International Journal of Pharmaceutics 441 (2013) 9–18

Contents lists available at SciVerse ScienceDirect

International Journal of Pharmaceutics

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

Formulation and optimization of desogestrel transdermal contraceptive patch

using crystallization studies
Vishal Sachdeva a , Yun Bai a , Agis Kydonieus b , Ajay K. Banga a,∗
a
Department of Pharmaceutical Sciences, College of Pharmacy and Health Sciences, Mercer University, Atlanta, GA 30341, USA
b
Agile Therapeutics, Princeton, NJ 08540, USA

a r t i c l e i n f o a b s t r a c t

Article history: Levonorgestrel (LNG) is the most commonly used progestin in contraception. In this study, we report
Received 20 August 2012 the use of an alternative progestin, desogestrel, for transdermal contraception. The drug was found to be
Received in revised form 5 December 2012 significantly more permeable as compared to LNG (p < 0.05). Crystallization studies were used to select
Accepted 12 December 2012
the best adhesive among acrylate (Duro-Tak 87-4098 and Duro-Tak 87-202A) and polyisobutylene (PIB,
Available online 20 December 2012
Duro-Tak 87-608A) pressure sensitive adhesives by determining the drug’s saturation solubility in them.
The use of copovidone and mineral oil as formulation excipients was investigated to increase drug loading
Keywords:
in the PIB adhesive. Physical characterization of the patches was performed using in vitro drug release,
Desogestrel
Contraceptive
content analysis, patch weight and thickness variations and rolling ball tack and peel adhesion studies.
Transdermal patch Optimized patches were evaluated for in vitro transdermal delivery across hairless rat skin. The saturation
Acrylate adhesive solubility of desogestrel was found to be approximately 49.3% (w/w) and 55.6% (w/w) in Duro-Tak 87-
Polyisobutylene adhesive 4098 and Duro-Tak 87-202A acrylate adhesives, respectively. The saturation solubility of desogestrel was
Crystallization significantly lower (3–4%, w/w) in the PIB adhesive. Mineral oil (10%, w/w) and copovidone (30%, w/w)
were found to be optimum for increasing drug loading and patch cosmetics. Results from the physical
characterization studies suggest that a uniform and reproducible 7 day drug-in-adhesive patch could be
developed.
© 2012 Elsevier B.V. All rights reserved.

1. Introduction
Progestogens are one of the most commonly used families of
drugs for contraception. These are known to produce their phar-
macological effect by suppressing pituitary gonadotropin secretion
resulting in follicular development inhibition and ovulation inhi-
bition and also by altering the cervical mucus to inhibit sperm
penetration. Progestogens used in the past were useful but in
an attempt to develop more potent drugs for improved and safe
contraception, scientists synthesized levonorgestrel by chemical
modification. However, improvement in progestogenic activity was
associated with simultaneous rise in intrinsic androgenic activity.
To reduce latter, scientists suggested reduction in progestogenic
dose or further chemical modification of the molecule. This led
to the development of a new generation of progestogens. Des-
ogestrel was the first of these to be developed for commercial
use in Europe in 1981 for oral contraception. In comparison to
levonorgestrel, desogestrel showed improved in vitro binding affin-
ity to progesterone receptors and greater progestogenic effect on
endometrial histology and ovulation inhibition. Further discussion
∗ Corresponding author. Tel.: +1 678 547 6243; fax: +1 678 547 6423.
E-mail address: banga ak@mercer.edu (A.K. Banga).
regarding the structural comparison between the two drugs can
be found in this publication (Lammers et al., 1998). Currently,
several oral combination contraceptive preparations containing
desogestrel and ethinyl estradiol are available in the US mar-
ket. These are marketed under several brands names including
Apri® ; AzuretteTM ; Caziant® ; Cyclessa® ; Desogen® ; EmoquetteTM ;
Kariva® ; Mircette® ; Ortho-Cept® ; Reclipsen® and VelivetTM . These
oral preparations are available as 21 or 28 tablet packages. The
tablets have to be administered at the same time daily and missed
tablets have to be followed by specific guidelines to compen-
sate for missed doses (Lexi-Drugs and Online., 2012). Minimum
daily dose of desogestrel for a woman with a regular menstrual
cycle is 60 ␮g/day (100 ␮g/day for levonorgestrel). However, most
marketed oral preparations consist of 150 ␮g drug (2.5 times the
required dose), to provide enhanced safety margin. This increased
drug loading in tablets has been reasoned as required because the
majority of the users have been reported to forget taking one or two
pills per cycle. Furthermore, the drug is known to undergo first pass
metabolism with oral bioavailability reported to be around 84%
and the elimination half life of desogestrel following a single dose
is known to be 12.4 ± 1.9 h (Lammers et al., 1998). Thus, despite
improved progestogenic activity, desogestrel in the currently avail-
able oral dosage forms is associated with the inconvenience of
taking one pill daily, risk of pregnancy due to non-compliance

