
Average mean diameter of coated pellets
UC
/C
146
0.8
0.85
0.9
0.95
1
1.05
1.1
App
aren
t vol
ume
(cm
3 )
0.2
0.22
0.24
0.26
0.28
0.3
0.32
0.34
0.36
Poro
sity
(A)
(B)
(C)
Figure 44. Plots of (A) apparent volume (B) porosity and (C) pellets volume fraction of compacts containing lactose 80M (■), micronized lactose (□), PH200 (▲) and PH105 (Δ) against the average mean pellet diameter of sugar pellets.
(µm)
0.330.340.350.360.370.380.390.4
0.41
350 400 450 500 550 600 650 700
Average mean pellet diameter
Pelle
t vol
ume
fract
ion
147
1998). Besides, the number of force transmission points increased as the pellet size
decreased. This allowed the compression stress to be distributed over a greater
number of contact points, thus reducing the stress exerted per unit point on pellet
coats.
From Figures 44B and C, the compact porosity and pellet volume fraction increased
and decreased respectively as the pellet size decreased. Hence, it was expected that
the pellet coat damage was the least for the smallest pellets by considering the
compact porosity and pellet volume fraction. However, an opposite trend was
observed. This unexpected finding was better explained by the main mechanisms of
coat damage for sugar pellets. The coat of the sugar pellet was largely damaged
through surface indentations and pellet core fragmentation. Since the smaller pellets
(Groups I and II) had greater specific surface area, a larger total surface area to
volume ratio of the pellet would be in contact with the filler particles. This resulted in
more indentations and thus a greater extent of coat damage. This also explains the
greater difference in UC/C values when they were compressed with the coarser grades
of fillers (PH200 and lactose 80M) (Figure 43) compared to the finer grades of fillers
(micronized lactose and PH105). In addition, the load stress values at break for the
smaller pellets (Groups I and II) were significantly lower (p < 0.05) than the bigger
pellets (Group III) according to one-way ANOVA (Table 12). Hence, the threshold to
fragment was lower for the smaller pellets, resulting in a greater extent of coat rupture
through pellet core fragmentation during compression. The difference in the UC/C
values of Groups I and II pellets was small or insignificant when they were
compressed with the same filler. This could be attributed to the smaller difference in
size and tensile strength between the two groups of sugar pellets.
148
PART 5: CONCLUSION
149
5. CONCLUSION
Particle size of lactose fillers played an important role in affecting the extent of coat
damage of pellets during compression. This was shown to be mainly due to the larger
and deeper indentations brought about by the pressure-concentrated lactose edges in
contact with the pellet surface. This mechanism of coat damage was exacerbated by
the close packing structure of coarser lactose particles. Through the addition of
micronized lactose to coarser grades of lactose, pellet coat damage could be
ameliorated. The maximum reduction in pellet coat damage was achieved when 75 %,
w/w of micronized lactose was present in the lactose filler blend. Hence, micronized
lactose can be employed to mitigate the extent of coat damage caused by coarser
lactose particles with careful consideration of its adverse effect on the overall flow
property.
Particle size of fillers was also found to be an important factor for a filler which
consolidates by plastic deformation (L-HPC). However, filler particle size was less
important in some other fillers such as MCC and crospovidone. The latter was
attributed to the ability of the filler particles to collapse internally when they were
pressing against the coated pellets. As a result, the pressure at the filler-pellet contact
points was alleviated and dissipated over a larger cross-sectional contact area of the
collapsed particles. This effect was more evident with MCC and cropovidone fillers
of coarser grades. For the finer filler grades (< 80 μm), lactose fillers were shown to
induce a lower extent of coat damage compared to MCC and crospovidone fillers of
equivalent particle size. This was attributed to the higher Hausner ratio and thus,
higher interparticulate friction for the lactose fillers compared to MCC and
crospovidone fillers. It was confirmed in the later part of the study that lactose fillers
150
underwent a longer phase of particle rearrangement than MCC fillers. As a result, this
allowed the pellets to undergo spatial adjustments at higher compression pressures,
reducing the forceful pellet-filler interactions. In addition, the compressibility studies
also showed that compressible materials could result in higher pellet coat damage due
to greater reduction in compact volume during compression. Hence, material
compressibility acted as a double-edged sword. On one hand, it could reduce the
extent of indentations caused by coarser filler particles through particle deformation.
On the other hand, the reduction of compact volume due to material compressibility
resulted in greater extent of pellet coat damage.
Depending on the compression pressure applied, the pellets underwent different
compression phases leading to different levels of coat damage. Overall, there were
three main phases, with two critical transition points. Minimal coat damage was
observed in the initial lag phase where the particles underwent mainly rearrangement.
The first critical point, which was found to be associated with the pellet volume
fraction range of 0.33 to 0.39 depending on the filler type, corresponded to the bond
percolation threshold of the compact components. Beyond this point, extensive
particle deformation occurred, resulting in acceleration of pellet coat damage with
increase in compression pressure and pellet volume fraction. This constituted the
second phase. The pellet volume fraction at the second critical point was
approximately 0.48 for all the fillers. It corresponded to the percolation threshold for
the coated pellets. Beyond the second critical point, pellet fusion and formation of
larger pellet aggregates occurred, resulting in longer disintegration time and thus,
slower drug release. This constituted the third phase. At almost all the compression
pressures tested, UC/C value for compacts containing micronized lactose was the
lowest, followed by PH105, PH200 and lactose 80M. Compared to the other filler
151
types, the coated pellets compressed with micronized lactose were less sensitive to the
changes in compression pressure, pellet volume fraction and compression speed.
