5
Analysis of APHIS Assessments

As indicated in Chapter 1, regulatory agencies charged with assessing the safety of transgenic plants face a daunting task. This is partly because environmental risk assessment for transgenic plants is new and partly because the social context in which regulatory decisions about transgenic organisms must now be made is dramatically different from the social context in which these agencies are accustomed to working.

The Animal and Plant Health Inspection Service (APHIS) began its involvement in the regulation of transgenic organisms in the mid-1980s. APHIS’s regulatory system has improved substantially since that time. As pointed out in the case study of two Bt corn petitions for nonregulated status (see Chapter 4), the scope of environmental issues addressed and the degree of rigor with which they were addressed both increased between 1994 and 1997. Furthermore, development of a notification process that focuses on plant ecology was an important step in effectively streamlining the field-testing process. The learning process at APHIS has not come without missteps, but the agency seems to use those missteps as opportunities for further improvement. The present analysis is designed to facilitate additional improvements in the APHIS system, which will be necessary to meet future challenges of assessing potential environmental risks of a large number of diverse and novel transgenic products. In analyzing the APHIS regulatory process, the committee searched for problem areas. The committee’s criticisms here are not meant to be an indictment of the system but rather a means to help improve an already functioning system. It is hoped that APHIS personnel will find this analysis useful in the spirit it is intended.



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 167
Environmental Effects of Transgenic Plants: The Scope and Adequacy of Regulation 5 Analysis of APHIS Assessments As indicated in Chapter 1, regulatory agencies charged with assessing the safety of transgenic plants face a daunting task. This is partly because environmental risk assessment for transgenic plants is new and partly because the social context in which regulatory decisions about transgenic organisms must now be made is dramatically different from the social context in which these agencies are accustomed to working. The Animal and Plant Health Inspection Service (APHIS) began its involvement in the regulation of transgenic organisms in the mid-1980s. APHIS’s regulatory system has improved substantially since that time. As pointed out in the case study of two Bt corn petitions for nonregulated status (see Chapter 4), the scope of environmental issues addressed and the degree of rigor with which they were addressed both increased between 1994 and 1997. Furthermore, development of a notification process that focuses on plant ecology was an important step in effectively streamlining the field-testing process. The learning process at APHIS has not come without missteps, but the agency seems to use those missteps as opportunities for further improvement. The present analysis is designed to facilitate additional improvements in the APHIS system, which will be necessary to meet future challenges of assessing potential environmental risks of a large number of diverse and novel transgenic products. In analyzing the APHIS regulatory process, the committee searched for problem areas. The committee’s criticisms here are not meant to be an indictment of the system but rather a means to help improve an already functioning system. It is hoped that APHIS personnel will find this analysis useful in the spirit it is intended.

OCR for page 167
Environmental Effects of Transgenic Plants: The Scope and Adequacy of Regulation APHIS’s regulatory process has never led to the release of a transgenic plant that clearly caused environmental damage. However, without systematic monitoring, the lack of evidence of damage is not necessarily lack of damage. Furthermore, based on the questions raised in Chapter 4 regarding the need for a stronger scientific basis of APHIS analysis, the committee recommends that the APHIS decision-making process could be made significantly more transparent, thorough, accurate, and scientifically robust by enhanced scientific peer review, active solicitation of public input, and development of determination documents with more explicit presentation of data, methods, analyses, and interpretations. As discussed in Chapter 2, there are two roles for risk assessment— technical decision support and creation of legitimacy in the regulatory process. The analysis here is divided into two main sections that relate to these two roles. First, the committee analyzes how APHIS has involved the public in the development of its risk assessment process and its specific rulings. Then, the technical approaches taken by APHIS to support its decisions are analyzed. ANALYSIS OF PUBLIC INVOLVEMENT The issue of public involvement in APHIS’s decision making is complex and must be accorded a somewhat summary, even ancillary, role in the present analysis. A thorough analysis would require a book of its own! This issue is an important one because public involvement in the regulatory decision-making process is desired on at least three counts (NRC 1996): Public involvement in government rule making is required by basic principles of democracy. Government authority ultimately rests on the consent of the governed, and it is desirable for public agencies to find appropriate ways to ensure that decisions are consistent with this principle. Opportunities for public involvement can broaden the basis of information on which regulatory decisions are made, improving the quality of decision making. Research on environmental risk indicates that public confidence in environmental policy making is particularly sensitive to the opportunity for concerned citizens to be involved in the decision-making process. Currently, APHIS policies for public involvement conform to a fairly narrow interpretation of those required by the Federal Administrative Procedures Act. It is useful to summarize involvement at two distinct levels: First, who is involved or has input into decision making in the notification, permit, and petition processes?

OCR for page 167
Environmental Effects of Transgenic Plants: The Scope and Adequacy of Regulation Second, who is involved or has input when policy for review of these decisions in being established? With regard to both questions, APHIS may choose to solicit advice from sources deemed useful by agency staff (although the short turnaround required for notifications may limit the time to obtain that advice). Also, the widest level of public input presently solicited by APHIS is achieved through publication in the Federal Register of intent to deregulate, issue permits, or alter internal APHIS procedures. APHIS generally receives comments on these Federal Register notices for a limited period and is required to issue responses to all comments that express disagreement with intended action. External Input into the Decision-Making Process Table 5.1 summarizes the current range of external participation in APHIS’s deliberative process for characterizing and evaluating risks asso TABLE 5.1 APHIS Involvement of Potential Participant Groups in the Process of Commercializing a Transgenic Plant or Its Products   Applicants External Experts Critics Consumers Users Petition for Nonregulated status   Hazard identification I I N N N Risk measurement I I-N N N N Making a decisiona F F F F F Risk management I-N N N N N Permit for release   Hazard identification I N N N N Risk measurement I N N N N Making a decisionb I I I I I Risk management I N N N N Notification   Hazard identification I N N N N Risk measurement I N N N N Making a decision I N N N N Risk management I N N N N The process for involving interested or affected parties is designated as F, formal process; I, informal process; N, no process available. aNotification of an impending decision must be published in the Federal Register, requesting comments from interested parties. bIssuance of a permit is under the authority of the Secretary of Agriculture and does not require publication in the Federal Register. When permits are announced interested or affected parties can request the related decision documents.

