C L I N I C A L A N D LA B O R A T O R Y I N V E S T I G A T I O N S DOI 10.1111/j.1365-2133.2005.06608.

‘Atrophic telogen effluvium’ from cytotoxic drugs and a

randomized controlled trial to investigate the possible
protective effect of pretreatment with a topical vitamin D3
analogue in humans
T.O. Bleiker, N. Nicolaou, J. Traulsen* and P.E. Hutchinson
Department of Dermatology, Leicester Royal Infirmary, Leicester LE1 5WW, U.K.
*Medical Department, Leo Pharma, Ballerup, Denmark

Summary

Correspondence Background Hair loss from cytotoxic drugs is classically ascribed to the loss of frac-
P.E. Hutchinson. tured hairs (anagen effluvium). Telogen hair loss has also been described but
E-mail: dermat@doctors.org.uk
some authors have denied any effect on the hair cycle. There are conflicting
reports on a protective effect of pretreatment with a vitamin D analogue on cyto-
Accepted for publication
7 November 2004
toxic drug-induced hair loss in rodents.
Objectives To investigate the process of cytotoxic hair loss and any protective effect
Key words: on the hair of pretreatment with topical calcipotriol.
anagen effluvium, randomized controlled trial, Methods Breast cancer patients who were about to receive cycles of chemotherapy
topical calcipotriol with cyclophosphamide 600 mg m)2, methotrexate 40 mg m)2 and 5-fluoroura-
cil 600 mg m)2 were recruited and randomized to receive calcipotriol scalp solu-
Conflicts of interest:
None declared. tion 50 lg mL)1 or vehicle. The solution was applied twice daily from 4 days
prior to chemotherapy and continued for 14 days in each treatment cycle. Shed,
plucked and cut hairs were sampled. Absolute shed rates, the proportion of major
hair types, the presence of proximal hair shaft changes, regrowth (using the new
anagen hair count) and hair density were assessed.
Results Ten patients receiving calcipotriol and 14 receiving vehicle completed three treat-
ment cycles and nine from both groups completed six cycles. There was no detectable
effect of calcipotriol on the proportion of patients experiencing minimal hair loss from
chemotherapy, shed rates, plucked telogen and fractured hair counts, the morphology
of shed and plucked hair, hair regrowth or hair density. Combining results of the treat-
ment groups, there was a large variation in the impact of chemotherapy on hair loss,
from total loss in five patients to no obvious loss in five. Excluding the latter, during che-
motherapy shed telogen hairs (mean 81% of shed hairs) predominated over fractured
(12%) and anagen hairs (6%) (P ¼ 0.0002). The major pathological change was prox-
imal hair shaft tapering, baseline mean 3% of shed hairs rising to 48% (P ¼ 0.0005)
during treatment, and there was a consequent decrease in normal telogen hairs, baseline
mean 98% of all telogen hairs falling to 55% (P ¼ 0.0005) during treatment. The
pathological tapered telogen hairs had normal or small, sometimes diminutive, bulbs.
Fracturing of hairs with diminutive bulbs produced typical ‘exclamation mark’ hairs.
Conclusions The cardinal effects of cytotoxic drugs found in this study were tapering
of the proximal hair shaft and premature entry of the follicle into telogen,
conflicting with the conventional view that affected hair follicles continue in
anagen. There was a resulting effluvium of a mixture of tapering telogen hairs and
fractured hairs. As entry into telogen is an integral part of the process, cytotoxic hair
loss may be regarded as a variant of the conventional ‘telogen effluvium’ and we
propose the term ‘atrophic telogen effluvium’. There was no obvious protective
effect on the hair loss of prior treatment with topical calcipotriol.

2005 British Association of Dermatologists • British Journal of Dermatology 2005 153, pp103–112 103
104 Mechanism of cytotoxic alopecia and effect of vitamin D, T.O. Bleiker et al.
Hair loss is a familiar and often dreaded complication of the
cytotoxic treatment of cancer. This is classically ascribed to the
loss of fractured hairs and the process is often termed anagen
effluvium, presumably on the assumption that the respective
follicles remain in anagen. However, there have also been
reports of associated loss of telogen hairs1,2 while some have
denied any effect on the hair cycle.3,4 A general mechanism of
action of cytotoxic drugs in cancer is the production of DNA
damage and apoptosis. Apoptosis has been demonstrated in
hair follicles following cytotoxic treatment5–8 and apoptosis
has also been demonstrated in normal catagen.7
A protective effect of pretreatment with a vitamin D ana-
logue on hair loss produced by cytotoxics in neonatal rats has
been reported.9 This was not confirmed in mice, but a more
rapid and complete regrowth was noted following topical vita-
min D analogue treatment.10 Partial inhibition of cytotoxic
drug-induced apoptosis in murine hair follicles has also been
reported.8 Topical calcitriol (1,25-dihydroxyvitamin D3) was
found to have no protective effects on hair loss in 12 breast
cancer patients receiving chemotherapy with 5-fluorouracil,
adriamycin and cyclophosphamide.11
The aims of the present study were to investigate the mor-
phology of the shed hair and formally to study any associated
alteration in the hair cycle resulting from cytotoxic therapy in
humans. The second aim was to investigate the protective
effect of a topically applied vitamin D analogue, calcipotriol,
against this cytotoxic drug-induced hair fall.
Subjects and methods
This study was a prospective, randomized, double-blind, vehi-
cle-controlled parallel group trial. It received local ethics com-
mittee approval.
Patients and chemotherapy regimen
The patients were newly diagnosed adult female breast cancer
patients who were scheduled to receive adjuvant chemother-
apy in the oncology department of our hospital. The patients
gave informed, signed consent. Exclusion criteria were any
pre-existing hair condition, including pattern alopecia, derma-
tological scalp conditions, pregnancy or breast feeding, con-
current treatment with a drug associated with an effect on
hair (anticoagulants, valproate, antithyroid drugs and reti-
noids) and hypercalcaemia.
The chemotherapy regimen was cyclophosphamide
600 mg m)2, methotrexate 40 mg m)2 and 5-fluorouracil
600 mg m)2 given on days 1 and 8 in a 4-weekly cycle to a
total of six consecutive cycles. If untoward chemotherapy-
induced toxicity occurred, treatment cycles were delayed until
recovery.
The patients were randomized to receive either calcipotriol
scalp solution 50 lg mL)1 (Leo Pharma, Ballerup, Denmark)
or vehicle of the calcipotriol scalp solution. The solution was
applied twice daily through multiple partings to the entire
scalp from 4 days prior to the first component of each

