Director, OHDSI Center Northeastern University & University of Oxford University of Oxford, NDORMS Portland, United States
Background: The persistence of COVID-19 symptoms (“long COVID”) is an established phenomenon following acute infection. The World Health Organisation (WHO) definition identifies cases as an ensemble of up to twenty-five key persisting symptoms after any SARS-CoV-2 infection. As the prevalence of SARS-CoV-2 reinfection rises globally, there are growing concerns about a potential association between long COVID and SARS-CoV-2 reinfection.
Objectives: To study the occurrence of persistent long COVID symptoms after reinfection vs primo-infection with SARS-CoV-2
Methods: We conducted a multinational cohort study using primary care records from England, United Kingdom (CPRD AURUM) and Catalonia, Spain (SIDIAP), both mapped to the OMOP Common Data Model. SIDIAP records were linked to hospital admissions data for completeness. SARS-CoV-2 infections were identified using antigen or PCR tests. The study period began on 1 January 2020 and ended at either last available date (CPRD: Jun 2021) or censored at the time when widespread COVID-19 testing was ceased nationally (SIDAP: Mar 2022). To be included, persons were required to be registered at a GP practice for at least 180 days prior to test with no history of the symptom in the 180 days prior to index. Twenty-five symptoms identified by the WHO definition of long COVID and persisting for >90 days after infection were studied. We defined long COVID as the presence of one or more WHO symptoms and performed sensitivity analyses to evaluate multi-symptom trends. We estimated rate ratios for long COVID symptoms following reinfection vs primo-infection. Primo-infections were matched 3:1 by age (+/- 5 years), test type (antigen or PCR) and date (week) to reinfections, identified by subsequent infection after a 6-week washout.
Results: We matched 39,492 first infections to 13,253 reinfections in Catalonia, and 2,126 first infections to 709 reinfections in the UK. Overall, follow-up was longer in SIDIAP than in CPRD (median: 358 days in SIDIAP, 151 days in CPRD). Baseline characteristics were broadly balanced after matching, with SMD < 0.1 for demographics and all co-variates except for dementia in SIDIAP (primo-infections: 3.4%, reinfections: 6.0%, SMD: 0.119). The risk of WHO long COVID symptoms was reduced for all individual after re-infection, compared to after primo-infection in both databases, with RR ranging from 0.08 (95%CI 0.04-0.17) for anxiety (in CPRD cohort) to 0.41 (0.37-0.45) for gastrointestinal issues (in SIDAP cohort).
Conclusions: Long COVID persistent symptoms were less frequent after reinfection than after matched primo-infection. It is unclear whether this is due to accumulated immunity, vaccine uptake, or change in dominant variants of SARS-CoV-2.
