The ARRIVE guidelines are a set of reporting suggestions for reporting animal work. These follow on from the CONSORT guidelines about reporting clinical trials in humans. The aim was to improve the transparency of animal studies to help them be more reproducible and translatable.

According to the authors the ARRIVE 2010 have not been fully adopted, will the 2019 revision improve this?

I suspect not, because again they request behaviours that people do not do and a standard that probably no one currently adheres too.

Is it wrong to have aspirations? Absolutely not and many of the elements help with experimental design. With time change occurs

In 2010 the guidelines were written, not be jobbing vivsectionists, but probably people with a 3Rs leaning and included 20 aspirational guidelines. I think no one was achieving at the time they were written. Likewise I suspect there are still nobody achieving these now. These were (1) common sense things that every body does, so not a problem (2) Things that people do but weren’t usually reported so not a problem to report them but sadly there were (3) things that essentially nobody does and were unlikely to be adopted as they would probably be a “reject” signal if discussed. (4) Nobody does as they did not incorporate current publication restrictions (word length and citation restrictions) and the fact that they were not clinical trials but animal studies using a common protocol used many times and the work evolved and you have reported them once so why do it over and over again. (5) Things you would have reported during the ethical review process and every one knows the stock answer and then there were (6) A few suggestions concerning the 3Rs, that no one was really going to adopt and shouldn’t have been part of the original submission. I use mice because they are…blah blah blah, blah not a cheap mammal etc. The 3Rs value of the experiment, when most experiments do not concern themselves with this stuff etc.

The guidelines have now been revised ARRIVE 2019 and are under review. This puts the guidelines into ten core reporting elements and eleven recomended elements. So again we have guidelines, which are aspirations. I hope industry review this paper. They work to better standards than the academic. I hope it does not simply gets reviewed and goes through on the nod.

The 2019 variant wants animal experiments to become like human trials and be registered like at clinical (USA) or EurdraCT (EU) before the trial is undertaken. In human trials it aims to stop data hacking (selecting a positive outcome as the main endpoint once you know which one gives a postive response after the trial has finished) and also indicates that drug X has been tested. However, this is not perfect, we see that trials are only reported retrospectively and not all trial outcomes are reported and many of the trials registered are not undertaken. However, there is yet to be a site where these animal studies are deposited as there is surely no point having them all over the place.

I know if animal studies have their protocol registered it may help reduce repetition of studies, but how many cannabinoid in EAE studies have been done since we reported they were a waste of time in 2003. A search of “cannabinoid and encephalomyelitis” picks up 115 papers and counting. Our negative data have been airbrushed from history by the positive studies, many using doses of tetrahydrocannabinoid that would make the animal unconsious for hours (but not reported). I find it hard to be bothered to review or read these. Editors want these as Weed gets them media attention which is good for their impact. The human data shows that we were likely to be correct, but we still we see animal study after animal study.

The papers suggest that you can deposit these at certain sites and keep them secret until after the work is published. But what is the point if it is not searchable? If embargoed until published, again what is the point. Our recent paper on NOD EAE used data from 1996 and 2016. A long time to embargo it. In EAE the outcome is the clinical score, however we have seen how this outcome is abused it has nothing to do with data hacking it is mis interpretation and misrepresentation. I personally do not believe that transparency is the central problem of lack of translatability, but I have discussed this before.

However, why would you want to declare to the world your experiment before it is done. It is not about the fear of data hacking, but it would be the fear that you will be scooped once someone says that’s a good idea and does it quicker than you, but importantly you have destroyed your chance of obtaining a patent. Public disclosure before the patent is filed and the patent is worthless. Filing the patent early wastes years of the drug development life and you can’t really file a patent until you know the result. Once filed you have 18 months before the patent is visible, then the costs start to mount. This is very important as the patent is what allows companies to develop agents. This is why the industry view is essential. It is a massively expensive business filing patents and the lawyers win, win, win. You have to file them in each country and you have to translate the patent into each language. It cost us $20,000 to get a patent translated in to Japanese for example. There was someone from a company on the author list surely they should have thought about the patent issue. Companies will not do it I suspect and they do not give a stuff about publishing animal data, it often only sees the light of day after the human trial is being undertaken. Who would know that FTY720 is fingolimod and BG12 was dimethyl fumarate.

However, I support the concept of improved reporting. I believe that EAE studies really need to develop some quality control as there is a lot of un-reproducible guff out there. This will not translate to another lab let alone humans.

This leads to unnecessary animal suffering, alot of false hope on your behalf and sadly useless clinical trials. This is why I have been critical, including being publically critical of some studies. I really want the animal science to improve. If it doesn’t it will be more quickly marginalised.

In the UK it is increasingly difficult to do animal work. As you have seen, funding has been cut from some of the large science centres in the UK, we have pro anti-vivisection ministers at the helm of science and animal work is being increasing regulated and stuff is moving West and Eastward, where animal welfare is lower on the agenda.

