Designing Humane Endpoints for Vaccine Efficacy Studies

By Dr. Zubair Khalid, DVM, MS, PhD · 2026-08-11

Key Takeaways

• Humane endpoints are critical for balancing scientific validity in vaccine efficacy studies with animal welfare, requiring prospective definition of criteria to trigger euthanasia or intervention before an animal reaches a moribund state, thereby preventing severe pain and distress.
• Endpoint criteria must be objective and measurable, integrating clinical signs (e.g., gait, respiratory effort, neurological deficits), physiological parameters (e.g., rectal temperature, body weight change), and species-specific disease manifestations, validated to reliably predict moribundity with acceptable sensitivity and specificity.
• Monitoring frequency must be tailored to disease kinetics, with rapid-progressing infections (e.g., African Swine Fever Virus) necessitating more frequent assessments than slower models, and observation schedules should include both routine and triggered checks.
• Scoring systems should convert clinical observations into quantifiable values, using a tiered approach where a total score or a maximum score for a single critical parameter (e.g., seizure activity, severe dyspnea) triggers euthanasia, with parameters weighted by welfare impact rather than solely diagnostic value.
• Documentation of all endpoint decisions, including date, time, individual parameter scores, total score, observer identity, and action taken, is essential for regulatory compliance, protocol refinement, and demonstrating consistent application of criteria across study groups.
• Refinement opportunities, such as the use of serological assays or in vitro methods as alternatives to lethal challenge where validated, and the judicious use of supportive care (e.g., analgesics, fluids) that does not confound study objectives, should be integrated into study design to reduce animal use and suffering.

Vaccine efficacy studies require deliberate infection of research animals with pathogenic organizms to measure protection against disease, death, or colonization. The central welfare problem is that the unvaccinated control group must develop measurable disease for the study to be scientifically valid, yet that same disease process can cause substantial pain and distress. Humane endpoints are predetermined criteria that trigger euthanasia or other alleviating intervention before an animal reaches a moribund state, while still permitting collection of the data the study was designed to generate. This article provides a framework for designing, validating, and implementing humane endpoints in vaccine challenge studies across species, with emphasis on the procedural decisions a veterinary researcher must make during protocol development and study conduct.

The reader is assumed to be a veterinary professional engaged in vaccine research, whether in an academic, government, or commercial setting. The article addresses the practical question of how to select endpoint criteria that satisfy both regulatory expectations and animal welfare obligations without compromising the statistical power or interpretability of the efficacy data. The principles discussed apply to rodent models, swine, non-human primates, and other species used in challenge studies, though species-specific considerations are noted where they materially alter the approach.

At a Glance

Scientific Basis for Humane Endpoints

Humane endpoints rest on the premise that an animal's physiological and behavioral state deteriorates in a predictable sequence during progressive infectious disease. If that sequence can be characterized, a threshold can be identified that reliably precedes death or irreversible suffering. The endpoint threshold must be early enough to prevent distress but late enough that the study retains its ability to distinguish vaccinated from unvaccinated animals. This balance is the central design problem in vaccine challenge studies.

The regulatory framework for humane endpoints derives from the recognition that laboratory animal testing can cause significant pain and distress, and that alleviation is permissible only when it does not interfere with study objectives. The National Research Council's guide for laboratory animal care states that animals should not be permitted to die from disease unless the study design specifically requires death as an endpoint, and even then, such designs require strong scientific justification and institutional approval. Institutional animal care and use committees evaluate proposed endpoints against this standard, and the scientific community has increasingly adopted the position that death as an endpoint should be replaced wherever a validated surrogate exists.

The statistical logic of humane endpoints deserves explicit attention. A vaccine efficacy study compares outcomes between vaccinated and control groups. If the endpoint is death, the outcome is binary and the analysis is straightforward. If the endpoint is a clinical score or a physiological threshold, the outcome becomes a time-to-event variable, and the analysis must account for censoring when animals are euthanized at the endpoint. Survival analysis methods, including Kaplan-Meier curves and Cox proportional hazards models, accommodate this structure. The key requirement is that the endpoint criteria are applied consistently across all groups, so that any bias in endpoint application affects all animals equally.

