CCAC guidelines on: choosing an appropriate endpoint in experiments using animals for research, teaching and testing

9. GUIDELINES FOR SELECTING APPROPRIATE ENDPOINTS IN SPECIFIC AREAS OF BIOMEDICAL RESEARCH AND TESTING
   
   For some specific areas of biomedical research and testing, more detailed guidelines for selecting an appropriate endpoint are provided in this section. Endpoint guidelines for animals used in monoclonal antibody production, cancer research, toxicology, infectious disease studies, and pain research are included. However, these are not the only areas where specific guidelines can be developed using the expertise of the attending laboratory animal veterinarian and the oversight of the ACC.



1. Monoclonal Antibody Production in Rodents
   
   Guideline: That as long as rodents continue to be used for monoclonal antibody production, the following endpoints be established:



• the increase in body weight due to the accumulation of ascites fluid in the abdomen and/or tumor growth should not produce pain and/or distress to the animal;


• depending on the condition of the mouse, a maximum of two taps of the ascites fluid are allowed, with the second tap being a terminal procedure. Ascites fluid taps should be done under general anesthesia.

It is widely recognized that the production of monoclonal antibodies (mAb) in rodents raises important concerns regarding the potential for severe pain and/or distress in the animals. In view of the rapidly developing and widely available in vitro methodology to produce monoclonal antibodies, the continued use of animals for mAb production is increasingly difficult to justify. As a result several countries have taken the position that alternative in vitro methods should be used for monoclonal antibody production, and that the mouse ascites method could be justified (to an ACC) only under special circumstances (e.g., when the in vitro method had failed to produce the required antibodies). Some countries recently acted to ban the mouse ascites method of mAb production (Shalev, 1998). The CCAC supports the use of in vitro methods for mAb production wherever possible.

At several stages of mAb production; from intraperitoneal injection of the priming agent, to the accumulation of ascites tumor fluid, and to collection of the ascites fluid, the animals may experience pain and/or distress. Therefore, when monoclonal antibodies are produced in vivo, clearly defined limits and endpoints placed on the application of the various manipulations, and close monitoring of the condition of the animals are required to minimize the potential for distress.

Following injection of hybridoma cells, routine care should include daily observations by appropriately trained staff for the first week (approximately), and before ascites fluid accumulation is evident (as indicated by the swelling of the abdomen). Any observations of unusual behavior or symptoms during this time should be addressed in a timely fashion. Pertinent signs of distress include: decrease in activity; hunched appearance; ruffled hair coat; respiratory distress; weight loss (which may be masked by the accumulating fluid in the abdomen). Once the ascites fluid accumulation has resulted in obvious abdominal swelling, the condition of the animal must be assessed at least twice every 24 hours at regularly spaced intervals.

The production of mAb in rodents provides an excellent example of how both replacement of mice with in vitro methods, and refinement of procedures with establishment of endpoints can reduce the distress experienced by the animals. For a review and bibliography of antibody production and alternative methods, see Smith, et al., 1997. Refinements and endpoints for mAb in rodents are discussed by Gillette, 1987; Marx, et al., 1997; McGuill & Rowan, 1989; Mueller, et al., 1986; and Workman, et al., 1998.


2. Cancer Research
   
   Guideline: For all cancer research in animal models, endpoints should be established that minimize the potential for pain and/or distress in the animals.
   
   Some recommended endpoints are:
   
   
   • the tumor mass should not proceed to the point where it significantly interferes with normal bodily functions, or causes pain or distress due to its location (solid tumors);
   
   
   • weight loss exceeding 20% of the body weight of a similar normal animal (taking into account the tumor mass);
   
   
   • ulceration/infection of the tumor site;
   
   
   • invasion of surrounding tissues by a localized tumor;
   
   
   • persistent self-induced trauma.
   
   Tumor burden should not exceed 5% of the animal's normal body weight for routine tumor passage or 10% for animals involved in therapeutic experiments (10% typically represents a subcutaneous flank tumor diameter of 17mm in a 25g mouse or 35mm in a 250g rat). Calibration curves should be established as part of the characterization of any new tumor system (Workman, et al., 1998)
   
   One of the scientific concerns that has been expressed about arbitrary assignment of an endpoint in cancer therapy studies is that early euthanasia may alter longevity or survival data which are important indicators of "successful" treatment. For example, the "successful" treatment of cancer in rats which resulted in them living a month longer, might be masked by early euthanasia based only on clinical observations. Workman, et al. (1998) recommend that the required information on response to therapy be obtained by tumor regrowth delay, clonogenic assay following tumor excision or an appropriate surrogate endpoint. In such cases finding the signs of disease and distress that point to an irreversible deterioration in the animal, is important.
   
