ARTICLE
The field of physiological and pathological changes of hair growth/loss
is an amazingly complex one from the technological point of view. Indeed
the asynchronous scalp hair root activity contributes to an apparently
stable global mass of hair albeit it is clearly associated with cyclical
activity of the hair follicle. Repeat cycles of hair follicle growth
involution and regrowth contribute to the continuous renewal of the existing
pool of hair. This global result depends, however, on the programmed activity
of every single contributing follicle in various skin including
scalp regions.
As global changes are the cumulative end result of discrete, barely
perceptible changes of the function and/or structure of individual hair
follicles, the understanding and measurement of the process should use
an analytical approach. The aim being to detect early changes at the single
hair level i.e. before the disorder becomes clinically noticeable.
This concept has been detailed earlier in view of diagnostic, prognostic
and monitoring methods for scalp hair loss [1, 2].
Amongst other methods reviewed elsewhere [3], the use of repeat photography
after clipping of a selected skin site the so-called phototrichogram
method (PTG) initiated by Saitoh [4] enables differentiation of
growing from non-growing hair at the skin surface. PTG might not be appropriate
for unselected subjects who attend the hair clinic especially those with
less natural contrast between hair and skin colour [5]. Also, hair diameter
and pigmentation are significantly decreased during the hair follicle
miniaturisation process of AGA [6]. This might significantly alter visibility
of hair and generate erroneous figures from scalp PTG readings.
Herein we report the results of a methodological study comparing the
results of hair measurements using three different techniques. By comparison
with the light microscopic examination of scalp biopsies and our previously
published PTG method [7], the new PTG method including hair contrast enhancement
(CE-PTG) which appears as a significant technological improvement for
scalp hair growth measurement especially in AGA.
Material and methods
Subjects
After approval of the study protocol by the Ethical Committee for Human
Investigation, 10 male subjects with androgenetic alopecia (AGA stages
III (n = 4), IV (n = 3), V (n = 3) according to the Norwood-Hamilton classification
[6] completed the study. The mean age of subjects was 36.8 years (range:
23-49 years). Subjects with scalp disorders other than AGA and those under
medication for any reason or taking any treatment known to influence hair
growth were excluded.
Study Protocol
Sampling methods
Each subject was asked to visit the clinic twice at 48 hrs interval.
On the first visit, hairs were clipped on a target site (1 cm2
area) located on the top of the head. The spot was chosen in a zone of
active progression of alopecia (i.e. not totally bald). We chose
it on the crossing of two imaginary lines: the longitudinal axis starts
from the external edge of the eyebrow in a caudal direction towards the
anterior-mid scalp zone and the transversal axis runs from one ear tip
to the other. Photographs (Fig.
1a), with original x 3 enlargement, were taken with the macro camera
(Medical Nikkor lenses) using the scalp immersion proxigraphy method described
in detail elsewhere [7]. All photographs were taken before and after application
of transient hair dye for contrast enhancement (CE, Fig.
1b) an adaptation of the method described by Blume et al. [8].
After 48 hrs, photographs were taken again before (Fig.
1c) and immediately after a new CE (Fig.
1d) to assess hair growth. Indeed, according to our experience, the
2 days interval between photographs is long enough for identification
of anagen hair follicles. They produce a new hair segment about 0.6 mm
in length (dotted lines in Fig.
1c and d), as opposed to resting or telogen hair follicles
(no growth between day 0 and day 2, full lines only on Fig.
1c and 1d).
After local anaesthesia, scalp biopsies (punch diameter: 4 mm) were
performed within the photographed site and fixed in Bouin's fluid. One
biopsy was made in each subject while in 3 of them we were able to obtain
a second punch biopsy from the same photographed site (Fig.
1e). Finally, another photograph of the investigated site was performed
after the biopsy (Fig. 1f).
This would allow the localisation of the biopsy site(s) on the photographed
scalp, evaluation of matching of hair on both photographs and biopsies
(Fig. 1g). A paired comparison
of "hair counts" and assessment of hair growth phase of individual hair
follicles by the different methods could eventually be made (see below
and flow chart diagram, Fig.
