ABOVE: © iStock.com, Pandagolik

In December of 2020, Brooke Keaton, a 41-year-old preschool teacher in Charlotte, North Carolina, and her family fell ill. At first, she was congested and had a cough, but soon she developed a high fever and felt debilitatingly weak. “I should have probably gone to the hospital,” she recalls, “but I have a now-four-year-old and a twelve-year-old and I wasn’t about to leave my two babies at home.” Keaton had COVID-19. After a grueling two weeks, she recovered, “but I still didn’t feel right,” she recalls. Since then, Keaton has been battling a host of symptoms that have put her out of work. She felt her heart race even when she was sitting down. And though some of her symptoms have improved, she still fatigues easily and suffers from insomnia and memory loss. “Brain fog doesn’t even begin to cover what I felt,” says Keaton. “I have to set reminders for myself like I have dementia.”

As with many people living with long COVID, when Keaton first went to the doctor, her symptoms were dismissed as anxiety. It took months of self-advocacy before she could see even a single specialist. Since October of last year, over the course of more than 70 doctor appointments, she’s been diagnosed with arthritis, a heart condition, and bursitis in her hips. The causes of some of her most debilitating symptoms, including her brain fog, remain unknown. 

The US Centers for Disease Control and Prevention (CDC) estimates that one or more of a host of persistent symptoms arise in 20 percent of people in the US who’ve recovered from a SARS-CoV-2 infection, even when those infections were mild. A meta-analysis performed in April puts this figure globally at 43 percent, and the US Census Bureau estimates that as many as 4 million people in the country are out of work as a result. Despite crippling cognitive and physical symptoms, doctors are unable to find anything wrong with many long COVID patients. More than 200 symptoms are now associated with the disease, and still “there are no established effective treatments for long COVID,” says Linda Geng, who directs the Stanford Post-Acute COVID-19 Syndrome Clinic. “Anecdotally, patients have reported improvement with a variety of different treatments, but then there are just as many patients who do not respond to those same treatments.” 

I have to set reminders for myself like I have dementia.

—Brooke Keaton

After a slow start that frustrated many long COVID patients, also called long-haulers, scientists are beginning to piece together why a coronavirus infection causes lasting symptoms in some patients. “There are multiple groups coming up with different kinds of hypotheses; I think it’s worth looking at all possibilities and not being focused on one single thing,” says Avindra Nath, an immunologist at the National Institutes of Health. 

One of those possibilities is that COVID-19 causes the immune system to go haywire, triggering a long-lasting inflammatory response that wreaks havoc on multiple organ systems. This long-lasting inflammation may be explained by the continued presence of virus or viral particles in organs such as the gut, but may also be due to the reactivation of dormant pathogens such as Epstein-Barr virus, a herpesvirus that is the most common cause of mononucleosis. Another theory is that COVID-19 sets off a response akin to an autoimmune disease that causes antibodies to target and destroy the body’s own tissues. Still another hypothesis blames microclots, tiny blood clots that have been found to be more common in long COVID patients.

Taken together, studies suggest that long COVID and its many symptoms may not have just one cause, and may not be just be one disease. “There may not be a single mechanism for all long COVID—there could be multiple mechanisms ongoing, and for some individuals one may be more dominant than the other,” says Nath. 

Illustration of woman holding head above four graphic representing four hypotheses behind long COVID
Four hypotheses as to what might be causing long COVID symptoms
The Scientist STAFF

Inflammation and immune dysfunction

By now, several studies have shown that inflammation is a hallmark of long COVID, at least in some patients. For instance, inflammatory cytokines and activated immune cells have been found in the blood of up to 82 percent of long COVID patients from one to eight months post-infection. Wes Ely, a physician at the Vanderbilt University School of Medicine who treats ICU patients, says that evidence points to long COVID involving a “dysregulation of the body’s immune system and autonomic nervous system that leaves the person with a new normal.”

Taken together, studies suggest that long COVID and its many symptoms may not have just one cause, and may not be just be one disease.

