Skip to main content

Epidemiological Studies of Exposed Southern Urals Populations

Final Report Summary - SOLO (Epidemiological Studies of Exposed Southern Urals Populations)

Executive Summary:
SOLO was a five year integrated, multi-disciplinary project to investigate the risks to human health of low and protracted radiation exposures, considering exposures to radionuclides within the body as well as external radiation. The project continued and developed upon previous follow-up studies of the health of two key population cohorts from the Southern Urals of the Russian Federation: the Mayak Worker Cohort (MWC) and the Techa River Cohort (TRC), the latter being exposed as a result of Mayak discharges of radioactivity to the Techa River. It also makes use of the Sellafield worker cohort (SWC) in the UK and, for the first time, has pooled cohorts to assess risks from exposures in utero (MWC and TRC) and from plutonium in workers (MWC and SWC).

Dosimetry
The reliability of risk estimates obtained from epidemiological studies is dependent on the quality of radiation dose information for the members of the cohorts being studied. Two important strands of work were undertaken to improve the reliability of external dose estimates for the Mayak and Techa cohorts. Personal dosemeter data for Mayak workers and dose estimates for Techa residents based on location were compared for validation with results obtained using EPR measurements on teeth and FISH chromosome aberration techniques on cultured white blood cells. In addition, detailed environmental measurements at a particular village on the Techa River were made to facilitate dose reconstruction for inhabitants. Important developments included improvements in the reliability of EPR measurements and insights into the use of FISH to measure internal doses from strontium-90 and plutonium-239 as well as external dose. The estimation of internal doses to the foetus were improved for plutonium-239 using measurements of urinary excretion in mothers who had been Mayak workers, and their offspring, and for strontium-90 by detailed modelling of data relating to Techa residents. Updated models and a common approach were applied to the calculation of plutonium doses to Mayak and UK Sellafield workers.

Cancer risks in adults from external radiation and internal plutonium
Analyses were completed of morbidity from lung, liver and bone cancer, all other solid cancers (as a single group) and from haematolymphatic cancers in the MWC. The results provided important support to previous observations of mortality, with risk estimates consistent with published values, and evidence that plutonium-239 intakes can cause lung, liver and bone cancer but that the radionuclide is ineffective as a leukaemogen. Analyses of lung cancer and leukaemia in a pooled cohort of Mayak and Sellafield plutonium workers showed consistency of results across the pooled cohort with raised risks of lung cancer but not leukaemia. The dose-response and risk estimates for plutoinium-239 induced lung cancer were consistent with comparable risk estimates for external gamma radiation when account is taken of the relative biological effectiveness of alpha particles emitted by plutonium-239.

Non-cancer risks in adults from external radiation and internal plutonium
Previous analyses of ischaemic heart disease (IHD) and cerebrovascular disease (CeVD) mortality and morbidity in the MWC were extended with longer follow-up and more cases to give greater statistical power. These analyses showed good statistical evidence for an association between external radiation exposure and morbidity but not mortality, an observation requiring further investigation. A linear-dose response relationship was observed for the incidence of IHD and CeVD in relation to exposure to external radiation but the dose-response for plutonium-239 exposure was more complex. Analysis of circulatory disease mortality in the pooled cohort of Mayak and Sellafield plutonium workers was complicated by differences in the background disease rates; risk estimates were mostly positive in relation to external exposure (although notably less so for MWC compared to SWC) but there was little evidence for a link with internal exposure. Analyses of risks of chronic respiratory disease incidence in the MWC recorded significant excess risk in relation to internal plutonium-239 exposure to the lung but no excess risk in relation to external gamma radiation exposure.

Cancer risks following in utero irradiation from external radiation and internal radionuclides
A pooled cohort of children was defined whose mothers either worked at Mayak or lived near the Techa River before and/or during pregnancy. External and internal radiation doses to these children were calculated based on dosimetry models developed during the project. The pooled cohort currently lacks the statistical power to show significant results, but the absence of a positive result at this stage is important in showing that risks of in utero exposure have not been dramatically underestimated, particularly for internal radionuclides.

