Abstract
Background
Analyses of Japanese Abomb survivors' cancer mortality risks are used to establish recommended annual dose limits, currently set at 1 mSv (public) and 20 mSv (occupational). Do radiation doses below 20 mSv have significant impact on cancer mortality in Japanese Abomb survivors, and is the doseresponse linear?
Methods
I analyse stomach, liver, lung, colon, uterus, and allsolid cancer mortality in the 0 – 20 mSv colon dose subcohort of the 1950–90 (grouped) mortality cohort, by Poisson regression using a timelagged colon dose to detect latency, while controlling for gender, attained age, and ageatexposure. I compare linear and nonlinear models, including one adapted from the cellular bystander effect for α particles.
Results
With a lagged linear model, Excess Relative Risk (ERR) for the liver and allsolid cancers is significantly positive and several orders of magnitude above extrapolations from the Life Span Study Report 12 analysis of the full cohort. Nonlinear models are strongly superior to the linear model for the stomach (latency 11.89 years), liver (36.90), lung (13.60) and allsolid (43.86) in fitting the 0 – 20 mSv data and show significant positive ERR at 0.25 mSv and 10 mSv lagged dose. The slope of the doseresponse near zero is several orders of magnitude above the slope at high doses.
Conclusion
The standard linear model applied to the full 1950–90 cohort greatly underestimates the risks at low doses, which are significant when the 0 – 20 mSv subcohort is modelled with latency. Nonlinear models give a much better fit and are compatible with a bystander effect.
Background
Studies of the Japanese Abomb survivors cohort have been the key source of radiation risk estimates used to establish environmental and occupational protection standards [1,2] despite the differences between instantaneous exposure in war to gamma and neutron irradiation, and chronic low dose occupational or public exposure in general or to alpha particles. There is an unresolved debate [3] on whether Excess Relative Risk (ERR) increases linearly with dose across a very large range of exposures including the low dose range.
Currently the International Commission on Radiological Protection (ICRP) recommends an annual occupational dose limit of 20 mSv (whole body dose), and many researchers believe that doses below this level have little or no impact on human health. The ICRP risk estimates are strongly based on analyses of the Japanese data for the full dose range assuming a linear dose response. On the other hand in the Life Span Study Report 12 (henceforth LSS12) Pierce et al. [4] adopted a linear ERR model for the Japanese data but found some evidence of a nonlinear convex doseresponse at lower doses for the category of allsolid cancers combined.
Around one third of the data in the 1950–90 mortality cohort analysed in LSS12 concerns persons who received 0 – 20 mSv instantaneous exposure in 1945. There is no reason to assume any particular relation between dose response in this region and across the entire followup. Instead, the low dose data may be analysed in its own right with or without assuming a linear dose response.
In experiments at cell level, low dose radiation induces significant and nonlinear effects on micronuclei, chromosome rearrangements and instability, point mutations, and changes in cell cycle protein levels. Studies reviewed by Mothersill and Seymour [5] show low doses have large effects on cells whose nuclei, or even the affected cells themselves, need not be directly traversed by radiation (bystander effects). Some studies find responses rising rapidly with dose and then reaching a plateau. Whilst these early nonlinear cellular events may be critical for tumour formation, the human body's defences may also greatly alter the initial dose response or dilute its significance, if the analysed outcome is a cancerous tumour emerging decades later and eventually identified as cause of death.
I analyse the five leading cancer sites (stomach, liver, lung, colon, and uterus) and the grouped category of allsolid cancers in the 0 – 20 mSv portion of the 1950 – 90 Japanese mortality cohort. Followup began 5 years after exposure but rather than assume this lag to be optimal, latency is estimated for individual cancers by fitting models which include a variable latent period. I test linearity in the dose response by modelling with two nested nonlinear models, each of which contains the linear model.
Methods
Life Span Study mortality cohort data (1950–1990) were obtained from the Radiation Effects Research Foundation [6] via the Comprehensive Epidemiologic Data Resource [7]. The file r12canc.dat includes anonymous information on a cohort of 86 572 survivors, presented as 16 612 grouped data cells crosstabulated by city, sex, total shielded kerma, exposure age and attained age categories, and dose category. Cell data includes a weighted adjusted colon dose, representing an average of individual instantaneous ("flash") doses weighted by personyears within the cell (and assuming neutron Relative Biological Effectiveness RBE = 10). Mean ageatexposure, mean attained age, and deaths from specific cancers or grouped cancer categories are also shown for each cell. Liver cancer (ICD9 155(0,1,2)) includes both deaths attributed to primary liver cancer, and deaths which were not specified as primary or secondary liver cancer. Cancer of the uterus refers to ICD9 180–182, including cervical cancer (ICD9 180).
The data uses the DS86 dosimetry [8], but a limited comparison with the current DS02 dosimetry [911] is undertaken.
Subcohorts are defined by restricting weighted adjusted colon dose as given in the cell data. The 0 – 20 mSv subcohort has 3011 cells comprising 1690391.75 personyears (py) observation. This subcohort consists of exactly those cells whose individual members received no more than 20 mSv adjusted colon dose (3009 cells have dose category 1 or 2, while two other cells each contain a single person at risk whose adjusted colon dose was 20 mSv). Other subcohorts defined below lack such simple interpretation.
Specified cancer deaths in the i^{th }cell are assumed to be Poisson distributed with expected value λ_{i}T_{i }where T_{i }is py observation and λ_{i }depends on control variables and a lagged dose D_{φ }defined for each cell as weighted adjusted colon dose (in 10 mSv units) if TimeSinceExposure ≥ φ and 0 otherwise. TimeSinceExposure = mean attained age – mean age at exposure.
The models below assume
λ = λ_{0 }(1 + ERR)
where y_{j }are control variables while ERR depends on D_{φ }and parameters θ_{k }(including φ) but not on control variables or their parameters.
Gender, log mean attained age, and 14 indicator variables defined by 0 ≤ ageatexposure < 5, 5 ≤ ageatexposure < 10,... and 65 ≤ ageatexposure < 70 were used as controls. Indicator variables defined from mean ageatexposure (cell data) are identical to those defined from the ageatexposure stratification of the dataset. An indicator defined by 70 ≤ ageatexposure would be redundant. Log mean attained age was an adequate alternative to the use of attained age categories.
Four models were fitted:
Linear ERR = βD_{φ}
Transient ERR = σD_{φ }[] with τ ≥ 0
Twophase ERR =βD_{φ }+ σD_{φ }[] with τ ≥ 0
Category ERR = β_{1}(dosecat1) + β_{2}(dosecat2) + β_{3}(dosecat3)
The usual Linear model is used here with lagged dose. The Control model (ERR = 0) is nested when β = 0.
The Transient model gives an asymptotic linear dose response at very low doses, decaying to zero at higher doses if τ > 0. The Linear model is nested when τ = 0, while σ = 0 nests the Control model.
