Abstract
BACKGROUND: Long COVID is a common complication of infection with severe acute respiratory syndrome coronavirus 2, but the prevalence and predictors of the condition remain poorly characterized. METHODS: We prospectively studied adults (≥18 years) with acute coronavirus disease 2019 (COVID-19) presenting to an urban safety net hospital and associated clinics between July 2020 and December 2022. Logistic regression models were used to evaluate the association between baseline demographic, clinical, and laboratory characteristics with long COVID status, defined as symptoms persisting at least 9 months after acute disease. Among unrecovered participants, we describe the prevalence of individual symptoms. RESULTS: We enrolled 222 participants, 162 (73%) of whom had known recovery status by 9 months. Median age was 54 years, half (55%) were female, and the majority of participants (78%) had at least 1 comorbidity at the time of COVID-19 diagnosis. Based on acute illness characteristics, the adjusted odds ratio for long COVID was 3.0 (95% confidence interval [CI], 1.1-8.0) among those with detectable nucleocapsid antigen and 3.6 (95% CI, 1.2-11) for those who required supplemental oxygen. Of the 41% of participants with symptoms persisting at least 9 months, central nervous system and psychological symptoms were most commonly reported, with 57% reporting functional limitations due to their persistent symptoms. CONCLUSIONS: The strong association with initial disease suggests a decreasing prevalence of long COVID as acute illnesses become milder. However, many contemporary patients still experience high viral burden with extended viral replication, even after vaccination. Our findings highlight the importance of properly characterizing long COVID as viral evolution shifts acute disease presentation.