Microbiologist Elies Bik is well-known for applying a close eye to studies, and has spent years anonymously submitting reports on plagiarism and image duplication to journal editors. Last year, she published an analysis of work, with troubling results – that 1 in 25 biomedical papers contains inappropriate duplications in images. She’s never stopped reading papers closely, and recently flagged one on Twitter for concerns about its methodology. Although Bik is not against animal research, she said the treatments rats faced in the paper made her queasy – and didn’t seem justified. She explains her reasoning below.
For the past three years, I have been reporting about new papers in the microbiome field through my blog Microbiome Digest. The stream of microbiome publications appears to get bigger every week — as research has tied intestinal microbiota to a range of health conditions such as atopic dermatitis, periodontitis, and even neurological conditions — so since last month the blog is now being run with the help of an amazing team of young microbiology researchers and science communicators.
Earlier in February, a paper by Yu et al. was published in the Journal of Pharmaceutical and Biomedical Analysis with the interesting title “Variations in gut microbiota and fecal metabolic phenotype associated with depression by 16S rRNA gene sequencing and LC/MS-based metabolomics.” The title suggested that stool samples from patients diagnosed with depression were different from that of controls, which sounded very interesting. However, when I read the abstract, it became clear that the study was performed in animals, by using the “chronic variable stress (CVS)-induced depression rat model”.
At that moment, I recognized the study; I had been asked to peer review the manuscript for a different journal in September 2016. The methods section of that manuscript described the treatments that were used to induce stress in these rats in great detail — and they were horrible.
The rats were treated with several different “stimuli” such as multiple electrical shocks within a minute, forced swimming in ice cold or hot water, withholding food or water for 48 hours, immobilization for five hours, and exposure to loud sounds or stroboscopic lights. Each animal received at least one or two of these treatments every day for 28 days. That’s nearly a month of torture.
I found these methods so appalling and inhumane that I refused to peer review the manuscript. I contacted the Editor, who agreed that this manuscript should be rejected. But apparently, the authors got it accepted for publication in JPBA.
In its published form, the methods section describing the rat treatments has been taken out completely, and instead the authors refer to an older, 2011, paper by the same group. The awfulness of the experimental setup in the current paper is therefore not very obvious, and might have not been noticed during peer review.
This paper has several very concerning issues.
Firstly and most obviously, the treatment of these rats was horrendous and out of proportion. Just reading the methods section made me feel nauseous, and I felt sorry for what these poor animals must have gone through. The paper states that the experimental procedures were approved by the Ethics Committee of the Beijing Institute of Medicinal Plant Development and the CAMS and PUMC (Chinese Academy of Medical Sciences Peking Union Medical College), but I cannot imagine that any institutional animal research board elsewhere would have approved this study.
I’d like to make clear here that I am not against animal experiments. In the microbiome field, animal studies have proven to be very valuable. For example, germ-free mice, which are born and raised in sterility, allowed researchers to study the effect of microbial colonization or lack thereof on the development of organs, tissues and the immune system, the in vivo interactions of different individual microbiome members, or the effects of inoculations with the stool from obese or lean humans. Such carefully designed and executed animal studies have greatly advanced the field. However, the experiments performed on the rats in the Yu et al. 2017 paper would qualify as torture, if done on humans. This should not have been published.
The second, huge problem with this particular rat model is that such treatments do not induce depression. Instead, they induce extreme stress. I can only imagine the fear that these animals must have felt when they heard the door of the animal facility open. The extreme physical and neurological demands and anxiety induced in these animals are very different than the persistent sadness and loss of interest that are associated with human depression. So the findings of these experiments will have very little, if any, value for patients diagnosed with depression.
The third flaw with this paper is that it does not make much sense to test the effect of depression on the fecal composition in an animal model. Instead, the authors could have studied stool samples from human depression patients and compared these to stool from healthy controls. That would have been “the real deal,” and a much more valuable contribution in the research of this human disease.
Taking these concerns together, this study used disproportionally inhumane animal suffering to study something that has no value for human health, and that could have easily been performed in humans. I would like to call upon the Editorial Board of the Journal of Pharmaceutical and Biomedical Analysis, and Elsevier, the publisher, to consider to retract this paper. Studies that include torturing of animals without any scientific reason should not be published.
I also very much hope that both editors and peer reviewers will be on the lookout for manuscripts that include this or other similarly awful type of animal experiments. Thresholds of the amount of acceptable animal suffering clearly differ from country to country. During my searches for image manipulations, I have regularly seen experiments in which live mice are dipped in boiling water, sewn together, or allowed to walk around with tumors as big as their own body. An approval by an institutional committee does not necessarily mean that peer reviewers and editors should find such animal experiments acceptable too. We can and have to push back on useless studies where the amount of animal suffering goes much beyond what is acceptable in the name of science.
If we can achieve that, then hopefully these poor rats have not died for nothing.
Note: Retraction Watch has contacted the corresponding author of the paper, as well as the publisher of the journal, Elsevier, who told us the handling editor is looking into the matter.
