~ Pergamon

Fd Chem. Tox ic. Vol. 32. No.8, pp . 769-775 . 1994

Copyright © 1994 Elsevier Science Ltd
OZ78-6915-(94)EOO53-P Printed in Great Britain. All right s reserved
0278-6915/94 $7.00 + 0.00

THE IDENTIFICATION AND CLASSIFICATION OF SKIN

IRRITATION HAZARD BY A HUMAN PATCH TEST
D. A. BASKETTER, E. WHllTLE, H. A. GRIFfITHS and M. YORK
Unilever Environmental Safety Laboratory, Sharnbrook, Bedford MK44 1LQ. UK

(Accepted 2 MQFch /994)

Abstract-There exist various regulatory instruments the purpose of which is to ensure that the intrinsic
toxic hazards associated with substances and preparations are identified. In the context of identification
of skin irritation potential , the method is normally the Draize test. Guidance notes provided by the OECO
and the EEC expect that corrosive substances will have been screened out by a variety of methods.
Substances or preparations which cause a sufficient degree of skin irritation will be classified as skin
irritants. The primary motivation behind the present work was to introduce the concept that it is possible
to assess the hazard potential of a substance or preparation to produce skin irritation in a human study.
In the example presented here, 20% sodium lauryl sulfate (SLS) has been chosen as the positive control.
With the protocol currently devised, occluded patch treatment with 20% SLS for up to 4 hr produces an
irritant response in just over half of the panel. An irritant response is taken as a clinically evident and
significant increase in erythema. oedema or dryness-a minimum of a + reaction on the ICORG scale.
At such a level of response with the positive control (both in terms of intensity and in proportion of the
panel), it is then possible to judge and/or to determine statistically, whether the test material has produced
a level of skin irritation which is similar to, greater, or lower than the positive control. In this way a human
patch test protocol can form a fundamental component of a strategy for the replacement of animals in
determination of skin irritation and corrosion potential. By use of a careful and progressive protocol and
by comparison of test data against a positive control it is both possible and practical to classify substances
and preparations in terms of their skin irritation potential using that endpoint in the species of concern,
man.

INTRODUCTlON physicochemical tests, structure-activity consider-

Within the EEC there exist various regulatory instru-
ments (e.g. the Dangerous Preparations Directive;
EEC, 1983) the purpose of which is to ensure that, by
suitable means, the intrinsic toxic hazards associated
with substances and preparations are identified. On
the basis of this identification process, suitable warn-
ing labels (risk phrases) may be required (EEC, 1991).
For the assessment of skin irritation potential, the
method to be followed is based on the Draize test
(Draize, 1944 and 1959) which uses 4 hr semi-
occluded patches of test material applied to the
clipped/shaved skin of a minimum of three rabbits.
The test is described in Annex V (EEC, 1992)and also
in the recently updated OECD Test Guideline 404
(OECD , 1993). Substances or preparations which
cause, on average, a defined degree of skin irritation
will be classified as 'Irritant' and will carry the risk
phrase R38 "May cause irritation by skin contact".
If the response is severe, the substance may fall into
one of two categories of corrosivity. However, the
guidance notes now pro vided by both the OECD and
the EEC expect that corrosive substances will have
been screened out by a variety of methods including
Abbreviations: ICDRG = International Contact Dermatitis
Research Group; SLS = sodium lauryl sulfate.
ations, in vitro assays and the use of a sentinel rabbit
in the Draize test. One of the most reliable in vitro
assays of skin corrosion potential is the in vitro skin
corrosivity test (Oliver et al., 1986 and 1988) which
has been the subject of a successful interlaboratory
validation (Botham et al., 1992).
However, this approach to the identification of
skin corrosion and irritation potential suffers from
two drawbacks. In the first place, the use of the
Draize rabbit test is still obligatory to determine
whether classification as a skin irritant is required .
Secondly, the rabbit is used as a surrogate for
humans, so interspecies differencesare not taken into
account . To address these problems, a hierarchical
strategy not involving rabbits, but rather the use of
human skin in vitro and in vivo has been proposed
(Basketter, 1994; Basketter et al., 1994). As a pre-
liminary screen, the adaptation of the in vitro skin
corrosivity test for human skin rather than rat skin
has been demonstrated successfully (Whittle and
Basketter, 1993a). Furthermore, it was shown that
certain interspecies differences may exist in the
identification of corrosive substances (Whittle and
Basketter, 1993b). Such a conclusion is not new and
has been made in the context of skin irritation/
corrosion many years ago (Nixon et al., (975).

