Ivermectin has protected hundreds of millions of people from river blindness. It is one of the most consequential medicines of the past half-century. In parts of Central Africa, however, the same dose that helps one person can seriously harm another. People who carry very large numbers of a second parasite, Loa loa, can develop a rare but potentially fatal brain disorder after taking it. This article explains why that happens and what the research shows about the risk threshold. It also describes how a simple blood measurement, paired with a strategy called "test-and-not-treat," has allowed health workers to treat nearly everyone safely while protecting the few who would be endangered.

Two worms, one medicine, and a dangerous overlap

Onchocerciasis, or river blindness, is caused by the filarial worm Onchocerca volvulus and spread by blackflies that breed near fast-flowing rivers. Adult worms live in nodules under the skin. They release millions of tiny larvae, called microfilariae, which migrate through the skin and eyes and cause intense itching, skin disease and, eventually, blindness. Recent global estimates put the number of infected people at roughly 20 million, more than a million of whom have some vision loss. Over 99 percent of cases are in Africa.

Ivermectin descends from avermectin, a compound produced by a soil bacterium, Streptomyces avermitilis. Satoshi Ōmura of Japan's Kitasato Institute isolated the organism from a soil sample, and William Campbell's team at Merck developed the drug. The two shared the 2015 Nobel Prize in Physiology or Medicine. It is a remarkable thing that a microbe in ordinary earth should yield a medicine that restores so much to so many. Since 1987 the manufacturer has donated ivermectin for onchocerciasis "for as long as needed." Ivermectin is approved for treating onchocerciasis, and it is given once or twice a year to whole communities.

The complication is geographic. Loa loa, the "African eye worm," is spread by deer flies of the genus Chrysops in the rainforests of Cameroon, Gabon, the Republic of Congo, the Democratic Republic of Congo, the Central African Republic and neighbouring areas. Many of these places also have river blindness. Millions of people carry Loa. Some of them have extraordinary numbers of its microfilariae circulating in their blood.

What post-ivermectin encephalopathy is and why it happens

Cases began to appear in Cameroon in the late 1980s and early 1990s as ivermectin distribution expanded. Previously healthy adults became severely ill within a day or two of a routine dose. Researchers including Michel Boussinesq and Jacques Gardon of France's Institut de Recherche pour le Développement, working with Cameroonian colleagues, documented a consistent pattern:

The leading explanation is mechanical and inflammatory. Ivermectin paralyses and kills microfilariae quickly. When tens of thousands of larvae per millilitre of blood die at once, they can lodge in the small vessels of the brain and retina and set off local inflammation. Autopsy findings and the retinal haemorrhages seen in patients support this view, although the full mechanism is not settled. Loa itself is not especially susceptible to ivermectin. The problem is the sheer burden of larvae cleared in a short time.

The outcome depends heavily on care. Many patients who received attentive nursing, meaning hydration, feeding, turning to prevent pressure sores and protection of the airway, recovered over weeks to months, often without lasting deficit. Deaths did occur, sometimes from complications of prolonged immobility rather than the brain injury itself. Some survivors were left with permanent impairment. Reviews of reports gathered by the donation programme confirmed that these events were rare relative to the hundreds of millions of doses given. Each one, however, was a person who had come forward in good faith to receive a medicine.

The number that matters: microfilarial density

The decisive insight came from careful measurement. In a study published in The Lancet in 1997, Gardon and colleagues followed people treated with ivermectin in the Lekié area of central Cameroon. Blood had been sampled before treatment, so the researchers could relate each person's Loa microfilarial density to how they fared afterwards. The risk of reactions rose steeply with parasite load. Marked functional impairment became notably more likely above roughly 8,000 microfilariae per millilitre. The risk of serious neurological reactions was concentrated among people above about 30,000 per millilitre.