Cytokines are a class of molecules that help boost inflammation as part of the innate immune response, and levels can increase after an initial COVID-19 diagnosis to incredibly high levels, which correlates with long-lasting tissue damage. In most patients, these levels fall again as the infection clears up, but in some, especially those who suffered severe disease, certain cytokines can remain elevated even after the infection has cleared, signaling that the immune system is still on high alert. Multiple studies have found increases in cytokines known as interleukins (ILs), such as IL-6, interferons (IFNs), and tumor necrosis factor (TNF) in long COVID patients compared with uninfected healthy controls or people who fully recovered from COVID-19. Several studies have also documented signs of immune cell dysfunction, including the presence of highly active T cells in long COVID patients, as well large numbers of cytotoxic T cells and naïve B cells, the latter of which behave as though they’ve never encountered a pathogen before. Yet another possibility is that SARS-CoV-2 proteins are superantigens, proteins that directly activate T cells and lead to massive inflammation and tissue damage. 

The reason behind this inflammation, and how it may be causing symptoms, has been more difficult to pin down, but animal studies—mainly on golden hamsters, a widely used model for respiratory diseases—provide some insight. A recent study compared the effects of SARS-CoV-2 infection in golden hamsters to those of the flu virus and found that only the former caused widespread inflammation and long-term damage to the kidneys, lungs, and brain. In particular, the olfactory epithelium (the lining inside of the nose) and the olfactory bulb (the part of the brain that processes smell) both suffered inflammation long after the virus was cleared from the hamsters’ systems. The authors of the paper write that the olfactory bulb neurons could be spreading inflammation to the surrounding tissues, and say that SARS-CoV-2 may have “rewired” neurons to continually express inflammatory markers. 

There’s been little evidence of SARS-CoV-2 attacking neurons directly, as it’s been found in very low levels in nonolfactory parts of the brain in postmortem patients, when it’s been found at all. Yet Michelle Monje-Deisseroth, a neuro-oncologist at Stanford University, says that even back in 2020, she worried that COVID-19 patients might suffer from brain fog and memory loss after infection. Although she stresses that neurons are likely not being directly infected or harmed by the virus, she points out that microglia, the supportive cells that feed neurons and help them communicate, could be entering a state where they do more harm than good. “What my group has been studying for a number of years in another disease context after systemic exposure to . . . chemotherapy and other inflammatory challenges is that microglia can become reactive,” she says. “As a result, cellular communication can be disrupted, and cognitive processing can be impaired.” In one study, Monje-Deisseroth confirmed that mice infected with SARS-CoV-2 showed increased microglia reactivity, especially in the hippocampus. She and her team also saw this effect in COVID-19 patients postmortem. Encouragingly, says Monje-Deisseroth, this is largely reversible in chemotherapy patients. 

See “Brain Fog Caused by Long COVID and Chemo Appear Similar

An autoimmune response

Another immune-related hypothesis for long COVID, which Nath says doesn’t necessarily preclude inflammation-related symptoms, is that the body starts targeting itself. Long COVID shares symptoms with some suspected autoimmune diseases—mainly ME/CFS, which is marked by post-exertional malaise and muscle pain—and multiple studies have found that long COVID patients develop self-antibodies that persist for a year, and perhaps longer. 

Antibodies identified in people who’ve recovered from an acute SARS-CoV-2 infection target self-antigens including phospholipids, transcriptional and nuclear proteins, interferons, and even CD8, a protein on some T cells. The generation of such autoantibodies, says Juan-Manuel Anaya, an immunologist at the University of Rosario in Bogotá, Colombia, might be spurred on by the initial inflammation, when B cells are churning out antibodies to fight the infection. The inflammatory milieu brought about during COVID-19 cases, especially during more severe cases, he says, causes a strong immune response that “favors the autoreactivity of B lymphocytes.” A recent study supports this hypothesis, finding that B cells are naïve, activated, and largely unregulated during the initial response to infection in severe COVID-19 cases. (Anaya was not involved in the study.) 