OCR for page 167
Environmental Effects of Transgenic Plants: The Scope and Adequacy of Regulation ciated with regulated transgenic plants. There are three possible paths to commercialization in which external input could be considered: An applicant can petition for nonregulated status, which means the transgenic plant can be released without regulatory oversight. (See Chapter 3 for details of the petitioning process.) At present this avenue is the most common pathway to commercialization. As of the end of 2000, APHIS had received a total of 73 petitions for deregulation. Of these, 52 petitions had been approved, 19 had been withdrawn, two were incomplete, and none had been denied. APHIS works with applicants to clarify the standards for approval. As a consequence, petitions are improved, sometimes substantially, prior to an APHIS decision. The lack of denials could be related to this clarification process, which among other functions serves to inform applicants of the likelihood of approval of petitions. An applicant can commercialize the products of a regulated article (but not the live plant itself) under the permitting system. (See Chapter 3 for details of the permitting process.) The permitting system is also used to accumulate data needed for commercialization through the first path. However, based on the relatively small number of field tests conducted under permit, this approach is rarely used. In the year 2000, only 36 field tests were performed under permit. A final path for developing commercial products from a regulated article is through use of the notification system. One drawback to this approach is that APHIS may, at any time, require permits for applicants who are growing a regulated article under notification. (See Chapter 3 for details of the notification process.) Notification is also currently the primary route for field testing to gather information needed for commercialization through the first path. In the year 2000, almost 900 field tests were conducted under notification, and several products have been commercialized. APHIS publishes its intent to deregulate or permit (if it conducts an environmental assessment for the latter) in the Federal Register. Comment periods are summarized in Chapter 3. In the case of notifications, the agency does not publish each notification separately or in a published list. Instead, it maintains an updated list of all submitted notifications (Federal Register 1992). APHIS periodically publishes a Federal Register notice announcing the list’s availability (but apparently has not done so for several years). On request, the agency provides the list directly to interested parties. APHIS has made arrangements with the National Biological Impact Assessment Program to make the notification list available online (see “Field Test Releases in the U.S.,” Information Systems for Biotechnology database: www.nbiap.vt.edu). Note that there is essentially no opportunity for either general public comment or external scientific review on notification decisions prior to a decision being made.

OCR for page 167
Environmental Effects of Transgenic Plants: The Scope and Adequacy of Regulation Finding 5.1: The notification process involves no public input. Beyond publication in the Federal Register, the legislative authority under which APHIS regulates transgenic plants does not stipulate formal mechanisms for involving external participants. However, the agency has developed guidelines for petitioning for nonregulated status and for submitting notifications. These guidelines provide for deliberative interactions between APHIS and the parties seeking to commercialize the product (see Chapter 3 for details). APHIS sometimes uses informal processes to contact outside experts to solicit information (see Chapter 4, squash case study), but it also has the authority to convene such groups formally to provide external advice. External Input into the Establishment of Policy Table 5.2 summarizes previous involvement of potential participant groups in APHIS policy-making processes, including developing risk analysis procedures and procedures for commercializing a transgenic plant or its products. Three of the primary processes are listed. As discussed in Chapter 3, APHIS once used a subcommittee of the U.S. Department of Agriculture’s Agricultural Biotechnology Research Advisory Committee (USDA-ABRAC) to provide advice for initiating the notification process. The agency has also held a number of national and regional meetings to discuss its policies with stakeholders. The rationale for APHIS risk management activities is determined largely by statutory authority. Because the U.S. Coordinated Framework for the Regulation of Biotechnology was developed under statutes put in place before the advent of transgenics, no public debate or congressional testimony specifically relating to the environmental risks or public confidence issues associated with transgenic plants occurred in connection with the creation of APHIS statutory authority. Effectiveness of Efforts to Solicit External Input The processes summarized in Tables 5.1 and 5.2 are somewhat effective in ensuring democratic decision making and improving the quality of APHIS’s decisions, as noted earlier. Certainly, publication in the Federal Register in compliance with the Federal Administrative Procedures Act is intended to provide at least some degree of democratic legitimization of the general rule-making process. Although theorists of democracy and governmental procedure continue to debate the adequacy of this approach to public participation, debate transcends APHIS policies and the appropriate focus of this report. In addition, APHIS has been somewhat successful at using public input to improve the quality of its decisions, as

