Skin Research and Technology 2016; 22: 276–283
Printed in Singapore
All rights reserved
doi: 10.1111/srt.12258
Classification by causes of dark circles and appropriate
evaluation method of dark circles
S. R. Park1, H. J. Kim1, H. K. Park1, J. Y. Kim1, N. S. Kim1, K. S. Byun1, T. K. Moon1,
J. W. Byun2, J. H. Moon2 and G. S. Choi2
1
Ellead Skin and Bio Research, Gyeonggi-do, Korea and 2Department of Dermatology, Inha University College of Medicine, Incheon, Korea
Background: Dark circles refer to a symptom that present
darkness under the eyes. Because of improvement in the qual-
ity of life, the dark circles have been recognized as one of
major cosmetic concerns. However, it is not easy to classify
the dark circles because they have various causes.
Methods: To select suitable instruments and detailed evalua-
tion items, the dark circles were classified according to the
causes through visual assessment, Wood’s lamp test, and
medical history survey for 100 subjects with dark circles. After
the classification, were newly recruited for instrument confor-
mity assessment. Through this, suitable instruments for dark
circle evaluation were selected. We performed a randomized
clinical trial for dark circles, a placebo-controlled double-blind
study, using effective parameters of the instruments selected
from the preliminary test.
Results: Dark circles of vascular type (35%) and mixed type
(54%), a combination of pigmented and vascular types, were
the most common. Twenty four subjects with the mixed type
dark circles applied the test product (Vitamin C 3%, Vitamin A
0.1%, Vitamin E 0.5%) and placebo on randomized split-face
“dark circles” is not a medical term
T HE TERM
but is being used by many people in gen-
eral to indicate pigment muscle under the eyes
(1). Dark circles can be caused by a variety of
conditions such as infection, inflammation,
allergies, and lifestyle factors. Possible causative
factors of the dark circles include excessive pig-
mentation; thin and translucent lower eye-lid
skin overlying the orbicularis oculi muscle; and
shadowing due to skin laxity and tear trough
(2).
The reason why the skin around the eye
appears dark is because the skin tissues around
the eyes are very thin. As the skin ages, the
lipid layer under the eyes sags and bulges out
to create shadows and make the area darker.
Excessive pigmentation under the eyes can be a
frequent cause of dark circles (3). Watanabe
276
© 2015 John Wiley & Sons A/S.
Published by John Wiley & Sons Ltd
Skin Research and Technology
for 8 weeks. The effective parameters (L*, a, M.I., E.I., quasi
L*, quasi a* and dermal thickness) were measured during the
study period. Result showed that the L* value of Chromame-
terâ, Melanin index (M.I.) of Mexameterâ and quasi L* value
obtained by image analysis improved with statistical signifi-
cance after applying the test product compared with the pla-
cebo product.
Conclusion: We classified the dark circles according to the
causes of the dark circles and verified the reliability of the
parameter obtained by the instrument conformity assessment
used in this study through the efficacy evaluation. Also based
on this study, we were to suggest newly established methods
which can be applied to the evaluation of efficacy of functional
cosmetics for dark circles.
Key words: dark circles – causes – classification – evaluation
Ó 2015 John Wiley & Sons A/S. Published by John
Wiley & Sons Ltd
Accepted for publication 29 July 2015
et al. studied periorbital biopsies of 12 Japanese
patients with dark circles, showing that all of
them had dermal melanosis via histology (1).
This is a common symptom frequently
described as tired eyes. Although dark circles
may be the result of facial aging, they can affect
both men and women of all ages and have a
variety of causes. They can be a significant cos-
metic concern, and many individuals seek treat-
ment for this common condition. There are
various cosmetic products and dermatological
treatments for those who are concerned about
the dark circles. Despite its prevalence, there
are few published studies about dark circles
and its pathogenesis (4). As the individual
anatomical basis for dark circles is not clearly
defined and it is not easy to classify the dark
circles because they have various causes.

