and for variation within humans. Guidance on the selection of appropriate safety
factors has been published (IPCS 1987 ), and it is generally dependent on the nature
of the critical effects and the endpoint, as well as the quality and quantity of the data
used to make the safety assessment. Safety factors range in value between 10 and
2,000. Typically, a safety factor of 100 is applied when long-term animal studies are
available. It allows for human beings to be, up to tenfold more sensitive than the test
animals, and for tenfold variation between individuals within the human population
(Rubery et al. 1990 ).
The value of an ADI for a chemical can differ amongst regulatory agencies due
to differences in regulatory policies, or the choice of safety factors, or both. The
ADI of a substance may be based on a toxicological, pharmacological or micro-
biological endpoint; and all relevant endpoints should be assessed when character-
ising the hazard. Derivation of the ADI is based on the critical endpoint observed, to
result from the lowest dose. A pharmacological ADI is derived from a NOEL
determined from pharmacological studies in laboratory animals or human pharma-
cological data, whereas a microbiological ADI is determined from studies into the
effects on intestinal microflora. The studies for generating microbiological data
may be conducted in humans, in vivo in gnotobiotic animals, or in vitro using
species and strains of microorganisms representative of the human gut flora. The
approaches used for establishing a microbiological ADI for residues in foods that
are derived from animals treated with an antimicrobial agent were reviewed by
Cerniglia and Kotarski ( 2005 ).
The residues of some compounds including pharmacologically active veterinary
drugs are associated with short-term (acute) toxicity. To protect consumer health
from such residues, some but not all jurisdictions establish an acute reference dose
(ARfD) as the appropriate health standard. In addition, all jurisdictions base the
ADI of substances on acute toxicity endpoints when it is appropriate to do so. Both
the ARfD and the ADI protect consumers against acutely toxic residues in food.
Allergenicity has not been a major issue with most of the veterinary drugs evaluated
to date. An exception is benzylpenicillin, for which allergy was the determining
factor in the safety evaluation conducted by the JECFA (WHO 1990 ). The evalua-
tion did not establish an ADI for benzylpenicillin and instead recommended that the
daily intake of penicillin from food be kept as low as practicable and always below
30 mg per person (0.5mg/kg) of the parent drug.
The NOEL-safety factor approach is not considered to be suitable for setting
acceptable intake levels for substances such as genotoxic carcinogens, whose
effects involve non-threshold mechanisms. Genotoxic carcinogens may cause
genetic alterations in target cells, either directly or indirectly, and they are assumed
to be harmful at any level of exposure. Evidence for genotoxicity is generally
obtained from mutagenicity testing. Genotoxic chemicals are banned from use
in food-producing animals in many countries, while in other countries; their use
is permitted at a level of risk that is sufficiently small to be deemed negligible.
In contrast, non-genotoxic carcinogens per se do not cause mutations. These
substances act at extra-genetic sites, causing enhanced cellular proliferation or
sustained hyperfunction or dysfunction at the target sites, or both. In principle,
270 P.T. Reeves