OCR for page 167
Environmental Effects of Transgenic Plants: The Scope and Adequacy of Regulation TABLE 5.2 Public Involvement in Policy Making for Risk Assessment and Management   Applicants External Experts Critics Consumers Users Permit for release (1987)   Hazard identification N N N N N Risk measurement N N N N N Making the decision to establish systema F F F F F Notification system and petition for nonregulated status (1993)   Hazard identificationb I I I I I Risk measurement I I I I I Making the decision to establish systemc F F F F F Expansion of notification system (1997)   Hazard identification N N N N N Risk measurement N N N N N Making the decision to establish systemd F F F F F The process for involving interested or affected parties is designated as F, formal process; I, informal process; N, no process available. aProposal was published in the Federal Register, where anyone could make comments. bDuring the 1993 proposal, APHIS co-convened an expert panel to review the proposal. APHIS also presented the proposal at several public meetings during which interested or affected parties could comment. These procedures were not repeated during the 1997 proposal. cProposal was published in the Federal Register where anyone could make comments. dProposal was published in the Federal Register where anyone could make comments. noted earlier in the virus-resistant squash case and the involvement of the ABRAC. However, the third desired outcome of public involvement— improving public confidence in the decision-making process—has not received sufficient attention. Effectiveness of External Input on Specific Decisions APHIS’s informal procedures for consulting with outside scientific experts and the record of Federal Register notices, public comments, and APHIS responses to comments suggest that the agency’s approach has been useful in some specific cases for assembling technical information needed to exercise its regulatory authority regarding permits and peti-

OCR for page 167
Environmental Effects of Transgenic Plants: The Scope and Adequacy of Regulation tions. Comments on a number of permits have elicited detailed technical responses from APHIS staff and on several occasions have resulted in requests for additional data or additional studies on the likely risks posed by issuance of a permit (see deregulation case study of virus-resistant squash in Chapter 4). The effectiveness of the public comment period for this purpose depends on an informal network of people and organizations that monitor the Federal Register and generate or recruit comments from people having relevant expertise. The committee notes that external input through the Federal Register publications, whether from scientists or the public at large, has dwindled over the years. In information provided by APHIS, the committee found that they received a total of 378 comments for the first 10 petitions they considered (from 92-196-01p to 94-319-01p); the greatest number was for the second petition involving virus-resistant squash (92-204-01p, see Chapter 4). In contrast, for the 10 most recently approved petitions (from 97-287-01p to 99-173-01p), APHIS received a total of 11 comments for 3 of them, and no comments for the others. It is possible that the decline in comments is due to improved APHIS decision making, or a decrease in interest on the part of external scientists and the public. In the case of decreasing numbers of negative comments, another possibility is that frustration with the process may have resulted in declining public involvement in this specific process. Indeed, information provided by members of the public interest community confirmed that their perceived lack of responsiveness of APHIS to the comments they provided during deliberations over the transgenic squash petition (92-204-01p) marked a “watershed” for them in which they felt their efforts proved a “waste of time” and that in the future their efforts were better spent on activities other than writing comments to APHIS. No matter what the actual cause, APHIS may be losing potentially valuable public input. The lack of input from outside scientists in precedent-setting decisions is especially problematic. In contrast, the Environmental Protection Agency (EPA) often convenes formal scientific advisory panels to provide information that will help improve the technical rigor of its decisions. The Federal Register comment process has several weaknesses for the purpose of eliciting public involvement. First, Federal Register notices are often quite technical and not written in a manner that is accessible to the lay public. Second, when public comments have been issued regarding questions about confidence in the regulatory process for biotechnology, APHIS’s responses have largely been perfunctory. For example, early in the history of permits a number of comments were submitted expressing the view that the U.S. regulatory approach was incomplete or had gaps. The APHIS response was either to simply note disagreement with these opinions or to reply that they were not relevant to the particular decision

OCR for page 167
Environmental Effects of Transgenic Plants: The Scope and Adequacy of Regulation at hand. These responses are adequate within a narrow interpretation of administrative procedures, but APHIS could have sought alternative methods of involving the public. As noted above, the agency has held public meetings to inform the public about its policies, but such meetings reach a very small audience. Recommendation 5.1: For precedent-setting decisions involving permits and petitions, APHIS should actively solicit external scientific review. External Input into the Establishment of Policy As reviewed in Chapter 3, the present regulatory system used by APHIS was established in a series of policy-setting decisions. Starting in 1987 with establishment of the permit system, with its risk-based process regulatory trigger, several subsequent decisions excluded certain transgenic organisms from regulation. Two significant policy-setting decisions were the 1993 decision to establish a notification system and a process to petition for nonregulated status (the deregulation decision) and the 1997 decision to expand the notification system. APHIS did not use any formal processes (i.e., processes that are part of APHIS written policy) beyond the use of the Federal Register to announce 60-day public comment periods for involving the public in these decisions. For the 1993 decision, the agency used informal mechanisms to solicit public input, which included several public workshops and a scientific review by a subcommittee of ABRAC to evaluate the scientific merits of the 1993 proposal for a notification system. Increased utilization of a system of external scientific review of important policy decisions could help in securing greater public confidence in the regulatory process. Recommendation 5.2: For changes in regulatory policy, APHIS should convene a scientific advisory group to review proposed changes. Input of the general public in the process of policy development has been limited to the same mechanisms used in soliciting input on specific regulatory decisions. Actively broadening the basis for public involvement in environmental decision making is a difficult and potentially expensive proposition. Among the strategies that have been used by other agencies are advisory committees with representation from a number of self-identified advocacy groups, and public hearings. Outside the United States other approaches for distributing information and eliciting responses from the interested public have included consensus councils, blue

