Professional Documents
Culture Documents
Woer TZ 2011
Woer TZ 2011
Woer TZ 2011
ABSTRACT: Electronic tongues are sensor array systems that are increasingly being used in
the field of pharmaceutics to provide taste assessment data of formulations. The applicability
of an electronic tongue in the development of a taste masked generic ibuprofen suspension,
starting from a commercial taste masked product, was evaluated in this study. The initial
screening study on 3 proprietary and 11 generic products showed that sensors of the taste sens-
ing system TS-5000Z could clearly detect differences between the products. The variation of
sensor responses were mainly caused by sodium salts, sweeteners, and preservatives, whereas
pH and viscosity did not affect sensor response. In addition, the presence of the particles (20–
100 :m) did not damage the sensor membranes. Based on this screening, and the known
qualitative composition of the proprietary formulations, the approximate quantitative com-
position of a proprietary formulation could be deduced and a taste masked generic formu-
lation could be developed using the electronic tongue data. Differences in sensor responses
between the proprietary and optimized generic formulation were smaller than 11 mV for each
sensor. Based on these results a rational approach of implementing an electronic tongue to
simplify the development of a taste masked generic formulation could be introduced. © 2011
Wiley-Liss, Inc. and the American Pharmacists Association J Pharm Sci 100:4460–4470,
2011
Keywords: pediatric; formulation; analysis; excipients; in vitro models; multivariate analysis
4460 JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 100, NO. 10, OCTOBER 2011
DEVELOPMENT OF A TASTE-MASKED GENERIC IBUPROFEN SUSPENSION QUERY 4461
calibration of the sensors. The final formulation could velop a new generic formulation approaching the pro-
then be compared with a pleasant tasting placebo for- prietary formulation. According to this new top-down
mulation to finally improve taste-masking efficiency. approach based on electronic tongue measurements,
An alternative approach is the development of the comparability of marketed products was evalu-
a generic taste-masked formulation utilizing an al- ated and the feasibility to monitor the influence of ex-
ready available taste-masked product. In this in- cipient variation on sensor responses was elaborated.
stance, the type of excipients used for the proprietary In addition, the ability to rationalize and facilitate
formulation is already known, whereas the quanti- the development of a generic product with the same
tative composition of excipients remains undisclosed. pleasant taste attributes was investigated.
The comparison between eight generic and one pro-
prietary product, with respect to their taste proper-
Experimental
ties, has already been successfully conducted based
on electronic tongue measurements.12 Nevertheless, Ibuprofen Marketed Suspensions
up until now, the development of a generic formula-
Fourteen ibuprofen suspensions (2%, 4%) from eight
tion oriented to a taste-masked proprietary product
pharmaceutical companies and seven different man-
and guided by electronic tongue investigations has
ufacturers commercially available in Germany were
not been performed. It is therefore an aim of this
investigated (Table 1). The qualitative composition of
study to show how the composition of a brand prod-
the marketed products is shown in Table 2 as listed
uct could be deduced and how to rationally develop
in the respective product leaflet. For taste evalua-
a taste-masked generic product based on electronic
tion, products were tested without centrifugation or
tongue measurements.
dilution.
To perform this study, oral ibuprofen suspensions
available on the German market were selected as Development of Generic Formulations
model formulations. According to a recently published
Ibuprofen (European Pharmacopoeia (Ph.Eur.) grade)
study, ibuprofen is one of the most frequently pre-
was generously provided by BASF (Ludwigshafen,
scribed drugs in pediatric primary care, used for the
Germany). Glycerol 85% (Ph.Eur. grade), polysorbate
treatment of fever, pain, or rheumatoid arthritis.13
80 (Ph.Eur. grade), quinine hydrochloride (Ph.Eur.
