Skip to main content

Antigen-based testing but not real-time PCR correlates with SARS-CoV-2 virus culture

Our take —

This study provides evidence that rapid antigen tests may have a better probability of detecting the most infectious or contagious individuals in the population as compared to RT-PCR tests. However, the rapid test missed almost one-quarter of the samples that were positive by RT-PCR, and the predictive value depends on prevalence and will decrease with decreasing prevalence. Given that antigen tests can provide a result within minutes, they could be a useful public health tool for improving timeliness of prevention measures particularly among those who might be most likely to transmit to others. More research is needed to understand how infectivity in the lab correlates to infectivity in the real world, as well as to better understand how well antigen tests may be able to detect infectious samples from pre- or asymptomatic individuals.

Study design


Study population and setting

38 samples with evidence of SARS-CoV-2 by RT-PCR (Lyra) were collected prospectively from individuals symptomatic for COVID-19 with onset of symptoms occurring in previous 0-7 days. Samples were tested by rapid antigen test (BD Veritor) and in laboratory-based cell culture to assess infectivity.

Summary of Main Findings

Rapid antigen test correctly identified 29/38 (76%) RT-PCR positive samples as antigen positive. 28/38 (73.7%) RT-PCR positive samples were able to infect cells in laboratory-based cell culture, and samples that were able to infect cells had on average more quantifiable RNA in them than samples that were not able to infect cells (p < 0.001). The rapid test was able to detect antigen in 27/28 (96%) of the samples that were able to infect cells. The positive predictive value, or the probability that a sample was infectious to cells given a positive test result, was 90% for the antigen test and 73.7% for the RT-PCR test.

Study Strengths

Samples were collected prospectively and were well characterized with known duration of time since symptom onset. Paired samples collected from the same individuals at the same time were used for antigen, RT-PCR, and infectivity studies allowing for direct comparison.


Small sample size limits the precision of the estimates provided. The limitation of study participants with symptoms for less than 8 days limits our understanding of the correlation between test results and infectiousness in pre- or asymptomatic individuals. The authors used infectivity in cells cultured in a lab, which is the best lab method to measure infectious virus, but is not necessarily the same as infectious or transmissible in the real world. The study only compares a single rapid test (BD Veritor) to a single RT-PCR test (Quidel Lyra), so further evaluation using other tests is necessary before these results can be extrapolated more generally to other antigen tests. Additionally, predictive value depends on disease prevalence, and as disease prevalence declines so will the predictive value of both tests.

Value added

This study integrated RT-PCR test results and quantification, with rapid antigen test results and cell culture infectivity in a single study with samples collected from symptomatic SARS-CoV-2 infected patients which allowed them to compare the results of the two types of tests to each other in the context of viral load and infectivity which hasn’t been done previously. Through this they were able to show that, at least in this study, among patients with symptom onset of one week or less, the rapid antigen test could be used to identify patients with infectious virus, which has the potential to be a useful public health tool.

This review was posted on: 2 November 2020