ASTM E3219-20

This international standard was developed in accordance with internationally recognized principles on standardization established in the Decision on Principles for the
Development of International Standards, Guides and Recommendations issued by the World Trade Organization Technical Barriers to Trade (TBT) Committee.
Designation: E3219 − 20
Standard Guide for
Derivation of Health-Based Exposure Limits (HBELs)
This standard is issued under the fixed designation E3219; the number immediately following the designation indicates the year of
original adoption or, in the case of revision, the year of last revision. A number in parentheses indicates the year of last reapproval. A
superscript epsilon (´) indicates an editorial change since the last revision or reapproval.
1. Scope which one daily dose is not for the 50 kg standard adult but the
dosage form is adjusted to a target population with a lower
1.1 This guide describes the scientific procedures underly-
body weight.
ing the integrative interpretation of all data concerning an
1.7 The primary scope of this guide is to ensure the safety of
active pharmaceutical ingredient (API) taking into account
human patients exposed to residual active substances and
study adequacy, relevance, reliability, validity, and compound-
intermediates via medicinal products. The general principles of
specific characteristics (for example, potency, toxicological
this guide can also be applied to the manufacture of veterinary
profile, and pharmacokinetics) leading to a numerical value for
medicinal products. However, there may be certain unique
the API, which is used further in the quality risk management
toxicological and pharmacological species-specific differences,
(ICH Q9) of cross contamination during the manufacture of
such as metabolism and sensitivity, as well as assumptions
different products in the same manufacturing facilities.
such as body weight for veterinary medicines that are not
1.2 This guide describes general guidance for calculating
addressed in this guide.
and documenting a health-based exposure limit (HBEL). It
1.8 This guide may be used independently or in conjunction
should serve the involved qualified experts as a reference for
with other proposed E55 standards published by ASTM Inter-
HBEL derivations and should harmonize the different ap-
national.
proaches and nomenclature to the greatest extent possible.
1.9 Units—The values stated in SI units are to be regarded
1.3 This guide should be used for calculating and document-
as standard. No other units of measurement are included in this
ing an HBEL, when required or necessary, for APIs (including
standard.
biologics), intermediates, cleaning agents, excipients, and other
1.10 This standard does not purport to address all of the
chemicals (that is, reagents, manufacturing residues, and so
safety concerns, if any, associated with its use. It is the
forth) used for cleaning validation and verification (Guides
responsibility of the user of this standard to establish appro-
F3127 and E3106). In scope is the cleaning and cross contami-
priate safety, health, and environmental practices and deter-
nation of surfaces of manufacturing equipment and medical
mine the applicability of regulatory limitations prior to use.
devices but does not include leachables/extractables (21 CFR
1.11 This international standard was developed in accor-
211.67, 21 CFR 610.11, 21 CFR 820.70, and 21 CFR 111.27).
dance with internationally recognized principles on standard-
1.4 The principles in this guide may also be used as a basis
ization established in the Decision on Principles for the
for setting occupational exposure limits.
Development of International Standards, Guides and Recom-
1.5 The principles in this guide may be applied during the mendations issued by the World Trade Organization Technical
Barriers to Trade (TBT) Committee.
development and commercial manufacturing of small or large
molecular weight medicines as well as isolated pharmaceutical
2. Referenced Documents
intermediates.
2.1 ASTM Standards:
1.6 Subsequent-product HBEL values may be set for spe-
E1262 Guide for Performance of Chinese Hamster Ovary
cific routes of exposure (for example, oral, inhalation, and
Cell/Hypoxanthine Guanine Phosphoribosyl Transferase
parenteral) when necessary (for example, because of differ-
Gene Mutation Assay
ences in bioavailability) and for specific patient populations
E3106 Guide for Science-Based and Risk-Based Cleaning
(for example, children) if formulations are manufactured in
Process Development and Validation
This guide is under the jurisdiction of ASTM Committee E55 on Manufacture
of Pharmaceutical and Biopharmaceutical Products and is the direct responsibility of For referenced ASTM standards, visit the ASTM website, www.astm.org, or
Subcommittee E55.14 on Measurement Systems and Analysis. contact ASTM Customer Service at service@astm.org. For Annual Book of ASTM
Current edition approved Feb. 1, 2020. Published April 2020. DOI: 10.1520/ Standards volume information, refer to the standard’s Document Summary page on
E3219-20. the ASTM website.
