# Humane Endpoints for Non-Rodent Laboratory Species ## Key Takeaways - Humane endpoints for non-rodent species (rabbits, dogs, non-human primates) require species-specific physiological baselines and clinical sign interpretation due to their larger size, longer lifespans, and complex behavioral repertoires, necessitating a departure from rodent-derived criteria. - Objective and reproducible endpoint determination relies on integrating multiple assessment modalities, including body weight trends (not just absolute loss), body condition scoring, behavioral indicators (e.g., social withdrawal, facial expressions), and neurological status, rather than single parameters. - A tiered decision framework is crucial, distinguishing between parameters that trigger intensified monitoring (e.g., 10% body weight loss) versus those mandating immediate veterinary assessment (e.g., 15% body weight loss with reduced fecal output) or euthanasia (e.g., 20% body weight loss or neurological signs). - Documentation of endpoint decisions, including the rationale, observer, date, time, and specific criteria met, is essential for protocol review, refinement, and demonstrating compliance with animal welfare regulations. - Recognizing species-specific vulnerabilities, such as rabbits' tendency to mask pain and rapid decompensation, dogs' potential for stoic behavior, and non-human primates' complex social dynamics, is critical for accurate and timely intervention. - The evidence base for non-rodent humane endpoints is limited, with much guidance derived from expert opinion and institutional experience rather than species-specific validation studies, highlighting the need for careful adaptation and prospective evaluation. --- Humane endpoints are the predetermined criteria at which an experimental animal is removed from a study, typically by euthanasia, to prevent or terminate avoidable pain, distress, or suffering. For non-rodent laboratory species, including rabbits, dogs, and non-human primates, endpoint determination carries additional complexity because of their larger body size, longer lifespans, more pronounced behavioral repertoires, and the higher degree of individualisation in their clinical assessment. This article provides veterinary researchers with a framework for designing, implementing, and refining humane endpoints in studies using these species. It addresses species-specific physiological baselines, clinical sign interpretation, pain and distress scoring, and the integration of endpoint criteria into study protocols and institutional oversight. The clinical question this article answers is practical: how does a veterinarian distinguish a reversible experimental effect from an irreversible decline that mandates intervention, and how are those criteria made objective, reproducible, and defensible to an institutional animal care and use committee? The answer draws on established welfare assessment principles, published refinement guidance, and the physiological and behavioral characteriztics that distinguish rabbits, dogs, and non-human primates from the rodent models more extensively covered elsewhere in the literature. Where the evidence base is thin, particularly for species-specific validated scoring systems, this article identifies the gaps explicitly. ## At a Glance | Parameter | Consideration | |---|---| | Baseline data | Collect individual pre-study body weight, temperature, food and water intake, and behavior for each animal | | Clinical scoring | Use species-appropriate, validated scoring tools where available, adapt rodent-derived systems only with explicit validation | | Body weight thresholds | Reference institutional and published standards, absolute percentage loss is insufficient alone for large species | | Pain assessment | Integrate behavioral, physiological, and facial expression indicators, species differ markedly in pain display | | Neurological status | For primates and dogs, mentation and motor function changes may precede systemic signs | | Social behavior | Withdrawal from group interaction is an early, sensitive indicator in social species | | Intervention triggers | Define two-tier criteria: monitoring intensification versus immediate euthanasia | | Documentation | Record endpoint decisions, rationale, and outcomes for protocol review and refinement | ## Conceptual Foundations of Endpoint Design ### The Purpose of a Humane Endpoint A humane endpoint serves two functions simultaneously. It protects the individual animal from experiencing a level of suffering that the study does not justify, and it preserves scientific validity by ensuring that data are collected before physiological decompensation confounds the measurements. The [National Research Council's Guide for the Care and Use of Laboratory Animals](https://grants.nih.gov/grants/olaw/guide-for-the-care-and-use-of-laboratory-animals.pdf) establishes