Athena Akrami has not been well since March 2020, when she first experienced symptoms of a SARS-CoV-2 infection. Although the initial infection cleared within weeks, more than a year later Akrami still experiences fatigue, joint pain, and frequent fluctuations in her body temperature. Puzzled by seemingly normal blood tests, her doctors had not been able to pinpoint any obvious source of her condition, she says, though they suspect it is related to her bout of COVID-19. Then, during a recent clinical check-up for fertility treatments, one of Akrami’s doctors recommended that they look for any imbalances in her immune system, specifically, that they test the levels of inflammatory cytokines. This time, the results were not normal: her cytokine levels were elevated.
The role of cytokine storms, which occur when the overproduction of cytokines sets off a positive feedback loop of immune responses, in aggravating COVID-19 infections was clear from the early days of the pandemic. Little is known, though, about how these molecules might mediate the enduring effects of the condition, which have come to be known as long COVID. In November 2020, Akrami, a neuroscientist and group leader at University College London’s Sainsbury Wellcome Centre, coauthored a preprint proposing that misfired cytokines, which would lead to a hyperactive immune system, could explain most of the symptoms of long COVID. Now, “we really need some sort of very precise, hypothesis-driven experimental designs” to test this idea, she says.
Akrami isn’t the only one calling for more research on long COVID, and a cytokine imbalance is just one of many proposed mechanisms thought to underlie the ailment, which can be marked by symptoms ranging from forgetfulness to rashes to chest tightness. Today, a year and a half since the pandemic started, data on COVID-19’s long-term effects remain scarce, but reports of the condition are anything but. The UK’s Office of National Statistics estimates that about 13.7 percent of people who have suffered a SARS-CoV-2 infection report long-term effects that lasted for at least 12 weeks after their diagnoses. These numbers could very well be an underestimation as a recent meta-analysis of 29 studies from more than a dozen countries deduced that nearly half of COVID-19 patients experience symptoms that last for a month or longer, with one study reporting a striking 87.4 percent.
No one really knows how or why long COVID happens, who is more likely to get it, or how to treat it.
No one really knows how or why long COVID happens, who is most susceptible to it, or how to treat it. Doctors can barely agree on a definition for the condition, with the only consensus being that to qualify as a case of long COVID, symptoms have to persist for more than 12 weeks after the onset of a SARS-CoV-2 infection. After months of diligent efforts from patient advocacy groups, long COVID is now recognized by the World Health Organization (WHO) as a legitimate condition, and earlier this year, the US National Institutes of Health announced a dedicated $1.15 billion fund to advance research on the matter.
“The main challenge at this point is to nail [it] down at a phenomenological level,” says Ryan Low, a computational neuroscientist at the Sainsbury Wellcome Centre, one of Akrami’s coauthors on the cytokine preprint, and a long COVID sufferer himself. “Just what are the relevant systems it involves? . . . What are the different signaling systems we care about? What organs are affected?”
What’s known about long COVID
In most cases, COVID-19 symptoms last a couple of weeks. During this time, known as the acute phase of infection, the body can respond in several ways. Some infections are mild or even asymptomatic, while others can elicit a cytokine storm that, if left unchecked, can lead to respiratory failure and other complications requiring hospitalization. In many cases, symptoms resolve with proper care and the patients regain their health. A sizeable fraction of COVID-19 patients, however, never fully recover, and experience a wide range of lingering symptoms.
Most commonly, these patients report fatigue, breathlessness, and “brain fog,” a catch-all term used to describe an inability to concentrate or recall information, and multiple surveys have identified effects on several organ systems in the body. Based on a survey published in July, for example, Akrami and her colleagues at the Patient-Led Research Collaborative identified 205 possible symptoms of long COVID. Essentially, they found that long COVID can affect almost any organ system to varying degrees in different people.
That COVID-19 should not be considered a disease of the lungs “was clear very early on, from anecdotal evidence from the clinicians,” says Louise Wain, a genetic epidemiologist at the University of Leicester in the UK. “We needed to look at all the different organ systems [and] at mental and cognitive effects of the virus and really take a . . . whole-person approach to how we followed up.”