The purpose of this study is to classify the
dark circles according to the causes and then
set up the objective methods for optimized eval-
uation of the dark circles. To select suitable
instruments and detailed evaluation items, the
dark circles were classified according to the
cause through visual assessment, Wood’s lamp
test and medical history survey for 100 subjects
with dark circles. We performed efficacy evalu-
ation of the dark circles after the pre-test for
classification by causes of dark circles. The flow
chart of the study design is shown in Fig. 1.
Materials and Methods
Conformity assessment through pre-test
The subjects were consisted of 100 healthy
Korean females (mean age, 41.3 years). The
study was conducted after obtaining Indepen-
dent Ethics Committee’s approval. Volunteers
aged between 19 and 56 years of age, with dark
circles and those who gave written informed con-
sent were enrolled in the study. To select the suit-
able instruments and detailed evaluation items,
the dark circles were classified according to the
cause through visual assessment, Wood’s lamp
test, and medical history survey. After the classi-
fication of dark circles by causes, we measured
the dark circle sites to identify any differences
between the dark circle and normal groups.
Measurement and instruments
Prior to the study, the subjects stood by at least
for 30 min in a controlled room at constant tem-
perature and humidity (20–24°C, 40–60% RH).
The subject’s frontal face was photographed
by Visia-CR (Canfield, Fairfield, NJ, USA) which
is equipped with chin supports and forehead
clamps that fix the face during the photograph-
ing process and maintain a fixed distance
between the subjects and the camera at all times.
Using the images from Visia-CR, the dark circles
9DVFXODUW\SH


3LJPHQWHGW\SH

0L[HGW\SH

Evaluation method of dark circles
were evaluated by the image analysis program
(Image-Pro Plus, Rockville, MD, USA). The quasi
L*, and quasi a* values were derived from the
Red, Green, and Blue values. The L* and a*
parameters were obtained from under the eyes
using Chromameter CR-400 (Minolta, Tokyo,
Japan). Mexameter MX18 (Courage and Khaz-
aka, Electronic GmbH, Cologne, Germany) was
used to measure the melanin index (M.I.) and
erythema index (E.I.) under the eyes. The dermal
density and skin thickness were analyzed using
Dermascan-C (Cortex Technology, Hadsund,
Denmark), a high resolution imaging machine
using 20 MHz supersonic waves which is very
useful in non-invasively observing changes to
the inner layers of the skin. The cutaneous blood
flow under the eyes was measured using Laser
Doppler Perfusion Imager (LDPI, PeriScan PIM
II; Perimed AB, J€arf€alla-Stockholm, Sweden) with
fixed conditions; blocking off all the lights and
analyzing by image analysis software (LDPIwin
version 2.6, Perimed AB).
Efficacy evaluation
This step was to confirm the appropriate evalu-
ation method of dark circles using the effica-
cious material. Based on the conformity test, we
performed efficacy evaluation on largest group,
dark circles of mixed type having both the vas-
cular and pigment components with the mix-
ture of retinol, vitamin C, and E. Twenty-four
healthy female patients between 20 and
59 years of age and clinically diagnosed with
dark circles of mixed type by dermatologist,
participated voluntarily in the clinical study.
We use the selected parameters from the pre-
liminary test for efficacy evaluation. The volun-
teers with dark circles applied the test product
(Vitamin C 3%, Vitamin A 0.1%, Vitamin E
0.5%) and placebo on randomized split-face
with double blind for 8 weeks. During the test-
ing period, the subjects visited the laboratory
three times (baseline, 4 weeks, and 8 weeks).
The subjects used the test products and placebo
every morning and night on under-eye area.
The test product was randomly applied on the
either of the half face (left or right) for each
subject. They were also instructed to avoid
9DVFXODU3LJPHQWHG
exposure to the sun, excessive exercise, and
(WF(\HEDJ
drinking alcohol prior to each visit. Visual
assessment was performed using our dark
Fig. 1. Classification of subjects according to causes of dark circles. circles grading scale from 0 to 5 (Fig. 2). To ver-

277
Park et al.

ify the reliability of the evaluation methods of
dark circles, the validation of the identically
designed trial was conducted at dermatology of
Inha University hospital.
Statistical analysis
The differences among three groups by cause of
dark circles (vascular type, mixed type, and
normal) were analysed using one-way ANOVA,
Welch test and Kruscal–Wallis test, with post-
hoc analysis based on Scheffe, Games-Howell,
and Mann–Whitney (Bonferrony correction)
post-hoc comparisons. Contrast test was per-
formed by repeated measures ANOVA (RM-ANOVA)
to compare the changes between the product
application side and the placebo application
side at different time points. RM-ANOVA was also
used to compare the changes of each site
between every time point. The inter-class corre-
lation coefficients were then calculated from the
between-reader error variance and the total
variance at visual assessment. Also, Spearman’s
q was used to correlate each parameter. Signifi-
cance was measured at P < 0.05.
Result
Conformity assessment
Classification of dark circles by causes
As a result of the classification of dark circles
among 100 females, the vascular type (35%) and
the mixed type (54%), a combination of pig-
mented and vascular dark circles, were most
common. The pigmented type, without vascular
component and structural type such as eye bags
accounted for 5% and 6% respectively (Fig. 1).
Therefore, to examine which equipment is
appropriate for efficacy evaluation of dark cir-
cles, we conducted preliminary test on the sub-
jects with the mixed type, vascular type, and
those without dark circles. Subjects classified as
having vascular type had dark circles below the
eyes that appeared red/purple. The tone was
darker inside than outside. When the skin was
pulled, the purplish tone tended to become dar-

