Regulatory Toxicology and Pharmacology 114 (2020) 104666

Contents lists available at ScienceDirect

Regulatory Toxicology and Pharmacology

journal homepage: www.elsevier.com/locate/yrtph

Acute toxicity “six-pack” studies supporting approved new drug applications T

in the U.S., 2015–2018
Joseph Manuppello∗, Kristie Sullivan, Elizabeth Baker
Physicians Committee for Responsible Medicine, 5100 Wisconsin Ave., NW, Suite 400, Washington, DC, 20016-4131, USA

A R T I C LE I N FO A B S T R A C T

Keywords: The acute toxicity “six-pack” is a battery of animal tests used to evaluate acute systemic toxicity by three routes
Toxicity tests of exposure, skin and eye irritation/corrosion, and skin sensitization. A perception exists that these tests are not
Acute required for pharmaceuticals. For the four years from 2015 through 2018, we tallied the number of corre-
Animal testing alternatives sponding tests submitted by sponsors in approved, original new drug applications, as reported by the U.S. Food
Animal use alternatives
and Drug Administration (FDA) in publicly available reviews. In 125 reviews, we identified 228 single dose acute
United States Food and drug administration
Investigational new drug application
toxicity studies, 62 in vivo local tolerance studies, and 32 in vivo skin sensitization studies, as well as 37 in vitro or
Lethal dose 50 ex vivo local tolerance studies. A total of 4798 animals were used in these studies; however, FDA's reporting was
Local lymph node assay inconsistent, and we estimate the actual number of animals used to be 8998. For the evaluation of single dose
New approach methodology (NAM) acute toxicity, we accessed two guidance documents with conflicting recommendations regarding routes of
In silico administration and number of species to be used. For the evaluation of local tolerance and skin sensitization,
most studies examined were conducted by routes other than that intended for human administration. Non-
animal methods used to evaluate skin sensitization were not reported.

1. Introduction 60 and 120 animals per test substance, was unnecessarily precise (FDA,
1988).
Acute toxicity is the toxicity produced by a substance when it is More recently, in its Predictive Toxicology Roadmap, FDA established
administered in one or more doses during a period not exceeding 24 h a comprehensive strategy to evaluate new methodologies and tech-
(FDA CDER, 1996). Animals used in acute toxicity tests may suffer se- nologies for their potential both to expand predictive toxicology cap-
vere pain and distress, as well as mortality, at the high dose levels abilities and to reduce the use of animals in the assessment of regulated
administered. A battery of animal tests that evaluates acute systemic products (FDA, 2017). Some FDA centers have also made efforts, either
toxicity by the oral, topical, and inhalation routes of exposure, as well internally or in collaboration with agency or industry partners, to de-
as skin and eye irritation/corrosion, and skin sensitization, is commonly velop, validate, and implement such methodologies for specific end-
referred to as the “six-pack”. Efforts to reduce the use of animals in points. For example, in its guidance for industry and review staff,
these tests have focused on their applications in the safety assessment of Nonclinical Safety Evaluation of Reformulated Drug Products and Products
agrochemicals and industrial chemicals. For the nonclinical assessment Intended for Administration by an Alternate Route, FDA's Center for Drug
of pharmaceuticals, a perception exists that these tests are generally not Evaluation and Research (CDER) recommends that reformulated topical
required and only rarely conducted in specific cases (Strickland et al., drug products be evaluated for eye irritation/corrosion potential by
2018; Prior et al., 2019). The U.S. Food and Drug Administration (FDA) appropriate in vitro or ex vivo methods (FDA CDER, 2015). As discussed
ceased to require the “classical” LD50 (median lethal dose) test more below, such New Approach Methodologies (NAMs) are also available
than three decades ago, recognizing that the test, which used between for the evaluation of skin irritation/corrosion and skin sensitization. In

Abbreviations: BCOP, Bovine Corneal Opacity and Permeability; CBER, Center for Biologics Evaluation and Research; CDER, Center for Drug Evaluation and
Research; EPA, U.S. Environmental Protection Agency; FDA, U.S. Food and Drug Administration; GPMT, guinea pig maximization test; ICCVAM, Interagency
Coordinating Committee on the Validation of Alternative Methods; ICH, International Conference on Harmonisation; IDE, Investigational Device Exemptions; IND,
Investigational New Drug; LD50, median lethal dose; LLNA, Local Lymph Node Assay; LOEL, Lowest Observed Effect Level; MLD, Minimum Lethal Dose; MNL,
Maximum Nonlethal Dose; NAM, New Approach Methodology; NDA, New Drug Application; NOAEL, No Observed Adverse Effect Level; OECD, Organisation for
Economic Cooperation and Development
∗
Corresponding author.
E-mail address: [email protected] (J. Manuppello).

https://doi.org/10.1016/j.yrtph.2020.104666
Received 7 February 2020; Received in revised form 13 April 2020; Accepted 17 April 2020
Available online 23 April 2020
0273-2300/ © 2020 The Authors. Published by Elsevier Inc. This is an open access article under the CC BY license
(http://creativecommons.org/licenses/BY/4.0/).
J. Manuppello, et al. Regulatory Toxicology and Pharmacology 114 (2020) 104666

addition, FDA participates in the Interagency Coordinating Committee For each review in which multiple studies of the same type were
on the Validation of Alternative Methods (ICCVAM), which recently reported, we noted whether these studies differed in the test article,
published A Strategic Roadmap for Establishing New Approaches to Eval- species of animal, route of exposure, or experimental protocol used and
uate the Safety of Chemicals and Medical Products in the United States copied any relevant descriptive text to a comment in the workbook
(ICCVAM, 2018). (Supplementary File, "Multiples" sheet and individual review sheets).
In order to measure any reduction in animal use, the number of On FDA's Search for FDA Guidance Documents web page (FDA,
animals used must first be determined; however, this can be hindered 2020b), we filtered by product (Drugs) and topic (Pharm/Tox) and
by the limited ability to quantify animals used for toxicity testing in the downloaded all guidance documents returned. We also searched for and
U.S. (ICCVAM, 2018). For pharmaceuticals, FDA's reviews of approved downloaded any additional guidance documents listed in CDER's Re-
new drug applications (NDAs), which are publicly available, can pro- search List of Guidance Documents under the Pharmacology/Toxicology
vide insight into the types of studies that sponsors conduct and the heading (FDA, 2020c). We reviewed relevant general guidance docu-
numbers and species of animals they use. For the four-year period from ments in detail and searched all guidance documents for key terms
2015 through 2018, we tallied the single dose acute (systemic) toxicity, (single dose, acute, local tolerance, irritation, sensitization, and sensi-
local tolerance (eye, skin, and vascular irritation/corrosion), and skin tivity).
sensitization studies that were submitted by sponsors in approved,
original NDAs, as reported by FDA in its reviews. For each study, we 3. Results
also collected additional information, as described below, including the
number and species of animals used, if reported. We then compared our For the four-year period from 2015 through 2018, we identified 125
results to the information needs and testing recommendations identi- FDA reviews of approved, original NDAs. Studies that were submitted
fied by FDA in its guidance to industry. Our results challenge the per- by sponsors, but not reviewed by FDA, were listed in 38 (30.4%) of
ception that the acute toxicity “six pack” tests are only rarely conducted these reviews (Supplementary File, “List” sheet).
for the nonclinical assessment of pharmaceuticals and identify needs for
clear guidance to industry and consistent reporting by FDA. 3.1. Single dose acute toxicity