However, its taste is described as being sour, bitter,
grade), and sodium saccharin (Ph.Eur. grade) were
and peppery.14 One possible approach of taste mask-
purchased from Caelo (Hilden, Germany). Citric
ing is the synthesis of a poorly soluble salt of the drug
acid (Ph.Eur. grade), domiphen bromide (analyt-
substance15 because only dissolved drug molecules
ical grade), and sodium citrate (Ph.Eur. grade)
are able to interact with taste buds, unlike solid drug
were obtained from Sigma–Aldrich (Steinheim, Ger-
particles. In the case of ibuprofen, the poor solubil-
many); sodium chloride (Ph.Eur. grade) from VWR
ity in water at neutral pH (26 mg/L) can be utilized
(Leuven, Belgium); maltitol 55/75 (production grade)
in order to prepare a suspension. However, ibupro-
from DHW GmbH (Rodleben, Germany); orange fla-
fen exhibits an unpleasant bitter and sour taste even
vor (production grade) from Nordmann, Rassmann
when dispersed in an aqueous system.14 Therefore,
GmbH (Hamburg, Germany); and xanthan gum
the remaining concentration of ibuprofen in the liquid
(Ph.Eur. grade) from Roeper (Hamburg, Germany).
phase of a suspension still needs to be taste masked.
Purified water was obtained by reverse osmosis. The
There are various oral ibuprofen suspensions on
quantitative composition of the different formulations
the German market containing 2% or 4% ibuprofen,
is shown in Table 3. For taste evaluation, prototype
three proprietary formulations, and 11 generics, dif-
suspensions were prepared and stirred afterward for
fering in their taste and acceptance by children. The
24 h in order to reach equilibrium. Particles were sub-
leading brand product “Nurofen 2% orange” is well
sequently removed by centrifugation (15 min, 3345g)
known to have a pleasant taste and a preferred accep-
from one part of the prototype suspension to compare
tance by children.14 Thus, a taste screening including
the suspension with its corresponding base contain-
all ibuprofen suspensions available on the German
ing dissolved ibuprofen.
market was performed in order to check whether an
electronic taste sensing system would be able to differ- Electronic Tongue Measurements
entiate the taste of the various products. On the basis
Sensors
of this initial feasibility study, a generic formulation
with similar taste attributes, meaning similar sensor Sensors and reference electrodes were purchased
signal patterns obtained by electronic tongue mea- from TecLabS Europe OHG (Essen, Germany). Inner
surements, compared with the Nurofen formulation solution (0.2 mL; see section Chemicals and Reagents
was developed. Starting from the known qualitative for Electronic Tongue Measurements) was filled into
composition of the original product, the qualitative each sensor prior to the beginning of experiments.
composition of excipients was varied in order to de- The reference electrode was completely filled up with
DOI 10.1002/jps JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 100, NO. 10, OCTOBER 2011
4462 WOERTZ ET AL.
inner solution. All sensors were preconditioned in tongue was equipped with eight lipid membrane sen-
standard solution for 1 day before measurement. sors labeled according to the different taste qualities
bitterness (three sensors), sweetness, sourness, salti-
Chemicals and Reagents for Electronic Tongue ness, umami, and astringency and four corresponding
Measurements reference electrodes. The underlying measurement
Potassium chloride (analytical grade) was acquired principle is potentiometric and sensor responses were
from Grüssing (Filsum, Germany). Tartaric acid obtained as mV values consequently. A sensor check
(Ph.Eur. grade) was purchased from Sigma–Aldrich was conducted routinely before every measurement to
Laborchemikalien (Seelze, Germany). Water was assure that sensors were stable in the correct voltage
demineralized by reverse osmosis. Distilled water was range. Each sample was measured five times with the
obtained by in-laboratory distillation of demineral- sweetness sensor and four times with the remaining
ized water. Absolute ethanol (purity 99.8%) was pur- sensors. One measurement cycle consisted of measur-
chased from VWR International. Hydrochloric acid ing a reference solution (Vr ), followed by the sample
(1 mol/L) and potassium hydroxide solution (0.1 mol/ solution (Vs ), a short (2× 3 s) cleaning procedure, and
L) were acquired from Merck (Darmstadt, Germany). measurement of the aftertaste (Vr ) in reference solu-
The inner solution for sensors and reference elec- tion. The aftertaste was measured by determining the
trodes consisting of 3.33 mol/L potassium chloride in change of membrane potential caused by adsorption
saturated silver chloride solution was provided by In- of the substance to the lipid membrane after the short
sent (Atsugi-shi, Japan). cleaning procedure. Both, sensor output for taste, also
called relative value (R), and sensor output for af-
Preparation of Standard, Washing, and Sample Solutions tertaste, also called CPA value (change of membrane
potential caused by adsorption) were calculated in re-
Two washing solutions for negatively and positively
lation to the preliminary determined sensor response
charged sensors, respectively, were prepared by di-
to the reference solution (Vr ).
