Ivermectin is one of the most widely used medicines in history, given billions of times to people and animals to treat parasitic worms and external parasites. This article explains three things that belong together: why ivermectin is so effective against worms, insects and mites; why it is generally safe for mammals, including people, at approved doses; and what a genetic quirk in collies and related herding dogs reveals about where that safety comes from. The short answer is that ivermectin's main target is a type of nerve channel that mammals do not have, and the mammalian brain is protected by a molecular pump that actively expels the drug. When that pump is missing, the margin of safety shrinks dramatically.

From a Japanese soil sample to a Nobel Prize

Ivermectin's story begins with a soil bacterium. In the 1970s, Satoshi Ōmura of the Kitasato Institute in Tokyo collected soil near a golf course on the Izu peninsula and cultured an organism later named Streptomyces avermitilis. Working with William Campbell and colleagues at Merck in the United States, researchers found that the bacterium produced a family of compounds, the avermectins, with remarkable activity against parasitic worms. The discovery was published in Antimicrobial Agents and Chemotherapy in 1979. Ivermectin is a slightly modified version of avermectin B1, chemically altered to improve its safety.

Ivermectin entered veterinary use in 1981 and was later approved for human use against river blindness (onchocerciasis). In 1987 its manufacturer began donating it for river blindness control programmes, which have since distributed it on a vast scale across Africa and Latin America. Ōmura and Campbell shared the 2015 Nobel Prize in Physiology or Medicine for the work. For readers of faith, there is something worth pausing over here: a single microbe in an ordinary patch of earth turned out to carry a compound that has relieved the suffering of millions of the world's poorest people. Careful science found it, but it was there to be found.

The target: glutamate-gated chloride channels in invertebrates

Nerve and muscle cells are controlled by ion channels, tiny gated pores that let charged particles flow across the cell membrane. In worms and arthropods (insects, mites, lice), one important family is the glutamate-gated chloride channel, or GluCl. When these channels open, negatively charged chloride ions flow into the cell, making it less excitable. Normally this is a brief, controlled signal.

Ivermectin binds to GluCls and locks them open. The result is a sustained flood of chloride, which effectively silences the affected nerve and muscle cells. In nematode worms, this paralyses the pharyngeal muscles the worm uses to feed and the body muscles it uses to move. The evidence for this mechanism is strong and comes from several independent lines:

Vertebrates, including humans, do not have glutamate-gated chloride channels. In our nervous system glutamate is mainly an excitatory messenger acting on different receptors. This is the first layer of ivermectin's selectivity: its primary target simply does not exist in us.

Why mammals are largely spared: the P-glycoprotein gatekeeper

Selectivity is not absolute. At higher concentrations, ivermectin can also act on related mammalian channels, particularly GABA-A receptors and glycine receptors, which are inhibitory chloride channels in the brain and spinal cord. If enough ivermectin reached the central nervous system, it could depress brain function, causing sedation, loss of coordination, tremor and, in severe cases, coma.

The reason this rarely happens is the blood-brain barrier. The capillaries supplying the brain are lined with tightly joined cells that restrict what passes from blood into brain tissue. Embedded in the blood-facing membrane of these cells is a protein called P-glycoprotein, encoded by the gene ABCB1 (historically called MDR1, for "multidrug resistance," because it was first studied in cancer cells that pumped out chemotherapy drugs). P-glycoprotein uses cellular energy to grab a wide range of fat-soluble molecules and push them back into the bloodstream before they can reach the brain. Ivermectin is one of its substrates.

The decisive evidence came, partly by accident, from the Netherlands Cancer Institute. In 1994, Alfred Schinkel and colleagues reported in Cell that mice engineered to lack the main brain form of P-glycoprotein (the mdr1a gene) appeared normal until their colony was treated with ivermectin for a mite infestation, a routine procedure in animal facilities. The knockout mice developed neurological signs and many died. Follow-up experiments showed these mice were on the order of 100 times more sensitive to ivermectin than normal mice, with brain concentrations of the drug many times higher. A naturally occurring P-glycoprotein deficiency in a laboratory mouse strain (CF-1), reported by Merck scientists in the late 1990s, showed the same pattern with a related avermectin.

So mammalian safety rests on two layers: the absence of the main target, and an active pump guarding the organ where secondary targets live. It is a fine example of how the body is built with protective redundancy.

What collies taught us: the MDR1 mutation

Veterinarians had noticed since the 1980s that collies could suffer severe, sometimes fatal, neurological reactions to ivermectin at doses that other dogs tolerated easily. The explanation came from Katrina Mealey and colleagues at Washington State University, who reported in Pharmacogenetics in 2001 that ivermectin-sensitive collies carried a four-base-pair deletion in the MDR1 gene. This deletion scrambles the genetic code and produces a severely truncated, non-functional P-glycoprotein. In effect, these dogs are a natural version of the Dutch knockout mice.

Later work, including a 2004 PNAS study by Mark Neff and colleagues, traced the mutation to a single ancestral dog that likely lived in Britain in the 1800s, before modern breed registries. It has since been found in a number of herding breeds and their relatives, including:

The numbers make the point vividly. Dogs with two copies of the mutation may show toxicity at around 100 micrograms per kilogram, whereas dogs with normal P-glycoprotein typically tolerate many times that. Signs include dilated pupils, drooling, wobbliness, tremors, apparent blindness, disorientation and coma. Importantly, the very low monthly doses used to prevent heartworm (around 6 micrograms per kilogram) are considered safe even in affected dogs. The danger lies mainly in higher doses used for mange or, commonly, accidental ingestion of horse or livestock paste. Affected dogs are also more sensitive to several other drugs that P-glycoprotein normally pumps out, including the antidiarrhoeal loperamide and certain chemotherapy agents. A simple cheek-swab genetic test now allows owners and veterinarians to know a dog's status in advance, a sensible step for any responsible owner of a herding breed.

What this means for human safety

In humans, ivermectin is approved by the US Food and Drug Administration in oral form for strongyloidiasis and onchocerciasis, and in topical forms for head lice and rosacea. Standard oral doses are around 150 to 200 micrograms per kilogram, usually given once. At these doses, its safety record across decades of mass treatment is very strong. A dose-escalation study by Guzzo and colleagues, published in the Journal of Clinical Pharmacology in 2002, gave healthy volunteers doses up to about ten times the highest approved dose without evidence of central nervous system toxicity.

The collie story, however, explains where caution belongs:

A separate and well-documented risk deserves mention because it is often confused with direct neurotoxicity. In people heavily infected with the eye worm Loa loa, found in parts of Central Africa, ivermectin can trigger a serious encephalopathy. Research in Cameroon, published in The Lancet in 1997 by Gardon and colleagues, linked this risk to very high blood levels of the parasite's larvae. The likely mechanism is a reaction to dying parasites rather than failure of P-glycoprotein, and it shapes how treatment programmes screen patients in those regions.

Stewardship, informed choice and the household

The science here supports informed confidence rather than either fear or carelessness. Ivermectin is a genuinely valuable medicine with a well-understood reason for its safety. That understanding also tells families what to watch for: keep livestock and pet formulations stored apart from human medicines, know whether a herding-breed dog carries the MDR1 mutation, and give your own physician a complete picture of other medicines, pregnancy and children's weights before any treatment. Medical freedom is best exercised by patients who understand how a drug actually works, working alongside a doctor who knows them.

Key takeaway: Ivermectin kills parasites by locking open chloride channels that mammals lack, and the mammalian brain is further shielded by the P-glycoprotein pump, so its safety depends on that pump working, as collies with the MDR1 mutation vividly demonstrate.