AbstractBackgroundWorkers employed at the Savannah River Site (SRS) were potentially exposed to a range of chemical and physical hazards, many of which are poorly characterized. We therefore compared the observed deaths among workers to expectations based upon death rates for referent populations.MethodsThe cohort included 18,883 SRS workers hired between 1950 and 1986. Vital status and cause of death information were ascertained through 2002. Sex‐specific standardized mortality ratios (SMR) were computed using U.S. and South Carolina mortality rates. SMRs were tabulated separately for monthly‐, weekly‐, and hourly‐paid men.ResultsMales had fewer deaths from all causes [SMR = 0.80, 90% confidence interval (CI): 0.78, 0.82], all cancers (SMR = 0.85, 90% CI: 0.81, 0.89), and lung cancer (SMR = 0.88, 90% CI: 0.82, 0.95) than expected based upon US mortality rates. The SMR for cancer of the pleura was 4.25 (90% CI: 1.99, 7.97) for men. The SMR for leukemia was greater than unity for monthly‐paid (SMR = 1.33, 90% CI: 0.88, 1.93) and hourly‐paid (SMR=1.36, 90% CI: 1.02, 1.78) men. Female workers had fewer deaths from all causes (SMR = 0.75, 90% CI: 0.69, 0.82) than expected, but more deaths than expected from cancer of the kidney (SMR = 2.58, 90% CI: 1.21, 4.84) and skin (SMR = 3.90, 90% CI: 2.11, 6.61).ConclusionsWhile the observed numbers of deaths in most categories of cause of death were less than expected, there are greater than expected numbers of deaths due to cancer of the pleura and leukemia, particularly among hourly‐paid male workers. It is plausible that occupational hazards, including asbestos and ionizing radiation, contribute to these excesses. Am. J. Ind. Med. 50:881–891, 2007. © 2007 Wiley‐Liss, Inc.