26
3 Health Effects of Radiation
3.3 Radiation Effects on the Body Following Chronic
Exposure to Radiation
The primary concerns with chronic exposure to relatively low levels of radiation are
that we will develop cancer. There are two primary competing hypotheses regarding
the risk of cancer induction from radiation exposure, the Linear, No-Threshold (LNT)
model and Threshold models.
The LNT hypothesis suggests that all radiation exposure is potentially harmful
(“no-threshold”), and that the risk of getting cancer from radiation is directly proportional to the dose received (“linear”). LNT is the most conservative radiation dose–
response model in that it predicts the highest risk from a given amount of radiation
exposure. This is one of the reasons that the LNT is the foundation of radiation regulations world-wide—since scientists are not certain as to how humans respond to
low levels of radiation exposure, it makes sense to control dose (and risk) according
to the most conservative model.
One concern with the LNT model is that it can be used to predict cancer risks down
to vanishingly small levels of exposure, and so it has been used to calculate expected
cancer rates from exposure to trivial levels of radiation. For example, say that the
risk of getting cancer from a given radiation exposure is five additional cancer deaths
for every 1 person-Sv. That means that exposing 100 people to 10 mSv each should
result in an extra five cancer deaths among those people. Or, exposing one million
people to 10 µSv each should also lead to five added cancer deaths. It’s easy to see
that we can use this model to predict added cancer deaths from any level of radiation
exposure, no matter how trivial, if enough people are exposed. By analogy, we can
also say that, since a 1000 kg rock will crush a person, throwing a million one-gram
rocks at a million different people will cause one or more deaths by crushing. In
reality, we know that this is not the case.
Because of this both the Health Physics Society and Roger Clarke, former head of
the International Commission on Radiation Protection (ICRP) have advised against
using the LNT model to predict risk at low levels of exposure. According to the
Health Physics Society [4] it is not scientifically appropriate to calculate a numerical
risk estimate from any exposure of less than 0.1 Sv. In a similar vein, the ICRP [5]
has suggested that, when looking at the risk from collective dose, if the most highly
exposed individual receives a trivial dose, then everyone’s dose should be treated as
trivial.
There are also studies that appear to show a threshold, below which no ill effects
from radiation exposure are noted; still other studies seem to indicate a reduction
in cancer from exposure to low levels of radiation—this latter effect is known as
hormesis. While there are epidemiological studies that appear to support each of
these models, the differences are not statistically significant, making it difficult to
claim that any specific hypothesis is correct. Thus, regulations and risk estimates
tend to be based on the LNT model as it predicts the highest risk for any given level
of exposure. These three models are shown in very simplified form in Fig. 3.3.
3 Health Effects of Radiation
3.3 Radiation Effects on the Body Following Chronic
Exposure to Radiation
The primary concerns with chronic exposure to relatively low levels of radiation are
that we will develop cancer. There are two primary competing hypotheses regarding
the risk of cancer induction from radiation exposure, the Linear, No-Threshold (LNT)
model and Threshold models.
The LNT hypothesis suggests that all radiation exposure is potentially harmful
(“no-threshold”), and that the risk of getting cancer from radiation is directly proportional to the dose received (“linear”). LNT is the most conservative radiation dose–
response model in that it predicts the highest risk from a given amount of radiation
exposure. This is one of the reasons that the LNT is the foundation of radiation regulations world-wide—since scientists are not certain as to how humans respond to
low levels of radiation exposure, it makes sense to control dose (and risk) according
to the most conservative model.
One concern with the LNT model is that it can be used to predict cancer risks down
to vanishingly small levels of exposure, and so it has been used to calculate expected
cancer rates from exposure to trivial levels of radiation. For example, say that the
risk of getting cancer from a given radiation exposure is five additional cancer deaths
for every 1 person-Sv. That means that exposing 100 people to 10 mSv each should
result in an extra five cancer deaths among those people. Or, exposing one million
people to 10 µSv each should also lead to five added cancer deaths. It’s easy to see
that we can use this model to predict added cancer deaths from any level of radiation
exposure, no matter how trivial, if enough people are exposed. By analogy, we can
also say that, since a 1000 kg rock will crush a person, throwing a million one-gram
rocks at a million different people will cause one or more deaths by crushing. In
reality, we know that this is not the case.
Because of this both the Health Physics Society and Roger Clarke, former head of
the International Commission on Radiation Protection (ICRP) have advised against
using the LNT model to predict risk at low levels of exposure. According to the
Health Physics Society [4] it is not scientifically appropriate to calculate a numerical
risk estimate from any exposure of less than 0.1 Sv. In a similar vein, the ICRP [5]
has suggested that, when looking at the risk from collective dose, if the most highly
exposed individual receives a trivial dose, then everyone’s dose should be treated as
trivial.
There are also studies that appear to show a threshold, below which no ill effects
from radiation exposure are noted; still other studies seem to indicate a reduction
in cancer from exposure to low levels of radiation—this latter effect is known as
hormesis. While there are epidemiological studies that appear to support each of
these models, the differences are not statistically significant, making it difficult to
claim that any specific hypothesis is correct. Thus, regulations and risk estimates
tend to be based on the LNT model as it predicts the highest risk for any given level
of exposure. These three models are shown in very simplified form in Fig. 3.3.