OCR for page 167
Environmental Effects of Transgenic Plants: The Scope and Adequacy of Regulation ribbon committees, standing committees, external contracts, and public jury approaches. Some European countries use standing committees of public-sector scientists or mixtures of public- and private-sector scientists to provide wider deliberations on critical issues. Others contract with public-sector scientists to provide detailed critiques of materials entering into the decision process. The 2001 revision of European Union’s Directive 90/220 specifies that applicants perform the risk assessment (unlike in the United States, where a government agency performs the risk assessment). This allows the European regulatory community to solicit evaluations of these assessments using external scientists in ways that would be unwieldy for APHIS. None of these efforts at involving the public can be undertaken without considerable cost in terms of both money to actually carry out the exercises and staff time and potentially needless delays in the regulatory process. Arguably, the effect of efforts undertaken outside the United States to involve greater segments of the public has been to considerably broaden the basis of decision making beyond that of the scientific disciplines deemed relevant to the interpretation and measurement of environmental risk by APHIS staff. Whether this other approach to decision making reflects greater sensitivity to alternative perspectives, including consumer and citizen desires, or an intrusion of inappropriate nonscientific viewpoints regarding what should be a science-based regulatory process is the subject of continuing debate. Finding 5.2: There is a need to actively involve more groups of interested or affected parties in the risk analysis process while maintaining a scientific basis for decisions. TECHNICAL ANALYSIS OF APHIS OVERSIGHT There is no question that the APHIS regulatory processes that have evolved to handle transgenic plants have resulted in a system that allows for thousands of field tests and dozens of deregulations of transgenic plants while at the same time allowing for scientific scrutiny of the regulated articles. Nonetheless, there are opportunities for improvement of this system. This section starts with general comments relevant to all APHIS decision-making processes—notification, permit, and petition. It then focuses on each process separately. The most important APHIS decisions regarding transgenic plants are those made during the notification and deregulation processes, because of the hundreds of field releases of transgenic plants under notification that occur every year and the fact that deregula-

OCR for page 167
Environmental Effects of Transgenic Plants: The Scope and Adequacy of Regulation tion removes a transgenic plant from any further APHIS oversight. Because of the major differences between the notification and the petition for deregulation processes, each of these processes is considered individually after making a few general comments about the overall APHIS regulatory process. Comments also are provided on the third, now relatively rare, permitting process. That commentary is placed between notification and petition because the permitting process shares certain features with both. It is noted that the permitting process may become more common as it is the only route to commercialization of products from plants intentionally grown to produce pharmaceutical compounds. General Comments and Concerns Geopolitical Scale The scope of APHIS oversight is limited to whether and how transgenic plants are moved and released in the United States. Although APHIS determinations sometimes include consideration of environmental impacts of deregulated genetically modified organisms outside the confines of the United States (see the case studies in Chapter 4), the agency is under no obligation to do so. Indeed, relevant scientific data may be unavailable for many countries that are centers of diversity for wild relatives of a transgenic crop. Yet these locations might be the very places where environmental impacts might occur due to transgene flow into those populations. Also, once a transgenic plant is deregulated, descendants of that plant may find themselves intentionally or unintentionally transported far beyond the borders of the United States, to radically different environments. A transgenic crop variety developed to fit into U.S. systems of agriculture may cause changes in the agricultural systems of other countries that would cause environmental degradation. One can imagine an argument being made by certain stakeholders, that if the U.S. government found a plant to be safe, that judgment should be good enough for a country without the resources to conduct its own environmental analysis. That would be wrong. Just because APHIS finds a transgenic plant to have no significant impact in the United States is not a guarantee that it will not have an impact elsewhere. To its credit, APHIS has held biosafety meetings in a number of developing countries to inform policymakers of the potential environmental effects of transgenic plants. Finding 5.3: APHIS’s environmental assessments should be interpreted as being confined to the evaluation of effects occurring in the United States.

OCR for page 167
Environmental Effects of Transgenic Plants: The Scope and Adequacy of Regulation CBI in APHIS Decision Documents Many people, including those generally supportive of biotechnology, decry the apparently large amount of data and information in submissions marked “CBI” (confidential business information). Under this CBI stamp, all manner of data are hidden from public view and even from independent scientific scrutiny. This form of business information protection is not unique to the biotechnology industry and those who monitor the applications for approval of new pesticides by the agrochemical industry are faced with similar challenges. This committee sometimes found that it could not provide an independent scientific assessment of APHIS rulings because of the broad use of CBI. Clearly, businesses have the right and the need to protect sensitive information from their competitors. But APHIS does not have a mechanism to distinguish what is truly competitive information (and worthy of confidence) from less sensitive information that might be of value in the public discourse and open to scientific analysis. The lack of such a mechanism may be related to issues much wider than the APHIS regulatory system. However, its absence creates several contradictions. Without a transparent mechanism to evaluate and judge confidential data from applicants, APHIS appears to accept without question an applicant’s assertion of what is CBI. Public credibility is eroded when the same information marked CBI in APHIS documents is not considered CBI and is open to public inspection in other jurisdictions, such as Canada or Europe. It must be understood that U.S. standards for what constitutes CBI are very broad. A company can claim that certain information is CBI as long as it can show there is any possibility that disclosure of the information could directly or indirectly harm its business. In this legal sense APHIS may have little power to limit CBI. Another interpretation of the degree of CBI in documents sent to APHIS by applicants is that the agency is not working to provide as much information as possible to the public. Regardless of the accuracy of either of these interpretations, the extent of CBI in these documents makes public assessment of APHIS decisions extremely difficult. Finding 5.4: The extent of CBI in registrant documents sent to APHIS hampers external review and transparency of the decision-making process. Resistance Risk and Non-target Effects For pesticidal plants, both APHIS and EPA examine risks of nontarget effects and the potential for pests to evolve resistance to the pesticidal component of the plant. The committee examined APHIS’s approach

