Skip to main content

Longitudinal assessment of diagnostic test performance over the course of acute SARS-CoV-2 infection.

Our take —

This study compared three diagnostic tests (saliva RT-qPCR, nasal RT-qPCR, and a rapid antigen fluorescent immunoassay) for detecting viable SARS-CoV-2 virus among 43 adult cases in the USA. Participants self-collected nasal swab and saliva samples daily for up to 14 days after their initial positive test. RTqPCR tests (on either nasal or saliva samples) are better than the Quidel Sofia rapid nasal antigen test for detecting SARS-CoV-2 earlier in the infectious period. However, serial testing with either a RTqPCR or rapid antigen test every three days still achieves high sensitivity (80% – 85%) and can identify SARS-CoV-2 infection early in the infectious period, which is valuable for preventing forward transmission.

Study design

Prospective Cohort

Study population and setting

In a university setting of students and employees who were required to submit saliva samples for SARS-CoV-2 RTqPCR testing every 2-4 days, 43 participants who had a positive test within 24 hours enrolled in the study. Saliva RTqPCR, nasal swab RTqPCR, and Quidel Sofia 2 SARS Antigen Fluorescent Immunoassay (nasal rapid antigen testing) were performed daily to assess the daily sensitivity of the testing methods relative to nasal virus culture.

Summary of Main Findings

The sensitivity of RTqPCR (nasal swab and saliva) and Quidel rapid antigen tests peaked when culturable virus was detected in nasal swab samples; however, sensitivity of RTqPCR tests were higher in the pre-culturable time than the rapid antigen test. These results show that RTqPCR-based testing is better for detecting early SARS-CoV-2 infection during the infectious period to potentially reduce forward transmission; however, all modalities achieve high sensitivity if used at least every 3 days.

Study Strengths

This study is one of the first to report on longitudinal sensitivity of RTqPCR and rapid antigen testing methods across the course of SARS-CoV-2 infection, compared to the viral culture from nasal swabs, which is a proxy for presumed infectious virus.  The evaluation of sensitivities of testing methods by frequency of testing was based on empirical data, rather than previous studies which have relied on modeling, and therefore can have strong potential public health value.


The sample size was limited to adults in a university setting, usually healthier and younger, so results may not be generalizable to other groups (eg. older adults or children) or the general population.  Viral culture is also used as a proxy for infectiousness of participants but is not a perfect measure of infectability.  Some samples may also be false negatives in viral culture due to sample handing methods (eg. freeze/thaw).  Participants were recruited based on having a positive RTqPCR result which may limit our understanding of the true sensitivity, particularly during the pre-PCR positive period early in infection as well as among samples with lower viral loads that might have had false-negative results.  Additionally, the study participants all had asymptomatic or mild disease and the authors did not report differences in sensitivity of the test methods for symptomatic versus asymptomatic participants and could not assess test sensitives among those with moderate or severe disease.

Value added

This study shows that the sensitivities of various SARS-CoV-2 detection methods change over the course of infection. RTqPCR methods (nasal swab and saliva based) tend to tend to overperform rapid antigen testing, especially in the early disease period which is important for early detection of infected individuals.  However, with testing every 3 days or less, both RTqPCR and rapid antigen testing achieve >98% sensitivity for any infection detection, and >80% sensitivity for detection of infection during the period when virus was culturable in nasal samples, the most likely infectious period. This suggests regular screening using rapid antigen testing at least every 3 days can achieve high sensitivity for detecting COVID-19 cases when rapid turnaround of RT-PCR results is not possible.

This review was posted on: 21 September 2021