2. Materials and methods In the 125 reviews examined, we identified 228 single dose acute
toxicity studies, 1.8 studies per review, on average. In 61 reviews, we
On its Drugs@FDA: FDA-Approved Drugs web page (FDA, 2020a), we identified no such studies; in 15, we identified one study; and in 49, we
searched FDA's reviews of approved, original NDAs by month from identified two through 12 studies. Twenty-two of these studies were
January 2015 through December 2018. In the tables returned, for “Type listed as submitted by sponsors but not reviewed by FDA
1" (new molecular entity) entries, we noted the approval date, drug (Supplementary File, “List” sheet).
name, and company name in a Microsoft Excel workbook (Supple- Oral administration was used in 112 (49.1%) of these studies, in-
mentary File). We accessed each Type 1 entry and noted the route in- travenous in 73 (32.0%), subcutaneous in 23 (10.1%), inhalation in
tended for human administration. From the “Approval Date(s) and eight (3.5%), intraperitoneal in seven (3.0%), topical and intramuscular
History, Letters, Labels, Reviews” table, we accessed “Reviews”, and in in two each (0.9%), and intrathecal in one (0.4%). Rats were used in
the drug approval package, we noted the availability of a Pharma- 112 studies (49.1%), mice in 56 (24.6%), dogs in 39 (17.1%), monkeys
cology/Toxicology Review either as a separate document or as a sum- in 19 (8.3%), and rabbits and cats in one each (0.4%). In addition, 163
mary in a Multi-Discipline Review. We downloaded each of these re- studies (71.5%) were conducted by the route intended for human ad-
views and, for any review including scanned documents, recognized ministration, while 65 studies (28.5%) were conducted by other routes
text using Adobe Acrobat. (Supplementary File, “Summary” sheet).
Most Pharmacology/Toxicology Reviews available as separate Fig. 1 shows the number of single dose acute toxicity studies con-
documents included a “Studies Submitted” section with a “Studies Not ducted by species and route of administration as well as by routes in-
Reviewed” subsection; we noted whether studies that were submitted tended or unintended for human administration. Rats were used more
but not reviewed were listed. If such studies were not listed, we copied frequently than any other species in studies conducted by all routes,
any explanatory text to a comment in the workbook (Supplementary except for intrathecal; monkeys were used in the only study conducted
File, “List” sheet). Pharmacology/Toxicology Reviews summarized by this route. Mice were used the next most frequently in studies by the
within Multi-Discipline Reviews did not include a “Studies Submitted” intravenous, subcutaneous, intraperitoneal and inhalation routes; mice
section, and studies not reviewed were not listed. and dogs were used the same number of times in studies by the oral
For each animal study listed, we noted the type of study (single dose route. Dogs were used the next most frequently in studies by the in-
acute toxicity, local tolerance, or skin sensitization), whether the study travenous, subcutaneous, and inhalation routes. Monkeys were used the
was not reviewed by FDA, the route of administration, and the species next most frequently in studies by the oral and intravenous routes. In
and number of animals used, if reported. addition, more studies were conducted by the route intended for human
In addition, for single dose acute toxicity studies, we noted whether administration in all species except rabbits; rabbits were used in only
each study was described as “dose range finding,” “dose escalation,” or one study, which was by the topical route.
similar and whether reporting included: For the evaluation of acute toxicity, we accessed two relevant
general guidance documents, Single Dose Acute Toxicity Testing for
• no observed adverse effect level (NOAEL) or lowest observed effect Pharmaceuticals (FDA CDER, 1996) and M3(R2) Nonclinical Safety Stu-
level (LOEL); dies for the Conduct of Human Clinical Trials and Marketing Authorization
• minimum lethal dose (MLD) or maximum nonlethal dose (MNL); for Pharmaceuticals (FDA CDER and FDA CBER, 2010). FDA CDER notes
• test substance-related lethality; and that its 1996 guidance was originally published as part of a proposed
• target organ(s) of toxicity. implementation document entitled U.S. FDA's Proposed Implementation
of International Conference on Harmonisation (ICH) Safety Working Group
For local tolerance studies, which include those for the evaluation of Consensus Regarding New Drug Applications and that its approach is
skin and eye irritation/corrosion, we noted the type of irritation eval- considered to be in general agreement with the ICH position on acute
uated and, for in vitro or ex vivo studies, the method used. For skin toxicity testing. FDA CDER and FDA CBER (Center for Biologics Eva-
sensitization studies, we noted whether each study was a murine local luation and Research) note that the 2010 guidance is a revision of the
lymph node assay (LLNA) or a guinea pig maximization test (GPMT). ICH M3 guidance originally published in 1997 and that the revisions

2
J. Manuppello, et al. Regulatory Toxicology and Pharmacology 114 (2020) 104666

Fig. 1. The number of single dose toxicity studies reported by each route of administration and by routes intended or unintended for human administration, as well as
by species. “Other routes” combines studies conducted by topical, intramuscular, and intrathecal routes of administration.