769
770 D. A. BASKETTER et al.

In the present paper, the use of human skin 'in vivo'
for the identification and classification of skin irri-
tation potential is explored using one material
classified as irritant and one material classified as
corrosive. A prototype human assay is described. The
data generated support the viability of the approach
and also confirm that the intraspecies differences
highlighted by the in vitro human skin corrosivity test
are not artefactual.
MATERIALS AND MEllfODS
All the work described was conducted according to
Good Laboratory Practice.
Volunteers. 24 volunteers (11 female, 13 male; age
range 21 to 49 yr), all of whom gave fully informed
written consent, took part. As the experiment pro-
ceeded, the number of active participants was re-
duced as the occurrence of irritant reactions to the
specific test materials precluded the need for further
treatment. All volunteers were healthy and free of
skin disease. Ethical approval for the study had been
obtained from the Colworth Ethical Committee.
Chemicals. Sodium lauryl sulfate (SLS; > 98%
pure) was obtained from BDH, (Poole, Dorset, UK).
A 20% (w/v) solution of SLS was prepared using
distilled deionized water. Commercial CS/C IO fatty
acid mix [55/45 (w/w); > 98% pure] was supplied by
Unichema, (Gouda, The Netherlands) and used un-
diluted. A volume of 0.1 ml was pipetted onto the
patch.
Human patch test method. Patches composed of
I em? Webril undercast padding, regular finish (code
3175 from Kendall Co., Boston, MA, USA) were
occluded and held in place by Leukosilk surgical tape
(BDF, Beiersdorf, Hamburg, Germany). They were
applied to the outer upper arm for 30 min (first five
panellists only), I, 2, 3 or 4 hr.. Treatment sites were
assessed for the presence of erythema, oedema and
dryness on a four-point scale and for the presence of
other clinical signs (e.g. scaling or glistening), at 24,
48 and 72 hr. The overall response could then be
placed into categories which were based on the
International Contact Dermatitis Research Group
(ICDRG) scale (Fregert, 1981) (Table I).
A score of ' +' or greater for erythema, dryness or
oedema at either 24, 48 or 72 hr in an individual
subject was taken as evidence of a positive irritant
response to the test substance. Panellists who demon-
strated this level of response with any test material
were not exposed further to that substance. The
testing schedule is shown in Table I.
It is assumed that panellists who experience a
reaction of' +' or greater at application times of less
than 4 hr would present an irritant response following
a 4 hr application. It would be unethical to continue
treatment in such subjects. In evaluating the results,
the critical point is to assess the number of panellists
who either had, or would have had, a positive irritant
response after a 4 hr patch treatment.
RESULTS
Although the protocol allowed for the phased
introduction of panellists and progressive increases in
patch application time, since no significant adverse
effects were seen and the study proceeded to com-
pletion the results for all panellists have been com-
bined in Table 2 (with the exception of the completely
negative readings at 30 min). This Table illustrates
the results in terms of the number of panellists
demonstrating specific levels of erythema at three
assessment times. The Table also indicates the num-
ber of panellists who demonstrated reactions that
were sufficiently high (+ or greater) to preclude
further testing with that substance and the numbers
of panellists absent from each assessment. The
threshold for a response to be judged as irritant was
a '+ ' reaction using a scale based on the ICDRG
scale (Fregert, 1981).No irritation reactions exceeded
a grade' +' response. Plate I shows a typical example

