Yes, most animals used in laboratory testing are killed at the end of a study. In the United States, roughly 95 percent of rodents and other purpose-bred research animals are euthanized rather than rehomed, because federal law, Institutional Animal Care and Use Committee (IACUC) protocols, and scientific design typically require tissue collection at the endpoint. A minority are reused across non-terminal studies or placed through adoption programs, which remain rare but are growing. The fate depends on species, study type, regulatory status, and institutional policy.
The Mechanism: Why Endpoints Usually Mean Death
Animal testing is not a single act. It is a sequence: acquisition, acclimation, treatment, observation, and endpoint. The endpoint is the moment data collection stops. In most protocols, the endpoint is terminal by design, because the science requires tissue that only exists after death.
Consider a standard toxicity study. Researchers dose animals for 28 or 90 days, then necropsy them to examine liver, kidney, and brain tissue under a microscope. Blood is drawn terminally. Organs are weighed and fixed. None of that is possible if the animal walks away alive. The necropsy is the data.
The mechanism has three drivers. First, regulatory requirements: the FDA's toxicology guidance and the OECD test guidelines for chemicals are built around terminal necropsy. Second, statistical design: terminal sampling removes the confound of recovery between measurement and analysis. Third, biosafety: animals exposed to infectious agents or hazardous compounds generally cannot re-enter the general population.
In plain terms: the animal is often the measurement instrument, and the study ends when the instrument is read.
A smaller share of studies are non-terminal. Behavioral, observational, and some pharmacokinetic studies can end with the animal alive. Even then, reuse is limited by protocol, by prior exposure history, and by the practical cost of holding animals between studies.
Quantified Context: How Many Animals, How Many Deaths
The scale is large and the death rate is high. The most-cited global estimate, from Taylor and colleagues in 2008, put annual vertebrate use at roughly 115 million animals. More recent analyses suggest the figure has shifted with outsourcing and model changes, but the order of magnitude holds.
In the US, the Animal Welfare Act (AWA) covers only certain species: dogs, cats, primates, rabbits, guinea pigs, hamsters, and a few others. Mice, rats, and birds bred for research are explicitly excluded from AWA counts. That exclusion is the single biggest reason official US numbers understate reality. AWA-reported use is typically around 800,000 to 1 million animals per year, while the true total including rodents is estimated in the tens of millions.
| Metric | Figure | Notes |
|---|---|---|
| Global annual vertebrate use | ~115 million (2008 estimate) | Taylor et al., widely cited |
| US AWA-covered animals | ~800,000–1,000,000/yr | Excludes mice, rats, birds |
| US total including rodents | Tens of millions/yr | Estimated, not officially counted |
| Rodents as share of research animals | ~90–95% | Mice and rats dominate |
| Estimated euthanasia rate at study end | ~90–95% | Species and study dependent |
| Rehoming rate (dogs, cats, primates) | Single-digit percent | Institution and species dependent |
The rehoming number is the hardest to pin. No federal database tracks post-study adoption. Survey work and institutional reports suggest that among AWA-covered species, adoption programs exist at a minority of facilities, and the proportion of eligible animals actually placed is typically in the single digits. Beagles from toxicity studies and primates from non-terminal behavioral work are the most common candidates.
Species Variation: Who Gets Rehomed and Who Does Not
The fate of an animal after a study is strongly predicted by species, not by the study's outcome alone.
Dogs and cats. Beagles are the standard canine toxicology model. Some facilities run adoption programs, often partnered with rescue organizations. A 2014 change to the AWA added a requirement for research facilities to place dogs and cats in adoption if a suitable home is found, rather than automatically euthanizing them. The rule raised placement rates but did not eliminate euthanasia, because behavioral or medical history can disqualify an animal.
Non-human primates. Primates are expensive to maintain and ethically sensitive to retire. A small number of sanctuaries, such as Chimp Haven for chimpanzees retired from NIH-supported research, accept animals. Most primates on terminal studies are euthanized. The 2015 NIH decision to retire remaining research chimpanzees is the clearest policy-driven exception.
Rodents. Mice and rats are almost never rehomed. Their numbers are too large, their health status (often specific-pathogen-free or experimentally modified) makes them unsuitable as pets, and their short lifespans make post-study placement impractical. For rodents, euthanasia at endpoint is near-universal.
Fish and amphibians. Increasingly used in research, these species have sparse welfare oversight and low rehoming rates. Most are euthanized by immersion in buffered tricaine or similar agents.
