For mortality associated with respiratory tract cancer
For mortality associated with respiratory tract cancer, the RR estimated by the intercept-only model presented a gradient in which the values were also highest within the central region of the city and lowest on the periphery, with the exception of the extreme south (Fig. 3a). Fig. 3b shows a change in the spatial variability of mortality without the influence of traffic density and the MHDI. In the central region, there were no longer any weighting areas with an RR of 2.0–3.0 and in the north and east regions have an increased in the number of those with an RR between 1.3 and 2.0. There was an overall reduction in the number of weighting areas with a low RR and in the extreme southern region of the city an RR between 1.3 and 2.0 reduced. The risk in the extreme southern could be explained by the lower influence of the neighborhood due to the larger areas, the reduced number of neighbors and the similarity between them. In addition, these weighting areas have a smaller population.
Discussion Our study showed that areas with high traffic density coincided with areas that presented higher crude RR values for the incidence and mortality for respiratory tract cancer. The ecological model showed that the spatial variability of the incidence was influenced by the traffic density and socioeconomic status. After adjustment for those covariates, there was an overall decrease in the RR values, although they increased on the periphery, the spatial pattern no longer being explained by the covariates. This finding, showing the influence of traffic, is in agreement with those of previous local studies that reported the influence that ozone and particulate AZD-5991 with an aerodynamic diameter ≤10 μm (PM10), both of which are important traffic-related air pollutants, have on the incidence of lung and laryngeal cancer in the city of São Paulo [7,29]. The spatial variability of mortality associated with respiratory tract cancer was also influenced by the covariates. The upper limit of the crude RR was lower than that of the incidence, fewer areas corresponding to that value, although this higher RR value for incidence and mortality were comparable in the most central region of the city. In the ecological analysis, adjusting for the covariates also promoted a reduction of the RR in the central region of the city, with a corresponding increase in the areas on the periphery. Although the 0.025 quantile of the posterior distribution of the effect of traffic density was less than 1 (0.99), it is highly likely that traffic density is a risk factor for mortality associated with respiratory tract cancer, the probability that its effect size was ≤1 being 0.05. In a review study that evaluated the influence of air pollution and its effects on public health in the city of São Paulo, the authors concluded that traffic-related air pollution is associated with increased mortality risk . There have been few studies evaluating the association between traffic density and the frequency of adult cancers. A study conducted in the United States used traffic density data from 168 counties and showed that the risk of developing adenocarcinoma or squamous cell carcinoma of the lung was 136% and 68% higher, respectively, for individuals residing in areas with a traffic density of approximately 937 vehicles per square mile than for those residing in areas with a traffic density of approximately 1 vehicle per square mile, with a dose-response gradient . Other studies have used the distance between the residence and a roadway with high traffic volume as an exposure variable. In a study conducted in Denmark  the authors detected an increased incidence of lung cancer among individuals living within 50 m of a major roadway (traveled by > 10,000 vehicles/day). In Italy, the overall risk of death from lung cancer was found to be significantly higher for individuals living within 25 m of a major roadway than for those living 500–1999 m from such a roadway . Another study conducted in Italy  reported that the mortality risk was higher among individuals living within 50 m of a roadway with high traffic volume than among those living ≥ 250 m from the same roadway.