1h).
Photographs
The exact area corresponding to the site of the biopsy was determined
on the photographs used for PTG and CE-PTG. Two subjects were excluded
from the study because of inadequate matching of the single biopsy site
and photographed site. Analysis of the photographed hairs was made exclusively
on the remaining 8 subjects with a total of 11 well documented files.
The first phototrichogram (PTG) was performed by combining pictures taken
at day 0 and day 2 i.e. just before CE (Fig.
1a and c). The second phototrichogram analysis was made from
pictures taken after contrast enhancement (CE-PTG) at day 0 and day 2
(Fig. 1b and d).
During preliminary studies (unpublished data), a protocol was developed
such that these two distinct phototrichogram analyses (PTG and PTG-CE)
could be performed on this set of documents without requiring distinct
or extra visits.
The following measures were performed: total hair density (all visible
hair/cm2), thick hair (thicker than a 40 mum ruler) and thin
hair (equal or less than a 40 mum ruler) density, and the percentage of
anagen hair i.e. ratio of (growing hair/all visible hair) x 100.
A preliminary study compared microscopically measured hair diameters (Caucasian
subjects with AGA, unpublished data) with calibration on scalp photographs.
The study showed that assessment of individual hair diameter on photographs
was accurate if the diameter of the same hair under the microscope was
less than, or equal to 40 mum. The measurements of hair diameters on fibres
thicker than 40 mum were significantly more variable because of their
elliptical section, a factor already known from other studies [9].
Scalp biopsies
Scalp punch biopsies were prepared for transverse sectioning i.e.
parallel to the scalp surface [10]. The sectioning was performed in 2
steps. All 11 specimens were first cut into two parts at ± 1mm below
the skin surface (Fig. 1i) i.e.
at the level of the opening of the sebaceous gland (level B). Transversal
sectioning (40 sections per sample, 6 mum thickness) was performed on
the upper segment and sections were stained with haematoxylin and eosin.
This would allow counting of the total number of hair follicles and the
evaluation of matching of hair follicles with hair observed on photographs.
For 2/11 biopsies (from 2 subjects, Hamilton stages IV and V), the vast
majority of individual follicles observed under the microscope could be
traced on photographs. Hence, in a second step, the 2 remaining parts
of the biopsies (under and above the level of the sebaceous gland or reticular
dermis level B) were completely sectioned. A total of 1,048 serial
sections were examined in order to obtain a microscopic overview (Fig.
1j) from the dermo-epidermal junction (level A) down to the deeper
dermis and hypodermis (level C). Hair diameters in sections were measured
by micrometry with a 1 µm resolution. The data obtained from these
2 completely sectioned biopsies were used for paired comparative analysis
of individual hair or hair follicle measures: total hair density (all
visible hair/cm2), thick (> 40 µm) and thin (¾
40 µm) hair, and the number of anagen hairs. The growth staging criteria
on histology were defined according to Whiting [11].
Statistical analysis
Descriptive statistics were made and paired t test was used whenever
this was appropriate. P values ¾ 0.05 were considered for significance
of paired differences.
Results
Hair density (n/cm2)
Hair density (Integer values shown in Table
I) observed on histological slides (level B) was higher as compared with
PTG (paired difference: 30 hair/cm2), but significantly less
than CE-PTG (paired difference: 32 hair/cm2; p < 0.05).
Between PTG techniques, the paired difference, as a function of presence
or absence of CE was 62-hair/cm2 (p < 0.05). In 9 out of
11 paired occurrences the difference was > 10% with an average of 30%
better hair detection after CE. After deleting the data of one subject
with the most dramatic density changes after CE (subject 5; + 247 hair/cm2
after CE), marginal changes of the average values of hair/cm2
(PTG: 189, CE-PTG: 233, histology: 203) were observed without affecting
statistically significant differences between methods. The average of
absolute paired differences after CE was however reduced (62 (n = 11)
to 44 (n = 10) hair/cm2).
