Study population and setting
The study population consisted of 96 patients (22 with mild disease, 74 with severe disease; median age 55 years, 60% male) with laboratory-confirmed SARS-CoV-2 infection from a single hospital in Zhejiang province, China, admitted between January 19 and February 15, 2020. Respiratory, serum, stool, and urine samples were collected daily from each admitted patient (total n=3497). A median of 18 respiratory samples, 7 stool samples, 7 serum samples, and 1 urine sample were collected per patient. The end of follow-up was March 20, 2020.
Summary of Main Findings
SARS-CoV-2 RNA was detected in 100% of respiratory samples at admission, 59% of stool samples, 41% of serum samples, and a single urine specimen. There was no difference in the proportion of samples with RNA detected between those with mild and severe disease. However, the median duration of detectable virus was longer in respiratory specimens from severely ill patients than in those with mild disease (21 versus 14 days). The median duration of detectable virus in stool (22 days) was significantly longer than in respiratory (18 days) or serum specimens (16 days). Patients with a severe disease had a higher viral load in respiratory specimens than those with mild disease, and the viral load remained high 3-4 weeks after disease onset. In contrast, in those with mild disease, the respiratory viral load declined after 2 weeks.
Samples were collected longitudinally from multiple body sites.
The number of patients was small (n=96). Detection of RNA may not reflect viable virus that can infect another individual. The role of fecal and bloodborne transmission is unclear. Because samples were only collected from those in hospital, there may be selection bias: patients who remain in hospital for a long time may not be reflective of those who were discharged, who died, or who were never admitted.
This study presents the most detailed examination to date of longitudinal patterns in viral shedding among COVID-19 patients.