Beyond the Frontiers of Science: The Food Quality Protection Act and Cumulative Risk Assessment

Ken Weinstein
Nicholas Gertler
Latham & Watkins
Washington, D.C.

Chris Wilkinson

In 1996, Congress enacted far-reaching changes to U.S. law governing pesticide residues in food. Known as "The Food Quality Protection Act" (FQPA) (P.L. 104-170 (Aug. 3, 1996)), the new law, among other things, required the U.S. Environmental Protection Agency (EPA or Agency) to do something that no regulatory authority had ever done, or had any notion how to do: determine the "cumulative risk" of substances that have a "common mechanism of toxicity."

This means, in essence, that EPA must first determine which pesticides can be grouped together based on the type and mechanism of toxicity they exert. The Agency must then develop appropriate methodology for cumulating the risks of various combinations of the "common mechanism groups" to which humans may be exposed. If the total combined exposure to the group exceeds that which is considered safe, decisions must be made to reduce uses and/or application rates of some pesticides within the group.

At the time that the FQPA was under consideration by Congress, no one - and certainly no regulatory agency in the world - had ever performed such a cumulative risk assessment (CRA). Scientists had not previously studied the issue and there were no protocols or established methods for conducting such assessments. Furthermore, there was a notable absence of the types of data (exposure and toxicology) needed to perform the assessments.

The novelty and challenge of conducting CRAs are matched by important regulatory implications. While the mandate to perform CRAs clearly places EPA in uncharted territory, it can potentially create significant pressure to cancel or limit the uses of economically important pesticides. For this reason, EPA should make certain that all concerns with its method are resolved before publishing a public CRA of a group of products that are of importance to agriculture.

Cumulative Risk Assessment of the Organophosphorus Pesticides

A. Development of the Organophosphorus CRA
The first group of pesticides that EPA determined share a common mechanism of toxicity were the organophosphorous (OP) compounds. The OPs exert their insecticidal and mammalian toxicity through their ability to inhibit the enzyme acetylcholinesterase (AChE) that plays a critical role in nerve transmission. There is general scientific agreement that this effect represents a common mechanism (or, more accurately, "mode") of toxicity by which most OPs operate. Inhibition of AChE can result in neurotoxicity through interference with the passage of nerve impulses in the nervous system. While certain dose levels of OPs may be toxic to humans, exposures below the "no-effect" level have, by definition, no toxic effect. As a result, the OPs have been safely and widely used around the world for more than four decades and their effectiveness and relatively low cost have made them valuable adjuncts to modern agriculture. Currently, there is no evidence to suggest that these products, either individually or collectively, have caused or are causing any adverse effects on humans who might be exposed to trace residues in food or water or to low levels resulting from residential uses.

Over the past two years, the Office of Pesticide Programs (OPP) has presented to the FIFRA Scientific Advisory Panel (SAP) for comment several documents describing its proposed methodology for conducting cumulative risk assessments. The difficulties inherent in this enterprise are clearly illustrated by the (still incomplete) evolution of OPP's thinking on how to measure "relative potencies" of the OPs.

EPA proposes to use the Toxicity Equivalence Factor (TEF) approach to CRA in which exposures to a group of OPs of different potencies are normalized to yield a total equivalent exposure to one of the group termed the "index compound" (IC). To do this, the relative potency factor (RPF) of each OP in the group is calculated as the ratio of its selected point of departure to that of the IC. (EPA initially selected chlorpyrifos as the IC and chose to base relative potencies on the ED50 values for steady-state inhibition of plasma cholinesterase (ChE) inhibition in male rats in studies of at least 30 days duration. More recently, the Agency selected mathamidophos as the IC and based relative potencies on the inhibitory activity towards red blood cell (RBC) AChE in male rats in studies >21 days duration.) The exposure to each individual chemical is then multiplied by its RPF to convert it into an equivalent IC exposure and the combined IC-equivalent exposures of the group are assessed in relation to the ADI of the IC. Although the methodology used for calculating RPFs for the OPs is of central importance to the entire CRA process, it is important to recognize that, to date, there is no scientific consensus on the most appropriate way this should be accomplished.