further harmonize recommendations in a number of areas. As the re- studies conducted by the intended route were reported, and this route
commendations cited below are identical in both the FDA (FDA CDER was not intravenous, additional studies by the intravenous route were
and FDA CBER, 2010) and ICH (ICH, 2009) versions of this guidance, reported in 14 reviews (29.8%). Overall, studies conducted by only one
we will refer to the latter for clarity. route were reported in 45 reviews (70.3%), while studies conducted by
In its 1996 guidance, FDA CDER states that the information ob- more than one route were reported in 19 reviews (29.7%). In addition,
tained from acute toxicity studies is useful in choosing doses for re- studies conducted in only one species were reported in 15 reviews
peated dose studies and for providing preliminary identification of (23.4%), while studies conducted in more than one species were re-
target organs of toxicity. 34 single dose acute toxicity studies (14.9%) ported in 49 reviews (76.6%; Supplementary File, “Summary” sheet).
were described as “dose range finding” or “dose escalation.” Target With regard to animal use, in its 1996 guidance, FDA CDER notes
organs of toxicity were identified in 8 studies (3.5%; Supplementary that studies should be designed so that the maximum amount of in-
File, “Summary” sheet). While FDA does not recommend specific test formation is obtained from the smallest number of animals (FDA CDER,
guidelines for evaluating acute toxicity, it does recommend that the test 1996). In the 141 studies for which animal numbers were reported,
substance be administered to animals so as to identify doses causing no 4279 animals were used, an average of 30 animals per study across
adverse effects and doses causing major (life-threatening) toxicity. species (Supplementary File, “Summary” sheet). Table 1 shows the total
NOAELs or LOELs were reported for 38 single dose acute toxicity stu- and average (per study) number of animals used, by species, as well as
dies (16.7%), while MLDs or MNLs were reported for 54 studies (23.7%; the estimated total number of animals used when these average num-
Supplementary File, “Summary” sheet). In its 2009 guidance, ICH states bers are applied to studies for which the number of animals used was
that lethality should not be an intended endpoint in studies assessing not reported. Extrapolation raises the estimated total to 7034 animals.
acute toxicity (ICH, 2009). Substance-related lethality was reported in
63 studies (27.6%; Supplementary File, “Summary” sheet). 3.2. Local tolerance
In its 1996 guidance, FDA CDER recommends that acute toxicity
studies in animals ordinarily be conducted using two routes of drug In the 125 reviews examined, we identified 62 local tolerance stu-
administration, the route intended for human administration and in- dies, 0.5 studies per review, on average. In 87 reviews, we identified no
travenous, if feasible, as well as in at least two mammalian species, such studies; in 23, we identified one study; and in 15, we identified
including a nonrodent species, when reasonable (FDA CDER, 1996). two through five studies. 12 of these studies were listed as submitted by
However, in its 2009 guidance, ICH recommends that acute toxicity sponsors but not reviewed by FDA (Supplementary File, “Summary”
information be obtained from appropriately conducted dose-escalation sheet).
studies or short-duration dose-ranging studies, and that when this in- Eye irritation was evaluated in nine of these studies (14.5%), skin
formation is available from any study, separate single-dose studies are irritation in 18 (29.0%), and vascular irritation in 35 (56.5%). In ad-
not recommended. In addition, ICH notes that while historically, acute dition, 21 studies (33.9%) were conducted by the route intended for
toxicity information has been obtained from studies in two mammalian
species, using both the clinical and a parenteral route of administration,
1
it now recommends that studies providing this information be limited Of the remaining eight reviews in which toxicity was reported to be eval-
to the general toxicity test species and to the route intended for human uated only by a route other than that intended for human administration, the
intended route was ophthalmic in three, and toxicity was evaluated by the
administration only (ICH, 2009).
intravenous route in each. In another three of these reviews, the intended route
Of the 64 reviews in which single dose acute toxicity studies were
was oral, and toxicity was evaluated by the intravenous route in two and by the
reported, 56 (87.5%) included studies conducted by the route intended topical route in one. In one of these reviews, the intended route was vaginal,
for human administration,1 and 27 (42.2%) included studies by a dif- and toxicity was evaluated by the subcutaneous route. In the last of these re-
ferent route. The intended route was the only route reported to have views, the intended route was topical, and toxicity was evaluated by the in-
been used in 37 (57.8%) of these reviews. Of the 47 reviews in which travenous route.

3
J. Manuppello, et al. Regulatory Toxicology and Pharmacology 114 (2020) 104666

Table 1 epithelium model EpiOcular (OECD, 2019a) in one study, and with
Reported animal use in single dose acute toxicity studies. human corneal epithelial cells in one study. Skin irritation was eval-
Studies Animal Animal Studies Not Animal Nos. uated with reconstructed human epidermis models in seven studies
Reporting Nos. Total, Nos. Per Reporting Total, with EpiDerm™ and six with EpiSkin™ (OECD, 2014). Mucous mem-
Animal Nos. Reported Study, Animal Nos. Extrapolated brane irritation was evaluated in one hen's egg chorioallantoic mem-
Average brane test (Supplementary File, “Summary” sheet).
Mice 36 1348 37 20 2097
Rats 65 2418 37 47 4166 3.3. Skin sensitization
Dogs 24 335 14 15 544
Monkeys 14 136 10 5 185 In the 125 reviews examined, we identified 32 skin sensitization
Rabbits 1 10 10 0 10
studies, 0.3 studies per review, on average. In 101 reviews, we identi-
Cats 1 32 32 0 32
fied no such studies; in 18, we identified one study; and in six, we
identified two through four. Eight of these studies were listed as sub-
human administration, while 41 studies (66.1%) were conducted by mitted by sponsors but not reviewed by FDA. Skin sensitization was
other routes (Supplementary File, “Summary” sheet).2 evaluated in 28 murine LLNAs, and four GPMTs (Supplementary File,
Fig. 2 shows the number of eye, skin, and vascular irritation studies “Summary” sheet).
by species as well as by routes intended or unintended for human ad- We found no general guidance for the evaluation of skin sensitiza-
ministration. Rabbits were used more frequently than any other species tion; however, in Immunotoxicology Evaluation of Investigational New
for all study types. In all of the local tolerance studies in which species Drugs, FDA CDER recommends that when a drug is intended for topical
other than rabbits were used, vascular irritation was evaluated. In ad- administration, its sensitizing potential be evaluated using an appro-
dition, more studies were conducted by routes other than that intended priate assay (FDA CDER, 2002). Only two of the 24 reviews (8.3%) in
for human administration in rabbits and rats; in all other species, stu- which the results of skin sensitization studies were reported are for
dies were conducted by the intended route only. drugs intended for topical administration (Supplementary File, “Sum-
For the evaluation of local tolerance, we accessed one relevant mary” sheet). FDA CDER identifies the most common methods for
general guidance document, M3(R2) Nonclinical Safety Studies for the evaluating the skin sensitizing potential of drugs to be the Buehler assay
Conduct of Human Clinical Trials and Marketing Authorization for and the GPMT, both of which it considers to be reliable and to have a
Pharmaceuticals, the same as that cited above for single dose acute high correlation with known human skin sensitizers. We found no re-
toxicity (FDA CDER and FDA CBER, 2010; ICH, 2009). ICH notes that it ports of Buehler assays in the reviews examined. FDA CDER also notes
is preferable to evaluate local tolerance by the route intended for that results obtained with the murine LLNA correlate well with tradi-
human administration as part of general toxicity studies, and stand- tional guinea pig tests, the results are quantitative rather than sub-
alone studies are generally not recommended (ICH, 2009). All of the jective, and CDER accepts this method for evaluating the skin sensi-
local tolerance studies identified were stand-alone studies.3 In addition, tizing potential of drugs intended for topical use.
ICH recommends that, to support limited human administration by In the six skin sensitization studies for which animal numbers were
nontherapeutic routes (for example, a single intravenous dose to assist reported, 215 animals were used, an average of 36 animals per study.
in the determination of absolute bioavailability of an oral drug), a local Extrapolation to studies for which the number of animals used was not
tolerance study in a single species is considered appropriate (ICH, reported raises the estimated total to 1147 animals (Supplementary
2009). File, “Summary” sheet).
Of the 38 reviews in which local tolerance studies were reported, 13
(34.2%) included studies in which irritation was evaluated by the route 4. Discussion
intended for human administration, and 29 (76.3%) included studies by
a different route. The intended route was the only route reported to We identified gaps in the information reported by FDA in many of
have been used in nine of these reviews (23.6%). Overall, studies its reviews of approved NDAs which likely resulted in underestimates of
conducted by only one route were reported in 30 reviews (78.9%), the number of single dose acute toxicity, local tolerance, and sensiti-
while studies conducted by more than one route were reported in eight zation studies submitted by sponsors as well as in the number of ani-
reviews (21.1%). In addition, studies conducted in only one species of mals used in these studies. FDA listed studies that were submitted but
animal were reported in 34 (89.5%) reviews, while studies conducted in not reviewed in only 30% of its reviews; if such studies were not listed,
more than one species were reported in four reviews (10.5%; we copied any explanatory text to a comment in the Microsoft Excel
Supplementary File, “Summary” sheet). workbook (Supplementary File, “List” sheet). In many cases, this text is
In the 22 local tolerance studies for which animal numbers were simply “Studies Not Reviewed: None” or similar. While it is possible
reported, 304 animals were used, an average of 14 animals per study. that in some cases, this indicates that all submitted studies were re-
Extrapolation to studies for which the number of animals used was not viewed, there are reasons to doubt this interpretation. Instead, it may
reported raises the estimated total to 817 animals (Supplementary File, indicate only that all studies judged to be “pivotal” were reviewed.
“Summary” sheet). For example, among the explanations for why studies were not re-
Local tolerance was evaluated by in vitro or ex vivo methods in 23 viewed, we find:
studies for eye irritation, 13 studies for skin irritation, and one study for
mucous membrane irritation. Eye irritation was evaluated with the • “Studies considered irrelevant for the nonclinical safety assessment
bovine corneal opacity and permeability (BCOP) test method (OECD, were not reviewed; "
2017) in 21 studies, with the reconstructed human cornea-like • “Studies Briefly/Not Reviewed: Selected ADME studies; studies by
IV or IP route of administration; several non-pivotal studies; "
• “Preliminary dose-range finding studies and non-pivotal studies for
2 deciding nonclinical safety … were not reviewed in detail; "
• “Dose range finding studies and non-pivotal studies were not re-
If the route intended for human administration was intravenous, sub-
cutaneous, intramuscular, or intraperitoneal, any vascular irritation study was
considered to be by the intended route. viewed; "
3
We did not note when local tolerance was evaluated as part of general • “Dose-ranging and non-GLP studies were examined, but not for-
toxicity studies, because this would have required examining all general toxi- mally reviewed; "
city studies reviewed, which is outside the scope of this article. • “Dose range-finding studies have not been reviewed."
4
J. Manuppello, et al. Regulatory Toxicology and Pharmacology 114 (2020) 104666

