Rapid systematic review of the sensitivity of SARS-CoV-2 molecular testing on saliva compared to nasopharyngeal swabs

BackgroundNasopharyngeal sampling has been the standard collection method for COVID-19 testing. Due to its invasive nature and risk of contamination for health care workers who collect the sample, non-invasive and safe sampling methods like saliva, can be used alternatively. MethodsA rapid systematic search was performed in PubMed and medRxiv, with the last retrieval on June 6th, 2020. Studies were included if they compared saliva with nasopharyngeal sampling for the detection of SARS-CoV-2 RNA using the same RT-qPCR applied on both types of samples. The primary outcome of interest was the relative sensitivity of SARS-CoV-2 testing on saliva versus nasopharyngeal samples (used as the comparator test). A secondary outcome was the proportion of nasopharyngeal-positive patients that tested also positive on a saliva sample. ResultsEight studies were included comprising 1070 saliva-nasopharyngeal sample pairs allowing assessment of the first outcome. The relative sensitivity of SARS-CoV-2 testing on saliva versus nasopharyngeal samples was 0.97 (95% CI=0.92-1.02). The second outcome incorporated patient data (n=257) from four other studies (n=97 patients) pooled with four studies from the first outcome (n=160 patients). This resulted in a pooled proportion of nasopharyngeal positive cases that was also positive on saliva of 86% (95% CI=77-93%). DiscussionSaliva could potentially be considered as an alternative sampling method when compared to nasopharyngeal swabs. However, studies included in this review often were small and involved inclusion of subjects with insufficient information on clinical covariates. Most studies included patients who were symptomatic (78%, 911/1167). Therefore, additional and larger studies should be performed to verify the relative performance of saliva in the context of screening of asymptomatic populations and contact-tracing.