&EPA
United States
Environmental Protection
Agency
Office of Pesticides and
Toxic Substances
Washington. D.C. 20460
EPA-560/11-82-001
October 1981
Pesticides and Toxic Substances
EYE
IRRITATION
TESTING
An Assessment of Methods and
Guidelines for Testing Materials
for Eye Irritancy
•
-------
EPA-560/11-82-001
October 1981
EYE IRRITATION TESTING
AN ASSESSMENT OF METHODS AND GUIDELINES
FOR TESTING MATERIALS FOR EYE IRRITANCY
by
K.J. Falahee
C.S. Rose
S.S. Olin
H.E. Seifried
Contract No. 68-01-6176
Project Officer - Charles Auer
Technical Advisors - Norbert P. Page, Daljit Sawhney
Assessment Division
Office of Toxic Substances
Washington, D.C. 20460
OFFICE OF PESTICIDES AND TOXIC SUBSTANCES
U.S. ENVIRONMENTAL PROTECTION AGENCY
WASHINGTON, D.C. 20460
-------
This report was prepared as an account of work sponsored by an
agency of the United States Government. Neither the United States
Government nor any agency thereof, nor any of its employees, contrac-
tors, subcontractors, or their employees, makes any warranty,
expressed or implied, nor assumes any legal liability or responsibility
for the accuracy or completeness of the information contained herein
nor for any third party's use or the results of such use of any informa-
tion, apparatus, product, or process disclosed in this report, nor repre-
sents that its use by such third party would not infringe privately owned
rights.
This report has been reviewed by the Office of Pesticides and Toxic
Substances, U.S. Environmental Protection Agency, and approved for
publication. Approval does not signify that the contents necessarily
reflect the views and policies of the United States Government or the
U.S. Environmental Protection Agency, nor does mention of trade
names or commercial products constitute endorsement or recommen-
dation of use.
in
-------
TABLE OF CONTENTS
Page
PREFACE vii
EXECUTIVE SUMMARY ix
1.0 INTRODUCTION 1
2.0 HISTORICAL PERSPECTIVE 1
3.0 SURVEY OF TEST PROTOCOLS 5
3.1 TEST AGENTS 5
3.1.1 Corneal Penetration 6
3.1.2 pH 7
3.1.3 Results of Skin versus Eye Irritation 10
3.2 TEST AGENT VOLUME 11
33 METHOD OF EXPOSURE 12
33.1 Liquids and Solids 12
33.2 Aerosols 13
33.3 Corneal Applicators 14
3.4 DURATION OF EXPOSURE 14
3.5 SELECTION OF SPECIES 18
3.5.1 Use and Characteristics of Different Species 18
3.5.2 Comparison of Monkey and Rabbit Responses 20
3.5.3 Other Species 22
3.6 TEST GROUP SIZE 25
3.7 USE OF ANESTHETICS 26
3.8 OBSERVATION PERIOD 28
4.0 ASSESSMENT OF THE IRRITATION REACTION 29
4.1 EXAMINATION TECHNIQUES 29
4.2 SUBJECTIVE SCORING 30
4.3 OBJECTIVE METHODS OF ASSESSMENT 34
43.1 Corneal Thickness 34
4.3.2 Intraocular Pressure 36
4.3.3 Corneal and Conjunctival Edema 38
43.4 Capillary Permeability 38
43.5 Histological Investigation 39
43.6 Photographic Observation 40
43.7 Additional Techniques 40
-------
TABLE OF CONTENTS (Continued)
Page
5.0 ADDITIONAL IN VIVO STUDIES 41
5.1 TESTS IN HUMANS 41
5.2 COMPLEMENTARY TESTS 43
6.0 IN VITRO TESTS 46
6.1 ISOLATED EYES AND CORNEAS 47
6.2 TISSUE CULTURE 49
7.0 EXTRAPOLATION OF ANIMAL DATA TO MAN 51
8.0 APPROACHES FOR THE FUTURE 55
9.0 CONCLUSIONS 56
10.0 RECOMMENDATIONS FOR FURTHER RESEARCH 62
11.0 REFERENCES 63
APPENDICES
Appendix A — Scale of Weighted Scores for Grading the Severity
of Ocular Lesions 73
Appendix B — Test for Eye Irritants 75
TABLES
Table 1. Criticisms of the Draize/FHSA Test for Eye Irritation 2
Table 2. Ocular Irritation Test Methods 4
Table 3. Corneal Thickness of Several Species of Animals 19
Table 4. Comparison of Response of the Rabbit and Monkey Eye
to Irritants 23
Table 5. Comparison of Response of Several Species to Eye
Irritants 25
Table 6. Summary of Tissue Culture Tests for Assessing Eye
Irritation 50
Table 7. Comparison of Animal and Human Responses 51
FIGURE
Figure 1. Tier Approach for Eye Irritation Testing 57
vi
-------
PREFACE
Recently, the Environmental Protection Agency (EPA) has been
working with the Interagency Regulatory Liaison Group (IRLG)
within the Federal government and with the international Organisation
for Economic Cooperation and Development (OECD) to formulate a
uniform set of toxicity testing guidelines. As part of this effort, EPA's
Office of Pesticides and Toxic Substances requested Tracer Jitco to
critically review, evaluate, and summarize the available information on
eye irritation testing.
In recent years, there has been increased public attention and
concern regarding the use of animals in scientific research. During the
past year, a number of bills were introduced in the U.S. Congress
related to the use of animals in research, including one to promote the
development of methods of research, experimentation, and testing that
minimize the use of live animals and another to establish a National
Center for Alternative Research. Public and scientific concern in this
area has been reflected in the formation of a number of special interest
groups and in several recent symposia which considered alternative
testing methods and procedures.
In the area of eye irritation testing specifically, much attention has
been directed to the use of the rabbit. The Cosmetic, Toiletry, and
Fragrance Association sponsored a workshop on ocular safety testing
in October 1980 in which alternatives to the Draize rabbit eye test were a
primary topic of discussion. In February 1981, the National Institutes
of Health sponsored a symposium on Trends in Bioassay Methodology
in which particular attention was focused on the feasibility of using in
vitro and mathematical methods to reduce the dependency on the use of
live animals in biological testing. Organizations represented at this
symposium included academic institutions, industry, animal welfare
groups, and government research and regulatory agencies. Progress in
the development of in vitro and other alternative test methods is dis-
cussed in the present report.
In conducting this assessment, Tracor Jitco focused on the guide-
lines recently developed and promulgated by the OECD and the IRLG.
On-line and manual searches of the relevant data bases were conducted
and the significant literature was acquired and reviewed. Input on
current developments was obtained through personal communication
with investigators in the field and staff in several government agencies.
The EPA provided the most recently developed versions of the OECD
guidelines and contributed valuable background information concern-
ing Federal and international efforts to standardize testing guidelines.
The various protocols for eye irritation testing were examined and
issues requiring further research based on these findings were identified.
Vll
-------
The authors would like to thank Van Seabaugh of the Consumer
Product Safety Commission for his advice during the preparation of
this document and for providing the photographs contained in Appen-
dix B of this report. Thanks are extended to Robert Osterberg of the
Food and Drug Administration, James Murphy and Robert Jaeger of
the EPA, and veterinary pathologist Robert Kovatch of Tracor Jitco
for their review and helpful suggestions. The editorial assistance of
Edward Cremmins, Marion Levy, and William Theriault (Tracor Jitco)
is also gratefully acknowledged.
Vlll
-------
EXECUTIVE SUMMARY
A primary objective of toxicity testing guidelines is to ensure that
the tests are scientifically sound, practical and cost effective, reliably
predict effects in humans, and conform to acceptable standards for the
care and use of laboratory animals. This report examines in detail the
various protocols for eye irritation testing with particular reference to
the guidelines published by the Interagency Regulatory Liaison Group
(IRLG) and the Organisation for Economic Cooperation and Develop-
ment (OECD). Characteristics of test agents which can influence the
irritation response were considered along with the effect of test vari-
ables such as volume of test agent, method and duration of exposure,
test group size, and species. Procedures for assessing the irritation
reaction were considered in detail. Some complementary tests that have
been developed to evaluate other responses of the eye to foreign sub-
stances (e.g., stinging) or that correlate with eye irritation also were
reviewed. Finally, a comprehensive review was made of recent progress
in the development of in vitro tests for eye irritation.
The historical antecedent for the eye irritation testing guidelines
recently published by both the OECD and the IRLG is the Draize test,
developed by Draize, Woodard, and Calvery in 1944. From 1964 to
1981, a modified Draize test has been specified by the Federal Hazard-
ous Substances Act as the preferred method for eye irritation testing.
The eye irritation test as specified in the OECD and the IRLG
guidelines utilizes a minimum of three albino rabbits. Substances are
instilled into the conjunctival sac of the rabbit eye at a volume of 0.1
milliliter for liquids or 100 milligrams for solids. Aerosols are tested by a
one-second spray directly into the eye from a distance of approximately
10 centimeters. Reactions are graded according to the severity of lesions
produced in the cornea, iris, and conjunctiva at intervals throughout a
72-hour observation period (or longer when needed to evaluate the
reversibility of a positive response).
In general, the specific experimental procedures recommended by
the OECD and IRLG were found to be adequately supported by the
experimental data evaluated in this review. Comments on the principal
characteristics of the test methods follow.
The OECD and IRLG assumption that substances found to be
severely irritating to the skin will also be severe eye irritants is valid for
the vast majority of substances for which comparative data are avail-
able. Similarly, chemicals of pH 11.5 or greater or pH 2 or less generally
are severe eye irritants, although the relationship at the acid end of the
scale is less well-defined than at the alkaline end. Thus, the OECD
suggestion that substances that are severe skin irritants or fall outside
the pH 2-11.5 range need not be tested for eye irritation is supported.
IX
-------
Initial testing for eye irritation with three animals normally will be
sufficient to identify substances that are non-irritating or maximally
irritating. Testing with additional animals will be necessary to reliably
characterize substances of intermediate degrees of irritancy.
An observation period of up to 72 hours, as specified in the
guidelines, should be adequate for the qualitative identification of eye
irritants. If a fuller characterization of the reversibility of persistent
irritation or eye damage is desired, longer observation periods will be
necessary.
The albino rabbit has been by far the species most commonly used
in eye irritation testing since Draize and coworkers proposed their
methods, and it remains the species of choice today. While the size and
physical characteristics of the rabbit eye do simplify the test procedures,
other factors are also important in the current preference for the rabbit.
Where comparisons are possible, the rabbit eye appears to be at least as
sensitive, and frequently more sensitive, to irritants than the human eye;
thus, test results permit a conservative extrapolation in human risk
assessment. In addition, evaluation of the significance of an irritant
response can be placed on a firmer comparative basis with the rabbit,
since eye irritation test scores are most meaningful when compared for a
series of substances, and most test results available for comparison are
from studies in rabbits.
The monkey eye is less sensitive than the rabbit eye to most
substances and apparently is a somewhat more accurate predictor of
human eye irritation response. The use of the rat for eye irritation
testing has not been adequately studied. The OECD and IRLG guide-
lines both recognize the value of testing in species other than the rabbit
for comparative purposes. However, limited information as well as
practical considerations (particularly with monkeys) will continue to
restrict this additional testing to those products for which a more
in-depth evaluation is desired.
The direct instillation of 0.1 milliliter or 100 milligrams of the test
substance into the conjunctival sac of the rabbit eye should be con-
tinued as the standard procedure, as recommended by OECD and
IRLG, to allow correlation of results with the historical data base.
However, the recent literature suggests that additional tests with
smaller volumes or more dilute solutions can be especially valuable in
ranking substances of similar irritancy and in correlating irritation
responses in the test species with those resulting from typical accidental
human exposures. Direct spraying of pressurized aerosols into the eye,
while not readily comparable with the standard instillation procedure,
has become the accepted method for tests of this type of product based
on relevancy to human exposure.
Separate tests to determine the effect of flushing the exposed eye
with water at a fixed interval after instillation of the test substance
-------
should remain optional, as provided for in the guidelines. Washing the
eyes following exposure is generally beneficial in reducing the response
to an irritant. The effects of this procedure are evaluated in this report.
Accuracy and reliability in the scoring of lesions have long been
major goals in eye irritation testing and in the development of testing
guidelines. The modified Draize scoring system is still considered ade-
quate for grading irritant responses and the accuracy of scoring can be
further enhanced by using a separate reference set of photographs of
graded eye responses (see Appendix B). Interpretation of eye irritation
scores for purposes of risk assessment requires a full evaluation of the
data and their significance in potential human exposure situations.
Objective methods of measuring eye irritation can improve the
uniformity and precision of the assessment process and enable the
evaluation of more subtle effects and changes in the eye. Techniques
that have been investigated include measurements of corneal thickness,
intraocular pressure, capillary permeability, and corneal and conjuncti-
val edema along with histological examination, photography, and spec-
ular and electron microscopy. These techniques should remain as
optional means of supplementing conventional assessments.
Considerable efforts are being directed towards the improvement
of the current animal test as well as the development of alternative
methods. In vitro tests are being developed in an effort to create a rapid
screening method and to reduce the use of animals in eye irritation
testing. Tissue culture tests using corneal epithelial cell lines or other
more common cell lines appear potentially beneficial as screening tech-
niques and possible adjuncts or alternatives to the in vivo rabbit eye test.
In vitro tests are limited, however, by their inability to demonstrate the
whole animal response.
Several issues were identified from this review that could benefit
from additional experimental research and development. These
include:
• The continued development of in vitro tests
• Validation of an acceptable topical ocular anesthetic for use in the
Draize test
• Better definition of the rabbit eye response to acids of pH less than
2
• Further investigation of test results using smaller volumes (or
dilutions)
• Comparison of techniques for administering the test agent
• More extensive evaluation of other species (particularly rodents)
for routine testing purposes.
A tiered approach to eye irritation testing is envisioned in which
substances would be initially screened for pH, dermal irritation, and
XI
-------
activity in in vitro eye irritation tests (when adequately validated). The
need for testing in intact animals could thus be limited to substances of
lower irritation potential or substances requiring a more complete
characterization of the response. This approach should improve the eye
irritation testing process in terms of cost-effectiveness and use of animal
resources.
In summary, it is concluded from this review and evaluation of eye
irritation testing methods that the 1981 OECD and IRLG guidelines are
consistent with current practices and are adequately supported by the
published work on this subject. The guidelines are by nature flexible
and require the thoughtful and considered scientific input of the investi-
gator in their application, interpretation, and evaluation. At the same
time, there is sufficient standardization in the basic procedures to
ensure general acceptability of the test results. The continuing develop-
ment and validation of improved and alternative test methods is
encouraged.
Xll
-------
1.0 INTRODUCTION
Animal tests for assessing eye irritation and injury resulting from
chemical exposure are designed to help predict the potential consequen-
ces of human ocular exposure to foreign materials. Data from such tests
have been used to assess the degree of risk from numerous substances
that present a potential exposure hazard during production, handling,
distribution, and use. In recent years, significant improvements have
been realized in the reliability, predictability, and reproducibility of eye
irritation testing methods. Several reviews have been published during
the past decade (Marzulli and Simon, 1971; Buehler, 1974; McDonald
and Shadduck, 1977; Ballantyne and Swanston, 1977).
The Environmental Protection Agency (EPA) has been working in
conjunction with the Interagency Regulatory Liaison Group (IRLG)
and the international Organisation for Economic Cooperation and
Development (OECD) to develop uniform guidelines for toxicity test-
ing. Guidelines for eye irritation testing have recently been published in
final form by both the IRLG and OECD. The present report, under-
taken at the request of the EPA Office of Toxic Substances, reviews past
and present test methodologies in relation to the guidelines. A number
of aspects were singled out for consideration: regulatory criteria;
methods and techniques; scoring and classification systems; species;
significant modifications in existing methods; new methodology, both
in vivo and in vitro', general predictability, sensitivity and accuracy of
the tests; humane considerations; and comparisons of data from ani-
mals and humans.
The present review has utilized a broad base of information includ-
ing the open literature, public comments submitted in response to
proposed guidelines published by the EPA and the IRLG, and personal
communication with investigators in government, academia, and
industry. The information provides a basis for judging the adequacy
and validity of present methods as well as a benchmark for comparison
and evaluation of new or alternate methods and test systems.
2.0 HISTORICAL PERSPECTIVE
Draize, Woodard, and Calvery (1944) described a test in which
albino rabbits were used to assess the potential of substances to cause
eye irritation. This technique (Draize et al., 1944; Draize, 1959), com-
monly referred to as the Draize test, was derived from a method
originally described by Friedenwald et al. (1944). The Draize test sub-
jectively and quantitatively scores eye irritation by observing corneal,
iridial, and conjunctival damage following instillation of a test material
into the rabbit eye. One-tenth milliliter of the test material is placed in
-------
the conjunctival sac of one eye of the albino rabbit with the other eye
serving as a control. A series of nine rabbits is used for each test. The
treated eyes of three rabbits remain unwashed. The treated eyes of two
other groups of three rabbits each are washed with 20 milliliters of
lukewarm water approximately at body temperature two or four
seconds after instillation of the test material. Ocular readings are made
at 24, 48, and 72 hours and at 4 and 7 days after treatment or as long as
injury persists. While the Draize test remains the basis of current eye
irritation testing methods, many modifications of the original proce-
dure have been developed and employed in the testing of chemical
substances for experimental and regulatory purposes. Some aspects of
the Draize test have also been questioned by various authors (Table 1).
Table 1. Criticisms of the Draize/FHSA Test for Eye Irritation
Humane considerations
Discrepancies in response of the rabbit and human eye: test is performed in
rabbits but results are applied to humans.
Method of exposure to test agent not comparable with means of human acci-
dental exposure in the case of aerosols, powders, and granular substances.
Volume of materials as tested produces exaggerated results in the rabbit eye
compared to findings in man.
Assessment of reaction is influenced by the test group size and the length of
the observation period.
Subjective nature of the scoring system.
Interlaboratory variability in tests with identical materials.
Difficulties in interpretation of test results.
Inability to correlate active inflammatory signs with permanent structural
change.
Rowan (1980; 1981)
Harriton (1981)
Buehler (1974)
Marzulli and Simon (1971)
Beckley (1965b)
Griffith et al. (1980)
Weltman et al. (1965)
Buehler (1974)
Heywood and James (1978)
Russel and Hoch (1962)
Rieger and Battista (1964)
Weil and Scala (1971)
Kay and Calandra (1962)
Ballantyne and Swanston (1977)
Aronson (1975)
In the Federal Hazardous Substances Act(FHSA) (Federal Regis-
ter, 1964) a modified Draize test was adopted as the official method of
the Food and Drug Administration (FDA) for eye irritancy evaluation.
A proposal for the modification of the FHSA method was put forth by
-------
the FDA in 1972 (Federal Register, 1972). This proposal, however, was
never finalized and was eventually withdrawn officially in 1979 (Federal
Register, 1979c).
In 1972, the responsibility for administering the FSH A was trans-
ferred to the Consumer Product Safety Commission (CPSC) upon the
signing into law of the Consumer Product Safety Act (15 U.S.C. 2051;
1972). To assist in the evaluation of eye irritation test results, the CPSC
in 1976 made available the "Illustrated Guide for Grading Eye Irritation
Caused by Hazardous Substances" (CPSC, 1976), which was based on
an earlier guide produced by the FDA (1965). Both guides are currently
out of print. In addition, the FHSA method of ocular irritation testing
has been published by the CPSC (Federal Register, 1979d). (The plates
from the "Illustrated Guide" [CPSC, 1976] are included in Appendix B
of the present report.)
At the request of the CPSC, the National Research Council of the
National Academy of Sciences (NAS) reviewed the FHSA rabbit eye
test and recommended a new protocol in the revision of NAS Publica-
tion 1138, "Principles and Procedures for Evaluating the Toxicity of
Household Substances" (NAS, 1977). The objectives of this document
were to revise and expand an earlier publication bearing the same title
(NAS, 1964) in light of advancements in scientific technique, as well as
to improve the accuracy and consistency of various toxicity testing
methods including ocular irritation testing.
The Environmental Protection Agency proposed testing guidelines
in 1978 for the development of ocular irritation data on pesticides under
the Federal Insecticide, Fungicide, and Rodenticide Act (FIFR A) (Fed-
eral Register, 1978). Similar guidelines were proposed in 1979 for
chemicaj substances and mixtures for which testing is required under
the Toxic Substances Control Act (TSCA) (Federal Register, 1979a).
To resolve differences in regulatory policy among federal agencies
requiring ocular irritation testing, the Interagency Regulatory Liaison
Group proposed and published test guidelines in 1979 (Federal Regis-
ter, 1979b). Public comments in response to the draft guidelines were
considered in developing the final guidelines (IRLG, 1981).
The guidelines which have been developed and published by the
Organisation for Economic Cooperation and Development (OECD,
1981) are similar to the published guidelines of the IRLG (1981). The
IRLG and OECD guidelines will be periodically updated to implement
changes in the "state-of-the-art." As indicated in a recent Federal
Register notice (Federal Register, 198la), the CPSC considers the test
methods developed by the IRLG and OECD to be suitable for obtain-
ing information on ocular irritation. Other test methods may be used
with prior consent of the agency. While final standards have not yet
been adopted by the EPA, the agency will consider proposed protocols
which are consistent with the IRLG or OECD guidelines (Federal
-------
Register, 1981a). On January 27, 1981, the EPA issued the following
statement: ". . . the policy of the United States ... is to pursue
consistency in its test standards with the OECD Test Guidelines. They
will have the same basic requirements so that data developed according
to either EPA or OECD procedures with respect to those requirements
should satisfy EPA needs. "(Federal Register, 198 Ib). The FDA will use
the IRLG guidelines as standard procedures for testing substances
regulated under the Federal Food, Drug, and Cosmetic Act; but data
derived using the OECD guidelines will also be acceptable. The FDA
further recognizes that special tests may be required in some cases
(Federal Register, 198la).
Specific details of the above protocols and guidelines are presented
in Table 2. The various testing criteria are discussed in the following
section.
Table 2. Ocular Irritation Test Methods
Test Species
Age/Weight
Sex
Number Animals/
Group
Test Agent;
Volume and
Method of
Instillation
Liquids
Solids
Aerosols (e)
Irrigation
Schedule
Draize et al.
(1944)
Draize (1959)
Albino rabbit
NSft/
NS
9
0.1 ml; direct instilla-
tion into lower
conjunctiva! sac
NS
NS
At 2 sec (3 animals)
and at 4 sec (3 ani-
mals) following instil-
FHSA (Federal
Register, 1964;
1979(1)
Same
NS
NS
6
Same as Draize
100 mg or 0.1 ml
equivalent when this
volume weighs less
than 100 mg; direct
instillation into lower
conjunctival sac
NS
Eyes may be washed
after 24 hr reading
HAS (1977)
Same /a]
Sexually mature/less
than 2 yr old
Either sex
4 (minimum)
Liquids and solids1 two
or more different doses
within the probable
range of human
exposure, (d/
Manner of application
should reflect probable
route of accidental
exposure.
Short burst at distance
approximating self-
induced eye exposure.
