Swabs Collected by Patients or Health Care Workers for SARS-CoV-2 Testing.

The early medical response to the COVID-19 pandemic in the United States was limited in part by the availability of testing. Testing guidelines for the SARS-CoV-2 virus required health care workers to collect a swab of the oropharynx (OP) or nasopharynx (NP). This potentially increased transmission risk to health care workers who lacked sufficient personal protective equipment (PPE) due to wide-spread shortages[1].

In other clinical experiences[2,3], obtaining a tongue, nasal, or mid-turbinate (MT) sample is faster, better tolerated, and causes less potential for sneezing, coughing, and gagging, than an NP swab. Additional recent evidence supports the validity of non-NP samples for SARS-CoV-2 detection[4,5]. If also collected by the patient, such methods would reduce health care worker high-exposure and preserve limited PPE.

We enrolled over 500 individuals seen in any one of five ambulatory clinics in the Puget Sound, Washington region with symptoms indicative of upper respiratory infection. Participants were provided instructions and asked to self-collect tongue, nasal, and MT samples, in that order, followed by an NP sample collected by a health care worker. All samples were submitted for RT-PCR testing at a reference laboratory which yielded qualitative results (positive/negative) and Ct values report for positive samples only. See the on-line supplement for additional inclusion criteria, sampling methods, and statistical design.

The study was powered to a one-sided test to determine if the non-NP sensitivities were significantly greater than 90%. For 80% power, 48 NP positive patients were needed given an assumed true sensitivity of 98%. Pairwise analyses were conducted between each patient-collected sample and the health care worker-collected NP samples, which served as the comparator across all. Of the 501 patients with tongue and NP samples, 450 were negative by both swabs, 44 were positive by both swabs, 5 were NP positive and tongue negative, and 2 were tongue positive and NP negative. Of the 498 patients with nasal and NP samples, 447 were negative by both swabs, 47 were positive by both swabs, 3 were NP positive and nasal negative, and 1 was nasal positive and NP negative. Of the 504 patients with MT and NP samples, 452 were negative by both swabs, 50 were positive by both swabs, 2 were NP positive and MT negative, and 0 were MT positive and NP negative.

Using health care worker-collected NP samples as the comparator, estimated sensitivities of the patient-collected tongue, nasal, and MT samples and their one-sided 95% confidence intervals were 89.8% (80.2 – 100.0), 94.0% (84.6 – 100.0), and 96.2% (87.7 – 100.0), respectively. While the sensitivity of the nasal and MT samples were greater than 90%, none of the patient-sample sensitivities were statistically significant when tested using a one-sided test of proportions (p-values 0.50, 0.24, and 0.11 for tongue, nasal, and MT, respectively). Ct values from the RT-PCR demonstrated Pearson correlations of 0.48, 0.78, and 0.86 between the positive NP results and the positive tongue, nasal, and MT results, respectively. Panel A shows plots of the Ct values for the patient collected sites against the NP site, with a linear regression fit super-imposed on the scatterplot. For patients that tested positive by both the NP and one of the test location swabs, the Ct values for the tongue, nasal, and MT swabs were less than the NP Ct values 18.6%, 50.0%, and 83.3% of the time, respectively, indicating that viral load may be higher in MT than NP, and equivalent between nasal and NP. See the on-line supplement for additional results.

This work demonstrates the clinical utility of using patient-collected tongue, nasal, or MT sampling to health care worker-collect NP sampling for diagnosis of COVID-19. Adoption of patient-collected sampling techniques should help reduce PPE use, and provide a more comfortable patient experience. Our analysis was cross-sectional, in a single geographic region, and limited to single comparisons to NP. Despite these limitations, we believe that self-collected samples for SARS-CoV-2 testing from sites other than NP is a useful approach during the COVID-19 pandemic.

Plots showing the Cycle Threshold (Ct) values of the tongue, nasal, and MT tests against those of the comparator NP test. The correlation coefficient is superimposed on each sub-figure along with a trend line estimated using a simple linear regression. Figure 1a) shows Ct values from the 43 patients that had positive tongue and NP results and available Ct values. Figure 1b) shows Ct values from the 46 patients that had positive nasal and NP results and available Ct values. Figure 1c) shows Ct values from the 48 patients that had positive MT and NP results and available Ct values.

Click here for additional data file.