SARS-CoV-2 could also be unlocking a dormant autoimmune disease, says Anaya, which is marked by the presence of autoreactive immune cells and autoantibodies to specific proteins. Some viruses have been implicated in the development autoimmune diseases. For example, researchers have linked Epstein-Barr virus to a whole host of autoimmune diseases including lupus, Sjorgen’s syndrome, and systemic sclerosis. And though they aren’t entirely sure how, scientists suspect that measles, hepatitis C, and a dozen other viruses wake up latent autoimmune diseases.  

Whether SARS-CoV-2 triggers a full-blown autoimmune disease or not, the self-antibodies B cells produce could come about as a result of molecular mimicry, in which the immune system attacks proteins in the body that resemble the virus, although Anaya says that there’s little evidence SARS-CoV-2’s proteins resemble any human proteins. Another possibility is that the virus damages the host’s tissue, and these destroyed cells become the targets of immune destruction.

A major difficulty with pinning down long COVID’s pathology is that studies of the disease often have conflicting results. For example, at least two studies that looked at the presence of proinflammatory markers found no significant increase in levels of autoantibodies in long COVID patients. This could be due to selection bias, Anaya explains, as many of his studies were performed on Latin American patients hospitalized for COVID-19, while many other studies were done on European populations. More than one researcher who spoke to The Scientist has commented that the size of most studies, which are typically around 100-200 patients, limiting their statistical power and raising the prospect of sampling bias.

Continuing infection

Another hypothesis behind the persistent immune effects is that SARS-CoV-2 never left: It, or remains of the virus, are still present somewhere in the body and causing inflammation and immune dysregulation. In a study published on September 2 in Clinical Infectious DiseasesHarvard biomedical engineer David Walt and his team found low levels of the virus’s spike protein in the blood of 70 percent of study participants, all of whom had had long COVID for at least 12 months. “There’s clearly some evidence for a persistent viral infection,” says Walt. He says that the virus could be circulating undetected in some tissues, such as the lung or gut, even if it’s cleared from circulation. Some viral infections take more than six months to get rid of, and it’s possible that “these viral reservoirs are highly recalcitrant to treatment,” Walt speculates. But there hasn’t yet been any evidence of active, replicating virus in long COVID patients.  

Walt says that his team’s finding is encouraging because it means that an antiviral might be able help some patients recover. “We’re hoping to get approval soon to start a clinical study where we can initiate a regimen of antiviral over a period of time.” 

A recent study led by Ami Bhatt, a physician and geneticist at Stanford University, and her team uncovered evidence that the virus might stick around in the gut: SARS-CoV-2 RNA in the feces of long COVID patients seven months after diagnosis. The presence of viral RNA in the gut also correlated with several long COVID symptoms. She’d long watched COVID-19 patients come into the clinic with severe gastrointestinal issues and knew that the virus infects gut cells; the new finding suggests it’s not easily evicted, she says. Another study on biopsies from patients previously diagnosed with inflammatory bowel disease also found SARS-CoV-2 RNA in gut cells three to eight months after infection, and its presence was once again correlated with multiple long COVID symptoms. 

If SARS-CoV-2, or even pieces of it, are sticking around for months after infection, it could alter immune function, researchers note. In a recent study from Yale immunologist Akiko Iwasaki and colleagues found that the blood plasma of 99 patients with long COVID contained so-called exhausted T cells 3 and 12 months after infection. But while that exhaustion could come from chronic exposure to SARS-CoV-2, it could also come from exposure to other viruses. In some study patients, the researchers found that the Epstein-Barr virus is reawakened and that this reactivation correlates with long COVID symptoms. Iwasaki and her colleagues also report that long COVID patients had lower cortisol levels than patients who had been infected but fully recovered. Cortisol is typically anti-inflammatory and is thought to contain the immune response, and may be a marker of immune dysregulation. Acute COVID-19, by contrast, is marked by higher-than-normal cortisol levels, as is commonly found early on in other diseases as well. While there is so far no evidence that cortisol can treat long COVID patients, low cortisol could be a potential marker for disease. 

