Methods and search results
We searched repetitively Pubmed, medRxiv and bioRxiv with keyword “coronavirus” limiting to results published between 1st December 2019 and 15th May 2020 and screened the titles and brief description of >8700 publications in total. We identified studies on SARS-CoV-2 (incl. name variations like “novel coronavirus”, “2019-nCoV”) in humans (all ages), written in English, Chinese or French, and excluded reviews, viewpoints, or news. We selected studies using the words “detection”, “testing”, “PCR”, “viral load”, “viral kinetics/dynamics/clearance/shedding”, “isolation”, “persistence”, “samples”, “bodily fluids”, “diagnosis”, “case report/series”, “case(s)”, “cluster”, “transmission”, “patients”, “neonate”, “child(ren)”, “pregnant”, “clinical characteristics/findings/manifestations/features/outcomes”, “infection”, “pneumonia”, “asymptomatic” in the title, brief description or abstract (if available) aiming to narrow down to clinical reports. This yielded 702 publications for in-depth abstract and full-text screening. Additionally, we scanned literature cited in these articles as well as suggested similar publications and COVID-19 resource collections on the publishers’ websites. Finally, we included 264 studies reporting on SARS-CoV-2 molecular detection and/or virus isolation in any laboratory specimens of COVID-19 cases. We excluded reviews, meta-analyses, news, guidelines, or modelling studies based on public data.
We aimed to summarize the current information on SARS-CoV-2 kinetics in relation to clinical syndrome, in different bodily fluids, while also noting any specifics in vulnerable groups (pregnant women, children and immunocompromised individuals). We extracted data on cases’ demographics (number of adults, children, pregnant and immunocompromised individuals; age; sex; severity of disease) and specimens tested for SARS-CoV-2 (number of patients with collected respiratory, gastrointestinal and blood sampling, oral fluid, tears, urine, cerebrospinal fluid, semen, vaginal fluid, breast milk and perinatal samples; all incl. sample types tested by PCR and virus cultures where available; number of patients with sequenced samples) and compiled a dataset containing 32515 COVID-19 cases (Supplementary Dataset). Even though nasopharyngeal swabs are the routine respiratory sample for viral infections, we also included publications with nasal/midturbinate swabs or those using the term “nasal” indiscriminately when describing nasopharyngeal swab collection technique. Where possible, we noted duplicate case reports in different publications. Due to varying disease severity definitions, we opted for a simplified approach: patients with symptoms or clinical course described as mild, moderate, common, typical, are referred to as “mild”; while severe, critically ill and/or admitted to intensive-care units (ICU), as “severe”. Additionally, we aggregated data from 217 adults with data points from 2777 samples with known collection day post symptoms onset (dps) (Figure 1, Figure 2, Table, and Supplementary Table).