0378-5173/$ – see front matter © 2012 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.ijpharm.2012.12.014
10 V. Sachdeva et al. / International Journal of Pharmaceutics 441 (2013) 9–18
with required regimen, unnecessary higher systemic drug expo-
sure, high first pass metabolism and short elimination half life.
Considering these limitations, delivery of desogestrel from a simple
transdermal patch appears to be more reasonable. This is because
transdermal delivery systems offer several advantages including
reduced systemic drug exposure, bypass of first pass metabolism,
suitability for drugs with short half lives, ease of self administration
and termination and widespread acceptance of another available
weekly patch product for delivery of a different steroid (Banga,
2011; Sitruk-Ware, 1995).
Transdermal patch technology has advanced tremendously
since the first scopolamine patch was introduced in the market in
1979. It is due to today’s advanced patch making technology that
nearly a billion patches are manufactured every year (Prausnitz and
Langer, 2008). These transdermal patches are classified into three
types: drug in adhesive (consists of drug directly dispersed into
adhesive polymer), reservoir (consists of a drug reservoir between
a backing membrane and rate controlling membrane with a skin
contacting adhesive layer) and matrix (consists of a drug reser-
voir in the center with a peripheral adhesive ring around the
edges) patches (Ghosh et al., 1997; Subedi et al., 2010; Tan and
Pfister, 1999). The selection of the most appropriate design of the
transdermal system would depend upon several factors such as
physicochemical properties of the drug, intended duration of appli-
cation, drug amount to be delivered, whether or not dose titration
will be required and others. Among the three designs mentioned
above, the drug in adhesive patches is the simplest and the most
commonly used design. In this, the drug is incorporated directly
into an adhesive and is layered in between the backing membrane
and a release liner (Lipp and Muller-Fahrnow, 1999). These patches
offer advantages of being simple and elegant with smaller and
thinner dimensions. From the manufacturing point of view, drug-
in-adhesive matrix type patches offer advantages such as lack of
drug leakage issues and ease of manufacturing over other trans-
dermal systems (Kim and Choi, 2002; Minghetti et al., 2007). It is
due to all these reasons why drug in adhesive patches have gained
increased popularity among patients in the last two decades (Jain
and Banga, 2010; Lipp and Muller-Fahrnow, 1999). Drug release
from a drug in adhesive patch depends directly upon the drug
concentration in the patch and follows first order kinetics. Upon
release from the patch the drug diffuses passively through the skin
layers to reach the systemic circulation, to produce the desired
therapeutic effect. In order to facilitate this passive diffusion of
the drug from the transdermal patches, a supersaturated state
would be preferred in these patches (Hadgraft, 1999; Latsch et al.,
2004; Lipp, 1998; Tan and Pfister, 1999). Supersaturation refers to
a state in which the drug amount solubilized in a given vehicle
exceeds its saturation solubility. In such a state, increase in the
drug’s thermodynamic activity is associated with an increase in
drug permeation. However, such systems are metastable (thermo-
dynamically unstable) and there is high probability that the drug
will crystallize within such systems during storage (Kim and Choi,
2002; Ma et al., 1996; Minghetti et al., 2007; Variankaval et al.,
1999).
Crystallization generally initiates with the formation of an
embryo by collision of single drug molecules. The continuation
of such collisions between the molecules results in an increase
in the critical radius of the embryo that does not re-dissolve
easily, resulting in the formation of a drug crystal. The crystal
growth then continues until a state of saturation is achieved (Lipp,
1998). The development of crystals in transdermal patches is unde-
sirable because it makes them unstable, could allow dangerous
amounts of drug to be delivered, influences skin permeation neg-
atively and reduces the esthetic appeal of the product leading to
reduced patient product acceptability (Kim and Choi, 2002; Ma
et al., 1996). Crystallization of sex steroids in drug-in-adhesive
patches has occurred in the past and has been reported in the
literature (Lipp, 1998; Lipp and Muller-Fahrnow, 1999; Ma et al.,
1996; Yu et al., 1991). Most recently, the rotigotine patch (Neupro® )
was recalled from the market in 2008 due to drug crystallization
issues, as snowflake like crystals were detected on the patch sur-
face (Chaudhuri, 2008). Thus, investigation into the crystallization
potential of a drug in a specific formulation and the use of alter-
nate approaches to prevent it are essential for the development
of an acceptable and efficient transdermal product. A number of
different approaches have been reported by which crystallization
in a patch can be eliminated. These include reducing drug loading
below the saturation solubility, pro-drug usage where the pro-
drug has higher solubility in the formulation, use of co-solvents
to increase solubility, adhesive modification to increase solubility
of the drug in the adhesive, and crystallization inhibitors (Lipp,
1998). Among these the use of additives such as crystallization
inhibitors and/or solublizers has been reported widely in the lit-
erature (Cilurzo et al., 2005; Jain and Banga, 2010; Kim and Choi,
2002; Kotiyan and Vavia, 2001; Lipp, 1998; Ma et al., 1996; Schulz
et al., 2011). These additives include polyvinylpyrrolidone (PVP)
and its derivatives, copovidone, crospovidone, dextrin derivatives,
mannitol, polyethylene glycol, polypropylene glycol, poloxamer,
Tween 80® , Labrasol® and glycerin (Jain and Banga, 2010; Kim
and Choi, 2002; Lipp, 1998; Ma et al., 1996; Schulz et al., 2011).
A number of mechanisms have been proposed in the literature by
which these additives might prevent crystallization. These include
additives getting adsorbed on the crystal surfaces, preventing the
crystal nucleation process, and co-precipitate formation between
the drug and the additives to form amorphous solids, i.e., forming a
solid “solution”, which is a more plausible explanation of this phe-
nomenon. Incorporation of these additives in the patch formulation
helps in maintaining higher amounts of drug in the patch and in
achieving higher drug flux across the skin for a longer time period.
The hypothesis of crystallization inhibition is not well established;
for example it was recently shown that PVP acts as a solubilizing
agent for levonorgestrel and not as a crystallization inhibitor (Jain
and Banga, 2010).
In this study, a 7 day transdermal drug in adhesive contraceptive
patch using desogestrel was prepared, optimized and evaluated. For
this, both slide and patch crystallization studies were performed to
determine the saturation solubility of the drug in the patch compo-
nents. The use of two acrylate adhesives and one polyisobutylene
(PIB) adhesive was investigated. In order to increase drug loading
in the PIB adhesive without causing crystallization, the use of two
additives – copovidone (Plasdone® S-630) and mineral oil were also
investigated. In vitro skin permeation studies were then performed
