The diagnostic accuracy of saliva in detecting novel coronavirus in symptomatic or asymptomatic people at risk of COVID-19 exposure: A systematic review and meta-analysis of diagnostic test accuracy (DTA) studies.

Objectives: The present systematic review aimed to evaluate the diagnostic accuracy of saliva in detecting novel coronavirus in symptomatic or asymptomatic people at risk of exposure to COVID-19. Materials and methods: The comprehensive searches were made using four electronic databases [MEDLINE (via PubMed), EMBASE, Google Scholar, and The Cochrane Central Register of Controlled Trials] till 31st April 2020 and clinical trials meeting the predefined inclusion criteria were included. The Real-time polymerase chain reaction (RT-PCR) test analyzing the salivary samples was considered as an index test and analyzing the nasopharyngeal/Oropharyngeal/nasal/throat swabs was considered as a reference standard test. The diagnostic accuracy was measured using sensitivity and specificity, positive and negative likelihood ratios, diagnostic odds ratio, and summary receiver operating curves (SROC). The risk of bias was analyzed using the revised tool for the quality assessment of diagnostic accuracy studies (QUADAS -2). The random effect models were used to pool sensitivity and specificity and the extracted data were analyzed and interpreted using MetaDisc software. Results: From the initial search of 160 studies, 10 studies were included for qualitative analysis, and 4 studies were included for quantitative analysis. The salivary diagnostic showed good sensitivity [0.89 (95% CI: 0.81-0.95; p=0.000; I2 = 92.5%)] and specificity [0.95 (95% CI: 0.92-0.98; p=0.000; I2 = 94.7%)] with “High” quality evidence. The area under the SROC curve was 0.97, inferring an ‘excellent’ diagnostic potential of saliva in detecting COVID-19 patients. Conclusion: The accuracy of salivary samples in detecting the 2019-nCoV in symptomatic and asymptomatic people at risk of COVID-19 exposure indicates a high sensitivity (0.89) and high specificity (0.95) with 'high' certainty of evidence.

研究目的：本项系统综述旨在评估唾液检测新型冠状病毒的诊断准确性，研究对象为存在新冠病毒暴露风险的有症状或无症状人群。
材料与方法：本研究通过MEDLINE（通过PubMed检索）、EMBASE、Google Scholar及Cochrane对照试验中心注册库四个电子数据库开展全面检索，检索时限截至2020年4月31日，纳入符合预设纳入标准的临床试验。以唾液样本为检测对象的实时聚合酶链反应（Real-time polymerase chain reaction，RT-PCR）作为指标检测方法，以鼻咽/口咽/鼻腔/咽喉拭子检测作为参考标准检测方法。本研究采用灵敏度（Sensitivity）、特异度（Specificity）、阳性似然比（Positive Likelihood Ratio）、阴性似然比（Negative Likelihood Ratio）、诊断比值比（Diagnostic Odds Ratio）、置信区间（Confidence Interval，CI）以及综合受试者工作特征曲线（Summary Receiver Operating Characteristic，SROC）评估诊断准确性；采用诊断准确性研究质量评价修订工具（QUADAS-2）分析偏倚风险。研究采用随机效应模型合并灵敏度与特异度，并使用MetaDisc软件对提取的数据进行分析与解读。
结果：初步检索共获取160项研究，其中10项纳入定性分析，4项纳入定量分析。唾液检测展现出良好的灵敏度[0.89（95%置信区间：0.81~0.95；P=0.000；I²=92.5%）]与特异度[0.95（95%置信区间：0.92~0.98；P=0.000；I²=94.7%）]，证据质量等级为“高”。综合受试者工作特征曲线下面积为0.97，提示唾液检测新冠病毒具备“优秀”的诊断潜能。
结论：针对存在新冠病毒暴露风险的有症状及无症状人群，唾液样本检测2019新型冠状病毒的诊断准确性体现为高灵敏度（0.89）与高特异度（0.95），证据确定性为“高”。