   Redgate, et al. (1991) in their examination of a brain tumor model in a rat (9L gliosarcoma in Fischer 344 rats), concluded that a weight loss period of more than six days had a high correlation with irreversible progression to death. In this model then, an endpoint that satisfied the scientific concerns could be established at the end of a six day period of consecutive weight loss, which in this case was about ten days before death of the animals.
   
   A thorough discussion of the many welfare factors that must be considered when cancer research is undertaken is contained in the publication of the United Kingdom Coordinating Committee on Cancer Research (UKCCCR) (Workman, et al., 1998), providing valuable information for investigators, and ACC. Additional criteria for selecting the endpoint in cancer research and in toxicology have been proposed by Montgomery (1987, 1990), Redgate, et al. (1991), and Tomasovic, et al. (1988).
   
   
3. Toxicological Studies and Toxicity Testing
   
   
   1. Acute toxicity testing
      
      Guideline: Before a protocol that includes safety/efficacy/toxicity testing with death as an endpoint for regulatory purposes can be accepted by the institution's animal care committee, there must be clear, written documentation obtained by the investigator from the appropriate regulatory agency that the proposed test is a necessary part of the submission for licensing/approval. The investigator must also demonstrate to the animal care committee that an alternative in vitro test will not be acceptable to the regulatory agency, and that this testing has not been previously done elsewhere.
      
      For most toxicological studies, the investigator is interested in the primary or secondary interaction(s) of the compound with the body's cells and tissues, and not the tertiary effects (such as dehydration, anorexia, etc.) which may cause the death of the animal. Thus, from a scientific viewpoint, it is important to collect as much data as possible during the early stages of the clinical effects. Any data collected from a moribund animal, or any unobserved animal death, represents lost information on the compound studied (Toth, 1997).
      
      Guidelines on Acceptable Testing Standards: Toxicity tests should be done according to the guidelines of the Organization for Economic Cooperation and Development (OECD), International Conference on Harmonization of Technical Requirements for Registration of Pharmaceuticals for Human Use (ICH), Health Canada (HC), or US Food and Drug Administration (FDA), using the minimum number of animals possible, and with all possible consideration for the relief of animal pain and/or distress.
      
      When initiating toxicological studies, there is usually little information available on the test compound. However, there may be data on the class of compounds, or compounds with similar chemical structures that can be used to guide the experimental design. Establishing doses and effects after single or repeated administration, toxicity and/or exaggerated pharmacological responses are the purpose of these studies.
      
      The use of pilot experiments in determining endpoints (Olfert, 1995; Browder, 1995; Everitt & Griffin, 1995) are of particular value when dealing with compounds of unknown effects. Using a small number of animals and beginning with low-end doses, the morbidity, the time course of effects and the frequency of observations required to set earlier endpoints could be determined. Data collected in pilot studies that support earlier endpoints in specific toxicity studies, if provided to the regulatory agencies, could also assist the regulatory agencies in validating and approving earlier endpoints. As noted previously, conducting a pilot experiment also provides the opportunity for all persons to become experienced with the expected particular signs and symptoms.
      
      Dose range finding studies are widely used in pharmaceutical toxicity research programs. Doses are selected based on pharmacology, pharmacokinetic characteristics, and previous experience with the class of drug under investigation. These studies also give an indication of potential target organs, and unexpected adverse effects that may be addressed in subsequent definitive studies, and significantly reduce the numbers of animals required.
      
      The pharmaceutical industry initiative in the form of the International Conference on Harmonization of Technical Requirements for Registration of Pharmaceuticals for Human Use (ICH) was organized to make recommendations on ways to achieve greater harmonization in the interpretation and application of technical guidelines and requirements for product registration in order to reduce or obviate the need to duplicate the testing carried out during the research and development of new medicines. There was agreement after the first ICH meeting (1991), and subsequent issuance of guidelines on single dose and repeat dose toxicity tests, that the LD₅₀ (Lethal Dose 50) be abandoned for use in pharmaceuticals. Although an assessment of acute toxicity is required, a calculated LD₅₀ is not (Federal Register, 61(166), Single Dose Acute Toxicity Testing for Pharmaceuticals; Revised Guidance).
      
