Ioannidis / Raoult: Dialogue of the Deaf?

This February 20, the IHU Marseille had as distinguished guest for a video conference: the prestigious Stanford Professor, John P.A. Ioannidis, MD, PhD

The theme was: “Epidemiology of COVID-19: proofs, risks and misunderstandings”

Professor Ioannadis made a presentation with powerpoint slides. He then took a few questions and also did dialogue with Professor Raoult.

When it comes to medical science, Professor John P.A. Ioannidis is extremely recognized. He holds the position of C.F. Rehnborg Chair in Disease Prevention, Professor of Medicine, of Epidemiology and Population Health, and (by courtesy) of Biomedical Data Science, and of Statistics; co-Director, Meta-Research Innovation Center at Stanford University. See a more detailed bio here.

The invitation may have been triggered by extremely controversial comments made recently by Professor John P.A. Ioannidis in a podcast two weeks ago about “Building a Meta-Research Career and Constructing COVID-19 Health Policy”

“We have a paper coming out in Nature Communications, which is a meta-analysis of hydroxychloroquine trials. We show a significant increase in mortality,” asserted Professor Ioannidis, adding:

“Probably we killed about 100,000 people with hydroxychloroquine as a treatment globally.”

The full podcast can be found at:

The relevant excerpt can be found here in audio:

Ioannidis is very respected globally, including by Professor Raoult, who presents him as the world’s number one when it comes to epidemiological analysis.

Brief Analysis of the Meta-Analysis Referred to by Professor Ioannidis

The study referred to by Professor Ioannidis in the interview, and forthcoming for publication in Nature Communications, can be found in the preprint “Mortality outcomes with hydroxychloroquine and chloroquine in COVID-19: an international collaborative meta-analysis of randomized trials, by Cathrine Axfors MD, PhD et al.”

The analysis covers 28 studies, half of which were published, the others unpublished. HCQ was evaluated in 26 trials (10,012 patients). There were only five trials (18%) in an outpatient setting. Two large trials (RECOVERY and SOLIDARITY) dominated, as they contributed to respectively 47% and 19% of all patients in the HCQ trials.

The analysis covers 10,319 patients and only takes into consideration randomized controlled trials, thereby excluding the major observational studies that were conducted about early outpatient treatment with HCQ by Raoult’s team at IHU-Marseille.

Regarding hospital level trials, dominant in the meta-analysis is the highly controversial RECOVERY trial by oxford University, which we have widely covered on – with high, probably toxic doses of hydroxychloroquine, administered late (3 days after hospital admission), administered alone, i.e. without zinc and other agents susceptible to improve outcomes. Over a thousand of innocent volunteer patients died under the watch of Profs Landry and Horby, with this low quality therapy (or the alternative placebo, given in the name of “science”).

The WHO SOLIDARITY trial is somewhat different, but also applies to hospital level treatment only, when the disease moves towards an inflammatory response and hydroxychloroquine is anyway known to be of little use.

From their analysis, relying mostly on those two trials, the authors conclude that “The average mortality was 10.3% (standard deviation 13.5%) in inpatient trials and 0.08% (standard deviation 0.18%) in outpatient trials.”

This brought the authors to conclude that “treatment with HCQ was associated with increased mortality in COVID-19 patients, and there was no benefit from treatment with CQ.”

While the authors indicate that their “findings have unclear generalizability to outpatients,” they still insist, without any caveat, that “medical professionals around the globe are encouraged to inform patients about this evidence.”

While the overall results of this meta analysis are not relevant to outpatient use, because dominated by hospital level treatment trials, it’s worth looking at the five included studies focusing on outpatient treatment.

In this meta analysis, out of the 5 outpatient studies, only three have over 10 participants, two using HCQ and one chloroquine: COVID-PEP (244) International (Boulware et al); BCN PEP CoV-2  (136) Spain; NCT04342650  (78) (chloroquine) Brazil.

The study by Boulware et al has been abundantly criticized and debunked. The study has many flaws: absence of proper diagnosis because reliance on self-assessment; inclusion of mostly young, Internet savy, population; delays in provision of the drugs sent by courier throughout North America; loss of many patients, administration of HCQ alone i.e. not in combination with agents such as Zinc known to improve its efficiency, reliance on necrological sections of newspapers to figure out if there were deaths, etc.

Despite all these issues, the Boulware study was published last August in the New England Journal of Medicine. Its main conclusion is that “After high-risk or moderate-risk exposure to Covid-19, hydroxychloroquine did not prevent illness compatible with Covid-19 or confirmed infection when used as postexposure prophylaxis within 4 days after exposure.”

