ARTICLE
Auteur(s) : Şebnem Aktan, Sevgi Akarsu,
Turna Ilknur, Melda Demirtaşoğlu, Şebnem Özkan
Department of Dermatology, Dokuz Eylul University School of
Medicine, 35340 İnciraltı, Izmir, Turkey
accepté le 17 Mars 2007
Female pattern hair loss (FPHL) is characterized by a diffuse
reduction in hair density which mainly affects the crown and
frontal scalp with retention of the frontal hairline. Diffuse
central thinning, frontal accentuation (“Christmas tree” pattern)
and, although uncommon, fronto-temporal recession/vertex loss are
the presenting patterns of FPHL. In addition, bitemporal thinning
is commonly associated with FPHL [1]. Ludwig graded FPHL into three
levels according to the severity of diffuse central thinning [2],
while progression of hair loss according to frontal accentuation
has been suggested by Olsen [3]. Clinical findings of a decrease in
density of the central scalp hair with sparing of the occipital
region, no patches of hair loss, negative hair pull on the sides
and back of the scalp and a normal appearing scalp are usually
sufficient to diagnose FPHL, although biopsy may be required in
difficult cases [4]. So far, many measurement methods ranging from
the most to the least invasive ones have been described in the
assessment of hair loss [5]. In recent years, epiluminescence
microscopy (ELM) has been combined with digital image analysis
(Trichoscan) to assess biological parameters of hair growth [6, 7]
and it is reported to be a valuable method in measuring hair
parameters [8, 9]. This technique flattens the target area onto the
scalp through direct contact with a digital ELM, enabling
determination of hair density and anagen/telogen ratio by using
image analysis software.Scalp hair loss patterns correspond to
visual representations of stages of hair loss and the relationship
between the clinical diagnosis of FPHL and objective measurements
of hair biologic parameters have been evaluated in only a few
studies.The aim of this study was to assess the relationship of
clinical presentations of FPHL with hair density and anagen hair
ratio in Turkish patients diagnosed clinically as FPHL and to
compare the measurements of hair with those of controls, with the
aid of ELM combined with digital image analysis (Trichoscan).
Visual counting of hairs using the photomacrographs of the same
images was used in parallel to evaluate hair density.
Materials and methods
Subjects: thirty-nine patients with a complaint of progressive hair
thinning or loss for more than 2 years, who were diagnosed
clinically as FPHL, were enrolled. Patients with postpartum
alopecia and other forms of alopecia and patients with
precipitating factors such as thyroid disease and anemia were
excluded. None of the subjects had signs or symptoms of
hyperandrogenism. The clinical severity of women with FPHL was
classified according to the global grading system as proposed by
Ludwig. Thirty-one healthy women with no complaint and symptoms of
hair thinning or alopecia or history of hair disease were taken as
the control group. Informed consent was obtained from all subjects
and the study was approved by the local ethics committee.
Clipping of hairs and ELM: In patients with FPHL, hair clipping
was performed in 2 localizations: one in a transitional area of
hair loss between normal hair and balding area in the mid-scalp and
the other in the occipital region. These two target spots were
defined and areas of 1.6 cm2 were clipped using a
hairliner. In control subjects the mid-scalp region was chosen for
clipping. The clipped hairs within the target area were dyed for 12
minutes with a solution (RefectoCil® Gschwentner,
Vienna, Austria) and thereafter the coloured area was cleaned with
an alcoholic solution (Actoderm®; Braunschweig, Germany)
as suggested [6] and digital images were obtained at 30-fold
(analyzed area 0.505 cm2) magnification by means of
a digital ELM system (Griscope, Adana, Turkey) while the area was
still wet. Images were taken immediately after clipping and two
days after clipping. The analysis was carried out using the
software (Grimed Software-Trichoscan Professional Grimed Version,
Adana, Turkey) on the images taken 2 days after clipping. In order
to perform the inter-rating reliability analyses for both of the
hair parameters which were automatically evaluated by the software
(hair density and anagen ratio), two consecutive images from the
same target area (transitional area of hair loss in the mid-scalp)
of each patient were recorded on the same day. Inter-rating
reliability of the method was assessed for both hair density and
anagen ratio to see if the results of this method regarding hair
density and anagen ratio are reproducible. The analysis was
performed by computing intra-class correlation coefficients (ICC)
between the ratings in two consecutive images taken from the same
site in mid-scalp region of each patient.
