At the request of several Bionet Toxicology readers,
I have briefly summarized a joint Society for Risk
Analysis/Resources for the Future Seminar, "10X
Safety Factor." Approximately 100 persons attended
the seminar on June 11, 1997 at Resources for the
Future in Washongton, DC.
James D. Wilson, a senior fellow at Resources for the Future, and an
organizer of the seminar, explained the intent of the session: to
explore the meaning and impact of a requirement in the recently passed
Food Quality Protection Act (FQPA) to apply an extra tenfold safety
factor to pesticides for the protection of children.
Richard Merrill, Professor at the University of Virginia Law
School, described the origins of the Delaney amendment. He
showed how it applied to pesticides under an interpretation of the
U.S. Food and Drug Administration (FDA) that a pesticide that
concentrated during processing of a food was a food additive. The
U.S. Environmental Protection Agency (EPA), instead, used risk-benefit
provisions to set tolerances for certain pesticides with carcinogenic
effects at high doses, based on the concept of negligible risk. When
the courts declared this practice illegal, the agricultural community
faced the elimination of approximately twelve pesticides. FQPA does
not repeal the Delaney amendment. Instead, it no longer applies to
processed food. The general food safety standard applies now. Under
FQPA, EPA will base the regulation of pesticides entirely on risk. In
his view, it is not clear whether the application of an extra tenfold
safety factor is mandatory or discretionary. He pointed out that the
current Supreme Court weighs the legislative history of a statute much
less than before.
Eric Olson, an attorney for the Natural Resources Defense Counsel
(NRDC) spoke next. NRDC supported passage of FQPA. In his view, the
statute reflects a societal concern for complete protection of
children, and application of a tenfold safety factor is mandatory.
In contrast, Thomas Carrato, a pesticide attorney with the Monsanto
Company, thought the application of the tenfold safety factor was
discretionary. To him, EPA receives ample data from registrants of
agricultural chemicals to detect potential effects specific to
James Lamb, an attorney-toxicologist at Jellinek, Schwartz and
Connolly, laid out some science-policy predicaments created by FQPA,
by focusing on how to set a reference dose. In his view, the decision
by an influential National Academy of Sciences (NAS) Committee was
based on policy -- the idea that children may be more sensitive to
pesticides than adults. He parsed EPA's existing practice about the
application of an extra safety factor, the NAS recommendation, and the
FQPA requirement. To him, all three differ from each other. Thus,
the legislative history of FQPA appears inaccurate, when it refers to
codifying either the existing EPA practice or the NAS recommendation.
James Rowe, an EPA scientist, described how EPA interprets the
approximately $50 million in data received to support registration of
a typical agricultural pesticide through a weight-of-the-evidence
approach. In his view, the data that EPA's scientists review give
them a good understanding of the nature of a pesticide's toxicity. He
described EPA's current policy for the application of extra safety
factors as case-by-case and, thus, discretionary.
Susan Hurt, a scientist at Rohm and Haas and representing an American
Crop Protection Association Committee, analyzed the safety factors for
pesticides that EPA has used in registration decisions after the
passage of FQPA. To her, EPA's decisions about safety factors are
reasonable, well considered, grounded in toxicological data, and
Fanny Ennever, an epidemiologist at Columbia University's School of
Public Health, analyzed the probability distributions for several
safety factors, using data gathered by Michael Dourson. For example,
to understand the safety factor for intraspecies variability, she
compared the slopes of acute toxicity dose-response curves in rats for
490 substances. Based on this analysis, a tenfold safety factor would
fail to protect a sensitive population only approximately one out of a
thousand times. She similarly evaluated safety factors for
interspecies variation, duration of exposure, low effect to no
observed effect levels, and chronic to developmental effect levels.
For example, the no effect dose for chronic effects is approximately
60% lower than the no effect dose for developmental effects. However,
the variation in this factor was high. Overall, a factor of ten seems
to work acceptably for most sources of uncertainty. Further, she
pointed out that combining several factors leads to addition
protection. For example, the combination of two factors, each
protecting against all but one in twenty events, will protect against
more than one in a hundred events. However, because of the wide
variance in the ratio of developmental to chronic effect levels, if
this source of uncertainty is construed as the extra tenfold safety
factor for children in FQPA, it accomplishes very little. For
developmental effects, conducting a separate test seems preferable,
instead of applying a safety factor to chronic effect data.
Several discussants summed up the previous talks.
Clark Carrington, a toxicologist at FDA, pointed out that one primary
benefit of a safety factor is that it preempts discussion. If a
safety factor becomes discretionary, scientists will intensely discuss
whether to apply it, thereby losing a primary benefit. Further, trying
to break down safety assessment into a series of small decisions
generates a fragmented decision process
Barbara Petersen, an exposure assessor at Novigen Sciences, reviewed
variability versus uncertainty in consumption and residue data for
pesticides. In comparison to the safety factors for nutrients in
food, the typical 100-fold safety factor for pesticides is large.
Resha Putzrath, a toxicologist at Georgetown Risk Group pointed out
that, as more sensitive subpopulations petition for special
consideration, we can expect even more safety factors.
Laurissa Rudinko, a toxicologist at Environ Corporation, compared two
views of safety factors, the first a difference between two bright
lines, the second a difference between two fuzzy distributions.
James D. Wilson asked the discussants to take a position
about the scientific interpretation of the new standard under FQPA --
whether a "reasonable certainty" of "no harm" might have a precise
statistical interpretation -- for example, a stated probability of
finding an effect more than one standard deviation different from
background. Given the diversity of responses, it would appear that a
uniform interpretation would be difficult to attain.
Paul Price, an Engineer at McLaren/Hart, made a presentation
showing that that the standard tenfold safety factor for intraspecies
variability already protected children.
In closing, David McCallum of Focus Group and an organizer of the
seminar, gave a general overview of the day's discussions, by pointing
out that policy and science are dissimilar processes. In his view,
the extra tenfold safety factor for children in FQPA is primarily a
policy issue. Opinion surveys reveal a deep public concern about the
effects of pesticide residues, yet the public also makes food
purchases based primarily on price and appearance. An extra safety
factor matters little in this context. In his view, FQPA will have
important unintended impacts, particularly in product safety.
This note was cross-posted to the Toxlist and Riskanal servers.
Daniel M. Byrd III, Ph.D., D.A.B.T.
Consultants in Toxicology, Risk Assessment and Toxicology
1225 New York Avenue, NW
Washington, DC 20005
ctraps at radix.net