Defining the Endpoint Threshold

The endpoint threshold must be anchored to a measurable, repeatable set of criteria that a trained observer can apply reliably. The published literature on experimental Streptococcus suis infection in pigs illustrates both the diversity of approaches and the current lack of standardization. Body temperature thresholds for fever in these studies range widely, with many protocols setting the threshold at 40.5°C, while monitoring intervals vary from once to three times daily. Only a minority of published studies use formal scoring systems, and those that do vary in which parameters they include. This heterogeneity complicates cross-study comparison and makes it difficult to establish best practices from the literature alone.

A defensible endpoint definition includes objective physiological parameters, such as rectal temperature, body weight change, and respiratory rate, combined with behavioral and clinical observations that are scored using a defined rubric. The rubric should be piloted before the study begins, using historical data or a small cohort, to confirm that the criteria are achievable and that they predict moribundity with acceptable accuracy. Retrospective analysis of health records from animals previously infected with the challenge agent can identify which parameters best predict the transition to a moribund state. For African swine fever virus in swine, regression analysis of historical data from 103 experimentally infected animals identified rectal temperature as a potentially useful predictor of moribundity, while viral strain and survival duration after inoculation were significant risk factors for death due to disease instead of euthanasia. This type of analysis, performed on data from previous studies, provides an evidence base for endpoint selection that is superior to expert opinion alone.

Species-Specific Considerations

The clinical manifestations of a given pathogen differ across species, and the endpoint criteria must reflect the species-specific disease course. In swine, Streptococcus suis infection produces meningitis, arthritis, and septicemia, each with distinct clinical presentations that require different monitoring parameters. Locomotion scoring and central nervous system signs are relevant for meningeal and joint disease, while fever and feeding behavior capture systemic illness. In non-human primates, the range of vaccine studies spans respiratory pathogens, hemorrhagic fevers, and other agents, each with its own clinical trajectory. The requirement for high-level biocontainment for some pathogens imposes additional challenges to implementing refinement, because animals may be observed only through protective equipment or video monitoring, which limits the sensitivity of clinical assessment.

Rodent models present a different set of challenges. The small body size of mice and hamsters limits the volume of blood that can be collected for monitoring, and clinical signs may be subtle until the animal is severely ill. Intracerebral injection of rabies virus in mice, used for vaccine potency testing, produces neurologic signs that progress rapidly, and the international workshop on rabies vaccine testing recommended that general anesthesia be used for the injection procedure and that humane endpoints be used routinely as the basis for euthanasia. For Leptospira vaccine potency testing in hamsters, the challenge infection produces unrelieved pain and distress, and workshop participants encouraged consideration of analgesics and earlier humane endpoints when the hamster vaccination-challenge assay is used.

Balancing Welfare with Scientific Validity

The tension between welfare and scientific validity is most acute in the control group, which must develop disease to demonstrate that the vaccine confers protection. If the endpoint is set too early, control animals may be euthanized before they develop measurable disease, and the study may fail to detect a difference between groups. If the endpoint is set too late, animals experience avoidable suffering. The resolution lies in careful characterization of the disease model before the efficacy study begins, using pilot studies or historical data to establish the time course of clinical signs and the relationship between clinical severity and the probability of death.

Refinement opportunities exist at multiple levels of study design. The international workshop on Leptospira vaccine potency testing identified the potential for eliminating the back-titration procedure in the hamster challenge assay, which could reduce animal use by 50% for each potency test, and for using cryopreserved Leptospira stock to replace continual passaging through hamsters. Serology assays were identified as a way to further reduce and refine animal use, though they should be considered only after attempting in vitro assays. For rabies vaccines, the mouse antibody serum neutralization test has been validated for adjuvanted veterinary vaccines and was recommended as a near-term priority for replacing the mouse challenge. These alternatives do not eliminate the need for humane endpoints in studies where challenge is still required, but they reduce the number of animals exposed to infection and thus the aggregate welfare burden.

Designing the Clinical Monitoring Protocol

The monitoring protocol must be defined before challenge begins and approved by the institutional animal care and use committee. Its structure determines whether endpoint criteria can be applied consistently across animals, observers, and time points. A protocol that is too rigid will fail when disease presents atypically. A protocol that is too flexible will produce inconsistent decisions and compromise both welfare and data integrity.

Selecting Observation Frequency

Observation frequency should be matched to the expected disease course and the speed with which clinical signs progress. For pathogens that cause rapid deterioration, such as African swine fever virus, daily observation is insufficient. The retrospective analysis of 103 experimentally infected swine found that predicting moribundity from clinical signs alone was difficult, supporting the use of frequent assessments and objective physiologic markers Clinical Indicators of Moribundity in Swine Experimentally Inoculated with. For slower disease models, twice-daily observation may be adequate, but the interval must never exceed what the fastest expected progression would allow.