      Similarly the OECD discourages the use of the classical LD50 test, and recommends alternatives that reduce animal numbers and use morbidity rather than mortality as the endpoint (OECD, 1987). The Interagency Research Animal Committee of the US (IRAC, 1993) Recommendations on LD₅₀ Testing state:
      
      
      1. "The Classical LD₅₀ test should only be conducted when specifically justified for reasons of scientific necessity and approved by the institutional animal care and use committee (IACUC).
         
      2. Toxicity testing procedures based on the principles of reduction and refinement (such as the Limit test) should be used until alternative test methods become validated."
         
         
      The choice of toxicity tests should attempt to meet the collective requirements of the many national regulatory agencies to minimize the need for repeat studies for a specific country. In each case, the client must check with the regulatory agency(ies) to determine what testing is required.
      
      Recalculation of previously published toxicological data should also be considered, to ensure that ED₅₀ (Effective Dose 50) or LD₅₀ values are accurate (Irvine, et al., 1992).
      
      Acute toxicity studies in fish. Acute toxicity tests with fin fish have been primarily for environmental protection from industrial effluents and the risk assessment of new pest control agents for agriculture and forestry. The concern has been primarily one of safety and there has been little interest in the pathology associated with the toxicity.
      
      If no information is available on the potential toxicity to fish of the material to be tested, a range finding study should be conducted with a minimal number of fish. On the appearance of peracute clinical signs of irritation or toxicity, the fish should be immediately removed to fresh water. If there is no substantive reduction in the clinical signs within five minutes of transfer to fresh water, the fish should be euthanized.
      
      Only when a concentration of the test materials which does not cause peracute clinical signs has been identified should a more substantial number of fish be exposed to this and lower concentrations to determine the acute toxicity. The minimum number of dilutions, minimum replications and minimum number of animals per dilution consistent with the accuracy and confidence interval required should be used.
      
      In situations involving the monitoring of an effluent far exceeding required safety limits, the tests should be limited to a pass-fail approach rather than quantitating the dilution which produces a specified endpoint (e.g., LC₅₀ - Lethal Concentration 50).
      
      During initiation of exposure, fish should be observed continuously for the first 45 to 60 minutes. Subsequent observations should be made on a geometric progression basis (e.g., 1.5h, 3h, 6h and 12h) and then a minimum of twice daily for the duration of the study. Frequency of observation should increase with the appearance of any clinical signs. On the appearance of severe clinical signs which are inconsistent with survival (e.g., gaping, forceful collisions with the side of the tank or aquarium, hyperactivity) the fish should be removed to fresh water. If the clinical signs persist without interruption for more than five minutes or are followed by loss of normal posture or position in the water column in the fresh water, the fish should be euthanized.
      
      Animal Care Committee oversight in acute toxicity studies with death as an endpoint. In addition to the points already raised, some additional principles that will assist the institutional ACC in fulfilling its obligations when death as an endpoint is part of a protocol have been delineated by Hamm (1995). As noted above, written documentation justifying the need for death as the endpoint must be provided to the ACC prior to initiation of the study. The written documentation should include assurances that alternate earlier endpoints were considered, and the reasons why analgesia cannot be provided. The frequency of monitoring by trained personnel, identification of signs of distress or illness, and treatment of animals found with abnormalities, should be provided to the ACC and be part of the conditions for experiment. Complete documentation of all observations and actions taken should be retained and be available to the ACC and attending veterinarian (Hamm, 1995).
      
      
   2. Chronic toxicity studies and studies in aging
      
      Guideline: Before a protocol that requires holding animals to an age close to or beyond the median survival age specific to the species or strain (e.g., chronic toxicity studies, carcinogenicity testing, or aging studies) is approved by the institution's animal care committee, the investigator in collaboration with the veterinary staff must establish the endpoint criteria for euthanasia of the animals, the persons responsible for monitoring the animals' condition, and the authority of the persons who will make the decision to euthanize.
      
      Some of the most controversial areas concerning endpoints are in chronic toxicity or aging studies, especially the carcinogenicity studies required for New Drug Applications (NDA). In these studies, animals may receive the test compound for up to two years by the route(s) of administration to be used by humans. Most often this is oral (gavage or mixed in the diet), but other routes may be used. The conflict occurs when investigators are concerned with survival rates. To be considered a negative study, the US Food and Drug Administration (FDA) and Environmental Protection Agency (EPA) require that survival be at least 50%. The investigator in collaboration with the veterinary staff can establish a criteria for euthanasia of animals. This should be defined prior to the start of the study, and addressed on a case-by-case basis.
      