This was a cohort study, not a randomised trial, and the number of serious events was small. Its central finding is nonetheless biologically coherent and has been consistent across later work. Two further features of Loa matter:

Programmes first responded at the community level. A questionnaire-based method called RAPLOA, developed with the World Health Organization's tropical disease research programme and published in the early 2000s, estimates how intense Loa transmission is from the proportion of people who recall seeing a worm crossing their eye. High-risk communities could then receive extra surveillance and prepared medical teams. That is useful, but it cannot tell which individual is at risk. In many areas where river blindness was only mild, the balance of risk led programmes to withhold mass ivermectin altogether. This left whole populations without protection and stalled elimination efforts.

From microscope to phone: test-and-not-treat

The obvious solution was to measure each person's microfilarial density before dosing. Traditional thick blood smears require trained microscopists and hours of work, which is impractical in village settings. A collaboration among Daniel Fletcher's bioengineering group at the University of California, Berkeley, Thomas Nutman's laboratory at the US National Institute of Allergy and Infectious Diseases, the Institut de Recherche pour le Développement and Joseph Kamgno's team at the Centre for Research on Filariasis and other Tropical Diseases in Cameroon produced a different approach. They built a smartphone-based video microscope, the LoaScope.

Described in Science Translational Medicine in 2015, the device takes a fingerprick of blood in a small capillary. It films the sample and uses software to count microfilariae by detecting their wriggling motion. A result comes back in a few minutes. In field validation its counts agreed well with conventional thick smears, particularly at the high densities that matter for safety.

The strategy was tested in the Okola health district of Cameroon, and the results were published in the New England Journal of Medicine in 2017 by Kamgno and colleagues. Health workers tested about 16,000 residents. Anyone with a Loa count at or above 20,000 microfilariae per millilitre was not given ivermectin. That cut-off was deliberately set below the 30,000 level to leave a safety margin. Roughly 2 percent of those tested were deferred. Everyone else was treated, and no serious adverse events occurred. Mild reactions such as itching and aches were reported and resolved, as is typical after ivermectin.

Because this was a single-arm implementation study and serious events are rare, it cannot prove the risk is zero. The design was nonetheless sound and the population large, and the result matched what the earlier density data predicted. Subsequent work in other Cameroonian districts has reported similar safety findings. Researchers continue to study cost, logistics and how well the approach holds up across repeated annual rounds. The underlying principle is simple and morally attractive. Instead of accepting a small statistical loss across a population, or denying everyone a beneficial medicine, the programme identifies the particular person at risk and protects them.

Caring for those who are deferred

Being excluded from ivermectin does not mean being abandoned. Several approaches exist or are under study, with differing levels of evidence:

Diethylcarbamazine, which does kill Loa, carries its own risk of encephalopathy at high densities. It is also unsuitable in onchocerciasis because it can provoke severe skin and eye reactions. It is therefore no shortcut. Untreated loiasis is not harmless either. A retrospective cohort study from Cameroon published in The Lancet Infectious Diseases in 2017 linked high Loa microfilarial densities with increased mortality. That finding argues for taking these patients' own health seriously, not merely setting them aside.

What this means for patients and families

The lessons extend beyond African treatment campaigns. Anyone who has lived in, worked in or travelled for long periods in the rainforest regions of Central Africa should tell their physician this before taking ivermectin for any reason. Tropical medicine guidance generally advises checking for Loa microfilariae in such patients beforehand. Families who have emigrated, missionaries, aid workers and their children are among those for whom this history matters.

Test-and-not-treat also models informed consent at its best. People are told what their blood shows and why a decision is being made. A clear, individual result replaces a blanket rule. This is medicine practised as stewardship, attentive to the person in front of the clinician, and it requires no choice between helping the many and protecting the few.

Key takeaway: Measuring each person's Loa loa microfilarial density before giving ivermectin, and deferring the small minority with very high counts, has allowed river blindness treatment to reach communities that were once considered too dangerous to treat, without reported serious adverse events in the landmark field study.