luting absolute ethanol-to-ethanol 30% with distilled
water and adding 100 mM hydrochloric acid in the
case of negatively charged sensors or 100 mM potas- R = Vs − Vr (1)
sium chloride and 10 mM potassium hydroxide for the
positively charged sensors. A standard solution serv-
ing as cleaning and reference solution was prepared
by dissolving 30 mM potassium chloride and 0.3 mM CPA = Vr − Vr (2)
tartaric acid in 1 L of distilled water.
The whole measurement procedure was performed
Electronic Tongue System and Measurement Setup
for all samples and repeated up to five times. For fur-
All measurements were performed with the taste ther data treatment, the two first runs were discarded
sensing system TS-5000Z (Insent). This electronic for measurements with the sweetness sensor and the
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 100, NO. 10, OCTOBER 2011 DOI 10.1002/jps
DOI 10.1002/jps
Table 2. Qualitative Composition of Investigated Ibuprofen Suspension Available on the German Market (According to the Product Leaflet)
Excipient Formulation F1 F2 F3 F4 F5 F6 F7
Ibuprofen (g) 2.0 2.0 2.0 2.0 2.0 2.0 2.0
Maltitol (g) 30.0 23.0 23.0 35.0 23.0 35.0 45.0
Sodium saccharin (g) 0.1 0.1 0.1 0.06 0.06 0.1 0.1
Sodium chloride (g) 0.75 – 0.375 0.375 0.375 0.375 0.375
Sodium citrate (g) 0.2 – 0.1 0.1 0.1 0.1 0.1
Orange flavor (gtt) 3 3 3 3 3 3 3
Citric acid (g) 0.05 0.05 0.05 0.05 0.05 0.05 0.05
Domiphen bromide (g) 0.01 0.01 0.01 0.01 0.01 0.01 0.01
Xanthan (g) 0.4 0.4 0.4 0.4 0.4 0.4 0.4
Glycerol 85% (g) 3.0 3.0 3.0 3.0 3.0 3.0 3.0
Polysorbate 80 (g) 0.025 0.025 0.025 0.025 0.025 0.025 0.025
Purified water (g) ad 100 ad 100 ad 100 ad 100 ad 100 ad 100 ad 100
first run was discarded for measurements with the camera D300 (Nikon, Tokyo, Japan) under polarized
remaining sensors, as recommended by the supplier, light.
to enable conditioning of the sensors. The method was
validated as described in a previous paper.16 Laser Light Diffraction
The median (x50 ) of ibuprofen particles was deter-
Evaluation of Results
mined by a laser diffraction system (Helos/KF-Magic,
The results were recorded as raw data (mV) of the Sympatec GmbH, Clausthal-Zellerfeld, Germany) via
relative measurement of the sample to the reference. wet dispersion method using a 50 mL cuvette. One
Sensor signal results were either evaluated directly or drop of ibuprofen suspension was stirred in a satu-
multivariate data analysis was performed. For mul- rated solution of ibuprofen for 5 min at 1000 rpm.
tivariate data analysis, raw data were pretreated by Measurements were performed in triplicate.
mean centering and scaling to unit variance. Data
processing, graphical illustration, and statistical in-
terpretation of the results were carried out using RESULTS
Excel 2007 (Microsoft, Redmond, Washington) and Screening of Ibuprofen Marketed Suspensions
SIMCA-P+ v11.5 (Umetrics AB, Umeå, Sweden). To
determine the distance between two samples (p, q) af- Three proprietary (Nuro 2% O, Nuro 2% S, and Nuro
ter multivariate data analysis, Euclidean distances 4% S) and 11 generic ibuprofen suspensions (2% and
were calculated including all the principal compo- 4%) were evaluated for pH, particle size, dynamic vis-
nents (PCs) (n) built for the model. cosity, and taste properties.