The Twophase model gives an asymptotic linear dose response at very low doses, and a possibly different asymptotic linear response at higher doses, with an exponential transition between the two. The Transient model is nested when β = 0, while σ = 0 (or τ = 0) nests the Linear model. The Control model is nested when β = σ = 0, or when τ = 0 and β + σ = 0. ERR ~ βD_{φ }when τ D_{φ }>> 1, while ERR ~ (β + σ)D_{φ }when τ D_{φ }<< 1. The ratio of asymptotic slopes is R = 1 + σ/β.
Typical doseresponse curves for these three models are shown in Figure 1.
Figure 1. Dose response prototypes for linear, transient, and twophase models. Typical dose response curves for the linear model ERR = βD where ERR is Excess Relative Risk and D is a dose variable (arbitrary units), the transient model ERR = σDe^{(τD)}, and the twophase model ERR = βD + σDe^{(τD)}. Here β, σ, τ are parameters with τ ≥ 0. The linear model is shown with β > 0. Two versions of the transient model are shown, with σ > 0 and with σ < 0. The transient model converges to 0 as dose → ∞. Two versions of the twophase model are shown, with σ >> β > 0 and σ >> β > 0; the first version occurs when β > 0 and σ/β > e^{2 }~ 7.39. Other possibilities (not shown) lack local maxima and minima, but the twophase model always converges to the linear model as dose → ∞.
The Category model uses indicator variables defined by cutpoints 0 < γ_{1 }< γ_{2}
dosecat1 = 1 if 0 < D_{φ }≤ γ_{1 }and = 0 otherwise
dosecat2 = 1 if γ_{1 }< D_{φ }≤ γ_{2 }and = 0 otherwise
dosecat3 = 1 if γ_{2 }< D_{φ }and = 0 otherwise
The Control model is nested by β_{1 }= β_{2 }= β_{3 }= 0. D_{φ }= 0 also defines the baseline. The model is applied here only with latencies obtained from the twophase model, and cutpoints γ_{1 }= 0.025 (0.25 mSv), γ_{2 }= 0.05 (0.5 mSv) which assign roughly equal py to dosecat1, dosecat2, and dosecat3.
When φ = 5 years, dosecat1 specifies 699 cells (620850.09 py). Hiroshima data is contained in 544 cells (573199.23 py), while 155 cells (47650.86 py) contain Nagasaki data. Mean attained age ranges from 8.16 years to 96.84 years, and mean ageatexposure ranges from 1 to 82.99 years. All individual doses within these cells are below 5 mSv. The weighted adjusted colon dose for the cell ranges from 0.004 mSv to 0.25 mSv. As φ increases, cells move to the baseline. For example at φ = 36.9 years, dosecat1 specifies 156 cells (104592.3 py) of which 115 cells (95871.94 py) refer to Hiroshima and 41 cells (8720.36 py) to Nagasaki.
Likewise when φ = 5 years, dosecat2 specifies 560 cells with 519832.84 py, of which 109 cells (95907.76 py) contain Hiroshima data and 451 cells (423925.08 py) contain Nagasaki data. Mean attained age ranges from 8.21 years to 93.32 years and mean ageatexposure ranges from 0.92 to 84.71 years. All individual doses within these cells are below 5 mSv. The weighted adjusted colon dose for the cell ranges from 0.25 mSv to 0.49 mSv. At φ = 36.9 years, dosecat2 specifies 115 cells (97536.25 py) of which 28 cells (21631.85 py) refer to Hiroshima and 87 cells (75904.4 py) to Nagasaki.
Finally when φ = 5 years, dosecat3 specifies 1579 cells with 548454.05 py, of which 780 cells (359098.67 py) contain Hiroshima data and 799 cells (189355.38 py) contain Nagasaki data. Mean attained age ranges from 8.12 years to 102.2 years and mean ageatexposure ranges from 0.93 to 91.56 years. All individual doses within these cells are below 20 mSv. The weighted adjusted colon dose for the cell ranges from 0.5 mSv to 20 mSv. At φ = 36.9 years, dosecat3 specifies 353 cells (97629.42 py) of which 176 cells (63005.21 py) refer to Hiroshima and 177 cells (34624.21 py) to Nagasaki.
Note that whilst the dosecat variables are well defined on the cells, they have no simple interpretation in terms of individual doses within those cells. A cell with D_{φ }≤ 0.025 (0.25 mSv) may contain individuals with lagged doses up to 5 mSv; it is only the weighted average which is constrained below 0.25 mSv.
Each model is also constrained by λ = λ_{0}(1 + ERR) ≥ 0 (equivalently ERR ≥ 1) in all cells, as the Poisson distribution is undefined if λ_{i}T_{i }< 0.
Poisson regression [12] is based on minimising the model Deviance, or equivalently minimising Σ_{i }[E_{i } O_{i}ln(E_{i})] where E_{i }and O_{i }are the expected and observed values.in each cell. While O_{i }is given by the data, E_{i }depends on the model and its parameter values. The data were placed on a spreadsheet and the Excel tool Solver (NewtonRaphson iteration) was used to optimise the parameters subject to defined constraints.
With the control model, the minimal Deviance is independent of φ. Once φ is fixed, the linear model has a unique minimum Deviance. With other models, local minima were compared to find the absolute minimum DevMod_{φ}. Searches were conducted by partitioning the τ axis as [0,1], [1,2]... [2^{j1 },2^{j}]... [2^{8},2^{9}], [2^{9},∞) and beginning iteration from (C, 0, 0, τ_{j}) where C represents the control model Maximum Likelihood Estimate parameter values, β and σ are initially set to 0 (β is constrained to 0 for the transient model); while τ_{j }is constrained to the j^{th }interval.
If model I is nested within model J by k parameter constraints, the likelihood ratio test for comparing the two models at φ, LRT_{JI,φ }= DevMod_{I,φ }– DevMod_{J,φ }is approximately χ^{2 }distributed on k df. At a given φ, LRT_{lincon }and LRT_{transcon }for comparing the linear and transient models with the control model each have 1 df, while LRT_{2pcon }for comparing the twophase model with the control model has 2 df. Note that LRT_{2ptrans}, LRT_{2plin}, and LRT_{translin }have 1 df as each nested pair is defined by a single parameter constraint, and that LRT_{2pcon }= LRT_{2plin }+ LRT_{lincon}.
For each model J (linear, transient, and twophase), I found best fits by choosing φ = φ_{m }to maximise LRT_{Jcon,φ }(equivalently to minimise DevMod_{J,φ}) either absolutely or with the constraint that ERR at D_{φ }= 0.025 be ≥ 0. The value of φ_{m }may vary with the model.
Profile Likelihood Confidence Intervals for ERR were computed at specified values of D_{φ}. For example, in the twophase model ERR depends on D_{φ }and T = (β,σ,τ) whilst Deviance depends on (φ,C,T) where C is the vector of control parameters. With φ fixed, let V_{φ }be the set of T which can be ruled out with 95% confidence, as all choices of C give Dev(φ,C,T) – DevMod_{φ }> K, where K ~ 7.8147 satisfies χ_{3}^{2}(K) = 0.95. Define Min_{φ }(respectively Max_{φ}) as the largest v (smallest w) for which the constraint ERR(D_{φ},T) < v (> w) implies T ∈ V_{φ}. The interval (Min_{φ}, Max_{φ}) is taken as a 95% CI for ERR(D_{φ}) in the twophase model.
Results
For each cancer site and for φ = 5, 6, ... 44, models were fitted to the 0 – 20 mSv male/female (M/F) data.
For the linear model only, I estimated ERR_{0.025,φ }by fitting the model to the restricted subcohort of cells with mean weighted adjusted colon dose ≤ 0.5 mSv; ERR_{1,φ }was obtained from fitting the 0 – 20 mSv data, as with all other modelling.
More precise optimal latencies φ_{m}, 95% CI's for and , and GoodnessofFit (GoF) statistics were determined for the nested models, including separate male and female results. The category model (M/F) was then fitted using the optimal latencies from the twophase model.
Results are now outlined for specific cancers (M/F).
Stomach
The linear model is never significant against the control model, but Figure 2 suggests two distinct latency periods: 5 ≤ φ ≤ 21, and 29 ≤ φ ≤ 41. Outside these periods, the nested models show no significant improvement on the control model. In both periods, the twophase and transient models improve on the control model. The twophase model significantly improves on the transient and linear models when 5 ≤ φ ≤ 21 and on the linear model when 29 ≤ φ ≤ 41. For the twophase model, ERR_{0.025,φ }and ERR_{1,φ }are both positive when 5 ≤ φ ≤ 21, and negative when 29 ≤ φ ≤ 41 (Figure 3).
