Coronavirus tests
PCR: understand, forgive
Tens of millions of COVID-19 tests have already been conducted in Russia, but still not everyone understands the difference between them and why their results may contradict each other. Let's figure it out together.
Coronavirus: the beginning
Many people wonder why so many analysis systems are being created for the coronavirus and why the results are so often false. Low-quality tests, a conspiracy of manufacturers and governments hiding statistics should not be blamed for everything. Many of the answers lie in the peculiarities of the different stages of the course of the disease.
These graphs show changes in the concentration of coronavirus particles and the antibodies produced in response to them. The analysis with the search for viral particles will show whether the patient is infected at the moment, and the antibody test helps to determine not only the infected person (after some time), but also the person who has already been ill (Siouxsie Wiles& Toby Morris/Wikipedia).
Symptoms begin to appear about a week after infection, when the number of viral particles increases as much as possible. During this period, a person is most contagious. A week after the onset of symptoms, the concentration of IgM subtype antibodies reaches a peak, which begin to be produced in response to the rapid multiplication of viral particles.
IgG antibodies begin to be produced approximately simultaneously with IgM, but their concentration grows more slowly. The maximum values can be seen about two weeks after the onset of symptoms. However, if IgM antibodies, which worked as a rapid response unit, are practically not detected three weeks after the first signs of the disease, the level of IgG, even if they later entered the border service, persists for a relatively long time after recovery.
Virus Search
Viral particles can be detected using the PCR method, or polymerase chain reaction. SARS-CoV-2 refers to RNA viruses, which means that it stores its genetic information - instructions for assembling the necessary proteins for its framework—the capsid - in an RNA molecule, and not in DNA, like all normal people (and most living organisms). The virus itself cannot copy its RNA and collect proteins, so it attacks the cell and exploits its production facilities. But in order to make a cell make proteins to order, you first need to "translate" your genetic instructions into a suitable "language", that is, into DNA. This process is called reverse transcription, because with conventional transcription, which is used much more often, everything happens the other way around: for the production of protein according to DNA "patterns", RNA is synthesized, based on the code of which the protein will be assembled. In the analysis, reverse transcription also has to be done, therefore, a type of such test used, including for coronavirus, is called RT-PCR (that is, PCR with reverse transcription).
In general, PCR is a standard laboratory method, very accurate and widely used. When developing a specific test, researchers study the genome of the virus and select fragments by which it can be recognized. But in the samples, the concentration of pieces of the genome will be quite low, and in order to detect them, many copies are made (amplification). For this purpose, a mechanism is used by which copies of DNA and living organisms are propagated. Reverse transcription in the analysis for COVID-19 should be done, because it is easier to create multiple copies of DNA this way.
The RT-PCR scheme: a seed (primer, indicated in yellow) is placed on the "tail" of the RNA, consisting of adenines. The seed is needed to connect other "letters" of the genetic code. Based on RNA, one DNA helix is synthesized, where each nucleotide is matched with its pair. Then the second one is completed to the chain (painted in blue in the figure). Then this DNA section is copied repeatedly (Jpark623/Wikimedia Commons).
For analysis for coronavirus, kits for so-called real—time RT-PCR are often used - a faster and automated version that allows you to determine the number of new copies after each cycle. PCR tests for coronavirus last on average from 40 minutes to 2-3 hours. As a site for recognition, they use gene N, gene E and other fragments of the viral genome (usually several sites to reduce the likelihood of errors).
False and false-
For PCR, different biological material can be used, which is relatively easy to collect: a smear from the nasopharynx, throat, saliva or sputum. It is believed that it is easiest to collect saliva: a person can spit into a test tube without the participation of doctors and laboratory technicians, which makes their work safer. For the patients themselves, including people in quarantine, this approach may be more convenient. However, scientists have doubts that the concentration of coronavirus particles in saliva will be sufficient for analysis (although there is evidence in favor of the fact that tests with saliva are even more accurate). But during the course of the disease, their concentration in the nasopharynx and throat also changes. After the first week, the virus mainly passes into the lungs and will no longer multiply so intensively where samples are usually taken from.
