An inverse relationship between smoking and COVID-19

Researchers at the University of Washington and Stanford University School of Medicine have reported a significant inverse relationship between current smoking and coronavirus disease 2019 (COVID-19) mortality rates across forty countries.

However, David Haynor and the team say the difference in mortality rates between countries with the highest and lowest national smoking prevalence seems too large to be mainly accounted for by the effects of smoking. They suspect unidentified confounding factors could still be responsible for any perceived protective effect of smoking.

However, the magnitude of the association observed and the wide-ranging implications this could have highlighted the importance of the further investigation, say the researchers.

A pre-print version of the paper is available in medRxiv*, while the article undergoes peer review.

Novel Coronavirus SARS-CoV-2 Colorized scanning electron micrograph of a cell (green) infected with SARS-COV-2 virus particles (purple), isolated from a patient sample. Image captured at the NIAID Integrated Research Facility (IRF) in Fort Detrick, Maryland. Credit: NIAID
Novel Coronavirus SARS-CoV-2 Colorized scanning electron micrograph of a cell (green) infected with SARS-COV-2 virus particles (purple), isolated from a patient sample. Image captured at the NIAID Integrated Research Facility (IRF) in Fort Detrick, Maryland. Credit: NIAID

The association has been made before

According to Haynor and colleagues, recent studies have shown that smokers are significantly under-represented among hospitalized COVID-19 patients in China, France, Italy, Germany, the UK, the USA, Israel, Iran, South Korea, Kuwait, Mexico, Spain, and Switzerland.

“The apparent substantial under-representation of smokers among COVID-19 inpatients consistently across thirteen countries is remarkable,” says the team. “This is surprising as smoking is generally associated with greatly exacerbating respiratory infections.”

Suggested mechanisms that may confer a protective effect of smoking include altered host cell expression of angiotensin-converting enzyme 2 (ACE2, the receptor the virus uses to infects cells); the anti-inflammatory activity of nicotine; the antiviral effect of nitric oxide; the effects of smoking on the immune system and vapor heat-related stimulation of immunity in the respiratory tract.

Daily smoking prevalence correlated inversely with national COVID-19 mortality rates of the 20 hottest countries. Pearson’s correlation without adjustments:
Daily smoking prevalence correlated inversely with national COVID-19 mortality rates of the 20 hottest countries. Pearson’s correlation without adjustments: R=-.718, p=.0002.

It is not clear whether confounding factors are involved

However, the researchers say variations in the studies’ results make it unclear whether confounding factors may have been contributing to the effect. For example, reports generally did not adjust for age and comorbidity and records were not necessarily accurate regarding smoking status.

“We, therefore, sought to test the association in a way that was not subject to any of these confounds,” writes the team.

The researchers also included countries with relatively similar temperatures, given that climate has previously been identified as an important factor in studies of  COVID-19. Haynor and colleagues observed that across 19 countries with mortality rates that were 50% higher than among others, all but two fell within a relatively narrow temperature range.

Daily smoking prevalence correlated inversely with national COVID-19 mortality rates of the 20 coldest countries. Pearson’s correlation without adjustments:
Daily smoking prevalence correlated inversely with national COVID-19 mortality rates of the 20 coldest countries. Pearson’s correlation without adjustments: R=-.567 p=.0046

“We hypothesized that if there was a protective effect of smoking, it might be possible to detect it outside of this moderate temperature band where temperature appeared to be a dominant factor and mortality rates were extreme,” write Haynor and colleagues.

What did the team do?

Using the Johns Hopkins Mortality Analysis database, the team selected 20 “hot” and 20 “cold” countries that had a minimum mortality rate of .03 deaths per 100,000 population.

“A minimum mortality threshold was required because extremely low mortality rates may reflect inadequate testing — furthermore, this limits the impact of floor effects in the analysis,” explains the team.

The researchers examined the relationship between mortality rates and national smoking prevalence after adjustment for known risk factors associated with COVID-19 mortality and after adjustment for independent variables including gender ratio, obesity prevalence, age over 65 years, and average ambient temperature.

The same correlation was identified

A significant inverse correlation between current daily smoking and the COVID-19 mortality rate was identified for hot countries, cold countries, and the two groups of countries combined.  

However, after adjusting for multiple confounders, this association remained significant for hot countries and the combined countries, but not for the cold countries.

In hot countries, for each percentage point increase in the smoking rate, mortality decreased by 0.147 per 100,000 population. The mortality rate was several times higher in countries with the lowest smoking prevalence compared with those that had the most significant smoking prevalence.  

When hot and cold countries were combined, the mortality rate decreased by 0.257 per 100,000 population.

The difference in mortality rate is “too large” to be accounted for by smoking

The researchers say they think the difference in mortality rate between the countries with the lowest and highest smoking prevalence is too large to be accounted for primarily by the effects of smoking.

“The reason is that even if we assume every smoker is 100% protected from developing COVID-19, there are too few smokers in the population to produce such a large effect, and it is reasonable to assume that there is a confounding influence,” they explain.

The team points out that differences in smoking prevalence may be linked to variations in political structures, economics, or behavioral tendencies that impact the acquisition, testing, diagnosis, treatment, or reporting of COVID-19.

“At this time, there is no clear evidence that smoking is protective against COVID-19, so the established warnings to avoid smoking should be emphasized,” warn Haynor and colleagues.

“However, the magnitude of the apparent inverse association of COVID-19 and smoking and its myriad clinical implications suggest the importance of the further investigation,” they conclude.

*Important Notice

medRxiv publishes preliminary scientific reports that are not peer-reviewed and, therefore, should not be regarded as conclusive, guide clinical practice/health-related behavior, or treated as established information.