that the veterinarian, in collaboration with the principal investigator, is responsible for defining these endpoints prospectively and for ensuring that all personnel can recognize them reliably. Endpoints must be specific enough to be applied consistently by different observers and flexible enough to accommodate individual variation. ### The Severity Spectrum and Cumulative Burden Non-rodent species used in research typically experience a cumulative burden that extends beyond the experimental procedure itself. Housing conditions, handling, restraint, sampling procedures, and the experimental manipulation each contribute to overall welfare compromise. The [NC3Rs resources on replacement, reduction and refinement](https://www.nc3rs.org.uk/) emphasize that refinement applies across the entire study lifecycle, also to the terminal phase. For rabbits, dogs, and non-human primates, this means that endpoint criteria must account for the interaction between experimental effects and husbandry-related stressors. A rabbit with mild experimental disease may cross an acceptable welfare threshold more quickly if it is also experiencing transport stress or social isolation. ### Species-Specific Baseline Physiology Endpoint interpretation depends on knowing what is normal for the species, the strain or breed, and the individual. Rabbits are prey species that mask signs of pain and illness until compromise is advanced. Their normal body temperature ranges higher than that of dogs or primates, and their gastrointestinal physiology makes anorexia a more rapidly dangerous sign because of the risk of ileus and gastric stasis. Dogs, as social carnivores, display pain through changes in posture, vocalisation, and interaction with familiar humans. Non-human primates show a spectrum of responses that varies with species, social rank, and rearing history, a dominant male macaque may display pain differently from a subordinate female, and both may differ from a singly housed animal. ## Physiological and Behavioral Indicators Across Species ### Body Weight and Body Condition Body weight is the most universally applied endpoint criterion, but its interpretation differs across non-rodent species. In rabbits, body weight can fluctuate with gut fill and hydration status, so a single measurement is less informative than a trend over consecutive days. The [body weight algorithm approach developed for rodent glioma models](https://pubmed.ncbi.nlm.nih.gov/32488031/) demonstrates that mathematical analysis of weight trajectories can predict endpoints earlier than absolute thresholds, and this principle transfers conceptually to larger species, though the algorithm itself has not been validated outside rodents. For dogs and non-human primates, body condition scoring provides a more robust assessment than weight alone because it is less affected by hydration and gut fill. The MSD Veterinary Manual provides species-specific body condition scoring systems that can be adapted for research settings. ### Pain and Distress Recognition Pain assessment in non-rodent species requires integration of multiple modalities. Rabbits show pain through tooth grinding, hunched posture, reduced grooming, and changes in fecal output. Dogs exhibit lameness, guarding, restlessness, and altered interaction with handlers. Non-human primates may show facial expression changes, particularly around the eyes and mouth, that are analogous to the grimace scales validated in rodents. The [consensus guidance on surrogate endpoints in acute lung injury models](https://pubmed.ncbi.nlm.nih.gov/33332817/) underscores that no single behavioral sign is reliable across animals or models, and that composite scoring systems outperform individual parameters. For all three species, the absence of obvious pain behavior does not confirm the absence of pain, particularly in prey species or in animals that have learned that pain display attracts handling. ### Neurological and Mentation Changes For dogs and non-human primates, changes in mentation and neurological function often precede systemic signs of decline. A dog that becomes withdrawn, fails to respond to familiar commands, or shows subtle proprioceptive deficits may be experiencing intracranial pathology or systemic metabolic disturbance. Non-human primates may show reduced responsiveness to environmental stimuli, changes in grooming behavior, or altered dominance interactions before any measurable change in body weight or temperature. These indicators require familiarity with each individual's baseline behavior, which argues for assigning a single primary observer to each animal where staffing permits. ## Endpoint Criteria by Species ### Rabbits Rabbits present particular challenges for endpoint assessment because they are prey species that mask signs