Copyright © ASTM International, 100 Barr Harbor Drive, PO Box C700, West Conshohocken, PA 19428-2959. United States
E3219 − 20
F619 Practice for Extraction of Materials Used in Medical ISO 17664 Processing of health care products - Information
Devices to be provided by the medical device manufacturer for the
F719 Practice for Testing Materials in Rabbits for Primary processing of medical devices
Skin Irritation
2.3 ICH Guidelines:
F748 Practice for Selecting Generic Biological Test Methods
ICH M7(R1) Assessment and Control of DNA Reactive
for Materials and Devices
(Mutagenic) Impurities in Pharmaceuticals to Limit Po-
F750 Practice for Evaluating Acute Systemic Toxicity of
tential Carcinogenic Risk (Step 4; 31 March 2017)
Material Extracts by Systemic Injection in the Mouse
ICH Q3A(R2) Impurities in New Drug Substances
F756 Practice for Assessment of Hemolytic Properties of
ICH Q3B(R2) Impurities in New Drug Products
Materials
ICH Q3C(R6) Impurities: Guideline for Residual Solvents
F763 Practice for Short-Term Intramuscular Screening of
(Final; 4 October 2019)
Implantable Medical Device Materials
ICH Q3D(R1) Guideline for Elemental Impurities (Step 4)
F813 Practice for Direct Contact Cell Culture Evaluation of
ICH Q9 Quality Risk Management (Step 4)
Materials for Medical Devices
ICH S9 Nonclinical Evaluation for Anticancer Pharmaceu-
F895 Test Method for Agar Diffusion Cell Culture Screening
ticals
for Cytotoxicity
2.4 Federal Regulations:
F981 Practice for Assessment of Compatibility of Biomate-
21 CFR 111.27 What requirements apply to the equipment
rials for Surgical Implants with Respect to Effect of
and utensils that you use?
Materials on Muscle and Insertion into Bone
21 CFR 211.42(d) Design and Construction Features
F1408 Practice for Subcutaneous Screening Test for Implant
21 CFR 211.46(d) Ventilation, air filtration, air heating and
Materials
cooling
F1439 Guide for Performance of Lifetime Bioassay for the
21 CFR 211.67 Equipment cleaning and maintenance
Tumorigenic Potential of Implant Materials
21 CFR 211.176 Penicillin contamination
F1903 Practice for Testing for Cellular Responses to Par-
21 CFR 610.11 General safety
ticles in vitro
21 CFR 820.70 Production and process controls
F1983 Practice for Assessment of Selected Tissue Effects of
Absorbable Biomaterials for Implant Applications 3. Terminology
F2382 Test Method for Assessment of Circulating Blood-
3.1 Definitions:
Contacting Medical Device Materials on Partial Throm-
3.1.1 acceptable daily exposure, ADE, n—this term for a
boplastin Time (PTT)
health-based exposure limit (HBEL) is synonymous with the
F2808 Test Method for Performing Behind-the-Knee (BTK)
term permitted daily exposure (PDE); see HBEL for details.
Test for Evaluating Skin Irritation Response to Products
3.1.2 accumulation, n—progressive increase in the amount
and Materials That Come Into Repeated or Extended
of a substance in an organism or part of an organism that
Contact with Skin
occurs because the rate of intake from all routes of exposure
F2888 Practice for Platelet Leukocyte Count—An In-Vitro
from repeated dosing exceeds the organism’s ability to remove
Measure for Hemocompatibility Assessment of Cardio-
the substance from the body, ultimately leading to a steady-
vascular Materials
state tissue concentration higher than that associated from a
F2901 Guide for Selecting Tests to Evaluate Potential Neu-
single dose.
rotoxicity of Medical Devices
3.1.3 adjustment factor, AF, n—numerical factor used in a
F3127 Guide for Validating Cleaning Processes Used During
quantitative risk assessment to represent or allow for the
the Manufacture of Medical Devices
extrapolation, uncertainty, or variability of an observed expo-
2.2 ISO Standards:
sure concentration and its associated health outcome in a
ISO 10993-1 Biological evaluation of medical devices --
particular laboratory species or exposed population to an
Part 1: Evaluation and testing within a risk management
exposure concentration for the target population (for example,
process
from animals to human patients and short-term exposure to
ISO 10993-4 Biological evaluation of medical devices – Part
chronic exposure) that would be associated with the same
4: Selection of tests for interactions with blood
delivered dose.
ISO 10993-6 Biological evaluation of medical devices – Part
6: Test for local effects after implantation
3.1.3.1 Discussion—Synonymous with the terms uncer-
ISO 10993-10 Biological evaluation of medical devices –
tainty factor (UF), modifying factor (MF), and safety factor
Part 10: Tests for irritation and skin sensitization
(SF). Ideally, AFs are based on quantitative chemical-specific
ISO 10993-11 Biological evaluation of medical devices –
Part 11: Test for systemic toxicity
ISO 10993-17 Biological evaluation of medical devices--
Available from International Conference on Harmonisation of Technical
Requirements for Registration of Pharmaceuticals for Human Use (ICH), ICH
Part 17: Establishment of allowable limits for leachable
Secretariat, 9, chemin des Mines, P.O. Box 195, 1211 Geneva 20, Switzerland,
substances
http://www.ich.org.
Available from U.S. Government Printing Office, Superintendent of
Available from American National Standards Institute (ANSI), 25 W. 43rd St., Documents, 732 N. Capitol St., NW, Washington, DC 20401-0001, http://
4th Floor, New York, NY 10036, http://www.ansi.org. www.access.gpo.gov.
