Does Long Covid depend on our beliefs?
Researchers found that testing positive for Covid-19 is less predictive of Long Covid symptoms than what a patient believes about having been infected.
Foreword: Long Covid is real. But we are still learning about it. Many people have already interpreted the study I am covering as evidence that Long Covid is a psychiatric or "imagined" illness. That’s simply not what the study says. It does, however, tell us that we have a lot to learn. I’m providing my honest assessment of the data, and yet I still worry some people who have Long Covid will find this study, and therefore my coverage of it, hurtful. As a physician, my goal is to shed light on what we can learn from the new data to help people recover. I anticipate and hope that many readers will alert me to things I haven’t considered and that the researchers did not think of either, via the comments section below. The fact is, we still don't truly understand Long Covid. So any conclusions any of us reaches are bound to be interim ones, at best. -JF
Recently, a group of French scientists conducted a bold experiment regarding Long Covid. The researchers studied nearly 27,000 participants, asking them in late 2020 and early 2021 whether they believed they had contracted Covid-19 during the first 9 months of the pandemic and about 18 categories of new persistent symptoms commonly reported by Long Covid patients. The novel achievement here, though, was that the survey results were then correlated to SARS-CoV-2 antibody test results, which indicated whether each participant had been infected during the study period. (The volunteers had already submitted blood tests and were sent the results before they took the surveys.) Synthesizing the data allowed the investigators to tease out whether Long Covid symptoms were associated with a belief in having contracted Covid-19 (via a previous positive test or a physician's diagnosis, or in the absence of either) or with actually having antibodies that reflected a prior infection.
The results of this study, published in JAMA Internal Medicine, are remarkable on several fronts. The top-line finding is that a belief in having had Covid-19 was statistically associated with a variety of Long Covid symptoms, ranging from difficulty breathing to poor concentration to fatigue, regardless of whether there was laboratory evidence of a prior infection. In other words, belief in prior infection mattered more than actual prior infection, as determined by the confirmatory antibody testing results.
This matters for two reasons. First, it's possible that some people have misattributed new symptoms to Long Covid, when another disease is to blame. In rare cases, those diseases could even be serious ones. For example, imagine someone with persistent nausea or fatigue convincing themselves (and even their physicians, who are indeed "suggestible") that they have Long Covid, when the real culprit is an undiagnosed auto-immune disease or cancer which that person had already begun to develop? Second, the findings imply that among those with asymptomatic or mild infections, Long Covid may be less common than we feared. That suggests that the extent of the psychiatric trauma of the pandemic remains incompletely understood, let alone treated. The data from this new study remind us that we have to investigate all symptoms carefully, even in the midst of a pandemic.
Indeed, among participants who believed they had had Covid-19, regardless of their test results, 15 of the 18 categories of Long Covid symptoms were statistically more frequent than in those who didn’t believe they had had it. Meanwhile, regardless of belief, 10 of the 18 categories of Long Covid symptoms were more likely to be present among those with positive antibody tests. Finally, when the researchers analyzed the data to factor out the influence of the participants' beliefs, only one category of symptom was statistically linked to a prior infection: loss of smell (“anosmia”). This is conspicuous because of all of the symptoms included, anosmia is perhaps the one “hallmark”, or most distinctive symptom of Covid-19. Other linked symptoms, like headache, dizziness, cough, and chest pain are common to many viral or post-viral syndromes and are not particularly native to Covid-19. The bottom line is that believing you had had Covid-19 was more predictive of Long Covid symptoms than was having verifiable evidence of prior disease.
•••
One aspect of this study that has received less attention is the striking disjunction between participants' beliefs about whether they had had Covid-19 and the results of their antibody tests. The study says that participants received their antibody test results before filling out the survey; it is not clear whether some participants simply disbelieved the antibody results, didn't look at them or understand them, or that they answered the survey questions based on their state of mind prior to receiving the results. Regardless, the gulf between what the participants thought had happened to them, and what researchers were able to determine had happened, was goodly in size.