Fig. 2. The number of local tolerance studies by type of irritation evaluated and by routes intended or unintended for human administration, as well as by species.

Fig. 3. Pharmacology/Toxicology Reviews summarized as a section within Multi-Discipline Reviews, or available as a separate document, in each month examined
beginning December 2016.

In addition, among the explanations for why reviews in which but not reviewed in those reviews in which such studies were not listed,
submitted but not reviewed studies were listed, we find: we estimate the number of submitted studies missed in this analysis to
be 50 single dose studies, 27 local tolerance studies, and 18 skin sen-
• “The following studies were not reviewed as they are non-pivotal, sitization studies.
less relevant or redundant at the current stage; " A total of 4798 rats, mice, dogs, monkeys, rabbits, pigs, guinea pigs,
• “The studies listed in the following table … were not reviewed, as minipigs, and cats were reported to have been used in the studies ex-
they were not pivotal to the nonclinical assessment …; " amined. However, the number of animals used was reported in only
• “The studies listed below were … not reviewed, as they were not 62% of single dose acute toxicity studies, 35% of local tolerance studies,
pivotal to the nonclinical assessment.…" (Supplementary File, “List” and 19% of skin sensitization studies. If, on average, the number of
sheet). animals used for each study type was similar in studies for which the
number of animals used was not reported, we estimate an additional
From these explanations, it seems clear that dose range finding 2755 animals were used for single dose studies, 513 animals for local
studies, in particular, were frequently submitted by sponsors but not tolerance studies, and 932 animals for skin sensitization studies, raising
reviewed by FDA. These studies are relevant to this analysis, because the estimated total number of animals used in the reviews examined to
FDA CDER includes choosing doses for repeated dose studies among the 8998.
useful information that can be obtained from acute toxicity studies, and In order to measure any reduction in the number of animals used in
ICH recommends that acute toxicity information be obtained from dose- the testing of new drugs, the number of animals used in that testing
escalation studies or dose-ranging studies. must be accurately determined. Unfortunately, FDA's reviews of ap-
Studies that were listed as submitted by sponsors but not reviewed proved NDAs became less helpful in this regard during the four years
by FDA included 22 single dose studies, 12 local tolerance studies, and examined. As noted above in Section 2.0, Pharmacology/Toxicology
eight sensitization studies. If similar numbers of studies were submitted Reviews summarized as a section within Multi-Discipline Reviews did

