In Reply—Repeated Testing in SARS-CoV-2 Infection

To The Editor:

We appreciate the points raised by Lippi et al regarding our article describing repeated testing for severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection.¹ In summary, the authors emphasize that repeated testing may be helpful in improving the negative predictive value of testing and ensuring that cases of COVID-19 are identified. The authors include evidence supporting the conclusion that identification of the SARS-CoV-2 virus is directly related to the number of nasopharyngeal swabs that are collected and also emphasize the importance of case-finding in control of the pandemic. In general, we agree that repeated testing may be helpful in certain situations of ongoing high suspicion for active infection where alternative approaches are not feasible; however, we believe that testing should not be applied indiscriminately in a resource-constrained situation.

The results of several studies have suggested that the number of unique patient specimens tested for SARS-CoV-2 is directly related to the positive identification of the virus and that there may be a high false-negative rate of molecular testing.^{2 , 3} The study by Zhang et al² reported 41 hospitalized patients with an initial negative polymerase chain reaction test who had at least 1 positive result on subsequent testing. However, the timing between tests was not reported in this article, which raises the possibility that some of the patients could have become infected after their first test. No laboratory test has 100% sensitivity, and we agree that the likelihood of detecting infected individuals will increase if they are tested more frequently. This characteristic of laboratory testing is not unique to SARS-CoV-2 but could be applied to molecular testing for many other infectious diseases. However, widespread indiscriminate repeated testing is not currently possible.

Unfortunately, supply chain challenges continue to limit the widespread availability of SARS-CoV-2 polymerase chain reaction testing in the United States. Tests should be used in an efficient manner and guided by principles of diagnostic stewardship.⁴ We agree that there may be a role for repeated testing in patients with high clinical suspicion of coronavirus 19, and where a positive result would change clinical management. However, in situations with low pretest probability and limited resources, repeated testing of a sample collected from the same anatomical site demonstrated a low yield (2%) in our study. The clinical stage of illness can be used to determine whether an upper or lower respiratory specimen may provide more useful information, with detection of SARS-CoV-2 in lower respiratory sources becoming more likely as the disease progresses. The data on test characteristics in asymptomatic patients remain limited. Further studies are needed to identify the utility of repeated testing in this population.

Even in the absence of widespread test availability, there are a variety of strategies that have been proposed to mitigate the risk of false-negative testing, including strict infection control measures, development of improved diagnostic tests, risk stratification of patients before testing, and the use of standardized protocols for the management of patients who test negative.⁵ These strategies, combined with strategic use of repeated testing, will preserve limited resources and provide the best route toward ending the SARS-CoV-2 pandemic.