International Journal of Pharmaceutics 511 (2016) 659668

Contents lists available at ScienceDirect

International Journal of Pharmaceutics

journal homepage: www.elsevier.com/locate/ijpharm

Achieving a robust drug release from extended release tablets using an

integrated continuous mixing and direct compression line
Satu Lakioa,* , Pirjo Tajarobia , Hkan Wikstrma , Magnus Franssona , Johan Arneheda ,
Tuomas Ervastib , Simo-Pekka Simonahob , Jarkko Ketolainenb , Staffan Folestada ,
Susanna Abrahmsn-Alamia
a
AstraZeneca Gothenburg, Pepparedsleden 1, SE-43183 Mlndal, Sweden
b
School of Pharmacy, Yliopistonranta 1, University of Eastern Finland, FI-70211, Finland

A R T I C L E I N F O A B S T R A C T

Article history:
Received 14 June 2016 In the present work the viability of integrated continuous mixing and compression processes for
Received in revised form 20 July 2016 manufacturing of extended release (ER) matrix tablets was investigated in terms of dissolution behavior.
Accepted 23 July 2016 The purpose was also to evaluate the combined effect of processing variables and compositional variables
Available online 26 July 2016 on the release robustness. The continuous process was provoked by a challenging formulation design,
including variable powder characteristics and compositions of high and low amount of poorly soluble and
Keywords: poorly owing drug substance (ibuprofen). Additionally a relatively low amount of two different ER
Continuous direct compression matrix former grades (standard granulation grade CR and direct compression grade DC2 of
Continuous mixing
hydroxypropyl methylcellulose, HPMC) was used to challenge the system. Robust ibuprofen release
Dissolution
was obtained faster when HPMC CR was used. However, robust release was also achieved when using
Extended release
Hydroxypropyl methylcellulose HPMC DC2 at high ibuprofen content, even though it took slightly longer time to reach the steady state of
Percolation the process. Due to its poor ow properties, HPMC CR would be very challenging to use in traditional
Robust drug release direct compression. The results showed that by using continuous processing it is possible to manufacture
and achieve robust performance of compositions that would not be possible with traditional batch
processing due to for instance poorly owability.
2016 Elsevier B.V. All rights reserved.

1. Introduction
The benets of continuous pharmaceutical manufacturing of
oral solid dosage forms (OSD) have been discussed a lot in the
literature (Vervaet and Remon, 2005; Schaber et al., 2011; Yin and
Clayton, 2014). Some of these advantages include lower processing
costs, signicant time savings, no need for traditional scale-up,
more compact manufacturing facilities, as well as more exible
output capacity. When considering continuous mixing and direct
compression some additional advantages can be also stated,
namely better control over the mixing process, no discharge issues
such as ratholing and no need for sampling to determine the end
point. In pharmaceutical industry, adaption of new manufacturing
technologies has generally been rather slow. There are many
reasons for this slow manufacturing evolution such as poor
intellectual property incentives, regulatory policies, lack of proven
* Corresponding author.
E-mail address: satu.lakio@astrazeneca.com (S. Lakio).
off-the-shelf equipment, and so on. For further background, see the
discussions by Plumb (2005), Nicholson (2014), Lee et al. (2015),
and references therein. The slow adaptation of new technologies is
therefore reected in that batchwise processes still remain as the
dominating platform for OSD drug product manufacture.
Manufacturing of extended release matrix tablets can be a
challenging task in both batch and continuous processing and goes
beyond manufacturing of immediate release (IR) OSD. The
performance of tablets based on this formulation principle relies
not only on assuring assay and proper distribution of the drug
substance but also on the ability to provide extended and robust
drug release in vivo. The latter depends mainly on the polymeric
matrix and its distribution in the composition, which also needs to
be considered during continuous powder processing.
For hydrophilic extended release tablets based on hydroxyl-
propyl methylcellulose (HPMC) continuous direct compression
may provide additional opportunities but is also associated with
some challenges. Historically wet granulation has been the most
common manufacturing method for HPMC based formulations,
but when using water as granulation liquid there is a challenge due

http://dx.doi.org/10.1016/j.ijpharm.2016.07.052
0378-5173/ 2016 Elsevier B.V. All rights reserved.
660 S. Lakio et al. / International Journal of Pharmaceutics 511 (2016) 659668