Because of this, the analysis taken in the first days of the appearance of symptoms, the accuracy of such a test will be higher than 94%, since the number of viral particles, as we remember from the graph, will be high (10 5-10 6 per milliliter), and even not the most highly sensitive systems for analysis will be able to detect them. Moreover, even with an asymptomatic course of the disease, there will be quite a lot of viral particles to detect. But by the tenth day, the accuracy drops sharply: the probability of a false negative diagnosis reaches 33%! It is not surprising that patients are checked three or four times before being released from quarantine.
In addition to real-time PCR, there are more urgent options, for example, based on loop isothermal amplification (LAMP). As you can understand from the name, it is also based on the reproduction of copies of the necessary fragments. But if in the classical versions of PCR a strictly defined temperature is needed for each stage of this copying (and the whole process takes place in a special device — a thermal cycler), with express tests everything takes place at the same temperature, which reduces the time.
There are also concepts such as sensitivity and specificity of tests. The first means that the analysis clearly shows the presence of the virus. So, a sensitivity of 90% will mean that 10% of infections will not be detected (a false negative result). Specificity implies that the test is selective and will not confuse the desired fragment of the virus with anything else. These cases will become false positive results. Low-specific tests will have low positive prognostic significance, especially if there are few patients in the population. So, if there are only 5% of cases, then testing 100 people with an error of 5%, we will get 5 really sick and 5 false positive results, and only half of our 10 diagnoses will be correct.
Based on this knowledge, it can be said that for people without symptoms, it would be better to use kits with high sensitivity, whereas in the presence of symptoms, the number of viral particles is usually higher, and even less sensitive methods can detect them.
Real-time PCR is considered the most sensitive. However, his almost flawless results in laboratory conditions, alas, do not agree with the accuracy in real clinics, which drops to 88 or even 66%.
What about us?
In Russia, back in April, eight variants of PCR tests became available, which passed accelerated registration. This changed the situation with the monopoly of the Vector Research Center, which initially supplied kits with low sensitivity (105 molecules per milliliter), requiring additional reagents for reverse transcription. The new kits have become more sensitive (from 500 to 10,000 molecules per milliliter in the sample), they have an internal control sample (a molecule that helps to detect that the reaction went wrong when the laboratory assistant made an error). The leaders of the market with the largest production (except for the SSC "Vector") were the Central Research Institute of Epidemiology of Rospotrebnadzor, "Vector-Best" (not to be confused with the SSC "Vector") and "DNA Technologies". The test from the Strategic Planning Center of the Ministry of Health also has a very high declared accuracy.
Private companies "Hemotest" and "Helix" use analysis from "Vector". Invitro makes more accurate tests with Vector-Best, but it's impossible to get to them for analysis just like that: they only accept samples in containers from state clinics. LabQuest offers tests of "DNA technology", "Litech" and an express test from "Generium", and the test of the Central Research Institute of Epidemiology is done in the CMD company, which goes to the house with a courier. Laboratories approved by Rospotrebnadzor can be found here.
The problem with Russian tests is that we do not have independent data on their comparative verification. So we have to trust only their stated sensitivity, keeping in mind that in general there is a risk of false negative results due to the imperfection of the methods available to us that best detect coronavirus in the first days of symptoms. And since it is not so easy and fast to take an analysis at will to a polyclinic, and the services of private companies are not cheap, then you can understand why early testing is ideal, but not always achievable. The option of obtaining biomaterial directly from the lungs cannot be called convenient and easily scalable (although sputum is used in some tests, which is more accurate than saliva).
Fortunately, as we have already mentioned, there are other tests and diagnostic methods that will help detect not viral particles, but various manifestations of the disease. For this, for example, antibody tests and computed tomography are done, which we will talk about in the second part of the article.
To be continued.
Portal "Eternal youth" http://vechnayamolodost.ru