HSE COVID19 Antigen Testing Working Group Antigen Test Validation Summary Report

Title: COVID 19 Antigen Test Validation Summary

Report Date: 4th June 2021

Report produced by the HSE COVID-19 Antigen Testing Working Group

Executive Summary:

The Irish Antigen Project Evaluation Working Group was established in October 2020, to perform independent and sitespecific validations in Ireland. It has undertaken desktop evaluations to identify assays suitable for further evaluation, and site-specific evaluations in acute hospitals, in meat processing plants and in community swabbing centres. The validation has focussed on symptomatic individuals in line with intended use, with one of the assays also extensively validated in a cohort of asymptomatic individuals. A total of seven antigen test have been validated/verified, 6 lateral flow tests (LFT’s) and one microfluidic device with reader.

The key findings from this validation work are as follows:

• Reverse transcription polymerase chain reaction (RT-PCR) is the gold standard test for detection of SARS-CoV2. It is the most sensitive technology for detection of SARS-CoV-2 virus components in a sample when the virus components are present at low levels. The currently available antigen diagnostic tests (ADTs) require higher levels of virus components in the sample to generate a positive result, when compared to RT-PCR.

• Although the threshold for virus detection with currently available SARS-CoV-2 ADTs is higher than for RT-PCR tests, ADTs may have applications in specific contexts, due to their suitability for deployment in different settings, faster turnaround times or because there is a clinical preference for a test that only detects the virus when it is present at higher levels.

• A preference for detection of virus only when present at high levels could arise if there was evidence that SARsCoV-2 was only significant in clinical or public health terms when present at high levels. Based on current evidence, people can be seriously ill and can be infectious or be about to become infectious – even when the virus is only detectable at low levels. Therefore testing systems that detect virus at the lowest possible level are generally preferred.

• When tested in a real-world setting, validation of the actual performance of ADTs within their intended use, in accordance with manufacturer’s instructions highlighted significant differences from manufacturers claimed test performance. This finding is in line with international experience, and highlights the importance of defining the performance characteristics of the ADTs compared to the gold standard method to make good decisions about how those tests should be used with maximum benefit and the least harm.

• The validation work focussed on verification of the performance of several ADTs as a diagnostic test, for use with both nasopharyngeal and nasal collected sample types in symptomatic individuals. The PanBio COVID-19 Rapid Test (Nasal) was evaluated as an asymptomatic screening test for individuals working in meat processing plants.

• Where a public health risk assessment indicates likely utility of ADTs, for example in a suspected outbreak or among vulnerable populations, testing a series of samples from suspected cases by a method such as ADT, that is readily deployed and provides near real time results can be valuable, even when that method does not reliably detect all infected or infectious people. PCR remains the most sensitive method for detecting infectious virus, however if deploying ADTs in an outbreak setting, nasopharyngeal ADTs with the highest sensitivity, demonstrated by our independent validations, would be preferred over ADTs that use nasal sampling.

• Well characterised ADTs may also have a role as a supplement to RT-PCR testing in the event of circumstances in which PCR capacity is not adequate to meet requirements. In such circumstances symptomatic people with a not-detected ADT would require further testing, either with PCR, or a second ADT 2-3 days later.

• Among asymptomatic workers in meat processing plants with positive RT-PCR results for SARS-CoV-2, 51.9% had a positive test with the validated nasal ADT test (i.e. a test sensitivity of 51.9%). For use with asymptomatic individuals, the sensitivity of the Abbot Panbio COVID-19 Ag Rapid Test (Nasal) even taking into account higher viral RNA levels, is below the minimum requirements set out by WHO and ECDC.

• The lower sensitivity of an individual test can be compensated for to some extent by frequent testing. Modelling studies by HIQA have predicted that if this test is employed in a serial screening programme in meat processing plants, with supervised self-swabbing and ADTs performed by trained individuals, once or twice a week this would be a viable alternative to the current procedure of monthly RT-PCR testing in terms of reducing spread of COVID-19. However, this is modelling data which implies full compliance with a testing regimen of once or twice per week, (by the food and business organisation (FB0) organising the testing, supervised self-swabbing and the workers being tested) which in the real-world setting may not be achievable. We know from the validation work and subsequent implementation of ADTs at meat processing plants that there are significant practical and logistical issues around implementation and operationalising of ADTs in this setting.

• A single ADT, even under optimal conditions of use, will not detect a significant proportion of people who would be identified by RT-PCR as infected and potentially infectious for others. On that basis it is not recommended as a single stand-alone test and is not the preferred method to maximise detection of infected or infectious people.

• ADTs are highly specific, which means that people who test positive by ADT almost always test positive also by RT-PCR. However, it is important to stress that even with a highly specific test (antigen or RT-PCR) the proportion of all positive tests that are false positives increases as number of infected people in the population tested declines. A test system with 99% specificity, is expected to generate an equal number of false positive and true positive test results, if only 1% of the people tested truly has the infection. If we consider the current estimated prevalence in Ireland which is 0.1%, only 1 in 10 positive results will be true positives. Hence as the prevalence of infection in the population tested decreases, ADT positive results will require confirmation with PCR, to prevent inappropriate isolation or broader public health actions on the basis of false positive ADTs.

Read more here.