to hard lump formation and associated poor compaction perfor- (IbuPS), ibuprofen load (Ibu%) and mixer speed. The design was a
mance. Instead direct compression may be used to avoid the issues full factorial with 19 runs (Table 1). However, low IbuPS and high
with water granulation. However, commercially available HPMC Ibu% combination was excluded due to extremely poor ow.
grades are mostly designed for wet granulation and are normally Modde 10 (Umetrics MKS AB, Ume, Sweden) was used both to
not suitable for traditional batch mixing followed by direct construct the experimental design and for evaluation of the
compression due to a small particle size and therefore poor powder dissolution data. The model was tted with multiple linear
ow characteristics. However, directly compressible grades of regression (MLR).
HPMC have recently become available and their performance has In all experiments, the concentrations of HPMC and PRUV
been investigated in several papers (Mohamed et al., 2013; Rogers were constant (32% and 2% w/w, respectively), the ibuprofen
et al., 2013; Heiman et al., 2014). In spite of some investigations on concentration was changed within three levels (2%, 15% or 22%
direct compression of hydrophilic matrices being reported w/w) as well as the concentration of mannitol (44%, 51% or 64%,
previously, to the best of our knowledge, integrated continuous w/w). The total feed rate of powders was kept constant at
manufacturing using continuous dry powder mixing and continu- 3.5 kg/h.
ous direct compression has only be published in our previous
paper (Ervasti et al., 2015). In that study, we demonstrated that
2.2. Process description
extended release matrix tablets can be successfully manufactured
using an integrated continuous mixing and direct compression
The continuous mixing and direct compression manufacturing
manufacturing line. In fact dissolution behavior of tablets
set-up used in the study is described in Fig. 1A. The feeding, mixing,
manufactured using continuous direct compression in general
and compressing parts are all integrated into a complete
has not been studied extensively before. Jrvinen et al. (2013)
continuous tablet manufacturing line. Three kinds of LIW feeders
measured paracetamol release after continuous direct compres-
(K-Tron, Types K-ML-D5-KT20 and K-CL-SFS-KT20 (two different
sion but their formulation contained only MCC and paracetamol
versions), Niederlenz, Switzerland) were used in the study. K-ML-
and thus the release was found to be driven by the compression
D5-KT20 fed HPMC and mannitol an also ibuprofen with high
force rather than disintegrant activity in the formulation. Thus API
content (1522%) and K-CL-SFS-KT20 fed PRUV and ibuprofen with
was released via disintegration rather than dissolution.
low content (2%). Raw materials were fed from the feeders directly
In the present work, the viability of integrated continuous
into the continuous mixer (Modulomix, Hosokawa Micron,
mixing and compression processes for manufacturing of extended
Doetinchem, the Netherlands). After mixing, the powder mixture
release matrix tablets was investigated mainly in terms of
was guided (ow controlled by gravity) into the hopper of the
dissolution behavior. The continuous process was challenged by
tablet press (PTK-PR1000, PTK CO., Ltd, Incheon, Republic of Korea).
a formulation design, including variable powder characteristics
This very simple set-up enabled conducting experiments without
and compositions of high and low amount of the drug substance
any conveyors, thus reducing the risk of segregation due to
(ibuprofen). Additionally a relatively low amount (32%, w/w) of the
transport. The feeders and the mixer were adjusted and controlled
extended release matrix former (HPMC) was used in order to
by an in-house software, and data from the feeders were acquired
maximize the amount of ller (mannitol) for better compaction
for later analysis. The tablets were compacted with a turret speed
and powder ow. At these low amounts of matrix former there is a
of 48 rpm using eight 7 mm concave punches to produce tablets
known risk that formulations may show poor release robustness
with a target weight of 150 mg. Compression force between 5 and
(Tajarobi et al., 2009). This has been shown to be even more
9 kN was used in this study. The detailed description about the
challenging when using the larger particle sized HPMC matrix
continuous line (including feeders) and used parameters can be
formers developed to be more suitable for direct compression
found in Ervasti et al. (2015).
(Heiman et al., 2014). Thereby the opportunity to evaluate the
The timeline of the process is described in Fig. 1B. To start
process using several different performance aspects was enabled.
processing, the feeders and the mixer were rst turned on (time
In addition, the design of present work gave the opportunity to
point 0) and the powder mixture was collected into the hopper of
evaluate the combined effect of continuous processing variables
the tablet press for 12 min to serve adequate ll level for the press.
and compositional variables on the release robustness.