pH Measurements
n
d(p, q) = (p i − qi )2 (3) The pH of the marketed ibuprofen suspension varied
i=1 between 3.28 for Ibuflam 2% and 4.7 for Nuro 2% O
(Table 4). A correlation between the univariate sensor
pH Measurement responses of the electronic tongue and the pH of the
formulations could not be found. For example, Dolor
The pH of all the suspensions was determined by a 2% and Dolor 4% were evaluated with different taste
digital Knick pH-Meter 766 Calimatic (Knick GmbH properties by all sensors while the pH of both products
& Company, Berlin, Germany). was approximately 4.0.
Viscosity Measurement Ibuprofen Particles
The viscosity of all the suspensions was determined Products from the proprietary formulations as well
with a Kinexus rheometer (Malvern Instruments as all generic products consisted of ibuprofen drug
Ltd., Worcestershire, United Kingdom) at shear rates particles (x50 = 20–100 :m) dispersed in the suspen-
of 0.1/s (representing storage conditions) and 100/s sion vehicle, except the Ibuflam 2% suspension. The
(representing stress while swallowing). presence of cellulose acetate phthalate and ethylcel-
lulose in this formulation indicated that a coating of
Microscopy
the drug particles was performed during manufactur-
Microscopic images of solid ibuprofen particles in sus- ing to improve the taste of the formulation (Table 2).
pension were taken with a light microscope DMLB Microscopic images (Fig. 1) of the suspension illus-
(Leica Microsystems, Wetzlar, Germany) and a digital trated coated ibuprofen particles. Although several
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 100, NO. 10, OCTOBER 2011 DOI 10.1002/jps
DEVELOPMENT OF A TASTE-MASKED GENERIC IBUPROFEN SUSPENSION QUERY 4465
Viscosity [Pa s]
DOI 10.1002/jps JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 100, NO. 10, OCTOBER 2011
4466 WOERTZ ET AL.
by PC-3 (∼14%) and PC-4 (∼6%), which is displayed than 11 mV for each sensor. Figure 4 additionally
here because they did not reveal further differences shows the differences between the prototype formula-
between the formulations The main differentiation, tions and their corresponding suspension bases after
therefore, between the samples has to be carried out PCA, including all sensors. About 85% of the propri-
with respect to the x-axis (PC-1), with additional in- etary formulation could be expressed by the first two
formation gained by the y-axis (PC-2). All formula- components (PC-1 = 66%, PC-2 = 19%), whereas the
tions were detected as being different from the pure remaining part is explained by PC-3 (∼11%) and PC-4
ibuprofen solution. In addition, three clusters exist (∼2%). As expected, the remaining 13% did not fur-
on the PCA map representing the main differences ther affect the main results and were therefore not
between the suspensions. displayed here. Again, it becomes obvious that F1
and F7 were closest to the proprietary formulation
Development of a Generic Ibuprofen Suspension Nuro 2% O. Distances calculated on the basis of the
(Top-Down Approach) four PCs showed the order of similarity of prototype
Figure 3 shows sensor responses of prototype formu- formulations with respect to the proprietary formula-
lations F1–F7 in comparison to the proprietary formu- tion: Nuro 2% O < F7 < F1 < F6 < F3 < F4 < F5 <
lation Nuro 2% O normalized to the response of the F2. In addition, formulation F1 approached the pro-
proprietary formulation used in every measurement prietary formulation quite well, but revealed a large
to assure comparability of the results. It can be seen distance to its suspension base at the same time (Ta-
how the influence of excipient variation was clearly ble 5). The pH of the prototype formulations decreased
reflected by the sensor responses. Although formu- with the removal of sodium chloride and sodium cit-
lation F1 containing the highest amount of sodium rate and increased with the addition of sweetening
chloride and sodium citrate appeared to be quite sim- agents ending up with the same pH compared with
ilar to the original formulation, the removal of these the proprietary formulation Nuro 2% O (Tables 4
substances lead to remarkable changes in sensor re- and 6).
