1. Why use animals at all?

1.1. Models of human normality and disease

Animals are used in research when the biological systems being studied are too complex to be modelled in any other way. For example, our immune systems are composed of very many different cell types, located in different parts of our bodies. We can study some of the properties of each of these individual cells in isolation, having harvested them from a dead animal, and this is indeed a very informative approach. However, at present, we can only see how they work together to produce a reaction, for example to a vaccine, by studying the immune system in an intact organism. Likewise, the control of blood glucose levels (which goes wrong in diabetes), involves not only the pancreas (where the hormone insulin is made) but also our muscles, fat stores and the brain.

More and more complex biological systems are being modelled successfully, so the range of non-animal subjects for productive research will continue to grow. It is a central ethical principle, also enshrined in UK law, that, when such opportunities to replace animals with other equally informative experimental systems arise, then these must be taken.

There is another important ethical principle embedded in the law, that research carried out using animals must have a foreseeable benefit and not be driven by curiosity alone.

1.2. Use in the University of Dundee

Animals are used in research in the University in two connected ways:

  1. As examples of normal physiological systems. How are these controlled (down to the molecular level)? How do they respond to changes in the environment? In what ways can these control mechanisms become disrupted and how does this lead to disease?
  2. As models of diseased or abnormal states. Can these be reversed, or at least slowed, by interventions that might form the basis of new treatments?

While the number of medical treatment and prevention options to improve human health have increased dramatically over the last few decades, there remain a number of very challenging areas, such as:

  • cancer
  • degenerative diseases of the nervous system (e.g., Alzheimer's disease and Parkinson's disease)
  • diseases associated with the global obesity epidemic (e.g., diabetes and cardiovascular disease)
  • "neglected diseases" (often involving parasitic infections and prevalent in some of the poorest countries in the world)
  • diseases involving imbalances in the immune system (e.g., multiple sclerosis, lupus and severe combined immunodeficiencies)

The University has important research programmes in all these areas. We need to have much better understanding of how the conditions first arise if new methods for prevention, early detection and treatment are to be developed. The initial "benefits" are therefore in terms of enhanced knowledge, with the expectation that this will become the basis of improvements in human health. For example, if we understand the molecular details of how certain parasites survive, reproduce and associate with their "host" organisms, we can then identify targets for the development of entirely new drugs that will disrupt these processes, while having only minor and acceptable effects on the human or animal host. This onward development might also occur in the University or almost anywhere else in the world, spurred on by the basic findings coming out of Dundee and collaborating institutes.

2. The 3Rs

The University constantly strives to improve animal welfare and to minimise animal suffering through implementation of the "3Rs". These are:

  • The REPLACEMENT of animals, wherever possible, with alternative systems with which to conduct research.
  • The REDUCTION of the number of animals used to the minimum consistent with the desired outcome, e.g., by detailed review of the existing scientific literature to avoid unnecessary repetition and by careful design of new experiments and rigorous statistical treatment of their results;
  • The REFINEMENT of procedures so that the least possible harm is done in collecting the scientific information;

The ethical review process acts as a central hub for the dissemination of best practice in the 3Rs within the University, so that an advance made by one individual or group can be swiftly adopted by as many others as possible. It also connects with similar processes in other establishments, to allow even wider sharing of best practice.

There is a National Centre for the 3Rs (https://nc3rs.org.uk), which exists to drive progress in all three of these principles across the UK. Some research projects in the University are aimed specifically at enhancing the 3Rs and others have been recognised (in one case by the award of a prize from the National Centre) as having made significant contributions in this area as well.

3. What do we mean by "animal"?

To a biologist, the animal kingdom encompasses a huge range of organisms, from single-celled microbes to humans. A more restricted definition is used in UK law, which defines an "animal" to be any vertebrate (i.e. with a backbone) except humans, and any cephalopod species (octopus, squid, cuttlefish and nautilus). The reasoning behind this choice is largely based on "sentience", i.e. our expectation that an animal can perceive events as being unpleasant, distressing and/or painful. This is, of course, a subjective judgement and, while "lower" species such as worms and insects should be used wherever possible instead, there may still be ethical questions that need to be addressed as to their welfare, and whether they in turn can be replaced in future by even less sentient systems.

