Development of a Taste-Masked Generic Ibuprofen Suspension:

Top-Down Approach Guided by Electronic Tongue Measurements

KATHARINA WOERTZ, CORINNA TISSEN, PETER KLEINEBUDDE, JOERG BREITKREUTZ
Institute of Pharmaceutics and Biopharmaceutics, Heinrich-Heine-University Duesseldorf, 40225 Duesseldorf, Germany

Received 11 January 2011; revised 10 March 2011; accepted 27 April 2011

Published online 23 May 2011 in Wiley Online Library (wileyonlinelibrary.com). DOI 10.1002/jps.22629

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

INTRODUCTION One tool to overcome these drawbacks is the so-

called electronic tongues or electronic taste sens-
The taste of an oral liquid formulation administered
ing systems. These sensor array-based systems non-
to a child has an important impact on the adherence
specifically measure the interaction of excipients in
to drug therapy. Infants and toddlers are especially
a liquid formulation. The ability to detect differ-
more sensitive to bitter taste than adults.1 As nu-
ent substances in the same formulation in parallel
merous active pharmaceutical ingredients (APIs) ex-
is a main advantage compared with other analyt-
hibit a bitter taste, the regular intake, particularly in
ical techniques.3 The data produced, sensor signal
the case of chronic diseases, might be affected by the
patterns obtained as voltage values, offer the per-
child’s dislike to the taste of the medicine. Taste is,
spective of comparing multicomponent formulations
therefore, a prevalent issue that has to be addressed
with respect to their taste attributes. Electronic taste
in formulation development.2 Monitoring taste dur-
sensing systems have already been used successfully
ing the development process has many challenges
in pharmaceutics4,5 such as the characterization of
such as objectivity of human taste sensation, ethical
APIs,6 taste-masked solid and liquid formulations,7–9
concerns due to possible toxicity of the drug, and the
as well as for quality control purposes.10 In addition,
lack of comparability between adult and juvenile taste
the use of an electronic tongue for the stepwise and
sensation.
rational development of a taste-masked formulation
starting from an unpleasant tasting API and adding
Correspondence to: Katharina Woertz (Telephone: +49-211- specific taste-masking excipients was recently intro-
8115693; Fax: +49-211-8114251; E-mail: katharina.woertz@
uni-duesseldorf.de) duced in a previous paper.11 This bottom-up approach
Journal of Pharmaceutical Sciences, Vol. 100, 4460–4470 (2011) was shown to be effective as a preliminary screening
© 2011 Wiley-Liss, Inc. and the American Pharmacists Association of suitable taste-masking excipients after extensive

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.

Table 1. Investigated Ibuprofen Suspensions Available on the German Market

Abbreviated Name Brand Name Pharmaceutical Company Manufacturer

Nuro 2% O Nurofen Junior Fiebersaft Orange 2% Reckitt Benckiser Deutschland GmbH BCM Limited/Almirall Hermal
GmbH/Reckitt Benckiser
Healthcare (UK) Limited
Nuro 2% S Nurofen Junior Fiebersaft Erdbeer 2% Reckitt Benckiser Deutschland GmbH BCM Limited/Almirall Hermal
GmbH/Reckitt Benckiser
Healthcare (UK) Limited
Nuro 4% S Nurofen Junior Fiebersaft Erdbeer 4% Reckitt Benckiser Deutschland GmbH Italfarmáco/Farmasierra
Manufacturing
Dolor 2% Dolormin für Kinder 2% McNeil GmbH & Company oHG McNeil Ibérica
Dolor 4% Dolormin für Kinder 4% McNeil GmbH & Company oHG McNeil Ibérica
Ratio 2% IBU-ratiopharm 2% Ratiopharm GmbH Merckle GmbH
Ratio 4% IBU-ratiopharm 4% Ratiopharm GmbH Merckle GmbH
CT 2% Ibuprofen-CT 2% Kindersaft CT Arzneimittel GmbH CT Arzneimittel GmbH
CT 4% Ibuprofen-CT 4% Kindersaft CT Arzneimittel GmbH CT Arzneimittel GmbH
AbZ 2% Ibuprofen AbZ 2% Saft AbZ-Pharma GmbH Merckle GmbH
AbZ 4% Ibuprofen AbZ 4% Saft AbZ-Pharma GmbH Merckle GmbH
Ibudolor 2% Ibudolor Kindersaft 2% STADA GmbH STADA GmbH
AL 2% Ibuprofen AL Saft für Kinder 2% ALIUD
R
PHARMA GmbH ALIUD
R
PHARMA GmbH
Ibuflam 2% Ibuflam Kindersaft gegen Fieber und Winthrop Arznemittel GmbH Winthrop Arznemittel GmbH
Schmerzen 2%

