Cancer mortality of cadmium workers.

Cancer mortality of cadmium workers SIR-In their paper on cancer mortality in cadmium workers Elinder etal summarised the "mortality and incidence data" for cancer of the prostate in six stud- ies and the mortality data for lung cancer in five stud- ies.1 The latest follow up data were combined to yield 28 observed and 17-2 expected cases of prostate can- cer with an SMR of 162 and a two tailed p = 0-02. For lung cancer, there were 195 observed and 161-4 expected deaths with an SMR of 121 and a two tailed p = 0-008. Based on this pooling of results, the authors concluded that "there is a clear impression of an increased incidence of these cancers among cad- mium workers compared with the expected figures from the general population." Doll has drawn attention to an error committed by an international workshop on the carcinogenicity of metals in 1981.2 The data available at that time for prostatic cancer in four series of cadmium workers were combined to yield 14 observed versus 5-6 expected cases with a one tailed p = 0-002. The conclusion of the workshop was that "exposure to cadmium had contributed to the development of prostatic cancer." The inclusion of the first set of data must be regarded as hypothesis forming and should be excluded from subsequent analysis to check the unexpected finding of excess prostatic cancer in cad- mium workers. If the first British series3 with four