May be conducted with
separate experimental
groups
IRLG (1981)
Same
Young adult/2.0 to
3.0 kg
Either sex
3 (preliminary
test) (cl: 6
Same as Draize
Same as FHSA
1 sec burst sprayed at
approx. 4 inches
Same as FHSA
OECD (1981)
Same
NS
NS
3 (minimum)
Same as Draize
Same as FHSA
1 sec burst sprayed at
10 cm
Same as FHSA; in
addition, for sub-
stances found to be
lation of test agent
irritating: wash at
4 sec (3 animals) and
at 30 sec (3 animals)
-------
Table 2. Ocular Irritation Test Methods (Continued)
Irrigation
Treatment
Examination
Schedule
Draize et al.
(1944)
Draize (1959)
20 ml tap water (body
temp.)
24 hr
48 hr
72 hr
4 days
FHSA (Federal
Register, 1964;
1979d)
Sodium chloride
solution U.S.P. or
equivalent
24 hr
48 hr
72 hr
HAS (1977)
NS
1 day
3 days
7 days
14 days
IRLG(1981)
Same as FHSA
24 hr {II
48 hr
72 hr
OECD (1981)
Wash with water for
5 min using volume
and velocity of flow
which will not cause
injury.
1 hr
24 hr
48 hr
72 hr
7 days
21 days
Use of
Fluorescein
Use of
Anesthetics
Scoring and
Evaluation
NS
NS
Draize et al.
(1944)^
May he applied after
the 24 hr reading
NS
CPSC (1976) (h)
May be used Same as FHSA
NS May be used
Draize et al. CPSC (1 976) (hj
(1944)^;, or a slit
lamp scoring system
Same as FHSA
May be used
CPSC (1976) (hi
(a) Tests should be conducted in monkeys when confirmatory data are required.
(b) Not specified.
fa) If the substance produces corrosion, severe irritation, or no irritation in a preliminary test with 3 animals, no further testing is necessary. Jf
equivocal responses occur, testing in at least 3 additional animals should be performed.
(d) Suggested doses are 0.1 and 0.05 ml for liquids.
(e) Currently no testing guidelines exist for gases or vapors.
fff Eyes may also be examined at 1 hr, 7, 14, and 21 days.
(g) Presented in Appendix A.
(h) Presented in Appendix B.
3.0 SURVEY OF TEST PROTOCOLS
3.1 TEST AGENTS
The physical and chemical properties of a test substance can signif-
icantly influence the type and extent of damage produced in the eye.
Occupational exposure to irritants can be in the form of liquids, solids,
mists, or aerosols, vapors, and gases. For example, eye exposure data
collected over a 12-month period from a chemical manufacturing plant
showed that 52% of reported eye injuries caused by chemicals involved
exposure to vapors and mists, 31% were caused by liquids, and 17% by
solids (McLaughlin, 1946). The irritation produced by a test substance
may be due to a combination of the physical form to which the eye is
-------
exposed (including particle size), and certain chemical and physical
properties such as pH, lipid/water coefficient, and ionization. If the
potential role of such properties in causing specific ocular reactions
were established, a baseline for making testing decisions could be
developed. For certain substances, this could entirely eliminate the need
for testing. The OECD guidelines, for instance, state: "Strongly acidic
or alkaline substances, for example with a demonstrated pH of 2 or less
or 11.5 or greater, need not be tested owing to their probable corrosive
properties." With both the IRLG and OECD procedures, substances
shown to be severe irritants in dermal irritation tests may be assumed to
be eye irritants and need not be tested in the eye. The scientific bases for
these exceptions are reviewed in the following sections.
The increasing number of studies on mechanisms underlying for-
eign compound-induced ocular irritation may yield a molecular
approach to the prescreening of test agents based on a relationship of
the physical/ chemical characteristics of these agents with their specific
biochemical/biophysical effects.
3.1.1 Corneal Penetration
Characteristics such as molecular size, configuration, osmotic
pressure, and lipoidal nature, as well as chemical properties (including
the degree of ionization following in vivo application) influence the
capacity of an irritant to penetrate the tissues of the cornea. Factors
affecting the penetration of substances into the rabbit cornea were
investigated by Swan and White (1942). Rabbit eyes were exposed to
test solutions of dye intermediates for three minutes, then rinsed with
normal saline solution. After excision of the corneas, the test substances
were isolated and coupled with other dye intermediates to form azo dyes
which were measured colorimetrically. The results indicated that 1)
penetration is inversely related to the degree of polarity of the molecule
and 2) penetration rate is not inversely related to molecular size.
Cogan et al. (1944) studied the permeability characteristics of the
beef cornea in vitro. Transfer of substances through the cornea was
found to be primarily dependent on lipid/water solubility. While lipid
solubility was necessary for substances to penetrate the corneal epithe-
lium, water solubility was a requirement for penetration of the corneal
stroma. Substances which penetrated both the epithelial cell layer and
the stroma possessed properties of both lipid and water solubility. An
approximate correlation was found between molecular size and penet-
rability through the corneal stroma, although penetration through the
epithelium-stroma combination could not be correlated with molecular
size. It was also found that charge or absence of charge alone could not
account for differences in permeability.
-------
Similarly, Cogan and Hirsch (1942) found that the penetration
capacities of weak electrolytes through excised beef cornea were
dependent on the lipid/water solubility of the dissociated and undisso-
ciated molecular forms. They concluded that corneal penetration is best
achieved when the dissociation constant is sufficiently low so that the
substance is present in the undissociated state at the hydrogen ion
concentration of the eye, and when the undissociated form of the
molecule possesses lipophilic characteristics.
The in vitro permeabilities to four organic liquids were studied in
human and animal corneas by Marzulli( 1965). Similar rates of penetra-
tion in the human cornea were found for dimethyl sulfoxide and 70%
isopropyl alcohol (mean values: 53.4 and 53.1 micrograms/square cen-
timeter/minute, respectively). Acetic acid (10%) and diethyl cresyl
phosphate penetrated the human cornea at a slower rate (mean values:
1.6 and 3.0 micrograms/square centimeter/minute, respectively). Cor-
neal penetration rates in the monkey, rabbit, dog, and swine, compared
to man were found to vary with the agent tested, precluding an interspe-
cies rank of corneal permeability.
In summary, 1) substances which possess maximum lipid and
water solubilities and are poorly ionized have the greatest capacity to
penetrate the cornea, and 2) the relative rates of corneal permeability
among different species are a function of the specific agent tested.
3.1.2 pH
McLaughlin (1946), in a report of 500 cases of chemical eye burns
resulting from accidental exposure, found that the most serious injuries
were caused by highly acidic or basic materials.
Friedenwald et al. (1944) investigated the acid/base effect of a
buffer solution in the rabbit cornea. Rabbit eyes were exposed to a
solution containing buffer (sodium citrate, potassium acid phosphate,
and boric acid, each at 0.028 molar, and sodium chloride at 0.038
molar) and hydrochloric acid or sodium hydroxide with a final pH
ranging from 1 to 13. Solutions were administered by allowing one drop
to fall directly onto the cornea every 2 seconds for a 10-minute period.
Ocular lesions were graded on a numerical scale of 0-40 and were
reported as the percentage of a maximum reaction (100%). The results
showed a sharp rise in the percent reaction when the pH of the solution
was greater than 11.5. A 100% reaction was observed with a pH of
approximately 13. At the acid extreme, the increase in percent reaction
per decrease in pH was less pronounced. The reaction increased from
approximately 3% to 33% as the pH of the solution decreased from 3 to
1.
The above results in corneas with intact epithelium were compared
to the effects produced on the corneal stroma following intracorneal
injection of the test solution or exposure of the cornea to the solution
-------
after mechanical removal of the epithelium. At the alkaline extreme, no
differences were found between effects on the intact cornea irrigated
with test solution and effects on the corneal stroma. This finding is in
keeping with the observed effects of swelling and desquamation of the
corneal epithelium following irrigation with solutions of sodium
hydroxide. A marked protective influence of the epithelial cell layer,
however, was evident in tests at low pH. The percent reaction in the
corneal stroma increased from approximately 5 to 70 as the pH of the
solution decreased from 5 to 1.
Carpenter and Smyth (1946) determined the pH of the surface and
aqueous humor of rabbit eyes exposed to 1.0% sodium hydroxide or a
5.0% aqueous solution of ethylene diamine. The treated eyes were
irrigated with physiological saline for 30 minutes prior to the measure-
ment of pH. With sodium hydroxide, the pH measurement of the
aqueous humor of the treated eye was 8.5 whereas a value of 8.2 was
found for the normal eye. The surface pH determinations of the treated
and normal eyes were 8.0 and 7.2, respectively. With the less highly
ionized alkali, ethylene diamine, values were: surface pH, 8.1; pH of
aqueous humor, 8.7. The increased alkalinity of the surface and aque-
ous humor of the treated eyes indicates that alkaline materials can bind
to and penetrate the rabbit cornea. Damage to internal ocular struc-
tures may be incurred where the buffering capacity of the aqueous
humor is exceeded by the alkalinity of materials contacting the eye.
In contrast to the above determination of pH 8.2 for aqueous
humor of the rabbit eye, Best and Taylor (1943) report a value of 7.1 to
7.3 for the corresponding human tissue. If reliable, these data present a
10-fold decrease in hydroxyl ion concentration which may have an
important bearing on the respective responses of the human and rabbit
eye to chemicals with pH-dependent ionization constants and, there-
fore, on the interpretability of rabbit eye data when these agents are
tested.
Carpenter and Smyth (1946) suggested that alkalies are initially
adsorbed to the cornea and subsequently penetrate into previously
unaffected tissue resulting in a progressive burn. Severe lesions fre-
quently develop as late complications although the injury may initially
appear mild (Hughes, 1946). The degree of severity depends more on the
pH of the solution and the duration of exposure than on the nature of
the cation. In contrast to the effects of exposure to alkalies, no evidence
for a similar progressive action of acid has been observed (McLaughlin,
1946). The early assessment of an acid burn is a measure of the long-
term damage to be expected (Potts and Gonasun, 1975). Friedenwald et
al. (1946) showed that the severity of an acid burn is dependent on both
the pH of the solution and the protein affinity and protein precipitating
ability of the anion. The severity of reaction produced by several acids
of varying pH adjusted with sodium hydroxide was studied in the rabbit
-------
cornea following intracorneal injection. Ocular lesions were expressed
on a scale of 0 to 100. With trichloroacetate solutions, as the pH
decreased from 3 to 2, the reaction grade increased from 0 to 40. With
chlorine solutions, decreasing the pH from 3 to 1 produced increases in
the severity of reaction from grade 0 to 86. Decreases in pH from 5 to 1
with a buffer solution (citrate-phosphate-borate) produced increases in
reaction from grade 0 to 74. With solutions of metaphosphate and
sulfosalicylate, the reaction grade increased from 0 to 48 as the pH
decreased from 7 to 1. Picrate, tungstate, and tannate solutions each
produced somewhat uniform reactions from pH 9 to 1. Irrigation of the
intact corneal epithelium with the same solutions produced reactions of
much less severity.
Similar to the findings of Friedenwald et al. (1944, 1946), an
investigation by Krueger (1959) showed that the corneal epithelium was
a principal barrier to the penetration of acids. The penetrabilities of
acids and alkalies were determined by measuring the pH of the anterior
chamber by an implanted electrode following topical application to the
cornea in isolated pig eyes. In eyes with intact epithelium, alkalies
penetrated to the anterior chamber at a much faster rate than did acids.
When the epithelium was removed, the penetration rates of acids and
alkalies were approximately the same.
According to Grant (1974), splashes of concentrated strong acids,
such as sulfuric, hydrochloric, nitric, phosphoric, and chromic acids,
and liquid sulfur dioxide, can be as injurious to the eye as splashes of
strong alkalies. He further states, however, "In tests on normal rabbit
eyes, acids are found significantly injurious only when the pH is as low
as 2.5 or lower. Applied to human eyes, solutions from pH 7 down to
pH 2 induce an increasingly strong stinging sensation, but on brief
contact cause no damage."
Guillot et al. (1981) studied the ocular irritation produced by 56
test substances, representing various chemical classes, according to the
methods published by the Journal Officiel Francaise and the OECD,
and the method proposed by the Association Francaise de Normalisa-
tion (AFNOR). Specific tests were conducted to validate the provision
(OECD, AFNOR) that excludes the testing of substances with pH 2 or
less. Measurements of pH were performed on the test agents; the pH of
solid materials was measured from a saturated solution of the pulver-
ized substance in distilled water. Four highly acidic substances (one
liquid and three solids) received the highest test rating: extremely
irritating. These were dimethylsulfate, pH 1.0; copper nitrate, pH 1.2;
aluminum nitrate, pH 0.8; and oxalic acid, pH 1.0. One solid product
with a pH of 2.1 (2,4-dichlorochlorosulfonyl-5-benzoic acid) was also
extremely irritating while three solids with pH values ranging from 2.5
to 2.7 varied in irritancy. The OECD and AFNOR methods also specify
that no test is required for substances with pH 11.5 or greater. Although
-------
no substances at this alkaline extreme were included in the study,
Guillot et al. reported that two substances with a pH of 10.7 and one
with a pH of 10.8 were very irritating, and one substance with a pH of
10.5 was extremely irritating.
The effect of pH on sensory irritation (stinging) has also been
studied. Laden (1973) attempted to relate the stinging potential of acidic
materials on human and animal skin to specific properties of the test
agents. Stinging potential in the rat was measured as the length of time
the abraded tail of the animal could be maintained in a test formulation
before being flicked out. Effects in humans were evaluated by applying
a small quantity of the test formulation to forearm skin which had been
abraded by repeated stripping with tape. No general correlations were
found between the stinging potential of solutions of acidic materials and
the hydrogen ion concentration, the tonicity, or the nature of the anion.
The sensory irritation of the eye to a specific solution containing
primary, secondary, and tertiary sodium phosphates (iso-osmotic with
the lacrimal fluid) at varying pH was studied in humans by Trolle-
Lassen (1958). Statistical analysis of the results showed no irritation
effects from the solution at pH values ranging from 7.4 to 9.6. The
incidence of irritation increased from 1.0% to 99% as the pH decreased
from 7.3 to 5.9 and as the pH increased from 9.7 to 11.4.
The above studies with alkali support the provision in the OECD
guidelines that precludes the need to test substances of pH 11.5 or
greater. Similar support is not unequivocally apparent for the OECD
proposed cut-off of pH 2 or less, although the evidence suggests that this
is at least in the right range.
3.1.3 Results of Skin versus Eye Irritation
Comparison of skin and eye irritation data indicates that substan-
ces which produce irritant effects on skin generally are also found to be
irritating to the eye. In most cases the severity of irritation produced in
the eye is similar to or greater than that observed in skin. For example,
of 195 chemicals given relative hazard ratings for both ocular and
dermal (no patch) irritation in rabbits, 45% were rated as more hazard-
ous to the eye than to the skin, 44% were given the same numerical
rating for the severity of both types of irritation, and 11% were rated as
more hazardous to the skin than to the eye (Sunshine, 1969). Three of
the 22 compounds comprising the latter group (bromomethane and two
mixtures of chlorinated biphenyls) showed the greatest deviation from
the normal pattern in that the relative hazard grades indicated markedly
greater irritation to the skin than to the eye under the conditions of the
tests. In every case, however, a compound found to be irritating to the
skin was also observed to produce a positive response in eye tests.
10
-------
Similarly, in tests of 6 shampoos, Jacobi and Ortmann (1971)
found results of irritancy in the rabbit eye to parallel results for the
closed patch test in rabbit skin.
Included in a table of range finding data for over 300 compounds
are the scores for both primary skin irritation on the uncovered rabbit
belly and rabbit eye injury (Smyth et al., 1962). While several com-
pounds were found to be irritating to the eye but not the skin, the
majority produced both types of irritation. A small percentage of
compounds (ca. 4%), however, were found to be minimally irritating to
the eye but were severely irritating to the skin.
Results of a toxicity survey of 145 detergents showed that all were
eye irritants, whereas 47% also were either primary skin irritants or
corrosives. (Seabaugh et al., 1977).
The above findings, as well as the generally recognized relationship
between the effects of chemicals on skin and mucous membranes,
support the IRLG and OECD guidelines provision that substances
found to be severely irritating in dermal irritation tests are assumed to
be ocular irritants and, therefore, need not be tested in the eye.
3.2 TEST AGENT VOLUME
The original Draize test specifies the instillation of 0.1 milliliter of
the test agent into the conjunctival sac of the rabbit eye. The procedures
of the FHSA, IRLG, and OECD stipulate a volume of 0.1 milliliter for
liquids and a volume equivalent to 0.1 milliliter or a weight of not more
than 100 milligrams in the testing of solids. According to the test
protocol recommended by the NAS, two or more different doses of the
test material which are within the range of probable human exposure
should be used. Suggested doses are 0.1 and 0.05 milliliters.
Griffith et al. (1980) studied the irritancy of several substances in
the albino rabbit eye with the objective of producing a degree and
persistence of irritation approximating that commonly found in human
experience. Twelve test materials were each administered in volumes of
0.01, 0.03, and 0.1 milliliters of liquid or dry weight equivalent. Only
hexane and 0.1% benzalkonium chloride produced no observable irrita-
tion at all tested volumes. The average maximum scores for the irritat-
ing materials were found to be lower for the smaller volumes of
material, and eyes treated with the smaller volumes tended to clear
completely in less time. Isopropyl alcohol, however, displayed a flatter
dose/ response pattern. The 0.01 milliliter dose was the most sensitive in
separating the irritability indices of the materials for all observation
periods except day 21, in which case 0.03 milliliter resulted in better
separation. The test materials were classified as innocuous or slightly
11
-------
irritating to severely irritating or corrosive to the human eye, based on
descriptions in the literature, occupational incidents, and reports of
consumer exposures. When the evaluations for humans were compared
with the experimental results for each test volume, the 0.01 milliliter
dose data best matched human experience. While recognizing the
importance of attempting to achieve greater precision and sensitivity in
the assessment of irritation, the authors view accuracy in prediction of
human responses as an even more important goal.
Bell et al. (1979) studied the effect of varying the percent activity of
test materials rather than the volume instilled in the eye. Shampoos with
approximately 15% active chemical (undiluted) usually produced mod-
erate tissue damage when diluted to 2.5% to 3.0%, and this dilution
more readily distinguished between preparations in terms of corneal
and iridial damage than did no dilution or lesser dilutions. Concentra-
tions of at least 7% active matter accentuated changes and masked
differences in irritant potential which had been detected at lower
concentrations.
Use of the standard volume (0.1 milliliter) has historically provided
reliable detection of the adverse effects produced by test agents. Recent
evidence suggests, however, that the use of smaller volumes (e.g., 0.01
milliliter) or appropriate dilutions of test agents can increase the ability
of the test to differentiate between the irritancy potential of different
substances. For substances which require extensive characterization,
testing with additional animals to assess the effects of smaller volumes
or dilutions may thus provide useful supplementary information.
3.3 METHOD OF EXPOSURE
3.3.1 Liquids and Solids
The procedures specified by Draize, the FHSA, IRLG, and OECD
call for the direct instillation of the test agent into the lower conjunctival
sac of the rabbit eye. In a survey of individuals involved in testing
products for potential eye irritation, Beckley (1965b) reported that the
sprinkling of a powder or granular substance on the surface of the
cornea rather than directly introducing it into the lower conjunctival sac
was recommended as more closely simulating accidental entry into the
human eye. According to the NAS, application to the eye should be
conducted in a manner that reflects the probable mode of accidental
exposure. When direct application is called for, the material should be
dropped or sprinkled directly onto the cornea rather than being instilled
into the lower conjunctival sac. The lids should be held open momen-
tarily to ensure contact of the substance with the cornea, then gently
12
-------
released. While direct corneal application may be justified at times in
the assessment of specific hazards, the overall effectiveness of conjuncti-
val sac instillation supports its continued use as the standard procedure.
3.3.2 Aerosols
The test guidelines of the IRLG and OECD include a specific
provision for the testing of aerosols. Aerosol products are administered
in a single, short burst for about one second at a distance of 10
centimeters (OECD) or about four inches (IRLG) directly in front of
the eye. This provision responds to a criticism reported by Beckley
(1965b) concerning the testing of aerosols in which 0.1 milliliter of the
collected aerosol liquid is directly instilled into the conjunctival sac.
Spray exposure to aerosols may cause eye damage not only due to
the active ingredients, but also to the physical impingement of the
particles as well as the cooling produced by the propellant (Idson, 1968).
MacLean (1967) reported cases in which accidental exposure to house-
hold sprays resulted in an "aerosol keratitis" (corneal reaction to the
impregnation of the epithelium with small particles by the force of a
pressurized chemical spray). This keratitis is generally delayed in onset,
mild, and transient; although recovery from the more severe forms may
take several months.
Giovacchini (1972) has reported using three different treatment
groups of rabbits in testing aerosols. Two groups received the liquid
concentrate. The eyes of one of these groups were subsequently washed.
In the third group, the material was sprayed into the open eye for one
second at a distance of six inches. Others have compared the results of
spray exposure to aerosols with the results of direct instillation. Expo-
sure of the rabbit cornea to several brands of hair sprays showed that all
of the products were eye irritants when dispersed as a spray (details of
exposure not given), while only some products were eye irritants follow-
ing direct instillation of the collected hair spray liquid (Marzulli and
Simon, 1971). Kay and Calandra (1962) have also examined the influ-
ence of exposure methods on results using aerosol hair sprays. Draize
instillation of 22 products resulted in a mean score of 4.5, whereas spray
instillation of 13 products at an unspecified distance from the eye
produced a mean score of 3.5. The former value places the products in
the classification of moderately to severely irritating, while the latter
classifies them as mildly to moderately irritating. Since unequal
numbers of unidentified products were tested by the two methods, it
cannot be concluded that the differences in the mean scores and result-
ing ratings were due only to the method of exposure. MacLeod (1969)
reported that spray exposure to tear gas in monkeys (into the open eye
for two seconds at a distance of four centimeters or directly in front of
the face for four seconds at a distance of six feet) produced effects of less
13
-------
severity and persistence than direct instillation. Further knowledge of
the manner in which humans are exposed to aerosol sprays may be of
help in selecting the appropriate procedures for the testing of these
products.
3.3.3 Corneal Applicators
Devices have been used to direct the administration of test material
to the cornea and exclude it from contact with the conjunctiva. In a
comparative evaluation of eye irritation in monkeys and rabbits,
Buehler and Newmann (1964) found that the use of a cup aspirator in
rabbits produced lower scores in tests with surfactant formulations than
those that were obtained by instillation into the conjunctival sac. These
scores more closely approximated the response seen in monkeys treated
by direct instillation onto the cornea. On the other hand, the application
of 1% sodium hydroxide with a cup aspirator in rabbits produced a
more severe corneal opacity (with permanent impairment of vision)
compared with the effect produced by instillation into the conjunctival
sac. Battista and McSweeney (1965) demonstrated that the use of a
corneal applicator in rabbits produced a greater and more uniform
corneal opacity and less variability between tests when compared with
instillation of the test agent into the conjunctival sac. A viscous solution
of strong alkali was used to determine the effective time required to
produce an opacity score of 2 after seven days. The result following
conjunctival sac instillation was 7.75±2.5 (mean±S.D.) seconds com-
pared with 3.9 ±0.87 seconds for applicator instillation. Although the
detection of irritants requires the evaluation of damage to the whole eye
and not to the cornea alone, the carefully controlled exposure provided
by application devices may be useful at times for characterizing specific
corneal effects.