inner solution. All sensors were preconditioned in
standard solution for 1 day before measurement.
Chemicals and Reagents for Electronic Tongue
Measurements
Potassium chloride (analytical grade) was acquired
from Grüssing (Filsum, Germany). Tartaric acid
(Ph.Eur. grade) was purchased from Sigma–Aldrich
Laborchemikalien (Seelze, Germany). Water was
demineralized by reverse osmosis. Distilled water was
obtained by in-laboratory distillation of demineral-
ized water. Absolute ethanol (purity 99.8%) was pur-
chased from VWR International. Hydrochloric acid
(1 mol/L) and potassium hydroxide solution (0.1 mol/
L) were acquired from Merck (Darmstadt, Germany).
The inner solution for sensors and reference elec-
trodes consisting of 3.33 mol/L potassium chloride in
saturated silver chloride solution was provided by In-
sent (Atsugi-shi, Japan).
Preparation of Standard, Washing, and Sample Solutions
Two washing solutions for negatively and positively
charged sensors, respectively, were prepared by di-
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-
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
tartaric acid in 1 L of distilled water.
Electronic Tongue System and Measurement Setup
All measurements were performed with the taste
sensing system TS-5000Z (Insent). This electronic
tongue was equipped with eight lipid membrane sen-
sors labeled according to the different taste qualities
bitterness (three sensors), sweetness, sourness, salti-
ness, umami, and astringency and four corresponding
reference electrodes. The underlying measurement
principle is potentiometric and sensor responses were
obtained as mV values consequently. A sensor check
was conducted routinely before every measurement to
assure that sensors were stable in the correct voltage
range. Each sample was measured five times with the
sweetness sensor and four times with the remaining
sensors. One measurement cycle consisted of measur-
ing a reference solution (Vr ), followed by the sample
solution (Vs ), a short (2× 3 s) cleaning procedure, and
measurement of the aftertaste (Vr
) in reference solu-
tion. The aftertaste was measured by determining the
change of membrane potential caused by adsorption
of the substance to the lipid membrane after the short
cleaning procedure. Both, sensor output for taste, also
called relative value (R), and sensor output for af-
tertaste, also called CPA value (change of membrane
potential caused by adsorption) were calculated in re-
lation to the preliminary determined sensor response
to the reference solution (Vr ).
R = Vs − Vr (1)
CPA = Vr
− Vr (2)
The whole measurement procedure was performed
for all samples and repeated up to five times. For fur-
ther data treatment, the two first runs were discarded
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)

