Veterinarians swear to protect animal welfare and relieve animal suffering. How can they participate in animal research that causes so much animal suffering? On the other hand, where better to meet the challenge of their oath than in animal laboratories? Many ethical questions of animal use echo across disciplines in veterinary medicine. Other questions are specific to the use of animals in laboratories:

• Does animal research provide enough useful knowledge that cannot reasonably be obtained via alternative methods?
  
• What special justifications might be necessary when experiments can cause significant pain and distress to the animals?
  
• How should ethics oversight committees factor in uncertainty about how much an experiment might harm the animals or how much benefit is likely to result?
  
• If animals are sufficiently like humans to serve as useful models of human biology, are they too similar to humans emotionally for their use in experiments to be justifiable?

An assortment of formal laws and informal guidelines reflect the applied ethics of laboratory animal work. Key features of the common ethic are:

• It is sentientist: sentience grants a human or nonhuman the right to moral consideration.
  
• It is animal welfarist: animal pain, suffering, and well-being, rather than autonomy, telos, or a right to life are animals’ interests of concern.
  
• It is consequentialist: significant potential benefits (usually to humans) may justify some harms to animals.
  
• It is speciesist: humans interests receive greater consideration than other sentient species’ interests.
  
• It is jurisdictional: veterinarians have a defined role in overseeing laboratory animal health and welfare.

In this chapter, I review this applied ethic in the broader context of veterinary ethics and animal ethics.

Laboratory Animal Ethics in the Context of Veterinary and Animal-use Ethics

Animal-use Ethics in Laboratories

Some aspects of laboratory animal care will seem familiar to veterinarians in other disciplines, but it helps to understand how scientists use animals to understand the veterinary and ethical issues in this field. That requires an understanding of the varied ways in which an animal may be used as a model.

Most animal research uses nonhuman animals to learn something about human biology. No one animal species can replicate all aspects of human biology. Some examples illustrate the breadth of laboratory animal use:

• For some people, the first image of animals in laboratories might be rabbits in restrainers for testing cosmetics and cleaners for eye injury (the Draize test) or dosing mice with compounds to find the dose that will kill 50% of a cohort (the median lethal dose [LD50] test).
  
• Most animal studies of SARS-Cov-2 virus (the causative agent of the Covid-19 pandemic) require an animal with human-like angiotensin-converting enzyme 2 (ACE-2) receptors on cells: such species choices include ferrets, monkeys, and hamsters (Munoz-Fontela et al. 2020). Mice engineered with the human ACE-2 receptor gene are now common in Covid labs.
  
• Healthy animals are used to study basic biology. To study individual brain cells’ reactions to light or other inputs, a scientist may surgically place recording electrodes in a rat or a monkey’s cerebral cortex.
  
• Scientists induce and study a wide variety of human-like injuries, illnesses, and treatments in animals and test treatments. Animals can be the beneficiaries too. This may be indirect where things we learn about human biology are applicable to other animals, in human settings or in the wild. Or more direct, as when scientists study mice, cats, dogs, or cattle with the specific goal of improving companion animal or food animal health (Bailey 2018).
  
• Laboratory animals may be used in hypothesis-driven experiments (i.e. research), but also as teaching models for veterinary and human medicine and as subjects in toxicology and safety testing.

There is thus a menagerie of species in laboratories, chosen through a constellation of factors – size, cost, genetics, safety, susceptibility to various microorganisms and diseases, and established knowledge in a particular field. Most animals in modern laboratories have been purpose-bred for research. United States Department of Agriculture (USDA) statistics on animal use in research in 2017–2018 totaled 780,070, including:

• 182,580 miscellaneous mammals (e.g. ferrets, voles, gerbils)
  
• 171,406 guinea pigs
  
• 133,634 rabbits
  
• 80,539 hamsters
  
• 70,797 nonhuman primates
  
• 63,094 pigs and sheep
  
• 59,401 dogs
  
• 18,619 cats. (USDA 2019)

These statistics do not include mice, rats, fish, or other species, with estimates of mouse and rat use in the United States ranging from as low as 15 to as high as 110 million per year (Carbone 2021).