Figure 2. Stomach cancer mortality in 0 – 20 mSv subcohort, LRT vs latency. LRT is the likelihood ratio test for comparison of the indicated pair of models evaluated at fixed latency φ, and is χ^{2 }distributed on 1 d.f. (red line shows LRT_{2plin }for comparing the twophase and linear models, green line shows LRT_{translin }for comparing transient and linear models, blue line shows LRT_{lincon }for comparing linear and control models) LRT values above 3.84 are significant at p = 0.05, above 6.63 at p = 0.01, above 10.83 at p = 0.001. Other LRT can be derived at each φ value, e.g. LRT_{2pcon }= LRT_{2plin }+ LRT_{lincon }and LRT_{transcon }= LRT_{translin }+ LRT_{lincon }while LRT_{2ptrans }= LRT_{2plin } LRT_{translin}. Latency φ (in years) is used in defining the lagged dose D_{φ }= DS86 weighted adjusted colon dose if TimeSinceExposure ≥ φ, D_{φ }= 0 otherwise. Models specify the Excess Relative Risk ERR as a function of D_{φ }while controlling for gender, log attained age, and ageatexposure categories. Control: ERR = 0 Linear: ERR = βD_{φ}; control model nested by β = 0. Transient: ERR = σD_{φ }e^{(τDφ) }with τ ≥ 0; linear model nested by τ = 0. Twophase: ERR = βD_{φ }+ σD_{φ }e^{(τDφ) }with τ ≥ 0; linear model nested by σ = 0; transient model nested by β = 0.
Figure 3. Stomach cancer mortality in 0 – 20 mSv subcohort, ERR vs latency. For each model (twophase in red, transient in green, linear in blue) fitted at fixed latency φ, series 1 (solid lines) shows the Excess Relative Risk ERR_{1 }(ERR at 10 mSv lagged dose); series 025 (dotted lines) shows ERR_{0.025 }(ERR at 0.25 mSv lagged dose). For latency φ and model definitions see Figure 2. For Linear 025 only, ERR_{0.025 }is estimated from the 0 – 0.5 mSv data.
For the linear model (Table 1) φ_{m }= 6.66 is optimal in the region of ERR ≥ 0, but the model is not significant against the control model. For the transient model (Table 2) φ_{m}= 11.85 is optimal within the range for which ERR_{0.025,φ }≥ 0, but ERR is not significantly positive. The optimal latency for the twophase model (Table 3) is φ_{m }= 11.89, when comparison with the control model has LRT_{2pcon }= 21.190 (2 df), and comparisons with the transient and linear models have LRT_{2ptrans }= 15.684 (1 df) and LRT_{2plin }= 19.643 (1 df). Each LRT is highly significant. The twophase model gives ERR_{0.025,φ }= 0.391 with 95%CI (0.077, 0.857) while ERR_{1,φ }= 0.459 (0.113, 0.942). When D_{φ }= Dmax = 0.012, ERR attains a local maximum of 0.539 (0.171, 1.037).
Applying the category model (Table 4) with φ = 11.89 gives LRT_{catcon }= 21.590 (3 df) and both β_{1 }and β_{3 }are significantly positive.
Table 4. Category model^{a}
The models pass GoodnessofFit tests with Dev, but Pearson ChiSquare exceeds df due to one case in a cell with T = 0.4 py. The twophase model is preferable to its nested alternatives by LRT comparisons, and clearly distinguishes the latency periods as regions of positive and negative ERR. At the optimal latency, R = 1 + σ/β= 262.62.
Liver
There are two distinct latency periods outside which the models are insignificant (Figure 4). When 10 ≤ φ ≤ 23 the twophase model improves on the transient, linear, and control models, and ERR_{0.025,φ }and ERR_{1,φ }are negative in all three models (Figure 5). When 32 ≤ φ ≤ 41 the twophase model improves on the transient, linear, and control models, and ERR_{0.025,φ }and ERR_{1,φ }are nonnegative in all three models. When 34 ≤ φ ≤ 43 the transient and linear models each improve on the control model with positive ERR, and when 34 ≤ φ ≤ 38 the transient improves on the linear model.
Figure 4. Liver cancer mortality in 0 – 20 mSv subcohort, LRT vs latency. Axes and model definitions as in Figure 2.
Figure 5. Liver cancer mortality in 0 – 20 mSv subcohort, ERR vs latency. Axes and model definitions as in Figure 3.
The optimal latencies are 38.58 (linear), 36.90 (transient), and 36.90 if the twophase model is optimised with ERR_{0.025,φ }≥ 0. At φ_{m }= 36.90 the twophase model has LRT_{2pcon }= 34.874 while LRT_{2ptrans }= 16.755 and LRT_{2plin }= 27.642, all highly significant comparisons. At the optimal latency, the twophase model gives ERR_{0.025,φ }= 1.099 (0.264, 2.374) while ERR_{1,φ }= 1.428 (0.481, 2.954). The local maximum ERR occurs when D_{φ }= 0.013. Estimates from the transient and linear models are comparable.
Applying the category model with φ = 36.90 gives LRT_{catcon }= 33.844 and β_{1}, β_{2}, and β_{3 }are all significantly positive. If "city" is included as a control, LRT_{catcon }= 34.939, β_{1 }= 1.35 (0.72, 2.23), β_{2 }= 0.73 (0.14, 1.60), β_{3 }= 1.39 (0.70, 2.34).
The models pass both GoF tests. The twophase model is preferable to its nested alternatives by LRT and gives wider periods of significant improvement on the control model, distinguished as regions of positive and negative ERR. At the optimal latency R = 204.11.
Lung
The linear model is never significant against the control model (Figure 6). The transient improves on the linear and control models when 5 ≤ φ ≤ 38. The twophase improves on the linear and control models when 5 ≤ φ ≤ 21 and when 24 ≤ φ ≤ 38, and on the transient model when 32 ≤ φ ≤ 38. When 5 ≤ φ ≤ 21 the transient and linear models give positive ERR_{0.025,φ }and ERR_{1,φ }(Figure 7) as does the twophase when 7 ≤ φ ≤ 21. When 22 ≤ φ ≤ 41 ERR_{0.025,φ }and ERR_{1,φ }are negative in the twophase and transient models.
Figure 6. Lung cancer mortality in 0 – 20 mSv subcohort, LRT vs latency. Axes and model definitions as in Figure 2.
Figure 7. Lung cancer mortality in 0 – 20 mSv subcohort, ERR vs latency. Axes and model definitions as in Figure 3.
Optimal latencies are 16.90 (linear), 13.60 (transient), and 13.60 if the twophase model is optimised with positive ERR_{0.025}. At φ_{m }= 13.60 the transient model has LRT_{transcon }= 15.701 while LRT_{translin }= 14.394, both highly significant. At this latency the transient model gives ERR_{0.025,φ }= 0.790 (0.195, 2.006) while ERR_{1,φ }= 0.796 (0.189, 1.985). The twophase model does not improve the transient and gives comparable ERR with somewhat wider CI's. It does improve on the linear model with LRT_{2plin }= 14.734 and gives R = 86.39.
Applying the category model with φ = 13.60 gives LRT_{catcon }= 19.304 and β_{1}, β_{2}, and β_{3 }are all significantly positive.
The 3 nested models pass both GoF tests. The transient model is preferable here.
Colon
None of the models give significant positive response for the M/F data (Figure 8, Figure 9). The transient model has occasional weak negative response.
Figure 8. Colon cancer mortality in 0 – 20 mSv subcohort, LRT vs latency. Axes and model definitions as in Figure 2.
Figure 9. Colon cancer mortality in 0 – 20 mSv subcohort, ERR vs latency. Axes and model definitions as in Figure 3.
Uterus
The linear model is never significant against the control model (Figure 10). When 12 ≤ φ ≤ 16 the transient and twophase models both improve on the control model, and ERR is negative in all three models (Figure 11). When 24 ≤ φ ≤ 33 ERR_{0.025,φ }and ERR_{1,φ }are positive in the twophase and transient models. At the optimal latency φ = 26.91 (in the region of positive ERR) both the twophase and transient models are weakly significant against the linear and control models. The 95%CI's do not exclude negative ERR.
Figure 10. Uterine cancer mortality in 0 – 20 mSv subcohort, LRT vs latency. Axes and model definitions as in Figure 2.
Figure 11. Uterine cancer mortality in 0 – 20 mSv subcohort, ERR vs latency. Axes and model definitions as in Figure 3.
Applying the category model with φ = 26.91 gives the weak result LRT_{catcon }= 8.077 with only β_{1 }significantly positive.
