PRINCIPLES OF TOXICOLOGY

512 EPIDEMIOLOGIC ISSUES IN OCCUPATIONAL AND ENVIRONMENTAL HEALTH
TABLE 21.1 Cholera Deaths in London (1984) by Water Supply
Water Company Population (1851) Cholera Deaths Rate per 1000 Population
Southwark 167,654 844 5.0
Lambeth 19,133 18 0.9
Sources: Snow (1855); Beaglehole et al. (1993).
record sources commonly used by epidemiologists include employment records, trade union files,
hospital records, motor vehicle registrations, and disease registries. All of these data sources have their
own individual advantages and limitations.
Since Snow, epidemiology has expanded from a method for the investigation of acute infectious
disease epidemics to a multi-faceted scientific discipline. Epidemiology now includes research into
the causes of chronic diseases such as cardiovascular disease and cancer. Epidemiologists often
specialize in particular areas of human health, such as nutrition, occupational and environmental health,
and genetics. Nevertheless, the basic epidemiologic principles have changed little since the time of
Snow and his colleagues.
Epidemiology has been used to investigate the possible associations between disease and exposures
in both the workplace and in the environment. Occupational epidemiologic studies established the
associations between asbestos and lung cancer, vinyl chloride and angiosarcoma of the liver, benzene
and leukemia, repetitive trauma, and carpal tunnel syndrome, as well as many other occupational
exposure–human health effects. Environmental epidemiologic studies have investigated the associations
between methyl mercury exposure and severe neurologic disease near Minamata Bay (Japan),
the effects of radiation in atomic bomb survivors, and the possible carcinogenic effects of electromagnetic
fields. The advantages and limitations of research in these two overlapping areas of epidemiology
are discussed below.
21.2 EPIDEMIOLOGIC CAUSATION
In science, proof that a given exposure causes human health effects is established by a hierarchy of
evidence. This evidence could be the existence of a medical literature with multiple individual case
reports, which associates human disease with a particular exposure. There could be toxicologic
evidence in experimental animals in which the particular exposure causes diseases in animals similar
to those seen in humans. Regardless, epidemiologic studies are considered to be the highest level of
scientific evidence for proving an association between a particular toxic exposure and human health
effects.
In epidemiology, proof of causality (or the association of a particular exposure with a particular
disease) is based on a variety of criteria. These criteria were first expounded by Hill in 1965, with
subsequent refinement and embellishment. These criteria include a temporal relation, plausibility,
consistency, strength, a dose–response relationship, and reversibility and/or preventability. In addition,
consideration must be given to the appropriateness of the design and to limitations, such as sample
size, in each individual epidemiologic study. Ultimately, evidence of causality is the body of
epidemiologic studies meeting all of these criteria.
When considering the possibility of an association between an exposure and a disease, the exposure
must precede the onset of disease. Evidence of a dose–response relationship is necessary; with an
increased dose of a chemical, the risk of disease is increased. The association between the exposure
and the disease must make scientific sense (e.g., have biological plausibility).
Statistical significance does not in itself signify a true association; the association must be
biologically plausible as well as statistically significant. If possible, the association should be
reproducible in toxicologic studies with laboratory animals and other systems such as in vitro systems.