Claim that French study showed long COVID doesn’t exist misinterprets the study and fails to account for its limitations

An article published on Substack in early November 2021, a platform that supports subscription newsletters, claimed that a French study^[1] demonstrated that long COVID is “mostly a mental disease”. Long COVID is a condition in which COVID-19 survivors continue to experience symptoms of illness weeks to months after recovering from the initial infection. These symptoms include fatigue, difficulty breathing, and difficulty thinking clearly. Anyone who was infected with SARS-CoV-2 can develop long COVID, although the condition tends to develop in people who had more severe disease^[2]. This type of condition is similar to another illness that is also observed following infection by bacteria or viruses, called myalgic encephalitis, more commonly known as chronic fatigue syndrome.

The study referenced was published by Matta et al. in the journal JAMA Internal Medicine^[1]. It conducted a survey of more than 26,000 people in France and examined whether a person’s belief that they had COVID-19 or their infection status, as determined by a serology test, was associated with persistent symptoms linked to long COVID. To do this, researchers first conducted a serology test using dried blood spot samples taken from volunteers between May and November 2020. Then between December 2020 and January 2021, all participants were asked to report whether they thought they had experienced COVID-19 infection along with physical symptoms of the disease in the past four weeks and that had persisted for at least eight weeks.

The authors reported finding that a person’s belief that they had COVID-19 was associated with the presence of persistent physical symptoms In the study’s Abstract, which provides a summary of the study’s findings, the authors concluded that “persistent physical symptoms after COVID-19 infection may be associated more with the belief in having been infected with SARS-CoV-2 than with having laboratory-confirmed COVID-19 infection”.

The study was interpreted in certain quarters to mean that “long COVID doesn’t exist” or that it is “driven mostly by hysteria”. The study It received more than 2,800 engagements on social media, including Facebook and Twitter, according to the social media analytics tool CrowdTangle. It was notably shared by groups expressing COVID skepticism and COVID denial (see examples here and here).

Such claims form a part of the false narrative that COVID-19 isn’t a public health concern—indeed, over the course of the pandemic, multiple claims dismissing the severity COVID-19 have made frequent appearances, ranging from the false claims that there aren’t actually many cases of infection because PCR tests are unreliable, that COVID-19 is like the seasonal flu, that COVID-19 didn’t actually lead to more deaths, and that most people classified as COVID-19 deaths didn’t actually die from COVID-19. Overall, these false claims prop up a fabricated narrative to justify the actions of certain groups that oppose public health measures like COVID-19 vaccination and mask-wearing.

To ascertain the accuracy of the claims made in the Substack article as well as the broader claim that the study shows long COVID doesn’t exist, Health Feedback reached out to the corresponding author of the study, Cédric Lemogne, a professor and head of the division of adult psychiatry at the Hôpital Hôtel-Dieu, Paris.

In an email to Health Feedback, Lemogne refuted these claims and pointed out several inaccuracies in the Substack article and social media posts.

“Our results do not demonstrate that long Covid doesn’t exist,” he clarified. “Instead, they strongly suggest that these persistent symptoms after COVID-19 (that are real symptoms for sure as they are experienced) may have other causes than SARS-CoV-2.” In other words, the findings suggest that persistent symptoms associated with long COVID aren’t unique to long COVID alone, and that one needs to rule out other causes before diagnosing long COVID. But the study wasn’t designed to determine whether the condition long COVID exists.

He also pointed out that the Substack article inaccurately attributed long-term symptoms to a patient’s belief that they had COVID-19: “Our study is observational, so symptoms may produce the belief, rather than the opposite. This is even stated in the Discussion section of our study.”

Experts interviewed by the U.K. Science Media Center regarding the study also highlighted this limitation. Kevin McConway, emeritus professor of applied statistics at the Open University, stated that “it’s never possible to be certain about what causes what in observational studies.”

Lemogne highlighted another inaccuracy in the Substack article, which claimed that patients didn’t know about their serology test results when they were asked about whether they believed they had COVID-19 and about their symptoms. In fact, the study’s Methods clearly stated that all participants knew about their serology test results by the time they were surveyed.

Health Feedback also reached out to other researchers studying long COVID who weren’t involved in the study. David Strain, a clinical senior lecturer at the University of Exeter Medical School who studies long COVID in addition to other chronic post-viral illnesses, said that the study’s design don’t support the claims made on social media.

“Long Covid is not based on any single symptom, but the aggregation of symptoms in recognisable patterns. This study did not make any attempts to explore whether total symptom burden was related to positive serology,” he explained.

He also added that because the number of people experiencing each symptom was very small, it is “unlikely to be able to confidently say there is no association, just that [the study is] underpowered to detect it.”

