OBJECTIVE:Lung, laryngeal and esophageal cancers have smoking as one of their main risk factors. The objective of this study was to evaluate the population attributed risk (PAR) of smoking for these forms of cancer. METHODS: The study was based in three case-control studies conducted in medium size cities in Brazil. Incident cases of lung cancer, laryngeal cancer and esophageal cancer seen at a hospital setting and diagnosed through biopsy were analyzed; controls were hospitalized patients with another diagnoses. Smoking was the exposure factor measured at three levels: non-smokers, former smokers and smokers, which were defined using a questionnaire applied by trained interviewers. For effect measure, odds ratio was used and the populational attributed risk for smoking was then calculated for a 95% CI. RESULTS: A total of 122 lung cancer cases and 244 controls, 50 cases of laryngeal cancer and 48 cases of esophageal cancer, and 96 controls for both of them were studied. The prevalence of smoking exposure was 34%, which is the overall prevalence of smoking in this city's adult population. Odds ratios (OR) for the PAR analysis were the adjusted OR for confounding variables from each study. Lung cancer PAR was 63% (95% IC, 0.58-0.68) for former smokers and 71% (95%IC, 0.65-0.77) for smokers. Larynx cancer PAR was 74% (95% IC, 0.70-0.78) and 86% (95%IC, 0.81-0.85) for former smokers and smokers, respectively. Esophageal cancer PAR was 54% (95%IC, 0.46-0.62) for smokers. CONCLUSION: Smoking is an avoidable risk factor and smoking cessation could be responsible for significant reductions in the incidence of these three forms of cancer.
Smoking; Lung neoplasms; Laryngeal neoplasms; Esophageal neoplasms; Attributable risk; Case-control studies