Characteristics of Dark circles grade (0-5)

0: Dark circle is not noticeable
1: Faint purple color is observed on the front and underneath the eyes. When the skin is pulled
with hand, the color darkens so that one can tell it is a dark circle.
2: Even the skin is not pulled, purple shadow is clearly observable. Slight pigmentation may
instead be observed (or both).
3: Deep purple shadows are observed underneath the eyes/moderate pigmentation is partially
observed (or both).
4: Overall, uniform and dark pigmentation is observed and occasionally it can be observed on
both above and underneath the eyes.
5: Uniform and very dark pigmentation is observed on both above and underneath the eyes.

Fig. 2. Dark circles grade and description.

278

ker, but did not become any more obvious
under Wood’s lamp. Dark circles which
appeared brown/gray were observed from the
subjects classified as having pigmented dark
circles. Dark circles became more distinguish-
able under Wood’s lamp. Subjects with the
mixed type showed the tone of dark circles sim-
ilar to the vascular dark circles, which appeared
red/purple. Solar lentigo and non specific pig-
mentation were also observable on dark circle
area. Under Wood’s lamp, pigmentation and
blemish around the dark circles became more
obvious (Fig. 3).
Evaluation of instrument suitability
Conformity assessment was performed on the
subjects with the most frequent patterns, the
vascular type (22 females; mean age,
41.8 years), the mixed type (24 females; mean
age, 43.0 years), and those without dark circles
(21 females; mean age, 38.4 years), to select
appropriate index of the dark circle evaluation.
The homogeneity between each group was
established by checking the M.I. and E.I. of the
cheek area before actual test. There were no sig-
nificant differences in M.I. and E.I. between
each group. As indicated in Tables 1 and 2, M.I.
value of the mixed type was higher than that of
Fig. 3. Characteristics of subjects based on dark circles causes. Photographs obtained with Visia-CR.
Evaluation method of dark circles
the vascular type and that of the normal group
with statistical significance (P < 0.05); however,
there was no significant difference in the values
between the vascular type and the normal
group. For E.I. value, the value of the mixed
type was higher than that of the normal group
with statistical significance (P < 0.05), but there
was no significant difference in the values
between the vascular type and the normal
group. L* value of the mixed type was lower
than that of the vascular type and that of the
normal group with statistical significance
(P < 0.05) and the value of the vascular type
was lower than that of the normal group with
statistical significance (P < 0.05). For a* value,
the value of the mixed type was higher than
the normal group with statistical significance
(P < 0.05) however, there was no significant dif-
ference in the values between the vascular type
and that of the normal group (Table 2). The
quasi L* value of the mixed type was the
lowest in 3 groups and the quasi L* value of
the vascular type was lower than the normal
group with statistical significance (P < 0.05).
Quasi a* values showed the contrast tendency
against quasi L* value with statistical signifi-
cance (P < 0.05). The dermal thickness of the
normal group was significantly higher than that
279
Park et al.

TABLE 1. Mean values  SD of each parameter on dark circles area TABLE 3. Mean values and SD measured on dark circles during
among the groups treatment with the test product and placebo in split-face randomly for
8 weeks
Without dark
circles Mixed type Vascular type Test material Placebo