Project Context and Objectives:
SOLO was a five year integrated, multi-disciplinary project to investigate the risks of radiation exposure. It involved dosimetrists, operational health physicists, statisticians and epidemiologists from nine institutions in the EU, Russian Federation and the USA.
The project primarily studied the health of two key radiation exposed populations from the Southern Urals of the Russian Federation. The aim was to derive better estimates of the risks of cancer and some non-cancer diseases from long term exposure to ionising radiation.
The Mayak nuclear facility situated in the Chelyabinsk region of the Russian Federation began operating in 1948 to produce plutonium for the Soviet atomic weapons program. Workers at the plant, who lived in the nearby city of Ozyorsk, were exposed to ionising radiation often at high levels over many years. Also exposed were residents of villages along the Techa river which flows close to the plant and into which radioactive waste was discharged. Both groups were also exposed as a result of an accident at Mayak in 1957.
Current and retired Mayak workers, who are still living in Ozyorsk, have annual medicals in which smoking and alcohol consumption information is recorded. This makes them particularly valuable for studying the effects of radiation on diseases that are affected by smoking and alcohol consumption, such as cancer and circulatory diseases.
Over the past two decades, considerable effort has been directed to identifying cohorts of exposed populations in the Southern Urals and to following them up. These cohorts are the following:
Mayak Worker Cohort (MWC): This consists of workers employed at one or more of the three main plants of Mayak PA since the start of operations in 1948. These plants were the nuclear reactor complex, the radiochemical plant and the plutonium production plant. The initial focus was on workers first employed in 1948-1958, since they had the highest exposures, both in terms of external gamma radiation and intakes of plutonium. However, more recent analyses have also considered workers first employed after 1958, whose exposures tended to be lower than those of earlier workers. Workers at the auxiliary plants have also been considered in a few analyses, although their exposures tend not to be well-identified.
Extended Techa River Cohort (ETRC): This includes persons who lived in one of the Techa River villages at some point during 1950-1960 and who were exposed to external and internal radiation as a consequence of radioactive releases from Mayak PA. The cohort contains residents who were born in 1949 and earlier and lived in either the five administrative raions in Chelyabinsk oblast (Kunashaksky, Krasnoarmeisky, Argayashsky, Sosnovsky and Kaslinsy) or in either Kataisky raion or Dalmatovsky in Kurgan oblast (i.e. the oblasts through which the Techa River flows). The study cohort comprises close to 30 thousand persons.
Techa River Offspring Cohort (TROC): This consists of persons born in 1950 or later in the afore-mentioned raions of Chelyabinsk or Kurgan oblasts, or in Chelyabinsk city, who have at least one parent who is a member of ETRC. The cohort currently contains about 21,500 individuals, of whom about 6,000 were exposed in utero.
Data on deaths and cancers cases in all three cohorts have been ascertained using, inter alia, vital statistics data, address bureau records, cause-of-death registers both in Ozyorsk and in raions along the Techa River autopsies (mainly for deaths in Ozyorsk) and records from oncology clinics and the URCRM clinic. Furthermore, morbidity data were collected for Mayak workers over the period that they were resident in Ozyorsk, even after they had stopped working at Mayak. Information on lifestyle factors such as smoking and alcohol consumption was also collected for Mayak workers whilst resident in Ozyorsk.
Project aims:
Verification of external dosimetry for Southern Urals radiation exposed populations:
The reliability of the risk estimates provided by epidemiological studies is highly dependent the quality and quantity of the data on which they are based. Among the most important factors for radiation effects studies are the accuracy of the estimates of the radiation dose and of the follow-up information from which the risk estimates are derived. The current dosimetry systems used to estimate external dose in the Mayak and Techa river cohorts are known as Mayak Worker Dosimetry System 2008 (MWDS-2008) and Techa River Dosimetry System 2009 (TRDS-2009). These systems have evolved over many years to be as accurate as possible but important issues remain.
An aim of the project was to develop and implement strategies for validating the MWDS-2008 and TRDS-2009 dosimetry systems that are used to calculate external doses for the Mayak worker cohort (MWC) and Techa River populations. The project also aimed to identify ways of improving these dosimetry systems.
The overall strategy for validating both these dosimetry systems was to perform comparative analyses of individual external dose estimates generated by the dosimetry systems, with measurements of absorbed dose in teeth by electron paramagnetic resonance (EPR) and dose to red bone marrow (RBM) obtained by measurement of translocations in peripheral blood lymphocytes with fluorescence in-situ hybridization (FISH). These ‘within individual’ comparisons were identified as the best strategy for dose validation, noting the variability of doses between individuals. In addition, the project undertook a programme of environmental measurements at one of the Techa River villages, Metlino, to estimate the historic anthropogenic exposures and thereby improve estimates of doses to residents.

Analysis of cancer and non-cancer diseases in the Mayak cohort:
Analyses of cancer risk in the occupationally exposed Mayak population provides independent estimates of risk with which to compare those used for the derivation of radiological protection standards that are based on the Japanese Atomic bomb survivor cohort. Of particular interest are cancers that may be associated with plutonium exposure i.e. those of the lung, liver and bones. Analyses of these and other cancers have been performed in proceeding projects but the precision of the Mayak estimates could be improved.
The project aimed to provide improved risk estimates of cancer incidence from the specific cancers for which prior evidence suggested an excess might be expected (lung, liver and bone cancer) and to examine if risks from other cancers were also raised in the Mayak cohort. This was achieved using extended follow-up and with an expanded cohort can provide greater precision.
The internationally agreed system of radiation protection does not include non-cancer diseases in its estimate of detriment from low dose exposures. Continued follow-up of circulatory disease mortality and morbidity among Mayak workers up to the end of 2010, particularly for those first employed after 1958, would provide valuable extra information about whether raised risks might be found not only at relatively high doses but also for external gamma doses below 1 Gy and possibly below 0.5 Gy, as well as for relatively low plutonium exposures. Such a finding – taken together with information from studies of the Japanese A-bomb survivors and medically-exposed groups – might suggest that circulatory diseases should be included in estimates of radiation detriment. Furthermore, the follow-up of the Japanese A-bomb survivors has reported not only raised mortality from circulatory disease but also from other non-cancer diseases such as respiratory disease (Preston et al., 2003).
The project aimed to examine mortality and morbidity from circulatory diseases in the extended cohort with additional follow-up compared to previous analyses and for the first time to examine risks of respiratory diseases.

Assessing the risks of plutonium exposure in a pooled cohort of Mayak and Sellafield radiation workers
Previous studies of the Mayak workers have provided estimates of the effects of plutonium on the risk of cancer and non-cancer diseases at high exposure levels. To use these estimates to quantify risk in a low exposure situation, as experienced by workers today, requires assumptions about the extrapolation of risk from high to low doses from plutonium exposure that are based on little evidence. Analysis of the Mayak data in conjunction with data from other workers exposed to lower levels of plutonium would provide evidence for the way in which the high dose risks should be extrapolated to low dose situations. Workers from the Sellafield plant in the UK could provide such an additional low exposure cohort. A pooled dataset of these two groups would be the most statistically informative ever on the association of mortality and cancer incidence with occupational exposures to plutonium and external doses but this is dependent on demonstrating good compatibility of the two groups.
The project aimed to perform a feasibility study of the potential to carry out analyses of risks from lung cancer, leukaemia and circulatory disease in a pooled cohort of plutonium workers from Mayak PA and the Sellafield plant.
Two main areas of compatibility for pooling were addressed, considering the feasibility of generating both harmonised dosimetry and health data across the two cohorts. In addition the statistical power was estimated to ensure the pooling provided an increase over analysing the two datasets separately. Given the analyses were deemed feasible and statistically powerful, a preliminary analysis of the disease risks would be carried out.

