Inaba Shinji, Ikeda Shuntaro, Tsuda Naoki, Sogabe Kyosei, Inoue Katsuji, Nogami Naoyuki, Ishii Eiichi, Yamaguchi Osamu
Department of Cardiology, Pulmonology, Hypertension, and Nephrology, Ehime University Graduate School of Medicine, Toon, JPN.
Department of Community Medicine, Pulmonology, and Cardiology, Ehime University Graduate School of Medicine, Toon, JPN.
Cureus. 2025 Aug 7;17(8):e89589. doi: 10.7759/cureus.89589. eCollection 2025 Aug.
Objectives In Japan, clinical diagnosis based solely on symptoms, without the use of test kits, has been adopted to enable the rapid identification of individuals infected with coronavirus disease 2019 (COVID-19). A history of close contact with COVID-19 patients is a prerequisite for such symptom-based diagnosis. However, the current diagnostic criteria lack objectivity. This study aimed to develop a symptom-based algorithm stratified by vaccination status to support more reliable clinical diagnosis of COVID-19 among individuals with high-risk exposure. Methods This retrospective, single-center study was conducted in Japan between April 2021 and May 2022. An algorithm for predicting COVID-19 infection was developed by comparing symptoms in COVID-19-positive and COVID-19-negative individuals with high-risk exposure. Analyses were stratified by vaccination status, given its potential influence on symptom presentation. Patients A total of 179 individuals with high-risk exposure to COVID-19 patients were included in the analysis. Results The most common setting of close contact was within households or among roommates (55.3%, 99/179), followed by workplace or school settings (26.3%, 47/179). The combination of all three symptoms-fever, sore throat, and cough-demonstrated 100% specificity but low sensitivity, irrespective of vaccination status. Among vaccinated individuals, the combination of sore throat and cough was a more reliable diagnostic indicator, whereas fever was more predictive among unvaccinated individuals. Conclusion The symptom-based diagnostic algorithm developed in this study demonstrated a sensitivity of 65.3% and a specificity of 88.5%, approaching the diagnostic performance of rapid antigen testing. This algorithm may facilitate simple, rapid, and accessible clinical diagnosis of COVID-19 in resource-limited or high-demand settings.
目的 在日本，已采用仅基于症状而不使用检测试剂盒的临床诊断方法，以快速识别感染2019冠状病毒病（COVID-19）的个体。与COVID-19患者的密切接触史是这种基于症状诊断的前提条件。然而，目前的诊断标准缺乏客观性。本研究旨在开发一种按疫苗接种状况分层的基于症状的算法，以支持对高风险暴露个体进行更可靠的COVID-19临床诊断。方法 这项回顾性单中心研究于2021年4月至2022年5月在日本进行。通过比较高风险暴露的COVID-19阳性和COVID-19阴性个体的症状，开发了一种预测COVID-19感染的算法。鉴于疫苗接种状况对症状表现的潜在影响，分析按疫苗接种状况进行分层。患者 共有179名高风险暴露于COVID-19患者的个体纳入分析。结果 最常见的密切接触场景是在家庭内部或室友之间（55.3%，99/179），其次是工作场所或学校环境（26.3%，47/179）。无论疫苗接种状况如何，发热、喉咙痛和咳嗽这三种症状同时出现时特异性为100%，但敏感性较低。在接种疫苗的个体中，喉咙痛和咳嗽同时出现是更可靠的诊断指标，而在未接种疫苗的个体中发热的预测性更强。结论 本研究开发的基于症状的诊断算法敏感性为65.3%，特异性为88.5%，接近快速抗原检测的诊断性能。该算法可能有助于在资源有限或需求高的环境中对COVID-19进行简单、快速且可及的临床诊断。
