Performance of nasopharyngeal swab and saliva in detecting Delta and Omicron SARS‐CoV‐2 variants

Abstract A prospective cohort study was conducted during the Delta and Omicron severe acute respiratory syndrome coronavirus type 2 (SARS‐CoV‐2) epidemic waves from paired nasopharyngeal swab (NPS or NP swab) and saliva samples taken from 624 participants. The study aimed to assess if any differences among participants from both waves could be observed and if any difference in molecular diagnostic performance could be observed among the two sample types. Samples were transported immediately to the laboratory to ensure the highest possible sample quality without any freezing and thawing steps before processing. Nucleic acids from saliva and NPS were prospectively extracted and SARS‐CoV‐2 was detected using a real‐time reverse‐transcription polymerase chain reaction. All observed results were statistically analyzed. Although the results obtained with NP and saliva agreed overall, higher viral loads were observed in NP swabs regardless of the day of specimen collection in both SARS‐CoV‐2 epidemic waves. No significant difference could be observed between the two epidemic waves characterized by Delta or Omicron SARS‐CoV‐2. To note, Delta infection resulted in higher viral loads both in NP and saliva and more symptoms, including rhinorrhea, cough, and dyspnea, whereas Omicron wave patients more frequently reported sore throat. An increase in the mean log RNA of SARS‐CoV‐2 was observed with the number of expressed symptoms in both waves, however, the difference was not significant. Data confirmed that results from saliva were concordant with those from NP swabs, although saliva proved to be a challenging sample with frequent inhibitions that required substantial retesting.

laboratories could not simply switch to antigen detection to meet the demand for testing. Rather, rtRT-PCR testing had to be elevated to new, higher levels of throughput and turnaround time, primarily through the use of fully integrated, automated, 5,6 and semiautomated high-throughput systems, 7 pooling of samples, [8][9][10] logistics, 11 and triage of samples through multiple platforms in simultaneous use.
Since the beginning of the pandemic, nasopharyngeal swab (NP swab or NPS) collected in viral transport medium has been considered the preferred specimen for molecular detection of SARS-CoV-2. 12,13 Although this sample is well suited for high-precision SARS-CoV-2 molecular diagnostics because it allows for easy processing, low inhibition, and high detection sensitivity with a simple and rapid collection procedure, it is not without drawbacks. [14][15][16][17][18][19] First, some patients complain about the procedure and find it uncomfortable or even painful. 14,15 Second, the collection procedure cannot be standardized and therefore varies from sample to sample. 16 Third, there have been problems due to shortages in the availability of swabs and viral transport media. 17 Finally, new insights into the tropism of SARS-CoV-2 led to the testing of other samples for their putative higher suitability. 18,19 Thus, oropharyngeal swabs (OPS) and saliva have been suggested as the best samples for molecular detection of SARS-CoV-2, 20 but other sample types such as alveolar lavage fluid, sputum, urine, serum/plasma, whole blood, nasal swabs (NS), corneal secretions, and even anal swabs and stool have been investigated. [21][22][23] Available studies seem not to show the uniformed performance of these samples, with the varying agreement in the detection rate of SARS-CoV-2 RNA, but they mostly agree that the NPS is still a better sampling method choice. 12 In September 2021, Slovenia experienced its fourth  wave caused by the SARS-CoV-2 Delta genomic variant. Delta was rapidly replaced by the Omicron genomic variant in the first 2 weeks of 2022, which was later the cause of the fifth and largest wave to date, according to a national report on the genetic variants of SARS-CoV-2 (data available only in the Slovenian language at: https://www. nlzoh.si/objave/sledenje-razlicicam-sars-cov-2-53/). Although cases detected by rtRT-PCR reached unprecedented numbers in January 2022, hospitals did not experience as great an influx of patients as during the Delta wave.
This study aimed to examine several features of SARS-CoV-2 infection in outpatients during the Delta and Omicron waves.
Specifically, we examined the clinical relevance of saliva samples compared with NPS in general and the genomic variant. We also compared viral load in outpatients infected with the Delta and Omicron genomic variants in NPS and saliva samples and related these data to days after symptom onset, symptom severity, and vaccination status.

| Study design, study population, and sample collection
A head-to-head comparative study included individuals visiting the largest COVID-19 swab collection center in Slovenia for routine NPS collection for SARS-CoV-2 testing. Written informed consent was obtained from all participants included, along with at least 1 ml of selfcollected posterior saliva sample following instructions and supervision by medical personnel onsite. In addition, a short questionnaire on clinical symptoms, such as rhinorrhea, cough, sneezing, sore throat, headache, body temperature, hoarseness, dyspnea, chest pain, duration of illness (in days), sex, age, and vaccination history. The NPS was collected in CITOSWAB VTM (nal von Minden GmbH) and paired saliva samples in an empty Saliva Collector (Biocomma Limited) without any buffer added. In the Delta wave (September-October 2021), 298 individuals, and in the Omicron wave (January 2022), 326 individuals participated.
All samples were processed and analyzed immediately after collection.  Rhinorrhea, cough, and dyspnea were more frequently reported symptoms among Delta wave outpatients, whereas sore throat was more frequently reported among Omicron wave outpatients, as shown in Table 2. Participants in both waves also reported other symptoms that were not already given in the questionnaire, such as joint and/or muscle pain, diarrhea, vomiting, nausea, loss of taste or smell, and sinusitis.
After repeated extraction and rtRT-PCR, a valid result was obtained for all samples, thus all were included in the comparison.

