Diagnosis by genes: truth and fiction
Genetic testing
The popularity of genetic testing in Russia and abroad creates a persistent myth in society about the possibilities of predicting the appearance of cardiac, oncological and other diseases. Doctor of Biological Sciences Yuri Aulchenko in a lecture on "Gazeta.Ru" spoke out against the propaganda of the omnipotence of genetic tests and told which of them and how much one can believe, and which ones can be considered outright commerce.
Yuri S. Aulchenko is a recognized expert in the field of statistical analysis of biological data, in particular, the analysis of polymorphic genomes in natural and artificial populations.
Over the past 15 years, he has worked at leading research centers in Europe, such as Erasmus MC Rotterdam, Wellcome Trust Centre for Human Genetics, Helmholtz Zentrum Munchen.
At the same time, he did not stop cooperating with his scientific alma mater – the Novosibirsk Scientific Center (Novosibirsk State University, Institute of Cytology and Genetics SB RAS).
Since 2011, he has been working as a senior researcher at the Institute of Cytology and Genetics SB RAS and Director of the consulting company "Yurii Aulchenko" consulting in Rotterdam (Netherlands). In addition, he is an honorary professor at the University of Edinburgh (UK).
Yuri Aulchenko is the author of more than 160 scientific papers published in international peer-reviewed journals, including Nature, Nature Genetics, the New England Journal of Medicine, Bioinformatics, the Journal of the American Medical Association, PLoS Genetics. These works have been cited more than 6,500 times. Under the guidance of Dr. Aulchenko, 6 PhD and 9 MSci projects were defended.
Currently, the team under the leadership of Dr. Aulchenko is developing methods for statistical analysis of molecular biological data in the context of the biology of common diseases. The research team's area of interest includes diseases such as type II diabetes, Alzheimer's disease, cardiovascular diseases, osteoporosis, hypertension, stroke, breast cancer, multiple sclerosis and so on.
Genetic tests can be divided into two main groups: tests to determine the presence of mutations for so-called Mendelian diseases and for complex (complex) diseases. Mendelian diseases, in which the presence of a mutation in most cases leads to the occurrence of the disease, are simpler to develop tests and the reliability of these tests is very high. In the case of complex diseases, the situation is much more complicated. The risk of such diseases can be increased both by the presence of mutations in certain genes and as a result of environmental factors; the presence of a separate mutation in a gene or the presence of an environmental risk factor does not always lead to the disease. Most common diseases – such as breast cancer, hypertension, coronary heart disease, stroke, osteoporosis, and so on – are complex.
For complex diseases, the first step of scientific verification of the test is to make sure that changes in the gene entail the appearance of a specific disease. At this step, it is important to check the statistical significance of such a relationship, because for many types of diseases there are not yet enough significant studies to reliably show the relationship between the presence of a mutation and the risk of the disease on their basis. Many tests for common diseases in the medical services market have neither an evidence-based scientific basis nor a convincing statistical sample. All this speaks, unfortunately, about the very low quality of such services, which today are positioned as advanced achievements of medical genetic science, as modern personalized medicine.
In addition, for any clinical tests in which the dependence of the risk of developing the disease on specific changes in the genome has already been confirmed, there is a very important concept characterizing the quality of classification, which is measured by the "area under the curve". It is based, in particular, on the volume of collected statistics of research results. The area under the curve is the probability of distinguishing a sick person (or one who will get sick) based on the test from a healthy one.
If this indicator is 0.5, then the probability of error is 50%. With the same success, you can just throw a coin and save a lot of money. Higher indicators, from 0.8 and higher, paradoxically, are also not yet a reason for testing, at least within the framework of insurance medicine. First, it is necessary to understand whether the desired disease is curable. For example, there is a fairly reliable test that determines the risk of blindness in old age. But there is no preventive treatment for this disease yet, medicine does not even know how to postpone the date of its onset. Consequently, there is no meaningfulness of this test, except for that group of patients who are ready to accept and use the information that they will soon become irreversibly blind. Last but not least, the meaning of testing depends on the availability and cost of preventive therapy and on the source of cost coverage. It's one thing when they are covered by the state or voluntary health insurance, and quite another when you pay out of your own pocket.
An example in defense of genetic testing can be a very expensive test in Europe, which determines with 90% certainty the risk of familial (Mendelian) forms of breast cancer. At high risk, the patient has the opportunity to remove the breast before the onset of the disease or take other preventive measures (for example, more frequent mammographic examination). In a number of European countries, the cost of this test, worth a thousand euros, is covered by insurance medicine. Of course, this test, like any genetic examinations, is not prescribed en masse, but with a hereditary tendency, which is determined by identifying cases of this disease among the patient's relatives.