M.I. 112.03  20.00 172.41  42.63 126.28  26.62 Mean SD Mean SD F P

E.I. 238.74  29.59 287.70  43.60 265.06  48.71
L* 63.47  2.08 57.68  2.46 60.10  3.02 Visual grade
a 11.53  1.42 13.02  0.95 12.38  1.73 0 weeks 3.27 1.20 3.23 1.23 0.792 0.459
Quasi L* 67.14  2.59 58.92  3.92 62.64  4.29 4 weeks 3.18 1.18 3.14 1.25
Quasi a 14.71  2.18 18.93  2.15 16.86  3.22 8 weeks 3.05 1.29 3.09 1.23
Dermal thickness 1.48  0.16 1.35  0.14 1.33  0.15 M.I.
Dermal density 44.46  7.08 39.62  8.39 41.22  6.50 0 weeks 156.53 40.84 160.63 48.25 9.956 0.000*
Blood flow 1.23  0.50 1.22  0.34 1.14  0.42 4 weeks 154.40 38.44 161.35 45.80
8 weeks 151.20 35.87 163.09 42.99
E.I.
0 weeks 305.75 55.75 310.05 59.55 0.151 0.861
TABLE 2. The result of statistical analysis among the three groups 4 weeks 314.47 51.43 321.50 56.68
8 weeks 297.59 51.69 304.73 56.96
Without dark Without dark L*
circles vs. circles vs. Mixed type vs. 0 weeks 58.89 2.42 58.81 2.67 14.049 0.000*
mixed type vascular type vascular type 4 weeks 59.07 2.45 58.85 2.64
8 weeks 59.37 2.40 58.95 2.66
M.I. 0.000* 0.128 0.000* a*
E.I. 0.000† 0.125 0.192 0 weeks 12.59 1.45 12.55 1.40 0.156 0.856
L* 0.000* 0.000* 0.009* 4 weeks 12.61 1.42 12.53 1.41
a* 0.001† 0.197 0.287 8 weeks 12.57 1.47 12.44 1.64
Quasi L* 0.000‡ 0.000‡ 0.006‡ Quasi L*
Quasi a* 0.000* 0.038* 0.040* 0 weeks 59.58 4.29 59.48 4.37 5.391 0.008†
Dermal thickness 0.023† 0.009† 0.917 4 weeks 60.24 4.39 59.62 3.92
Dermal density 0.099 0.362 0.766 8 weeks 60.28 4.65 59.40 4.57
Blood flow 0.998 0.758 0.778 Quasi a*
*Welch post-hoc Games-Howell, P < 0.05. 0 weeks 18.59 2.79 18.95 2.65 0.044 0.957
†One-way ANOVA post-hoc Scheffe, P < 0.05. 4 weeks 18.65 2.80 19.01 2.70
‡Kruscal–Wallis post-hoc Mann–Whitney Bonferroni correction, 8 weeks 18.85 2.90 19.26 3.20
P < 0.017. Dermal thickness
0 weeks 1.39 0.19 1.38 0.15 14.471 0.000*
4 weeks 1.50 0.15 1.41 0.12
of the mixed type and the vascular type; how-
8 weeks 1.60 0.14 1.41 0.14
ever, no significant differences in the values
*One-way ANOVA post-hoc Scheffe, P < 0.001.
between the vascular type and the mixed type †One-way ANOVA post-hoc Scheffe, P < 0.05.
were observed. There were no significant differ-
ences in the dermal density among the three (P < 0.05) (Fig. 4a). On the other hand, no signifi-
groups. The results of blood flow showed no cant difference in change of E.I. value was
significant differences among the three groups. observed between the treated side and the con-
According to the results of the conformity test, trol side. In comparison of the changes between
the values of Mexameter, Chromameter, image the groups there were statistically significant
analysis, and Dermascan-C were selected for increases in the L* value and quasi L* value after
the efficacy evaluation of dark circles. 8 weeks on treated side (Fig. 4b,c). However,
there were no statistically significant differences
in a* value and quasi a* value between the trea-
Efficacy evaluation ted side and the control side. As a result, the der-
Efficacy evaluation test was done by formulating mal thickness of treated side increased more
a dark circle alleviating test product and evaluat- than the control side after 8 weeks with statistical
ing its efficacy via double-blind test. Table 3 significance (P < 0.05) (Fig. 4d). For M.I. value,
shows the results before and after application of significantly negative correlations were found
test product. The visual scorings of dark circles between L* value and quasi L* value. However,
on both the treated side and the control side were the dermal thickness showed no significant cor-
decreased; however, there was no significant dif- relations with other parameters (Table 4). To
ference between 2 groups. The M.I. value of the confirm the test results, Inha University hospital,
treated side decreased more than the control side the validation institution, performed the identi-
after 8 weeks with statistical significance cal study; the resulting pattern was the same.

