Many major issues including multiple impossible numbers, blinding failure, randomization failure, and many protocol violations, as detailed below. Submit Updates or Corrections

Also see: Fraudulent Trial On Ivermectin Published By The World's Top Medical Journal. Big Pharma Reigns - Part 2 and Part 1, and 10 Questions for the TOGETHER Trial Investigators.

Apr 5: The paper was silently updated, with no indication or explanation of the changes. Changes include: age range, placebo description, per-protocol count, and death counts (details below).

May 5: The paper was silently updated again. A new summary notes that authors attempted to screen for previous ivermectin use, contradicting both the discussion section, where authors claim they ensured no use for COVID-19, and the exclusion criteria and interview forms, which do not specify ivermectin use.

Delayed >6 months. The paper was delayed over 6 months with no explanation. The companion fluvoxamine arm, completed at the same time, was published Aug 23, 2021. The paper was submitted to NEJM in Sep 2021 [vimeo.com (B)]. COI forms suggest that additional authors were added after submission and the corresponding author changed from Prof. Mills to Dr. Rayner [doyourownresearch.substack.com], whose conflicts include Pfizer, Merck, the Gates Foundation, and the Australian Goverment.

No response to data request. The trial registration states that data was to be available at termination and upon request [clinicaltrials.gov], however authors have not responded to a request for the data. Even funders of the trial have been unable to access the data [odysee.com]. Requests can be sent to thetogethertrial@gmail.com, let us know the outcome.

Three different death counts. In the original paper, Table 3 shows 21 and 24 deaths, while Table S6 shows 20 and 25 [twitter.com (B)]. In Table 3, death and grade 5 events showed the same 21/24 numbers, but different effect sizes, with 0.81 being closer to the 20/25 counts and the previously reported number. This is consistent with one death being moved between arms after manuscript generation, but not updated in Table S6 or the Table 3 AE RR. This cannot be explained by the safety population excluding patients with zero doses because the AE control deaths are higher. In email, a co-principal investigator suggested that the discrepancy was due to one being COVID-19 deaths and the other being all-cause deaths [stevekirsch.substack.com]. That explanation does not fit the data because one arm increases while the other arm decreases. Both co-principal investigators report in the paper that "they had full access to all the trial data and vouch for the accuracy and completeness of the data and for the fidelity of the trial to the protocol." A third set of death counts, 20 and 24, with RR 0.84, was presented by a co-principal investigator on Mar 18, 2022 [vimeo.com (C)]. In total, 4 different death relative risks have been presented: Mar 18, 2022 presentation: 0.84, Mar 30 paper: 0.88 (T3), 0.81 (T3 AE), and 0.80 (TS6, presented as 20 and 25 only without group sizes). 6 days after publication, the paper was updated, with no information given on what was changed. In this version, a "respiratory, thoracic and mediastinal disorders" death was removed from the control arm and an "infections and infestations" death was added to the ivermectin arm. The paper still indicates RR 0.81 for death AE.

Trial was not blind. Ivermectin/placebo blinding was done by assigning a letter to each group that was only known to the pharmacist. If a patient received a 3-dose treatment, investigators immediately know that the patient is more likely to be in the treatment group than the control group, because 3-dose placebo was relatively rare (~46% from PP). If a patient received non-3-day treatment, investigators immediately know that the patient is not an ivermectin treatment patient. Moreover, by observing the frequency of allocations, investigators can easily determine which letter corresponds to active ivermectin 3-day treatment, thereby removing all blinding. For example, consider 3-dose-ivermectin and 3-dose-placebo being identified by the letters G and K. If allocations to date have been G:11 and K:20, there is a very high probability that K is ivermectin. Note that this blinding failure is only obvious because the journal required the authors to restrict to the 3-day placebo group. Also note that it would have been trivial to avoid if desired, for example by using a unique identifier for all medication bottles. Note that there may be additional reasons for blinding failure, for example the paper specifies identically shaped bottles, but does not appear to specify identical appearance tablets [doyourownresearch.substack.com (B)].