OCR for page 167
Environmental Effects of Transgenic Plants: The Scope and Adequacy of Regulation target species can be regulated by notification. For example, toxins in pollen that harm non-target species could be regulated by notification. As indicated above, it is often difficult to ascertain many details about a tested plant from the published notifications. The committee did, however, find one case where enough information was available to result in questioning the rigor with which the restrictions on the notification process are enforced. This case is presented in BOX 5.1 as an example, with the caveat that the committee did not have access to all of the information available to APHIS. The committee concluded that the plant tested had the potential for toxicity to a broad array of organisms both in the field and after the plant was harvested. Based on this information, the commit- target species that feed on the plant, not dispersed plant parts, such as seeds, pollen, or plant residue. Even using this restricted definition, it is difficult to understand how the avidin-producing corn could qualify for commercialization under notification because the corn kernels are toxic to many diverse insect species (Kramer et al. 2000). Allergenicity is not one of the characteristics considered in accepting a notification package. While avidin has no history of causing allergies (Langeland 1983), the notification procedure used by APHIS does not require any reporting about potential allergenicity, and any such risks would not be considered as a part of a notification package. In other words, even if avidin were allergenic, the applicant could grow and commercialize the product under the APH-IS notification system without any additional oversight. Based on conversations with John Howard of ProdiGene and APHIS personnel, avidin corn appears to be well contained in current commercial plantings. All of the avidin corn is planted by a single contractor on less than 5 acres of land. Furthermore, biological containment of the avidin transgene is reinforced by its toxicity to specific corn tissues. Avidin-producing plants that express high levels of avidin are male sterile, presumably due to the toxicity of the avidin to pollen-producing tissues of the corn plant. These containment methods are more stringent than required under the APHIS notification procedure. Under the notification system, an applicant only has to show that progeny from any gene flow to other crop plants will not survive in the environment. Processing corn for animal feed or human food is one way suggested by APHIS to ensure that these crop plants will not survive in the environment. Such procedures clearly will not ensure that risky transgene products will not enter human foods if the transgenic crop is commercialized under the APHIS notification system. It is worth noting that the first field tests of avidin-producing corn were carried out in 1993 based on an APHIS permit that had been applied for in 1992, before the notification system was established. While there do not seem to be any unacceptable environmental risks in the current production of avidin corn, this case points to a lack of rigor in the notification process that could lead to problems with future transgenic plants. For example, it would be possible, under notification, to grow thousands of acres of a transgenic crop that produced a substance that was allergenic or toxic to livestock or humans after seed was harvested. There is a clear need for APHIS to reassess its notification process given the novel products likely to be produced by transgenic crops in the future.

OCR for page 167
Environmental Effects of Transgenic Plants: The Scope and Adequacy of Regulation tee does not understand why APHIS has repeatedly allowed field testing of this transgenic plant under the notification system. The fact that the substance coded for by the transgene has the potential to be an allergen was of special interest. Examination of APHIS guidelines revealed that plants with allergenic properties can be grown under notification. The committee questions the wisdom of allowing such plants to be grown under the streamlined notification system. Regulation of novel compounds under notification raises the more general question of how APHIS determines that a compound is not toxic to non-target species in the field, given the limited information that must be provided by applicants. The committee’s comments on the APHIS procedures are not meant to imply that there is any risk in this specific case where the applicant seems to have taken extra steps to ensure human and environmental safety. Finding 5.7: It appears that a transgenic plant with toxic properties to non-targets was grown to create a commercial product under the notification process. APHIS personnel must process the notification applications very rapidly (see Chapter 3). The total number of APHIS personnel available to process notifications, permits, and petitions appears to be insufficient. Currently there are about 10 permanent APHIS biotechnology evaluations staff. From the time the committee began its study in July 2000 until July 2001, APHIS always had staff vacancies, and staff turnover rates appear to be high. This situation results in a heavy per-person workload. The heavy workload coupled with no public feedback may detrimentally affect the rigor of the determinations. The number of APHIS personnel who conduct field visits to sites under notification and permit is also small. This includes the permanent APHIS Biotechnology, Biologics, and Environmental Protection (BBEP) staff plus APHIS field personnel who are not all trained to understand the implications of the evaluations they are making. (To the credit of APHIS, it is clear that these field personnel are, on average, better trained today than they were five years ago. Turnover in field personnel makes maintaining adequate training levels difficult.) To maintain compliance with performance standards, sufficient numbers of appropriately trained personnel must be allocated to visit field sites. Only a subset of notification sites receives an inspection visit. In field seasons such as 2001, when APHIS field inspectors have an emergency priority (of inspecting for foot and mouth disease), few may be available to visit notification sites. Recommendation 5.4: Because of the large number of field tests conducted, resources for compliance monitoring are necessary to maintain a suitable number of well-trained APHIS field inspectors.

OCR for page 167
Environmental Effects of Transgenic Plants: The Scope and Adequacy of Regulation Technical Analysis of the Permitting Process As noted earlier, the permitting process is currently a relatively rare APHIS decision-making process for transgenic plants, with only a few dozen such decisions per year. Prior to introduction of the notification process, the permitting process was the only regulatory mechanism for conducting field trials, so the total number of permit decisions associated with field tests numbers over a thousand. Similar to the petition process, permit applications contain more than the simple list required for notification—that is, they provide some scientific information and analysis. The information provided by the applicant is primarily for the purpose of assuring appropriate confinement and disposal of the transgenic plants. Most permit applications, like all notification applications, are not listed in the Federal Register during the process of APHIS assessment unless they include an Environmental Assessment as required under the National Environmental Protection Act. From 1996 to the present, APHIS appears to have conducted environmental assessments for about 6% of the permit applications processed. If APHIS does not issue an environmental assessment reporting on its environmental risk assessment, there is no potential for feedback from stakeholders except the applicant. And even when an assessment is issued, feedback from interested and affected parties could be frustrated by information that is limited because of CBI. Plants grown under permit share certain characteristics with plants grown under notification. There are no restrictions to the acreage under which they may be grown, and they may be grown to produce nonliving commercial products. As noted elsewhere in this report, it is anticipated that transgenic plants intentionally grown to produce commercial pharmaceutical products will be grown under permit. Thus, problems associated with the scale of the plantation discussed for notification may also apply to plants grown under permit. However, while APHIS personnel visit only a subset of notification field sites, all permit sites are visited. The committee’s findings and recommendations under the section on notifications generally apply to the permit process as well. Technical Analysis of the Petition Process If APHIS approves a petition for deregulated status, deregulation is absolute. APHIS generally indicates that it cannot “conditionally,” “temporarily,” or “partially” deregulate. For example, APHIS cannot deregulate an article but require monitoring of it. Similarly, APHIS deregulates a regulated article and all of its descendants. As noted in Chapter 3, trans-

