Although vinyl chloride is an established cause of liver angiosarcoma, the evidence is inconclusive on whether it also causes other neoplastic and nonneoplastic chronic liver diseases as well as neoplasms in other organs. Furthermore, the shape of the dose-response relation for angiosarcoma is uncertain. We have extended for approximately 8 years the mortality and cancer incidence follow-up of 12,700 male workers in the vinyl chloride industry in four European countries. All-cause mortality was lower than expected, whereas cancer mortality was close to expected. A total of 53 deaths from primary liver cancer (standardized mortality ratio 2.40, 95% confidence interval = 1.80-3.14) and 18 incident cases of liver cancer were identified, including 37 angiosarcomas, 10 hepatocellular carcinomas, and 24 liver cancers of other and unknown histology. In Poisson regression analyses we observed a marked exposure response for all liver cancers, angiosarcoma, and hepatocellular carcinoma. The exposure-response trend estimated for liver cancer in analyses restricted to cohort members with cumulative exposures of <1,500 parts per million-years was close to that estimated for the full cohort (relative risk of 2.0 per logarithmic unit of cumulative dose). No strong relation was observed between cumulative vinyl chloride exposure and other cancers. Although cirrhosis mortality was decreased overall, there was a trend with cumulative exposure.