Abstract
Saliva has been a COVID-19 diagnostic specimen of interest due to its simple collection, scalability, and yield. Yet COVID-19 testing and estimates of the infectious period remain largely based on nasopharyngeal and nasal swabs. We sought to evaluate whether saliva testing captured prolonged presence of SARS-CoV-2 and potential infectiousness later in the disease course. We conducted an observational study of symptomatic COVID-19 patients at University Hospital in Newark, NJ. Paired saliva and nasal specimens from 96 patients were analyzed, including longitudinal analysis of paired observations from 28 of these patients who had multiple time-points. Saliva detected significantly more cases of COVID-19 beyond 5 days (86.1% [99/115] saliva vs 48.7% [56/115] nasal, p-value < 0.001), 9 days (79.4% [50/63] saliva vs 36.5% [23/63] nasal, p-value < 0.001) and 14 days (71.4% [20/28] saliva vs 32.1% [9/28] nasal, p-value = 0.010) of symptoms. Additionally, saliva yielded lower cycle thresholds across all time periods, indicative of higher viral loads in saliva. In the longitudinal analysis, a log-rank analysis indicated that the survival curve for saliva was significantly different from the curve for nasal swabs (p<0.001) with a median survival time for saliva of 18 days compared to 13 days for nasal swabs. We additionally performed saliva viral cultures among a similar COVID-19 patient cohort and noted patients with positive saliva viral cultures between 7 to 28 days of symptoms. Findings from this study suggest that SARS-CoV-2 RNA persists longer and in higher abundance in saliva compared to nasal swabs, with potential of prolonged propagating virus. Testing saliva may thus increase yield for detecting potentially infectious virus even beyond the first five days of symptomatic COVID-19.