Revisiting the “painful truth of Long COVID“
This commentary article was originally published on Todd Davenport’s Substack. The Sick Times has republished it with permission.
There is a recurring pattern in the history of medicine that still pops up from time to time. When our understanding of a disease lags far behind the need for explanations, biology becomes a matter of one’s own perception. This inevitably shifts the burden of proof back onto the people we should be helping; those who suffer with the disease.
A recent essay on Long COVID published in WIRED magazine leans into this familiar trope, reframing the condition as a matter of unrealistic expectations. The essay suggested that perhaps Long COVID is less about the well-documented biological disruptions taking place and more about how the brain interprets them. That if we can just change the brain’s interpretation, then we might change the experience and fix the problem.
The premise seems powerful. But it puts the cart before the horse. It’s an oversimplification that confuses a complex multisystem pathophysiology with psychogenesis. This essay did a speed-run of one of medicine’s oldest mistakes within the span of a mere 7,000 words.
Given the ever-increasing number and quality of studies documenting what we have come to understand about the pathophysiology of Long COVID, it is very difficult to think Long COVID is principally a disorder of perception. Instead of a disorder of perception, the existing scientific literature describes a multi-system disease of bioenergetic failure, immune dysregulation, and vascular dysfunction.
My words from the WIRED essay, in this (proper) context:
“People with these conditions [such as Long COVID] get one of two labels in the popular imagination: they get ‘lazy’ and they get ‘crazy.’”
Unfortunately, the WIRED essay did nothing to dispel this false dichotomy. It told the story of Long COVID as a mind-invented illness in need of microdosing a little optimism. This essay will discuss Long COVID as a failed physiological struggle to re-establish equilibrium after disruption from the SARS-CoV-2 virus.
The brain operates within a physiological milieu
The argument in favor of Long COVID mind-body treatments seems initially plausible because it comes from two true claims. To the uninitiated reader, the logic built on these claims looks like careful, impartial thinking because the initial claims are true.
Here’s why the logic doesn’t stack up.
The first claim is that the brain constructs experience. Following a mind-body conceptualization of the Long COVID, pain, fatigue, and breathlessness must then be interpretations. Neuroplasticity, a real trainable structural and functional adaptation of neurons and neuronal connectivity, suggests a nervous system maladaptation that may be unraveled with the right combination of stimuli.
But the nervous system does not function like a computer. It does not independently process signals from other parts of the body in isolation. Instead, the brain functions within the context of physiological constraints. The physiological constraints of Long COVID occur at multiple levels. Immune dysregulation reshapes neural signaling through cytokines, neuroinflammation, and poor viral clearance. Vascular dysfunction impairs blood flow, which limits oxygen delivery and removal of wastes. Mitochondrial impairment causes wild fluctuations in the energy available for neural functioning. Abnormal autonomic signaling scrambles homeostatic responses.
A brain operating in the context of inflammation, hypoperfusion, dysautonomia, and energy deficit will not produce the same perceptions and experiences as one functioning in a more normal internal environment. That is because perception is not simply constructed; it is biophysically constrained. These biophysical constraints, in turn, are the root causes of the symptoms, signs, and disablement of Long COVID.
One important characteristic of Long COVID among many is a disturbance in energy metabolism. Studies consistently show mitochondrial dysfunction and impaired oxidative metabolism in Long COVID. When the body cannot reliably generate energy, exertion becomes destabilizing. These findings should reframe our understanding of exhaustion from a vague perceptual problem to the consequence of an energy supply that cannot meet the energy demand.
The essay suggested brain retraining for Long COVID is like stroke rehabilitation. This caught my attention as a longtime physical therapy faculty member who is trained, licensed, board certified, and experienced as a physical therapist and who teaches and researches physical therapy. However, my opinion was not quoted for this part of the essay.
There are two main differences between how stroke and Long COVID affect the brain. First, a stroke is a localized injury that can be remodeled around, where Long COVID is a generalized phenomenon of inflammation and poor perfusion. Second, rehabilitation starts after the acute stroke is over. In brain retraining for Long COVID, we are heaping load on top of dysfunction, because the core pathophysiology characterizing Long COVID has not been stopped. It is, again, putting the cart before the horse.