of pain and illness. A rabbit that is severely compromised may continue to eat and appear alert until decompensation is rapid and profound. The assessment sequence should therefore prioritize fecal output, food intake, and gastrointestinal motility over more conspicuous behavioral signs. Daily monitoring should include fecal pellet count and consistency, since reduced or absent fecal output is often the earliest reliable indicator of ileus, pain, or systemic illness. Abdominal palpation for gastric dilatation and auscultation for borborygmi should be performed when gastrointestinal disease is suspected. Body weight should be measured at least every 48 hours in studies where weight loss is an expected sequela, and more frequently during acute phases. Dental disease, pododermatitis, and urine scald are common husbandry-related conditions that can confound endpoint assessment. These conditions should be scored separately from experimental disease signs, and their presence should trigger a review of housing and nutrition instead of automatic euthanasia. However, if a rabbit develops anorexia with weight loss exceeding 15% of baseline body weight despite supportive care, euthanasia is generally indicated regardless of the underlying cause. Respiratory rate and effort are critical parameters in rabbit models of pneumonia, pleuritis, or pulmonary fibrosis. Nasal discharge, ocular discharge, and piloerection are useful adjunctive signs. Neurological signs such as head tilt, nystagmus, or paresis should prompt immediate veterinary assessment, as these may indicate encephalitozoonosis, otitis interna, or experimental disease progression. ### Dogs Dogs used in research are typically purpose-bred animals with known health histories, which facilitates accurate baseline assessment. Their social nature and trainability allow for more nuanced behavioral evaluation than is possible in most other laboratory species. However, this same sociality can mask pain, as dogs may continue to interact with familiar handlers despite significant discomfort. The clinical score should incorporate appetite, activity level, interaction with handlers, and response to being approached. A dog that withdraws from social contact, refuses favourite foods, or becomes aggressive when handled is demonstrating significant distress. Gait assessment is particularly valuable in orthopedic, neurological, and oncology models. Lameness should be scored using a defined scale that distinguishes weight-bearing from non-weight-bearing lameness. Cardiovascular parameters including heart rate, pulse quality, mucous membrane color, and capillary refill time should be assessed in models where hemodynamic compromise is anticipated. Respiratory rate and effort are similarly important in pulmonary and cardiac models. Body temperature should be measured when sepsis, inflammation, or thermoregulatory dysfunction is suspected. Dogs are uniquely positioned among laboratory species for telemetric monitoring of physiological parameters. Implanted transmitters can provide continuous data on heart rate, blood pressure, body temperature, and activity without repeated handling. This technology is particularly valuable for detecting deterioration that occurs between scheduled observations, but it does not replace direct clinical assessment. ### Non-Human Primates Non-human primates require the most individualised approach to endpoint assessment of any laboratory species. Their cognitive complexity, social structure, and behavioral repertoire demand that endpoints be tailored to the individual animal, the specific study, and the species involved. A cynomolgus macaque may display different signs of distress than a rhesus macaque or a marmoset, and group-housed animals may behave differently when observed alone versus with their social group. Behavioral assessment should include posture, facial expression, vocalisation, self-directed behavior, and social interaction. A non-human primate that isolates itself from group members, refuses to engage in normal social grooming, or displays stereotypic behavior is experiencing significant distress. Facial expression scoring, including orbital tightening and ear position, has been validated in macaques and should be incorporated into routine assessment. Food intake and body weight are reliable indicators of health status, but they must be interpreted in the context of social hierarchy. A subordinate animal may have reduced food access even when healthy, while a dominant animal may maintain food intake despite significant illness. Body condition scoring should be performed using a species-appropriate scale that accounts for muscle mass over the scapulae, spine, and pelvis. Non-human primates are particularly susceptible to stress-related gastrointestinal disturbances, including diarrhea and gastric