In contrast to sugar pellets, the extent of coat damage of MCC pellets was not
dependent on the filler particle size. The compressibility of the filler seemed to be the
dominating factor affecting the extent of coat damage. Except micronized lactose, the
extent of coat damage of MCC pellets was lower than that of sugar pellets. These
observations were attributed to the greater mechanical strength of MCC pellets and
differences in the main mechanism of pellet deformation. Sugar pellets would
undergo brittle fragmentation under pressure while MCC pellets would undergo
densification and shape deformation under pressure. Consequently, MCC pellets were
more resistant to surface indentations. In addition, the coat damage for MCC pellets
was more gradual than sugar pellets as they were less brittle. Unlike sugar pellets,
coat damage for MCC pellets was mainly dependent on the extent of core
densification during compression rather than surface indentation and pellet
fragmentation. Hence, the most suitable filler might not be the same for different
coated pellet cores, especially when the pellets deform differently.
A greater extent of coat damage was found in sugar pellets of smaller particle size.
This was attributed to the greater specific surface area of smaller pellets in contact
with the filler particles. Furthermore, the mechanical strength of the smaller pellets
was lower than that of larger pellets. The smaller pellets exhibited a lower threshold
to fragmentation during compression.
152
PART 6: REFERENCES
153
6. REFERENCES
Abrahamsson, B., Alpsten M., Jonsson U.E., 1996. Gastro-intestinal transit of a multiple-unit formulation (metoprolol CR/ZOK) and a non-disintegrating tablet with the emphasis on colon. International Journal of Pharmaceutics, 140, 229-235.
Abbaspour, M.R., Sadeghi, F., Garekani, H.A., 2008. Design and study of ibuprofen disintegrating sustained-release tablets comprising coated pellets. European Journal of Pharmaceutics and Biopharmaceutics, 68, 747-759.
Alderborn, G., 2007. Tablets and compaction. In: Aulton, M.E. (Ed.), Aulton’s pharmaceutics: The design and manufacture of medicines. Churchill Livingstone, U.S., pp 441-482.
Altaf, S.A., Hoag, S.W., Ayres, J.W., 1998. Bead compacts. I. Effect of compression on maintenance of polymer coat integrity in multilayered bead formulations. Drug Development and Industrial Pharmacy, 24, 737-746.
Altaf, S.A., Hoag, S.W., Ayres, J.W., 1999. Bead compacts. II. Evaluating of rapidly disintegrating nonsegregating compressed bead formulations. Drug Development and Industrial Pharmacy, 25, 635-642.
Alvarez-Lorezo, C., Gómez-Amoza, J.L., Martínez-Pacheco, R., Souto, C., Concheiro, A., 2000. Evaluation of low-substituted hydroxypropylcelluloses (L-HPCs) as filler binders for direct compression. International Journal of Pharmaceutics, 197, 107-116.
Armstrong, N.A., 2008. Introduction. In: Watt, P.R., Armstrong, N.A., (Ed.), Tablet and capsule machine instrumentation. Pharmaceutical Press, U.K., pp 1-9.
Armstrong, N.A., Palfrey, P.L., 1989. The effect of machine speed on the consolidation of four directly compressible tablet diluents. Journal of Pharmacy and Pharmacology, 41, 149-151.
Asano, T., Tsubuku, S., Sugawara, S., Miyajima, M., Sato, H., Yuasa, H., Kanaya, Y., 1997. Changes in volume and compression energy upon compression of calcium silicate tablets. Drug Development and Industrial Pharmacy, 23, 679-685.
Aulton, M.E., Dyer, A.M., Khan, K.A., 1994. The strength and compaction of milllispheres. The design of a controlled-release drug delivery system for ibuprofen in the form of a tablet comprising compacted polymer-coated millispheres. Drug Development and Industrial Pharmacy, 20, 3069-3104.
Baker, R.W., 1987. Principles of membrane diffusion. In: Controlled release of biologically active agents. John Wiley & Sons, U.S., pp 22-38.
Bansal, P., Vasireddy, S., Plakogiannis, F., Parikh, D., 1993. Effect of compression on the release properties of polymer coated niacin granules. Journal of Controlled Release, 27, 157-163.
154
Barra, J., Falson-Rieg, F., Doelker, E., 1999. Influence of the organization of binary mixes on their compactibility. Pharmaceutical Research, 16, 1449-1455.
Bashaiwaldu, A.B., Podczeck, F., Newton, J.M., 2004. Application of dynamic mechanical analysis (DMA) to the determination of the mechanical properties of coated pellets. International Journal of Pharmaceutics, 274, 53-63.
Béchard, S.R., Leroux, J.C., 1992. Coated pelletized dosage form: effect of compaction on drug release. Drug Development and Industrial Pharmacy, 18, 1927-1944.