Like Retraction Watch? Consider making a tax-deductible contribution to support our growth. You can also follow us on Twitter, like us on Facebook, add us to your RSS reader, sign up on our homepage for an email every time there’s a new post, or subscribe to our daily digest. Click here to review our Comments Policy. For a sneak peek at what we’re working on, click here.
“Instead, the authors could have studied stool samples from human depression patients and compared these to stool from healthy controls. That would have been “the real deal,” ”
Such a study, unfortunately, would not have been “the real deal” as it could not establish causality. It would have been accepted in a high impact journal, fit very well with the current zeitgeist and received a good reception from the microbiomics community. But it would have most likely been interpreted as bad microbiome causing depression. This could have led to a protracted, expensive and ultimately futile effort to fix the microbiome to cure depression.
That’s very much true, however, the study showing mere correlation should be done before the animal study. The correlation study (which can be done without animals) should be a necessary prerequisite for the animal study.
(and then actual model of depression should be used, of course)
But this assumes that there is a purely biological cause for “depression” (whatever that is; it is not established yet what is even meant by that term which is constantly changing over time, just like all the supposed biologically-based ‘mental disorders’), and so what you are proposing is simply applying theory that has never been scientifically substantiated. That is not a solution. It might actually be that there is some other cause for the phenomenon, one that cannot be ascertained by simple cause/effect models, but that needs methodology that can adequately handle very complex phenomena. I would say in science there are three crises: 1) lack of good, innovative methodology, 2) lack of good, innovative theory development (which requires the ability to critically analyze rather than unquestioningly adopt existing theory under peer pressure), and 3) the ongoing crisis in torturing animals “in the name of science” for essentially dubious purposes in the end.
Unfortunately, these authors only learned from the best. Hans Selye, who was often called “the father of stress”, published 39 papers on the subject of “calciphylaxis”. His lab “perfected” various animal models that were subjected to combinations of elaborate surgical interventions with chemical poisoning to show what “stress” can do. This work even made it to the pages of Science magazine in 1961, where a combination of chemicals was demonstrated to induce rat to completely shed its skin, which the authors called “cutaneous molt”: http://science.sciencemag.org/content/134/3493/1876.long. The horrific picture of a rat coming out of its own calcified skin was not perceived as cruelty; on the contrary, it was a major scientific breakthrough. Not clear, though, if all that animal suffering was for any benefit to very rare human patients with calciphylaxis.
I have served for years on our institutional animal welfare committee and occasionally we deal with protocols that use animal models of depression, which make use of procedures such as forced swimming (in warm water!) or similar procedures. All of these protocols receive extensive scrutiny and have to be backed up by a strong scientific rationale. Procedures like the ones described in Yu et al would have zero chance of being approved.
Ethical use of animals in research requires a judgement of the balance between the harms done to the animals and the benefits accruing from the research. That’s the job of the ethics committee and it would be wrong for an article with such approvals to be retracted. On a practical note, journals simply cannot review the judgements of every ethics committee for every piece of work.
The techniques used here are used not infrequently to model certain aspects of human diseases including aspects of depression. Of course, depression is a human disease and so a model can only ever be a model, but they can be instructive none the less. As Siim points out, the experiment here claims evidence for causality, that human observational studies cannot address.
One way of enhancing benefit is to increase the value (reliability) of the research, and here there is a problem – although it is a problem shared by a large number of studies in biomedicine, not just those involving animals. While they claim randomisation, they did not describe how this was achieved; they do not report that they blinded their handling of the animals in the days post intervention or that they blinded the samples when they assessed the outcome; and they do not report why they picked 8 animals per group (rather than 6, or 10, or 60).
These are more important weaknesses and are things that we should be doing our utmost to fix. if there are concerns about the decision making of certain ethics committees, that is a something the journal should take up with them. The scientists cannot choose to which ethics committee they submit.
Your statements raise a lot of questions about the medical research industry. Because it is an industry devoid of wisdom, those outside of the industry must question its practices and hold it accountable for its consequences. For one, questioning what benefits for humans research on animals yields; this is a topic that needs a lot of investigation and discussion. Even when there is some benefit for some humans, the medical industry is unable to evaluate the benefit in any terms other than financial, which is why the medical industry has used tens of thousands of primates, directly contributing to the point of them being pushed to the brink of extinction, just to try to find new “products” to increase revenues. Another question we must be asking is why is it that so many laboratory researchers avoid the topic of ethics altogether, by saying it is the responsibility of the IRB, not the scientists, to determine whether a study is ethical. That kind of thinking makes the animal laboratory scientist into an unquestioning robot worker (or torturer if it takes pleasure in his job, which many do) who just does what it is told by those who design the studies for the profit of large medical corporations, and is “off the hook” because it is just following orders. This is the reason the Nuremberg ethics were created in the first place, to declare that no scientist (and no military officer) can be “off the hook” because its superiors were giving orders out of a sense of personal gain or profits. You are describing the discarding of the Nuremberg ethics which is the main document that is supposed to guide research ethics in all of the human sciences.