280

Discussion
Hyperpigmentation and vascular stasis of eye-
lids have been considered as possible causes of
dark circles, in addition to accentuated shadows
created by complex light reflection at a concave,
sunken orbital area (5). The amount of blood in
the lower eyelid is larger and velocity of blood
flow is slower than that in the cheek. Therefore
stasis, or congestion, of cutaneous blood tends
to occur and dark circles may appear (6). More-
over, dermis in the eyelid displays lower
amount of collagen, elastin, and glycosamino-
glycan than dermis at other sites (5). The color
of cutaneous blood at the lower eyelid thus
seems to easily influence skin color. Ohshima
and Takiwaki reported that the mean value of
M.I. in thinner collagen gel samples were signif-
icantly higher than in thicker collagen gel sam-
ples in vitro dark circles model (7). Another
cause of dark circles is shadowing due to sun-
ken eyes from aging. But it is very difficult to
improve dark circles using cosmetics. The
causes of dark circles are numerous and yet
there has been no clear method to measure
them.
The result of this study showed that 89%
(vascular type 35%, mixed type 54%) of subjects
were founded to have vascular component in
their dark circles and only 6% was classified as
pigmented type without vascular component.
Also, large portion of the subjects of the age
over 30 have been classified into the mixed
type because they had developed melasma or
solar lentigo over vascular dark circles. Causes
of dark circles including dermal thinning and
pigmentation under eye became severe along
age. According to the results of the conformity
test, the parameter of pigmentation component
(L*, quasi L*, M.I.) and the parameter of vascu-
lar component (E.I, a*, quasi a*, dermal thick-
TABLE 4. Correlation among the instrumental assessments for efficacy
evaluation
R slower compared with that in the pigment
Parameter 0 weeks
L* vs. M.I. 0.794*
L* vs. Quasi L* 0.854*
M.I. vs. Quasi L* 0.742*
L* vs. Dermal thickness 0.347
M.I. vs. Dermal thickness 0.362
Quasi L* vs. Dermal thickness 0.21
*Correlation is significant at the 0.001 level.
Evaluation method of dark circles
ness) were significantly different between the
mixed type groups and the normal group. But
the vascular type showed significant difference
in the parameter (E.I, a*, quasi a*, dermal
thickness) related to vascular component
between the normal group. Parameter for pig-
mentation (L*, quasi L*, M.I.) and parameter
for vascular component (quasi a*,) are more
severe in mixed type group than the vascular
type group. It is considered that the average
age of the subjects of the mixed type was
higher than those of the vascular type. It is
deduced that some vascular type progress to
mixed type with age.
We used a cream containing vitamin C, vita-
min E, and retinol for efficacy evaluation. Vita-
min C and its derivatives, such as magnesium
ascorbyl phosphate and AG, have along history
as topical skin lightening agents and are known
to inhibit melanogenesis in human melanocytes
(8). In addition, vitamin C has also been shown
to regulate collagen production (5, 9). Vitamin E
can improve the cardiovascular system to
reduce lipid peroxidation, which results in the
reduction or clearing of bruising on the lower
eyelid (10). Retinol and vitamin C have also
been documented to induce collagen synthesis.
While, a study of human skin showed clinical
improvement in melasma and senile freckles
when u vitamin C (magnesium ascorbyl phos-
phate) (10–13). In this clinical trial, we expected
regulation of collagen synthesis by the test pro-
duct (Vitamin C 3%, Retinol 0.1%, Vitamin E
0.5%) in addition to inhibition of melanogenesis
and concealing the color of blood stasis, which
would improve the appearance of dark circles.
The parameters for pigmentation M.I, L*, and
quasi L*, showed statistically significant
changes, but the parameters for blood flow such
as a*, E.I. and quasi a* showed no changes after
8 weeks. Dermal thickness which was pre-
sumed as a cause if it was thinning was
increased. The changes in the vascular compo-
nents for the test products seemed to come
4 weeks 8 weeks
component. The change in dark circles within
8 weeks by visual grading was found to be dif-
0.797* 0.812* ficult to observe. We also compared the correla-
0.844* 0.858*
tion coefficients of all parameters. For M.I.
0.694* 0.763*
0.099 0.267 value, negative correlations were found
0.063 0.172 between L* value and quasi L* value. However,
0.004 0.318 the dermal thickness showed no significant
negative or positive correlations with other
281
Park et al.

