Along with physical fatigue, “brain fog” has become one of the best-known manifestations of the condition known as Long Covid. Yet it’s still unclear why some people infected with SARS-CoV-2 develop cognitive problems—which can include trouble concentrating and remembering—and others don’t. Now, a large study of people hospitalized with COVID-19 early in the pandemic has identified two proteins involved in blood clotting, fibrinogen and D-dimer, that are associated with cognitive deficits up to 1 year post-infection.
The findings, published today in Nature Medicine, are an “important advance” for scientists’ understanding of how Long Covid develops, says Steven Deeks, a physician-scientist at the University of California, San Francisco (UCSF) who was not involved in the work. Combined with previous research, the study supports the theory that blood clots triggered during the acute stages of viral infection could lead to lingering symptoms such as brain fog. Deeks cautions, however, that the study’s focus on unvaccinated people who had severe COVID-19 may limit relevance for the wider population; many people with Long Covid had milder initial infections—they aren’t represented in the research.
The ongoing hunt for mechanisms underlying Long Covid has yielded several possible culprits, including excess blood clotting, rampant inflammation, and persistence of SARS-CoV-2 in the body. Several studies have also proposed potential biomarkers—usually molecular features of a patients’ blood or other tissues—that appear during or after infection and could help predict who is most at risk of persistent illness. But few clinical studies have focused on disentangling cognitive deficits associated with the condition, says Maxime Taquet, a clinical psychiatrist at the University of Oxford. “Biomarkers that are general for Long Covid [might] not necessarily apply to brain fog.”
To dig into this aspect of the illness, Taquet and colleagues used data from the post-hospitalization COVID-19 (PHOSP-COVID) study, which collected information from thousands of adults hospitalized with SARS-CoV-2 across the United Kingdom in 2020 and 2021. Participants were given blood tests upon hospitalization, and, at 6 months and 12 months postadmission, answered questionnaires, and took cognitive tests.
Analyzing data for more than 1800 patients, the team found that people who had higher-than-average levels of fibrinogen in their blood when they were hospitalized later scored worse on tests measuring skills such as memory and attention, and rated their general cognitive abilities as poorer in questionnaires. At the 6-month follow-up, for example, the high-fibrinogen group assessed their cognitive problems as being about 0.7 points worse on a 7-point scale compared with the low-fibrinogen group. Differences in objectively assessed cognitive skills were slighter—around 0.7 points worse on a 30-point scale—possibly because these tests are geared toward detecting major impairment such as dementia, not smaller deficits, Taquet says.
Raised blood levels of the second protein, D-dimer, were also predictive of cognitive problems at 6 months and 12 months, although only on the subjective questionnaires: 6-month scores were about 1.5 points worse for the high-D-dimer group. Those people were also more likely to report fatigue and difficulty breathing, and to say they had trouble carrying on with work.
The team’s analysis of a large U.S. electronic health database corroborated those findings, linking heightened fibrinogen or D-dimer during COVID-19 hospitalization to diagnoses of dementia, mild cognitive impairment, and other cognitive issues 6 months later.
This isn’t the first time fibrinogen and D-dimer have been linked to COVID-19: Previous studies have found elevated levels of the proteins, along with excess clotting, in hospitalized patients. Prepandemic research has also connected high fibrinogen to cognitive problems and dementia, Taquet says, which indicates it may be a general marker of cognitive deficits. (His team confirmed that connection in a subanalysis of electronic health records from before 2020.) D-dimer showed no such association with these deficits outside COVID-19 patients, perhaps suggesting a role specific to coronavirus infections, he adds.
The study doesn’t establish exactly how the proteins might be causing damage, although Taquet speculates that fibrinogen could be forming blood clots that disrupt circulation in the brain, or may even directly interact with receptors in the nervous system. D-dimer may be more likely to reflect clotting in the lungs, he says—which could help explain its link to breathing problems.
Overall, the study was “was rigorously designed and analyzed,” says Tanayott Thaweethai, a biostatistician at Massachusetts General Hospital. “They were looking to understand these cognitive deficits in a really detailed way.”
The researchers tried to rule out possible confounding factors in the PHOSP-COVID data set. For example, people with more severe cases of COVID-19 may simply be more prone to long-term cognitive problems. Yet fibrinogen and D-dimer could still predict cognitive deficits even when differences in severity, assessed through a patient’s need for supplemental oxygen and other factors, were taken into account. Taquet and colleagues also controlled for preexisting cognitive deficits, although they did so using surveys that asked people to recall their cognitive performance pre-COVID-19—a less reliable measure than assessments administered on the spot, Thaweethai says.
The question of how to diagnose Long Covid has been controversial, and the new study’s results are unlikely to lead to a definitive test any time soon. In addition to being unvaccinated and hospitalized, people in the study were exposed to early SARS-CoV-2 variants, so it’s unclear whether the blood protein associations will hold for patients infected more recently. Moreover, a biomarker that predicts cognitive problems in 6 months or 12 months—as fibrinogen and D-dimer did—isn’t as useful as a test that could clearly diagnose a current case of Long Covid, Deeks says.
On the treatment front, there’s little evidence that targeting the two proteins, or blood clotting in general, could prevent Long Covid. Anticoagulants have been proposed as a possible treatment but have yet to be tested in a rigorous clinical trial and could be dangerous in some patients.
Nevertheless, the findings are likely to spur more research into these clotting proteins and the biological pathways they’re involved in. Fibrinogen in particular “keeps popping up in all these studies,” says Deeks, who is helping to run a Long Covid study at UCSF. After reading the new paper, he’s determined to “double down on trying to figure out why.”