Assessing the risks of in utero exposure in a pooled cohort of the offspring of female Mayak workers and of female members of the Techa River cohort
Data on the effects of in utero exposure are sparse. Studies of the effects of prenatal X-ray exposures have found raised risks of childhood cancer following in utero exposure but studies of atomic bomb survivors exposed in utero have been less clear about possible raised risks of cancer in childhood, although they do show a raised risk of solid cancer in adulthood. In particular, there is little information on the risk of cancer following protracted external and/or internal exposure in utero, particularly many years after exposure.
The project aimed to assess the risk of cancer following in utero irradiation in a pooled cohort of subjects born to female Mayak workers and to the females in the Techa River cohort. To accomplish this task the cohort of in utero offspring of the female members of the Techa River Cohort needed to be created and the in utero exposed offspring of female Mayak workers identified. The feasibility to combine these two datasets should be assessed – in particular the likely power of a pooled analysis. Then an analysis of cancer risks could be carried out for cancer types for which a sufficient number of incidences were observed.
Project Results:
Subproject 1: External dosimetry for Southern Urals Populations
The objectives of the subproject were to develop and implement strategies for validating the Mayak Worker Dosimetry System 2008 (MWDS-2008) and the Techa River Dosimetry System 2009 (TRDS-2009) that are used to calculate external doses for the Mayak worker cohort (MWC) and Techa River populations. The subproject also aimed to identify ways of improving these dosimetry systems. This work is important as the quality in terms of the statistical power of epidemiological studies based on these cohorts is related to the accuracy and reliability of radiation dose estimates.

The overall strategy for validating both these dosimetry systems was to perform comparative analyses of individual external dose estimates generated by the dosimetry systems, with measurements of absorbed dose in teeth by electron paramagnetic resonance (EPR) and dose to red bone marrow (RBM) obtained by measurement of translocations in peripheral blood lymphocytes with fluorescence in-situ hybridization (FISH). These ‘within individual’ comparisons were identified as the best strategy for dose validation, noting the variability of doses between individuals.

Further work on TRDS-2009 was undertaken by detailed measurements of absorbed dose in bricks from buildings in the abandoned village of Metlino, particularly from the church tower. Using these measurements and by reconstructing the current exposure geometry and that applying in 1949-1956, integral air kerma values at the Techa river shoreline in Metlino were estimated and compared with the external exposure estimates derived in the dosimetry system.


WP1: Validation of MWDS-2008
Development of the strategy
The strategy was based on the analysis of absorbed dose ratios between film badge doses, tooth enamel doses (as defined by EPR) and RBM doses (as defined by FISH) and allowed for the identification of exposures at workplaces and the separate estimation of the energy of photon exposures. The specifics of the strategy are described in the Deliverable 1.1.1. In summary, cumulative air kerma was calculated as the sum of the products of individual annual film badge doses and mean dose conversion factors (as defined in the dosimetry system) of the relevant exposure scenarios for each year. Cumulative air kerma from the EPR or FISH measurements were calculated as sums of the separate estimates for the three periods of monitoring with different types of film badges. In each period, the cumulative air kerma was calculated as the product of the fraction of absorbed dose in enamel or RBM in the period and mean dose conversion factors of the relevant scenarios of the period. The mean dose conversion factors for enamel and RBM were calculated using photon spectra and exposure geometry of a scenario as provided in the documentation for the dosimetry system. The distribution of air kerma ratios was evaluated between EPR and FISH, and the dosimetry system and between FISH and EPR.

EPR for Mayak workers
Tooth samples and data on tooth specification, exposure histories and scenarios of exposure that were required for the dose reconstruction were collected for 60 individuals who were Mayak PA workers and residents of Ozyorsk. The samples were transferred to either HMGU or ISS for the measurement of absorbed dose by EPR and the individual external dose estimation and the calculation of dose conversion factors for absorbed dose in tooth enamel to air kerma for specific exposure scenarios from MWDS-2008. A Monte Carlo approach was then used to derive distributions of dose conversion factors using scenarios from MWDS-2008. Deliverable 1.1.2 describes the preliminary analysis of the comparison of external doses and includes descriptions of the MWDS-2008 external exposure scenarios at the Radiochemical, Plutonium and the Reactor plants. Distributions of dose conversion factors depending on the characteristics of the radiation fields have been calculated for all scenarios that were relevant for the workers included in this project for the three types of film badges (IFK, IFKPb and IFKU) used between 1948 and 1992, and for incisors, canines and molars of upper and lower jaw and for RBM. Data on neutron and beta doses was also obtained in order to analyze their possible influence on EPR, FISH and film badge results.

FISH for Mayak workers
A group of Mayak workers was identified with characteristics of specific exposure scenarios, who had, according to their film badges, estimates of external dose that exceeded 0.5 Gy and for whom EPR measurements were available. For these individuals (40 Mayak workers and 10 Ozyorsk residents) samples of peripheral blood were taken. Lymphocytes from these samples were cultured and chromosomal slides prepared. FISH analysis, based on painting three pairs of chromosomes, was performed in three separate laboratories. Data on the yield of translocations in the three laboratories were pooled. Individual doses for each worker were estimated from the pooled data using calibration data for comparison with doses reconstructed from the film badge and EPR measurements.

Evaluation of experimental results and recommendations for improving MWDS-2008
The EPR/FISH/film badge measurements derived here were added to the database of similar measurements created under the SOUL project. For comparison of film badge, EPR and FISH based absorbed doses, values of air kerma were calculated using dose conversion factors corresponding to the MWDS-2008 exposure scenario in which workers were assumed to be exposed.

