May 20, 2020
0 0

Evil critic: COVID’s cigarette smoke

“There are almost no smokers among patients with coronavirus infection, so smoking protects against coronavirus!” - Such messages are full of media and social networks. Is this really so, let's understand.

People with bad habits often try to find some rational justification for them. For example, I do not overeat, but create a “safety cushion” for a heart attack. Well, yes, according to some studies, the dependence of survival in this acute condition on BMI has a rather bizarre form: for BMI less .... , 5 the risk of death is maximum, then it gradually decreases, and with BMI = 22 begins to grow again, even if standardized by age, gender and diabetes. And alcohol, as you know, it is not entertainment for the sake of health, but for health, because drinkers will live moderately nemeryanno. When coffee was still a carcinogen, one could talk about its incredible effect on memory. The benefit of research in Pabmed is more than enough, you can find confirmation of any point of view, there would be a desire and a bit of free time.

The situation with smoking is about the same, for example , at one time the argument "Yes, I smoke, but this is to protect against Parkinson" was very popular. Interestingly, this is almost true. When the pandemic started and how preprints, preprufs, and studies on small samples fell from the cornucopia, sofa analysts were immediately excited by the opened prospects - you could pull an owl even on the globe of the Samara region (“Saratov!” Immanuel Vitorgan’s voice immediately sounded in my head. “We talked about the Saratov region ”). And, of course, findings about the positive role of smoking in the history of COVID were immediately swayed by famous media personalities.

In March, the information voiced by Alexander became a cause for enthusiasm Nevzorov, they say, proved that only 1.4% of smokers are among COVID cases in China. Even with a cursory acquaintance with the original source, it turned out that I did not win “Volga”, but 3 rubles, but I lost, not in the lottery, but in cards — that is, everything is as usual. But the resonance was impressive, he even circled the globe a couple of times, like a shock wave from the explosion of a Soviet hydrogen bomb.

Where did this figure come from? From a publication in the journal Allergy . There really was talk about Chinese patients, however, all about of them hospitalized for coronavirus infection. Indeed, only 1.4% of them were current smokers. And another 5% of the former. But to draw some far-reaching conclusions on such a small sample, to put it mildly, is incorrect. What the authors themselves state:

“The relationship between smoking and coronavirus infection is not clear , and the exact reasons for the lower incidence of COVID - 16 the current smokers are still unknown. "

But who cares about the opinion of the authors, if we have Nevzorov?

But that's not all. If we take a closer look at the data in Table 1 in this article, we will find out that among former smokers, a mild course occurred in 1.3% of cases, and a severe course in 5.2% of cases, among current smokers a mild course in 04, 8%, and heavy - in 10, 9%, that is, it absolutely does not mean that smoking makes it any easier fate with COVID - 16. In the next Chinese study, published in the much more prestigious of The New England Journal of Medicine , in a sample of 802 smoking patients have already been detected 05, 6%, but they did not raise him to the banners.

The next wave went after the publication of the preprint of the French doctors. In the media and social networks, of course, it was called a full-fledged study, although it was not even published in the medical preprint archive, but on a free review service, and the rating of this work was immediately set low, at the time of writing, 2, 79 of 5. It is clear that the media and smokers shouted with one voice, they say, now, now they are not some Chinese, now whole French have proved that coronavirus is bypassing smokers! But why didn’t colleagues share their optimism?

The fact is that research is teeming with spaces. One of the most notable - not all hospital patients are included in it, for example, those lying in intensive care units were ignored. That is, this could affect the overall figures, and, in particular, the assessment of the severity of COVID in smokers and non-smokers. Well, the little things came up, for example, the majority of patients in this particular hospital are doctors, and they generally smoke less than the rest of the population of France, plus a positive answer to the question of smoking is becoming less and less socially acceptable. In general, questions remain and unambiguous conclusions about the protective effect of smoking cannot be drawn from this study.

Another bucket of oil was added to the fire by the results of a preprint of British researchers under the control the famous Ben Goldacre (Ben Michael Goldacre). In the framework of the OpenSAFELY project, they analyzed the data 87695 of adult residents of England, today it is the largest cohort study on risk factors increased mortality from COVID.

Fragment of the final "Christmas tree" chart:

Фрагмент итогового «ёлочного» графика

The person who first sent the link to this preprint wrote:

“Now is the sample sufficient?” And the result is unequivocal: former smokers risk significantly more than current ones. Moreover, the current effect is clearly visible protective effect. You will not argue with Goldacre, he seems to be from yours ”,

sarcastically completed his appeal smoking and not about to quit reader

(he introduced himself like that ).

I do not understand why you can not argue with Goldacre. Well, yes, he is “of ours”, if we talk about sane doctors, adherents of evidence-based medicine, but it’s not just pleasant to discuss with such colleagues, but also productive. I had to communicate with Goldacre for a long time, years 04 ago, but then he made a very favorable impression, including in the sense of perception of criticism. I hope he has not lost these qualities over the past time, like some more-to-us-non-colleagues.

But back to work. What do the authors write about smoking?