ulceration. Stool consistency and frequency should be monitored daily. Vomiting, which is uncommon in healthy non-human primates, should be treated as a significant clinical sign requiring immediate assessment. ## Monitoring Frequency and Documentation The frequency of monitoring should be proportional to the expected rate of deterioration. Studies involving rapidly progressive disease, surgical intervention, or pharmacological manipulation may require monitoring every 4 to 6 hours during critical phases. Stable animals in long-term studies may be adequately monitored once daily, provided that clear criteria for escalating monitoring frequency are defined in advance. | Parameter | Rabbit | Dog | Non-Human Primate | |-----------|--------|-----|-------------------| | Body weight | Every 48 hours minimum, daily during acute phases | Daily during acute phases, 2 to 3 times weekly in chronic studies | Daily during acute phases, 2 to 3 times weekly in chronic studies | | Food intake | Daily, quantify if possible | Daily, note refusal of preferred foods | Daily, account for social hierarchy | | Fecal output | Daily, count pellets | Daily, note consistency and frequency | Daily, note consistency and frequency | | Behavioral score | Twice daily during acute phases | Twice daily during acute phases | Twice daily, include social interaction assessment | | Pain assessment | Twice daily using species-specific grimace scale | Twice daily using validated pain scale | Twice daily using facial expression and posture | | Respiratory rate | Daily, more frequent in pulmonary models | Daily, more frequent in pulmonary or cardiac models | Daily, more frequent in pulmonary models | | Temperature | When indicated by clinical signs | When indicated by clinical signs | When indicated by clinical signs | Documentation should include the date and time of each assessment, the observer's identity, the parameters measured, and the scores assigned. Any deviation from the expected clinical course should be recorded, along with the veterinary assessment and the rationale for any change in monitoring frequency or endpoint criteria. The [Guide for the Care and Use of Laboratory Animals](https://grants.nih.gov/grants/olaw/guide-for-the-care-and-use-of-laboratory-animals.pdf) emphasizes that veterinary staff must have the authority to implement humane endpoints without requiring investigator approval, and this authority should be documented in institutional protocols. ## Decision Frameworks and Escalation A tiered decision framework is more useful than a single threshold for euthanasia. The first tier comprises parameters that trigger increased monitoring frequency and veterinary assessment. The second tier comprises parameters that trigger immediate veterinary examination. The third tier comprises parameters that mandate euthanasia without delay. For example, in a rabbit study, a 10% body weight loss from baseline might trigger twice-daily monitoring and veterinary review. A 15% weight loss combined with reduced fecal output would trigger immediate veterinary examination. A 20% weight loss, or any weight loss combined with neurological signs or respiratory distress, would mandate euthanasia. The same framework applies across species, but the specific thresholds differ. Dogs may tolerate greater body weight loss than rabbits before decompensation, while non-human primates may deteriorate more rapidly once clinical signs appear. The [MSD Veterinary Manual](https://www.msdvetmanual.com/) provides species-specific reference ranges for physiological parameters, and these should be used to establish baseline values for individual animals. ## Equipment and Technical Considerations The equipment required for endpoint assessment varies with species and study design. Scales accurate to 0.1% of body weight are essential for all species. For rabbits and non-human primates, a sling or restraint device may be needed for safe weighing and examination. Dogs can typically be weighed on a platform scale. Telemetric monitoring systems are available for all three species and can provide continuous physiological data. These systems require surgical implantation, which itself carries welfare implications and should be justified in the protocol. The [NC3Rs](https://www.nc3rs.org.uk/) provides guidance on refinement of surgical procedures and postoperative care that applies equally to telemetric implantation and other surgical interventions. Imaging modalities, including radiography, ultrasonography, and computed tomography, may be used to assess disease progression and to inform endpoint decisions. These techniques require anesthesia or sedation in most cases, and the welfare impact of repeated imaging must be weighed against the information gained. In some studies, a single imaging time point may be sufficient to confirm disease progression