Table I. Treatment schedule and assessment of response for patch tests for skin irritation
in humans
Test schedule
Day I Treat I panellist (pilot) for 30 min
Day 3 Treat pilot subject for I hr and 4 further panellists for 30 min
Day 8 Treat pilot subject for 2 hr and the 4 panellists for I hr
Day 9 Treat pilot subject for 3 hr
Day 10 Treat 4 panellists for 2 hr and main group of panellists for I hr
Day 15 Treat pilot subject for 4 hr
Day 16 Treat 4 panellists for 2 hr
Day 17 Treat main group of panellists for 2 hr
Day 22 Treat 4 panellists for 4 hr
Day 23 Treat main group of panellists for 3 hr
Day 29 Treat main group of panellists for 4 hr
Assessment of respease
o No reaction
+ Weakly positive reaction (usually characterized by mild erythema across
most of the treatment site)
++ Moderately positive reaction (usually distinct erythema possibly spreading
beyond the treatment site)
+++ Strongly positive reaction (strong spreading erythema with oedema)
A score of + or greater at 24. 48 or 72 hr in an individual subject was taken as evidence of
a positive irritant response to the test substance. Panellists who demonstrated this level
of response with any test material were not exposed further to that substance.
Plate I. Typical erythematous response to 20% SLS. This is the level of irritation which would be taken
as sufficient to regard the subject as having experienced a positive reaction, i.e. +.

771
 Human patch test for skin irritation hazard 773

Table 2. Summary of individual erythema scores

l-hr exposure'[ 2-hr exposure 3-hr exposure 4-hr exposure
Grade
Test substance Scoring time (hr)t... 24 48 72 24 48 72 24 48 72 24 48 72
Distilled 0 22 21 4 20 20 16 19 18 14 14 16 14
water + 0 0 0 0 0 0 0 0 0 0 0 0
A 2 3 20 2 2 6 I 2 6 5 3 5
Not treated 0 2 4 5
20% SLS 0 20 20 4 15 17 15 16 16 12 9 9 8
+ 2 I 0 5 3 I 0 0 1 3 4 3
A 2 3 20 2 2 6 I I 4 4 3 5
Not treated 0 2 7 8
C,.IO fatty 0 19 19 4 16 17 14 16 15 12 10 12 II
acid + 3 2 0 3 2 1 0 0 0 I 1 I
A 2 3 20 2 2 6 I 2 5 5 3 4
Not treated 0 3 7 8
SLS = sodium lauryl sulfate
·Grades: 0 = no irritation reaction; + = weak erythematous response; A = subjects absent at that evaluation.
tScoring time after initiation of patch application.
tDuration of patch exposure.
§Not treated: no. of subjects not treated with that exposure duration to the particular substance. Subjects who were no longer treated with
SLS or fatly acid mix were also not treated with distilled water.

