Smoking and COVID-19 severity

Clinical data

There have been a few studies looking at the relationship between smoking and disease severity. Some of the meta-analyses are summarized here:

• A meta-analysis from February 15, 2020 included 10 studies from China, with a total of 76,993 patients. Six of the 10 studies included data on smoking, and in these studies (n unknown), smoking was noted in approximately 7% of patients (well below the community prevalence of smoking in China, which approaches 50%), however the authors admit that “high and significant heterogeneity was found between the 6 included studies” and no information was provided about quality of data within the studies. In addition, there is no data presented on patient outcomes.[7]
• A meta-analysis from March 9, 2020 included 5 studies from China totalling 1,399 patients. Only one study (Liu) showed a statistically significant association between smoking and severity; the others trended toward an association but were not significant.[8]
• A meta-analysis from March 17, 2020 included 5 studies from China (4 in Wuhan and one in mainland China) totalling 1,549 patients. An unadjusted analysis found that compared to nonsmokers, smokers face a higher (approximately double) risk of developing severe complications (including ICU admission or invasive ventilation) and dying from COVID-19 infections.[9]
• A meta-analysis from April 1, 2020 included 13 studies from China totalling 5,960 patients. The study looked at reported smoking rates among hospitalized patients and did not adjust for confounding factors. The studies included reported smoking rates of 1.4-12.6%.[10]
• A meta-analysis from April 6, 2020 included 12 studies from China (10), Korea (1) and the U.S. (1), with a total of 9,025 COVID-19 patients. This analysis found that 5.5% of the severely ill patients (n=878) had a history of smoking. Furthermore, 17.8% of patients with a history of smoking experienced disease progression compared to 9.3% nonsmokers, and found that smokers have 2.25 times the odds of severe COVID-19 outcomes than nonsmokers. The authors do not appear to have corrected for other patient characteristics, and note that there was variability in how the individual studies defined “smoking,” particularly whether former smokers were included. However, this does appear to be the most robust meta-analysis to date. See table of included studies at the end of this document.[11]

It is notable that many of these meta-analyses depended on the same source articles for their analyses, though they did not completely overlap. Included in all meta-analyses, Guan et al was the largest study with 1,099 patients in China, which found that 12.4% of current smokers (n=137) were either admitted to the ICU, intubated, or died, compared to only 4.7% of never smokers (n=927).[12]

Is smoking protective?

There have been several news articles and attention given to this idea. The argument is largely based on the reported rates of smoking:

In the above studies, although smokers tended to be more likely to have severe disease than nonsmokers, most articles cite a smoking prevalence rate of 3-18% (most < 10%), which is well below documented smoking rates in China. For example, smoking may be protective overall, but only smokers with other confounding comorbidities (which may or may not be caused by smoking) are being admitted to the hospital, so they are sicker than non-smokers?

In one study found outside China, Miyara et al studied 482 COVID positive patients in France (343 inpatients and 149 outpatients, excluding patients in the ICU) and similarly found that the number of patients who reported smoking (approximately 5%) was much lower than the rate of smoking in the general population (approximately 25%), after standardization by age and gender.[13] The authors argue that other studies from China and the U.S. that show increased disease incidence and severity among smokers are flawed due to (1) failure to include outpatients, (2) failure to adjust for age and gender, (3) failure to take into account the baseline smoking prevalence in the study community, and/or (4) missing data about smoking status.

ACE2 Receptor hypothesis:

Lungs exposed to smoke appear to have an increased number of ACE2 receptors through a number of proposed mechanisms, including upregulation of ACE2 receptors and proliferation of ACE2 receptor-heavy cell types.[14][15][16] It is also believed that the ACE2 receptor is used by SARS-CoV-2 to gain cell entry;[17][18] this is considered to be a potential mechanism for increasing the risk of developing COVID-19 after viral exposure among smokers.

Additional information:

Smith et al demonstrated increased ACE2 receptor levels in lung tissues of both mice exposed to smoke, as well as human smokers; they also found that former smokers had about a 30% reduction in ACE2 levels compared to current smokers.[15] Leung et al examined ACE2 expression in the small airway epithelia of smokers and COPD patients and found an increased expression of ACE2 receptors in the airway of current (but not former) smokers and individuals with COPD.[19] Although there are a few hypotheses, the mechanism for which smoking upregulates ACE2 remains to be demonstrated.

Nicotinic acetylcholine receptors hypothesis:

There is a physiologic basis that smoking may have a protective effect in the form of nicotinic acetylcholine receptors (nAChR). Nicotine may also lessen the hyperinflammatory response of COVID-19 by engaging in the ‘cholinergic anti-inflammatory pathway’,[20] a physiological mechanism that inhibits release of macrophage TNF and systemic inflammatory response.[16] Acetylcholine (ACh) has auto/paracrine control of macrophage cytokines production through their nAChR.[21] Dysregulation of macrophage nAChRs leads to overproduction of cytokines, similar to the cytokine storm syndrome which leads to hyperinflammatory syndrome as seen in a subgroup of COVID-19 patients.[20]

Observation of structural evidence supports the hypothesis that SARS-Cov-2 virus is a nAChR blocker, and nicotine may be able to control/prevent the disease by competing with SARS-Cov-2 in the binding of nAChR.[20] Currently, in France, there are plans to test nicotine patches on COVID-19 hospitalized patients and in the general population, and of note, sales of nicotine replacement products have had to be regulated after this announcement.[20] Interestingly, Ivermectin, which recently had been shown to inhibit replication of SARS-CoV-2, is a nAChR modulator.[22]

Comorbidities:

Although there are relatively few studies published looking at the direct relationship between smoking and COVID-19 acquisition/outcomes, it is also crucially important to note that smoking is associated with cardiovascular disease, chronic kidney disease, hypertension, chronic lung disease such as COPD - these comorbidities have all been abundantly demonstrated to increase morbidity and mortality from COVID-19 (whether or not they were caused by smoking).

Vaping:

Vaping’s association with the progression of COVID-19 has not been studied intensively for comprehensive interpretation. However, there is evidence that aerosol from vaping can harm lungs at cellular and organ levels and worsen immune response.[23] However, the sample size of the study was small to establish compelling evidence. Additional evaluation on current and former vaping smokers should be valuable for future investigation.