OCR for page 167
Environmental Effects of Transgenic Plants: The Scope and Adequacy of Regulation genes can be intentionally moved by crossbreeding into other varieties of the same species or even other species. Because the plants are fully deregulated, one can imagine that the transgenes could be used for purposes other than originally intended. For example, without any further APHIS consideration, a selectable marker transgene that confers herbicide tolerance in a deregulated article could be transferred to another variety or another species to create a new variety to be sold as a herbicide-tolerant product. As a hypothetical example, if a male-sterile poplar tree with a Bt gene were deregulated, it would be possible to cross the Bt gene into a male-fertile poplar without further regulation. Although APHIS generally indicates that it cannot “partially” deregulate an article, there is some inconsistency in the agency’s documents. For example, in the case of the 1997 Bt herbicide-tolerant cotton petition discussed in Chapter 4, the APHIS assessment formally considered only two alternative actions: “no action,” which would mean refusal to grant nonregulated status or a determination of a “finding of no significant impact” which would result in complete deregulation. APHIS found no significant impact in this case and deregulated this article in the entire United States. This resulted in a ruling that differed with one by the EPA that restricted the growing of cotton in some areas of Florida and in Hawaii. This is a problematic outcome because the U.S. Coordinated Framework for the Regulation of Biotechnology indicates that all regulatory agencies should regulate transgenic plants with similar scrutiny. Finding 5.8: APHIS and EPA made different decisions on the planting range of Bt cotton in the United States. The use of two alternatives in the APHIS assessment of this cotton petition contrasts with an APHIS environmental assessment of a Bt corn petition that was also reviewed in 1997 (96-317-01p). In the corn assessment, APHIS presents three alternative actions. In addition to the two actions listed in the cotton case, the additional action listed is to “approve the petition with geographical limitation.” If APHIS had used this approach with cotton, it could have provided a ruling more similar to that of EPA. Finding 5.9: APHIS appears to be inconsistent in reporting its authority to deregulate on a geographic basis in the United States. The need for a case-by-case assessment of the environmental effects of transgenic plants has been emphasized repeatedly (NRC 1987, 1989, 2000c) because even the same transgenic trait in two different plants may pose different concerns. Because each plant/trait/environment combination is different (see Chapter 2) and because our understanding of genetics and ecology is still developing, it is important for APHIS to resist a

OCR for page 167
Environmental Effects of Transgenic Plants: The Scope and Adequacy of Regulation rigid “checklist” approach suggesting specific data needed for petitions. While such data requirements might be convenient from the regulator’s viewpoint, this approach could be perceived as onerous busywork from the applicant’s perspective, and as an indication of bureaucratic rubber stamping from the activist communities’ perspective. Beyond simple identification information, the only information provided to APHIS should be useful in identifying and determining risk. The data required of APHIS should be associated with the product (phenotype), not the process. Details of the molecular biology of the transgenic organism are helpful only when they define the phenotype of that organism in the context of exploring a potential risk. For example, it is important that APHIS be provided with the complete sequence of expressed transgenes in order to provide assurance that only the coding sequences of interest will be expressed in the transgenic plant. Evidence to date indicates that molecular gene transfer techniques often give complicated insertions of the transgene into the genome. Transgenics can be selected, however, that give Mendelian ratios and behave genetically in a normal single-gene fashion. Nevertheless, instead of a straightforward insertion of an unmodified DNA sequence, the transgene may be inserted in a rearranged manner, as multiple copies, or both. Recent studies have shown that host DNA of unknown origin may separate transgenes or parts thereof (Takano et al. 1997; Kohli et al. 1998, 1999; Pawlowski and Somers 1998; Jackson et al. 2001). An open reading frame (ORF) can be created (Somers, 2001, University of Minnesota, personal communication) at some point in the process of transfer or insertion. The possibility exists that an ORF could lead to expression, although the likelihood is remote that such an RNA or subsequent protein product, if produced, could have any negative consequences. As cloning technologies improve, however, the ideal would be to clone all transgene components and be certain that no unexpected gene products would be produced. At present, cloning the complex insertion(s) is not a trivial matter. However, it would seem prudent to encourage the sequencing of inserts whenever possible and to include such information for review by the appropriate regulatory agency. Striving for a gene transfer technology that provides simpler inserts is important for the future. Agrobacteriummediated gene transfer may provide simpler insertions than particle bombardment, and new ideas on how to further reduce disruptive events in the transformation process are emerging (Koprek et al. 2001). Gene replacement technologies may become common in the future and perhaps alleviate the concern of producing new ORFs. Of course, spontaneous gene mutation happens at a low frequency, so any sequence can occasionally be modified to produce a new ORF. How often this happens in nontransgenics is not known.

