Data analyses
For all health outcomes considered, a preliminary analysis was carried out in order to compare mortality and morbidity of the entire cohort with an external reference area. SMR (Standardized Mortality Ratio) and SIR (Standardize Incidence Ratio) were calculated to compare mortality and cancer incidence of the cohort with the regional population or with the local health district population for hospital admissions.
In the integrated database of the cohort, for each individual we had demographic information, SES level, exposure levels at the residence for heavy metals and NO2, date of entry, date of exit and several outcomes (mortality, incidence, and hospitalization). Person-years at risk were calculated by calendar period, gender and age. Age and socioeconomic status adjusted rate ratios (RR) and 95% Confidence Intervals (95% CI) of the association with heavy metal exposure were estimated with Poisson regression separately for males and females and for all the health outcomes considered, using the lowest exposure category as the reference. The choice of the relevant outcomes was based on the literature review. Analyses for soft tissue sarcoma were performed also by combining the results for men and women. To evaluate the possibility of confounding from traffic-related air pollution, we ran an additional analysis using NO2 levels as an adjustment factor. We considered a priori that the associations to be noted were those with an increasing trend in the adjusted rate ratios across the exposure categories and/or a statistically significant rate ratio in the third or fourth quartiles when compared with the first one.
The statistical package STATA was used for all analyses [20].