Exposure data There are five main sectors in the glass manufacturing industry: flat glass, container and pressed ware, art glass, special glass (e.g. optical, ophthalmic, electronic) and fibre glass (which is not considered here). The basic steps in the manufacture of glass products are melting, fining, homogenization, annealing and forming. Art and special glasses are produced by pot processes, involving manual batch handling. Art glass production has changed little with time and, for the most part, still involves blowing by mouth. During the twentieth century, the production of flat glass and container glass has evolved from traditional batch processes to highly automated processes. The modern production of flat glass is the most highly automated and usually utilizes tank melting with the continuous feeding of batch ingredients and the float (Pilkington) process for forming. The production of containers and pressed ware has also become increasingly mechanized, with mechanical blowing or pressing of the molten glass. Exposure to lead, arsenic and antimony oxides occurs primarily in sectors of the industry where traditional, non-mechanized techniques are used, such as in the production of crystal and other art glasses. Other potential exposures in glass manufacture include silica, asbestos, other metal oxides and polycyclic aromatic hydrocarbons. Human carcinogenicity data Four cohort studies of workers involved in glass manufacture - at a plant in Italy producing glass containers, among ceramics and glass workers in Austria, at two glass factories in Finland and among art glass-workers in Sweden - found increased risks for lung cancer. Population-based case-control studies in Sweden and Canada also found increased risks for lung cancer in glass-workers; a population-based case-control study in China found a significantly increased risk for lung cancer in female glass-workers and a nonsignificantly decreased risk in male glass, ceramics and enamelled product workers. None of the studies was specifically informative with respect to work in the flat-glass manufacturing industry. It is unlikely that the increased risk for lung cancer can be explained by nonoccupational risk factors such as smoking, in view of the consistency and magnitude of the findings, which were obtained in studies of various designs in different countries. When smoking habits were addressed in one of the studies, the estimated relative risk for lung cancer was increased. In general, no distinction was made in these studies between different components of the glass manufacturing industry. The only subgroup of glass-workers for whom specific findings were available was glass-blowers. Population-based case-control studies in Sweden on glass-workers and in Canada on people exposed to glass dust found small increased risks for stomach cancer, whereas in three cohort studies of glass-workers in Italy, Austria and Finland the risks for stomach cancer were not increased; in two of the cohort studies, the numbers of cases were small. Only the cohort study in Finland and the case-control study in Sweden specifically reported findings on stomach cancer in glass-blowers; both showed stronger increases in risk in glass-blowers than in glass-workers in general. The three cohort studies in Austria, Finland and Sweden showed little evidence of an increased risk for intestinal cancer. A Swedish population-based case-control study of colon cancer found a small increase in risk in glass-workers in general but a stronger increase in glass-blowers. Two population-based case-control studies, in Canada and the USA, showed nonsignificantly increased risks for urinary bladder cancer in glass-workers, but the numbers of cases were small. An Italian cohort study showed an increased risk for laryngeal cancer in glass-workers. In the Finnish cohort study, an increased risk was seen for basal-cell carcinomas of the skin in male workers. The evidence that favours a causal association between exposures in the glass manufacturing industry and cancer is: a reasonably consistent association with lung cancer in all four cohort studies; a similar though less consistent association with lung cancer in three case-control studies; a larger lung cancer risk than can reasonably be explained by nonoccupational confounding factors; the presence of human lung carcinogens in some components of the glass manufacturing industry; and the finding of an increased risk for stomach cancers in several cohort and case-control studies. Findings that limit the interpretation of causality include: the poorly characterized and heterogeneous exposures of workers in the glass manufacturing industry, which are likely to result in a weak or null association between exposure and cancer risk in some studies; the absence of demonstrated dose-response relationships; the fact that, in some studies, risk estimates were made for the combination of glass-workers and workers in other industries, thereby diminishing the degree to which results can be interpreted for the glass manufacturing industry itself; and the relatively few studies of workers in the glass manufacturing industry. Other relevant data A single study reported an increased frequency of chromosomal aberrations in peripheral blood lymphocytes of subjects working in a glass factory in the Czech Republic. Evaluation There is limited evidence that occupational exposures in the manufacture of art glass, glass containers and pressed ware are carcinogenic. There is inadequate evidence that occupational exposures in flat-glass and special glass manufacture are carcinogenic. Overall evaluations The manufacture of art glass, glass containers and pressed ware entails exposures that are probably carcinogenic to humans (Group 2A). Occupational exposures in flat-glass and special glass manufacture are not classifiable as to their carcinogenicity to humans (Group 3).