Patient counts do not match previously released enrollment graph. Authors claim the ivermectin and control patients were all from on or after March 23, 2021, however independent analyses of the enrollment graph (contained in this presentation [dcricollab.dcri.duke.edu]) require including patients prior to this date to reach the reported numbers [doyourownresearch.substack.com (C), longhaulwiki.com]. The enrollment graph shows much higher enrollement to ivermectin near the start of the trial. The only way that the number of placebo patients can be the same as the number of treatment patients is if placebo patients were taken from an earlier period [Marinos], which creates a nonconcurrent control group [nejm.org] and substantial confounding by time as below.

Conflicting placebo arm counts across IVM/FLV arms. The IVM placebo arm has 679 patients and the FLV arm has 756. The 679 should be shared between the arms, with 77 extra patients for FLV. For FLV, there were 34 placebo patients requiring mechanical ventilation, for IVM there was only 25, indicating that 9 of the extra 77 placebo patients for FLV had mechanical ventilation, a much higher percentage during a period that had lower deaths and CFR (and included vaccinated patients). Placebo all-cause hospitalization shows 95/679 for IVM and 99/756 for FLV, i.e., only 4 of the extra 77 patients were hospitalized, but the paper reports an additional 9 patients with mechanical ventilation.

For FLV, there were 11 grade 1 AEs, for IVM there were 12, with 77 less patients.

For FLV, there were 50 grade 3 AEs, for IVM there were also 50, meaning the 77 extra patients had 0% grade 3 AEs vs. an expected 7.4%

For FLV, there were 54 CKD patients according to Figure 3 and eTable1 (2 according to Table 1). For IVM there was 5.

DSMC not independent. Reviewer 1 of the protocol notes that the DSMC is not independent [gatesopenresearch.org]. Prof. Thorlund is Vice President of the contract research organisation (CRO, Cytel), professor at the sponsoring university, and an author of the protocol. Dr. Häggström is an employee of the CRO. [doyourownresearch.substack.com (C), twitter.com (C)] reveals many other conflicts. Prof. Thorlund has written >100 papers with Prof. Mills. Prof. Singh has written 29 papers with Prof. Mills. Prof. Orbinski has written 9 papers with Prof. Mills. The first version of the web site showed Prof. Mills and Prof. Thorlund as joint leads. Emails pointed to a company MTEK Sciences, founded by Prof. Mills and Prof. Thorlund (MTEK is hypothesized to stand for Mills, Thorlund, Edward, Kristian). MTEK received grants from the Gates Foundation. MTEK also employed Dr. Häggström. MTEK was acquired by Cytel in 2019. Dr. Häggström works for the Gates Foundation. Two members of the DSMC have published a paper with members of a well known anti-ivermectin research group [Thorlund] and Dr. Hill, whose meta analysis has reports of external influence [c19ivermectin.com, c19ivermectin.com (B), twitter.com (D)]. The trial protocol reports that "an independent DSMC will be established, composed of scientists of unrivalled reputation and expertise, without involvement with this research protocol."

Unequal randomization, significant confounding by time. The trial reports 1:1:1:1 randomization, however independent analysis shows much higher enrollment in the ivermectin treatment arm towards the start of the trial [c19ivermectin.com (C), longhaulwiki.com]. This introduces very significant confounding by time due to the major change in the distribution of variants. [Zavascki] show dramatically higher mortality for Gamma vs non-Gamma variants (28 day mortality from symptom onset aHR 4.73 [1.15-19.41]). Many more patients were randomized to ivermectin vs. placebo in the first few weeks, for example the first week shows 82 ivermectin vs. 28 placebo patients, 2.9x higher. The period of excess ivermectin enrollment coincides closely with a period of significantly higher deaths and CFR in Brazil.

Missing time from onset patients show statistically significant efficacy. For the known time since onset subgroups, both groups show worse results than the overall results [twitter.com (E)], with the missing 317 patients showing significant efficacy RR 0.51, p = 0.02 (compared to 1.00 and 1.14 for known patients).