Beckert, T.E., Lehmann, K., Schmidt, P.C., 1996. Compression of enteric-coated pellets to disintegrating tablets. International Journal of Pharmaceutics, 143, 13-23.
Beckert, T.E., Lehmann, K., Schmidt, P.C., 1998. Compression of enteric-coated pellets to disintegrating tablets: uniformity of dosage units. Powder Technology, 96, 248-254.
Bodmeier R., 1997a. Tableting of coated pellets. European Journal of Pharmaceutics and Biopharmaceutics, 43, 1-6.
Bodmeier, R., Guo, X., Paeratakul, O., 1997b. Process and formulation factors affecting the drug release from pellets coated with the ethylcellulose- Pseudolatex Aquacoat. In: McGinity, J.W. (Ed.), Aqueous polymeric coatings for pharmaceutical dosage forms. Marcel Dekkar, New York, pp. 55-80.
Bodmeier, R., Paeratakul, O., 1994. Mechanical properties of dry and wet cellulosic and acrylic films prepared from aqueous colloidal polymer dispersions used in the coating of solid dosage forms. Pharmaceutical Research, 11, 882-888.
Bolhuis, G.K., Chowhan, Z.T., 1996. Materials for direct compaction. In: Alderborn, G, Nystrom, C. (Ed.), Pharmaceutical powder compaction technology. Marcel Dekker, New York, pp 419-493.
Busignies, V., Leclerc, B., Porion, P., Evesque, P., Couarraze, G., Tchoreloff, P., 2006. Investigation and modelling approach of the mechanical properties of compacts made with binary mixtures of pharmaceutical excipients. European Journal of Pharmaceutics and Biopharmaceutics, 64, 51-65.
Cal, S., Rodriguez-Puente, B., Souto, C., Concheiro, A., Gómez-Amoza, J.L., Martinez-Pacheco, R., 1996. Comparison of a spray-dried α-lactose monohydrate with fully hydrated roller-dried β-lactose. International Journal of Pharmaceutics, 136, 13-21.
Campos-Aldrete, M.E., Villafuerte-Robles, 1997. Influence of the viscosity grade and the particle size of HPMC on metronidazole release from matrix tablets. European Journal of Pharmaceutics and Biopharmaceutics, 43, 173-178.
Cantor, S.L., Hoag, S.W., Augsburger, L.L., 2009. Formulation and characterization of a compacted multiparticulate system for modified release of water-soluble drugs- Part 1 Acetaminophen. Drug Development and Industrial Pharmacy, 35, 337-351.
155
Celik, M., Maganti, L., 1994. Formulation and compaction of microspheres. Drug Development and Industrial Pharmacy, 20, 3151-3173.
Center for Drug Evaluation and Research, 1995. Immediate Release Solid Dosage Forms: Scale-Up and Post Approval Changes (SUPAC-IR): Chemistry, Manufacturing and Controls, In Vitro Dissolution Testing and In Vivo Bioequivalence Documentation.
Center for Drug Evaluation and Research, 1997: Modified Release Solid Oral Dosage Forms: Scale-Up and Post Approval Changes (SUPAC-MR): Chemistry, Manufacturing and Controls, In Vitro Dissolution Testing and In Vivo Bioequivalence Documentation.
Chambin, O., Rota, A., Rochat-Gonthier, M.-H., Pourcelot, Y., 2005. Performance of multilayered particles: Influence of a thin cushioning layer. Drug Development and Industrial Pharmacy, 31, 739-746.
Chopra, R., Alderborn, G., Podczeck, F., Newton, J.M., 2002. The influence of pellet shape and surface properties on the drug release from uncoated and coated pellets. International Journal of Pharmaceutics, 239, 171-178.
Cole, G.C., 1995. Introduction and overview of pharmaceutical coating. In: Cole, G.C. (Ed.), Pharmaceutical coating technology. Taylor & Francis, London, pp. 1-5.
Costa, P., 2001. An alternative method to the evaluation of similarity factor in dissolution testing. International Journal of Pharmaceutics, 220, 77-83.
Costa, P., Sousa Lobo, J.M., 2001. Modeling and comparison of dissolution profiles. European Journal of Pharmaceutical Sciences, 13, 123-133.
Costa, P., Sousa, J.J.S., Pais, A.A.C.C., Formosinho, S.J., 2003. Comparison of dissolution profiles of ibuprofen pellets. Journal of Controlled Release, 89, 199-212.
Dashevsky, A., Kolter, K., Bodmeier, R., 2004. Compression of pellets coated with various aqueous polymer dispersions. International Journal of Pharmaceutics, 279, 19-26.
Debunne, A., Vervaet, C., Mangelings, D., Remon, J.P., 2004. Compaction of enteric-coated pellets: influence of formulation and process parameters on tablet properties and in vivo evaluation. European Journal of Pharmaceutical Sciences, 22, 305-314.
Doelker, E., 1993. Comparative compaction properties of various microcrystalline cellulose types and generic products. Drug Development and Industrial Pharmacy, 19, 2399-2471.
Edge, S., Kibbe, A., Kussendrager, K., 2006. Lactose, monohydrate. In: Rowe, R.C., Sheskey, P.J., Owen, S.C., (Ed.), Handbook of pharmaceutical excipients. Pharmaceutical Press, Washington, DC, pp. 389-395.
