Ivermectin
[Note: Since this newsletter was posted, on 9/9/2021, the evidence has become clearer that ivermectin has no benefits for the prevention or treatment of COVID-19. See, for example, here and here.]
Ivermectin. If you have a parasitic infestation such as head lice or scabies, your doctor might give you a prescription for it. And, you would take the medicine, because knowing that you have parasites on your body is, at minimum, really creepy. (No pun intended.)
In this scenario, neither your political beliefs nor those of your doctor would come into play. In contrast, the use of ivermectin to treat COVID-19 is, like everything else about the pandemic, extremely politicized. Support for ivermectin comes almost exclusively from the right. Opposition to it can be seen across the political spectrum, albeit concentrated on the left.
The purpose of this newsletter is to discuss research on ivermectin's effectiveness as a COVID-19 treatment. I'll be as impartial as I can concerning the politics, though my liberal biases will be evident.
When I started writing this newsletter, here's what I thought I was going to conclude: Ivermectin doesn't help COVID-19 patients, off-label use of the drug is dangerous, and those who promote its use are ignorant or mendacious.
My conclusion turns out to be more nuanced than that. Although studies favoring ivermectin as a COVID-19 treatment are weak at best, I will argue that it's too soon to dismiss the drug as ineffective. I will also complain about how both liberals and conservatives have been discussing it. Ivermectin deserves – and is receiving – careful scientific scrutiny. It's not a miracle drug. It's not bleach either.
Broadly, there are three perspectives on the use of ivermectin to treat COVID-19:
Perspective #1: Ivermectin reduces COVID-19 symptoms and negative outcomes (e.g., death).
Perspective #2: Ivermectin has no impact on COVID-19 symptoms and outcomes.
Perspective #3: We don't have enough data yet to say definitively whether ivermectin reduces COVID-19 symptoms and negative outcomes.
If you lean left, politically speaking, you probably favor perspective #2, because authoritative groups, such as the FDA and the WHO, recommend that ivermectin not be used to treat COVID-19, except in clinical trials, and this recommendation has been amplified in liberal media outlets. There, ivermectin is often dismissed as snake oil (or “horse paste”, as Nobel prizewinner Paul Krugman referred to it last week), and support for ivermectin as a COVID-19 treatment has been linked to willful spread of misinformation by the far right.
The problem with dismissing ivermectin as horse paste is that more than 10 peer-reviewed studies claim to show that it benefits COVID-19 patients.
The methodologies underlying these studies (as well as those that report no benefits) are highly problematic, as outlined by the National Institutes of Health (NIH) back in February, as well as by a variety of independent researchers, including a team who published a review of randomized controlled trials in July. Problems include small sample sizes (often less than 100), nonrandom assignment of patients to ivermectin versus comparison groups, sloppy medical procedures, poorly defined variables, and so on. The consensus view is that no large-scale, well-designed study on treating COVID-19 with ivermectin has been published yet. Such studies are currently in progress. Meanwhile, the NIH, the WHO, and many independent researchers assert that current evidence is not sufficient to draw any conclusions about ivermectin as a COVID-19 treatment. The NIH (the largest funder of biomedical and health research in the world, and home to the world’s largest biomedical research institution) is an interesting example, because it formerly took perspective #2 and recommended against off-label use of ivermectin. Since February 2021, however, the official NIH position is that there's “insufficient evidence...to recommend either for or against the use of ivermectin for the treatment of COVID-19”.
In other words, in contrast to what one often sees in the liberal press, the mainstream research community mostly represents perspective #3 – i.e., we can't say one way or the other yet whether ivermectin benefits COVID-19 patients.
After looking at the studies I've come to agree with this perspective. Although none of these studies seems adequate, methodologically speaking, here's why I wouldn't entirely dismiss the drug:
1. Some studies show that ivermectin helps with COVID-19, some don't. But comparison groups almost never fare better than ivermectin groups. In other words, when significant findings are reported, they more commonly favor ivermectin than whatever treatment or placebo it’s compared to. Otherwise, group differences tend to be nonsignificant rather than favoring the comparison group.
2. Among nonsignificant findings, ivermectin more often than not outperforms whatever it's compared to. Although the rules of the game (i.e., the inferential statistics game) say that nonsignificant findings don't count, researchers choose an arbitrary cut-off point for what is versus isn't significant. A large number of consistent findings that are close to that cut-off point constitute a hint.
3. The NIH and at least some independent researchers seem to be slightly biased in their complaints about pro-ivermectin studies. Although I agree with most of these complaints, a few of them seem excessive. Here's an example: A 2020 study on 72 COVID-19 patients in Bangladesh found that compared to a placebo, ivermectin significantly reduced the time to virological clearance – i.e., how many days passed until no more SARS-CoV-2 virus could be detected by means of a PCR swab test. Virological clearance took 9.7 days on average for the ivermectin group versus 12.7 days for the placebo group. However, the NIH faulted the study, stating that PCR results are “not an appropriate surrogate for clinical efficacy” and concluding that the study tells us nothing about the clinical efficacy of ivermectin. I'm not a biologist, but I find this troubling. The NIH's concerns are that (a) the PCR test isn’t 100% accurate, and that (b) the duration and severity of COVID-19 symptoms are somewhat independent of how long the virus remains in one's body. (Ideally, the efficacy of a drug should be measured in terms of symptom reduction and health outcomes rather than just viral presence.) Fine. But the study shows that ivermectin kills the virus more rapidly than a placebo does. Inaccuracies in PCR test data should be seen at the individual level rather than systematically affecting the results for either group (see here). Thus, even if this study doesn't reliably inform us about clinicial efficacy, surely it tells us something about the potential benefits of ivermectin.
