SARS-CoV-2 viral load is associated with increased disease severity and mortality fajnzylber2020sars
The paper quantified SARS-CoV-2 viral load from participants with a diverse range of COVID-19 disease severity, including those requiring hospitalization, outpatients with mild disease, and individuals with resolved infections. Blood was collected from hospitalized participants, non-hospitalized symptomatic individuals seeking care at a respiratory infection clinic, and participants who had recovered from known COVID-19 disease. Nasopharyngeal swabs, oropharyngeal swabs, sputum, and urine were collected from hospitalized participants. Data were obtained from the supplementary materials.
Analytes
Nasopharyngeal_SARSCoV2_N
Nasopharyngeal swabs were collected in 3 mL of phosphate-buffered saline. Viral concentrations were quantified using RT-qPCR targeting the N gene in the nasopharyngeal swab samples.
Fitted by censored maximum likelihood. The red line is the median individual; the shaded region is the full range of a simulated cohort drawn from the fitted population, so it shows what simulating from this dataset would produce rather than a confidence interval, with dashed lines at the central 95%. Open triangles are non-detects, drawn at the censoring limit and entering the fit as "below this value" rather than being dropped. See the modelling methods for what these estimates do and do not support.
Oropharyngeal_PBS_SARSCoV2_N
Oropharyngeal swabs were collected in 3 mL of phosphate buffered saline. Viral concentrations were quantified using RT-qPCR targeting the N gene in oropharyngeal swab samples.
Fitted by censored maximum likelihood. The red line is the median individual; the shaded region is the full range of a simulated cohort drawn from the fitted population, so it shows what simulating from this dataset would produce rather than a confidence interval, with dashed lines at the central 95%. Open triangles are non-detects, drawn at the censoring limit and entering the fit as "below this value" rather than being dropped. See the modelling methods for what these estimates do and do not support.
Oropharyngeal_VTM_SARSCoV2_N
Oropharyngeal swabs were collected in viral transport medium. Viral concentrations were quantified using RT-qPCR targeting the N gene in oropharyngeal swab samples.
Measurements only — no shedding model is fitted to this analyte, usually because it is sampled once per participant, leaving no trajectory to fit, or because nothing was ever detected. Open triangles are non-detects, drawn at the assay's censoring limit. See the modelling methods for what these estimates do and do not support.
Sputum_SARSCoV2_N
Sputum samples were collected and viral concentrations were quantified using RT-qPCR targeting the N gene.
Fitted by censored maximum likelihood. The red line is the median individual; the shaded region is the full range of a simulated cohort drawn from the fitted population, so it shows what simulating from this dataset would produce rather than a confidence interval, with dashed lines at the central 95%. Open triangles are non-detects, drawn at the censoring limit and entering the fit as "below this value" rather than being dropped. See the modelling methods for what these estimates do and do not support.
Plasma_SARSCoV2_N
Plasma samples were collected and viral concentrations were quantified using RT-qPCR targeting the N gene.
Measurements only — no shedding model is fitted to this analyte, usually because it is sampled once per participant, leaving no trajectory to fit, or because nothing was ever detected. Open triangles are non-detects, drawn at the assay's censoring limit. See the modelling methods for what these estimates do and do not support.
Urine_SARSCoV2_N
Urine samples were collected and viral concentrations were quantified using RT-qPCR targeting the N gene.
Measurements only — no shedding model is fitted to this analyte, usually because it is sampled once per participant, leaving no trajectory to fit, or because nothing was ever detected. Open triangles are non-detects, drawn at the assay's censoring limit. See the modelling methods for what these estimates do and do not support.