296 JOURNAL OF COSMETIC SCIENCE
APPLICABILITY IN DIFFERENT MATRICES
Apart from emulsions A and B, additional glucose-free formulations of different viscosity,
transparency, conductivity and pH, “shampoo A,” and “gel A,” were spiked with glucose
between 0 and 7*10−3% w/w. The ingredients present in the two formulations at
concentrations above 0.001% w/w can be seen in Table I, together with certain properties of the
formulations. Obtained calibration curves showed good linearity (R2 =0.994 for shampoo
A, (Figure 2, right) and R2 =0.989 for gel A). Back-calculated glucose concentrations in
the standards were acceptable (within ±14.1% of their nominal concentration for shampoo
A standards) and marginally acceptable (within ±41.0% of their nominal concentration)
for gel A standards.
Our findings altogether, support the applicability of the proposed quantification approach
in different cosmetic matrices irrespective of their transparency, oil composition (up to
27.9% w/w oil), pH (5.4 – 6.8), viscosity (3 – 1,000k mPa), and conductivity (2.92 – 14.56
mS/cm). It is to be noted that the GOD/POD enzymatic system remains functional even
at rather extreme conditions in cosmetic formulations (such as a surfactant concentration of
∼20% w/w (shampoo A, Table I) and ethanol concentration of 62% w/w (gel A, Table I).
Known protein denaturants (such as urea, surfactants) commonly present in cosmetic
formulations have been shown to diminish the activity of at least one of the enzymes: For
example, rather mediocre GOD activity inhibition was reported (by less than 10% to 25%
at 25°C and pH 6.4) in the presence of urea and the anionic surfactant sodium n-dodecyl
sulfate (at ≤2% w/w, each).20 Much more significant was the inhibition in the presence
at ≤2% w/w of the cationic surfactant Dodecyl Trimethyl Ammonium Bromide.20 We
did not observe a significant activity compromise under the conditions employed (total
Table V
Technical Parameters for the Quantitative Determination of Glucose With the Proposed Methodology in
Emulsion A
Intermediate precision Accuracy Linearity
Mean [Glu],
10−3% w/w
CV %Measured [Glu],
10−3% w/w
Spiked [Glu],
10−3% w/w
Bias %R2 Linear range,
10−3% w/w
1.52 11.56 (n =3) 1.52 (n =3) 1.37 +11.06
≥0.991 (n =3) 0.00 – 6.37
4.75 1.68 (n =3) 4.75 (n =3) 4.55 +4.37
CV: coefficient of variation [Glu]: Glucose concentration n: number of replicates R2: correlation coefficient.
Table IV
Parameters of the Linearity of the Response Between Measured Signal and Glucose Concentration
(Equation, Pearson’s Coefficient of Determination (R2)) Upon Monitoring Blue Channel Output Alteration
with Time After Processing
Time After Initiation of Processing of Last Sample (Min) Linear Equation R2
5 –8206 × +185.1 0.997
9 –9581 × +180.4 0.981
13 –10121 × +180.3 0.976
21 –10665 × +179.8 0.977
26 –10834 × +182.4 0.962
60 –10529 × +170.8 0.945
R2: correlation coefficient.

297 IN SITU ANALYSIS OF GLUCOSE IN COSMETIC FORMULATIONS
surfactant concentration of up to ∼20% w/w, urea concentration of 15% w/w, ethanol of
62% w/w (Table I)). It might be possible that the presence of polymers in the formulations
offers a certain stability to the enzymatic system, despite the presence of denaturants.21 The
above provide an indication of the applicability potential of the assay to a good variety of
cosmetic matrices.
ACCURACY UPON ANALYSIS OF COMMERCIAL PRODUCTS
Mock glucose-containing formulations were then prepared in glucose-free commercial
creams (emulsion B, emulsion C (O/W, brown-colored hand and foot cream) and emulsion
D (O/W, blue-colored body butter)). The concentration of spiked glucose in each matrix,
as well as the color/transparency of each matrix, is indicated in Table VI. All mock
glucose-containing formulations were then treated as “unknown” samples. The mock
unknown samples were either analyzed directly, or were diluted appropriately, to lower
glucose levels within the linear region of the method (dilution indicated in Table VI).
As diluent, untreated emulsion B (glucose-free) was used. They were then submitted
to the standard addition protocol adapted to the proposed analysis format. For spiked
concentration calculation with the standard addition method, the blue channel output
of emulsion B treated with the WR was taken as zero-analyte signal, in the case of the
1:1,000 and 1:10,000 dilutions.
For the 1:40 and 1:10 dilutions, emulsion B treated with the WR did not provide a suitable
“zero signal” since there was a significant effect of the colored matrix. Instead, extrapolation
to nonzero-analyte concentration (a standard containing 0.001 – 0.005% w/w glucose in
the corresponding matrix) was used. A good linear relationship was obtained between light
intensity and glucose concentration in emulsion, as indicated by the R2 values displayed in
Table VI. Our results also demonstrate an acceptable bias. An image of the colored wells
upon application of the standard addition method is given in Figure 3, as well as of the
initial emulsion.
These data provide an initial indication in favor of the validity of the approach for
quantifying glucose content in commercial cosmetic matrices within the frame of QC.
They also demonstrate that the proposed analysis format successfully takes into account
Table VI
Technical Parameters for the Quantitative Determination of Glucose in Commercial Products Using the
Standard Addition Procedure
Matrix Color Transparency Dilution applied Spiked [Glu], %w/w R2 Bias %
emulsion B (sensitive
skin cream)
White opaque Nondiluted 0.0091 0.993 −15.3
emulsion C (hand and
foot cream)
Brown opaque 1:1,000 3.12 0.997 +4.9
emulsion C (hand and
foot cream)
Brown opaque 1:40 0.08 0.991 +3.2
emulsion C (hand and
foot cream)
Brown opaque 1:10 0.02 0.999 +13.12
emulsion D (body
butter)
Blue opaque 1:10,000 8.26 0.995 +15.9
R2: correlation coefficient.