Growth phases
Thorough examination of sections at the reticular dermis (level B) of
the 11 biopsies did not give enough information to study the growth phase
of all visible hair follicles.
Accordingly, comparison between growth phase data obtained through PTG,
CE-PTG and histology was performed on the 2 completely sectioned biopsies.
This combines the 3 histological levels A, B and C. Clearly identified
hairs and hair follicles along with their growth phases are shown for
each method in Table II
and a typical result is shown in Fig.
2.
Major difficulties were encountered with histological analysis when
we attempted growth staging of follicles subject to extreme miniaturisation
i.e. those located exclusively in level A. This contrasts with
the easiness of staging terminal type follicles, a majority of which reached
the deeper dermis (level C) and were found to be anagen stage. Most of
them produced thick hair fibres well visible on CE-PTG except during the
initial stages when the tapered end of a newly synthesised anagen hair
sprouts at the scalp surface. Hence growth of a thin hair at the scalp
surface may for a short while reflect a much thicker hair at the dermal
level. All catagen and telogen follicles observed under the microscope
were identified as such on CE-PTGs. The former appear as minimally growing
hair fibres reflecting the "squeezing out" phenomenon of the old hair
fiber during catagen. One observation pointing to the precision of the
method is worth mentioning: a thick telogen hair on CE-PTG appeared to
correspond with a thin anagen hair at histology (level C). This follicle
at the initial stages of hair re-growth (anagen 3-4, tapered hair at level
C) was traced higher up in the dermis (level B) where the outer root sheath
interconnected with the thick telogen club, the ultimate stage of the
previous cycle before exogen. Hence, a typical telogen germinal
unit generated different but complementary views between histology and
CE-PTG.
Thin and thick hair
The average density (standard deviation) of thin and thick hair is also
reported in Table I. The
hair diameter analysis at level B of 11 biopsies confirms our previous
statement that a lower proportion of thin hair was detected at this level
in transverse sections (33.6%) as compared with PTG (42.4%) and CE-PTG
(51.6%). In order to detect why such variation was present and because
of the more accurate detection of thinning hair by CE-PTG, a paired analysis
of hair calibration was performed between values obtained from CE-PTG
and serial sections obtained from the 2 biopsies.
From a total pool of 91 records including all hairs visible either on
CE-PTG and/or biopsy, 16 were discarded. This decision was made because
the hair fiber was lost during sectioning of the biopsy (n = 13) or the
hair seen on the edges of deeper part of the biopsies could not be traced
in the upper levels (n = 3). Hence a paired analysis was performed on
75 hair fibres and their corresponding follicles. All 33 hairs identified
as thick on CE-PTGs had diameters > 40 µm in the biopsy. The smallest
diameter of a thin hair pointed on CE photographs was found to
be 8 µm in diameter on histology. Out of 30 hairs categorised as
thin on CE-PTGs, 3 were slightly thicker on microscopy of the corresponding
hair follicle (i.e. 42, 42 and 44 µm). This, together with
changes of hair diameter during cycling (see previous section), may explain
that thin hair is slightly over-rated with CE-PTG as compared with histology.
With the notable exception of one 30 µm hair that did not reach
the scalp surface (anagen 4-5; histology level A) and therefore could
not be seen on photographs, CE-PTG missed the detection of 12 other much
finer thin hairs. Indeed their diameter ranged from 8 to 17 µm; 6
of those missing thin hairs on CE-PTGs appeared as black dots on CE
photographs. This was due to dye retention in the follicular ostia of
hair follicles which were described histologically as severely miniaturised
(visible at level A only) or as empty i.e. containing no hair fiber
at any level. The other 6 very fine hairs did not reach the scalp surface
(n = 3) or might have been hidden to the observer by other hair fibres
originating from adjacent follicles ending in the same ostium (n = 3).
Conversely, although extremely miniaturised follicles can be traced at
level A it is almost impossible to ascribe functional activity (see previous
section). This could be addressed only during a follow-up study using
repeat CE-PTG provided that the hair shows up from time to time at the
scalp surface a situation that may be considered clinically as
not relevant.