Allsolid
When 22 ≤ φ ≤ 23 the twophase model is significant against transient, linear, and control models (Figure 12) and ERR is negative in all models (Figure 13). When 39 ≤ φ ≤ 44 the transient model improves on the control model, and when 42 ≤ φ ≤ 43 the transient and twophase improve on the linear and control models. When 37 ≤ φ ≤ 44 ERR is positive in all models. The optimal latency in all three nested models is φ_{m }= 43.86, at which all are significant against the control model and the transient and twophase are significant against the linear model.
Figure 12. Allsolid cancer mortality in 0 – 20 mSv subcohort, LRT vs latency. Axes and model definitions as in Figure 2.
Figure 13. Allsolid cancer mortality in 0 – 20 mSv subcohort, ERR vs latency. Axes and model definitions as in Figure 3.
At φ_{m }= 43.86 the transient model has LRT_{transcon }= 16.781 and LRT_{translin }= 11.753, both highly significant, with ERR_{0.025,φ }= 0.271 (0.070, 0.509) while ERR_{1,φ }= 0.300 (0.034, 0.633). The twophase model has LRT_{2pcon }= 17.925, LRT_{2plin }= 12.896, both highly significant, while LRT_{2ptrans }= 1.143 is not significant. ERR_{0.025,φ }= 0.269 (0.051, 0.538) while ERR_{1,φ }= 0.272 (0.044, 0.662). The linear model has LRT_{lincon }= 5.029 and ERR_{1,φ }= 0.259 (0.030, 0.522). If applied to the 0 – 0.5 mSv data, the optimal latency is φ_{m }= 41.78 at which LRT_{lincon }= 5.562 and ERR_{0.025,φ }= 0.175 (0.028, 0.344).
At φ = 43.86 the category model has LRT_{catcon }= 16.311 while β_{2 }and β_{3 }are significantly positive.
The models pass both GoF tests. The transient model is preferable here, for simplicity.
Comparison with LSS12
The ERR values found here by applying nonlinear models to the 0 – 20 mSv subcohort and optimising latency are several orders of magnitude above those derived by extrapolating from results for a linear model applied with fixed 5 year lag to the entire dose range of Abomb survivors, as in LSS12 [4] whose Tables AII (Male/Female), AIII (Male) and AIV (Female) show ERR/Sv (organ dose). If a linear model is applied to a single dataset, ERR for d mSv will be (d/1000)(ERR/Sv). Extrapolations of this type are inherent in the ICRP estimate of the risks arising from low doses, which underpin its recommended annual dose limits.
Comparisons are shown in Table 5. Significant discrepancies occur at 10 mSv (ERR_{1}) and 0.25 mSv (ERR_{0.025}) with the twophase model for stomach, liver, and lung; and with the transient model for liver and lung.
Table 5. Comparison with extrapolations from LSS12^{a}
Estimates of ERR_{1 }from the nonlinear models are 2 to 3 orders of magnitude above extrapolations from LSS12; for ERR_{0.025 }the discrepancy is 3 to 4 orders of magnitude.
Dosimetry
As a first step towards understanding how dosimetry errors may affect these results, the models were fitted to portions of the 0 – 20 mSv data for the liver. Models were refined to control for "city", although this had little impact. There are 173 data cells with weighted adjusted colon dose = 0. If these cells are deleted from the 0 – 20 mSv subcohort and analysis is restricted to the remaining 2838 cells, fitting the twophase and linear models for the liver (M/F) at φ = 36.90 years gives LRT_{2pcon }= 37.639, LRT_{2plin }= 30.095, and ERR_{1 }= 1.478 (0.52, 3.04). Thus the results for the liver are not caused by any special features of the zerodose cells.
Alternatively, choose 0 = x_{0 }< x_{1 }< x_{2 }< ... < x_{9 }< x_{10 }= 20 mSv. Define S_{i }as the set of data cells in the 0 – 20 mSv subcohort for which the weighted adjusted colon dose does not fall in the interval [x_{i1}, x_{i}]. The x_{i }may be chosen so that the S_{i }have roughly equal py of observation. Then, fitting the twophase and linear models for the liver (M/F) with φ = 36.9 to the reduced datasets S_{i }gives LRT_{2plin }and ERR_{1 }values as shown in Table 6. LRT_{2plin }values are ≥ 18.94 (1 df) and ERR_{1 }is fairly stable, varying from 1.21 to 1.74.
Table 6. Twophase and linear models fittted to subsets of the liver data^{a}
The DS86 dosimetry used here and in LSS12 was reinvestigated by a Joint Working Group using physical measurements, resulting in a new dosimetry system DS02 [11]. A dataset using DS02 and enabling comparison with DS86 was released by RERF last year and can be downloaded as DS02can.dat from the RERF website [9]. However, the cancer mortality fields in this dataset only show deaths from solid cancers (combined), liquid cancers, and leukaemia. Investigation of the stomach, liver, and lung as individual sites is not possible from this public dataset. Furthermore, the stratification of the new dataset differs from that in LSS12 and it is not possible to simply read off the DS02 values for cells in the LSS12 data.
Nonetheless for those cells in the DS02 dataset which have a DS86 dose, the weighted adjusted DS02 colon dose using neutron RBE = 10, shown as "cola02w10", can be compared with the corresponding DS86 dose variable shown as "cola86w10".
Only 11 cells, with 13.83 py and no solid cancer deaths, have cola02w10 < 0.005 Sv and cola86w10 ≥ 0.005 Sv. Only 5 cells, with 11.24 py and no solid cancer deaths, have cola02w10 ≥ 0.005 Sv and cola86w10 < 0.005 Sv. For cells with cola86w10 ≥ 0.005 Sv the ratio θ = cola02w10/cola86w10 varies from 0.216 to 1.925 with mean 1.08 and standard deviation 0.082. There are only 7 cells, with 10.47 py and no solid cancer deaths, for which θ < 0.5. Likewise 92 cells, with 195.84 py and 3 solid cancer deaths, have θ > 1.5. In the subcohort with 0.005 Sv ≤ cola86w10 ≤ 0.02 Sv, θ varies from 0.216 to 1.27 with mean 1.026 and standard deviation 0.077. There are 15 cells with 78.5 py and 2 solid cancer deaths, for which θ < 0.8, and 9 cells with 9.88 py and 1 solid cancer death, for which θ > 1.2
Thus if DS02 is an accurate estimate of the flash dose, there is virtually no misclassification of DS86 between the categories "above 5 mSv" and "below 5 mSv", and above 5 mSv the DS86 dose is a reasonable estimate, though 8% below DS02 on average.
The scope in this paper for misclassification of the dose is therefore reduced if doses below 5 mSv are taken as baseline and doses from 5 to 20 mSv are taken as a single category which is then analysed with latency. While that approach is too crude to detect nonlinearity, it gives very similar results for the liver to those found with the linear model.
As shown in Table 1 the optimal latency for the liver (M/F) using the linear model is φ = 38.58, for which comparison with the control model has LRT_{lincon }= 11.21, β = ERR_{1 }= 0.69 (0.25, 1.26). Now define E = 1 if TimeSinceExposure ≥ 38.58 and 5 mSv ≤ colon dose ≤ 20 mSv, E = 0 otherwise. As TimeSinceExposure is known, errors in E can only arise if doses below 5 mSv were misclasssified as above 5 mSv, or vice versa. Fitting the model defined by ERR = γE to the liver (M/F) data for the 0 – 20 mSv dose range and comparing with the control model (γ = 0) gives LRT = 11.04, γ = 0.67 (0.24, 1.21). Thus the results for the linear model can be reproduced with a twocategory model using a cutpoint of 5 mSv, and these categories are almost identical whether defined by DS86 or DS02.
As a further test, the liver data were modelled in the extended DS86 dose range 5 mSv – 500 mSv, where DS86 and DS02 are in reasonable agreement. The twophase model is a significant improvement on the linear model at latency φ = 36.9 years, with LRT_{2plin }= 10.37 (1 df) and ERR_{1 }= 0.74 (0.10, 1.79).
Similar results for the liver are obtained from the 0 – 500 mSv dose range. The twophase model is a significant improvement on the linear model at latency φ = 36.9 years, with LRT_{2plin }= 16.86 (1 df) and ERR_{1 }= 0.76 (0.15, 1.66).
