The drug Ivemectin has been used by billions of people for decades and thus has better safety data than almost any other medical intervention. During the last year, several small trials, mostly at frontline hospitals in poor countries have shown it to be effective in reducing deaths from Covid. But when I posted this video where Dr. Tess Lawrie discusses the evidence (meta-analysis) about Ivermectin for reducing Covid deaths, Facebook tagged it as partly misleading and slapped this fact-check on it. This article is a rebuttal to that fact-check. Towards the end, I also include my rebuttal to other arguments I have seen at other places. The main theme across various arguments is that biology is very very complex and nothing in medicine/health has bulletproof evidence. It is always possible to criticise any medical intervention. Thus, to avoid biases, it is very important to not use double standards, which is what the detractors of Ivermectin seem to be doing. Below, I quote paragraphs from the fact-check which are followed by my response. Criticisms/contributions are welcome: please make an issue here or submit a merge request to this file.

Clinical trials on the effectiveness of ivermectin against COVID-19 have also been done, some of which were indeed published and included in Lawrie’s meta-analysis. However, there are several issues with Lawrie’s meta-analysis. It wasn’t peer-reviewed, unlike published studies in scientific journals, and several of the studies used by the FLCCC and included in Lawrie’s review have also not been peer-reviewed[3-5].The lack of peer review means that scientists with relevant expertise, such as epidemiologists and biostatisticians, haven’t independently reviewed the research. Peer review is an important step in the scientific publishing process, as it helps authors identify and correct substantial errors or shortcomings in their studies. It can also help to determine the quality of a given study. For example, some publications may contain very strong scientific evidence and novel discoveries, while others may have only weak research to back up their hypotheses. Overall, peer review can help prevent the spread of exaggerated or unsupported scientific claims.

Peer review is neither necessary nor sufficient for validity of a document: these days, especially in health/medicine, its most important utility is perhaps in being a sign of existing biases: see the Lancet paper from last year completely dismissing the lab-leak hypothesis as a conspiracy theory, or the Lancet-published fraudulent surgispere study dismissing hydroxychloroquine

(BTW, Tess’s paper did pass peer review and was published later)

The focus on ivermectin as a potential drug against COVID-19 started with an in vitro study published by Caly et al. in 2020 showed that ivermectin had antiviral effects against the virus that causes COVID-19, SARS-CoV-2, in cell cultures[6]. Although this result provided researchers with a justification for continuing studies of the drug, this study in itself didn’t provide evidence of ivermectin’s clinical efficacy against COVID-19. Furthermore, the study used very high concentrations of ivermectin that were many times higher than the doses approved for use in people

A recent HUMAN double blind Randomized-Controlled-Trial (RCT) from Israel shows that even the dose actually being used in many countries is effective in reducing viral load. It was a small trial (n<100), and yet, it was able to show that after day 4 of treatment, the Ivemectin group had consistently higher rate of recovery as adjudicated by negative PCR tests.: image In the medrxiv disqus comments section for this paper, some people have complained that on some days, the difference was not statistically significant (95% confidence interval crosses 1.0), e.g. day 4 and day 10. But note that even on those days, the 95% confidence interval lies almost totally above 1.0 (Ivermectin effective) and this is most likely an artifact of the small trial size. It is incorrect to get hung up on the arbitrary 95% threshold. Instead of declaring Ivermectin useless for those days, the patients should be told that it is extremely safe and, it is 85% (but not 95%) likely that the reduction in viral load is not a chance finding: my guess is that most patients would opt to get Ivermectin. (the fact-check was probably written before this trial’s results were announced, but it was never updated) In any case, given the evidence of reducing deaths, the most important end-to-end outcome, it is not necessary to have the mechanistic evidence fully correct: Ivermectin may be helping by some other mechanism, e.g. improving the immune response.

For example, some limitations of these clinical studies included small sample size and vague definitions of disease severity. For instance, if the severity of symptoms is poorly defined, it becomes difficult to objectively assess the level of improvement provided by a drug treatment, if any.

The video talks about reduction of deaths, arguably the most objective and most important efficacy metric.

Furthermore, some of the studies would draw claims upon observations, which were not statistically significant, i.e. it could not be ruled out that all the results occurred only by chance.

It is true that most of the trials were small and thus the reduction of deaths were not statistically significant. However, it is dishonest for them to not mention that when we combine all the trials – as is the state-of-the art for making guidelenes in evidence based medicine – the reduction in deaths does become statistically significant.

Furthermore, some studies did not have a placebo control at all. Without a placebo control, we cannot exclude the bias created by the placebo effect. The placebo effect occurs when a participant’s belief that she or he is receiving an effective treatment produces a positive effect on their symptoms, even though they have actually been given a placebo with no active effect on their illness.