El-Gazayerly, O.N., Rakkanka, V., Ayres, J., 2004. Novel chewable sustained-release tablet containing verapamil hydrochloride. Pharmaceutical Development and Technology, 9, 181-188.
156
Esbensen, K.H., Guyot, D., Westad, F., Houmøller, L.P., 2001. Principal component analysis (PCA) – Introduction and in practice. In: Multivariate data analysis – In practice, CAMO ASA, Trondheim, Norway.
European Agency for the Evaluation of Medicinal Products (EMEA), Human Medicines Evaluation Unit, 1999. Note for guidance on quality of modified release products: A: oral dosage forms B: transdermal dosage forms section I (quality), CPMP/QWP/604/96.
Fell, J.T., and Newton, J.M., 1971. Effect of particle size and speed of compaction on density changes in tablets of crystalline and spray-dried lactose. Journal of Pharmaceutical Sciences, 60, 1866-1869.
Ford, J.L., Mitchell, K., Rowe, P., Armstrong, D.J., Elliot, P.N.C., Rostron, C., Hogan, J.E., 1991. Mathematical modeling of drug release from hydroxypropylmethylcellulose matrices: Effect of temperature. International Journal of Pharmaceutics, 71, 95-104.
Ford, J.L., Rubinstein, M.H., Hogan, J.E., 1985. Formulation of sustained release promethazine hydrochloride tablets using hydroxypropylmethylcellulose matrices. International Journal of Pharmaceutics, 24, 327-338.
Habib, Y.S., Augsburger, L.L., Shangraw, R.F., 2002. Production of inert cushioning beads: effect of excipients on the physicomechanical properties of freeze-dried beads containing microcrystalline cellulose produced by extrusion-spheronization. International Journal of Pharmaceutics, 233, 67-83.
Hahm, H.A., Augsburger, L.L., 2008. Orally disintegrating tablets and related formulations. In: Augsburger, L.L., Hoag, S.W. (Ed.), Pharmaceutical dosage forms. Tablets. Volume 2. Rational design and formulation. Informa Healthcare, USA, pp. 293-311.
Hancock, B.C., Rowe, R.C., 1998. The thermal expansion of pharmaceutical solids. S.T.P. Pharma Sciences, 8, 213-220.
Harris, M.R., Ghabre-Sellassie I., 1997. Aqueous polymeric coating for modified release oral dosage forms. In: McGinity, J.W. (Ed.), Aqueous polymeric coatings for pharmaceutical dosage forms. Marcel Dekkar, New York, pp. 81-100.
Haslam, J.L., Forbes, A.E., Rork, G.S., Pipkin, T.L., Slade, D.A., Khossravi, D., 1998. Tableting of controlled release multiparticulates, the effect of millisphere size and protective overcoating. International Journal of Pharmaceutics, 173, 233-242.
Hausner, H.H., 1967. Friction conditions in a mass of metal powder. International Journal of Powder Metallurgy, 3, 7-13.
Heckel, R.W., 1961a. Density-pressure relationships in powder compaction. Transactions of the Metallurgical Society of AIME, 221, 671-675.
Heckel, R.W., 1961b. An analysis of powder compaction phenomena. Transactions of the Metallurgical Society of AIME, 221, 1001-1008.
157
Heng, P.W.S, Chan, L.W., Ong, K.T., 2003. Influence of storage conditions and type of plasticizers on ethylcellulose and acrylate films formed from aqueous dispersions. Journal of Pharmacy and Pharmaceutical Science, 6, 334-344.
Hersey, J.A., Rees, J., 1971. Deformation of particles during briquetting. Nature Physical Science, 230, 96.
Higuchi, T., 1961. Rate of release of medicaments from ointment bases containing drugs in suspension. Journal of Pharmaceutical Sciences, 50, 874-875.
Higuchi, T., 1963. Mechanism of sustained-action medication. Theoretical analysis of rate of release of solid drugs dispersed in solid matrices. Journal of Phamaceutical Sciences, 52, 1145-1149.
Hogan, J.E., 1995. Film-coating materials and their properties. In: Cole, G.C. (Ed.), Pharmaceutical coating technology. Taylor & Francis, London, pp. 6-52.
Holman, .L.E., 1991. The compaction behaviour of particulate materials. An elucidation based on percolation theory. Powder Technology, 66, 265-280.
Huffine, C.L., Bonilla, C.F., 1962. Particle-size effects in the compression of powders. A.I.Ch.E. Journal, 8, 490-493.
Johansson, B., Wikberg, M., Ek, R., Alderborn, G., 1995. Compression behaviour and compactability of microcrystalline cellulose pellets in relationship to their pore structure and mechanical properties. International Journal of Pharmaceutics, 117, 57-73.
Johansson, B., Nicklasson, F., Alderborn, G., 1998. Effect of pellet size on degree of deformation and densification during compression and compactability of microcrystalline cellulose pellets. International Journal of Pharmaceutics, 163, 35-48.