Medically known as post-acute sequelae of SARS-CoV-2 infection or persistent post-COVID syndrome, long COVID can occur regardless of the severity of acute infection. Much of the early documentation of symptoms and their prevalence came from patient research groups, and by now government health agencies such as the UK National Health Service and the US Centers for Disease Control and Prevention (CDC) have released care and treatment guidelines for long COVID. There is no single treatment for the condition, and current therapies, such as breathing exercises and psychological aid, only target some specific symptoms.
Given the lack of coherent data on long COVID, several scientists across the globe have launched efforts to better understand the condition. In the UK, for example, researchers involved with the Post-hospitalization COVID-19 (PHOSP-COVID) study work closely with clinicians to conduct long-term follow-ups of COVID-19 patients who were once hospitalized. And in Baltimore, the Johns Hopkins COVID Long study aims to capture the experiences of up to 25,000 COVID-19 patients in the US, including both people who did and people who did not require hospitalization.
It’s important to capture large numbers of patients because “one person’s recovery is very different than others’,” says Johns Hopkins Bloomberg School of Public Health epidemiologist Priya Duggal, a coinvestigator on the school’s long COVID project. It’s also important to include patients who might have been missed by advocacy groups due to lack of access or representation, she adds. So far, “white women have formed very large groups in social media circles, and they are well represented as having long term symptoms,” she says. “But we don’t know if race, for example, is a factor and we’re just not hearing it in certain voices because certain racial groups aren’t talking about it in social media circles.”
Indeed, with data only just beginning to roll in, it is difficult to predict who will get long COVID. The PHOSP-COVID group, in a recent preprint that analyzed around 1,000 patients, found that female sex and severe acute infections were risk factors for a longer recovery period. Although COVID-19 itself has different impacts on people of different ethnicities and races, it is not certain whether the same applies to long COVID. Wain is keen to see what the long-term follow-ups will yield, but warns that it might take a while.
“I think there’s certainly some interesting risk factor signals coming out,” Wain says. “[But] I think in terms of understanding whether they are features of the data itself, whether they’re related to factors in exposure, or whether they’re actually intrinsic biological mechanisms that are driving the difference . . . we’re not there yet.”
Long COVID Studies Around the Globe
Dozens of studies around the world are investigating the long-term effects of SARS-CoV-2 infection. Below is a sample of some trials that have been included in clinical trial registries.
© ISTOCK.COM, GMAST3R
Study/Consortium Name and Identifier
Target Cohort Size
|Weill Cornell Medicine
|To be determined
|Gene expression data
|Long COVID Registry
|To be determined
|Symptom prevalence, preexisting conditions, blood samples
|COVID Human Genetic Effort
|To be determined
|Genetic sequencing data, immune profiles
|Johns Hopkins University
|Symptom prevalence, preexisting conditions
|When target cohort size is met
|MOIST (Multi-Organ Imaging with Serial Testing in COVID-19 Infected Patients)
|University of Alberta
|MRI scan imagery, blood sample
|Study of Biomarkers in the Long-term Impact of Coronavirus Infection in the Cardiorespiratory System (NCT04353245)
|University of Sao Paulo General Hospital
|Ergospirometry (respiration and gas metabolism)
|1 year, 3 months
|Long-term COVID-19 Immune Response in a Vulnerable Neighbourhood in Argentina (NCT04673279)
|Hospital Italiano de Buenos Aires
|PHOSP-COVID (Post-hospitalization COVID-19 study)
|University of Leicester
|Symptom prevalence, lung function, blood samples
|> 1 year
|COVERSCAN (Mapping Organ Health in Recovery from COVID-19 Disease due to SARS-CoV-2 infection) (NCT04369807)
|MRI scan imagery
|CV-COVID-19 (Long-term Effects of Coronavirus Disease 2019 on the Cardiovascular System) (NCT04359927)
|Hospital Clinic of Barcelona
|IPEA (Immune-mediated pathogenesis and development of autoimmunity in neurological manifestations of COVID-19) (DRKS00024434)
|University Medical Cologne
|Immune profile, neurological function
|Post COVID-19 Functional Status in Egypt (NCT04479293)
|Symptoms, overall wellness
|COLDSTER (Comparison of Two Corticosteroid Regimens for Post COVID-19 Diffuse Lung Disease) (NCT04657484)
|Postgraduate Institute of Medical Education and Research
|CT scan imagery, oxygen saturation, quality of life, adverse effects
|DALT-COV (Depression and Anxiety in Long Term Coronavirus Disease) (NCT04893668)
|Neurotransmitter levels, depression, anxiety
|POTS-LCS (Postural orthostatic tachycardia syndrome [POTS] in Long COVID syndrome: A detailed profiling study) (ACTRN12621000476831)
|University of Adelaide
|Electrocardiogram, respiratory function, quality of life, depression, fatigue, blood/urine samples
|1 year, 2 months
What causes long COVID?