My own words from the essay, this time in proper context:
“If we overload the nervous system, say with neuroplasticity training [or] cognitive behavioral therapies, then we’ll end up with the same problems that we did with physical exertion … If these were therapies that worked, they would’ve worked by now, the several trials that have been done would’ve yielded very convincing results. And we just haven’t seen them.”
People living with Long COVID are not over-reacting. Signs and symptoms are not a misinterpretation or a misattribution. They are important signals of exceeding physiological limits. The nervous system, in distress itself, enforces those limits. All this internal chaos reflects a nervous system protecting itself from further physiological compromise. Effective training can only begin after putting a stop to the physiological compromise.
All this internal chaos reflects a nervous system protecting itself from further physiological compromise. Effective training can only begin after putting a stop to the physiological compromise.
Correlation is (still) not causation
The second true but mischaracterized claim is that some people with Long COVID improve while doing mind-body interventions. Improvement associated with brain retraining was held up as a rationale there must be something to it. That we must discuss it. That we must study it. That failing to do both is hurting people and holding back progress. That we’re missing an important part of the mechanism. And all of this is happening unfairly because of gatekeeping on the part of overly vocal patients, advocates, and researchers.
A main theme from the essay was that some patients are heard while others are unfairly rejected. The essay deployed the important clinical lesson to Believe patients as an interesting rhetorical device that I commonly see used among mind-body proponents and apologists.
Believe patients was used as a rhetorical cudgel against those who would ignore people whose improvements they’ve associated with mind-body techniques. “Well, if you’re listening to patients, why aren’t you listening to these patients?”
And Believe patients served as a rhetorical poison pill insulating the essay against critique. Criticizing the arguments becomes, by convention, attacking patients, and standing in the way of progress. Critique is taken as indirect proof the argument must be true. It displaces the blame back onto patients by serving as the basis for policing tone and content. Dissenters even become, according to some social media commenters, “bad faith, hysterical, shit-stirring chaos agents.” “See? You’re criticizing patients. You’re making this conversation difficult. It must be your fault why we’re not getting anywhere.”
Sidestepping this clever rhetorical device and re-entering the very practical realities of clinical and scientific life, our guiding philosophy is to Believe patients—and to always “trust but verify.” Patients are always the experts of their lived experience. They may not be (although frequently are) the experts in the underlying physiology. Through this lens, there are two important reasons why broadly attributing causation to observed improvements with mind-body interventions doesn’t hold up.
The first reason is methodological. As the essay pointed out, research into mind-body interventions is characterized by a high risk of bias. This indicates expectation effects and regression to the mean are difficult to disentangle from physiological change. Imagine a study. A researcher asks someone how they feel at the beginning of the study, and then spends the next two months telling them their pain is coming from their brain and that they are feeling better. When the researcher asks them again how they feel at the end of the study, we would not be surprised that a non-trivial number of people would say their pain is coming from the brain and they feel better. After all, the researcher spent those two months training the participant what the acceptable outcome of the study would be. Without objective measures obtained in the context of a careful methodology, claims about efficacy and causation quickly fall apart.
The second reason is conceptual. Improvement with mind-body interventions does not mean those interventions resolved the underlying pathology. In heterogeneous conditions with variable courses like Long COVID, some degree of improvement may be expected in some individuals regardless of intervention. A key question, regardless of the intervention under study, is whether the intervention modifies some aspect of the disease process. And, at present, there is no evidence that mind-body interventions reverse any of the core biological abnormalities identified in Long COVID. Happier thinking has not yet been shown to reduce viral loads, calm cytokine storms, improve endothelial function, nor make more productive mitochondria.
In science and in clinical work, we do not always have the luxury of simply taking things at face value. I am always filled with gratitude when people improve and regain at least some measure of their pre-illness lives back. But I am frustratingly slow to attribute causation to why they improve. My habit of trusting but carefully verifying has sometimes made me unpopular among researchers and patients alike. Yet, as a researcher who knows what he studies and has felt completely free to do it for a long time, I always encourage caution in attributing causation, especially when it comes to championing a unifying explanation that would explain all cases of improvement, even when it’s unpopular.