5
J. Manuppello, et al. Regulatory Toxicology and Pharmacology 114 (2020) 104666

not include a “Studies Submitted” section, and studies not reviewed studies by more than one route. In addition, of the reviews for which
were not listed in these reviews. The first Multi-Discipline Review we the intended route was not intravenous, 30% included studies by the
identified was for an NDA approved in December 2016, and the intravenous route in addition to the intended route. FDA also re-
availability of this type of review, rather than a Pharmacology/Tox- commends that acute toxicity studies be conducted in at least two
icology Review as a separate document, became more frequent over mammalian species. Of the reviews in which such studies were re-
time (Fig. 3). ported, 77% included studies in more than one species. These ob-
Finally, only reviews of approved NDAs are publicly available. servations suggest that many sponsors are following recommendations
While for the time period examined, we found 125 such reviews for new made by FDA CDER in its 1996 guidance rather than those made by ICH
molecular entities, according to FDA-TRACK, FDA's agency-wide per- in its 2009 guidance.
formance management system, FDA received 1970 original investiga- FDA CDER also notes that studies should be designed so that the
tional new drug (IND) applications during the same time period, nearly maximum amount of information is obtained from the smallest number
16 times the number of available NDA reviews (FDA, 2020d). There- of animals. FDA cites its LD50 test policy, in which it states that the
fore, it seems certain that the single dose acute toxicity, local tolerance, “classical” LD50 test requires from 60 to more than 120 animals per test
and skin sensitization studies examined represent only a small part of substance to attain a statistically precise median number (FDA, 1988).
all such studies conducted to support new drugs during this time period FDA recognized that this test was unnecessarily precise and ceased to
and that the number of animals used in the studies examined represents require it. We calculate the average number of both mice and rats used
a correspondingly small part of the total number of animals used. Based in those studies for which animal numbers were reported to be 37.
on this analysis, the total number of animals used in studies to support While this is lower than that required by the classical LD50 test, as
new drugs during this time period can be roughly estimated to be described by FDA, it is still a relatively large number compared to that
140,000. used to meet other regulatory requirements. For example, in its Health
These observations suggest a need for FDA to adopt a standard Effects Test Guidelines; OPPTS 870.1100; Acute Oral Toxicity, the U.S.
format for its reviews of approved NDAs that includes a list of all studies Environmental Protection Agency (EPA) recommends use of the Up-
submitted as well as the species and number of animals used in those and-Down Procedure for the evaluation of acute systemic toxicity, with
studies, preferably both for studies that were and were not reviewed. which for most applications, testing will be completed with only four
Ideally, such minimal animal use information would be made public not animals after the initial reversal in animal outcome, or between six and
only for reviews of approved NDAs but also for reviews of INDs. fifteen animals in total (EPA, 2002).4
We found little association between the recommendations made by We must emphasize that the two general guidance documents ac-
FDA CDER in its guidance to industry and the information reported in cessible for evaluating single dose acute toxicity include conflicting
reviewed studies. In its 1996 guidance, FDA CDER states that the in- recommendations, and many sponsors appear to be following re-
formation obtained from acute toxicity studies is useful in choosing commendations made in the older guidance. Clearly, the availability of
doses for repeated dose studies and in providing preliminary identifi- two conflicting guidance documents can only increase sponsors' un-
cation of target organs of toxicity. Further, FDA CDER recommends that certainty over FDA's expectations. We recommend that FDA CDER re-
the test substance be administered to animals so as to identify doses vise its 1996 guidance to be consistent with the more recent ICH gui-
causing no adverse effects and doses causing major (life-threatening) dance, in particular with regard to reducing the overall number of
toxicity. However, this information is reported for relatively few of the stand-alone studies by encouraging the use of acute toxicity information
studies examined. Only 15% of these studies were described as “dose obtained from other studies, using no more than one route of admin-
range finding” or “dose escalation” (as noted above, it seems certain istration and one species of animal if stand-alone studies are conducted,
that such studies were underreported); target organs of toxicity were and discouraging the inclusion of lethality as an intended endpoint.
identified in 4%; NOAELs or LOELs were reported in 17%; and MLDs or Alternatively, FDA CDER could simply retire its 1996 guidance in favor
MNLs were reported in 24%. This raises the question of why the ma- of the more recent ICH guidance.
jority of these studies were conducted, if not to provide the information FDA could further reduce the number of animals used for the eva-
that FDA CDER identifies as useful. Indeed, the more recent ICH gui- luation of acute toxicity by recommending the use of NAMs to meet its
dance states that when acute toxicity information is available from any regulatory needs. For example, Prieto et al. (2013) found that the in
study, separate single-dose studies are not recommended. This guidance vitro cytotoxicity test based on neutral red uptake by BALB/c 3T3 cells
also states that lethality should not be an intended endpoint in studies has high sensitivity, although the model does not perform well with
evaluating acute toxicity; however, test substance-related lethality was chemicals needing biotransformation or acting via specific mechanisms
reported in 28% of the studies examined (Supplementary File, including neurotoxicity or cardiotoxicity. Other organ-specific models
“Summary” sheet). These observations are consistent with those of may be used to refine this approach, including those based on HPCT-
Robinson et al. (2008) who found that acute toxicity data was generally 1E3 hepatocytoma cells (Kneuer et al., 2007), human BJ fibroblasts
not used to terminate drugs from development, to support dose selec- (Mannerström et al., 2017), and a variety of neuronal cells (Forsby
tion for repeat dose studies in animals, or to set doses for the first et al., 2009). Organ-specific toxicity is better investigated with human-
clinical trials. These authors concluded that acute toxicity studies are relevant in vitro or in silico models or a combination of both; cardio-
not needed prior to first clinical trials in humans (Robinson et al., toxicity is an area that has seen particularly fruitful work (e.g., Sirenko
2008), and ICH cites this article in its 2009 guidance. In addition, et al., 2017; Passini et al., 2017; Kurokawa et al., 2018). In addition,
Chapman et al. (2010) found that acute toxicity studies conducted later first-in-human microdosing studies can be used to determine kinetic
in drug development are not used for managing overdose of pharma- properties (Burt et al., 2018), and microphysiological systems show
ceuticals and are of little value in treating human poisoning from great promise in determining both toxicity and kinetics (e.g., Jang
chemicals. et al., 2019). Finally, ICCVAM recently organized a global collaboration
Oral and intravenous routes of administration were used more fre-
quently than other routes in the single dose acute studies examined,
accounting for 49% and 32% of studies, respectively. In its 1996 gui- 4
The test consists of a single ordered dose progression in which animals are
dance, FDA CDER recommends that acute toxicity studies in animals dosed, one at a time, at a minimum of 48-h intervals. The first animal receives a
ordinarily be conducted using two routes of administration, the route dose a step below the level of the best estimate of the median lethal dose. If the
intended for human administration and intravenous, if feasible, while animal survives, the dose for the next animal is increased by 3.2 times the
ICH recommends that such studies be limited to the intended route original dose; if it dies, the dose for the next animal is decreased by a similar
only. Of the reviews in which such studies were reported, 30% included dose progression.