The UK law provides additional protection to non-human primates (monkeys, apes, etc.) and to species commonly thought of as companion animals, i.e. cats, dogs and horses. In other words, having made a case why the use of animals in research is necessary, someone must then have further justification why only members of these species are suitable.

4. Species in use in the University

The University currently has facilities for the housing of mice, rats, some other small rodent species and amphibia. It currently has no facilities suitable for housing other legally recognised "animal species", though some research is done on insects (fruit flies) and on nematode worms. At present, there is little or no field research carried out on wild animals or birds.

5. Legal regulation

In the UK, the use of animals for scientific and education purposes is regulated under the Animals (Scientific Procedures) Act 1986 (ASPA). The law was modified slightly in 2012 to bring it into complete compliance with updated EU legislation. Although the UK has now left the European Union, no significant changes to the national legislation have occurred or are expected as a result.

ASPA controls the performance of "regulated procedures" on protected animals (see above). A regulated procedure is defined as one that

  • is carried out for a scientific or education purpose, and
  • might cause the animals pain, suffering, distress or lasting harm.

The threshold is set at the likely pain or distress caused by the skilled insertion of a hypodermic needle (i.e. equivalent to the fleeting pain we experience when receiving a single vaccination).

While the law allows for the performance of regulated procedures on living animals for education purposes, no such procedures are currently carried out at the University. The only use of regulated procedures here is therefore for a scientific purpose.

There are some procedures on living animals that are not regulated. These include:

  • procedures which cause less pain, suffering, distress or lasting harm than the threshold. Examples include observational or behavioural studies;
  • Certain methods for the humane killing of animals, listed in Schedule 1 to ASPA. These might be used, for example, to kill animals prior to harvesting tissues or cells for detailed laboratory study.

The University breeds and uses genetically altered mice, as these are invaluable tools for studying the effects of individual genes or groups of genes on normal physiology and on the development and treatment of disease. These animals, however normal their welfare and behaviour appear to be, are regarded simply by existing as being on a regulated procedure. In other words, the benefit of the doubt is given to the animal.

Legal regulation is by the granting of licences by the Home Office and by the scrutiny of our facilities by the Home Office inspectors. There is an Establishment Licence, whose holder accepts overall responsibility for the University's compliance with the law. Project licences are issued to senior researchers and describe the regulated procedures to be performed, the scientific necessity for those procedures, their expected effects on animal welfare and the steps that will be taken to reduce these effects to a minimum. Personal licences are granted to individuals who actually perform the regulated procedures. There are approximately 45 project licence-holders and about 100 personal licence-holders in the University.

6. Ethical review

The ethical issues arising from the use of animals in scientific research can usefully be debated in terms of the balance between harms and benefits. "Harms" in this context relate to the deviations from normal welfare experienced by the animals undergoing regulated procedures. "Benefits", as described above, relate to the potential for better human or animal health in the future.

The "harm / benefit" analysis is carried out in the University's ethical review process. This involves a detailed analysis of research proposals by a welfare and ethical use of animals committee, which comprises a mixture of scientists and non-scientists and includes people with specific animal welfare expertise (e.g. a veterinary surgeon). As well as the consideration of applications for projects to use animals, there are processes for monitoring progress and compliance of on-going research programmes, overseeing the training and competence of University members who work with animals and maintaining the standards of housing in which animals are kept.

7. Animal numbers

The numbers of regulated procedures carried out in the University over the last few years are shown in the table and as a graph.

TOTAL 43656 34895 33649 35860 29512 20262 22502
Mouse 41953 33861 32725 34347 28421 19928 22336
Rat 938 608 539 1201 988 256 155
Frog 699 392 373 312 103 78 11
Hamster 66 34 12 0 0 0 0


There was a substantial fall in the number of animals used in 2020. This was due to the general restrictions on research activity in the University; part of its response to the Covid-19 pandemic. Numbers rose again in 2021, but not to pre-pandemic levels.

