An independent analysis of the data of the 30423 patients treated at the IHU-Méditerranée Infection gives even better results than those initially announced by the IHU scientists

Gathered around Professor Christian Perronne, Valère Lounnas (PhD in theoretical chemistry applied to proteins and former researcher at EMBL-Heidelberg), Pr. Eleftherios Gkioulekas (PhD, School of Mathematical and Statistical Sciences, University of Texas), Dr Marc Rendell, (MD Rose Salter Medical Research Foundation), Dr Alexis Lacout and Xavier Azalbert (econometrician, editorial director of France-Soir) independently and autonomously analysed the data of 30,423 patients cared after by the IHUM (Institut Hospitalo-Universitaire-Méditerranée Infection).

Two highlights: 

• The independent study of data from 30,423 patients treated at the IHUM, shows that the dual therapy used at the IHUM reduces the number of patients in intensive care and death by 58%.
• The AP-HM (Marseille Hospital), under the responsibility of François Crémieux, has given its consent to the use of the data for the purpose of scientific publications. This puts an end to the unsubstantiated claims of Professor Molimard who described this study as a "wild trial".

Professor Perronne declares:  

“The team did a very precise job, combining the mathematical and methodological rigor necessary for the analysis of a big data database with the essential clinical skills."

The link for the study is here. The document in pdf format can be downloaded by clicking here.

A study authorized by the AP-HM directed by François Crémieux

The patients analysed are those from the infectious diseases unit of the AP-HM (Assistance Publique - Hôpitaux de Marseille) treated at the IHU-Méditerranée during the acute phase of the pandemic (from March 2020 to December 2021). This important clarification is necessary to establish the responsibilities of each party, particularly in terms of care and clinical research. The AP-HM is therefore the administrative manager for care, data collection and its use. And it is with this in mind that the AP-HM, now represented by its Director General François Crémieux, has given its approval for the use of data for scientific publications in the interest of medical science. Authorizations 2020-151 and 2020-152, which France-Soir has been able to consult, establish that the studies of Pr Million and Pr Lagier on Covid are registered in the registry of treatment activities of the AP-HM for the period from April 27, 2020 extended to December 31, 2021. This is, of course, one of the missions of an internationally recognized centre of expertise such as the IHUM. These elements contradict the unsubstantiated assertions of Professor Molimard of the University of Bordeaux, an outspoken opponent of the IHU's work, who stated without factually substantiating it that these data constituted a "wild trial". Contacted by France-Soir several times, François Crémieux and Mathieu Molimard were never available to answer these specific points. In a letter dated February 2024, which we have been able to consult, the AP-HM withdrew the authorization to use the data, thereby confirming that there had indeed been an authorization. The results of the IHU study are so much in favour of early treatment, confirmed by the independent analysis of Lounnas et al, that the authorities seem to be going out of their way to ensure that these data are dequalified for research purpose with authoritative and circular arguments.

The results of the independant study, now published in a peer review journal confirm the results of the UHUM scientists

The data of 30,423 patients, after control by a bailiff, were published on a server and uploaded by the multidisciplinary team who spent several months analyzing them, methodically reviewing the various potential biases as well as all the factors that may affect the evaluation of the treatment within the limits of the data provided. It should be remembered that it was in October 2023 that the IHU researchers had republished their study  showing results clearly in favor of early treatment. It was republished as the researchers were placed under pressure with their first publication by the French authorities – the authors collectively decided to withdraw the preprint to protect their younger colleagues.

The results are indisputable in favor of the early treatments implemented at the IHU-Méditerranée (hydroxychloroquine + azithromycin or azithromycin with or without ivermectin) on the main endpoint combining transfer to intensive care and death.

The efficacy of dual therapy amounted to a 58% reduction in the risk of going to intensive care or death.

This is the first critical analysis published in a peer-reviewed journal, which confirms the results obtained by the IHU-Méditerranée researchers.

Summary of the research paper

"A cohort of 30,423 patients with covid-19, treated between March 2020 and December 2021 at the IHU-Méditerranée Infection in Marseille (France), was retrospectively analyzed in terms of intent to treat and disease aggravation factors, in order to quantify efficacy against the composite endpoint of transfer to intensive care or death, within two months (56 days) of admission.

Within the limits of the data and statistical models used, after adjusting for sampling biases, multivariate logistic regression analyses were performed to determine efficacy as measured by an unadjusted and adjusted odds ratio (OR) for the subset of patients who received hydroxychloroquine plus azithromycin combination therapy (HCQ-AZ) or no specific therapy (i.e., no HCQ, no AZ and no ivermectin) (24,943 patients).

Azithromycin without hydroxychloroquine but combined with ivermectin in 31.3% of cases was significantly active in the absence of specific treatment (unadjusted OR = 0.720, 95% CI = [0.574, 0.905] p = 0.005 and adjusted OR = 0.727, 95% CI = [0.608; 0.870] p < 0.001).