2005 British Association of Dermatologists • British Journal of Dermatology 2005 153, pp103–112
monthly treatment cycle to 2 days after the final component,
so that the lotion was used uninterruptedly for 14 days in
each treatment cycle.
Assessment visits
All patients were assessed at baseline (2 weeks prior to the
onset of chemotherapy), 13–16 days after the final chemo-
therapy dose in each treatment cycle and 6 weeks after the last
administration of chemotherapy.
Hair sampling
Shed hair
The patients were asked to make formal 5-day shed collec-
tions. They washed their hair on day 0 and then over the next
5 days groomed normally and the shed hairs were collected in
separate envelopes, each day, as found in the brush or comb
or on the pillow. One collection was performed at baseline
and then there were two consecutive collections starting
immediately after the final dose of chemotherapy in each
treatment cycle.
Plucked hair
Approximately 75 hairs were plucked using needle-holding
forceps, with adhesive tape with a fabric backing (Elasto-
plast) covering the jaws to minimize slipping. This proce-
dure was performed at baseline and at visits following the
second, fourth and sixth chemotherapy cycles. The plucked
sites were standard for each visit. A central point of the scalp
was defined as the intersection of a line representing the
anterior–posterior midline of the vertex and a line joining the
pinnae, when the head tilt was such that the lower margins of
the ear lobules were horizontal with the angles of the mouth.
All sites were 5 cm lateral to the central point. The baseline
site was on the left and 3 cm anterior to the mid-point and
the second site was on the left and 1 cm posterior. The third
and fourth sites were at contralateral sites (on the right), the
third being anterior.
Cut hair
A perspex template with a central cut-out area measuring
0Æ5 · 1Æ0 cm was placed on the scalp and the area marked
out with a marker pen. All the hairs contained within this
area were then removed by cutting with fine sharp scissors.
This was performed at baseline and after chemotherapy
cycles 3, 5 and 6 and on follow-up. The scalp ‘landmarks’
were as for the plucked samples and all sites were again
5 cm lateral to the anteroposterior mid-line. The baseline site
was left 1 cm anterior to the mid-point, chemotherapy cycle
3, left 3 cm posterior to the mid-point, cycle 5, right 1 cm
anterior, cycle 6, right 3 cm posterior and follow-up as for
baseline.
 Mechanism of cytotoxic alopecia and effect of vitamin D, T.O. Bleiker et al. 105
Hair examination
For each patient all samples were randomized and the investi-
gator was blinded as to which visit collections were from.
Shed hair rate
The shed hair collections were counted and the daily mean
calculated.
Proximal hair shaft morphology of shed hair and plucked
hair (including telogen counts)
A randomly selected subsample of 30 shed hairs and all of the
plucked hairs in each sample were laid out on glass slides and
were secured with transparent adhesive tape which was not
placed over the proximal ends. The morphology of the prox-
imal hair shaft was examined microscopically. The hairs were
classified as telogen, fractured or anagen, any abnormality in
the root or proximal hair shaft was documented, and the pro-
portion of each subtype was calculated.
New anagen hair count (estimation of regrowth)
Both the cut and the plucked hair obtained above were used.
The number of hairs < 5 cm in length, having fine tapering
distal tips, and the total number of hairs were counted. The
percentage represented by the short hairs (the new anagen
hair count, NAH12) was calculated.
Hair density
The number of hairs in the cut hair samples was counted and
the number of hairs cm)2 calculated.
Statistical analysis
Ideally, the primary response measure would be differences in
shed hair rates. However, this could not be used due to lack of
knowledge of the distribution properties of shed hair counts
during chemotherapy. Therefore the proportion of patients
exhibiting no or minimal hair loss was used, which is a less sen-
sitive but clinically relevant measure. The expected proportion
experiencing this level of loss has been reported as approxi-
mately 45%.13 A population size of 22 patients is required to
detect an increase to 90% (power 80%, significance 0Æ05). Dif-
ferences between groups for the primary response parameter
were analysed by contingency tables. Otherwise, unpaired
groups were analysed by the Mann–Whitney U-test and paired
groups by the Wilcoxon signed rank test. Correlations were ana-
lysed by the Spearman’s rank correlation coefficient (R).
Results
Twenty-four patients were enrolled and received at least three
chemotherapy treatment cycles (10 calcipotriol, 14 vehicle),

