Does Long COVID Cause Anxiety, or Does Anxiety Cause Long COVID? What the Sequence Shows
Two prospective cohorts measured opposite directions, and each found an association. What the order of events does and doesn't say about your symptoms.
Photo: Aurélia Chevreul-Gaud / Unsplash
Researchers have measured both directions in people, and each study found an association. Adults with long COVID in a Michigan cohort were about twice as likely to report COVID-related post-traumatic stress symptoms more than a year later. A separate group of researchers examined the reverse relationship and found that psychological distress before infection was associated with more long COVID afterward. Your symptoms aren't imagined under either finding, and the order of events is what makes that clear.
What the Michigan cohort followed
The Michigan COVID-19 Recovery Surveillance Study enrolled 3,492 adults whose COVID-19 had been confirmed by PCR between March 2020 and May 2022. Everyone was surveyed twice. The first survey came a median of 4.4 months after their infection started, and the second a median of 18.4 months after.
Long COVID was defined at that first survey in ordinary words: not recovered to your usual state of health 90 days or more after the infection began. At baseline, 17.0% of the group fit that description.
At follow-up, 10.1% reported COVID-related post-traumatic stress symptoms. Those were measured with a six-item version of the PTSD Checklist for Civilians, anchored to each person's own COVID-19 diagnosis. That is a symptom questionnaire rather than a structured clinical interview, which means it identifies people who screen positive and not people who carry a diagnosis.
The researchers adjusted for sociodemographic characteristics, pre-existing conditions, how severe the acute illness had been, the survey mode, and the phase of the pandemic in which someone was diagnosed. In that fully adjusted model, the risk of COVID-related PTSD symptoms was 2.08 times higher in the long COVID group, with a 95% confidence interval of 1.65 to 2.63.
The study that examined the reverse relationship
A team publishing in JAMA Psychiatry in 2022 asked the question from the other end. They used three ongoing cohorts: the Nurses' Health Study II, Nurses' Health Study 3, and the Growing Up Today Study. Depression, anxiety, worry about COVID-19, perceived stress, and loneliness were all measured in April 2020, before anyone included in the analysis had been infected.
Of 54,960 participants, 6% reported a positive test during follow-up. Among those 3,193 people, every form of distress measured beforehand was associated with a higher risk of symptoms lasting four weeks or longer. Probable anxiety carried a risk ratio of 1.42. Probable depression was 1.32, and perceived stress in the highest quartile versus the lowest was 1.46. Participants with two or more types of distress before infection were at nearly 50% higher risk, at 1.49.
This cohort was 96.6% female, and 38.0% were active health care workers. Both the infections and the later symptoms were self-reported. That limits how well the numbers describe anyone outside that group.
Why both can be true without the illness being psychological
In clinic, I watch this go wrong from two opposite directions, and it is the same patient who pays for it either way.
One version is the doctor who hears that stress is involved, concludes the whole illness is psychological, and sends the person home. Usually that's a short appointment and a normal set of labs rather than unkindness, and it still leaves the person with nothing to work with.
The other version is harder to see. It is the clinician who cares a great deal, who is determined not to be accused of dismissing anyone, and who therefore won't raise the nervous system at all. The patient leaves with a physical workup and no acknowledgment that a year and a half of being unwell has done something to them. The person sitting between those two versions isn't served by either one.
A stress response is a physiological event. Stress hormones circulate and cells carrying receptors for them respond, which happens whether or not anyone has decided the illness is psychological. Describing something as a stress response is a statement about what the body is doing, and it isn't a statement that the person made it up.
What makes the difference here is the order of events. Long COVID was recorded at a median of 4.4 months, and the trauma symptoms at a median of 18.4 months. The reverse finding doesn't cancel that. Prior distress being associated with more long COVID puts distress alongside the other risk factors researchers examine, and it says nothing about whether the later symptoms are genuine. Both studies are observational, so neither the Michigan result nor the JAMA Psychiatry one establishes cause.
Where this leaves your own care
The Michigan authors concluded that mental health screening, monitoring, and interventions need strengthening among adults with long COVID. That reads as an addition to physical care rather than a replacement for it.
There is a distinction that tends to get lost here. Treating the psychological consequences of a long illness isn't the same as treating the illness itself with therapy. In ME/CFS, which parts of CBT helped and which one was linked to worse fatigue has been studied directly, and the answer wasn't uniform across the components. Support for what being sick has done to you is reasonable. Presenting that support as the treatment for the physical illness is a separate claim, and this pair of studies gives you no reason to accept it.
If getting anyone to take the physical illness seriously has been the harder problem, what 577 patients reported about getting care describes what actually stood in the way for them. There is more on post-viral illness in the chronic fatigue and ME/CFS resource hub.
All of this is educational, and it isn't medical advice. Nobody can tell you what applies to your own case without knowing your history. If chronic fatigue or EBV reactivation is part of your picture, the EBV Reactivation Treatment Algorithm is a step-by-step flowchart for working through it.
FAQ
Do these studies prove that long COVID causes anxiety?
No, they show association rather than cause. In the Michigan cohort, adults who had long COVID at about four months were 2.08 times as likely to report COVID-related post-traumatic stress symptoms at about eighteen months. A prospective sequence like that is stronger evidence about direction than a single snapshot, and it still leaves room for other factors influencing both.
Can stress before an infection make long COVID more likely?
In one large cohort it was associated with it. Depression, anxiety, worry about COVID-19, perceived stress, and loneliness measured in April 2020 were each associated with more symptoms lasting four weeks or longer among people infected afterward. The risk ratios ran between 1.32 and 1.46, and participants with two or more types of distress were at 1.49. That group was 96.6% female, and 38.0% were active health care workers.
Is long COVID a psychological illness?
No. Long COVID is defined by symptoms that persist after a confirmed infection, and the Michigan study measured trauma symptoms as something that came after it. Psychological symptoms alongside a physical illness are common across many chronic conditions, and their presence isn't evidence that the physical illness began in the mind.
Should someone with long COVID be screened for PTSD symptoms?
The authors of the Michigan study concluded that mental health screening, monitoring, and interventions need strengthening for adults with this condition. Whether it applies to you is a question for your own clinician. A positive result on a screening questionnaire identifies someone who should be assessed properly, and it isn't a diagnosis on its own.
Does treating the anxiety improve the physical symptoms?
That wasn't tested in either the Michigan cohort or the JAMA Psychiatry analysis. Both were observational studies of association, which means they can describe who is at higher risk and not what happens when you intervene. Anyone telling you that treating the anxiety resolves the physical illness is going past what these two results support.
References
- 1.
- 2.
