The findings of our study can be summarized as follows: (a) for the detection of SARS-CoV-2, saliva is a stable specimen with acceptable specificity and sensitivity at the early stages of infection; (b) saliva specimen is appropriate for pooling, with accurate diagnostic performance.
Saliva has been identified as a reliable testing specimen for SARS-CoV-2 using the RT-PCR approach in several recent studies 6–8. In the updated European Centre for Disease Prevention and Control (ECDC) recommendations, saliva is also mentioned as a convenient specimen for SARS-CoV-2 testing 9,10. The FDA has approved methods for SARS-CoV-2 testing using saliva in several laboratories 11,12. Nevertheless, saliva has been included in IFCC COVID-19 Guidelines on Molecular, Serological, and Biochemical/Haematological Testing as a promising sample type 13.
The high expression of ACE2 receptors in salivary gland cells is the main factor for virus affinity that could lead to active replication and transmission by saliva droplets expelled into the air during coughing or sneezing 14,15.
It is essential to establish standardized sample collection procedures, safe logistics, and reliable testing methods that meet performance requirements. Our stability testing confirmed that saliva samples stayed equally positive up to 24h.
In the analytical context, saliva can have the same or even better performance than NPS 16. It has been previously confirmed to be highly sensitive and specific at the early stage of infection 0–14 days after onset of symptoms and in asymptomatic cases 16,17. Therefore, it is essential to take into consideration the relationship between the dynamics of viral load, Ct values, and the number of days after the onset of symptoms. As described in prior publications, the Ct value is important in determining the infectiousness of a patient sample. Ct values above 34 do not emit infectious virus particles, and values between 27–34 show low viability of the viral load. Samples with a Ct value of 13–17 show positive virus viability 18. Ct values may vary depending on the different assays by up to 5 cycles for the same sample 19. Variation also appears due to the quality of the specimen obtained and the different treatment methods used to prepare samples for testing 20. As expected, we got some variation between the NPS and saliva sample types. In our pilot testing using 44 positive saliva samples we found slightly higher (13%) Ct values in pools, compared to individual tests. Similar findings have been published by other researchers 21. This observed slight difference still allows for the safe use of saliva pooling for surveillance purposes with sufficient diagnostic accuracy. Nevertheless, pooling strategies are described in the latest update of COVID-19 testing strategies and objectives published by the ECDC 10. There is evidence of benefits from pooling in low prevalence countries with a low proportion of positive samples – up to 5%. Recent publications indicate that for populations with a prevalence of less than 1% the testing of saliva pools of 10 or 20 samples is more beneficial, and our independent calculations also bear this out. Comparing the correlation between different pool sizes (usefulness and efficiency) with test positivity rate, we concluded that a pool size of 10 is more efficient than a pool size of 5. The calculations are based on test price, reimbursement conditions, and the number of tests per sample (when using a pooling strategy) 22,23.
The pooling strategy for self-collected saliva samples is the optimal solution that saves resources and reduces the testing time 24. In our real-life field test on a “perifocal” population, we successfully tested nearly a third of the citizens of the town of Kuldiga in three days. This allowed for fast identification of asymptomatic SARS-CoV-2 cases and those with mild symptoms to enable timely contact tracing and outbreak containment. The use of nasopharyngeal swabs in such a situation would have required more staff and time with possible patient compliance issues.
Additionally, improvements may reasonably be expected in terms of organizing the distribution of self-sampling kits to patients as well as the logistics of returning the kits to the laboratory that would further increase the usefulness of screening using a pooled testing strategy.
There were some limitations in our study: the size of the cohort of participants was limited, the selected laboratory-developed RT-PCR method for evaluation was used, also a lack of multicenter observation. Future studies extending the cohort are needed to confirm current findings and provide the implications in clinical practice.
In conclusion, our findings confirm the stability of SARS-CoV-2 RNA in saliva, showing acceptable performance in terms of specificity and sensitivity at the early stages of infection, and the advantages of testing using a pooling strategy in a low prevalence population. Ct values and detected/missing gene information are favorable for the interpretation of the results from several aspects, e.g. epidemiological investigation, determination if the patient is infectious at the current stage, etc. Self-sampling makes the procedure faster, safer, and requires fewer resources. The targeted distribution of test kits among a population with a known outbreak significantly increased the positivity rate. Saliva pool testing on a large scale provides an additional tool to take timely measures and contain outbreaks.