2005 British Association of Dermatologists • British Journal of Dermatology 2005 153, pp103–112
20 received four cycles (10 calcipotriol, 10 vehicle), 19
received five cycles (nine calcipotriol, 10 vehicle), 18 received
six cycles (nine calcipotriol, nine vehicle) and 14 attended for
the final follow-up (eight calcipotriol, six vehicle).
Shed hair
Five patients were considered not to have been materially
affected by the chemotherapy. During chemotherapy, at no
time was the mean shed rate in these patients greater than
baseline, and telogen hairs represented 97Æ5% of the shed
hairs, anagen 2% and fractured hairs 1% (cf. chemotherapy-
affected individuals below). However, there was a microscop-
ically detectable effect of chemotherapy in that a mean of 16%
of all shed hairs (99Æ8% of which were telogen) were tapered
at the proximal end during chemotherapy, although this did
not translate into hair loss. The five patients comprised two
(20%) receiving calcipotriol and three (21%) receiving vehicle
(P ¼ 0Æ75).
Mean (range) shed rate at baseline for the whole group of
patients receiving calcipotriol was 69 (21–260) hairs per day
and for those receiving vehicle, 56 (12–216) (Fig. 1). One of
the patients in the vehicle group experienced very rapid (after
the first chemotherapy cycle) and heavy shedding but this
occurred outside the collection period. It was therefore not
possible to collect this hair and the patient has been excluded
from the shed hair results. The maximal mean shed rate
occurred after the second chemotherapy cycle and there was a
large variation between individuals in both treatment groups.
The mean was 223 (39–745) hairs per day in the calcipotriol
group and 196 (24–701) in the vehicle group (P ¼ 0Æ56). In
both groups there was a similar progressive fall in mean shed
rate following subsequent treatment cycles and mean shed rate
following the fifth cycle was similar to baseline rates in both
groups.
Telogen hair was the predominant shed hair type through-
out treatment (Fig. 2) As a proportion of all shed hair, telogen
hair at baseline represented a mean of 91% (39–100%) and
96% (83–100%) in the calcipotriol and vehicle groups,
300
250
Shed rate (hairs day–1)
200
150
100
50
0
Base cycle 1 cycle 2 cycle 3 cycle 4 cycle 5 cycle 6 Follow-
line up
Fig 1. Mean total shed rates through the study: calcipotriol- and
vehicle-treated groups. d, Vehicle; s, calcipotriol. Error bars: mean ±
SEM.
106 Mechanism of cytotoxic alopecia and effect of vitamin D, T.O. Bleiker et al.

the calcipotriol group and 2% in the vehicle group and the

telogen
100% means during chemotherapy were 5% and 4%, respectively
(P ¼ 0Æ1) (Fig. 2).
80%
The range of morphologies of shed telogen and fractured
% main hair types

60% fractured hair following chemotherapy is illustrated in Figure 3. The

40%
20%
0%
Baseline Cycle 1 Cycle 2 Cycle 3
Fig 2. Main hair types of shed hair through the study: calcipotriol-
and vehicle-treated groups. d, Vehicle; s, calcipotriol.
respectively, falling to means during treatment of 86%
(60–99%) and 86% (58–99%), respectively. There was no
difference between groups in the level of decrease (P ¼ 0Æ07).
Fractured hairs were rare at baseline [calcipotriol 3% (0–12%)
and vehicle 2% (0–8%)], but represented a mean of 9%
(0–34%) of the shed hairs in the calcipotriol group and 7%
(0–30%) in the vehicle group (Fig. 2) during chemotherapy
(P ¼ 0Æ8). At baseline 6% of shed hairs had anagen roots in
anagen
characteristic change was tapering of the proximal root shaft,
which varied in the degree of reduction in cross-sectional size
and in the length of hair shaft over which this occurred. At
baseline, 2% (0–7%) of hairs were tapered in the calcipotriol
group and 3% (0–9%) in the vehicle group (P ¼ 0Æ1), while
Cycle 4 Cycle 5 Cycle 6 Follow-up
the mean proportions during chemotherapy were 43% (1–
76%) and 42% (2–83%) (P ¼ 0Æ8). The tapering occurred in
both telogen hair and in proximally fractured hair. In the
tapered telogen hairs the size of the bulb could be normal or
small, sometimes diminutive (Fig. 3E). Fracturing of the shaft
of telogen hairs with small bulbs produced characteristic
‘exclamation mark’ hairs. At baseline normal telogen hairs rep-
resented 98% (92–100%) of all telogen hairs in the calcipotri-
ol group (Fig. 4) and 98% (90–100%) in the vehicle group
(P ¼ 0Æ8). This proportion fell rapidly following the first two
chemotherapy cycles to 61% (26–100%) and 62% (11–
100%), respectively (P ¼ 0Æ8) and then remained relatively
consistent throughout the rest of chemotherapy (Fig. 4). Mean

A B C D E

F G H I

Fig 3. Morphologies of the hair roots and proximal hair shaft following chemotherapy. (A–E) Telogen hairs: (A) normal; (B–E) varying patterns
of proximal shaft tapering with normal-sized telogen bulbs (B,D) or diminutive bulbs (C,E). (F–I) Fractured proximal hair shafts: (F) normal
appearing; (G–I) differing patterns of proximal hair shaft tapering.

2005 British Association of Dermatologists • British Journal of Dermatology 2005 153, pp103–112
 Mechanism of cytotoxic alopecia and effect of vitamin D, T.O. Bleiker et al. 107

normal 100% normal

% forms of telogen hairs
100%

% forms of telogen hairs

tapered normal
bulb 75% tapered normal
bulb
75%

50% tapered small 50%

bulb tapered small
bulb
25%
25%
0%
Baseline Cycle 1 Cycle 2 Cycle 3 Cycle 4 Cycle 5 Cycle 6 Follow 0%
up Baseline Cycle 1 Cycle 2 Cycle 3 Cycle 4 Cycle 5 Cycle 6 Follow-
up

Fig 4. Telogen hair forms through the study: calcipotriol- and

vehicle-treated groups. d, Vehicle; s, calcipotriol. Fig 6. Components of shed telogen hair: groups combined and
unaffected individuals omitted.

