Mortality and cancer incidence in New Zealand meat workers : a thesis presented in partial fulfilment of the requirements for the degree of Doctor of Philosophy in Epidemiology at Massey University, Wellington, New Zealand

Background: Several studies have suggested increased risks of cancers of the lung and lymphohaematopoietic tissue associated with work in the meat industry. The evidence for lung cancer is reasonably consistent, although few studies have controlled for smoking. Increased risks of lymphohaematopoietic cancers have been found consistently in case-control studies, including several conducted in New Zealand, but not in cohort studies. This project aimed to ascertain whether there is an increased risk of these cancers in workers employed in the New Zealand meat processing industry, and to identify what exposures are associated with any increased risks. Methods: Two cohorts, 4,064 individuals assembled from union records and 6,647 individuals assembled from company records, were followed from 1988 until 2000. Exposure status was assigned according to a job-exposure matrix. The observed number of deaths and cancer registrations was compared with expected numbers using five-year age-specific rates for the New Zealand population. Subgroup analyses evaluated the effect of duration of exposure to selected agents. Results: Vital status was determined for 93% (union) and 92% (company) of the total possible person-years. In the union cohort, mortality from all causes (SMR 0.86) and all cancers (SMR 0.88) were reduced, with no elevation observed for the cancers of a priori interest. Mortality from all causes (SMR 1.12) and all cancers (SMR 1.12) were elevated in the company cohort, with a significant excess of lung cancer (SMR 1.79) and an excess of non-Hodgkin's lymphoma (SMR 1.45). Subgroup analyses showed significant trends of increasing risk with duration of exposure to biological material. Conclusions: The union cohort exhibited a strong healthy worker effect, with no increase in mortality or cancer incidence. By contrast, excess risks for all cause and cancer mortality and incidence, and for lung and lymphohaematopoietic cancers, were observed in the company cohort. This is unlikely to be due to confounding by smoking, and the strong dose response relationship suggests the effect is related to occupational exposures