British Journal of Dermatology (1991) 124, 361-364.
ADONIS 000709639100084G
Isotretinoin treatment alters steroid metabolism in
women with acne
M.RADKMAKER. M.WALLACE.* W.CUNLIFFEt AND N.B.SIMPSON
Departments of Dermatologii and "Biochemistrii. Glasgow Roijal Inftrmarij.
•^Department of Dermatology, Leeds General Infirmarij. l^eds. U.K.
Glasgow
Accepted for publication 1 October 1990
Summary
The effect of isotretinoin (Roaccutane) on serum steroids and urinary steroid metabolites was
investigated in seven female patients receiving the drug for treatment of severe acne over a 16-week
period. Serum concentrations of deiiydroepiandrostcrone suiphate (DHAS). androstenedione (A2).
and free androgen index (FAl) were not significantly altered. There was a significant fall in
testosterone during treatment and a signiticant reduction in the 24 h urinary excretion of
androsterone. tetrahydrocortisone (THE) and tetrahydrocortisol (THF) from week 8 onwards and for
aetiocholanolone and allo-THF from week 12 (P<() 05). Although pretreatment levels of urinary
steroid metabolites were not abnormal, the ratios of the 5a/5// metabolites (androsterone:aetiocholanolone and allo-THF:THF) were at the upper limit of the reference range and were lowered after
treatment, suggesting that 5a-reductase activity is sensitive to isotretinoin.
Tsotretinoin suppresses sebaceous gland activity by
blocking the differentia lion of basal cells into sebocytes.'
The exact mechanism is not known but because these
glands are androgen-responsive the effects of isotretinoin on androgens have been investigated. A number of
studies suggest that sebaceous gland inhibition is not
hormonally related but may cause secondary changes in
the activity of the skin steroid 5a-reductase.^ To obtain
more information on the response of steroid Sa-reduction to treatment with isotretinoin we measured urinary
steroid metabolites by capillary column gas liquid
chromatography and serum androgen levels (testosterone, androstenedione (A2). and dehydroepiandrosterone sulphate (DHAS)) by radioimmunoassay during
the course of 16 weeks of therapy with isotretinoin
(Roaccutane).
Methods
Subjects
Seven women, median age 26-7 years (range 18-44
years) with persistent acne that was non-responsive to
antibiotics were studied. All were using barrier methods
of contraception and had a normal menstrual pattern
with regular cycles of between 21 and 35 days. There
was no evidence of hirsutism, hair loss or virilism.
Correspondence: Dr M.Rademaker. Department of Dennatology.
Glasgow Royal Infirmary. (liasgow CA OSF. U.K.
Pretreatment liver function and fasting serum lipids
were normal and remained so during treatment. All
seven women were given a 16-week course of isotretinoin (Roaccutane. Roche Products Ltd) at a dose of i
mg/kg/body weight per day. The mean total acne grade'
before starting was 3 5 (95% confidence interval, 0 3 6 6) with a mean facial gradeof 1 4 (01-2-7). Twentyfour-hour urine samples were collected Immediately
before starting the drug and at 4-weekly intervals
thereafter. Facial and total acne were also scored at
monthly intervals.'
Hormone analysis
Sex hormone binding globulin (SHBG) was measured by
time-resolved fluoroimmunoassay ('Delfia' SHBG kits.
Pharmacia). Testosterone. DHAS and A2 were measured by conventional 'in house" radioimmunoassays.
Derived free testosterone levels were calculated from the
total testosterone and SHBG according to the method of
Nanjee and Wheeler.'* Urinary steroid profiles analysis
was performed by capillary column chromatography by
a method similar to that of Bevan ct al.'' Separation of
steroid methyloxime-trimethylsilyl ether derivatives was
carried out on a model 438 gas chromatograph (PackerBecker BV. Delft. The Netherlands) fitted with a 'Flexsil'
2 5 mxO-52 mm capillary column coated with OV-1
(Phase-Sep. Queensferry. Clwyd). Measurements were
made of the most abundant androgen metabolites
361
362
M.RADEMAKER et al
common side-effects of isotretinoin (dry lips and eyes)
but remained well throughout the treatment period and
experienced no signiiicant alteration of their liver function tests or fasting serum lipids. Menstrual patterns did
not change.
The effects of isotretinoin treatment on serum androgens and SHBd concentrations are shown in Table 1.
The mean baseline concentrations of testosterone.
DHAS. and SHBG were normal while a modest elevation
in A2 was observed (due to an abnormally high A2 level
of 1 6 1 nmol/1 in one patient suspected of suffering from
polycystic ovarian disease: others were all within the
normal range 0-6-8-8 nmol/I). The mean androstenedione concentration fell during treatment: tbis fall was
not significant and was largely due to normalization of
serum A2 in the patient with an elevated pretreatment
level. Over the 16-week treatment period tbere was a
significant fall in serum testosterone at weeks 8 and 16
(P<0-05).