using optimized patches. Physical characterization of these patches
was also performed using drug release studies, patch content anal-
ysis, patch thickness and weight variation studies, rolling ball tack
test and peel adhesion test.
2. Materials and methods
2.1. Materials
Desogestrel, copovidone (Plasdone® S-630), backing mem-
brane (2 mil polyester with an ethylene vinyl acetate copolymer,
ScotchpakTM 9732 from 3M) and release liner (3 mil fluoropolymer-
coated polyester film, ScotchpakTM 9744 from 3M) were provided
by Agile Therapeutics, Inc. (Princeton, NJ, USA). Mineral oil was
obtained from Sigma–Aldrich (St. Louis, MO, USA). Acrylate PSA
adhesives (Duro-Tak 87-4098 and Duro-Tak 87-202A) and a poly-
isobutylene (PIB) adhesive (Duro-Tak 87-608A) were obtained as
gift samples from Henkel Corporation (Dusseldorf, Germany). PEG-
400, gentamycin sulfate, tetra hydrogen furan (THF), and HPLC
 V. Sachdeva et al. / International Journal of Pharmaceutics 441 (2013) 9–18
grade methanol were purchased from Fisher Scientific (Pittsburgh,
PA, USA). Deionized water was used for the HPLC analysis.
2.2. Skin for permeation studies
Hairless rat skin was used to evaluate the permeation of des-
ogestrel from the final optimized patch. Skin was isolated from
hairless rats (male, 8–10 weeks old and 350–400 g in weight) that
were obtained from Charles River (Wilmington, MA, USA). All the
animals were allowed to get acclimatized for at least 1 week prior
to their use in any experiment. All the studies were performed as
per the protocol approved by the Institutional Animal Care and
Use Committee (IACUC) at Mercer University. Hairless rats were
euthanized by carbon dioxide asphyxiation prior to the perme-
ation experiment and its abdominal skin was carefully excised
using a pair of scissors and forceps. The underlying subcutaneous
fat was removed from the excised skin and the abdominal skin
thus obtained (∼1 mm thick) was used for the permeation experi-
ments.
2.3. Slide crystallization studies
For these studies, desogestrel or levonorgestrel was dissolved
in THF (400 mg/ml). A drop of this solution was then transferred
using a pipette on a glass slide. The slide was then placed under the
hood for air drying at room temperature to allow the organic sol-
vents to evaporate. Drug crystals thus obtained on the slide were
observed under a polarized microscope (Leica DM 750) for nine
consecutive days and again after a month. Crystal images were
taken at 10× or 100× magnification (as specified) using a DFC-
280 camera which was attached to the microscope. Similar studies
were performed to determine the saturation solubility of the drug
in the additives copovidone (drug:copovidone – 100:0, 99.5:0.5,
99:1, 98.5:1.5, 98:2, 97:3, 96:4, 95:5, 92.5:7.5, 90:10, 87:13, 84:16,
80:20, 70:30, 60:40, 50:50, 30:70, 20:80, 10:90, 0:100, w/w) and
mineral oil (10%, w/w). For this the drug and the additive were
mixed together in THF in different (w/w) ratios and the slides were
observed for crystals. For saturation solubility of desogestrel in
acrylate PSA adhesives (Duro-Tak 87-4098 and Duro-Tak 87-202A)
and PIB PSA adhesives (Duro-Tak 87-608A) both slide and patch
crystallization studies were performed. For the slide crystallization
studies with adhesives, drug and adhesive were mixed in different
ratios to achieve a concentration range, diluted with equal amount
of methanol and mounted on glass slides. The highest concentra-
tion at which no crystals were observed was considered as the
drug’s saturation solubility in the respective adhesive (Foreman
et al., 2002). Several drug concentrations in adhesive were inves-
tigated for Duro-Tak 87-4098 (7.7, 10.4, 29.4, 45.5, 55.6 and 62.5%,
w/w), Duro-Tak 87-202A (39.4, 46.7, 49.3, 56.5, 62.8 and 66.1%,
w/w) and Duro-Tak 87-608A (2, 4, 7.5, 10 and 20%, w/w). The solid
content (left following evaporation of solvent) percentages used to
calculate the wet weight of adhesives were 38.5% for Duro-Tak 87-
4098, 40% for Duro-Tak 87-202A and 38% for Duro-Tak 87-608A.
For patch crystallization studies, patches were prepared using the
procedure described below at various drug concentrations in adhe-
sive and observed for crystallization for at least one month. The
concentrations investigated for drug in PIB alone patches were 2,
3, 4, 7.5, 10 and 20% (w/w) and for drug in 10% mineral oil in PIB
patches were 3.7, 4.4, 5, 7.5 and 10% (w/w). Slides or patches pre-
pared using exactly the same procedure but without drug served
as corresponding controls.
2.4. Drug in adhesive patch preparation
The drug in adhesive transdermal patches was prepared for crys-
tallization and permeation studies using the procedure described
11
below. Pre-determined amounts of drug, adhesive, ethyl acetate
and/or additives (copvidone/mineral oil) were weighed into a glass
container with lid and sealed using a parafilm to minimize loss
of organic solvents. PIB consist of 38% solid content and the wet
weight of this adhesive was back calculated accordingly for weigh-
ing into the glass container. The formulation was stirred for 2 h
using a magnetic stirrer to form a homogenous mixture. The mix-
ture was then casted on the release liner using a Gardner film
casting knife (BYK-AG-4300 series, Columbia, MD, USA) and the
cast sheet was dried in an oven at 60 ◦ C for 17 min. The pur-
pose of this drying step was to evaporate organic solvents and
for this reason the drug to adhesive ratio for patch formulations
was calculated based on adhesive’s dry weight. Following drying,
the entire sheet was laminated using a backing membrane which
was placed on the cast film using a roller, ensuring no air pock-
ets were formed. This sheet was observed for crystallization by
visual inspection and under polarized microscope (Leica DM 750)
for nine consecutive days and again after a month. This longer dura-
tion of observation of a month was essential because crystallization
sometimes did not occur immediately following patch preparation
because the diffusion coefficient of steroids is known to be low
in high viscosity adhesive matrix systems (Lipp, 1998). Following
each microscopic evaluation, the patches were heat sealed in Barex
pouches (PET/LDPE/AL foil/Barex) (American Packaging Corpora-
tion, Rochester, NY, USA) and stored at room temperature. Crystal
images were taken using a DFC-280 camera which was attached
to the microscope. The sheets showing no crystal formation dur-
ing the duration of observation (at least 1 month) were used for
permeation studies. For all permeation and physical characteriza-
tion studies, patches of the desired size were cut out of the above
mentioned prepared sheets.
2.5. Permeation studies
The 7 day permeation studies were performed using in vitro
Franz diffusion cells (PermeGear, Inc., Hellertown, PA, USA) hav-
ing an effective diffusion surface area of 0.64 cm2 (n ≥ 3). In order
to compare the permeability of levonorgestrel and desogestrel,
saturated solution of each drug was prepared separately in PEG-
400. These served as corresponding donor solutions. The receptor
phase consisted of PEG 400 having gentamycin sulfate (80 mg/L).
Gentamycin sulfate was added to the receptor phase to prevent
microbial growth during the 7 day study (Chisty et al., 2002; Valiveti
et al., 2004). During the entire study the receptor phase was main-
tained at 37 ◦ C with constant stirring at 600 rpm. Freshly excised
