[Skip to Content]
[Skip to Content Landing]
October 2, 1987

Carcinogenicity of Ethylene Oxide

Author Affiliations

Colorado State University Fort Collins

Colorado State University Fort Collins

JAMA. 1987;258(13):1733. doi:10.1001/jama.1987.03400130047024

To the Editor.—  With regard to a recent communication by Hertz-Picciotto et al1 about ethylene oxide and leukemia, a few comments are offered. The premise of their comments is that exposure to ethylene oxide produces leukemia in both men and women, as evidenced by the study of Hogstedt et al.2 They proceed to evaluate how well the dose-response relationship observed in a single study of rats, if applied to humans, can predict the outcomes of two studies of humans, one positive2 and one negative.3 They conclude that because both predicted and observed numbers of leukemia deaths fall within the 95% confidence limits of predicted numbers, support for the carcinogenicity of ethylene oxide is demonstrated.The authors fail, however, to recognize the constraints placed on the proper interpretation of the data by Hogstedt et al in light of the many serious methodological problems embodied therein, which were