Editorial: In Musculoskeletal Research, Too Many Animals are Being Harmed for Too Small a Return

A few months ago, Clinical Orthopaedics and Related Research® ran an article showing that more than 40% of papers based on animal models in one large country’s national meetings over a 9-year time period were never published, and almost half of those that were published never got cited or were cited only once [4]. In the experiments those authors analyzed, orthopaedic researchers took the lives of nearly 10,000 animals to fuel this inefficient machine. As I said in an Editor’s Spotlight essay [2] that covered the article, too much animal research is done in surgery, and I think much of it is ill-considered. If one is going to take a life—any life—as part of a scientific experiment, it seems to me that the research question should not merely be answerable, but it must also be important enough to justify the taking of life or causing pain to animals. The recent CORR® paper [4] shows that much musculoskeletal research fails to clear that bar. Other work suggests that the problem extends beyond orthopaedic animal research, and seems generally to be true across the specialty of surgery. It appears that surgical animal research is much less impactful than any caring person might like it to be [9], and that animal studies consistently overestimate treatment-effect sizes when investigators seek to replicate the interventions in humans [6]. As the Editor-in-Chief of a large surgical journal, I am, of course, in the business of publishing animal research, and have done so many, many times. I recognize there are some areas of research that can’t be studied without using animals, including but not limited to certain topics in immunology, as well as mechanistic studies about healing processes; in those areas and others, animal experiments are the only way to begin to get the answers that our patients need. I also know that some animals are raised specifically and deliberately for experimentation, which may soften the ethical blow in the minds of some. Further, the US Food & Drug Administration (FDA) generally requires testing in animals before it will approve many kinds of trials in humans, and the FDA often will ask for two or more animal models to be used in such preclinical research [11]. I understand from individuals who do this type of work that it is almost impossible to advance a drug trial without first obtaining results in animal models; in fact, the FDA recently revised its rules on the subject of when animal experiments can be used instead of efficacy trials in humans [10]. For those reasons and others, animal research is not going away anytime soon—nor should it. But those explanations do not justify the amount of animal research being done in surgery and musculoskeletal science, so much of which is either inconsequential [4, 9] or misleading [6]. As I mentioned, if we are going to take animals’ lives, we should be fairly confident that the questions we’re asking are important, that the answers to those questions will inform future research, and, ideally, that the discoveries made on the backs of animals will lead to meaningful improvements in human health. Considerable evidence suggests that surgical and musculoskeletal researchers often get this wrong. Much, perhaps most, surgical and musculoskeletal animal research has not accomplished what even its proponents might have hoped it would in terms of influence on future research [1, 4, 9] or in terms of estimates of how well interventions tested in animals might perform in humans [6]. Most cardiovascular research in animals lacks what I’d consider foundational methodological elements, such as randomization, blinding, and sample size calculations [8]; there is little reason to think it’s much different in our specialty. Fewer than one in 10 animal studies in another recent analysis (which included many studies on orthopaedic questions) would have been considered ethically permissible when analyzed under a rubric that considered animal suffering, the likelihood of benefit to humans, and the importance of the research as determined by subsequent impact through citations [7]. That paragraph had a lot of negatives in it; perhaps it’s better to focus on what we can do, rather than what we shouldn’t. I will start and end my suggestions with what scientists can do, since they have the largest role to play. I believe that: Scientists should focus on the fundamentals of study design before taking animals’ lives. Specifically, scientists need to ask testable questions on important topics and identify and use the most-realistic models so as to maximize the chance that their discoveries will deliver meaningful improvements to human health. Researchers also should determine the smallest number of animals needed to answer those questions, and start the process with a careful review of the ARRIVE (Animal Research: Reporting of In Vivo Experiments) guidelines [5] so that no important methodological elements are missed. Thoughtfully designed, well-executed studies should fare better in peer review, and are more likely to be used by future clinician-scientists in the translational work that must follow. Readers of this journal who sit on institutional review boards, university ethics committees, and, especially, institutional animal care and use committees should apply much more stringent standards when assessing the potential clinical relevance of animal studies, given how often studies on animals get performed but never get published (and how many of those that get published never get cited [1, 4, 9] or result in misleading estimates of treatment effects in humans [6]). Those institutional boards (and the orthopaedic surgeons who sit on them) should insist first that authors justify the need for an animal model, and in that justification, that authors first explain why other approaches (such as in vitro work or computer simulations) are insufficient. Going forward, we at CORR® will ask authors to provide such justifications in the animal studies we publish. Institutional boards must require stricter adherence to the ARRIVE guidelines [5], which so often seem like an afterthought in animal studies I read in orthopaedic journals rather than a guideline that ensures a measure of methodological rigor and a best-practices level of humane animal treatment. This may be a smaller issue in animal studies funded by large grants, since federal agencies already apply numerous strict standards [3]. But in my experience, this problem is real and pervasive in smaller self-funded or privately funded projects, which investigators use as steps on the way to the larger grants. Because of this, increased institutional oversight is critical. But by far my most-important suggestion is that scientists who believe their research questions justify the causing of pain or the taking of animals’ lives need to self-assess from a posture of radical modesty, given how many scientists have had the same well-intentioned but badly mistaken thought. A close colleague whom I deeply admire once asked me whether adopting a stricter standard on animal research would result in more animal studies that have been performed not being published. He felt that sacrificing animal life and having nothing to show for it itself is an ethical lapse. No doubt he is correct. But I think the far larger problem would be for journals to continue to publish these papers using the same standards we’ve used to this point, which have had so little effect and which have caused so much harm.