and justify euthanasia. ## Species-Specific Endpoint Tables The following tables provide endpoint criteria for each species. These criteria should be adapted to the specific study, the expected disease course, and the individual animal. The [WOAH terrestrial animal health standards](https://www.woah.org/en/what-we-do/standards/codes-and-manuals/terrestrial-code-online-access/) provide general principles for animal welfare assessment that are applicable to research settings. ### Rabbit Endpoint Criteria | Parameter | Increased Monitoring | Veterinary Assessment Required | Euthanasia Indicated | |-----------|---------------------|-------------------------------|----------------------| | Body weight loss | 10% from baseline | 15% from baseline | 20% from baseline | | Fecal output | Reduced pellet count | No fecal output for 24 hours | No fecal output with abdominal distension | | Food intake | Reduced | No intake for 24 hours | No intake for 48 hours | | Respiratory rate | 20% above baseline | 40% above baseline | Respiratory distress with cyanosis | | Behavior | Mild lethargy | Marked lethargy, hunched posture | Unresponsive, recumbent | | Neurological signs | Mild head tilt | Progressive head tilt, ataxia | Seizures, paresis, coma | ### Dog Endpoint Criteria | Parameter | Increased Monitoring | Veterinary Assessment Required | Euthanasia Indicated | |-----------|---------------------|-------------------------------|----------------------| | Body weight loss | 10% from baseline | 15% from baseline | 20% from baseline | | Appetite | Reduced intake | Refuses food for 24 hours | Refuses food and water for 48 hours | | Gait | Mild lameness | Non-weight-bearing lameness | Recumbent, unable to stand | | Cardiovascular | Heart rate 20% above baseline | Heart rate 40% above baseline, weak pulse | Pale mucous membranes, prolonged capillary refill time | | Behavior | Withdrawn | Reluctant to interact | Unresponsive, vocalising, aggressive when handled | | Respiratory | Mild tachypnoea | Labored breathing | Respiratory distress with cyanosis | ### Non-Human Primate Endpoint Criteria | Parameter | Increased Monitoring | Veterinary Assessment Required | Euthanasia Indicated | |-----------|---------------------|-------------------------------|----------------------| | Body weight loss | 10% from baseline | 15% from baseline | 20% from baseline | | Food intake | Reduced | No intake for 24 hours | No intake for 48 hours | | Stool quality | Soft stool | Diarrhea for 24 hours | Hemorrhagic diarrhea, vomiting | | Behavior | Social withdrawal | Isolation from group, self-directed behavior | Unresponsive, self-injurious behavior | | Facial expression | Mild orbital tightening | Moderate orbital tightening, ear flattening | Severe grimace, eyes closed | | Neurological signs | Mild tremor | Ataxia, paresis | Seizures, coma | These criteria assume that supportive care, including analgesia, fluid therapy, and nutritional support, has been provided where appropriate. The [AVMA practice resources](https://www.avma.org/resources-tools) provide guidance on euthanasia methodology and confirmation of death that should be followed when endpoint criteria are met. ## Recognized Complications and Failure Modes Endpoint protocols fail in predictable ways. The most common failure is the delayed endpoint, where an animal crosses a threshold but is not euthanized because monitoring intervals are too long or because the threshold itself is ambiguous. In non-human primates, social housing complicates detection because subordinate animals may hide clinical signs, while dominant animals may show aggression that is mistaken for disease. In dogs, stoic behavior can mask pain until the animal is profoundly compromised. Rabbits present a distinct problem: they are prey species that suppress overt signs of illness, and their rapid decompensation once clinical signs appear leaves little margin for intervention. A second failure mode is the premature endpoint, where an animal is euthanized for a reversible condition that could have been managed. This occurs most often with body weight loss that reflects dehydration instead of disease progression, or with transient postoperative anorexia. The discriminating check is response to supportive care: an animal that stabilizes or improves within 24 to 48 hours of fluid therapy and analgesia was not at its true endpoint. A third failure mode is the unvalidated surrogate. A physiological or behavioral parameter that correlates poorly with the actual disease process will generate both false positives and false negatives. For example, telemetric heart rate variability may change in response to handling stress instead of disease progression, and a clinical score that weights a single sign too heavily will miss animals whose primary manifestation is different. The body weight course