BackgroundNasopharyngeal antigen Rapid Diagnostic Tests (RDTs) and saliva RT-PCR have shown variable performance to detect SARS-CoV-2.MethodsIn October 2020, we conducted a prospective trial involving patients presenting at testing centers with symptoms of COVID-19. We compared detection rates and performance of RDT, saliva PCR and nasopharyngeal (NP) PCR.ResultsOut of 949 patients enrolled, 928 patients had all three tests. Detection rates were 35.2% (95%CI 32.2-38.4%) by RDT, 39.8% (36.6-43.0%) by saliva PCR, 40.1% (36.9-43.3%) by NP PCR, and 41.5% (38.3-44.7%) by any test. For those with viral loads (VL) ≥106 copies/ml, detection rates were 30.3% (27.3-33.3), 31.4% (28.4-34.5), 31.5% (28.5-34.6), and 31.6% (28.6-34.7%) respectively.Sensitivity of RDT compared to NP PCR was 87.4% (83.6-90.6%) for all positive patients and 96.5% (93.6-98.3%) for those with VL≥106. Sensitivity of STANDARD-Q®, Panbio™ and COVID-VIRO® Ag tests were 92.9% (86.4-96.9%), 86.1% (78.6-91.7%) and 84.1% (76.9-89.7%), respectively. For those with VL≥106, sensitivities were 96.6% (90.5-99.3%), 97.8% (92.1-99.7%) and 95.3% (89.4-98.5%) respectively. Specificity of RDT was 100% (99.3-100%) compared to any PCR. RDT sensitivity was similar <4 days (87.8%) and ≥4 days (85.7%) after symptoms onset (p=0.6). Sensitivities of saliva and NP PCR were 95.7% (93.1-97.5%) and 96.5% (94.1-98.1%), respectively, compared to the other PCR.ConclusionsThe high performance of RDTs allows rapid identification of COVID cases with immediate isolation of the vast majority of contagious individuals. RDT could be a game changer in primary care practices, and even more so in resource-constrained settings. PCR on saliva can replace NP PCR.ClinicalTrial.gov Identifier: NCT04613310