I've provided three reasons why I wouldn't dismiss the possibility that ivermectin can help COVID-19 patients. In short, I opt for perspective #3 (“We don't have enough data yet to say definitively whether ivermectin impacts COVID-19 symptoms and outcomes.”) Meanwhile, conservative and liberal commentators have each taken a different path away from this perspective.
Some conservatives push perspective #1 (“Ivermectin reduces COVID-19 symptoms and negative outcomes.”). Most of them do so without a clear presentation of evidence. To my knowledge, the only systematic attempt by the right to be scientific is a 4-page summary of evidence published in late August by the Front Line COVID-19 Critical Care Alliance (FLCCC). Unfortunately, this document merely summarizes some of the methodologically limited studies I mentioned earlier, but with few details about methodology, no critical evaluation, and almost no acknowledgment of the studies that don't find benefits of ivermectin. The FLCCC also refers to a meta-analysis that's now being corrected because the largest study reviewed was retracted in July owing to plagiarism and outrageously faulty data (e.g., over a third of the participants listed as COVID-19 mortalities had died even before the study began; further details here).
In remarks to the U.S. Senate on December 8, 2020, the founder and co-president of the FLCC, Pierre Kory, called ivermectin a “wonder drug” with “miraculous effectiveness” in the treatment of COVID-19. These are deeply irresponsible statements. Even if you ignored the studies that show no benefits of ivermectin, and you ignored the methodological limitations of all the pro-ivermectin studies and took their results at face value, the effects are consistently small. The most dramatic example I could find was a 2020 study in which COVID-19 patients receiving ivermectin recovered in 10.1 days on average, vs. 17.9 days for those receiving standard therapy. I realize that if you're a patient, that's not a small difference, but it doesn't seem miraculous (and, again, you have to overlook some obvious flaws, like small sample size, and failure to adjust the data for patient mortality). Other pro-ivermectin studies show much smaller effects.
In contrast to conservatives, liberals often embrace perspective #2 (“Ivermectin has no impact on COVID-19 symptoms and outcomes...”). There's a subtle but important problem with this perspective. To illustrate, compare the following statements:
Statement 1: “There's no clear evidence that X is helpful.”
Statement 2: “X is not helpful.”
In some cases, the first statement implies the second one. If you've spent decades carefully studying X, and you haven't found clear evidence that it's helpful, you should probably conclude that it isn't. (Or, to put it less crudely: The longer and more carefully you've studied whether X is helpful, the more confidently you can infer Statement 2 from Statement 1.)
In other cases, the first statement does not imply the second one. If you began to study whether X is helpful only 5 minutes ago, and you haven't found anything yet, you shouldn't infer that it's not helpful. Statement 1 will remain true until you obtain data that either support or contradict Statement 2. And, if you find that X helps, but the evidence is weak (and other evidence suggests that it's not helpful) you should stick with Statement 1 and wait for additional data before concluding anything.
What's tricky here is that scientists don't have clear rules for when to infer Statement 2 from Statement 1. At some point, if there's enough evidence against X being helpful (and there's enough reason to doubt any evidence that it is) we conclude that X isn't helpful. But what counts as “enough” is a subjective call, and not everyone will agree on when it’s achieved.
With respect to ivermectin, liberal commentators seem to assume that there's already “enough” evidence to infer Statement 2, even though scientists are awaiting the results of more than 30 well-designed studies. For example, in a New York Times story, updated this Tuesday, you can read that ivermectin “has not been shown to be effective in treating Covid”. Although it's true that ivermectin hasn't been shown to be effective, it hasn't been shown to be ineffective either. Perhaps the journalist meant to imply this, but the article only mentions studies that fail to demonstrate effectiveness, not those that do.
Part of the problem may simply be that experts and authoritative organizations don't always use the clearest possible language. For example, the FDA's position is presented on a web page entitled “Why you should not use ivermectin to treat or prevent COVID-19”, where it’s noted that “Currently available data do not show ivermectin is effective against COVID-19.” This is like Statement 1. However, as I keep stressing, saying that the data “do not show ivermectin is effective” is different from saying that the data “show that ivermectin is not effective.” I think the FDA should spell out what’s meant more explicitly. The NIH does a better job of this (“insufficient evidence...to recommend either for or against the use of ivermectin for the treatment of COVID-19”), as does the WHO (“The current evidence on the use of ivermectin to treat COVID-19 patients is inconclusive.”).
Bottom line? More research is needed before we can be reasonably certain that ivermectin is or isn't beneficial for COVID-19 patients. At the moment, there are hints of small benefits, but the evidence is far from reliable. (The only thing we're clear about is that taking too much ivermectin is bad for you, as shown by data from poison control centers.) So, feel free to ignore those who insist it's a miracle cure – and those who dismiss it as horse paste.
Thanks for reading!