Unknown onset patients were enrolled, subgroup results opposite of previous trials. After imputation, the percentage of patients in the late treatment subgroup went from 46% to 56%. 87% of the unknown patients were predicted to be in the late group. This is reasonable and expected — patients that do not recall when the onset was are more likely to have had onset further in the past. What is not clear is how these patients could be enrolled in the trial, how many of these patients had onset >7 days, how this very late 317 patient subgroup could show much greater efficacy as above, and why authors did not report this result, analyze this in greater detail, or recommend further research.

Side effect profile consistent with many treatment patients not receiving authentic ivermectin and/or control patients receiving ivermectin. The side effects (e.g., gastrointestinal side effects were lower in the ivermectin arm) suggest that many ivermectin patients may not have received authentic ivermectin, or that placebo patients may have taken ivermectin. For comparison, there was a 3.6 times greater incidence of diarrhea in the treatment arm in [Lim].

A local Brazilian investigator reports that, at the time of the trial, there was only one likely placebo manufacturer, and they reportedly did not receive a request to produce identical placebo tablets [doyourownresearch.substack.com (B)]. They also report that compounded ivermectin in Brazil is considered unreliable. The protocol reports that "the study medication used will come from pharmaceutical plants that hold a commercial authorization for their production, already approved by ANVISA."

Vaccine inclusion changes. The trial changed from including vaccinated patients to excluding them on Mar 21, 2021 [clinicaltrials.gov (B), twitter.com (F)], and on Jul 5 the exclusion was changed to specify >14days. As above, meeting the reported placebo counts likely requires taking placebo patients from the earlier period, which has very significant confounding due to variant changes, and additionally confounding due to vaccination. Both of these favor the placebo group. The original vaccine inclusion criterion is unusual, but is shown in both the protocol and the clinicaltrials.gov record [clinicaltrials.gov (C), togethertrial.com]. This may have been chosen to reduce the potential for efficacy. For more on vaccination inclusion see [twitter.com (G)].

Incorrect conclusion. The conclusion states that ivermectin "did not result in a lower incidence of [hospitalization] or of [ER observation >6hr]". This is incorrect, hospitalization was 17% lower, which is not statistically significant with the sample size and typical statistical analysis. For the Bayesian analysis the authors use, the ITT probability of superiority for ivermectin was 79.4%, which is a positive result, the opposite of the conclusion.

Ivermectin use widespread in the community. Recent ivermectin use was not in the exclusion criteria. Ivermectin was available OTC, was recommended by the government for COVID-19, and had nine times higher sales [twitter.com (H)]. Authors claim they ensured patients did not use ivermectin via "extensive screening", but do not explain why this was not an exclusion criterion, or how this unwritten exclusion was ensured even though there is extensive missing data related to written exclusion criteria. Similar unwritten exclusions were not mentioned for other arms [twitter.com (I)], a primary investigator previously stated such an exclusion should not be an issue [twitter.com (J)], and it is not mentioned in the interview sheets [osf.io]. After publication, a co-principal investigator reportedly wrote that "even if some patients did access IVM, the fact that it is blinded should still maintain balance", which is incorrect, placebo patients taking ivermectin are expected to improve, treatment patients that already have significant tissue distributions may have positive, neutral, or negative responses to additional treatment.

Single-dose recruiting continued after change. The trial had requested moving to 3-dose treatment by Feb 15/19, when only 19 patients had been recruited, however the trial continued recruiting an additional 59 patients to single dose treatment [twitter.com (K)].

Per-protocol population different to the contemporary fluvoxamine arm. Table 2 per-protocol numbers show 92% per-protocol patients for ivermectin and only 42% for control. This appears to be a post-hoc change selecting only 3-day placebo patients, while similar selection does not appear to have been done for the companion fluvoxamine trial (showing 74% and 82% per-protocol patients for fluvoxamine and control) [Reis].

