No room for nuance. It’s positive or negative. The result of a Covid-19 screening test by RT-PCR is binary, without any indication of the amount of virus present – the viral load – and without prioritization.
For Sylvie Van der Werf, head of the national reference center for respiratory viruses at the Institut Pasteur, this has something to do with “Problematic” : “A positive result close to the detection threshold, therefore with very little virus, is returned with the same weight as another with a Ct of 12 or 15, synonymous with certain positivity and potentially very high contagiousness. ”
Which is fraught with consequences if one uses RT-PCR to assess the contagiousness of individuals: an analysis published by the New-York Times on August 29 estimates that, on sets of cases tested positive – and therefore placed in isolation – this summer on the east coast of the United States, 85% to 90% were not contagious ! While these figures cannot be generalized, they illustrate a major pitfall in the use of diagnostic RT-PCR tests as contagiousness tests.
The “Ct”, or “Cycle Threshold” is the number of amplification cycles necessary to reach a fluorescence threshold value, which makes it possible to declare that the sample is positive for Sars-CoV-2. This is the principle of PCR: duplicating viral genetic sequences contained in a sample taken from a patient during successive amplification cycles, until they can be detected using fluorescent markers.
If there is a lot of virus in the original sample, it will only take a small number of cycles to reach the fluorescence threshold: the Ct will be small. Conversely, a low viral load will require a large number of cycles: the Ct will be high. As practiced today, RT-PCR tests consider as positive any sample with a Ct up to the maximum specified by the supplier of the machine and the reagent kit, which is often more than 40.
In other words, even a very low viral load gives a positive result. This high sensitivity is welcome for a diagnosis, but it gives erroneous information to identify a contagious person.
Lower the Ct? Yes but how much?
Faced with the too high sensitivity of RT-PCR tests for a contagiousness test, some call for reducing the threshold beyond which a patient is considered positive. This is the case of researchers interviewed for the New York Times article.
This is also the case of Bernard La Scola, professor of microbiology at the Institut hospitalo-universitaire (IHU) Méditerranée infection in Marseille, in favor of a revision of the rules of interpretation before reporting the test result: “I would be of the opinion to define a Ct beyond which the patient is no longer considered positive, he adds. The result must be useful and help to make relevant decisions. “
How much should the threshold Ct be lowered beyond which a patient is no longer considered positive? Or more accurately, considered to be at low risk of transmitting the virus. Some research carried out since the start of the pandemic offers some initial leads.
Comparison with a viral culture
Several studies have investigated the correlation between the Ct obtained in a sample positive for Sars-CoV-2 and the possibility of culturing the virus present in the sample on cells in vitro – a necessary but not sufficient condition for the sample to be infectious.
An article pre-published on the site medRxiv last July summarized it. “No viral culture was obtained from samples […] with Ct values less than 24 or 34, the researchers say. The possibility of cultivating virus decreases as Ct increases. “
Considered in this summary, a study carried out at the IHU Méditerranée infection in Marseille was published on April 27 in the journal European Journal of Clinical Microbiology & Infectious Diseases. First author, Bernard La Scola explains that out of the 183 samples analyzed, it is no longer possible to cultivate the virus in vitro when the Ct exceeds 34. “We have continued since and today have many more samples which confirm these results”, he assures.
More recently, similar findings were reported by researchers at the UK Public Health Agency in an August 13 article in Eurosurveillance : “The probability of cultivating virus drops to 8% in samples for which the Ct is greater than 35.”
Ct not very well tuned
However, one difficulty prevents setting in stone a value of Ct beyond which a patient could be considered as non-contagious: “Cts are not absolute values”, underlines Mr La Scola.
“For the same sample, they will be different from one laboratory to another and cannot be compared systematically”, adds Laurence Prots, director of the Cerballiance Alpes Maritimes Microbiology Functional Unit, for whom wanting to draw conclusions concerning contagiousness from Ct is a “False good idea”.
Indeed, the values of Ct depend on many parameters which go from the quality of the sample to the extraction method to purify the RNA present in the sample taken, through the target gene considered, the reagents and the PCR machine. …
Symptoms, timeline and medical biology
In addition, a major pitfall of this type of reasoning comes from the timing of the test for infection. A patient tested too early – before symptoms appear – will have a low viral load and a high Ct.
Interpreting his test as negative would be a major mistake since this same patient could have a much higher viral load and be very infectious 48 hours later, when symptoms appear.
“The contagiousness of a patient must be assessed from clinical information, his history, and the results of medical biology”, insists Mrs. Prots.
Quantitative RT-PCR to get everyone in agreement?
The variability of Ct and the difficulty of comparing the results from one laboratory to another are also illustrated in the summary of the scientific literature cited above: the Cts beyond which it becomes difficult to cultivate the virus on cells. in vitro vary from 24 to 34 …
The scientists consulted still wonder about the value of 24, which seems particularly low. “In our laboratory, with a Ct of 25 we manage to cultivate the virus in 70% of cases, remarks M. La Scola. Obviously, the patients are contagious. “
To remove the uncertainties and compare the results from one laboratory to another, a solution could come from the use not of Ct, but of the sample’s viral load – expressed in number of copies of the genome per milliliter – obtained through quantitative RT-PCR analysis. What, perhaps, to make everyone agree.