Nuro 2% Ibudolor 2%, AL

O Nuro 2% S Nuro 4% S Dolor 2% Dolor 4% Ratio, CT, AbZ 2% Ratio, CT, AbZ 4% 2% Ibuflam 2%
API Ibuprofen Ibuprofen Ibuprofen Ibuprofen Ibuprofen Ibuprofen Ibuprofen Ibuprofen Ibuprofen
Sugars Sorbitol Glucose Glucose Glucose
Maltitol Maltitol Maltitol Sucrose Sucrose Sucrose Sucrose Maltitol Sucrose
Sweeteners Sodium Sodium saccharin Sodium saccharin Acesulfame K – Sodium saccharin Sodium saccharin Sodium saccharin
saccharin
Thaumatin
Flavors Cream Cream
Orange Strawberry Strawberry Not specified Not specified Strawberry Strawberry Strawberry not specified
Preservatives Citric acid Citric acid Citric acid Citric acid Citric acid Citric acid Citric acid Citric acid Citric acid
Domiphen Domiphen Sodium benzoate Sodium Sodium Potassium sorbate Potassium sorbate Parabenes
bromide bromide benzoate benzoate
Stabilizing agents HPMC Maize starch Maize starch Cellulose acetate
phthalate
Sodium Sodium citrate Sodium citrate Ethylcellulose
citrate
Sodium Sodium chloride Sodium chloride Modified starch
chloride
Xanthan Xanthan Xanthan Xanthan Xanthan Xanthan Xanthan Xanthan Xanthan
Other Azorubin Allura Red Allura Red
AC (colorant) AC (colorant)
Glycerol Glycerol Glycerol Glycerol Glycerol Glycerol
Polysorbate Polysorbate 80 Polysorbate 80 Polysorbate Polysorbate Polysorbate 80 Polysorbate 80 Polysorbate 80 Polysorbate 80
80 80 80
Liquid phase Purified Purified water Purified water Purified Purified Purified water Purified water Purified water Purified water
water water water
DEVELOPMENT OF A TASTE-MASKED GENERIC IBUPROFEN SUSPENSION QUERY

JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 100, NO. 10, OCTOBER 2011

4463
4464 WOERTZ ET AL.

Table 3. Composition of Prototype Generic Suspensions According to Nuro 2% O

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
remaining sensors, as recommended by the supplier,
to enable conditioning of the sensors. The method was
validated as described in a previous paper.16
Evaluation of Results
The results were recorded as raw data (mV) of the
relative measurement of the sample to the reference.
Sensor signal results were either evaluated directly or
multivariate data analysis was performed. For mul-
tivariate data analysis, raw data were pretreated by
mean centering and scaling to unit variance. Data
processing, graphical illustration, and statistical in-
terpretation of the results were carried out using
Excel 2007 (Microsoft, Redmond, Washington) and
SIMCA-P+ v11.5 (Umetrics AB, Umeå, Sweden). To
determine the distance between two samples (p, q) af-
ter multivariate data analysis, Euclidean distances
were calculated including all the principal compo-
nents (PCs) (n) built for the model.
camera D300 (Nikon, Tokyo, Japan) under polarized
light.
Laser Light Diffraction
The median (x50 ) of ibuprofen particles was deter-
mined by a laser diffraction system (Helos/KF-Magic,
Sympatec GmbH, Clausthal-Zellerfeld, Germany) via
wet dispersion method using a 50 mL cuvette. One
drop of ibuprofen suspension was stirred in a satu-
rated solution of ibuprofen for 5 min at 1000 rpm.
Measurements were performed in triplicate.
RESULTS
Screening of Ibuprofen Marketed Suspensions
Three proprietary (Nuro 2% O, Nuro 2% S, and Nuro
4% S) and 11 generic ibuprofen suspensions (2% and
4%) were evaluated for pH, particle size, dynamic vis-
cosity, and taste properties.

pH Measurements
 n

d(p, q) =  (p i − qi )2
i=1
pH Measurement
The pH of all the suspensions was determined by a
digital Knick pH-Meter 766 Calimatic (Knick GmbH
& Company, Berlin, Germany).
Viscosity Measurement
The viscosity of all the suspensions was determined
with a Kinexus rheometer (Malvern Instruments
Ltd., Worcestershire, United Kingdom) at shear rates
of 0.1/s (representing storage conditions) and 100/s
(representing stress while swallowing).
Microscopy
Microscopic images of solid ibuprofen particles in sus-
pension were taken with a light microscope DMLB
(Leica Microsystems, Wetzlar, Germany) and a digital
(3) The pH of the marketed ibuprofen suspension varied
between 3.28 for Ibuflam 2% and 4.7 for Nuro 2% O
(Table 4). A correlation between the univariate sensor
responses of the electronic tongue and the pH of the
formulations could not be found. For example, Dolor
2% and Dolor 4% were evaluated with different taste
properties by all sensors while the pH of both products
was approximately 4.0.
Ibuprofen Particles
Products from the proprietary formulations as well
as all generic products consisted of ibuprofen drug
particles (x50 = 20–100 :m) dispersed in the suspen-
sion vehicle, except the Ibuflam 2% suspension. The
presence of cellulose acetate phthalate and ethylcel-
lulose in this formulation indicated that a coating of
the drug particles was performed during manufactur-
ing to improve the taste of the formulation (Table 2).
Microscopic images (Fig. 1) of the suspension illus-
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