The observation schedule should specify both routine checks and triggered checks. Routine checks occur at fixed intervals. Triggered checks occur when an animal crosses a predefined threshold, such as fever, and require more frequent assessment until the animal either recovers or reaches an endpoint. The review of experimental Streptococcus suis infection in pigs found that monitoring intervals in published studies varied between daily and three times daily, with few studies applying formal scoring systems A critical review on experimental Streptococcus suis infection in. This variation reflects differing pathogen kinetics, but it also indicates that many protocols default to convenience instead of to disease biology.

Selecting Parameters

Each parameter in the monitoring protocol must have a clear rationale linking it to the pathogenesis of the challenge agent. Parameters fall into four categories: general condition, specific clinical signs, physiologic measurements, and behavioral indicators.

General condition parameters include coat condition, posture, activity, and response to handling. These are nonspecific but often the first indicators of deterioration. Specific clinical signs depend on the pathogen and target organ system. For S. suis, these include central nervous system signs such as head tilt, circling, and recumbency, as well as lameness from arthritis A critical review on experimental Streptococcus suis infection in. For respiratory pathogens, respiratory rate, effort, and auscultation findings take priority. For pathogens causing systemic disease, the combination of fever, reduced feed intake, and lethargy may precede more severe signs.

Physiologic measurements add objectivity. Rectal temperature is the most commonly used and the most frequently standardized. The S. suis review noted that most studies set the fever threshold at 40.5°C, but this threshold was applied inconsistently and without clear justification A critical review on experimental Streptococcus suis infection in. Body weight is another objective parameter, but it changes slowly and is more useful for subacute or chronic models than for acute disease. The ASFV analysis identified rectal temperature as a potential predictor of moribundity, though the relationship was not strong enough to serve as a sole criterion Clinical Indicators of Moribundity in Swine Experimentally Inoculated with.

Behavioral indicators include response to external stimuli, interaction with cage mates, and willingness to move. These are particularly relevant for social species such as pigs and non-human primates. Social housing, environmental enrichment, and training for cooperation with procedures are features of contemporary good practice that improve both welfare and the reliability of behavioral assessment Opportunities for Refinement of Non-Human Primate Vaccine Studies.

Parameter Selection by Pathogen Class

The choice of primary parameters determines the sensitivity of the endpoint system. A protocol that relies only on temperature will miss animals that deteriorate without fever. A protocol that relies only on behavior will miss animals that are physiologically compromised but still mobile.

Building the Scoring System

A scoring system converts clinical observations into a numeric value that can be compared against a threshold. The system must be validated for the specific challenge model before it is used to make euthanasia decisions. Published scoring systems for S. suis infection in pigs are inconsistent in their inclusion of parameters such as body temperature, feeding behavior, and respiratory signs A critical review on experimental Streptococcus suis infection in. This inconsistency makes cross-study comparison difficult and can lead to animals being maintained longer than necessary in one study and euthanized prematurely in another.

Score Structure

Each parameter is scored on a scale, typically 0 to 3, where 0 is normal and 3 is severe. The scores are summed to produce a total. The threshold for euthanasia is set at a total score that corresponds to a defined level of suffering, with the additional requirement that any single parameter reaching its maximum score triggers euthanasia regardless of the total. This two-tier structure prevents an animal with one catastrophic sign from being kept alive because other parameters are normal.

The scoring system must define each score level in operational terms. A score of 1 for respiratory effort must specify what the observer sees, such as increased abdominal effort at rest. A score of 2 must specify a further change, such as open-mouth breathing. Vague descriptors such as "moderate distress" are not acceptable because they cannot be applied consistently.

Weighting Parameters

Not all parameters carry equal welfare weight. A scoring system that sums unweighted scores treats a mild fever as equivalent to a seizure. Weighting can be applied by multiplying parameter scores by a factor before summation, or by setting separate thresholds for individual parameters. The latter approach is simpler and more transparent. For example, a protocol might specify that any animal with a seizure lasting more than 30 seconds is euthanized immediately, regardless of total score.

