Accurate molecular diagnosis of respiratory viruses depends on obtaining specimens from anatomical sites with active viral replication. In many settings, nasopharyngeal or joint nose-and-throat swabs (NTS) are used for RT-qPCR testing. Nasopharyngeal swabbing requires trained personnel and is invasive, while joint NTS is less invasive and can be used for self-collection. During the COVID-19 pandemic, nose-only swabbing was often preferred by patients because it is easier and less uncomfortable. The Rhinoswab (RhinoMed) is an anterior nasal swab developed to collect nasal discharge from the anterior nares with minimal invasiveness and was introduced in 2020.
The primary objective was to estimate diagnostic concordance between Rhinoswab anterior nasal samples and joint NTS for RT-qPCR detection of influenza A, influenza B, and SARS-CoV-2 among symptomatic outpatients in Hong Kong.
This analysis used data from an ongoing outpatient surveillance study conducted between 03 May 2023 and 29 April 2024. The study enrolled symptomatic patients presenting for medical care with febrile acute respiratory illness, defined as at least two respiratory symptoms within three days of symptom onset. Paired joint NTS and Rhinoswab samples were collected from consenting participants and tested by RT-qPCR for influenza A, influenza B, and SARS-CoV-2. Human ribonuclease P (RNase P) served as an internal control. A cycle threshold (Ct) value below 40 was considered positive.
Of 531 patients invited to provide paired samples, 488 (91.9%) consented and provided both a joint NTS and a Rhinoswab sample; 43 (8.1%) opted out of Rhinoswab collection. Among those who opted out, 16/43 (37.2%) declined after viewing the Rhinoswab demonstration video. The cohort included a wide age range (a junior Rhinoswab was used in 246 of 488 Rhinoswabs, 50.4%), the mean age was 25 years, and 55.5% of participants were female. Participants completed a questionnaire that captured vaccination history and demographics. The study protocol was approved by the Institutional Review Board of the University of Hong Kong and written informed consent was obtained.
Sampling order was standardized: rapid antigen tests were performed first (not part of this analysis), followed by a joint NTS (CLASSIQSwabs™, Copan), and then the Rhinoswab. Rhinoswab collection was performed by trained study staff following manufacturer instructions; the swab was inserted into the anterior nares, left in place for 15 seconds while the participant breathed normally, then moved back and forth for an additional 15 seconds, and the loop was snapped into the collection tube. All swabs were transported in viral transport medium to the laboratory and tested using a ViiA7 RT-qPCR system (ThermoFisher). RNA extraction methods and RT-qPCR targets were consistent with previously described protocols. Human RNase P was included to monitor specimen adequacy.
Only participants with paired joint NTS and Rhinoswab samples were included in analyses. Agreement was assessed by concordance (percent of paired swabs with identical positive/negative results) and Cohen’s kappa with 95% confidence intervals. Concordance was stratified by age group, time since symptom onset, and recent vaccination status (influenza or COVID-19 vaccination within the prior year). Diagnostic performance of the Rhinoswab relative to joint NTS was summarized with sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) with 95% confidence intervals. Correlation of Ct values between paired swabs was evaluated using Pearson’s correlation coefficient. Differences in Ct (Rhinoswab minus NTS) were plotted against NTS Ct values to examine variability across viral load ranges. Analyses were performed in R.
Between May 2023 and April 2024, 488 paired samples were analyzed. Influenza A was detected in 136 participants by both swab types. Influenza B was detected in 42 participants by both swab types, and SARS-CoV-2 was detected in 32 participants by both swab types.
Concordance between Rhinoswab and joint NTS was high across the three viral targets: 98.0% (95% CI: 96.3% to 99.0%) for influenza A, and 99.0% (95% CI: 97.6% to 99.7%) for both influenza B and SARS-CoV-2. Cohen’s kappa values indicated near-perfect agreement: 0.95 (95% CI: 0.92 to 0.98) for influenza A, 0.94 (95% CI: 0.88 to 0.99) for influenza B, and 0.92 (95% CI: 0.85 to 0.99) for SARS-CoV-2. Concordance was consistent across age groups, time since symptom onset, and vaccination status.
Cycle threshold values from Rhinoswab samples were positively correlated with those from joint NTS for all viral targets and for the RNase P internal control. Pearson correlation coefficients were 0.76 for influenza A, 0.818 for influenza B, and 0.729 for SARS-CoV-2; RNase P showed a moderate correlation (r = 0.616). When plotting the difference in Ct values (Rhinoswab minus NTS) against NTS Ct, there was no evidence of systematically greater disagreement at higher or lower Ct values. Most discordant paired samples (16 of 18, 88.9%) had Ct values ≥30, consistent with low viral loads near the assay detection limit.
In this outpatient symptomatic cohort in Hong Kong, Rhinoswab anterior nasal sampling demonstrated high concordance with joint NTS for RT-qPCR detection of influenza A, influenza B, and SARS-CoV-2. Agreement metrics (concordance and Cohen’s kappa) were consistently high, and Ct values from Rhinoswab correlated well with those from NTS. The majority of discordant results occurred at high Ct values (≥30), suggesting that discrepancies were concentrated among specimens with low viral load.
Given their minimal invasiveness and the observed diagnostic agreement, Rhinoswabs appear to be a viable alternative specimen type for supervised respiratory virus testing in symptomatic ambulatory patients. The study supports consideration of anterior nasal sampling where ease of collection and patient comfort are priorities, while recognizing that very low viral load specimens may yield discordant results.
Limitations reported in the source include that participants could opt out of Rhinoswab collection (8.1% opted out), and that Rhinoswab collection followed NTS collection, which could theoretically affect specimen quality; however, these factors did not prevent strong concordance across analyses. Detailed limitations and broader generalizability considerations are provided in the full article.
All data and analysis code supporting the study findings are publicly available in a GitHub repository maintained by the authors. Funding was provided by grants received by one author from the Health and Medical Research Fund and the Research Grants Council of the Hong Kong SAR Government; funders had no role in study conduct or reporting. Competing interests reported by the authors are disclosed in the source article.