6
J. Manuppello, et al. Regulatory Toxicology and Pharmacology 114 (2020) 104666

to build predictive in silico models for acute oral systemic toxicity based to sponsors that the evaluation of local tolerance by routes not intended
on a large dataset of rodent studies and targeted towards regulatory for human administration and the evaluation of skin sensitization for
needs identified across federal agencies. When the models were com- drugs not intended for topical administration are not recommended. In
bined, consensus predictions demonstrated excellent performance when addition, FDA should consider whether its recommendation that the eye
compared to the animal data (Kleinstreuer et al., 2018a). The Colla- irritation/corrosion potential of reformulated topical drug products be
borative Acute Toxicity Modeling Suite is available both as standalone evaluated by appropriate in vitro or ex vivo methods can be extended to
software and through EPA's CompTox Dashboard (EPA, 2020). other classes of drugs.
The only relevant general guidance accessible for the evaluation of While the significant acceptance of in vitro and ex vivo methods for
local tolerance was M3(R2) Nonclinical Safety Studies for the Conduct of the evaluation of eye and skin irritation/corrosion is encouraging, it is
Human Clinical Trials and Marketing Authorization for Pharmaceuticals disappointing that similar progress is not evident in the evaluation of
(ICH, 2009). ICH notes that it is preferable to evaluate local tolerance skin sensitization, especially considering the increasing acceptance of
by the route intended for human administration as part of general such methods in other sectors and the length of time for which NAMs
toxicity studies, and stand-alone studies are generally not re- have been available. Kleinstreuer et al. (2018a,b) used murine LLNA
commended. We found that the intended route was the only route used and human skin sensitization data to evaluate the performance of six
in just 24% of reviews in which local tolerance studies were reported. defined approaches, comprising eight non-animal testing strategies, for
While we did not note when local tolerance was evaluated as part of both hazard and potency characterization. Multiple non-animal testing
general toxicity studies, we identified a total of 62 stand-alone local strategies incorporating in vitro, in chemico, and in silico inputs de-
tolerance studies, raising concern that sponsors frequently overlook this monstrated equivalent or superior performance to the LLNA when
recommendation. ICH also recommends that, to support limited human compared to both animal and human data for skin sensitization
administration by nontherapeutic routes, a local tolerance study in a (Kleinstreuer et al., 2018b). In April 2018, EPA released a draft interim
single species is considered appropriate. While it is possible that this science policy, Use of Alternative Approaches for Skin Sensitization as a
exception accounts for the 41 studies conducted by routes other than Replacement for Laboratory Animal Testing (EPA, 2018). This draft policy
those intended for human administration, it seems unlikely that this is builds upon substantial work done by the Organization for Economic
the case. In addition, more than one species was reported to be used for Cooperation and Development (OECD) to establish the adverse outcome
the evaluation of local tolerance in four reviews (Supplementary File, pathway (AOP) for skin sensitization and to develop guidelines for
“Summary” sheet). several methods that assess the ability of chemicals to activate the first
In Nonclinical Safety Evaluation of Reformulated Drug Products and three key events in this AOP (OECD, 2018a; OECD, 2018b; OECD,
Products Intended for Administration by an Alternate Route, FDA CDER 2019b). Under this interim policy, EPA will accept submissions in
recommends that for reformulated topical drug products that contain which two defined approaches combining these methods are used. We
substances that have not been evaluated for eye irritation/corrosion recommend that FDA develop general guidance for the evaluation of
potential, that potential should be evaluated by appropriate in vitro or skin sensitization and consider whether a similar approach can be ap-
ex vivo methods (FDA CDER, 2015). Specifically, the in vivo rabbit plied to the testing of new drugs.
ocular irritation test method is no longer recommended. While this As noted in sections 3.1-3.3, multiple single dose studies were re-
recommendation is limited to a specific class of drug products, we ported in 49 reviews, multiple local tolerance studies in 15, and mul-
found that the BCOP test, an ex vivo method, was used more than twice tiple skin sensitization studies in nine. In many cases, these studies
as frequently as the rabbit test in the reviews examined. In addition, differed in the test article, route of administration, species of animal, or
almost as many skin irritation tests used reconstructed human epi- experimental protocol used (Supplementary File, “Multiples” sheet). In
dermis as used rabbits (Fig. 4). several cases, reporting was inadequate to determine how studies dif-
FDA addresses the evaluation of skin sensitization only in its gui- fered. By beginning with problem formulation and proceeding through
dance, Immunotoxicology Evaluation of Investigational New Drugs, in a decision-making framework that accounts for all available informa-
which it recommends that when a drug is intended for topical admin- tion and considers how NAMs can address information needs at each
istration, its sensitizing potential be evaluated using an appropriate step, Integrated Approaches to Testing and Assessment (IATA) can help
assay, such as the Buehler assay, the GPMT, or the murine LLNA (FDA sponsors avoid the conduct of such similar studies. In particular, in-
CDER, 2002). While the GPMT and murine LLNA were the only assays formation derived from read-across, quantitative structure–activity re-
reported in 24 reviews, only two of these reviews were for drugs in- lationship models, and in vitro or ex vivo methods can supplement
tended for topical administration (Supplementary File, “Summary” available information, minimizing the conduct of in vivo studies.
sheet). In their article following a 2017 workshop, Towards Global
These observations suggest a need for FDA to communicate clearly Elimination of the Acute Toxicity Six-Pack, Prior et al. (2019) state that
these tests are primarily conducted for the classification and labelling of
industrial chemicals and agrochemicals. They note that they are largely
banned for testing of cosmetics and rarely required for testing of
pharmaceuticals (Prior et al., 2019). Likewise, in their report on the
status of acute systemic toxicity testing requirements by U.S. regulatory
agencies, Strickland et al. (2018) omitted FDA CDER, noting that it does
not request nonclinical acute systemic toxicity data (Strickland et al.,
2018). These authors state that FDA CDER's 1996 guidance document,
Single Dose Acute Toxicity Testing for Pharmaceuticals, is outdated and no
longer used by FDA, noting that any acute toxicity information needs
are satisfied by repeated-dose toxicity studies, and single dose acute
toxicity studies may not even be reviewed by CDER. While our results
confirm that single dose acute toxicity studies are sometimes not re-
viewed by CDER, even if acute systemic toxicity data are not requested,
sponsors nevertheless frequently submit such data. Regarding whether
FDA CDER's 1996 guidance is no longer used, we note that it is still
accessible, and we found no indication that it is not to be used. Further,
Fig. 4. Local tolerance studies in vivo and in vitro or ex vivo. in its 2005 guidance, Nonclinical Studies for the Safety Evaluation of

