新冠病毒抗原快速篩檢的診斷表現與在疫情爆發即時監測的應用

Abstract

背景 2021年5月台灣爆發第一波新冠病毒感染疫情，黃金診斷標準核酸檢測昂貴耗時且量能不足，醫療照護機構開始使用抗原快篩作為診斷新冠病毒感染參考，政府開設應用抗原快篩的社區篩檢站，然而台灣缺乏本土快篩試劑於真實臨床情境使用下的診斷表現資料，而抗原快篩應用於社區篩檢站的角色與效果尚待研究。

方法 本研究利用台灣北部兩個體系五間醫院於2021年5月至10月間病患抗原快篩與核酸檢測配對資料，研究抗原快篩臨床使用診斷效力，以及其陽性預測值隨疫情階段改變的變化，並分析抗原快篩偽陰性之核酸檢測CT值與抗原快篩真陽性之差異。本研究同時分析該波疫情中社區篩檢站之每日檢驗工具陽性率數據，分析檢驗陽性率不同延遲值與每日病例數之相關性，並建立病例數之線性預測模型。

結果 全研究區間45,735 對同日採檢抗原快篩與核酸檢測配對結果中，抗原快篩診斷表現敏感度為0.552、特異度0.995、陽性預測值0.614、陰性預測值0.994、診斷勝率為258.30；抗原快篩陽性預測值由疫情成長期的0.653降低至低度流行期的0.026，抗原快篩陽性之核酸檢測CT值顯著較抗原快篩偽陰性組低。 本研究發現社區快篩站檢驗陽性率與每日病例數高度相關且顯示出領先趨勢，抗原篩檢陽性率較核酸檢測陽性率率先反應病例數變化，而陽性率線性模式是預測病例數之良好模型。

結論 新冠病毒抗原快篩為一低敏感度，高特異度的檢測，臨床上無法取代核酸檢測用以診斷新冠病毒感染。雖然敏感度低，但是抗原快篩陽性代表病患病毒量高，傳染性較強。社區篩檢站使用快篩做大規模篩檢，除了迅速找出感染者以執行個人防疫措施之外，篩檢站檢測陽性率是疫情變化的領先趨勢指標。

Background Real-time Reverse Transcriptase Polymerase Chain Reaction (RT-PCR) detecting the RNA of SARS-CoV-2 is the gold standard for the diagnosis of COVID-19 disease. However, it is expensive, labor, and time-consuming. During the first wave of the Covid outbreak in May of 2021 in Taiwan, healthcare providers started utilizing SARS-CoV-2 antigen rapid tests (RAT) as a diagnostic tool to facilitate COVID-19 diagnosis, and the government activated the RAT-based mass screening program in the community. Although the RAT is used enormously in daily clinical practice, the diagnostic performance of RAT in clinical settings is lacking. And the role of RAT as a mass screening tool in community Covid screening program is not yet studied.

Methods We analyzed Covid RAT and RT-PCR paired results from the 1st of May to the 31st of October in 2021 from five branch hospitals of two hospital systems in northern Taiwan. The duration of enrollment included the whole Covid-19 outbreak period. Using RT-PCR as the gold standard, we calculated the sensitivity, specificity, positive predictive value, negative predictive value, and diagnostic odds ratio, regarding different outbreak stages. The cycle threshold value (CT value) from RT-PCR-diagnosed COVID-19 patients was compared between RAT positive (true positive) group and RAT negative (false negative) group. Using the daily test outputs from the4 largest community screening stations near the outbreak center, we calculated the Pearson correlation coefficient between the 3-day moving average of the nationwide COVID-19 case number and the plus to minus seven days lag value of the polled 3-day positivity rate.

Results We examined 45,735 RAT and RT-PCR paired results during the 6-month study period. The sensitivity of RAT compared to RT-PCR is 0.552, specificity 0.995, positive predictive value 0.614, negative predictive value 0.994, and diagnostic odds ratio 258.30. The RAT positive predictive value decreased from 0.653 to less than 0.026 as the outbreak was controlled. The median RT-PCR CT value of the RAT-positive group is significantly lower than that of the RAT-negative group (p-value < 0.0001). The positivity rate output from community screening stations is strongly associated with the number of confirmed cases, with the LAG3 of the RAT positivity rate yielding the highest Pearson correlation (r=0.9683).

Conclusion SARS-CoV-2 rapid antigen tests are specific but not adequately sensitive to exclude the diagnosis of COVID-19. Although its sensitivity is not satisfactory, COVID-19 patients with negative RAT results generally carry lower viral load and are less contagious. In addition to individual diagnostic value, the daily positive rate output from the RAT-based community screening program is a leading indicator of outbreak trend and a good predictor of case number.