Here, again, is one of those times.
I am always filled with gratitude when people improve and regain at least some measure of their pre-illness lives back. But I am frustratingly slow to attribute causation to why they improve.
Uncertainty does not always require equipoise
Exercise is frequently used as a tool to microdose optimism in mind-body interventions. The thinking is that if your get-and-go has got up and left, the best way to get it back is to go after it. Eventually your body will catch up. Systems will recondition. You will become more confident. Life will return to normal again.
Seems intuitive, right? Not so fast.
Data from two-day cardiopulmonary exercise testing (CPET) already has demonstrated that patients with Long COVID, on average, exhibit a predictable and reproducible impairment in energy metabolism. These findings include decreases in the body’s oxygen consumption and work rate that we do not see in people who are otherwise healthy but out of shape. And, based on the measurements the studies used, none of these findings are subject to cognitive control.
Any claim that exercise still needs to be tested is not made from a position of impartiality. These claims are made from the decision to ignore the existing literature. Continuing to treat this issue as an unresolved empirical question simply re-manufactures a sense of mystery. And that manufactured mystery, in turn, is used to justify trials that expose participants to an intervention we already know commonly causes harms.
The manufactured mystery of exercise in Long COVID is further facilitated by critical flaws in research design. Studies consistently fail to adequately screen for people with Long COVID most at risk for post-exercise worsening. (We describe this problem in a manuscript we’re preparing.) By enrolling heterogeneous samples without distinguishing between who may or may not worsen, these studies mix people with different physiological responses into a single group. The result is combining responders and non-responders without really knowing who’s who. In these studies, even modest improvements in the responder subgroup can cover up worsening in the non-responder subgroup. Selected objective outcomes, such as activity counts from wearable sensors, run the risk of being self-fulfilling prophecies. Thus, average benefits may reflect statistical artifacts instead of genuine efficacy. Even trials aiming to measure post-exertional worsening mostly do not use the tools created to measure what they say they are measuring.
All of this matters from a practical perspective. The consequence of poor-quality exercise research is a body of evidence that seems to support exercise but consistently diverges from the lived experience of many patients with Long COVID. I haven’t met someone with Long COVID who has not yet tried exercising before consulting with me, a physical therapist. Most patients knew something was wrong even before someone could give it the label of Long COVID, because they worsened with exercise first. If exercising was the cure, the evidence we have for it should look a lot more definitive by now.
All of this also matters from an ethical perspective. The justification for broadly prescribing or testing exercise should necessarily become more stringent because we should be building on existing knowledge that a large subset of patients has a compromised physiological response to exertion. Manufactured mystery does not permit us to ignore known risks simply because benefits may exist for some. And it does not permit us to reframe a physiological limitation into a perceptual problem to sustain a research agenda. Evidence of potential harms could be why the essay noted institutional review boards now carefully scrutinize exercise trials. After all, weighing benefits and harms, particularly involving vulnerable participants, is their one job.
Speaking from experience in this clinical and research space, it can be an uneasy thing to hold two seemingly opposing ideas in your head. But it is something we must do here. We can rely on our rapidly accumulating knowledge base and use it to make decisions supporting our patients, even while acknowledging there is much we have yet to uncover. And in the process, we must avoid manufacturing mystery where it does not genuinely exist.
Any claim that exercise still needs to be tested is not made from a position of impartiality. These claims are made from the decision to ignore the existing literature.
The mind-body trap in Long COVID
Long COVID, and many of the patients living with it, are currently ensnared in what I call the mind-body trap. Sometimes that trap is set in essays written for large outlets. But, most often, the trap is sprung more quietly every day in countless dead-end visits with clinicians.
The mind-body trap is the false hope of maybe improving someday with the right mindset or cognitive training. Conversely, the trap is that failure to improve is because your thoughts just weren’t happy enough or didn’t perform the training correctly. A key part of how the trap works is everyone having it both ways, except for the person living with the disease—who gets none.