For comparison, the results for the liver obtained from the 0 – 20 mSv dose range at latency φ = 36.9 years, are LRT_{2plin }= 29.67 (1 df) and ERR_{1 }= 1.46 (0.50, 3.00). Similar results when "city" is omitted from the model are shown in Table 3.
The DS02 public dataset does not show the liver but does allow modelling of allsolid cancers. The subcohort S defined by cola02w10 ≤ 0.02 and TimeSinceExposure ≤ 45.39 years (the maximum value in the 0 – 20 mSv subcohort of LSS12 data) has 1682335.39 py and 4363 cases, roughly comparable to the LSS12 subcohort analysed with DS86 in this paper. Models were defined as previously, but using cola02w10 in place of the DS86 weighted adjusted colon dose. None of the cells in S contain Nagasaki factory workers. Controls for "city" and "ground distance category" (proximal or distal) were introduced, though they had little impact. If the linear model is fitted to S with latency φ = 43.29 years then LRT_{lincon }= 3.79 (1 df). For the twophase model LRT_{2pcon }= 10.49 (2 df) and LRT_{2plin }= 6.70 (1 df), again showing nonlinearity. ERR_{1 }= 0.223 (0.07, 0.58). For the DS86 results, Table 3 shows ERR_{1 }= 0.272 (0.04, 0.66) at latency 43.86 years.
The estimates of ERR_{1 }for allsolid cancers are similar whether derived from the 0 – 20 mSv subcohort of LSS12, or the comparable 0 – 20 mSv subcohort of the DS02 data.
The DS02 data contains an additional 10 years of followup. Nonlinearity is still present in the extended cohort. Define the subcohort T by cola02w10 ≤ 0.02 without restricting TimeSinceExposure. At latency 43.29 years the linear model is indistinguishable from the control model (LRT = 0.47) but the twophase model has LRT_{2pcon }= 9.49 (2 df) and LRT_{2plin }= 9.02 (1 df). ERR_{1 }= 0.08 (0.07, 0.26).
Discussion
This paper begins from the publicly available data, and I do not know if a comparable analysis of the anonymous individual data would show similar results.
In reanalysing portions of the 1950–90 grouped data, the approach here has four features.
1) To predict risks at 10 mSv, the 0 – 20 mSv data is analysed directly.
2) A variable lag period is used to analyse latency.
3) Dosimetry data is not reduced to categories before modelling. Dose is taken as a numerical variable, defined on the grouped data cells, but results are also tested in a 4category model with baseline defined by D_{φ }= 0 and cutpoints which roughly equipartition the py in nonbaseline categories.
4) Linearity of the dose response is tested by nesting within more complex models.
With a fixed 5 year lag, none of the cancers considered here show significant effects in the 0 – 20 mSv dose range using a linear model. Allowing latency to vary in this model gives significant positive responses for the liver and allsolid cancers.
Latency reflects biochemical changes required after initial radiation if mutant cells are to progress and form a tumour eventually identified as cause of death, and historical changes in environmental factors which interact with radiation for a particular cancer. Thus latency may be organ and gender specific.
Rothman [13] illustrates how ignoring latency may mask important effects, whether or not the original exposure was brief. Analysis using the lagged dose D_{φ }is a simple approach depending on only one parameter. The response might be clearer by modelling the effect of D*w, with w some more general function of TimeSinceExposure. Such models have been applied to lung cancer mortality in uranium miner cohorts [14,15].
If the linear model were appropriate throughout the low dose region, we might expect ERR_{0.025,φ }~ 0.025(ERR_{1,φ}) whatever data were used to estimate each ERR. In fact, the ERR values are often comparable. Two nonlinear models give significant improvements in the 0 – 20 mSv dose range for the stomach, liver, lung, uterus and allsolid cancers, and for various gender specific sites. These improvements are strong, for example p < 0.001 when comparing the twophase and linear models (M/F) for stomach, liver, lung, and allsolid; and p < 0.000001 for the liver.
Unlike the linear and control models, the transient and twophase models require extensive computation as the Deviance may have multiple local minima at any choice of latency φ (fixed when fitting the model). Computation involves a search for local minima, selection of the minimum Deviance at φ, and then a comparison amongst these minima for different φ values. The optimal φ is chosen to give the absolute minimum Deviance, with or without the constraint that ERR_{0.025 }be nonnegative. The search is streamlined by restricting φ to 5, 6, ... 44 and later refined to consider all φ (to 2 decimal places) in a range which appears likely to contain the optimal value.
I do not know of any general analytical method which might limit the total number of local minima at fixed φ in this data. Instead, the τ axis is partitioned (see Methods). Fitting the model with τ constrained to an interval typically yields a τ value at either endpoint, reflecting the constraint, except for those intervals which contain τ values at which the Deviance attains a local minimum.
Searches begin from the control parameter values which optimise the control model, while τ is confined to the relevant interval and β,σ are initially set to 0. Conceivably, this choice of initial conditions may cause the NewtonRaphson iteration to miss some local minima, though testing other initial conditions did not detect any other solutions. In any case the minimum Deviance at any particular φ can be no higher than the values found here, so any missing minima could only strengthen the evidence of nonlinearity.
Fitting the twophase model to the lung data illustrates these issues. When φ = 13 three local minima are detected. At τ = 4.51, Dev = 1040.35. At τ = 13.86, Dev = 1040.87. At τ = 175.63, Dev = 1050.25. The minimum Dev at φ = 13 is thus 1040.35. The linear model has Dev = 1053.64. Thus LRT_{2plin }for comparing the twophase and linear models is 1053.64 – 1040.35 = 13.29, and it is this value which is displayed in Figure 6 when φ = 13. Likewise the ERR value computed at this minimum Dev is displayed in Figure 7 when φ = 13. The resulting graphs indicate the region to be searched for an optimal choice of φ, subject to the constraint ERR ≥ 0. This optimum is φ = 13.6, the value shown in Table 3. At this latency there are again 3 local minima, two of which have similar Dev. At φ = 13.6 the minimum at τ = 4.45 with Dev = 1038.58 has ERR_{1 }= 0.88 while the local minimum at τ = 13.54 with Dev = 1039.26 has ERR_{1 }= 1.10. Note that while τ varies widely without appreciable change in Dev, ERR_{1 }is much more stable. In this example, the 95%CI is (0.12, 3.36). For this reason, ERR is a much better focus for analysis than the model parameters themselves.
Confidence intervals for ERR are often somewhat wider in the twophase model than in the simpler transient model which is sufficient to describe the lung and allsolid (M/F, F) data. However, as well as improving the fit for the stomach, liver and allsolid (M) the twophase model gives a more coherent account of the latency regions of significant positive or negative dose response.
Linear extrapolation of the LSS12 results shows almost no response at the doses considered here. Whilst LSS12 uses organ doses and a different system of controls, these factors do not account for the large discrepancy in risk estimates. Alternative controls affect the estimates by a factor of 2 or less, and the use of organ doses has even less impact. Significant discrepancies arise when the dose range is restricted to 0 – 20 mSv and latency is included in the analysis.