Jones, B.E., 2007. Hard gelatin capsules. In: Aulton, M.E. (Ed.), Aulton’s pharmaceutics: The design and manufacture of medicines. Churchill Livingstone, Philadelphia, pp 515-526.
Kállai, N., Luhn, O., Dredán, J., Kovács, K., Lengyel, M., Antal, I., 2010. Evaluation of drug release from coated pellets based on isomalt, sugar, and microcrystalline cellulose inert cores. AAPS PharmSciTech, 11, 383-391.
Kim, H., Fassihi, R., 1997. Application of binary polymer system in drug release rate modulation. 2. Influence of formulation variables and hydrodynamic conditions on release kinetics. Journal of Pharmaceutical Sciences, 86, 323-328.
Korsmeyer, R.W., Gruny, R., Doelker, E., Buri, P., Peppas, N.A., 1983. Mechanisms of solute release from porous hydrophilic polymers. International Journal of Pharmaceutics, 15, 25-35.
Kühl, P., Mielck, J.B., 2002. Tabletting of pellet-matrix systems: ability of parameters from dynamic and kinetic models to elucidate the densification of matrix formers and of pellets. International Journal of Pharmaceutics, 248, 101-114.
158
Kuny, T., Leuenberger, H., 2003. Compression behaviour of the enzyme β-galactosidase and its mixture with microcrystalline cellulose. International Journal of Pharmaceutics, 260, 137-147.
Landin, M., Martinez-Pacheco, R., Gbmez-Amoza, J.L., Souto, C., Concheiro, A., Rowe, R.C., 1994. Dicalcium phosphate dihydrate for direct compression: Characterization and intermanufacturer variability. International Journal of Pharmaceutics, 109, 1-8.
Lánsky, P., Weiss, M., 1999. Does the dose-solubility ratio affect the mean dissolution time of drugs? Pharmaceutical Research, 16, 1470-1476.
Lehmann, K.O.R., 1997. Chemistry and application properties of polymethacrylate coating systems. In: McGinity, J.W. (Ed.), Aqueous polymeric coatings for pharmaceutical dosage forms. Marcel Dekkar, New York, pp. 101-176.
Leu, R., Leuenberger, H., 1993. The application of percolation theory to the compaction of pharmaceutical powders. International Journal of Pharmaceutics, 90, 213-219.
Leuenberger, H., Leu, R., Bonny, J.D., 1992. Application of percolation theory and fractal geometry to tablet compaction. Drug development and industrial pharmacy, 18, 723-766.
Leuenberger, H., Leu, R., Bonny, J.D, 1996. Application of percolation theory and fractal geometry to tablet compaction. In: Alderborn, G, Nyström, C. (Ed.), Pharmaceutical powder compaction technology. Marcel Dekker, New York, pp 133-164.
Li, S.P., Feld, K.M., Kowarski, C.R., 1997. The effect of polymer coating systems on the preparation, tableting, and dissolution properties of sustained-release drug pellets. Drug Development and Industrial Pharmacy, 23, 623-631.
Liu, J.P., Chow, S.C., 1996. Statistical issues on the FDA conjugated estrogen tablets bioequivalence guidance. Drug Information Journal, 30, 881-889.
Liu, J.P., Ma, M.C., Chow, S.C., 1997. Statistical evaluation of similarity factor f2 as a criterion for assessment of similarity between dissolution profiles. Drug Information Journal, 31, 1255-1271.
Liu, L., Khang, G., Rhee, J.M., Lee, H.B., 2000. Monolithic osmotic tablet system for nifedipine delivery. Journal of Controlled Release, 67, 309-322.
Lundqvist, A.E.K., Podczeck, F, Newton, J.M., 1997. Influence of disintegrant type and proportion on the properties of tablets produced from mixtures of pellets. International Journal of Pharmaceutics, 147, 95-107.
Lundqvist, A.E.K., Podczeck, F, Newton, J.M., 1998. Compaction of, and drug release from, coated drug pellets mixed with other pellets. European Journal of Pharmaceutics and Biopharmaceutics, 46, 369-379.
159
Maarschalk, K.V., Vromans, H., Bolhuis, G.K., Lerk, C.F., 1996. The effect of viscoelasticity and tabletting speed on consolidation and relaxation of a viscoelastic material. Journal of Pharmaceutics and Biopharmaceutics, 42, 49-55.
Maarschalk, K.V., Vromans, H., Groenendijk, W., Bolhuis, G.K., Lerk, C.F., 1997. Effect of water on deformation and bonding of pregelatinized starch compacts. European Journal of Pharmaceutics and Biopharmaceutics, 44, 253-260.
Maganti, L., Celik, M., 1994. Compaction studies on pellets: II. Coated pellets. International Journal of Pharmaceutics, 103, 55-67.
Moore, J.W., Flanner, H.H., 1996. Mathematical comparison of dissolution profiles. Pharmaceutical Technology, 20, 64-74.
Muschert, S., Siepmann, F., Leclercq, B., Carlin, B., Siepmann, J., 2009. Prediction of drug release from ethylcellulose coated pellets. Journal of Controlled Release, 135, 71-79.
Nicklasson, F., Alderborn, G., 1999. Modulation of the tabletting behaviour of microcrystalline cellulose pellets by the incorporation of polyethylene glycol. European Journal of Pharmaceutical Sciences, 9, 57-65.