EPA's RPF methodology has evolved significantly in response to peer review by the SAP. In September 2000, EPA presented to the SAP a case study on 24 OPs to demonstrate hazard assessment methods for determining cumulative risk. (See U.S. EPA, "Cumulative Risk: A Case Study Of The Estimation Of Risk From 24 Organophosphate Pesticides," Nov. 9, 2000 (Pilot CRA). The Pilot CRA follows a number of earlier EPA exchanges with the SAP. In September and December 1999, EPA presented guidance documents and early case studies to the SAP for discussion of the hazard and exposure aspects of cumulative risk. See U.S. EPA, "Issues Pertaining To Exposure Assessment And Cumulative Risk," Presented to the FIFRA Science Advisory Panel Dec. 9, 1999; U.S. EPA, "Proposed Guidance On Cumulative Risk Assessment Of Pesticide Chemicals That Have A Common Mechanism Of Toxicity," August 27, 1999. On the basis of these discussions, the SAP recommended that EPA develop more complex case studies to better demonstrate the implementation of EPA's cumulative risk methodology. In response, EPA presented a hazard evaluation of the 24 OPs to the SAP in September of 2000. See U.S. EPA, "Endpoint Selection And Determination Of Relative Potency In Cumulative Hazard Assessment: A Pilot Study Of Organophosphorus Pesticide Chemicals," Sept. 5, 2000.) The document described a proposed approach for determining the relative potencies of 24 OPs based on chronic dose-response data of plasma ChE inhibition.

In September 2001, EPA presented to the SAP a revision of the 2000 methodology based largely on the recommendations of the earlier SAP review (U.S. EPA Office of Pesticide Programs, Preliminary Cumulative Hazard and Dose-Response Assessment for Organophosphorus Pesticides: Determination of Relative Potency and Points of Departure for Cholinesterase Inhibition, July 31, 2001 (hereinafter Relative Potency document).). The SAP has reviewed EPA's latest proposed approach and has issued a guarded endorsement along with recommendations for how it might be further modified. The Agency is expected to issue a preliminary CRA of the OPs on the basis of these recommendations.

In developing its latest (September 2001) RPF methodology for peer review, EPA abandoned its original approach and adopted an entirely new method for RPF measurement based on fitting the ChE inhibition data to a simple exponential model. Major changes in EPA's methodology include:

• use of data on RBC AChE inhibition in male rats instead of data on plasma ChE inhibition, a less reliable surrogate of AChE in the nervous system (use of RBC ChE to evaluate the biological effects of OP is consonant with a broad consensus of scientific opinion and international regulatory policy as well as with EPA's own published guidelines regarding the use of ChE as an endpoint for regulation of OPs); and
• determination of RPFs from the slope of simple exponential plots of RBC AChE inhibition data instead of using ED50 data for inhibition of plasma ChE.

B. A Critical Assessment of EPA's CRA Methodology
EPA presented its latest RPF methodology to the SAP on September 6 and 7. Adding urgency and importance to the SAP's review was the fact that EPA had agreed to a Dec. 1, 2001, deadline for issuing a "preliminary" CRA of the OPs in order to settle litigation brought by the Natural Resources Defense Council and related organizations. (The agreement allows for some slippage of this deadline if EPA encounters "substantial scientific difficulties.") Groups representing growers and the agrochemical industry opposed this non-statutory deadline, concerned that it provided insufficient time for EPA to develop a valid CRA methodology. Among those awaiting the SAP report was a federal judge in San Francisco, who had conditioned approval of the settlement on a favorable pronouncement as to EPA's methodology by the Agency's science advisors.

As described below, EPA's RPF methodology raises significant concerns, many of which were acknowledged by the SAP. In a report released on Sept. 11, 2001, the SAP identified a number of problems with EPA's methodology but stated nonetheless that "it is possible that a draft risk assessment . . . could be completed by December 2001, and strongly encouraged pursuit of this goal." (FIFRA Scientific Advisory Panel Report on EPA's "Preliminary Cumulative Hazard and Dose-Response Assessment for Organophosphorus Pesticides: Determination of Relative Potency and Points of Departure for Cholinesterase Inhibition," Sept. 11, 2001 (hereinafter SAP Report) at 22 (emphasis added).) However, several SAP members expressed serious concerns regarding EPA's ability to complete a scientifically valid CRA on such a short timeframe and significant doubt as to the wisdom of forcing the Agency to try to meet this non-statutory deadline. In the words of SAP member Dr. Gene McConnell at the September 5th SAP meeting,

[I]t's just unfortunate that you're in this corner where you have to make these decisions by the first of December, and I'm just concerned that if we don't have all of these issues in there, and it's probably not practical, you don't have time to get all these issues I'm concerned about in there . . . an inappropriate risk assessment is worse than no risk assessment.