Multiple conflicting randomization protocols. [twitter.com (L)] reviewed the randomization protocol, finding three different algorithms, and conflicting versions in the papers.

Time of onset, required for inclusion, missing for 317 patients. For the companion fluvoxamine arm, 24% of patients had an unknown time from onset, including 179 of the control patients [Reis]. In this trial, 0 patients have an unspecified time from onset in Table 1, due to imputation. However, Figure 2 reveals that the time from onset is unknown for 317 patients, similar to the fluvoxamine paper. However, time from onset is required for the inclusion criteria. According to Figure 2, age and BMI also show missing values.

Conflicting comorbidity counts. The companion fluvoxamine arm ran from Jan 20 to Aug 5, 2021, while this trial ran from March 23 to Aug 6, 2021 — most control patients should be shared, with an additional 10% for fluvoxamine from the earlier start. The Aug 6 presentation, which has a date of 9:38am Aug 6 local trial time [doyourownresearch.substack.com (D)], shows 678 placebo patients, indicating that either 0 or 1 placebo patients were randomized on Aug 6. Zero patients should have been randomized on Aug 6, because authors cannot add patients after unblinding.

The fluvoxamine control arm shows 16/756 control patients with asthma. The ivermectin control arm has a subset of these patients (679), but shows a much higher prevalence of asthma (60 patients). This might be possible due to imputation if there was a very high percentage of missing data, however imputation does not appear to be a good explanation. For example, placebo CKD goes from 2 to 5 (FLV->IVM). First, it is not logical to impute CKD on patients based on the other variables. Second, the protocol specifies imputation only with up to 20% missing data, making it unlikely that imputation would add 150% of CKD patients. Third, the degree of change between FLV and IVM varies dramatically, with IVM reporting 666%, 275%, 150%, and 43% more patients for CPD, asthma, CKD, and CCD, without any clear explanation for similar differences in the percentage of missing data (all were collected on the same interview form).

Conflicting target enrollment. There are conflicting target enrollment numbers. The protocol showed 800 patients per arm as of Mar 21, 2021 (after the trial started) [static1.squarespace.com, twitter.com (M)], the co-principal investigator reported 800 per arm in an interview published June 14, 2021 [halifaxexaminer.ca], and the protocol changed to 681 on June 22 [static1.squarespace.com (B)]. However, the trial record from Jan indicates 2724 (681*4) patients [clinicaltrials.gov (C)], suggesting that the 800 goal was later, and was kept for fluvoxamine but reverted for ivermectin. The fluvoxamine arm which started two months earlier was terminated at the same time, and was terminated due to superiority [Reis] after 741/756 patients. Note that Gamma was declining significantly around the termination point, which likely favors improved efficacy if the trial continued, given the late treatment and dosage used. The co-principal investigator reports three different reasons for stopping the trial [twitter.com (N)]: a) because they ran out of money, b) because third parties were not supportive, and c) it was done by the DSMC and was out of their control.

Reportedly terminated for futility although futility threshold not reached. The trial was reportedly terminated due to futility [twitter.com (O)], however the futility thresholds were 20%, 40% and 60%, and all published probabilities are >60% (ITT 79.4%). Additionally, the fluvoxamine arm did not have the higher 60% threshold, only using 40%. Note the DSMC was not independent as below.

Screening to treatment delay. Most Together Trial protocols show an additional day delay in already late treatment for most patients. The Aug 5, 2021 protocol published with the metformin paper [ars.els-cdn.com], shows treatment administration one day after screening, baseline, and randomization (Table 2, schedule of study activities). This can also be found in the protocol dated Mar 11, 2021 [drive.google.com]. The protocol attached to the ivermectin paper, dated Feb 15, 2021, shows a different schedule, stating that the treatment should be administered on the same day of randomization. There is no explanation of when this change was made, how the overlapping metformin and ivermectin arms could use different schedules, or how this change was implemented (there are many tasks in the screening and baseline visits). There is no reporting for how many patients received treatment on the same day. The form for the first treatment visit asks if there were clinical events including >6hr ER visits since the baseline visit, which would not be possible if this visit was immediately after randomization. Time of first treatment was recorded [osf.io], but no information has been reported. According to [Forrest], WhatsApp messaging and video was used for recruitment, raising the question of how medication was delivered in cases where recruitment was done online.

