Radiation epidemiology and health effects following low-level radiation exposure

Abstract

Radiation epidemiology is the study of human disease following radiation exposure to populations. Epidemiologic studies of radiation-exposed populations have been conducted for nearly 100 years, starting with the radium dial painters in the 1920s and most recently with large-scale studies of radiation workers. As radiation epidemiology has become increasingly sophisticated it is used for setting radiation protection standards as well as to guide the compensation programmes in place for nuclear weapons workers, nuclear weapons test participants, and other occupationally exposed workers in the United States and elsewhere. It is known with high assurance that radiation effects at levels above 100–150 mGy can be detected as evidenced in multiple population studies conducted around the world. The challenge for radiation epidemiology is evaluating the effects at low doses, below about 100 mGy of low-linear energy transfer radiation, and assessing the risks following low dose-rate exposures over years. The weakness of radiation epidemiology in directly studying low dose and low dose-rate exposures is that the signal, i.e. the excess numbers of cancers associated with low-level radiation exposure, is so very small that it cannot be seen against the very high background occurrence of cancer in the population, i.e. a lifetime risk of incidence reaching up to about 38% (i.e. 1 in 3 persons will develop a cancer in their lifetime). Thus, extrapolation models are used for the management of risk at low doses and low dose rates, but having adequate information from low dose and low dose-rate studies would be highly desirable. An overview of recently conducted radiation epidemiologic studies which evaluate risk following low-level radiation exposures is presented. Future improvements in risk assessment for radiation protection may come from increasingly informative epidemiologic studies, combined with mechanistic radiobiologic understanding of adverse outcome pathways, with both incorporated into biologically based models.