The Boulware study was revisited, and two independent recalculations of its results point to a significant positive effect of hydroxychloroquine, i.e. the very contrary conclusion reached by Boulware et al. and published in the New England Journal of Medicine!

The data of the study, hailed by the media as proof that the “orange man” was wrong, actually suggest he was right! 🙂


A first re-analysis was by Wiseman et al., who presented as interim findings that: 

“HCQ may reduce the development of COVID-19 by as much as 65% (RR 0.35, CI 0.13-0.93, p=0.044) when received within 3 days of exposure (RR 0.83 at 3-5 days; RR 1.37 at 5-7 days).”

The second re-analysis was by Brazilian statistician Márcio Watanabe, from the Universidade Federal Fluminense: 

“We conclude their randomized, double-blind, placebo-controlled trial presents statistical evidence, at 99% confidence level, that the treatment of Covid-19 patients with hydroxychloroquine is effective in reducing the appearance of symptoms if used before or right after exposure to the virus.”

Note that Watanabe communicated to the undersigned that NEJM did not respond to his submission to the journal, for this re-analysis to be published as a Letter to the Editor. NEJM still stands with the findings published by Boulware et al, even if these seem to be false, or at the very least misleading.

It must be noted that, among the outpatient studies, only the one by Boulware et al, despite these statistical flaws, carries weight in the meta-analysis co-authored and hailed by Ioannidis.

The second study, by Oriol Mitjà looked at the reduction in the viral load following HCQ treatment in the first days of symptoms, in a population of health care workers. There were 136 patients in the intervention arm. It was a relatively young population with a mean age of 41.6 years and mostly female (72.1%). In this study, there were no death or mechanical ventilation, either for the treatment or control group. Hospitalization occurred for 7.1% in the control group, and 5.9% in the treatment group. 

This outpatient study is of very little use, as it does not focus on truly at risk people (say above 65) and does not involve an effective early treatment combining hydroxychloroquine with agents such as zinc.

As for the Brazilian study, using chloroquine alone as therapeutic agent, it was discontinued shortly after several deaths occurred, and it is subject to a federal criminal investigation, as toxic doses of chloroquine may have been used.

Here is the analysis by Dr Simone Gold about this particular aspect of the study.

The two other outpatient studies referred to in the meta-analysis are not relevant because being extremely small (yet surprisingly not disregarded by the authors of the meta-analysis): NCT04333654 (5) (HcQ) International; NCT04340544  (8) (HCQ) Germany.

One can conclude that this meta-analysis co-authored by Prof Ioannidis, at least in its pre-print format, is not only flawed but also totally irrelevant to outpatient treatment.

The analysis does not disqualify at all the use of hydroxychloroquine, in the context of a quality package of care and a multi-sequencing of drugs, as described in the landmark publication by McCullough et al titled “Multifaceted highly targeted sequential multidrug treatment of early ambulatory high-risk SARS-CoV-2 infection (COVID-19),” and implemented by many medical doctors worldwide, including at Professor Raoult’s hospital in Marseille.

The Presentation by Professor Ioannidis and the Q&As

The conference by Professor Ioannidis can be viewed at the link below. It touches about key aspects of the response to the pandemic. He stresses how the disease is mostly lethal to older people, and emphasizes that risk stratification can easily be implemented.

He rightly emphasizes that we may have sought to protect those at low risk and failed to protect those at risk, i.e. older people – something that has been stressed at lengths in this blog since April 2020.

He deplores that not enough emphasis has been put in protecting the vulnerable groups of society, especially in nursing homes, using terms such as “massacres” to refer to what happened, and continues to happen, in these aged care facilities.

In his recommendations, he stresses much heightened protection measures in nursing homes to avoid infections, among various other measures (see screenshot of slide).

During his powerpoint presentation, he does not refer to early outpatient treatment, and does not refer to either hydroxychloroquine (despite his claim that it killed 100,000 people), ivermectin (despite the considerable attention currently given to this therapeutic agent) or even budesonide, despite the recent Oxford randomized study clearly showing its effectiveness for outpatient treatment.

During the Q&A session, it was not Professor Raoult, but someone in the audience who asked about the elephant in the room, i.e. early outpatient treatment.

Question: “What do you think about the early outpatient treatment?”

Response: “Oh goodness uh I think it’s unclear to me how much we can achieve with early outpatient treatment.”

“What I do believe is that we should avoid overcrowding hospitals. I think that somehow the response to the pandemic has been centered on hospitals. A lot of attention has fallen on how many ICU beds do we have, at kind the end of the chain.”