analysis algorithm developed for an intracranial rat glioma model predicted endpoints in 97% of animals without observer-dependent confounding, but the same algorithm cannot be assumed to transfer across species or tumor models without validation. | Observation | Likely cause | Discriminating check | |---|---|---| | Weight loss with normal appetite | Dehydration, measurement error, or early disease | Skin turgor, hematocrit, repeat weighing at same time of day | | Lethargy in a group-housed primate | Social suppression, subclinical disease, or environmental stress | Isolate temporarily, compare behavior with baseline video | | Tachypnoea in a rabbit | Pain, respiratory disease, or heat stress | Rectal temperature, thoracic auscultation, response to analgesia | | Vocalisation on handling in a dog | Pain, fear, or learned avoidance | Observe undisturbed behavior, assess response to gentle palpation | | Sudden death without preceding signs | Rapidly progressive disease or missed subtle signs | Review monitoring records, consider more frequent assessment | ## Common Errors and Corrective Action Less experienced clinicians tend to rely on a single parameter, most often body weight, instead of a composite score. Body weight is useful but lags behind acute deterioration, and it is confounded by tumor mass, ascites, and edema. The corrective action is to pair weight with at least one behavioral and one physiological parameter, and to document the trend instead of the absolute value. A second error is treating the score sheet as a checklist instead of a decision tool. A score sheet that is completed but never acted upon provides no welfare benefit. The corrective action is to define, before the study begins, which score combinations trigger intervention, which trigger euthanasia, and who has authority to make that decision outside of business hours. A third error is failing to account for the cumulative burden of multiple procedures. An animal that has undergone repeated blood sampling, imaging under anesthesia, and a surgical biopsy may be below the threshold for any single endpoint criterion yet be experiencing substantial cumulative distress. The severity spectrum framework requires that the total burden, also the current clinical state, inform endpoint decisions. ## Limitations of the Evidence Base The published evidence on humane endpoints is heavily weighted toward rodents. Systematic reviews and consensus processes, including the modified Delphi study on surrogate endpoints in acute lung injury models, have focused on small animal models. The extrapolation of these findings to rabbits, dogs, and non-human primates rests on clinical judgment instead of controlled data. Species-specific endpoint validation studies are scarce, and much of the practical guidance is derived from expert opinion and institutional experience instead of prospective evaluation. Expert opinion differs on several points. There is no consensus on whether a 20% body weight loss threshold should be adjusted for species, age, or study duration. Some clinicians advocate for a 15% threshold in rabbits because of their rapid metabolic decompensation, while others argue that a 20% threshold with more frequent monitoring is acceptable. Similarly, the role of behavioral assessment in non-human primates is contested: some groups rely heavily on facial expression and posture scoring, while others consider these too subjective for reliable endpoint determination. The evidence base is also limited by poor reporting. Studies frequently describe endpoints as "humane" or "in accordance with institutional guidelines" without specifying the criteria used, which prevents replication and meta-analysis. The OBSERVE guidelines for rodent cancer models explicitly call for standardized reporting of refinement measures, and the same principle should apply to non-rodent studies. ## Referral, Consultation, and Reporting Veterinarians should seek specialist consultation when an animal's condition does not fit the predefined endpoint criteria, when a novel model produces unexpected clinical signs, or when the attending clinician is uncertain whether a clinical sign reflects disease, treatment effect, or an unrelated condition. Laboratory animal medicine specialists, clinical pathologists, and species-specific clinicians can provide valuable input. For non-human primates, consultation with a veterinary behaviorist may be warranted when social dynamics confound clinical assessment. Regulatory reporting is required when an animal experiences unanticipated pain or distress that exceeds the approved severity classification, when an animal dies before the humane endpoint is applied, or when a protocol deviation results in prolonged suffering. Institutional animal care and use committees and national oversight bodies must be notified according