The USDA reports animals in “pain categories.” For FY 2018, about 60% of animals were in Category C (no significant pain or distress; includes euthanasia of healthy animals); about 32% were in Category D (procedures that could cause significant pain or distress, but receive anesthetics, analgesics, or other drugs for alleviation); with about 7% in Category E (significant pain or distress, but drugs withheld if they could interfere with the experiment) (USDA 2019).

Public opinion polls and current laws and regulations reflect a societal ethic that scientists should have limited license to use animals in laboratories (Ormandy and Schuppli 2014; Pew Research Center 2018). Human interests preempt animal interests in this setting but with some limits. Thus, concern for animals’ welfare can require scientists to invest time, energy, and money. Highly invasive or painful studies in species people care most about (such as dogs and primates) done for trivial purposes receive the lowest levels of public support, and regulations reflect these priorities. The application of this limited license to use animals mostly takes the form of a cost–benefit analysis in which most of the costs are the animals’ diminished welfare and most of the benefits are knowledge that may help current and future humans but only rarely do the experimental animals in the laboratories benefit.

For laboratory animal use to be justifiable two conditions are necessary but neither alone is sufficient:

1. Human interests must sometimes be able to take precedence over nonhuman animals’ interests.
   
2. Animal research must produce useful, valid knowledge (Carbone 2012b; Beauchamp et al. 2019).

In other words, laboratory animal research must be rejected if human-privileging speciesism is immoral or the data from such experiments are not reliable.

The consensus among scientists is that animal research produces valuable knowledge (Nuffield Council on Bioethics 2005; Ainsworth 2006). Thousands of scientists use millions of animals with the expectation that much animal data will “translate” to human biology and medicine (Carbone 2021). Others are skeptical that animal studies are useful, or at least, are skeptical that they are as useful as scientists expect them to be (Greek and Greek 2004; Hansen and Greek 2013). Critics explain their concerns in evolutionary terms (LaFollette and Shanks 1996). While all animals, humans included, share many biological similarities, there are also many differences. Therefore, it is not always predictable which animals will be most like humans at the level necessary for extrapolation from nonhuman to human.

There are many stakeholders, costs, and benefits associated with animal research (Table 11.1).

Regulation of Animal Use in Research, Testing, and Education

Legal regulation of animal laboratories varies globally. Most laws focus on some sort of cost–benefit analysis and standards for minimizing animal suffering. In the United States, laboratories generally operate outside of local anti-cruelty laws, and most are regulated by some combination of laboratory-specific laws at the federal and state level.

At the federal level, two main laws cover laboratory animal welfare (Carbone 2004). The Health Research Extension Act of 1985 mandates oversight by the National Institutes of Health Office of Laboratory Animal Welfare (NIH OLAW) for institutions that receive federal research grants from the NIH or other government agencies. It covers all vertebrate animals and relies on self-regulation by institutions via annual reports and self-reports. Rather than a complex set of regulations, it relies on a guidance document from the US National Academies of Science, the Guide for the Care and Use of Laboratory Animals (Institute for Laboratory Animal Research 2011). The evolution of the Guide from its first edition in 1963 to the current 2011 edition reflects evolving thoughts and priorities in laboratory animal care (Animal Care Panel 1963). OLAW’s mandate to assure animal welfare in federally funded research means that many animals, for example used by private pharmaceutical companies that do not receive federal research dollars, are outside its jurisdiction.

The Animal Welfare Act (AWA), first passed in 1966, has comparable standards to the Guide and OLAW, with a set of Animal Welfare Regulations that covered institutions must follow (USDA Animal Care 2019). Passed in reaction to stories of pet theft for laboratories, its early years focused on animal acquisition, husbandry, and veterinary care up to the point at which the animals start on an experiment. Since the 1980s, the AWA now has standards for how animals are used in experiments. Enforcement is via self-reports, as well as unannounced inspections, both routine and “for cause,” by veterinarians with the USDA. The AWA covers most mammal species used in laboratories regardless of funding, but alone among Western nations, excludes most rats and mice, who combined likely comprise 99% of mammals in US labs (Carbone 2021). It also excludes frogs, birds, and – possibly most numerous – fish. The USDA publishes inspection reports on its website, along with annual self-reports of animal use, in three “pain categories” (USDA 2019, 2020)

Complementing the AWA and NIH rules, US and international laboratories can apply for accreditation via an independent organization, the American Association for Accreditation of Laboratory Animal Care (AAALAC) International (AAALAC International 2021a). This voluntary program includes triennial site visits by practicing laboratory animal veterinarians and research specialists as well as self-reporting. In the United States, it relies mainly on the Guide as its standard of accreditation, and its accreditation reports are confidential and not available to the public. Its coverage includes all vertebrates, as well as “higher-level” invertebrates such as cephalopods (AAALAC International 2021b).