But, as we have seen, the mind-body trap in Long COVID is simply two large logical fallacies in a trench coat.
Well before COVID, the mind-body narrative long has misinterpreted the role of behavior in diseases contained within the diagnostic realm of Long COVID. The concept of kinesiophobia, or fear of movement, has been used to suggest that patients avoid activity in ways that perpetuate disability. Kinesiophobia suggests that, over time, the avoidance becomes the self-perpetuating reason for the disability, along with attendant pain, anxiety, and depression. These ideas form much of the basis for contemporary brain retraining methods, echoed in Sarno and others. However, the empirical record does not consistently support the kinesiophobia model.
Over two decades ago, a study in myalgic encephalomyelitis/chronic fatigue syndrome, one prominent subset of Long COVID, showed no meaningful relationship between fear of movement and exercise capacity, activity limitation, or participation restriction in people with chronic fatigue syndrome. We and others consistently have measured physiological deficits during maximal effort physical exertion, followed by prolonged symptoms and delayed recovery. Even our new data in Long COVID builds on themes that were already very well established before COVID and addresses the severe flaws of previous studies.
A good barometer of someone’s commitment to defending mind-body tropes is how they view the PACE trial. The more positively folks view the PACE trial, the more committed they seem to be in defending behavior interventions as a cure. This essay elided the catastrophic and well-documented methodological problems and research skulduggery associated with the PACE trial in favor of uncritically highlighting the rosy perspectives of the investigators about their own work. The essay treated the PACE trial’s exclusion from contemporary clinical guidelines as indirect proof of its validity because it was never retracted. Even before COVID, we pointed out this pitfall of failing to retract this poor quality work for newcomers trying to interpret this complex literature.
But there is no major conspiracy to be seen here. Sometimes the clinical and research establishment just gets it right and moves on.
People trying to live with Long COVID are not inactive because they are irrational. They are inactive because activity makes them worse. Activity avoidance, in this context, is not maladaptive. Abstaining from exercise is not the risky proposition that less-experienced investigators assert. It is self-protection. The kinesiophobia model doesn’t hold because it reverses causation in Long COVID, treating the consequence as the cause and placing the cart, once again, before the horse.
Abstaining from exercise is not the risky proposition that less-experienced investigators assert. It is self-protection.
Some concluding thoughts
History cautions us to avoid leaping from this is complex and we lack treatments to we must need behavioral therapies. I can not imagine the distress of living with Long COVID, even though I see it in my friends, colleagues, patients, and research participants every day. Patients often need psychosocial supports. But they do not need therapies suggesting their distress is the main cause.
Not all that long ago, diseases like cancer, multiple sclerosis, and rheumatoid arthritis had few effective treatments. Their mechanisms were incompletely understood. We now understand enough about their biology to have developed treatments that modify their disease processes, and in some cases, prevent and cure them. Based on these historical precedents, I expect that consciously sidestepping the mind-body trap in Long COVID will yield the same results.
It is true that Long COVID is complex. There is a lot to know already and still a lot of unanswered questions. Yet, the vastness of science in Long COVID is not a blank slate to be painted over with the opinions of a select group of people without independent verification. Again, we “trust but verify.” And, so far, the bulk of the verifiable evidence tells us that to treat Long COVID principally as a behavioral problem is to misunderstand it entirely.
We don’t know everything. But we know enough to start helping patients, and to ask what research questions will get us to tests and cures.
Misperceiving Long COVID as a problem of misperception has real consequences. It encourages interventions that push beyond physiological limits. It reframes protective strategies as maladaptive. It unfairly calls into question the believability and motives of millions of patients who are suffering. And in so doing, it distracts the narrative away from help and toward additional marginalization.
Cognition is not the root cause of Long COVID. If we want progress for the patients we serve, we need to avoid the tempting logical fallacies of the mind-body trap. Instead, we should consider cognition in its proper context and continue following the physiology where it leads.
This commentary article was originally published on the Substack of Todd Davenport, a physical therapist and exercise scientist who specializes in Long COVID and ME.