The University makes extensive use of genetically altered mice in its research programmes, primarily to develop our understanding of how certain genes affect the development of disease and the response to therapies. While the great majority of these animals apparently experience completely normal welfare throughout their lives, the law requires us to view their maintenance and breeding as being, in themselves, regulated procedures. In 2021, a total of 12,995 mice were recorded as being used for the maintenance of established breeding colonies of genetically altered mice, and a further 168 mice for the generation of new colonies. The number of regulated procedures in mice in that year, excluding those in breeding programmes, was 9,173.

The University also uses some animals for scientific purposes where licences are not required. In 2021, 2,176 mice and 733 rats were used in this manner. All the animals were genetically normal. Some of the mice and all of the rats were killed humanely to provide tissues and cells for detailed laboratory study. The remaining mice were genetically normal animals that were generated in the breeding of genetically altered mice.

8. What actually happens to the animals

8.1. Prospective severity

Every experimental protocol listed in a project licence has a severity category attached to it, representing the greatest deviation from normal welfare that is expected, even in only one animal. The categories are:

  • Mild
  • Moderate
  • Severe
  • Non-recovery

The lower threshold in the "mild" category is set at a deviation equivalent to the transient discomfort experienced during the single skilled insertion of a hypodermic needle. If the greatest disturbance to the animal is expected to be less than this, then the procedure is not regulated and does not require licensing.

The borders between "mild" and "moderate" and between "moderate" and "severe" are not so well defined in general. Each protocol, however, will describe the relevant humane end-points that will be applied and that therefore, for that protocol only, define the severity category. For example, an experimental intervention that is expected to cause the animals to lose weight might have a humane end-point set at a loss of 20%. So long as this loss is expected to take several days, then a "moderate" severity category might be appropriate.

The experience of unremitting severe harm is regarded as always being unacceptable and there can be no scientific benefit that would out-weigh it. This level of suffering must never be the expected outcome of an experimental intervention.

The "non-recovery" category refers to procedures in which the first action is to anaesthetise the animal and then it is not allowed to recover. Apart from any minor discomfort during induction of the anaesthesia, the animal will therefore feel no pain at all.

It is important to note that there must never be any unnecessary pain or distress, whether it be "moderate" or even "mild". Experimental procedures must always be refined to cause the least deviation from normal welfare, regardless of the limits set on severity in the protocols. It should therefore be the case that the great majority of animals on which a particular procedure is carried out will experience much less deviation from normal welfare than that defined by the severity category (see below).

In the University of Dundee at the present time, there are no experimental protocols with a category of "severe". Most project licences have a mixture of protocols in the "mild" and "moderate" categories, and few have one or more "non-recovery" protocols as well.

8.2. Actual severity

When a procedure has been completed, then the actual lifetime experience of the animal must be assessed and assigned a severity, which may be less than, equal to, or more than the category that was applied prospectively to the experimental protocol. For the reporting of actual severity, an additional category is allowed, making the choice between one of the following:

  • Below threshold
  • Mild
  • Moderate
  • Severe
  • Non-recovery

"Below threshold" means that, as far as anyone could tell, the animal actually experienced normal welfare throughout its life and was killed humanely. It often applies to the breeding and maintenance of genetically altered mice in the "mild" category, where no experimental intervention was necessary during the animals' lives (perhaps because tissues and cells were harvested post mortem for scientific study in the laboratory).

In most cases, the assignment of actual severity is reasonably straightforward, using the humane endpoints in the project licence as a guide. So, in the weight-loss example given above, if the animals actually lost very little weight over the duration of the experiment, an actual severity of "mild" might be appropriate, even though the category assigned to the protocol was "moderate".

It is important to understand that the aim of the severity classification is to assess the harm caused by the scientific intervention only. As background sickness is very rare in modern high-health rodent colonies, it is usually safe to assume that any deviation from normal welfare during a scientific procedure (or in the maintenance of a line of genetically altered animals) is related to the scientific intent and must therefore be included in the assessment.

A particular difficulty, however, arises when animals that appeared to have been in good health all their lives are unexpectedly found dead. "Background mortality" is never zero in large colonies of mice or rats so, if an animal is found dead, one cannot be certain that the event was related to the scientific procedure or the genetic makeup. Background deaths are often associated with giving birth, the early days of post-natal life, or old age (much as in humans), but can occur at any stage of an animal's life.

