Clinical accuracy of instrument-based SARS-CoV-2 antigen diagnostic tests: A systematic review and meta-analysis

Abstract

Abstract Background During the COVID-19 pandemic, antigen diagnostic tests were frequently used for screening, triage, and diagnosis because they are faster and less expensive than molecular tests, despite being less sensitive terms of analytical and clinical performance. Novel instrument-based antigen tests (iAg tests) hold the promise of outperforming their instrument-free, visually-read counterparts. Here, we provide a systematic review and meta-analysis of the SARS-CoV-2 iAg tests’ clinical accuracy. Methods We systematically searched MEDLINE (via PubMed), Web of Science, medRxiv, and bioRxiv for articles published before November 7th, 2022, evaluating the accuracy of iAg tests for SARS-CoV-2 detection. We performed a random effects meta-analysis to estimate sensitivity and specificity and used the QUADAS-2 tool to assess study quality and risk of bias. Results We compiled the data from 117 studies involving 95,181 individuals and assessing the clinical accuracy of 24 commercial COVID-19 iAg tests. The studies varied in risk of bias but showed high applicability. Of 24 iAg tests from 99 studies assessed in the meta-analysis, the pooled sensitivity and specificity compared to molecular testing of a paired NP swab sample were 76.7% (95% CI 73.5 to 79.7) and 98.4% (95% CI 98.0 to 98.7), respectively. Out of the eight tests with sufficient data for test-specific meta-analysis, only LumiraDx satisfied the WHO's sensitivity and specificity standards. Higher sensitivity was noted in individuals with high viral load (99.6% [95% CI 96.8 to 100] at Ct-level ≤ 20) and within the first week of symptom onset (84.6% [95% CI 78.2 to 89.3]), but did not differ between tests conducted as per manufacturer’s instructions and those conducted differently, or between point-of-care and lab-based testing. Conclusion Our study reveals that iAg tests' pooled sensitivity increases with lower Ct-values, a proxy for viral load, or within the first week of symptom onset, enabling reliable identification of most COVID-19 cases. While these characteristics are similar to those of instrument-free antigen diagnostics, different types of iAg tests might allow standardized result interpretation to curb human error, automated reporting, upscaling test runs or additional test-specific functions, such as the indication of antigen levels. Depending on their setup, they can be used at the point of care or at a laboratory with high-throughput.