do “confidential documents” from the Pfizer laboratory call into question the efficacy and safety of its vaccine?

No, these documents have not “leaked”

Contrary to what is advanced by several Internet users, these documents have not “leak” but were released as a result of US legal proceedings, under the Freedom of Information Act. This 1966 law obliges federal agencies to release their documents to anyone who requests them.

In January, a federal judge therefore ordered the Food and Drug Administration (FDA), US food and drug administration, to make public within eight months the information on which it had relied to approve the Pfizer vaccine against Covid-19. The FDA never said it wanted to keep these documents confidential, but initially claimed a period of seventy-five years, which she justified by the few 450,000 pages of documents to be published, as explained by the media Bloomberg Law*.

The request was supported by the Public Health and Medical Professionals for Transparency collective, a group of researchers and scientists, many of whom have questioned the interest of vaccines against Covid-19. The collective has published some of the documents in free access on its website*.

No, they don’t reveal that “1,223 people” died during clinical trials.

Several publications, in French and in English*, claim that these documents show that “1,223 people have died” after receiving an injection of the Pfizer vaccine “during the tests”. These deaths would have affected “3%” from “42,000 trial participants”. These messages are often accompanied by a screenshot of a table.

These figures are well taken from a 38-page report (PDF)*, which is among the documents published on the Public Health and Medical Professionals for Transparency platform. But it does not concern clinical trials conducted by the laboratory. It is, as the title of the document indicates, the“cumulative analysis of reported adverse effects” to February 28, 2021, after the authorization of the Pfizer vaccine in several countries. The figure of 42,086 therefore does not refer to the number of participants in clinical trials, but to the number of pharmacovigilance reports reported on this date worldwide. You can read there that “most cases (34,762) were from the US (13,739), UK (13,404), Italy (2,578), Germany (1,913), France (1,506) , Portugal (866) and Spain (756); the remaining 7,324 cases were spread across 56 countries.”

Among the adverse effects noted, a large number concerned headaches (24.1% of cases), fever (18.2%) or fatigue (17.4%). In addition, the table actually reports 1,223 deaths. But the “methodology” part of the document specifies that the reports have been reported “regardless of the assessment of causation”. In other words: among these reports, the causal link with vaccination is not necessarily established. The figure of 3%, calculated by Internet users, can therefore simply return the number of death reports out of the total number of reports. This is in no way a mortality rate potentially linked to vaccination.

No, Pfizer does not advise against vaccination for pregnant or breastfeeding women.

Other publications, shared thousands of times, assure that Pfizer would have recognized the risks associated with the vaccination of pregnant and breastfeeding women. These messages are accompanied by a document in English, in which it is mentioned that the vaccine from Pfizer “not recommended during pregnancy”. Regarding breastfeeding, it is added that“we don’t know if the vaccine [de Pfizer] is excreted in breast milk” and that one “Risk to newborns/infants cannot be excluded.”

In reality, this document does not come from the American laboratory, but from an opinion issued by the British Medicines Agency*, at the very start of the vaccination campaign in the United Kingdom, in December 2020. At the time, these precautions were explained by the lack of data, pregnant women not having been part of clinical trials. Since then, studies conducted in real life, based on the medical monitoring of vaccinated people, have delivered their conclusions. “The safety of the vaccine in pregnant women is now fully established”reassures Sandrine Sarrazin, Inserm research fellow at the Marseille-Luminy immunology center.

“At the same time, numerous data have shown that being pregnant is an important comorbidity factor in the event of Covid-19 infection.”

Sandrine Sarrazin, researcher in immunology

at franceinfo

As for breastfeeding, “Iobservational data” don’t have “no evidence of adverse effects in newborns/infants” after the injection in the mother of an RNA vaccine, like that of Pfizer, exposes the National Agency for the Safety of Medicines (ANSM). On the contrary, “we now know that the antibodies generated by the mother pass into breast milk and provide effective protection to the newborn”, adds Sandrine Sarrazin.