2. Materials and methods

Table 1
D-optimal experimental design worksheet.
Two different grades of hydroxypropyl methylcellulose (HPMC)
Run Speed (rpm) HPMC PS (mm) Ibu PS (mm) Ibu (%)
were used as matrix formers: standard wet granulation grade,
Methocel K100 Premium LV CR (CR, Dow Chemical Company, N1 300 77 30 2
N2 1200 77 30 2
Midland, Michigan, USA) and a new direct compressible grade,
N3 300 120 30 2
Methocel K100 Premium LV DC2 (DC2, Dow Europe GmbH, N4 1200 120 30 2
Bomlitz, Germany). For the centre points 50:50% (w/w) blend of CR N5 300 77 66 2
and DC2 was manufactured. The model drug substance was N6 1200 77 66 2
ibuprofen (Zhengzhou Debao Fine Chemical Co. Ltd, Henan, China). N7 300 120 66 2
N8 1200 120 66 2
Two different ibuprofen particles sizes (IbuPS) were used: 1) as N9 300 77 66 22
received from the supplier (large IbuPS), and 2) small IbuPS N10 1200 77 66 22
prepared by ball milling (Retch S1, Haan, Germany). Mannitol N11 300 120 66 22
(Parteck M200, Merck KGaA, Darmstadt, Germany) was added as a N12 1200 120 66 22
N13 300 77 30 15
soluble ller and sodium stearyl fumarate (PRUV, Moehs,
N14 1200 77 49 22
Barcelona, Spain) as a lubricant. N15 300 120 49 22
N16 1200 120 30 15
2.1. Design of experiment N17 750 100a 49 15
N18 750 100a 49 15
N19 750 100a 49 15
The study was performed as an experimental design with four
factors: HPMC particle size (HPMC PS), ibuprofen particle size a
A 50:50% (w/w) mixture of HPMC CR and DC2 was used in center points.
 S. Lakio et al. / International Journal of Pharmaceutics 511 (2016) 659668
Fig. 1. The direct compression continuous manufacturing set-up used in the experiments. A. The whole continuous manufacturing line. B. A timeline describing the process
phases.
The complete continuous processing was started at 12 min when
tablet press was turned on. The duration of each run was 32 min
(12 min hopper lling + 20 min continuous direct compression).
In addition to continuous manufacturing some tablets were also
compressed to study the effect of compression force on the release
behavior. The aim was to conrm that the observed burst release
was not an effect of varying compaction forces. The powder
mixtures were taken out after 20 min continuous mixing and then
compacted at two levels of compaction pressures (low: 34 kN and
high: 68 kN). Eccentric tableting machine Kilian SP300 (Kilian &
Co GmbH, Cologne, Germany) was used to compress tablets. Tablet
properties were kept the same as in continuous line, i.e. 7 mm
concave punches and a tablet weight of 150 mg. Wintab 2
(AstraZeneca in house) was used to control the forces and to
collect data.
2.3. Tablet analysis
The weight and tensile strength of the tablets were measured
(n = 10) with a MultiCheck Turbo III (Erweka GmbH, Heusen-
stamm, Germany). Using this data tablet weight variations were
calculated.
2.4. Assay and content uniformity
Tablets from all batches were sampled every two minutes from
the start of the tablet press with a sampling time of 1020 s. An
Accura transmission Raman system (Horiba Scientic, Lille, France)
equipped with a 785 nm diode laser was used to measure
ibuprofen content (n = 3). The spectral range was 135
2240 cm 1, and all tablets were measured using 4 acquisitions
with a 60 s integration time, i.e. a total acquisition time of 4 min.
The Raman measurements and assay determination are described
in more detail in Ervasti et al. (2015). The predicted content in% (w/
w) ibuprofen was converted into% label claim (LC) by multiplying
with the individual tablet weight and dividing by the nominal
content in mg. Assay was calculated for each batch and process
time point (every two min) as the average of the three predicted
tablet contents. Content uniformity was estimated by pooling
tablets from the four last time points and then calculating the
relative standard deviation (RSD). By pooling the contents from the
four last time points, all in the steady-state region, n was increased
to 12, which is a reasonable number considering that n = 10 is often
used for regular content uniformity analyses.
661
2.5. Dissolution
Dissolution testing of drug release from tablets was carried out
in a USP dissolution apparatus II (Hanson research SR8-Plus,
Chatsworth, CA, USA) using ultraviolet (UV) detection with bre
optics (Varian Cary 50 Bio, C Technologies Inc., Bridgewater, NJ,
USA). A bre-optic probe was inserted directly into each vessel and
the UV absorption measured in situ. The USP II method was
modied by incorporating a quadrangular stationary basket above
the paddles, where tablets were placed. Dissolution testing was
done at 37  C in 1000 ml of phosphate buffer (pH 7.4) (I = 0.1) or in
1000 ml of 0.1 M HCl (pH 1.0) with a paddle speed of 50 rpm.
Dissolution proles were adjusted to 0% released at time point 0,
noise-ltered (Savitsky-Golay, order 0 and window 5) and then
normalized to 100% released based on the level at the end of the
dissolution. The number of individual proles (measured tablets)
was 13, depending data available for every batch. The average
values at 32 min process time point (after 20 min tableting) was
used in the modeling. The time to reach 80% released (T80) was
linearly interpolated between data points encompassing 80%
released.
2.6. Confocal laser scanning microscopy
The distribution of the matrix former (HPMC) in N6 (HPMC
grade CR) and N15 (HPMC grade DC2) after 32 min processing (i.e.
after 20 min tableting) was investigated using confocal laser
scanning microscopy (CLSM). The CLSM system consisted of a
Nikon D-ECLIPSE C1 confocal system, a Nikon Eclipse TE2000-E
inverted microscope and a Nikon Plan Apochromat 10.0x/0.45/4.00
dry objective. The wavelength of the excitation laser was 408 nm