Exposure scenarios and periods of use of different film badge types were sub-divided into 15 groups: five exposure clusters defined by the radiation sources at the Mayak facilities (Radiochemical, Plutonium, Reactor plant) and predominant exposure geometry (anterior-posterior, AP, and isotropic, ISO), and by the three periods of use of film badge types IFK, IFKPb and IFKU. In spite of the different radiation sources involved in the various MWDS-2008 scenarios there was only moderate variability of dose conversion factors within (due to workers moving within workplaces) and between different scenarios with the same exposure geometry and same detector type (IFK, IFKPb, IFKU, molar, RBM). With the same exposure geometry and detector type, the standard deviation of the mean dose conversion factors was below approximately 20%. Larger differences in the dose conversion factors were obvious between different detector types and exposure geometries. For ISO geometry, the dose conversion factors varied by 80% from 0.6 (IFK) to 1.0 (IFKU), and for AP geometry by 140% from 0.8 (RBM) to 1.9 (IFK). For film badges and tooth EPR based estimates, dose conversion factors in the AP scenarios were about 80% larger than in ISO scenarios.

Air kerma estimates derived from EPR based absorbed dose estimates for 243 workers and derived from FISH based absorbed dose measurements for 125 workers were compared with air kerma estimated from MWDS-2008 (based on film badges). The majority (90%) of the workers involved in this study were considered in the MWDS-2008 as being externally exposed in ISO exposure geometry. Workers at the Radiochemical and Plutonium plants were exposed in scenarios with AP exposure geometry but only a few measurements were available from them and thus only tentative conclusions can be reached for these scenarios. Deliverable 1.1.3 describes the detailed results of the analysis.

For approximately 20% of the workers, significant additional photon exposure outside of the workplace or non-photon exposure was identified from the EPR or FISH measurements. For the other 80%, their main exposure was identified as due to photon exposures in the workplace. For 40% of these workers, their exposures as measured by EPR or FISH, were in close agreement with estimates from the MWDS-2008 scenarios. For the remaining workers, the lack of agreement was considered to be probably due to the workers having exposure to softer or harder photons than were assumed in the MWDS-2008 exposure scenarios. If this explanation is correct, exposures with softer photons were most frequent before 1954, whereas exposures with harder photons were most frequent after 1960.

For workers whose main occupational dose was due to photon exposures in the workplace, the air kerma estimates from MWDS-2008, enamel and RBM were in agreement within a factor of 2 with 95% confidence. These workers fell mainly into two groups, one comprising about 55% of the workers had very close agreement between air kerma estimates from MWDS-2008, enamel and RBM. The second group including about 25% with likely exposure in scenarios with soft photons had air kerma estimates from MWDS-2008 of about half those from enamel and RBM.

For a limited number of workers from Radiochemical and Plutonium plants with AP exposure who were involved in the production of highly purified plutonium dioxide after 1960, underestimation of the air kerma from MWDS-2008 by about 50% was observed. This could be the result of these workers having the majority of their exposure to photons with energy levels below 100 keV from plutonium and americium isotopes, and with less or no contribution from exposure to fission products as currently assumed.

The recommendations for further improvement of the external dosimetry system MWDS-2008 and the reduction of the uncertainties in individual dose estimates for MPA workers are described in the Deliverable 1.1.4. Improvements in dose estimates of up to a factor of 2 can be expected. The main source of current discrepancies was identified as likely to be due to inadequate consideration of soft photon exposures. To overcome this issue it will be necessary to update the exposure scenarios for the personnel involved in the production of plutonium dioxide in particular for those who were employed in this process after 1960. Furthermore, it was found that the treatment of non-routine exposures during repair work will require improvement. During repair work, releases of fission products with short half-lives including noble gases took place. These accidental releases are not taken into account in the current dose reconstruction methodology although according to preliminary calculations they could contribute appreciable fractions of the overall dose for some workers.

WP1.2 Validation of TRDS-2009

Development of strategy
The strategy was the same as that used to validate MWDS-2008 in that dose estimates derived from TRDS-2009 were compared to estimates based on EPR and FISH measurements. The specifics of the strategy are described in Deliverable 1.2.1.

EPR for Techa river residents
Enamel samples were prepared from 93 teeth from 50 permanent residents of the upper Techa region. EPR measurements were performed in three different laboratories on these samples to estimate absorbed dose. In addition, an estimate of the contribution to the total EPR-based absorbed enamel dose from strontium-90 (90Sr) was measured using the thermoluminesce method. The results of the EPR measurements from different laboratories were harmonized to account for method-specific biases and performances (uncertainties and detection limits). Combining the data obtained in this study with those data collected previously resulted in a total of 173 measurements from exposed donors being available, supported by assessments of the contribution to enamel dose from 89,90Sr incorporated into teeth. Details are provided in Deliverable 1.2.2.

FISH for Techa River Residents
Two groups of donors were investigated by FISH: (1) residents of the upper Techa region with an expected high contribution of external exposure; (2) residents of the lower Techa region who were predominantly exposed via ingestion of radionuclides (mostly 89,90Sr) from the river water and local foodstuffs. For these donors, samples of the peripheral blood were taken, lymphocytes were cultured and chromosomal slides were prepared. Data on the yield of translocations from three participating laboratories were pooled in order to obtain an adequate number of cells scored (802 genome-equivalent cells per donor) and from that a sufficiently accurate absorbed dose estimate. Combining the data collected in this study with that under the SOUL project gave FISH based measurements for a total of 97 exposed persons on which the analyses could be based.

Linear dependence between the radiation-induced translocation frequency and individual red bone marrow (RBM) dose from incorporated 89,90Sr, calculated with TRDS-2009, was observed for 26 donors who were predominantly exposed via ingestion of radionuclides. This allowed the quantification of the contribution of 89,90Sr to total radiation-induced translocation frequency for residents of the upper Techa region. The analysis of translocations from external gamma-exposure was performed on the remaining 71 donors. These results are described in Deliverable 1.2.3.