7
J. Manuppello, et al. Regulatory Toxicology and Pharmacology 114 (2020) 104666

Pharmaceutical Excipients, FDA CDER and FDA CBER cite this 1996 and increasing the efficiency of human drug development. Altern Lab Anim 46,
guidance to support their recommendation that acute toxicology studies 335–346. https://doi.org/10.1177/026119291804600603.
Chapman, K., Creton, S., Kupferschmidt, H., Bond, G.R., Wilks, M.F., Robinson, S., 2010.
be performed in both a rodent species and a nonrodent species (FDA The value of acute toxicity studies to support the clinical management of overdose
CDER and FDA CBER, 2005). Finally, many reviews included a “Phar- and poisoning: a cross-discipline consensus. Regul. Toxicol. Pharmacol. 58, 354–359.
macology/Toxicology Filing Checklist” (e.g., FDA, 2015) in which one https://doi.org/10.1016/j.yrtph.2010.07.003.
EPA, 2002. Health effects test guidelines OPPTS 870.1100 acute oral toxicity. https://
of the questions asked is: “Are all required … and requested IND studies www.regulations.gov/contentStreamer?documentId=EPA-HQ-OPPT-2009-0156-
… completed and submitted (carcinogenicity, mutagenicity, ter- 0003&contentType=pdf, Accessed date: 15 January 2020.
atogenicity, effects on fertility, juvenile studies, acute and repeat dose EPA, 2018. Interim science policy: use of alternative approaches for skin sensitization as a
replacement for laboratory animal testing. https://www.regulations.gov/
adult animal studies, animal ADME studies, safety pharmacology, etc.)?” contentStreamer?documentId=EPA-HQ-OPP-2016-0093-0090&contentType=pdf.
[Emphasis added] This would seem to indicate that acute systemic EPA, 2020. CompTox chemicals dashboard (predictions). https://comptox.epa.gov/
toxicity studies are required or requested in at least some cases. dashboard/predictions/index, Accessed date: 27 January 2020.
FDA, 1988. LD50 test policy. Fed. Regist. 53, 39650–39651.
FDA, 2015. Orkambi (lumacaftor/ivacaftor) oral tablet. https://www.accessdata.fda.gov/
5. Conclusions drugsatfda_docs/nda/2015/0206038Orig1s000PharmR.pdf, Accessed date: 6
February 2020.
Using publicly-available information, we tallied the single dose FDA, 2017. FDA's predictive toxicology Roadmap. https://www.fda.gov/media/109634/
download, Accessed date: 15 January 2020.
acute toxicity, local tolerance, and skin sensitization studies that were FDA, 2020a. Drugs@FDA: FDA-approved drugs. https://www.accessdata.fda.gov/scripts/
submitted by sponsors in approved, original NDAs, as reported by FDA, cder/daf/, Accessed date: 15 January 2020.
from 2015 through 2018. We found that sponsors submit such studies FDA, 2020b. Search for FDA guidance documents. https://www.fda.gov/regulatory-
information/search-fda-guidance-documents, Accessed date: 15 January 2020.
frequently, challenging the perception that they are generally not re- FDA, 2020c. Research list of guidance documents. https://www.fda.gov/drugs/
quired or conducted for pharmaceuticals. We found little association guidances-drugs/research-list-guidance-documents, Accessed date: 15 January 2020.
between the information reported by FDA in its reviews of these studies FDA, 2020d. FDA-TRACK: agency-wide program performance. https://www.accessdata.
fda.gov/scripts/fdatrack/view/track.cfm?program=cber&status=public&id=
and the information identified as useful by FDA in its guidance to in- CBER-All-IND-and-IDEs-recieved-and-actions&fy=All, Accessed date: 15 January
dustry. Sponsors frequently conducted these studies by routes other 2020.
than those intended for human administration and in more than one FDA CDER, 1996. Guidance for industry. Single dose acute toxicity testing for pharma-
ceuticals. https://www.fda.gov/media/72288/download, Accessed date: 14 January
species of animal. For single dose acute toxicity studies, we accessed 2020.
two conflicting guidance documents, and our results suggest that many FDA CDER, 2002. Guidance for industry. Immunotoxicology evaluation of investigational
sponsors are still following the older guidance. While we found sig- new drugs. https://www.fda.gov/media/72228/download, Accessed date: 15
January 2020.
nificant use of in vitro and ex vivo methods for the evaluation of skin and
FDA CDER, 2015. Nonclinical safety evaluation of reformulated drug products and pro-
eye irritation/corrosion, the use of NAMs for the evaluation of skin ducts intended for administration by an alternate route. Guidance for industry and
sensitization was not reported, despite their increasing acceptance by review staff. https://www.fda.gov/files/drugs/published/Nonclinical-Safety-
regulators in other sectors. Evaluation-of-Reformulated-Drug-Products-and-Products-Intended-for-
Administration-by-an-Alternate-Route.pdf, Accessed date: 15 January 2020.
In order for FDA to reduce the use of animals in the assessment of FDA CDER, FDA CBER, 2005. Nonclinical studies for the safety evaluation of pharma-
new drugs, our analysis suggests a need for FDA to update its guidance ceutical Excipients. https://www.fda.gov/media/72260/download, Accessed date:
for industry such that its recommendations both reduce the number of 15 January 2020.
FDA CDER, CBER, 2010. M3(R2) nonclinical safety studies for the conduct of human
studies conducted by sponsors and identify appropriate NAMs that can clinical trials and marketing authorization for pharmaceuticals. https://www.fda.
replace animal studies. In particular, FDA CDER should revise its 1996 gov/media/71542/download, Accessed date: 15 January 2020.
single dose acute toxicity guidance to be consistent with the more re- Forsby, A., Bal-Price, A.K., Camins, A., Coecke, S., Fabre, N., Gustafsson, H., Honegger, P.,
Kinsner-Ovaskainen, A., Pallas, M., Rimbau, V., Rodríguez-Farré, E., Suñol, C.,
cent ICH guidance or retire this guidance. In addition, FDA should Vericat, J.A., Zurich, M.G., 2009. Neuronal in vitro models for the estimation of acute
emphasize that the evaluation of local tolerance and skin sensitization systemic toxicity. Toxicol. Vitro 23, 1564–1569. https://doi.org/10.1016/j.tiv.2009.
by routes not intended for human administration is not recommended. 07.017.
ICCVAM, 2018. A strategic Roadmap for establishing new approaches to evaluate the
In order for FDA to measure progress toward this goal, it must improve
safety of chemicals and medical products in the United States. https://ntp.niehs.nih.
and standardize its reporting of animal use, ideally extending that re- gov/iccvam/docs/roadmap/iccvam_strategicroadmap_january2018_document_508.
porting beyond approved NDAs to include INDs. So doing would allow pdf, Accessed date: 15 January 2020.
ICH, 2009. Guidance on nonclinical safety studies for the conduct of human clinical trials
more new drugs to be evaluated across a broader range of potential
and marketing authorization for pharmaceuticals M3(R2). https://database.ich.org/
effects in less time while using fewer animals and reducing costs. sites/default/files/M3_R2__Guideline.pdf, Accessed date: 15 January 2020.
Jang, K.J., Otieno, M.A., Ronxhi, J., Lim, H.K., Ewart, L., Kodella, K.R., Petropolis, D.B.,
Declaration of competing interest Kulkarni, G., Rubins, J.E., Conegliano, D., Nawroth, J., Simic, D., Lam, W., Singer, M.,
Barale, E., Singh, B., Sonee, M., Streeter, A.J., Manthey, C., Jones, B., Srivastava, A.,
Andersson, L.C., Williams, D., Park, H., Barrile, R., Sliz, J., Herland, A., Haney, S.,
The authors declare that they have no known competing financial Karalis, K., Ingber, D.E., Hamilton, G.A., 2019. Reproducing human and cross-species
drug toxicities using a Liver-Chip. Sci. Transl. Med. 11https://doi.org/10.1126/
interests or personal relationships that could have appeared to influ- scitranslmed.aax5516. eaax5516.
ence the work reported in this paper. Kleinstreuer, N.C., Hoffmann, S., Alépée, N., Allen, D., Ashikaga, T., Casey, W., Clouet, E.,
Cluzel, M., Desprez, B., Gellatly, N., Göbel, C., Kern, P.S., Klaric, M., Kühnl, J.,
Martinozzi-Teissier, S., Mewes, K., Miyazawa, M., Strickland, J., van Vliet, E., Zang,
Acknowledgements Q., Petersohn, D., 2018. Non-animal methods to predict skin sensitization (II): an
assessment of defined approaches. Crit. Rev. Toxicol. 48, 359–374. https://doi.org/
Funding for the work reported in this article was provided by the 10.1080/10408444.2018.1429386.
Kleinstreuer, NC, Karmaus, A, Mansouri, K, Allen, DG, Fitzpatrick, JM, Patlewicz, G.,
Caroll Petrie Foundation. 2018. Predictive models for acute oral systemic toxicity: a workshop to bridge the gap
from research to regulation. Comput. Toxicol. 8, 21–24. https://doi.org/10.1016/j.
Appendix A. Supplementary data comtox.2018.08.002.
Kneuer, C., Lakoma, C., Honscha, W., 2007. Prediction of acute toxicity in HPCT-1E3
hepatocytoma cells with liver-like transport activities. Altern Lab Anim 35, 411–420.
Supplementary data to this article can be found online at https:// https://doi.org/10.1177/026119290703500410.
doi.org/10.1016/j.yrtph.2020.104666. Kurokawa, Y.K., Shang, M.R., Yin, R.T., George, S.C., 2018. Modeling trastuzumab-re-
lated cardiotoxicity in vitro using human stem cell-derived cardiomyocytes. Toxicol.
Lett. 285, 74–80. https://doi.org/10.1016/j.toxlet.2018.01.001.
References Mannerström, M., Toimela, T., Sarkanen, J.R., Heinonen, T., 2017. Human BJ fibroblasts
is an alternative to mouse BALB/c 3T3 cells in in vitro neutral red uptake assay. Basic
Burt, T., Vuong, L.T., Baker, E., Young, G.C., McCartt, A.D., Bergstrom, M., Sugiyama, Y., Clin. Pharmacol. Toxicol. 121 (Suppl. 3), 109–115. https://doi.org/10.1111/bcpt.
Combes, R., 2018. Phase 0, including microdosing approaches: applying the Three Rs 12790.