      Investigators should familiarize themselves with the current ICH recommendations for specific testing requirements prior to drafting their research/testing protocols.
      
      The housing and care of aging animals requires special considerations for their comfort and well-being. Rats over 18 months of age, for example, tend to be large and develop weight related problems such as foot pad lesions, obesity, and tumors. The use of solid bottom cages, along with appropriate bedding material will improve the comfort level for the animals. Aged animals may also require special attention regarding the way food and water are presented to them.
      
      
4. Pain Research
   
   Because it is an inherent aspect of studies of pain in humans and animals that some pain must be produced, such experiments raise special ethical concerns. This difficulty was recognized early, as the study of pain became a more distinct discipline. The following guiding principles have been extracted from the literature cited in the accompanying discussion in this section.
   
   Guidelines:
   
   
   • the animals should be exposed to the minimal pain necessary for the purposes of the experiment;
     
   • the duration of the pain must be as short as possible and the number of animals involved kept to a minimum;
     
   • threshold levels of pain stimuli rather that supra-threshold levels should be used whenever possible;
     
   • if models of acute pain, or acute pain tests are being used, where the pain is not terminated by the animal's reaction, but may extend beyond the time necessary to obtain results, the pain should be terminated as quickly as possible;
     
   • tests other than avoidance tests are strongly discouraged;
     
   • animal models experiencing chronic pain should be provided with adequate analgesia at all times. Exceptions to this should be restricted to those times justified to the institutional animal care committee by the investigator with evidence that the analgesics will interfere with the aims of the investigation.
   
   
   The International Association for the Study of Pain (IASP) Ethical Guidelines for Investigations of Experimental Pain in Conscious Animals (Zimmermann, 1983) contain seven points relating to the justification of the experiment, and to the potential severity and duration of pain produced in the experimental animal. These seven points are:
   
   
   1. "It is essential that the intended experiments on pain in conscious animals be reviewed beforehand by scientists and lay-persons. The potential benefit of such experiments to our understanding of pain mechanisms and pain therapy needs to be shown. The investigator should be aware of the ethical need for a continuing justification of his investigations.
      
   2. If possible, the investigator should try the pain stimulus on himself; this principle applies for most non-invasive stimuli causing pain.
      
   3. To make possible the evaluation of the levels of pain, the investigator should give a careful assessment of the animal's deviation from normal behavior. To this end, physiological and behavioral parameters should be measured. The outcome of this assessment should be included in the manuscript.
      
   4. In studies of acute or chronic pain in animals, measures should be taken to provide a reasonable assurance that the animal is exposed to the minimal pain necessary for the purposes of the experiment.
      
   5. An animal presumably experiencing chronic pain should be treated for relief of pain, or should be allowed to self-administer analgesic agents or procedures, as long as this will not interfere with the aims of the investigation.
      
   6. Studies of pain in animals paralyzed with a neuromuscular blocking agent should not be performed without a general anesthetic or an appropriate surgical procedure that eliminates sensory awareness.
      
   7. The duration of the experiment must be as short as possible and the number of animals involved kept to a minimum."
      
      
   For animal models of acute pain, careful design and conduct of the experiment will mean that results are obtained using the fewest numbers of animals necessary. An appropriate endpoint needs to be defined. The type of pain test proposed will influence the selection of the endpoint.
   
   As noted in the IASP statement, it is incumbent on the investigator to be thoroughly familiar with the behavior of the animal species being used, and to document behavioral and physiological changes being observed in animals subjected to stimuli causing pain. This is also important so that unexpected severe levels of pain are recognized if they occur. Some of the tests for acute pain have been extensively used on humans, and those experiences should guide investigators regarding the nature of the pain produced.
   
   Many pain tests use a threshold pain level, one that triggers avoidance or withdrawal behavior in the animal. If the stimulus intensity is above the threshold needed to elicit a response, special care must be taken in its application. Threshold levels of pain stimuli rather than supra-threshold levels should be used whenever possible. Franklin & Abbott (1989) indicate that in many instances supra-threshold pain stimuli are unnecessary.
   
   In some pain tests the animals can terminate the painful stimulus by their action. For example, in the tail flick test, and the hot plate test, the animal's first response is taken as the endpoint, and the animal is removed from the stimulus. In other tests (e.g., the acute tissue injury-inducing tests, which include the writhing test and the formalin test), the painful stimulus presumably goes on longer than the test period, and the animals cannot escape from it. Using avoidance tests would be preferable to using tests where the pain continues after the results are obtained. If models of acute pain, or acute pain tests are being used where the pain is not terminated by the animal's reaction, then the pain should be terminated as quickly as possible. This may require humanely killing the animals as soon as the test is completed (e.g., the writhing test), or administering analgesic drugs.
   