to local requirements. The timing and format of such reporting varies by jurisdiction, and the attending veterinarian should be familiar with the applicable standards. The attending veterinarian has the authority to override a protocol and euthanize an animal at any time, regardless of the study design, when welfare considerations demand it. This authority should be exercised transparently, documented thoroughly, and communicated to the study team without delay. ## Frequently Asked Questions ### How Should I Prioritize Endpoint Monitoring When Staffing and Time Are Limited? Prioritize the highest-risk periods and the most informative parameters. For rabbits, dogs, and non-human primates, the immediate postoperative period and the first 24 to 48 hours after any intervention carry the greatest cumulative burden. During these windows, allocate staff to twice-daily assessments. Outside these windows, daily scoring of body condition, mentation, and food intake captures most deteriorations. If full clinical scoring is impossible, weigh the animal daily and record whether it ate and defecated. Weight loss precedes visible clinical decline in many models, and an algorithm based on body weight trajectory can predict endpoints more consistently than intermittent observation by different scorers, as shown in a rat glioma model [body weight algorithm predicts humane endpoint in an intracranial rat glioma model](https://pubmed.ncbi.nlm.nih.gov/32488031/). Document any reduced monitoring frequency in the protocol and justify it prospectively. ### What Do I Do When the Planned Euthanasia Method or Monitoring Equipment Is Unavailable? Use the least invasive alternative that still provides reliable data, and change the endpoint criteria accordingly. If telemetric physiological monitoring is unavailable, replace it with serial body weight, clinical score, and behavior-based assessments. These measures detect deterioration adequately when performed consistently by trained observers [body weight algorithm predicts humane endpoint in an intracranial rat glioma model](https://pubmed.ncbi.nlm.nih.gov/32488031/). If the preferred euthanasia agent is unavailable, consult the current [AVMA professional practice resources](https://www.avma.org/resources-tools) for acceptable alternatives and adjust the endpoint to ensure the animal does not reach a state where the substitute method could cause distress. Record the substitution and the rationale in the animal care and use documentation. Do not delay euthanasia while waiting for a specific drug or device. ### How Do Endpoint Criteria Differ Between a Purpose-Bred Beagle and a Client-Owned Dog in a Non-Clinical Study? Purpose-bred research dogs have documented health histories, known genetic backgrounds, and standardized housing, so baseline values for weight, temperature, and behavior are reliable. Client-owned dogs entering a study carry unknown comorbidities, variable socialisation, and prior medication histories, which can confound endpoint interpretation. For client-owned animals, establish a longer acclimation period and obtain a thorough history before assigning baseline thresholds. Use a higher index of suspicion for pain or distress when behavior changes, because prior experience shapes individual responses. The [Guide for the Care and Use of Laboratory Animals](https://grants.nih.gov/grants/olaw/guide-for-the-care-and-use-of-laboratory-animals.pdf) emphasizes individualised veterinary assessment over rigid numerical cut-offs. In both populations, document the rationale for any deviation from the approved endpoint criteria. ### What Records Must I Keep to Demonstrate That Endpoints Were Applied Correctly? Maintain a per-animal log with date, time, observer identity, clinical score components, body weight, and any intervention performed. Record the specific criteria that triggered euthanasia and the time from criterion fulfilment to euthanasia. Include photographs or video where skin lesions, ocular changes, or gait abnormalities are part of the criteria, because these support retrospective review. Record all deviations from the approved protocol, including delayed recognition, and the corrective action taken. The [NC3Rs resources on replacement, reduction and refinement](https://www.nc3rs.org.uk/) provide templates for score sheets and monitoring records. Institutional oversight bodies expect these records to demonstrate that endpoints were applied consistently and that staff were competent. Keep records for the duration required by your institution, which often exceeds the study period. ### How Should I Present a Humane Endpoint Recommendation to a Principal Investigator Who Resists Early Euthanasia? Frame the recommendation around data quality, not sentiment. An animal that reaches a moribund state produces physiological data that are