The placebo effect is indeed a problem for trials designed to measure subjective outcomes like pain-reduction. But very objective and hard outcomes, e.g. death, which this video is about, are much less prone to such biases. Until we have large blinded trials to evaluate reduction in deaths – of which there are NONE – these Randomzed Controlled Trials (RCTs) provide a very useful signal about the causal effects of ivermectin on deaths when given to patients with knowledge of what they are getting, as it happens in real life. Doing blinded trials can be very hard in a pandemic setting when people are dying in hospitals, where the main priority is to save lives, and doing irrefutable trials is a secondary priority, unlike prevention trials like the vaccine trials.

It’s also worth pointing out the double standard here. Masks were mandated. Where is the double-blind RCT showing that it is effective. You cant even do a blind RCT for masks (people cant wear masks and not know it). The only RCT (Danish trial, n=4000) of masks for Covid could not find any statistically significant benefit of masks.

Many of the claims (e.g. prevention of transmission, prevention of ASYMPTOMATIC infections) confidently being made about mrna vaccines do NOT even have unblinded RCTs supporting them and thus they suffer from even bigger problems like possible confounding: the possibility of health-conciuous people being overrepresented in the vaccinated group. The Pfizer Covid vaccine trial was double blinded, but given the strong symptoms in vaccine recepients, it is likely that the subjects and their evaluators could guess reasonably whether they got the vaccine or the placebo, which is specially problematic because they evaluated outcomes much less objective and much less important than reduction in deaths: many large vaccine trials (e.g. Pfizer trial) recruited too few high-risk subjects to show a statistically significant reduction in deaths. Worse, at 2-3 months, they unblinded the trials and offered the vaccine to even the placebo group : see the section “Unblinded and without a control group—what about safety?” in the linked article. (This paragraph is not to undermine the importance of vaccines: they have great evidence about reducing symptomatic infections. But ivermectin is needed for the infections that happen until/despite global vaccination: even the US as 1000s of vaccine breakthrough cases, 2% of which are fatal https://www.latimes.com/science/story/2021-05-25/2-percent-of-breakthrough-infections-cause-covid-deaths-cdc).

Contrary to the video’s claim, a double-blind, placebo-controlled study by López-Medina et al. showed that ivermectin is unlikely to be effective in patients with mild COVID-19[7]. This study was well-designed and doesn’t possess many of the limitations present in the clinical trials cited in Lawrie’s meta-analysis. Since this study was double-blinded, neither the trial participants nor the researchers knew if they had received or administered the placebo or the drug being tested. This design helps to eliminate bias. That being said, the authors concluded that larger trials need to be conducted to better understand the effect of ivermectin on COVID-19.

It is incorrect to conflate “too small to measure statistically-significant reduction in deaths” with “ineffective in reducing deaths”. Just like many other Ivermectin-for-covid RCTs, this trial was too small to measure a statistically-significant reduction in deaths (0 deaths in Ivermectin arm, 1 in control): on combining such trials, most of which showed non-statistically significant reduction in deaths, the results become statistically significant. It is also incorrect to cherry-pick just one trial: there were many other RCTs, which the video reviews. Not only were there other double-blind RCTs, it is incorrect to ignore the unblinded RCTs for very objective and hard outcomes such as death,as expained in the above paragraph. (Also, some have raised other unreliability concerns about the Lopez-Medina trial https://osf.io/u7ewz, but I haven’t tried to verify them.)

Both Lawrie and the FLCCC cited three ongoing clinical trials testing the effects of ivermectin as a treatment for COVID-19 that produced inconclusive results, as well as a report by Juan Chamie, a data analyst with no training in biology or medicine.

Why is that even relevant? This video is by Tess Laurie, a medical doctor who also has a PhD in medicine. Also, You need to evaluate the argument, not the credentials of who said it.

On the 22nd of March 2021, the European Medicines Agency released a statement advising against the use of ivermectin for prevention or treatment of COVID-19 outside randomised clinical trials.

Appeal to authority has no place in science.