“This is very unfortunate because this is a community infection. It needs to be managed in the community we need to do our best to keep pretty much most of the management in the community and avoid hospitals being overwhelmed.”

“I’m not sure that, of all the people who get admitted, we do really much for them.”

“We have some options for treatment. It’s very difficult to document how effective they are and or how ineffective they are as you realize especially for early treatment the outcomes are going to be pretty good in the vast majority of patients anyhow. So it’s very difficult to substantiate this.”

“There’s huge polarization, there’s huge debate, there’s a lot of people have nothing to do with science who have been talking about it.”

“I know that Didier (Raoult) has suffered in that mix of journalism and politics and then upheaval that has ensued”

“I just think that we should try to keep the infection away from hospitals as much as possible and much of the time there’s nothing to do other than being calm and wait.”

Despite not endorsing early outpatient treatment, Ioannidis notes: “the earlier the treatment, the least likely you will have bad outcomes.”

Additionally, he warns about a risk of over reaction and of over-treatment: “the earlier the treatment, the higher the risk of over reaction” he declared.

He does not mention at all the important issue of the long haulers, which suffer long term symptoms from the disease – a condition that typically does not occur for people being treated early.

Ioannidis mentions there are “zillions of trials” but a small fraction of them has been completed. “There is still a lot to learn.”

“We need very solid data before I can say we know what to do in that case”

Professor Didier Raoult then explained what was done at IHU-Marseille regarding early treatment.

“We have treated something like 14,000 people, among those 11,000 as outpatients. Our fatality for outpatients is 1 out of 1000, so it’s quite low.”

“For patients that come, more late they come, more high is the fatality rate.”

Ioannidis did not acknowledge or ask questions about this long case series of outpatients and the very low fatality rate achieved. This is probably for the same reason Raoult’s work was not included in the meta-analysis: the observational, non-randomized nature, of the Raoult’s findings.

There are many epidemiologists valuing observational studies. Ioannidis does not appear to be one of them.

Further in the discussion, Ioannidis talks about people in a later stage of the disease.

“For most people who die from COVID-19, There is very little that can be done. You know, once you get to that advanced stage of the disease, you know that dexamethasone probably works so we have even randomized controlled trial for that.”

“It’s very difficult to think of other interventions that have a very clear ability to save lives once you are at that late stage.”

“For early treatment, it’s a different thing, but as I said, almost all those people do very well anyhow, so it’s very difficult to identify specific interventions with clear benefits in that population.”

Professor Raoult did not challenge Professor Ioannidis assertion that “it’s very difficult to identify specific interventions with clear benefits” for outpatient treatment, while he and his team have treated such 11,000 outpatients since the beginning of the pandemic.

The outpatient treatment protocol at IHU-Marseille has evolved considerably, with the addition of zinc, anti-coagulants, corticosteroids, etc. as outlined in this recent video.

To Conclude

Professor Ioannidis kept his position that there is no outpatient treatment for C19 showing clear benefits, yet he did not reiterate or explain how he found that 100,000 people would have supposedly died because of hydroxychloroquine.

Professor Raoult probably chose to avoid challenging his guest speaker about his most controversial claim, and only stressed as defence that the fatality rate among outpatients treated at IHU Marseille was only 1 out of 1,000.

Somewhat bizarrely, Professor Ioannidis emphasizes the need to treat the disease early, in the community, to avoid overcrowding hospitals, but he does not provide any form of solution about it, as he claims there is no evidence outpatient treatment works.

Professor Raoult, who has treated thousands of outpatients at IHU-Marseille, did not even try to convince Professor Ioannidis that he and his team are actually saving lives through the early therapies they provide to thousands of outpatients.

Many relevant elements were not brought into the discussion, including the study showing a reduction in mortality in nursing homes in the Marseille area, at the initiative of Prof. Raoult and IHU Marseille, the study by Derwand et al of Dr Zelenko’s patients, and the study by McCullough et al, of Dr Procter’s patients, showing a considerable reduction in the need for hospitalization with early treatment, to name just those.

So here is a question: was this videoconference a real academic discussion, going into the depth of the issue and attempting to find concrete responses and solutions, or was this kind of a (diplomatic) dialogue of the deaf?

Maybe it’s time now for Professor Harvey Risch, Dr Peter McCullough and others to invite Professor Ioannidis to debate this critical question of early outpatient treatment.

It’s doubtful that they will let Professor Ioannidis get away with his 100,000 hydroxychloroquine deaths claim, and his assertion that “it’s very difficult to identify specific interventions with clear benefits” for outpatient treatment.