   Animal models of chronic pain are of particular concern, since the cost to animals where chronic pain is induced (adjuvant-induced arthritis in rats, for example) can be very high. It is well-known that such disease states in humans are often accompanied with severe, unremitting pain and distress. Unless particular attention is paid to the animal models used in such a study, pain and/or distress could easily extend beyond that necessary for the purposes of the research. Franklin & Abbott (1989) suggest that pain produced in adjuvant-induced arthritis in rats approaches an intensity such that, based on the rating proposed by Morton & Griffiths (1985), relief (analgesics) should be given. The principal investigator, the institutional veterinarian, and the ACC should ensure that such animals are afforded every consideration for limiting or easing their discomfort and pain.
   
   All animals with chronic pain should receive special attention not only to relieve their pain, but also with respect to their husbandry and housing. A great deal can be done, beyond the routine care given to normal laboratory animals, to make these special animal models comfortable (e.g., gentle handling, improved access to food and water, housing in solid bottom cages with deep, soft bedding). The expertise of the laboratory animal veterinarian and the animal health technicians should be consulted, and should be integral to the discussions of these aspects by the ACC.
   
   
5. Infectious Disease Studies, Vaccine Trials, etc.
   
   Guideline: For all infectious disease research, including virulence tests in animal models, endpoints should be established that minimize the potential for pain and/or distress in the animals.
   
   Some studies in infectious disease (e.g., tests to establish the virulence of an infectious organism) are still being conducted with mortality as the proposed endpoint (also referred to as the Rodent Protection Test). The use of PD₅₀ (Protective Dose 50) tests in mice may be required when anti-infective studies are done.
   
   Soothill, et al. (1992) found that in mice infected with bacteria, hypothermia of 34^oC was predictive of mortality. Siems & Allen (1989) recommended that the endpoint in a disease model (chronic infection with systemic Candida albicans) be set at (among other measurements) the point when the animals lose more than 20% of body weight, or when the body temperature drops more than 4^oC; both of which are easily monitored. The magnitude of these changes from normal would also give a maximum score on the Morton & Griffiths (1985) proposal, indicating severe negative effects on the animal.
   
   The UK Rodent Protection Test Working Party considered the following as some of the general signs exhibited by rodents with systemic infections; ruffled or "spikey" fur, weight loss, ocular discharge, lethargy, hunched back, ataxia, tremor, hypothermia, cyanosis (Acred, et al., 1994). Recommended clinical scoring systems useful for certain bacterial Rodent Protection Tests (RPT), for the Candida albicans RPT, and for the Herpes simplex RPT were provided by the RPT Working Group (Acred, et al., 1994).
   
   The use of observational checklists with sufficiently frequent observation times (see part 5 Determining the Required Frequency of Animal Observations) will assist the investigator in detecting the point at which there are signs of progressive deterioration of the animal's condition leading to death if no action is taken (Acred, et al., 1994).
   
   
6. Specific Animal Models with the Potential for Significant Levels of Pain and/or Distress
   
   Guideline: Any pain and/or distress, or deficits in function that negatively affect the animal's well-being, not scientifically "necessary" for the study, should be alleviated or minimized. Cost or convenience should not deter from this. Further, as soon as the study is complete, the pain and/or distress should be alleviated through treatment or euthanasia.
   
   Some animal models have inherent or induced functional deficiencies with the potential for pain and/or distress. There is a responsibility to attend to the special needs of these animals, beyond the care provided to normal animals. Animal models with inherited deficits (distressed phenotypes), transgenic animals, ischemia or seizure models, stereotaxic manipulations, are some examples of the kinds of laboratory animals for whom this recommendation is made.
   
10. SUMMARY
    
    All those involved in the use of animals for research, teaching and testing; ACC members, investigators and their research staff, the laboratory animal veterinarian and the animal health technicians, have responsibilities for the humane care and use of experimental animals. In working with investigators to establish appropriate endpoints, the ACC should ensure that the appropriate criteria are used by the principal investigator to determine the endpoint. Through the use of observational checklists and animal condition scoring systems, the most objective, and more humane, endpoints can be identified. Responsibilities for observing and monitoring the animal's condition must be clearly delineated. Persons involved in establishing and validating endpoints in invasive experiments are encouraged to present and publish these data, to support efforts at continually refining the animal use practices that occur in biomedical research.