8
J. Manuppello, et al. Regulatory Toxicology and Pharmacology 114 (2020) 104666

OECD, 2014. Test No. 431: in Vitro Skin Corrosion: Reconstructed Human Epidermis cardiotoxicity. Front. Physiol. 8, 668. https://doi.org/10.3389/fphys.2017.00668.
(Rhe) Test Method. OECD Guidelines for the Testing of Chemicals, Section 4. OECD Prieto, P., Cole, T., Curren, R., Gibson, R.M., Liebsch, M., Raabe, H., Tuomainen, A.M.,
Publishing, Paris. https://doi.org/10.1787/9789264224193-en. Whelan, M., Kinsner-Ovaskainen, A., 2013. Assessment of the predictive capacity of
OECD, 2017. Test No. 437: Bovine Corneal Opacity and Permeability Test Method for the 3T3 Neutral Red Uptake cytotoxicity test method to identify substances not
Identifying I) Chemicals Inducing Serious Eye Damage and Ii) Chemicals Not classified for acute oral toxicity (LD50 > 2000 mg/kg): results of an ECVAM vali-
Requiring Classification for Eye Irritation or Serious Eye Damage. OECD Guidelines dation study. Regul. Toxicol. Pharmacol. 65, 344–365. https://doi.org/10.1016/j.
for the Testing of Chemicals, Section 4. OECD Publishing, Paris. https://doi.org/10. yrtph.2012.11.013.
1787/9789264203846-en. Prior, H., Casey, W., Kimber, I., Whelan, M.4, Sewell, F., 2019. Reflections on the progress
OECD, 2018a. Test No. 442D: in Vitro Skin Sensitisation: ARE-Nrf2 Luciferase Test towards non-animal methods for acute toxicity testing of chemicals. Regul. Toxicol.
Method. OECD Guidelines for the Testing of Chemicals, Section 4. OECD Publishing, Pharmacol. 102, 30–33. https://doi.org/10.1016/j.yrtph.2018.12.008.
Paris. https://doi.org/10.1787/9789264229822-en. Robinson, S., Delongeas, J.L., Donald, E., Dreher, D., Festag, M., Kervyn, S., Lampo, A.,
OECD, 2018b. Test No. 442E: in Vitro Skin Sensitisation: in Vitro Skin Sensitisation Nahas, K., Nogues, V., Ockert, D., Quinn, K., Old, S., Pickersgill, N., Somers, K., Stark,
Assays Addressing the Key Event on Activation of Dendritic Cells on the Adverse C., Stei, P., Waterson, L., Chapman, K., 2008. A European pharmaceutical company
Outcome Pathway for Skin Sensitisation. OECD Guidelines for the Testing of initiative challenging the regulatory requirement for acute toxicity studies in phar-
Chemicals, Section 4. OECD Publishing, Paris. https://doi.org/10.1787/ maceutical drug development. Regul. Toxicol. Pharmacol. 50, 345–352. https://doi.
9789264264359-en. org/10.1016/j.yrtph.2007.11.009.
OECD, 2019a. Test No. 492: Reconstructed Human Cornea-like Epithelium (RhCE) Test Sirenko, O., Grimm, F.A., Ryan, K.R., Iwata, Y., Chiu, W.A., Parham, F., Wignall, J.A.,
Method for Identifying Chemicals Not Requiring Classification and Labelling for Eye Anson, B., Cromwell, E.F., Behl, M., Rusyn, I., Tice, R.R., 2017. In vitro cardiotoxicity
Irritation or Serious Eye Damage. OECD Guidelines for the Testing of Chemicals, assessment of environmental chemicals using an organotypic human induced plur-
Section 4. OECD Publishing, Paris. https://doi.org/10.1787/9789264242548-en. ipotent stem cell-derived model. Toxicol. Appl. Pharmacol. 322, 60–74. https://doi.
OECD, 2019b. Test No. 442C: in Chemico Skin Sensitisation: Assays Addressing the org/10.1016/j.taap.2017.02.020.
Adverse Outcome Pathway Key Event on Covalent Binding to Proteins. OECD Strickland, J., Clippinger, A.J., Brown, J., Allen, D., Jacobs, A., Matheson, J., Lowit, A.,
Guidelines for the Testing of Chemicals, Section 4. OECD Publishing, Paris. https:// Reinke, E.N., Johnson, M.S., Quinn Jr., M.J., Mattie, D., Fitzpatrick, S.C., Ahir, S.,
doi.org/10.1787/9789264229709-en. Kleinstreuer, N., Casey, W., 2018. Status of acute systemic toxicity testing require-
Passini, E., Britton, O.J., Lu, H.R., Rohrbacher, J., Hermans, A.N., Gallacher, D.J., Greig, ments and data uses by U.S. regulatory agencies. Regul. Toxicol. Pharmacol. 94,
R.J.H., Bueno-Orovio, A., Rodriguez, B., 2017. Human in silico drug trials demon- 183–196. https://doi.org/10.1016/j.yrtph.2018.01.022.
strate higher accuracy than animal models in predicting clinical pro-arrhythmic