E3219 − 20
toxicokinetic (TK) or toxicodynamic (TD) data or both and 3.1.10.1 Discussion—The effect shall be relevant for the
consider factors such as interspecies extrapolation, duration of target population (for example, unintended exposure to a
exposure, intraspecies variability, severity of effect, and others. patient or a healthy employee), that is, it is both statistically
Often, default AF values are used because of the absence of significant and clinically relevant. In this context, “critical
chemical-specific TK and TD data. For the purposes of this
effect” means the lead effect is undesired but not necessarily
guide, the terms “pharmacokinetic (PK)” and “pharmacody- harmful in nature. The critical effect may result in the lowest
namic (PD)” are essentially synonymous to “toxicokinetic” and
HBEL; however, there are exceptions.
“toxicodynamic” in the context of HBEL setting.
3.1.11 drug allergy, n—immunologically mediated drug hy-
3.1.4 adverse effect, n—test-item-related change in the
persensitivity reaction.
morphology, physiology, growth, development, reproduction,
3.1.11.1 Discussion—Of the four types of hypersensitivity
or life span of an animal that likely results in an impairment of
reactions, Type I, an immediate IgE-mediated, hypersensitivity
functional capacity to maintain homeostasis or an impairment
reaction is the most common and is a true allergic reaction (9,
of the capacity to respond to an additional challenge or both.
10). T-cell mediated (Type IV) hypersensitivity reactions are
(1-3)
delayed-type reactions and are the second most common.
3.1.4.1 Discussion—A biologically significant pharmaco-
3.1.12 genotoxicity, n—also genetic toxicity; the effect that
logical effect should be considered adverse when establishing
results from damage to DNA and altered genetic expression.
an HBEL for an unintended contaminant or residue.
3.1.5 benchmark dose/benchmark concentration, BMD/
3.1.12.1 Discussion—The four types of genetic change are
BMC, n—mathematically derived dose of a substance that
gene mutation (change in DNA sequence within a gene),
produces a predetermined change in the response rate of an
chromosome aberration (changes in the chromosome
adverse effect relative to the background response of this effect.
structure), aneuploidy/polyploidy (increase or decrease in the
(4-6)
number of chromosomes), and epigenetics (external changes to
DNA such as methylation).
3.1.5.1 Discussion—The BMD or BMC refer to central
3.1.13 general assessment factors, n—factors used to evalu-
estimates. The benchmark dose lower limit (BMDL) and
ate the quality and relevance of scientific and technical
benchmark lower concentration (BMCL) refer to the corre-
information.
sponding lower limit of a one-sided 95 % confidence interval
on the BMD or BMC, respectively.
3.1.13.1 Discussion—Five general assessment factors in-
3.1.6 benchmark response, BMR, n—predetermined change
clude soundness, applicability and utility, clarity and
in the response rate of an adverse effect relative to the
completeness, uncertainty and variability, and evaluation and
background response rate of this effect (for example, 10 %
review (11), with the level of quality assurance applied to the
response for quantal (“yes/no”) or continuous data). (4-6)
information is commensurate with the intended use of the
3.1.6.1 Discussion—The BMR is the basis for deriving
information and the degree of confidence necessary in that
BMDs and BMCs.
information (12).
3.1.7 bioavailability, n—fraction of a substance that reaches
3.1.14 generic drug, n—drug product that is comparable to
the systemic circulation after administration or exposure.
a brand/reference listed drug product in dosage form, strength,
3.1.8 carcinogen, n—agent that is capable of increasing the route of administration, quality and performance
incidence of malignant neoplasms, reducing their latency, or characteristics, and intended use.
increasing their severity or multiplicity.
3.1.14.1 Discussion—Biosimilars are generic biologics.
3.1.8.1 Discussion—The induction of benign neoplasms
3.1.15 hazard characterization (dose-response assessment
may, in some circumstances, contribute to the judgment that
in U.S. EPA risk assessment framework), n—qualitative and,
the agent may be carcinogenic. The terms “neoplasms” and
wherever possible, quantitative description of the inherent
“tumor” are used interchangeably (7). Carcinogens that are
property of an agent or situation having the potential to cause
likely causing tumors by interaction with deoxyribonucleic
adverse effects (13). It is a description of the potential adverse
acid (DNA) (genotoxic) are distinguished from carcinogens
health effects attributable to a specific compound, the mecha-
causing tumors by other mechanisms not involving genotoxic-
nisms by which the agent exerts its toxic effects, and the
ity (non-genotoxic).
associated dose, route, duration, and timing of exposure.
