Discover what the latest science from the RECOVER Initiative means for our ability to understand, diagnose, prevent, and treat Long COVID.
This page contains descriptions of findings from RECOVER research studies. These descriptions use plain language and a format that is easy to understand.
If you want to learn more about the scientific discoveries described here, you can also browse and search the complete list of RECOVER Publications.
This RECOVER study looked at how well the Pfizer COVID-19 vaccine, called BNT162b2, protected children and teens from long-term health problems after having COVID-19, also known as Long COVID. Researchers studied the electronic health records (EHRs) from more than 385,000 children and adolescents across 13 hospitals and clinics during the Delta and Omicron waves. In the Delta wave, researchers studied the EHRs of 112,590 adolescents ages 12 to 20. Of these, 88,811 adolescents were vaccinated before getting COVID-19. In the Omicron wave, they studied the EHRs of 188,894 children ages 5 to 11 years old, and 84,735 adolescents ages 12 to 20. Of these, 101,277 children, and 37,724 adolescents were vaccinated before getting COVID-19. In the Delta period, the researchers found that vaccinated adolescents were 95% less likely to develop Long COVID. In the Omicron period, the vaccine lowered the risk of Long COVID by 60% in children and 75% in adolescents. The study also showed that the vaccine’s main benefit in preventing Long COVID in children and adolescents came from preventing them from getting COVID-19 in the first place.
RECOVER researchers studied how COVID affects people over time. They wanted to know if parts of the SARS-CoV-2 virus (which causes COVID-19), called antigens, stay in people's blood long after a COVID infection. They also wanted to see if these antigens are linked to Long COVID. Long COVID is when a person has 1 or more symptoms that last for at least 3 months after getting COVID. First, the researchers took blood samples from adults who had COVID. These samples were taken a few days after getting sick and up to 14 months later. Then, they looked for 3 types of antigens that are usually in the blood after having COVID: the S1 subunit of the spike protein, the nucleocapsid protein, and the full-length spike protein. The full-length spike protein was the most common antigen found in adults who had COVID, especially in those who had blood taken between 4 and 7 months after getting COVID.
The study also looked at 34 common Long COVID symptoms, like trouble breathing, muscle pain, and brain fog. Many participants said that they had at least 1 Long COVID symptom for 1 month or more after getting sick. In the group with the most common kinds of symptoms, 2 out of 5 people had antigens in their blood. In the group without symptoms, about 1 out of 5 people had antigens in their blood. This study shows that the 3 antigens researchers found might be linked to Long COVID. People with symptoms were twice as likely to have these antigens in their blood for a long time. The results also show that parts of the SARS-CoV-2 virus can stay in the blood for up to 14 months after an infection. This is true for people who have long-term symptoms. But, it is also true in some people who do not have long-term symptoms.
This RECOVER study focused on a problem that happens when scientists study rare events. When scientists have very few cases of rare events to study, the usual ways of studying data can give misleading results. One way to fix this is to combine data from many hospitals or clinics so there are more cases to study. But, sharing detailed patient information raises concerns about privacy, so patient privacy must be protected when combining data from multiple healthcare centers.
In this study, RECOVER researchers created a new tool called ODAP-B, which combines information more accurately without sharing patient details. This is because it only needs a summary of patient data, not full health records. By using ODAP-B, researchers can better understand what increases the chance of a rare event happening while protecting patient privacy.
Metformin is a drug that many doctors prescribe as treatment for type 2 diabetes mellitus (T2DM). Past studies have found that taking metformin before and while having COVID-19 helps lower the chances of getting very sick with COVID-19. These studies also found that taking metformin can lower the chance of a person being sick many months after getting COVID-19, which is called Long COVID. In this study, RECOVER researchers were interested in understanding if people with diabetes who were on different medications got Long COVID or died. They looked at electronic health records (EHRs), or computer-based versions of patient health records, for 88,342 adults with T2DM. The study looked at 2 groups. The first group used metformin to control their diabetes, while the second group used other medications to control their diabetes. The researchers found that the chances of getting Long COVID or dying were a little bit lower in the group that took metformin compared to the group taking other medications. Since being on metformin may help only a little bit with Long COVID for adults with T2DM, people taking diabetes medications do not need to change their treatment plans unless told by their doctor.
