Smoking just one cigarette, one too many

​ 

A recent systematic review and meta-analysis confirmed that smoking even one cigarette/d is really bad, with about ½ the attributable risk for cardiovascular disease as smoking 20 cigarettes/d (see http://www.bmj.com/content/bmj/360/bmj.j5855.full.pdf ).

Details:

--141 cohort studies in 55 publications were included

Results:

--men:

    --coronary heart disease (CHD): pooled relative risk was 1.48 for smoking one cigarette per day and 2.04 for 20 cigarettes per day

        -- but 1.74 and 2.27 among studies in which the relative risk had been adjusted for multiple confounders

        --ie, men smoking 1 cigarette/d had 46% of the excess CHD risk of smoking compared to ​20 cigarettes/d (64%, if use the fulling adjusted risk) 

        ​--and, men smoking 5 cigarettes/d had 57% of the excess risk of smoking 20 cigarettes/d

    --stroke: pooled relative risk​ was 1.25 for smoking one cigarette per day and 1.64 for 20 cigarettes per day

        -- but 1.30 and 1.56 among studies in which the relative risk had been adjusted for multiple confounders

        --ie, men smoking 1 cigarette/d had 41% of the excess risk for stroke compared to smoking 20 cigarettes/d (53%, if use the fulling adjusted risk)

        ​--and, men smoking 5 cigarettes/d had 52% of the excess risk of smoking 20 cigarettes/d

--women:

    --pooled relative risk for coronary heart disease was 1.57 for smoking one cigarette per day and 2.84 for 20 cigarettes per day

        -- but 2.19 and 3.95 among studies in which the relative risk had been adjusted for multiple confounders

        ​--ie, women smoking 1 cigarette/d had 31% of the excess CHD risk of smoking compared to 20 cigarettes/d (38%, if use the fulling adjusted risk)

        --and, women smoking 5 cigarettes/d had 43% of the excess risk of smoking 20 cigarettes/d

    --stroke: pooled relative risk​ was 1.31 for smoking one cigarette per day and 2.16 for 20 cigarettes per day 

        -- but 1.46 and 2.42 among studies in which the relative risk had been adjusted for multiple confounders

        --ie, women smoking 1 cigarette/d had 34% of the excess risk for stroke compared to smoking 20 cigarettes/d (36%, if use the fulling adjusted risk)​​

        ​--and, women smoking 5 cigarettes/d had 44% of the excess risk of smoking 20 cigarettes/d

Commentary:

--there are 1 billion smokers around the world, with higher prevalence in developing countries: 90% of men and 11% of women (though the political subtext here is that tobacco manufacturers, including in the US, shifted their target to the developing world as people here became more intent on stopping smoking and smoking rates started to plummet)

--the Health Survey for England in 2013 and 2014 found that 26% of current smokers wanted to cut their smoking amount but not stop and 40-41% said they smoked less than the prior year

 --the % of smokers consuming 1-5 cigarettes/d has risen from 18.2% in 2009 to 23.6% in 2014.  Similar pattern in US: those consuming <10 cigarettes/d increased from 16% in 2005 to 27% in 2014

--and, among 24,658 US adolescents, 10% considered light smoking non-harmful and only 35% of light-smokers thought that their smoking was associated with lots of harm

--the data in the above meta-analysis however confirmed that the risk of either coronary heart disease or stroke was not much different if smoking 5 vs 1 cigarette/d

--consistent with other studies, women seemed to have a higher risk of coronary heart disease at lower levels of smoking than men

--the data on lung cancer risk from smoking seems to be linear: the less smoked the lower the risk (ie, smoking 1 cigarette/d carried a risk of about 1/20th of smoking 20/day). but prior studies have not found this to be true with coronary heart disease, with, for example, a 30% increased risk in second-hand exposure in never-smokers, vs a 2-3x risk in smokers (they found that the risk for second-hand exposure was similar to the risk of smoking 1 cigarette/d).

--various sensitivity analyses did not contest the overall results (eg, one cannot assume that the cigarette consumption remained constant throughout the study, and was likely to decrease as with the prevalent trends. but assessment of only the earlier years of the study, prior to likely large changes in cigarette consumption, did not change the conclusions much. and, from an other angle: there were 3 studies done of heavy smokers (>15/d) decreasing their consumption by at least 50% (verified by carbon monoxide or cotinine concentrations) finding no clear reduction in coronary heart disease 5-21 years later. this was in contrast to cardioprotection in those who quit smoking. but there was a large reduction in lung cancer risk with smoking reduction, as expected.

--confounding is always a concern in these types of observational studies. one might expect that the heavier smokers also had generally less healthy lifestyles, but that should decrease the excess risk of smoking only 1 cigarette/day if those people adopted healthier lifestyles with the lower cigarette consumption. and, of course, there are large concerns about combining large quantities of data from very different studies which controlled for different variables in groups of patients who probably did not smoke exactly the same amount every day and may well have mis-stated their actual cigarette consumption.... but overall the data are reasonably consistent, and the fact that in using the same data there was a linear relationship with lung cancer reinforces their conclusions about the non-linearity for heart disease (though not consistently analyzed in this systematic review/meta-analysis, several of the large studies included in this analysis also looked at lung cancer, confirming large reductions with decreased cigarette consumption)

--how can this heart disease connection be explained mechanistically? cigarette smoking has many different effects on the body which might lead to clinical heart disease, including endothelial dysfunction, platelet activation/aggregation, arterial stiffness, oxidative stress, inflammation, heart rate variability, lipid changes, prothrombotic changes (smoking leads to significant increases in plasminogen activator inhibitor-1) and that many of these effects can be elicited with brief exposure to second-hand smoke. as a likely related issue, coronary heart disease is also associated with air pollution, esp with fine particulate matter <2.5 mm in diameter, in a dose-response manner. also, the studies on smoking cessation suggest a quite dramatic decreased risk within 6 months, which also suggests that a major part of the mechanism is not simply atherosclerosis but by some of these more transient phenomena (platelet changes, prothrombotic factors, inflammation, endothelial dysfunction)

So, I think the real message here is that any smoking is bad. though it is certainly good for patients to cut down in cigarette consumption (and, of course, we should be very supportive of these efforts), it is important that the patients understand the ultimate goal is stopping smoking entirely, that decreasing smoking to very low levels does decrease the risk of cardiovascular disease a lot, but the residual risk is still in the 50% range.

to be added to the email list: contact me at gmodest@uphams.org