3.1.9 clinically relevant, adj—biologically meaningful
3.1.16 health-based exposure limit, HBEL, n—dose that is
change in patient health in response to exposure.
unlikely to cause an adverse effect if an individual is exposed,
3.1.10 critical effect, n—first adverse effect, or its known
by any route, at or below this dose every day for a lifetime.
precursor, that occurs in the increasing dose/concentration
scale after appropriate adjustment for interspecies differences
3.1.16.1 Discussion—The HBEL, being based on the criti-
and interindividual variability. (8)
cal effect, should be protective of all populations by all routes
of administration and should be the result of a structured
scientific evaluation of all available pharmacological and
toxicological data including both non-clinical and clinical data
The boldface numbers in parentheses refer to a list of references at the end of
this standard. (14, 15).
E3219 − 20
3.1.17 intermediates, n—materials produced during steps in 3.1.29 pharmacodynamics, n—derived from toxicodynam-
the synthesis of an active pharmaceutical ingredient (API) that ics; describe and quantify the sequence of cellular and molecu-
shall undergo further molecular change or processing resulting lar events at the target site leading to a pharmacological
in an API. response to a drug.
3.1.18 in silico, adj—expression used to mean “performed 3.1.30 pharmacokinetics, n—derived from toxicokinetics;
describe and quantify the time course of absorption,
on computer or via computer simulation.”
distribution, biotransformation, and excretion of a drug.
3.1.19 in vitro, adj—studies that are performed with cells or
3.1.31 point of departure, PoD, n—dose-response point that
biological molecules outside their normal biological context,
marks the beginning of a low-dose extrapolation to derive an
for example, proteins evaluated in solution or cells in artificial
HBEL. (8)
culture medium.
3.1.20 lowest observed adverse effect level, LOAEL, 3.1.31.1 Discussion—This point can be a NOAEL/NOEL,
n—lowest exposure level in a study in which there were
LOAEL/LOEL, or BMDL for an observed effect (18).
statistically or biologically significant changes in frequency or
3.1.32 potency (activity), n—expression of the relative re-
severity of adverse effects between the exposed population and
sponse of an agent as compared to a given or implied standard
its appropriate control group. (8)
or reference.
3.1.21 lowest observed effect level, LOEL, n—lowest dose 3.1.33 qualified expert, n—individual with specific educa-
or exposure level in a study in which a statistically or tion and training in toxicology/pharmacology/
biologically significant effect is observed in the exposed pharmacotherapy and risk assessment methods that can apply
population compared with an appropriate unexposed control the principles of toxicology to deriving an HBEL.
group that demonstrated an effect between the exposed popu-
3.1.34 reliability, n—inherent quality of an effect value in a
lation and its appropriate control group. (8)
test report or publication relating to a clearly described
experimental design, performance of the experimental
3.1.22 margin of safety, MOS, n—ratio of the HBEL to the
procedures, and reporting of the results to provide evidence of
estimated exposure. (13)
the reproducibility and accuracy of the findings. (19, 20)
3.1.23 mechanism of action, n—detailed description, often
3.1.35 risk assessment, n—systematic process to organize
at the molecular level, of the means by which an agent causes
and analyze scientific knowledge and information used to
a disease or other adverse effect. (16)
characterize the potential adverse effects of human exposures
3.1.23.1 Discussion—The term “mechanism of action” im-
to an agent, including uncertainties inherent in the process of
plies a more detailed understanding and description of events,
inferring risk. (13, 21, 22)
often at the molecular level, than is meant by mode of action
3.1.35.1 Discussion—According to the National Research
(17).
Council paradigm, risk assessment consists of four steps: (1)
3.1.24 mode of action, n—sequence of key events and
hazard identification, (2) dose-response assessment, (3) expo-
processes, starting with interaction of an agent with a cell,
sure characterization, and (4) risk characterization (21).
proceeding through operational and anatomical changes, and
3.1.36 severity, n—extent to which an effect impairs the
resulting in the adverse effect. (16, 17)
functional capacity of an organism, that is, the degree of
3.1.24.1 Discussion—A “key event” is an empirically ob- adversity.
servable precursor step that is itself a necessary element of the
3.1.36.1 Discussion—This continuum is a composite of
mode of action or a biologically based marker for such an
many variables, including degree of impairment to the
element (17).
organism, magnitude, organ effected, incidence, reversibility,
3.1.25 no observed adverse effect level, NOAEL, n—highest
pathologic severity, and other factors that give an indication of
exposure level at which there are no biologically significant
the severity. Examples of severe effects include
increases in the frequency or severity of adverse effects
carcinogenicity, teratogenicity, neurotoxicity, and death.
between the exposed population and its appropriate control;
3.1.37 threshold of toxicological concern, TTC, n—TTC
some effects may be produced, but they are not considered
approach is a screening and prioritization tool for the safety
adverse or precursors of adverse effects. (8)
assessment of chemicals when hazard data are incomplete and
3.1.26 no observed effect level, NOEL, n—exposure level at
human exposure can be estimated and, thus, for deciding
which there are no statistically or biologically significant
whether exposure to a substance is so low that the probability
increases in the frequency or severity of any effect between the
of adverse health effects is low and that no further data are
exposed population and its appropriate control. (8)
necessary.