Interactions between dual therapy (HCQ-AZ) and model covariates were systematically explored. No interaction between HCQ-AZ treatment and vaccination was detected.  Statistically significant favorable interactions were detected between HCQ-AZ treatment and male sex, age categories ≥ 50 years, the UK variant and where the variant has not been determined, obesity, chronic obstructive pulmonary disease (COPD), cancer and immunodeficiency, confirming the high efficacy of this early treatment. No statistically significant adverse interactions of HCQ-AZ with any other covariate were detected.

The limitations of the models and their implications for the results are discussed in detail as well as the limitations of the IHU data

The conclusions of the research are: “State-of-the-art statistical analysis of the IHU Méditerranée data demonstrated the efficacy of the empirical treatment using a combination of hydroxychloroquine and azithromycin, given as an early treatment. Taking into account the very large size of the observational single institution cohort of patients coherently treated, together with the quality of the statistical approach we used, these results pose a serious challenge to those who have continuously denied the potential efficacy of hydroxychloroquine-based treatment of Covid-19 patients during the pandemic. We have confirmed the validity of the intent-to-treat approach, in times of urgency, with a combination of medications reasonably thought, in early 2020, as having a potential efficacy on the disease at hand. This work should provide an incentive for other independent researchers to conduct further analysis, possibly with more advanced methods.”

The non-profit Bonsens.org financed the publication costs and a computer at 829 euros to carry out the work.

Given the results of the IHU study and the independent analysis by Lounnas et al, a parliamentary fact-finding comission becomes essential in order to assess how many lives could have been saved with a systematic use of early treatments. The toxicity risks of hydroxychloroquine, azithromycin, ivermectin are exceedingly low, as has been observed over decades for these molecules (WHO's Vigibase lists only 114 deaths attributable to HCQ over a period of 50 years and tens of billions of doses). In addition, Emmerich confirmed this result by demonstrating in a peer-reviewed study that the state of Para in Brazil, having systematized early treatment, had 5.5 times fewer deaths than the neighboring state of Amazonas, which did not systematize them. The risk-benefit analysis is clearly in favor of the use of early treatments.

Estimation of the number of lives that could have been saved

Traduction en Anglais. It is therefore necessary for experts to look at the number of lives that could have been saved with the systematization of early treatment. 
In France, 167,642 deaths have been attributed to Covid (source: worldometer). Assuming that the cause of death is really Covid and not with Covid, and applying the 58.1% efficacy rate of early treatment obtained in the Lounnas et al study, a first estimate of the number of lives saved amounts to 97,400 (58.1%*167,642 deaths). The recalculated mortality rate in France would therefore amount to 1,038 deaths per million inhabitants (d/Mh) (167,642 - 97,400)/67,697 - the French population) against 2,476 d/Mh

Any estimate derived from a mathematical model needs to be calibrated to the real world in order to validate its relevance.  
With this aim in mind, let's go back to Emmerich's study which compares the Para state to Amazonas. In Para, early treatment has been systematized and there are 296 d/Mh or 2,575 deaths for a population of 8.7 million inhabitants for the period from November 11, 2020 to March 15, 2021. 
In France on November 11, 2020, there were 42,535 Covid deaths and 90,832 on March 15, 2021 (source worldometer). There were 48,297 deaths in the period. Applying the 58.1% efficacy rate, it is estimated that 28,061 deaths could have been prevented in the period. Covid deaths over the period would therefore have amounted to 20,236 and the number of deaths per million inhabitants would be 298.93 (20236/67.697). This figure is statistically no different from the figure of 296 d/mh in the state of Para.  This initial calibration work can be refined.

The total number of deaths attributed to Covid worldwide stands at 7,003,621 (source: Worldometer). By applying the efficacy rate of early treatment obtained in the Lounnas et al. study, we can estimate that 4,069,104 deaths could have been prevented. A conclusion that should be of great interest to the authorities, as well as the public in their quest to evaluate the effectiveness of health measures around the world.

Note: Questions asked to Mr. François Crémieux were as follows:  I am preparing an article on the publication on the 30,000 patients treated at the University Hospital of Marseille, which was the subject of a first retracted publication and a second publication by the professors of the IHUM.  Professor Molimard has spoken to the media explaining that this study would have required RIPH1 approval and states that "it is a wild trial".

1 – As this study is carried out on patients at university hospitals and therefore under your full responsibility, have you given permission to use this retrospective database? If so, on what date?  2 – Has IGAS asked you about this and made any recommendations on this subject?

3 – The professors of the IHUM have been attacked for this study, does the CHU defend them as the head of the AP-HM institution, the institution that employs them?

4 – When Mr. Molimard declares that it is "a wild essay", what can you say about it? Given that it calls into question your responsibility and the conclusions of IGAS?

5 – Is it planned by the Ministry of Health to carry out an audit to answer the question posed by MP Dupont-Aignan to the Minister of Health on the validity of the IHU results in terms of care and performance of treatment with hydroxychloroquine and azithromycin?