3.4 DURATION OF EXPOSURE
The role of the length of exposure and the efficacy of flushing the
eyes with water following treatment in modifying and ameliorating the
severity of the irritation reaction are determined when the original
Draize test methods are used. Floyd and Stokinger (1958) found that
the severe irritation produced by several organic peroxides (methyl
ethyl ketone peroxide, cumene hydroperoxide, and t-butyl hydroperox-
ide) could be prevented in all cases by washing the eyes of the rabbits
with water four seconds after instillation of the test solutions. Olson et
al. (1962) demonstrated that washing within 30 seconds after applica-
tion reduced the severity of reaction to surfactants in the rabbit eye.
14
-------
Gaunt and Harper (1964), using the Draize methodology, studied
the potential irritancy of 10 commercially available shampoos. Treat-
ment with each sample was as follows: sample instilled, no further
treatment; 10% dilution of sample instilled, no further treatment; sam-
ple instilled, eye irrigated with 20 milliliters water at two seconds; and
sample instilled, eye irrigated with 20 milliliters water at four seconds.
Dilution of the sample with water prior to its instillation into the eye
resulted in marked decreases in the scores. While irrigation provided an
ameliorative effect in most cases, in one case a more severe effect was
produced; one shampoo induced severe corneal opacity in two out of six
rabbits when the eyes were irrigated with water. Little difference in
benefit was observed between the two and four second washing proce-
dures. These results show that a test protocol using one group of
animals which receive the undiluted test agent with no further treatment
and another group of animals which have their eyes washed at a fixed
interval following exposure to the test material may provide valuable
information on the eye irritation capacity of test materials.
Guillot et al. (1981), in the study mentioned earlier, performed tests
with irrigation according to the OECD protocol, whereby substances
found to be irritating may subsequently be tested in separate groups
receiving irrigation at 4 and 30 seconds after instillation. Twelve of 23
substances rated irritating to very irritating received lower classifica-
tions of irritancy with rinsing at 30 seconds. For eight of the remaining
11 substances for which rinsing at 30 seconds produced no classification
change, a lower classification of irritancy resulted when rinsing was
initiated at four seconds. For 7 of 15 severely to extremely irritating
substances, rinsing at 30 seconds resulted in a lowering of classification.
With four of these seven substances, rinsing at four seconds produced a
further lowering of classification.
A modification of the Draize test procedures was proposed by
McDonald and Shadduck (1977) for evaluation of "dermatologic type
products." Both eyes of albino rabbits are used, three rabbits compris-
ing each test group. Irrigation with 200 milliliters of tap water for one
minute follows treatment at 20 seconds, 5 minutes, and 24 hours using
separate test groups. Six sham control eyes each are washed at time 0
and at 24 hours. An additional six eyes serve as untreated controls and a
commercial product control is included among the test materials. Eyes
are examined and graded by a slit-lamp scoring system at 1, 24,48, and
72 hours and on days 7 and 14; the test may be terminated at any point if
all treated eyes have returned to normal. The wash times were selected
because ocular changes for a variety of products have been found to
differ depending on the length of exposure prior to washing. Of 75
shampoos tested, 80% produced ocular reactions of greater severity,
higher incidence, and longer duration in animals undergoing irrigation
15
-------
at 5 minutes following instillation compared with those undergoing
irrigation at 20 seconds or 24 hours. In experiments with a skin cleanser,
washing at 20 seconds following exposure entirely prevented the irrita-
tion observed at wash times of 5 minutes and 24 hours following
exposure. The early washing, however, resulted in more severe conjunc-
tival congestion with an acne scrub product.
Bayard and Hehir (1976) investigated different durations of expo-
sure to test agents in the rabbit eye irritation test. Twelve rabbits per
treatment group were exposed to an acid, base, or alcohol for five
minutes. An additional 12 rabbits were exposed for 24 hours. Results
showed that the effects of a 5-minute exposure were not significantly
different from those for a 24-hour exposure at the 24-hour examination,
but were significantly less severe for opacity, iritis, chemosis, redness,
and overall scoring when examined at 48 and 72 hours.
Davies et al. (1976) investigated the maximum delay time for
remedial irrigation in the rabbit eye following instillation of a 10%
solution of sodium lauryl sulfate. Corneal opacity or a dulling of the
normal corneal luster was produced in every rabbit when the delay
between instillation and washing was 20 seconds or longer. Eyes
exposed to the test agent for 20 or 30 seconds displayed reactions that
were similar to those in eyes exposed for up to two minutes and only
slightly better than the reactions observed following no irrigation.
Reactions were fewer and less severe with exposure times under 20
seconds. No corneal opacity was produced after four seconds exposure
although a dulling of the corneal luster was observed in four animals.
The range of critical exposure before corneal damage is produced in the
rabbit eye after instillation of a 10% solution of sodium lauryl sulfate
was, therefore, concluded to be 4-10 seconds. Irrigation with 100 milli-
liters of water produced no further reduction in the irritant response
observed with a 20 milliliter wash volume.
Using sodium hydroxide, acetic acid, and hydrochloric acid,
Battista and McSweeney (1965) investigated ocular irritation reactions
as a function of exposure time in rabbits. Appropriate rinsing schedules
and a corneal applicator were used, varying exposure conditions quan-
titatively between no-effect and maximum-effect levels. From the
resulting time response curves, the relative irritancy of the different
samples could be compared by estimating the exposure time required to
produce equivalent lesions. The following concentrations and approxi-
mate exposure times before rinsing produced an estimated opacity
score of grade one at the seven-day observation point: sodium hydrox-
ide: 1.0% at 26 seconds; acetic acid: 21% at 2 seconds, 10.5% at 28
seconds, and 5.25% at 65 seconds; and hydrochloric acid: 3.7% at 9
seconds and 1.9% at 70 seconds. With this method, interpretation and
comparison of data are reduced to a single time measurement for
achieving whatever level of response is chosen. In addition, an estimate
16
-------
of the maximum exposure time that will be tolerated by the eye for a
given test material can be obtained.
Seabaugh et al. (1976) compared 32 chemicals for effects of length
of exposure with a two-minute irrigation period. Irritancy changes,
recorded during a seven-day observation period, were compared for
exposure times ranging from 30 seconds to 24 hours. For 30-second
versus 24-hour exposures, decreases in irritancy were observed with
20% of the chemicals, increases with 5%, and no change with 75%. In
comparisons of 2-minute and 24-hour exposures, 15% decreased, 10%
increased, and 75% showed no change. For 5-minute compared with
24-hour exposures, 13% decreased, 10% increased, and 77% did not
change. For six of the chemicals tested, both a 2-minute and a 15-
minute irrigation was evaluated. In general, there were no apparent
benefits from washing rabbits eyes for either 2 or 15 minutes at any of
the shorter chemical exposure times as compared with the 24-hour
exposure. According to the authors, these data indicated that irrigation
of rabbit eyes does not contribute to the outcome of eye irritancy tests in
most cases when compared with the official FHSA procedure.
Gupta and Schiavo (1976) have reported a rapid testing procedure
in rabbits which differentiates low levels of ocular irritation by multiple
exposures to test agents. Two drops of a test solution are instilled into
the lower conjunctival sac of either eye hourly, nine times a day for four
consecutive days, according to a balanced incomplete block design.
Two different solutions are tested in each animal, using one eye for each
material. With this method, 10 rabbits were used to test four ophthalmic
formulations of benzalkonium chloride in different concentrations.
Normal saline was included as a control. Each of the 10 possible
combinations of the five formulations occurred once. Eyes were exam-
ined before treatment and before the first instillation on each day of
treatment. Ocular reactions were graded by the FHSA method of
assessment with the addition of conjunctival discharge scores. Gross
observations and a statistical analysis of the data revealed that the
formulation containing the highest concentration of benzalkonium
chloride caused the most ocular irritation. Previous testing of this
solution in rabbits using standard laboratory procedures and the Draize
method of grading ocular irritation failed to detect the reactions. Thus,
the repeated instillation of ophthalmic formulations permitted an eval-
uation of irritancy with particular relevance to human exposure
conditions.
The FHSA, IRLG, and OECD methods modified the original
Draize procedure by changing post-treatment washing from a require-
ment to an optional wash after the 24-hour observation period. The
OECD protocol further states that additional tests employing irrigation
may be indicated for certain substances found to be irritating. In these
cases, the eyes of six rabbits are washed for five minutes using a volume
17
-------
and velocity of flow which will not cause injury. Three of the rabbits
receive irrigation at four seconds after instillation; the other three
rabbits at 30 seconds. According to the NAS procedures, irrigation is
similarly not required for determining the inherent irritancy of a
substance.
3.5 SELECTION OF SPECIES
The use of the rabbit in eye irritation testing has been recently
criticized by animal welfare groups. The rationale for and problems
with this species selection as compared with other alternatives are
reviewed in the following sections.
3.5.1 Use and Characteristics of Different Species
The albino rabbit is the test animal specified in all eye irritation
testing guidelines and regulations for a number of reasons. The animal
itself is readily available, docile, easily handled, relatively inexpensive,
and easy to maintain. The rabbit eye is large and the corneal surface and
bulbar conjunctival areas are both large and easily observed. The iris is
unpigmented, allowing ready observation of the iridial vessels
(McDonald and Shadduck, 1977). There is also a wealth of comparative
irritancy data that has been generated using rabbit eyes. The large
conjunctival sac offers a convenient "pocket" for instillation of a mea-
sured volume of test material. On the other hand, limited comparative
data from controlled exposures of humans and rabbits generally show
that responses of the rabbit eye are much more severe and long lasting.
There are also several examples of cases in which rabbit tests have failed
to alert investigators to effects that appeared later in humans. Rieger
and Battista (1964) described cases in which the Draize test erroneously
classified a widely distributed shampoo as irritating and failed to pre-
dict the inherent danger in the use of a neutralizer or to select the milder
(by human assessment) of two relatively mild products.
The rabbit eye possesses several anatomical and physiological
differences from the human eye which may contribute to differences in
response. The cornea is thinner, Bowman's membrane, if it exists at all,
is essentially unrecognizable, there is a well developed nictitating mem-
brane, the fur surrounding the eye and eye lids is thick, the blink reflex is
not well developed, pH of aqueous humors are different (rabbit: pH 8.2,
human: pH 7.1-7.3), the rabbit eyelids are looser, and the tearing
mechanism is less effective (NAS, 1977; McDonald and Shadduck,
1977). In addition, despite the general marked improvement in stan-
dardization of laboratory animals, there are still significant variations
among laboratory rabbits. Weltman et al. (1965) have commented that
age, strain, and sex differences may be important in ocular irritation
tests, although no specific data were presented.
18
-------
In an effort to improve the reliability of tests in predicting human
eye irritation, other species have been examined including the guinea
pig, rat, mouse, hamster, chicken, dog, cat, and monkey. With the
possible exception of the monkey, none of these species has been
studied thoroughly enough yet to substitute for the rabbit as the stan-
dard test species.
Among non-human primates, the rhesus monkey has been used
most frequently. Rhesus monkeys, however, are expensive, difficult to
obtain, and are even less uniform than rabbits. Handling and restraint
can be difficult and their pigmented irises complicate the observation of
some irritation responses.
The corneal thickness of different animal species varies considera-
bly (Table 3). These differences in thickness are important, since the
capacity of an irritant to penetrate the cornea is inversely related to total
corneal thickness (Marzulli and Simon, 1971). Corneal penetration
rates of various substances in vitro have shown interspecies variations
(Marzulli, 1965). Kuhlman (1959) has shown species variation in the
enzyme concentrations of rabbit, cat, and rat corneal epithelium.
Table 3. Corneal Thickness of Several Species of Animals
Species
Thickness
(Millimeters)
Reference
Cat
Dog
Rhesus Monkey
Rabbit
Mouse
Human
0.62 Marzulli and Simon (1971)
0.55 Marzulli and Simon (1971)
0.52 Marzulli and Simon (1971)
0.37 Marzulli and Simon (1971)
0.10 Davson (1962)
0.51 Maurice and Giardini (1951)
19
-------
Recognition of the need for additional or alternate animal models
in eye irritation testing is illustrated by the criteria which have been
suggested for determining the significance of ocular irritation in the
assessment of dermatologic products (Aronson, 1975). The author
states that an irritation reaction classified as significant (any ocular
irritation reaction which leads to or can potentially lead to permanent
or functional change) requires evidence of irritation in both humans
and an animal species. When human data are not available, presump-
tive evidence may be demonstrated in two different animal species.
Initial testing should be performed in the albino rabbit and, when
necessary, analogous studies should be conducted, preferably in a
primate. Similarly, the NAS (1977) recommends the use of the monkey
as a second test species when confirmatory data are required.
While the rabbit eye is characterized by certain differences in
anatomy and physiology in comparison with the human eye, it also
possesses functional attributes which provide a reliable system for
assessing the irritant properties of chemicals. The generally greater
responsiveness of the rabbit eye increases the possibility of detecting
compounds which are eye irritants and provides for a conservative risk
assessment. Continuing research is needed, however, on alternate spe-
cies and test methods for better characterizing substances which pose
special hazards to man.
3.5.2 Comparison of Monkey and Rabbit Responses
Similarities in the anatomy and physiology of the monkey and
human eye suggest that the monkey might predict more accurately than
the rabbit model the potential of substances to produce eye irritation in
humans. Reactions to eye irritants have, therefore, been compared in
the monkey and rabbit (and in some cases, in humans). The eye irrita-
tion responses of the rhesus monkey and the albino rabbit were com-
pared in a comprehensive study by Buehler and Newmann(1964). Four
materials were tested by direct instillation into the conjunctival sac of
the rabbit eye and onto the cornea of the monkey eye. Test substances
were the cationic surfactant benzalkonium chloride, a liquid anionic
surfactant formulation (37% coconut alkyl-sulfate, 12% coconut etha-
nolamide, and 18% ethanol), a nonionic/cationic surfactant mixture
(20% nonylphenoxypolyethyleneoxy ethanol and 2% lauryl isoquinoli-
nium bromide), and sodium hydroxide. The rabbit eye was also tested
by application of these same materials with a cup aspirator to expose
the corneal surface without subjecting the bulbar and palpebral con-
junctiva to a corresponding insult. Resulting corneal effects in the
monkey after direct instillation were restricted to superficial changes
such as physiologic edema and slight surface alterations, whereas
20
-------
instillation into the rabbit conjunctival sac produced a hyper-response
characterized by varying degrees of opacity, pannus formation, and
tissue buildup. With the exception of sodium hydroxide, the response of
the rabbit eye more closely approximated the response of the monkey
eye when the cup aspirator was used. Differences in species response
were attributed to anatomical/physiological differences and to the
effect of conjunctival irritation on corneal response.
A preliminary study comparing ocular responses in the rhesus
monkey with those in albino rabbits and humans indicated that the
rhesus monkey more nearly reflected human responses than did the
rabbit (Beckley et al., 1969). Instillation of a 5% soap solution in rabbit
eyes showed almost no corneal epithelial effect. The same material
caused corneal epithelial damage in both monkeys and humans. While a
detergent composition produced irritation in all three species, the rabbit
response was greater than the human or monkey response.
Benke et al. (1977) observed that two surfactant formulations
produced severe irritation when tested undiluted in rabbit eyes. Moder-
ate irritation was produced when 10% dilutions were used. In monkey
eyes, however, the formulations were found to be considerably less
irritating even when undiluted and unrinsed.
A comparative study of responses of the rabbit and monkey eye
was reported by Green et al. (1978). A number of substances were
investigated in tests with and without irrigation. Reactions were graded
throughout a 21-day observation period by external observation in
addition to the use of a slit lamp scoring system. Responses to each
substance were classified based on the most severe response that was
seen in any one of a group of animals receiving the same treatment.
Histological observation and photographic documentation of the ocu-
lar lesions were also performed. The rabbit eye was found to be more
sensitive to injury than the monkey eye for most of the chemicals
studied, the principal exception being 5% sulfuric acid. The difference
in scores between the two species was greatest with 1% sodium hydrox-
ide in which case the rabbit response was much greater than that of the
monkey. The scores for the two species also followed different patterns
with time. With most substances the monkey eye showed maximum
response at one hour whereas the maximum response of the rabbit eye
was observed at 24 hours. The difference in scores between the two
species depended somewhat on the method of grading. Analysis of
variance snowed the general increase in the rabbit eye response to be
greater compared with the monkey eye when scores were obtained using
the slit lamp as opposed to external observation. No pattern was seen in
the comparative responses of the two species to irrigation (at two
minutes following instillation of the test agent), though rinsing gener-
ally produced lower scores. According to the numerical scores, rabbit
eyes treated with 1% sodium hydroxide benefitted markedly from
21
-------
rinsing but rinsing was less beneficial to monkey eyes treated with the
same agent. Classification of responses indicated that, in fact, irrigation
produced a marked increase in the response of the monkey eye to a 5%
solution of sulfuric acid. In the rabbit eye, however, irrigation lowered
the classification of the response to the same agent.
Table 4 presents the comparative response of the rabbit and mon-
key eye to test agents. For most test substances, the degree of irritation
produced in the rabbit eye is greater than that produced in the monkey
eye. A comparison of animal and human responses is presented in
Table 7. In general, the decreased response of the monkey eye compared
with the rabbit eye more closely approximates the response found in
humans. Testing in the monkey, therefore, may be valuable for predict-
ing the degree of irritation produced by substances contacting the
human eye. Use of the rabbit is advantageous for maximizing the
probability of detecting substances that may be eye irritants for man.
3.5.3 Other Species
Except for studies of monkeys, apparently little effort has been
made to establish and validate the use of species other than the rabbit in
eye irritation testing. According to Grant (1974), guinea pigs and rats
are less satisfactory than rabbits, though no details are given.
Sanderson (1959) reported using the guinea pig to assess ocular
irritation as part of a screening process to evaluate a number of organic
solvents and emulsifying agents for possible use as injection solvents in
toxicity testing. A 10 microliter drop of the compound was placed on
the corneal surface of one eye with the untreated eye serving as a
control. No details were provided regarding time course of observation
or whether any specific scoring system was used. The toxicity of glycerol
formal was reported in detail. Administration of this agent to the eye of
the guinea pig caused only slight temporary irritation without pupillary
constriction or obscuration.
Leuenberger (1973) used the rat to study the severity of lesions and
ultrastructural changes in corneal epithelium following topical applica-
tion of various anesthetics. All of the drugs produced similar effects.
Since others have observed non-uniformity in the effects produced by
these agents in different test systems, the rat model apparently possesses
limited discriminatory power.
The rat model was studied for sensitivity to three comercial pro-
ducts established as irritants by the Draize rabbit eye test (W. Troy,
personal communication, 1981). In the rat eye all three products were
found to be non-irritating. Mel'nikova and Rodionov (1979) found that
the rabbit eye was more sensitive than the eye of the rat or guinea pig to
the irritancy produced by the antibiotic, grisein. Rat eyes treated with
20% to 30% solutions of the antibiotic returned to normal at seven davs
22
-------
Table 4. Comparison of Response of the Rabbit and Monkey Eye to Irritants
Test Agent
More
Sensitive
Species Reference
Surfactant formulations Rabbit (a)
1% Sodium hydroxide Rabbit
Cytarabine hydrochloride Similar (b)
Liquid detergent Rabbit
Various materials Rabbit
5% Soap solution Monkey
Detergent Rabbit
Chloroacetophenone Rabbit (c)
Iodine solution Rabbit
NS (dj Rabbit
Surfactant formulations Rabbit
Commercial shampoos, cationic detergents Rabbit
Variety of substances Rabbit
5% Sulfuric acid (e) Monkey
Buehler and Newmann (1964)
Buehler and Newmann (1964)
Elliot and Schut (1965)
Beckley(1965a)
Carter and Griffith (1965)
Beckley et al. (1969)
Beckley et al. (1969)
MacLeod (1969)
Hood et al. (1971)
Giovacchini (1972)
Benke et al. (1977)
Gershbein and McDonald (1977)
Green et al. (1978)
Green et al. (1978)
(a) Response of the rabbit to test agent instilled with a corneal applicator more closely
approximated the monkey response.
(b) Peak effects were observed in rabbits at 7 days and in monkeys at 8-12 days.
(c) In some cases, the two species showed similar effects.
(d) Not specified.
(e) Eyes of both species were irrigated.
23
-------
but reactions in rabbit eyes persisted for three weeks following instilla-
tion of a 10% solution. A 30% solution in guinea pig eyes brought about
a progressive clouding of the cornea and areas of corneal opacity.
Because the main objective of the study was not the comparison of
ocular responses in different species, these comparisons were not fully
reported.
An evaluation of the corneal irritancy of four commercial sham-
poos and two cationic detergents in various animal species was con-
ducted by Gershbein and McDonald (1977). The following order of
decreasing corneal sensitivity to the shampoos was inferred from the
findings: rabbit, hamster, mouse, rat, guinea pig, dog, cat, rhesus
monkey, and chicken. The mouse and hamster corneas appeared to be
more sensitive to the two detergents than the rabbit, the sensitivity of
the guinea pig and rat corneas was lower, and the remaining species
appeared less sensitive although the results were based on only a small
number of eyes. The corneal lesions of the albino rabbit in the above
tests were generally more persistent and cleared with greater difficulty
compared with those of the other larger species. The rabbit also dis-
played the highest sensitivity to conjunctival damage by the various test
agents among the larger species examined. The authors suggest that
current eye data from the various species may not be strictly relevant to
man, and they recommend caution in extrapolating rabbit-eye findings
to human experience.
Results of eye irritation tests with a liquid detergent in the rabbit,
dog, monkey, and man were reported by Beckley (1965a). Corneal
reaction was found to be slightly greater in the dog than in the rabbit,
with the monkey cornea showing the least sensitivity. Iridial and con-
junctival sensitivity was greatest in the rabbit, and least in the monkey.
Results of three separate studies in humans with the same test agent
showed that the irritation response was characterized by varying
degrees of conjunctivitis. No effects were found in human iris or cornea.
Giovacchini (1972) observed that proposed products which were
irritating in the standard rabbit eye test were less irritating when tested
in monkeys and dogs. Studies in humans showed that the irritation
responses more closely resembled those of the monkey and dog rather
than the rabbit.
The sensitivity of various species to eye irritants is summarized in
Table 5. In view of the generally greater response to ocular irritants in
the rabbit compared with other species, including man, it would seem
reasonable to further explore the predictive value, discriminatory
power, and practicality of other animal test species.
24
-------
Table 5. Comparison of Response of Several Species to Eye Irritants
Test Agent
Decreasing Order of
Sensitivity of Species Tested
Reference
Liquid detergents
NS
Commercial
shampoos
Cationic detergents
Grisein
Rabbit; Dog (aj; Monkey
Rabbit; Dog, Monkey
Rabbit; Dog, Monkey
Rabbit; Hamster; Mouse; Rat; Guinea Pig;
Dog; Cat; Monkey; Chicken
Mouse, Hamster; Rabbit; Guinea Pig, Rat;
Dog, Cat, Monkey, Chicken
Rabbit; Guinea Pig, Rat
Beckley (1965a)
Carter and Griffith
(1965)
Giovacchini (1972)
Gershbein and
McDonald (1977)
Gershbein and
McDonald (1977)
Mel'nikova and
Rodionov (1979)
(a) Corneal effects were greatest in the dog.