and captured channels were: blue 450/35 nm, green 515/30 nm
and red 605/75 nm. Raman imaging was then used to assign tablet
components to the CLSM data.
In order to expose a cross-section near the tablet centre for
CLSM and Raman imaging, approximately half of the tablet was
milled away using a tablet mill. The resulting surface was nished
using a microtome with glass knives followed by diamond knives
to progressively remove excess material and smooth the surface.
This concluded that the area imaged by CLSM and Raman was a
cross-section near the tablet centre.
The CLSM images were processed using Fiji/ImageJ (Schindelin
et al., 2012; Schneider et al., 2012). At rst individual overlapping
CLSM images (68/tablet) were joined by stitching. Then
thresholding was performed to separate ibuprofen, and HPMC
662 S. Lakio et al. / International Journal of Pharmaceutics 511 (2016) 659668
from mannitol and cavities using the Color Threshold function in
FIJI/ImageJ. Thereafter color thresholding was used to separate
ibuprofen and HPMC and blue pixels were counted to be ibuprofen.
Hereafter only the HPMC component was analyzed. Images were
centered and then cropped to be 6.5  2.5 mm and watershed
segmenting was performed for HPMC domains. Finally, distribu-
tion analysis using BioVoxxel toolkit (Brocher, 2015) was executed,
excluding particles that were cut from the edges of the image, and
then the relative nearest neighbor distance from watershed
segmented data was calculated.
3. Results and discussion
A more detailed analysis related to feeders, tableting process
and tablet properties can be found in our previous paper (Ervasti
et al., 2015). Since this paper focuses mainly on the release
performance of extended release tablets through continuous
mixing and direct compression manufacturing other aspects are
discussed only shortly.
3.1. Feeder data
It is crucial to ensure that the feed rate of each bulk material is
controlled accurately during continuous powder mixing and direct
compression (Marikh et al., 2005; Pernenkil and Cooney, 2006;
Portillo et al., 2008). In addition the powder properties are known
to affect the time to reach the steady state. In this study it was
found that using high dose ibuprofen feeding (15% and 22% w/w)
the settling time for the feeders was in all cases less than 5 min. At
low dose ibuprofen feeding (2% w/w), however, the settling time
was notably longer (1017 min). This is consistent with observa-
tions by Besenhard et al. (2016) and Llusa et al. (2016). The
inuence of powder/wall interaction increases due to higher
surface-to-volume ratio.
When studying the dissolution behavior of HPMC matrix tablets
it is important to analyze the feeder data of HPMC in addition to the
API feeder data. There was no signicant difference between CR
and DC2 regarding feeding behavior. When studying feeding of
HPMC more carefully it was noticed that when there was more
material (>4 kg) in the feeder the frequency of variance in the feed
rate was higher. However the amplitude of the variance
(approximately 1.7%) was similar for all runs regardless of the
lling degree of the feeder and HPMC grade. In addition the HPMC
feeders reached the steady state only a few seconds after starting
the feeding.
3.2. Tablet analysis
Tablet weight was stable for all batches within 45 min after the
start of the tableting. The weight of tablets manufactured from CR
HPMC grade was generally higher. With DC2 HPMC grade the tablet
weight was lower due to larger particle size and therefore
incomplete lling of the die but the stability was overall better.
It was also noted that higher mixer speed resulted in higher tablet
weights due to improved ow properties, thus improved die lling.
The tablet weight variation was low within each run; the difference
between minimum and maximum weights was always less than
7.3 mg (mean 3 mg) during steady state. Notably, it was still
possible to manufacture tablets from powder mixtures with
relatively poor owability (e.g. N9, N10, N13 and N14).
Generally, the tensile strength of HPMC CR grade tablets was
higher. The combination of low Ibu% and the use of HPMC CR grade
gave the highest tensile strength tablets. With HPMC DC2 grade,
the tablet tensile strengths were overall lower. There was more
variation over time in HPMC CR tensile strengths compared to
HPMC DC2. A general conclusion in relation to all tablets
manufactured was that the changes in the tensile strength seen
throughout the trials were not relevant for the overall tablet
performance, since the tensile strength was high enough to
withstand handling of all runs.
3.3. Ibuprofen release performance at pH 7.4
In line with previous results, when using HPMC CR more robust
and sustained release behavior for low dose compositions was
achieved (Tajarobi et al., 2009; Heiman et al., 2014). After reaching
the steady state (after 20 min tableting), compositions containing
large amounts of ibuprofen (22% w/w) gave a relatively robust and
sustained release independent of mixer speed and particle sizes of
HPMC and ibuprofen (Fig. 2). The dissolution curves for these
compositions were grouped together and showed the slowest
release performance. The release behavior of the low dose
compositions containing HPMC CR had a slightly faster release
than the corresponding high dose compositions, e.g. the interme-
diate dissolution proles of Fig. 2. The reason for this was most
probably the relatively higher amount of easily soluble ller in the
low dose compositions resulting in faster hydration and erosion of
the matrix (Tajarobi et al., 2009). The release behavior for low dose
compositions containing HPMC DC2 (N3, N4, N7 and N8) at pH 7.4,
