ABSTRACT
Background: Testing using nasopharyngeal swabs (NPS) samples is the cornerstone for the control of the COVID-19 pandemic, but the procedure is uncomfortable and generates anxiety, especially in children. We aimed to evaluate the adequacy of oral saliva swab analysis using RT-PCR comparing to NPS by RT-PCR and Antigen Rapid Test (AgRT) on NPS in children. Methods: Cross-sectional multicenter diagnostic study nested in a prospective, observational cohort (EPICO-AEP) carried out between February and March 2021 at 10 hospitals in Spain. Participants were children 0 to 18 years old with symptoms compatible with SARS-CoV-2 infection of ≤5 days of duration attending at emergency departments. Three samples were collected, two NPS (for AgRT and for RT-PCR) and one oral saliva swab for RT-PCR. In patients with discordant results, new NPS was collected for viral culture and original samples were tested for viral RNA subgenomic (sgRNA) study. Results: 1174 children were included in the analysis, aged 3.8 years (IQR, 1.7-9.0), 647/1174 (55.1%) were male and 760/1174 (64.7%) presented fever 1 day before emergency department admission (IQR 1.0-2.0). Overall, 73/1174 (6.2%) patients tested positive in at least one of the techniques. Sensitivity for RT-PCR in oral saliva swab was 72.1% (95%CI, 59.7-81.9) and specificity 99.6% (95%CI, 99.0-99.9);AgRT in NPS was 61.8% (95%CI, 49.1-73.0) and 99.9% (95%CI, 99.4-100). Kappa index for RT-PCR oral saliva swab was 0.80 (95%CI, 0.72-0.88), and for AgRT was 0.74 (95%CI, 0.65-0.84) vs RT-PCR in NPS. A Bayesian model was used to estimate the accuracy assuming that RT-PCR in NPS is not a perfect gold standard. In this model, sensitivity for RT-PCR oral saliva swab was 84.8% (95%Cr 71.5-93.6), and for AgRT, it was 72.5% (95%Cr, 58.8-83.6). Specificity for RT-PCR oral saliva swab was 99.7% (95%Cr, 99.2-99.9), and for AgRT it was 99.9% (95% Cr, 99.6-100). The Cts were higher in oral saliva swabs compared with NPS;being Ct (NPS)=0.5 x (Ct saliva) + 4.5 (p=0.027). Overall, 4 (10.8%) patients with discordant results had a positive culture. In 3 of the 4 patients, the discordance consisted of positive result on oral saliva swab and nasopharyngeal swabs RT-PCR but negative by antigen rapid diagnostic test. No patient had (+) culture, (+)NP, (-)oral swab. Conclusion: RT-PCR on oral saliva swab is an accurate option for SARS-CoV-2 testing in children. A friendlier technique for younger patients, who must be tested very frequently, may help to increase the number of patients tested.
ABSTRACT
Background: The accuracy of rapid antigen tests (RAT) SARS-CoV-2 for in children is unknown. Our aim was to determine the diagnostic accuracy and concordance of the RAT PanBioTM (Abbott) compared to RT-PCR in nasopharyngeal smear (NPS) samples, in symptomatic pediatric population. Methods: This is a descriptive, retrospective, multicentre clinical study nested in a prospective, observational, multicenter cohort study. We included pediatric patients aged 0 to 16 years with symptoms consistent with COVID-19 of ≤5 days of evolution, attended in the Emergency Departments of the seven centers involved. A total of two consecutive NPS were obtained from each patient: one was employed to perform the RAT and the other to perform RT-PCR. Sample size for a non-inferiority study was calculated considering 80% power, for a 5% prevalence and a 90% sensitivity, using RT-PCR as the gold standard reference. A confusion matrix was displayed. Non-inferiority of sensitivity and specificity between diagnostic tests was assessed using the McNemar's test. The agreement between the two methods was calculated using Cohen's kappa index. Results: A total of 1620 patients were tested in 7 hospitals. The overall sensitivity for RAT PanBioTM was 45.4% (95%CI, 34.1-57.2), and specificity was 99.8% (95%CI, 99.4-99.9) (Figure 1). The positive predictive value (PPV) for this 4.8% prevalence was 92.5% (95%CI, 78.6-97.4). The negative predictive value was 97.3 % (95%CI, 96.8-97.8). Positive likelihood ratio (PLR) was high - 233.8 (IC 95%, 73.5-743.3), and negative likelihood ratio (NLR) was low - 0.54 (95%CI, 0.44-0.67). Conclusion: Compared to RT-PCR, the sensitivity of the RAT PanBioTM was low in children with <5 days of symptoms of COVID-19. The specificity and PLR were good, and the NLR and concordance with RT-PCR were only moderate. These results suggest that the test is very good when the result is positive, and that the test has only a limited value when the result is negative. In relation with screening and public health policy, these results should be interpreted considering also rapidness, availability and false positives ratio compared to RT-PCR or other tests.