| Viral loads by sample type and genomic variant
Both Delta and Omicron wave outpatients had statistically significantly higher viral loads (p < 0.001) in NPS compared to saliva. A mean of 5.08 log 10 RNA copies/µl (Delta wave) and 4.56 log 10 RNA copies/µl (Omicron wave) was detected in NPS compared with a mean of 3.85 log 10 RNA copies/µl (Delta wave) and 3.26 log 10 RNA copies/µl (Omicron wave) in saliva. A similar statistically significant difference was observed when comparing the same sample types (NPS vs. saliva) between waves. Statistically, significantly higher mean viral loads were observed in NPS (p < 0.001) and saliva (p < 0.001) in Delta wave compared to Omicron wave outpatients.
The results are shown in Figure 1A.
When the SARS-CoV-2 viral load was normalized on the human reference gene (RNA SARS-CoV-2/1000 cell copies) still Delta and Omicron wave outpatients had statistically significantly higher mean RNA SARS CoV-2/1000 cell copies in NPS (p < 0.001) compared to saliva. A mean of 6.58 log 10 RNA SARS-CoV-2/1000 cell copies (Delta wave) and 5.89 log 10 RNA SASR-CoV-2/1000 cell copies (Omicron wave) was detected in NPS compared with a mean of 5.92 log 10 RNA SARS-CoV-2/1000 cell copies (Delta wave) and 5.5 log 10 RNA SARS-CoV-2/1000 cell copies (Omicron wave) in saliva. A similar statistically significant difference was observed when comparing the same sample types (NPS vs. saliva) between waves.
Statistically, significantly higher mean viral loads were observed in NPS (p < 0.001) and saliva (p = 0.003) in Delta wave compared to Omicron wave outpatients. The results are shown in Figure 1B.

| Viral load temporal dynamics by sample type and genomic variant
The dynamics of log 10 RNA copies/µl were assessed in both sample types for Delta and Omicron wave outpatients at 1, 2, 3, 4, 5, and 6 days after symptom onset. The number of SARS-CoV-2-positive outpatients whose samples were collected later than 6 days after symptom onset was small (10 in the Delta wave and 3 in the Omicron wave) and they were therefore excluded from further analysis. The results confirm that mean log 10 RNA copies/µl were consistently higher in NPS compared with saliva in both waves, regardless of the day after symptom onset. This pattern was less pronounced only in outpatients in the Omicron wave whose samples were collected 1 day after symptom onset. In this group, mean log 10 RNA copies/µl in saliva approached the viral load in NPS but decreased more rapidly in subsequent days compared with the mean log 10 RNA copies/µl in NPS, as shown in Figure 2. Compensation for the amount of human DNA in the sample (normalization) did not affect the overall pattern, as shown in Figure 2C,D. When the viral load was normalized in outpatients in the Omicron wave whose samples were collected 1 day after symptom onset, in this group, the mean log 10 RNA SARS-CoV-2/1000 cell copies in saliva was higher than the mean log 10 RNA SARS-CoV-2/1000 cell copies in NPS. However, the viral load in saliva decreased more rapidly in subsequent days compared with mean log 10 RNA SARS-CoV-2/1000 cell copies in NPS, as shown in

F I G U R E 1 Comparison of viral loads in nasopharyngeal (NP) swabs (blue) and saliva (S) samples (red) between Delta and Omicron waves.
Non-normalized (A) and normalized (B) viral loads are presented as log 10 (SARS-CoV-2 RNA copies/µl) or as log 10 (SARS-CoV-2 RNA copies/ 1000 cell copies), respectively. The box represents the first and third quartile, the line in the box, the median, the cross, the mean value, and the whiskers, the minimal and maximal value, excluding outliers that are presented as individual dots. Outliers are defined as values that deviate >1.5 times the interquartile range from the box limits. Statistical significance for individual comparisons is shown as p-values. SARS-CoV-2, severe acute respiratory syndrome coronavirus 2.

| 4707
When looking for correlations between viral load and the specific symptom, statistically significantly higher log 10 RNA copies/µl were observed only in NPS for outpatients that reported rhinorrhea compared with outpatients that did not report this symptom

AUTHOR CONTRIBUTIONS
The manuscript has been read and approved by all authors and was not submitted, published, and accepted for publication elsewhere.