Table 4. pH and Viscosity of Ibuprofen Suspensions

Available on the German Market

Viscosity [Pa s]

Product pH 0.1/s 100/s

Nuro 2% O 4.70 34.6 0.2
Nuro 2% S 4.50 38.5 0.2
Nuro 4% S 4.20 22.9 0.2
Dolor 2% 4.04 13.6 0.2
Dolor 4% 4.01 13.6 0.2
Ratio 2% 3.51 47.1 0.2
Ratio 4% 3.50 70.7 0.3
CT 2% 3.46 48.0 0.2
CT 4% 3.48 34.6 0.2
AbZ 2% 3.46 68.3 0.3
AbZ 4% 3.52 45.6 0.2
Ibudolor 2% 3.52 12.3 0.2
AL 2% 3.47 9.3 0.1
Ibuflam 2% 3.28 9.0 0.2

small drug crystals were observed for the proprietary

formulation (Fig. 1b), larger noncrystalline particles
were seen next to some small drug crystals for the Ibu-
flam 2% suspension (Fig. 1a). However, the presence
of the ibuprofen drug crystals leads to the assump-
tion that little of the ibuprofen was dissolved in the
suspension.
Viscosity
The viscosity of the different ibuprofen suspensions
was tested at different shear rates to simulate stor-
age conditions as well as mechanical stress (Table 4).
All suspensions were subject to decreasing viscosity
(0.1–0.3 Pa s) at higher shear rates and immediate
reconstruction after shearing. For low shear rates,
the viscosities of suspensions differed between 9 and
70 Pa s. However, no correlation between electronic
tongue sensor responses and the different viscosities
was found.
Electronic Tongue Measurements
It is generally recommended by the supplier of the
electronic tongue to remove particles from the liquid
vehicle because the sensor membranes could be dam-
aged by the mechanical stress caused by the particles.
This removal of particles should not have an impact
on the reliability of taste assessment, as the interac-
tion with taste buds on the human tongue happens
with dissolved molecules. Therefore, only the amount
of drug constituting the saturated solution is respon-
sible for taste transmission itself. However, the pres-
ence of particles can influence the overall palatability
of the formulation as the mouthfeel might be differ-
ent. But, this difference in texture is not detected by
the electronic tongue sensors. Therefore, the influence
of particles should be tested by other methodologies
but can be neglected for electronic tongue measure-
ments. Because of the high level of stabilization of
Figure 1. Microscopic pictures of ibuprofen suspensions:
Ibuflam 2% (a) and Nuro 2% O (b).
some marketed suspensions, particles could not be
removed from the suspension base, neither by cen-
trifugation nor by filtration. Thus, all formulations
were investigated in their original composition. As
a result, particles surprisingly did not influence the
stability or functionality of the sensor membranes,
and standard deviations between the four measure-
ment cycles were in an acceptable range (0.03–15 mV
within a range of sensor response values of 20 to
−140 mV), the measurement time was not prolonged
due to further stability checks, and the sensor check
was passed again after measurement of the suspen-
sions. Figure 2 shows a principal component analysis
(PCA) map representing the sensor signal patterns
of the 14 different ibuprofen suspensions, demineral-
ized water, a saturated ibuprofen solution, and 1 mM
solution of quinine hydrochloride as an external stan-
dard. Measurements were normalized to the quinine
hydrochloride standard to improve comparability of
the results. Because of the different characteristics
of the samples, information was distributed by both
PCs. Although PC-1 represented about 53% of the to-
tal information, 23% of the information was displayed
by PC-2. The remaining information was explained