3.1.27 over-the-counter (OTC) drugs, n—medicines sold
3.1.37.1 Discussion—The TTC is not applicable when
directly to the consumer without a prescription from a health-
compound-specific assessment and toxicity data are available
care professional.
or are required under existing regulations (23, 24).
3.1.28 permitted daily exposure, PDE, n—this term for a 3.1.38 toxicodynamics, n—describe and quantify the se-
health-based exposure limit (HBEL) is synonymous with quence of cellular and molecular events at the target site
acceptable daily exposure (ADE); see HBEL for details. leading to an adverse response to a chemical.
E3219 − 20
3.1.39 toxicokinetics, n—describe and quantify the time Society of Pharmaceutical Engineers (ISPE) (29) in which it is
course of absorption, distribution, biotransformation, and ex- mentioned that relevant residue limits should be based on a
cretion of chemicals. toxicological evaluation.
3.1.40 uncertainty, n—refers to a lack of knowledge about 4.5 Key Concepts—This guide applies the following steps:
specific factors, parameters, or models. (25) (1) hazard characterization, (2) identification of the critical
effect(s) including dose-response assessment, (3) determina-
3.1.40.1 Discussion—It is important to characterize ad-
tion of one or several points of departure (PoD)s, (4) applica-
equately variability and uncertainty in a risk assessment.
tion of PoD-specific AFs, and (5) calculation of HBELs
“Uncertainty includes parameter uncertainty (measurement
including justification of selected HBEL (18) (see Fig. 1).
errors, sampling errors, systematic errors), model uncertainty
(uncertainty due to necessary simplification of real-world 5. Procedure
processes, mis-specification of the model structure, model
5.1 The procedure proposed in this guide for determination
misuse, use of inappropriate surrogate variables), and scenario
of an HBEL is based on the methods for establishing the
uncertainty (descriptive errors, aggregation errors, errors in
permitted daily exposure (PDE) as described in EMA guidance
professional judgment, incomplete analysis).” (25) See also
(14), the acceptable daily exposure (ADE) value as described
Ref (26) for a generic list of common types of uncertainties in
in ISPE guidance (29), as well as principles outlined in the
inputs and methodologies.
scientific literature.
3.1.41 variability, n—refers to observed differences attrib-
5.2 The establishment of an HBEL is a process that requires
utable to true heterogeneity or diversity in parameter values
expertise and needs to be done by a qualified expert and, if
over time, space, or different member of a population (for
possible, should be peer reviewed by relevant subject matter
example, in cumulative exposure dose or dose rate to an
experts. A curriculum vitae (CV) should be available on
individual or group of individuals or in response to exposure).
request that demonstrates the educational background (for
(25, 26)
example, toxicology, pharmacology, medicine, or other health-
3.1.41.1 Discussion—It is an inherent property of a popula- related disciplines), certifications such as the Diplomate of the
tion being evaluated and, while it can be better characterized American Board of Toxicology (DABT) or European Regis-
with more data, it usually cannot be reduced and cannot be tered Toxicologist (ERT), years of experience in the field, and
publications related to the field. While all are not required for
eliminated.
a “qualified expert,” the appropriate documentation in these
areas demonstrates the expertise to work in this area. Typically,
4. Significance and Use
certification registries require an academic degree in a relevant
4.1 Guidelines for unintended human exposure to active
subject, basic knowledge of the major areas of toxicology, at
pharmaceutical ingredients (APIs) are required by various
least five years of relevant toxicological experience, suitability
global regulations as part of international quality requirements,
for registration (for example, by published works, reports, or
needed as good product stewardship, and are considered the
assessments), and current professional engagement in the
industry standard.
practice of toxicology (30, 31).
4.2 Application of the approach described within this guide
5.3 Documentation describing the procedure to derive an
applies a scientifically justified, data-driven, approach to de-
HBEL should be described by the qualified expert in a
riving safe limits for unintended exposures to individual
monograph. The purpose of a monograph is to communicate
substances. These limits can then be further used to calculate
effectively with the stakeholders and document the scientific
cleaning limits used in quality risk assessment for the manu-
data and methods underlying the HBEL derivation to enable its
facture of pharmaceuticals. The HBEL approach considers
inspection by the regulators. An example template for an
substance-specific properties (type of effect, potency,
HBEL monograph is available in Appendix X1; however, the
pharmacology, safety profile, and so forth). Specific ap-
general format may vary.
proaches are applicable to different categories of substances
5.4 Hazard Identification and Characterization:
and in specific stages in drug development.
