We are clearly in a second epidemic wave that is even larger than the one at the start of 2020. Admittedly, the medical care of patients has greatly improved, however given the presence of Sars-CoV-2 in the population we remain in a far from ideal situation particularly regarding the diagnostic capacity we have at a national level. For this reason, the authorities have taken the decision to develop a new diagnostic tool which, according to them, meets the increase in requests and limits the time it takes to render results: antigenic tests.
What are the indications, limits of use and precautions to be taken with these new tests? Decryption.
Unlike target amplification genomic research (RT-PCR) techniques, antigen tests look for one or more proteins of the virus through specific antibodies without amplification. The main advantages are a rendering time of theoretically around 30 minutes (unlike several hours for conventional PCR) without the need for very sophisticated equipment. The main drawback that we will develop is a lower sensitivity than RT-PCR.
An insensitive but specific test
What does that mean ? Quite simply, a positive test will allow the diagnosis to be made with certainty … but on the contrary, a negative result will not be conclusive and should be interpreted with all the necessary caution.
Indeed, studies by the CNR and the APHP Henri Mondor Hospital show a maximum sensitivity (ability to detect patients) of 60%, that is to say that at best, when the test is performed in the first 4 days of symptoms, it will detect 6 out of 10 sick patients. In other words, 4 patients with COVID will be tested negative even though they are carriers of the virus.
This sensitivity drops to less than 20% in asymptomatic people: 8 out of 10 patients with COVID will test negative. It is therefore understandable that the HAS recommendations, still currently, are to reserve these tests for situations where it is difficult to obtain an RT-PCR test (reference test because it is very sensitive and very specific).
This lack of sensitivity poses a real ethical problem: performing a less reliable test when there is a more efficient test is a real loss of opportunity for the patient. This is why the HAS recommendations are clear: these tests should only be performed in symptomatic patients if RT-PCR cannot be obtained within 48 hours. However, our group has been reporting the results for several months in 24 hours in 90% of cases and in less than 48 hours for all files (even non-priority ones).
Announced and uncontrolled performance
Another singularity and peculiarity that is difficult to understand. The sensitivity and specificity specifications which are used to authorise the tests (the HAS requires a sensitivity greater than 80% and a specificity> 99%) are provided by the suppliers themselves, without control of the authorities, nor even the validation of method file which made it possible to obtain these figures. Therefore, the enormous difference observed between the announced sensitivity and that which is measured in real life in the field can be explained. In fact, in the Henri Mondor study, 6 tests announced as compliant do not exceed 60% sensitivity and are therefore in reality non-compliant.
The high proportion of false negative tests could have significant consequences for the spread of the epidemic. In fact, a patient infected with the virus and who is asymptomatic will in most cases have a negative antigenic test. Reassured by this result, the patient will relax on barrier gestures, for example, not wearing a mask during family reunions. The phenomenon is likely to be drastic because massive antigenic test screening campaigns are scheduled in time for the end of year celebrations, for example in the Auvergne Rhône-Alpes region which aims to test 8 million people. How many people infected with the virus will be falsely tested negative by these insensitive tests? No doubt tens of thousands who will no longer protect themselves and no longer protect others, especially their families during the holidays by being falsely reassured.
These tests are carried out entirely manually and are therefore not very compatible with large volumes and high speeds. It will be difficult to carry out mass screening because the human resources needs are colossal. In addition, with manual processing there always remains greater human error, reversal of the result etc. For the past 20 years, laboratories have eliminated all manual handling as it was responsible for 90% of result errors. Automation and computer connections have eliminated manual results entry and tube reversals, making results extremely reliable and allowing full traceability. How will the recalls of defective batches be managed on these manual tests? How will the results be reported on the central SI-DEP platform, which enables precise monitoring of the epidemic? Automated tests are under development, but they will require automated labs. We will therefore lose the advantage of the speed of rendering results.
A specificity to watch out for!
We have noticed specificity problems when we have to use swabs other than those supplied by the manufacturer with clearly false positives (negative samples in PCR). This can be a concern if there is a shortage of collection material in favor of another that has not been assessed.
Why is the government not following scientific recommendations?
All scientific studies have the same conclusions: antigenic tests significantly underestimate the number of sick patients. The HAS, the French Society of Microbiology (retrospective study on 3,000 samples), the Henri Mondor hospital (AP-HP), … all warn against the misuse of these tests. Since the decree of November 17, the conditions for carrying out these tests have been further relaxed, in complete contradiction with scientific recommendations: now, all symptoms can be detected, even frail people (over 65 or with comorbidity) .
The massive screening campaign to control the epidemic must be based on the most sensitive tests possible: the tests used must be able to detect healthy, asymptomatic carriers who can transmit the disease without knowing it. However, antigenic tests do not have the sensitivity necessary to perform this task correctly. Are they therefore unnecessary? No. They have a real purpose, as the HAS recommendations indicate, when it is difficult to have a RT-PCR test quickly. We can also imagine the value of these rapid tests in ad hoc situations where the speed of rendering exceeds the lack of sensitivity. For example, in a patient who must integrate an EHPAD and who has a fever: an antigen test will immediately let us know whether the patient should be placed in isolation. The same goes for a patient admitted to the emergency room: the rapid response will allow the patient to be referred more effectively.
Conversely, their misuse leads to falsely reassuring patients carrying the virus who may no longer respect barrier gestures as they believe they are negative.
We therefore ask that the results of antigenic tests be clearly accompanied by an interpretation which takes into account their low sensitivity: any negative result should not be rendered “negative” because it may mislead the patient.