British Journal of Dermatology (1982) X07, 583-590.

A double-blind study of the effects of

13-cis-retinoic acid on acne, sebum
excretion rate and microbial population

KATHRINE KING*t+, D.H.JONESt, DIANA C.DALTREY* AND

W.J.CUNLIFFE$
^School of Health & Applied Sciences, Leeds Polytechnic, fUniversity Department of Immunology and
^Departments of Dennatology, The General Infirmary, Leeds, and St James's (University) Hospital, Leeds

Accepted for publication 10 March 1982

SUMMARY
Forty-eight acne patients were treated orally with 13-cis-retinoic acid in a double-blind dose
response study. There was a marked clinical improvement with a concomitant reduction in
sebum excretion rate (SER) and production rate of free fatty acids (FFA). Microbial numbers
decreased significantly; the decrease in propionibacteria was greater than that of aerobic
bacteria. The decline in micro-organisms occurred after the reduction in sebum and FFA
production. This suggests that the effect of the drug upon micro-organisms is secondary to the
change in sebum excretion but it is nevertheless an important factor in the resolution of the aene.

Despite the overall efficiency of acne treatment, about 5",, of patients are unresponsive. Studies
of these patients indicate that systemic treatment with 13-cis-retinoic acid (a synthetic
derivative of vitamin A) causes a significant reduction in the production of sebum and a
significant clinical improvement with prolonged remission (Peck et al.., 1979). There is also
limited m vivo data to suggest that this drug alTects the skin micro-organisms (Jones, Cunliffe &
Cove, 1981; Weissmann, Wagner & Plewig, 1981). Thus the aim of this study was to investigate
and correlate the effects of 13-cis-retinoic acid on the acne, the microbial skin population, the
SER, the production rate of FFA and the levels of corneocyte shedding.

PATIENTS AND METHODS

Patients
Forty-eight patients (twenty-four male, twenty-four female, mean age 25 years) with a history of
acne which would not respond to conventional therapy were included in the trial. All patients
had received several courses of antibiotics, often large doses for 6 months or more. They all
stopped other treatment at least i month prior to the start of the study.
0007-0963/82/1 ioo-os83$O2.cx] iQ 1982 British Association of Dermatologisu
583
584 K.King et al.
Treatment schedule and clinical assessment
Patients were randomly allocated to three groups and given 10, 05 or 01 mg/kg bodyweight
13-cis-retinoic acid daily for 16 weeks. Patients were seen prior to treatment and thereafter at
monthly intervals for evaluation. The acne was assessed using standard lighting and the severity
was graded on a o~io scale on the face, back and chest. All patients were told to take adequate
contraceptive precautions.

Estimation of sebum excretion rate (SER)

The SER (/ig sebum per cm^ skin per min) was monitored before and during treatment using the
gravimetric method of Strauss & Pochi (1961) modified by CunlifFc & Shuster (1969).

Estimation of the production rate of free fatty acids (FFA)

Sebum collected by the absorbent paper technique was reconstituted in heptane and the FFA
content determined by titration against tetra-n-butyl ammonium hydroxide in redistilled
methanol (Dole & Meinertz, i960). Thymol blue was used as indicator. The production rate of
FFA was expressed as ng FFA produced per cm^ skin per min.

Microbiological sampling
At each visit skin micro-organisms were sampled from the right cheek using the 'scrub'
technique (Williamson & Kligman, 1965). The method of Miles & Misra (1938) was used to
determine the number of bacterial colony-forming units present per cm' skin. Aerobic bacteria
were counted after 48 h incubation at 37"C on a heated blood agar (blood agar base, Oxoid,
+ 7*^,, horse blood). Propionibacteria were enumerated after seven days anaerobic incubation at
37 C in a GasPak jar (atmosphere 10",, CO?, 90",, H^) on a reinforced clostridial medium
(Oxoid) + 1-5% agar (Oxoid) containing 6 /Jg/ml furazolidone (Eaton Laboratories).
Numbers of the yeast Piiyrosporum spp. were estimated microscopically using a membrane
filtration method (Mulvany, 1969).
Results were expressed as logm («+ i) colony forming units per cm^ skin.
A total count of corneocytes per sq. cm removed during 'scrub' sampling was estimated using
a haemocytometer (McGinley, Marples & Plewig,