The weighting should reflect the welfare impact of each sign, not its diagnostic value. A sign that is diagnostically important but causes little suffering, such as mild lymphadenopathy, should not trigger euthanasia. A sign that causes significant suffering, such as dyspnea or neurologic impairment, should trigger euthanasia even if it is not pathognomonic.

Documenting Endpoint Decisions

Every endpoint decision must be documented in a way that allows retrospective review. The record should include the date and time of each observation, the individual parameter scores, the total score, the identity of the observer, and the action taken. This documentation serves three purposes: it provides an audit trail for regulatory review, it allows the protocol to be refined based on actual disease progression, and it supports the scientific validity of the study by demonstrating that endpoints were applied consistently.

The observation record should be designed before the study begins. A paper form or electronic spreadsheet with preprinted parameter names and score definitions reduces the risk of observer error. The form should include a comments field for observations that do not fit the scoring system, such as unexpected clinical signs or equipment failures.

Observer training is a prerequisite for reliable scoring. All personnel who will score animals must be trained on the scoring system using video or photographic examples, and their scoring must be compared against a reference standard. Inter-observer reliability should be assessed before the study starts and periodically during the study. Discrepancies should be resolved by discussion and, if necessary, by adjusting the score definitions.

Refining Endpoints During the Study

The endpoint protocol should be treated as a living document. If the first animals in a study reach endpoints earlier or later than predicted, the protocol should be reviewed. Animals that reach endpoints earlier than predicted may indicate that the scoring system is too sensitive, that the challenge dose is too high, or that the predicted disease course was incorrect. Animals that survive longer than predicted may indicate that the scoring system is missing signs of deterioration.

Any protocol change during a study must be approved by the institutional animal care and use committee and documented in the study records. Changes that affect the scientific endpoints, such as altering the euthanasia threshold, may compromise the study's validity and should be made only when the welfare benefit clearly outweighs the scientific cost. The use of humane endpoints is a prerequisite for regulatory acceptance of vaccine potency tests, and deviations from approved protocols must be justified Humane endpoints for laboratory animals used in regulatory testing.

Supportive Care Without Confounding

Supportive care can reduce suffering without compromising study objectives, but it must be planned prospectively. Analgesics, fluids, and nutritional support may be appropriate for some challenge models and inappropriate for others. The decision depends on whether the treatment would mask the clinical signs that define the endpoint or alter the immune response under study. Workshop participants addressing Leptospira vaccine potency testing encouraged consideration of analgesics and earlier humane endpoints in the hamster challenge assay Report on the international workshop on alternative methods for. This recommendation reflects a broader shift toward providing supportive care whenever it does not interfere with the scientific readout.

The protocol must specify which supportive measures are permitted, which are prohibited, and under what conditions each applies. A measure that is permitted for one study may be prohibited for another, depending on the vaccine's mechanism of action and the challenge model's endpoints. The decision should be made during protocol development, not in response to an animal's deterioration.

Recognized Complications and Failure Modes

The most common failure in vaccine efficacy studies is not the absence of endpoints but their misapplication. A scoring system that is too permissive allows animals to progress to moribundity before euthanasia criteria are met, while an overly strict system triggers euthanasia before the study can distinguish vaccine efficacy from natural disease progression. Both failure modes compromise the scientific objective and the welfare mandate.

Early detection depends on distinguishing predictable clinical trajectories from aberrant ones. In swine models of African swine fever virus, rectal temperature has shown utility as a predictor of moribundity, but viral strain and survival duration after inoculation are independent risk factors for death due to disease instead of euthanasia. A pig that spikes to 41.5°C on day 3 post-inoculation and a pig that reaches the same temperature on day 9 are not equivalent, the monitoring protocol must account for expected disease kinetics.

A second recognized failure mode is observer drift. When multiple technicians score animals across a multi-week study, individual interpretation of ambiguous parameters such as "reduced responsiveness" or "mild dehydration" diverges over time. This is detected by scheduled inter-observer reliability checks, where two trained observers independently score the same animal and their results are compared. Discrepancies above a pre-set threshold trigger retraining and recalibration.

A third failure mode is the unplanned interaction between supportive care and endpoint criteria. Analgesic administration can mask pain-related parameters, while fluid therapy can transiently improve hydration scores without altering the underlying disease trajectory. The protocol must specify whether supportive care is permitted and how it affects endpoint scoring, otherwise clinical judgment becomes inconsistent across animals.