Of the 914 participants who believed that they had been infected, just under half (453, or 49.5%) actually had SARS-CoV-2 antibodies. Yes, it’s possible that some of the 461 who believed they had been infected but who had no antibodies were correct. Yes, the antibody tests only detect 87% of previous infections. But accounting for that means that it is statistically likely that this would only add back around 60 people to the group of participants who correctly believed they had previously been infected. That would still mean that around 44% of those who believed they had had Covid-19 apparently had not. Granted, only 1.5% to 1.7% of the participants in the entire study ultimately fell into the category of believing they had previously had Covid-19 despite having no laboratory evidence to support that conviction. But if that rate applies to the entire population of France, there could be hundreds of thousands of people there who believe they have some degree of infection-derived protection from the virus but who are likely incorrect.
On the flip side, out of the 1,091 participants who tested positive for antibodies, 58% said on the survey that they did not believe they had been infected. In other words, only 42% of those who were infected knew it. Again, it is unclear why this number remained so low even after participants received their antibody results. Do people have so little faith in science, or so little ability to read test results? Or again, were they answering based on their belief prior to receiving the test results (even though I have verified with the study authors that the participants were instructed not to)? One thing we can say is that the discord does comport with other data suggesting that the number of actual infections is, in many regions, at least twice the number of reported ones. Meanwhile, I confess unease that some have criticized this study on the basis that antibody tests aren't reliable enough to rely on. That alone is cause for reflection. While antibody tests are not perfect, they are not that bad.
•••
Where does this study leave us? Is Long Covid real? Yes. There’s more than enough data from other studies to tell us that Covid-19 does not leave the body unscathed. Nor do I believe that anosmia is the only symptom of Long Covid. (And despite the results above, that is not what the authors of this new study really say.) Remember, 10 categories of symptoms were statistically more likely in the participants who had Covid-19 antibodies. It’s possible, if not likely, that the severity of those symptoms was simply greater among those with laboratory evidence of prior infection. We’ve seen this kind of thing before. In the vaccine trials, around 25% of trial subjects who unknowingly received a placebo shot reported fatigue, compared to over half among those receiving an actual Pfizer or Moderna shot. I’d venture that not only the rate but also the severity of the reported fatigue was greater among those that got the real shot. I feel fatigue on many days. But I was really, really tired the days after I got vaccinated.
Also, while I can’t speak to every Long Covid symptom, my personal experience treating Covid-19 is that the chest X-rays and CT scans I’ve seen, even among patients with very mild symptoms, are often so impressive that I can’t imagine that these patients would not have some downstream problems, whether triggered by some future infection or an asthma or emphysema flare-up.
But it’s also likely that for at least some if not many Long Covid patients, the trauma of the pandemic—fortified by a belief of having been infected, accurate or not—is a major contributor to the frequency of symptoms.
•••
This study does not say that Long Covid symptoms are simply imagined. The stress that we have all been under is apt to wreak all kinds of havoc on the body, some of which may generate both measurable and unmeasurable effects on the body. And in fact, I even wonder whether there may have even been a degree of “Long Covid denial” among the participants in the study. Again, 58% of the participants found to have Covid-19 antibodies (which was 2.4% of the study participants overall) reported that did not believe they previously had Covid-19, despite being provided with antibody evidence to the contrary. Why is that? Did some of these patients simply downplay or ignore symptoms that were there? My personal version of this experience is that I did not have any symptoms from a groin hernia until it was pointed out to me during a pre-employment check-up. From that day forward, I had mild discomfort every day until I had an operation to repair it. Ignorance was bliss, but only to a point. Eventually I was going to need that operation.
It’s likely that Long Covid has more than one cause. For many, the cause is SARS-CoV-2 itself. For others, the symptoms likely result from the trauma of the pandemic, alone or in combination with other diseases (medical or in some cases psychiatric) that have gone undiagnosed during the pandemic, an alarming alternative that the authors hypothesized in the paper. I do wonder whether informing patients that at least some Long Covid symptoms are driven by beliefs would help reassure a subset of them, and therefore help them recover sooner. So, yes, in some cases, Long Covid symptoms might be controlled or improved by changing our perceptions around it. Even if that’s the case, though, there will still be much work to be done to help the remaining Long Covid patients, of which, unfortunately I anticipate there will be many.
•••
❓💡🗣️ Questions? Comments? Join the conversation below!
Follow me on Twitter, Instagram, and on Facebook and help me share accurate frontline medical information!
📬 Subscribe to Inside Medicine here and get updates from the frontline at least twice per week.
Thanks to Dr. Céline Gounder and to the lead author of the JAMA Internal Medicine study, Dr. Cédric Lemogne for insights and comments.