(a) (b)

(c) (d)

Fig. 4. Changes in mean values ± SD measured at dark circles during the treatment with the test product (Vitamin C 3%, Vitamin A 0.1%, Vita-
min E 0.5%) and the placebo in split-face randomly for 8 weeks. (a) M.I. value measured by Mexameter (b) L* value measured by Chromameter
(c) Quasi L* value measured by image analysis obtained with Visia-CR (d) Dermal thickness measured by Dermascan-C. **One way ANOVA post-
hoc Scheffe, P < 0.05; ***One way ANOVA post-hoc Scheffe, P < 0.001.

parameters. Therefore, the dermal thickness
alone may not be enough to be used as the
major parameter for efficacy evaluation but only
as a reference indicator. In this study significant
improvement in parameters for dark circles was
observed in 8 weeks. Furthermore, we con-
firmed that the results of dermatology by Inha
University hospital, the validation institution
showed a similar tendency. Therefore, this
evaluation method is considered to be reliable
for dark circle evaluation.
In conclusion, we classified the dark circles
according to the causes and verified the reliabil-
ity of the parameters obtained by instrument
conformity assessment used in this study
through the efficacy evaluation. Also, based on
this study, we have suggested new methods
and scoring grade which can be applied to the
evaluation of efficacy of for dark circles. It is
important in the development of cosmetics and
dermatological treatment methods in differenti-
ating the main cause of the dark circles before
treatment and select appropriate methods
corresponding to the cause.
Acknowledgments
This research was supported by a grant
(13172KFDA460) from Korea Food and Drug
Administration in 2013.

References
1. Watanabe S, Nakai K, Oshushi T.
Condition known as “dark rings
under the eyes” in the Japanese
population is a kind of dermal
melanocytosis which can be suc-
cessfully treated by Q-switched
ruby laser. Dermatology Surg 2006;
32: 785–789.
2. Lowe NJ, Wieder JM, Shorr N,
Boxrud C, Saucer D, Chalet M.
Infraorbital pigmented skin. Pre-
liminary observations of laser
therapy. Dermatol Surg 1995; 21:
767–770.
3. Roh MR, Chung KY. Infraorbital
dark circles: definition, causes, and
treatment options. Dermatol Surg
2009; 35: 1163–1171.
4. Freitag FM, Cestari TF. What
causes dark circles under the eyes.
J Cosmet Dermatol 2007; 6: 211–
215.

282

5. Oresajo C, Dickens M, Znaiden A.
Eye area problems puffiness, bags,
dark circles and crowsfeet. Cosmet
Toiletries 1987; 102: 29–34.
6. Matsumoto M, Kobayashi N,
Osamu H, Arai S. Study on the
mechanisms associated with dark
circles. J Soc Cosmet Chem 2000;
34: 152–159.
7. Ohshima H, Takiwaki H. Evalua-
tion of dark circles of the lower
eyelid: comparison between reflec-
tance meters and image processing
and involvement of dermal thick-
ness in appearance. Skin Res
Technol 2008; 14: 135–141.
8. Hakozaki T, Takiwaki H, Miya-
moto K, Sato Y, Arase S. Ultra-
sound enhanced skin-lightening
effect of vitamin C and niaci-
namide. Skin Res Technol 2006; 12:
105–113.
9. Chung JH, Youn SH, Kwon OS,
Cho KH, Youn JL, Eun HC. Regu-
lations of collagen synthesis by
ascorbic acid, transforming growth
factor-beta and interferon-gamma
in human dermal fibroblasts cul-
tured in three-dimensional collagen
gel are photoaging-and aging-inde-
pendent. J Dermatol Sci 1997; 15:
188–200.
10. Mitsuishi T, Shimoda Y, Mitsui Y,
Kuriyama Y, Kawana S. The
effects of topical application of
phytonadione, retinol and vitamins
C and E on infraorbital dark cir-
cles and wrinkles of the lower eye-
lids. J Cosmet Dermatol 2004; 3:
73–75.
11. Elson ML. Topical phytonadione
(vitamin K1) in the treatment of
actinic and traumatic purpura.
Cosmet Dermatol 1995; 8: 25–27.
Evaluation method of dark circles
12. Lou WW, Quintana AT, Gerone-
mus RG, Grossman MC. Effects of
topical vitamin K and retinol on
laser-induced purpura on nonle-
sional skin. Dermatol Surg 1999;
25: 942–944.
13. Lupo MP. Antioxidants and vita-
mins in cosmetics. Clin Dermatol
2001; 19: 467–473.
Address:
S. R. Park
Ellead Skin and Bio Research
325, Hwangsaeul-ro, Bundang-gu
Seongnam-si
Gyeonggi-do 463-824
Korea
Tel: +82 31 709 9070
Fax: +82 31 703 9071
e-mail: srom.pak@gmail.com
283