Molecular diagnosis on nasopharyngeal swabs (NPS) is the current standard for COVID-19 diagnosis, but saliva may be an alternative specimen to facilitate access to diagnosis. We compared analytic performances, feasibility and acceptability of NPS, saliva, and oral-self sampling swab for the detection of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2). A prospective, multicenter study was conducted in military hospitals in France among adult outpatients attending COVID-19 diagnosis centers or hospitalized patients. For each patient, all samples were obtained and analyzed simultaneously with RT-PCR or transcription-mediated amplification method. Clinical signs, feasibility, and acceptability for each type of sample were collected. A total of 1220 patients were included, corresponding to 1205 NPS and saliva and 771 OS. Compared to NPS, the sensitivity, specificity, and kappa coefficient for tests performed on saliva were 87.8% (95% CI 83.3–92.3), 97.1% (95% CI 96.1–98.1), and 0.84 (95% CI 0.80–0.88). Analytical performances were better in symptomatic patients. Ct values were significantly lower in NPS than saliva. For OS, sensitivity was estimated to be 61.1% (95% CI 52.7–69.4) and Kappa coefficient to be 0.69 (95% CI 0.62–0.76). OS was the technique preferred by the patients (44.3%) before saliva (42.4%) and NPS (13.4%). Instructions were perceived as simple by patients (> 90%) for saliva and OS. Finally, the painful nature was estimated to be 0.9 for OS, on a scale from 0 to 10, and to be 5.3 for NPS. Performances of OS are not sufficient. Saliva is an acceptable alternative to NPS for symptomatic patient but the process required additional steps to fluidize the sample.

鼻咽拭子 (NPS) 的分子诊断是 COVID-19 诊断的当前标准，但唾液可能是促进诊断的替代标本。我们比较了 NPS、唾液和口腔自采样拭子在检测严重急性呼吸综合征冠状病毒 2 (SARS-CoV-2) 方面的分析性能、可行性和可接受性。在法国的军队医院对 COVID-19 诊断中心的成人门诊患者或住院患者进行了一项前瞻性多中心研究。对于每位患者，所有样本均采用 RT-PCR 或转录介导的扩增方法同时获得并进行分析。收集了每种类型样本的临床症状、可行性和可接受性。总共包括 1220 名患者，对应于 1205 NPS 和唾液和 771 OS。与 NPS 相比，唾液检测的敏感性、特异性和 kappa 系数分别为 87.8% (95% CI 83.3–92.3)、97.1% (95% CI 96.1–98.1) 和 0.84 (95% CI 0.80–0.88)。有症状患者的分析表现更好。NPS 中的 Ct 值显着低于唾液。对于 OS，敏感性估计为 61.1% (95% CI 52.7–69.4)，Kappa 系数为 0.69 (95% CI 0.62–0.76)。在唾液 (42.4%) 和 NPS (13.4%) 之前，OS 是患者 (44.3%) 首选的技术。患者 (> 90%) 认为关于唾液和 OS 的说明很简单。最后，OS 的疼痛性质估计为 0.9，范围从 0 到 10，NPS 为 5.3。操作系统的性能不够。对于有症状的患者，唾液是 NPS 的可接受替代品，但该过程需要额外的步骤来流化样品。