All three oversight programs converge on the central role of an in-house animal oversight committee for institutions under their purview. Institutional Animal Care and Use Committee (IACUC) is the common name for these ethics committees in the United States: such committees are central to laboratory animal welfare oversight around the world. In the United States, the committee must include a veterinarian and a nonaffiliated member, usually a nonscientist. Under US rules, at least one veterinarian must have a seat on the committee, with no special voting priority, i.e. one member, one vote.

The main functions of these committees include self-policing and self-inspection of animal care facilities and animal use, including investigation of in-house allegations and concerns about animal treatment. Most of their time and effort goes to prospectively reviewing scientists’ proposals (usually called protocols) to conduct animal experiments. Committees review the reasons for doing the experiments, the qualifications of the people doing the work, the choice of species, the number of animals, and the potential for pain and distress and for mitigating or preventing pain and distress.

The Three Rs Alternatives and Six Principles of Animal Research Ethics

A long-standing set of organizing principles for scientists, veterinarians, and ethics committees is the consideration of Russell and Burch’s “Three Rs” of alternatives in animal research (Russell et al. 1959). Alternatives is an unfortunate word choice insofar as it suggests substituting animals with something different, but in this context, its definition is broader: it refers to alternatives in animal research, not just alternatives to animal research.

Replacement alternatives are conceptually the most straightforward: find ways to generate research data without using animals at all. Candidates for replacement include studying cells in tissue culture (in vitro techniques), developing computer simulations, making better use of human epidemiological data and human volunteers, or using inanimate models in teaching (see Chapter 18 for a thorough discussion of nonanimal methods). Scientists also seek to replace so-called “higher” animals when possible, by switching from dogs to mice, or from mice to fruit flies, though the moral consistency in replacing sentient dogs with sentient mice, sometimes referred to as “relative replacement,” is questionable (Tannenbaum and Bennett 2015).

Reduction comprises efforts to lower the numbers of animals required. This often means rethinking statistical tests to use just the number necessary for statistically valid results.

Refinement constitutes all the myriad ways to reduce the potential for pain and distress, and to increase positive life experiences and happiness (Carbone 2004). Scientists may seek humane endpoints, such as euthanizing animals before they develop severe disease, or they may expand their use of anesthetics and painkillers. Other refinements include replacing surgery with noninvasive manipulations or improving the housing for animals in their experiments.

Russell and Burch’s framework categorizes pain and distress as direct versus contingent. For example, scientists induce direct pain in an experiment in order to study pain neurobiology; scientists in other fields perform surgeries for experiments, but pain is an unwanted, or contingent, side-effect of (for example) placing electrodes in the animal’s brain. Pain, distress, and suffering can result from experimental procedures, but also from the daily housing of animals, with boredom, isolation, fighting, noise, bright lights, and cramped quarters all possible housing-related stressors.

Cost–benefit analyses in animal research weigh benefits (mostly to humans) against harms/costs (mostly to the animals). Beauchamp and DeGrazia have built on the Three Rs framework in their Principles of Animal Research Ethics (2019). Three of their six principles address the benefits of animal research and the other three build on the animal-harm reduction strategies of the Three Rs (Table 11.2). Consistent with long-standing practice, their animal welfare principles call for meeting animals’ basic needs throughout their lives, not just during an actual experiment. They continue the principle that any harms to animals must be limited to those that are absolutely necessary to obtain important data, implicitly allowing that some experiments will cause pain or distress, or will confine animals or deprive them of various pleasures and freedoms (United States Interagency Research Animal Committee 1985). Their most radical-looking principle suggests setting upper limits on harm to animals, but they do allow for the current practice that proposed upper limits may be exceeded when research needs strongly justify that (Carbone 2019b).

