After examining the perspectives of government agencies who oversee the implementation and monitoring of performance standards, the workshop featured two sets of panelists who presented various end-users’ perspectives. In the first session, Neil Lipman, Executive Director of the Center of Comparative Medicine and Pathology, Professor of Veterinary Medicine in Laboratory Medicine and Pathology at Weill Medical College of Cornell University, and Laboratory Member of Memorial Sloan Kettering Cancer Center, gave the perspective of someone working in academia; Mary Ann Vasbinder, Head of Corporate 3Rs Responsibility and Training Strategy in the Office of Animal Welfare, Ethics, and Strategy at GSK, provided a pharmaceutical industry perspective; and John Bryan II, Public Service Assistant and Wildlife Veterinarian at the Southeastern Cooperative Wildlife Disease Study (SCWDS), gave a wildlife biologist’s perspective. In the second session, Bart Carter, Director of Animal Resources and Attending Veterinarian for the University of Texas Southwestern Medical Center, provided an agricultural perspective; Kenneth Litwak, Laboratory Animal Advisor to the Animal Welfare Institute, spoke from a public interest perspective; and John Bradfield, Senior Director at the Association for Assessment and Accreditation of Laboratory Animal Care (AAALAC) International, gave the perspective of an international accrediting and assessment organization. Paul Locke, Associate Professor of Environmental Health Sciences at the Johns Hopkins University Bloomberg School of Public Health and Distinguished Visiting Professor of Animal Law and Science at Lewis and Clark Law School, then gave a brief synopsis of the six presentations. An open discussion, moderated by David Anderson, ended the workshop’s first day.
THE PERSPECTIVE OF U.S. ACADEMIA1
The Center of Comparative Medicine and Pathology (CCMP) is unusual, said Neil Lipman, in that it supports two separate and distinct institutions: Weill Cornell Medical College and Memorial Sloan Kettering Cancer Center, each with its own budget and its own IACUC. Typically, the facility houses a quarter of a million animals on any given day, has about 53,000 rodent cages, and is staffed by some 250 people working in 10 vivaria in New York and Doha, Qatar, with resources smaller institutions may not have. Lipman characterized CCMP as mouse-centric and highly harmonized, as all facilities use the same standard operating procedures. In addition, CCMP has a large multidisciplinary anatomic and clinical pathology laboratory to perform certain types of testing and a postdoctoral training program to provide the labor and skills to develop and assess a performance standard. The culture at CCMP, Lipman added, is to make evidence- and outcome-based decisions, while management is hypothesis and data-driven whenever possible.
From his perspective, engineering standards have both advantages and disadvantages, as previously mentioned. Major advantages include the implementation of a nominal baseline; some level of standardization and consistency; and ease and economy of implementation. Ideally, engineering standards are based on objective, scientific data, which Lipman said is not always the case. In contrast, poor-quality data leads to tight and inflexible engineering standards, which restrict progress and slow the development of new knowledge. They often are empirical and anthropomorphic, he said, and influenced by politics rather than sound information. When not well prescribed, they can also result in unnecessary costs. Performance standards allow greater flexibility to also align with science to improve animal welfare and care. They can be tailored to an individual situation or institution, and they can refine and improve engineering standards. Despite these characteristics, performance standards are expensive to properly develop, a challenge for resource-limited Animal Care and Use Programs. They can also be subjective, making it difficult to objectively determine the appropriate outcomes for assessment.
Lipman discussed two examples of CCMP’s operations to illustrate its three primary program drivers: assure animal health and welfare, support science, and provide high-quality customer service. The first involved
_________________________
1 This section is based on the presentation by Neil Lipman, and the statements are not endorsed or verified by the National Academies of Sciences, Engineering, and Medicine.
identifying and genotyping pre-weanling mice at an age that permits additional tissue collection. “There are limited methods available to reliably and permanently identify pre-weanling mice, and we have historically used toe-clipping at our institution,” said Lipman, who acknowledged that some feel that this technique is not humane. In fact, the U.K. Joint Working Group on Refinement of Production of Genetically Engineered Mice recommended using toe clipping as a last resort and performed under local anesthesia (BVAAWF/FRAME/RSPCA/UFAW Joint Working Group, 2003).
In 2010, two European publications (Castelhano-Carlos et al., 2010; Schaefer et al., 2010) indicated that toe-clipping may be the preferred method for neonatal mice up to age 7 days based on physiological and behavioral observations. These publications were referenced in the 8th edition of the Guide, which, Lipman believes, established a de facto engineering standard that warranted further study to create a new performance standard to minimize pain and distress; limit short- and long-term physiologic and behavioral effects; be safe for both animal and operator; be fast, easy, and permanent; and be cost-effective.
Lipman and his colleagues (Paluch et al., 2014) conducted a study to evaluate the effects of toe-clipping on the welfare of pre-weanling animals. Four groups of C57Bl/6J mice were toe-clipped on postnatal day (PND) 7 and four groups on PND 17. One group received and one did not receive a topical vapocoolant anesthetic, while two served as control groups. All were handled and restrained by an experienced investigator. The pups were observed for their reactions immediately after surgery, and at 1, 3, 5, 8 and 12 hours after clipping. The team conducted developmental assessments beginning at PND 6 (to obtain baseline values prior to clipping on PND 7) through PND 21 and behavioral assessments when the animals were adults. The results (Figures 4-1 and 4-2), he explained, showed no observable reactions in the PND 17 group compared to the control, but some signs of distress in the PND 7 group. As shown in Figure 2, the use of vapocolant created difficulties in both age groups. There were no differences in the developmental and behavioral results or weight gain among any of the groups. As anatomical studies showed the digits to be fully innervated and mostly calcified by PND 7, additional pain after PND 7 should not be expected, said Lipman. As a result of this study, CCMP revised its institutional guidelines and established new procedures, eliminating the use of vapocoolant; extending the recommended age for toe-clipping to greater than or equal to PND 17; limiting the procedure to one digit per day; and recommending the toe-clip site to be the distal first third of the first
phalanx. When toe clipping and tail biopsy are both needed, the procedure should be done between PND 14 and PND 17.