RESULTS

The clinical response (Fig. i) was marked (/'<o 01, unpaired Wilcoxon rank sum test). There
was a gradual reduction in acne grade over the 16 weeks treatment with afinalimprovement of
70%. There was no difference in effectiveness between the doses. The SER (Fig. 2) was
significantly reduced at 4 weeks (P < o 001), the reduced level being maintained thereafter. The
production rate of FFA followed a similar pattern (Fig. 3) with a significant reduction after 4
weeks treatment (P<oooi). Neither the reduction in the SER nor the reduction in FFA
production rate was dose dependent (Student t-test).
The changes in bacterial fiora per cm' of skin are shown in Figs 4 and 5. The numbers of
aerobic bacteria were significantly decreased by the eighth week on all doses (P<OO5).
However, after 16 weeks of therapy the count of 10 mg/kg group had risen and the overall
reduction in count (o-t6 weeks) was no longer significant (paired Wilcoxon rank sum test). The
anaerobic population was significantly smaller after 12 weeks treatment on all doses (P<005,
paired Wilcoxon rank sum test) and remained significantly reduced until the cessation of
 Effects of 13-cis-retinoic acid 585

• O-I mg/kg 1 s.e. /? = 13

0 - 5 m g / k g ± s.e. n = Z \

I'O mg/kg 1 s.e. / i : 14

2-0

1-5

1-0

O5

B 16
Weeks
FIGURE 1. Acne grade on the face before and during treatment with r3-cis-retinoic acid.

therapy. There was no evidence of a dose response in the reduction of either the aerobic or
anaerobic bacteria (unpaired Wilcoxon rank sum test).
The numbers of the yeast Pityrosporum spp. are difficult to estimate, and it is not always a
member of the skin flora. The data (Fig. 6) represent only those twenty-six patients in whom
Pityrosporum spp. was observed before the commencement of treatment. A significant decrease

TABLE I. The tiumber of corneocytes shed from cheek

skin during 'scrub' sampling before and during treatment
with 13-cis-retinoic acid

Mean log,o (corneocytes) cm ^ ± s.e.

Treatment
time in weeks 10 mg/kg 0-5 mg/kg O-I mg/kg

0 461 ±008 4'72±0'O7 4-63 + 0-08

4 4-86 ±0-06 4-88+0-04 477±O'O7
8 4-97 + 006 4-91 +003 487 + 0-09
12 4-93+003 4-79 ± 005 4 73 ± o-o8
16 4-84 ±0-02 4-71 ±0-04 463 ± 0-09
586 K.King et al.
2-0
O-I mg/kg ± s.e. n = l3
0 5 mg/kg ±s.e. rt=2l
I'O mg/kg ± s.e. /?= 14
1-5

10
S

E
Seb

0-5

0 4 8 12 16
weeks
FIGURE 2. Sebum excretion rate before and during treatment with 13-cis-retinoic acid.

2501-

« 200

O-I mg/kg ± s.e. /? = I3

0 ' 5 mg/kg ± s.e. n = 2l
150
1-0 mg/kg ± s.e. n = l 4

100

50 -

8 12 (6
Weeks
FIGURE 3. Production rate of free fatty acids before and during treatment with 13-cis-retinoie acid.
 Effects of 13-cis-retinoic acid 587

5'Or-

4-0

P<.00\

Z-0
01 mg/kg 1 s.e. /1 = I3
0-5 mg/kg l s . e . / ) = 21
1-0 mg/kg ± s.e. n - 14

16
Weeks
FIGURE 4. Population of aerobic skin bacteria before and during treatment with i3-cis-retmoic acid.

5-01-
01 mg/kg Is.e. ^ = 13
0-5 mg/kg ts.e./7:2l
4-0 - mg/kg *s.e./7 = 14

3-0

2-0

12 16
Weeks
FIGURE 5. Population of anaerobic skin bacteria before and during treatment with 13-cis-retinoic acid.
588 K.King et al.
50
i O'l mg/kg + s.e. n
^ 0-5mg/hg ± s.e. n
• 1-0 mg/kg + s.e. n

I 3-0
I

1-0 -
P<Q-O\

8 16
Weeks
FIGURE 6. Population of Pityrosporum spp. before and during treatment with 13-cis-retinoic acid.

inthedensity of this organism was observed from week 8 onwards (P<o 05, paired Wilcoxon
rank sum test); again there was no evidence of a dose response (unpaired Wilcoxon rank sum
test).
There was no significant change in numbers of corneocytes removed during sampling
throughout the course of treatment (Table i).