Evaluation of experimental results and implications for new dosimetry system

The comparative analyses of pooled EPR and FISH data with the doses derived from TRDS-2009 were performed in which the contribution of local confounding exposure from strontium radioisotopes incorporated in teeth and bones was subtracted using independent measurements of 90Sr-body burden and 90Sr concentration in dental tissues. The dose estimates derived from the EPR and FISH measurements were found to be consistent. At 7km from the release site the mean dose values were 510 and 550 mGy for FISH (based on 23 subjects) and EPR (based in 11 subjects) respectively. For villages located between 54-70km from the discharge site the mean values were 160 mGy (based on 34 subjects) for EPR and 130 mGy (based on 23 subjects) for FISH. The upper bound of individual estimates for both methods was equal to 2.2 –2.3 Gy.

Doses from external exposure were calculated for all donors using the Techa River Dosimetry System (TRDS-2009) from the data on contamination of the Techa River floodplain, and simulation of air kerma above contaminated soil while also taking into account age-dependent life-styles and individual residence histories of the donors. For comparison with the results of EPR and FISH, the estimates of dose from external sources were supplemented by the calculation of dose from internal exposure due to 137Cs incorporated in soft tissues of the donors. The calculations were performed in cooperation with U.S. colleagues involved in JCCRER project 1.1.

The EPR and FISH-based dose estimates were generally in agreement with the estimates of external exposure plus 137Cs internal exposure calculated with the TRDS-2009 within uncertainty bounds. The maximum difference between the EPR and FISH derived doses and the TRDS-2009 dose was observed for villages located 48-50km from the release site. The results will be used to update and improve the modules for external dose calculations in the TRDS. Details of the analysis of the results are provided in Deliverable 1.2.4.

WP1.3

Sampling, field and luminescence measurements
Brick samples were collected from the church tower and from the ruins in Metlino village. Dose rate mapping and in-situ gamma spectrometric measurements and an assessment of the exact church tower geometry were performed. Luminescence measurements of the samples and an assessment of the background and anthropogenic dose were then performed. TLDs were left at the site for one year to measure the current background dose.

A total of 44 brick samples were measured from 17 exposed and 2 background samples collected from three walls and at three different heights of the church tower. The SW and SE walls face towards the former Techa River shoreline, the NW wall faces away from the river, towards a contaminated area created by the installment of the reservoir lake in recent years. Anthropogenic doses from bricks in the NW wall ranged from 2.6 to 0.8 Gy and showed the same decrease with height as the annual doses measured by TLDs (12.2.to 5.1 mGy y-1). In contrast, doses from the SW and SE wall ranged from 0.9 to 1.1 Gy, with no clear height dependence, whereas the corresponding annual doses in TLD again showed a decrease with height, with values between 3 and 1.4 mGy y-1.

Dose rates in air 1m above ground were mapped up to 50m to the northwest and 10m to the southwest of the church tower and also at the sample positions. Measurement of sediment activity and water depth along four transects across the reservoir lake allowed the identification of the former two to three Techa river beds and of an uncontaminated part of a peninsula between the river arms. For the remaining area a homogenous contamination in the order of 1GBq m-2 was detected. Measurements were conducted by specialists of Kurchatov Institute, Moscow.

The effective half-life of 137Cs was assessed by the re-analysis of historic and contemporary water activity measurements and by using the brick doses of samples from the NW wall with the TLD values. The two approaches yielded half-lives of 19.1 ± 5.3 y and 18.7 ± 2.0 y, respectively.

Five exposed and three background bricks were sampled from the ruins of two former buildings (one a storage house) in 2013. Estimated ages of the brick slices varied between 121± 12 to 154 ± 17 years for the storage house and 108 ± 10 to 139 ± 12 years for the second building. Anthropogenic doses in bricks varied between 80±40 and 140±30 mGy, depending on the assumed sample age. No clear dose depth profile was identified. Annual doses measured by TLD ranges from 0.65 to 1 mGy y-1. Dose rate mapping in front of the ruins revealed a relatively homogenous dose rate in air, varying between 0.3 and 0.7 µSv h-1.

Radiation transport calculations
Models of the site geometry and source activity and distribution for the time period before the evacuation and for the period after the installment of Reservoir 10 were derived. Monte Carlo simulations of radiation transport were performed to calculate transition matrices in order to convert anthropogenic doses measured in bricks to integral air kerma values at reference locations.

Dose conversion coefficients to relate dose in TLD to dose in brick were calculated for the two different housing materials employed (Al and Cu). A highly detailed geometric model of the church tower and the NW wall of the granary was established and integrated into the Monte Carlo Neutral-Particle (MCNP) transport code. The source geometry of the present day contamination was constructed based on the dose rate mapping around the church and by fitting the source strength to the measured TLD doses at the brick sample positions. Using the effective half-life of 137Cs, doses accumulated in bricks since evacuation and installment of the reservoir lake could be calculated for all sample positions (so called “reservoir dose”) and ranged between 0.2 and 2.0Gy. By subtracting the reservoir dose from the measured anthropogenic doses, doses in bricks accumulated from 1949-1956 could then be inferred. This yielded doses between 0.3 and 0.9Gy for the SW and SE walls and doses that are compatible with zero dose for the NW wall. The latter result justifies the approach for calculating the effective half-life of 137Cs described above.

A model of the geometry in Metlino for the time-period before the evacuation (1949-1956) was created based on a recently declassified historic map on dose rate measurements from 1951-1954 and on the measurements conducted on the reservoir lake (water depth and sediment activity). Three source areas were assumed: a highly contaminated shoreline, a medium contaminated floodplain and an adjacent low contaminated floodplain. From the historic dose rate measurements, ratios of relative source strengths of the three areas of 71:3.6:1 were estimated. Using radiation transport calculations, dose conversion factors to convert doses in brick to air kerma at shoreline were derived.