Endothelial damage

Some researchers argue that COVID-19 is not just a respiratory disease, but a vascular one. Endothelial cells express ACE2, the receptor that SARS-CoV-2 uses to bind to and infect cells, and COVID-19 is associated with a number of acute clotting syndromes. Thus, some scientists suggest that vascular issues are at the root of some long COVID symptoms or cases. 

COVID-19 is known to cause damage to endothelial cells, and that damage is thought to persist long after the acute disease has passed. Long COVID is also associated with endothelial dysfunction, a type of coronary artery disease. One study showed that blood flow to the heart was lower than normal in some patients months after a coronavirus infection. And at least one study has found evidence of antibodies attacking endothelial cells in the brain during acute infection. 

“Usually the endothelial damage [caused during acute infection] would be gone once the virus creating the problem is gone,” says Vanderbilt’s Ely. But he says that even after their initial SARS-CoV-2 infection, patients of his have suffered from a number of blood-related ailments, including thrombophilia, suggesting that their endothelial cells may be suffering sustained damage and inflammation. “Sometimes even after the virus is gone, the damage just keeps going. It’s like a domino effect.” 

Sometimes even after the virus is gone, the damage just keeps going. It’s like a domino effect.

—Wes Ely, Vanderbilt University School of Medicine 

A team led by Etheresia Pretorius, a physiologist at Stellenbosch University in South Africa, and Douglas Kell, a biochemist at the University of Liverpool, also reports that long COVID’s symptoms may be blood-related. Based on studies they’ve done with cohorts of COVID-19 and long COVID patients, the researchers have claimed that evidence is overwhelming that microclots—tiny, fibrous clumps of blood cells—are responsible for many long COVID symptoms.  

Kell has long argued that COVID-19 is a coagulopathy, affecting how the blood clots in various ways. In one study, Kell, Petrorius, and colleagues linked long COVID to amyloid blood clots, composed of an abnormal amyloid version of fibrinogen. Another finding from that study was that SARS-CoV-2 can induce microclots in normal blood in a dish. According to a new study by the team, several long COVID patients have reported success with anticlotting medication. Nath, who did not participate in the research, says that the endothelial hypothesis is “more controversial” than others, due to a lack of supporting evidence from other groups. 

The path to a treatment

Most scientists who spoke with The Scientist agree that there could be multiple causes of the disease, and the proposed hypotheses are not mutually exclusive. “It’s likely not just one thing, one disease,” says Nath. Ely agrees. “I don’t think there’s gonna be a unifying hypothesis here.” 

This, of course, poses a significant hurdle to researchers and doctors treating long COVID. With no proven treatments for these individuals, “we urgently need a randomized clinical trial,” Anaya says.  

Several large, randomized clinical trials on potential treatments are in progress around the globe, some of which aim to identify subgroups of long COVID patients that may benefit from a particular treatment or present with a subset of symptoms. One of those, the NIH’s $1.5 billion RECOVER study, is set to test more than a dozen different immunosuppressants, immune enhancing drugs, corticosteroids, and antiviral drugs in an effort to correct the immune dysfunction that happens in long COVID. Nath says that these treatments affect multiple arms of the immune system, but that diagnostics and treatment for long COVID eventually might benefit from a more specific approach. As such, biomarkers that pinpoint the problem could help treat the disease, he adds. “If you knew exactly what part of the immune system is affected, then you could precisely target it.” A similar trial in the UK, STIMULATE-ICP, seeks to identify subgroups of long COVID patients. It launched in June and aims to enroll 4,500 people. 

For Keaton and many others with long COVID, answers and interventions can’t come soon enough. Right now, she says she’s focused on her family, and making it through day by day. Keaton’s youngest child recently started preschool. “I taught children her age,” she says, but “I’m to the point now where in the middle of trying to teach her certain things . . . I lose the words. I forget the songs . . . I couldn’t even teach her how to button her clothes.”

She says that it’s been a struggle to get disability payments, and her first application was denied. As she’s without work, Keaton says that right now she can’t afford to try treatments that have reportedly worked for some patients.

Keaton says that the recent challenges have “been a true test for,” her and her family. “But that’s not uncommon in my life with long COVID. And I’m still waiting on answers.”