Stewart and Kneale [16] found evidence of selection bias in the LSS 1950 – 1985 cohort, thought to reflect the fact that only those victims able to survive from 1945 – 1950 were eligible to enter the cohort. The test group used by Stewart and Kneale to detect bias included less than 4% of the total cohort but had nearly 30% of high doses (> 1000 mSv). It is plausible that such bias would be reduced in the 0 – 20 mSv subcohort, but I cannot test this from the publicly available RERF data.
Pierce and Preston [17] analysed allsolid cancer incidence in Japanese survivors from the 1958 – 1994 tumour registry data for the range 0 – 500 mSv, using a linear ERR model based on colon dose and a categorical model with cutpoints 0, 5, 20, 100, 200, 250, 300, 400 mSv (colon dose). Estimates correspond to ERR ~ 0.006 at 10 mSv. Likewise, if the linear model here is applied to allsolid cancer mortality (1950–90) in the 0 – 500 mSv dose range with latency 5 years, ERR_{1 }= 0.004 with LRT = 8.58.
Pierce and Preston focus on survivors who were exposed relatively near the hypocentres of the Abombs and exclude distal survivors ( ≥ 3 km distant) on the grounds that they had higher baseline cancer rates and that some lifestyle cancer risk factors correlate with urbanrural distinctions, though cigarette smoking had almost no correlation with estimated dose or distance from the hypocentre. Excluding the distal group lowered the baseline by about 5% in their data. Although that is significant in relation to the estimates of ERR in the RERF studies, it is marginal compared to the ERR values found here with the latency models.
The 0 – 20 mSv subcohort contains many proximal as well as distal survivors (10,159 proximal survivors in the incidence dataset had doses below 5 mSv). The results here may of course reflect other risk factors which may correlate and/or interact with radiation dose, but which could only be approached through the individual data. Investigation of possible confounders should also consider latency and nonlinear models such as those analysed here, given their clear superiority to the linear model with fixed 5 year lag, for the lowdose grouped data.
This paper is based on the DS86 dosimetry and the results could reflect dosimetry errors, arising from incorrect estimation of the flash dose or by omitting other radiation sources. The doses received in Hiroshima and Nagasaki include the flash dose (used here), induced radioactivity in building materials or soil which persisted for several weeks, "black rain" which fell in the immediate aftermath of the bombings, natural background radiation, global fallout from atmospheric weapons tests, occupational and medical exposures. The public data does not include any individual occupational or medical exposures. Natural background and global fallout should not be correlated with the exposures arising directly from the bombs in 1945, and would be expected to bias results towards the null. Doses from induced radioactivity and "black rain" could be relevant, but currently available RERF public datasets do not include either of these two additional sources.
Errors in the flash dose itself are unlikely to explain the results. Nonlinearity and large values of ERR at 10 mSv persist when the zerodose data is deleted from the 0 – 20 mSv subcohort and likewise when ten intervals spanning 0 – 20 mSv are used to delete dose ranges from the data.
The DS02 dataset shows that there is virtually no misclassification between the DS86 categories "below 5 mSv" and "above 5 mSv". For the liver at latency 36.9 years, it makes little difference whether dose is taken as a categorical variable defined by the 5 mSv cutpoint in the data stratification, or as a numerical value analysed with the linear model.
DS02 and DS86 are in reasonable agreement above 5 mSv. If datacells with DS86 dose below 5 mSv are excluded from the 0 – 20 mSv dose range, the twophase model is no longer a significant improvement on the linear model for the liver. However, very similar evidence of nonlinearity for the liver arises in the 5 mSv – 500 mSv dose range where DS86 is a reasonable approximation to DS02, and in the 0 – 500 mSv dose range considered by Pierce and Preston [17].
Analysis of solid cancers using DS02 in the 0 – 20 mSv dose range gives estimates of ERR comparable to those derived from DS86.
These various arguments suggest that the results are unlikely to be explained by errors in dosimetry of the flash dose, although DS02 like DS86 is subject to some uncertainty due to random errors in specifying individual location and shielding. Separately, induced radioactivity and "black rain" represent additional doses not reported by DS86 or DS02.
Induced radioactivity appears unlikely to fully explain the results. According to the RERF website [6] "The closer to the hypocentre, the higher was the dose [from induced radioactivity]. Past investigations suggested that the maximum cumulative dose at the hypocentre from immediately after the bombing until today is 0.8 Gy in Hiroshima and 0.3–0.4 Gy in Nagasaki. When the distance is 0.5 km or 1.0 km from the hypocentre, the estimates are about 1/10 and 1/100 of the value at the hypocentre, respectively." The issue was examined in detail in the DS86 Final Report Chapter 6 and an Appendix to this Chapter [8]. The cumulative dose from induced radioactivity decreases exponentially with distance from the hypocentre.
From the DS02 dataset, the minimum distance from the hypocentre amongst cells with DS86 doses below 20 mSv is 2.081 km while for cells with DS86 doses below 500 mSv it is 1.210 km. For the 5 – 500 mSv range the total cumulative impact of induced radioactivity would be below 8 mSv in Hiroshima and below 4 mSv in Nagasaki. Consider the unlikely possibility that induced radioactivity adds 8 mSv to those cells which contain Hiroshima liver cancer deaths, and 4 mSv to those cells which contain Nagasaki liver cancer deaths, while leaving all other cells unaffected. Under this extreme assumption, the twophase model at latency φ = 36.9 years has solutions with ERR_{1 }< 0 but the linear model gives LRT_{lincon }= 11.03 (1 df) and ERR_{1 }= 0.029 (0.010, 0.054). Conversely, suppose that induced radioactivity adds 8 mSv to those Hiroshima cells which do not contain liver cancer deaths, and 4 mSv to those Nagasaki cells which do not contain liver cancer deaths, while leaving all other cells unaffected. Then the linear model at latency φ = 36.9 years gives LRT_{lincon }= 4.63 (1 df) and ERR_{1 }= 0.016 (0.001, 0.036), which is still 4.3 times higher than the LSS12 estimate. These two extreme assumptions may perhaps provide limits on the scope for induced radioactivity to affect the linear model in the 5 – 500 mSv subcohort, where the corresponding results without the addition of induced radioactivity are LRT_{lincon }= 7.25 (1 df) and ERR_{1 }= 0.021 (0.005, 0.044). For the 0 – 20 mSv dose range the total cumulative impact of induced radioactivity would be below 0.08 mSv in Hiroshima and below 0.04 mSv in Nagasaki. Whatever consequence this may have for the twophase model, its effect on the linear model is negligible.
"Black rain" fell primarily at some distance from the hypocentres. According to the RERF website, "Because of wind, the rain mainly fell in northwestern Hiroshima (KoiTakasu area) and in eastern Nagasaki (Nishiyama area). The maximum estimates of dose due to fallout are 0.01–0.03 Gy in Hiroshima and 0.2–0.4 Gy in Nagasaki. The corresponding doses at the hypocentres are believed to be only about 1/10 of these values." According to Chapter 6 of the DS86 Final Report [8] the rainfall was concentrated around 3000 m from the hypocentres in both cities. From the DS02 dataset, the maximum distance from the hypocentre amongst cells with DS86 doses above 5 mSv is 2.683 km, somewhat closer to the hypocentres than the main rainfall areas. Maximum doses from rainfall were much lower in Hiroshima, though RERF do not give enough detail to estimate an upper bound for the dose from "black rain" in Hiroshima at distances below 2.683 km. However, if the Hiroshima data for the liver is analysed separately for the 5 – 500 mSv subcohort, the twophase model remains a significant improvement over the linear model at latency φ = 36.9. LRT_{2plin }= 5.718 (1 df) and ERR_{1 }= 0.674 (0.06, 1.92) while for the linear model LRT_{lincon }= 7.946 and ERR_{1 }= 0.027 (0.007, 0.055).