Evaluation of experimental and modelling results and implications for the dosimetry system
From samples from the SW and SE walls of the church, air kerma values at the shoreline were estimated at 26 [17-35] Gy and 45 [35-45] Gy respectively. The former agrees well with the TRDS-2009 calculations of 19.8Gy whereas the latter is a factor of approximately 2 higher. Given the complexity of the reconstruction and the high degree of incomplete knowledge of the exact exposure geometry for the historic period, especially for the area facing the south-eastern wall, the experimental and modelling results can be regarded as an overall validation of the external doses predicted by the dosimetry system.

Estimates of a possible excess dose in bricks due to noble gas releases revealed values compatible with zero absorbed dose for all samples and with an upper limit of 30 or 80 mGy, depending on chosen background value. A significant external exposure of the population of Metlino due to noble gas releases can thus not be confirmed.

SOLO sub-project 2: Epidemiological studies of Mayak workers
Studies of both cancer and non-cancer diseases have been carried out in the Mayak worker cohort under previous EC and current US funded projects. This project expands these analyses using an enlarged cohort with extended follow-up which improved their statistical power to detect risks at low doses. In addition, analyses of respiratory diseases were performed for the first time.
WP2.1 Non-cancer morbidity and mortality among Mayak workers
Analyses of risks due to circulatory diseases among Mayak PA workers first employed during 1948-1982
Incidence and mortality risks from ischemic heart disease (IHD) (410–414 ICD-9 codes) and from cerebrovascular diseases (CeVD) (430–438 ICD-9 codes) were examined in the cohort of Mayak PA workers first employed during 1948-1972 and followed-up to 31 December 2005 using the “Mayak Worker Dosimetry System 2008” (MWDS-2008) and compared with previously published risk estimates in the same cohort based on the Doses-2005 dosimetry system. Significant associations were observed between doses from external gamma-rays and IHD and CeVD incidence and also between internal doses from alpha-radiation and IHD mortality and CeVD incidence. The estimates of ERR/Gy were consistent with those estimates from the previous studies based on the Doses-2005 system apart from the relationship between CeVD incidence and internal liver dose where the ERR/Gy based on MWDS-2008 was just over three times higher than the corresponding estimate based on Doses-2005 system. These results have been published.
Risks of circulatory diseases in the extended cohort of 22,377 Mayak PA workers employed up to 1982 and followed to 31 December 2008 were also analysed using the MWDS-2008 dosimetry system. Effects studied included incidence and mortality from CVD and IHD and mortality from all circulatory diseases (390 – 459 ICD-9 codes). By the end of the follow-up period, there were 7,225 and 8,717 incident cases of IHD and CVD respectively, 2,848 and 1,578 deaths from IHD and CVD respectively, and 5,010 deaths from circulatory diseases.
There were significantly increasing trends in CVD incidence with total absorbed dose from external gamma rays and total absorbed dose to the liver from internal alpha-particle radiation exposure after taking into account relevant non-radiation factors. The ERR/Gy estimates were 0.46 (95% CI 0.37 0.57) and 0.28 (95% CI 0.16 0.42) respectively, based on a linear dose-response model. The results of the categorical analyses of CVD incidence were in good agreement with a linear dose response for external gamma ray doses but for internal alpha-particle doses the picture was less clear. For the first time, an excess risk of CVD mortality was seen in workers with internal alpha-particle doses to the liver greater than 0.1 Gy compared to those workers who were exposed to doses of less than 0.01 Gy. A significant increasing trend for CVD mortality with internal alpha-particle dose was revealed in the sub cohort of workers exposed at liver doses <1.0 Gy after having adjusted for non-radiation factors, ERR/Gy = 0.84 (95% CI, 0.09 1.92).
There were significant increasing trends in IHD incidence in relation to total absorbed dose from external gamma rays (ERR/Gy = 0.18; 95% CI 0.11 0.27) and IHD mortality in relation to internal alpha-particle radiation to the liver (ERR/Gy = 0.30; 95% CI 0.10 0.59) after taking into account relevant non-radiation factors. The results of the categorical analyses were consistent with a linear dose response. The increasing linear trend in IHD incidence with internal absorbed dose to liver was seen among male (ERR/Gy = 0.20; 95% CI 0.07 0.36) but not female workers.
Taking all circulatory diseases as a single group, there was a significantly increasing trend in circulatory disease mortality with increasing total absorbed external gamma-ray dose among Mayak workers exposed to doses < 4 Gy (ERR/Gy = 0.07; 95% CI 0.01 0.14) which doubled after additionally adjusting for internal dose. There was also a statistically significant increasing trend with total absorbed liver dose from internal alpha-radiation using the full dataset (ERR/Gy = 0.27; 95% CI 0.12 0.48) and where the data set was restricted to workers with internal dose < 1 Gy (ERR/Gy = 1.20; 95% CI 0.66 1.89) but the estimate for internal alpha-radiation using the unrestricted data set was sensitive to adjustment for external dose.
Overall the risk estimates obtained in this study in relation to external exposure are in good agreement with estimates obtained for other large studies of nuclear worker cohorts as well as the atomic bomb survivor cohort (the Life Span Study).
Analysis of respiratory disease among Mayak PA workers first employed during 1948-1958 and 1948-1982
Chronic bronchitis (CB) was identified as the predominant chronic respiratory disease in the Mayak worker cohort. A detailed verification of the diagnoses of CB in the cohort of Mayak workers was undertaken for those first employed at one of the main facilities during 1948-1958 and followed-up to 31 December 2005. This showed that sufficient incident cases were verified to justify an analysis of disease incidence but the few deaths associated with CB did not justify a study of mortality.