To date, it is easier to list examples of available reliable tests than to compile an extensive list of questionable commercial services. Most of the good tests are for rare diseases. We have reliably learned to predict familial forms of Alzheimer's disease, early familial forms of cancer and cardiovascular diseases. Since such forms are rare, carrying out genetic tests does not make sense if there have been no cases of early onset of this disease in the family.
The detection of non-hereditary (sporadic) forms of cancer in recent years has shifted slightly towards greater reliability, but the probability of error in their predictions is still high, so scientists evaluate their effectiveness with a cautious formulation: "the test has potential." This means that the statistics are still small, but, years later, their reliability may significantly increase. Genetic tests are not yet able to predict non-hereditary (acquired) heart attacks and strokes.
Genetic tests can be effective for selecting the type of drugs or their dosage. For example, 4 drugs have been approved for the treatment of lung cancer in the United States, which are prescribed after genetic tests. Another example is the treatment of insulin-dependent diabetes of the first type. According to statistics, about 5% of patients with this diagnosis can be treated with pills – they do not need insulin injections. With high confidence, these patients can be identified using genetic testing or using glycosylation biomarkers. There are effective genetic tests to determine the personal dose of warfarin, a blood–thinning drug that patients constantly take after cardiovascular surgery to prevent the formation of blood clots. The required dosage in different patients sometimes differs significantly. In this case, too much blood thinning can cause severe internal bleeding. For some HIV-infected patients, one of the drugs that prevent the development of AIDS is deadly – a genetic test allows you to identify this group of patients.
Personalized medicine and genetic testing really play a big role in diagnosis and treatment. But as a genetic scientist, I am afraid that scientific results, many of which are not confirmed by clinical practice and statistics, put on commercial rails, introducing into the market of medical services what has not proven its effectiveness. On the one hand, as a rule, this is just a prognostic test that can cause direct harm only to the patient's wallet, but not to his health. However, the situation poses a serious threat to the future of medical genetics.
If most of the genetic tests that patients pay for today do not confirm their effectiveness, it may someday lead to the complete rejection of this method by medical institutions. Instead of developing this direction, it will be reduced. And, first of all, those to whom this type of diagnosis could really help will suffer from this. To sort out this chaos of truth, half-truth and fiction, I divided all genetic tests into three categories.
The first is the tests confirmed scientifically and repeatedly tested practically. In addition, these tests make economic sense – for example, on the basis of them, preventive treatment can be prescribed, which is cheaper than the treatment of the disease.
The second group consists of tests that have a good scientific basis, but are little tested practically or, perhaps, do not make sense within the framework of insurance medicine (for example, there is no medicine and no way to slow down the onset of the disease or treatment is cheaper than diagnosis).
And the third is absolutely unsubstantiated. For ease of understanding, let's call the first group "surgery", that is, the field of medicine where the doctor is fully responsible for the diagnosis: either you remove an inflamed appendix, or you get peritonitis and death. The third group clearly claims the title of "urine therapy", which, on its own initiative (and under its own responsibility), treats a certain group of patients who firmly believe in its effectiveness. In the second one there are more or less working tests, and there are still very poorly tested ones. This category is more vague and requires a very legible approach. But who should figure it out?
It is possible to organize a clear division only at the level of the entire healthcare system by including the first group of tests in services paid for by compulsory medical insurance, the second – in voluntary insurance (VMI), when the insurance company, based on scientific and clinical data, and assessing the risks, decides for itself what it is willing to pay for. The third group of tests can be left on the condition that the patient will pay for them independently. In Europe, this system is designed exactly like this and balances well on the real possibility of a conflict between the state, doctors, patients and insurance companies. Doctors need payment for services, and the state and the insurance company need a healthy citizen/client and cost savings. If the diagnosis is not detected in time, expensive treatment will follow. To make a diagnosis that does not exist – the cost of unnecessary treatment. Therefore, if the test is included in the list of compulsory insurance, in Europe you can be sure of its effectiveness. When it is included in the list of VMI services, a private insurance company takes responsibility for its quality, which is unprofitable to pay for quackery, risking its reputation. The patient himself is responsible for the rest of the tests – he is free to spend his money as he pleases.
While there is no such system in Russia, the patient risks buying a "urine therapy" service on the doctor's prescription. After all, genetic examinations are not yet included in compulsory insurance. It is doubly disappointing that for the organization of such a system in our country there are quite enough qualified specialists in medical genetics who are able to separate the wheat from the chaff. I really want to believe that their knowledge will still serve the domestic healthcare in the foreseeable future.
Portal "Eternal youth" http://vechnayamolodost.ru19.09.2012