The video also misleads viewers by suggesting that safety data regarding a drug’s use for one purpose applies to different uses. Although scientists have known about ivermectin for more than forty years and it has been used to treat parasitic infections, this isn’t relevant to the safety of ivermectin in COVID-19 patients. Simply because a drug is safe to use for one disease doesn’t mean that it’s safe to use for another disease. Physiological changes produced by one disease doesn’t translate to the same changes in the case of another disease. The differences in physiological states can mean that the drug is safe to use for certain diseases, but not others. Indeed, many drugs have contraindications, or conditions in which the drug shouldn’t be used because it can harm the patient

In theory, this is legitmate concern about safety: it is true that a drug that is otherwise very safe may worsen a new disease, although such a possibility seems to be very unlikely: at least the long history of use Ivermectin rules out the possibility of it having any inherent toxicity which is the main safety problem with most medical interventions. Also, no drug/vaccine is 100% safe. Where are the trials used to recommend paracetamol/tylenol/acetaminophen and NSAIDs like ibuprofen when the pandemic started? A drug (oral Ivermectin) that has been distributed at a population level (to billions of people) for decades to prevent river blindness and has been prescribed even to kids for decades (e.g. for head lice, scabies), has much better safety evidence than the mrna vaccines which has not even been studied in humans for more than a few months. In contrast, dexamethasone, which the experts actually recommend based on 1-month safety evidence in a Covid-specific RCT, turned out to be suspected to cause deadly fungal infections in India in the longer term. Pre-covid, it was ALREADY known that mechanistically, steroids like dexamethasone temporarily “slows down” the immune system, and may cause increased susceptibility to infections.

In the video, Lawrie also discussed the affordability of ivermectin and how it can be easily and massively produced. Some have argued that the reason ivermectin isn’t being deployed against COVID-19 is that it is cheap and wouldn’t generate as much profit for pharmaceutical companies. Simultaneously, we can conclude that a cheap drug, thus widely affordable, which would be administered to tens to hundreds of millions worldwide, would be a hugely lucrative business. Interestingly, ivermectin’s manufacturer (under the name brand Strocemol) Merck doesn’t recommend its use against Covid-19 due to the lack of scientific evidence and safety data.

Merck does not own the patent to Ivermectin anymore. Other manufacturers can it for dirt cheap, as they are already doing it in India. In fact, when making that statement, it was very unethical of Merck to not disclose that they are working on a lucrative treatment for Covid ($1.2 billion deal with FDA). This lucrative deal would be pointless if the world recognizes Ivermectin’s evidence of effectiveness, and Merck can no longer profit from Ivemectin: off-patent, Indian companies can manufacturer it at much less cost than would be profitable to Merck.

Elswhere, people claimed that there could be publication bias. When I asked them, none of them could produce a single published trial (even a preprint) excluded by the meta-analysis of Tess. Some said that some of them may have turned out to be negative and the author may have not put up even a preprint. This is possible but highly unlikely given the immense anti-ivermectin bias in the academic elite and the mainstream media: anybody with a negative trial would become famous overnight and be on nytimes frontpage next day. Why would anybody with a negative trial not publish even a preprint? On the contrary, it seems there may be a publication bias in the other direction: people with studies showing cheap effective treatments seem to have difficulty/delays in even getting their preprints up

Some other research group has published another meta-analysis of Ivermectin where they excluded all RCTs that only have preprints and have not been formally published, citing possibility of publication bias. Although peer review can catch some obvious mistakes, given how politicised academia (peer-reviewers) has become on topics like lab-leak, hydroxychoroquine, Ivermectin, Covid vaccines, the subset of peer-reviewed articles likely suffers from confirmation bias. Also, exclusion of non-peer-reviwed reduces the overall number of deaths in the combination of the considered RCTs thus the reduction of deaths barely misses statistical si (relative risk of death in the Ivermectin group: 0.37, 95%CI 0.12 to 1.13). As explained above, given the strong bias against cheap off-patent repurposed drugs, the publication bias is probably in the other direction, especially if we add peer-review as a requirement, given how politicized Covid-19 science has been. Also, even after excluding non-formally-published RCTs, it is preposterous to conclude that “IVM is not a viable option to treat COVID-19 patients.”: the 95% CI is almost completely below 1 (reduction in deaths). Again, if you tell a patient that Ivermectin is extremely safe with 4 decades of safety data and it is 80% (but not 95%) likely that the observed reduction in deaths is not a chance finding, most Covid patients would take it. And if you do not exclude RCTs for questionable reasons, the likelihood rises to well over 95%, as explained in Tess’s video and her meta-analysis. (Also, in the medrxiv disqus comments section for this meta-analysis paper with a heavy anti-ivermectin bias, many have pointed numerical errors, but I have not verified them.)

Some people have raised concerns about the Elgazzar study. The authors havent yet responded yet. (Also, the fact that some typos were copy pasted in the raw data of their study isnt necessarily an evidence of data fabrication: spreadsheet software often offer existing entries as autofill suggestions, which leads of copying of errors even without actually doing copy-paste). However, I have not read those concerns carefully and there may have been some data fabrication. Note that the meta-analysis described in the paragraph above already did not include this and many other (non-peer reviewed) studies and yet found a mortaility benefit of Ivermectin with high confidence but not has high as 95%. Also is a video discussing what happens if we only remove the Elgazzar study from Tess’s meta-analysis.