Analysis of cancer mortality and incidence among nuclear weapons workers at the Rocky Flats Plant

We studied a cohort of 16,258 production-era workers at Rocky Flats to assess relations between radiation and chemical exposures and lung cancer mortality, to estimate standardized mortality and incidence ratios, and to examine the effect of censoring on cancer incidence rates.; Multivariate Poisson regression modeling of the production-era cohort produced non-statistically significant elevated plutonium systemic deposition rate ratios for lung cancer. In this model, external penetrating radiation with a 10-year lag period, birth cohort, and employment duration were identified as important confounders. We then identified a cohort of plutonium-exposed workers and repeated the Poisson modeling. Lung cancer rate ratios for subjects with plutonium systemic depositions of ≥150 Bq where positively statistically significant when adjusted for external penetrating radiation with a 10-year lag period, birth cohort, and age at first plutonium exposure. Employment duration, although statistically significant in previous analyses, was not statistically significant in this model.; Standardized mortality ratios (SMRs) were calculated for both the production-era and plutonium exposure cohorts. Most estimates, including those for all cause mortality, all cancers, and lung cancer were significantly less than unity. Elevated, but non statistically significant SMRs were consistently observed for cancers of the rectum, brain and other central nervous system, and connective tissue, along with unspecified neoplasms of the nervous system and unspecified anemias.; Standardized incidence ratios (SIRs) were estimated, with results similar to those observed for the SMRs. Lung and prostate cancer incidence rates and SIRs were compared to lung and prostate cancer mortality rates and SMRs to determine the degree to which left and right censoring method affected SIRs. We found that left censoring methods produced the greatest influence on incidence rates and SIRs. However, these differences were less in the SIRs, which are adjusted for age and calendar time.