and cleaned hairless rat abdominal skin was obtained on the day
of the experiment. This isolated skin was placed in between the
donor and the receptor compartments and the entire set up was
then secured in place using a clamp. Donor solution (0.5 ml) was
then loaded into the donor cells using a pipette and the top was
covered using a parafilm and a silver foil. Samples (0.5 ml) were
withdrawn at predetermined time points (24, 48, 72, 96, 120, 144,
168 h) and replaced with equal volume of fresh receptor fluid. The
samples obtained were analyzed for drug content (levonorgestrel
or desogestrel) using HPLC. Using exactly the same protocol as
described above, permeation experiments (n ≥ 4) were then per-
formed using the final optimized desogestrel patches across the
hairless rat abdominal skin. The only difference was that instead
of using the saturated desogestrel solution as donor, desogestrel
containing patches were used. Transdermal patches, big enough to
cover the receptor compartment top were cut out of the cast sheets,
the release liner was removed and the patches were placed on the
skin such that the adhesive side of the patch was facing the stratum
corneum side of the skin. The donor cell was then placed and the
entire set up was secured using a clamp. All samples obtained were
analyzed using HPLC.
12 V. Sachdeva et al. / International Journal of Pharmaceutics 441 (2013) 9–18
2.6. In vitro drug release study
The 7 day patch release studies were performed (n = 6) using
in vitro Franz diffusion cells (Jain and Banga, 2010; Jain et al., 2003).
Patches of 1 cm2 were cut out of the prepared patch sheets and
the backing membrane side of these patches was then glued to
parafilm using a cyano-acrylate adhesive to allow easy handling
and mounting of the patches on the Franz diffusion cells. The recep-
tor compartment consisted of PEG 400 having gentamycin sulfate
(80 mg/L) and was maintained at 37 ◦ C with constant stirring at
600 rpm. The release liner was removed from the patches and the
active portion of the patch was placed on the receptor compart-
ment (adhesive side facing receptor fluid) ensuring absence of any
air bubbles in between the patches and the receptor fluid. The donor
cell was then placed on the receptor compartment and the entire
set up was secured using a clamp. Samples (0.5 ml) were taken at
predetermined time points (1, 3, 4, 6, 8, 10, 12, 24, 36, 48, 60, 72,
84, 96, 108, 120, 132, 144, 156, 168 h) and replaced with equal vol-
ume of fresh receptor fluid. The samples obtained were analyzed
for desogestrel using HPLC.
2.7. Weight and thickness variation of optimized patches
Weight variation of the prepared patches was also determined
by cutting 32 individual patches, 1 cm2 in surface area and recor-
ding their weight. The average weight of the backing membrane and
release liner having exactly the same area was then subtracted from
the weight of each patch to obtain the actual weight of the contents
in the active portion of the patch. The average weight of each patch
along with the standard error was reported. The thickness of the
patches was measured using an Absolute Digimatic caliper (Model
# CD-6-CS, Mitutoyo, Tokyo, Japan) and was reported. Six 1 cm2
patches were cut from the patch sheets and the thickness of the
individual patches was measured.
2.8. Rolling ball tack test
To determine the tack of optimized desogestrel patches, a RBT-
100 rolling ball tack tester (Chem Instruments, Fairfield, OH, USA)
was used. The equipment provides a convenient method for quan-
tifying an adhesive’s ability to adhere to another surface. The
equipment consists of a metallic inclined trough and a ball. The test
involved placing the metallic ball at the top of trough’s inclined sur-
face and allowing it to roll down the incline by pressing a spring
lever. The distance traveled by the ball between the end of the
trough and the point where the ball stopped on the adhesive was
measured. The test was performed in quadruplicate and the average
distance traveled by the ball on the patch was reported.
2.9. Peel adhesion test
The bond strength of the optimized patches was determined
using a PA-1000-180 180◦ peel adhesion tester (Chem Instruments,
Fairfield, OH, USA). The equipment provides a convenient method
for quantifying the force required to pull a flexible patch away from
a non-flexible material (stainless steel) that is positioned parallel
to the patch. One end of the test specimen strip was placed in the
load (10 lb) cell grip and the other end of the test specimen was
made to adhere to the test platform. The minimum, average and
maximum force registering in the process of peeling the patch from
the stainless steel were determined and recorded.
2.10. Quantitative analysis
Analysis of the amount of drug in the samples was performed
using a chromatographic method described in the literature with
few modifications (Matejicek and Kuban, 2007). The Alliance high
performance liquid chromatography (HPLC) system (Waters Corp.,
MA, USA) equipped with a photodiode array detector (Waters
2996) was employed. Phenomenex RP C6 Luna 5␮ column (Phen-
omenex, Torrance, CA) set at 35 ◦ C was employed for gradient
elution method. The mobile phase consisted of methanol and
water. The gradient method was initiated with the use of a 70:30
(methanol:water) solution, followed by a change of the mobile
phase composition to 100% methanol over the next 7 min. This
methanol:water (100:0) composition was maintained till the 10th
minute and then the mobile phase composition was changed again
to a composition of 70:30 (methanol:water) by the 12th minute.
The run time of each injection was 15 minutes and the injection
volume was 100 ␮l. The flow rate of the mobile phase throughout
the run was 1.5 ml/min. The wavelengths used for the detection of
levonorgestrel and desogestrel were 244 nm and 210 nm, respec-
tively, and the retention times for the two drugs were around 6 min
and 8.5 min, respectively. The standard curve was linear over the
range of 0.5–100 ␮g.
2.11. Statistical analysis
All the results presented in the graphs are an average of at least
n = 3 trials and the error bars represent the standard errors (SE).
Student t-test and analysis of variance (ANOVA) were used to deter-
mine statistically significant differences. The p-value used in this
study was 0.05.
3. Results and discussion
3.1. Desogestrel vs. levonorgestrel permeation studies
The objective of this study was to investigate the development
of a 7 day drug-in-adhesive contraceptive patch using desogestrel
as the progestin molecule. Transdermal delivery of desogestrel
has been described in the patent literature (Lipp et al., 1995) but
research studies have not been reported in the past to the best of
the author’s knowledge. Therefore, it was essential to first deter-
mine desogestrel’s passive permeation profile across the skin. In
these studies, permeation of levonorgestrel was also performed
simultaneously for comparison as it is one of the most commonly
used progestins for contraception. In vitro permeation studies were
performed across hairless rat skin using Franz diffusion cells. A
solution of PEG-400 saturated with either drug (levonorgestrel or
desogestrel) was used as donor solution in the permeation stud-
ies. The average flux and the cumulative amount are shown in
Fig. 1A and B respectively and the values were significantly higher