Compounding the uncertainty over doses from "black rain", individuals may have travelled into a rainfall area even if they were outside it when the bombs exploded. RERF's estimates refer to external doses, omitting any inhaled or ingested radiation. In any case the conclusions are provisional until a dataset for specific solid cancers showing DS02 flash doses, induced radioactivity and "black rain" becomes available. Potential confounding by other risk factors cannot be excluded. The neutron RBE of 10 used here, as in LSS12 and elsewhere, may be inappropriate. Despite all these reservations, which also affect the LSS studies, it is striking that both the 0 – 20 mSv and 5 – 500 mSv subcohorts show nonlinear dose response for the liver and the twophase model gives comparable estimates from both subcohorts for Excess Relative Risk at 10 mSv with latency 36.9 years. "Black rain" and induced radioactivity have quite different impacts within these two subcohorts.
I do not know whether these results and the optimal latencies are specific to the Abomb cohort. In their study of gamma radiation and mortality in the Oak Ridge workforce, Frome et al. [18] find a 'Lowdose β' value of 2.9 with LRT = 3.12 (p = 0.08) for the combined digestive category using a multiplicative model (without latency) with dose restricted to 0 – 640 mSv. Since their unit dose is 1 Sv this corresponds to ERR_{1 }= 0.029.
Nonlinear models are applied here without assuming any cellular mechanism, and their success in fitting the cohort data does not prove that any particular cell mechanism operated there. However, the twophase model adapts and simplifies a model derived by Brenner et al. [19] to explain the 'oncogenic transformation frequency' of cells exposed to broad beam irradiation by α particles:
TF = νq<N> + σ [1e^{(k<N>)}] [e^{(q<N>)}]
TF is the number of transformed cells per surviving cell (Excess Relative Risk of transformation); ν is the transformation frequency for cells struck directly by exactly one α particle, q the surviving fraction of cells struck directly by exactly one α particle, <N> the mean number of α particles striking each cell in the broad beam irradiation (a dose variable), k the number of cells receiving the bystander signal emitted from a cell struck directly by one or more α particles, and σ the (presumed) hypersensitive fraction of bystander cells which are transformed on receipt of any bystander signal. Note that the first term in the TF expression refers to direct effects, whilst the second term refers to bystander effects. As a function of <N>, TF is approximately linear with slope ν q + σ k at very low doses and approximately linear with slope ν q at higher doses.
The twophase model is asymptotically linear with slope σ + β at very low doses and asymptotically linear with slope β at higher doses. It generates curves of the same qualitative shape as the TF cell model.
In the cell model, the ratio of asymptotic slopes is R = 1+(σ k/ν q). From targetted microbeam experiments on C3H 10T^{1}/2 mouse fibroblast cells, Brenner et al. estimate ν = 1.3 × 10^{4}, σ = 6.4 × 10^{4}, and q = 0.8. Thus R ~ 1 + 6.2 k. In a subsequent paper Brenner and Sachs [20] suggest that k ~ 50 (so R ~ 311) for the human lung, from modelling doserate effects in (male) uranium miners exposed to radon.
In the twophase model, the ratio of asymptotic slopes is R = 1 + σ/β. For the male lung, with optimal latency 13.55 years, R = 202.22. For the lung, the twophase model is not significant against the transient and the confidence region for R is infinite and includes negative values. But for the stomach (M/F) R = 262.62 with 95% CI (102.41, 975.09), and for the liver (M/F) R = 204.11 (58.29, 792.85). Thus the low dose Abomb data is compatible with a bystander model using roughly comparable values of k, the number of cells receiving the bystander signal.
The flash dose comprised gamma and neutron doses. An appropriate cell model for bystander effects following gamma radiation may differ from that developed for α particles, where the impact of a single track on one cell is quite large. However fission neutrons, like α particles, are high LET (linear energy transfer) and might elicit a similar bystander signal. Induction of genomic instability in unirradiated bystander cells has been demonstrated for neutron irradiation in mice [21].
Nonlinear dose response curves can also arise from a hypersensitive population subset [22]. Sharp et al. [23] analysed primary liver cancer mortality in relation to Hepatitis B and C in Japanese Abomb survivors and found a very strong supermultiplicative interaction between HVC and radiation dose, but only for the high and medium dose ranges. For doses below 18 mSv, there was no significant interaction. However, Sharp et al. did not include latency and truncated all liver doses below 3 mSv. I could not analyse primary liver cancer because the data in r12canc.dat shows all liver cancers. Persistent inflammation including chronic liver disease has been detected in Japanese survivors 40 years after the bombing and correlated with radiation dose [24].
Conclusion
All the models considered here show unexpectedly large and significant results in the 0 – 20 mSv dose range when a lagged colon dose is taken as the main predictor. Nonlinear models allowing for asymptotic effects as dose approaches 0 improve the fit and give a much higher slope for the dose response curve near 0 than at higher doses.
Whether or not analyses of this cohort can be transferred internationally, the results here raise questions in Japan. In any case, significant results for doses below 20 mSv are directly relevant to the current ICRP recommendations limiting annual occupational exposure to 20 mSv (whole body dose). Analysis of the 0 – 20 mSv dose range gives responses several orders of magnitude above extrapolations from LSS12 (and other Life Span Studies).
Low dose effects which depend on latency or are detected by nonlinear models can still cause significant risks.
Declaration of competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
GD conceived of and designed the study and carried out the statistical analysis.
Acknowledgements
Special thanks to David Richardson for extended correspondence and suggesting the latency parameter approach, Steve Wing for raising dosimetry questions during peer review, and Tom Sorahan for comments on a draft. Thanks also to Dale Preston, John Cologne, Mike Thorne, Keith Baverstock, David Brenner, Beate Ritz, Teresa Coons, Rudi Nussbaum, Joe Abramson, and Ric Swindell for comments, criticism and encouragement at earlier stages.
This work was not funded.
The data described in this report were obtained from the Radiation Effects Research Foundation (RERF) in Hiroshima, Japan via the Comprehensive Epidemiologic Data Resource. RERF is a private foundation that is funded equally by the Japanese Ministry of Health and Welfare and the US Department of Energy through the US National Academy of Sciences. The conclusions reached in this report are mine and do not necessarily reflect the scientific judgement of RERF or its funding agencies.
References