Nordstrom, J., Klevan, I., Alderborn, G., 2009. A particle rearrangement index based on the Kawakita powder compression equation. Journal of Pharmaceutical Sciences, 98, 1053-1063.
Nokhodchi, A., Rubinstein, M.H., 1998. The compaction properties of mixtures of ibuprofen and hydroxypropylmethyl cellulose. S.T.P. Pharma Sciences, 8, 343-348.
O’Connor, R.E. and Schwartz, J.B., 1993. Drug release mechanism from a microcrystalline cellulose pellet system. Pharmaceutical Research, 10, 356-361.
O’Donnell, P.B., McGinity, J.W., 1996. Mechanical properties of polymeric films prepared from aqueous polymeric dispersions. In: McGinity, J.W. (Ed.), Aqueous polymeric coatings for pharmaceutical dosage forms. Marcel Dekker, U.S., pp. 517-548.
O’Hara, T., Dunne, A., Bulter, J., Devane, J., 1998. A review of methods used to compare dissolution profile data. Pharmaceutical Science & Technology Today, 1, 214-223.
Paronen, P, Ilkka, J., 1996. Porosity-pressure functions. In: Alderborn, G, Nyström, C. (Ed.), Pharmaceutical powder compaction technology. Marcel Dekker, New York, pp 55-76.
Patel, S., Kaushal, A.M., Bansal, A.K., 2007. Effect of particle size and compression force on compaction behaviour and derived mathematical parameters of compressibility. Pharmaceutical Research, 24, 111-124.
Patel, S., Kaushal, A.M., Bansal, A.K., 2010. Mechanistic investigation on pressure dependency of Heckel parameter. International Journal of Pharmaceutics, 389, 66-73.
160
Peppas, N.A., 1985. Analysis of fickian and non-fickian drug release from polymers. Pharmaceutica Acta Helvetiae, 60, 110-111.
Peppas, N.A., Sahlin, J.J., 1989. A simple equation for the description of solute release. III. Coupling of diffusion and relaxation. International Journal of Pharmaceutics, 57, 169-172.
Picker, K.M., 2000. A new theoretical model to characterize the densification behaviour of tableting materials. European Journal of Pharmaceutics and Biopharmaceutics, 49, 267-273.
Picker, K.M., 2004. “Soft tableting”: A new concept to tablet pressure-sensitive materials. Pharmaceutical Development and Technology, 9, 107-121.
Pillay, V., Fassihi, R., 1998. Evaluation and comparison of dissolution data derived from different modified release dosage forms: an alternative method. Journal of Controlled Release, 55, 45-55.
Pinto, J.F., Podczeck, F., Newton, J.M., 1997. The use of statistical moment analysis to elucidate the mechanism of release of a model drug from pellets produced by extrusion and spheronisation. Chemical and Pharmaceutical Bulletin, 45, 171-180.
Podczeck, F., 1993. Comparison of in vitro dissolution profiles by calculating mean dissolution time (MDT) and mean residence time (MRT). International Journal of Pharmaceutics, 97, 93-100.
Polli, J.E., Rekhi, G.S., Augsburger, L.L., Shah, V.P., 1997. Methods to compare dissolution profiles and a rationale for wide dissolution specifications for metoprolol tartrate tablets. Journal of Pharmaceutical Sciences, 86, 690-700.
Porter, S.C., 1997. Use of Opadry, Sureteric, and Surelease for the aqueous film coating of pharmaceutical oral dosage forms. In: McGinity, J.W. (Ed.), Aqueous polymeric coatings for pharmaceutical dosage forms. Marcel Dekker, New York, pp 419-440.
Ragnarsson, G., Sandberg, A., Jonsson, U.E., Sjögren J., 1987. Development of a new controlled release metoprolol product. Drug Development and Industrial Pharmacy, 13, 1495-1509.
Riepma, K.A., Veenstra, J., de Boer, A. H., Zuurman, K., Lerk, C. F., Vromans, H., 1991. Consolidation and compaction of powder mixtures: II. Binary mixtures of different particle size fractions of α-lactose monohydrate. International Journal of Pharmaceutics, 76, 9-15.
Rinaki, E., Dokoumetzidis, A., Macheras, P., 2003a. The mean dissolution time depends on the dose/solubility ratio. Pharmaceutical Research, 20, 406-408.
Rinaki, E., Valsami, G., Macheras, P., 2003b. The power law can describe the ‘entire’ drug release curve from HPMC-based matrix tablets: a hypothesis. International Journal of Pharmaceutics, 255, 199-207.
161
Roberts, R.J. and Rowe, R.C., 1986. The effect of the relationship between punch velocity and particle-size on the compaction behaviour of materials with varying deformation mechanisms. Journal of Pharmacy and Pharmacology, 38, 567-571.
Rowe, R.C., 1980. The expansion and contraction of tablets during film coating- a possible contributory factor in the creation of stresses within the film? Journal of Pharmacy and Pharmacology, 32, 851.
Rowe, R.C., 1997. Defects in aqueous film-coated tablets. In: McGinity, J.W. (Ed.), Aqueous polymeric coatings for pharmaceutical dosage forms. Marcel Dekker, New York, pp 419-440.