percentage normal telogen hairs during chemotherapy was

59% (36–99%) for the calcipotriol group and 64% (26–92%) 12% (0Æ5–40%) (P ¼ 0Æ0002), and anagen hairs, 6% (0–
for the vehicle group (P ¼ 0Æ4). The abnormal telogen hairs 20%) (P ¼ 0Æ0002). Tapering of the proximal hair shaft was
were all tapered, with either normal or small telogen bulbs. found in a mean of 48% (26–77%) of all shed hairs through
Tapered telogen hairs with normal-sized bulbs represented a chemotherapy (Fig. 5). Both percentage tapered hairs and
mean of 31% (1–63%) and with small bulbs 9% (0–30%) of fractured hairs were significantly increased compared
the telogen shed hairs during chemotherapy in the calcipotriol with baseline [3% (0–8%) (P ¼ 0Æ0005) and 3% (0–12%)
group and 27% (7–56%) (P ¼ 0Æ2) and 10% (0–31%) (P ¼ (P ¼ 0Æ0025), respectively]. Normal telogen hairs, as a pro-
0Æ9), respectively, in the vehicle group. Tapered telogen hairs portion of all telogen hairs (Fig. 6), fell rapidly during the
with normal-sized bulbs were rare at baseline in both groups first two courses of chemotherapy from a mean of 98% (92–
(2%) and with small bulbs nonexistent. 100%) pretreatment and then remained fairly constant
As there were no significant differences in shed hair results throughout the rest of treatment. The mean proportion
between treatments, the data were combined and patients not through chemotherapy was 55% (26–92%) (P ¼ 0Æ0005).
materially affected by chemotherapy (n ¼ 5) were omitted During this period the abnormality in the rest of the telogen
(leaving a population of n ¼ 18), for further investigation of hairs was tapering of the proximal hair shaft. The mean ratio
mechanisms of the effects of chemotherapy. Maximal shed- through chemotherapy of normal-sized telogen bulbs to small
ding, mean 256 (66–745) hairs per day, occurred after treat- telogen bulbs in these hairs was 3Æ6 : 1.
ment cycle 2, the mean total shed rate during chemotherapy Table 1 shows how the hair subtypes contributed to the
was 125 hairs per day and the maximum shed rate recorded chemotherapy-induced increase in total shed hair rate at the
was 1232 hairs per day (after chemotherapy cycle 3). More time of maximal mean shedding, i.e. after chemotherapy cycle
severe shedding tended to occur earlier and of the five patients 2. With increase in shed rate there was a decrease in the pro-
who subsequently shed all the scalp hair (NAH reaching portion of normal telogen hairs (P ¼ 0Æ02) and increases in

1Æ0), one experienced major shedding starting 1 week and the proportion of fractured hairs (P ¼ 0Æ05) and all types of
one within 1 month after the first chemotherapy treatment tapered hairs (P ¼ 0Æ03), especially telogen hairs with small
cycle. The mean proportions of the main hair types through bulbs (P ¼ 0Æ01). This is illustrated by one patient in the con-
chemotherapy (Fig. 5) were telogen hairs, 81% (58–99%), trol group who experienced very heavy shedding after chemo-
which were significantly more frequent than fractured hairs, therapy cycle 2 (mean 1066 hairs per day) and had a reduced
proportion of telogen hair (67%), increased proportion of
fractured hairs (26%) and increased proportion of tapered
hairs (81%) (cf. mean rates for all patients, Fig. 6). Although
100%
the proportion of normal telogen hairs decreased with increase
80% telogen
in total shed rate, there was also a significant increase in the
% all shed hairs

60% absolute number of normal telogen hairs (P ¼ 0Æ01).

tapered

40%
fractured
Plucked hair
20% anagen

0%
The range of proximal hair shaft morphologies was the same
eli
n e
cle
1
cle
2
cle
3
cle
4
cle
5
cle
6
w-
u p as in shed hair, which included tapered telogen and fractured
s
Ba Cy Cy Cy Cy Cy Cy llo
Fo hairs and disproportionately small telogen bulbs. Prior to
treatment, the percentage total telogen hairs (Fig. 7) was very
Fig 5. Main hair types of shed hair through the study: groups similar between treatment groups: calcipotriol mean 17%
combined and unaffected individuals omitted. (6–36%), vehicle 17% (3–41%) (P ¼ 0Æ7) (Fig. 7). There

2005 British Association of Dermatologists • British Journal of Dermatology 2005 153, pp103–112
108 Mechanism of cytotoxic alopecia and effect of vitamin D, T.O. Bleiker et al.

Table 1 The relationship of change in total

Shed hair Plucked hair shed rate, from baseline to chemotherapy
cycle 2, and change in the proportion of hair
Spearman Spearman types over the same period in shed and
correlation correlation plucked samples, between individuals
coefficient P-value coefficient P-value
Proportion total telogen )0Æ13 0Æ3 0Æ28 0Æ1
Proportion total anagen )0Æ35 0Æ07 )0Æ51 0Æ02
Proportion total fractured 0Æ41 0Æ05 0Æ30 0Æ2
Proportion normal telogen )0Æ49 0Æ02 )0Æ13 0Æ3
Proportion tapered telogen, 0Æ23 0Æ2 0Æ42 0Æ04
normal bulb
Proportion tapered telogen, 0Æ55 0Æ01 0Æ52 0Æ01
small bulb
Proportion tapered fractured 0Æ44 0Æ03 0Æ11 0Æ3
Proportion total tapered 0Æ47 0Æ03 0Æ55 0Æ009
Number of normal telogen 0Æ53 0Æ01

50% 50%

40% 40%

% of plucked hairs
% of plucked hairs

telogen
30% 30%
fractured
20% 20% tapered
total telogen
10% fractured 10%
normal telogen
0%
0%
Baseline Cycle 2 Cycle 4 Cycle 6
Baseline Cycle 2 Cycle 4 Cycle 6

Fig 8. Main hair types of plucked hair through treatment: groups

Fig 7. Plucked telogen counts and plucked fractured hair counts:
combined and unaffected individuals omitted.
calcipotriol- and vehicle-treated groups. d, Vehicle; s, calcipotriol.