(b) Not specified
3.6 TEST GROUP SIZE
The choice of test group size involves balancing economic consid-
erations and practicality with the benefits of using large test populations
for increased accuracy and reliability of test results. Weltman et al.
(1965) demonstrated that considerable differences may occur between
irritation scores of different sized groups. As the population decreased
from 24 to 4 animals, the Draize scores showed a greater degree of
variability and spread. The authors suggest that increasing the test
group size in the standard Draize procedure would help avoid chance
classification errors when small populations are employed.
Results of the study by Bayard and Hehir (1976), in which 12
rabbits comprised each treatment group, indicated that a reduction in
the number of animals would decrease the ability of a test to differen-
tiate degrees of irritation since a large variability between individual
rabbits was noted. Hey wood and James (1978), however, point out that
while reducing the number of animals would decrease the ability of the
test to differentiate degrees of irritation, increasing the group sizes
would not increase precision in scoring the ocular reaction. Nonethe-
less, accuracy in ranking the irritancy of test substances evidently could
25
-------
be improved by using larger numbers of animals, particularly for bor-
derline responses.
Guillot et al. (1981) compared the mean ocular irritation ratings in
two groups of three rabbits each with those obtained using six rabbits.
Classification differences due to test group size resulted for 25 of 56
substances tested. In only two cases, however, was a test substance
classified as non-irritating based on results in one group of three rabbits
while testing with three additional rabbits and with six rabbits resulted
in a rating of "slightly irritating." Thus the data showed that the use of
three animals in a preliminary test was adequate in differentiating a
positive from a negative response for roughly 96% of a wide variety of
substances. The importance of test group sizes in evaluating ocular
irritancy, therefore, appears to have been adequately addressed in the
methods currently required for regulatory purposes as well as those
which have been recommended.
The FHSA method reduced the Draize requirement from nine to
six animals per group. According to the IRLG and OECD, initial
testing should be performed on at least three animals, but additional
animals may be required to clarify equivocal responses. The NAS
procedures recommend that a minimum of four animals be used per
dose level, unless unequivocal evidence of severe irritation or corrosiv-
ity can be obtained using a smaller test group.
3.7 USE OF ANESTHETICS
The purpose of using anesthetics during eye irritation testing is to
reduce unnecessary pain and stress to the experimental animal without
altering the test results. While historically ocular anesthetics have not
been used, investigations into the use of these drugs have begun in
recent years in response to the increasing concern for the humane
treatment of test animals. Although it does not specifically preclude
their use, the FHSA testing procedure makes no provision for the use of
anesthetics. Both the IRLG and OECD guidelines provide the option to
use a local anesthetic prior to the instillation of the test substance.
Concern has been expressed that the use of anesthetics may inter-
fere with the test results. A comment received in response to the publica-
tion of the IRLG guidelines suggests that the use of anesthetics could
inhibit or reduce a normal blink reflex. This could cause corneal drying
and result in irreversible damage from a test agent. According to Grant
(1974), local anesthetics increase the permeability of the corneal epithe-
lium to drugs and chemicals.
Four topical ocular anesthetics were studied in the rabbit eye by
C. Hoheisel (personal communication, 1981). Effects of a number of
irritants were compared in the presence and absence of ocular anesthe-
sia induced by proparacaine hydrochloride, tetracaine hydrochloride
26
-------
benoxinate, and lidocaine. Whereas 0.5% sodium hydroxide caused
only conjunctiva! redness without iritis or opacity in control eyes, it
produced corneal opacity in proparacaine-treated eyes. The use of
proparacaine resulted in the reclassification of a 10% liquid dishwash-
ing detergent solution from the nonirritant to irritant category under
FHS A regulations. With other test agents, proparacaine did not affect
the irritation scores sufficiently to require reclassification as eye irri-
tants. Lidocaine did not sufficiently relieve pain and benoxinate inter-
fered with test scores. When dosed in a single application, tetracaine did
not adequately raise the pain threshold. When dosed in two applica-
tions 10-15 minutes apart, however, its pain relief properties improved
and the test scores were not altered.
Butacaine sulfate was studied in rabbits by Ulsamer et al. (1977).
Parameters measured following instillation of test agents to anesthe-
tized and unanesthetized eyes included opacity, corneal water content,
dry weight, and electrophoretic protein patterns. No significant differ-
ences in dry weight or in corneal electrophoretic protein patterns were
noted with any irritants tested. While neither opacity scores nor mois-
ture content produced a clear pattern of statistically significant differen-
ces for anesthetized versus unanesthetized corneas, the following effects
were observed. With 10% acetic acid and 10% ammonia, the unanesthe-
tized eyes had lower opacity scores. Following the administration of 5%
acetic acid and 1% ammonia, the anesthetized cornea had a higher
water content. Anesthesia produced a lower water content with 3%
sodium hydroxide.
Using 10% sodium lauryl sulfate as the irritant and 0.5% propara-
caine as the anesthetic in rabbits, Heywood (1977, as cited in Heywood
and James 1978), showed that no statistically significant differences
could be detected between the anesthetized and unanesthetized cornea.
The intensity of the reaction apparently increased following anesthesia
but there was no evidence of prolonged effect. In the monkey, anesthe-
sia had a very marked effect on irritants in which all reactions were
intensified and prolonged. The results indicated that use of a local
anesthetic is feasible when carrying out eye irritation tests in rabbits, but
is contraindicated in the monkey. In contrast to the findings of
Heywood, Bell et al. (1979) have reported the tendency of anesthetic use
to delay recovery from ocular damage in the rabbit, although the
specific type of anesthetic used was not identified.
Studies of the effects of anesthetic agents alone on the eye may
contribute toward identifying an anesthetic suitable for ocular irritation
testing. Gundersen and Liebman (1944) reported the effects of a number
of topical anesthetics on the regeneration of the guinea pig corneal
epithelium. The anesthetics were repeatedly administered to the eyes
following mechanical abrasion of the cornea. Of the anesthetics tested,
27
-------
0.5% tetracaine hydrochloride and 1.0% phenacaine hydrochloride
were the least toxic to the regenerating epithelium. The inhibitory effect
of 0.5% tetracaine was almost entirely eliminated by increasing the
tonicity of the solution in which it was administered. Tests with other
anesthetics showed an increasing inhibitory effect in the following
order: 1.0% butacaine sulfate, 4.0% larocaine hydrochloride, 4.0%
cocaine hydrochloride, and 10% cocaine hydrochloride. The effects
were also studied in cats and rabbits, but, according to the authors, the
results obtained in these species were equivocal. It was concluded that
the concentration and tonicity of anesthetic agents can exert a signifi-
cant effect in the corneal wound healing process.
In a similar study, Bykov and Semenova (1972) showed that 2.0%
lidocaine caused less delay in epithelial healing than 2.0% cocaine or
0.5% tetracaine in the rabbit cornea. Pfister and Burstein (1976)
reported that no disruptive effects were observable by scanning electron
microscopy following single-dose application of 0.5% proparacaine or
0.5% tetracaine to rabbit eyes. A preparation of 4.0% cocaine, however,
produced significant plasma membrane injury and loss of surface cells.
An ultrastructural examination of rat corneal epithelium following
repeated topical applications of 0.4% and 1.0% benoximate, 4.0%
cocaine, 2.0% lidocaine, and 0.5% tetracaine was reported by
Leuenberger (1973). With each of the anesthetics, scanning and trans-
mission electron microscopy revealed a loss of microplicae and micro-
villi on the corneal surface, corresponding to a fine corneal stippling
observed with the use of a slit lamp. No basic differences in the type of
ultrastructural changes or severity of lesions were found for the differ-
ent anesthetics.
Overall, these results suggest that further research is needed to
adequately demonstrate the acceptability of an anesthetic for use in eye
irritation testing. At present, tetracaine appears most promising.
3.8 OBSERVATION PERIOD
The original Draize test involved grading reactions at intervals of
24 hours to 7 days or longer for persistent injuries. Observation at 24,
48, and 72 hours is required by the FHS A and has been proposed by the
IRLG. The OECD guidelines recommend observation at 1, 24,48, and
72 hours. While the FHSA method does not specify the use of extended
observation periods, the IRLG states that the eyes may also be exam-
ined at 7, 14, and 21 days, at the option of the investigator. The OECD
recommends an extended observation period to determine the progress
and persistence of the lesions, but observation normally need not
28
-------
exceed 21 days after instillation. The NAS recommends an observation
period of 1 to 21 days. The following public comment was offered in
response to the proposed IRLG guidelines: "Extended observation (at
least three weeks) with emphasis on rate of recoveries of lesions rather
than on their interim intensity must be emphasized."
According to Green et al. (1978), whether a material has caused
structural change suggestive of permanent impairment of vision cannot
be determined until enough time has elapsed to permit healing. Based
on the results of their investigation of responses in the rabbit and
monkey, they concluded that (1) the observation period must extend up
to three weeks for adequate evaluation of the effects produced by
introducing chemicals into the eyes of animals, (2) some injuries may
not be evident until after three days have passed and many lesions
undergo significant healing within 21 days; and (3) it is not possible to
predict the ultimate outcome of the injury using observations made only
during the first three days after instillation of the chemical.
Guillot et al. (1981) found that, for 16 of 56 chemicals tested for eye
irritation in rabbits, reactions observed up to 72 hours after instillation
were less severe than those observed after an observation period of 7
days. Both methods were adequate, however, in differentiating a posi-
tive from a negative response in all cases.
A 72-hour observation period following instillation appears ade-
quate to define the irritation potential of a test material. Extended
observation periods increase the capacity of the test to detect delayed
reactions and to characterize the reversibility of lesions. Observation
periods of up to three weeks, therefore, are suggested when fuller
characterization of the irritation response is required.
4.0 ASSESSMENT OF THE IRRITATION REACTION
Key variables in the complex process of assessing the results of eye
irritation tests are the techniques used to examine the eye, the types of
responses that are recorded, the scoring system used, and the evaluation
and interpretation of the lesions produced.
4.1 EXAMINATION TECHNIQUES
Assessment of the ocular reaction can be facilitated by the use of a
slit lamp. Slit lamp biomicroscopy allows for accurate differentiation of
clinical signs in the transparent structures of the eye (cornea, aqueous
humor, lens, and vitreous body) by providing observation in optical
section with the slit beam and the microscope (Gelatt, 1981; Aronson,
29
-------
1975). All slit lamp biomicroscopes have two basic components: an
illuminating system of adjustable slit width and a binocular microscope
for magnification. The light can be moved freely and directed at the eye
in the frontal plane as well as at various angles to the eye and the
biomicroscope. The table mounted slit lamp biomicroscope offers max-
imum flexibility in illumination and magnification (up to 40x) as well as
photographic capability. The portable hand held model possesses slit
widths of generally 0 to 10 millimeters and magnification of 5 to 20x.
Examination of the location and extent of ocular lesions and the
measurement of corneal thickness are accomplished with the slit lamp
biomicroscope. Green et al. (1978) have found slit lamp examination to
be more sensitive than gross external observation for detecting long-
lasting damage to the eye. Baldwin et al. (1973) have described the use of
photography, using a photoslit lamp, in documenting ocular pathologi-
cal changes. Eye irritation reactions have also been examined with the
use of an ophthalmoscope (Weltman et al., 1965; Guillot et al., 1981).
Fluorescein staining is also a valuable aid in defining corneal
epithelial damage. Fluorescein allows visualization of very small super-
ficial lesions of the cornea which might otherwise be completely over-
looked. When the superficial layers of the epithelium are damaged, the
dye is readily taken up by the remaining deeper layers and it will
fluoresce when ultraviolet light is cast on the area (NAS, 1977). Inter-
pretation is facilitated by rinsing the eye with an irrigating solution to
remove excess and nonabsorbed fluorescein.
Fluorescein dye is available in two commercial forms that are
suitable for ocular testing: sterile ophthalmological solutions contain-
ing 0.25% to 1.0% sodium fluorescein and individual hermetically
sealed paper strips containing fluorescein. In response to the proposed
IRLG guidelines, one commenter felt that the possibility of contamina-
tion of the fluorescein solutions was a concern. The commenter sug-
gested that contamination could be circumvented by using the
fluorescein strips.
4.2 SUBJECTIVE SCORING
Perhaps the greatest difficulty with eye irritation testing is the
subjective scoring and evaluation of corneal damage. The original
Draize test ranked the total ocular irritancy of test materials on a scale
of 0 to 110. This score was based on the total of the individual subjective
scores for the cornea, iris, and conjunctiva, each multiplied by a factor
(Appendix A). The resultant weighting or "biasing" of the total score
was designed to compensate for the generally perceived importance of
each graded tissue in the normal visual process. Serious lesions such as
30
-------
pannus, phlyctena, and rupture of the eyeball are also reported. The
methods established by FHSA and recommended by the IRLG attrib-
ute greater significance to the individual scores obtained in the exam-
ined ocular tissues. These two methods and the method proposed by the
OECD all utilize the same system for grading ocular lesions. This
system, as described in the "Illustrated Guide for Grading Eye Irritation
by Hazardous Substances" (CPSC, 1976), is presented in Appendix B.
In the FHSA and IRLG protocols, a minimum of one positive
response observed in any of the examined tissues of at least four rabbits
indicates a positive test result. If only one animal exhibits a positive
reaction, the test result is regarded as negative. When two or three
animals exhibit a positive reaction, the IRLG guidelines provide for
either designating the substance as an irritant or, as is required with the
FHSA method, the test is repeated using a different group of six
animals. The second test is considered positive if three or more animals
exhibit a positive reaction. The FHSA method further requires a third
test if only one or two animals in the second test exhibit a positive
reaction. If any animal in the third test exhibits a positive reaction, the
substance is regarded as an irritant. The FHSA and IRLG methods
differentiate an irritant from a non-irritant but do not rank substances
according to the degree of irritancy which they produce.
According to the OECD proposed method, the individual scores
for ocular reactions do not represent absolute scores for the irritant
properties of a material. Rather, they should be viewed as reference
values and are meaningful only when supported by a full description
and evaluation of the observations.
For scoring of ocular irritant effects, the National Research Coun-
cil of the National Academy of Sciences (N AS, 1977) proposes both the
original scoring scheme of Draize and a technique using a slit lamp.
With the latter technique, the total individual score obtained at any one
observation period is the sum of all the ratings for the cornea, iris, lid,
and conjunctiva. The NAS also recommends including persistence of
response as a criterion in the classification of test substances in terms of
irritancy.
Kay and Calandra (1962) developed a system for rating eye irri-
tants based on the Draize scoring method. The extent and persistence of
irritation and the overall consistency of the data form the basic criteria
for the assignment of ratings from "non-irritating" to "maximally irri-
tating." Guillot et al. (1981) used a similar system for the interpretation
of test results. A mean index of ocular irritation (Draize scoring system)
is obtained for six treated rabbits at each observation time. The highest
mean index of ocular irritation obtained is used to sub-classify the test
substance between the limits of "non-irritant" to "extremely irritant. "A
final rating is then assigned based on the rate of recovery in a specified
number of animals.
31
-------
Ballantyne and Swanston (1977) have discussed the methods of
scoring and grading eye effects. They cited three major problems with
the Draize scoring system. The system describes only selected effects on
three tissues and as a result may yield an incomplete picture of the
reaction of an eye to a given chemical. Marked effects on tissues
considered less important are de-emphasized by the grading system
which can lead to difficulties in interpretation. Thirdly, the results are
reported not as the grades of effects seen but as biased scores arithmeti-
cally derived from the grades observed. The scoring system specified by
the FHSA was criticized for specifying effects on only three tissues and
for the resulting uneven distribution of the scores for the various effects
observed. The scoring system of Ballantyne et al. (1974) was cited as an
attempt to obtain a more uniform approach to grading. This system
records the scores for lacrimation, blepharitis, chemosis, injection of
conjunctival blood vessels, iritis, keratitis, and vascularization of the
cornea, each on a six point scale.
Another criticism of the Draize scoring system has been its inabil-
ity to correlate active inflammatory signs with permanent structural
change. Aronson (1975) has described a technique for observing
parameters in animals identical to those observed in human studies and
which introduces a classification of ocular structural change occurring
as the result of active inflammation. All observations are performed
with a slit lamp. The degree of inflammation required to produce
structural change is determined for the cornea. For example, a corneal
infiltrate greater than 2.0 millimeters in diameter must be present for
approximately four days before neovascularization of the cornea
occurs. The author notes that examination using a number of ocular
irritants would be necessary to establish whether the relationship
between active inflammation and structural change was statistically
reproducible.
Interlaboratory variation in eye irritation scores also has been
studied. In 1962, Russell and Hoch reported rabbit eye irritation results
of two variations of a shampoo formula tested by five different labora-
tories. Good agreement was observed between scores for rabbits given
identical treatment within the same laboratory, but differences in scores
and the resulting irritancy classifications of the tested samples were
found between laboratories. The authors recommended that proce-
dures be standardized and that a standard substance be developed
which could be used by any testing laboratory to determine whether it is
scoring at a "standard" level.
In a similar study, seven liquid materials were evaluated in the
Draize rabbit eye irritation test by 10 laboratories (Marzulli and
Ruggles, 1973). Each laboratory was furnished with the test methodol-
ogy and eye irritation score sheets to be used. Each substance was tested
in six albino rabbits of either sex. A volume of 0.1 milliliter of the test
32
-------
material was instilled onto the cornea of one eye of each animal and four
parameters (corneal opacity, iritis, conjunctival redness and chemosis)
were evaluated at 24, 48, and 72 hours and 7 days. Inclusion of all four
criteria for eye irritation resulted in a high degree of consistency among
evaluators in separating an irritant from a non-irritant. However, the
test was inadequate if only a single parameter was used to make the
evaluation. Marzulli and Ruggles concluded that the Draize test, as a
first step, is capable of providing reliable and reproducible information
regarding the eye irritant properties of test materials.
Weil and Scala (1971) in a collaborative study documented sub-
stantial variation in ocular scores within and among 24 laboratories
(nine consultant; six food or industrial chemical; seven toilet goods and
cosmetics; and two governmental laboratories). Reference procedures
similar to the Draize method and reference materials were supplied to
each of the participating laboratories. Certain materials that were rated
by some laboratories as the most irritating among those tested were
rated as the least irritating by others. Weil and Scala concluded that in
most cases the variation between laboratories resulted primarily from
the reading of the reaction as opposed to variation in interpretation and
performance of the procedures, although the latter factor was also
considered a component of the interlaboratory variability. The authors
commented that "unconscious bias or definite tendencies to over or
underread reactions or misinterpret the meaning of descriptive terms
can be counteracted only by having several or many laboratories per-
form similar tests."
A statistically based experimental design has been devised by
McDonald and Shadduck (1977) to examine variability in subjective
scoring among a group of investigators (analyst uniformity) and within
the same investigator (analyst precision). The design provides for four
groups of rabbits consisting of 24 eyes per group. Test groups receive
topical ocular treatments that result in reactions ranging from non-
irritating to severely irritating. A first group of 48 eyes is tested in the
morning and a second group in the afternoon. Observation of each
group is confined to one hour in order to minimize the effect of time on
ocular reactions. To assess analyst precision, each eye is observed twice
without prior knowledge of treatment. The degree of reproducibility is
evidenced by the tabulated scores for each ocular parameter. For
example, a conjunctival swelling score of two was given on 12 occasions
in both trials by the same investigator, who on seven other occasions
gave a score of three in the first trial and a score of two in the second trial
for a given treatment regimen. The precision of each of three analysts
was found to range from 73% to 88% for conjunctival congestion. A
reproducibility of generally 80% or better was observed for swelling,
light reflex, corneal intensity, iritis, and flare. This was considered to be
33
-------
good precision, for a subjective scoring system. For discharge, percent-
ages ranged from 57-79 which was considered inadequate. Analyst
uniformity for three investigators was considered to be very good for all
ocular parameters except congestion and discharge. Correlation coeffi-
cients were 0.90, 0.97, and 0.92 for corneal intensity and 0.63,0.59, and
0.61 for congestion for investigator A vs B, A vs C, and C vs. B,
respectively. The authors conclude that this experimental procedure
provides a sufficient number of eyes for statistical evaluation and is an
effective means of monitoring and improving investigator reproducibil-
ity. Related tests to assess individual reader consistency (Bayard and
Hehir, 1976) showed that each of three readers reproduced a positive or
negative score more than 90% of the time.
Historically, scoring systems based on subjective examination
have been universally used and generally are considered adequate;
although, as shown by the above studies, the use of these systems has led
to inaccuracies in the assessment of eye irritants. Without careful
standardization of the details of the test method as well as the observa-
tion and scoring system, significant interlaboratory variation is possi-
ble. With proper controls, subjective scoring systems generally produce
valid and useful information in a screening test and, with the use of
simple aids such as a standardized pictorial guide for scoring lesions (see
Appendix B), their accuracy can be measurably improved. Total evalu-
ation of each score obtained at each observation period is necessary.
Substances which are likely to come in contact with the human eye due
to anticipated patterns of production and use may require further
evaluation by more sophisticated methods.
4.3 OBJECTIVE METHODS OF ASSESSMENT
To overcome the disparities that can result from the subjective
scoring of eye irritation reactions, objective techniques have been com-
bined with and compared to the Draize examination procedure. These
techniques have included the measurements of corneal thickness,
intraocular pressure, corneal and conjunctival weight, and capillary
permeability, as well as performance of histological analyses. Objective
methods of assessment may be useful as aids to facilitate the evaluation,
and they have the advantage of being more readily standardized.
4.3.1 Corneal Thickness
Burton (1972) described the use of corneal thickness measurements
for assessing the corneal irritation produced in rabbits by materials of
differing degrees of irritancy. Corneal thickness was measured a few
hours prior to instillation and thereafter at intervals of 24,48,72 and 96
hours. For comparison, reactions were also scored by the'Draize
34
-------
method. Corneal swelling was determined by dividing the corneal thick-
ness value obtained on the day of observation by a value obtained prior
to treatment. This proportionality assumes that changes in refractive
index and radius of curvature (both of which influence corneal thick-
ness measurements) are a function of the degree of swelling which at
least partially compensates for individual variations in thickness of the
untreated cornea. Burton noted that, since the cornea swells with a high
degree of uniformity, measurement of corneal thickness at the corneal
apex alone is sufficient. Photographs of treated eyes show that the
thickness of the slit image is an approximate measure of the thickness of
the cornea; greater swelling is associated with greater irritation as
assessed by the Draize method; and measurements of slit image thick-
ness can distinguish subtle differences between reactions given the same
subjective score.