Mean delay. The reported mean number of days from symptoms to randomization likely only includes known onset patients and therefore is likely to significantly underestimate the actual average, in addition to not including the time between randomization and treatment.

Viral load not reported. The protocol has change in viral load as an outcome, however only viral clearance is reported, and without any details (for example, using a high Ct value would have limited relevance).

Incorrect dose reporting, many patients at higher risk due to BMI may have received lower per kg doses, and show lower efficacy. The paper reports 400μg/kg for 3 days, however the protocol indicates that this was only up to 90kg, meaning that the dose received for higher-risk high BMI patients was even further reduced from dosage which is already far below clinician recommendations for the dominant variant [twitter.com (P)]. 50% of patients had BMI ≥30. Much greater efficacy was seen in the low BMI subgroup (RR 0.77 vs 0.98).

Plasma concentration below known effective value. [Krolewiecki] show an antiviral effect only with plasma concentrations above 160ng/mL. Figure S5 shows that the authors expected the mean concentration to be well below this level [twitter.com (Q)]. Dosage requirements are likely to vary significantly depending on many factors including the variant encountered, time of administration, mode of administration, patient genetics, concomitant medications, SOC, and the distribution of the infection in different tissues. However, the dose used is far below what is recommended by clinicians for post-infection treatment with the Gamma variant — about 2.5 - 6.5x lower, depending on the recommendation and which estimate of fasting/fed administration is used. The trial used fasting administration, however Merck's product information reports that "administration of 30mg ivermectin following a highfat meal resulted in an approximate 2.5-fold increase in bioavailability relative to administration of 30mg ivermectin in the fasted state." [merck.com]. Moreover Dr. Craig Rayner, a senior investigator on the trial, previously published research indicating that a higher dose is required [sciencedirect.com], raising the question of why the dose and fasting administration was chosen, especially for the initial single dose regimen.

Primary outcome easy to game, selected after ivermectin one dose arm. The subjective "emergency room visit for >6 hours" criterion shows higher risk (RR 1.16), while hospitalization is lower (RR 0.83 all-cause, RR 0.84 COVID-19). The primary outcome results were set on March 21, 2021, after the single dose ivermectin arm. Given the known public biases of some investigators, this may have been specifically chosen to reduce efficacy. Authors claim that the 6hr threshold did not include waiting time, however the emergency visit form has no mention of waiting time, only recording presentation and discharge times [osf.io].

Including contraindicated chronic kidney disease patients. "Stage IV chronic kidney disease or on dialysis" was an inclusion criterion, however ivermectin is contraindicated with kidney disease [Arise, en.wikipedia.org, Nunes] (not always recognized, and may be less critical with very low dose use for other conditions). According to Table 1 there were only 7 CKD patients, however two conflicting numbers are provided in the fluvoxamine paper [Reis]: Table 1 reports 2 CKD placebo patients - it's not clear how CKD was imputed for 3 more patients in the smaller IVM group. Moreover, Figure 3 shows 54 placebo CKD patients for FLV.

Antigen test requirement. The protocol indicates that patients with a negative test may be included if they become positive a few days later, potentially resulting in a long unreported delay between randomization and treatment, depending on how investigators interpreted the protocol. The requirement for a positive antigen test excludes the possibility of early treatment in many cases - tests have very high false negative rates in the early stages of infection, and symptoms may appear before the test becomes positive.

Inconsistent subgroup analysis. The presented subgroup analysis is inconsistent with plans and with the fluvoxamine paper, including not presenting pre-specified subgroups, presenting subgroups that were not pre-specified, presenting different subgroups to the contemporary fluvoxamine paper, and modifying subgroup definitions [twitter.com (R)].