Common Errors in Endpoint Design

Less experienced investigators frequently make errors at the design stage instead of during execution. The most common is selecting parameters that are easy to measure but poorly correlated with the actual disease process. Rectal temperature is convenient and widely used, but it is a lagging indicator in many infections and can be normal during the terminal phase of some diseases. The published literature on experimental Streptococcus suis infection in pigs shows substantial inconsistency in clinical monitoring, with most studies setting fever thresholds at 40.5°C but varying widely in monitoring frequency and scoring parameters.

A second error is failing to pilot the scoring system. A scoring system developed from the literature may not transfer to a specific challenge model, a specific pathogen strain, or a specific animal source. A pilot cohort of two to three animals, observed intensively and scored daily, will reveal whether the parameters discriminate between mild and severe disease and whether the thresholds are set at appropriate points.

A third error is treating the endpoint protocol as fixed once approved. Humane endpoints require ongoing refinement as data accumulate during the study. If the first cohort of animals reaches euthanasia criteria earlier than predicted, the protocol should be reviewed, not simply continued. Conversely, if animals are surviving longer than expected with poor welfare, the thresholds may be too permissive.

Limitations of the Evidence Base

The evidence supporting humane endpoint design in vaccine efficacy studies is uneven across species and pathogens. The most developed literature comes from rabies vaccine potency testing, where international workshops have recommended general anesthesia for intracerebral virus injections and routine use of humane endpoints for the mouse challenge test. Similarly, Leptospira vaccine potency testing in hamsters has been the subject of international harmonisation efforts, with recommendations for analgesics and earlier humane endpoints.

For other pathogen classes, the evidence base is thinner. The review of S. suis infection models notes that published scoring systems are inconsistent in their inclusion of parameters such as body temperature, feeding behavior, and respiratory signs, and that only a few studies apply scoring systems at all. Expert opinion still differs on whether composite scoring systems are superior to a small number of well-validated single parameters, and on whether weight loss thresholds should be absolute or relative to baseline.

There is also genuine uncertainty about the transferability of endpoints between production systems and regions. A threshold appropriate for a specific pathogen strain in one geographic region may not apply to another strain with different virulence characteriztics. The WOAH terrestrial animal health standards provide a framework for disease-specific considerations, but they do not prescribe species-specific endpoint criteria.

Referral, Consultation, and Reporting

Specialist consultation is warranted when the study involves a pathogen with high morbidity and mortality, when the species is one with limited published endpoint data, or when the study is conducted under high-level biocontainment where observation is logistically constrained. Laboratory animal veterinarians with species expertise should be involved in protocol design from the outset, not consulted after welfare problems emerge.

Regulatory reporting obligations vary by jurisdiction and by funding source. Institutional animal care and use committees require reporting of unexpected mortality, unrelieved pain or distress, and deviations from approved protocols. The AVMA professional practice resources provide guidance on veterinary obligations in research settings, while the NC3Rs resources on the 3Rs offer practical frameworks for refinement that can inform protocol revisions. When a study must be stopped early because endpoints were inadequate, the failure should be documented and shared, as this information is essential for improving future study design.

Frequently Asked Questions

How do I justify a more refined endpoint protocol to an animal ethics committee or institutional review board?

Present the refinement as a scientific improvement, also a welfare concession. Explain that moribund animals produce physiologically distorted data, including aberrant cytokine profiles and bacteremia, that can obscure vaccine efficacy signals. Reference the National Research Council guide for laboratory animal care as the standard for endpoint justification. Show that your proposed criteria are evidence based, citing published scoring systems for your model species where they exist. Include a pilot data table or literature summary demonstrating that your endpoint threshold does not precede the earliest measurable vaccine-specific response. Offer a contingency plan for unscheduled euthanasia and state your expected mortality distribution under the refined protocol.

What should I do when the ideal monitoring equipment is unavailable in a high-containment facility?

Prioritize parameters that require no specialized equipment. Rectal temperature, body condition scoring, and gait assessment need only a thermometer and trained observation. The NC3Rs refinement resources emphasize that simple, repeatable clinical signs often outperform complex instrumentation in biocontainment settings. If telemetry or automated activity monitoring is unavailable, increase observation frequency and use two independent observers to reduce subjectivity. Validate your simplified score against the gold standard in a small pilot cohort before the main study. Document equipment limitations in the protocol and note how they affect endpoint sensitivity. Consider whether video recording through containment windows permits retrospective scoring without additional staff exposure.

