Smoking bans in public places result in a reduced incidence of acute myocardial infarction

Since 2004, when Sargent, Shepard and Glantz published the first results of a community smoking ban in Helena (MT, USA) [1] that observed a 40% decrease in the incidence of acute myocardial infarction (AMI) but a return to baseline rate after judicial suspension of the ban, 13 additional communities, states and countries have reported similar experiences with smoking bans, ranging from a 50% decrease to a 2% increase in incident AMIs [2–15]. These reports are all ecologic in design, vary widely in population size (29,636 to >18 million individuals) and duration of postban observation (3 months to 3 years), and differ in event definition. Most do not identify age, gender or smoking status of cases. In late 2009, these reports were subjected to two separate meta-analyses [16,17], which reported summary incidence rate ratios of 0.83 (95% CI: 0.80–0.87) and 0.92 (95% CI: 0.87–0.99), respectively. The meta-analyses differed in their inclusion criteria. Interstudy variance was primarily due to the wide range of observation times. Importantly, both studies noted a 15% per year reduction in incident AMIs as estimated by metaregression. Subgroup analysis of individual studies suggested that benefit was accrued primarily by nonsmokers, and was largest among women and younger persons.

Not all authors support community smoking bans. Interestingly, three out of four reports from Italy show no overall benefit [8,9,12]. An unpublished working paper by the National Bureau of Economic Research, which matched national hospitalization databases to US communities with smoking ordinances, failed to identify a benefit [101]. These investigators suggested that observed decreases in relative risk might be attributed to the natural temporal variance of AMI incidence. I am not aware of other studies or editorial opinions in the scientific literature that do not support community smoking bans. Although bans have nonetheless been opposed on social, market and legal grounds [102].

I propose that the hypothesis that community smoking bans reduce the risk of AMI has now been proved, based upon the 15 cited studies and the two meta-analyses that satisfy nearly all of the widely accepted criteria for proof of causation [18,19].

Is there evidence from true experiments in humans? As a randomized trial with humans would be unethical, the currently published ecologic studies and meta-analyses are the most scientifically valid study designs available and must suffice. The large, historical and contemporaneous controlled, prospective study by Pell et al., which measured both the smoking status and passive exposure of cases [11], is particularly well designed, fulfills nearly all the Newcastle-Ottawa quality criteria and is particularly compelling [103]. It demonstrated a 17% reduction in admissions for acute coronary syndromes.

Is the association strong? The strength of association, as measured by the incidence rate ratios of 0.83 and 0.92 in the meta- analyses [16,17], is reasonably strong, remembering that the measurement scale for risk ratios below 1.0 is not linear.

Is the association consistent from study to study? There is fair consistency among the 15 original reports, with incidence rate reductions ranging from -8 to -50% in 12 studies [1–7,10,11,13–15]. Three Italian studies observed incidence rate changes from 0 to 2% [8,9,12]. Both meta-analyses noted that most of the interstudy variance was attributable to variation in duration of post-ban observation, which ranged from 3 months to 3 years [16,17].

Is the temporal relationship correct? All studies determined AMI incidence before and after the institution of smoking bans, and most observed a decrease in incidence after the bans. However, in at least some populations studied, the incidence of AMI was already decreasing before the ban. Several studies also utilized parallel contemporaneous control populations, none of which experienced AMI incidence reductions to the extent seen in the ban communities. Particularly convincing in this regard is the Helena study, in which AMI incidence decreased by 40% during the ban and returned to the pre-ban level after judicial suspension of the ban [1].

Is there a dose–response relationship? Both meta-analyses noted a linear decrease in incidence associated with increasing ban duration at a rate of 15% per year [16,17].

Does the association make epidemiologic sense? It has been established that second-hand smoke exposure increases the risk of AMI by 25–31% [20,21]. A 15% per year decline in AMI incidence with a smoking ban follows intuitively and was predicted by modeling [22,23].

Does the association make biological sense? In animals and humans, tobacco smoke produces chemical and biological changes, such as increased platelet activation, lipid metabolism and endothelial function, which are involved in the pathogenesis of AMI [24–27].

Is the association specific? No. Factors other than environmental tobacco smoke are associated with AMI. Certainly, lowering lipids [28] and blood pressure [29] also decreases the risk of AMI.

Is the association analogous to a previously proven causal association? Among several possible examples, a reduction in the incidence of AMI is similarly produced by lowering exposure to dietary cholesterol through the Mediterranean-style diet [30].

The currently available scientific evidence is sufficient for governmental entities at all levels to move ahead with widespread and comprehensive bans of smoking in public places. As of October 2009 in the USA, 821 municipalities in 38 states have established bans. Such bans could prevent more than 100,000 AMIs per year in the USA alone. It is now time for all communities, states and countries around the world to act.

Financial & competing interests disclosure

The author has no relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties.

No writing assistance was utilized in the production of this manuscript.