Animal testing involves exposing live animals to chemicals, ingredients, or finished products to measure safety before those products reach humans. The methods range from applying a substance to an animal’s skin or eyes and watching for irritation, to feeding animals increasing doses of a chemical to find the amount that becomes lethal. These tests are used across industries including pharmaceuticals, pesticides, industrial chemicals, and (in some countries) cosmetics.
Skin and Eye Irritation Tests
One of the most well-known animal tests is the Draize test, developed in the 1940s and still in use today. It measures how irritating a substance is to skin or eyes. New Zealand White rabbits are the standard species for this test, chosen because their large eyes make observation easier and because they’re relatively inexpensive.
In the eye version, a small amount of the test substance (0.1 mL of liquid or 0.1 g of solid) is placed directly onto the cornea and inner eyelid of one eye of a conscious, restrained rabbit. The other eye is left untreated as a comparison. The rabbit is not given anesthesia. Researchers then check for redness, swelling, cloudiness, bleeding, and discharge at intervals of 1, 2, 24, 48, and 72 hours after exposure, and sometimes for up to 21 days. Each sign of damage to the cornea, iris, and surrounding tissue is scored on a numerical scale, with a maximum possible score of 110. Up to six rabbits may be used per substance, though testing stops at three if severe damage occurs early.
The skin version follows a similar logic: a substance is applied to a shaved patch of skin, and researchers score the resulting redness, swelling, and tissue damage over several days.
Lethal Dose Testing
Acute toxicity tests determine how much of a substance it takes to cause serious harm or death. The most recognized version is the LD50 test, which stands for “lethal dose, 50%,” the dose that kills half the test animals. While older versions of this test used large groups of animals, the current method approved by the Organisation for Economic Co-operation and Development (OECD) uses fewer.
In the updated procedure, animals (typically rats) are fasted and then given a single oral dose of the substance, usually through a tube inserted into the stomach. A single animal is dosed first, at a level estimated to be near the lethal threshold. If that animal survives, the next animal receives a higher dose. If it dies, the next receives a lower one. Animals are dosed one at a time, usually 48 hours apart, and observed closely for the first four hours, then daily for 14 days total. At the end of the study, the animals are killed and examined internally for organ damage. The lethal dose is then calculated statistically.
Repeated Exposure and Long-Term Studies
Many products people use daily, from medications to pesticides, involve repeated contact over months or years. To simulate this, animals are given the same substance over extended periods. These repeated-dose studies come in several tiers. Short-term studies last up to 14 days. Subchronic studies run up to three months. Chronic studies extend beyond three months and can last up to two years, essentially the lifespan of a lab rodent.
Rats are commonly used for six-month chronic studies, while dogs are used for nine- to twelve-month versions. Throughout these studies, researchers monitor body weight, food intake, blood chemistry, organ function, and behavior. At the end, animals are euthanized and their organs are examined under a microscope to identify which tissues were damaged. These studies reveal whether a substance accumulates in the body, which organs it targets, and whether it causes cancer or reproductive problems.
Routes of Exposure
Testing doesn’t always mean feeding a substance to an animal. Regulators require safety data across multiple exposure routes that mirror how humans might encounter the product. For pesticide registration in the United States, the EPA requires at least six categories of acute toxicity data:
- Oral: the substance is swallowed, typically by rats
- Dermal: applied to shaved skin to test absorption and irritation
- Inhalation: animals breathe in the substance as a vapor or aerosol
- Eye irritation: applied directly to rabbit eyes
- Skin irritation: applied to skin and observed for reaction
- Skin sensitization: tests whether repeated skin contact triggers an allergic response
Beyond acute effects, regulators also require 90-day feeding studies in rodents, lifetime cancer studies in both rats and mice, and reproductive toxicity studies that track effects across multiple generations of animals.
Which Animals Are Used
Rats and mice make up the vast majority of laboratory animals worldwide, but they fall outside USDA reporting requirements in the United States, which means exact numbers are difficult to pin down. Among the species the USDA does track, fiscal year 2024 data show roughly 106,400 rabbits and about 60,400 nonhuman primates were used in research facilities across the country.
Rabbits are the default species for skin and eye irritation studies. Rats and mice dominate toxicity, cancer, and chronic exposure research. Dogs, particularly beagles, are used in longer-term repeated-dose studies. Nonhuman primates are reserved primarily for pharmaceutical development. A typical development program for an antibody-based drug uses around 144 nonhuman primates across all required safety studies.
Oversight and Approval
In the United States, every institution that conducts animal testing must have an Institutional Animal Care and Use Committee, or IACUC. This committee reviews and approves every proposed experiment before any work with animals can begin. Approval requires either a full committee vote with a quorum present or a designated member review. The committee evaluates whether the research justifies the use of animals, whether pain and distress are minimized, and whether fewer animals or alternative methods could achieve the same result.
Protocols that carry a higher risk of causing pain or distress receive additional scrutiny. The USDA categorizes animal use into tiers: procedures with no pain, procedures where pain is managed with anesthesia or painkillers, and procedures where pain is not minimized because doing so would compromise the experiment. That last category is the smallest but most controversial. In fiscal year 2024, roughly 2,600 rabbits and about 1,200 nonhuman primates fell into that category.
Where Animal Testing Is Banned
The European Union has the most comprehensive ban on animal testing for cosmetics in the world. The ban rolled out in phases: testing finished cosmetic products on animals became illegal in September 2004, and testing individual cosmetic ingredients became illegal in March 2009. A final marketing ban took effect in March 2013, making it illegal to sell any cosmetic product in the EU if it or its ingredients were tested on animals, regardless of where the testing took place. This applies even when no non-animal alternative exists.
The ban covers cosmetics only. Pharmaceuticals, pesticides, and industrial chemicals sold in the EU still require animal testing data for regulatory approval. Several other countries and regions have adopted similar cosmetics bans, but in much of the world, including China (for certain imported products), animal testing for cosmetics remains either required or permitted.
Non-Animal Alternatives Gaining Ground
Validated alternatives are gradually replacing some animal tests. Three-dimensional reconstructed human skin and cornea models, grown from human cells in a lab, can now assess skin irritation and corrosion without animals. These tissue models are accepted under OECD testing guidelines, giving them regulatory weight in multiple countries. Computer modeling can predict how a substance moves through the human body, and cell-based assays can screen for specific toxic effects like hormone disruption.
The FDA has signaled openness to reducing animal requirements for drug development. For antibody-based drugs, the agency has outlined scenarios where a single animal species study combined with human cell-based data could replace the traditional two-species approach. In cases where the drug target exists only in humans, a battery of lab-grown tissue tests and computer models could potentially substitute for animal studies entirely. These shifts are still emerging, and most new drugs, pesticides, and chemicals continue to require animal data before reaching the market.