PREPRINT: National Smoking Rates Correlate Inversely with COVID-19 Mortality
Recent studies show cigarette smokers are markedly under-represented among patients hospitalized for COVID-19 in over a dozen countries. It is unclear if this may be related to confounding factors such as age distribution, access to care, and inaccurate records. We hypothesized that these concerns could be avoided by studying smoking prevalence in relation to COVID-19 mortality. Since climate has been identified as a factor in COVID-19, we studied groups of countries with relatively comparable temperatures. Methods: The 20 hottest and 20 coldest countries in the Johns Hopkins Mortality Analysis database with a minimum mortality rate of .3 deaths/100,000 were selected on the basis of the average temperatures of their largest city. Mortality rates were determined as of May 1, 2020 and correlated with national smoking rate adjusting for sex ratio, obesity, temperature, and elderly population. Results: A highly significant inverse correlation between current daily smoking prevalence and COVID-19 mortality rate was noted for the group of hot countries (R=-.718, p = .0002), cold countries (R=-.567, p=.0046), and the combined group (R=-.324, p=.0207). However, after adjustments only the regression for hot countries and the combined group remained significant. In hot countries, for each percentage point increase in smoking rate mortality decreased by .147 per 100,000 population (95% CI .102- 192, p=.0066). This resulted in mortality rates several-fold elevated in the countries with the lowest smoking rates relative to the highest smoking rates. In the combined group, mortality decreased by .257 per 100,000 population (95% CI .175-.339, p=.0034). Discussion: These findings add support to the finding of an inverse relationship between current smoking and seriously symptomatic COVID-19. However, we conclude that the difference in mortality between the highest and lowest smoking countries appears too large to be due primarily to the effects of smoking per se. A potentially beneficial effect of smoking is surprising, but compatible with a number of hypothetical mechanisms which deserve exploration: 1) Studies show smoking alters ACE2 expression which may affect COVID-19 infection or its progression to serious lung pathology. 2) Nicotine has anti-inflammatory activity and also appears to alter ACE2 expression. 3) Nitric oxide in cigarette smoke is known to be effective in treating pulmonary hypertension and has shown in vitro antiviral effects including against SARS-CoV-2. 4) Smoking has complicated effects on the immune system involving both up and down regulation, any of which might alone or in concert antagonize progression of COVID-19. 5) Smokers are exposed to hot vapors which may stimulate immunity in the respiratory tract by various heat-related mechanisms (e.g. heat shock proteins). Studies of steam and sauna treatments have shown efficacy in other viral respiratory conditions. At this time there is no clear evidence that smoking is protective against COVID-19, so the established recommendations to avoid smoking should be emphasized. The interaction of smoking and COVID-19 will only be reliably determined by carefully designed prospective study, and there is reason to believe that there are unknown confounds that may be spuriously suggesting a protective effect of smoking. However, the magnitude of the apparent inverse association of COVID-19 and smoking and its myriad clinical implications suggest the importance of further investigation.