OCR for page 167
Environmental Effects of Transgenic Plants: The Scope and Adequacy of Regulation This committee recommends that as much transgene sequence information as is reasonably feasible be reported as part of an application. Because the transgene region might be quite complicated, a protein product may or may not be produced, and the product may be inconsequential, any recommendation to require sequencing must remain flexible until much easier DNA sequencing and other technologies are available. For these reasons, the committee recommends that APHIS should require reporting of full DNA sequences of transgenes as they are integrated in the plant genome unless the applicant can provide scientific or technological justification not to do so. Recommendation 5.5: APHIS should require reporting of full DNA sequences of transgenes as they are integrated in the plant genome unless the applicant can provide scientific or technological justification not to do so. APHIS’s guidelines for applications for determination of nonregulatory status provide a suggested format for applicants to follow. Some of the features of this format are described in Chapter 3. While APHIS does not exactly give the applicant a checklist, it does come close to that in some places. In reading the guidelines this committee was often unsure of why some kinds of information were requested at all and was also unsure about what kind of evidence would be needed to respond to some of the questions asked. For example, the document states that: Applicants must report any differences noted between transgenic and nontransgenic plants that are not directly attributed to the expected phenotype. Differences observed could include changes in leaf morphology, pollen viability, seed germination rates, changes in overwintering capabilities, insect susceptibilities, diseases resistance, yield, agronomic performance, etc. Applicants must also note the types of characteristics that were compared between transgenic and nontransgenic plants and found to be unchanged. While this paragraph offers some general guidance, what is missing is information on the standards of evidence. Does an assessment of overwintering capability mean performing quantitative tests in a dozen distinct habitats for three years, or does it simply mean not noticing any major change in overwintering while conducting yield trials? While leaf morphology could be measured in the process of typical agronomic trials, measuring disease resistance could require special testing with a set of disease organisms under the environmental conditions best suited for a disease outbreak of each organism. The applicant could use common sense in determining what to do or could consult with APHIS personnel. The problem is that the answer to the question may depend on the APHIS

OCR for page 167
Environmental Effects of Transgenic Plants: The Scope and Adequacy of Regulation scientists’ own value system and on their knowledge of the phenotype in question. Of course, the importance and likelihood of unexpected changes in disease resistance will depend on the plant being examined. A similar concern about evidential standards is raised in the case study of virus-resistant squash presented in Chapter 4. Here the question was how to determine if gene flow of the virus resistance-conferring gene from the transgenic variety to weedy relatives of squash could change their level of weediness. In this case it was clear that regulatory personnel and external experts differed in what they considered sufficient evidence. The four case studies present in Chapter 4 also indicate that the evidence considered to be sufficient differs among petitions. As indicated above, some variation is explained by the case-specific nature of the risk assessments and the APHIS learning process from 1994 through 1997. However, some of the variation in rigor is hard to explain scientifically. Finding 5.10: From the committee’s assessment of APHIS guidelines and case studies, it appears that the agency does not provide a sufficient guide for evidential standards to applicants preparing petitions for deregulation. It also is not clear from the committee’s analysis that APHIS personnel have a system for matching evidential standards to the potential level of hazard and risk. Development of such evidential standards is not an easy task and may be best accomplished by multiple external experts in specific areas. As indicated in Chapter 2, a consensus of multiple external experts is likely to be more rigorous than the expert judgment of regulatory personnel because disagreements among external experts typically lead to more robust risk assessments. Furthermore, no single person or small groups of people is likely to have expertise in all of the areas needed to assess the risks of transgenic plants. The heavy workload of APHIS personnel was discussed earlier. Regarding expertise among APHIS-BBEP personnel, the committee’s assessment of their professional backgrounds indicates an appropriate number of the personnel with training in molecular biology and related fields, but too few with any formal training in ecology or population genetics. Greater diversity in areas of expertise could be helpful to this group. Finding 5.11: APHIS is understaffed, and the committee questions the match between the scientific areas of staff training and staff members’ responsibilities. Recommendation 5.6: APHIS needs to improve the balance between the scientific areas of staff training and the job responsibilities of the BBEP unit by increasing staff and making appropriate hires.

OCR for page 167
Environmental Effects of Transgenic Plants: The Scope and Adequacy of Regulation Even if the APHIS staff were larger and more balanced in expertise, an external scientific peer review process would still likely raise and address issues that would be missed by the staff. Peer review could help assess whether the data presented by an applicant were adequate to address specific risks or whether more data were necessary. This would be especially important when novel traits were being assessed. A bona fide analytical-deliberative process should involve risk identification and data evaluation by the applicant, regulators, and interested and affected parties. In many ways the squash case study exemplifies the best analytical-deliberative approach of the case studies presented. The risks are identified by the applicant and addressed. APHIS’s multiple Federal Register announcements and requests for external expert advice created an iterative process of analysis and deliberation. The fact that the process was controversial is, in part, a product of public involvement. The committee has identified more unanswered questions for the corn and cotton case studies discussed earlier, even though they are more recent than the squash case study because, in part, they were not as controversial and were processed without public or external scientist involvement. Apparently, with more stakeholder involvement and external expert information, fewer questions are left unanswered. The information necessary to determine whether a transgenic plant poses an environmental risk is like a moving target. Experience and information accumulated with time have changed our collective judgments of environmental impacts. For example, seven years ago when the first Bt corn petition was reviewed, APHIS did not consider adequately the potential effects of pollen from Bt corn to harm non-target Lepidoptera (see Chapter 4). Publication of a paper showing that Bt pollen was a hazard to monarch butterfly larvae aroused great public concern (Losey et al. 1999). Recent studies now show that, although Bt pollen is hazardous to monarch butterfly larvae, the risk is likely to be low but more research is still needed (BOX 2.1). Changes in both scientific information and social values affect the amount of attention given to specific risks. As discussed in Chapter 1, society’s view of the interaction between agricultural and nonagricultural systems has been changing, and more attention is now focused on these interactions. One issue that has recently been getting more attention is the effects of large-scale planting of transgenic crops. There is concern that some environmental effects would not necessarily be detected in smallscale field tests. Indeed, there is a general problem of implicitly extrapolating from the results of small-scale field trials on the order of tens of acres to potential effects of large-scale commercial “grow outs” on the order of millions of acres. The committee did not encounter questions

