Ethics of Research involving animals
Issues concerning the welfare of laboratory animals in toxicity testing
9.26 We have commented on the numbers and types of animals most commonly used in toxicity testing (paragraph 9.8). We also observed that some toxicity tests may extend over several months or years in contrast to most animal experiments conducted for biomedical research. For rodents, age-dependent health problems, with concomitant stress, will usually occur with increased frequency towards the end of tests. Loss of animals can compromise study validity and confound the interpretation of data, especially from carcinogenicity studies.11 This may sometimes encourage investigators to minimise animal loss by avoiding euthanasia as far as possible, which may result in increased pain and distress to the animals.
9.27 It is impossible to fully predict the pain and suffering that individual animals might experience during toxicity testing. However it is possible to assess the likelihood that pain and distress will occur under a particular set of conditions and exposures. The following aspects of toxicity testing can give rise to adverse consequences for the welfare of test animals, the extent of which depends on the test and species involved: (i) transport (see paragraph 4.36); (ii) housing and husbandry (see paragraphs 4.37–4.43);12 (iii) dosing and sampling procedures (which might be repeated) (see paragraphs 4.49–4.52); (iv) the length of the observation period and (v) the toxic consequences of dosing. The adverse effects on animals that may arise specifically in toxicity tests, as opposed to other forms of animal research, are due mainly to dosing procedures and the toxic effects of the treatments.13
9.28 Dosing can involve the repeated administration of test material by a variety of routes of exposure, including gavaging (stomach intubation or forced feeding), injection, skin painting and inhalation. Some types of administration are likely to be very stressful to animals, especially when they are repeated and are of relatively long duration (see paragraphs 4.45 and 9.28). In addition, dosing into the eye and inhalation exposure involve restraint for several minutes or hours.
9.29 The right choice of dosing vehicle and volume is an important means of refining toxicity testsfrom both scientific and welfare perspectives. This is particularly so regarding the maximum amounts that should be administered to the eye and orally by gavage.14 The use of low dosing volumes is a very effective way of reducing stress during topical ocular administration. Thus, the traditional dosing volume of 0.1 ml can be reduced by a factor of 10 or even 20 in eye-irritation studies. During gavaging, volumes of 1–50 ml/kg are usually administered, depending on the species being used. The administration of large volumes through this route can modulate the patterns of absorption, thereby affecting toxicity. For example, volumes nearing or exceeding the stomach volume will result in the delivery of some of the substance to the small intestine.
9.30 Stress can also be induced by physiological changes accompanying oral dosing. For example, alterations to gastric secretion and motility, as well as increases in heart rate and blood pressure, can occur. There can also be changes in biochemical parameters, such as levels of stress hormones. Furthermore, under conditions where animals are fed in laboratories ad libitum, as is the usual situation, gavaging of large volumes may result in aspiration of the test substance due to the presence of food in the stomach and duodenum. The volume of the gastrointestinal tract for receiving administered material is reduced and injury to the lungs may ensue. Recent research showed that gavaging rats with corn oil, but not the test substance or water, resulted in stress which was volume-dependent, as manifested by corticosterone levels (a hormone released in response to stress).15 The authors recommended that dosing volumes for rats should not exceed 10 ml/kg. It is important to consider this information in the light of other best-practice guidelines on dosing.16 At the same time, views differ as to how widespread the gavaging of large volumes ad libitum is in practice, and some researchers comment that significant steps have been made to refine the method.17
9.31 In metabolism studies, animals are housed in metabolism cages and might have external tubes implanted into their bile ducts.18 During toxicokinetic studies in dogs, it is not unusual for the same animals to be reused after a suitable period of time, as such animals are thought to suffer less stress than those used for the first time.
Effects due to toxicity
9.32 The usual practice in toxicity testing is to induce overt toxicity in some animal groups, in order to ensure that, where toxicity is not observed in other exposed groups, the effects are not due to any inherent defect in the methodology. Thus, some form of harm to animals is an integral part of animal-based toxicity testing and is viewed by those conducting such tests as being unavoidable to achieve the scientific objectives of the work.
9.33 Toxicity can arise from reversible or irreversible effects, and can affect a range of different organs to different degrees. The adverse effects of substances on animal physiology can range from minor changes, such as reduced weight gain, small physiological alterations or changes in the levels of circulating hormones, to severe effects such as organ function loss (a major cause of acute toxicity), leading to death. Intermediate levels of toxicity, such as those destroying tissue and adversely affecting tissue function, could result in pain and suffering. Similarly, the development of tumours during carcinogenicity testing, or intestinal swelling during sub-chronic or chronic testing, might also lead to pain and discomfort.
