S1 Data -

Background
Nasopharyngeal swab (NPS) remains the recommended sample type for Severe Acute Respiratory Syndrome Coronavirus-2 (SARS-CoV-2) diagnosis. However, the collection procedure causes discomfort and irritation to the patients, lowering the quality of the sample and exposing healthcare workers to risk. Furthermore, there is also a shortage of flocked swabs and personnel protective equipment in low-income settings. Therefore, this necessitates an alternative diagnostic specimen. The purpose of this study was to evaluate the performance of saliva against NPS for SARS-CoV-2 detection using RT-qPCR among COVID-19 suspected patients at Jigjiga, Eastern Ethiopia.
Methods
Comparative cross-sectional study was conducted from June 28 to July 30, 2022. A total of 227 paired saliva and NPS samples were collected from 227 COVID-19 suspected patients. Saliva and NPS samples were collected and transported to the Somali Regional Molecular Laboratory. Extraction was conducted using DaAn kit (DaAn Gene Co., Ltd China). Veri-Q RT-qPCR was used for amplification and detection (Mico BioMed Co, Ltd, Republic of Korea). The data were entered into Epi-data version 4.6 and analyzed using SPSS 25. McNemar’s test was used to compare the detection rate. Agreement between NPS and saliva was performed using Cohen’s Kappa. The mean and median of cycle threshold values were compared using paired t-tests and the correlation between cycle threshold values was measured using Pearson correlation coefficient. P value < 0.05 was considered statistically significant.
Results
The overall positivity rate of SARS-CoV-2 RNA was 22.5% (95% CI 17–28%). Saliva showed higher sensitivity (83.8%, 95% CI, 73–94.5%) than NPS (68.9%, 95% CI 60.8–76.8%). The specificity of saliva was 92.6% (95% CI, 80.6% - 100%) compared to NPS (96.7%, 95% CI, 87% - 100%). The positive, negative, and overall percent agreement between NPS and saliva was 83.8%, 92.6%, and 91.2% respectively (κ = 0.703, 95% CI 0.58–0.825, P = 0.00). The concordance rate between the two samples was 60.8%. NPS showed a higher viral load than saliva. There was low positive correlation between the cycle threshold values of the two samples (r = 0.41, 95% CI -1.69 to -0.98, P >0.05).
Conclusion
Saliva showed a higher detection rate for SARS-CoV-2 molecular diagnosis than NPS and there was significant agreement between the two specimens. Therefore, saliva could be suitable and easily obtainable alternative diagnostic specimen for SARS-CoV-2 molecular diagnosis.