FIGURE 4-1 Immediate post-clipping observations (note: no reactions were observed in PND 17 groups 1 and 3)
SOURCE: Lipman slide 14
FIGURE 4-2 Additional data from post-clipping observations
SOURCE: Lipman slide 18
The second example Lipman discussed focused on developing a humane, safe, and efficient method of euthanizing mice that would induce rapid unconsciousness and death; avoid excitement; minimize fear, distress, and anxiety; minimize changes in the animal’s environment; and would be efficient and reproducible. At CCMP, staff worked with a caging manufacturer to develop a sophisticated engineering paradigm based on the use of a single type of ventilated cage that allows room air to be replaced with carbon dioxide. This system was validated, including videographic behavioral assessment, and has now been used successfully for over nine years (McIntyre et al., 2007). Since the release of the latest edition of the Guidelines for Euthanasia by the American Veterinary Medical Association (AVMA 2013) a conflict exists with regard to the optimal flow rate for carbon dioxide used to euthanize rodents that CCMP employs. “We have a system that we have validated that does not meet this recommendation, so what do we do?” said Lipman. The CCMP team compared its performance with the AVMA Guidelines recommendation and showed that the CCMP system produced significantly lower distress levels in both mice and rats.
Even though this information was shared with the system’s manufacturer, this performance standard was not widely adopted because it was not compliant with the AVMA guidelines, Lipman noted. The CCMP team then developed a new protocol for introducing carbon dioxide into the cages at a lower displacement rate to meet the AVMA guidelines. Today CCMP can use either standard.
In Lipman’s opinion, research funds to develop science-based standards for the care of laboratory animals have been few and are getting fewer. To answer some of the larger questions, such as those about cage space, it will be necessary to create multi-institutional consortia to secure the funds to pay for these large studies, he added.
THE PHARMACEUTICAL INDUSTRY PERSPECTIVE2
A unique feature of the pharmaceutical industry, said Mary Ann Vasbinder, is its obligation to answer to multiple regulatory organizations worldwide, creating a dynamic working environment, particularly given that views on animal welfare differ across the globe. As a result, a pharmaceutical company has to consider what kind of performance standards are maximally applicable across global operations, have the flexibility to accommodate its needs across all of the markets in which it operates, and have aligned outcomes so there is a sense of satisfaction from doing the right thing for the animal.
One difficulty in developing a performance standard, she said, is getting everyone involved to agree on outcomes and expectations, particularly when those can change over time. At GSK, an acceptable performance standard is outcomes-based, has associated metrics for success, and derives from facts and data. When there are not enough facts and data, expert advice comes into play, but using experts comes with its own challenges, said Vasbinder, particularly when experts have different opinions based on their culture and background. At GSK, areas considered for performance standards include veterinary care, environmental enrichment, and acclimation to study, training, exercise, and socialization.
As an example of how GSK develops and implements a performance standard, Vasbinder reviewed a project on dog housing, which the company undertook because every site within the company handled the issue of cage size differently. After developing a management plan for a
_________________________
2 This section is based on the presentation by Mary Ann Vasbinder, and the statements are not endorsed or verified by the National Academies of Sciences, Engineering, and Medicine.
project that would examine dog housing, exercise, socialization, and environmental enrichment, her team lined up sponsors who provided research funds and agreed to build the cages the team would design based on the research findings. The team also created a communication plan to convey information to its key stakeholders on a regular basis and get buy-in as the project proceeded. After completing a literature review, talking to others in the community to learn what they had done, and developing a benchmarking system, Vasbinder and her colleagues conducted pilot studies with flexible housing options suitable for dogs and miniature pigs. From these pilot studies, the team realized it needed advice from behaviorists and people experienced with caging, so Vasbinder recruited five individuals to list and then prioritize the things they thought would be most valued to a dog in its home environment. Social interaction, both between dogs and between dogs and humans, the opportunity to exercise, and the ability to have environmental enrichment were perceived to be more important than cage size.
Using the experts’ recommendations, the team built cages that would increase visibility for the dogs, optimize interactions with staff, and have a softer look than with standard cages. The trial cages incorporating these features were built to accommodate between four and six dogs, based on unpublished studies suggesting this number created a good balance between social interactions and fighting behavior. The cages were also designed, based on a strong recommendation from a behavioral expert, with enough space to enable dogs to trot within their enclosure. To benefit staff, the cages had to be easily cleaned and sanitized.
Vasbinder’s team then looked at the different exercise programs the company’s facilities were using and decided they wanted a dedicated area outside of the kennel and the main room for exercise. They also decided the dogs needed access to complexity – to be able to climb on things, to interact with toys, and have social interactions with people and other animals and with enough space to run. The team developed a performance standard for exercise of 15 minutes based on observing that when staff let the dogs out to exercise, the dogs were all lying on the floor or interacting socially rather than exercising after 15 minutes. Currently, the team is working on a performance standard regarding acclimation to the environment and another on how to introduce new dogs into the facility and condition them to a study.
The resulting cages have horizontal bars to increase visibility; Dutch doors to allow staff to pet the dogs, medicate them, and give them treats without letting them out of their cages; and benches the dogs can jump onto to further increase visibility. The cages have what Vasbinder called
alleyways through which dogs can see other dogs, and the cages can be connected to one another. The team further found that flooring with some grip to it improved the health of the dogs’ feet. The cages also have areas where a dog can escape social interactions and be on its own should it need a break.
