Virological characteristics and the rapid antigen test as deisolation criteria in immunocompromised patients with COVID-19: A prospective cohort study

Abstract

There are limited data supporting current Centers for Disease Control and Prevention guidelines for the isolation period in moderate to severely immunocompromised patients with coronavirus disease 2019 (COVID-19). Adult COVID-19 patients who underwent solid organ transplantation (SOT) or received active chemotherapy against hematologic malignancy were enrolled and weekly respiratory samples were collected. Samples with positive genomic real-time polymerase chain reaction results underwent virus culture and rapid antigen testing (RAT). A total of 65 patients (40 with hematologic malignancy and 25 SOT) were enrolled. The median duration of viable virus shedding was 4 weeks (interquartile range: 3–7). Multivariable analysis revealed that B-cell depletion (hazard ratio [HR]: 4.76) was associated with prolonged viral shedding, and COVID-19 vaccination (≥3 doses) was negatively associated with prolonged viral shedding (HR: 0.22). The sensitivity, specificity, positive predictive value, and negative predictive value of RAT for viable virus shedding were 79%, 76%, 74%, and 81%, respectively. The negative predictive value of RAT was only 48% (95% confidence interval [CI]: 33–65) in the samples from those with symptom onset ≤20 days, but it was as high as 92% (95% CI: 85–96) in the samples from those with symptom onset >20 days. About half of immunocompromised COVID-19 patients shed viable virus for ≥4 weeks from the diagnosis, and virus shedding was prolonged especially in unvaccinated patients with B-cell-depleting therapy treatment. RAT beyond 20 days in immunocompromised patients had a relatively high negative predictive value for viable virus shedding.