When it was first described, long COVID was thought to be a complication of post-ICU syndrome, a condition that arises from long periods of hospitalization and that can delay recovery. But later reports of long COVID in nonhospitalized patients and in those with mild infections made it clear that the condition was not just fallout from hospitalization. While nobody knows what causes long COVID, there are currently several hypotheses, including overreactive immune systems, reservoirs of viral particles in different organ systems, genetic factors, and blood clots.
Blood clots in particular have received a lot of attention as a possible driver of long COVID, in large part because of the growing body of evidence that acute SARS-CoV-2 infection could trigger cell signaling events that lead to clot formation. A recent review even argued that COVID-19 should be treated as a disease of the vascular endothelium and not as a respiratory disease.
The blood clots seen in some COVID-19 patients persist for several weeks after recovery from acute infection, possibly due to injured blood vessel cells, overactive platelets, or errant immune cells. The persistence of these clots in long COVID patients could eventually cause fibrosis of different organs and explain many of the symptoms experienced by this group of patients, says Resia Pretorius, a physiologist at Stellenbosch University in South Africa. Pretorius, who with colleagues recently posted a preprint on the presence of small, hardy blood clots in the vessels of both acute and long COVID patients in South Africa, says that she strongly believes that these clots are “one of the key reasons why some of these individuals with long COVID might still be suffering from all of these . . . various symptoms that have been noted.”
A dysregulated immune system is another prime suspect for many long COVID symptoms. The immune profiles of COVID-19 patients have shown cascading effects of the virus on the immune system, likely mediating variations in severity of the disease. Similar work on long COVID patients has started uncovering immune dynamics in these patients that are distinct from acute COVID-19 patients. Greg Szeto, an immunologist at the Allen Institute for Immunology in Seattle, and his team recently profiled the immune signatures of five long COVID patients and, in a recent preprint, described results that echo these findings. In fact, during the acute phase, the long COVID patients seemed to have lower immune responses than did infected patients who later recovered fully. They then developed hyperinflammation that persisted for months. The team is now working on an expanded cohort to validate these preliminary results.
Such persistent inflammation might have been a direct result of increased levels of cytokines such as IL-1b, which has been found in the sera of some long COVID patients. Cytokine release syndrome, a condition in which an uncontrolled release of the proinflammatory molecules can trigger a cytokine storm, has also occurred in a subset of COVID-19 patients. In Akrami’s case, she and her colleagues suspect that poorly regulated cytokines underpin most if not all of her symptoms. Variations in the immune response are also reflected in the populations of B and T cells in COVID-19 patients who were followed for six months after hospitalization, as reported by a recent study from the University of Manchester. The study found that patients who took longer to recover after being discharged had higher levels of cytotoxic T cells and certain cytokines.
Russell Low, a retired physician and coauthor on the cytokine preprint with Akrami (and the father of the other coauthor, Ryan Low), notes that blood clotting is not totally separate from cytokine and immune dysregulation. “Clinically, they’re going on simultaneously,” he says. “They’re not independent processes,” but researchers tend to treat them as separate events.
The difficulty in teasing apart the causes of long COVID is exacerbated by genetic variations among patients. Apart from the PHOSP-COVID study, Wain is also a part of the COVID-19 Host Genetics Initiative, which fosters studies on the genetic basis of COVID-19. Wain says that she hopes to eventually tie in the data from both projects to get to a comprehensive picture of the long-term effects of COVID-19. Similar efforts are underway at the COVID Human Genetic Effort, an arm of which focuses on finding genes that might play a role in the development of long COVID.
