In our 24-hour media landscape, incomplete science can make its way into the media and the public consciousness long before scientists have definitive answers. Consider ivermectin.

Proponents call it a miracle cure, while detractors say it’s nothing more than a horse and cow dewormer. To learn what this drug really is and what it can-and can’t-do, Bottom Line Health spoke with F. Perry Wilson, MD, MSCE, associate professor of medicine and director of the clinical and translational research accelerator at Yale University School of Medicine.

What is ivermectin and what are its approved indications?

Ivermectin is an antiparasitic agent used in mammals, including horses, cows, and humans. You may already give it to your dog as a heartworm preventive medication. In people, the oral form is approved by the U.S. Food and Drug Administration (FDA) for strongyloidiasis and river blindness, both of which are caused by internal parasites. A topical form is approved for scabies and rosacea. When used appropriately, it is safe and effective.

Why is the FDA now issuing safety warnings about a drug it approved 50 years ago?

The FDA’s mission is to evaluate the safety and efficacy of drugs or medications for specific indications. All drugs have side effects, and the FDA has to weigh those side effects against the benefits of the drug. For parasitic diseases, the benefits of ivermectin outweigh the risks. Approval for one disease, however, doesn’t translate to approval for another. Because ivermectin has not yet been proven to treat COVID-19, there are no known benefits to offset any risks.

Further, people are now taking veterinary formulations of ivermectin, which are more concentrated, at doses intended for large animals. Overdose of the drug can lead to nausea, vomiting, diarrhea, allergic reactions, dizziness, problems with balance, seizures, coma, facial or limb swelling, neurologic adverse events, a sudden drop in blood pressure, a severe skin rash potentially requiring hospitalization, liver injury, and possible death.

Even at human-level doses, ivermectin can interact with a wide variety of medications, so it should only be prescribed by a physician, not purchased over-the-counter from a feed or farm supply store.

A study published in Antiviral Research reported that ivermectin destroyed 99.98 percent of SARS-CoV-2 (the virus that causes COVID-19) in a test tube. Is that proof that it works?

No, this kind of early test-tube study is only the first of many steps in the drug-development process. Scientists first work in test tubes to get a sense of what might work. In a test tube, bleach would kill COVID-19, but that doesn’t make it a safe or appropriate medication in humans. This study looked at ivermectin’s ability to reduce the amount of the virus that causes COVID-19. In one sense, it was successful: If you put enough ivermectin into the test tube, it will stop the virus from replicating. However, the amount needed to have that effect was about 100 times the amount you’d have in your blood if you took a safe dose of ivermectin. It’s not an achievable concentration in a human: It would be lethal.

A meta-analysis that combined data from 24 randomized trials concluded that ivermectin reduced mortality from COVID-19. What are the scientific community’s concerns with that meta-analysis?

That review included three types of studies: unpublished, preprint, and peer reviewed:

Unpublished data is data that no one has seen. You’re essentially taking the word of the authors that their results are accurate.

Preprints are published papers that are not reviewed by other scientists before publication. That means no one checks the data for errors. Preprints are a useful way to quickly disseminate study results, but they can contain errors that a review would have caught.

Peer-reviewed papers are the most reliable because an outside scientist deeply reads the information and asks pointed questions of the author to clarify the data and findings. A peer review can catch simple errors, like a mistyped number, or more serious problems, like poor study design, incorrect conclusions, or fraudulent data.

In the ivermectin meta-analysis, two of the studies that really drove the positive results had significant issues with their randomization that would have been caught had they been peer reviewed. In a randomized trial, you expect to see similar characteristics like age and sex in each group of people in the study. If group A has 50 women and 50 men, you would expect group B to have about the same. It’s like a coin toss: If you toss a coin 100 times, you’ll get about 50 heads and 50 tails.

For the two studies in question, there were significant irregularities in the randomization. For the first, a trial of patients hospitalized with COVID-19 in Iran, 29 percent of the study participants tested negative (PCR-negative) for the virus. But in the group of patients who took ivermectin, the negative rate was 20 percent, while in the control groups, it jumped to 47 percent. If you flipped a coin 5,000 times, you’d only get such a large discrepancy once, throwing the validity of the study results into question.

Another study from the meta- analysis, the Elgazzar trial from Egypt, compared ivermectin with hydroxychloroquine and reported that ivermectin dramatically reduced deaths. Because hydroxychloroquine is another type of treatment, this study had no control group. Further, there were multiple significant and extremely statistically improbable imbalances in the allegedly randomized groups. That wasn’t the only problem with this study: It later came to light that the report was rife with plagiarism and fraudulently manipulated data. The paper was retracted. When those two studies are removed from the meta-analysis, there is no evidence of a benefit to using ivermectin for COVID-19.

Does that mean ivermectin doesn’t work?

No. At this time, we don’t have any evidence that it does work, but there’s an important difference in saying, “we don’t have any evidence” and “this drug doesn’t work.” We can safely assume that ivermectin is not a miracle cure because, if it were, the smaller studies would have shown a bigger effect, but we need more studies to learn if it has at least a modest effect. There are several such studies going on right now, including the Activ-6, with 15,000 participants, and another from researchers at the University of Minnesota.

How can people become better consumers of medical information?

Because of the jargon in medical studies, there’s a lot of room for things to be taken out of context. People can cherry-pick findings and apply motivated reasoning: That’s when you have a conclusion that you believe is true, so you search for the data that supports your belief and ignore the data that doesn’t. We all do this, including scientists, so we need to watch out for it.

I’m really interested in teaching people how to appropriately read the medical literature. You don’t need a medical degree, but I do think you need a little bit of training on how study design works. That’s why I developed a free online course called “Understanding Medical Research: Your Facebook Friend is Wrong.” You can find it at

Related Articles