Immunological MS model studies have recieved little support for over a decade nearly two, yet it is still going strong in Germany (with alot of transgenic work) and USA. There is more coming out of Far East, some with dubious ethical standards.

Anyway the ARRIVE guidelines 2010 were asking for behavioural change. They were a lot to do in one leap for some and as the “big swingers” were often the problem people who were opinon-leading bad habits. I doubt some things have changed. Sample size calculations are seldom reported still. Likewise, the guidelines in their entirity did not and still do not incorporate/accomodate publication practise. If you go to journals were they have a few thousand word and a thirty reference limit you simply cannot report everythng.

We decided we would do one paper where essentially everything requested of the ARRIVE guidelines were reported and would cite that in every paper. So one reference and few words. Animal studies are not clinical trials and you do the same protocol with different agents over and over again. You are not going to waste words repeating this over and over again.

This new paper fails to report on whether and where the ARRIVE guidelines have helped to deliver change. However they acknowledge that “Despite considerable levels of endorsement by funders and journals over the years, adherence to the guidelines has been inconsistent, and the anticipated improvements in the quality of reporting in animal research publications have not been achieved”. Sadly they have not reported a re-analysis of animal papers. We looked two years in and there was not much change.

The only way to get real change is for the journals to pick them up and enforce change, this way it is now standard that an ethical review statement is given. However it is often hollow as you still see dubious stuff reported that clearly do not consider animal welfare, yet it is reported to be ethically reviewed. We still see papers reporting practises that were outlawed in the UK 25-30years ago, like bleeding mice from their eye sockets. However, the ARRIVE guidelines were probably too big of a jump and there is nothing consistent. Eg. if you submit to the Nature Stable journals, each one has its own requirements making it a pain in the bum if a papers get rejected from one and you want to submit to another. Some journals play lip service to the principles and other use it as a way to boost their impact factor by making you cite certain papers. This was a British led thing and frankly many non-British scientists can’t be bothered by enforcing these during the review process and indeed many scientist have no idea what they are. An American group came out with 7 guidelines.

I once had a paper where I ask the authors to reubmit their paper with attention to the requirements of ARRIVE, their response was “we did the experiments according to the ARRIVE guidelines”. So they couldn’t be bothered to even read them as their answer to my request was perhaps the only wrong one.

The ARRIVE guidelines 2019: updated guidelines for reporting animal researchNathalie Percie du Sert, Viki Hurst, Amrita Ahluwalia, Sabina Alam, Marc T. Avey, Monya Baker, William J.Browne, Alejandra Clark, Innes C. Cuthill, Ulrich Dirnagl, Michael Emerson, Paul Garner, Stephen T.Holgate, David W. Howells, Natasha A. Karp, Katie Lidster, Catriona J. MacCallum, Malcolm Macleod, OlePetersen, Frances Rawle, Penny Reynolds, Kieron Rooney, Emily S. Sena, Shai D. Silberberg, ThomasSteckler, Hanno Würbel bioRxiv 703181; doi:

Reporting animal research: Explanation and Elaboration for the ARRIVE guidelines 2019 Nathalie Percie du Sert, Viki Hurst, Amrita Ahluwalia, Sabina Alam, Marc T. Avey, Monya Baker, William J.Browne, Alejandra Clark, Innes C. Cuthill, Ulrich Dirnagl, Michael Emerson, Paul Garner, Stephen T.Holgate, David W. Howells, Natasha A. Karp, Katie Lidster, Catriona J. MacCallum, Malcolm Macleod, OlePetersen, Frances Rawle, Penny Reynolds, Kieron Rooney, Emily S. Sena, Shai D. Silberberg, ThomasSteckler, Hanno Würbel doi:

Why write this now…at least it gives an altmetric and links to the paper before it is published in final form so the debate can start. What do you think? Do you think studies should be registered?

About the author



    • Almost all studies listed on a public register are examining drugs that have a current patent or the patent has lapsed. Patents are generally filed as late in pre-clinical development as possible. This maximises the patent life of the drug once (if) it is registered. This ensures exclusivity and $$$$.

      If pre-clinical studies are reported, and the drug doesn’t yet have a patent, you are potentially inviting competition. If the competition manufacture a similar drug and patent it before you, your program could be sunk before it begins.

    • The patents are filed before the clinical trials are reported or they do the trials, the ones filed after the trials are undertaken are generally about dosage and you find that the information in the registration sites are not given until after the patent is filed.

  • It remind me of a paper i once read about

    TBI mouse models,.where they let some weight fall into the head of a living mouse to smash it and them (only them )

    The mouse gets euthanize


  • Patent issue aside, yes I think the pre-clinical studies should be registered (for all the reasons you detail above). Even with registration, the standards of approving bodies will differ (even with the guidelines in play).

    • How do you do this on a searchable database? In the Netherlands for example each experiment needs a DEC form so they are registered but no-one can search them

By MouseDoctor



Recent Posts

Recent Comments