9.34 The adverse effects which are used to define the MTD range from the very mild, which include non-clinical signs of lethargy or effects on weight, to the more substantial, such as convulsions. For example, various tests of toxicity often require signs to be scored, such as changes in the condition of the coat and eyes, as well as other signs of ill-health. Many of these conditions might be expected to reflect pain and suffering to differing degrees.
9.35 There is general confusion among toxicologists as to exactly what defines an MTD, ‘severe distress’, ‘obvious pain’, a ‘moribund condition’ and other descriptions of animal welfare. Some have argued that the relevant OECD test guidelines need to be revised accordingly.19 Several of the OECD test guidelines are vague on issues such as environmental enrichment, where for example group housing is not specified when it would be possible.20 All these ambiguities can act as potential sources of avoidable suffering for the animals.
9.36 Other examples of toxicity endpoints that are likely to be painful and stressful include skin irritation and corrosion where single doses are applied to shaved areas of the backs of rabbits. Exposure can extend over four hours, and the animals may experience ulceration of the skin as well as swelling and itching. In sensitisation testing, multiple dosing is practised, and in addition to the above signs, the skin may crack and peel. Other signs that can be observed during acute, sub-acute and chronic toxicity testing include both external and internal bleeding, diarrhoea, loss of appetite, vomiting (in non-rodents), aggression, salivation, changes in blood pressure, coma, convulsions, lateral recumbency and tremors, loss of fur and hair, dehydration, or nasal discharge. Some of the less drastic effects of toxicity can arise merely from the act of dosing.
9.37 Very severe adverse effects can become manifest extremely rapidly as a result of neurotoxicity following dosing. For example, during the mouse bioassay for diarrhoetic shellfish toxins, atypical results21 can arise which cause rapid death, following signs of substantial distress from shock and extensive trauma, accompanied by violent and rapid leg and body movements and agonal breathing (abnormal and uncertain respiration often characterised by gasping for breath), collapse and finally death from heart failure.22
General observations concerning the assessment of animal welfare in toxicity studies
9.38 It is difficult to assess accurately either the individual or the collective burden of suffering that is sustained by animals used in toxicity testing. Many toxicity procedures do not usually result in more than some discomfort to most of the animals concerned, at least in the case of rodents. Moreover, only certain test groups of animals will be subjected to tests leading to overt signs of toxicity during an experiment. These groups of animals comprise the concurrent positive controls (animals treated with a chemical known to have adverse effects as a comparator on the sensitivity of the test substance) and those animals that receive high doses in dose-response studies. However, in such cases it is likely that significant pain and distress could result, depending on the type of toxicity elicited. All animals used in toxicity testing are routinely killed immediately at the end of experiments for examination (see paragraphs 3.47–3.49).
9.39 The fact that animals can suffer stress during toxicity testing has been investigated in studies in rats by assessing stress and discomfort from clinical and pathological observations.23 A substantial proportion of the animals suffered from serious discomfort, with some having obvious clinical signs, such as impaired locomotion and anaemia. Most of these animals only displayed non-specific clinical signs and the development of humane endpoints was confounded. The difficulty of interpreting data where overt toxicity is induced can be exacerbated by the fact that dosing of very high levels of test material might be required, with accompanying adverse welfare consequences for animals, including death. Death as an endpoint in toxicity testing, particularly when caused by the above conditions (the administration of ‘heroic’ doses), can be a misleading indication of hazard, since it might well not reflect any direct biological effects of the test material. Rather, death in such circumstances can be due to indirect effects such as dehydration leading to a heart attack. Similar effects can be caused by starvation which might occur when food becomes unpalatable during dietary administration of the test substance.
9.40 It has also been stressed that the design of toxicity experiments should be related to the way in which the resulting experimental data are going to be used.24 Thus, if it is intended to label a substance as hazardous on the basis of adverse reactions detected in one or a few animals, there is little point in subjecting additional animals to treatment and potential toxicity. The use of pilot studies in which the unknown effects of a treatment can be assessed in a few animals prior to conducting a full-scale experiment are also desirable in order to reduce numbers of animals used, and the potential suffering. This approach is, unfortunately, not routinely practised by toxicologists.
9.41 It is important that those who care for and subject animals to toxicity testing should become aware of the behavioural, emotional and physiological conditions and requirements of the animals (see paragraph 4.18). The ability of animals to anticipate negative events such as experimental procedures can increase anxiety levels and alter hormonal production which might also compromise the scientific quality of the data.