on the other hand, was non-robust and showed a very rapid burst
like release, e.g. the most rapid release proles of Fig. 2. For the
high dose compositions containing HPMC DC2 the release was
slower and more robust than the corresponding low dose
compositions. The difference in behavior observed was also
conrmed through the MLR model which indicated that the time
to reach 80% release (T80) was shorter when DC2 was used, e.g.
faster release (Fig. 3). The MLR model also supported the
observation that the effect of HPMC PS was more pronounced
when using lower Ibu%. R2 and Q2 for the MLR model were 0.987
and 0.972, respectively, indicative of an excellent model.
Ibuprofen (as well as mannitol) can also have a negative effect
on gelling of the HPMC which can lead to poor release robustness.
Lapidus and Lordi (1968) showed that drug release can be
inuenced by the ionic content of the dissolution medium, and
suggested that inorganic ions with a high afnity for water could
dehydrate and result in a salting out of the polymer. Also the drug
itself can provide such salting out of the HPMC affecting its
hydration and is manifested in a lowering of the cloud point of
HPMC (Ridell et al., 1999). Such potential weakening of the gel
structure has obvious implications for drug diffusion and gel
strength (Mitchell et al., 1993a, 1993b).
Fig. 2. Ibuprofen release proles for individual tablets after 20 min tableting. Solid
lines represent ibuprofen content of 22% and dashes lines 2%. Lines with ball marks
represent HPMC CR grade.
 S. Lakio et al. / International Journal of Pharmaceutics 511 (2016) 659668
Fig. 3. A. Coefcient plot for time to reach 80% release (T80) at pH 7.4. B. Interaction effects plot (at pH 7.4) showing the effect of HPMC particle size (HPMC PS) and ibuprofen
content (Ibu%) on T80 during dissolution study.
3.3.1. The effect of mechanical properties of the tablets on release
performance
The tensile strengths and also the weights of tablets composed
of DC2 were generally lower than those of tablets containing CR
grade. Theoretically the release from low dose tablets containing
HPMC DC2 could be faster than HPMC CR tablets. However, the
variation in weight of these tablets was very small and could not
explain the dramatic change in release behavior we observe for the
two grades (Fig. 2), rather the same absolute rate (weight per time
unit) of drug release normalized to the tablet surface area would be
expected (Siepmann et al., 2002). Also, in order to conrm that the
burst release observed for low dose compositions containing DC2
was not an effect of varying compaction forces, powder mixtures
for some compositions were taken out after 32 min processing and
compacted at two compaction force levels (low: 34 kN and high:
68 kN). Release data for non-bursting tablets (N6 with HPMC CR)
were compared with that of bursting tablets (N4 with HPMC DC2)
(Fig. 4). The results conrmed that the compaction pressure had a
minor effect on release and that the bursting release was not an
effect of variable porosity or tensile strength of the tablet. Hence,
the difference between the low dose compositions was purely
Fig. 4. Effect of compaction force on bursting (N4) and non-bursting (N6) low dose
ibuprofen compositions (2% w/w). Open markers represent low compaction force
and lled markers high compaction force.
663
related to the compositional aspects and that the bursting
compositions contained HPMC DC2 that had a larger particle size.
3.3.2. Compositional effects on release performance
The tendency for faster release from tablets containing DC2 was
most probably related to a less effective ability to form a
homogeneous gel layer at the tablet/dissolution media interface.
For the low dose compositions this effect was aggravated by the
relatively high amount of easily soluble mannitol and hence
burst release was observed. Similar observations have been made
previously for HPMC DC grades at compositions where the amount
of HPMC was close to the percolation threshold (Heiman et al.,
2014). The review of Carabello (2010) nicely summarizes factors
affecting drug release from HPMC matrices in the light of the
percolation theory. A general observation made in the review is
that higher polymer percolation thresholds will be obtained for
higher relative particle sizes of the matrix forming polymer, owing
to the lower ability of the coarser particles to percolate the matrix.
The exact position of the polymer percolation threshold also
depends on the components in the formulation (llers, drug
substance etc.). Normally, the volume fraction of HPMC is a more
relevant characteristic to consider than the polymer content.
According to the literature (Rogers et al., 2013) the densities of the
DC2 and CR grades are similar. However, to make a fair comparison,
the volume fraction of the swelled HPMC in the tablet at a certain
time point very early in the dissolution process where a percolating
gel is formed, should be considered. With larger sized HPMC
particles the total surface area available for hydration is lower than
that of small particle sized HPMC at a xed polymer amount.
Hence, a faster swelling and therefore a lower percolation
threshold is expected for HPMC CR than for HPMC DC2. To
determine the percolation threshold of HPMC CR and DC2 at pH
7.4, high ibuprofen dose tablets with varying amounts of HPMCs
were manufactured in a batch mode. The amount of ibuprofen and
PRUV were kept constant (22% and 2%, respectively) and content of
HPMC was varied from 10 to 70% (w/w) and therefore also content
of mannitol was varied accordingly (666%, w/w). It was noticed
that the percolation threshold, the HPMC concentration below