Other scientists raised concerns about the study’s use of a serology test to confirm COVID-19, as serology testing isn’t sensitive enough to detect past or present infection. Nisreen Alwan, an associate professor in Public Health at the University of Southampton, also explained in a Twitter thread that “having Long Covid is in itself associated with weaker antibody response to infection”^[3]. This suggests that previously infected people could have been incorrectly classified in the study as being negative for prior or present SARS-CoV-2 infection.

Furthermore, a study documented how a third of people infected with SARS-CoV-2 don’t generate a measurable level of antibodies against the virus as tested by serology^[4]. This further compounds the problem of misclassifying previously infected people as negative for SARS-CoV-2 infection.

Another potential limitation of the study is that it didn’t account for the potential impact of false positives, said Conway. “With the assumptions [the researchers] make, there would be rather a lot of false positives […] about four in every ten people who test positive for antibodies. That is, quite a big proportion of those who had a positive serology result would not in fact have been infected—and in a study that is comparing how closely test results align to symptoms, false positives can matter too,” he explained.

Overall, claims that the study by Matta et al. shows long COVID is a mental illness or that it doesn’t exist misrepresent the conclusions of the study and fail to account for its limitations, such as its use of a low-sensitivity test to detect a person’s history of infection.

Regarding the claims in the Substack article:

The article states that “because it looks like “thinking you had covid” produces pretty much all the long term symptoms. having actual covid does not.” But the word “produces” suggests causal links, whereas our study is observational, so symptoms may produce the belief, rather than the opposite. This is even stated in the Discussion section of our study.

The article also claims that “before any knew their serology outcome, they were asked about whether they thought they had had covid and then about a battery of long term symptoms.” This is wrong as they all knew their serology results before being asked whether they thought they had had COVID-19. This is clearly stated in our Methods and mentioned later in the Discussion section^[1].

Regarding the claim that our study shows long COVID doesn’t exist:

Our results do not demonstrate that “Long Covid doesn’t exist”. Instead, they strongly suggest that these persistent symptoms after COVID (that are real symptoms for sure as they are experienced) may have other causes than SARS-CoV-2.

Regarding the claim that our study shows long COVID is “mostly driven by hysteria”:

I do not understand the interpretation of this person. Our results may be explained by the fact that persistent symptoms after COVID could be erroneously attributed to the virus while having another cause. This cause might be another disease (e.g. cancer, depression, another infection, etc.). The persistent symptoms could also be explained by the influence of belief on perception. I guess this is what is referred to as “hysteria” here, but the use of this term has been abandoned in medicine for decades.

Also, I do not understand how it is that our study shows that “most ‘long haulers’ are physiologically induced”, as claimed. This interpretation makes no sense to me.

To make a long story short, our results suggest that symptoms attributed to long COVID may have other causes, warranting a careful evaluation. These other causes might involve cognitive and behavioral mechanisms, but also organ damage due to other causes than SARS-CoV-2.

Firstly, long Covid is not based on any single symptom, but the aggregation of symptoms in recognisable patterns. This study did not make any attempts to explore whether total symptom burden was related to positive serology. Further, the very small numbers of people experiencing each symptom means it is unlikely to be able to confidently say there is no association, just that they are underpowered to detect it.

Secondly, the authors of the study seem to imply—based on the statement “persistent physical symptoms after COVID-19 infection may be associated more with the belief in having been infected with SARS-CoV-2 than with having laboratory-confirmed COVID-19 infection”—that the belief is the trigger of the symptoms. This suggests somehow that this is purely psychosomatic disease. Given the widespread acknowledgement and acceptance of long Covid, it is far more likely that people who experienced persistent “long symptoms” would attribute the initial causative infection to COVID-19 rather than any other virus, given the stigma and lack of awareness of other causes of post-viral fatigue, as experienced for decades by people suffering with Myalgic Encephalomyelitis (ME).

And finally, the claim that long Covid is no worse than long flu (although it’s way more likely that they mean long-term infection by the Epstein-Barr Virus, the cause of glandular fever and most commonly attributed as the cause of ME) actually begs the question: “Why have we not been doing more for those with ME?”

Whether you have just been run over by an Audi, VW or Skoda, the nature of the vehicle that hit you is irrelevant when the focus needs to be on mending the broken leg it caused.

REFERENCES

• 1 – Matta et al. (2021) Association of Self-reported COVID-19 Infection and SARS-CoV-2 Serology Test Results With Persistent Physical Symptoms Among French Adults During the COVID-19 Pandemic. JAMA Internal Medicine.
• 2 – Sudre et al. (2021) Attributes and predictors of long COVID. Nature Medicine.
• 3 – García-Abellán et al. (2021) Antibody Response to SARS-CoV-2 is Associated with Long-term Clinical Outcome in Patients with COVID-19: a Longitudinal Study. Journal of Clinical Immunology.
• 4 – Liu et al. (2021) Predictors of Nonseroconversion after SARS-CoV-2 Infection. Emerging Infectious Diseases.