0Æ0003) and tapered hairs from 0Æ2% (0–2%) to 38% (1Æ5–

were very few fractured hairs in either group: calcipotriol 4%
(0–13%) and vehicle 3% (0–10%) (P ¼ 0Æ7). After chemo-
therapy cycle 2, the mean percentage of telogen hairs rose in
both groups to similar levels: calcipotriol 35% (16–77%) and
vehicle 32% (9–90%) (P ¼ 1), as did fractured hairs: 28%
(2–65%) and 25% (0–60%), respectively (P ¼ 0Æ8). Levels of
change similar to those found in fractured hairs occurred in
tapered hairs which were, at baseline, calcipotriol 1% (0–2%),
vehicle 0% (P ¼ 0Æ1) and after chemotherapy cycle 2, 28%
(0–74%) and 29% (0–72%), respectively (P ¼ 1). After sub-
sequent treatment cycles percentage total telogen (and also
normal telogen hairs) decreased more rapidly in the calcipotri-
ol-treated group, but not significantly.
At the chemotherapy cycle 2 assessment there were no sig-
nificant differences between the two treatment groups and the
data have been combined, with patients not materially affected
by chemotherapy omitted (as for shed hairs above) (Fig. 8).
The proportion of all hair types, except for anagen hairs, rose
to a maximum after chemotherapy cycle 2 and then decreased
after subsequent treatment cycles. Mean total telogen count
rose from 16% (4–41%) at baseline to a maximum of 34%
(16–77%) (P ¼ 0Æ0003) after chemotherapy cycle 2, fractured
hair count from 4% (0–15%) to 33% (1Æ5–65%) (P ¼
74%) (P ¼ 0Æ0003). After chemotherapy cycle 2, 61%
(0–100%) of tapered hairs were fractured compared with
18% (0–50%) in the shed hairs through chemotherapy treat-
ment (P ¼ 0Æ0003). Normal telogen hairs (showing no patho-
logical proximal tapering) as a proportion of all telogen hairs
fell from 100% to 54% (25–100%) (P ¼ 0Æ05) over the same
time period. Telogen hairs with small bulbs were rare, reach-
ing a maximum of 2% of all plucked hair (after chemotherapy
cycle 2).
Table 1 shows the correlations between change in shed rate
from baseline to chemotherapy cycle 2 and change in propor-
tion of plucked hair types over the same time period. There
were similar trends as for the proportions of hair types in
shed hair. The positive relationship between shed rate and the
proportion of tapered hairs was again particularly strong
(P ¼ 0Æ009).
New anagen hair counts
NAHs derived from both cut and plucked specimens for both
treatment groups are shown in Figure 9. Mean NAH increased
with successive chemotherapy treatment cycles but somewhat
earlier in the calcipotriol-treated group, at cycle 4; however,

2005 British Association of Dermatologists • British Journal of Dermatology 2005 153, pp103–112
 Mechanism of cytotoxic alopecia and effect of vitamin D, T.O. Bleiker et al. 109

vehicle-treated patients was 23%. Mean density returned to

70%
baseline towards the end of chemotherapy, as would be
60% expected from the NAH results.
50%
% NAH

40%
30%
20%
10%
0%
Base cycle 1 cycle 2 cycle 3 cycle 4 cycle 5 cycle 6 Follow-
line
Fig 9. Mean new anagen hair count (NAH) through the study:
calcipotriol- and vehicle-treated groups. d, Vehicle; s, calcipotriol.
Error bars: mean ± SEM.
NAH was not significantly greater in the calcipotriol-treated
than in the vehicle-treated group at this time point (P ¼
0Æ06). The maximal mean NAH occurred after cycle 5 in the
vehicle group: 45% (9–100%), and after cycle 6 in the calci-
potriol group: 50% (15–99%). There was no significant dif-
ference between these maxima (P ¼ 0Æ7).
Combining treatment groups and omitting patients not
materially affected by chemotherapy, as above, percentage
NAH rose through the treatment cycles and reached a maxi-
mum of 57% (15–100%) from a baseline of 15% (6–29%)
(P ¼ 0Æ0007). In five patients with very severe shedding NAH
reached virtually 100% (97Æ5–100%). There was a strong rela-
tionship between increased shed rate at chemotherapy cycle 2
(compared with baseline) and increased NAH 3 months later,
at chemotherapy cycle 5 (compared with baseline) in treat-
ment groups combined: R ¼ 0Æ78 (P ¼ 0Æ0001).
Hair density
Mean hair density results through the study are shown in Fig-
ure 10. The greatest decrease for both the calcipotriol- and
500
Hair density (hairs cm –2)
Discussion
A protective effect of topical calcipotriol scalp solution has
been investigated in patients receiving multiple cycles of che-
motherapy. No effect on the proportion of patients apparently
unaffected by the chemotherapy, on shed hair rates, shed hair
types, plucked hair types, hair density or speed of regrowth
up
was detected. Jimenez and Yunis9 examined the protective
effect of topical calcitriol on cytotoxic drug-induced hair loss
in neonatal rats. These were treated for 5 days prior to receiv-
ing either etoposide, cyclophosphamide, or an adriamycin–
cyclophosphamide combination. There was a marked protect-
ive effect on hair loss in the calcitriol-treated groups. Paus
et al.10 investigated a protective effect on cyclophosphamide-
induced alopecia of topical calcitriol in a murine model with
hair follicles in anagen IV. No effect on initial hair loss was
found but regrowth was reported to be earlier, more dense
and of more normal colour in the calcitriol-treated animals
compared with controls. This was ascribed to diversion, by
the calcitriol, of the hair follicles into dystrophic catagen while
the controls entered dystrophic telogen. These results obvi-
ously conflict with those of Jimenez and Yunis.9 The authors
maintained that the murine model was more in line with the
human situation, but it is also of note that very much larger
doses of cyclophosphamide were used in the murine study:
120 vs. 35 mg kg)1. Schilli et al.,8 using the same murine
model, demonstrated a decrease in the number of apoptotic
cells in the bulbs in animals pretreated with topical calcipotri-
ol and another vitamin D analogue, KH1060 (Leo Pharma). In
humans, a trial of topical calcitriol in 12 breast cancer
patients, at three concentrations, and two controls showed no
protective effect on the hair loss induced by chemotherapy
with a combination of 5-fluorouracil, adriamycin and cyclo-
phosphamide.11
As there was no demonstrable protective effect of calcipotriol
in the present study, the data have been combined to explore
the mechanism of cytotoxic drug-induced hair loss. There was
a large variability between individuals in shedding rates pro-
duced by the chemotherapy. Five of the patients did not signifi-