for desogestrel as compared to levonorgestrel (p < 0.05). Average
cumulative amount of desogestrel and levonorgestrel at the end of
7 days was found to be 389.4 ± 6.2 ␮g/cm2 and 1.8 ± 0.1 ␮g/cm2 ,
respectively (Fig. 1B). These results suggest that desogestrel can
passively permeate through skin without the use of permeation
enhancers and its permeability was significantly higher than that
of levonorgestrel. Mathematical algorithms that predict the per-
meability of drugs through skin, based on the physicochemical
properties such as partition coefficient (log P), molecular weight
and melting point have been described in the literature (Barratt,
1995; Potts and Guy, 1992). These models are more directional
than precise in their predictions. For example one of the algorithms
uses only log P and molecular weight to predict permeation. How-
ever, the values of log P and molecular weight of desogestrel and
levonorgestrel are almost identical (log P 4; MW 310.47 Da vs. log
P 3.8; MW 312.45 Da) which would predict similar permeability
between the two progestins. Other algorithms that include melting
point would predict that desogestrel will have higher permeability
 V. Sachdeva et al. / International Journal of Pharmaceutics 441 (2013) 9–18
Fig. 1. Desogestrel vs. levonorgestrel permeation studies. Average drug flux (A, left) and average cumulative amount (B, right) plots for desogestrel (rounds) and levonorgestrel
(diamonds) across hairless rat skin from PEG solution saturated with drug. The error bars indicate the mean standard error (SE).
due to its lower melting point (Drug Bank and Database., 2012).
It is evident from the experimental results that the use of deso-
gestrel for the development of a transdermal contraceptive patch
is not only of interest due to its higher progestogenic activity and
reduced androgenic activity but also due to its better skin perme-
ation profile over that of levonorgestrel, which would allow one to
develop a much smaller and more elegant patch.
3.2. Crystallization studies
Having confirmed desogestrel’s skin permeability potential, the
next aim was to identify its saturation solubility in the patch to
avoid crystallization. Drug crystallization in drug in adhesive trans-
dermal patches is a critical issue as discussed earlier. Thus, it was
essential to perform the crystallization studies at an early product
development stage and adopt suitable approaches to prevent its
occurrence in the final product. Slide crystallization studies were
performed first to determine the saturation solubility of the drug
in different patch excipients. A similar study was performed with
levonorgestrel for comparison. The use of these types of studies for
identifying crystal formation and determining the saturation solu-
bility of drugs has been reported in the literature before (Foreman
et al., 2002; Jain and Banga, 2010). In the absence of any additive,
drug crystals were observed for both pure desogestrel (Fig. 2A) and
levonorgestrel (Fig. 2B) within 2 days following evaporation of the
organic solvent.
3.2.1. Drug crystallization in acrylate adhesives
The next aim of the study was to select an appropriate adhesive
and determine the drug’s saturation solubility in the adhesive and
all other components comprising the patch. Selection of an adhe-
sive for the development of a transdermal patch is critical to ensure
proper contact between the patch and the skin throughout the use-
ful life of the patch. A number of different classes of the pressure
sensitive adhesives are available for use in transdermal patch devel-
opment because they adhere to skin with light pressure and leave
no residue upon removal, are easy to use, are stable to environ-
mental changes and have good appearance. A number of different
factors should be considered while selecting an adhesive including
biocompatibility (non-irritant and non-sensitizing to skin), formu-
lation compatibility (inert to drug/formulation components and
allow for desired drug solubility), delivery system compatibility
(allow required drug/enhancers diffusivity and permeability) and
adhesive properties (adequate tack, cohesive strength and skin
adhesion). The three most commonly used pressure sensitive adhe-
sives are PIB, acrylates (polyacrylate copolymers) and silicones
(polysiloxanes). Three different adhesives were investigated for
13
the preparation of desogestrel transdermal patches (two acrylate
adhesives; Duro-Tak 87-4098 and Duro-Tak 87-202A and one PIB
adhesive; Duro-Tak 87-608A). Chemically, acrylate adhesives are
formed by the copolymerization of acrylic acid, acrylic esters, and
functional monomers such as vinyl acetate whereas PIB adhesives
are homopolymers of isobutylene (Tan and Pfister, 1999). More
specifically, Duro-TakTM 87-4098 was a vinyl acetate/2-ethylhexyl
acrylate copolymer without any reactive functional groups (non-
curing). Duro-Tak 87-202A was an all acrylic copolymer containing
a high level of hydroxyl functional acrylate monomer and was
one of the most hydrophilic grades available. Duro-Tak 87-608A,
a fully saturated hydrocarbon polymer without any functional
groups consisted of poly-isobutylene rubber solution and was very
hydrophobic. This information about the compositional differences
of the adhesives was obtained from the product’s technical data
sheet supplied by the manufacturer. The saturation solubility of
desogestrel in these adhesives was determined using the slide
method discussed earlier as well as crystallization studies on com-
plete patches. Determination of the saturation solubility of the drug
in the adhesives/polymers is critical as it determines the maximum
amount of drug that can be incorporated into the patch to ensure
maximum drug delivery without concern for long term instabil-
ity and crystallization. Therefore, to predict saturation solubility
of drug in adhesives/polymers Foreman et al. proposed an alter-
native and relatively faster method involving slide crystallization
studies (Foreman et al., 2002). This approach of studying crystal-
lization offers several benefits including simplicity of the method,
determination of saturation solubility in individual or combina-
tion of excipients and use of very small amounts of drug and/or
excipients.
Using Foreman’s method the upper limit of the saturation sol-
ubility of desogestrel in Duro-Tak 87-4098 was found to be less
than 56.5% (w/w) and it was taken as 49.3% (w/w) (Fig. 3). This was
based on the observation that slides having 56.5, 62.8 and 66.1%
(w/w) drug in Duro-Tak 87-4098 adhesive developed drug crys-
tals within the observation time period of 1 month whereas the
slide containing 49.3% drug (Fig. 3) and lower drug amounts did
not crystallize.
In an attempt to identify an adhesive with lower saturation sol-
ubility with an intention to reduce drug loading in the final patch,
another acrylate adhesive (Duro-Tak 87-202A) was studied. The
saturation solubility of desogestrel in this adhesive was found to
be even higher i.e. between 55.6% (w/w) and 62.5% (w/w) as drug
crystals were seen in the slides having 62.5% (w/w) (Fig. 4) con-
centration or higher (not shown) but not at 55.6% (w/w). Among
the two acrylate adhesives investigated, the saturation solubility of
desogestrel in Duro-Tak 87-4098 was lower suggesting that more
14 V. Sachdeva et al. / International Journal of Pharmaceutics 441 (2013) 9–18