of a weak (+) erythematous reaction to 20% SLS
after 3 hr of occluded application. As expected, dis-
tilled water failed to cause any erythematous re-
sponses. Following the 1 hr exposure, 20 out of 22
panellists (two were absent) demonstrated no reac-
tion to the 20% SLS, but two demonstrated a weak
erythematous response. These two individuals were
not treated further with SLS. Of the remaining
panellists tested for 2 hr with 20% SLS, 16 showed no
effect, but five demonstrated a weak response and
were not treated further. The 3-hr exposure resulted
in one further panellist reacting and being withdrawn
and at 4-hr exposure a further four panellists had a
.+' erythematous reaction. Thus, ultimately, 12 of
the 24 subjects demonstrated an erythematous reac-
tion following exposure to 20% SLS for up to 4 hr.
The level of erythematous reaction to the CS/IO fatty
acid mix was similar to that to SLS. Three panellists
had a reaction after I hr of exposure and were not
treated further. Four more had an erythematous
reaction after 2 hr of exposure. Ultimately, nine
panellists experienced an erythematous irritant reac-
tion to the fatty acid mix.
Responses other than erythema were infrequent.
Those that were of sufficient significance to merit a
.+' score have been added to those who achieved a
positive erythema score and are recorded in Table 3
which provides an overall summary. Essentially, one
panellist had an odd 'glazing' of the skin in response
to 20% SLS, whilst one other had a '+' grade of
dryness in response to the fatty acid mix. Thus,
overall 54% (13/24) of the panel reacted to 20% SLS,
42% (10/24) to fatty acid and none to distilled water.
The response to SLS was in accord with experience
with an earlier panel, where nine out of 17 individuals
(53%) reacted to occluded application ofSLS for up
to 4 hr (D. A. Basketter and H. A. Griffiths, 1993).
The slightly lower response to the fatty acid mix was
not statistically significant (Fisher's exact test).
All reactions resolved in a few days with none
exceeding a moderate level of irritation (+ +), even
beyond the formal scoring time. On patch removal,
it was noted that a small number of the panellists had
an erythematous response to the fatty acid mix which
was short lived. These were considered to be weak
non-immune immediate contact reactions (contact
urticaria). Panellists made no significant comments
on subjective responses (e.g. stinging, itching).
DISCUSSION
The primary motivation behind the present work
was to introduce the concept that it is possible to
assess the hazard potential of a substance or prep-
aration to produce skin irritation in a human study.
At present the only accepted route in regulatory
toxicology is to carry out the Draize rabbit skin test
(Draize, 1959). This method uses semi-occluded
patches of undiluted test material applied for 4 hr to
the shaved dorsal skin of three rabbits (OECD, 1993).
Notwithstanding the objections to the use of animals,
it should be readily apparent that skin irritation is an
endpoint which can be reproduced and assessed in a
statistically significant number of members of the
species of concern, humans. Our proposed strategy is
to eliminate corrosive substances at an early stage

Table 3. Summary of distribution of positive reactions

No. of subjects with reaction/total no. of subjects

Test substance Erythema Dryness Other Total

Distilled water 0/24 0/24 0/24 0/24 (0%)
20% SLS 12/24 0/24 1/24 13/24 (54%)
C'1I0 fatty acid 9/24 1/24 0/24 10/24 (42%)
SLS = sodium lauryl sulfate