Assessments of the first generation of cages led to some modifications that further improved animal health and reduced stereotypic behavior. Vasbinder noted that staff morale improved as well. Even though these new cages took longer to clean, staff enjoyed the ability to interact more with the animals. In reviewing the project, she said the performance standards she and her team developed have worked well and the outcomes from this project were positive, despite the lack of published literature on which to base some of the decisions the team made. Many of the decisions came down to observations, listening to the advice of experts, and applying common sense, but Vasbinder wondered if common sense is an acceptable rationale for making decisions on performance standards.
GSK, she explained, has a set of core principles and policies for animal care driving much of what she and her colleagues do to create a better environment for the dogs and for the people who care for them. GSK, she said, is studying the characteristics of good exercise for dogs and plans to publish the findings.
Most wildlife research, John Bryan noted, is done for the sake of the wildlife and is not about using animals as a model for human biology. As a result, the methodology used to develop performance standards in the wildlife setting is different than in the laboratory setting. As a wildlife veterinarian, Bryan believes the definition of a performance standard that Patricia Turner presented fits nicely with the performance standards that are used in wildlife veterinary medicine, particularly because that definition includes the word “discretion.”
What ties wildlife research to biomedical research, he added, is the desire to develop the best possible standards reflecting the beliefs and values of a civilized society. One big difference, though, is that the IACUCs reviewing wildlife research have to deal with the fact that virtually every project is unique in its operations. This requires that IACUC members look
_________________________
3 This section is based on the presentation by John Bryan II, and the statements are not endorsed or verified by the National Academies of Sciences, Engineering, and Medicine.
at each submission from scratch, given that from project to project species change, climate changes, terrain changes, and seasons change. Seasons, for example, can affect how a particular species metabolizes drugs, its reproductive activity, and its foraging and other behaviors. In addition, there are occupational health issues associated with the research being conducted in the field, where conditions can change dramatically over the course of a study.
For an IACUC with responsibility for oversight of wildlife research, on which Bryan has served, the key considerations for a wildlife performance standard are that they represent a standard or guideline, describe a desired outcome, and most importantly, provide flexibility in achieving that outcome while respecting the discretionary authority residing with the IACUC, the animal care and use procedure managers, and the principal investigators. Bryan noted that no two wildlife IACUCs will come up with the same interpretation of a proposed performance standard, since each IACUC is composed of members with varied experience.
Some wildlife performance standards, he said, are designed de novo, but many are designed using resources and best practice tools employed in the assessment of animal activities. The point of the design process is to achieve the desired outcome of the highest possible standards of animal care in the context of compliance and regardless of circumstance, Bryan explained. The traditional resources an investigator can draw upon when designing a performance standard are the Animal Welfare Act and the Guide. The Animal Welfare Act defines a field study as one conducted on free-living wild animals in their natural habitat as long as it does not involve an invasive procedure, have the potential to harm an animal, or materially alter the behavior of an animal under study. If a study does meet the definition of a field study, the Animal Welfare Act says it is exempt from the requirement for review by an IACUC. The Guide, meanwhile, states that “it does not purport to be a compendium of all information regarding field biology and methods used in wildlife investigations, but the basic principles of humane care and use apply to animals living under natural conditions. IACUCs engaged in the review of field studies are encouraged to consult with a qualified wildlife biologist.”
While these two sources of information do mention wildlife research, they provide little in the way of guidance for those who are conducting wildlife research. Professional organizations, such as the American Fisheries Society, the American Society of Mammologists, and the Ornithological Council, have issued guidance documents that wildlife researchers and IACUCs use to help reach decisions on a proposed performance standard. Most of what goes into developing new
performance standards comes from knowing the ecology of the system being studied, having experience working in the field, and being flexible by developing contingency plans for when nature does something unexpected.
Ecology, said Bryan, is the cornerstone of wildlife project oversight and it cannot be underestimated as a key variable. A system’s ecology comprises the environment, climate, season, and species; understanding how they relate to one another is critical for developing a sound performance standard. For example, the best time to study wolves is in winter because the snow makes it possible to track them, so proposing studying wolves in the middle of summer may raise concerns for a wildlife IACUC. Similarly, a project proposing to chemically immobilize and tag grizzly bears in winter will raise flags because bears are hibernating then. So, too, would a project that proposes to dart a bear emerging from hibernation in March in the shoulder, given that after months of hibernation, the only place on a bear with excess fat is the rump. In those cases, a wildlife IACUC would pull out a performance standard and suggest changes to the investigator. A big challenge for wildlife IACUCs, said Bryan, is that since the wildlife field is so large there will be many projects it needs to review for which none of the members have any experience relevant to those projects. What the IACUC has to do then is find an outside expert with the experience needed to make a sound decision. One step Bryan took when he was chair of the U.S. Park Service IACUC was to create an archive of projects the IACUC had reviewed as a resource the committee could use when reviewing a proposal.
Being a wildlife veterinarian, said Bryan, is much more akin to being an ecologist than being a veterinarian who works with small or farm animals because the animal cannot be considered in isolation. Working with predators, for example, requires thinking about territoriality and social structure. Wolves are extremely territorial, and darting a wolf in an area that is too close to a rival pack’s territory can be dangerous for both the animal and the biologists. He recounted a principal investigator who came before the IACUC with a project to study mountain goats, but had not considered this research would be conducted at altitude, perhaps as high as 12,000 feet. None of the team members had experience working at altitude.
Flexibility is a key feature for a wildlife IACUC, just as it is for the biologist conducting research. “The flexibility that you build into wildlife IACUC oversight has to be or should be minimally as dynamic as the project itself, while inflexibility in oversight begs catastrophe,” stressed Bryan. It is hard to apply strict engineering-type standards to wildlife
research, he said, because of the complexity of the environment within which any wildlife project operates and the number of unknown unknowns present in the system. Wildlife IACUCs, he said, need to have the room to be flexible so they can exercise ingenuity and be responsive to the investigator’s needs.
