427 - Variation in SARS-CoV-2 molecular diagnostic test performance in symptomatic versus asymptomatic populations

Background: Growing recognition of the importance of asymptomatic and pre-symptomatic transmission for severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2) has led to a substantial expansion of testing from symptomatic to asymptomatic patients, and particularly those with risk factors for infection. Viral burden in asymptomatic individuals can differ from symptomatic patients, which can impact test performance. We therefore evaluated the impact of expanded testing indications upon the sensitivity and specificity of molecular diagnostic assays for SARS-CoV-2.

Methods: We performed a retrospective review of laboratory results from 5,122 emergency room patients and inpatients tested for SARS-CoV-2 between 05/03/2020 and 06/13/2020 using the Hologic Panther Fusion and the Cepheid Xpert assays at the Brigham & Women’s Hospital in Boston, MA. Descriptive analyses were performed for trends in testing volume, rates of positivity and cycle thresholds (Cts) over time based on symptom status. We calculated the proportion of new diagnoses made on a patient’s first test as an indirect measure of sensitivity. We calculated the proportion of first tests that are positive with a Ct value < 35 as an indirect measure of specificity.

Results: The overall rate of positivity over the study period was 8.7% (599/7,510 tests; 440/4,795 people) and declined by 1.8% (95% CI -2.2% - -1.4%, P< 0.0001) each week. Relative to tests in symptomatic people, the asymptomatic population had a higher mean Ct value (35.1 vs 32.3; P < 0.0001). Ct values increased by 0.7 (95% CI -0.1 - +1.4, P=0.07) and 0.8 (95% CI +0.3 - +1.4; p=0.01), sensitivity declined by 4% (95% CI -9% - +1%, P=0.08) and 12% (95% CI -20% - -5%, P-0.01) and specificity declined by 8% (95% CI -3% to 20%; P=0.13) and 9% (95% CI 7% - 11%; P=0.0002), over the time period of the study for asymptomatic and symptomatic patients, respectively.Figure 1: Trends in Ct values by symptoms
Figure 2: Trends in diagnosis by first versus second test by symptoms
Figure 3: Trends in proportion of people with their first test having Ct < 35 by symptoms


Conclusion: We show that the proportion of patients with low SARS-CoV-2 viral loads has increased as testing has expanded to the asymptomatic population and as transmission wanes in the community. This negatively impacts the performance of molecular assays by increasing the risk of false negatives and the detection of non-viable virus. Decision algorithms based on molecular assay results may need re-evaluation in light of these dynamics.