5.4.1 The purpose of the hazard identification and charac-
4.3 The basis for the HBEL derivation is all available
terization is to identify the health effects caused by a chemical
substance-specific data. Interpretation of these data considers
agent. It involves evaluating the quality and relevance of the
the quantity and robustness of the database and the reliability
available scientific and technical information on the chemical
and relevance of the data. Typically, adjustment factors (AFs)
agent, including the mechanism(s) by which an agent exerts its
are used to address variability and uncertainty in different
toxic effects; the associated doses, and the route, duration, and
parameters to determine a safe human exposure limit, although
timing of exposure. The U.S. Environmental Protection
alternative, purposefully conservative, approaches [for
Agency (EPA) has described the five general assessment
example, threshold of toxicological concern (TTC), read-
factors it typically considers in evaluating such data: (1)
across] may be used as appropriate.
soundness; (2) applicability and utility; (3) clarity and com-
4.4 This guide supports, and is consistent with, elements of pleteness; (4) uncertainty and variability; and (5) evaluation
the European Commission (EU) Guidelines for Good Manu- and review (11).
facturing Practice for Medicinal Products for Human and 5.4.1.1 The evaluation of all substance-specific information
Veterinary Use (27, 28) and guidance from the International should result in a comprehensive characterization of the
E3219 − 20
NOTE 1—This figure represents an example where three possible PoDs have been selected based on three distinctive critical effects, followed by
PoD-specific application of AFs and calculation of three HBELs.
FIG. 1 Process Underlying the Calculation and Final Selection of an HBEL
hazards and understanding of the safety profile of a substance. may be more relevant when deriving an HBEL based on data
Evaluation of the quality and validity of toxicological data are from secondary literature searches than from using proprietary
frequently conducted following the reliability scoring catego- innovator data, which are typically based on original good
ries and codes developed by Klimisch et al (19). Such an laboratory practice (GLP) guideline studies. It is recommended
evaluation is to ensure that the data being used to identify if using the Klimisch criteria that the studies used to derive the
potential critical effects are of sufficient quality and validity to critical effect should have a Klimisch score of either 1 (reliable
address the hazards of the chemical. Determining data quality without restriction) or 2 (reliable with restriction). If data with
E3219 − 20
a Klimisch score of 4 (reliability not assignable) are used, a 5.4.4 Literature searches for hazard characterization should
justification should be provided. Data with a Klimisch score of be performed or reviewed by a toxicologist or other qualified
risk assessment expert. Verifying the reliability of this infor-
3 (not reliable) should not be used. In lieu of the original
mation remains a responsibility of the qualified expert. A
studies, secondary data sources that extract information from
highly reliable studies (such as found in product package qualified expert can efficiently determine the literature search
strategy based on the type of compound (data-rich or data-
inserts, investigators brochures, and so forth) are acceptable to
poor). The qualified expert can also determine where the data
use for identifying the critical effect. The ToxRTool Excel
gaps occur and may either try to obtain the data, fill in the gaps
spreadsheet is a useful tool for evaluating studies and scoring
as well as possible (for example, read-across, mechanism of
their reliability using the Klimisch criteria (32).
action, and so forth), use approaches such as the Threshold of
5.4.1.2 Data quality evaluation of human epidemiological
Toxicological Concern (TTC), or apply a larger AF because of
studies is far more complex given the wide variety of study
increased uncertainty from lack of data (18). Ideally, high-
designs (for example, randomized clinical trials, nonrandom-
quality clinical datasets are available and should be evaluated
ized cohort studies, case-control studies, case-crossover
as they are generally more relevant than nonclinical studies for
studies, cross-sectional studies, and pharmacovigilance
most adverse health effect endpoints (exceptions being devel-
studies), each with a potential for biases (that is, confounding,
opmental and reproductive toxicity, carcinogenicity) to the
information bias, and selection bias) that could introduce
calculation of a human HBEL.
systematic errors in a study, a variety of critical appraisal tools,
5.4.5 The following end points are typically available for
and elaborate methods to synthesize multiple study results
review on a commercial stage API:
through systematic reviews and meta-analyses. Nonetheless,
5.4.5.1 Nonclinical Data—A variety of dose-response and
there is no consensus or “gold standard” tool for these
mechanistic nonclinical data are collected during API devel-
evaluations and no single tool that works across study types
opment to support a drug filing. These include single-dose
(33-35). Use of human data from clinical epidemiological
safety pharmacology studies (for example, cardiac,
studies that follow good epidemiologic practice guidelines (for
neurobehavioral), repeated-dose studies (including develop-
example, GRADE, PRISMA, and CONSORT) or high-quality
mental and reproductive toxicity studies), local tolerance,
systematic reviews (for example, Cochrane Database Systemic
sensitization, and carcinogenicity studies. During data
Reviews) are to be preferred.
collection, factors related to the mechanism of action such as
5.4.2 Drugs that have recently become off-patent have been
target receptors, potency, pharmacological effect(s), and the
evaluated and approved according to the up-to-date methods
indication(s) for the drug product will have been characterized.
