In this time of health care reform debate in the U.S., it is fashionable to focus on what is wrong. Unmentioned is arguably the most important health triumph in the past 50 years: the steep decrease in deaths from cardiovascular disease (1- 3). Between 1950 and 2005, the age-adjusted, all-age death rate from heart disease decreased from 587 per 100,000 persons to 211 per 100,000, a spectacular improvement (4). As often happens with such remarkable successes, the credit gets shared. One factor with a major claim to causality is the decrease in smoking prevalence as the result of strong tobacco-control programs (5- 6).

In that regard, Goldman (2) estimates that 25% of the decrease in deaths from cardiovascular disease is from decreased disease incidence and 75% is from reducing deaths in persons with known disease. Of that 75% decrease, approximately one-half is due to risk factor reduction. Obviously, smoking figures in both primary prevention (decreased disease incidence) and secondary prevention (risk factor reduction). Indeed, the decrease in smoking prevalence is another modern health triumph. The prevalence of smoking among men in the U.S. has decreased from 57% in 1955 to 21% in 2007, whereas smoking among women declined from a high of 34% in 1965 to 18% in 2007 (5). The 2007 overall smoking prevalence rate of 19.8% marked a modern low (7).

There are 2 routes whereby tobacco smoke can hurt a patient's heart—by smoking directly or by inhaling someone else's smoke. When smoking rates were high, it was hard to avoid secondhand smoke, and it was also unclear that secondhand smoke was damaging because the assumption was that—as with smoking and lung cancer—sustained exposure was required for harm. But now that smoking is less common and smokers are increasingly marginalized (8- 9), evidence has emerged that secondhand smoke exposure is nearly as harmful to the heart as is chronic active smoking. Barnoya and Glantz (10) recently reviewed the many ways in which this harm occurs. Secondhand smoke exposure alters platelet function, causes endothelial dysfunction, increases arterial stiffness, decreases levels of high-density lipoprotein, increases markers of inflammation, increases arterial intima-media thickening, increases infarct size, causes oxidative stress and mitochondrial damage, decreases heart rate variability (thus increasing the risk of malignant arrhythmias), and increases insulin resistance (10). It is hard to imagine substances that would be more cardiotoxic. Furthermore, these adverse effects are observed at very low exposure doses (11).

So, what can cardiologists do to keep their patients from being exposed to tobacco smoke? Obviously, they should counsel patients who smoke to quit by serving as a resource to help that happen; directing them to affiliated cessation systems such as those of Kaiser Permanente, Group Health Cooperative, the Mayo Clinic, or the Veterans Administration system; or referring them to toll-free telephone counseling available through 1-800-QUIT NOW (12- 13). However, cardiologists now must also attend to the risks of secondhand smoke exposure.

This issue of the Journal contains a meta-analysis by Myers et al. (14) of published studies that examine changes in rates of acute myocardial infarction (AMI) after the institution of second-hand smoke bans in specific locales. The review takes advantage of natural experiments in which public smoking was banned, and rates of AMI were compared before and after that ban, sometimes including comparison sites without a ban. The inclusion criteria restricted the analysis to only 11 peer-reviewed articles representing 10 sites: 5 from the U.S., 1 from Canada, 3 from Italy, and 1 from Scotland. Observation times for measuring the effect ranged from as short as 2 months to as long as 3 years. The populations covered in the studies varied from a high of 19 million people (New York State) to a low of 29,000 (Bowling Green, Ohio). In the meta-analysis, overall results were weighted by the population size included in each study, as well as the duration of observation after the ban was imposed.

The overall mean decrease in AMI incidence after ban imposition was 17%. All studies showed at least 1 subgroup with decreases, and 9 had overall substantial declines, ranging from −9% to −50%. Two sites, both in Italy, had essentially no change. Differences were greater in sites from the U.S. and in those with longer observation periods, with AMI incidence decreasing by 26% with each year of post-ban observation. As shown in the authors' Figure 3, the longer the observation period after the ban, the lower the incidence rates.

Why is this article important? First, by adding published reports outside of the U.S., it builds on evidence of a previous meta-analysis that showed a decrease in post-ban AMI incidence of 19% (15). When the first observation of the effect of smoking bans on community disease incidence in Helena, Montana, was reported, it was unclear whether this relationship was an artifact or real, especially because the comparison site had a large (46%) increase in AMI frequency (16). To some, those results seemed too good to be true. By combining all of the published studies on the effects of such bans, Myers et al. (14) add to the evidence that these bans do protect the hearts of those prone to coronary disease. Further, in the 5 reports that used nonban comparison sites, the nonban sites all had much smaller decreases in AMI prevalence. Finally, the findings that the effect is stronger the longer the period of observation and that the largest declines occurred in the U.S.—where smoking prevalence is lower and thus the proportion of nonsmokers who could be exposed to secondhand smoke is greater—support the hypothesis that smoking bans are beneficial. As the authors note, the smaller effect sizes all occurred in studies outside the U.S. plus the New York State study. All these sites had short observation periods.

What are the limitations of this study? Publication bias is always a potential problem, although the authors were unable to surface any reports that went unpublished. More problematic is the reality that in the >3,000 communities and 33 states that restrict public smoking in the U.S. (17), most have not tracked changes in AMI incidence. The same holds for other countries. Another potential limitation is the vigor with which such bans are enforced. Given the lower smoking prevalence rates in the U.S. and the stigma attached to smoking in that country, it is likely that such bans are more easily accepted there and better enforced than in Italy and Scotland, where the published reports showed either a lower (Italy) or average (Scotland) effect of smoking bans. A related question is whether these decreases merely reflect the secular trends of decreasing AMI frequency. However, the rate of decrease is much greater than those secular trends, and those studies with comparison sites all showed lower rates of decline, or in the case of Helena, a 46% increase.

Could there have been a coding effect, whereby the diagnosis of AMI somehow changed in communities with bans? That seems unlikely. More troubling is the fact that the effects are much smaller in the reports from large population sites, with the exception of Scotland. The authors tested whether geographic region (U.S. vs. non-U.S.), population size, or length of post-ban observation period affected the incidence reduction rate and concluded that region and observation period did but size did not. How robust that conclusion is remains to be seen. It is likely that the larger the population encompassed by a ban, the less uniform enforcement of the ban might be.

The take-away messages for cardiologists are clear. A 17% risk reduction for AMI is not trivial. It is prudent to assume that exposure to secondhand smoke is almost as dangerous to persons with diagnosed or latent coronary disease as active smoking (10). Therefore, cardiologists should expand their clinical repertoire to include screening and counseling for secondhand smoke exposure, just as they screen for lipid disorders. In their roles as health advocates, they should also support bans on public smoking, as well as other tobacco control measures such as tax increases on cigarettes, counter-marketing campaigns, and expanded cessation services such as telephone quitlines (6,18- 19). These initiatives lower the probability of young people initiating smoking, increase the rate at which smokers quit, and lower the frequency of smoking among those not yet willing or able to quit. Cardiologists not only have much to celebrate about the spectacular decreases in cardiovascular disease, they also have the opportunity to do much more. Further decreased exposure to tobacco smoke, such as occurs with public smoking bans, is a keystone to such progress.