400
cantly increase the shed rate while others experienced torrential
300 loss with maximal loss of > 1000 hairs per day, the greatest
shed rate recorded being 1232 hairs per day. The five patients
200 who were not obviously affected were excluded from further
investigation. The mean maximal hair shedding occurred after
100 chemotherapy cycle 2. The average latency of the chemother-
apy effect would therefore appear to be 4–5 weeks. More seri-
0 ously affected patients experienced very heavy shedding earlier:
Base cycle 1 cycle 2 cycle 3 cycle 4 cycle 5 cycle 6 Follow-
line up one patient of the five who developed total loss had a massive
shed starting 2 weeks after the first chemotherapy cycle and
Fig 10. Mean hair density through the study: calcipotriol- and another after 3–4 weeks. The major component of the increase
vehicle-treated groups. d, Vehicle; s, calcipotriol. Error bars: mean ± in shed hairs overall was telogen hair, compared with frac-
SEM. tured and anagen hairs, in both cases (P ¼ 0Æ0002) (Fig. 5).

2005 British Association of Dermatologists • British Journal of Dermatology 2005 153, pp103–112
110 Mechanism of cytotoxic alopecia and effect of vitamin D, T.O. Bleiker et al.
Fractured and anagen hairs represented a relatively small com-
ponent of the increased shedding, fractured hairs predomin-
ating (Fig. 5). Of the telogen hairs (Fig. 6), the proportion of
normal telogen hairs fell to approximately 50% for most of the
chemotherapy treatment, the remainder being replaced with
telogen hairs with tapered proximal shafts with either normal-
sized telogen bulbs or pathologically small telogen bulbs. Frac-
turing of tapered hairs with small telogen bulbs produced
typical ‘exclamation mark’ hairs. The major pathological
change in all hair types was tapering of the proximal shaft,
which occurred in approximately 50% of all shed hairs
throughout the chemotherapy treatment (Fig. 6).
At the time of maximal shedding, the mean plucked telogen
count (Fig. 8) approximately doubled compared with baseline
(P ¼ 0Æ0003). As with shed hair, there was a marked increase
in tapered hair, rising from practically zero to 38% (P ¼
0Æ0003), but unlike in the shed hair there was also a marked
increase in the proportion of fractured hairs, which represen-
ted 33% of all plucked hairs (P ¼ 0Æ0003). It is highly likely
that many of these were broken by the plucking procedure,
being made more vulnerable to fracturing by tapering prox-
imal shafts. The evidence for this is that 61% of plucked
tapered hairs were fractured compared with only 18% of shed
tapered hairs (P ¼ 0Æ0003). It is also likely that these fractured
hairs were in telogen or about to enter telogen, as evidenced
by the fact that most (
80%) shed hairs throughout chemo-
therapy were telogen hairs.
Thus the hallmark of the effect of cytotoxics was a large
increase in shed telogen hairs; tapered hairs of all types, in
both shed and plucked hair; and fractured hairs, particularly
in plucked hair. More severely affected patients in terms of
increased shedding produced greater proportions of tapered
and fractured hairs, with a consequent decrease in the propor-
tion of normal telogen hairs, but also an absolute increase in
normal telogen hair shed rate (Table 1).
The NAH is a measure of regrowth produced by hairs
entering the next growth cycle12 and is expressed as the pro-
portion of hairs having the tapering distal tips as produced at
the onset of anagen. That the chemotherapy-induced shed hair
was being replaced by regrowth in the next cycle is evidenced
by the large increase in NAH, from a baseline of 15% to a
maximum of 57% (P ¼ 0Æ0007); the strong correlation
between shed rate and subsequent NAH, R ¼ 0Æ78 (P ¼
0Æ0001); and the fact that in five patients with very severe
shedding NAH reached virtually 100%.
Amalgamating these results, the major effect of chemother-
apy was, firstly, a progressive tapering of the hair shaft. This
is likely to be the result of a progressive decreased cellular
output from the hair matrix. Fracturing of the hair shaft may
occur at this stage, resulting in shedding of fractured hairs.
The tapering was followed by entry of a large majority, if not
all, of the affected follicles into telogen, as evidenced by the
NAH data. This resulted in shedding of telogen hairs, which
represented the majority of shed hairs in the present study.
The size of the telogen bulb was either normal or diminutive.
Although the major effect of chemotherapy was to produce