which tablet burst was observed, for HPMC CR was lower (between
20 and 30%) than for HPMC DC2 (between 30 and 40%) (Fig. 5).
Hence, for the compositions investigated in the continuous
manufacturing line the tablets with HPMC CR were well above
the percolation threshold whereas those with HPMC DC2 were
very close to it.
664 S. Lakio et al. / International Journal of Pharmaceutics 511 (2016) 659668
Fig. 5. Determination of percolation threshold for HPMC CR (2030%) and DC2 (30
40%) at pH 7.4.
3.3.3. Confocal laser scanning microscopy
Confocal laser scanning microscopy (CLSM) combined with
Raman imaging of dry tablets was used to gain a better
understanding of the HPMC distribution. As the measurements
were performed in dry at state the HPMC was not swelled and
gelled to the same extent as it was in wetted state during
dissolution measurements. Even so, image analysis in the dry state
can provide guidance to characteristics for the hydrated state.
Raman imaging enabled the separation of HPMC from other
components (ibuprofen and mannitol) and, hence, helped in
assigning all tablet components for the CLSM data. HPMC CR (in
N6) was found to be better dispersed in tablets after 32 min
processing than HPMC DC2 (in N15) (Fig. 6). This observation was
also conrmed by calculating the relative nearest neighbor
distance values (RNND). RNND after 32 min processing was 1.91
for N6 and 1.08 for N15. Larger RNND value referred to better
dispersion of HPMC. This result showed that the CLSM and Raman
methods can be valuable tool for studying HPMC distribution in
matrix.
3.3.4. Effects of processing time on release performance at pH 7.4
In the previous section we conrmed that different release
proles can robustly be generated by composition. Another
interesting aspect in the continuous mixing and direct compres-
sion is the processing time and its inuence on the release
performance.
The processing time was found to have an effect on the release
rate for most of the runs at pH 7.4 (Fig. 7). Tablets collected at early
processing times released ibuprofen markedly faster than tablets
collected at later time points and the variability in release between
individual tablets was higher at earlier time points. This effect was
more pronounced when HPMC DC2 was used (Fig. 7). Thus, under
Fig. 6. HMPC particles in CLSM gures in white A. N6 after 20 min and B. N15 after 20 min tableting. Analyzed area 6.5  2.5 mm.
these poor mixing conditions HPMC was not capable to form a
strong homogeneous gel layer at the tablet/dissolution media
interface fast enough to prevent burst at the amount used in this
study (32% w/w). This effect observed was most probably not an
effect of the different porosity or density of the tablets since these
factors have been found to have very low effect on the release from
HPMC based matrices, especially at the relatively low drug
loadings of the present study (Timmins et al., 1992; Velasco
et al., 1999). Also, the absence of compaction pressure effect on
release shown in Fig. 4 supports this conclusion.
The effect of processing time on release proles of N5, N6, N11,
N12, N14 and N15 was studied in more detail (Fig. 7). These runs
were selected since they could enlighten the effect of mixing
speed, different HPMC grade and ibuprofen load on release
performance.
Interestingly, the processing time did not have a signicant
effect on the dissolution behavior when a combination of CR grade
and high mixing speed was used apart from the rst time point of
the low dose composition (4 min after starting the tableting)
(Figs. 7B and E). At low mixing speed, though, for the same HPMC
grade some small changes over time were observed and a steady
state reached after about 10 min tableting (Fig. 7A). The effect of
processing time was more pronounced when HPMC DC2 was used
(Figs. 7C, D and F). At high mixing speed, though, tablets with
HPMC DC2 grade (run N12) stabilized earlier giving relatively
robust dissolution performance already after 6 min tableting
(Fig. 7D). At low mixing speed for the same composition (run
N11) the release was not found robust until at 32 min processing
time (e.g. after 20 min tableting) (Fig. 7C) even though compaction
properties and assay stabilized much earlier. The time dependence
of some compositions was most likely linked to uneven distribu-
tion of matrix former in the beginning of the processing.
The assay was robust despite of material properties and process
parameters and for most of the runs the assay stabilized relatively
fast to an acceptable level. For N11 there was no signicant
variability in assay over time and the values were never larger that
10% of the label claim. Even so, the dissolution behavior of N11
was time dependent (Fig. 7C). It should be also noted that only
assays of N5 and N6 had more than 10% variation from the label
claim. A slightly higher mean value of assay over time was
observed for these samples but only for N5 a clear decrease of assay
over time was observed. It is important to notice that, in spite of
this, the release rate varied the least over time with these
compositions. Hence the variation of the release performance over
time was not connected to variation in assay.
3.4. Ibuprofen release performance at pH 1
The release behavior of ibuprofen was different in pH 1 than in
pH 7.4 (Fig. 8). The release was signicantly slower at low pH and
the robustness was signicantly better, e.g. the variability with
time was lower than at pH 7.4. Ibuprofen is an acid and is therefore
in its neutral state at pH 1 which leads to lower solubility. A low
solubility drug formulated in an extended release matrix
 S. Lakio et al. / International Journal of Pharmaceutics 511 (2016) 659668 665