DOI 10.1002/jps JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 100, NO. 10, OCTOBER 2011
4466 WOERTZ ET AL.
Figure 2. Screening of taste properties of 14 ibuprofen suspension available on the German
market; PCA map built by sensors: bitterness 1 (+ aftertaste), sweetness, sourness, saltiness,
umami (+ aftertaste), and astringency (+ aftertaste).
by PC-3 (∼14%) and PC-4 (∼6%), which is displayed
here because they did not reveal further differences
between the formulations The main differentiation,
therefore, between the samples has to be carried out
with respect to the x-axis (PC-1), with additional in-
formation gained by the y-axis (PC-2). All formula-
tions were detected as being different from the pure
ibuprofen solution. In addition, three clusters exist
on the PCA map representing the main differences
between the suspensions.
Development of a Generic Ibuprofen Suspension
(Top-Down Approach)
Figure 3 shows sensor responses of prototype formu-
lations F1–F7 in comparison to the proprietary formu-
lation Nuro 2% O normalized to the response of the
proprietary formulation used in every measurement
to assure comparability of the results. It can be seen
how the influence of excipient variation was clearly
reflected by the sensor responses. Although formu-
lation F1 containing the highest amount of sodium
chloride and sodium citrate appeared to be quite sim-
ilar to the original formulation, the removal of these
substances lead to remarkable changes in sensor re-
sponses (F2–F3). Furthermore, the varying amount
of sodium saccharin and maltitol was seen by the
differing sensor responses to formulations F4–F7. It
was further seen that sensors umami, saltiness, sour-
ness, bitterness 1, and bitterness 2 were most sensi-
tive to changes in the sodium chloride/sodium citrate
amount, whereas sensors umami, saltiness, sourness,
sweetness, bitterness 1, and bitterness 3 detected the
changes in maltitol content. The variation of sodium
saccharin had an influence on all sensors except sen-
sor bitterness 3.
Although the sensor responses of most sensors var-
ied for the formulations F1–F6 compared with the
proprietary formulation, the difference between for-
mulation F7 and the original formulation was smaller
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 100, NO. 10, OCTOBER 2011
than 11 mV for each sensor. Figure 4 additionally
shows the differences between the prototype formula-
tions and their corresponding suspension bases after
PCA, including all sensors. About 85% of the propri-
etary formulation could be expressed by the first two
components (PC-1 = 66%, PC-2 = 19%), whereas the
remaining part is explained by PC-3 (∼11%) and PC-4
(∼2%). As expected, the remaining 13% did not fur-
ther affect the main results and were therefore not
displayed here. Again, it becomes obvious that F1
and F7 were closest to the proprietary formulation
Nuro 2% O. Distances calculated on the basis of the
four PCs showed the order of similarity of prototype
formulations with respect to the proprietary formula-
tion: Nuro 2% O < F7 < F1 < F6 < F3 < F4 < F5 <
F2. In addition, formulation F1 approached the pro-
prietary formulation quite well, but revealed a large
distance to its suspension base at the same time (Ta-
ble 5). The pH of the prototype formulations decreased
with the removal of sodium chloride and sodium cit-
rate and increased with the addition of sweetening
agents ending up with the same pH compared with
the proprietary formulation Nuro 2% O (Tables 4
and 6).
DISCUSSION
Evaluation of Marketed Products
The results from the electronic tongue measurements
clearly showed that the formulations could be dis-
tinguished according to their excipients and man-
ufacturer (Fig. 2). One group consists of the three
Nurofen products. The clear differentiation to the re-
maining suspensions can be explained by the pres-
ence of sodium chloride and sodium citrate for sus-
pension stabilization. In addition, sodium ions are
known for masking bitter taste.17 The difference be-
tween Nurofen 2% and Nurofen 4% can be explained
DOI 10.1002/jps
 DEVELOPMENT OF A TASTE-MASKED GENERIC IBUPROFEN SUSPENSION QUERY 4467

Figure 3. Approaching the original product (Nuro 2% O) by variation of the quantitative

composition (F1–F7). Univariate data evaluation.