confounded by agonal change, and tissue collected after prolonged distress may not reflect the experimental condition under study. Reference the consensus that surrogate endpoints should replace death as the primary outcome in animal models [surrogate humane endpoints in small animal models of acute lung injury](https://pubmed.ncbi.nlm.nih.gov/33332817/). Offer to validate the endpoint by correlating the clinical score with the primary outcome measure in a small pilot cohort. Propose a staged escalation, where the investigator is notified at a pre-terminal score and euthanasia occurs at a defined threshold. This preserves scientific objectives while meeting welfare obligations. Document the discussion and the agreed criteria in the protocol amendment. ### How Do I Set Endpoints for a Species When Published Reference Values Are Sparse? Build the endpoint from species-specific biology and pilot data instead of extrapolating from rodents. For rabbits, monitor fecal output and gastrointestinal motility closely because ileus is a common terminal pathway. For non-human primates, prioritize social withdrawal and changes in facial expression, which experienced observers detect earlier than weight loss. Use the [MSD Veterinary Manual](https://www.msdvetmanual.com/) for species-specific normal physiology and clinical signs. Run a small pilot cohort to establish baseline ranges for your colony under your housing conditions, then set thresholds at two standard deviations from that baseline. Consult the [WOAH terrestrial animal health standards](https://www.woah.org/en/what-we-do/standards/codes-and-manuals/terrestrial-code-online-access/) where disease-related endpoints intersect with notifiable conditions. Review the criteria after the first five animals and adjust if animals are reaching endpoints earlier or later than predicted. ## Related Clinical & Scientific Guides * [Refining IACUC Protocols to Minimize Animal Pain and Distress](/knowledge/veterinary-medicine/laboratory-animal-science/refining-iacuc-protocols-minimize-animal-pain-distress) * [Health Monitoring Programs for Laboratory Animal Facilities](/knowledge/veterinary-medicine/laboratory-animal-science/health-monitoring-programs-for-laboratory-animal-facilities) * [Anesthetic Risk Assessment in Laboratory Animals: Preoperative Evaluation](/knowledge/veterinary-medicine/laboratory-animal-science/anesthetic-risk-assessment-in-laboratory-animals-preoperative-evaluation) ## References and Further Reading - [Opportunities for improving animal welfare in rodent models of epilepsy and seizures.](https://pubmed.ncbi.nlm.nih.gov/26376175/). 2016. - [Analgesia in clinically relevant rodent models of sepsis.](https://pubmed.ncbi.nlm.nih.gov/27909191/). 2016. - [Surrogate Humane Endpoints in Small Animal Models of Acute Lung Injury: A Modified Delphi Consensus Study of Researchers and Laboratory Animal Veterinarians.](https://pubmed.ncbi.nlm.nih.gov/33332817/). 2021. - [OBSERVE: guidelines for the refinement of rodent cancer models.](https://pubmed.ncbi.nlm.nih.gov/38992214/). 2024. - [Body weight algorithm predicts humane endpoint in an intracranial rat glioma model.](https://pubmed.ncbi.nlm.nih.gov/32488031/). 2020. - [Malaysia's commemorative activities for the second Asia Laboratory Animal Day (ALAD).](https://pubmed.ncbi.nlm.nih.gov/41502386/). 2025. - [Guide for the Care and Use of Laboratory Animals, 8th Edition](https://grants.nih.gov/grants/olaw/guide-for-the-care-and-use-of-laboratory-animals.pdf). National Academies Press, 2011. - [NC3Rs Resources on Replacement, Reduction and Refinement](https://www.nc3rs.org.uk/). NC3Rs. - [MSD Veterinary Manual, Professional Edition](https://www.msdvetmanual.com/). MSD Veterinary Manual. ## Related Articles - [Defining Humane Endpoints in Rodent Cancer Models](/knowledge/veterinary-medicine/laboratory-animal-science/defining-humane-endpoints-in-rodent-cancer-models) - [Humane Endpoint Determination for Rodent Models of Sepsis](/knowledge/veterinary-medicine/laboratory-animal-science/humane-endpoint-determination-for-rodent-models-of-sepsis) - [Designing Humane Endpoints for Vaccine Efficacy Studies](/knowledge/veterinary-medicine/laboratory-animal-science/designing-humane-endpoints-for-vaccine-efficacy-studies) - [Evaluating Body Condition Scoring for Laboratory Rodent Welfare](/knowledge/veterinary-medicine/laboratory-animal-science/evaluating-body-condition-scoring-laboratory-rodent-welfare) - [Selecting Bedding Materials for Laboratory Rodent Welfare](/knowledge/veterinary-medicine/laboratory-animal-science/selecting-bedding-materials-laboratory-rodent-welfare) > This article is educational professional reference material for veterinary audiences. It is not a substitute for veterinary diagnosis, individual clinical judgment, current product labeling, or applicable regulatory requirements.