“Both current and former smoking were related with a higher risk in models adjusted only by age and gender, but in the fully adjusted model, there was weak evidence of a slightly lower risk for current smokers (fully adjusted risk ratio OP 0, 83, CI 0, 40 - 0 , 93)) . In a subsequent analysis, we added separate covariates to the model with age, gender, and smoking to investigate this further: the change in RR was apparently largely due to adaptation to chronic respiratory diseases (RR 0, 86, 0, 43 -1,04 after adjustment) and social deprivation (OP 0, 93, 0, 86-1,04 after adjustment). Other individual adjustments did not eliminate the positive relationship between current smoking and outcome. ”

AND here I will not argue with Goldacre myself, but I will give the floor to another iconic character from "ours", the editor of the medical portal Medscape Perry Wilson. His tongue is suspended on the right side, so here's a comparison for you according to Wilson:

“This all the same, what to say: if all the people who were not wearing seat belts would not have flown out through the windshield, they would have been more likely to survive the accident. ”

The fact is that in the analysis of the British n / a Goldacre and arterial hypertension has a protective effect. Interestingly, it is exactly the same in absolute value as current smoking. Does this mean that high blood pressure prevents death from COVID’s? And what does everyone need to get hypertension immediately? And for people with high blood pressure, relax and get out of xkcd dusk self-isolation? Of course not.

Everything is the same with smoking. Because for them, the conclusion is as follows:

“If all these smokers there was no respiratory disease, they would actually have a better chance of surviving with COVID - 18 than non-smokers. "

If only yes. Who would argue, of course. However, Goldacre associates do not claim to establish cause and effect relationships, so there is simply nothing to argue with. Well, the media and social networks ... Better mournfully silent, yeah. And remember a fragment of one excellent comic strip on finding patterns in large data sets:


Are there meta-analyzes and systematic reviews on this topic? Has already. The first systematic review was published in March; it included only five cohort studies - one prospective and four retrospective studies with a total sample of 1549 patients. And the results can hardly be called encouraging or supporting the hypothesis of “protective smoking”: smokers have 1.4 times more chances (OR = 1.4, 93% CI: 0, 88 - 2, ) get sick COVID - 16 in severe form and about 2.4 times more likely to be hospitalized in ICU, smokers are also more likely to need mechanical ventilation or die compared to non-smokers (OP = 2.4, 88% CI: 1, 22-4,02).)

The first meta-analysis also appeared in March, but the authors managed to scrape together only five more or less qualitative studies with 1517 by the participants. The final Christmas tree rhombus did not move from the line of lack of differences, so it was not possible to show that smoking positively or negatively affects the severity of COVID. In May, a larger meta-analysis appeared, it included 16 peer-reviewed studies with a total number of participants 2020. And here the general rhombic on the "Christmas tree" has clearly shifted, and not in favor of smokers:

In an article on the website of the University of California, the scientists of which conducted this meta-analysis, one of the authors explained why such a low number of smokers among patients with COVID in some studies may be related:

“Low prevalence may actually be the result of an underestimation of smoking, especially considering the difficult conditions that often arise when caring for people re servation of health systems ".

What is it? About what colleagues working in the “red zones” around the world, including my son, are talking about: a large flow of patients is being processed, many are simply not able to say something intelligible because of pronounced shortness of breath, so the question is smoking sick or No interests physicians in the least. And often the information about this does not fall into the medical history, or they write, as in the film "DMB":

- We have not been able to find out what the brakes are.

- Write in the ascribed “Fedya”, his face is thick, he will do it. 17425445

That is, errors in the interpretation of the results obtained may depend, among other things, on defects in the collection of primary information for quite objective reasons. Moreover, there are problems even now, what can we say about the initial stage of the pandemic. And before the advent of SARS-CoV-2, this was a well-known, studied and well-documented problem, here is one of the publications on the topic, here is the second.

Another important aspect: not everyone is tested. There is, for example, a preprint from three Canadian researchers who completed the British Biobank and obtained all the positive COVID tests from there, trying to break them down by smoking status. After adjusting for covariates, it turned out that smokers are slightly more likely to get a “+” in the coronavirus test, but the difference with non-smokers is not significant. That is, again, the protective effect is not visible from this angle either.

And the last thing I would like to draw attention to: the physiological or biochemical base. I didn’t have to see a single sane explanation of how smoking could be protected except “Well, somehow everything is sterilized itself and the virus dies.” But on the opposite side of the explanations and analogies at least flood. For example, in another Canadian study, it was found that on the membranes of the epithelial cells of the airways of smokers and people suffering from COPD (which is clearly associated with smoking), an excess of the already known ACE2 protein is expressed through which coronavirus penetrates the cells. This is not a training alarm, sorry, not a mouse study, but a very human one, so its results are most likely to apply to all smokers, not just Canadian ones. What was confirmed by a group of American researchers who published similar data in the journal Cell . In addition, it is well known that smokers as a whole more often suffer from infections due to damage to the immune system by components of tobacco smoke, there are reviews on this subject, here is an example of a sample 2004, there are also works with a detailed description of various types of infections, including purely viral infections.

So in favor of the hypothesis "Smoking protects against COVID" in any of its variations of evidence is not scratched. The French want to run nicotine patches on the doctors as a possible means of protection, but it will be important to build the right research design, otherwise we will get another study about seat belts and a windshield. But about this set of circumstances, we seem to have figured out everything. More evidence to the god of evidence!

Article Categories:

Leave a Reply