by the different preservative used as well as the pres-
ence of thaumatin. The second group consists of the
two “Dolormin” formulations, which were detected as
being similar with respect to the y-axis, and different
with respect to the x-axis. The formation of the cluster
can be explained by the different preservative and sta-
bilizing agent compared with the other formulations,
whereas the difference between the 2% and 4% sus-
pension can be explained by the different sweetening
agents used for the suspension vehicle. The last and
biggest group is represented by the remaining for-
mulations showing a marginal further subdivision.
The right part of the group consists of suspensions
containing exactly the same excipients marketed by
different companies, whereas on the left side formu-
lations, AL 2%, Ibuflam 2%, and Ibudolor 2% can be
found. The reason for this is the different nature of
sweetening agents used for these formulations. The
formulation containing coated particles, Ibuflam 2%,
also contained pure ibuprofen crystals (Fig. 1), lead-
ing to the assumption of the presence of dissolved
ibuprofen as well. Furthermore, only a small differ-
ence was found between suspensions containing 2%
or 4% ibuprofen within the same suspension vehicle.
This showed that the presence of particles did not
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.

Figure 4. Approaching the original product (Nuro 2% O) by variation of the quantitative

composition (F1–F7 and B1–B7). Multivariate data evaluation; PCA map built by sensors:
bitterness 1, 2, 3, sweetness, sourness, saltiness, umami, and astringency.

