More than three years after the COVID-19 pandemic began, more than three-quarters of adults in the United States, approximately 260 million people, have contracted the infectious disease. While most recover within a few weeks, approximately one out of every 10 people develops long COVID. Dozens of symptoms are associated with this mysterious condition, some of which include fatigue, breathlessness, joint or muscle pains, chest pain, irregular heart beat, kidney pain, tinnitus, nausea, shaking in the hands, depression, and “brain fog.”
What causes long COVID is still vigorously debated. Researchers have proposed a few possibilities, including reservoirs of SARS-CoV-2 particles that persist in patients’ bodies, abnormal immune responses to an acute infection, and blood-vessel damage.
The viral genome lies dormant in a person’s blood cells.
But recent research published in Nature identified certain biomarkers related to immune and endocrine system function detectable in the blood that may help us understand what causes long COVID, as well as how to diagnose and potentially even treat it. “The ability to find distinguishing biomarkers is a big deal,” says Alexandra (Sasha) Tabachnikova, a Ph.D. student in immunobiology at Yale School of Medicine who is a co-author on the study.
The researchers analyzed data from 268 participants, comparing patients who had suffered four months or more of long COVID symptoms, with a group who had never had COVID at all, but had been vaccinated, and another group who had fully recovered from COVID. They took blood samples from all participants, and analyzed these samples to measure which kinds of immune cells were most active. They also exposed study participants’ blood samples to a variety of viral proteins, such as SARS-CoV-2 and Epstein-Barr, as well as proteins normally present in human cells, to measure immune response. Once they had identified differences in the patients’ blood, they used machine learning to better understand which ones would best allow them to differentiate between patients with and without long COVID.
Tabachnikova and her colleagues found striking patterns in the blood from patients with long COVID. They detected abnormal activity in immune cells called T-cells, which help fight infection, reactivation of viruses latent in the blood, such as Epstein-Barr, and precipitous declines in cortisol, a hormone that regulates stress.
Epstein-Barr virus causes mononucleosis and is found in the blood of 95 percent of humans worldwide. After a person is first infected with Epstein-Barr, the virus may stop producing viral particles, but the genome lies dormant in a person’s blood cells. During instances of stress and reduced immune cell function, as when a person contracts COVID, the virus may reactivate and begin proliferating.
The finding aligns with research showing that people with detectable levels of Epstein-Barr virus at the time of COVID diagnosis are more likely to develop long COVID symptoms like fatigue, mucousy coughs, and memory problems, compared to those without detectable levels of the virus.
The researchers also found that long COVID patients had higher levels of IL-8 and galectin-1, proteins involved in regulating inflammation, as well as much lower levels of cortisol. Decreases in cortisol levels are also associated with chronic fatigue syndrome, which could explain why some of the symptoms long COVID sufferers experience are similar to those found in chronic fatigue syndrome, says Tabachinkova.
Based on the biomarkers they identified, the researchers designed a machine learning algorithm that was able to differentiate between people with and without long COVID in their study population with 96 percent accuracy. While the study population was small, the research paints a clearer picture of an enigmatic condition. Appreciating the biological mechanisms involved is another step closer to developing effective treatments. 