Among the cohort of 12,210 workers first employed at one of the main plants of the Mayak nuclear facility during 1948–1958 and followed up to 31 December 2005, 1,175 incident cases of CB were verified. The analyses of non-radiation factors revealed that, as expected, the underlying risk of CB incidence increased with increasing attained age and was higher among smokers compared with never-smokers. An unexpected finding was of a high baseline CB rate before 1960 compared to later years. The cause of this is not clear but a number of factors may play a role including the later diagnosis of disease that was already present at first employment. Based on the post 1960 follow-up data, a statistically significant linear dose response relationship with cumulative external gamma-ray dose (ERR/Gy = 0.14 95% CI 0.01 0.32) was found. Using the same subset, but with an additional restriction to workers with cumulative internal lung dose below 1 Gy, a statistically significant linear dose response relationship with internal alpha-radiation lung dose was also found (ERR/Gy = 2.70 95% CI 1.20 4.87). In both cases, adjustment was made for non-radiation factors, including smoking and either internal or external dose as appropriate. There are no similar incidence studies with which to compare results. However, the most recent data from the atomic bomb survivor cohort (the Life Span Study) showed statistically significant excess mortality risk for respiratory diseases of 22% per Gy and this value is within the confidence bounds of the estimate of the risk from this study in relation to external dose.
CB incidence radiation risk was re-analysed in the extended cohort additionally including workers employed during 1959-1982 with follow-up extended to the end of 2008 based on the MWDS-2008 dose estimates and which also incorporated quantitate smoking information and information on prior exposure to risk factors for CB before employment at Mayak. During the follow-up period, 2,135 incident cases of CB were verified. As in the previous study, increased risk of chronic bronchitis was related to increasing attained age and was higher among smokers and those who reported that they were exposure to other occupational risk factors for CB prior to employment at Mayak PA. It was also found that the high baseline chronic bronchitis risk in the first decade of follow-up among workers employed before 1960 was not seen among workers first employed in subsequent decades. Based on follow-up data after 1960, a statistically significant linear dose response relationship with cumulative external gamma-ray dose (ERR/Gy = 0.14 95% CI 0.02 0.26) and internal lung dose (ERR/Gy = 1.14 95% CI 0.41 2.18) was found. From the same subset but with an additional restriction to workers with cumulative internal lung dose below 1 Gy the ERR/Gy estimate was slightly higher (ERR/Gy = 1.77 95% CI 0.72 3.16). In both cases, adjustment was made for non-radiation factors (gender, attained age, smoking status and plant) and either internal or external dose as appropriate. The risk estimates obtained for the cohort of earlier workers employed before 1959 and the extended cohort including recent workers agree well. However, after introducing a smoking index adjustment (instead of smoking status) and adjustment for exposure to harmful factors before employment at Mayak PA, the ERR/Gy in relation to external exposure was not statistically significant.
WP2.2 Cancer incidence among Mayak workers
Maintenance of the Medical-Dosimetry Register
The Medical-Dosimetry Registry has now been completed for 22,377 workers first employed between 1948-1982 at the reactors, radiochemical, and plutonium production facilities of Mayak PA. The follow-up of vital and health status was extended by 4 years to the end of 2008 and was verified for 95% of the workers. By the end of 2008, 41% had migrated from Ozyorsk and over 50% of the cohort had died. Cancer registration data could only be collected for residents in Ozyorsk for whom the information was obtained from the medical documents of oncology and pathology offices. All diagnoses were coded in compliance with the International Classification of Disease, ICD-9 and ICD-O. Quality control checks on the information including verification of histological subtypes were performed. A report on number of cases and sources of information for verification was prepared (Deliverable 2.2.1).
Analysis of cancer morbidity risk in the organs of main deposition for plutonium: lung, liver, skeleton
The cohort of 22,373 Mayak PA workers was analysed for those first employed at one of the three main plants between 1948 -82 and followed to the end of 2004 using organ doses from MWDS-2008. Among workers monitored for plutonium, there were 331 cases of lung cancer, 37 cases of liver cancer, and 10 cases of malignant neoplasms of bone and associated connective tissue.
There was clear evidence for a linear association between internal plutonium lung dose and the risk of lung cancer. For males, a significant internal plutonium dose response for all histological types of lung cancer was evaluated (adenocarcinoma, squamous-cell and other epithelial); the estimated ERR/Gy for adenocarcinoma was the largest (ERR/Gy= 32.5; 95% CI: 16.3; 71.9) about 11-fold higher than that for squamous-cell lung cancer (ERR/Gy=3.1 (95% CI: 0.3; 9.1). The relationship between liver cancer risk and plutonium exposure was best described by a linear-quadratic (LQ) function, but the size of the quadratic component reduced after restricting the dataset to workers with internal doses <2 Gy. Hepatocellular cancer was the most frequently observed type of liver cancer associated with internal plutonium exposure and hemangiosarcomas were exclusively observed only at high internal plutonium doses (>4 Gy). For malignant neoplasms of bone and associated connective tissues, there was no statistically significant trend in relation to internal plutonium dose, but a statistically significantly higher risk (RR=13.7; 95% CI= 3.0; 58.5) was found among unmonitored female plutonium workers who were employed in the most hazardous plutonium production facility commissioned prior to 1950. This work is now published (Labutina et al. Health Physics 2013, Deliverable 2.2.2).
Assessment of morbidity risk for solid cancers except for the lung, liver, skeleton
