Leukaemia after cervical cancer irradiation in Denmark.

A matched case-control study among cervical cancer patients in Denmark was conducted to evaluate the possible influence of radiation therapy on leukaemia risk. Among 25718 women treated between 1943-1977, 27 cases of leukaemia occurred one or more years after diagnosis. Four controls were matched to each case on age, calendar year of diagnosis and survival without developing a second cancer. Detailed medical records were abstracted, and individual radiation doses to active bone marrow were determined by medical physicists. All leukaemia diagnoses were classified by one haematopathologist. A 2.5-fold risk of radiogenic leukaemia was found, although the excess was not statistically significant (95% Cl = 0.5-12.1). The pattern of risk by dose was somewhat compatible with animal experiments, ie, the risk rose, peaked, and then decreased to near normal levels at the highest dose categories. The average dose to the total bone marrow was 783 rads, and thus a much greater risk of leukaemia was predicted on the basis of estimates from other exposed populations. Assuming that the pelvic bone marrow is essentially destroyed by high-dose radiotherapy and should, perhaps, be disregarded for the purpose of risk estimation, risk coefficients consistent with other studies can be computed. Nonetheless, the very low risk of radiogenic leukaemia found among cervical cancer patients suggests that the induction of leukaemia by radiation is a complex function of dose, and that very high doses may result in a decreased risk, possibly associated with cellular sterilization or other mechanisms.