This RECOVER study looked at how a child or young adult’s body weight, measured by body mass index (BMI), might impact their risk of developing Long COVID. Researchers studied medical records from over 172,000 children and young adults across 26 hospitals between March 2020 and May 2023. They found that children and young adults with obesity (a high BMI) were 25% more likely to have long-term health problems after getting COVID-19 than those with a healthy weight. The findings also show that children with severe obesity (a very high BMI) were about 42% more likely to experience these long-term problems. They also checked if these links were different for children and young adults, but did not find significant differences. This study highlights that the BMI of children and young adults before getting COVID-19 is an important factor in their risk for developing Long COVID.
In this RECOVER study, researchers wanted to find out which health conditions were common in people with long-term effects of COVID-19, known as Long COVID. They looked at the medical records of over 14 million patients who had and did not have COVID-19 using a method called topic modeling. Topic modeling is a method to find patterns in groups of documents. Researchers used topic modeling to look at lists of health problems that millions of patients faced over time, including before and after they had COVID-19. This allowed them to find the conditions that were more common in people with Long COVID. Researchers found that health problems related to the lungs, heart, brain, and immune system were more common in people with Long COVID compared to those who didn’t have Long COVID or had not had COVID-19. Common problems included chronic fatigue (feeling tired all the time), headaches, anxiety, and heart issues. They also found that some problems were more common in people of a certain age or sex (whether they are male or female). The study shows that Long COVID can affect many parts of the body and highlights how health problems can differ between groups. It also shows that age and sex can affect the chance of getting certain health conditions after COVID-19.
Many research centers have general formal guidelines about whether and how employees can participate in research studies. However, these guidelines don’t always discuss the risks and benefits of researchers, staff, and community partners enrolling in studies they are involved in, which can create ethical questions. It is very important to consider these ethical questions for studies that aim to enroll large numbers of diverse participants.
All authors of this article are part of the Researching COVID to Enhance Recovery (RECOVER) Initiative, a large study enrolling participants at multiple locations across the country. They used what they learned from RECOVER to address what’s currently missing from existing policy and ethics papers by creating a set of guidelines for how research studies can enroll researchers, staff, and community partners. The authors conclude that people should not be excluded from participating in their own research studies if certain requirements are met and there are rules to ensure everyone’s safety.
RECOVER researchers know that there is a lot of important information about COVID-19 and Long COVID in electronic health records (EHRs), especially in the notes that doctors write during care. The researchers used a computer tool called Natural Language Processing (NLP) to find signs and symptoms of different conditions. However, NLP doesn’t always work well, especially with new health problems like COVID-19 or Long COVID, which are always changing. To improve this, researchers created a new NLP system that could collect detailed information about Long COVID signs and symptoms from EHRs. After testing the system at other medical sites, they found it worked well across different locations. The study showed that the new NLP system could successfully find the information they were looking for about COVID-19 and Long COVID in EHRs. The NLP is now available to other researchers and is being used to collect information that can support additional studies on COVID-19 and Long COVID.
This RECOVER study looked at 25 different blood tests to see if any of the tests could be used to tell if someone might have Long COVID. Long COVID is when someone feels sick for a long time after having COVID-19. Researchers took blood from people who did and did not have COVID-19. They took blood from both groups 6 months after they got infected or after they joined the study. Researchers found small differences between the blood of people in each group. People who had COVID-19 had slightly lower platelet counts (platelets help stop bleeding), a bit more sugar in their blood, and signs of possible kidney problems. But, these differences were very small and might not be very important in finding out if someone has Long COVID. The study found that none of the 25 blood tests could show if someone has Long COVID because the test results were too similar between people with Long COVID and those without it.