DISCUSSION
13-Cis-retinoic acid undoubtedly aids the resolution of acne in patients who bave previously
failed to respond to conventional therapy. Our clinical data confirms the studies of Farrell,
Strauss & Stranieri (1980) and Strauss et al. (1981) who found no evidence of a dose response.
The clinical improvement is thought primarily to be due to the reduced SER (68-76%
reduction after i month of therapy); Farrell et al. (1980) report similar findings. The sebaceous
glands are known to atrophy after 3 weeks of treatment (Landthaler ei al., 1980). Sebaceous
gland activity in man is androgen mediated, but studies (Sansone-Bazzono, Seeler &
Cummings, 1980) suggest that 13-cis-retinoic acid is not an anti-androgen but acts directly
upon gland differentiation.
Although the therapy is clinically successful, it is not without side-effects. Many of our
patients complained of cheilitis and dryness of the skin during treatment. It was thought that the
loosening of the stratum corneum implied by these side-effects, and confirmed experimentally
by Elias et al. (1981), might result in increased removal of skin squames, and hence
micro-organisms, during 'scrub' sampling. In fact changes in the number of corneocytes
removed at each sampling time were negligible. The 'scrub' technique is probably gentle
compared to the tape stripping and friction blisters used by Elias. We are therefore satisfied that
our microbiological data were not affected by an artefact in the sampling technique.
 Effects of 13-cis-retinoic acid 589
13-Cis-retinoic acid caused a reduction in the number of micro-organisms on the skin. This
extends the limited observations of Jones et al. (1981) and Weissmann et al. (1981). The
microbial reduction is probably secondary to the changes in sebum excretion. The population
decline was greatest in those micro-organisms associated with lipid-rich areas, i.e. propionibac-
teria and pityrospora. The density of Propionibacteriwn acnes correlates with the rate of
sebaceous secretion (McGinley et al. 1980). Recently, it has been suggested that the presence of
triglyceride is particularly important (Webster, Ruggieri & McGinley, 1981). Numbers of
propionibacteria began to drop after 4 weeks of therapy; by this time the SER and hence
available triglyceride was very low and perhaps a limiting factor.
Free fatty acids are produced mainly by microbiai breakdown of sebum triglycerides
(Marples, Downing & Kligman, 1971) and their production rate may be regarded as a measure
of microbial function. A reduction in the amount of triglyceride available for breakdown, a
reduced microbial population and change in microbial function, would all be expected to
contribute to the reduced level of FFA production.
The relationship between microbial numbers and acne severity is unclear (Leyden et al.,
1975; Cove, Cunliffe & Holland, 1980a) but micro-organisms probably contribute to the clinical
picture by production of biologically active substances under certain micro-environmental
conditions (Holland, Cunliffe & Roberts, 1978). A reduction in population density or an
alteration in microbial function might therefore contribute to clinical resolution.
The reduction in the numbers of aerobic bacteria was less marked than that in the anaerobes
and yeasts. Most aerobes isolated in this study belonged to the Micrococcaceae. These organisms
may also depend upon some substrate within sebum for growth, e.g. coagulase negative
staphylococci may utilize oleic acid from sebum as a substitute for biotin which is lacking on the
skin surface (Cove, Holland & Cunliffe, 1980b}. Additionally complex interactions may occur
between aerobic and anaerobic populations on the skin, changes in the size and function of one
component of the flora could conceivably affect the others.
There are other ways in wbich 13-cis-retinoic acid could affect the microbial population.
Propionibacteria and pityrospora are found mainly within the infundibulum of the piloseba-
ceous unit which shrinks during treatment, resulting in less space for micro-organisms.
Secondly, 13-cis-retinoic acid may act directly upon microbial cells. So far, work by Weissmann
et al. (1981) has failed to demonstrate such an effect. However, the drug may inhibit the
microbial exocellular enzymes of importance in the skin environment. Lastly, vitamin A is
known to be an immunopotentiator (Jurin & Tannock, 1972) and one synthetic retinoid
CRo-io-9359)isnow thought to have immunomodulatory effects (Tsambaos& Orfanos, 1981).
If 13-cis-retinoic acid has a similar role it could modify the host response to microbial antigen
and this might also contribute to clinical improvement.

ACKNOWLEDGMENTS
We wish to thank Roche Products Limited (in particular Dr A.J.Miller) for providing the drug
and financial support, also. Miss V.B.Cmiech and Mr R. A.Forster for their technical expertise,
and our many consultant colleagues for referring the patients to us.