Hi Malcolm,
While I of course agree that we should do as much as we can to improve methodological standards, including by being more demanding when we peer-review, I disagree that one should not judge on the ethical (and social) acceptability of animal research when peer-reviewing. This is defined by the balance between harm, benefit, and now also ‘likelihood of benefit’ (proposed as a third factor in EC guidance for project evaluation, turning the ‘Bateson Square’ into a cube) so all factors are, in my view, indissociable.
I won’t dwell on the specifics of this particular case, but make the general case that journal editors and reviewers have their share of responsibility for the ethical treatment of animals, not only by applying their judgement on what is acceptable (granted, this can be subjective, but arguably everything else in peer-review is, also) but also by helping define through peer-review what is acceptable. The same goes for scientific standards, of course.
Also, I have discussed elsewhere that (reported) ethics review is not a guarantee of ethical treatment of animals (https://goo.gl/gpBS5K). This may result from low standards or outright deception by authors. And even when proof of ‘ethics review’ is presented it often happens (and I’ve witnessed that first hand) that scientists will ask committees to do so ‘a posteriori’, usually because they were prompted by a journal to do so. This makes it the more important for journals to be, as proposed by Bernard Rollin, the true “guardians of the gates for animal welfare”.
As scientists we carry out animal experiments under a tacit (and often explicit) social licence that is based on public trust that our work is both relevant and humane. If we demand of authors of the manuscripts we review to uphold high standards to ensure the former, can’t see why we should not do the same regarding the latter.
The international community needs to come up with international ethics standards for animal and human research. At the very least, journals (editors and reviewers) of a country should not publish research that is against their country’s standards. Otherwise, researchers will bypass their ethics committee and conduct/outsource their research to the country with the lowest standards.
I would say that this applies to a lot of other ethical issues as well – workers’ rights, financial transparency etc. Unethical medical research taking place in a small number of countries that permit it is similar to offshore tax havens, sweatshops etc. – people in Country A where something is banned profit from it being carried out in Country B.
I agree that institutional ethics boards are doing a good job stopping overuse of animals and unnecessary cruelty. But, in my opinion, they are often in a difficult position with regard to the issue of clinical relevance of the animal models offered for their consideration. In this respect, I am curious if a modern ethics board would have stopped the aforementioned Dr. Selye from creating a model of “dermatomyositis”, especially after he would have admitted that he intended to create an experimental disease “without attempting to decide…whether it is closely related to the dermatomyositis of man”. https://www.ncbi.nlm.nih.gov/pubmed/?term=Selye+AND+dermatomyositis
Apparently, only animals in the experimental group suffer greatly, and all proper controls are set, thus increasing substantially the number of animals required.
Fully agree with Dr. Elies Bik. There is absolutely no need of using animals in such studies. But policy makers and rule setters don’t have much knowledge about these things and that is why rules differ from region to region.
Uniform international ethics would deal with the problem more efficiently but we don’t have an scientific body that is global to set the rules. In such a scenario peer-reviewers and publishers have a great role to play to stop these unethical practices.
What the article sadly lacks is a critical examination of whether the animal model is predictive of human outcome with respect to drug testing and human disease. Animal researchers continue to receive public funding as a result of public ignorance and an ethical approval process that does not require proof that animals models are evidence based with respect to human disease and drug testing.
You are absolutely right.
The procedure to induce (or rather to attempt to induce) a depressive state in rats seems to be an extreme variant of highly doubtful validity of the Unpredictable Chronic Mild Stress procedure (disclosure: my colleagues and I have used UCMS in mice). Note the word “Mild”. The stressors used in the above study are anything but mild. If this is a model of anything at all, it’s more likely PTSD than depression… Five hours of restraint stress or 48 hours of water deprivation are quite extreme. (The latter must be close to being deadly…) Swimming in 45C water (even for only 5 minutes) seems extreme, too. It must raise the core temperature of the animals to more than unpleasant levels. The stroboscopic lights could have been used for, say, 10 minutes instead of 4 hours (!). Swimming in 15C water for 5 min seems unpleasant, but not as extreme as the other stressors and more in the range of what is normal in this kind of research. So are the foot shocks (2 shocks of 1 mA of 2 sec within a minute), although I am assuming that they did this for only one minute at a time (i.e. 2 shocks/session; the 2011 article is not clear on this). For mice food deprivation for 48 hours would be extreme and result in a decrease of body weight of at least 20%. I don’t know how this works out in rats (I work with mice myself, they have a faster metabolism than rats). The 110dB noise stimulus is what is routinely employed in studies of the startle response or pre-pulse inhibition in both mice and rats. Unless this was applied for a long time (again, the 2011 article is not clear about this), this would not be extreme. In all, I agree that this is not an acceptable procedure.
I am not a scientist. I have one comment, is there a PROPORTIONALY HUMANE animal suffering as opposed to “disproportionally inhumane animal suffering”?