used to assess their safety and efficacy. Conversely, drugs that
The compilation of all relevant toxicological data of the
have been off-patent for decades may not have been assessed
substance should permit the identification of the critical ef-
with the same rigorous methodology, especially in the preclini-
fect(s) and the dose-response relationships of the observed
cal phase. This may result in a data gap for certain potential
effects in relevant nonclinical species and relevant routes of
adverse health effect end points that need to be addressed while
exposure. Some consideration for identifying the critical effect
assessing the data quality and reliability during calculation of
could include the type of effect measured, severity and revers-
an HBEL. In addition, nonclinical evaluations may be abbre-
ibility of effect, human relevancy of the effect, duration of
viated for certain indications, such as oncology, thus also
exposure (generally more weight is applied to longer versus
resulting in data gaps (ICH S9). To assure consistency of the
shorter studies), species selected, route of administration,
HBELs, it is important to select the PoD that is reliable, while
number of animals tested, type of endpoints measured, and
appropriately modifying certain AFs to address potential data
appropriate statistical analysis.
gaps.
5.4.5.2 Human Data—As described in 5.4.5.1, a variety of
5.4.3 Another gap that is present when assessing certain
epidemiology data may be collected during development and
older generic drugs is accessibility of the primary data. In
post-approval in patients and often healthy human volunteers
many cases, only a summary is available, with no details about
that support the safety and efficacy profile of an API. Where
NOAELs identified during the nonclinical and clinical trial
available, these human data are often of higher relevance than
studies, the route of administration, or the doses is accessible.
animal data for the same endpoints, for example, the
In the absence of access to the nonclinical and early clinical
pharmacokinetics, pharmacological effects, and adverse clini-
trial data, human data (for example, late-stage clinical trials,
cal effects (36). Characteristics of a robust clinical dataset for
post-marketing surveillance/pharmacovigilance, and occasion-
an API could include:
ally case reports) may be used as the PoD since a suitably large
(1) Information on pharmacological effects and its dose
number of patients and patient populations may have been
dependence, the indication, and range of therapeutic doses
treated over the intervening years since approval. In those
(including those for sensitive subpopulations);
cases, it may be sufficient to select the PoD based on the
(2) Adverse effects observed at therapeutic doses and,
clinical doses used to treat human patients. However, it is
optimally, also the dose dependence of these effects, including
important to note whether susceptible subpopulations have
adverse effects at sub- and supra-therapeutic doses;
been identified or purposefully excluded from the treatment (3) Pharmacokinetics in humans including all available
(for example, women of child-bearing potential because of
absorption, distribution, metabolism, and excretion (ADME)
developmental toxicity concerns). parameters in healthy and patient populations; and
E3219 − 20
(4) Information on effects and precautions/ preferred to the traditional NOAEL/LOAEL approach as it
contraindications for specific subpopulations, such as patients corrects for these limitations (4-6, 43-48). The BMDL is
with severe renal or liver impairment, pregnant women,
typically considered equivalent to the NOAEL. The BMDL is
children, or the elderly.
dependent on the benchmark response (BMR), which is based
on the sensitivity of the study, and in many cases, the BMR is
5.5 Identification of the Critical Effect(s):
considered to be 10 % over background of effect. The goal of
5.5.1 The purpose of this step is to identify the effect most
a BMD is to fit a model to the dose-response data, and it
likely to be relevant for the target population (patients) and
represents an acceptable alternative to the NOAEL assessment
target route of exposure (oral, parenteral, other). The “critical
factor approach for deriving an HBEL (42, 49).
effect” has been defined as the “most sensitive adverse effect
that is considered relevant to humans” (37) or the “first 5.6.3 The typical dosing schedule for a pharmaceutical
clinically significant adverse effect that is observed as the dose
should be considered during HBEL extrapolation. For APIs
increases” (29, 38) and “the first adverse effect, or its known
administered at least twice daily, the HBEL is expressed as the
precursor, that occurs to the most [relevant or] sensitive species
total daily human therapeutic dose. However, the potential for
as the dose rate of an agent increases” (39). The critical effect
acute health effects from a single dose (C -mediated effects)
max
shall be clinically relevant (1-3, 40). To evaluate the clinical
has to be considered, as a single dose may have a clinically
relevance of an adverse effect, the similarity of effects between
relevant effect that is the critical effect. For APIs administered
animal species and humans and demonstration of homology
with dosing intervals greater than once daily (for example,
between the animal model and humans are evaluated (41).
routine dosing schedules such as once weekly or once monthly
5.5.2 For an API with a favorable therapeutic index, there is
as is commonly seen for biologic drug products and some small
a large margin between doses that cause a pharmacological
molecules), generally a PoD as a prorated daily dose can be
effect and doses that cause adverse effects. In such cases, the
used (that is, the single dose divided by the number of days
critical effect is often identified as the intended pharmacologi-
between doses). For APIs that are not routinely chronically
cal activity. This follows the assumption that all effects, both
administered to an individual patient, but rather on an ad hoc
intended pharmacology and unintended toxicity, are considered
basis (for example, vaccinations, surgical or certain medical
adverse in a potential cross-contamination scenario. In this
procedures), the PoD should be evaluated on the basis of the
context, “critical effect” means the lead effect that is undesired
available data with AFs incorporated as appropriate to reflect
but not necessarily adverse in nature. In the context of setting
potential chronic exposure for derivation of an HBEL. In these
HBELs, pharmacological effects are considered adverse (37).