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
Figure 5. Proposed top-down approach for rational de-
velopment of taste-masked oral liquids using electronic
tongues.
data analysis showed the importance of the concen-
tration level of these substances. Although formula-
tion F1 had a similar sensor response to the pro-
prietary formulation, formulation F2, containing no
sodium chloride and sodium citrate, appeared to have
the largest PCA distance. However, the distance of the
formulation F1 to its corresponding placebo was com-
parably high, leading to the assumption that sodium
chloride and sodium citrate influenced the solubility
of the ibuprofen over time. Therefore, a concentration
level in between was chosen for all other formulations
(F3–F7). For formulation F3, mV values of sensors
umami, saltiness, sourness, and astringency were al-
ready close to the proprietary formulation, whereas
sensor responses of sensors for sweetness and bitter-
ness were still different. By variation of the concen-
tration level of sodium saccharin and maltitol, sensor
responses became similar to the proprietary formula-
tion and it became obvious that both substances were
involved in the sensor’s responses. The ionic structure
of sodium saccharin mainly influenced the sensor re-
sponses (F4 vs. F6), but the addition of maltitol was
necessary as well (F3 vs. F6). Therefore, a high level
of both sweeteners was needed to finally obtain a sim-
ilar sensor signal pattern compared with the propri-
etary formulation (F7). The small remaining distance
between formulation F7 and the proprietary formula-
tion showed that formulation F7 still did not exactly
match the sensor signal patterns of the proprietary
formulation. However, this might be explained by the
difference in excipient origin, for example, orange fla-
vor, leading to slightly different sensor responses.
CONCLUSION
A new optimized generic ibuprofen suspension with
similar sensor signal patterns was developed starting
from an already existing taste-masked brand product.
To strengthen these results, a taste test of the new
generic formulation should be performed by a trained
DOI 10.1002/jps
4469
human taste panel in the future. However, reliability
and validity of electronic tongue data have already
been shown in previous studies.11,18 Therefore, this
newly introduced protocol offers the possibility of de-
veloping a pleasant tasting generic formulation based
on electronic tongue sensor signal patterns according
to a rational top-down approach (Fig. 5). The influ-
ence of excipient variation on sensor response values
is therefore one major prerequisite to apply this ap-
proach. In addition, excipients influencing the taste of
a formulation should be detectable by the electronic
tongue sensors. If these conditions are fulfilled, the
creation of a larger prediction set containing data
on various excipients would be possible to rationally
break down an existing formulation without the need
of human taste testing. Another opportunity would be
to determine the optimal taste-masking excipient con-
centration for the particular API and the most appro-
priate excipient with the lowest concentration could
be chosen with the benefit to reduce the amount of
potential harmful excipients and to save costs.
ACKNOWLEDGMENTS
This project has been financially supported by AIF
(Arbeitsgemeinschaft industrieller Forschungsvere-
inigungen “Otto von Guerike” e. V.; project no. 15980),
which is gratefully acknowledged. Our further grat-
itude goes to Insent, Atsugi-chi, Japan and TecLabS
Europe, Essen, Germany for providing technical sup-
port and advice.
REFERENCES
1. Mennella JA, Beauchamp GK. 2008. Optimizing oral medica-
tions for children. Clin Ther 30:2120–2132.
2. Davies EH, Tuleu C. 2008. Medicines for children: A matter of
taste. J Pediatr 153:599–604.
3. Woertz K, Tissen C, Kleinebudde P, Breitkreutz J.
Taste sensing systems (electronic tongues) for pharma-
ceutical applications. Int J Pharm (in press) doi:10.1016/
j.ijpharm.2010.11.028.
4. Vlasov Y, Legin A, Rudnitskaya A. 2002. Electronic
tongues and their analytical application. Anal Bioanal Chem
373:136–146.
5. Kobayashi Y, Habara M, Ikezazki H, Chen R, Naito Y, Toko
K. 2010. Advanced taste sensors based on artificial lipids with
global selectivity to basic taste qualities and high correlation
to sensory scores. Sensors 10:3411–3443.
6. Uchida T, Tanigake A, Miyanaga Y, Matsuyama K, Kunitomo
M, Kobayashi Y, Ikezaki H, Taniguchi A. 2003. Evaluation
of the bitterness of antibiotics using a taste sensor. J Pharm
Pharmacol 55:1479–1485.
7. Kayumba PC, Huyghebaert N, Cordella C, Ntawukuliryayo
JD, Vervaet C, Remon JP. 2007. Quinine sulphate pellets
for flexible pediatric drug dosing: Formulation development
and evaluation of taste-masking efficiency using the electronic
tongue. Eur J Pharm Biopharm 66:460–465.
8. Li L, Naini V, Ahmed SU. 2007. Utilization of a modified
special-cubic design and an electronic tongue for bitterness
masking formulation optimization. J Pharm Sci 96:2723–2734.
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,
Ohwaki T. 2010. A new method for evaluating bitterness of
medicines in development using a taste sensor and a dis-
integration testing apparatus. Chem Pharm Bull 58:1009–
1014.
10. Abdullah MZ, Rahman ASA, Shakaff AYM, Noor AM. 2004.
Discrimination and classification of eurycoma longifolia jack
in medicinal foods by means of a DSP-based electronic taste
sensor. Trans Inst Meas Control 26:19–39.
11. Woertz K, Tissen C, Kleinebudde P, Breitkreutz J. 2010. Ra-
tional development of taste masked liquids guided by an elec-
tronic tongue. Int J Pharm 400:114–123.
12. Tokuyama E, Matsunaga C, Yoshida K, Mifsud J-C, Irie T,
Yoshida M, Uchida T. 2009. Famotidine orally disintegrating
tablets: Bitterness comparison of original and generic prod-
ucts. Chem Pharm Bull 57:382–387.
13. Neubert A, Verhamme K, Murray ML, Picelli G, Hsia Y, Sen
FE, Giaquinto C, Ceci A, Sturkenboom M, Wong ICK, TEDDY
Network of Excellence. 2010. The prescribing of analgesics and
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 100, NO. 10, OCTOBER 2011
non-steroidal anti-inflammatory drugs in paediatric primary
care in the UK, Italy and the Netherlands. Pharmacol Res
62:243–248.
14. Reader S, Shaw H, Hails S 2006. A taste-testing study in
healthy volunteers (children) to investigate children’s pref-
erence for ibuprofen or placebo suspension. Paediatr Perinat
Drug Ther 7:54–58.
15. Ayenew Z, Puri V, Kumar L, Bansal AK. 2009. Trends in phar-
maceutical taste masking technologies: A patent review. Re-
cent Pat Drug Delivery Formulation 3:26–39.
16. Woertz K, Tissen C, Kleinebudde P, Breitkreutz J. 2010. Per-
formance qualification of an electronic tongue based on ICH
guideline Q2. J Pharm Biomed Anal 51:497–506.
17. Keast RSJ, Breslin PAS. 2002. Modifying the bitterness of
selected oral pharmaceuticals with cation and anion series of
salts. Pharm Res 19:1019–1026.
18. Krause J. 2008. Novel paediatric formulations for the
drug sodium benzoate. Ph.D. Thesis. Duesseldorf, Germany:
Heinrich-Heine-Univerity.
DOI 10.1002/jps