(androsterone, aetiocholanoione) and glucocorticoid
metabolites (telrahydrocortisone (THE), tetrahydrocortisol (THF). and allo-tetrahydrocortisol (allo-THF)).
Ratios of 5a: 5/^ pairs (i.e. androsterone:aetiocholanoione and allo-THF:THF ratios) were calculated to provide an index of exogenous 5a-reductase activity.
Statistics
Results are presented as mean ± standard error of the
mean (95% confidence interval). Statistical analyses
were by the non-parametric tests of analysis of variance
and by the Wilcoxon sign-rank test.
Results
Acne was virtually abolished in all patients by the end of
the 4-month treatment period: the mean total acne score
was 0-25 ( -O-O5-O-5) and the mean facial acne score
was 0-07 (0()2-0-12). All patients experienced the
Weeks
Hormone (reference range)
Baseline
Testosterone
( 1 - 3 2 nmol/I)
1-9 ±0-3
l-7±O-2
I)ehydrt>epiandrosterone
sulphate (2-11 |Umol/l)
7-4±l-8
Sex hormone binding
globulin
(30-120 nmol/n
Androstenedione (nt»ol/l)
(0-6-^8-8 nmol/1)
12
16
l-2±0-l'
1 5 ±0-2
l-3±01*
10-7 ±3-8
7-7±l-6
7-9 ±2-0
7-2±l-7
79±19
71±17
71±21
7O±21
71±2O
9 4±2 3
8 1 ±1-4
7-5±l-l
8-9±2 4
7-4±l-8
3-4±l-4
2-8 ±0-7
2-5±0-6
3-5 ±1-5
2-4±]l
Table 1. Serum steroid concentrations
(meaniSEM) of seven female patients with
severe acne during a 16 week cour.se of
isotrctinoin therapy (1 mg/kg/day). 'r<{)()
compared with baseline
Free androgen index
Weeks
Hormone (reference range)
Baseline
12
16
Androsterone
(!O29±457/(g)
1205±286
1090±226
596±12r
538±664'
664±194'
Aetiocholanoione
O335±525/ig)
7iO±l57
609 ±207*
74O±5O''
1086±241
1175 ±288
Tetrabydrocortisone (THE)
(1221±946/ig)
27O8±527
285S±759
I480±234' 1553±322
1587±292'
Tetrahydroc-ortisol (THF)
(16l7±525;jg)
1331 ±218
1298±246
778±161- 835±214
750±133"
AIlo-Tetrahydrocortisol
(allo-THF)
(1008 ±491 ;ig)
1293 ±302
12}2±439
572 ±184
548 ±117"
450±10]-
Table 2. Total 24 hour urinaty excretion of
steroids (mean±SFAl. jUg/24 h) of seven
female patients with severe acne during
a 16-week course of isotretinoin therapy
(1 mg/kg day). • p < 0 05 compared with
baseline
ISOTRETINOIN TREATMENT AND STEROID METABOLISM
L2
1.1
1.0-
S 0.8 •
0.7-
0.6
0.5-'
8
12
16
Week
Figure 1. Ratio of 5a: 5^ urinary steroid metabolites (mean ± SEM) in
.seven female patients with severe acne during a 16-week course of
isotretinoin therapy U mg/kg/day) ( ' P < 0 - 0 5 from baseline). • .
androsterone:aetiochoianolone; A, aUo-THF:THF.
Total 24-hour excretion of the selected steroid metabolites are shown in Table 2: the baseline concentrations
ofthe patients were within the expected limits for normal
women. Over the study period, a marked reduction in
urinary steroid metabolites was observed with significant falls of androsterone. THE and THF from week 8
onwards and aetiocholanolone and allo-THF from week
12 (P<nnO5).
Baseline ratios of both androsterone: aetiocholanolone (M2±()-()9) and allo-THE:THF (0-89±0-n5)
were at the upper limit of normal (normal mean values
O-65±O2. range O - 3 5 - M 5 and ()-62±()2. range
( ) 2 - l - 0 . respectively). These high ratios fell significantly during treatment (Fig. 1).
Discussion
A number of studies suggest that sebaceous gland
involution induced by isotretinoin is not androgenmediated. This conclusion is based on the lack of any
dramatic effect of the drug on circulating precursor
androgens. Lookingbill et al^ found that serum androgens testosterone. DHA and DHAS were not altered by
treatment although an increase in free testosterone was
found and Matsuoka et nl.^ found no change in circulating concentrations of cortisol. androstenedione. DHA
and DHAS. Some of these findings are confirmed in this
study where no significant change in DHAS. androstenedione and free androgen index occurred on treatment.
363
We did however find a small but significant fall in the
serum level of testosterone.
