. This paper is highly flawed, with multiple basic errors, invalid assumptions, highly biased discussion, and failure to correct any of the issues in over a month.
. Authors appear to have used a mismatched dataset, incorrectly including data from neighboring municipalities
. Author's R code appears to incorrectly select patients
. Authors also do not appear to have read the original paper in full, missing details of the population analyzed
. They also appear to have missed the distinction between an individual's residential city and the location of their medical care
. Authors assumptions are also invalid - participation for patients with symptoms may depend on multiple factors. For example, patients with more severe symptoms may be more likely to perceive serious risk and more likely to seek known effective treatment. Authors also only discuss potential biases in one direction - for example ignoring bias due to dropping out of the program. The assumption of no effect for treatment is contradicted by 17 out of 17 prophylaxis studies.
While it is possible the errors were accidental, two factors increase the severity.
The extreme mismatch in deaths with the original paper should have been a sign to check. The cause of the mismatch should have been relatively easy to determine, and in case authors were still stuck, they could have asked the original authors. Such an extreme and easily identifiable error in mortality numbers undermines author's credibility.
Further, authors were notified of the mortality error within 12 hours of announcing the preprint on Twitter and acknowledged the error, initially claiming a rapid correction would be forthcoming. However, the paper has still not been withdrawn as of Oct 16, over two months after publication, and is still available without any warning, note, or even partial correction. Leaving known highly incorrect information for so long raises ethical concerns.
Discussion in this paper is highly biased, supporting concern over bias from the team. Authors claim that
, but in reality all 17 of 17 prophylaxis studies at the time showed statistically significant benefits of ivermectin, including all 4 of 4 RCTs.
Authors reference none of the other prophylaxis studies, and cherry pick a few non-prophylaxis studies, and even then they ignore the signs of efficacy in those studies. While it's common for authors to misrepresent research, excluding all 16 other studies and cherry picking a few non-prophylaxis studies stands out as one of the most extreme cases.
Authors cite only 4 of the 99 studies, all with major issues, and none reporting prophylaxis results. Further, the 4 studies cited all show signs of efficacy -
showed 61% lower hospitalization for ivermectin vs. placebo (not reported by authors, without statistical significance); the co-principal investigator of
the posterior probability ivermectin was effective was 99%, 98%, 97% for mean time unwell and clinical progression @14 and 7 days, all exceeding the pre-specified threshold for superiority (clinical progression results were changed without explanation between the preprint and journal version, the primary outcome was not reported, with a new post-hoc highly biased outcome created).
All 4 of the cited studies are low quality studies with major issues. For example, one trial reports impossible data, refuses to release data despite pledging to, and had randomization failure, blinding failure, and numerous protocol violations
.
Author's summary is discordant with reality - 60 studies show statistically significant positive results for one or more outcomes (including 23 RCTs)
.
Further, author's discussion of preclinical data is highly cherry-picked and highly misleading, with inaccurate interpretation of the concentration required based on
(see discussion) and other studies, and no mention of the majority of proposed mechanisms and supporting data.
Ivermectin is an inhibitor of importin-α/β-dependent nuclear
import of viral proteins
Götz, Kosyna, Wagstaff, Wagstaff (B),
a SARS-CoV-2 3CL
pro inhibitor
Mody,
binds to glycan sites on the SARS-CoV-2 spike protein preventing interaction
with blood and epithelial cells and inhibiting hemagglutination
Boschi,
exhibits dose-dependent inhibition of lung injury
Abd-Elmawla, Ma,
may inhibit SARS-CoV-2 induced formation of fibrin clots resistant to
degradation
Vottero,
shows protection against inflammation, cytokine storm, and mortality in an LPS
mouse model of severe infection/inflammation that shares key pathological
features of severe COVID-19
DiNicolantonio, Zhang,
may be beneficial in severe COVID-19 by binding IGF1 to inhibit the promotion
of inflammation, fibrosis, and cell proliferation that leads to lung damage
Zhao,
may minimize SARS-CoV-2 induced cardiac damage
Liu, Liu (B),
has immunomodulatory
Munson and anti-inflammatory
DiNicolantonio (B), Yan properties, and has an extensive and very
positive safety profile
Descotes.
