Skip to main content

Saliva is more sensitive for SARS-CoV-2 detection in COVID-19 patients than nasopharyngeal swabs

Our take —

This small cohort study, available as a preprint and thus not yet peer reviewed, suggests that self-collected saliva is more sensitive to detect SARS-CoV-2 by RT-PCR than nasopharyngeal swabs in hospitalized COVID-19 patients; however, it is important to note that both sampling methods resulted in false-negatives. Large longitudinal studies with paired specimens and greater diversity in patient severity (including pre-symptomatic and asymptomatic infections) are needed to further evaluate the comparative effectiveness of saliva and nasopharyngeal swabs for the detection of SARS-CoV-2 infection.

Study design

Prospective Cohort

Study population and setting

Two independent cohort studies in Yale New Haven Hospital enrolled 44 hospitalized COVID-19 patients and 98 healthcare workers without symptoms but with occupational exposure to COVID-19 patients. COVID-19 patients provided self-collected saliva and/or worker-administered nasopharyngeal swabs (n=121 samples); 29 patients contributed 38 paired samples of both specimen types. Healthcare workers provided self-collected saliva and/or nasopharyngeal swabs approximately every 3 days (n=244 total samples collected) with 33 healthcare workers contributing matched samples. Both specimen types were tested by RT-PCR for the quantitative detection of SARS-CoV-2.

Summary of Main Findings

Hospitalized COVID-19 patients had a significantly higher viral load of SARS-CoV-2 in saliva than in nasopharyngeal swabs in the overall cohort and among participants with paired specimens. Among the 29 patients with paired specimens, there were 8 false-negative tests for nasopharyngeal swabs compared to saliva and 3 false-negative tests for saliva specimens compared to nasopharyngeal swabs. In a small subset of patients, SARS-CoV-2 viral loads in saliva and nasopharyngeal swabs generally decreased with time from symptom onset. In the high-risk healthcare worker cohort, there were two healthcare workers for which SARS-CoV-2 was detected in their saliva but not their nasopharyngeal swabs. In both cohorts, saliva was more sensitive than nasopharyngeal swabs for the detection of human RNase P, an internal control that indicates proper sample collection.

Study Strengths

Study strengths include the use of two longitudinal cohorts and collecting paired specimen types.


Although the study used a longitudinal design, it did not serially sample the same individual across all stages of infection (i.e., hospitalized patients were not sampled when they were pre-symptomatic). The study had a small sample size with very few pre-symptomatic and asymptomatic individuals that were confirmed to have been infected with SARS-CoV-2 infection. It is unclear why so many participants did not have matching specimens given the prospective design. Sample storage conditions also differed between the two cohorts examined. The analysis also did not account for within-person correlation due to repeat sampling.

Value added

This is one of the earlier reports demonstrating that SARS-CoV-2 detection is more sensitive in self-collected saliva than nasopharyngeal swabs among hospitalized COVID-19 patients.

This review was posted on: 17 June 2020