Stability of saliva specimens
To confirm the stability of saliva samples, we performed 36 serial tests on 8 primary SARS-CoV-2 positive saliva samples collected from patients 48 hours after their NPS tested positive. The data are summarized in Tables 1-3. Testing was randomized by different time intervals, the time of day, and the shifts of technicians. It includes repeated testing of two samples after 22h (Table 2) and serial repeats of one sample (Table 3) with Ct value reporting.
All results confirmed a positive result after 22-25h of the first test being conducted. Samples were stored refrigerated at 2-6⁰C between tests.
Comparison of saliva and nasopharyngeal/oropharyngeal samples
Between May 13 and September 21, 2020, 431 tests on saliva samples were performed, and the results were compared with paired NPS RT-PCR test results. Using NPS RT-PCR as the reference, there were 104 PCR positive samples and 327 negative patients, including 125 pediatric (age 5-17 years (average 11.3 years)) samples. The observed RT-PCR positive results were divided into groups according to the number of days after the onset of symptoms: 0-14 days (22 patients); 15-30 days (52 patients); 31-70 days (20 patients) and 10 asymptomatic patients. The results of the comparison of the mentioned subgroups are presented in Table 4.
Sensitivity and specificity for the respective groups were: for 0-14 days 90% (95%CI 68.3%-98.7%) and 100% (95%CI 22.4%-100%); for 15-30 days 64.7% (95% CI 46.5%-80.3%) and 83.3% (95%CI 58.9%-96.4%); for 31-70 days 71.4% (95%CI 29.0%-96.3%) and 84.6% (95%CI 54.6%-98.1%); for asymptomatic 71.4% (95%CI 29.0%-96.3%) and 66.7% (95%CI 0.8%-90.6%). The mean Ct values for the groups were 33.9 for NPS and 33.8 for saliva (0-14 days), 34.8 for NPS and 34 for saliva (15-30 days), and 36.5 for NPS and 39 for saliva in the asymptomatic patient group.
Pooling of saliva
To evaluate the testing of pooled saliva samples, 37 pools were constituted. There were 15 pools of 5 samples (4 negatives + 1 positive), 13 pools of 10 samples (9 negatives + 1 positive), and 9 pools of 20 samples (19 negatives + 1 positive). The RT-PCR testing for each pool was repeated twice. The total number of measurements was 74. All results were 100% positive.
When using pooling, the number of tests needed per sample was calculated as follows: the total number of tests needed to test all samples is divided by the total number of samples. For example, if 500 samples are being tested in pools of 10, it means 50 tests are initially needed. With a positivity rate of 1%, a maximum of five pools will contain a positive specimen (if each positive sample is in a different pool). The 50 samples from these five pools will have to be retested individually (each of the five positive pools contains 10 samples). Therefore, a total of 100 tests are needed to evaluate all 500 samples and the number of tests per sample is 100/500 = 0.2.
The positivity rate in Latvia on September 23, 2020, was 0.5%.
The approximate positivity rate at which the pooling of 5 samples becomes more efficient than the pooling of 10 samples is 3% (2-4% depending on test price, laboratory load, tests per sample).
Field study
To evaluate the convenience of saliva pool testing by RT-PCR in field conditions, the town of Kuldiga (total population 10,352) with an ongoing outbreak of COVID-19 in a textile factory was selected. In collaboration with local authorities, 4,100 saliva self-sampling kits were distributed to inhabitants at a specially established distribution point. There were four persons dedicated to the distribution of the kits and four drivers to deliver the samples by car to the testing laboratory in Riga (150 km one way). During the study period, 3,660 saliva samples were collected (response rate 91.5%), delivered to the laboratory, and tested in pools of 10 samples. There were a total of 366 pools. In the first round of testing 43 pools were found to be positive and samples from these pools were re-tested individually. The saliva samples from 44 patients were confirmed as positive by RT-PCR. Mean Ct values of pooled samples were 13% higher than individually tested Ct (31.6 versus 27.6). The positivity rate of the tested population was 1.2%. In total 796 RT-PCR tests were performed. At the time of the sample collection, 68.2% (30/44) of patients did not report any symptoms on the questionnaire form.
Nasopharyngeal swab tests and saliva tests were compared with the number of new COVID-19 cases. The findings illustrate that the second peak in new cases occurred 6-7 days after the initial peak. Such findings correspond with the mean incubation period of the SARS-CoV-2. The results reflect a decrease in new cases in the following observation period (Figure 1 Comparison of confirmed new cases with testing in field study).
Contactless self-service terminals and saliva testing on national level
In total, (between September 29, 2020 and April 10, 2021) 415 673 RT-PCR tests of saliva were conducted. Tests were used for screening purpose of local population in selected municipalities, healthcare workers and other essential teams, social care centers, schools, pre-school groups, participants of special meetings, sport competitions and award ceremonies. Contactless self-service terminals were accepted in hospitals, factories and municipalities as an additional way to submit self-collected saliva for testing. The order-sample-result flow for contactless solution is represented Figure 2 Process Flow. Until April 10. 2021, 6 selfservice terminals have been installed. This option was used for 14294 saliva samples (3,4% of all saliva samples). The total number of performed RT-PCR tests on saliva samples in our laboratory are presented in Table 5. Percentage of saliva samples used for testing inceased continuously month after month reaching 67% of all SARS-CoV-2 tests performed in our laboratory in March 2021.