Validation of snort-spit saliva in detecting COVID-19 using RT-PCR and Rapid Antigen Detection Test

Ryner Jose D Carrillo; Abigail D Sarmiento; Mark Anthony C Ang; Michelle H Diwa; Cecille C Dungog; Daniel I Tan; Jan Alexis C Lacuata; Jacob Ephraim D Salud; Ramon Antonio B Lopa; John Mark S Velasco; Regina P Berba; Charlotte M Chiong

Return

Validation of snort-spit saliva in detecting COVID-19 using RT-PCR and Rapid Antigen Detection Test

Author: Ryner Jose D. Carrillo ^{1
,

2} ; Abigail D. Sarmiento ¹ ; Mark Anthony C. Ang ^{1
,

2} ; Michelle H. Diwa ^{1
,

2} ; Cecille C. Dungog ¹ ; Daniel I. Tan ¹ ; Jan Alexis C. Lacuata ¹ ; Jacob Ephraim D. Salud ¹ ; Ramon Antonio B. Lopa ^{1
,

2} ; John Mark S. Velasco ³ ; Regina P. Berba ^{1
,

4} ; Charlotte M. Chiong ^{1
,

3
,

5}
Author Information

1. Philippine General Hospital, University of the Philippines Manila
2. College of Medicine, University of the Philippines Manila
3. National Institutes of Health, University of the Philippines Manila
4. College of Medicine, University of the Philippines Manila
5. College of Medicine, University of the Philippines Manila
Publication Type:Journal Article
MeSH: Saliva; SARS-CoV-2; Reverse Transcription; Reverse Transcriptase Polymerase Chain Reaction
From: Acta Medica Philippina 2021;55(2):211-215
CountryPhilippines
Language:English
Abstract: Objective. To determine the diagnostic accuracy of self-collected snorted and spit saliva in detecting COVID-19 using RT-PCR (ssRT-PCR) and lateral flow antigen test (ssLFA) versus nasopharyngeal swab RT-PCR (npRT-PCR). Methods. One hundred ninety-seven symptomatic subjects for COVID-19 testing in a tertiary hospital underwent snort-spit saliva self-collection for RT-PCR and antigen testing and nasopharyngeal swab for RT-PCR as reference. Positivity rates, agreement, sensitivity, specificity, and likelihood ratios were estimated. Results. Estimated prevalence of COVID-19 using npRT-PCR was 9% (exact 95% CI of 5.5% - 14.1%). A higher positivity rate of 13% in the ssRT-PCR assay suggested possible higher viral RNA in the snort-spit samples. There was 92.9% agreement between ssRT-PCR and npRT-PCR (exact 95% CI of 88.4% to 96.1%; Cohen’s Kappa of 0.6435). If npRT-PCR will be assumed as reference standard, the estimated Sensitivity was 83.3% (exact 95% CI of 60.8% to 94.2%), Specificity 93.9% (exact 95% CI of 89.3% to 96.5%), Positive predictive value of 57.7% (exact 95% CI of 38.9% to 74.5%), Negative predictive value of 98.2% (exact 95% CI of 95% to 99.4%), positive likelihood ratio of 3.65 (95% CI of 7.37 to 24.9), negative likelihood ratio of 0.178 (95% CI of 0.063 to 0.499). There was 84.84% agreement (95% exact CI of 79.1% to 89.5%; Cohen’s Kappa of 0.2356) between ssLFAvs npRT-PCR, sensitivity of 38.9% (exact 95% CI of 20.3% to 61.4%), specificity of 89.4% (exact 95% CI of 84.1% to 93.1%), PPV of 26.9% (95% CI of 13.7% to 46.1%), NPV of 93.6% (exact 95% CI of 88.8% to 96.4%), LR+ of 3.67 (95% CI of 1.79 - 7.51), LR – of 0.68 (95% CI of 0.47 - 0.99). Conclusion. Our data showed that snort-spit saliva RT-PCR testing had acceptable diagnostic performance characteristics and can potentially be used as an alternative to the standard nasopharyngeal/oropharyngeal swab RT-PCR test for COVID-19 in certain situations. However, our data also showed that snort-spit saliva antigen testing using lateral flow assay did not offer acceptable performance.
Full text:2779-Article Text-24436-1-10-20210428.pdf