Although similarly expressed by other SAP members at the public meeting, this sentiment was muted in the SAP's written report.

Nonetheless, the SAP's report was not an unqualified endorsement of EPA's methodology, but rather an encouragement for the Agency to refine and improve its approach prior to applying it. The SAP noted that "substantial additional analysis will be required to support the current empirical approach to dose-response modeling, necessitating a significant commitment of resources if this deadline is to be met." (SAP Report at 22.)

EPA's Model Does Not Account for Threshold Effects
The model appears to be seriously flawed and inconsistent with the known biology of the inhibition of AChE by OP pesticides. Like the linear multistage model for cancer risk assessment, EPA's proposed use of a simple exponential model to fit dose-response data for AChE inhibition predicts that any finite dose of an OP pesticide is associated with a decrease in AChE. It fails to acknowledge the well-established fact that OPs exhibit clear dose thresholds below which no inhibition will occur.

Existing Data Are Unsuitable for CRA
The data to which EPA applied its RPF methodology are of questionable value for this purpose. Currently available hazard data (AChE inhibition data) are fundamentally unsuitable for CRA. These data are of central importance to the process since they are the basis of the RPFs used to cumulate exposures. Most of the data were generated specifically for the purpose of providing clear no observable adverse effect level (NOAEL) values for regulatory standard setting. They were never intended for measuring slope-functions of dose-response curves as they are currently being used in the context of CRA. The data are inadequate because:

• many of the data are outdated - some are 15-20 years old and of uncertain quality
• often there are relatively few data points to generate a dose-response curve, and treatment groups are small;
• there are inconsistencies in the methodology used for endpoint measurement and different rat strains were used;
• very few AChE inhibition data are available to allow dose-response analysis of acute OP pesticide exposures; and
• very few data are available for dermal or inhalation exposure routes.

Assumption of Steady-State Levels of Inhibition
A major flaw in the proposed CRA procedure is the use of what are assumed to be steady-state levels of RBC AChE inhibition measured in long-term (> 21 days) studies to assess the potential risks associated with short-term or acute daily exposures. As a result of large differences in pharmacokinetic (PK) behavior among OP pesticides (there are large differences in the PK behavior of different OP pesticides, particularly in relation to their metabolic activation and inactivation by cytochrome P-450-mediated oxidases and their degradation by esterases and glutathione transferases), the patterns of RBC AChE inhibition (intensity, duration, and time to peak effect) will vary substantially. At very low doses, in particular, the active OP is unlikely to reach the AChE target, due to "first pass" effects at major portals of entry into the body and in the liver. As a result, the true contribution to any cumulative effect on AChE often will be zero. The effects observed following long-term (> 21 days) continual exposures to relatively high doses of OP pesticides bear little or no relationship to those that can be expected from shorter term or acute exposures to low doses. A model based solely on such data will always overestimate acute effects by some real, but unknown, amount.

Conclusion
Given the novelty and challenge of conducting a CRA - something that no regulatory agency in the world has done before - EPA can hardly be faulted for wielding a methodology that is not yet ready for application. Due to the continuing unsettled nature of EPA's approach, however, it is important that the development of a valid CRA not be subordinated to the desire to meet arbitrary deadlines. Given the importance and precedential value of EPA's first ever CRA, EPA should make certain that all concerns with its method are resolved before a definitive CRA is published.

This article was contributed by Nicholas Gertler and Kenneth Weinstein, specialists in environmental law with Latham & Watkins, Washington, D.C., and Christopher Wilkinson, a consultant toxicologist and former member of EPA's FIFRA Scientific Advisory Panel.