OCR for page 167
Environmental Effects of Transgenic Plants: The Scope and Adequacy of Regulation about such changes in scale in APHIS assessments, but large-scale changes are the very kind of environmental changes that make up the primary impacts of modern agriculture (see Chapter 1). Finding 5.12: APHIS assessments of petitions for deregulation are largely based on environmental effects considered at small spatial scales. Potential effects from scale-up associated with commercialization are rarely considered. The focus on small-scale field and laboratory tests is evident in the types of tests typically conducted to assess impacts on non-target organisms. For crops that produce Bt toxins, the general types of tests used to assess non-target effects are laboratory toxicology tests similar to those used to assess conventional insecticides. The organism being tested is exposed to the toxin itself (or a closely related toxin) at concentrations 10 to 10,000 times higher than the organism will experience in the field. If the right organism is tested using an appropriate method of exposure (but see Hilbeck et al. 2000, NRC 2000c, Marvier 2001), such testing can be valuable. However, these elevated-dose, acute toxicity tests can miss the effects of a chemical on biological processes other than the one that causes the acute toxicity. For example, testing of many conventional pesticides at maximum tolerated doses did not reveal their estrogenic activity, which can be seen at much lower doses. Finding 5.13: The “toxicology-type” risk assessments used by petitioners are useful but not sufficient to assess the non-target risks of pesticidal crops. In addition to laboratory testing, there have been some short-term, small-scale field tests that have examined the effects of transgenic crops on invertebrate biodiversity. It would be possible for applicants to conduct more comprehensive field evaluations of transgenic plants for environmental effects prior to petitioning for nonregulated status than is currently done. A number of the case studies in Chapter 4 point out specific ways that these studies could be improved. However, even more comprehensive and relevant precommercialization field testing would still be limited to detecting effects on organism abundances and field characteristics that occur on small-time and spatial scales. Therefore, there is a need for a system of testing and monitoring after commercialization, when large-scale plantings begin. Development of such postcommercialization assessments is important, complex, and likely to be quite expensive. The next chapter is devoted to examining the conceptual and practical aspects of developing such postcommercialization assessments.

OCR for page 167
Environmental Effects of Transgenic Plants: The Scope and Adequacy of Regulation Changes in APHIS Oversight over Time As emphasized repeatedly throughout this report, the regulation of transgenic organisms is relatively new, and APHIS should be commended for being the first regulatory agency in the world to develop a regulatory framework for the oversight of transgenic organisms. Being first, however, also has its disadvantages because there were no prior regulatory models specific to transgenic crops from which to adapt the system of oversight. Consequently, APHIS-BBEP has had to adapt its procedures at the same time it was creating them—a challenging and sometimes unsettling process. For example, during 1987, the first year of operation under the newly approved regulatory system, only a handful of permits were received, and each could receive considerable attention from the staff. By 1991, APHIS-BBEP was receiving literally hundreds of permit applications, which greatly stressed the agency’s technical capacity, and the number of petitions was anticipated to increase in the future. Moreover, APHIS-BBEP found that the majority of the applications concentrated on corn, soybean, cotton, potato, tomato, and tobacco. In addition, it was found that confinement procedures required for each of these crops had many similarities. Consequently, in 1992 APHIS-BBEP proposed the development of a notification system so that these common applications, which had already undergone an environmental review under permitting, could receive less oversight. This system was implemented in 1993 and is the cornerstone of the APHIS regulatory process. In 1997 they expanded this notification system, which allows many transgenic plants to be planted without any environmental review. The committee discussed some of the scientific concerns and inconsistencies associated with this expanded process. This notification system provides great regulatory flexibility for APHIS and in principle has considerable scientific validity. The 1993 system, in particular, is an outstanding example of how a regulatory agency can learn from experience and adapt its regulatory procedures. As discussed in Chapter 2, transgenic crops have been assessed in the United States by APHIS assuming a static non-adaptive risk analysis framework, without systematic regard to the potential for mistakes, in either management or risk assessment. Certainly the system is designed to limit mistakes, but they do occur, and their significance is unknown outside the specific context in which they occurred. The committee discussed later in Chapter 2 several models that could be used to formally supplement the management framework in APHIS’s risk assessment practices. One such model—the fault tree-analysis—could be used to assess risks that could occur through failure of the current oversight system. Conducting such an analysis would provide APHIS

OCR for page 167
Environmental Effects of Transgenic Plants: The Scope and Adequacy of Regulation with considerable information to improve its regulatory procedures and could contribute to a more formal methodology for a learning system. One caveat regarding these models is that the cost of such an analysis is unknown, but might be high. Recommendation 5.7: APHIS needs to formalize its learning process. CONCLUSION The best regulatory decision making depends on using the best information available. The committee has identified two interweaving factors that could increase the amount of information available for the regulation of transgenic plants. First, it has identified that the flow of information to and from external scientists during the APHIS decision-making process could be improved if external input were more actively sought by the agency and if the impediments to flow of data by “confidential business information” were reduced. Second, the committee perceives that the small APHIS staff and their high workload precludes the opportunity to develop the flexibility and breadth necessary to deal with the complex, diverse, and evolving challenges of a growing workload, new products, new environmental questions, and increase in the spatial scale of commercial production.