Page 44 - Veterinary Toxicology, Basic and Clinical Principles, 3rd Edition

Concepts in Veterinary Toxicology Chapter | 1 11

VetBooks.ir 10 −1 1 Region where morbidity and mortality has important implications for
assessing risks such as from air pollutants like particulate
excess health
risk can be
matter and ozone. For example, the Clean Air Act (CAA,
Supralinear measured 1970) requires the EPA administrator to set National
10 −2
Ambient Air Quality Standards (NAAQS) for these pollu-
U-shaped
Added risk 10 −3 Region where tants at levels that are protective of public health with
an adequate margin of safety. If the exposure response
extrapolation
−4
relationship for the pollutant in question has a threshold,
10
is required
then the setting of the standard is relatively straightfor-
Linear
10 −5 Threshold
Sublinear ward, thereby identifying the threshold concentration and
Range of ability to measure compounds setting the NAAQS at a lower concentration for a given
10 −6
0 10 −4 10 −3 10 −2 10 −1 1 10 1 10 2 average time and statistical form. If a linear, no-threshold,
exposure response relationship is apparent, the question
Exposure (ppm)
arises as how to determine when the NAAQS is set low
FIGURE 1.3 Schematic rendering of exposure response relationships enough. McClellan (2011) has addressed that topic in a
for various toxicants.
paper drawing on the guidance of Supreme Court Justice
Stephen Breyer in the legal case of Whitman v. American
discussion of the concept of hormesis in which very low- Trucking Associations (2001). McClellan (2011) has
level exposures have positive effects with negative effects emphasized the importance of “distinguishing between
observed only at higher exposure levels (Calabrese and (a) the science that informs the setting of the standard and
Baldwin, 2003; Calabrese and Blain, 2005; Calabrese (b) the policy judgments inherent in selecting the stan-
et al., 2007). Technically, in hormesis there is a beneficial dard.” This discussion is extended in a recent commentary
effect at some low level of exposure, which decreases (McClellan, 2016a) on estimates of attributable risk for
with increasing exposure/dose and at yet higher levels ambient particulate matter.
adverse effects become apparent. The concept of hormesis The early development of Threshold Limit Values
has been well known for decades to veterinarians who are (TLVs) for control of occupational exposures by The
aware that certain agents, such as vitamins and minerals, American Conference of Governmental Industrial
are essential for life at low concentrations and can pro- Hygienists (ACGIH), organized in 1938, assumed the
duce toxicity with excess intake. existence of thresholds in exposure response relation-
As an aside, there has been an on-going debate for ships. The initial data used in establishing TLVs were
decades as to whether linear exposure response relation- provided primarily by opportunistic studies of occupation-
ships, especially for cancer, are realistic, i.e., an added ally exposed human populations. In the absence of human
level of exposure, regardless of how small, results in a data, data from controlled exposure studies in laboratory
calculable monotonic increase in cancer risk. It has been animals were used. This necessitated the use of safety
argued by some that the linear exposure response model factors to account for (1) interindividual variability,
is appropriate for regulatory purposes for assessing cancer (2) interspecies extrapolation, and (3) duration of the
risks because every dose of a new agent is added to a study, as will be discussed later. The original safety fac-
background of genetic and other damage in somatic cells tors were formally proposed by Lehman and Fitzhugh
arising from multiple agents and endogenous factors. (1954) of the FDA. In 1970, the newly formed USEPA
I have discussed these issues in a recent book chapter on began using the same factors. However, the EPA identi-
radiation toxicity (McClellan, 2014). That discussion fies them as uncertainty factors apparently out of a desire
draws on the extensive epidemiological data available to avoid use of a potentially contentious word: safety.
on human populations exposed to radiation. After World War II an increased public concern about
In recent years the debate over the nature of the the occurrence of cancer emerged. This was stimulated by
exposure response relationship has been extended from multiple factors. Extensive research conducted during and
cancer as an endpoint to noncancer endpoints. Arguments after the war on the effects of both external ionizing radi-
for the use of linear nonthreshold exposure response ation and internally deposited radionuclides emphasized
relationships for noncancer endpoints have been advanced the importance of cancer as a radiation-induced disease.
by White et al. (2009). An alternative view has been Concern for radiation-induced cancer was further height-
advanced by Rhomberg et al. (2011a,b), namely, that lin- ened when the intensive follow-up of Japanese A-bomb
ear low-dose extrapolation for noncancer health effects is survivors revealed an increase, first in hematopoietic
the exception, not the rule. Extension of the debate over neoplasms, and, later in solid cancers. These findings
the nature of the exposure response relationship to non- soon led to abandoning a threshold approach to evaluating
cancer endpoints, such as respiratory or cardiovascular radiation risks in favor of using a probabilistic or