Research highlight

Medical research: Modelling the timing of SARS-CoV-2 transmission

Nature Medicine

April 16, 2020

Patients with the respiratory disease COVID-19 may begin to shed, or excrete, infectious SARS-CoV-2 coronavirus 2–3 days before the first symptoms appear, suggests a modelling study reported in Nature Medicine.

Several factors can affect the efficacy of control measures designed to prevent the spread of SARS-CoV-2. These include the time between successive cases in a chain of transmission (serial interval) and the period between exposure to an infection and the appearance of symptoms (incubation period). If the serial interval is shorter than the incubation period, this would indicate that transmission may have happened before overt symptoms develop. As a result, control measures that take place at the point at which symptoms appear may have a reduced effect in controlling the spread of infection.

Eric Lau and colleagues investigated the temporal patterns of viral shedding in 94 patients with COVID-19 admitted to Guangzhou Eighth People’s Hospital in China. Throat swabs were collected from these patients from when symptoms first appeared until 32 days later. In total, 414 swabs were analysed, and the authors found that the patients had the highest viral load at the onset of symptoms.

Separately, they modelled COVID-19 infectiousness profiles from a different sample of 77 ‘transmission pairs’ from publicly available data. Each transmission pair comprised two patients with COVID-19 with a clear epidemiological link, with one patient highly likely to have infected the other. From here, the authors inferred that infectiousness started 2.3 days before symptoms appeared and peaked at 0.7 days before their appearance. They estimated that 44% of secondary cases were infected during the pre-symptomatic stage, with infectiousness predicted to decrease quickly within 7 days.

The authors note that the limitations of the study include the reliance of the research on patient recall of the onset of symptoms, which may have introduced bias, as there may be a delay in the recognition of the first symptoms.

doi: 10.1038/s41591-020-0869-5

Return to research highlights

PrivacyMark System