The mortality of workers involved in the production of stainless and alloyed steel from 1968 to 1992 was studied, in order to investigate the risk of lung cancer due to exposure to metals, i.e. iron oxides, chromium and/or nickel compounds. The study design was a historical cohort mortality study and a nested case-control study concerning lung cancer. Standardized mortality ratios (SMRs) were computed using regional mortality rates as an external reference for comparing observed and expected numbers of deaths, adjusting for age, sex and calendar time. Conditional logistic regression was used to estimate odds ratios (ORs). Occupational exposure was assessed through the complete job histories of cases and controls and a specific job-exposure matrix. The cohort comprised 4,288 male and 609 female workers. The observed overall mortality was significantly lower than expected [649 deaths; SMR = 0.91; 95% confidence interval (CI) 0.84-0.98]. No significant SMR was observed for mortality from lung cancer (54 deaths; SMR = 1.19; CI 0.88-1.55). The case-control study was based on 54 cases and 162 individually matched controls. Smoking habits were available for 71%. No lung cancer excess was observed for exposure to (1) metals and/or their compounds, i.e. iron (OR = 0.94, CI 0.48-1.86), chromium and/or nickel (OR = 1.18, CI 0.62-2.25), and cobalt (OR =0.64, CI 0.33-1.25), (2) acid mists (OR = 0.43, CI 0.17-1.10), and (3) asbestos (OR = 1.00, CI 0.54-1.86). With respect to exposure to polycyclic aromatic hydrocarbons (PAHs) and silica, which are often found together in workplaces, (1) high and statistically significant lung cancer excesses were observed, the ORs being 1.95 (CI 1.03-3.72) and 2.47 (CI 1.28-4.77) respectively, (2) quantitative exposure parameters revealed upward trends reaching statistical significance (P < 0.05), and (3) adjustments for tobacco consumption did not reveal any confounding factors from smoking. This study failed to detect any relationship between lung cancer and exposure to iron, chromium, nickel and/or their compounds. High and statistically significant relative risks, along with increasing trends, were observed for simultaneous exposure to PAHs and silica.