Committee on the Biological Effects of Ionizing Radiations (CBEIR): Health Effects of Exposure to Low Levels of Ionizing Radiation – BEIR V. Washington, D.C.; 1990.

ICRP: 1990 Recommendations of the International Commission on Radiological Protection Publication 60, Annals of the ICRP 21 (1–3). Oxford; 1991.

Nussbaum RH: The linear nothreshold doseeffect relation: is it relevant to radiation protection legislation?
Med Phys 1998, 25(3):291299. PubMed Abstract  Publisher Full Text

Pierce DA, Shimizu Y, Preston DL, Vaeth M, Mabuchi K: Studies of the Mortality of Atomic Bomb Survivors. Report 12, Part I, Cancer: 1950–1990.
Radiat Res 1996, 146(1):127. PubMed Abstract  Publisher Full Text

Mothersill C, Seymour C: Radiationinduced bystander effects: past history and future directions.
Radiat Res 2001, 155(6):759767. PubMed Abstract  Publisher Full Text

RERF (Radiation Effects Research Foundation) [http://www.rerf.or.jp/] webcite

CEDR (Comprehensive Epidemiologic Data Resource) Data File Set JALSSA03 File ID: R12CANC [http://cedr.lbl.gov/cgibin/spiface/find/cedrfile/?FILE = 1040] webcite
accessed 19 September 2002

USJapan Joint Reassessment of Atomic Bomb Radiation Dosimetry in Hiroshima and Nagasaki (DS86): Volume 1 and Volume 2 [http://www.rerf.or.jp/shared/ds86/ds86a.html] webcite

DS02 Risk Estimation: Solid Cancer and Leukemia Mortality Data [http://www.rerf.or.jp/eigo/scidata/ds02.htm] webcite

Preston DL, Pierce DA, Shimizu Y, Cullings HM, Fujita S, Funamoto S, Kodama K: Effect of recent changes in atomic bomb survivor dosimetry on cancer mortality risk estimates.
Radiat Res 2004, 162(4):37789. PubMed Abstract  Publisher Full Text

Cullings HM, Fujita S, Funamoto S, Grant EJ, Kerr GD, Preston DL: Dose estimation for atomic bomb survivor studies: its evolution and present status.
Radiat Res 2006, 166(1 Pt 2):21954. PubMed Abstract  Publisher Full Text

Statistical methods in cancer research. Volume II – The design and analysis of cohort studies. IARC Sci Publ. 1987, (82):1406.

Rothman KJ: Induction and latent periods.
Am J Epidemiol 1981, 114(2):253259. PubMed Abstract  Publisher Full Text

Committee on the Biological Effects of Ionizing Radiations (CBEIR): Health Risks of Radon and Other Internally Deposited AlphaEmitters – BEIR IV. Washington, D.C.; 1988. PubMed Abstract  Publisher Full Text

Langholz B, Thomas D, Xiang A, Stram D: Latency analysis in epidemiologic studies of occupational exposures: application to the Colorado plateau uranium miners cohort.
Am J Ind Med 1999, 35(3):246256. PubMed Abstract  Publisher Full Text

Stewart AM, Kneale GW: Abomb survivors: factors that may lead to a reassessment of the radiation hazard.
Int J Epidemiol 2000, 29(4):708714. PubMed Abstract  Publisher Full Text

Pierce DA, Preston DL: RadiationRelated Cancer Risks at Low Doses among Atomic Bomb Survivors.
Radiat Res 2000, 154(2):178186. PubMed Abstract  Publisher Full Text

Frome EL, Cragle DL, Watkins JP, Wing S, Shy CM, Tankersley WG, West CM: A mortality study of employees of the nuclear industry in Oak Ridge, Tennessee.
Radiat Res 1997, 148(1):6480. PubMed Abstract  Publisher Full Text

Brenner DJ, Little JB, Sachs RK: The bystander effect in radiation oncogenesis: II. A quantitative model.
Radiat Res 2001, 155(3):402408. PubMed Abstract  Publisher Full Text

Brenner DJ, Sachs RK: Do low doserate bystander effects influence domestic radon risks?
Int J Radiat Biol 2002, 78(7):593604. PubMed Abstract  Publisher Full Text

Watson GE, Lorimore SA, Macdonald DA, Wright EG: Chromosomal instability in unirradiated cells induced in vivo by a bystander effect of ionizing radiation.
2000, 60(20):560811. PubMed Abstract  Publisher Full Text

Baum JW: Population heterogeneity hypothesis on radiation induced cancer.
Health Phys 1973, 25(2):97104. PubMed Abstract

Sharp GB, Mizuno T, Cologne JB, Fukuhara T, Fujiwara S, Tokuoka S, Mabuchi K: Hepatocellular carcinoma among atomic bomb survivors: significant interaction of radiation with hepatitis C virus infections.
Int J Cancer 2003, 103(4):531537. PubMed Abstract  Publisher Full Text

Neriishi K, Nakashima E, Delongchamp RR: Persistent subclinical inflammation among Abomb survivors.
Int J Radiat Biol 2001, 77(4):47582. Publisher Full Text