A more significant inhibitory effect was observed' on
production of 5ci androgens (dihydrotestosterone and
3a-androstenediol glucuronide in women; 3a-androstenediol glucuronide in men) and this was thought to be
due to the reduction in sii^e of the sebaceous gland. We
have previously reported significant but small reductions in serum dihydrotestosterone in women on isotretinoin.' Whether or not isotretinoin has a direct effect on
steroid 5a-reduction in vivo has. until now. been uncertain. In vitro studies of testosterone metabolism in
hamster flank organ showed lack of inhibition of 5areductase by isotretinoin" but in female rats isotretinoin
treatment caused a decrease in steroid Sa-reductase
activity in both skin and liver.^ In cultured human
prostatic cells 5a-reductase can be inhibited by retinoic
acid through competitive inhibition of cofactor action
(NADPH)'" yet in men circulating dihydrotestosterone is
unchanged throughout isotretinoin therapy.In the present study marked reductions in the major
androgen and glucocorticoid metabolites were found at
the same time as decreased levels of 5a-reduction. 5areduction of androgens occurs in the peripheral tissues
as well as the liver while glucocorticoids are mainly 5areduced in liver. It is therefore unlikely that these
changes can be accounted for solely by drug-induced
reduction in the size ofthe sebaceous gland and a direct
effect on liver 5a-reduction is postulated. A clue as to
how these aspects of steroid metabolism are related is
given by recent reports on steroid metabolism in women
with hirsutism and polycystic ovarian disease."'- In
such patients it is postulated that increased 5a-reduction
in the liver enhances hepatic cortisol metabolism and
circulating cortisol concentrations are maintained
within normal limits by increasing the secretion of
corticotrophin which simulates adrenal production of
both glucocorticoids and androgens. It may be that, by
inhibiting 5a-reduction within the liver, isotretinoin
reduces cortisol excretion, and thereby reduces corticotrophin secretion, the stimulus for adrenal steroid
production. Alterations in steroid metabolism of a
similar nature occur in patients suffering from acute
intermittent porphyria but in this case 5a-reduction may
be secondary to endogenous rather than exogenous
factors.'^
Further work is required to verify and extend these
findings. It is of course possible that the observed effects
on steroid metabolism are related to the as yet unknown
mechanism of action of retinoids in controlling cell
differentiation.
364
M.RADEMAKER et al
References
1 Landthaler M, Kummermehr J. Wagner A. Plewig G. Inhibitory
effects of I i-(/.\ retinoic acid in human sebaceous glands. Arch
Dermatol Res 19S(); 269: 297-^09.
2 Lookingbill DP. Hemers I,M. Tigelaar RE. Shalita AR. rJTect of
isolretinoin on serum levels of precursor and peripherally derived
androgens in patients with acne. Arch Dermatol 1988: 124:
540-J.
3 Burke BM. Cunliffe Wf.Theassessmentof acne vulgaris: the Leeds
technique, Br I Dermaio! 1984: l i l : 8 J - 9 2 .
4 Nanjee MN. Wheeler MJ. Plasma free testosterone—is an index
sufficient.^ Ann Clin Hiochem I98S: 22: i87-9(),
5 Bevan BR. Savvas M. Jenkins JM et al. Abnormal steroid excretion
in gcstational trophobia.stic disease complicated by ovarian thecalutein cysts. / Clin Pathol 1986: 39: 6 2 7 - i 4 .
(i Matsuoka LY. Wortsman J, Lifrak OT et al. Inflect ol' isotretinoin in
atne is not mediated by adrenal androgens. / Am Acad Dermalol
1989:20: 128-9.
7 Simpson NB. Cunliffe WJ. Rademaker M el al. The sebaceous gland
8
9
10
11
12
13
as a source of androgen—an hypothesis. In: Acne and Re/atcd
Disorders. (Marks R. Plewig G eds). I/indon: M Dunitz Ltd. 1989.
Gomez EC. Moskowitz RJ. Effect of 11-cis-rt'tinoic acid on hamster
flank organ. / Invest Dernmtol 1980: 74: 392-7.
Gracf V. Ritter B. Goerz G. Eiffect of synthetic retinoids on 5areductasesin rat liver and skin. Proc of Sth Inl Cong of Endocrinologii
{July 1988. Kyoto. Japan), p 423.
HalgunsetJ, Sunde A. LundmoPI. Retinoic acid (RA): an inhibitor
of 5a-reductase in human prostatic cancer cells. / .'iteroid Biochem
1987; 28: 731-6.
Buchanan LM. Wallace AM. Thomson J. Abnormal steroid
metabolism in hirsute women, Scott Med / 1990: 35: 155.
Stewart PM, Shackleton CHL. Beastall GH. Inwards CRW, 5areductase activity in polycystic ovary syndrome, hincet 1990:
335:431-3.
Herrick AL. Wallace AM. McCoU KEL et al. Association of elevated
sex hormone binding globulin with abnormal steroid metabolism
in acute inteimittant porphyria, / F.ndoiTinol 1988: Suppl 119:
114.