Since 2004, when the first report of a significant drop in acute MI rate related to the smoking ban in public and work places was published for the town of Helena, Montana , there has been an increasing number of demonstrations of this topic in various countries. The current study provided an up to date meta-analysis of the existing literatures on the effectiveness of smoking-free legislation on the risk of acute MI. Similar to the pooled estimates published in previous meta-analyses [30, 31], this study of 18 studies (44 estimates) indicated a significant reduction of acute MI rate after the implementation of smoking bans in public and work places. To our knowledge, our study was the first to examine the association between the acute MI reduction and the corresponding smoking prevalence decrease, which provided further evidence of the association of environmental tobacco smoke with acute MI.
The findings from this meta-analysis indicated that the smoking-free legislations were associated with a reduced risk of acute MI in general population. Overall, the risk of acute MI decreased by 13%, a moderate effect size compared with previous estimates of a 27% decrease by Dinno , an updated 19% reduction by Glantz  and 17% by two other studies [31, 33]. However, our estimate was consistent with a 15% drop by one study . And this was supported by Pechacek and Babb, which estimated 18% to 19% as the maximum impact that could be expected to be associated with smoking ban . Some studies reported separate estimates for different subpopulations, only those with significant reductions were included by previous meta-analyses, which might be one reason for the relatively higher estimate by previous meta-analyses.
We noticed that the results published so far showed a large variation: studies with smaller population in the United States [10, 11, 28, 35] usually reported larger reductions ranging from 27% to 40%, while larger studies usually reported relative modest reductions: 5% in New Zealand , 8% in New York , 13% in four Italian regions , and 13% in Switzerland . At the same time, some subgroup estimates did not find significant reduction (for example, among people aged 60 years old and above in Piemont, Italy , population aged 30–54 years old in Christchurch, New Zealand ). The discrepancy among these studies could in part be attributable to different compliance with the legislation and different changes in smoking behavior among different countries.
Our meta-regression analysis showed that the reduction in acute MI risk was greater in studies with higher smoking prevalence rate decreases, suggesting that the protective effect of legislation could be directly attributed to the corresponding smoking rate decrease. This was corroborated by the study in Arizona, US, which demonstrated a 3.9% decrease in smoking prevalence, inducing a 13% reduction of acute MI hospitalization .
There were several possible pathways linking the smoking ban regulations to acute MI reduction. The smoking-free policy could reduce the amount of cigarette consumption among the active smokers, encourage smoking quitting, enhance the awareness of the public about the harmfulness of smoking, and more importantly reduce environmental tobacco smoke exposure among the passive smokers. Thus, the beneficial effects could be expected to be in a long-run fashion, besides the short-term effects .
The implementation of smoking ban law had been reported to result in significant decline in environmental tobacco smoke [16, 38]. For example, in New York, after the implementation of the statewide smoking restriction law, population exposure to environmental tobacco smoke declined by nearly 50% and cotinine levels in the saliva from New York State adults declined from 0.078 ng/ml to 0.041 ng/ml [11, 16]. In UK and Ireland, a 95% reduction in ETS among the general population was reported 9 months after the smoking ban . In Scotland, the indoor fine particulate matter (PM2.5) concentration declined by 86% two weeks after the Scottish legislation . In Ireland, salivary cotinine concentration among hotel workers fallen by 69% and air nicotine concentration dropped by 83% .
This finding strengthened our knowledge that passive smoking was a serious risk factor for acute myocardial infarction and that its elimination reduced the acute MI occurrence [41, 42]. One percentage decrease in smoking prevalence was estimated to reduce 2.8 percentage of acute MI rate on average. And this association was biologically possible. Exposure to ETS could rapidly induce platelet aggregation, thrombosis, endothelial dysfunction and inflammation, and these effects were comparable to those suffered by active smokers [41, 42]. These effects were estimated to increase the risk of acute myocardial infarction by around 10% globally , which was in line with the current estimate. As this risk factor was relatively easily modifiable, the expansion of the smoke-ban policy in public and working places were expected to have a pivotal public health significance.
The strength of the present study included the up to date pooled estimation of the relationship of acute MI risk with the smoke-free legislations and the examination of the correlation of this reduction with the corresponding smoking rate decreases. Meanwhile, a few limitations of this study should be noted. All the included studies were the ecological studies in study design nature, which caused uncertainty in causal inference. Nonetheless, when taken in aggregate, these studies offered evidence that smoking bans were followed by significant reduction in the rate of acute MI in the general population. A recent study reported contradicting results on the impact of smoking ban on acute MI in 387 US counties when adjusting for non-linear secular trend in acute MI compared with the model adjusting for linear secular trend . However, most of the studies included in this meta-analysis did not consider non-linear trend of acute MI rate , which might have caused concern in our estimate. Future studies should consider the non-linear secular trend in acute MI occurrence. Temporal trend of acute MI rate might be influenced by factors other than smoking exposure, such as long-term trend, other air pollutants, atmospheric temperature, influenza epidemics, changes in diagnostic standards and preventive strategies, which were not taken into account in this meta-analysis. It was also possible that the reduction in acute MI risk could be due to the reduced amount of cigarettes smoked. It should also be noted that the smoking rate used in the analysis came from diverse sources, where the definition of smoking might be different, however, we used the difference of smoking prevalence before and after the smoking free legislation as the indicator, so this should not have influenced the result seriously.