light smoking and mortality

One increasingly common issue in primary care is what to say to patients who smoke only a few cigarettes a day. A recent observational study suggests that even smoking <1 cigarette per day as a consistent, long-term habit is associated with significantly increased mortality (see smoking small amt and mortality jamaintmed2016 in dropbox, or doi:10.1001/jamainternmed.2016.7511).

Details:
-- the National Institutes of Health – AARP Diet and Health Study is a huge database of 168,140 men and women aged 59 to 82 from 2004-2005 that has very specific information on cigarette smoking, including patient recall of their average number of cigarettes per day (CPD) during nine periods of time, starting with <15 yo to >70 yo (most of the time intervals were every 5 years until age 30 then every 10 years after that) along with number of CPD broken down into six categories (0, <1, then intervals of 10 CPD until 30, then >30 CPD). This database includes people from six states (California, Florida, Pennsylvania, New Jersey, North Carolina, Louisiana) and two metropolitan areas (Atlanta Georgia, and Detroit Michigan).
-- Mean age 71, 58% men. at the 2004-5 baseline, there were 22,333 current smokers/156,405 former smokers/111,473 never smokers. Mean follow-up 6.6 years, during which 37331 people died.
-- Of 19,853 current smokers with complete information, 1341 (6.8%) reported smoking <1 CPD and 6036 (30.4%) smoked 1-10 CPD at baseline.

 Results:
-- most people who smoked <1 CPD or 1-10 CPD at the 2004-5 baseline did smoke more CPD earlier in their lives. But, 159 (9.1%) and 1493 (22.5%) of these individuals reported consistently smoking those small amounts throughout their lives.
-- lifelong consistent smokers with <1 CPD or 1-10 CPD tended to start smoking at a later age, were more likely to be non-Hispanic black or Hispanic, and have lower educational levels, less alcohol use, and less history of MI or emphysema.
-- As compared to never smokers,
    -- current consistent smokers at the baseline exam who smoked <1 CPD had a 64% increased all-cause mortality [HR, 1.64 (1.07-2.51)]
    -- current consistent smokers at the baseline exam who smoked 1-10 CPD had an 87% increased all-cause mortality [HR 1.87 (1.64-2.13)]
    -- these associations were similar in men and women, and were found in an array of smoking-related causes of death, but especially strong for:
        -- lung cancer: HR 9.12 (2.92-28.47) for those who smoked <1 CPD; and HR 11.61 (8.25-16.35) for those who smoked 1-10 CPD.
        -- respiratory disease deaths: HR 6.00 (4.05-8.89) for those who smoked 1-10 CPD, but there were too few respiratory deaths in those smoking <1 CPD for statistical power.
        -- cardiovascular deaths: HR 2.78 (1.49-5.18) for those who smoked <1CPD; HR 1.50 (1.13-1.99) for those who smoked 1-10 CPD
    -- The former smokers who had consistently smoke <1 CPD or 1-10 CPD, and those who quit at a younger age, had progressively lower mortality risks; those who quit after age 50 had a higher mortality, about 43% for each of these “light” CPD groups

Commentary:
-- smoking is one of the biggest public health challenges, associated with 5 million deaths per year globally. It has also been found in several analyses to be the single most powerful risk factor for cardiovascular disease.
-- the issue of exposure to small amounts of cigarette smoke is increasingly common in the US. The percent of daily smokers who smoke <10 CPD has increased from 16 to 27% from 2005 to 2014. And, the percent of people who do not smoke every day has increased from 19 to 23%. However, there are studies finding that, especially in young people, they believe that low levels of smoking is safe.
-- In this study, as others, former smokers have a significantly decreased mortality as compared to current smokers. From other studies, these data are particularly robust for cardiovascular disease, where there is a dramatic decrease after about six months of smoking cessation. This may be in part because of the rapid decrease in the prothrombotic fibrinogen levels, which are really elevated in current smokers.
-- the authors make a strong point that smoking duration is associated with much more adverse health effects than smoking intensity. However, on my review of their references (including Flanders WD. Cancer Res. 2003:63(19): 6556), these studies did not look at those who had stopped smoking, where the cardiovascular effects (quickly, as above) and lung cancer affects (over about 15 years) revert to levels close to those of non-smokers. Also, in these old studies, people were smoking many more cigarettes than currently. However, it does make sense to impress upon current "light" smokers that duration of smoking is likely to be more important than the fact that they are only smoking a few cigarettes, and using this to reinforce the importance to quit.

so, my general approach previously had been to assume that even light smoking led to lung cancer, based on data from secondhand smoking studies, and that there is likely increased cardiovascular mortality, based on the fact that smoking even small amounts seems to rapidly increase fibrinogen levels (I have not seen robust data on this last point). The above observational study, however, gives pretty good support to the conclusion that even smoking fewer than one cigarette per day over the long-term is associated with increased all-cause as well as cardiovascular mortality. And, my experience in doing motivational interviewing even in these "light" smokers has been pretty effective.