The principles of social benefit flesh out how scientists, veterinarians, research funders, and ethics committees might better evaluate the necessity of using animals. They argue that scientists should pursue alternatives but allow that some alternatives may be so impractical or expensive as to be unreasonable or unavailable, in which case some animal use might be permissible. They contend that “net benefit” is not just hoping that studying cancer in mice may produce useful knowledge, but also the possibility that animal studies may draw resources from other potentially valuable research projects or can be misleading. “Net benefit” should account for translatability, a growing concern among scientists in various fields, as well as their critics (Garner 2014; Fernandez-Moure 2016; Mogil 2019). If animal research is to be of value, its animal data must “translate” to human biology, successfully predicting benefits and harms. For example, by the time a drug has cleared all of the years of preclinical animal studies and is ready for human clinical trials, highly translatable animal data will yield a strong chance that drug is safe and effective in people. Different readers will interpret a commonly cited statistic – that some 90% of drugs at this stage fail in human trials – differently (Kola and Landis 2004; Hay et al. 2014). Is 10% success an encouraging reflection on the value of animal studies? Is 90% failure a condemnation that animal studies are misleading? Is it a number to prompt examination of the failures in order to better animal models?

The principle of sufficient value to justify harm holds that some projects or some scientific questions are more important than others. In a medical setting, for instance, it may be of more value to conduct research on fatal brain cancers of children than on largely cosmetic conditions in adults.

These principles would work best if people with sufficient expertise and a diversity of perspectives and values simultaneously evaluated a scientist’s plan in terms of the harms to animals, the potential value of the project, and the scientific quality and merit (Figure 11.1). At present in the United States, these three evaluations occur separately, in different bodies’ hands and in no set sequence. Thus, an animal ethics committee or IACUC may evaluate a project’s animal welfare issues before or after a funding agency evaluates the scientific merit, and neither will know the outcomes of the other’s review (Carbone 2019b).

The Ethics of Uncertainty in Laboratory Animal Use

Ethical harm–benefit review of proposed animal experiments starts with establishing several important facts. It requires technical knowledge to assess the possibility and severity of potential animal harms, to prevent or treat pain and suffering, and to evaluate if pain mitigations, whether drugs or nursing care, will adversely affect the experiment (Peterson et al. 2017; Bleich et al. 2020). It requires a scientific evaluation of the likelihood the work will produce reliable data that can be translated to humans and other animals. With these facts in hand, a scientist or their ethics committee might then determine if the goals of the project are of sufficient value to justify any harms to the animals.

Ethics committees frequently operate in uncertainty. Animal ethics committees have a breadth of members but evaluating the merit of the proposed methods requires a panel with expertise in the particular model, to help the scientist avoid harming animals in an experiment that does not even provide clear data. Without communication among these various review bodies, the harm–benefit analysis is weak. The potential impact of the research is similarly hard to state with certainty, especially in basic science without apparent immediate application, leading some to question whether evaluation of likely benefits is possible (Niemi 2020).

If animals are at risk of suffering from experiments, then only projects with the best chances of success in pursuit of important goals should receive approval. In the face of uncertain harms and benefits, the committee faces the ethical decision of whether to err more toward precaution about harming animals, or the risk of delaying potentially valuable science.

Veterinarians in Animal Laboratories

Veterinarians work wherever there are laboratory animals. Their numbers are small: of about 100,000 American Veterinary Medical Association member veterinarians who list their primary field of activity, only about 1100 (well under 2%) list laboratory animal care as their primary responsibility (American Veterinary Medical Association [AVMA] 2020). The size and mission of the research program often determines the numbers and specialties of veterinarians. Veterinarians may specialize in one of several roles at large institutions or may fill many roles simultaneously at large or small laboratories. Table 11.3 lists some of these roles that veterinarians play directly, or in a supervisory capacity.

The relationship between veterinarians, scientists, and animals is complex. Veterinarians have partial authority over how scientists use animals (Case Study 11.1). In laboratories covered by US laws, veterinarians are guaranteed a voting membership on the animal ethics committee, but most decisions to grant or revoke research privileges are at the committee level. In cases of individual animals’ health and welfare, veterinarians have more autonomy, including the authority to prescribe treatment or euthanasia even if the scientist using the animals disagrees (Institute for Laboratory Animal Research 2011). Despite that authority for on-the-spot decisions for individual animals, in my experience, veterinarians usually have less authority over animal health and welfare problems at the group or project level, which can lead to professional and moral distress.