Accuracy of Self-Collected Saliva and Anterior Nasal Swab Samples Used to Test for SARS-CoV-2 Infection

Close up of a young woman having a nasal swab test done by his doctor
Researchers compared the results of self-collected saliva and anterior nasal swab samples vs nasopharyngeal swab samples collected by healthcare workers used to test for SARS-CoV-2 infection.

Results of self-collected anterior nasal swab (ANS) and saliva samples obtained within 3 to 7 days of COVID-19 symptom onset showed a sensitivity of 80% or higher, according to findings from a cross-sectional study published in Emerging Infectious Diseases.

Researchers from the Centers for Disease Control and Prevention recruited patients (N=1076) who were eligible for SARS-CoV-2 infection testing from a hospital in Georgia between August and November 2020. Patients were educated about the proper self-collection of ANS and saliva samples. Results of self-collected samples were compared with the results of nasopharyngeal swab (NPS) samples collected by healthcare workers (HCW).

Of the patients included in the study, 51.9% were men, 57.1% were aged 50 years and older, 81.2% were Black, and 74.9% had at least 1 chronic condition. A total of 8.0% of NPS samples were positive for SARS-CoV-2 infection.

Among patients who had both ANS and NPS samples collected (n=995), 96.8% of tests were concordant (k, 0.73; 95% CI, 0.64-0.82). The rates of test agreement corresponded with a sensitivity of 59% (95% CI, 47%-70%), specificity of 100% (95% CI, 92%-100%), and negative predictive value of 97% (95% CI, 95%-98%).

Among patients who had both saliva and NPS samples collected (n=921), 97.1% of tests were concordant (k, 0.76; 95% CI, 0.67-0.85). The rates of test agreement corresponded with a sensitivity of 68% (95% CI, 55%-78%), specificity of 99% (95% CI, 99%-100%), and negative predictive value of 90% (95% CI, 96%-98%).

In regard to results of the self-collected samples, the NPS samples had a concordance rate of 97.2% (k, 0.79; 95% CI, 0.71-0.86). Results of both self-collected samples had a combined sensitivity of 71% (95% CI, 60%-81%), specificity of 99% (95% CI, 99%-100%), negative predictive value of 98% (95% CI, 96%-98%), and positive predictive value of 92% (95% CI, 82%-97%).

Among men, Black patients, and those aged 18 to 29 years and 50 to 59 years, results of self-collected saliva samples showed significantly increased sensitivity compared with self-collected ANS samples(all P <.05). However, results of self-collected saliva samples from Hispanic patients showed significantly decreased sensitivity (P <.05).

Cycle threshold (Ct) values were more strongly correlated between ANS and NPS samples (r, 0.75) compared with saliva and NPS samples (r, 0.53). Ct values were increased among patients who reported symptoms of COVID-19 infection (P =.03), as well as those who were tested within 1 week of symptom onset (P <.01). Results of ANS and saliva samples that were self-collected within 7 days of symptom onset had sensitivities of 80% and 85%, respectively.

This study was limited by its lack of power to detect significant differences in sensitivity among the 3 types of samples tested.

These data indicated that the sensitivity of self-administered tests was decreased compared with NPS samples collected by HCWs. In addition, accuracy was impacted by patient characteristics and time since symptom onset.

Reference

Smith-Jeffcoat SE, Koa M, Hoffman A, et al. Effects of Patient Characteristics on Diagnostic Performance of Self-Collected Samples for SARS-CoV-2 Testing. Emerg Infect Dis. 2021;27(8):2081-2089. doi:10.3201/eid2708.210667

This article originally appeared on Infectious Disease Advisor