On the relationship between smoking bans and incidence of acute myocardial infarction

During the last few years several studies have reported a substantial reduction of acute myocardial infarction (AMI) in the general population few months after the enforcement of comprehensive smoking bans. We reviewed the consistency and plausibility of this association, investigating the effect of the Italian law, entered into force on January 10, 2005. We compared the AMI incidence on the first year after the ban with the period before (2000-2004) in the Tuscany population aged 30-64 years. The analysis was performed with a Poisson model of the monthly time-series, adjusting for seasonality and comparing different models with linear and non-linear long-term trends. While the model with linear time trend estimated a decrease of 5.4% (RR 0.95; 95% CI: 0.89-1.00), this effect completely disappeared once the linearity assumption was relaxed (RR 1.01; 95% CI: 0.93-1.10). The model with non-linear terms showed a significantly improved fit (P-value = 0.01). The estimate of the effect of the ban seems to be highly sensitive to the model specification and to the effects of unaccounted factors which could modify the trend of AMI incidence, such as changes in the prevalence of other risk factors or the modification of diagnostic criteria. Several arguments which are put forward to inspect the causal relation between smoking bans and AMI indicate that the plausible effects could be lower than the estimates reported so far.