sponses (F2–F3). Furthermore, the varying amount
of sodium saccharin and maltitol was seen by the
differing sensor responses to formulations F4–F7. It DISCUSSION
was further seen that sensors umami, saltiness, sour-
Evaluation of Marketed Products
ness, bitterness 1, and bitterness 2 were most sensi-
tive to changes in the sodium chloride/sodium citrate The results from the electronic tongue measurements
amount, whereas sensors umami, saltiness, sourness, clearly showed that the formulations could be dis-
sweetness, bitterness 1, and bitterness 3 detected the tinguished according to their excipients and man-
changes in maltitol content. The variation of sodium ufacturer (Fig. 2). One group consists of the three
saccharin had an influence on all sensors except sen- Nurofen products. The clear differentiation to the re-
sor bitterness 3. maining suspensions can be explained by the pres-
Although the sensor responses of most sensors var- ence of sodium chloride and sodium citrate for sus-
ied for the formulations F1–F6 compared with the pension stabilization. In addition, sodium ions are
proprietary formulation, the difference between for- known for masking bitter taste.17 The difference be-
mulation F7 and the original formulation was smaller tween Nurofen 2% and Nurofen 4% can be explained
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 100, NO. 10, OCTOBER 2011 DOI 10.1002/jps
DEVELOPMENT OF A TASTE-MASKED GENERIC IBUPROFEN SUSPENSION QUERY 4467
by the different preservative used as well as the pres- containing exactly the same excipients marketed by
ence of thaumatin. The second group consists of the different companies, whereas on the left side formu-
two “Dolormin” formulations, which were detected as lations, AL 2%, Ibuflam 2%, and Ibudolor 2% can be
being similar with respect to the y-axis, and different found. The reason for this is the different nature of
with respect to the x-axis. The formation of the cluster sweetening agents used for these formulations. The
can be explained by the different preservative and sta- formulation containing coated particles, Ibuflam 2%,
bilizing agent compared with the other formulations, also contained pure ibuprofen crystals (Fig. 1), lead-
whereas the difference between the 2% and 4% sus- ing to the assumption of the presence of dissolved
pension can be explained by the different sweetening ibuprofen as well. Furthermore, only a small differ-
agents used for the suspension vehicle. The last and ence was found between suspensions containing 2%
biggest group is represented by the remaining for- or 4% ibuprofen within the same suspension vehicle.
mulations showing a marginal further subdivision. This showed that the presence of particles did not
The right part of the group consists of suspensions influence the measurement and that the saturated
DOI 10.1002/jps JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 100, NO. 10, OCTOBER 2011
4468 WOERTZ ET AL.
solution together with the excipients determined the Table 5. Distances Between Prototype Generic Formulations
sensor responses. Furthermore, against previous ex- (F) and Their Corresponding Suspension Bases (B) and Distances
Between Prototype Formulations and Nuro 2%O (Principal
pectations, sensor membranes were not damaged by Component Analysis, R2 ( PC-1–PC-4) = 0.99%)
the particles. This offers the opportunity to measure
suspensions without preliminary time-consuming fil- Distance to Suspension Distance to Nuro
tration or centrifugation. However, to investigate Formulation (F) Base 2% O
other particle containing systems in the same man- F1 3.99 2.83
ner, different particle sizes and shapes could have an F2 1.61 6.16
influence on the stability of the membranes and need F3 2.09 3.54
F4 0.39 4.68
to be considered before. The pH differed between the F5 1.08 5.62
formulations (Table 4), but did not correlate with the F6 0.64 3.21
univariate sensor responses. In addition, a possibly F7 0.81 1.92
increased dissolved amount of ibuprofen at higher
pH values was not detected by the sensors because Table 6. pH of Prototype Generic Ibuprofen Suspensions
sensor responses were dominated by the presence of (F) and Their Corresponding Suspension Bases (B)
taste-masking excipients. Therefore, the influence of
pH on electronic tongue sensor responses could be ne- Formulation (F)/Suspension Base (B) pH
glected here. The differences in formulation viscosity F1 4.96
also had no effect on sensor responses. The increase B1 5.06
F2 3.54
in viscosity of a formulation is one possible approach
B2 3.67
to mask unpleasant taste as the interaction of dis- F3 4.59
solved molecules with taste buds is reduced.15 As it B3 4.68
was known from previous studies that xanthan gum F4 4.66
does not influence the sensor responses, the different B4 4.67
F5 4.21
velocities of particle sedimentation due to higher vis-
B5 4.21
cosity could lead to differences in interaction of the F6 4.67
particles with the sensor membranes. However, the B6 4.67
differences in viscosity were too small to influence F7 4.70
the electronic tongue’s sensor responses. Concluding, B7 4.70
the electronic tongue was able to differentiate ibupro-
fen marketed suspension in terms of their excipients,
whereas the presence of particles, difference in vis- a generic formulation with similar taste properties.