Comparison: Fate of Animals at Study End
| Species | Typical terminal endpoint | Reuse potential | Rehoming/retirement | Primary driver |
|---|---|---|---|---|
| Mice/rats | Near-universal euthanasia | Low | Rare | Numbers, health status, cost |
| Dogs (beagles) | Euthanasia or adoption | Moderate | Growing, single digits | AWA placement rule (2014) |
| Cats | Euthanasia or adoption | Moderate | Growing, single digits | AWA placement rule (2014) |
| Non-human primates | Mostly euthanasia | Moderate | Very rare, sanctuary-based | Cost, biosecurity |
| Rabbits | Euthanasia | Low | Occasional | Study design |
| Fish/amphibians | Euthanasia | Minimal | Rare | Limited oversight |
The table makes a pattern visible: the more an animal resembles a pet and the fewer of them a study uses, the more likely it is to be rehomed. The more it resembles a model organism used in bulk, the more likely it is euthanized.
Timeline: The Life Cycle of a Research Animal
| Stage | Typical duration | What happens | Fate hinge |
|---|---|---|---|
| Breeding/acquisition | Weeks to months | Purpose-bred, health-screened | Source facility policy |
| Acclimation | 1–2 weeks | Housing, handling, baseline data | Welfare check |
| Study conduct | 1 day to 2 years | Dosing, observation, sampling | Terminal vs non-terminal design |
| Endpoint | Study-defined | Necropsy or recovery | Protocol |
| Post-study decision | Days | Reuse, rehome, retire, or euthanize | IACUC protocol and adoptability |
| Euthanasia | Minutes | Approved agent (e.g., CO2, pentobarbital) | AVMA guidelines |
The American Veterinary Medical Association (AVMA) publishes euthanasia guidelines that specify acceptable agents and methods by species. These guidelines are the operative standard in most US facilities.
Misconceptions and Non-Obvious Details
Misconception one: "All animals die at the end." Not true, but close. Non-terminal studies exist, and reuse across studies is real. The 90–95 percent figure is an estimate of the dominant pattern, not a legal mandate.
Misconception two: "Adoption means the animal was saved from testing." Adoption happens after testing ends. The animal was tested; it simply was not killed. The distinction matters for how you read "rescued beagle" stories.
Misconception three: "Euthanasia is unregulated." In the US, IACUC oversight and AVMA guidelines govern method and timing. A 2019 NIH workshop specifically examined reducing pain at the endpoint, which is an active welfare issue rather than a settled one.
The non-obvious detail is a numbers paradox. The species most protected by law (dogs, cats, primates) make up a few percent of animals used but a large share of public attention. The species least protected by law (mice, rats) make up roughly 90 percent of use and almost 100 percent of euthanasia volume. Reform efforts aimed at rehoming dogs therefore touch a small slice of the total.
Another caveat: "reuse" is not always humane. Reusing an animal in a second study can extend cumulative stress and exposure. Some IACUCs cap the number of lifetime protocols per animal for this reason.
What to Actually Do If This Concerns You
- Ask institutions for their adoption and retirement policy, not just their general welfare statement. Specifics reveal actual practice.
- Support sanctuary and rehoming organizations that accept research animals, since placement capacity is the binding constraint.
- Advocate for the species that get least attention: rodent welfare standards and fish oversight are where the volume is.
- Prioritize validated non-animal methods, such as organ-on-chip and in silico models, as replacements where the science permits.
- Read AWA reports alongside independent estimates, because official counts exclude most animals.
- Push for endpoint-reduction practices, which lower suffering without abandoning study validity.
Limitations and Open Questions
The euthanasia rate is estimated, not audited. No national system counts post-study outcomes for rodents, and rehoming statistics are facility-reported. The 90–95 percent figure should be read as a well-supported estimate, not a verified census.
Rehoming policy is uneven across states, countries, and institutions. The EU's Directive 2010/63 includes a rehoming preference for animals that can be placed, but implementation varies widely. Cross-country comparisons are therefore rough.
Whether rehoming always benefits the animal is an open welfare question. A laboratory beagle placed in a home may face adjustment stress, and follow-up data on post-placement outcomes are thin. The honest position is that rehoming reduces killing but does not automatically equal good welfare. That distinction, more than any single number, is what the phrase "are animals killed after animal testing" really turns on.