2005 British Association of Dermatologists • British Journal of Dermatology 2005 153, pp103–112
hair shaft tapering and subsequent entry of the follicle into
telogen, there was an increased loss of normal telogen hairs
with increase in total shed rate (P ¼ 0Æ01) (Table 1) and this
may represent a concurrent classical telogen effluvium.
The question arises as to how representative are these
results of chemotherapy, where a limited number of drugs
was used at a fixed dose, for cytotoxic therapy in general.
Crounse and van Scott4 investigated a wide range of cytotoxics
and concluded that the follicular response patterns were com-
mon to all and were not drug specific. This would appear
likely, as cytotoxics have a common mechanism of action, i.e.
DNA damage resulting in cycle arrest and apoptosis. In the
present study there was a very wide range of severity of
impact between patients: some patients were barely affected
and others experienced total loss. With greater doses of che-
motherapy, there is no reason to suspect that the basic
response of hair shaft tapering culminating in telogen would
be different, as evidenced in the present study by the patients
who experienced total hair loss. However, more dramatic
tapering might result in a greater tendency to fracture in situ,
resulting in a greater proportion of shed fractured hair com-
pared with tapering telogen hairs as was found in the more
severely affected patients in the present study.
In previous studies, hair loss from cytotoxics has largely
been attributed to the loss of fractured hair.1–4,14,15 This hair
is sometimes described as dystrophic16 or dystrophic anagen
hair or tapered broken hair.14 Few authors1,2 have commented
on coincidental loss of telogen hair. The present study con-
firmed that tapered fractured hairs were characteristic of the
process, but tapered telogen loss was more frequent and
equally characteristic. Also, some studies3,4 have included data
only on plucked hair samples and have extrapolated the results
to shedding. This may be misleading, as the traction involved
in plucking may produce fracturing in hairs which would not
naturally be shed as fractured hairs. This is illustrated by the
present study, where fractured hairs were common in plucked
hair (33%) and relatively rare in shed hair (12%) and tapered
hairs were usually fractured in plucked hair (61%) but not in
shed hair (18%), as discussed above. Two hair shaft changes
causing hair shedding have previously been described: an
acute constriction3,4 and a progressive tapering.3,4,14 Only
progressive tapering was observed in the present study. This
may be a function of cytotoxic dosage. Smaller doses of cytox-
ics result in lesser degrees of DNA damage which may induce
cell cycle arrest, in either G1–S or G2–M depending on the
nature of the DNA damage, while more profound damage, as
would be anticipated from aggressive modern chemotherapy,
results in apoptosis. It could be argued that cycle arrest causes
acute constriction whereas apoptosis results in progressive
tapering. Such acute constrictions, produced by lower cyto-
toxic doses, might be envisaged to result in hair shaft fractur-
ing and subsequent continued growth of the affected follicle
in anagen, as previously suggested.3,4 However, in the present
study some patients received low ‘effective’ doses of chemo-
therapy such that there was minimal shedding but acute con-
striction of the hair shaft was not seen. Apoptosis has been
 Mechanism of cytotoxic alopecia and effect of vitamin D, T.O. Bleiker et al. 111
demonstrated in hair follicles following cytotoxic treatment.5–7
Few authors have commented on an effect of cytotoxics on
the hair cycle, except where an effect has been denied,3,4 and
a delay in the next anagen cycle in mice has been described.10
Again, where an effect was denied in humans the study was
of plucked hairs only and there was no direct evidence against
a cycle effect. We believe that the present study provides
incontrovertible evidence that there is a major effect on the
hair cycle.
The hair loss from cytotoxic drugs is often referred to in
the literature as anagen effluvium.17 This is normally applied
to situations of fractured hair loss and there is a tacit
assumption that the relevant follicles remain in anagen. How-
ever, the present study has shown that follicles have either
entered or are in the process of entering telogen and that a
more prominent component of the shed hair may be tapered
telogen hairs, as in the present study. It is emphasized by
many authors that the impact of cytotoxics is on actively
growing anagen hairs.3,4,10,14 It is also a fact, however, that
in a classical telogen effluvium, the causative follicular insult
is also impacting anagen follicles in the first instance, after
which these enter telogen and hairs are lost as such. The dif-
ference in the loss induced by cytotoxics is that the hair shaft
first tapers prior to entering telogen. We are aware of no rig-
orous proof in the literature that hairs affected by such taper-
ing ever remain in anagen to produce a regrowth. Whether
the hair shaft breaks is then a function of the balance
between loss of tensile strength, consequent on the degree of
hair tapering of the hair shaft, and prevailing external trac-
tion forces. As part of the process of cytotoxic follicular dam-
age there may, in fact, be a concurrent classical telogen
effluvium as mentioned above. A classical telogen effluvium
is caused by a systemic insult, but only a limited population
of scalp follicles is vulnerable (up to
50%17), perhaps
because they are older in terms of the time they have been
in anagen (anagen age) and are approaching normal telogen.
Cytotoxics might be envisaged to trigger the natural apoptotic
process of telogen immediately in such hairs, while other
hairs, robust against this signal, remain in anagen to endure
the effects of apoptosis of the matrix which eventually also
results in telogen. We would therefore maintain that the so-
called anagen effluvium is a special example of a telogen
effluvium and would propose the term ‘atrophic telogen
effluvium’. This embraces two fundamental aspects of the
process: ‘atrophy’ (narrowing) of the hair shaft and entry of
affected follicles into telogen. This would be defined as a
process involving progressive hair shaft tapering, an efflu-
vium of predominantly a variable mixture of tapered frac-
tured and telogen hairs and followed by premature entry of
affected follicles into telogen.
There is a striking similarity between the process described
here and hair loss in alopecia areata. A prominent component
has been shown to be premature entry of follicles into telogen
and fracturing of the proximal shafts of the telogen hairs.18
The exclamation mark hair of alopecia areata is a telogen hair
with a diminutive telogen bulb whose shaft has broken.