In the Burton study, corneal swelling observed in 100 rabbits
treated with a variety of materials was compared with the Draize scores
recorded on the same day. Corneal swelling correlated closely with both
the Draize opacity score and total corneal score. A relationship between
conjunctival reaction assessed subjectively and the degree of corneal
swelling also was observed, but the correlation was not as good as that
for the total corneal score. To investigate the relationship between the
persistence of reaction and corneal swelling, an analysis was made of the
results of experiments on 100 rabbit eyes to which 34 different materials
had been applied and in which swelling in excess of 130% of the
untreated thickness had been noted. The maximum corneal thickness
recorded during the first four days of observation was used to express
corneal swelling. Subjective assessment (without a slit lamp) classified
reactions as persistent in 57 of the 100 rabbit eyes observed. The mean
corneal swelling of this group was 164.14% +18.08%. The remaining 43
were judged not to have reacted persistently. These eyes had corneas
swollen to a mean extent of 156.72% +14.96% (P < 0.05 compared
with the value for substances classified as causing persistent reactions).
Persistent corneal damage was, therefore, associated with greater cor-
neal swelling. Burton concluded that the corneal thickness method of
assessing corneal damage would be effective in minimizing laboratory
discrepancies if the method of measurement and possibly the age/ size
range of the rabbits used were standardized to some degree.
Draize scores in relation to objective assessments of ocular irrita-
tion, including measurement of corneal thickness, were also investi-
gated by Conquet et al. (1977). Seven organic solvents ranging from
non-irritating to very irritating were tested in rabbits. Two hours after
instillation of the test solvent, the Draize score did not reveal any
corneal changes, although significant corneal thickness variations com-
pared with those of the controls were found for four of the seven
compounds tested (P < 0.05). At 24 hours, both methods detected
35
-------
corneal changes, but for two of the solvents (dimethylsulfoxide and
Carbitol®) corneal thickness measurements differed from control mea-
surements (P < 0.05), whereas the Draize corneal scores did not differ
from control scores. Both corneal thickness measurements and Draize
total scores, however, identified the same compounds as irritants. Gra-
dations of irritancy were similarly assessed by the two methods at two
hours (correlation coefficient, r = +0.86) and 24 hours (r = +0.99).
There was no significant correlation between the two methods at days 3,
7, and 11 (r = +0.71, +0.48 and +0.72, respectively). The authors
conclude that the Draize procedure did not permit adequate detection
of increased corneal thickness and, therefore, suggest that specific
corneal thickness measurements be made in addition to the standard
Draize test.
In rabbit eye tests with a 10% solution of sodium lauryl sulfate, a
maximum response in corneal thickness was observed at two hours
(0.50 +0.04 millimeters compared to baseline value of 0.34 +0.08),
whereas a maximum response in clinical parameters (redness, swelling,
discharge) was observed at four hours (Walton and Heywood, 1978).
Burton et al. (1981) used corneal thickness measurements to assess the
irritancy of chemicals applied to enucleated rabbit eyes. The in vitro
measurements, expressed as corneal swelling, showed a generally good
agreement with published results of eye irritation in rabbits.
The above studies and a review by Heywood and James (1978)
indicate that the measurement of corneal thickness may be a valuable
supplement to the standard eye irritation scoring system. Measure-
ments of corneal thickness will increase the time and expense of testing
but they also provide the advantage of using the same eye as its own
control.
4.3.2 Intraocular Pressure
The technique of applanation tonometry has been used to measure
intraocular pressure in the assessment of eye irritation. This technique
measures the force required to produce a degree of flattening of the
cornea (Heywood and James, 1978).
Walton and Heywood (1978) performed a subjective clinical
assessment, corneal thickness measurements, and applanation tonome-
try throughout a 24-hour period on rabbit eyes treated with a 10%
solution of sodium lauryl sulfate. A calibrated Perkins hand-held
tonometer (applanation) was used to measure intraocular pressure.
Prior to this measurement, the cornea was anesthetized with one drop
of 0.5% proparacaine hydrochloride. A small drop of evaporated milk
was applied to the tonometer prism to visualize the applanation rings.
The eyes were thoroughly irrigated with normal saline after each of
three tonometer readings.
36
-------
The intraocular pressure increased more quickly and returned to
near normal sooner than the other factors measured. The highest value
was reached at 30 minutes after instillation of the irritant (32.17 ±2.11
millimeters Hg; P < 0.001 compared to baseline value of 19.33 ± 1.47).
At 24 hours, intraocular pressure remained elevated above baseline
(P < 0.05). The untreated eyes also experienced a rise in intraocular
pressure (P < 0.05 at 2 hours). Increases in corneal thickness could be
correlated to increases in intraocular pressure. While swelling persisted
throughout the observation period, corneal thickness showed maximal
response after two hours. The maximal response in clinical parameters
was observed after four hours, at which time the corneas were dull and
there was evidence of iritis. Walton and Heywood (1978) conclude that
measurement of intraocular pressure may be of value in the assessment
of local eye irritation but it significantly increases the number of vari-
ables associated with the assessment. This measurement methodology
requires local anesthesia, introduction of a second material in order to
visualize the applanation rings, and thorough irrigation following the
procedure, all of which could influence the long-term effects of the
primary irritant.
Ballantyne et al. (1972) investigated the effects of different concen-
trations of various materials on intraocular tension in the rabbit. Mea-
surements were made with a hand-held applanation tonometer.
Ammonia, butyl carbitol, w-chloracetophenone and o-chloro-
benzylidene malononitrile all caused significant dose response increases
in intraocular tension at 10 minutes. Tensions returned to near normal
by 1 hour. With ammonia and w-chloracetophenone, higher concen-
trations produced marked increases in the rise of intraocular tension per
incremental increases in concentration. With butyl carbitol and
o-chlorobenzylidene malononitrile, the rise in tension per increase in
concentration was relatively uniform over the range studied. While
there was no evidence that these changes were a measure of irritancy,
the authors suggest that a relationship may exist between the rise in
tension and the severity of the subsequent reaction.
Intraocular pressure measurements were included in an objective
evaluation of the ocular irritation produced by topically administered
nitrogen mustard in rabbits (Maul and Sears, 1976). Nitrogen mustard
elicited dose/response increases in intraocular tension. Investigators
in an earlier study utilizing applanation tonometry (Smith and
Mickatavage, 1963) felt that intraocular pressure measurements as
ordinarily determined in such animals as rabbits, dogs, and cats were
not reliable. Such measurements, therefore, were limited to human test
subjects as an additional correlation with local toxicity responses. The
use of intraocular pressure measurements in the assessment of
responses in eye irritation testing requires further study.
37
-------
4.3.3 Corneal and Conjunctiva! Edema
In tests with seven organic solvents, Conquet et al. (1977) measured
the corneal and conjunctival water content of treated rabbit eyes.
Animals were killed at 2 and 24 hours following instillation of the test
agent. At two hours, corneal edema (percent dry weight/wet weight)
was different from controls for five of the compounds (P < 0.05).
Draize corneal scores, however, indicated no reaction with any solvent
at the two-hour observation period. At 24 hours, the degree of corneal
edema produced by three solvents (different from control; P < 0.05)
corresponded with a positive reaction assessed by the Draize corneal
score. A significant linear correlation was also found between corneal
thickness and corneal edema measurements at both 2 hours (r = +0.95)
and 24 hours (r =+0.93). Conjunctival changes were detected by
edema measurement and the Draize procedure. A significant correla-
tion was found between both methods at 2 hours (r = +0.94) and 24
hours (r = +0.96).
Lallier et al. (1975) determined that corneal and conjunctival
weights (edema) are precise and repeatable measurements. Based on
results with organic solvents, these measurements were suggested to be
of use for the quantitative assessment of ocular irritation. The data
reported by Lallier et al. (1975), however, show that the percentage dry
weight (cornea) never fell more than 10% below the control value with
even the most severe irritants (Heywood and James, 1978). Heywood
and James (1978) also cited the findings of Wright etal.( 1976) in which
the water content of the cornea showed no direct relationship to opac-
ity. The value of corneal and conjunctival edema measurements in eye
irritation testing is not readily apparent.
4.3.4 Capillary Permeability
Changes in capillary permeability have been assessed by determin-
ing the concentration of dye per gram of dry weight of conjunctiva
or per milliliter of aqueous humor following intravenous injection
(Conquet et al., 1977). Following instillation of various test solvents, a
significant linear correlation was found between capillary permeability
in the conjunctiva and the Draize conjunctival score at 2 hours
(r = +0.96) and at 24 hours (r = +0.98).
While Lallier et al. (1975) reported that the increase in the capillary
permeability at the blood/aqueous humor level appears to be the first
event occurring in eye irritation, Conquet et al. (1977) found that the
diffusion of dye in the aqueous humor did not allow a comparison with
scores achieved by other methods, particularly the Draize iris score.
Maul and Sears (1976) observed increases in the amount of intrave-
nously injected dye recovered from the conjunctiva and iris of rabbit
eyes following treatment with nitrogen mustard.
38
-------
4.3.5 Histological Investigation
Histological evaluation of ocular lesions entails the description
and classification of lesions by a pathologist and is, therefore, a subjec-
tive assessment. However, the lesions are examined at the microscopic
level and are systematically recorded in considerably greater detail than
with a macroscopic subjective assessment. The information obtained on
structural change provides a more complete evaluation of ocular dam-
age. In practice, the histological evaluation of eye damage is sufficiently
sensitive and reliable for this technique to be included with the objective
methods.
Heywood and James (1978) summarize the parameters which
should be evaluated in a histological examination of eye irritation: "The
normal cornea should show microvillae in the superficial cells of the
corneal epithelium, microtubules in Descemet's membrane and vesicles
in the endothelium, indicating active transport systems. There should
be no oedema and the collagen fibres in the stroma should be of normal
size and show no distortion. In the conjunctiva, there should be normal
microvillae and the epithelium cells should be rich in endoplasmic
reticulum. Normal goblet cells are of considerable importance, as they
are responsible for the mucin in the precorneal film. Alterations in any
of these parameters must be considered of significance."
Histological observation was included by Weltman et al. (1965) in
a comparative evaluation of techniques used to assess eye irritation. In
rabbit eye tests with two shampoos, a correspondence was noted
between the eye irritation scores determined at the time of death and the
degree of histologic damage. Approximately two-thirds of the treated
eyes showed considerable erosion in the stratified squamous layers of
the cornea during the first three days following instillation of the test
preparations. By the seventh day, all the eyes treated by both substances
were histologically normal in appearance, consistent with the evalua-
tion by the Draize procedure.
Tonjum (1975) studied the effects of benzalkonium chloride on the
rabbit corneal epithelium. Within two minutes of application of the
irritant, the superficial cells displayed small holes and a loss of microvil-
lae. Histological analysis was performed by Mel'nikova and Rodionov
(1979) in determining the ocular pathology in the rabbit and guinea pig
eye produced by the antibiotic grisein.
Histological analysis was also used to classify the responses of the
rabbit and monkey eye to a variety of test substances (Green et al.,
1978). A discrepancy was found, however, between the results of histo-
pathological and slit lamp examinations. At 21 days following treat-
ment, histopathologic analysis detected a few lesions which were not
detected by the slit lamp. These findings were considered consistent
with the supposition that pathologic examination is the more sensitive
39
-------
method. In several cases, however, histopathologic examination failed
to detect changes that had been judged by the slit lamp to be evidence of
injury. This was attributed to a possible failure to section the eye
through an area containing a small lesion, though differences in inter-
pretation of the evidence presented by the two methods were also
thought to be a possible factor in the discrepancy.
4.3.6 Photographic Observation
An investigation of ocular irritation by Weltman et al. (1965)
included the use of photography. Immediately before autopsy of rabbits
scheduled for histological analysis of the ocular tissues, the eyes were
photographed twice. The first picture was taken with the treated eye at
rest and the second one with the upper and lower lids held apart.
Histologic observations were made from different sections of each eye.
Examination of the photographs revealed that the external, physical
appearance of the test and control eyes corresponded to the conditions
shown by the various histologic sections.
Photographs of slit images obtained from untreated and treated
rabbit corneas are included in a report by Burton (1972). The "Illus-
trated Guide for Grading Eye Irritation Caused by Hazardous Substan-
ces" (CPSC, 1976) contains photographs of rabbit eye lesions which
correspond to the various grades of irritancy. Baldwin et al. (1973) have
suggested the use of photographic recordings of ocular lesions for
instructional purposes. Photographs and descriptions of reactions in
the rabbit and monkey eye included in a report by Green et al. (1978)
were considered by the authors to contribute a substantial data base for
comparing other substances with those employed in their investigation.
4.3.7 Additional Techniques
In an objective evaluation of the ocular irritation produced by
nitrogen mustard in the rabbit eye, Maul and Sears (1976) examined
changes in pupil diameter and the level of protein in the aqueous humor.
The latter changes were determined by refractometry of samples
obtained through corneal paracentesis. Miosis occurred in the pupil
within minutes after application of the irritant at concentrations of
0.1% to 10%. The irritant also elicited dose/response increases in pro-
tein concentration in aqueous humor (P < 0.01 at 0.1% nitrogen mus-
tard compared with control). These results, together with
measurements of intraocular pressure and hyperemia of the conjunctiva
and iris, allowed an analysis of the possible mechanism and pathway of
nitrogen mustard-induced ocular irritation in rabbits. The results sup-
port the hypothesis that the effects in the eye are produced by a specific
transmitter/ receptor mechanism for one or more substances released
by pain fibers in response to the induced irritation. Pupillometry was
40
-------
also studied by Krejci and Krejcova (1974). Rabbit eyes (and cultures of
rabbit corneal epithelium) were treated with N-butyl gallate and epi-
nephrine. Changes in pupil diameter were compared with development
of degenerative changes in tissue culture.
Hey wood and James (1978) reported that measurements of corneal
curvature are precluded because of inaccuracies that result from the
disturbing of the precorneal film and the superficial epithelial cells of
the cornea by irritants. The technique of enzyme histochemistry has
been used by Gasset et al. (1974) to compare the effects of three
ophthalmic preservatives on rabbit eyes. NADH2-oxidoreductase
staining was used to test the viability of the endothelium. Szary (1977)
combined slit lamp observations and aqueous humor measurements of
protein, sialic acid, seromucoid, and proteolytic activity to evaluate the
intensity of inflammatory changes in the anterior chamber of the rabbit
eye. Increased ocular inflammatory changes have been correlated with
increased ocular temperatures measured with an infrared thermometer
(Ashford and Lamble, 1974).
Specular microscopy has been used to visualize the endothelial and
epithelial layers of corneas both in vivo and in vitro. The use of clinical
specular microscopy in ophthalmological studies has been reviewed by
Sugar (1979). Scanning and transmission electron microscopy may
prove useful for evaluating cellular toxicity of topically instilled pro-
ducts (McDonald and Shadduck, 1977). The major drawback of elec-
tron microscopy in the evaluation of ocular tissue damage is the
requirement for observation of chemically fixed tissues in the dehy-
drated state (Burstein, 1980). Cryofixation has been described as an
alternative to chemical fixation of surface tissues (rabbit corneal
endothelium) for viewing by electron microscopy (Burstein and
Maurice, 1978).
5.0 ADDITIONAL IN VIVO STUDIES
5.1 TESTS IN HUMANS
Studies in human volunteers are valuable for predicting the irri-
tancy of products used in and around the eyes. These studies typically
follow the eye irritation test in animals and provide a direct evaluation
of a particular product's safety.
A one-day method was devised by Callahan et al. (1979) for
studying acute ocular irritation effects in humans from an over-the-
counter ophthalmic preparation. Sixty-one subjects of either sex partic-
ipated in the study. Prior to dosing, the eyes of each subject were
examined with a slit lamp and given a baseline rating. One eye received
41
-------
two drops of the active solution (containing 0.05% tetrahydrozoline)
and the other eye received two drops of a placebo (the same vehicle but
without tetrahydrozoline). This procedure was repeated every three
hours until four doses had been administered. Five minutes after each
dosing, the investigator's clinical findings, as well as the participants'
subjective assessments of burning, stinging, or itching, were given
numerical ratings relative to the degree of severity. Neither the investi-
gator nor the subject knew which eye had the active solution. The
clinical parameters investigated were lacrimation, palpebral and bulbar
conjunctival hyperemia, and lid edema. The test product proved to be
non-irritating. The majority of clinical scores achieved during the test-
ing period were for mild palpebral and conjunctival hyperemia. As
would be expected from the vasoconstrictive properties of tetrahydroz-
oline, however, fewer of the eyes which received the active solution
containing tetrahydrozoline showed conjunctival hyperemia. No
cumulative dose/response effect was apparent; subjects with positive
clinical scores or subjective symptoms from either solution were fairly
evenly distributed throughout the four evaluation periods. Since only a
few reactions were observed, an additional five-day irritancy protocol
was not initiated. The authors concluded that a carefully monitored
safety study is necessary for detection of subtle changes that may
constitute or lead to unforeseen ocular damage in the user of an oph-
thalmic preparation.
Van Abbe (1973) reported that both rabbit eye tests and human
usage trials failed to predict eye irritation under certain product use
conditions. A hairdressing for men, which was marketed after initial
screening tests indicated that it was safe, was reported to be an eye
irritant in 1 in 8,500 purchases. In every instance, the complainant had
been exposed to rain or snow after using the product. In some of these
cases, ophthalmologists noted pitting of the corneal epithelium and a
positive response to fluorescein. Reevaluation showed that rabbit and
monkey eye tests were negative for both the diluted and undiluted
product. However, definite irritation was observed when rinsings (sim-
ulating exposure to rain or snow) taken from the human head after
normal use of the hairdressing were concentrated by freeze-drying and
instilled into rabbit eyes. Transient corneal wrinkling or pitting was
noticeable one to two hours after initial instillation and became more
prominent after a second instillation. Corneal pitting was not seen after
24 hours. When single ingredients of the product were tested in aqueous
solution, only one ingredient, 13-mole ethoxylated oleyl alcohol, gave a
definite irritant response. Transient corneal wrinkling or pitting at one
to two hours and a slight reddening of the conjunctiva at 24 and 48
hours were observed. Much greater irritancy was shown on animal skin
than in the eye. When 0.01 milliliter of the product was applied daily for
four days in concentrated form or diluted up to 1:4 with water, irritation
42
-------
occurred in intact and abraded rabbit skin and on the ears of female
CF/1 mice. The same product did not have an effect on human skin in
patch testing or in normal use.
In experiments designed to corroborate the animal findings,
human volunteers used the product daily for five days. No irritation was
reported after rinsings from the head were instilled in one eye of the user
four times daily for five days. Although aqueous solutions of the
13-mole ethoxylated oleyl alcohol did produce a burning sensation,
lacrimation, and erythema, the delayed response reported in complaints
from users was not seen. Van Abbe suggests that, in the application of
rabbit eye test results to humans, investigators should perhaps pay more
attention to corneal pitting or wrinkling. A second instillation of test
agent may also be necessary, since in this case a positive response was
obtained in the rabbit eye only after two instillations.
Schuck et al. (1966) devised a method for measuring the eye
irritancy produced in humans by photo-oxidation products of polluted
air. Conditions of photochemical air pollution were simulated in a large
reaction chamber equipped with light sources capable of generating an
intensity equivalent to noonday sunlight. The eyes of the subjects were
exposed to the chamber atmosphere via welding masks mounted in the
sides of the chamber. Subjects indicated the intensity of irritation by
adjusting a scalepointer throughout a five-minute exposure period and
an investigator recorded their rate of blinking. A set of 70 experiments
was conducted in which formaldehyde was produced in the parts per
million range by irradiation of ethylene. A linear relationship was found
between reported eye irritation and formaldehyde concentrations
greater than 0.3 parts per million. A reverse relationship was observed
for lesser concentrations. Most subjects experienced the same irritation
intensity at concentrations of 0.05 parts per million of formaldehyde as
they did at 0.5 parts per million. In some subjects the relationship
between blink rate and reported severity of irritation appeared to be
inverse. In similar experiments, peroxyacyl nitrates were produced by
the irradiation of propylene. The authors suggest that eye irritation
produced by photochemical air pollution can be predicted by the level
of formaldehyde and peroxyacetyl nitrate.
5.2 COMPLEMENTARY TESTS
Irritation profiles that assess sensory irritation (sting/burn) in
addition to primary irritation may be useful in determining the ocular
hazards posed by substances as well as the acceptability of commercial
products to consumers. Tests which have been used to evaluate sensory
irritation include the blepharospasm (closing of the lids) test, the mouse
writhing test, and the mouse upper respiratory tract test.
43
-------
The concentration of the riot control agent o-chlorobenzylidene
malononitrile (CS) required to produce a blepharospasm response in
50% of a group of test subjects (EC50) was compared in guinea pig,
rabbit, and man (Ballantyne and Swanston, 1973). The human eye was
about seven times more sensitive to this irritant than the guinea pig eye
and about 18 times more sensitive than the rabbit eye. In a similar study,
Ballantyne and Swanston (1974) also reported species differences in
blepharospasm response to dilute solutions of dibenzoxazepine. The
human eye was demonstrated to be 40 times more sensitive than the
guinea pig eye and 90 times more sensitive than the rabbit eye. Because
of the species variation, the investigators concluded that results from
animal threshold studies could not readily be extrapolated to man.
In view of the inadequacy of the Draize test for predicting discom-
fort or stinging in human eyes, Shanahan and Ward (1975) investigated
the mouse writhing test as an animal model for estimating the relative
sting potential of eight experimental and three commercial shampoos.
Male mice were injected intraperitoneally with 0.2 milliliter of graded
shampoo concentrations in water. The writhing response was charac-
terized by contraction of the abdominal musculature and was often
accompanied by tension and flexion of the trunk and extension of one
or both of the hind limbs. One writhe per mouse was considered a
positive response as were other overt symptoms indicative of pain or
discomfort. Results were compared with the sting response from test
material dilutions in human eyes and with the irritation produced by
full strength concentrations in the Draize rabbit eye test. The mouse
writhing test successfully identified four shampoos which were signifi-
cantly more irritating than the reference shampoo. Tests in humans also
revealed that these four shampoos produced a high degree of eye sting.
The mouse writhing test also determined that six of the shampoos were
equivalent in mildness to the reference shampoo. Sting evaluations in
human eyes similarly showed these six to be the mildest. Only one of the
shampoos produced significant, visible irritation in the Draize rabbit
eye test. Shanahan and Ward suggested that the mouse writhing test
would be a valuable adjunct in predicting the potential for ocular
discomfort and stinging.
Dossou and Pamart (1979) also used the mouse writhing test to
determine the sting capacity of shampoos. Serial dilutions of each of 22
shampoos were investigated following intraperitoneal injection in male
and female Swiss mice. The number of mice showing a reaction and the
DC50 (concentration necessary to produce writhing in 50% of a test
population) were determined for each product. The baby shampoos
generally were more mild than the adult shampoos. The mouse writhing
test was also performed on selected primary constituents of the sham-
poo formulas. DC50 measurements were higher for male mice than for
females. The sting capacity appeared to coincide with the irritancy of
44
-------
the ingredients as determined by the rabbit eye test. In a related study of
13 shampoos, however, no correlation was apparent between Draize
scores and the DC50 values found in the mouse writhing tests.
