An odd correlation between cigarette smokers and COVID-19

Charles Corfield, President & CEO, nVoqHealthcare Leave a Comment

In our fourth installment of Charles Corfield’s pandemic blog, he discusses an odd correlation between cigarette smokers and COVID-19. Please check back every Monday as Charles shares his thoughts on what is trending with COVID-19.

About the Author

Charles Corfield, President & CEO, nVoq

On the research front there have been several studies reporting results on anti-viral therapies. In sum, they were “meh”. A study of less than 1,000 patients comparing hydroxychloroquine to supportive treatment found no difference. The authors note that their hospital has now removed hydroxychloroquine from its list of things to try. A few more anti-virals – Remdesivir, Lopinavir, Ritonavir, and Duranavir – also seem to be duds. I note that in cell cultures Remdesivir has been consistently effective at blocking/reducing infection, and it makes me wonder if the unimpressive clinical studies may be another clue about the virus. Suppose that Remdesivir has actually worked just fine in patients in that it eliminated the virus, but their clinical issue (by that point) was not the virus, but their reaction to it (the “cytokine storm” you have read/heard about). In other words, by the time patients are sick enough to go to hospital and get Remdesivir, the inflammatory train has already left the station.

Which brings me to an illuminating study from 2004, but first a couple of observations that researchers have made about COVID-19 and smoking. A Chinese team which examined hospitalizations at a number of sites in China was struck by the under-representation of smokers. Patients with a history of smoking were in the single digits percent versus 25% of population which smokes (roughly half the men and a few of the women). Another study in France noted much the same: at hospital in Paris, people who smoked every day, were underrepresented. This is counter-intuitive and not surprisingly other researchers have looked at published studies of the outcomes for hospitalized patients, and found that smokers do worse, which conforms to intuition.

By the way, the differing results are not necessarily in conflict; rather, they reflect something called “Bayesian statistics” where the outcome is conditioned on selection criteria. In this case, it may well be true that smokers in the general population are less likely to develop symptoms which will send them to hospital. If they do end up in hospital, they do much worse than non-smokers. If smoking has a protective effect, it makes you wonder if nicotine plays a role (albeit nicotine is one of many compounds in cigarette smoke). I searched for papers with nicotine and terms consistent with inflammation and turned up a paper from 2004 where the researchers induced peritonitis in mice and then gave some of them nicotine. 80% of the untreated mice died of the infection compared to 50% of the mice that got nicotine. In case that makes you wonder if nicotine is an antibiotic, it isn’t. Rather, the nicotine latches onto (acetylcholine) receptors which play a role in the regulation of (some) inflammatory responses. In this study the activation of the receptors damps down one (or more) of the inflammatory cascades.

The paper demonstrated that in mice with a severe infection a good fraction of the mortality was due to the inflammatory response and not the infection. If the inflammatory cascade in COVID-19 is like that in those mice, then, it is not surprising that nicotine has a protective effect. This would be motivation to explore whether nicotine has a useful role in managing COVID-19 outcomes and it would be easy to test -- think nicotine patches, rather than wards filled with cigarette smoke. Although the gold standard for studies is “large enough number of subjects, randomized, double-blinded, placebo-controlled”. I don’t think you can blind the patients as to whether they are getting a real or fake nicotine patch. Therefore, we would want to see a strong positive result in nicotine’s favor, sufficient to overcome the concern that any positive result was simply due to the placebo effect. On the historical front, genetic material in organisms tends to drift with time and the number of mutations (differences) from the original progenitor increases. Researchers can use this as a rough and ready clock to estimate how long organisms have been around. Four coronaviruses account for 20-30% of common colds. They began to infect humans at different times within the last eight centuries. E.g. HCoV-NL63 jumped to humans sometime between the 13th and 15th centuries, while HCoV-OC43 made the leap around 1890. You might be interested to learn that 1889/1890 was a pandemic year.

Leave a Reply