cases, or where the dosing schedule is intermittent or
5.5.3 Each identified critical effect will generally necessitate
otherwise, the PoD is not a dose at steady state, the PK AF may
application of different AFs, meaning that the effect occurring
be used instead of daily dose averaging. Where applicable, PK
at the lowest dose identified might not always lead to the
or PD can also be used to inform the derivation of a daily dose
derivation of the lowest HBEL value. It is recommended that
or a pharmacologically ineffective dose that can be used as a
each of the relevant, reliable, critical effects should be used to
PoD (18, 50).
derive an HBEL (18).
5.6.4 The body weight and other dosing parameters (for
5.6 Determination of the PoD:
example, body surface area for a topical drug) may change
5.6.1 The PoD determination builds upon the data
depending on the route of exposure being evaluated for
collection, dose-response assessment, and identification of the
establishing a limit, as well as for the regulatory jurisdiction.
candidate critical effects (28). It has the dimension of a dose
For the general population, the body weight used can be
(for example, mg/kg or mg/person). The PoD for the critical
conservatively set to a small adult person of 50 kg (14, 29, 51,
effect is used to derive the lowest HBEL relevant for human
52). Other jurisdictions may use alternative values for adults
exposure. In determining the PoD, all relevant end points
and different pediatric populations (53-56). The European
including nonclinical and clinical data shall be evaluated.
Medicines Agency (EMA) has stated the “derivation of limits
Ideally, the PoD is based on the no-observed-adverse-effect
will need to take account of the dose to be administered, which
level (NOAEL) or the no-observed-effect level (NOEL) of the
will be influenced by the body weight of the species to be
most sensitive or relevant species or both for the critical
treated” (14). If assessing alternative populations or exposure
effect(s) [ICH Q3C(R6)(37). When a NOAEL or NOEL are not
routes (for example, infants), consult an appropriate reference
available, the lowest-observed-effect level (LOEL) or the
(53-56). In a draft document, EMA has suggested consideration
lowest-observed-adverse-effect level (LOAEL) can be used as
of body weight values for three pediatric populations: 0.5 kg
a PoD.
for a prematurely born newborn, a 3.5 kg newborn, and 10 kg
5.6.2 The NOAEL approach has its limitations including:
for a child (57).
(1) the identification of NOAEL values is limited by the doses
5.6.5 Regarding body surface area, the EPA guidance pro-
tested; (2) the NOAEL may not represent a true 0 % response
vides mean and 95th percentile estimates of the total body
(that is, because of sample size considerations, a study may not
surface area for children and adults ((56), Table 7.1). For
have sufficient power to detect an adverse effect “signal”); (3)
adults, mean total body surface area values are on the order of
the NOAEL is highly dependent on sample size; (4) NOAELs
2 2 2
are not always available; and (5) it does not consider the 2 m (20 000 cm ). The FDA assumes a 1.62 m body surface
area for a human adult of 60 kg, therefore, for a 50 kg adult, the
dose-response curve or data variability and, thus, “wastes” data
(6, 42). Where feasible, the benchmark dose (BMD) method is body surface area would be 1.35 m (58).
E3219 − 20
5.6.6 For certain protein therapeutics, guidance is available peer review of the available data should be applied to yield a
for first-in-human (FIH) dose selection, which represents a consensus in the selection of each AF. Care should be taken not
dose that is expected to have no clinical effect (59-61). This to adjust for the same uncertainty in two factors. For the
includes estimating the minimum anticipated biological effect selection of the following AFs, the rationale should be pro-
level (MABEL) from PK/PD modeling. The FIH dose may vided in detail and justified for each calculation.
serve as a surrogate PoD pending collection of clinical data. 5.7.5 Sources of variability, uncertainty, and additional
adjustments that are typically addressed in a quantitative risk
5.7 Application of AFs:
assessment include, but may not be limited to, the list in Table
5.7.1 The purpose of the application of AFs is to adjust for
1.
uncertainty and variability in the various parameters measured
5.7.6 The list in Table 1 is rather a compilation of terminol-
in the critical study compared to effects that may occur in the
ogy on factors and does not indicate that all should be used for
population targeted by the HBEL assessment. Synonyms in-
each PoD. The specific AFs used on an organizational basis
clude assessment factors, uncertainty factors, safety factors,
should be described procedurally to demonstrate consistency
and modifying factors [ICH Q3C(R6)] (14, 62). Eq 11 in 6.1
between documents.
provides the basic equation for the determination of an H
...