Fig. 2. Drug crystals, pure desogestrel (A, left) and pure levonorgestrel (B, right), observed on slides at 100× magnification on the second day following evaporation of the
organic solvent.

Fig. 3. Desogestrel crystallization study in acrylic adhesive Duro-Tak 87-4098. Photomicrograph captured at 10× magnification showing absence of drug crystals at 49.3%
(w/w) concentration and presence of drug crystals at 56.5, 62.8 and 66.1% (w/w) concentration.

Fig. 4. Desogestrel crystallization study in acrylic adhesive Duro-Tak 87-202A. Picture captured at 10× magnification showing absence of drug crystals at 55.6% (w/w)
concentration and presence of drug crystals at 62.5% (w/w) concentration.
 V. Sachdeva et al. / International Journal of Pharmaceutics 441 (2013) 9–18
efficient use of the drug could be made using Duro-Tak 87-4098 as
the PSA in the patch.
3.2.2. Drug crystallization in PIB adhesive
The third adhesive investigated was the PIB adhesive (Duro-Tak
87-608A). The PIB adhesive was tested in slides and patches at dif-
ferent drug concentration including 2, 3, 4, 7.5, 10 and 20% (w/w).
Crystals were observed at 7.5, 10 and 20% (w/w) concentrations
within 9 days (Fig. 5) while crystals appeared at 4% (w/w) concen-
tration on slide after 3 weeks (Fig. 5, bottom left). No crystals were
seen at 2% (w/w) or 3% (w/w) concentrations suggesting that the
saturation solubility of the drug in PIB was between 3 and 4% (w/w)
concentrations. These results indicate that slide/patch crystalliza-
tion studies can be helpful in the development of drug-in-adhesive
formulation. The findings discussed above indicate that the satu-
ration in the patch could be achieved with reduced drug amount
when PIB is used as the patch adhesive. This is beneficial from both
the manufacturing and environmental safety point of view. The lat-
ter is one of the concerns that the FDA has expressed for safety
issues FDA (2010) and the FDA has been encouraging the transder-
mal manufacturing companies to develop products that minimize
the amount of drug that is left behind after the useful life of the
patch. PIB has been reported to have broad acceptance for various
FDA-regulated applications including transdermal drug delivery
devices (Tan and Pfister, 1999). Other benefits that make PIB a
better adhesive for a desogestrel transdermal system include its
inertness, stability, flexibility and its long term adhesive properties
needed for the development of a 7 day patch. The last two bene-
fits have been attributed to the amorphous characteristics and low
glass transition temperature of PIB (Tan and Pfister, 1999). The use
of PIB has been reported to be more preferable for lipophilic drugs
with reduced polarity and low solubility parameter profile, which
is the case with desogestrel (Tan and Pfister, 1999). Considering the
above mentioned benefits, PIB was selected for the preparation of
patches for the remaining studies.
3.2.3. Drug crystallization in additives (Plasdone® S-630 and
mineral oil)
Incorporation of additives to increase drug loading was
attempted as the saturation solubility in the PIB adhesive alone
was low (3–4%, w/w concentration). Some increase in drug loading
was considered to be beneficial in order to keep the drug concen-
tration in the patch fairly constant over the 7 day period of patch
use. The two additives investigated were copovidone (Plasdone® S-
630) and mineral oil. Slide crystallization studies were performed
again to determine the saturation solubility of the drug in copovi-
done. In this experiment, desogestrel and copovidone were mixed
in THF at different (w/w) ratios and observed on slides for crys-
tallization. Crystals formed on the slides for some of the ratios
(drug:copovidone – 100:0, 99:1, 95:5, 92.5:7.5, 90:10 and 80:20)
are shown in Fig. 6.
It is clear from the images that the number and the size of the
crystal was reduced and the time to initial observation of crystal
formation increased with increasing amount of copovidone. For
example the first crystals in the 87:13 and 84:16 slides were found
within a month’s time period whereas the first crystals in the 80:20
slide were seen only after 2 months. Slides having drug and copovi-
done in 70:30, 60:40, 50:50 and lower (w/w) ratios did not show
crystals even after a period of 6 months. The exact saturation sol-
ubility could not be determined, but it is somewhere between 70
and 80% (w/w) concentration. Using a conservative approach, the
lowest percentage, i.e., 70% (w/w) was assumed as the saturation
solubility of the drug in copovidone to ensure no crystallization
would occur in the optimized patches. The reduction in crystalliza-
tion achieved with copovidone has been reported in the literature
as well (Jain and Banga, 2010). However, in our studies as indicated
15
Table 1
Final patch composition. Final composition of optimized patches weighing 300 mg
and having an area of 15 cm2 . Patches of 1 cm2 were cut out of these sheets and used
for various studies.
Constituents Patch weight Patch weight after
before drying (mg) drying (mg)
PIB+10% mineral oil
PIB 678.7 256.4
Mineral oil 30 30
Copovidone 1 1
Desogestrel 12.6 12.6
Total weight of sheet 300
PIB – polyisobutylene.
above and the studies with levonorgestrel [12], the prevention of
crystallization is due to the solubility of the respective progestins
in the copovidone.
Besides copovidone, the use of mineral oil as a solublizer was
also investigated to improve desogestrel solubility in the PIB adhe-
sive. Other intended benefits of incorporating mineral oil in the
patch were to soften the drug patch, increase the value of the diffu-
sion coefficient and decrease the resistance offered by the patch
matrix to the diffusion of the drug through it, especially since
steroids have been known to have low diffusion coefficients in
such high viscosity adhesive matrix systems (Lipp, 1998). Similar to
copovidone, it was essential to determine the saturation solubility
of desogestrel in the mineral oil. PIB patches were prepared con-
taining 10% mineral oil and the drug amount was varied at 3.7, 4.4,
5, 7.5 and 10% (w/w) concentrations. After ten months of observa-
tion the only patch that did not show crystal formation was the one
containing 3.7% (w/w) drug (Fig. 7), indicating that the saturation
solubility of desogestrel is between 3.7 and 4.4% (w/w).
Based on the saturation solubility numbers obtained earlier
for the drug in 10% mineral oil in PIB adhesive (4%) and copovi-
done (70%), an optimized patch sheet was prepared (Table 1). For
this, desogestrel equaling 90% of the saturation solubility of drug
obtained for each patch component (adhesive, copovidone and
mineral oil) was weighed and transdermal patch was prepared. The
purpose of adding 90% of drug with respect to its saturation solu-
bility value instead of 100% was to take into account deviations due
to non-ideal conditions and thus minimize the probability of drug
crystallization. On the other hand a high drug amount (90%) will
ensure a high concentration gradient across the skin throughout
the useful life of the patch.
3.3. Permeation studies using optimized patches
The optimized patches were used to evaluate desogestrel per-
meation profile across hairless rat skin. Rodents serve as a useful
alternative animal skin model for in vitro and/or in vivo studies. This
is due its small size, availability, convenient handling and housing
benefits. However, rodent’s and human’s skins have substantial
differences in physicochemical properties including thickness,
follicle density, skin layers composition, and number of sebaceous
glands. Therefore future studies should be performed with human
cadaver skin to more adequately assess the feasibility of this patch
for commercial purposes. Fig. 8A and B shows the average flux and
cumulative amount of desogestrel delivered following permeation
across the hairless rat skin from the optimized patches as well
as from the saturated PEG-400 solution (discussed earlier). The
average cumulative amount of desogestrel delivered at the end of
7 days from the patch was found to be 93.4 ± 7.1 ␮g/cm2 and the
average flux was found to be 0.7 ± 0.1 ␮g/cm2 /day, respectively.
The saturated PEG solution showed significantly higher average
cumulative amount of drug delivered as well as flux values when
compared to that delivered from the optimized patches (p < 0.05).
16 V. Sachdeva et al. / International Journal of Pharmaceutics 441 (2013) 9–18