FCT 32/8-F
774 D. A. BASKETTER
(Basketter, 1994; Whittle and Basketter, 1994). Sub-
sequently, materials may then be evaluated in an
ethically approved, cautiously progressive human
study protocol. The central tenet of this protocol is
that substances will be compared, side by side, with
a meaningful positive control. This approach is de-
signed to assess the presence or absence of hazard. It
is different in design and intent to those human
studies which are performed to assess either efficacy
or risk (e.g. Jenkins and Adams, 1990).
In the example presented here, 20% SLS has been
chosen as the positive control. There are several
facets to this. First, SLS is widely recognized as
irritant to skin and it is classified as 'Irritant' (EEC,
1983). Secondly, while it is not reported as a cause of
substantial or excessive skin reactions in most indi-
viduals following normal use of products containing
SLS, it is recognized that excessive use may result in
skin irritation. The EEC Dangerous Preparations
Directive (EEC, .1988) requires that preparations
containing 20% or more of SLS must themselves be
classified as a skin irritant. Therefore to ensure that
the human patch test protocol being developed was
of sufficientsensitivity to identify substances or prep-
arations that would be on the borderline for classifi-
cation as irritant, 20% aqueous SLS was adopted as
a positive control.
With the protocol currently devised, 20% SLS
produces an irritant response in about half of the
panel. The ultimate response rate of 54% in the
example study shown is fairly typical, but may change
with patch type (e.g. Hilltop chambers give an ap-
proximately 80% response rate). Occlusion has been
shown to be necessary to produce a significant level
of reaction to this level of SLS (Dillarstone and Paye,
1993). An irritant response is taken as a clinically
evident and significant increase in erythema, oedema
or dryness-a minimum of a '+' reaction on the
ICDRG scale identified at either the 24, 48 or 72 hr
assessment times. This relatively modest intensity of
reaction (see Plate I) is also a signal for cessation of
treatment with the causative substance in that panel-
list. At such a level of response with the positive
control (both in terms of intensity and in proportion
of the panel), it is then possible to judge and/or to
determine statistically whether the test material has
produced a level of skin irritation which is similar to,
greater, or lower than the positive control.
The observation of transient erythematous reac-
tions immediately after patch removal is a potential
complicating factor. Their disappearance at the 24 hr
assessment is indicative of an urticarial response
which should not be interpreted as irritation. As the
primary purpose of the study is to assess irritation
potential, then the response at 24, 48 and 72 hr must
be used to decide whether further testing is appropri-
ate. Nevertheless, subjects exhibiting possible urticar-
ial reactions should not be re-patched for a longer
duration of exposure until the nature of the response
has been resolved by assessment at 24 hr.
et al.
In the present example, it is quite evident that the
undiluted CS/CIO fatty acid mix is similar in irritation
potential to 20% SLS. Furthermore, undiluted
nonanoic acid (C9 fatty acid) produced no more than
irritation when applied under occlusion to human
skin for 48 hr (Wahlberg and Maibach, 1980).Conse-
quently it is worth noting that this fatty acid mix is
formally classified as corrosive (R34) on the basis of
a rabbit Draize test. However, it was not corrosive in
the in vitro corrosivity test using human skin (Whittle
and Basketter, 1993b). Thus it seems that the rabbit
skin response to irritation/corrosion may yield mis-
leading results. It cannot be argued that the rabbit
result provided a 'safety margin' because of the close
similarity of the results for the fatty acid and for 20%
SLS-unless the latter is now to be regarded as
corrosive, which is clearly a nonsense. Thus it would
seem that the correct classification for this fatty acid
blend should be 'Irritant' with the R38 risk phrase.
For ethical reasons, the protocol described is cau-
tious; some irritants may produce a relatively sharp
increase in irritant response with only a short increase
in exposure. However, it is possible to evaluate more
than one test substance in a single panel and for
preparations for which the irritant potential is 'pre-
dictable', the protocol may be reduced accordingly.
The future strategy in this laboratory is to refine and
standardize a human patch protocol that is suitable
for the production of data enabling the classification
or non-classification of substances and/or prep-
arations as skin irritants. Coupled with an appropri-
ate in vitro assay(s), such a route could avoid any use
of animals in the assessment of skin corrosion and
