295 IN SITU ANALYSIS OF GLUCOSE IN COSMETIC FORMULATIONS
WR/g emulsion to each preparation, in the order of decreasing glucose concentration. The
blue intensity component of the pictures of the microstrips was monitored. Parameters of
the linearity of the response between measured signal and glucose concentration (slope,
Pearson’s coefficient of determination (R2)) are displayed in Table IV. From Table IV data it
is evident that optimum time point for picture capturing is at five minutes (after reaction
initiation in the last treated preparation). For a limited time after that (no longer than
20 minutes), the linearity of response remains satisfactory. Within this period, sensitivity
increases with time (increase in negative slope of curve) at the compromise of the goodness
of fit, however. In all experiments that follow, pictures were taken at around five minutes
after processing initiation of the last sample/standard. In any case, pictures were taken no
longer than 25 minutes from the beginning of processing of the entire series of standards/
samples.
METHOD VALIDATION IN AN O/W EMULSION
The proposed analysis format was then validated with respect to the useful analytical
range, and its reproducibility and accuracy were determined upon analysis of two QCs in
emulsion A (Table V). The values of coefficient of variation and bias for the two QCs indicate
acceptable technical characteristics for the specific analysis purpose in a cosmetic emulsion.
For all standards (other than the zero standard), standard concentration was back-calculated
within ±20.5% of their nominal concentration in emulsion A. These technical parameters
support the reliability of the methodology for the extraction-free glucose quantification
in emulsion A, with a lower limit of linearity of 0.91*10−3% w/w. When compared to
chromatographic methods for the analysis of glucose extracted from various matrices, the
proposed method is superior to certain such methods, or inferior to others in terms of
lower limit of linearity12, which in any case is very sufficient here, for the analytical needs
of the cosmetic industry. The proposed method is however inferior in terms of precision or
accuracy when compared to literature chromatographic methods.18,19
Figure 2. Linear relationship between color intensity and glucose concentration in emulsion B (left) or
shampoo A (right) standards. Below each graph, a picture of the corresponding wells is given, which was used
to extract the graph data (where glucose concentration is indicated (in 10−3% w/w)).

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.