The recommendations of the British authorities have also evolved. The UK Medicines Agency* now recommends vaccination for pregnant women “when the potential benefits outweigh the potential risks to mother and fetus”. Le National Health Service* (NHS), the British public health system, recalls that pregnant women run “a higher risk of becoming seriously ill from Covid-19” and that it is therefore “important to get vaccinated to protect them and their child.

In France, vaccination against Covid-19 “is strongly recommended for pregnant women”writes the ANSM. She adds that“to date, no signal has been identified in pregnant and breastfeeding women with all the vaccines against Covid-19 available in France.”

No, Pfizer’s vaccine does not have an effectiveness limited to 12%, or even 1%

“You were sold a 95% efficiency of the injection, but the ‘Pfizer documents’ reveal 12% efficiency over the first seven days, then 1%”, another Internet user assures us. This false information comes in part from a blog post* published on April 3 by Sonia Elijah, a “former BBC journalist evolving today in the Trumpist and anti-vax spheres”Explain The world.

In reality, this is not based on one of the “Pfizer documents”, but on a note from December 2020 directly available on the FDA website*. It reads that the effectiveness of Pfizer’s vaccine against infection is 95%, seven days after the second injection. A figure calculated from the number of confirmed Covid-19 cases in the vaccinated group (eight cases) and in the group that received a placebo (162 cases).

Screenshot in support, Sonia Elijah argues that it would be more relevant to observe the number of cases of Covid-19 “suspected but not confirmed”because even if these people have not tested positive, they have “true symptoms”, she wrote. From the number of suspected cases in the vaccinated group (1,594 cases) but also in the placebo group (1,816 cases), it deduces a vaccine efficacy of 12%. However, it is misleading to rely on these cases, since nothing confirms that these people followed were indeed all suffering from Covid-19 and not from another disease with similar symptoms (cold, flu, etc.).

Above all, numerous studies in the real population, based on the vaccination of millions of people, have since certified the effectiveness of vaccines, in particular against severe forms of the disease. The emergence of variants has certainly reduced the protective effect of vaccination against infection, but it is wrong to imply that in December 2020, Pfizer’s vaccine had an effectiveness limited to 12%.

As for the figure of 1%, this rumor is related to the existence of two ways of evaluating theeffectiveness of a vaccine. The first, the measure of relative risk reduction (RRR), is the method “most commonly used”explains Sandrine Sarrazin. It compares the risk of infection of a vaccinated group relative to the risk of an unvaccinated group of the same size. So, if the vaccinated had half as many positive cases as the non-vaccinated, then the RRR is 50%. This is how laboratories have determined the effectiveness of their vaccines during clinical trials.

The second method, less used, is the absolute risk reduction (ARR), which calculates the difference in risk of being infected between the two groups. The scientific and medical popularization platform Health Desk* warns against a defect of this technique: used in the case of a clinical trial, it “may make a vaccine less promising than the effect it would have on the risk [d’être infecté] of a person over a longer period”. As proof, Health Desk gives this example: imagine a clinical trial of a vaccine based on a sample of 20,000 participants and a control group of 20,000 members. No one would test positive in the vaccinated group, but 200 would in the unvaccinated group. The RRR of the vaccine would then be 100%, but the RAR would only be 1%.

Dyears a commentary on the use of statistics, published in April 2021 in the magazine The Lancet Microbe*, researchers calculated the following RARs for the main vaccines against Covid-19 : “1.3% for the AstraZeneca vaccine, 1.2% for Moderna, 1.2% for Johnson & Johnson (…) and 0.84% ​​for Pfizer-BioNTech”. But the authors do not use these statistics to conclude that vaccines are ineffective. As detailed on the fact-checking site of the German press agency DPA, the relative reduction in risk and absolute risk reduction are simply “two statistics that are based on the same data, but do not describe the same thing”.

*All links followed by an asterisk refer to content in English.