2005 British Association of Dermatologists • British Journal of Dermatology 2005 153, pp103–112
Exactly similar hairs were evident in the present study. This
would lend support to the suggestion that apoptosis of the
hair matrix is induced in alopecia areata,19 but here presuma-
bly by cytotoxic immune attack.
Atrophic telogen effluvium as defined here is a pattern
response to cytotoxic drugs. However, there are other than
therapeutic sources of such agents and certain accidental and
criminal poisons and plant poisons are recognized as apopto-
sis-inducing agents, e.g. arsenic,20 colchicine21 (from Gloriosa
superba), minosine22 (from Leucaena glauca) and abrin23 (from
Abrus precatori). Such sources should be considered in patients
presenting with atrophic telogen effluvium with no obvious
cause.
In conclusion, cytotoxic drugs produce tapering of the
proximal hair shaft, which is likely to be the result of apopto-
sis, and premature entry of the follicle into telogen. This
results in an effluvium of a mixture of tapering telogen hairs
and fractured hairs, if the tapering hair shaft happens to break.
There is also concurrent loss of normal telogen hairs, the
cause of which may have a different mechanism. We propose
the term ‘atrophic telogen effluvium’. A similar mechanism
may be involved in alopecia areata and certain accidental and
criminal poisons and plant poisons. There was no detectable
protective effect of topical calcipotriol.
Acknowledgments
This was a single-centre study sponsored by Leo Pharma. We
thank Nicolas Taub, Lecturer in Medical Statistics, Leicester
University, for helpful statistical comment.
References
1 Zaun H. Tierexperimentelle Untersuchungen zur Pathophysiologie
der ‘gemischten Alopecie’. Arch Klin Exp Dermatol 1964; 221:75–
84.
2 Orfanos CE, Gerstein E. Haarausfall nach Zytostatica. Arztl Kosmetol
1976; 6:96.
3 van Scott EJ, Reinerton RP, Steinmuller R. The growing hair roots
of the human scalp and morphological changes therein following
amethoperin therapy. J Invest Dermatol 1957; 29:197–204.
4 Crounse RG, van Scott EJ. Changes in scalp hair roots as a measure
of toxicity from cancer chemotherapeutic drugs. J Invest Dermatol
1960; 35:83–90.
5 Cece R, Cazzaniga S, Morelli D et al. Apoptosis of hair follicle cells
during doxorubicin-induced alopecia in rats. Lab Invest 1996;
75:601–9.
6 Goldberg MT, Tackaberry LE, Hardy MH et al. Nuclear aberrations
in hair follicle cells of patients receiving cyclophosphamide.
A possible in vivo assay for human exposure to genotoxic agents.
Arch Toxicol 1990; 64:116–21.
7 Lindner G, Botchkarev VA, Botchkareva NV et al. Analysis of apo-
ptosis during hair follicle regression (catagen). Am J Pathol 1997;
151:1601–17.
8 Schilli MB, Paus R, Menrad A. Reduction of intrafollicular apopto-
sis in chemotherapy-induced alopecia by topical calcitriol-analogs.
J Invest Dermatol 1998; 111:598–604.
9 Jimenez JJ, Yunis AA. Protection from chemotherapy-induced alo-
pecia by 1,25-dihydroxyvitamin D3. Cancer Res 1992; 52:5123–5.
112 Mechanism of cytotoxic alopecia and effect of vitamin D, T.O. Bleiker et al.
10 Paus R, Schilli MB, Handjiski B et al. Topical calcitriol enhances
normal hair regrowth but does not prevent chemotherapy-induced
alopecia in mice. Cancer Res 1996; 56:4438–43.
11 Hidalgo M, Rinaldi D, Medina G et al. A phase I trial of topical top-
itriol (calcitriol, 1,25-dihydroxyvitamin D3) to prevent chemother-
apy-induced alopecia. Anticancer Drugs 1999; 10:393–5.
12 Berth Jones J, Hutchinson PE. Novel cycle changes in scalp hair are
caused by etretinate therapy. Br J Dermatol 1995; 132:367–75.
13 Bonadonna G, Brusamolino E, Valagussa P et al. Combination che-
motherapy as an adjuvant treatment in operable breast cancer.
N Engl J Med 1976; 294:405–10.
14 Maibach HI, Maguire HC. Acute hair loss from drug-induced abor-
tion. N Engl J Med 1964; 270:1112–13.
15 Harms M. [Alopecia and hair changes following colchicine ther-
apy]. Hautarzt 1980; 31:161–3.
16 Simpson NB. Diffuse alopecia: endocrine, metabolic and chemical
influences on the follicular cycle. In: Diseases of the Hair and Scalp
(Rook A, Dawber R, eds), 2nd edn. Oxford: Blackwell Science,
1991; 136–66.

2005 British Association of Dermatologists • British Journal of Dermatology 2005 153, pp103–112
17 Kligman AM. Pathologic dynamics of human hair loss. I. Telogen
effuvium. Arch Dermatol 1961; 83:175–98.
18 Eckert J, Church RE, Ebling FJ. The pathogenesis of alopecia areata.
Br J Dermatol 1968; 80:203–10.
19 Tobin DJ, Fenton DA, Kendall MD. Cell degeneration in alopecia
areata. An ultrastructural study. Am J Dermatopathol 1991; 13:248–
56.
20 Li YM, Broome JD. Arsenic targets tubulins to induce apoptosis in
myeloid leukemia cells. Cancer Res 1999; 59:776–80.
21 Goh EL, Pircher TJ, Lobie PE. Growth hormone promotion of tub-
ulin polymerization stabilizes the microtubule network and pro-
tects against colchicine-induced apoptosis. Endocrinology 1998;
139:4364–72.
22 Mikhailov I, Russev G, Anachkova B. Treatment of mammalian
cells with mimosine generates DNA breaks. Mutat Res 2000;
459:299–306.
23 Griffiths GD, Leek MD, Gee DJ. The toxic plant proteins ricin and
abrin induce apoptotic changes in mammalian lymphoid tissues
and intestine. J Pathol 1987; 151:221–9.