Varying Accuracy of COVID-19 Antibody Tests
That was the essential conclusion of a meta-analysis conducted by an international team of researchers and published in July in The BMJ. The authors reviewed more than 5,000 studies. And he ultimately included just 40 in his analysis. As part of their work, they conducted 49 risk of bias assessments, for each population and method evaluated in the studies, and found "high or unclear risk of bias" in 98% and evidence of bias in the interpretation of the serological test in 73. %
Notably, only 4 of the 40 studies included outpatients, and only 2 studies evaluated point-of-care tests.
According to the investigators, the pooled sensitivity of ELISAs measuring IgG or IgM was 84.3% (range 75.6% to 90.9%). However, the pooled sensitivity for lateral flow immunoassays (LFIA), the logical point-of-care test, was 66% (range 49.3% to 79.3%), while it was 97.8% % (range 46.2% to 100%) for chemiluminescents. immunoassays (CLIA).
Meanwhile, the pooled specificities for all platforms ranged from 96.6% to 99.7%. Of the samples used to assess specificity in the included studies, 83% were from populations tested before the start of the pandemic or included participants not suspected of having COVID-19. Among the LFIAs evaluated in the included studies, the combined sensitivity of the commercial kits (65.0%; range 49.0% to 78.2%) was lower than that of the non-commercial tests (88.2%; range from 83.6% to 91.3%).
"Existing evidence on the diagnostic accuracy of serologic testing for COVID-19 was found to be characterized by high risks of bias, heterogeneity, and limited generalizability to point-of-care testing and outpatient populations," the researchers wrote. . "Future studies to evaluate serological tests for COVID-19 should be designed to overcome the main limitations of the existing evidence base... Higher-quality clinical studies evaluating the diagnostic accuracy of serological tests for COVID-19 are urgently needed." 19".
More recently, in a review published in September in PLOS Pathogens, researchers evaluated 10 SARS-CoV-2 antibody test platforms (1 CLIA, 2 ELISAs, and 7 LFIAs) using "highly specific in-house ELISAs...developed for the detection of anti -spike, -receptor binding domain and -nucleocapsid antibodies (N)” and 110 positive samples for SARS-CoV-2 and 50 pre-pandemic negative.
Like their counterparts on the BMJ team, they found "wide variation" in the performance of different platforms, with specificity ranging from 82% to 100% and overall sensitivity from 60.9% to 87.3%. However, in direct comparisons on identical sample sets, they also found performance to be "highly dependent" on the timing of sampling, with sensitivities of over 95% observed in several tests when evaluating samples more than 20 days after the start of sampling. symptom
“Although LFIAs [generally] lack the semi-quantitative information provided by ELISAs, they have a clear utility advantage over ELISA or [CLIA]-based technologies,” the PLOS Pathogens team noted. “The pre-selected tests, combined with confirmatory reflex tests using our in-house ELISA, are now being taken to extended validations as part of a pilot clinical service at [our hospitals]. These incremental steps that maintain the scope of different technologies, while continuing to define multiple different use cases, will help determine the clinical utility and cost-effectiveness of COVID-19 serological testing in healthcare settings, both in the hospital and at home. community.