ABSTRACT
SARS-CoV-2 infectious virions have been reported in exhaled breath, but their source remains elusive: breath sampling systems used to date do not separate breath aerosols by size, fail to prevent salivary/fomite contamination, or aerosol size evolution before sample capture. We hypothesised that sampling end-tidal, oral exhaled breath condensate (EBC), after separating large droplets by inertial impaction 4cm from the lips, would quantify viral loads in distal lung-derived fine aerosols (FA). We used a collector (PBM-HALE ) that captures mechanically aerosolised viruses to sample adult participants for <30 min under informed consent;cases symptomatic for <5 days (n=30) or >5 days (n=12), positive by nasopharyngeal swab RT-PCR (Ct>=13.1), were sampled in clinical triage 'red zones', or COVID-19 wards with no mechanical ventilation or open windows. Salivary alpha amylase activity (Salimetrics LLC), or SARS-CoV-2 viral load (VIASURE SARS-CoV-2 (ORF1ab and N gene)) after QIAsymhpony DSP midi extraction, was quantified in 0.2mL FA EBC fractions. No salivary alpha amylase activity was detected in healthy participant FA EBC (>1:1,750 dilution of paired saliva vs assay detection limit (n=300)). No SARS-CoV-2 RNA was detected in FA EBC (1.18mL +/- 0.32 total volume) among any COVID-19 cases (Aug 2020-Jan 2022) at limits of detection of 120 genomes/mL FA EBC or 4.72 genomes/min exhalation. No pre-extraction spike-in control reaction inhibition was observed. No ambient contamination of the alveolar FA EBC was detected with this sampling device. The alveolar fraction of orally exhaled tidal breath lacks detectable SARS-CoV-2 viral load.