Journal reference:
Sally Robertson

Written by

Sally Robertson

Sally first developed an interest in medical communications when she took on the role of Journal Development Editor for BioMed Central (BMC), after having graduated with a degree in biomedical science from Greenwich University.


Please use one of the following formats to cite this article in your essay, paper or report:

  • APA

    Robertson, Sally. (2020, June 15). An inverse relationship between smoking and COVID-19. News-Medical. Retrieved on November 28, 2022 from

  • MLA

    Robertson, Sally. "An inverse relationship between smoking and COVID-19". News-Medical. 28 November 2022. <>.

  • Chicago

    Robertson, Sally. "An inverse relationship between smoking and COVID-19". News-Medical. (accessed November 28, 2022).

  • Harvard

    Robertson, Sally. 2020. An inverse relationship between smoking and COVID-19. News-Medical, viewed 28 November 2022,


  1. Tom Og Tom Og United States says:

    Countries that have a high percentage of smokers, AND who SEEM TO HAVE a lower rate of death from Covid-19, have that lower rate of death in their population, BECAUSE SO MANY OF THE SMOKERS ARE ALREADY DEAD.

    The people who MIGHT HAVE DIED from Covid-19, ALREADY DIED YEARS AGO, from smoking.

    Here's a quick (and probably widly inaccurate) demographic exercise:

    Fact: US deaths from lung cancer and other respiratory ailments, each year: 600,000.
    Fact: People who die from lung cancer, etc., DIE AT YOUNGER AGES, than average.

    Math: Number of people who die each year from smoking, etc.: 600,000

    Breakdown of people who die EACH YEAR, from smoking illnesses:
    for age 0-9: None
    age 10-19: None
    ages 20-29: 1,000
    ages 30-39: 60,000
    ages 40-49: 100,000
    ages 50-59: 200,000
    ages 60-69: 200,000
    ages:70-71: 39,000
    Total:          600,00 PER YEAR!

    600,000 PER YEAR, times 71years = 42,600,000 people DIED BEFORE THE COVID-19 ARRIVED.

    From age 72, and upward, the "smoking related deaths" are no longer relevant, because the age 0f 72 is the average age of life expectancy.

    The coutries with many smokers, also have fewer people who are alive, to become infected by the coronavirus.

    The "previous smokers" are already dead.


    Tom Og

    • Simon Hill Simon Hill Australia says:

      Your statistics suit your predetermined agenda. So,

      Yes, tobacco smoke in cigarettes is a killer. But, nicotine seems to be the ingredient that stops/inhibits the Covid-19 connecting to ACE2, the receptor the virus uses to infects cells. Don't get me wrong, I'm not saying smoke. I am saying, think about nicotine.
      Tobacco is part of the nightshade family. So are eggplants, (green) tomatoes, cauliflower, bell peppers and others that contain a small amount (in comparison to tobacco plants) of nicotine. Now, who has the most consumption of eggplants - China followed by India and both these countries have a small infection rate given their population.
      Another way I was thinking about it is garlic which is good (seemly) for cold viruses. Garlic has a natural pesticide similar to "nightshade" family of plants i.e. nicotine. Assuming that COV-19 has kinda originated from bats and the bats eat insects, the insects can't eat nightshade plants, is it possible that when COV-19 encounters nicotine the virus gets blocked??
      I am focusing on nicotine - not smoking.

      • MG MG United States says:

        Firstly, Tom Og, you seem angry, stop SHOUTING.

        Secondly, You would have to eat 10kg of eggplants to get 1 cig worth of Nicotine. I'm not ruling it out though.

        I also looked at Covid deaths per million population vs percentage of people who smoke by country. I see a positive correlation, in fact, taking 18 notable countries. I also took a smaller group of countries that have similar cultures in Europe and depending on which I included I either get a positive correlation or no correlation.

        I was hoping to see a negative correlation because it would be awesome to have another tool to fight Covid. I'm not too excited about the data.

    • enoughalready911 enoughalready911 United States says:

      Not how stats work pal.  They ARE alive, they have covid and they smoke. Therefore all DEAD previous to covid would NOT be calculated in.
      You presume the numbers are total population but the study is LIVING smokers with covid. Their co-morbidities SHOULD have made them MORE prone to death when infected with covid.  

      Look smoking is nasty and bad for your health.  But in this case it’s possible that smoking tends to 1)tone down the overreaction and inflammation in the lungs 2) the nicotine may HELP with inflammation 3)the thicker mucus in a smoker MAY help protect the lung tissue 4) the cilia may be conditions to lifting foreign matter out of the lungs with a smoker. 5)drawing in air past a lit coal on the cigarette may actually work as a small furnace melting the waxy and fragile outer shell of the virus destroying it.  Thus getting FEWER active viruses into the patient.

  2. Vladimir Breskin Vladimir Breskin Australia says:

    During adaptation to use of fire Pre-human and Human inhaled smoke for 2.5 million years. Smoke from fire fumigated environment of human habitat, let rid of parasites and created new micro floral makeup, rebuild immune system, impacted on physiology (loss of hair/fur cover, exposed skin as the new organ) and cultural behaviour. Addiction to fire, its’ light and heat lasted till mid-20th century. Spread of the habit to smoke tobacco coincided with introduction of electricity and decline of natural smoke of fire in environment. Inhaling of smoke is anthropologically normal.
    When we 2.5 million years (only proven time) were living with smoke and somebody tells you “it is not good, please stop”, remember that consequences should be same as – if you did not inhaled smoke for 2.5 million years - and somebody today tells you “smoke and inhale smoke”.
    Blind paranoia can kill.

The opinions expressed here are the views of the writer and do not necessarily reflect the views and opinions of News Medical.
Post a new comment
You might also like...
Scientists investigate the evolutionary origin of SARS-CoV-2 using recombination-aware phylogenetic analysis