As an example of how a wildlife IACUC puts these concepts into action, Bryan described a project the Park Service IACUC reviewed involving studying desert bighorn sheep. The IACUC immediately identified this project as an extremely challenging one in terms of the environment, the terrain, the climate, the species, the capture technique, and the handling procedures. The IACUC held a full committee review of the project to apply as much expertise as possible to the review. It quickly concluded this project was going to be highly dynamic and require the IACUC to be ready to exercise flexibility and ingenuity given the opportunity for almost radical change in the project at a moment’s notice. Nonetheless, the IACUC also accepted this was an important project and it would have to keep in touch with the principal investigator and monitor the study.
In fact, almost from the start the project needed to adapt to conditions in the field. The original capture and handling plan was not working, but because the principal investigator and IACUC had considered this might be a possibility, a backup plan was ready, and IACUC members were prepared to meet and discuss what was going to be a significant change when the principal investigator called in by satellite phone. In real time, the IACUC approved an amendment and filed the paperwork in the project folder and the committee archive, enabling the project to continue without delay. The key, said Bryan, was the IACUC was staffed with wildlife researchers who knew wildlife research can be chaotic, and so when the IACUC created its standard operating procedures for this project it built in flexibility to respond to dynamic change while also achieving compliance and following existing rules. Rigid standards, he said, would have resulted in catastrophe. In closing, Bryan said the definition of performance standards as written in the Guide fits well with wildlife IACUC oversight. This definition allows for a skeletal framework of discretionary authority and then the flexibility to carry out and address the needs of the investigators, the IACUC, and the institution that it serves while staying in compliance with the rules and existing standards. Certainly, he said, there are some details of wildlife project oversight that differ from what might be thought of as traditional guidelines, and those are important and need to be emphasized.
While most published performance standards for the care of agricultural species address production parameters, many of these species are used for biomedical research, said Bart Carter. These uses include medical device testing, drug development, surgical training, imaging technique refinement, oncology treatment testing, and human nutritional research. One noteworthy species, said Carter, is the miniature swine, which behaves similarly to commercial swine and is susceptible to all swine diseases, but of a size similar to that of a dog.
Agricultural species used in biomedical research have a few characteristics Carter said are worth noting. For example, they are herd animals and want to be together in groups. Commercial breeds grow rapidly and require large amounts of feed, even when young, and as a result, these animals will grow during the course of a study, something that needs to be considered when creating standards for these animals. Many agricultural species can become aggressive at sexual maturity, and these animals can be large, which not only has ramifications for housing needs but for the safety of animal facility staff and researchers. “If an animal that weighs 250 to 300 pounds objects to something that you want to do, it can push back hard enough that someone may be injured,” said Carter. While the size of these animals is often what makes them useful research subjects of imaging or physiology studies, size also can create challenges that have to be addressed.
There are three basic types of housing for agricultural animals in a research setting. The traditional biomedical facility is constructed to be easily sanitized, with sealed surfaces and stainless steel. It provides easy access to resources such as surgical suites, imaging centers, and enrichment items such as toys, but is likely to have limited space with a fixed pen size and limited ability for these large animals to get exercise or to exhibit species-specific behaviors.
An agricultural facility is designed for production agriculture with species-specific handling capabilities and a variety of housing conditions. The environmental conditions can vary from being fully air-conditioned, heated, and ventilated to being little more than a barn. Typically, this type of facility can house animals at a reduced cost compared to a biomedical facility, but it frequently has floors and other surfaces that are not easily sanitized. Most often, they are located remotely from the rest of the
_________________________
4 This section is based on the presentation by Bart Carter, and the statements are not endorsed or verified by the National Academies of Sciences, Engineering, and Medicine.
campus and thus animals must be transported for access to imaging and surgical facilities.
The outdoor pen or pasture is the third type of housing, and it offers the advantage of providing plenty of space, exercise opportunities, and the ability for the animals to exhibit herd behavior. The disadvantage of this type of housing is that there is little control of the environment and the animals can be exposed to other wildlife that can transmit parasites and infectious diseases. Observations after surgery, for example, become more difficult and again, it is likely that transportation will be needed to get the animals to imaging and surgical facilities.
Developing an agricultural animal performance standard, said Carter, requires everyone involved in the process – the veterinarians, the IACUC, the husbandry staff, and the research staff – to be familiar with the basic needs of the species of concern. One way for individuals who are less familiar with agricultural species to become informed is to read what is known as the “Ag Guide,” the Guide for the Care and Use of Agricultural Animals in Research and Teaching, 3rd Edition (Federation of Animal Science Societies, 2010), which AAALAC adopted in 2011 as a primary reference for agricultural species. This peer-reviewed and thoroughly referenced document, developed by the Federation of Animal Science Societies, contains information on both normal and abnormal behavior of agricultural species and on animal well-being. Though this book focuses on production agriculture, it has value for biomedical researchers, said Carter, as it can serve as a source of what is considered appropriate management for an agricultural species when developing a performance standard. It can be particularly valuable, he added, when trying to learn about a given species’ normal behaviors.
One challenge in creating a performance standard for agricultural animals is specifying how to maintain herd animals inside a biomedical facility while meeting their enrichment and socialization needs, particularly during the post-procedural care, said Carter. Moving them into a biomedical facility reduces their ability to engage in normal intraspecies behaviors. The reverse situation – moving animals from a biomedical facility to a more traditional farm setting – might seem easier, but these animals may not have been housed in that manner for some time, if ever, so it is important to consider what steps will be necessary to acclimate them to a less structured environment.
An acceptable performance standard for agricultural animals, said Carter, allows the animals to remain dry, be at a comfortable temperature, and remain free of urine and fecal contamination, and it also allows for appropriate data collection by the research staff.