Missing analysis. Authors do not provide time from onset analysis for either mortality or hospitalization, only the combined measure including the ER visits where anomalous results are seen. Authors do not provide per-protocol or mITT results for mortality or hospitalization. Per-protocol mortality results were provided for the companion fluvoxamine trial.

Missing outcomes. Many outcomes specified in the protocol appear to be missing, including the co-primary outcome of COVID-19 mortality (only all-cause mortality is provided, specific AE details not provided), time to clinical failure, days with respiratory symptoms, mortality due to pulmonary complications, cardiovascular mortality, COVID-19 symptom scale assessment, WHO clinical worsening scale assessment, and 14 day mortality.

Imputation error. In the paper authors use imputation in Table 1 but not in Figure 2. Authors also released a version of Figure 2 with imputation [togethertrial.com (B)], where the numbers for age and BMI now match the imputed numbers in Table 1. However, the time from onset numbers are very different, with the treatment arm showing 302 patients for 0-3 days, and the imputed version of Figure 2 showing 367 [doyourownresearch.substack.com (E)].

Missing age information. According to Figure 2, 98 patients are missing age information [twitter.com (S)].

Unexpected differences in missing data. Age is unknown for 98 patients, however according to Figure 2, BMI is missing for only 11 patients, smoking status is unknown for only 2 patients, lung disease is unknown for only one patient, and cardiovascular disease is known for all patients.

Mid-trial protocol changes. There were several mid-trial protocol changes on July 5, 2021 [clinicaltrials.gov (D)]. The number of patients for viral load analysis was reduced, only for the ivermectin arm. All-cause, cardiovascular, and respiratory death outcomes were deleted (all-cause was reported). Exclusions were modified to allow enrolling patients vaccinated within the last 14 days. Inclusion criteria were modified to allow enrolling healthy young people — the criterion "fever >38C at baseline" was added, allowing enrollment independent of increased risk.

Statistical analysis plan dated after trial start. The statistical analysis plan appears to be dated after the trial started [twitter.com (T)].

Per-protocol placebo results very different. The 3-dose placebo appears to have been much more effective [Marinos, twitter.com (U)]. This could be consistent with placebo patients accidently receiving treatment.

Imputation protocol violation. The protocol specifies multiple imputation with up to 20% of missing data, however imputation was done with time from symptom onset, which has >23% missing data [twitter.com (V)].

Two different per-protocol counts. Figure 1 shows 228 per-protocol for the control arm, while Table 2 shows 288. This was modified in the Apr 5 update without explanation.

Conflicting adverse event counts. Table 3 and Table S6 adverse event counts do not match for any grade, e.g., grade 1/2 in Table S6 shows 82 for IVM, while Table 3 shows 65 [twitter.com (W)]. The Apr 5 update changed the grade 5 events without explanation, however the other grades remain conflicting.

Possibly the largest financial conflict of interest of any trial to date. Disclosed conflicts of interest include: Pfizer, Merck, Bill & Melinda Gates Foundation, Australian Government, Medicines Development for Global Health, Novaquest, Regeneron, Astrazeneca, Daichi Sankyo, Commonwealth Science and Research Organization, and Card Research. Many conflicts of interest appear unreported. For example, Unitaid is a sponsor [Harper, togethertrial.com (C)].

Analysis done by a company that receives payment from and works closely with Pfizer. All analyses were done by Cytel. Cytel is a statistical modelling company that helps pharmaceutical companies get approval — they work very closely with Pfizer [cytel.com]. Cytel's software and services are used by the top 30 pharmaceutical companies [cytel.com (B)].

A co-principal investigator works for Cytel and the Gates Foundation [empendium.com]: "The majority of the time I work for a company called Cytel, where I design clinical trials, predominantly for the Bill & Melinda Gates Foundation".

Reportedly, the first author's center is funded by pharmaceutical companies, and independent investigators tried to participate in the trial but were denied [odysee.com (B)].