irritation (Basketter, 1994; Basketter et al., 1994). To
this end, several patch test systems and positive
control substances continue to be evaluated. Early
results from this work will be published shortly
(Dykes et al., 1994).
In conclusion, we have presented a human patch
test protocol which forms a fundamental component
of a strategy for the replacement of animals in the
determination of skin irritation and corrosion poten-
tial. By the use of a careful and progressive protocol
and by comparison of test data against a positive
control, we believe it is both possible and practical to
classify substances and preparations in terms of their
skin irritation potential using that very same end-
point in the species of concern, man.
Acknowledgements-The authors would like to express their
appreciation to Andrea Dickens for statistical advice, to
Andrew Shaw for the photography and to George Holland,
Chris Newsome and Mark Chamberlain for valuable discus-
sions.
REFERENCES
Basketter D. A. (1994) A strategic hierarchical approach to
acute toxicity testing. Toxicology in Vitro, 8, 855-859.
Basketter D. A., Whittle E. and Chamberlain M. (1994)
Identification of irritation and corrosion hazards to skin:
 Human patch test for skin irritation hazard
an alternative strategy to animal testing. Foodand Chemi-
cal Toxicology, 32, 539-542.
Botham P. A., Hall T . J., Dennett R., McCall J. C,
Basketter D. A., Whittle E., Cheeseman M., Esdaile D. J.
and Gardner J. (1992) The in vitro skin corrosivity test.
Results of an interlaboratory trial. Toxicology in vitro, 6,
191-194.
Dillarstone A. and Paye M. (1993) Antagonism in concen-
trated surfactant systems . Contact Dermatitis 28, 198.
Draize J. H. (1959). Dermal toxicity. In Appraisal of the
Safety of Chemicals in Foods, Drugs and Cosmetics,
pp. 46--59. Association of Food and Drug Officials of the
United States , Littleton, CO, USA.
Draize J. H., Woodard G . and Calvary H. O. (1944)
Methods for the study of irritation and toxicity of sub-
stances applied topically to the skin and mucous mem-
branes. Journal of Pharmacology and Experimental
Therapy 82, 377-390.
Dykes P. J., Black D. R., York M., Dickens A. and Marks
R. (1994) A stepwise procedure for evaluating irritant or
corrosive materials in normal volunteer subjects. Human
and Experimental Toxicology . In press .
EEC (1983) Commission Directive of29 July 1983 adapting
to technical progress for the fifth time Council Directive
67/648/EEC on the approximation of the laws, regu-
lations and administrative provisions relating to the
classification , packaging and labelling of dangerous sub-
stances . Official Journal of the European Communities
L2S7, I.
EEC (1988) Council Directive of 7 June 1988 on the
approximation of the laws, regulations and administrative
provisions of the Member States relating to the classifi-
cation, packaging and labelling of dangerous prep-
arations. Official Journal of the European Communities
Ll8,14.
EEC (1991) Annex I to Commission Directive 91/325/EEC
of 1st March 1991 adapting to technical progress for the
775
twelfth time Council Directive 67/548/EEC on the ap-
proximation of the laws, regulations and administrative
provision relating to the classification, packaging and
labelling of dangerous substances. Official Journal of the
European Communities LlSO, 8 July 1991, 34.
EEC (1992) Annex to Directive 92/69/EEC. Part B, Method
B.6. Acute Toxicity (Skin Sensitization). Official Journal
of the European Communities L383A, 131-136.
Fregert S. (1981) Manual of Contact Dermatitis, Munks-
gaard, Copenhagen.
Jenkins H. L. and Adams M. G . (1990) Progressive evalu-
ation of skin irritancy of cosmetics using human volun -
teers. International Journal of Cosmetic Science 11.
141-149.
Nixon G. A., Tyson C. A. and Wertz W. C. (1975)
Interspecies comparisons ofskin irritancy. Toxicology and
Applied Pharmacology 31, 481-490.
OECD (1993) Updated Test Guidelines 404, adopted July
12th, 1992. Organisation for Economic Cooperation and
Development, Paris.
Oliver G . J. A., Pemberton M. A. and Rhodes C. (1986) An
in vitro skin corrosivity test-modifications and vali-
dation. Food and Chemical Toxicology 24, 507-512.
Oliver G . J. A., Pemberton M. A. and Rhodes C. (1988) An
in vivo model for identifying skin-corrosive chemicals. I.
Initial validation. Toxicology in Vitro 2, 7-17 .
Wahlberg J . E. and Maibach H. l. (1980) Nonanoic acid
irritation-A positive control at routine patch testing.
Contact Dermatitis 6, 128-130.
Whittle E. and Basketter D. A. (1993a) The in vitro corro-
sivity test: development of the method using human skin .
Toxicology in Vitro 7, 265-268.
Whittle E. and Basketter D. A. (1993b) The in vitro corro-
sivity test: comparison of in vitro human skin with in vivo
data. Toxicology in Vitro 7, 269-274.
Whittle E. and Basketter D. A. (1994) In vitro skin corrosiv-
ity test using human skin . Toxicology in Vitro 8,861-863.