Complicating the development of a new performance standard for agricultural species is the need to accommodate their size, growth rate, and temperament. The health status of individual animals can be important when they are mixed with other animals. Respiratory disease outbreaks, for example, can quickly infect numerous animals, and post-surgical care can be difficult if they are catheterized and need regular monitoring. The scientific goals of the study can also complicate the development of performance standards.
To illustrate what goes into developing a performance standard for agricultural animals used in research, Carter discussed a few examples. The first hypothetical example was of a study involving five 100-pound pigs in a biomedical facility with 25 square-foot pens. The Ag Guide says pigs of that size need 10 square feet apiece, so the question is whether to house them as two sets of two animals and one animal by itself, which meets the space requirement but does not meet the need for herd behavior or socialization. The key here is to determine if placing three pigs in one pen gives them enough room to lay down, stretch out, and remain clean and dry. If so, that might be the time for a performance standard that allows for three animals in a pen to accommodate the socialization needs, rather than the exact space requirement, with an additional feeder to accommodate the third pig.
Carter discussed a different situation when a biomedical facility needs to farrow a sow as the piglets are required for a particular study. Standard agricultural practice would be to put the mother in a farrowing crate for a week before she is due and allow the piglets to be with her until they are weaned at three weeks. That, however, would place her in the farrowing crate for four weeks, so the crate must be of appropriate size to allow her to lay on either side, stand up, and rest without bumping her head on the feeder. During that period, sanitation would be limited to not disturb the piglets. The advantages of this system are many: the sows adjust to the farrowing crate with little training; the odds of the piglets surviving are increased; and it is easy for staff to access the animals. One question that arises, given there will be a farrowing crate in the middle of a biomedical facility, is how to deal with sanitation needs that differ from those of the more traditional biomedical research animals also housed in the facility.
In the farm setting, Carter said often there are agricultural pigs in one area of a large facility and biomedical pigs in another area, raising the question of whether the one group of animals negatively impacts the other if they are being treated differently. This is when the IACUC, veterinarians, researchers, and animal husbandry staff have to work together to answer project-specific questions, such as whether the two
groups can be mixed in the same room or if the presence of the two groups is going to alter sanitation schedules. There will be a need to account for the thermal comfort of the animals, the humidity, and floor space. When creating performance standards relating to husbandry, the conditions must be such that they keep the animals’ stress level low.
The answers to these questions are not always straightforward. For example, the thermal comfort zone for cattle is -15°C to 25°C, which is the range in the Ag Guide and in the Animal Welfare Act. For animals adapted to life outside and brought into a facility in winter, however, heat stress can develop at temperatures as low as 10°C, suggesting that a performance standard may be needed for this situation. Staff, said Carter, need to be familiar with the signs of heat stress and are aware of the normal behaviors and normal requirements for that species.
Achieving adequate ventilation is another issue that often arises when dealing with agricultural species. Tunnel ventilation, which uses wind chill to regulate temperature, is common in agricultural facilities, but it is a much different system than the one found in biomedical facilities for which the regulations mandate a set number of air changes per hour. Tunnel ventilation systems, in contrast, have set points for temperature, humidity, and even ammonia levels, and the system adjusts itself to establish what is known to be a comfortable condition for the animals in that facility. In a large building designed for agricultural species, the ventilation rate can be more accurately determined by these complex tunnel systems than one that simply produces the set number of air changes per hour, given ventilation needs change depending on the number of animals in the facility, their age, what they are eating, the amount of waste they produce, how that waste is handled, and the atmospheric conditions outside of the facility.
For poultry, group housing on a solid surface is the preferred system. There are situations, though, where individual egg production collection may be needed, such as when chickens are used in place of rabbits for polyclonal antibody production. Moving laying chickens into and out of pens or cages can be stressful and cause injury, raising the question of whether it is worth moving the animals regularly to clean and sanitize the cage or leave them in the cage for the duration of their egg-laying period and just clean the pans and floor underneath, said Carter. That is a question for the IACUC to address.
The last example he discussed involved the pasture or confined lot in a farm setting. From an animal’s perspective, this is where they may like to be as there is plenty of space for them to exhibit normal behavior and for timid animals to get away from aggressive ones. For such a setting,
there should be sufficient dry space available and the ability for the animals to regulate their temperature, such as shade in the summer or having a windbreak in the winter to block prevailing winds. The challenge, said Carter, is to evaluate an area such this. Features to examine, he suggested, would include feeders and waterers; general physical plant maintenance; drainage; waste management; how surgical procedures not related to a study, such as castrations and dehornings, are done; and how medical emergencies at the facility are handled. The discussion arises as to whether these should be performed as normal agricultural management or whether they require special conditions. His advice is that if routine surgical procedures are performed as normal agricultural practices, the methods used should be those that are least likely to cause pain or distress to the animal and should include pain prevention. These techniques should be written in a standard operating procedure reviewed by the attending veterinarian and IACUC. Medical emergencies would be dealt with as sterile surgeries.
For any study that requires working with agricultural animals, there is always the option of collecting tissue from a slaughterhouse, said Carter. IACUC protocols and oversight are not required, and neither are performance standards. He noted one scenario in which a researcher needs fresh blood or tissue and no other source is available. “Should they obtain an animal solely for that purpose or can an arrangement be made to collect the sample from a local farmer?” asked Carter, who strongly recommended the latter course of action.
Handling and transporting farm animals is another consideration, one covered in detail in the ILAR publication Guidelines for the Humane Transportation of Research Animals (NRC, 2006), said Carter. He noted, too, when developing a performance standard for agricultural animals outside of a traditional biomedical facility, the animals may require vaccinations and deworming, and they may be exposed to infectious diseases.
