ABSTRACT
Background Nasopharyngeal samples (NPS) are the mainstay of COVID-19 diagnosis. However, the extent to which assay signals relate to exhaled virus is unknown. We investigated the use of novel, non-invasive face-mask sampling (FMS) to detect exhaled SARS-CoV-2 RNA in two studies. Methods In an outbreak study (cohort 1), we performed FMS and NPS for 21 consecutive days after diagnosis on six healthcare workers who were screened positive for SARS-CoV-2. In a second hospitalised cohort (cohort 2), we performed FMS on 47 patients within 24 hours of a positive diagnosis. COVID-19 severity was graded according to WHO recommendations. Findings In cohort 1, SARS-COV-2 was detected by FMS in 10/40 (25%) samples (4/6 individuals), with no correlation between NPS and FMS RNA signals. All samples were negative by day 14 post diagnosis. Sustained FMS positivity with higher viral RNA signals showed a trend towards disease severity. In cohort 2, 19/47 (40%) individuals exhaled SARS-CoV-2 RNA extending over five orders of magnitude. FMS positive participants were older (positive: median age [IQR] 71 [61-84] vs negative: 61 [45-73], p=0.04) with more comorbidities (positive: 2 [1-3] vs negative: 1 [0-2], p<0.001) and have active cough (positive: 68% vs negative: 24%, p=0.003) and breathlessness (positive: 74% vs negative: 32%, p=0.005) during sampling, compared to FMS negative patients. Of five patients who were FMS positive and asymptomatic at time of sampling, two died of severe COVID-19 pneumonia within one month of follow up. Interpretation FMS detects exhaled SARS-COV-2, with stronger signals in those who develop severe disease.