Rue, P.J., Rees, J.E., 1978. Limitations of the Heckel relation for predicting powder compaction mechanisms. Journal of Pharmacy and Pharmacology, 30, 642-643.
Salako, M, Podczeck, F., Newton, J.M., 1998. Investigations into the deformability and tensile strength of pellets. International Journal of Pharmaceutics, 168, 49-57.
Santos, H., Veiga, F., Pina, M.E., Sousa, J.J., 2004. Compaction, compression and drug release characteristics of xanthan gum pellets of different compositions. European Journal of Pharmaceutical Sciences, 21, 271-281.
Sarisuta, N., Punpreuk, K., 1994. In vitro properties of film-coated diltiazem hydrochloride pellets compressed into tablets. Journal of Controlled Release, 31, 215-222.
Sawicki, W., Lunio, R., 2005. Compressibility of floating pellets with verapamil hydrochloride coated with dispersion Kollicoat SR 30 D. European Journal of Pharmaceutics and Biopharmaceutics, 60, 153-158.
Schmid, W., Freyer-Picker, K.M., 2009. Tableting and tablet properties of alginates: Characterization and potential for soft tableting. European Journal of Pharmaceutics and Biopharmaceutics, 72, 165-172.
Schultz, P., Kleinebudde, P., 1997. A new multiparticulate delayed release system. Part I: Dissolution properties and release mechanism. Journal of Controlled Release, 47, 181-189.
Shah, U. and Augsburger, L., 2001. Evaluation of the functional equivalence of Crospovidone NF from different sources. I. Physical characterization. Pharmaceutical Development and Technology, 6, 39-51.
Shah, V.P., Tsong, Y., Sathe, P., Liu, J.P., 1998. In vitro dissolution profile comparison- Statistics and analysis of the similarity factor, f2. Pharmaceutical Research, 15, 889-896.
Shangraw, R.F., 1990. Compressed tablets by direct compression. In: Leiberman, H.A., Lachman, L., Schwatz, J.B. (Ed.), Pharmaceutical dosage forms: Tablets. Marcel Dekker, New York, pp. 195–246.
Shimizu, T., Sugaya, M., Nakano, Y., Izutsu, D., Mizukami, Y., Okochi, K., Tabata, T., Hamaguchi, N., Igari, Y., 2003. Formulation study for lansoprazole fast-
162
disintegrating tablet. III. Design of rapidly disintegrating tablets. Chemical and Pharmaceutical Bulletin, 51, 1121-1127.
Sonnergaard, J.M., 1999. A critical evaluation of the Heckel equation. International Journal of Pharmaceutics, 193, 63-71.
Sousa, J.J., Sousa, A., Moura, M.J., Podczeck, F., Newton, J.M., 2002. The influence of core materials and film coating on the drug release from coated pellets. International Journal of Pharmaceutics, 233, 111-122.
Stauffer, D., Aharony, A., 1992. Introduction: Forest fires, fractal oil fields, and diffusion. In: Introduction to percolation theory. Taylor & Francis, Britain, pp1-14.
Tang, L., Schwartz, J.B., Porter, S.C., Schnaare, R.L., Wigent, R.J., 2000. Drug release from film-coated chlorpheniramine maleate nonpareil beads: effect of water-soluble polymer, coating level, and soluble core material. Pharmaceutical Development and Technology, 5, 383-390.
Tang, E.S.K., Chan, L.W., Heng, P.W.S., 2005. Coating of multiparticulates for sustained release. American Journal of Drug Delivery, 3, 17-28.
Tirkkonen, S., Paronen, P., 1993. Release of indomethacin from tabletted ethylcellulose microcapsules. International Journal of Pharmaceutics, 92, 55-62.
Torrado, J.J., Augsburger, L.L., 1994. Effect of different excipients on the tableting of coated particles. International Journal of Pharmaceutics, 106, 149-155.
Torrado-Duran, J.J., Torrada-Valeiras, J.J., Cadorniga, R., 1991. Microaggregated egg albumin particles containing paracetamol for tableting processes. Drug Development and Industrial Pharmacy, 17, 1305-1323.
Train, D., 1956. An investigation into the compaction of powders. Journal of Pharmacy and Pharmacology, 8, 745-761.
Tunon, A., Borjesson, E., Frenning, G., Alderborn, G., 2003b. Drug release from reservoir pellets compacted with some excipients of different physical properties. European Journal of Pharmaceutical Sciences, 20, 469-479.
Tunon, A., Grasjo, J., Alderbron, G., 2003a. Effect of intragranular porosity on compression behaviour of and drug release from reservoir pellets. European Journal of Pharmaceutical Sciences, 19, 333-344.
Türkoğlu, M., Varol, H., Celikok, M., 2004. Tableting and stability evaluation of enteric-coated omeprazole pellets. European Journal of Pharmaceutics and Biopharmaceutics, 57, 279-286.
Twitchell, A.M., 2007. Mixing. In: Aulton, M.E. (Ed.), Aulton’s pharmaceutics: The design and manufacture of medicines. Churchill L ivingstone, U.S., pp 152-167.
