Mortality and Tobacco Consumption: the Population Attributable Fraction Method as a Tool to Estimate Damage

Abstract

Do all risk factors influence the incidence of events in the same way? The answer is “no”, and one of the main reasons mentioned by some cardiovascular medicine specialists is the INTERHEART study (1), which shows that the magnitude of the association among nine conditions or cardiovascular risk factors is heterogeneous with respect to the incidence of a first acute myocardial infarction (AMI). For example, the odds ratios (ORs) observed were 1) 1.91 for history of arterial hypertension; 2) 2.37 for diabetes; 3) 2.87 for smoking (smokers vs. non-smokers); 4) 3.25 for an elevated ApoB/ApoA1 ratio, among others. This information has had healthcare consequences: it provided us with evidence to establish a hierarchy or prioritization of health problems in order to avoid overwhelming patients and support their long-term efforts. However, the INTERHEART study also provided population health information by widely spreading the concept of risk or population attributable fraction (PAF) –a “measure to estimate the proportion of cases that are attributed to a given exposure”– among healthcare professionals. (2) Thus, considering the contribution of each of these risk factors to the incidence of a first AMI event, the hierarchy defined by the ORs was modified since the PAF was 1) 49.2% for an elevated ApoB/ApoA1 ratio; 2) 35.7% for smoking; 3) 17.9% for history of arterial hypertension; and 4) 3.9% for diabetes. Same data, but different information: on the one hand, a more “clinical-related” perspective and, on the other hand, an epidemiologicalrelated perspective. The epidemiological perspective is used to understand population health problems, evaluate the results of the actions (or lack of actions) and propose improvements. The epidemiological approach has had some relevant milestones in our country and in the region, such as the publications of the CARMELA Study (3,4) and the National Survey of Risk Factors conducted