The Gates Foundation is a founding partner of GAVI, which took out Google ads telling people not to use ivermectin [twitter.com (X)], and a major funder of Unitaid, which may have modified the results of the Hill meta analysis in a way that prevented adoption [c19ivermectin.com, c19ivermectin.com (B), twitter.com (D)].

Associated with MMS Holdings. The trial is associated with MMS Holdings [dcricollab.dcri.duke.edu], whose mission includes helping pharmaceutical companies get approval and designing scientific studies that help them get approval. One of their clients is Pfizer [mmsholdings.com].

Certara. One of the senior investigators was Dr. Craig Rayner, President of Integrated Drug Development at Certara - another company with a similar mission to MMS Holdings. They state on their website that: "Since 2014, our customers have received over 90% of new drug and biologic approvals by the FDA." One of their clients is Pfizer [certara.com].

Local variant shows very different characteristics. The trial took place in an area of Brazil where the Gamma variant was prominent. Brazilian clinicians report that this variant is much more virulent, and that significantly higher dosage and/or earlier treatment is required, as may be expected for variants where the peak viral load is significantly higher and/or reached earlier [Faria, Nonaka].

Funding list incorrect. The paper does not include the Bill and Melinda Gates Foundation or Unitaid as funders, however the Mar 25, 2021 protocol shows the Gates Foundation [static1.squarespace.com] and the web site shows Unitaid [togethertrial.com (C)].

Single dose arm results missing. Results for the single dose ivermectin arm have not been reported.

Anomalous results from the same region. A local Brazilian investigator reports that the study was conducted in almost the same time and location as the Brazilian component of the molnupiravir trial. Notably, molnupiravir's EUA relied on the unusually higher efficacy observed in Brazil.

Designed by Cytel. The trial was designed by Cytel, a company that helps pharmaceutical companies get approval and that works very closely with Pfizer [cytel.com, cytel.com (C)]. Cytel's software and services are used by the top 30 pharmaceutical companies [cytel.com (B)].

Bayesian probability of superiority hidden in appendix. The bayesian probability of superiority figure, featured in the main paper for FLV, MET, HCQ, was hidden in the appendix for IVM [twitter.com (Y)].

Placebo unspecified. The placebo appears to be unspecified in the paper and protocol. The initial trial announcement indicated the placebo was vitamin C [cytel.com (C)], which would be an active treatment according to the results of 43 studies (mortality RR 0.75 [0.61-0.91]). The metformin arm reports using talc, however fluvoxamine and ivermectin do not appear to report details of the placebo, which could potentially be different, for example based on manufacturer limitations for matching active treatment tablets. Authors do not specify the appearance of the placebo tablets, suggesting that they may not match the treatment tablets, providing an additional reason for blinding failure.

Previous protocol changes. There are two previous published protocols, both are called "version 1", we refer to them as 1A (3/11/21 [drive.google.com] ) and 1B (8/5/21 [gatesopenresearch.org]. 1B deletes subgroup analysis by treatment delay, and deletes a statement requiring prior approval for amendments. 1B adds the statement: "we hypothesize that younger patients will benefit more than older patients."

Patients 50 years old were assigned to different groups in Table 1 and Figure 2 (≤50 vs. <50).

Greater efficacy was seen for patients >50 (RR 0.77) vs. patients ≤50 (RR 1.01).

The following comments are prior to the publication. We note that authors claim they have not included patients prior to the time period for the 3 dose ivermectin patients, however this conflicts with previously reported data as per the analyses above.

The trial randomization chart does not match the protocol, suggesting major problems and indicating substantial confounding by time. For example, trial week 43, the first week for 3 dose ivermectin, shows ~3x patients assigned to ivermectin vs. placebo [reddit.com]. Treatment efficacy can vary significantly over time, for example due to overall improvement in protocols, changes in the distribution of variants, or changes in public awareness and treatment delays. [Zavascki] show dramatically higher mortality for Gamma vs non-Gamma variants (28 day mortality from symptom onset aHR 4.73 [1.15-19.41]), and the prevalence of the Gamma variant varied dramatically throughout the trial [ourworldindata.org]. This introduces confounding by time, which is common in COVID-19 retrospective studies and has often obscured efficacy (many retrospectives have more patients in the treatment group earlier in time when overall treatment protocols were significantly worse).