Tye, C.K., Sun, C.Q., Amidon, G.E., 2005. Evaluation of the effects of tableting speed on the relationships between compaction pressure, tablet tensile strength, and tablet solid fraction. Journal of Pharmaceutical Sciences, 94, 465-472.
163
Veeco Metrology Group, 1999. Surface parameter. In: Wyko surface profilers technical reference manual. U.S., pp 2-1 - 2-24.
Vergote, G.J., Kiekens, F., Vervaet, C., Remon., J.P., 2002. Wax beads as cushioning agents during the compression of coated diltiazem pellets. European Journal of Pharmaceutical Sciences, 17, 145-151.
Vromans, H., DeBoer, D., Bolhuis, G.K., Kussendrager, K.D., Bosch, H., 1985. Studies on tableting properties of lactose. Part 2. Consolidation and compaction of different types of crystalline lactose. Pharmaceutisch Weekblad Scientific Edition, 7, 186.
Wagner, K.G., Krumme, M., Schmidt, P.C., 1999. Investigation of the pellet-distribution in single tablets via image analysis. European Journal of Pharmaceutics and Biopharmaceutics, 47, 79-85.
Wagner, K.G., Krumme, M., Schmidt, P.C., 2000a. Pellet-containing tablets. Examination of distribution and deformation behaviour. S.T.P. Pharma Sciences, 10, 327-332.
Wagner, K.G., Krumme, M., Beckert, T.E., Schmidt, P.C., 2000b. Development of disintegrating multiple-unit tablets on a high-speed rotary tablet press. European Journal of Pharmaceutics and Biopharmaceutics, 50, 285-291.
Walker, E.E., 1923. The properties of powders. Part VI. The compressibility of powders. Transactions of the Faraday Society, 19, 73-82.
Westermarck, S., Juppo, A.M., Kervinen, L., Ylirrusi, J., 1999. Microcrystalline cellulose and its microstructure in pharmaceutical processing. European Journal of Pharmaceutics and Biopharmaceutics, 48, 199-206.
Wheatley, T.A., Steuernagel, C.R., 1997. Latex emulsions for controlled drug delivery. In: McGinity, J.W. (Ed.), Aqueous polymeric coatings for pharmaceutical dosage forms. Marcel Dekkar, New York, pp 1-54.
Wikipedia contributors, Random close pack. In: Wikipedia, The Free Encyclopedia. http://en.wikipedia.org/wiki/Random_close_pack (accessed March 10, 2008)
Yang, L., Fassihi, R., 1995. Zero-order release kinetics from a self-correcting floatable asymmetric configuration drug delivery system. Journal of Pharmaceutical Sciences, 85, 170-173.
Yao, T., Yamadam, M., Yamahara, H., Yoshida, M., 1997. Tableting of coated particles. I. Small particle size chitosan as an agent protecting coating membrane from mechanical damage of compression force. Chemical and Pharmaceutical Bulletin, 45, 1510-1514.
Yao, T., Yamadam, M., Yamahara, H., Yoshida, M., 1998. Tableting of coated particles. II. Influence of particle size of pharmaceutical additives on protection of coating membrane from mechanical damage during compression process. Chemical and Pharmaceutical Bulletin, 46, 826-830.
164
York, P., 1978. Particle slippage and rearrangement during compression of pharmaceutical powders. Journal of Pharmacy and Pharmacology, 1978, 6-10.
York, P., 1979. A consideration of experimental variables in the analysis of powder compaction behaviour. Journal of Pharmacy and Pharmacology, 31, 244-246.
Yuasa, H., Takashima, Y., Omata, T., Kanaya, Y., 2001. Studies on stress dispersion in tablets. III. Suppression of fracture of coated film by an excipient during the preparation of tablets containing coated particle. S.T.P Pharma Sciences, 11, 221-227.
165
LIST OF PUBLICATIONS
Poster presentations
• W. C. Chin, L. W. Chan, P. W. S. Heng, A study on the phases of pellet coat damage during compression, AAPS Annual Meeting and Exposition, Washington D.C., USA, 23-27 October 2011
• W. C. Chin, L. W. Chan, P. W. S. Heng, Utilization of micronized lactose in reducing the extent of pellet coat damage during compression, AAPS Annual Meeting and Exposition, Los Angeles California, USA, 8 - 12 November 2009
• W. C. Chin, L. W. Chan, P. W. S. Heng, Exploring methods to reduce damage to pellet coat under compression, Coating Workshop, Lille, France, 11 September 2008
• W. C. Chin, L. W. Chan, P. W. S. Heng, Influence of diluent particle size and compression force on coat integrity of pellets, Asian Association of Schools of Pharmacy Conference, Makati City, Philippines, October 25 - 28, 2007
Journal publication
• X., Lin, W.C. Chin, K. Ruan, Y. Feng, P.W.S. Heng. Development of potential novel cushioning agents for the compaction of multi-particulates by co-processing micronized lactose with polymers. European Journal of Pharmaceutics and Biopharmaceutics (article in press).
Oral presentation
• W.C. Chin, L.W. Chan, P.W.S. Heng, Influence of pellet core properties on the extent of coat damage during compression, PharmSciFair, Prague, Czech, 13-17 June 2011