Alarie (1966) investigated the sensory irritant potential of various
substances using the mouse upper respiratory tract test. This test deter-
mines irritancy by monitoring the respiration rate in restrained mice
exposed to aerosolized materials. Low exposure concentrations are
used to determine sensory irritation without cellular damage. Irritation
was found to correspond quantitatively and reproducibly with a
decrease in respiratory frequency and an increase in the duration of
expiration. The sensory irritant effects of 52 compounds were tested in
humans and mice. After humans were exposed to the test agents,
responses were evaluated in terms of eye, throat, skin, nose, or chest
burning; conjunctivitis; lacrimation; coughing; and gagging. Com-
pounds found to be either non-irritating or irritating in humans were
similarly classified by the mouse upper respiratory tract test. Thus, such
tests in mice appear to permit a qualitative extrapolation to humans of
the sensory irritant potential of materials in the human eye.
The irritancy levels of several widely used surfactants were ranked
comparatively in animal test determinations of sensory irritation, pain
response, and tissue damage (Ciuchta and Dodd, 1978). Rabbit blepha-
rosphasm and mouse writhing tests were performed, as were the stan-
dard Draize rabbit eye and skin irritancy tests. In addition, the sensory
irritation potential of the surfactants was objectively evaluated by the
mouse upper respiratory tract test. For this test, respiration rates were
determined in chambered mice exposed to aerosolized 15% solutions of
the test surfactants. Both the mouse writhing test and upper respiratory
tract test permitted classification of the test materials in terms of
sensory irritation potential. Because the surfactants induced ocular
anesthesia, the blepharospasm test did not lend itself to sensory evalua-
tion of the test agents. Almost all of the surfactants received basically
the same scores in the rabbit eye irritancy tests. The rabbit skin irritancy
test was more effective in differentiating the irritancy of test agents,
though pattern shifts were noted at different concentrations.
The ocular anesthetic effect of certain products has been investi-
gated further in humans and rabbits. Prompted by the possibility that
users of soaps and surfactants capable of causing corneal anesthesia
may unknowingly damage their eyes, Harris et al. (1975) studied the
anesthetic effect on human and rabbit corneas of five common bath and
facial soaps and four commercial shampoos. Shampoos were tested at
full concentration and at dilutions of 1:2 and 1:4. Soaps were tested as
10% solutions. For purposes of comparison, one group of rabbits and
one group of humans received 0.5% tetracaine. The degree of corneal
anesthesia was measured with an esthesiometer, consisting of a nylon
filament of controllable length with numbers that correspond to the
45
-------
pressure exerted on the cornea. Anesthesia was assessed by the numeri-
cal value at which a blink reflex was elicited. The mean difference in
esthesiometer readings between treated and untreated eyes at each time
interval was analyzed. All of the test compounds produced anesthesia in
rabbits that was similar to that produced by tetracaine, but much longer
lasting. Tetracaine-induced anesthesia lasted between 20 and 30 min-
utes in human subjects and was comparable with that seen in rabbits.
None of the soaps or shampoos, however, produced corneal anesthesia
in humans. The results indicated that findings of corneal anesthesia
induced by surfactants in rabbits may not be extrapolated to the human
eye.
While results of the mouse writhing test (Shanahan and Ward,
1975) and the mouse upper respiratory tract test (Alarie, 1966) correlate
with the human sensory response to irritants, Heywood and James
(1978) stated that animal systems cannot be used to measure the phe-
nomenon of stinging in humans. According to Buehler(1974), methods
are needed for measuring both pain and anesthetic effects that may
occur when foreign chemicals are introduced into the eye but at present
there appears to be no satisfactory animal model for these
investigations.
6.0 IN VITRO TESTS
In eye irritation testing in vitro techniques have only recently
gained prominence as possible alternatives to the use of animals.
Although there is little published information on in vitro methods
specifically designed to assess eye irritation, some techniques have been
developed and reported as adjuncts to standard animal tests. Methodol-
ogies under study employ isolated eyes and corneas, mammalian cell
cultures, and other test systems.
An early attempt at the practical application of findings observed
in vitro was reported by Shapiro (1956) who studied the swelling and
dissolution of rabbit corneas following immersion in sodium hydroxide
solutions. Swelling and destruction of the cornea was found to occur
more rapidly at higher temperatures as well as at higher concentrations.
In vivo application of alkali for varying periods showed that corneal
swelling and damage increase rapidly as a factor of exposure time.
Shapiro, therefore, suggested that, for alkali exposure, immediate and
continued irrigation of the cornea with cold water would retard the
destructive action of the alkali on the corneal stroma.
While not strictly using an in vitro test, J. Leighton (personal
communication, 1981) is currently developing a test system using the
inflammatory response of the chick chorioallantoic membrane as a
measure of the eye irritation potential of a test substance. Burstein
46
-------
(1980) described the use of in vitro methods for determining the poten-
tial ocular effects of ophthalmological materials. Test systems include
whole corneal explants, which may be maintained in vitro for several
hours; corneal perfusion, which allows corneas to function in vitro for
up to a few hours; and cultured cells, which are isolated from corneal
tissues of animals. Transmission and scanning electron microscopy and
specular microscopy have been used to monitor changes in corneal cells
and in both epithelial and endothelial surfaces of isolated corneas.
6.1 ISOLATED EYES AND CORNEAS
Methods for the in vitro incubation of the rabbit cornea, either as a
part of the whole eye or as an isolated specimen, have been investigated
by Mishima and Kudo (1967). Edelhauser et al. (1976) have studied the
corneal swelling rate and ultrastructure in isolated rabbit and human
corneas perfused in vitro with intraocular irrigating solutions.
As reported in a Workshop on Ocular Safety Testing sponsored by
the Cosmetic, Toiletry, and Fragrance Association (CTFA, 1980),
Johnson has described the apparatus and procedure for maintaining the
isolated rabbit eye in vitro. A reasonably good correlation was observed
when the in vitro effects of 11 chemicals on corneal thickness and
anatomical characteristics in the isolated rabbit eye were compared
with published results of in vivo eye irritation tests of the same chemi-
cals. The correlation was best with moderate to severe irritants.
Focusing on the importance to visual acuity of the endothelial
layer of the cornea, Hull (1979) described studies in which three differ-
ent groups of drugs were tested for their effects on this layer in perfused
rabbit corneas. A five-minute perfusion with commercially available
epinephrine 1:1000 caused rapid swelling of the corneas. The endothe-
lial cells were markedly swollen and were characterized by cytoplasmic
vacuoles and loss of organelles. The toxic agent was not the epinephrine
itself, but rather 0.1% sodium bisulfite, which is included as a preserva-
tive. Diluting the epinephrine five-fold to 1:5000 prevented corneal
swelling even with a three-hour perfusion. Hull recommended a 1:5
dilution of the epinephrine 1:1000 agent, used for pupil dilation during
cataract surgery, to prevent corneal edema and damage to the corneal
endothelium.
Similarly, benzalkonium chloride, a commonly used preservative
in ophthalmic preparations, produced marked swelling of the corneal
endothelial surface when tested at the clinically used concentration of
0.01%. Perfusion with a 6.5 x 10~4% solution of benzalkonium chloride
produced a high degree of swelling in the endothelial cells and their
cytoplasmic organelles. A dilution of 6.5 x 10~6% was required to
prevent corneal swelling and to preserve the normal endothelial mosaic
pattern and cellular integrity.
47
-------
The in vitro effects of each of two commercially available miotic
compounds, 0.01 % carbachol and 1.0% acetylcholine, were investigated
following perfusionfor 15 minutes. Corneal swelling did not result from
the acetylcholine preparation, but the carbachol preparation produced
a swelling rate of 53 microns/ hour during the first hour after perfusion,
which decreased to about 5 microns/ hour for the remaining two hours.
No evidence of alteration of endothelial cell ultrastructure was revealed
by the microscopic observations. These results led the author to advise
caution in the current intraocular use of the carbachol preparation.
An economical and efficient utilization of animal resources in
developing an eye testing system was described by Carter et al. (1973).
Tests were conducted with bovine eyes obtained from a commercial
slaughterhouse. Exposure of the dissected corneas to detergents in
physiological solution produced an increase in hydration and changes
in sodium, potassium, and chloride concentrations.
Specific in vitro tests may also be useful in elucidating the mecha-
nisms of ocular irritation. Green et al. (1979) used cat eye preparations
to investigate a possible mechanism by which dibenz(b,f)l,4-oxazepine
produces irritant effects in the cornea. The irritancy of dibenz(b,f)l,4-
oxazepine and three of its methylated derivatives was measured by
determining the threshold concentrations of the compounds required to
produce sensory discharges in the ciliary nerve of the excised cat eye.
The guinea pig blepharospasm test was also performed. Both tests
yielded a similar ranking of the order of irritancy of the compounds
studied. Dibenz(b,f)l,4-oxazepine was the most potent irritant and
ll-methyldibenz(b,f)l,4-oxazepine was non-irritating. These results
support the hypothesis that the irritant effects of dibenz(b,f)l,4-
oxazepine in the cornea depend upon the compound's ability to bind to
epithelial cell membranes, since the lower irritancy of the methylated
derivatives correlates well with their lower affinity for the proposed
binding site.
Burton et al. (1981) described an in vitro method using rabbit eyes
for screening substances for eye irritancy. The eyes were removed
immediately after death, placed in a temperature controlled chamber,
and superfused with isotonic saline. Test substances reported in the
literature to be severely, moderately, or non-irritating were applied to
the eyes and then removed by rinsing. An exposure period of ten
seconds was selected because it provided the best discrimination
between irritant and non-irritant substances. Corneal thickness mea-
surements and slit-lamp observations were used to assess the effects.
Results with the test substances in vitro showed a generally good
agreement with published observations of eye irritation in man and
animals for the same substances. Deficiencies of the method were also
pointed out. These included the inability of the test system to assess the
potential for recovery or to evaluate possible effects on the conjunctiva.
48
-------
The authors stated that live animal experiments may still be necessary
before a new chemical is used by man but that they should not be carried
out before the possibility of severe irritation has been excluded using an
in vitro test.
6.2 TISSUE CULTURE
During the CTFA Workshop on Ocular Safety Testing (1980)
mentioned earlier, Chan described techniques used to isolate and cul-
ture three distinct cell lines from the rabbit cornea: epithelial cells,
fibroblasts, and endothelial cells. Characteristics of the epithelial cell
culture showed a good correlation between the behavior of these cells in
vitro and in vivo. Chan suggests that this cell line offers the potential to
evaluate toxicity by quantitating changes in the rate of cell division and
migration, in the rate of terminal differentiation, and in the ability to
produce multi-layer formations in culture.
The effects of a variety of antiglaucoma drugs were studied in
cultures of corneal epithelium derived from rabbits and freshly enu-
cleated human eyes (Krejci and Harrison, 1970). Test solutions were
diluted 1000-fold because preliminary tests showed that higher concen-
trations (1:10, 1:100) caused rapid cell death in corneal epithelium in
tissue culture. All control and treated corneal specimens were observed
for growth of epithelial sheets and development of degenerative
changes. The results were considered to be potentially beneficial in
choosing the least toxic drugs for topical application to glaucomatous
eyes complicated by abnormal corneal epithelium.
The suitability of other cell types has been investigated in correla-
tions with primary eye irritation test results. The findings of Bell et al.
(1979) in studies of human buccal mucosa cells exposed to a test
material for 10-15 minutes closely resemble results obtained in the
rabbit eye. The in vitro studies utilized phase contrast microscopy to
determine the proportion of treated cells in which nuclei were no longer
clearly visible. The authors note that these in vitro tests did not encom-
pass the full range of diverse factors which influence eye irritation,
including the ability of a test agent to penetrate the cornea.
A recent attempt to devise a simple, straightforward in vitro
method as an alternative to the Draize test has been reported by Simons
(1980). Since a severe irritant acts directly on cells in vivo, cytotoxicity is
determined using the in vitro technique. Three shampoos that were
established by the Draize test as being of high, moderate, or low
irritancy were tested on an established line of mouse embryo cells
chosen for its sensitivity. After the cells were cultivated in a growth
medium for 18 hours, each shampoo, diluted in growth medium, was
added at various concentrations to the cell-medium suspension and
49
-------
incubated for another 48 hours. The cells were then collected and the
total number of viable cells was counted. Statistical analyses of the data
showed that, although the test system differentiated the most irritating
shampoo from the least irritating one, it could not differentiate the
moderately irritating shampoo from the least irritating one. These
results do suggest a basis for an in vitro system for the screening of
severe irritants as an alternative to the Draize test. While the discrimina-
tory power of the test needs improvement, the development of the
technique may be a significant step in the search for a reliable in vitro
method.
An in vitro assay which can be directly applied to tests of the
irritancy of surfactant chemicals has been described by McCormack
(U.S. Dept. of Health and Human Services, 1981). The surfactants
induce the release of radioisotopically labelled serotonin into the
medium of cultured peritoneal mast cells from the rat. While not all
chemical compounds or classes of compounds lend themselves equally
well to this test, the results with surfactants corresponded closely with
those obtained by the Draize test.
The above test systems are summarized in Table 6.
Table 6. Summary of Tissue Culture Tests for Assessing Eye Irritation
Cell Type
Measurement of Irritation
Comparison with
Results in Rabbit
Eye Tests
Reference
Rabbit Cornea
Epithelial Cells
Rabbit and Human
Cornea Epithelial
Cells
Human Buccal
Mucosa Cells
Changes in the rate of cell division
and migration, the rate of terminal
differentiation, and the ability to
produce multilayer formations in
culture
Development of degenerative
changes, i.e., changes in growth of
epithelial sheets
Proportion of treated cells in which
nuclei are no longer clearly visible
Similar
Mouse Embryo Cells Cytotoxicity
Rat Peritoneal Release of radioisotopically labelled
Mast Cells serotonin into the medium
Not specified
Similar
Similar (discriminatory
power not perfected)
Similar (with surfactants)
Chan (CTFA,
1980)
Krejci and
Harrison
(1970)
Krejci (1976)
Bell et al.
(1979)
Simons
(1980)
McCormack
(U.S. Dept of
Health and
Human Ser-
vices, 1981)
50
-------
7.0 EXTRAPOLATION OF ANIMAL DATA TO MAN
A comparison of animal and human response in eye irritation tests
is presented in Table 7. These studies include varying numbers of
chemicals and, therefore, are not all of equal significance. It is evident,
however, that the rabbit eye is at least as sensitive or more sensitive than
the human eye in a large majority of studies on substances in which data
from both species exist. This emphasizes the generally conservative
nature of rabbit eye test results when applied to human risk assessment.
Comparative results from other studies also show that the responses of
the monkey and dog eye usually correlate more closely with the degree
of response seen in humans.
Table 7. Comparison of Animal and Human Responses
Irritation
Type
Test Agent
Sensitivity of
Animal Response
Test Species Compared to Man Reference
Primary Irritation Shampoo Rabbit (a) More sensitive
Detergent Rabbit Less sensitive
containing
neutralizer
NS (b) Rabbit Less sensitive
Anionic
surfactant
formulation
Anionic
surfactant
formulation
Rabbit (direct More sensitive
instillation)
Rabbit (appli- Similar
cator instilla-
tion)
Anionic Monkey Similar
surfactant
formulation
Rieger and Battista
(1964)
Rieger and Battista
(1964)
Rieger and Battista
(1964)
Buehler and Newman
(1964)
Buehler and Newman
(1964)
Buehler and Newman
(1964)
Soap solution Rabbit (a) Less sensitive
Sap from
Euphorbia
Rabbit (a) Variable
Gaunt and Harper
(1964)
Crowder and Sexton
(1964)
51
-------
Table 7. Comparison of Animal and Human Responses (Continued)
Irritation
Type
Primary Irritation
(Continued)
Test Agent
Sap from
Euphorbia
Liquid
tietergent
Liquid
detergent
Shampoos
Various
materials
Various
materials
NS
Cytarabine
hydrochloride
Cytarabine
hydrochloride
Detergents
and phos-
phates
5% soap
solution
5% soap
solution
Detergent
Detergent
NS
Test Species
Dog (a)
Rabbit
Dog
Rabbit (a)
Rabbit (a)
Monkey (a)
Dog (a)
Rabbit
Monkey
Monkey
Rabbit
Monkey
Rabbit
Monkey
Rabbit
Sensitivity of
Animal Response
Compared to Man
Similar
More sensitive
More sensitive
More sensitive
More sensitive
Similar
Similar
Similar
Similar
Similar
Less sensitive
Similar
More sensitive
Similar
More sensitive
Reference
Crowder and Sexton
(1964)
Beckley (1965a)
Beckley (1965a)
Bonfield and Scala
(1965)
Carter and Griffith
(1965)
Carter and Griffith
(1965)
Carter and Griffith
(1965)
Elliot and Schut
(1965)
Elliot and Schut
(1965)
Coulston and Serrone
(1969)
Beckley et al. (1969)
Beckley et al. (1969)
Beckley et al. (1969)
Beckley et al. (1969)
Giovacchini (1972)
52
-------
Table 7. Comparison of Animal and Human Responses (Continued)
Irritation
Type
Primary Irritation
(Continued)
Sensory Irritation
Blepharospasm
Response
Mouse Writhing
Test
Mouse Upper
Respiratory
Tract Test
Test Agent
NS
NS
Hairdressing
Hairdressing
Hairdressing
rinsings
Aqueous
solution of
hairdressing
ingredient
Various
materials:
0.1 ml
instilled
0.01 ml
instilled
ortho-Chloro-
benzylidene
Malononitrile
Dibenz-
oxazepine
Dibenz-
oxazepine
Shampoo
Various
materials
Test Species
Monkey
Dog
Rabbit (a)
Monkey (a)
Rabbit
Rabbit
Rabbit (a)
Rabbit (a)
Rabbit
Guinea pig
Rabbit
Guinea pig
Mouse
Mouse
Sensitivity of
Animal Response
Compared to Man
Similar
Similar
Less sensitive
Less sensitive
More sensitive
Similar
More sensitive
Similar
Less sensitive
Less sensitive
Less sensitive
Less sensitive
Similar
Similar
Reference
Giovacchini (1972)
Giovacchini (1972)
Van Abbe (1973)
Van Abbe (1973)
Van Abbe (1973)
Van Abbe (1973)
Griffith et al. (1980)
Griffith et al. (1980)
Ballantyne and
Swanston (1973)
Ballantyne and
Swanston (1973)
Ballantyne and
Swanston (1974)
Ballantyne and
Swanston (1974)
Shanahan and Ward
(1975)
Alarie (1966)
53
-------
Table 7. Comparison of Animal and Human Responses (Continued)
Sensitivity of
Irritation Animal Response
Type Test Agent Test Species Compared to Man Reference
Corneal Soaps, Rabbit
Anesthesia shampoos
Induced by Test
Agents
Commercial Rabbit
shampoo
formulations
Commercial Monkey
shampoo
formulations
More sensitive Harris, et al. (1975)
More sensitive Giovacchini (1972)
Similar
Giovacchini (1972)
(a) Study in humans not conducted. Comparison obtained from available human experience by
the authors cited.
(b) Not specified
Other studies reported in Table 7 show that the instillation of
volumes smaller than 0.1 milliliter into the conjunctival sac of the rabbit
eye produces results of greater predictive value. Testing substances
under the conditions of intended human use may also permit a more
accurate indication of the human response. This was illustrated by the
previously mentioned work of Van Abbe (1973) in which initial conven-
tional testing of a corfimercial hair dressing in rabbits failed to detect
subsequent effects reported by humans. Similarly, the value of approxi-
mating human use conditions in the testing of over-the-counter oph-
thalmological preparations is recognized by the Food and Drug
Administration (Federal Register, 1980). The proposed test procedure
for determining the safety of these materials is a modified Draize test
which permits the sample to be instilled in the manner and frequency of
intended use.
Experimental methods for determining the adverse ocular effects
of substances are primarily valuable as comparative tools. No single
species or experimental procedure is universally predictive of the
human response. When available, human studies, either controlled
54
-------
product use trials or industrial exposure surveys, are especially valuable
since they avoid the uncertainties inherent in the extrapolation of
n YI i rv\ o 1 yl*-t + n 4- n —« *-. u
animal data to man.
8.0 APPROACHES FOR THE FUTURE
The use of intact animals for eye irritation testing is presently the
only valid method for determining ocular safety. At the same time,
research is underway on alternative test systems. The objective is to
develop methods that are predictive of the response of the human eye to
irritants, are practical and cost effective and reduce the current depen-
dence on in vivo testing. Since testing in laboratory animals appears
necessary for the foreseeable future, humane considerations require
that methodologies at least minimize both the number of animals
needed in the assessment process and the possibility of causing undue
pain or discomfort. The investigation of topical anesthetic drugs is
representative of the concern to reduce potential discomfort of the test
animals.
A scheme similar to the following might be considered in the future
to determine the eye irritation potential of a substance (see Figure 1).
Initially determine the pH of the test material. Substances of pH 11.5 or
greater need not be tested further since it can be assumed that they
would be damaging to the eye. The same exclusion will apply to highly
acidic substances of pH approximately 2 or less, although this pH limit
is less well-defined than that at the alkaline end of the scale. Additional
testing of selected acids in the pH 1-3 range should help to establish this
lower limit more precisely.
Materials that are neither highly alkaline nor highly acidic are then
tested for dermal irritation. A substance that proves to be severely
irritating or corrosive to the skin would probably be severely irritating
or damaging to the eye. Testing in the eye, therefore, would not be
necessary. If the substance proves to be only slightly to moderately
irritating to the skin, further testing should be considered.
In vitro methods of testing for eye irritation potential will continue
to be developed. When fully validated, these tests may provide valuable
alternative or supplementary methodologies of testing materials for
ocular safety. Approaches involving tissue cultures of corneal epithelial
cells or more common cell lines appear most promising. The major
barrier to be overcome is the limited correlation which has been seen to
date between in vivo effects and the results of in vitro tests. In addition,
a primary criticism to the use of in vitro tests for eye irritation is their
inability to measure the whole body reaction, e.g., the reversibility of
lesions, delayed responses, and effects on specific ocular structures.
55
-------
Further research is necessary to determine which responses in vitro will
form the basis for the exclusion from testing or a requirement for testing
in the intact animal. The critical measurement will be that which detects
compounds that have the capability to produce ocular damage. Those
scoring as severe irritants may not need to be further tested or, if
exposure to the human eye is probable, the substances may more
effectively be tested in the rabbit using smaller volumes or dilutions.
Similarly, for substances scoring less severely, the use of in vitro results
for test exclusions or for selecting conditions of the intact animal test
may depend on anticipated product use and exposure hazards. Cur-
rently, in vitro tests are not adequately validated for use in the risk
assessment process.
Substances which are not screened out on the basis of pH, dermal
irritation test results, or in vitro test results would be tested in vivo for
eye irritation. These substances typically would be expected to cause
reactions ranging from non-irritating to, at the most, moderately
irritating.
This tier approach for eye irritation testing would not necessarily
simplify the overall evaluation process. A battery of tests as depicted in
Figure 1, however, which are straightforward, inexpensive, and quick
to perform, could be an extremely valuable tool if properly validated.
Use of a tier approach could significantly improve the eye irritation risk
assessment process with no loss in precision, predictability, or accuracy
of the data and could enhance the acceptability of the methods used in
the process.
9.0 CONCLUSIONS
Establishing the degree of risk associated with exposure of the eye
to a substance requires the design and execution of well-planned tests.