Fig. 5. Desogestrel crystallization study in PIB adhesive Duro-Tak 87-608A. Picture captured at 100× magnification showing presence of drug crystals at 10% (w/w) concen-
tration in patch, 7.5% (w/w) concentration in patch, 4% (w/w) concentration on slide (bottom left) and 4% (w/w) concentration in patch (bottom right).

This suggests that there is a greater resistance for drug diffusion
through the adhesive matrix of the patch when compared to the
drug diffusion through the PEG-400 solution.
3.4. Physical characterizations of optimized patches
Following skin permeation evaluation, other important phys-
ical characteristics of the patches was determined. Patches were
characterized for their drug release profile, weight and thickness
variation, content analysis and tack and peel adhesion.
The in vitro release profile of the drug observed during the 7 day
study is shown in Fig. 9. The average cumulative amount released
at the end of the 7th day was 519.1 ± 20.1 ␮g/cm2 , representing
62% of the drug contained in the patch. A steady and continuous
release of the drug was observed following a parabolic release,
which is the expected release profile and indicates that the drug

Fig. 6. Saturation solubility of desogestrel in copovidone. Pictures captured at 100× magnification showing presence of crystals in the slides at 100%, 99%, 95%, 92.5%, 90%
and 80% (w/w) concentration of drug in drug-copovidone mixture. No crystals were seen on slides with lower drug in copovidone (w/w) ratios.
 V. Sachdeva et al. / International Journal of Pharmaceutics 441 (2013) 9–18 17

Fig. 7. Desogestrel crystallization study in PIB adhesive with mineral oil. Pictures captured at 100× magnification showing absence of crystals in the patch containing 10%
(w/w) mineral oil and 3.7% (w/w) drug in PIB and presence of crystals in the patch containing 10% (w/w) mineral oil and 10% (w/w) drug in PIB.

A B
7 Patch 450 Patch
Average Flux (µg/sq.cm/hr) ± SE

Saturated PEG solution Saturated PEG soultion

Average Cumulative Amt. of

400
6
350

(µg/sq.cm) ± SE
5
300
4 250
3 200
150
2
100
1 50
0 0
0 50 100 150 200 0 100 200
Time (hours) Time (hours)

Fig. 8. Desogestrel permeation studies. Average drug flux (A, left) and average cumulative amount (B, right) plots for desogestrel delivered across hairless rat skin from PEG
solution saturated with drug (diamonds) and optimized patches (squares). The error bars indicate the mean standard error (SE).

was uniformly distributed throughout the patch. Content analy- was 15.8 ± 0.1 mg. The test for thickness variation indicated that the
sis (n = 7) was also performed by extracting the drug from 1 cm2 average thickness of the patch was 0.3 ± 0.0 mm including the back-
patches using 10 ml methanol and shaking at 400 rpm for 2.5 days. ing and the release liner. The thickness of the release liner and back-
HPLC analysis of the drug extract indicated uniformity of drug con- ing membrane without the drug-adhesive-layer was found to be
tent in the patches with a standard error of less than three percent. 0.1 ± 0.0 mm. The above results indicate that the optimized patches
Test of weight variation conducted for the patches (n = 32) showed were uniform in weight and thickness as well as drug content.
that the average weight of the patch (1 cm2 ), excluding the weight The peel adhesion and the tack of the optimized patches were
of the release liner and backing membrane, was 18.7 ± 0.4 mg. The also evaluated. Excellent adhesion of a transdermal system is cru-
average weight of the backing and release liner, each of 1 cm2 area, cial for contraception, because proper surface contact with the skin
is required to ensure constant delivery of the drug (Sitruk-Ware,
1995). Peel adhesion test was performed to determine adhesive
property of the patch. The average force required to peel the patch
Average Cumulative Amt. of (µg/sq.cm) ±

600
from the stainless steel test surface was found to be 2237.6 ± 89.8 g
with a minimum and a maximum average force of 2033.9 ± 113.5 g
500
and 2564.3 ± 99.1 g, respectively. Rolling ball tests were performed
to determine the tack of the optimized patches. The average dis-
400
tance traveled by the metallic ball in the rolling ball tack test was
300 was found to be 1.8 ± 0.0 cm. The results for the peel adhesion and
SE