In view of the low hair density values obtained by Trichoscan
analysis in healthy controls, a visual counting method was also
used to validate our findings. Visual counting was carried out on
photomacrographs of the same images used for digital image
analysis. Hair densities were determined by counting hairs in the
target area photographed with x 30 enlargements (the actual size
49.94 mm2) and dividing the hair count by area.
Hair densities obtained by digital image analysis and visual
counting were compared.
Statistical analysis: Groups were compared statistically using
t- test, paired samples t- test, ANOVA (Tukey’s test) and
correlation analysis test. Inter-rating reliability analyses were
performed with SPSS 13.0 by computing the intra-class correlation
coefficient (ICC) (two-way mixed effect model- consistency
definition).
Results
The mean ages of the patient and control groups were 37.5 ± SD13.1
and 35.0 ± SD13.9 respectively (p > 0.05). The distribution of
the ages of the patients was as follows: 10.2% (4/39) below 20,
23.1% (9/39) in the third decade, 23.1% (9/39) in the forth decade,
15.4% (6/39) in the fifth decade and 28.2% (11/39) over 50 years of
age. Mean of the duration of hair thinning recognized by the
patients and the mean age of the onset of hair thinning was 5.1 ±
SD3.5 and 32.5 ± SD13.0 years respectively. Patients with FPHL
included 27 with Ludwig grade I, 11 with Ludwig grade II and 1 with
Ludwig grade III. The demographic characteristics and hair
densities of the patients measured by the two methods are shown in
table 1. Thirteen patients described a
history of first degree relatives with pattern hair loss.
Fronto-temporal thinning was observed in 19 patients.
Hair density and anagen ratio in the mid-scalp region of the
patients were significantly lower than those of the controls (table 2). In subjects with FPHL, mid-scalp
hair density by either of the methods and anagen ratio were not
significantly correlated with age or duration of hair loss. Groups
of women with Ludwig I and Ludwig II/III hair loss, and controls
were compared regarding the hair parameters using ANOVA and the
results are summarized in table 3. Hair
density and anagen ratio measurements of the mid-scalp regions of
the patients (n = 39) were found to be significantly lower than
those of their occipital regions (table
1).
ICC for hair density and anagen ratio determined by the digital
image analysis were computed as 0.925 (95% CI: 0.862-0.960) and
0.805 (95% CI: 0.628-0.898) respectively, both indicating good
agreement among the ratings.
The mid-scalp hair densities (mean ± SD) determined by visual
counting and digital image analysis were as 212.8 ± 32.9 vs 141.7 ±
21.6 (p < 0.001) in controls (n = 31), and 168.0 ± 41.0 vs 127.2
± 29.1 (p < 0.001) in patients with alopecia (n = 39). Hair
densities determined by visual counting showed a significant
positive correlation with the differences between the hair
densities measured by the two methods, in controls (r = 0.867, p
< 0.001) and in patients with alopecia (r = 0.689, p <
0.001). The digital image analysis was also calculated to
underestimate mid-scalp hair density by 27.4% in the total group,
by 32.6% in controls, and by 23.2% in patients with alopecia.
Table 1 The demographic characteristics and mid-scalp
hair densities of the patients
|
Ludwig
|
|
Median age
|
- Mid-scalp (vertex) hair density
- (n/cm2)
|
- Occipital
- hair density
- (n/cm2)
|
p
|
|
I
|
27
|
36.5 (17-62)
|
132.6 ± 29.8a
|
140.3 ± 26.5a
|
< 0.05
|
|
178.4 ± 41.2b
|
202.0 ± 29.6b
|
< 0.001
|
|
II / III
|
12
|
39.9 (25-63)
|
115.0 ± 24.3a
|
132.2 ± 29.4a
|
< 0.05
|
|
144.3 ± 30.2b
|
180.5 ± 27.3b
|
< 0.001
|
ameasurements by digital image analysis (Trichoscan).
bmeasurements by visual counting.