The law gives us some helpful guidance on how to assign severity to these events. If it seems that animals are found dead more often in genetically altered colonies than in "normal" animals, then these events are associated with the genetic alteration and must therefore be included in our assessment of severity. However, such events are always rare, so the colonies may have to be quite large before we can be reasonably sure of this. In the meantime, we will usually give the benefit of the doubt to the animals; if they were genetically altered, then we assume that the events were associated with that. As we gather more data, we may find that there is no evidence for such a link, in which case we no longer consider the events in our assessment of actual severity.

Then we have to consider the severity of the event itself. Competent humane killing of an animal is not assigned a severity, so the question is the extent of any pain or distress the animal may have experienced since last being observed as being in apparent good health. If we know or can show from post mortem analysis that death was relatively quick, then we may not have to consider it as particularly severe. However, most often we do not know this for sure, in which case we again give the animal the benefit of the doubt and assign a "severe" category to the event.

8.3. What if something unexpected happens?

Science is all about investigating the unknown and therefore the results of experiments cannot be predicted accurately in advance. We make our best assessments as to what we expect the welfare effects of a scientific intervention will be, but sometimes we are confronted by unexpected events. These may be of a different nature to the expected ones, e.g. we observe a mobility problem, as well as or instead of, an expected weight loss. Or they may be of an increased magnitude or speed, e.g., we observe a weight lost of greater than 20% when we expected a lesser one. Either type of unexpected event can result in an assignment of an actual severity greater than that originally predicted.

All such expected events have to be dealt with in the best interests of the animal and then reported to the Home Office, who will determine which of the following options should apply:

  • This particular type of experiment should be stopped (most likely because the event jeopardises the scientific validity and therefore there is no longer a justification to proceed);
  • If the effect can now be managed and minimised within the original severity category (i.e., it now becomes an expected one), then work may be allowed to proceed. Again a Home Office inspector has to be convinced that, while the animal may experience no greater harm, the scientific justification is still strong enough to out-weigh it;
  • If the work is considered of sufficient scientific value, then the inspector may allow a temporary or permanent increase in the severity category of the protocol. This is never a decision taken lightly; all steps to avoid or minimise the effect must still be taken.

8.4. What happens in Dundee

In 2021, of the 22,336 procedures carried out on mice, 15,484 (69.3%) were assigned an actual severity of "below threshold". As above, these were genetically altered animals in breeding programmes, where there was no evidence at all of anything other than normal welfare. A "mild" actual severity applied to a further 5,276 mice (23.6%) and a "moderate" severity to 1,181 mice (5.3%). These experiments did involve expected and justified deviations from normal welfare, within the limits laid down in the project licences. A total of 311 mice (1.4%) underwent procedures conducted entirely under irreversible anaesthesia, formally classified as "non-recovery".

In the mouse breeding programmes, 63 animals were found dead; 0.5% of the 13424 animals in those programmes. For some of these animals, there appeared to be a clear genetic link. In other cases, we could not be certain that the events were not related to the genetic alteration, so the rules laid out above were followed and all these events were assigned to an actual "severe" category.

In the mice used for scientific interventions, 21 events (0.2% of the 9,341 animals used) were assigned to an actual "severe" category. None of these were expected and all were investigated and reported to the Home Office at the time.

In rats, 18 procedures were assigned a "mild" actual severity, and 54 a "moderate" one. There were 83 "non-recovery" procedures. All 11 procedures carried out on frogs were assigned an actual severity of "mild".

While the "severe" events constituted only 0.4% of the total number of procedures and may have been over-, rather than under-reported, it was very important to investigate each one to ensure that, if it could be avoided in future, it would be.

The percentages for 2021 were very similar to those in pre-pandemic years.

9. Relevant websites

  1. The Institute of Animal Technology (http://www.iat.org.uk). This is the professional body for animal care staff working in the scientific domain. The University is a corporate member.
  2. The Concordat on Openness on Animal Research (http://concordatopenness.org.uk). The University is a signatory to this important agreement.
  3. The University has endorsed the ARRIVE guidelines (http://www.nc3rs.org.uk/page.asp?id=1357), which are intended to improve the reporting of experiments involving animals in academic publications.
  4. The Home Office (https://www.gov.uk/research-and-testing-using-animals).