Two retrospective cohort studies were conducted to assess the risk of cancer among workers exposed to chloroprene (2-chloro-1,3-butadiene) (CP). One is a study of incidence and mortality among 2314 production workers employed in the CP production plant in Yerevan, Armenia, between 1940 and 1988. The cohort was followed up for cancer incidence for the years 1979-1990 and for cancer mortality for 1979-1988. The second study is a mortality study among 5185 shoe manufacturing workers in Moscow who used polychloroprene latex and glue. Shoe workers were employed between 1940 and 1976, and followed from 1979 through 1993. The standardized incidence ratios (SIR) and standardized mortality ratios (SMR) were calculated using the Armenian and Moscow population as reference. An internal comparison analysis based on Poisson regression modeling was conducted. In the Yerenan cohort, incidence and mortality from all cancers were below expectation, but increased incidence (SIR, 3.27; 95% confidence interval (CI), 1.47-7.27), and mortality (SMR, 3.39; CI, 1.09-10.5) from liver cancer were noted. A dose-response relationship was suggested between the risk of liver cancer and indices of CP exposure. For the entire Moscow cohort, all-cause mortality was close to expectation and all-cancer mortality was increased. There was an increase in the mortality from liver cancer (SMR, 2.4; CI, 1.1-4.3), kidney cancer (SMR, 1.8; CI, 0.9-3.4), and leukemia (SMR, 1.9; CI, 1.0-3.3). Mortality from liver cancer and leukemia was associated with various indicators of CP exposure. A similar, although less consistent, pattern was found for kidney cancer. The association between CP exposure and risk of leukemia may be due to concomitant exposure to benzene. The results for liver cancer point towards a carcinogenic effect of CP.