Factors important in the selection of the test method and the specific
procedures to be employed include (1) economic and practical limita-
tions, (2) the ability of the test to predict human responses, and (3)
humane concerns in the use of test animals. The use of any animal
species in this process involves a delicate balancing of pragmatic consid-
erations with the requirements for reliable test results. The specific
details of the test methods, as we have seen, can be tailored to the
information needs of the investigator and/ or the regulating authority.
These details include test population size, number of doses, number of
examination periods, and the addition of separate groups to receive eye
irrigation. Factors such as the intended use of a new product or ingre-
dient and the conditions under which it will be produced and distributed
may also influence the selection of a particular test method.
56
-------
pH MEASUREMENT
pH<2OR> 11.5
pH>2AND<11.5
DERMAL IRRITATION TEST
PRESENT
SEVERELY
IRRITATING
TO CORROSIVE
NON-IRRITATING TO
MODERATELY IRRITATING
FUTURE
1
"IN VITRO" TESTS* I •
RABBIT EYE
IRRITATION TEST
NO
FURTHER
TESTING
*/n vitro tests will require further research
and validation to determine their overall
usefulness in the risk assessment process.
Figure 1. Tier Approach for Eye Irritation Testing
57
-------
The guidelines recommended by the Organisation for Economic
Cooperation and Development and the Interagency Regulatory Liai-
son Group will contribute to standardization in eye irritation testing
and thereby should improve the consistency and quality of the data
generated for federal regulatory purposes. An evaluation of these guide-
lines in relation to the available literature indicates, in general, a solid
base of support for the recommended procedures.
SPECIES
In spite of several concerns, the rabbit remains the species of choice
for eye irritation studies. Both the OECD and IRLG guidelines recom-
mend that testing be performed using healthy adult albino rabbits. The
literature, which consists mostly of rabbit data, also generally supports
the primary use of the rabbit in eye irritation testing as a sensitive test
species that provides a conservative extrapolation to man. The monkey
eye is less sensitive than the rabbit eye to most substances and appar-
ently is a more accurate predictor of human eye irritation. Cost, han-
dling, and availability, however, restrict testing in the monkey. Other
species have also been examined. These include the guinea pig, rat,
mouse, hamster, chicken, dog, and cat. The use of the rat in eye
irritation testing has not been adequately studied.
PREPARATIONS
Exclusion of substances from testing in the rabbit eye, based on pH
or the results of dermal irritation tests, appears to be warranted. In fact,
the determination of the pH and/or dermal irritation potential of a test
substance normally should be performed before eye irritation testing is
conducted. As to the actual pH ranges which qualify for exclusion,
there is a strong basis at the alkaline end of the pH scale for assuming
that substances of pH 11.5 or greater will be severely irritating or
corrosive. At the acid end of the scale, although the current guideline
exclusion of substances of pH 2 or less is certainly in the right range, the
evidence is less clearcut and further investigation seems indicated.
Substances which score as severe skin irritants or corrosive agents
in standard dermal irritation tests generally can be assumed to be at
least severely irritating to the eye and need not be tested, as recom-
mended in the guidelines. Review of a large number of test results
suggests that this relationship will be valid for roughly 95% of a wide
variety of chemicals. The classification of a substance as an eye irritant
based on dermal irritation test results could be confirmed if question-
able by performing the eye irritation test.
58
-------
PROCEDURES
Initial testing for eye irritation with three animals normally will be
sufficient to identify substances that are non-irritating or maximally
irritating. Testing with additional animals will be necessary to fully
characterize substances of intermediate degrees of irritancy (OECD) or
to distinguish a positive from a negative test result for substances
producing equivocal results in the initial three animals (IRLG). The
conservative extrapolation in risk assessment provided by the generally
high sensitivity of the rabbit eye and by the procedure of instilling a
relatively large volume of material permits the use of only three animals
in the initial test.
The volume of material recommended for testing by both the
OECD and the IRLG is 0.1 milliliter (or 100 milligrams). This recom-
mendation in large part is a reflection of the historical use of this volume
from the time of Draize to the present. The continued use of this dose
for comparative purposes seems justified. A recent study indicates,
however, that the response in the rabbit eye to smaller volumes more
closely approximates the response encountered in man from typical
accidental exposure. Although not specifically addressed by the OECD
and IRLG guidelines, the issue of multiple doses and smaller volumes
has been raised by the National Academy of Sciences. The NAS sug-
gests that the use of two or more different doses would generate more
information and permit the determination of dose response characteris-
tics of a test material. The use of smaller volumes (or greater dilution) of
the test agent also is reported to permit better discrimination among
substances of similar irritancy. Thus, in addition to the standardized
volume, the use of smaller test volumes (or dilutions) should be consid-
ered if fuller characterization of the irritant properties of the test
material is desired.
Similarly, direct instillation of the test material into the conjuncti-
val sac has been established historically as the standard method of
exposure in eye irritation tests. While direct application to the cornea
probably more closely simulates accidental exposure in humans, this
technique has been used only occasionally and requires further evalua-
tion and standardization. Instillation into the conjunctival sac as
recommended in the guidelines appears to be the most practical and
reliable method at present. With proper control and validation, direct
corneal application could be considered as an optional technique on a
case by case basis.
The method of application of pressurized aerosol products appar-
ently has not been extensively investigated. Direct timed application of
a pressurized spray to the eye at a specific distance (as recommended in
both the OECD and IRLG guidelines) has evolved as the preferred
method for assessing the hazard posed by such products.
59
-------
The IRLG guidelines do not provide specifically for separate test
groups to assess the effect of irrigation with water following instillation
of the test material. An optional provision does allow for washing the
eyes of the test animals after the 24-hour examination, if desired. The
generally beneficial effects of washing the eyes with water following
exposure to an irritant have been demonstrated in a number of studies.
The timing of the wash following exposure (i.e., the duration of expo-
sure) also appears to be critical in preventing or ameliorating the irritant
response. For substances found to be irritants, the use of a separate
group (or groups) to accurately assess the role of irrigation in modifying
the irritant response appears to be a useful adjunct. This option is
provided for in the OECD guidelines. Test results on the effects of
washing also can be used to evaluate the appropriateness of standard
emergency procedures in the event of accidental human eye exposure to
specific substances.
OBSERVATION
The recommended observation periods of up to 72 hours following
instillation specified in the OECD and IRLG guidelines are adequate to
define the irritation capacity of a test material. The value of extending
the period of observation to determine the progress and reversibility of
corneal involvement and/ or the irritation reaction is emphasized by the
OECD guidelines. In both sets of guidelines, however, the use of
observation periods in excess of 72 hours is left to the discretion of the
investigator. The general philosophy is one of flexibility sufficient to
fully evaluate the hazards posed by test substances. As stated in the
OECD guidelines, "The duration of the observation period should not
be fixed rigidly but should be sufficient to evaluate fully the reversibility
or irreversibility of the effects observed."
Aids such as fluorescein dye and various optical instruments can be
used to facilitate examination of ocular reactions. The guidelines do
provide for optional use of these aids, which can be advantageous in
identifying and characterizing subtle changes and differences in
response. The IRLG guidelines also recommend that both eyes of the
test animals be examined using optical instruments, fluorescein, ultravi-
olet light, or other appropriate means to assess any existing abnormali-
ties prior to testing.
EVALUATION
Given the empirical nature of eye irritation testing, the achieve-
ment of greater consistency and reliability in the scoring of lesions has
long been a major goal. Modifications of the original Draize test,
therefore, have included significant changes in the scoring system.
Besides the original system proposed by Draize, a scoring system that is
60
-------
now widely used is a modified version adopted by the Food and Drug
Administration in the Federal Hazardous Substances Act of 1964 and
published shortly thereafter in the "Illustrated Guide for Grading Eye
Irritation by Hazardous Substances."This system eliminates the use of
the multiplication factors in the original Draize system which were
designed to attribute more significance to the portions of the anterior
eye considered most critical to vision. This modified Draize system is
recommended for use by both the OECD and IRLG guidelines and
provides an adequate basis for characterizing ocular lesions and irrita-
tion. Development of the illustrated guide by the FDA contributed
significantly to the reliability of the scoring process. This guide was
revised and expanded by the CPSC in 1976. The method and plates
from the guide are incorporated as Appendix B in the report as a
reference and aid in the evaluation of eye irritation results. The OECD
recommends the use of such a guide to promote standardization and
uniform interpretation and reporting of the data.
Both the OECD and IRLG provide narrative descriptions of the
range of irritation responses that correspond to individual numerical
scores, but neither assigns descriptive terms (e.g., slightly irritating,
moderately irritating) to the overall scores. In the IRLG method, the
scores are used to differentiate an irritant from a non-irritant. The
IRLG, however, does note that the three highest scores for corneal
opacity are indicative of corrosive effects when opacities persist to 21
days. As pointed out by the OECD guidelines, individual scores should
be viewed as reference values and not as absolute scores. The final
classification of the irritant capacity of a substance should be based on a
full description and collective evaluation of all aspects of the responses
of the test animals.
APPROACHES FOR THE FUTURE
The use of intact animals for eye irritation testing is currently the
only valid method for reliably determining ocular safety. The appropri-
ateness of the OECD and IRLG guidelines for these tests has been
generally confirmed in this review. Alternative methods are being devel-
oped, however, to address practical, economic, and animal welfare
concerns regarding the testing process. The available data suggest that a
tier approach may be feasible. Substances would be initially screened
for pH, dermal irritation, and possibly for activity in in vitro tests when
adequately validated. These preliminary tests should identify most
severe irritants and significantly reduce the need for testing these sub-
stances in the eyes of intact animals.
61
-------
10. RECOMMENDATIONS FOR FURTHER
RESEARCH
In the course of the development of this report, a number of issues
were identified which could not be resolved due to lack of adequate
information. Several of these areas which could benefit from further
developmental research are listed as follows:
• The pH at which highly acidic agents can be considered severely
irritating or corrosive has not been established as clearly as that for
highly alkaline materials. The recommendation that substances of
pH 2 or less need not be tested due to their potential irritant
properties needs further validation.
• The continued use of a single volume (dose) for eye irritation
testing is based primarily on historical precedent. Insertion of a
provision in the OECD/IRLG guidelines for the use of smaller
volumes or dilutions in addition to the recognized standard dose
appears to merit consideration. Additional research is warranted
to further correlate findings on the degree of response of the rabbit
eye to various test volumes with human eye response to accidental
exposure.
• The eye irritation response produced by a test substance depends in
part on the method of its application. Additional studies should be
conducted to characterize the differences which can occur when a
substance is instilled into the conjunctival sac as opposed to direct
application to the cornea with and without applicator aids.
• Little information exists regarding the general suitability of rodent
species in eye irritation testing. Further research on species such as
the guinea pig and rat is clearly justified to determine if there are
acceptable and practical alternatives to the use of the rabbit and to
provide added flexibility in test procedures.
• Use of topical ocular anesthetics in the test process would increase
the acceptability of eye irritation testing in animals. Further work
is necessary to establish the ability of these drugs to reduce pain
without significantly influencing the test results.
• Continuing research is needed to develop and validate a battery of
adequately sensitive in vitro tests.
• Additional research on the testing of aerosols might be directed
toward a correlation of effects with those produced by standard
conjunctival sac instillation of the collected liquid. An understand-
ing of the influence of spray pattern and aerosol characteristics
would also be useful in interpreting results.
62
-------
11. REFERENCES
Alarie, Y. (1966) Irritating properties of airborne materials to the upper
respiratory tract. Arch. Environ. Health 13:433-449.
Aronson, S.B. (1975) The role of ocular irritation in evaluation of
dermatotoxicity. In: Maibach, H. (ed.), Animal Models in
Dermatology. Churchill Livingston, London and New York. 23-
35.
Ashford, J.J. and Lamble, J.W. (1974) A detailed assessment procedure
of anti-inflammatory effects of drugs on experimental
immunogenic uveitis in rabbits. Invest. Ophthalmol. 13:(6)414-
421.
Baldwin, H.A.; McDonald, T.O.; Beasley, C.H. (1973) Slit-lamp
examination of experimental animal eyes. II. Grading scales and
photographic evaluation of induced pathological conditions. /.
Soc. Cosmet. Chem. 24:181-195.
Ballantyne, B. and Swanston, D.W. (1973) The irritant potential of
dilute solutions of ortho-chlorobenzylidene malononitrile (CS) on
the eye and tongue. Acta. Pharmacol. et Toxicol. 32:266-277.
Ballantyne, B. and Swanston, D.W. (1974) The irritant effects of dilute
solutions of dibenzoxazepine (CR) on the eye and tongue. Acta.
Pharmocol. et Toxicol. 35:412-423.
Ballantyne, B. and Swanston, D.W. (1977) The scope and limitations of
acute eye irritation tests. In: Current Approaches in Toxicology.
Porton Down/Wilts., Engl., Med. Div., Chem. Res. Establ. 139-
157.
Ballantyne, B.; Gazzard, M.F.; Swanston, D.W. (1972) Effects of
solvents and irritants on intraocular tension in the rabbit. J.
Physiol. 226:12P-14P.
Battista, S.P. and McSweeney, E.S. (1965) Approaches to a
quantitative method for testing eye irritation. /. Soc. Cosmet.
Chemists. 16:119-131.
Bayard, S. and Hehir, R.M. (1976) Evaluation of proposed changes in
the modified Draize rabbit irritation test. Toxicol. Appl.
Pharmacol. 37(1): 186.
Beckley, J.H. (1965a) Comparative eye testing: man vs. animal.
Toxicol. Appl. Pharmacol. 7:93-101.
Beckley, J.H. (1965b) Critique of the Draize eye test. Am. Perfume.
Cosmet. 80:51-54.
63
-------
Beckley, J.H.; Russell, T.J.; Rubin, L.F. (1969) Use of the rhesus
monkey for predicting human response to eye irritants. Toxicol.
Appl. Pharmacol. 15:1-9.
Bell, M.; Holmes, P.M.; Nisbet, T.M.; Uttley, M.; Van Abbe, N.J.
(1979) Evaluating the potential eye irritancy of shampoos. Int. J.
Cosmet. Sci. 1(2): 123-131.
Benke, G.M.; Brown, N.M.; Walsh; M.J.; Drotman, R.B. (1977) Safety
testing of alkyl polyethoxylate nonionic surfactants. I. Acute
Effects. Fd. Cosmet. Toxicol. 15(4):309-318.
Best, C.H. and Taylor, N.B. (1943) The Physiological Basis of Medical
Practice. Baltimore, Williams and Wilkins.
Bonfield, C.T. and Scala, R.A. (1965) The paradox in testing for eye
irritation. A report on thirteen shampoos. Proc. Sci. Sect. Toilet
Goods Assoc. 43:34-43.
Buehler, E.V. (1974) Testing to predict potential ocular hazards of
household chemicals. In: Winek, C.L. (ed.) Toxicology Annual
1974, Marcel Dekker. 53-69.
Buehler, E.V. and Newmann, E.A. (1964) A comparison of eye
irritation in monkeys and rabbits. Toxicol. Appl. Pharmacol.
6:701-710.
Burstein, N.L. (1980) Corneal cytotoxicity of topically applied drugs,
vehicles and preservatives. Surv. Ophthalmol. 25(1): 15-30.
Burstein, N.L. and Klyce, S.D. (1977) Electrophysiologic and
morphologic effects of ophthalmic preparations on rabbit cornea
epithelium. Invest. Ophthalmol. Visual Sci. 16(10):899-911.
Burstein, N.L. and Maurice, D.M. (1978) Cryofixation of rabbit
corneal endothelium. Invest. Ophthalmol. ARVO Suppl., 212.
Burton, A.E.G. (1972) A method for the objective assessment of eye
irritation. Fd. Cosmet. Toxicol. 10:209-217.
Burton, A.B.G.; York, M.; Lawrence, R.S. (1981) The in vitro
assessment of severe eye irritants. Fd. Cosmet. Toxicol. 19:471-
480.
Bykov, N.F. and Semenova, D.I. (1972) The effect of local anesthetics
on reparative regeneration of corneal epithelium. Oftalmol. Zh.
27:173-175.
Callahan, D.R.; Parish, L.C.; Nevyas, H. J. (1979) A method to evaluate
the safety of an OTC ophthalmic solution in normal human
volunteers. J. Soc. Cosmet. Chem. 30:67-71.
64
-------
Carpenter, C.P. and Smyth, H.F. (1946) Chemical burns of the rabbit-
cornea. Am. J. Ophthalmol. 29:1363-1372.
Carter, L.M.; Duncan, G.; Rennie, O.K. (1973). Effects of detergents on
the ionic balance and permeability of isolated bovine cornea. Exp.
Eye Res. 17:409-416.
Carter, R.O. and Griffith, J.F. (1965) Experimental bases for the
realistic assessment of safety of topical agents. Toxicol. Appl.
Pharmacol. 7:60-73.
Ciuchta, H.P. and Dodd, K.T. (1978) The determination of the
irritancy potential of surfactants using various methods of
assessment. Drug and Chemical Toxicology l(3):305-324.
Cogan, D.G. and Hirsch, E.G. (1944) The cornea. VII. Permeability to
weak electrolytes. Arch. Ophthalmol. 32(4):276-282.
Cogan, D.G.; Hirsch, E.G.; Kinsey, V.E. (1944) The cornea. VI.
Permeability characteristics of the excised cornea. Arch.
Ophthalmol. 31:408-412.
Conquet, G.D.; Laillier, J.: Plazonnet, B. (1977) Evaluation of ocular
irritation in the rabbit: objective versus subjective assessment.
Toxicol. Appl. Pharmacol. 39:129-139.
Consumer Product Safety Act. 15 U.S.C. 2051; 1972
Consumer Product Safety Commission (1976) "Illustrated Guide
for Grading Eye Irritation by Hazardous Substances."Directorate
for Engineering and Science. Washington, D.C.
Cosmetic, Toiletry, and Fragrance Association (1980) Workshop on
Ocular Safety Testing, Washington, D.C., Oct. 7.
Coulston, F. and Serrone, D.M. (1969) The comparative approach to
the role of nonhuman primates in evaluation of drug toxicity in
man: a review. Ann. New York Acad. Sci. 162(1):681-704.
Crowder, J.I. and Sexton, R.R. (1964) Keratoconjunctivitis resulting
from the sap of candelebra cactus and the pencil tree. Arch.
Ophthalmol. 72:476-484
Davies, R.E.; Kynoch, S.R.; Liggett, M.P. (1976) Eye irritation tests -
an assessment of the maximum delay time for remedial irrigation.
/. Soc. Cosmet. Chem. 27(7):301-306.
Davson, H. (1962) The Eye. Vol. 1, Vegetative Physiology and
Biochemistry. Academic Press, New York.
Dossou, K.G. and Pamart, B. (1979) Method for evaluating the eye
irritation caused by shampoos and certain surface-active agents in
the mouse. Labo- Pharma- Probl. Tech. 27(286):319-324.
65
-------
Draize, J.H. (1959) Dermal Toxicity. In: Appraisal of the Safety of
Chemicals in Foods, Drugs and Cosmetics. The Association of
Food and Drug Officials of the United States. Austin, Texas. 46-
59.
Draize, J.H.; Woodard, G.; Calvery, H.O. (1944) Methods for the study
of irritation and toxicity of substances applied topically to the skin
and mucous membranes. /. Pharmacol. Exp. Therap. 82:377-390.
Edelhauser, H.F.; Van Horn, D.L.; Schultz, R.O.; Hyndiuk, R.A.
(1976) Comparative toxicity of intraocular irrigating solutions on
the corneal endothelium. Am. J. Ophthalmol. 81(4):473-481.
Elliot, G.A. and Schut, A.L. (1965) Studies with cytarabine HC1 (CA)
in normal eyes of man, monkey, and rabbit. Amer. J. Ophthalmol.
60:1074-1082.
Federal Register. (1964) Sept. 17. Federal Hazardous Substances Act.
Title 21. C.F.R. 13009 - 191.12 Test for Eye Irritants.
Federal Register. (1972) Proposed Rule Making. April28.37(83):8534-
8535.
Federal Register. (1978) Proposed Rules. Aug. 22. 43(163):37336-
37351.
Federal Register. (1979a) Proposed Rules. July 26. 44(145):44070-
44071.
Federal Register. (1979b) Aug. 15. Vol. 43, Pg. 49015
Federal Register. (1979c) Proposed Rules. Sept. 25. Vol. 44, No.
187.
Federal Register (1979d) Consumer Product Safety Commission. Code
of Federal Regulations Title 16, part 1500.42, U.S. Government
Printing Office, Washington, D.C.
Federal Register. (1980) Proposed Rules. May 26. 45(89):30000-30050.
Federal Register. (1981a) Interagency Regulatory Liaison Group.
Jan. 22. 46(14):7075-7077
Federal Register. (1981b) Proposed Rules. Jan. 27. 46(17):8992-8993.
Floyd, E.P. and Stockinger, H.E. (1958) Toxicity studies of certain
organic peroxides and hydroperoxides. Amer. Indust. Hyg. Assoc.
J. 19:205-212.
Food and Drug Administration (1965) "Illustrated Guide for Grading
Eye Irritation by Hazardous Substances." Government Printing
Office, Washington, D.C.
Friedenwald, J.S.; Hughes, W.F.; Herrmann, H. (1944) Acid-base
tolerance of the cornea. Arch. Ophthalmol. 31(4):279-283.
66
-------
Friedenwald, J.S.; Hughes, W.F.; Herrmann, H. (1946) Acid burns of
the eye Arch. Ophthalmol. 35:98-108.
Gasset, A.R.; Ishii, Y.; Kaufman, H. E.; Miller, T. (1974) Cytotoxicity
of ophthalmic preservatives. Am. J. Ophthalmol. 78(1):98-105.
Gaunt, I.F.; Harper, K.H. (1964) The potential irritancy to the rabbit
eye mucosa of certain commercially available shampoos. J. Soc.
Casmet. Chem. 15:209-230.
Gelatt, K.N. (1981) Ophthalmic examination and diagnostic
procedures. In: Textbook of Veterinary Ophthalmology. Gelatt,
K.N. (ed.), . Lea and Febiger, Philadelphia. 206-261.
Gershbein, L.L. and McDonald, J.E. (1977) Evaluation ofthecorneal
irritancy of test shampoos and detergents in various animal
species. Fd. Cosmet. Toxicol. 15:131-134.
Giovacchini, R.P. (1972) Old and new issues in the safety evaluation of
cosmetics and toiletries. CRC Crit. Rev. Toxicol. 1:361-378.
Grant, W.M. (1974) Toxicology of the Eye. Thomas, Springfield, 111.
Green, D.M.; Balfour, D.J.K.; Muir, A. (1979) Effect of methyl
substitution on the irritancy of dibenz(b,f)l,4-oxazepine (CR).
Toxicology 12(2): 151-153.
Green, W.R.; Sullivan, J.B.; Hehir, R.M.; Scharpf, L.F.; Dickinson,
A.W. (1978) A systematic comparison of chemically-induced eye
injury in the albino rabbit and rhesus monkey. The Soap and
Detergent Association, New York.
Griffith, J.F.; Nixon, G.A.: Bruce, R.D.; Reer, P.J.; Bannan, E.A.
(1980) Dose-response studies with chemical irritants in the albino
rabbit eye as a basis for selecting optimum testing conditions for
predicting hazard to the human eye. Toxicol. Appl. Pharmacol.
55(3):501-513.