Discussion
The aim of the present study was to evaluate the accuracy of measurements
performed with 3 different techniques in scalp hair of male subjects with
AGA. The values of density indicate a significantly better detection of
hair after contrast enhancement (CE-PTG) as compared with the more conventional
PTG; in 9 out of 10 patients the difference was greater than 10%. In a
more detailed paired analysis, all thick hair growing or resting
was detected with the improved CE-PTG method; some hair thicker
than 40 µm might however be categorised as thin but this error occurs
within less than 10% diameter range (up to 44 µm); this error does
not have major clinical significance. Indeed such follicles might be either
thick in early anagen and the definitive thickness of the hair shaft will
appear in a short time at the scalp surface or they already are on their
way to more severe miniaturisation. A minority of thin hairs (less than
10% of the total hair sample in our paired analysis) could not be detected
because they remain inside the scalp (incipient or extremely short anagen
phase), or hidden, adhering to or behind other thicker hair, usually emerging
from the same follicular orifice. The error can once more be considered
as clinically not relevant as in the latter it represents a very small
fraction of barely visible hair while in the former case the hair is not
or will never become visible at the scalp surface. From the biological
perspective, it appears that mismatching remained within acceptable experimental
error. The threshold of the photographic system appears to be adequate
as indicated by the detection of extremely thin hairs (8 µm diameter).
Histological analysis of transversal sections of scalp punch biopsies
is an alternative but needs a substantial amount of work considering that
appropriate sampling is a must and that serial sectioning from top to
bottom is necessary for the generation of fully reliable figures. In addition
to the destructive nature of the sampling, the procedure does not allow
monitoring of individual scalp hair follicles over time, and the small
size gives rise to wider variation. Finally some hair fibres disappear
during the sectioning leaving "ghosts" that are unsuitable for measurements.
Nevertheless, histological analysis of scalp samples in general remains
essential. In fundamental hair research it provides a better understanding
of the hair follicle structure (e.g. streamers or fibrous tracts
and telogen germinal units). It also documents interaction of the hair
root with the interadnexial dermis (i.e. inflammatory infiltrate)
and this is of primary diagnostic importance. The method has been used
for therapeutic assays in humans with minoxidil [12] where hair density
appears to be of critical prognostic importance [13] and with finasteride
where efficacy was documented in male but not in female AGA [14].
The PTG on the human scalp has been used in a series of clinical research
programmes [15] and appears an appropriate choice for monitoring of small
scalp fields over long periods of time in properly selected male and female
candidates with a AGA [16-18]. More recently the method documented improved
hair growth in male subjects with AGA after 6 months and 1 year with a
1 mg/day treatment with finasteride [19]. The natural contrast between
hair and scalp is usually mentioned as a selection criterion for participants
to such trials. However the progression of AGA is associated with decreased
pigmentation of the thinning hair [6] which may subsequently escape photographic
detection. In the long term one has also to take into account greying
as those hairs are totally invisible on photographs (personal unpublished
data).
Our study shows that CE-PTG detects greater numbers of scalp hairs and
more specifically hair affected by the thinning process of AGA in Caucasian
men. These observations about thinning hair remain valid whatever the
natural contrast between thick hair and scalp. Therefore, we suggest that
changes in the duration of anagen and in hair diameter might be more accurately
measured using this improved method which is to become our standard operation
modality as well for diagnostic and prognostic evaluation as well as for
therapeutic monitoring.
CONCLUSION
Acknowledgements
A preliminary study comparing non invasive measurements and scalp histology
has been presented at the Society of Investigative Dermatology (abstract
in Journal of Investigative Dermatology, 106: 950, 1996). The initial
examinations of scalp sections were performed together with David Whiting
who encouraged the pursuing of this research topic. The serial sectioning
was made at the histopathology laboratory of the RHMS- Tournai (Dr. Nathalie
Renard). The author acknowledges the technical assistance of Dominique
Debauque and Caroline Tételin during scalp photography and sampling,
and Christine de Hosté, Céline Geveaux and Ghasan Shaker
who undertook the painstaking phototrichogram analysis and serial section
observations. The scientific advice of Bernadette de Brouwer, Thérèse
Leroy has been appreciated during the manuscript preparation.