According to this analysis [reddit.com], the total number of patients for the ivermectin and placebo groups do not appear to match the totals in the presentation (the numbers for the fluvoxamine arm match) — reaching the number reported for ivermectin would require including some of the patients assigned to single dose ivermectin. Reaching the placebo number requires including placebo patients from the much earlier ivermectin single dose period, and from the early two week period when zero ivermectin patients were assigned. If these earlier participants were accidently included in the control group, this would dramatically change the results in favor of the control group according to the changes in Gamma variant prevalence.

An investigator from Brazil notes that the gamma variant became prevailing in the state of Minas Gerais later than in the rest of the country, with the time when gamma prevailed for the trial locations being more closely aligned with the start of the ivermectin arm [ufmg.br]. Due to the substantial differences in disease course and risk between the variants, authors need to consider only patients recruited during the same time period.

Treatment delay is currently unknown, however the protocol allows very late inclusion and a companion trial reported mostly late treatment. Overall mortality is high for 18+ outpatients. Results may be impacted by late treatment, poor SOC, and may be specific to local variants [Faria, Nonaka, Sabino]. Treatment was administered on an empty stomach, greatly reducing expected tissue concentration [Guzzo] and making the effective dose about 1/5th of current clinical practice. The trial was conducted in Minas Gerais, Brazil which had substantial community use of ivermectin [otempo.com.br], and prior use of ivermectin is not listed in the exclusion criteria.

This trial uses a soft primary outcome, easily subject to bias and event inflation in both arms (e.g., observe >6 hours independent of indication). There is also an unusual inclusion criteria: "patients with expected hospital stays of <= 5 days". This is similar to "patients less likely to need treatment beyond SOC to recover", and would make it very easy to reduce the effect seen. This is not in either of the published protocols.

RCTs have a fundamental bias against finding an effect for interventions that are widely available — patients that believe they need treatment are more likely to decline participation and take the intervention [Yeh], i.e. RCTs are more likely to enroll low-risk participants that do not need treatment to recover (this does not apply to the typical pharmaceutical trial of a new drug that is otherwise unavailable). This trial was run in a community where ivermectin is widely known and used.

The same trial's results for a previous treatment were initially reported as RR 1.0 [0.45-2.21] [ajtmh.org], while the final paper reported something very different — HR 0.76 [0.30-1.88] [jamanetwork.com].

Trial design, analysis, and presentation, along with previous public and private statements suggest investigator bias. Design: including very late treatment, additional day before administration, operation in a region with high community use, specifying administration on an empty stomach, limiting treatment to 3 days, using soft inclusion criterion and a soft primary outcome, easily subject to bias. Analysis: authors perform analysis excluding events very shortly after randomization for fluvoxamine but not ivermectin, and report viral load results for fluvoxamine but not ivermectin. Presentation: falsely describing positive but not statistically significant effects as "no effect, what so ever" [Amrhein, odysee.com (C)]. Prior statements: [odysee.com (C)].

The local Brazilian investigator also reports that nitazoxanide was tested in the same location, however very few patients reportedly experienced urine discoloration, while 100% are expected to experience this side effect; and that 6-hour observation is a poor choice because it is almost impossible to stay less than 6 hours in Brazil.

For additional issues see: [covid19criticalcare.com, doyourownresearch.substack.com (F), longhaulwiki.com, Marinos, Marinos (B), stevekirsch.substack.com, trialsitenews.com, twitter.com (AA), twitter.com (AB), twitter.com (AC), twitter.com (Z)]. Protocols, approvals, and statistical analysis plans can be found here [togethertrial.com (D)].

NCT04727424.