Guillot, J.P.; Caillard, L.; Gonnet, J.F.; Clement, C. (1981) Chemicals -
Ocular and Cutaneous Local Tolerance-"Cosmetic," A.F.N.O.R.
and O.E.C.D. Protocols. Institut Francais de Recherches et Essais
Biologiques. Centre de Lyon, Les Oncins, b.p. 109, 69210
L'Arbresle.
Gundersen, T. and Liebman, S.D. (1944) Effect of local anesthetics on
regeneration of corneal epithelium. Arch. Ophthalmol. 31:29-33.
Gupta, S. and Schiavo, D.M. (1976) A rapid testing procedure to
determine low levels of ocular irritation. Lab. Anim. Sci.
26(3):468-472.
67
-------
Harris, L.S.; Kahanowicz, Y.; Shimmyo, M. (1975) Corneal anesthesia
induced by soaps and surfactants, lack of correlation in rabbits and
humans. Ophthalmologica 170:320-325.
Harriton, L. (1981) Conversation with Henry Spira: Draize test activist.
Lab Animal 10(1): 16-22.
Heywood, R. and James, R.W. (1978) Towards objectivity in the
assessment of eye irritation. J. Soc. Cosmet. Chem. 29:25-29.
Hood, C.I.; Gasset, A.R.; Ellison, E.D., Kaufman, H.E. (1971) The
corneal reaction to selected chemical agents in the rabbit and
squirrel monkey. Amer. J. Ophthalmol. 71:1009-1017.
Hughes, W.F. (1946) Alkali burns of the eye. 1. Review of the literature
and summary of present knowledge. Arch. Ophthalmol. 35:423-
449.
Hull, D.S. (1979) Effects of epinephrine, benzalkonium chloride, and
intraocular miotics on corneal endothelium. South. Med. J.
2(11):1380-1381.
Idson, B. (1968) Topical toxicity and testing./. Pharm. Sci. 57(1): 1-11.
Interagency Regulatory Liaison Group. Jan. 1981. Testing
Standards and Guidelines Work Group. Recommended Guideline
for Acute Eye Irritation Testing.
Jacobi, O. and Ortmann, E. (1971) Irritations of the skin and the
mucosae, by shampoos. Berufsdermatosen. 19:(2)75-83.
Kay, J.H. and Calandra, J.C. (1962) Interpretation of eye irritation
tests. /. Soc. Cosmet. Chem. 13:281-289.
Krejci, L. (1976) Influence of ocular drugs on the corneal epithelium.
Cesk. Oftalmol. 32(3): 163-168.
Krejci, L. and Harrison, R. (1970) Antiglaucoma drug effects on
corneal epithelium. Arch. Ophthalmol. 84:766-769.
Krejci, L. and Krejcova, H. (1974) Changes in pupil diameter and
effects on corneal tissue cultures from topically administered N-
butyl gallate (COMT inhibitor) and epinephrine. Opthal. Res.
6:15-22.
Krueger, R. (1959) Experimented und klinische Beobachtungen zur
Behandlung der Alkaliveratzung mit Ascorbinsaure. Bericht. d.
Opthalm. Ges. 62:255-258.
Kuhlman, R.E. (1959) Species variation in the enzyme content of the
corneal epithelium. /. Cell. Comp. Physiol. 53:313-326.
Laden, K. (1973) Studies on irritancy and stinging potential. /. Soc.
Cosmet. Chem. 24:385-393.
68
-------
Lalher, J.; Plazonnet, B.; LeDouarec, J.C.; (1975) Evaluation of ocular
irritation in the rabbit: development of an objective method of
studying eye irritation. Proc. European Society of Toxicology.
17:336-350.
Leuenberger, P.M. (1973) Ultrastructural changes of corneal
epithelium after topical anesthesia. Albrecht v. Graefes Arch. klin.
exp. Ophthal. 186:73-90.
MacLean, A.L. (1967) Aerosol keratitis - a common epithelial foreign-
body reaction to household chemical sprays. Amer. J.
Ophthalmol. 63(6): 1709-1719.
MacLeod, I.F. (1969) Chemical Mace: ocular effects in rabbits and
monkeys. /. Forensic Sci. 14:( 1)34-47.
Marzulli, F.N. (1965) New data on eye and skin tests. Toxicol. Appl.
Pharmacol. 7:79-85.
Marzulli, F.N. and Ruggles, D.I. (1973) Rabbit eye irritation test:
collaborative study. /. Assoc. Off. Anal. Chem. 56(4):905-914.
Marzulli, F.N. and Simon, M.E. (1971) Eye irritation from topically
applied drugs and cosmetics: preclinical studies. Am. J. Optom.
48:61-79.
Maul, E. and Sears, M.L. (1976) Objective evaluation of experimental
ocular irritation. Invest. Opthalmol. 15(4):308-312.
Maurice, D.M. and Giardini, A. A. (1951) A simple optical apparatus
for measuring the corneal thickness, and the average thickness of
the human cornea. Brit. J. Ophth. 35:169-177.
McDonald, T.O. and Shadduck, J.A. (1977) Eye Irritation. In:
Advances in Modern Toxicology, Marzulli, F.N. and Maibach,
H.I. (eds.), Vol. 4, Hemisphere, Washington and London. 139-191.
McLaughlin, R.S. (1946) Chemical burns of the human cornea. Amer.
J. Ophthalmol. 29(11): 1355-1362.
Mel'nikova, E.A. and Rodionov, G.A. (1979) Experimental pathology
of the rabbit and guinea pig eye caused by the antibiotic grisein.
Vrachebnoe Delo. 6:85-88.
Mishima, S. and Kudo, T. (1967) in vitro incubation of rabbit cornea.
Invest. Ophthalmol. 6(4):329-339.
National Academy of Sciences - National Research Council (1964)
Principles and Procedures for Evaluating the Toxicity of
Household Substances. Publication 1138, Washington, D.C.
69
-------
National Academy of Sciences - National Research Council (1977) In:
Principles and Procedures for Evaluating the Toxicity of
Household Substances prepared by Committee for the Revision of
NAS Publication 1138 for the Consumer Product Safety
Commission. Eye Irritation. 41-54.
Olson, K.J.; Dupree, R.W.; Plomer, E.T.; Rowe, V.K. (1962)
Toxicological properties of several commercially available
surfactants. /. Soc. Cosmet. Chem. 13:469-476.
Organisation for Economic Cooperation and Development. "Acute
Eye Irritation/Corrosion." OECD Publications and Information
Center, Suite 1207, 1750 Pennsylvania Ave., N.W., Washington,
D.C. 20006.
Pfister, R.R. and Burstein, N. (1976) The effects of ophthalmic drugs,
vehicles, and preservatives on corneal epithelium: a scanning
electron microscope study. Invest. Ophthalmol. 15(4)-.246-258.
Potts, A.M. and Gonasun, L.M. (1975) Toxicology of the eye. In:
Toxicology - The Basic Science of Poisons. Casarett, L.J. and
Doull, J. (eds.) Macmillan Publishing Co., Inc. New York, New
York. 275-309.
Rieger, M.M. and Battista, G.W. (1964) Some experiences in the safety
testing of cosmetics. /. Soc. Co\., Chem. 15:161-172.
Rowan, A.N. March 1980. The Draize Test: A critique and proposals
for alternatives. Institute for the study of animal problems.
Washington, D.C.
Rowan, A. (1981) The Draize test: Political and scientific issues.
Cosmetic Technology. 3(7):32-48.
Russell, K.L. and Hoch, S.G. (1962) Product development and rabbit
eye irritation. Proc. Sci. Sect. Toilet Goods Assoc. 37:27-32.
Sanderson, D.M. (1959) A note on glycerol formal as a solvent in
toxicity testing. /. Pharm. Pharmacol. 11:150-156.
Seabaugh, V.M.; Bayard, S.P.; Osterberg, R.E.; Porter, W.K.;
McCaulley, D.F.; Hoheisel, C.A.; Hehir, R.M.; Bierbower, G.W.
(1977) Detergent toxicity survey. American J. Public Health.
67(4):367-369.
Seabaugh, V.M.; Osterberg, R.E.; Hoheisel, C.A.; Murphy, J.C.;
Bierbower, G.W. (1976) A comparative study of rabbit ocular
reactions to various exposure times to chemicals. Soc. Toxicol.
15th Annu. Meet., Atlanta, Ga., March, paper 227.
Schuck, E.A.; Stephens, E.R.; Middleton, J.T. (1966) Eye irritation
response at low concentrations of irritants. Arch. Environ. Health.
13:570-575.
70
-------
Shanahan, R.W. and Ward, C.O. (1975) An animal model for
estimating the relative sting potential of shampoos. J. Soc.
Cosmet. Chem. 26:581-592.
Shapiro, H. (1956) Swelling and dissolution of the rabbit cornea in
alkali. Am. J. Ophthalmol. 42:292-298.
Simons, P.J. (1980) An alternative to the Draize test. In: The use of
alternatives in drug research. Rowan, A.N. and Stratmann, C.J.
(eds.) pp. 147-151, University Park Press, Baltimore, Md.
Smith, J.L. and Mickatavage, R.C. (1963) The ocular effects of topical
digitalis. Am. J. Ophthal. 56:889-894.
Smyth, H.F.; Carpenter, C.P.; Weil, C.S.; Pozzani, U.C.; Striegel, J.A.
(1962) Range-finding toxicity data: List VI. Amer. Indust. Hyg.
Assoc. J. 23:95-107.
Sugar, A. (1979) Clinical specular microscopy. Surv. Ophthalmol.
24(l):21-32.
Sunshine, I. (ed.) (1969) Handbook of Analytical Toxicology. Sec. A,
Vol. 1. The Chemical Rubber Co., Cleveland, Ohio. 585-670.
Swan, K.C. and White, N.G. (1942) Corneal permeability. I. Factors
affecting penetration of drugs into the cornea. Am. J. Ophthalmol.
25(9): 1043-1058.
Szary, A. (1977) The course of carrageenin-induced iridocyclitis in
rabbits. Arch. Immunol. Ther. Exp. 25(6):885-890.
Trolle-Lassen, C. (1958) Investigations into the sensitivity of the human
eye to hypo- and hypertonic solutions as well as solutions with
unphysiological hydrogen ion concentrations. Pharmaceutish
Weekblad. 93:148-155.
Tonjum, A.M. (1975) Effects of benzalkonium chloride upon the
corneal epithelium studied with scanning electron microscopy.
Acta. Ophthalmologica. 53:358-366.
Ulsamer, A.G.; Wright, P.L.; Osterberg, R.E. (1977) A comparison of
the effects of model irritants on anesthetized and nonanesthetized
rabbit eyes. Society of Toxicology, 16th Annual Meeting, Abs.
143.
U.S. Department of Health and Human Services (1981) "Trends in
Bioassay Methodology: In Vivo, In Vitro and Mathematical
Approaches". Initial Summary. Public Health Service, National
Institutes of Health.
Van Abbe, N.J. (1973) Eye irritation: studies relating to responses in
man and laboratory animals. J. Soc. Cosmet. Chem. 24:685-692.
71
-------
Walton, R.M. and Heywood, R. (1978) Applanation tonometry in the
assessment of eye irritation. J. Soc. Cosmet. Chem. 29(6):365-368.
Weil, C.S. and Scala, R.A. (1971) Study of intra- and interlaboratory
variability in the results of rabbit eye and skin irritation tests.
Toxicol. Appl. Pharmacol. 19:276-360.
Weltman, A.S.; Sharber, S.B.; Jurtshuk, T. (1965) Comparative
evaluation and the influence of various factors on eye irritation
scores. Toxicol. Appl. Pharmacol. 7:308-319.
Wright, P.L.; Ulsamer, A.G.; Osterberg, R.E. (1976) Effects of model
eye irritants on corneal proteins and water content. Society of
Toxicology, 15th Annual Meeting, Abs. 226.
72
-------
APPENDIX A
Table A. Scale of Weighted Scores for Grading the Severity of Ocular Lesions
(Draize et al., 1944)
I. Cornea
A. Opacity-Degree of Density (area which is most dense is taken
for reading)
Scattered or diffuse area—details of iris clearly visible 1
Easily discernible translucent areas, details of iris clearly visible.. 2
Opalescent areas, no details of iris visible, size of pupil barely
discernible 3
Opaque, iris invisible 4
B. Area of Cornea Involved
One quarter (or less) but not zero 1
Greater than one quarter—less than one half 2
Greater than one half—less than three quarters 3
Greater than three quarters—up to whole area 4
Score equals A x B * 5 Total maximum = 80
II. Iris
A. Values
Folds above normal, congestion, swelling, circumcorneal injec-
tion (any one or all of these or combination of any thereof),
iris still reacting to light (sluggish reaction is positive) 1
No reaction to light, hemorrhage; gross destruction (any one or
all of these) 2
Score equals A x 5 Total possible maximum = 10
III. Conjunctivae
A. Redness (refers to palpebral conjunctivae only)
Vessels definitely injected above normal 1
More diffuse, deeper crimson red, individual vessels not easily
discernible 2
Diffuse beefy red 3
73
-------
B. Chemosis
Any swelling above normal (includes nictitating membrane) 1
Obvious swelling with partial eversion of the lids 2
Swelling with lids about half closed 3
Swelling with lids about half closed to completely closed 4
C. Discharge
Any amount different from normal (does not include small
amounts observed in inner canthus of normal animals) 1
Discharge with moistening of the lids and hairs just adjacent to
the lids 2
Discharge with moistening of the lids and considerable area
around the eye 3
Score (A + B + C) x 2 Total maximum = 20
The maximum total score is the sum of all scores obtained for the cornea, iris and
conjunctivae.
74
-------
APPENDIX B
TEST FOR EYE IRRITANTS (16 CFR 1500.42)
Six albino rabbits are used for each test substance. Animal facili-
ties for such procedures shall be so designed and maintained as to
exclude sawdust, wood chips, or other extraneous materials that might
produce eye irritation. Both eyes of each animal in the test group shall
be examined before testing, and only those animals without eye defects
or irritation shall be used. The animal is held firmly but gently until
quiet. The test material is placed in one eye of each animal by gently
pulling the lower lid away from the eyeball to form a cup into which the
test substance is dropped. The lids are then gently held together for one
second and the animal is released. The other eye, remaining untreated,
serves as a control. For testing liquids, 0.1 milliliter is used. For solids or
pastes, 100 milligrams of the test substance is used, except that for
substances in flake, granule, powder, or other paniculate form the
amount that has a volume of 0.1 milliliter (after compacting as much as
possible without crushing or altering the individual particles, such as by
tapping the measuring container) shall be used whenever this volume
weighs less than 100 milligrams. In such a case, the weight of the 0.1
milliliter test dose should be recorded. The eyes are not washed follow-
ing instillation of test material except as noted below.
The eyes are examined and the grade of ocular reaction is recorded
at 24, 48, and 72 hours. Reading of reactions is facilitated by use of a
binocular loupe, hand slit-lamp, or other expert means. After the
recording of observations at 24 hours, any or all eyes may be further
examined after applying fluorescein. For this optional test, one drop of
fluorescein sodium opthalmic solution U.S.P, or equivalent is dropped
directly on the cornea. After flushing out the excess fluorescein with
sodium chloride solution U.S.P. or equivalent, injured areas of the
cornea appear yellow; this is best visualized in a darkened room under
ultraviolet illumination. Any or all eyes may be washed with sodium
chloride solution U.S.P. or equivalent after the 24-hour reading.
An animal shall be considered as exhibiting a positive reaction if
the test substance produces at any of the readings ulceration of the
cornea (other than a fine stippling), or opacity of the cornea (other than
a slight dulling of the normal luster), or inflammation of the iris (other
than slight deepening of the folds (or rugae) or a slight circumcorneal
injection of the blood vessels), or if such substance produces in the
conjunctivae (excluding the cornea and iris) an obvious swelling with
partial eversion of the lids or a diffuse crimson-red with individual
vessels not easily discernible.
75
-------
The test shall be considered positive if four or more of the animals
in the test group exhibit a positive reaction. If only one animal exhibits a
positive reaction, the test shall be regarded as negative. If two or three
animals exhibit a positive reaction, the test is repeated using a different
group of six animals. The second test shall be considered positive if
three or more of the animals exhibit a positive reaction. If only one or
two animals in the second test exhibit a positive reaction, the test shall
be repeated with a different group of six animals. Should a third test be
needed, the substance will be regarded as an irritant if any animal
exhibits a positive response.
GRADING (CPSC, 1976)
When grading by the official method, each structure of the eye is
considered independently. Damage is scored on the basis of the inten-
sity of response of the most severely affected portion; the area of
involvement is not taken into consideration (Table B). However, for
purposes of illustration, the plates in this appendix are not necessarily
graded on the most severely affected portion.
For completeness in judging the severity of the reaction caused by
the test material, corneal lesions such as necrosis (death of tissue),
pannus (subepithelial proliferation, pigmentation and accompanying
vascularization of the cornea), and corneal bulging should be noted. If
any of these conditions are observed during the test period, severe
ocular damage is indicated.
Although not part of the official method, grading the eyes at 1 hour
and at 7, 14, and 21 days after instillation of the test substance is often
helpful in judging the severity of the reaction. (On Plate 6 note the 4
chemosis at 1 hour and 4 opacity at 7 days which would not have been
observed during the official test period.)
As described in 16 CFR 1500.42, an optional method for reading
reactions is to place one drop of fluorescein sodium ophthalmic solu-
tion (U.S.P. or equivalent) in the eye. After the excess fluorescein is
flushed out with sodium chloride solution (U.S.P. or equivalent),
injured areas of the cornea appear yellow. This is best visualized in a
darkened room under ultraviolet illumination. Plate 3 illustrates ocular
damage as evidenced by fluorescein staining.
76
-------
Table B. Grades for Ocular Lesions (CPSC, 1976)
Cornea Grade (a)
No ulceration or opacity 0
Scattered or diffuse areas of
opacity (other than slight dul-
ling of normal luster), details
of iris clearly visible 1 *
Easily discernible translucent
areas, details of iris slightly
obscured 2
Nacreous areas, no details of
iris visible, size of pupil barely
discernible
Complete corneal opacity, iris
not discernible 4
Iris
Normal 0
Markedly deepened folds, con-
gestion, swelling, moderate cir-
cumcorneal injection (any of
these separately or combined);
iris still reacting to light
(sluggish reaction is positive) .. 1*
No reaction to light, hemor-
rhage, gross destruction (any
or all of these) 2
Conjunctive Grade (a)
Redness (refers to palpebral
and bulbar conjunctivae
excluding cornea and iris)
Vessels normal 0
Some vessels definitely injected . 1
Diffuse, crimson red, individ-
ual vessels not easily discern-
ible 2*
Diffuse, beefy red 3
Chemosis
No swelling 0
Any swelling above normal
(includes nictitating mem-
brane)
1
Obvious swelling with partial
eversion of lids 2*
Swelling of lids about half
closed 3
Swelling with lids more than
half closed 4
(a) Asterisks indicate lowest grade considered positive.
77
-------
PLATE 1
A normal eye and grades 1 through 3 redness
Normal Eye
1 Redness
2 Redness
3 Redness
78
-------
PLATE 2
Grades 1 through 4 cornea! opacity
1 Opacity
2 Opacity
3 Opacity
4 Opacity
79
-------
PLATE 3
The same eyes as in Plate 2 stained with fluorescein and shown under
ultraviolet illumination
1 Opacity
2 Opacity
3 Opacity
4 Opacity
80
-------
PLATE 4
Grades 1 and 2 iritis
1 Iritis
2 Iritis
2 Iritis
81
-------
PLATES
Grades 1 through 4 chemosis
1 Chemosis
2 Chemosis
3 Chemosis
4 Chemosis
These photographs may not accurately represent chemosis because the eyes have
been held open to show other aspects of irritation.
82
-------
PLATE 6
The same eye prior to instillation through 1 hour, 24 hours, 48 hours, 72 hours,
and 7 days after product instillation. Each eye is scored for redness, opacity,
iritis, and chemosis.
Normal Eye
1 Hour
2-3 Redness
1 Iritis
>2 Opacity
4 Chemosis
1 Opacity
>3 Chemosis
48 Hours
3 Redness
2 Iritis
>1 Opacity
3 Chemosis
72 Hours
7 Days
3 Redness
2 Iritis
> 1 Opacity
> 2 Chemosis
3 Redness
2 Iritis
4 Opacity
2 Chemosis
83
-------
REPORT DOCUMENTATION 1, REPORT NO.
PAGE EPA 560/11-82-001
EYE IRRITATION TESTING: An Assessment of Methods and Guidelines
for Testing Materials for Eye Irritancy
1981
Falahee. K.J.. Rose. C.S., Olin. S.S., Seifried. H.E.
9. Performing Organiz
Tracor Jitco, Inc.
1776 East Jefferson Street
Rockville, MD 20852-4081
10. Project/Task/Work Unit No.
Work Assignment 02
11. Contract(C) or Grant(G) No.
to 68-01-6176
(G)
Office of Toxic Substances
U.S. Environmental Protection Agency
Washington, D.C. 20460
13. Type of Report & Period Cove
Final Technical
Report
15. Supplementary Note;
16. Abstract {Limit- 200 words)
The ocular safety of materials is determined primarily by observing the irritation
produced by test agents instilled directly into the rabbit eye. The object of this study
was to review the scientific basis for published guidelines, particularly those recently
developed by the Organization for Economic Cooperation and Development (OECD) and the
Interagency Regulatory Liaison Group (IRLG). These guidelines are essentially the same
and recommend instillation of 0.1 ml material into the rabbit eye with observation for at
least 72 hr. To increase cost-effectiveness and optimize use of animals, the guidelines
reduce the number of animals and permit exclusions based on very high (>_ 11.5) or low
(<_ 2) pH, and demonstrated dermal irritation. This is based on the high probability that
agents meeting these criteria will be severe eye irritants. Data reviewed support
excluding strong alkalies (pH >_ 11.5). Exclusions of acids below pH 2 is not as well sup-
ported. Data indicate that the majority of severe dermal irritants would also be eye
irritants. The monkey appears the most predictive animal model for human response; howevei
the rabbit still remains the species of choice due to practical considerations and the
large data base which is useful for comparative purposes. Generally greater sensitivity oi
the rabbit eye compared to the human allows a conservative extrapolation to human risk-
The development of alternative methodologies (e.g. in vitro tests) and the use of topical
anesthetics is discussed. A tier strategy to eye irritation testing is proposed which
screens substances on the basis of pH, dermal irritation and results from other tests.
17. Document Analysis a. Descriptor
Eye; Toxicology; Bioassay; Evaluation; Experimental•Design; Toxic Tolerances;
Laboratory Animals; In Vivo Analysis; In Vitro Analysis; Assessments
o. Identifiers/Open-Ended Terms
Eye Irritation Testing Methods; Draize Method; Eye Irritants; Rabbits; Tier
Testing; OECD Testing Guidelines; IRLG Testing Guidelines.
COSATI Field/Group
18. Availability Statement
Document is available to the public through the
National Technical Information Service,
VA 991 SI
20. Security Class {This Page)
21. No. of Pages
100
(See ANSI-Z39.18)
OPTIONAL FORM 272 (4-77)
(Formerly NTIS-35)
84
------- |