cosity, and different pH did not significantly influence Within the scope of evaluating the influence of the
the measurement results. different taste-masking excipients on the sensor re-
sponses, the amount of sodium citrate and sodium
chloride was varied first. These excipients primarily
Evaluation of Generic Ibuprofen Suspensions
serve as stabilizing agents but are also known to influ-
On the basis of the results of the initial feasibility ence bitter taste. In addition, their influence on sen-
study, the proprietary formulation “Nurofen 2% or- sor responses was shown in the initial screening of
ange” was chosen as a target formulation to develop marketed suspensions. Univariate and multivariate
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 100, NO. 10, OCTOBER 2011 DOI 10.1002/jps
DEVELOPMENT OF A TASTE-MASKED GENERIC IBUPROFEN SUSPENSION QUERY 4469
DOI 10.1002/jps JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 100, NO. 10, OCTOBER 2011
4470 WOERTZ ET AL.
9. Harada T, Uchida T, Yoshida M, Kobayashi Y, Narazaki R, non-steroidal anti-inflammatory drugs in paediatric primary
Ohwaki T. 2010. A new method for evaluating bitterness of care in the UK, Italy and the Netherlands. Pharmacol Res
medicines in development using a taste sensor and a dis- 62:243–248.
integration testing apparatus. Chem Pharm Bull 58:1009– 14. Reader S, Shaw H, Hails S 2006. A taste-testing study in
1014. healthy volunteers (children) to investigate children’s pref-
10. Abdullah MZ, Rahman ASA, Shakaff AYM, Noor AM. 2004. erence for ibuprofen or placebo suspension. Paediatr Perinat
Discrimination and classification of eurycoma longifolia jack Drug Ther 7:54–58.
in medicinal foods by means of a DSP-based electronic taste 15. Ayenew Z, Puri V, Kumar L, Bansal AK. 2009. Trends in phar-
sensor. Trans Inst Meas Control 26:19–39. maceutical taste masking technologies: A patent review. Re-
11. Woertz K, Tissen C, Kleinebudde P, Breitkreutz J. 2010. Ra- cent Pat Drug Delivery Formulation 3:26–39.
tional development of taste masked liquids guided by an elec- 16. Woertz K, Tissen C, Kleinebudde P, Breitkreutz J. 2010. Per-
tronic tongue. Int J Pharm 400:114–123. formance qualification of an electronic tongue based on ICH
12. Tokuyama E, Matsunaga C, Yoshida K, Mifsud J-C, Irie T, guideline Q2. J Pharm Biomed Anal 51:497–506.
Yoshida M, Uchida T. 2009. Famotidine orally disintegrating 17. Keast RSJ, Breslin PAS. 2002. Modifying the bitterness of
tablets: Bitterness comparison of original and generic prod- selected oral pharmaceuticals with cation and anion series of
ucts. Chem Pharm Bull 57:382–387. salts. Pharm Res 19:1019–1026.
13. Neubert A, Verhamme K, Murray ML, Picelli G, Hsia Y, Sen 18. Krause J. 2008. Novel paediatric formulations for the
FE, Giaquinto C, Ceci A, Sturkenboom M, Wong ICK, TEDDY drug sodium benzoate. Ph.D. Thesis. Duesseldorf, Germany:
Network of Excellence. 2010. The prescribing of analgesics and Heinrich-Heine-Univerity.
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 100, NO. 10, OCTOBER 2011 DOI 10.1002/jps