THE PUBLIC INTEREST PERSPECTIVE5
The Animal Welfare Institute, said Kenneth Litwak, was founded in 1951 to take the middle ground between researchers and antivivisectionists. The organization is dedicated to reducing animal suffering caused by people by seeking better treatment of animals everywhere, be
_________________________
5 This section is based on the presentation by Kenneth Litwak, and the statements are not endorsed or verified by the National Academies of Sciences, Engineering, and Medicine.
it in the laboratory, on farms, in homes, or in the wild, and it is a strong supporter of the Animal Welfare Act. With respect to research animals, the Institute seeks to improve the housing and handling of animals and to encourage the development and implementation of alternatives to animal experimentation. The Animal Welfare Institute produces many publications, including Comfortable Quarters, a new edition of which is planned for the summer of 2015. This volume was written by veterinarians, technicians, and scientists in the research field and describes the state of the art for housing for the most common laboratory animal species.
The Institute maintains a refinement and enrichment database it updates quarterly, and it gives out refinement and enrichment grants that aim to identify better ways to handle laboratory animals and enrich their environment. It also sponsors the Laboratory Animal Refinement and Enrichment Forum, an online venue in which over 300 veterinarians, technicians, and scientists discuss methods of refinement and answer questions about performance standards. Three volumes of discussions from this forum have been published, with a fourth coming out shortly.
The history of the Animal Welfare Institute’s involvement with performance standards started in 1985 with the Improved Standards for Laboratory Animals Act, which laid out the minimum requirements for canine exercise and promoted the psychological well-being of primates. In the final regulations, USDA allowed facilities to develop a plan to implement those requirements. A lawsuit filed six years later by the Animal Legal Defense Fund and others claimed USDA had yet established a means of determining if institutions were following these plans, if the plans were in compliance with regulations, or if the plans accomplished their stated goals. Another lawsuit filed five years after that by the same plaintiffs charged that there had been an unreasonable delay in promulgating standards to promote the psychological well-being of primates. An initial ruling held that the regulations had not set standards, but that ruling was overturned on appeal in 2000 (Animal Legal Defense Fund, 2000).
The Animal Welfare Institute’s chief concern regarding regulation of performance standards is identifying who is responsible for determining if they are appropriate. The least desirable situation from Litwak’s perspective is when there is a performance standard but no engineering standard, which is the case with the psychological well-being of nonhuman primates. The performance standard holds dealers, exhibitors, and research facilities responsible for developing, documenting, and following an appropriate plan for environment enhancement to promote the
psychological well-being of nonhuman primates (CFR 2012). The plan must be in accordance with currently accepted professional standards, and it must require the physical environment be enriched by providing a means of expressing non-injurious species-typical behavior (ibid). The problem with this performance standard is vague language, said Litwak, regarding who defines what a professional standard is. Also vague is how many and what type of enrichment is needed and who defines what enrichment is acceptable. Even more confusing, he said, is understanding what is normal and abnormal. All of this, said Litwak, gets to the crux of the issue for his organization, which is there are multiple interpretations possible, making enforcement difficult to impossible.
Next on the desirability scale is when there are performance and engineering standards but they are not synchronized. In the case of cage space for non-human primates, the engineering standard requires primary enclosures to meet a minimum space requirement, while the performance standard states there has to be sufficient space to make a normal postural adjustment for freedom of movement. As an example of how these two standards conflict, Litwak used the case of a 20-pound Cynomolgus macaque whose typical body size would be 15 to 22 inches and with a tail 16 to 26 inches long. The engineering standard says floor space should be 4.3 square feet, but a monkey of that size would not be able to make a normal postural adjustment in a cage of that size, begging the question of which standard applies, particularly when the definition of a normal postural adjustment is unstated.
The ideal situation from the Animal Welfare Institute’s perspective would be the 1999 USDA draft policy on environmental enhancement to promote psychological well-being of non-human primates. This exhaustively written and researched draft policy described typical behavioral needs of primates and included the critical elements that would need to be addressed in an environmental enhancement plan as well as strategies to address those needs. The draft also included a list of relevant literature, talked about the specific needs of the most common primate species, and provided many examples of specific enrichment techniques. Unfortunately, said Litwak, this policy was never enacted.
In closing, Litwak said when considering engineering standards or performance standards, there is always the caveat there can be exemptions when required by a particular research protocol or in the judgment of the attending veterinarian, both with the approval of the IACUC.
From his organization’s perspective, though, “performance standards may be acceptable in support of an engineering standard, if properly
synchronized; however, they are not acceptable as alternatives to the engineering standards.” The Animal Welfare Institute, he added, does appreciate those individuals in research institutions who are providing more than minimum engineering standards and recognizing that all animals require an enrichment environment.
AN ACCREDITING ORGANIZATION’S PERSPECTIVE6
The publication of the 7th edition of the Guide, said John Bradfield, was a game-changer for AAALAC International (AAALAC) because it codified the concept of performance-based assessments and standards in the management and operation of research animal facilities and provided flexibility on how to establish and operate Animal Care and Use Programs. This made AAALAC’s job more difficult as there is no single right answer for how to promote animal welfare. Bradfield noted that the Guide is one of three primary reference AAALAC uses in its assessments - the other two are the Ag Guide and Appendix A of ETS 123 (Council of Europe 2006), which also contain performance-based language.
As an example of how the 8th edition of the Guide extended the performance-based approach and in so doing created stumbling blocks for AAALAC, Bradfield cited the following section:
“Solid-bottom caging, bottles, and sipper tubes usually require sanitation at least once a week. Some types of cages and housing systems may require less frequent cleaning or disinfection; such housing may include large cages with very low animal density and frequent bedding changes, cages containing animals in gnotobiotic conditions with frequent bedding changes, individually ventilated cages, and cages used for special situations.”