9.42 Several factors are expected to increase the numbers of animals being used in toxicity testing, as well as the severity of testing, including:
- the High Production Volume chemicals testing programme in the USA;25
- the new Registration, Evaluation and Authorisation of Chemicals (REACH) legislation in the EU26 (see Box 9.2);
- pesticide regulations in the EU that require more-extensive testing;
- the development and attempted validation of several animal tests to screen chemicals for endocrine (hormone)-disrupting activity;27 and
- the very substantial increase in the generation and utilisation of novel GM animal strains in toxicity studies.28
9.43 Finally, it must be acknowledged that toxicity tests in laboratory animals have limitations as a means of identifying hazards for human health, and managing risks to human health (see also Box 9.3). The example given in Box 9.4 also shows that different species may respond differently to the same compound. It has been argued that such problems fundamentally undermine the scientific and ethical justification for using animals to assess chemical safety. We have considered these questions briefly in paragraphs 8.39–8.41 and return to issues raised by the scientific validity of using animals in Chapter 10.
Footnotes11 Roe FJC (1993) Influence of animal species, strain, age, hormonal, and nutritional status, in Experimental Toxicology, The
Basic Issues, 2nd Edition, Anderson D and Conning D (Editors) (Cambridge: The Royal Society of Chemistry), pp23–34.
12 Morris T, Goulet S and Morton D (2002) The international symposium on regulatory testing and animal welfare:
recommendations on best scientific practices for animal care in regulatory toxicology ILAR J 43, Supplement: S123–5;
Hawkins P, Morton DB, Bevan R et al. (2004) Husbandry requirement for rats, mice, dogs and non-human primates used in telemetry procedures Seventh Report of the BVAAWF/FRAME/RSPCA/UFAW Joint Working Group on Refinement, Part B. Lab
Anim 38: 1–10.
13 Stephens ML, Conlee K, Alvino G and Rowan AN (2002) Possibilities for refinement and reduction: future improvements within regulatory testing ILAR J 43, Supplement: S74–9.
14 Brown AP and Levine BS (1999) Relationship Between Dosing Vehicles, Dose Volume, and Stress. Report prepared for the US
National Toxicology Program Unpublished report.
15 Brown & Levine (1999) ibid.
16 Morton DB, Jennings M, Buckwell A et al. (2001) Refining procedures for the administration of substances Lab Anim 35: 1–41.
17 See, for example: Brown AP, Dinger N, Levine BS (2000) Stress produced by gavage administration in the rat Contemp Top
Lab Anim Sci 39:17-21.
18 Gangolli SD and Phillips JC (1993) The metabolism and disposition of xenobiotics, in Experimental Toxicology, The Basic
Issues, 2nd edn, Anderson D and Conning D (Editors) (Cambridge: The Royal Society of Chemistry), pp130–201.
19 Koeter HBWM (1999) The OECD Test Guidelines Programme and animal welfare concern: how to avoid major animal suffering, in Humane Endpoints in Animal Experiments for Biomedical Research, Hendriksen CFM and Morton DB (Editors)
(London: Royal Society of Medicine Press), pp13–14.
20 Combes RD, Gaunt I and Balls M (2004) A scientific and animal welfare assessment of the OECD health effects test guidelines
for the safety testing of chemicals under the European Union REACH system. ATLA 32: 163-208.
21 These effects are ‘atypical’ in the sense that they arise very rapidly, usually within minutes of administration of the toxin (in most other cases effects more commonly occur within a timespan of several hours).
22 Combes RD (2003) The mouse bioassay for diarrhetic shellfish poisoning: a gross misuse of laboratory animals and of scientific methodology Alternat Lab Anim 31: 595–610.
23 Van Vlissingen JMF, Kuijpers MHM, van Oostrrum ECM et al. (1999) Retrospective evaluation of clinical signs, pathology and related discomfort in chronic studies, in Humane Endpoints in Animal Experiments for Biomedical Research, Hendriksen CFM and Morton DB (Editors), (London: Royal Society of Medicine Press), pp89–94.
24 Morton DB (2002) The importance of non-statistical experimental design in refining animal experiments for scientists,
IACUCs, and other ethical review panels, in Applied Ethics in Animal Research: Philosophy, regulation, and laboratory
applications, Gluck JP, DiPasquale A and Orlans FB (Editors) (West Lafayette, IN: Purdue University Press), pp149–78.
25 Nicholson A, Sandler J and Siedle T (2004) An evaluation of the US High Production Volume (HPV) chemical-testing programme. A study in (ir)relevance, redundancy and retro thinking ATLA 32 Supplement 1: 335–41.
26 Combes R, Dandrea J and Balls M (2003) A critical assessment of the European Commission’s proposals for the risk assessment and registration of chemical substances in the European Union ATLA 31: 353–64.
27 Combes RD and Balls M (2003) How much flexibility is possible when validating new in vivo and in vitro toxicity test methods? Alternat Lab Anim Exp 31: 225–32.
28 van Zeller A-M and Combes RD (1999) Transgenic mouse bioassays for carcinogenicity testing: a step in the right direction?
Alternat Lab Anim Exp 27, Supplement 1: 839–46.