rolling ball tack tests confirm that the adhesive and tack properties
200 of the patches were not compromised because of the incorpora-
tion of the additives. The use of rolling ball tack test and 180◦ peel
100 adhesion test for development of transdermal systems has been
reported in the literature. For the rolling ball tack test, interpre-
0 tation is based on the distance traveled by the ball. The greater
distance traveled by the ball indicates toward the reduced tack of
0 50 100 150 200
the adhesive. In the work performed by Minghetti et al. 1999 the
Time (hours)
authors assumed a tack value of zero, if the distance traveled by
Fig. 9. In vitro drug release studies. Average cumulative amount of desogestrel
the ball was more than 25 cm. For peel adhesion test, the force
released. The error bars indicate the mean standard error (SE). required to peel away a patch from a rigid surface is determined
18 V. Sachdeva et al. / International Journal of Pharmaceutics 441 (2013) 9–18
to evaluate the peel property. The authors used these tests in their
study to evaluate the adhesive properties of the prepared patches
(Minghetti et al., 1999). We should mention here that adhesion to
stainless steel is not a very good surrogate of adhesion to skin and
further studies will need to be performed on human subjects to sub-
stantiate the results. A number of other factors might influence the
adhesive properties of such a systems including environmental fac-
tors (temperature, humidity), conditions of patch application site
(hairy vs. non-hairy, wet vs. dry, smooth vs. rough, presence of cos-
metic formulations such as oily creams) and hydration caused due
to prolonged skin occlusion. Hormones as well as excipients such as
copovidone and mineral oil are known not to cause irritation or skin
sensitization when used in transdermal delivery systems. How-
ever, consideration to all of the above mentioned factors should be
given during further product development to ensure wider patient
acceptability.
4. Conclusion
The results presented in this manuscript indicate that it is fea-
sible to develop a transdermal drug-in-adhesive patch using the
relatively more potent and permeable progestin, desogestrel. Both
PIB and acrylic adhesives can be used to transdermally deliver des-
ogestrel; however due to the lower solubility of desogestrel in PIB,
it was determined that a patch using PIB would be more efficient in
terms of drug usage. The use of copovidone was found to be bene-
ficial in increasing the drug loading and the crystallization studies
reported were beneficial in determining the saturation solubility of
the drug in the patch at an early stage of product development and
thus eliminating the issue of patch crystallization. Further studies
to evaluate the performance of the optimized transdermal deso-
gestrel contraceptive patch however will be required.
Conflict of interest
This project was funded by Agile Therapeutics and one of the
authors, A. Banga, has served as a consultant to Agile in the past.
Acknowledgements
We want to thank Agile Therapeutics for funding this research
project. We would also like to acknowledge Miss Suprita Tawde for
her valuable suggestions during the project.
References
Banga, A.K., 2011. Transdermal and Intradermal Delivery of Therapeutic Agents:
Application of Physical Technologies. Taylor and Francis, London.
Barratt, M.D., 1995. Quantitative structure–activity relationships for skin permeabil-
ity. Toxicol. In Vitro 9, 27–37.
Chaudhuri, K.R., 2008. Crystallisation within transdermal rotigotine patch: is there
cause for concern? Expert Opin. Drug Deliv. 5, 1169–1171.
Chisty, M.N., Bellantone, R.A., Taft, D.R., Plakogiannis, F.M., 2002. In vitro evaluation
of the release of albuterol sulfate from polymer gels: effect of fatty acids on drug
transport across biological membranes. Drug Dev. Ind. Pharm. 28, 1221–1229.
Cilurzo, F., Minghetti, P., Casiraghi, A., Tosi, L., Pagani, S., Montanari, L., 2005. Poly-
methacrylates as crystallization inhibitors in monolayer transdermal patches
containing ibuprofen. Eur. J. Pharm. Biopharm. 60, 61–66.
Drug Bank and Database 2012. http://www.drugbank.ca/drugs/DB00304 (accessed
07.11.12).
FDA, 2010. Guidance for Industry. In: Residual Drug in Transdermal and Related
Drug Delivery Systems. Food and Drug Administration, pp. 1–3.
Foreman, P., Hansen, A., Silverberg, E., Manegold, T., Yang, H., Li, J., Jacobson, S., 2002.
Predicting Drug Solubility in Transdermal Adhesives. American Association of
Pharmaceutical Scientists (AAPS), National Starch and Chemical Company.
Ghosh, T.K., Pfister, W.R., Yum, S.I., 1997. Transdermal and Topical Drug Delivery
Systems. Interpharm Press Inc., Buffalo Groove, IL.
Hadgraft, J., 1999. Passive enhancement strategies in topical and transdermal drug
delivery. Int. J. Pharm. 184, 1–6.
Jain, P., Banga, A.K., 2010. Inhibition of crystallization in drug-in-adhesive-type
transdermal patches. Int. J. Pharm. 394, 68–74.
Jain, S.K., Chourasia, M.K., Sabitha, M., Jain, R., Jain, A.K., Ashawat, M., Jha, A.K., 2003.
Development and characterization of transdermal drug delivery systems for
diltiazem hydrochloride. Drug Deliv. 10, 169–177.
Kim, J.H., Choi, H.K., 2002. Effect of additives on the crystallization and the perme-
ation of ketoprofen from adhesive matrix. Int. J. Pharm. 236, 81–85.
Kotiyan, P.N., Vavia, P.R., 2001. Eudragits: role as crystallization inhibitors in drug-
in-adhesive transdermal systems of estradiol. Eur. J. Pharm. Biopharm. 52,
173–180.
Lammers, P., Blumenthal, P.D., Huggins, G.R., 1998. Developments in contracep-
tion: a comprehensive review of desogen (desogestrel and ethinyl estradiol).
Contraception 57, 1S–27S.
Latsch, S., Selzer, T., Fink, L., Kreuter, J., 2004. Determination of the physical state of
norethindrone acetate containing transdermal drug delivery systems by isother-
mal microcalorimetry X-ray diffraction, and optical microscopy. Eur. J. Pharm.
Biopharm. 57, 383–395.
Lexi-Drugs Online 2012. http://online.lexi.com.proxy.mercer.edu/lco/action/doc/
retrieve/docid/patch f/6867. Lexi-drugs online (accessed 07.11.12).
Lipp, R., Gunther, C., Riedl, J., Tauber, U., 1995. Desogestrel containing trans-
dermal application agent. Canadian patent document no. CA 2183544,
Canadian Intellectual Property Office. http://brevets-patents.ic.gc.ca/opic-cipo/
cpd/eng/patent/2183544/summary.html (accessed 07.11.12).
Lipp, R., 1998. Selection and use of crystallization inhibitors for matrix-type trans-
dermal drug-delivery systems containing sex steroids. J. Pharm. Pharmacol. 50,
1343–1349.
Lipp, R., Muller-Fahrnow, A., 1999. Use of X-ray crystallography for the characteri-
zation of single crystals grown in steroid containing transdermal drug delivery
systems. Eur. J. Pharm. Biopharm. 47, 133–138.
Ma, X., Taw, J., Chiang, C.M., 1996. Control of drug crystallization in transdermal
matrix system. Int. J. Pharm. 142, 115–119.
Matejicek, D., Kuban, V., 2007. High performance liquid chromatography/ion-
trap mass spectrometry for separation and simultaneous determination of
ethynylestradiol, gestodene, levonorgestrel, cyproterone acetate and deso-
gestrel. Anal. Chim. Acta 588, 304–315.
Minghetti, P., Cilurzo, F., Montanari, L., 1999. Evaluation of adhesive properties of
patches based on acrylic matrices. Drug Dev. Ind. Pharm. 25, 1–6.
Minghetti, P., Cilurzo, F., Pagani, S., Casiraghi, A., Assandri, R., Montanari, L.,
2007. Formulation study of oxybutynin patches. Pharm. Dev. Technol. 12,
239–246.
Potts, R.O., Guy, R.H., 1992. Predicting skin permeability. Pharm. Res. 9, 663–
669.
Prausnitz, M.R., Langer, R., 2008. Transdermal drug delivery. Nat. Biotechnol. 26,
1261–1268.
Schulz, M., Fussnegger, B., Bodmeier, R., 2011. Influence of adsorbents in transdermal
matrix patches on the release and the physical state of ethinyl estradiol and
levonorgestrel. Eur. J. Pharm. Biopharm. 77, 240–248.
Sitruk-Ware, R., 1995. Transdermal application of steroid hormones for contracep-
tion. J. Steroid Biochem. Mol. Biol. 53, 247–251.
Subedi, R.K., Oh, S.Y., Chun, M.K., Choi, H.K., 2010. Recent advances in transdermal
drug delivery. Arch. Pharm. Res. 33, 339–351.
Tan, H.S., Pfister, W.R., 1999. Pressure-sensitive adhesives for transdermal drug
delivery systems. Pharm. Sci. Technolo. Today 2, 60–69.
Valiveti, S., Kiptoo, P.K., Hammell, D.C., Stinchcomb, A.L., 2004. Transdermal perme-
ation of WIN 55,212-2 and CP 55,940 in human skin in vitro. Int. J. Pharm. 278,
173–180.
Variankaval, N.E., Jacob, K.I., Dinh, S.M., 1999. Crystallization of beta-estradiol in an
acrylic transdermal drug delivery system. J. Biomed. Mater. Res. 44, 397–406.
Yu, J., Chien, T., Chien, Y.W., 1991. Transdermal dual-controlled delivery of testos-
terone and estradiol: (I) impact of system design. Drug Dev. Ind. Pharm. 17,
1883–1904.