In a recent cohort study in aluminium foundries and remelting plants an unexpectedly high risk of lung cancer was found in workers in sand foundries. On the basis of present and, historical measurement data, we developed a statistical model for exposure to total dust and crystalline quartz for different jobs and time periods. Cumulative dose estimates of total dust and crystalline quartz were calculated and used in a nested case-control study in the cohort.
From the cohort of foundry workers (n=4,435), 51 cases oflung cancer were identified. After exclusion of cases and controls with nonspecific job titles, 46 cases remained. The final analysis was performed on 31 cases and 233 controls with one year or more of employment. Measurement data for total dust and crystalline quartz had been previously collected in each company in an industrial hygiene survey, parallel to the cohort study. Historical measurement data from the 1960s and onwards were added, totalling, 207 total dust and 103 crystalline quartz exposure observations. Regression models, using the determinants of job title, timeperiod, type of foundry and size of production, were developed for assessing historical total dust and crystalline quartz air concentrations. These estimates were used to calculate individual cumulative exposure in the case-control study.
In the multiple linear regression analysis, the determinants explained much of the variations in dust level (r2=0.58). The explained variation in crystalline quartz was much lower (r2=0.13). The regression coefficients for the type of foundry, time-period, and size of production were statistically significant for total dust. The maximum predicted dust and quartz levels were 10 mg/m3 and 0.07 mg/m3, respectively. On the basis of the regression analysis, grouped data about the variables type of found1y, time-period and size of production were used in the final model to calculate cumulative dust exposures. However, the type of foundry was excluded as regards crystalline quartz. The calculated cumulative dust and quartz exposures averaged 23 mg/m3*year and 0.31mg/m3*year, respectively. The odds ratios, in the nested case-control study, showed no significant dose-response trends for both dust at1d ctystalline quartz: for exposure to dust, the odds ratios increased from 1 (reference) to 1.3 (95% CI 0.41-4.3), and to 2.2 (95% CI 0.73-6.5) by increasing cumulative exposures, as assessed in two different categories, and for exposure to quartz from I (reference) to 1.2 (95% Cl 0.37-3.9), and to 2.3 (95% CI 0.79-6.8).
Despite the moderate historical levels of total dust and crystalline quartz assessed for this cohort of aluminium foundry workers, based on a limited exposure data set, we found a tendency to a dose-response for lung cancer. Although data are limited, these findings give some support to the causative role of (quartz) dust in the previously observed high risk of lung cancer in this cohort.