Wain says that the best way to understand how and why long COVID occurs would be to pool data from different studies and “bring together all the international datasets that are available to provide the statistical power.” Such efforts would not only help decipher the mechanisms of long COVID, but also help develop treatments, she says. Pretorius, who says she has been facing challenges in finding financial support for her South Africa-based study, echoes the call for global cooperation. “I think we must just have a big concerted effort to try to figure out what’s happening.”
Bodily Systems Affected by Long COVID
Infection with SARS-CoV-2 can lead to symptoms that last for weeks, months, or even years, and scientists are only just beginning to understand why. The long-term effects of COVID-19 can affect nearly every organ system. The type, severity, and duration of the symptoms vary across individuals, and researchers still know little about the mechanisms of long COVID.
Why is long COVID so difficult to study?
The heterogeneity of long COVID is, for a lot of researchers, unlike anything they have worked on before. The symptoms, severity, duration, and degree of initial infection all vary across individual patients. It’s also difficult to figure out who might be at risk, given the broad range of comorbidities and demographics involved.
Szeto points out that to truly understand the condition, it is important to acknowledge this heterogeneity. “Right now, we talk about it as one big thing. But my firm belief is it’s a variety of things.” Akrami agrees and emphasizes the value of stratifying cases. “I think the very first thing that needs to be answered is . . . can we really find subgroups of long COVID?”
Some researchers are attempting to do exactly this. For instance, the Patient-Led Research Collaborative survey clustered symptoms based on time of onset and persistence, while the PHOSP-COVID study clustered symptoms based on severity. In doing so, both studies were able to demarcate subsets of long COVID experiences, providing a starting point for identifying the signaling pathways or organ systems that mediate specific effects.
Working in reverse, Hossein Estiri, a clinical data scientist at Harvard Medical School, and his team have been using an algorithm on electronic health records to predict who is more likely to experience a severe COVID-19 infection. They recently repurposed the algorithm, now published as a preprint, to figure out whether, given a patient’s list of symptoms, it was likely that the individual had tested positive for COVID-19 in the past. In the future, they say, this could help physicians diagnose long COVID based on symptoms without a previous diagnosis of COVID-19. To truly help patients, however, Estiri notes that the onset of infection, variant of the virus, and treatment given during the acute phase should factor into the diagnosis and treatment.
Several researchers, Akrami and Russell Low included, are also looking to other post-viral syndromes to try and understand long COVID. The symptoms of myalgic encephalomyelitis/chronic fatigue syndrome (ME/CFS), a condition marked by debilitating exhaustion and cognitive problems, are very similar to those of long COVID, and ME/CFS can also arise after viral infections. “I think the underlying pathophysiology [of ME/CFS], while complex and incompletely understood, is similar” to that of long COVID, says Low. Several research efforts are now exploring possible links between ME/CFS and long COVID.
As with any new condition, the data on long COVID is constantly evolving, says Shruti Mehta, an epidemiologist at the Johns Hopkins Bloomberg School of Public Health in Baltimore, and a coinvestigator of the COVID Long study. Mehta, who has been designing surveys since April 2021 to collect data on long COVID, says this needs to be taken into account as researchers design future studies, noting that she has had to update her surveys regularly “just because everything keeps changing.”
To keep up with this dynamic landscape, the International Severe Acute Respiratory and emerging Infection Consortium (ISARIC), a global network of clinical research teams, has been working with the WHO to standardize data collection. The group, which offers a modular approach that allows investigators to add and remove questions based on the scenario, recently announced an effort to develop a list of primary symptoms of long COVID to help researchers narrow in on the most important parameters to be measured.
The need of the hour is to find the etiology of the condition and ensure equitable access to care, says Louise Sigfrid, a clinical research fellow at the ISARIC, “so that we can target treatments to improve outcomes.”
Clarification (September 4, 2021): This story has been updated from its original version to reflect the fact Greg Szeto of the Allen Institute for Immunology and colleagues recently posted an updated preprint that included data from five, not three, long COVID patients. The Scientist regrets any confusion.