Table 2 Mid-scalp hair density and anagen ratio in
patients and controls (mean ± SD)
|
Hair parameters
|
Patients (n = 39)
|
Controls (n = 31)
|
p
|
|
Hair density (n/cm2)
|
- Digital image analysis
- (Trichoscan)
|
127.2 ± 29.1
|
141.7 ± 21.6
|
< 0.05
|
|
Visual counting
|
168.0 ± 41.0
|
212.8 ± 32.9
|
< 0.001
|
|
Anagen ratio (%)
|
65.0 ± 9.5
|
82.7 ± 4.8
|
< 0.001
|
Table 3 Hair density and anagen ratio in Ludwig I and
Ludwig II/III patient groups and controls (mean ± SD)
|
|
Ludwig I (n = 27)**
|
Ludwig II/III (n = 12)***
|
- p < 0.05
- ANOVA (Tukey’s test)
|
|
- Digital image analysis
- (Trichoscan)
|
141.7 ± 21.6
|
132.6 ± 29.8
|
115.0 ± 24.3
|
* / ***
|
|
Visual counting
|
212.8 ± 32.9
|
178.4 ± 41.2
|
144.3 ± 30.2
|
|
|
82.7 ± 4.8
|
65.6 ± 7.6
|
63.7 ± 13.0
|
|
Discussion
In general, hair density ranges from 175 to 450
hairs/cm2 in normal non-balding subjects [9]. Birch et
al. [10] detected the mid-scalp hair density to be 293
hairs/cm2 in normal woman aged 35 years by using target
area macrophotographs. The total hair density on the vertex of 20
healthy women were identified as 300 ± 20 hairs/cm2 by
phototrichogram using videomicroscopy [11]. In our study group the
mean hair density of the healthy volunteers was assessed as 141.7 ±
21.6 hairs/cm2 by Trichoscan whereas it was calculated
to be 212.8 ± 32.9 hairs/cm2 by visual counting on the
macro-images. In another study, normal women were reported to have
a hair density of equal to or greater than 120/cm2 in a
Japanese population by phototrichogram [12]. Likewise, hair density
in Asians was reported to be significantly lower than in whites and
blacks [13]. On the other hand Nakazawa et al. [14] examined the
scalp and hair of 299 Japanese females and reported the mean hair
density in the vertex area to be 230/cm2 in the thirties
and 199/cm2 in the sixties. The differences in the
measurements may be due to different evaluation methods being used.
The relatively low hair density in our control group measured by
Trichocan may be attributed to the characteristics of our study
group as well as to the performance of the Trichoscan method.
Recently Van Neste and Trüeb evaluated the performances of
Trichoscan and contrast-enhanced phototrichogram (CE-PTG) analysis
on the same skin sites comparatively and reported Trichoscan to
underestimate hair density [15]. The lack of detection of thinner
hair by Trichoscan has also been described by other users with hair
density being underestimated by 22% [15]. Likewise Trichoscan was
observed to underestimate hair densities in both our healthy
controls and patient groups. In addition, hair density determined
by counting was found to show a significant positive correlation
with the amount of underestimation (the difference in the hair
densities between the two methods) by Trichoscan.
We did not observe any correlation between hair parameters and
patient age or duration of hair loss progression. On the other
hand, Van Neste [16] evaluated hair parameters of females showing
the Ludwig pattern hair loss by CE-PTG analysis and found that
increasing age and clinical severity were associated with decreased
hair density.
Birch MP et al. [10] analyzed the relationship between clinical
diagnosis of FPHL and objective measurements of hair density and
they reported mean hair densities to be lower in patients with FPHL
than the non-balding group although an extensive overlap was
recorded. Their subjects with Ludwig I hair loss were found to have
a significantly lower mean hair density than healthy women. In
addition they reported a significant difference in hair density
between women with Ludwig I and Ludwig II/III hair loss. In our
study group, the hair densities of women with FPHL were also
detected to be significantly lower than those of controls both by
visual counting and Trichoscan. The visual counting detected
significant differences among the mid-scalp hair densities of all
groups including controls, Ludwig I, and Ludwig II/III groups,
while digital image analysis revealed a significant difference only
between controls and Ludwig II/III group.
Ueki R et al. [12] examined and categorized hair growth patterns
of female subjects complaining of diffuse hair loss using
phototrichogram analysis and reported a positive correlation
between macroscopic observation method of the Ludwig classification
and the grading of hair growth patterns. Hair density was also
stated to be the most important parameter in assessing chronic
diffuse alopecia in Japanese female subjects. On the contrary, the
anagen hair ratio was found to be relatively unimportant in grading
hair patterns. In our study group, the anagen hair ratios of our
subjects with Ludwig I and Ludwig II/III hair loss were found to be
significantly lower than those of healthy controls. However no
significant difference was detected between the anagen ratios of
women with Ludwig I and Ludwig II/III hair loss.
