Our ability to understand and mitigate the spread of SARS-CoV-2 depends largely on antibody and viral RNA data that provide information about past exposure and shedding. Five months into the outbreak there is an impressive number of studies reporting antibody kinetics and RNA shedding dynamics, but extensive variation in detection assays, study group demographics, and laboratory protocols has presented a challenge for inferring the true biological patterns. Here, we combine existing data on SARS-CoV-2 IgG, IgM and RNA kinetics using a formal quantitative approach that enables integration of 3,214 data points from 516 individuals, published in 22 studies. This allows us to determine the mean values and distributions of IgG and IgM seroconversion times and titer kinetics, and to characterize how antibody and RNA detection probabilities change during the early phase of infection. We observe extensive variation in antibody response patterns and RNA detection patterns, explained by both individual heterogeneity and protocol differences such as targeted antigen and sample type. These results provide a robust reference for clinical management of individual patients, and a foundation for the mathematical models and serological surveys that underpin public health policies.

Our ability to understand and mitigate the spread of SARS-CoV-2 depends largely on antibody and viral RNA data that provide information about past exposure and shedding. Five months into the outbreak there is an impressive number of studies reporting antibody kinetics and RNA shedding dynamics, but extensive variation in detection assays, study group demographics, and laboratory protocols has presented a challenge for inferring the true biological patterns. Here, we combine existing data on SARS-CoV-2 IgG, IgM and RNA kinetics using a formal quantitative approach that enables integration of 3,214 data points from 516 individuals, published in 22 studies. This allows us to determine the mean values and distributions of IgG and IgM seroconversion times and titer kinetics, and to characterize how antibody and RNA detection probabilities change during the early phase of infection. We observe extensive variation in antibody response patterns and RNA detection patterns, explained by both individual heterogeneity and protocol differences such as targeted antigen and sample type. These results provide a robust reference for clinical management of individual patients, and a foundation for the mathematical models and serological surveys that underpin public health policies.

No abstract