The 7th edition required facilities to clean and sanitize cages, bottles, and sipper tubes at least once a week, which could be easily assessed. The 8th edition added the proviso about cage types and housing conditions for which cleaning and sanitizing may take place less often, but the phrase “may require less frequent” has no defined time element and that triggered a debate within the AAALAC Council on how to assess this
_________________________
6 This section is based on the presentation by John Bradfield, and the statements are not endorsed or verified by the National Academies of Sciences, Engineering, and Medicine.
performance-based standard. Bradfield said the Council developed a performance standard to assess a facility’s performance standard, measuring any issues with animal health and wellbeing, and whether cages are cleaned at least once a week. If the cleaning interval was longer than one but less than two weeks, which for ventilated rodent cage racks is a practice standard, the IACUC would need to be aware of the situation and consider what the potential impacts might be on the microenvironment and the animals, but the IACUC review need not be rigorous. However, if the cleaning interval was longer than two weeks, the IACUC needs to conduct a more rigorous evaluation validating the cage microenvironment is satisfactory, preferably including an objective, data-based analysis to show why the practice does not negatively impact animal health and well-being.
This example, said Bradfield, illustrates the mental gymnastics going through the minds of the site visit team for one simple topic in the Guide. Looking at the Guide from a higher level, it is clear what is expected but virtually silent on how to do it. Bradfield said this was an intentional decision on the part of the Guide’s authors to provide the targets and let facilities choose how to best meet those targets.
A second important concept arising from the 8th edition of the Guide is that professional judgment plays an important role in how site visits proceed. In AAALAC’s case, said Bradfield, it is not the professional judgment of one person, but of a body of people that come to an agreement. As he explained, a good performance standard to the AAALAC Council is precise, detailed, and has a defined goal, and it should spell out the assessment criteria and methods used to determine that the goal is met.
To illustrate the exercise AAALAC expects facilities to go through when establishing a performance standard, Bradfield discussed a hypothetical facility primarily housing rodents. This facility is about to undergo an AAALAC inspection and management is unsure if its sanitation policy meets the standards of the Guide. The facility team could start by consulting the Guide to determine the appropriate benchmark or goal for the performance standard. In this case the goal is for washing times to reduce or eliminate potential pathogens. From that goal, the facility would develop precise definitions for how to sanitize the cages based on the published literature, staff experience, and input from the manufacturer. This definition could include a preference for using mechanical washers whenever possible as well as a specific protocol for how to wash cages by hand when necessary. Assessment criteria could include husbandry and sanitation logs, visual inspection of the cages,
temperature-sensitive tapes for ensuring that wash water is hot enough, and bioluminescence and microbiological assays with established pass/fail criteria. The methods of evaluation could include standard operating procedures for day-to-day monitoring by facility staff; periodic reviews of microbiological monitoring records to see if pass/fail criteria are met; supervisor checks of cages, logs, and other records; and preventive maintenance of machinery.
Bradfield then briefly addressed the issue of global or program-wide exceptions to the Guide. There are times when institutions need to develop or evolve a practice that technically may not meet some criteria in the Guide, and when applied across a program AAALAC calls it a global exception or a program-wide exception. These arise from some unique feature or unique need of an institution and they require a site-specific, data-driven analysis of the practice demonstrating the institution is meeting an equivalent standard in the Guide in a manner that upholds the well-being of the animals and scientific integrity.
Bradfield concluded his presentation with a real-life story from a site visit to a large institution with many dogs bred in-house primarily for dental research. This institution had a goal of housing the dogs in social groups of compatible animals and adopting them out when the study was finished. The nature of the studies benefitted from having dogs grouped in this particular manner, but the inspection team noticed that pen size was slightly smaller than recommended in the Guide. When asked about this, the researchers said they explored two options. The engineering approach would have had them remove one dog from each pen according to Guide standards, while the performance-based approach took into account the housing method that worked best for the dogs and the science. The facility team defined their goal, which was that group size and age were important scientific factors in the housing strategy. They also recognized the importance of social housing and compatibility. To develop their assessment criteria, the team did a study in which they compared group-housed dogs according to the Guide standards and according to their plan, which was one half of a dog too many. Technicians and staff counted over the course of a year the incidents of stereotypic and aberrant behaviors observed and their impact on the scientific outcomes. At the end of one year, the IACUC, the attending veterinarians, and the investigators analyzed the data and found the performance-based strategy produced a greater number of socially compatible groups, fewer stereotypic behaviors, a higher rate of producing litters, and less variable scientific data while using fewer dogs. As a result of the effort these investigators and facility staff put into
creating this performance standard and meeting the three primary components that AAALAC wants to see (i.e., “professional input, sound judgment and a team approach”), the site visit team determined it was a reasonable performance-based approach to housing.
SUMMARY OF THE AFTERNOON SESSIONS
Paul Locke noted that the presentations represented different approaches to operationalizing performance standards. The first two and the last speaker – Lipman, Vasbinder, and Bradfield – illustrated the evidence-based approach of setting up hypothesis-driven research to show whether a performance standard works. The advantage of this approach is that it mimics how research is done by combining scientific testing, data collection, methodology development, and evaluation in one package.
The next two speakers – Bryan and Carter – were faced with different situations that in many ways are more challenging and where the one-package model could not be applied. These situations, said Locke, require an approach based on collecting evidence and constantly evaluating it with no clear hypothesis to test. While these are unique situations all of these speakers clearly pointed out the goal is to ensure that animal welfare is held to the highest standards and the best science is done with the fewest number of animals.
The last two speakers, Litwak and Bradfield, provided what Locke characterized as interesting views of how the system works or does not work. Locke noted that Bradfield pointed out that performance standards, in many ways, can be opaque, and while a lot of effort is spent explaining performance standards, it may be that the community is not getting everyone to understand how performance standard evaluations are conducted and assessed. The question Locke was left with after hearing these six presentations is whether the community can do a better job of thinking about the concept of performance standards. “The burden really is and should be on the IACUCs and the institutions to make sure the performance standards are measurable and can be evaluated and can be explained,” said Locke.