How do endpoint criteria differ between purpose-bred laboratory animals and production animals such as pigs?

Production species present distinct challenges because normal physiological values vary with breed, age, and management system. Fever thresholds in pigs, for example, range widely across published studies, with many groups using 40.5°C despite considerable inter-individual variation. The critical review of Streptococcus suis infection models highlights inconsistent scoring systems across swine studies and recommends harmonization. Production animals also show stoic behavior, masking pain until late stages. Weight loss is a more reliable indicator in growing pigs than in mature laboratory rodents. Locomotion scoring carries greater weight in pigs because arthritis and meningitis are common endpoints. Always establish baseline values for your specific population during acclimation instead of relying on textbook ranges.

What records must I keep for endpoint decisions to satisfy regulatory and sponsor expectations?

Maintain a prospective endpoint log separate from raw clinical data. Each entry should record the date and time of observation, the observer identity, individual parameter scores, the total score, and the specific criterion that triggered euthanasia. Document any score that approached but did not reach the threshold, with justification for continued observation. The WOAH terrestrial animal health standards require traceability for studies supporting vaccine licensure. Retain photographs or video of representative clinical signs where feasible. Record deviations from the protocol immediately, including the rationale and the attending veterinarian's approval. Archive these records with the study data so that auditors can reconstruct every euthanasia decision independently of the primary dataset.

How do I explain humane endpoints to a sponsor or collaborator who wants survival as the primary efficacy outcome?

Frame the discussion around data quality and regulatory acceptability. Survival as a crude binary outcome ignores the timing and mechanism of death, and animals that die overnight provide no clinical data. The workshop report on rabies vaccine potency testing documents international movement toward earlier endpoints in challenge models. Offer time-to-meet-endpoint as a statistically robust alternative that preserves survival analysis while meeting welfare obligations. Show that endpoint-based analyzes correlate with survival in historical data from your model. If the sponsor remains resistant, propose a composite outcome with survival as a secondary measure and endpoint as primary, then present both analyzes. Regulatory authorities increasingly expect refined endpoints, so framing the change as aligning with current standards often resolves the objection.

Can I provide analgesics or supportive care to animals on a vaccine efficacy study without invalidating the results?

Yes, with careful design. The workshop report on Leptospira vaccine potency testing explicitly encourages consideration of analgesics in challenge assays. Choose agents that do not modulate the immune response under investigation. Non-steroidal anti-inflammatory drugs may affect fever, which is often a scoring parameter, so document their use and adjust temperature thresholds accordingly. Fluid therapy and nutritional support do not typically confound vaccine readouts. Administer supportive care only after defining, in advance, which interventions are permitted and how they will be recorded. If an animal requires intervention that might alter the immune response, remove it from the efficacy analysis but report it separately. The AVMA professional practice resources provide guidance on analgesic selection across species.

Related Clinical & Scientific Guides

• Refining IACUC Protocols to Minimize Animal Pain and Distress
• Health Monitoring Programs for Laboratory Animal Facilities
• Anesthetic Risk Assessment in Laboratory Animals: Preoperative Evaluation

References and Further Reading

• Report on the international workshop on alternative methods for human and veterinary rabies vaccine testing: state of the science and planning the way forward.. 2012.
• A critical review on experimental Streptococcus suis infection in pigs with a focus on clinical monitoring and refinement strategies.. 2023.
• Clinical Indicators of Moribundity in Swine Experimentally Inoculated with African Swine Fever Virus.. 2021.
• Humane endpoints for laboratory animals used in regulatory testing.. 2002.
• Opportunities for Refinement of Non-Human Primate Vaccine Studies.. 2021.
• Report on the international workshop on alternative methods for Leptospira vaccine potency testing: state of the science and the way forward.. 2013.
• Guide for the Care and Use of Laboratory Animals, 8th Edition. National Academies Press, 2011.
• NC3Rs Resources on Replacement, Reduction and Refinement. NC3Rs.
• MSD Veterinary Manual, Professional Edition. MSD Veterinary Manual.

Related Articles

• Defining Humane Endpoints in Rodent Cancer Models
• Humane Endpoints for Non-Rodent Laboratory Species
• Humane Endpoint Determination for Rodent Models of Sepsis
• Anesthesia Equipment for Small Laboratory Animals: Setup and Maintenance
• Anesthesia for Laboratory Rabbits: Protocols and Monitoring

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.