Background: The adoption of saliva as a specimen type for SARS-CoV-2 mass surveillance can significantly increase population compliance with decreased exposure risk for healthcare workers. However, studies evaluating the clinical performance of saliva compared to nasopharyngeal swab (NPS) samples have demonstrated conflicting results regardless of the collection being in healthcare or community settings. Further, pooled testing with saliva remains a challenge owing to the ambiguous sensitivity, limit of detection (LoD), and processing challenges. To overcome these limitations, SalivaAll protocol was developed and validated as a cost-effective measure that must be used on saliva collected in health care or community settings with pooling utility for SARS-CoV-2 mass surveillance.
Methods: The study evaluated 429 matched NPS and saliva samples collected from 344 individuals in either healthcare or community setting. In phase I (protocol U), 240 matched NPS, and saliva samples were tested for SARS-CoV-2 detection by RT-PCR. In phase II (SalivaAll protocol), 189 matched NPS and saliva samples were tested, with an additional sample homogenization step for saliva before RNA extraction, followed by RT-PCR. Eighty-five saliva samples were evaluated with both protocols (U and SalivaAll). Subsequently, adopting SalivaAll protocol, a five-sample pooling strategy was evaluated for saliva samples based on FDA recommendations.
Results: In phase I, 28.3% (68/240) samples tested positive for SARS-CoV-2 from either saliva, NPS, or both. The detection rate was lower in saliva compared to NPS samples (50.0% vs. 89.7%). In phase II, 50.2% (95/189) samples tested positive for SARS-CoV-2 from either saliva, NPS, or both. The detection rate for SARS-CoV-2 was higher in saliva compared to NPS testing (97.8% vs. 78.9%). Of the 85 saliva samples evaluated by both protocols, 57.6% (49) tested positive for SARS-CoV-2 with either protocol U, SalivaAll, or both. The detection rate was 100% for samples tested with SalivaAll, whereas it was 36.7% with protocol U. Also, the LoD with SalivaAll protocol was 20 copies/ml. The pooled testing approach demonstrated a 95% positive and 100% negative percent agreement.
Conclusion: This single-site study demonstrated the variability of results reported in the literature for saliva samples, and found that the discrepancies are explained by processing challenges associated with saliva samples. We have optimized a protocol for saliva samples that results in higher sensitivity compared to NPS samples and also breaks the barrier to using pooled saliva testing for SARS-CoV-2.
SalivaAll: Clinical validation of a sensitive test for saliva collected in healthcare and community settings with pooling utility for SARS-CoV-2 mass surveillance.