The follicular changes in pattern hair loss include progressive
reduction in the duration of anagen, follicular miniaturization and
prolongation of the latent period of the hair cycle [1, 17]. Low
hair density in FPHL is suggested to be due to the prolongation of
the latent period of the hair cycle, so that there is an increase
in the number of “empty” follicles [10]. The hair cycle in which
the hair follicles remain empty has been termed as kenogen [18].
Androgenetic alopecia is also characterized by a progressive
reduction in hair shaft diameter and de Lacharriere et al. [19]
showed a link between hair diameter diversity and hair follicle
miniaturization. Thus hair diameter diversity was suggested to be
the main accurate clinical sign reflecting follicle miniaturization
process.
In conclusion, diminished hair density seems to be the main sign
in the clinical diagnosis of hair loss, although digital image
analysis was found to be less competent than visual counting in
detecting hair density. Thus, the increase in hair density in the
assessment of various treatment modality effects may be
underestimated by digital image analysis.
Acknowledgments
Financial support: none. Conflict of interest: none.
References
1 Olsen EA, Messenger AG, Shapiro J, et al.
Evaluation and treatment of male and female pattern hair loss. J Am
Acad Dermatol 2005; 52: 301-11.
2 Ludwig E. Classification of the types of androgenetic
alopecia (common baldness) occurring in the female sex. Br J
Dermatol 1977; 97: 247-54.
3 Olsen EA. Female pattern hair loss. J Am Acad Dermatol
2001; 45: S70-S80.
4 Olsen EA. The midline part: An important physical clue to
the clinical diagnosis of androgenetic alopecia in women. J Am Acad
Dermatol 1999; 40: 106-9.
5 Van Neste MD. Assessment of hair loss: clinical relevance
of hair growth evaluating methods. Clin Exp Dermatol 2002; 27:
358-65.
6 Hoffman R. Trichoscan: combining epiluminescence
microscopy with digital image analysis for the measurement of hair
growth in vivo. Eur J Dermatol 2001; 11: 362-8.
7 Hoffman R. Trichoscan: A novel tool for the analysis of
hair growth in vivo. J Invest Dermatol Symp Proc 2003; 8:
109-15.
8 Chamberlain AJ, Dawber RPR. Methods of evaluating
hair growth. Australas J Dermatol 2003; 44: 10-8.
9 Pierard GE, Pierard-Franchimont C, Marks R,
Elsner P. EEMCO guidance for the assessment of hair shedding
and alopecia. Skin Pharmacol Physiol 2004; 17: 98-110.
10 Birch MP, Messenger JF, Messenger AG. Hair
density, hair diameter and the prevalence of female pattern loss.
Br J Dermatol 2001; 144: 297-304.
11 D’amico D, Vaccaro M, Guarneri F,
Borgia F, Cannavo SP, Guarneri B. Phototrichogram
using videomicroscopy: a useful technique in the evaluation of
scalp hair. Eur J Dermatol 2001; 11: 17-20.
12 Ueki R, Tsuboi R, Inaba Y, Ogawa H.
Phototrichogram analysis of Japanese female subjects with chronic
diffuse hair loss. J Invest Dermatol Symp Proceed 2003; 8:
116-20.
13 Lee HJ, Ha SJ, Lee JH, et al. Hair counts
from scalp biopsy specimens in Asians. J Am Acad Dermatol 2002; 46:
218-21.
14 Nakazawa Y, Komori Y, Kio K, Shimada Y,
Tajima J, Murayama N, Tsuchida K, Lino M,
Tajima M. (Abstract) Characterization of scalp and hair of the
female with each period. J Dermatol Sci 2006; 42: 187-8.
15 Van Neste D, Trüeb RM. Critical study of hair
growth analysis with computer-assisted methods. J Eur Acad Dermatol
Venereol 2006; 20: 578-83.
16 Van Neste D. Female patients complaining about hair
loss: documentation of defective scalp hair dynamics with
contrast-enhanced phototrichogram. Skin Res Technol 2006; 12:
83-8.
17 Courtois M, Loussouam G, Hourseau C,
Grollier JF. Ageing and hair cycles. Br J Dermatol 1995; 132:
86-93.
18 Rebora A, Guarrera M. Kenogen: A new phase of the
hair cycle? Dermatology 2002; 205: 108-10.
19 de Lacharriere O, Deloche C, Misciali C,
et al. Hair diameter diversity. Arch Dermatol 2001; 137:
641-6.
|
Printable version
Version PDF