Fig. 7. Release proles of individual tablets collected at different tableting time points. A. N5, B. N6, C. N11, D. N12, E. N14 and F. N15.

Fig. 8. Release proles of individual tablets collected at different tableting time points at pH 1 (red) and pH 7.4 (black). A. N11. B. N12 and, C. N15. (For interpretation of the
references to colour in this gure legend, the reader is referred to the web version of this article.)
666 S. Lakio et al. / International Journal of Pharmaceutics 511 (2016) 659668
formulations generally results in release to a larger extent
determined by erosion compared to a more highly soluble drug
where the diffusion mechanism is more dominant (Tajarobi et al.,
2009; Heiman et al., 2014). For a composition such as the one used
in this study, the release mechanism is at low pH mainly
determined by erosion and it appeared to be more robust with
regards to variabilities in the distribution of the matrix former than
was observed at high pH. This was probably mainly related to a
much slower hydration rate of the matrix making gelled matrix
mechanically stronger. The robustness was better at low pH since
the effects of the percolation threshold are less dramatic when the
API has lower solubility (Tajarobi et al., 2009). Another reason for
the improved robustness was that ibuprofen did not have a
possibility to disrupt the gel structure at low pH, when not
extensively dissolved, in that extent than it could do in higher pH,
when to a larger extent dissolved.
According to the MLR model (R2 = 0.984, Q2 = 0.980) mixing
speed did not have an effect on release behavior at pH 1 (Fig. 9A).
The amount of ibuprofen (Ibu%) had the most signicant effect on
release behavior since when the Ibu% was higher there were more
components with lower solubility in the matrix. Also particle sizes
of ibuprofen (IbuPS) and HPMC (HPMC PS) and the interaction term
of HPMC PS and Ibu% had an effect. When Ibu% and IbuPS increased
and HPMC PS decreased the T80 was increased. This can be seen
clearly on the contour plots in Fig. 9B. For example if the
formulation would have Ibu% of 20% and the aim would be to reach
T80 > 350 min, Ibu PS 66 mm would need to be used to reach this
goal. However, if Ibu% would be lower (<10%) and the aim would be
to achieve T80 < 200 min, then also HPMC PS would have an effect
on T80 with all ibuprofen particle size.
3.4.1. Effects of processing time on release performance at pH 1.0
Interestingly processing time did not have an effect on release
behavior at pH 1 unlike at pH 7.4. For example when looking at runs
N11 and N15 that were signicantly affected by processing time at
pH 7.4 (Fig. 7C and F) there were no processing time effect at all
(Fig. 8). Also MLR model indicated that process time had no
inuence when considering time to reach 80% ibuprofen release
Fig. 9. Time to reach 80% ibuprofen release (T80) at pH 1 after 20 min tableting. A. Coefcient plot for time to reach 80% release (T80). B. Contour plot representing the effect of
HPMC particle size (HPMC PS), ibuprofen particle size (IbuPS) and ibuprofen concentration (Ibu%) on the time to reach T80.
(Fig. 9A). The release at pH 1 was slower at all processing time
points than the release at pH 7.4.
As mentioned before when pH is low the solubility of ibuprofen
is lower and therefore the system is not so sensitive to variabilities
in hydration and gelling of the HPMC matrix and instead to a larger
extent determined by the drug substance hydration properties.
4. Discussion
A homogeneous distribution of the matrix former (HPMC) is
required to provide a robust release performance. Our data clearly
conrmed that when the HPMC particle size was small (CR) and
the mixer speed was high this target was easier to reach. Also, more
robust release was obtained at high drug load and when the pH
was such that the solubility of the drug was lower. Therefore, when
using large HPMC particles (DC2) to provide improved powder
ow properties and more robust processing it is important to
assure that the time allowed for reaching steady state, also with
regards to release robustness, is sufciently long and/or that the
mixing speed is sufciently high. In order to more rapidly provide a
robust release from tablets containing large sized HPMC, a higher
content of matrix former may be needed. It was also noticed that
ibuprofen was homogeneously distributed signicantly faster than
matrix former, especially for HPMC DC2. Thus it is essential to
control also matrix former distribution during manufacturing.
When using higher amount of ibuprofen, a decent release
prole could be reached using both grades of HPMC and both
mixing speeds. However, time to reach equilibrium varied.
Reaching the steady state took longer for runs containing DC2.
The runs that contained DC2 had better ow properties and
therefore provided better robustness in other tablet properties
(assay, tablet weight and tensile strength) even though the tensile
strength of the tablets was slightly lower than with CR grade.
The manufacturing of extended release matrix tablets can be
a challenging task. The matrix former has to be homogeneously
distributed to ensure proper functionality of the tablet. This
usually demands quite long mixing times in batch processing.
However continuous mixing enables a more practical and time-
 S. Lakio et al. / International Journal of Pharmaceutics 511 (2016) 659668
saving approach. When the steady state has been achieved the
functionality of the matrix tablets is ensured. Even though
reaching the steady state can take some time, the quality may
after that remain equal as long as the process is continued.
With batch manufacture, every mixing process is a separate
process and needs to be started all over again every time.
Normally, HPMC CR would be very challenging to use in
traditional batch mixing followed by direct compression due to
its poor ow properties. However, one important benet of
continuous mixing followed by direct compression as demon-
strated here is that it enables the use of more poorly owing
materials than would be possible through a combination of
batch mixing and direct compression.
The same tablet quality can be achieved with batch and
continuous processing (Leuenberger, 2001; Ervasti et al., 2015).
Still robust and consistent homogeneity can be achieved signi-
cantly faster with the continuous processing developed here and
thus the batch-to-batch variation can be minimized. In addition
the data presented in this continuous process emphasize that
adaptive control is a pivotal means for robust operating generating
consistent quality output (see Ervasti et al., 2015 for further
discussion).
Reaching steady state regarding output quality is crucial in
continuous processes and it should be dened as a state where all
quality attributes are in a predetermined design space that does
not change over time. For extended release formulations the rate
of release is a key to performance and needs to be assured
adequately.
5. Conclusions
Homogeneous distribution of the matrix former (HPMC) is
required to provide a robust in vivo performance. Our data clearly
indicate that with small HPMC particle size (HPMC CR) and high
mixer speed this target was easily reached for the compositions
studied. When using larger particle size HPMC (DC2) to gain the
advantage of improved powder ow properties and more robust
processing conditions, it was found important to assure that
sufcient time is given to reach steady state, also with regards to
release robustness.
The ibuprofen release was signicantly slower and more robust
at high drug load than at low. In addition release was also slower at
pH 1 than at pH 7.4. This was due to the lower solubility of