David Anderson posed the first question, from an online participant, to Vasbinder, about the requirement that dogs get 15 minutes of exercise daily and whether that requirement changed if the dogs are getting social
exercise in their cage. Vasbinder replied the requirement is 15 minutes of social exercise per dog, but that some dogs cannot tolerate a social exercise environment. For those animals, there is a separate space where they exercise individually with the help of staff. Her preference is for a long exercise space that enables the dogs to truly run.
The second question from the online audience was addressed to the entire panel and asked if modern veterinary skills and the use of appropriate therapeutics need to be incorporated in discussions about performance standards. For example, the questioner asked, should there be a requirement for veterinarians to maintain a current license to practice and prescribe drugs in their working jurisdiction? Bryan, a licensed veterinarian, answered there have been some interesting issues regarding Drug Enforcement Administration licensure over the use of certain drugs in the field, and these have been resolved recently. He also said that when he was a federal veterinary officer the government recognized his Georgia and Colorado licensures as sufficient to practice on federal lands, and he and all of his colleagues who were veterinarians maintained their licensures and met the requirements for continuing education credits.
Robert Dysko from the University of Michigan Medical School asked how often studies validating a performance standard need to be repeated. Vasbinder said programs need reevaluation because there can be differences between practice and perception and that GSK addresses this issue by having outside consultants review its programs. Her program formed an advisory panel to help plan and carry out the evaluations, which it conducts every three to four years. Bradfield said AAALAC’s approach is to pose the question, “What is your strategy for periodic reevaluation of any performance standards that are in place?” with no answer in mind except that AAALAC does anticipate some periodicity in evaluations. At his former institution, for example, the IACUC reviewed standards monthly for one particular study involving infectious diseases, but a performance standard that was more innocuous and for which the conditions did not change would require far less frequent reevaluations.
Malak Kotb, from the University of North Dakota, asked if anyone had looked at the effect of sterilizing cages on the gut microbiota of the animals and, if so, whether this had any experimental effects. Bradfield said if there are studies in which the cage environment is critical for establishing the microbiota, there might need to be an institutional or study-specific, precise definition of clean.
Judy MacArthur Clark commented that Bryan and Carter both spoke about how they addressed conflicts between engineering standards and
performance standards, but she wondered if these conflicts represented instances where the engineering standard was wrong. Bradfield said this question captured the angst with which the AAALAC Council deliberates on these issues. He added he would never suggest the solution one institution established as working best for its animals would apply to another institution or scenario. The engineering standard, he added, is a starting point and a critical factor. He noted that AAALAC is clear on the point that limited resources and cage size alone are not acceptable justifications for housing animals too densely while recognizing there can be unique circumstances in which a broader consideration, other than the pure engineering approach, is more appropriate for the animals and the science.
Carter said it is important to remember that the IACUC and veterinary and animal care staff have the opportunity to evaluate a performance standard and weigh in on whether they think it is working. Vasbinder noted there is an opportunity to be creative about the needs of each species by modifying their housing to create more space within a cage to allow the animals to interact with their environment and give them control over whether they are interacting with each other or not. In the case of dogs, for example, a flat roofed dog house the dogs can jump onto or hide within could be one form of added enrichment not requiring more space per cage.
AAALAC, said Bradfield, takes into consideration the amount of due diligence that an institution puts into developing, testing, and validating its performance standards. In the case of the institution in his example, it had made an intentional decision based on its data about the best way to house its dogs, and the AAALAC Council determined this institution’s IACUC had done what an IACUC should do and it had provided a clear explanation of its decision. The wildlife area, said Bryan, is one in which each and every project is judged on its own, and it is the norm with wildlife IACUCs to require investigators to come up with sufficient justification for any variances from existing engineering standards. He also said he did not think engineering standards have a big future in the wildlife area.
Lipman commented on the assumption that the engineering standard is correct, saying he questions whether this is always the case. Steven Niemi pointed out most if not all engineering standards used to be standard practice before they became engineering standards, raising the question of whether there were arbitrary elements in them before they became codified as engineering standards. Whether they are correct or
not, his concern is that the community is creating a higher bar for doing something that differs from these engineering standards.
Niemi then asked Bradfield if AAALAC expects an institution to generate internal data to validate a globally applied change from the Guide, and Bradfield said this is what AAALAC expects with one caveat: if an institution can demonstrate that the change it wants to make is identical to that of another institution AAALAC would agree that outside data may indeed apply. However, in AAALAC’s experience this seldom occurs.
Given the long time between editions of the Guide, said Lipman, it is the intent of the Guide committee that as new information becomes available, it could serve as the basis for a new universal performance standard with repeated assessments. Bradfield’s example involving dog housing made Vasbinder wonder what would have happened if the IACUC had approved the request to keep litters together without further research based on a 3Rs rationale. Bradfield acknowledged that was an interesting scenario but guessed the AAALAC Council would have found that to be a less compelling case, albeit a valid rationale. What AAALAC Council wants, said Bradfield, is for an IACUC’s decision to be fully informed, careful, and thoughtful.
Cathy Liss asked Vasbinder if there was a backup plan in case neither group of dogs was doing well in her study. Vasbinder said if the study had not produced an acceptable outcome, the next step would have been to characterize what good exercise looks like from a behavioral standpoint, a study her team is conducting anyway. Such a study would have given her team a desired outcome that would serve as the vehicle for developing an approach to reach that outcome.
Liss commended the emphasis the panelists placed on social housing, and she asked them to comment on how each of their institution weighs the importance of social housing versus other aspects. Vasbinder said the behaviorists involved with GSK’s program put social housing as the highest priority for dogs and primates. Bradfield noted the 8th edition of the Guide emphasizes this topic in particular, but it is a complex issue in part because of a lack of knowledge. With rabbits, for example, there is no data on what is an appropriate social grouping, and it took years to understand how to form compatible groups of adult male Rhesus monkeys. Carter said social housing is becoming a higher priority item among investigators and there are fewer situations where animals are housed individually.
This page intentionally left blank.
