Sunday, 16 June, 2024
HomeCoronavirusCOVID Symptom Study app: Attributes and predictors of Long-COVID

COVID Symptom Study app: Attributes and predictors of Long-COVID

An analysis by researchers at Kings College London, using data from the COVID Symptom Study app, shows that one in 20 people with COVID-19 are likely to suffer symptoms for 8 weeks or more (so-called ‘long COVID’), potentially adding up to many hundreds of thousands in the UK and millions worldwide.

Led by Dr Claire Steves and Professor Tim Spector at King’s, this study focused on data from 4,182 COVID Symptom Study app users who had been consistently logging their health and tested positive for COVID-19 through swab PCR testing.

The team found that older people, women and those with a greater number of different symptoms in the first week of their illness were more likely to develop long COVID. The researchers have used this information to develop a model that can predict who is most at risk of long COVID based on their age, gender, and count of early symptoms. Statistical tests showed that this simple prediction was able to detect more than two thirds (69%) of people who went on to get long-COVID (sensitivity), and 73% effective at avoiding false alarms (specificity).

The team then tested this model against an independent dataset of 2,472 people who reported a positive coronavirus antibody test result with a range of symptoms and found that it gave similar predictions of risk.

The research could be used to help target early interventions and research aimed at preventing and treating this condition.

The research also provides insight into this poorly understood phenomenon and the experiences of people living with long COVID, and identifies two main symptom groupings. One was dominated by respiratory symptoms such as cough and shortness of breath, as well as fatigue and headaches, and the second form was clearly multi-system, affecting many parts of the body, including the brain, gut and heart.

Long COVID sufferers more commonly reported heart symptoms such as palpitations or fast heartbeat, as well as pins and needles or numbness, and problems concentrating (‘brain fog’). People with long COVID were also twice as likely to report that their symptoms had come back again after recovering (relapse) compared with those having short COVID (16% vs 8.4%). Insights learned so far are being used to make the COVID Symptom Study app better for studying long-COVID.

Overall, the team found that while most people with COVID-19 reported being back to normal in 11 days or less, around one in seven (13.3%, 558 users) had COVID-19 symptoms lasting for at least 4 weeks, with around one in 20 (4.5%, 189 users) staying ill for 8 weeks and one in 50 (2.3%, 95 users) suffering for longer than 12 weeks. These are conservative estimates, which, because of the strict definitions used, may underestimate the extent of long-COVID.

Extrapolating out to the general UK population, which has a different age and gender makeup compared with the COVID Symptom Study app users, the team estimated that around one in seven (14.5%) of people with symptomatic COVID-19 would be ill for at least 4 weeks, one in 20 (5.1%) for 8 weeks and one in 45 (2.2%) for 12 weeks or more.

Long COVID affects around 10% of 18-49-year-olds who become unwell with COVID-19, rising to 22% of over 70s. Weight also plays a role, with people developing long COVID having a slightly higher average BMI than those with short COVID. Women were 50 percent more likely to suffer from long COVID than men (14.5% compared with 9.5%), but only in the younger age group. The researchers also found that people with asthma were more likely to develop long COVID, although there were no clear links to any other underlying health conditions.

“These new findings highlight the importance of the large-scale population health data from the COVID Symptom Study app in understanding this disease and how it affects people. We are hugely grateful to our millions of loggers for their contributions to understanding COVID, as it’s with their help we can shed light on the scale and nature of the long-COVID”, says Dr Carole Sudre, from the School of Biomedical Engineering & Imaging Sciences

Dr Claire Steves, clinical academic and senior author from King’s said: “It’s important we use the knowledge we have gained from the first wave in the pandemic to reduce the long-term impact of the second. This should pave the way for trials of early interventions to reduce the long-term effects. Thanks to the diligent logging of our contributors so far, this research could already pave the way for preventative and treatment strategies for Long-COVID. We urge everyone to join the effort by downloading the app and taking just a minute every day to log your health.”

Professor Tim Spector, COVID Symptom Study lead and professor of genetic epidemiology from King’s said: “COVID-19 is a mild illness for many, but for one in 50 symptoms can persist for longer than 12 weeks. So, it’s important that, as well as worrying about excess deaths, we also need to consider those who will be affected by long COVID if we don’t get the pandemic under control soon. As we wait for a vaccine, it is vital that we all work together to stem the spread of coronavirus via lifestyle changes and more rigorous self-isolating with symptoms or positive tests.”

Reports of "Long-COVID", are rising but little is known about prevalence, risk factors, or whether it is possible to predict a protracted course early in the disease. We analysed data from 4182 incident cases of COVID-19 who logged their symptoms prospectively in the COVID Symptom Study app. 558 (13.3%) had symptoms lasting >28 days, 189 (4.5%) for >8 weeks and 95 (2.3%) for >12 weeks. Long-COVID was characterised by symptoms of fatigue, headache, dyspnoea and anosmia and was more likely with increasing age, BMI and female sex. Experiencing more than five symptoms during the first week of illness was associated with Long-COVID, OR=3.53 [2.76;4.50]. Our model to predict long-COVID at 7 days, which gained a ROC-AUC of 76%, was replicated in an independent sample of 2472 antibody positive individuals. This model could be used to identify individuals for clinical trials to reduce long-term symptoms and target education and rehabilitation services.

Carole H Sudre, Benjamin Murray, Thomas Varsavsky, Mark S Graham, Rose S Penfold, Ruth CE Bowyer, Joan Capdevila Pujol, Kerstin Klaser, Michela Antonelli, Liane S Canas, Erika S Canas Molteni, Marc Modat, M Jorge Cardoso, Anna May, Sajasurya Ganesh, Richard Jorge Cardoso, Anna May, Sajaysurya Ganesh, Richard Davies, Long H Nguyen, David Alden Drew, Christina M Astley, Amit D Joshi, Jordi Merino, Neli Tsereteli, Tove Fall, Maria F Gomez, Emma Duncan, Christina Menni, Frances MK Williams, Paul W Franks, Andrew T Chan, Jonathan Wolf, Sebastien Ourselin, Timothy Spector, Claire J Steves


[link url=""]Kings College London[/link]


[link url=""]MedRxiv abstract (not peer reviewed)[/link]

MedicalBrief — our free weekly e-newsletter

We'd appreciate as much information as possible, however only an email address is required.