Leukemia, Lymphoma, and Multiple Myeloma Mortality in the Russian Mayak Worker Cohort 1948-2015.

Increased leukemia incidence or mortality is a well-known effect of acute radiation exposure. Less is known about the risks associated with protracted exposure, such as those arising in occupational exposure settings. We used excess relative risk models to investigate the strength of evidence for and the shape of the dose response for mortality from leukemia (excluding chronic lymphocytic leukemia, CLL) in the Russian Mayak Worker Cohort. The cohort includes 25,757 workers followed for cancer mortality from 1948 to 2015 who were subject to both external low-dose-rate gamma ray and internal exposures (from alpha particles emitted by inhaled plutonium). The red bone marrow external dose estimates were based on individual readings from film badges or TLD dosimeters. The mean external marrow dose was 584 mGy for workers hired before 1959 and 105 mGy for those hired between 1959 and 1982. Internal exposures were described using red bone marrow alpha-particle dose for workers who were subject to plutonium monitoring and potential plutonium exposure categories for unmonitored workers. The mean marrow dose from internal exposure for monitored workers was 2.1 mGy for those hired before 1959 and 0.16 mGy for those hired between 1959 and 1982. Radiation effects were described using both excess relative rate (ERR) and excess absolute rate (EAR) models. The excess relative rates for the 96 deaths from non-CLL leukemia were described using a time-since-exposure-dependent quadratic response in cumulative external low LET dose with effect modification by attained age. While the largest ERRs [2.45 at 1 Gy, 95% confidence interval (CI) 0.33 to 11.9] were associated with external doses received between 2 and 5 years before death, there was also a significant increase in rates for doses received 5 or more years before death (ERR 0.28 at 1 Gy, 95% CI 0.06 to 0.72) and an indication of increased rates associated with doses within two years of death (ERR 1.47 at 1 Gy, 95% CI 0.06 to 10.6). Uncertainties in these excess relative rate estimates from the primary models were adjusted for dose uncertainty. Excess absolute rate (EAR) models were also used to describe the leukemia death rates. The pattern of the EAR variation with time-since-exposure was like that for the ERR Doses received 2 to 4 years before death had the largest EAR (ERR 4.78 per 10,000 person years per Gy2, 94% CI 1.75 to 10.7) with increased rates for doses received within two years of death (3.66, 95% CI 9.26 to 11.2), and for doses received 5 or more years before death (0.34, 95% CI 0.11 to 0.7). However, while the ERR decreased with increasing attained age, there was no indication of an attained age dependence in the EAR. The external-exposure radiation-associated leukemia risk appeared to be largely from acute myeloid leukemia. There was no evidence of external exposure effects on the risks of death from chronic lymphocytic leukemia, lymphoma, or multiple myeloma. There was no evidence of internal exposure effects on the rates of leukemia or other lymphohematopoietic malignancies. These analyses extend earlier studies of leukemia mortality in the Mayak worker cohort, with additional years of follow-up, utilize the latest bone marrow dose estimates, and include an assessment of the effect of shared dose uncertainty on risk uncertainty. Our results show significant excess risk for non-CLL leukemia mortality with a complex interaction between attained age, time since exposure, and age at exposure. The highest risk per unit dose was associated with exposures received 2-5 years before the time at risk. In addition, for a given total cumulative dose, the risk decreases rapidly with increasing attained age. We discuss the differences between the patterns of risk related to acute exposures in the Life Span Study of survivors of the atomic bombings in Hiroshima and Nagasaki, and chronic exposure in the Mayak Worker Cohort as well as differences between risk estimates in our study and others involving prolonged low-dose external gamma ray exposure.