ibuprofen in acidic conditions and a higher relative amount of
poorly soluble components in the high drug load compositions.
The processing time was generally found to have an effect on
the release rate at pH 7.4. Tablets collected at early processing
times released ibuprofen markedly faster than tablets collected at
later time points at earlier time points especially when using
HPMC DC2. Interestingly processing time did not have a signicant
effect on release behavior at pH 1.
The present work highlights the advantage with a continuous
manufacturing line through the possibility to tailor the process
parameters based on the functionality related characteristics of the
incoming raw material, here HPMC particle size. In particular, it is
shown that continuous mixing and direct compression can offer
advantages over batchwise processing for manufacturing of tablets
from poorly owing powder blends. By using continuous
processing it is possible to manufacture and achieve robust
performance of compositions that would not be possible with
traditional batch processing.
Acknowledgements
TE, SPS and JK gratefully acknowledge the nancial support as
well as the provision of equipment by the PROMIS Centre
667
consortium (supported by the Finnish Funding Agency for
Technology and Innovation (Tekes), the Regional Council of
Pohjois-Savo, the North Savo Centre for Economic Development,
Transportation and Environment, and participating industrial
partners). Peter Forsberg (AZ) is thanked for helping in continuous
tablet manufacturing and in-vitro dissolution tests. Johan Brei-
tholtz (AZ) and Fredrik Winge (AZ) are acknowledged for their
assistance in setting up the dissolution method. Jesper Hagberg is
acknowledged for performing the dissolution studies related to the
percolation threshold.
References
Besenhard, M.O., Karkala, S.K., Faulhammer, E., Fathollahi, S., Ramachandran, R.,
Khinast, J.G., 2016. Continuous feeding of low-dose APIs via periodic micro
dosing. Int. J. Pharm. 509, 123134.
Brocher, J., 2015. The BioVoxxel toolbox. ImageJ User and Developer Conference,
September 2015, Madison (WI), USA.
Carabello, I., 2010. Factors affecting drug release from hydroxypropyl
methylcellulose matrix systems in the light of classical and percolation theories.
Expert Opin. Drug Deliv. 7 (11), 12911301.
Ervasti, T., Simonaho, S.P., Ketolainen, J., Forsberg, P., Fransson, M., Wikstrm, H.,
Folestad, S., Lakio, S., Tajarobi, P., Abrahmsn-Alami, S., 2015. Continuous
manufacturing of extended release tablets via powder mixing and direct
compression. Int. J. Pharm. 495 (1), 290301.
Heiman, J., Tajarobi, F., Gururajan, B., Juppo, A., Abrahmsn-Alami, S., 2014. Roller
compaction of hydrophilic extended release tabletscombined effects of
processing variables and drug/matrix former particle size. AAPS PharmSciTech
16 (2), 267277.
Jrvinen, M.A., Paaso, J., Paavola, M., Leivisk, K., Juuti, M., Muzzio, F., Jrvinen, K.,
2013. Continuous direct tablet compression: effects of impeller rotation rate,
total feed rate and drug content on the tablet properties and drug release. Drug
Dev. Ind. Pharm. 39 (11), 18021808.
Lapidus, H., Lordi, N.G., 1968. Drug release from compressed hydrophilic matrices. J.
Pharm. Sci. 57, 12921301.
Lee, S.L., OConnor, T.F., Yang, X., Cruz, C.N., Chatterjee, S., Madurawe, R.D., Moore, C.
M.V., Yu, L.X., Woodcock, J., 2015. Modernizing pharmaceutical manufacturing:
from batch to continuous production. J. Pharm. Innov. 19.
Leuenberger, H., 2001. New trends in the production of pharmaceutical granules:
batch versus continuous processing. Eur. J. Pharm. Biopharm. 52 (3), 289296.
Llusa, M., Mohr, S., Baumgartner, R., Paudel, A., Koscher, G., Khinast, J., 2016.
Continuous low-dose feeding of highly active pharmaceutical ingredients in
hot-melt extrusion. Drug Dev. Ind. Pharm. 42 (8), 13601364.
Marikh, K., Berthiaux, H., Mizonov, V., Barantseva, E., 2005. Experimental study of
the stirring conditions taking place in a pilot plant continuous mixer of
particulate solids. Powder Technol. 157, 138143.
Mitchell, K., Ford, J.L., Armstrong, D.J., Elliott, P.N.C., Rostron, C., Hogan, J.E., 1993a.
The inuence of concentration on the release of drugs from gels and matrices
containing methocel. Int. J. Pharm. 100, 155163.
Mitchell, K., Ford, J.L., Armstrong, D.J., Elliott, P.N.C., Hogan, J.E., Rostron, C., 1993b.
The inuence of drugs on the properties of gels and swelling characteristics of
matrices containing methylcellulose of hydroxypropyl methylcellulose. Int. J.
Pharm. 100, 165173.
Mohamed, F.A.A., Roberts, M., Seton, L., Ford, J.L., Levina, M., Rajabi-Siahboomi, A.R.,
2013. Production of extended release mini-tablets using directly compressible
grades of HPMC. Drug Dev. Ind. Pharm. 39, 16901697.
Nicholson Price II, W., 2014. Making do in making drugs: innovation policy and
pharmaceutical manufacturing. B.C.L. Rev. 55, 491562.
Pernenkil, L., Cooney, C.L., 2006. A review on the continuous blending of powders.
Chem. Eng. Sci. 61, 720742.
Plumb, K., 2005. Continuous processing in the pharmaceutical industry. Changing
the mind set. Chem. Eng. Res. Des. 83, 730738.
Portillo, P.M., Ierapetritou, M.G., Muzzio, F.J., 2008. Characterization of continuous
convective powder mixing process. Powder Technol. 182, 368378.
Ridell, A., Evertsson, H., Nilsson, S., Sundelf, L., 1999. Amphiphilic association of
ibuprofen and two nonionic cellulose derivatives in aqueous solution. J. Pharm.
Sci. 88, 11751181.
Rogers, T.L., Hewlett, K.O., Theuerkauf, J., Balwinski, K.M., 2013. Assessing how the
physical properties and enhanced powder ow of HPMC affect process control
during direct compression of matrix tablets. Tablets Capsules 1422.
Schaber, S.D., Gerogiorgis, D.I., Ramachandran, R., Evans, J.M.B., Barton, P.I., Trout, B.
L., 2011. Economic analysis of integrated continuous and batch pharmaceutical
manufacturing: a case study. Ind. Eng. Chem. Res. 50 (17), 1008310092.
Schindelin, J., Argande-Carreras, I., Frise, E., Kaynig, V., Longair, M., Pietzsch, T.,
Preibisch, S., Rueden, C., Saalfeld, S., Schid, B., Tinevez, J.-Y., White, D.J.,
Hartenstein, V., Eliceiri, K., Tomancak, P., Cardona, A., 2012. Fiji: an open-source
platform for biological-image analysis. Nat. Methods 9, 676682.
Schneider, C.A., Rasband, W.S., Eliceiri, K.W., 2012. NIH Image to ImageJ: 25 years of
image analysis. Nat. Methods 9, 671675.
Siepmann, J., Streubel, A., Peppas, N.A., 2002. Understanding and predicting drug
delivery from hydrophilic matrix tablets using the sequential layer model.
Pharm. Res. 19 (3), 306314.
668 S. Lakio et al. / International Journal of Pharmaceutics 511 (2016) 659668

Tajarobi, F., Abrahmsn-Alami, S., Hansen, M., Larsson, A., 2009. The impact of dose
and solubility of additives on the release from HPMC matrix tabletsidentifying
critical conditions. Pharm. Res. 26 (6), 14961504.
Timmins, P., Delargy, A.M., Minchom, C.M., Howard, J.R., 1992. Inuence of some
process variables on product properties for a hydrophilic matrix controlled
release tablet. Eur. J. Pharm. Biopharm. 38, 113118.
Velasco, M.V., Ford, J.L., Rowe, P., Rajabi-Siahboomi, A.R., 1999. Inuence of drug:
hydroxypropyl methyl cellulose ratio, drug and polymer particle size and
compression force on the release of diclofenac sodium from HPMC tablets. J.
Control. Release 57, 7585.
Vervaet, C., Remon, J.P., 2005. Continuous granulation in the pharmaceutical
industry. Chem. Eng. Sci. 60, 39493957.
Yin, J., Clayton, J., 2014. The Role of powder characterization in continuous
manufacturing experimental work on wet granulation demonstrates the use of
dynamic powder characterization to support continuous processing for tablet
manufacturing. Pharm. Technol. 38 (6) .