Risk estimation and assessment

The manifestations of diseases and damages (e.g. cancer) caused by ionising radiation cannot be distinguished from so-called spontaneously occurring diseases. Possible causation by radiation can only be established when the diseases consistently occur statistically significantly more frequently in groups of individuals exposed to radiation and over different groups of individuals than in non-exposed control groups.

Epidemiological studies

Important epidemiological studies for the determination of the radiation-induced risk of cancer have been conducted in the following groups of individuals:

• Survivors of the atomic bomb explosions at Hiroshima and Nagasaki,
• Patients exposed to radiation for diagnostic and therapeutic purposes (e.g. the Canadian fluoroscopy cohort),
• Occupationally exposed persons (e.g. the Wismut uranium miners cohort),
• Residents in the surrounding areas of nuclear installations with high radioactive releases (Hanford (USA), Mayak (Russia)),
• Residents from the surrounding areas of damaged nuclear power plants (Chernobyl and Fukushima) and individuals employed in the cleanup,
• Individuals affected by above ground atomic bomb tests (e.g. veterans of the nuclear tests in Nevada, USA).

Estimating the risk of genetic effects

Genetic radiation effects have not been definitely demonstrated in human populations. To date, no increased rate of hereditary diseases in the descendants of the irradiated atomic bombs survivors in Hiroshima and Nagasaki in comparison with that for the general Japanese population has been determined. However, experimental studies on animals have shown that radiation can induce genetic alterations, so-called "mutations" in germ cells. Hence, the estimates for the genetic radiation risk for humans stem from these experimental studies.

Estimating the risk of leukaemia and cancer

The estimates for the radiation-induced risk of leukaemia and cancer are mainly based on the evaluation of the data of Japanese atomic bomb survivors. This group was exposed to a high dose rate (the entire dose in a split second), however, the dose was only high in a small percentage of those affected.

From the 105,000 individuals included in the analysis from 2007, 35,000 individuals received doses between 5 and 200 mGy. Less than 3 % of the atomic bomb survivors have received doses of more than one Gy. This shows that the atomic bomb survivors are not a highly exposed group of individuals. Nevertheless, it is important to note that – unlike in the exposure scenarios relevant today - they received their doses within a very short period (in a split second) and were thus exposed to high dose rates.

With atomic bomb survivors there are two types of investigations. Whereas the first type focuses on the causes of death (so-called mortality follow-up), the second type investigates newly diagnosed diseases (incidence follow-up). For current issues the findings from the incidence follow-up are more relevant. In the group of the 105,000 survivors in Hiroshima and Nagasaki, 850 additional radiation-induced cases of solid tumors were observed until the end of 1998. Added to that were about 85 deaths due to leukaemia which could only be determined from the approx. 85,000 individuals included in the mortality follow-up. Further, the lifetime risk to develop cancer is of interest. The data available on the observation period have to be extrapolated to the lifetime to determine the lifetime radiation risk.

For relatively low radiation exposure levels, as they are determined today in the environment and workplace, a further extrapolation of the results from the Japanese atomic bomb survivors is needed. If the effect of small doses or low dose rates is to be estimated, the epidemiological results, which are mainly available for high dose rates, have to be converted to exposure situations with low doses and chronic exposure. Furthermore, the rate of so-called "spontaneous" cancer and leukaemia diseases in a particular population observed plays an important role for being able to estimate the additional radiation-induced risk.

The extrapolations to lifetime risk have to consider the relevant low dose range and the spontaneous cancer rate within a population. They are therefore subject to estimation uncertainties. They are based on specific model assumptions, among other things on assumptions on cancer development and on the dose-response relationship. They also include assessments by panels of experts. It is necessary to verify these assumptions regularly and to adjust risk estimates to the respective current level of knowledge.

Assessment of the scientific evidence on radiation effects

Based on the data of the atomic bomb survivors of Hiroshima and Nagasaki evaluated so far, the United Nations Scientific Committee on the Effects of Atomic Radiation (UNSCEAR) estimates the lifetime risk coefficient for radiation induced fatal leukaemias and cancer at about 7 % per Sv for a population of all ages. This means that, if 1,000 individuals are each exposed to 100 mSv then it can be expected that seven individuals will additionally die from cancer or leukaemia (source: UNSCEAR 2006). According to statistical results about 250 individuals out of 1,000 die from cancer or leukaemia without radiation exposure. The ICRP estimates the radiation-induced cancer risk after radiation exposures at low dose rates (such as those occurring in everyday life and in the workplace) at 5.5 % per Sv for the total population and at 4.1 % per Sv for adults (source: ICRP 103, 2007).

On the basic assumption that low doses and chronic exposures are less effective than high doses and acute exposures (given the same total dose), the risk coefficients for radiation protection in the low-dose range have been divided by the factor two in accordance with the recommendations of the ICRP on dose limit values. The intention behind this is to particularly take into account the repair and recovery capacity of irradiated cells at low values of the dose and dose rate. The reduction does not result directly from the observational data for cancer in humans and is based on model assumptions founded on findings from laboratory experiments. The BfS considers the scientific justification for this reduction of the risk coefficients for low doses and chronic exposures insufficient.

Risk assessment

Risk assessment is an overall attempt to ensure realistic risk estimates supported by international and national panels of experts. Safety considerations are not included at this level of risk assessment.

It is currently not possible to make reliable estimates for the risk of developing diseases other than cancer and leukaemia after radiation exposure. However, the evaluations of atomic bombings survivors in Japan, exposed population groups in the former Soviet Union and radiotherapy patients indicate that radiation-induced cardiovascular diseases may also occur more frequently. The assumption that cataracts (clouding of the lens in the eye) belong to the deterministic radiation effects is currently being called into question. Here also, new evidence has emerged indicating that cataracts already occur at tenfold lower doses than until recently assumed (0.5 Gy compared to 5 Gy). There is an ongoing debate that there might be no threshold dose for these diseases and that they should be regarded as stochastic radiation effects like malignant neoplasms.

Reports about the effects of radiation on groups of individuals who have been exposed to low radiation doses can be found in scientific literature. For example observations of an increased occurrence of leukaemias, cancer, hereditary diseases or damage to the unborn child (teratogenic damage) are reported. Some studies refer to individuals living in the surrounding areas of nuclear installations, to individuals who have been exposed to radiation for medical purposes or to individuals living in areas of the Earth with relatively high natural radioactivity. The vast majority of these studies are so-called ecological studies which use data aggregated over time or space. These studies are considered less valid as compared to studies with individual person data on health and exposure. Many of these studies additionally have methodological weaknesses, such as small sample sizes, lack of appropriate controls, insufficient dosimetry, unconsidered influence of further risk factors other than radiation and social factors. Furthermore, there is a tendency to rather report or publish study findings when they indicate an increased risk (publication bias).