REFERENCES
COVE, J.H., CUNLIPFE, W.J. & HOLLAND, K . T . (1980a) Acne vulgaris: is the bacterial populaiion size significant? British
Journal of Dermatology, 102, 277.
 590 K.King et al.
COVE, J . H , HOLLAND, K.T. & CUNLIFFE, W.J. f 1980b) The vitamin requirements of siaphylococci isolated from human
skin. The Journal of Applied Bacteriology, 49, 29.
CUNLIFFE, W.J. & SHUsrER, S. (1969) The rate of sebum excretion in man. British Journal of Dermatology, Si^ 697.
DoLi-, V.P. & MEINEKTZ, H . (i960) Micro-determination of long chain fatty acids in plasma and tissues. Journal of
Biological Chemistry, 235, 2595.
ELIAS, P.M., FRITSCH, P.O., LAMPE, M . , WILLIAMS, M.L., BROWN, B.E., NHMANIC, M , & GBAYSON, S . (1981) Reiinoid
effects on epidermal structure, differentiation, and permeability. Laboratory Investigation, 44, 531.
FAKRELL, L.N., STRAUSS, J . S . & STRANIERE, A.M. (1980) The treatment of severe cystic acne with 13-cis-retinoic acid.
Journal of the American Academy of Dermatology, 3, 602.
HOLLAND, K.T., CUNUFFE, W.J. & ROBERTS, C D . (1978) T h e role of bacteria in acne vulgaris; a new approach. Clinkat
and Experimental Dermatology, 3, 253.
JONES, D.H,, CUNLIFFE, W.J. & DOVE, J.H. (1981) 13-Cis-retinoic acid in acne (a double-blind study of dose response).
In: Retinoids, Advances in Basic Research and Therapy (Ed. by C.E.Orfanos et al), p. 255. Springer-Verlag, Berlin.
JURIN, M. & TANNOCK, I . F . (1972) Influence of vitamin A on immunological response. hmm»wlogy,zi,2%y
LANDTHALER, M . , KUMMERMEHR, J., WAGNER, A. & PLEWIG. G . (1980) Inhibitory effects of 13-cis-retinoic acid on
human sebaceous glands. Archives of Dermatological Research, 269, 297.
LEYDEN, J.J., MCGINLEY, K.J., MILLS, O.H. & KLIGMAN, A.M. (1975) Propionibacteria levels in patients v^^ith and
without acne vulgaris. Journal of Investigative Dermatology, 65, 382.
McGiNLEV, K.J., MARPLES, R.R. & PLEWIG, G . {1969) A method for visualising and quamitating the desquamating
portion of the human stratum corneum. Journal of Investigative Dermatology, 53, 107.
MCGINLEY, K.J., WEBSTER, G.F., RUGGIERI, M R . & LEYDEN, J.J. (1980) Regional variations in density of cutaneous
propionibacteria: correlation of/*, acnes populations with sebaceous ^ecreiion. Journal of Clinical AUcrobiologv, 12,
672.
MARPLES, R.R., DOWNING, D . T . & KLIGMAN, A.M. (1971) Control of free fatty acids in human surface lipids by C.
acnes. Journal of Investigative Dermatology, 56, 127.
MILES, A.A. & MISRA, S.S. (193S) The estimation of the bacteriocidal powev of blood. Journal of Hygiene, $&, 7^2.
MULVANY, J . G . (1969) Membrane filter techniques in microbiology. Methods in Microbiology, 1, 205.
PECK, G . L . . O L S E N , T . G . , YODER, F.W., STRAUSS, J.S., DOWNING, D . T . , PANDYA, M . , BUTKUS, D . & ARNAUD-BATTAN-
DiER, J. {1979) Prolonged remission of cystic and conglobate acne with 13-cis-retinoic acid. The New England
Journal of Medicine, 300, 329.
SANSONE-BAZZONO, G . , SEELER, A . K . & CUMMINGS, B . (1980) The effect of 13-cis-retinoic add on the metabolism of
human sebaceous glands in vitro. Clinical Research, 28, 252.
STRAUSS, J.S. & POCHI, P . E . (1961) The quantitative gravimetric determination of sebum production. Journal of
Investigative Dermatology, 36, 293.
STRAUSS, J.S., THOMSEN, R.J., FARRELL, L . N . & STRANIERI, A.M. (1981) Oral retinoids in effects on human sebaceous
glands and nodulocystic acne. In: Retinoids, Advances in Basic Research and Therapy (Ed. by C.E.Orfanos ct al.), p.
237. Springer-Vcrlag, Berlin.
TSAMBAOS, D. & ORFANOS, C.E. (1981) Ultrastructural evidence suggesting an immunomodulatory activity of an oral
retinoid. British Journal of Dermatology, 104, 37.
WEBSTER, G.F., RUGGIERI, M.R. & MCGINLEY, K . J . (1981) Correlation of P. acnes populations with the presence of
triglycerides on nonhuman skin. Applied and Experimental Microbiology, 41, 1269.
WEISSMAN,, A., WAGNER, A. & PLEWIG, G . (1981) Reduction of bacterial skin flora during oral treatment of severe acne
with 13-cis-retinoic acid. Archives of Dermaiological Research, 270, 179.
WILLIAMSON, P. & KLIGMAN, A.M. (1965I A new method for the quantitative investigation of cutaneous bacteria.
Journal of Investigative Dermatology, 45, 498.