A Medical School Grant from Merck and Cy (USA) to Skinterface financially
supported this study.
REFERENCES
1. Van Neste DJJ, Rushton DH. Hair Problems in Women. Clin
Dermatol 1997; 15: 113-25.
2. Van Neste DJJ. Hair growth evaluation in clinical dermatology.
Dermatology 1993; 187: 233-4.
3. Shaker G, Van Neste D. Hair: scalp hair growth evaluation
methods. In: Gabard B, Elsner P, Surber C, Treffel P, eds. Dermato-pharmacology
of topical preparations: a product development-oriented approach. Berlin:
Springer, 1999: 241-54.
4. Saitoh M, Uzuka M, Sakamoto M. Human hair cycle. J Invest
Dermatol 1970; 54: 65-81.
5. Rushton DH, De Brouwer B, De Coster W, Van Neste D. Comparative
evaluation of scalp hair by phototrichogram and unit area trichogram analysis
within the same subjects. Acta Derm Venereol (Stockh) 1993; 73:
150-3.
6. Camacho F, Randall VA, Price VH. Hair and its disorders.
Biology, Pathology and Management. London: Martin Dunitz Ltd, 2000.
7. Van Neste DJJ, Dumortier M, De Brouwer B, De Coster W. Scalp
immersion proxigraphy (SIP): an improved imaging technique for phototrichogram
analysis. J Eur Acad Dermatol Venerol 1992; 1: 187-91.
8. Blume U, Ferracin I, Verschoore M, Czernielewski JM, Schaefer
H. Physiology of the vellus hair follicle: hair growth and sebum excretion.
Br J Dermatol 1991; 124: 21-8.
9. Rushton DH. Chemical and morphological properties of scalp
hair in normal and abnormal states. PhD Thesis, University of Wales,
Cardiff, UK, 1988.
10. Whiting DA. The value of horizontal sections of scalp biopsies.
J Cutan Aging Cosm Dermatol 1990; 1: 165-73.
11. Whiting DA. Scalp biopsy as a diagnostic and prognostic tool
in androgenetic alopecia. Dermatologic Therapy 1998; 8: 24-33.
12. Headington JT. Histological findings in androgenic alopecia
treated with topical minoxidil. Br J Dermatol 1982; 107 (suppl.)
22: 20-1.
13. Whiting DA. Diagnostic and predictive value of horizontal
sections of scalp biopsy specimens in male pattern androgenetic alopecia.
J Am Acad Dermatol 1993; 28: 755-63.
14. Whiting DA, Waldstreicher J, Sanchez M, Kaufman KD. Measuring
reversal of hair miniaturization in androgenetic alopecia by follicular
counts in horizontal sections of serial scalp biopsies: results of Finasteride
1 mg treatment of men and postmenopausal women. J Invest Dermatol
1999; 4/3: 282-4.
15. Guarrera M, Ciula MP. A quantitative evaluation of hair loss:
the phototrichogram. J Appl Cosmetol 1986; 4: 61-6.
16. Guarrera M, Rebora A. Anagen hairs may fail to replace telogen
hairs in early androgenic female alopecia. Dermatology 1996; 192:
28-31.
17. Guarrera M, Cipriani C, Rebora A. Delayed telogen replacement
in a boy's scalp. Dermatology 1998; 197/4: 335-7.
18. Courtois M, Loussouarn G, Hourseau C, Grollier JF. Ageing
and hair cycles. Br J Dermatol 1995; 132: 86-93.
19. Van Neste D, Fuh V, Sanchez-Pedreno P, Lopez-Bran E, Wolff
H, Whiting D, Roberts J, Kopera D, Stene JJ, Calvieri S, Tosti A, Prens
E, Guarrera M, Kanojia P, He W, Kaufman K. Finasteride increases anagen
hair in men with androgenetic alopecia. Br J Dermatol 2000; 143:
804-10.
|
Printable version
