With a substantial proportion of the European population having had COVID-19 at some point in the previous years, long-term complications are becoming increasingly prevalent. Studies suggest that about one out of ten people do not fully recover from their acute disease, and present ongoing symptoms for three months or longer.
In summer 2022, the EHDEN Evidence Generation Fund funded a large-scale multinational network study to characterise Long COVID and PASC in patients across Europe.
Via an open call, all EHDEN Data Partners were invited to participate in a study-a-thon hosted by the Health Data Sciences Group, University of Oxford. We were delighted to have eleven EHDEN Data Partners participate, and three clinical experts seeing Long COVID patients at clinics in Catalonia every day joining us for this project!
After intense preparations, the day arrived, and we were excited to welcome researchers to the beautiful Worchester College in April 2021 in Oxford. Analytic code was sent out to Data Partners prior to the face-to-face meeting, which allowed us to collaboratively discuss and interpret results during the week.
During the week, we worked on three different aims to better understand Long COVID:
Firstly, we aimed to appraise twenty-five WHO suggestive symptoms for the diagnosis of Long COVID and to identify which of these would be particularly discriminative for Long COVID compared to control populations. Secondly, we aimed to characterise drug use in people with Long COVID, and lastly, we derived clusters of symptoms among patients with Long COVID and compared them across databases.
We defined Long COVID as a combination of a record of COVID-19, confirmed by a positive test or a clinical COVID-19 diagnosis, and one or more symptoms listed in the WHO’s clinical case definition recorded between 90 and 365 days after infection. People who had the same symptom recorded in the six months before SARS-CoV-2 infection were not considered as Long COVID cases.
We found great heterogeneity in the presentation of Long COVID symptoms across populations, settings and data sources. Meta-analytic estimates suggested a list of most specific symptoms (e.g., altered smell/taste, dyspnea, fatigue/malaise, cough) vs less specific ones (e.g. gastrointestinal issues, abdominal pain), which calls for a further evaluation of the current WHO definition of Long COVID.
A variety of drugs were prescribed in Long COVID patients in the acute phase of COVID-19, highlighting the complexity of the disease and variety of symptoms associated with it.
Clustering of patients with one single Long COVID symptom were common. Other clusters included a mix of Long COVID symptoms, but were less consistent across databases. Therefore, future (ongoing) research is needed to identify clusters of patients with more than three symptoms.
After five intense days of multidisciplinary discussions, we summarised our findings in two abstracts, which will be presented as a poster and a presentation at the OHDSI Europe Symposium 1-3 July in Rotterdam.
Overall, we spent a very productive, collaborative and inspiring week together – thanks to everyone who made this experience so special! We are continuing to work closely with all the Data Partners to update our understanding of Long COVID and the comprehensive scientific findings will be disseminated in the near future.
Annika Jödicke
Senior Research Associate
in Pharmacoepidemiology
University of Oxford