The 1948-82 cohort with follow-up to 2004 contained a total of 1447 cases of solid cancers other than those of the lung, liver or skeleton. A weak association was found between cumulative exposure to external gamma-ray and the incidence of other solid cancers (ERR/Gy= 0.07; 95% CI: 0.01 – 0.15) but this association lost its significance after adjusting for internal plutonium exposure. There was no indication of any association with plutonium exposure for other solid cancers. Among 16 selected individual cancer sites, there was a statistically significant association with external exposure for lip (ERR= 1.74; 95% CI: 0.37; 6.71) oesophageal (ERR= 1.49; 95% CI: 0.27; 6.09) and stomach cancer (ERR= 0.17; 95% CI: 0.003; 0.40) and with plutonium exposure for pancreatic cancer (ERR= 1.58; 95% CI: 0.17; 4.77). This work is now published (Hunter et al. Br J Cancer, 2013, Deliverable 2.2.3).
Assessment of morbidity risk for leukaemia, lymphomas and multiple myeloma
The 1948-82 cohort with follow-up to 2004 contained a total of 143 cases of all types of lymphatic or haematopoietic cancers. Of those, 77 cases were of leukaemia and 31 cases of non-Hodgkin lymphoma (NHL) and 11 cases of multiple myeloma (MM). The incidence of leukaemia excluding CLL showed a non-linear dose response relationship for external gamma exposure with exponential effect modifiers based on attained age, time since exposure or age at exposure. Among the major subtypes of leukaemia, the excess risk of AML was the highest within the first 2-5 years of external exposure and decreased substantially thereafter, but remained statistically significantly raised. In comparison, excess risk of CML first occurred five years after exposure and decreased from ten years after exposure, although the association was not statistically significant. No evidence was found of an association between leukaemia and occupational exposure to internal plutonium nor was there any indication of a relationship between either external gamma or internal plutonium radiation exposure and the incidence of Hodgkin or non-Hodgkin lymphoma or multiple myeloma. This work has been submitted for publication to the journal Cancer Research (deliverable 2.2.4).
Report on cancer incidence in the 1948-1982 cohort followed-up to 31.12.2008 using MWDS -2008
The extended follow-up of the 1948-82 cohort raised the overall number of incident cancers for analysis to 2,903, of which 330 or 11% occurred in the extra follow-up period between 2005 – 2008. Overall, 94% were solid cancers, and 6% were lymphatic or haematopoietic cancers.
The previously identified increased risk of lung and liver cancer in relation to internal plutonium-exposure was confirmed. The results obtained for lung cancer with regard to histological types indicated that all cell types (excl. small-cell cancer) were associated with plutonium exposure, but as for the previous analysis the excess relative risk for adenocarcinoma was greater than that for squamous-cell cancer, small-cell cancer or other epithelial cancers. For liver cancers, the hepatocellular histological type of carcinoma was associated with internal plutonium exposure, with hemangiosarcomas mainly observed at high internal plutonium doses (>4 Gy).
In relation to external exposures, a statistically significantly increasing trend was found for the incidence of lung cancer, solid cancer other than lung, liver and bone cancers, all leukaemia, leukaemia excluding CLL and for the leukaemia subtype AML (Deliverable 2.2.5).
WP2.3 Cause-of-death register of Ozyorsk
The object of this WP was to collect information to extend the Ozyorsk Cause of death register from 2008 to the end of 2010, to analyse the mortality structure of the register and to perform an inter-institutional comparison of cause-of-death coding in the register.
Analysis of the mortality structure of Ozyorsk population as of December 31, 2010
The mortality of Mayak PA workers and non-workers was compared to Russian national figures in the period 1998-2010 to show the current mortality situation. To examine changes over time, mortality in the period 1973-2010 was analysed as workers and non-workers could be identified during this period. These analyses were done using age-standardized incidence rates (European standard population), and applying joinpoint regression techniques.
In the recent period 1998-2010, there were 2043, 3614, 419, and 2601 deaths among male workers, male non-workers, female workers, and female non-workers respectively, leading to age-standardised mortality rates of 1332, 1733, 441, and 597 per 100,000 person-years for these groups. Mortality rates were generally lower than in the rest of the Russian Federation for both workers and non-workers. Comparing workers to non-workers, deaths from circulatory diseases and from all other causes of death were markedly lower (ratios = 0.75 0.74) in both sexes. Among the workers, rates of cancer were generally lower than national rates in men (ratio= 0.89) while they were comparable to the national rates in women (ratio= 0.96). For specific cancer sites, exceptions to the general pattern among men were cancer of other and unspecified sites (ratio =1.5) bone, skin and connective tissue cancers (ratio = 1.25) while lung cancer rates were similar to national figures. Among women, exceptions to the general pattern were found for respiratory tract and lung cancer (ratio 1.6) and genitourinary cancers (ratio 1.1). Rates were markedly lower for leukaemia (ratio = 0.4-0.6). Overall mortality increased by about 1% per year during the period 1973-2010 for men but was relatively stable among women. However, overall cancer death rates in men decreased by 1% per year among workers and by 0.6% in non-workers. The trends for cancer of the stomach (2% for workers and 3 % for non-workers) and of the lung (2.8% in both groups) were steeper. Among women, cancer death rates were less variable, but there was also a decrease in stomach cancer; however, the decrease in lung cancer was only in female workers. Breast cancer increased in both groups by the same proportion, whereas genitourinary cancers (cervix and uterus) increased among workers and decreased in non-workers (Deliverables 2.3.2 2.3.3 and 2.3.4).
The inter-institutional comparison of cause-of-death coding
A randomly selected sample of 251 death certificates (approximately 1%) from Mayak workers first employed at the main facilities in 1948-1982 was re-coded using ICD-9 separately by specialists from SUBI and for validation by experienced coders of the Danish Cancer Society Research Center. Of the 246 eligible death certificates, 233 were identically classified based on a main disease category matching (94.7%) and 182 based on matching of the first three digits of the ICD code (74.0%). The Kappa coefficient for the inter-rate reliability was 0.92 showing good agreement. There were 13 mismatches identified. Following a round-table discussion among the coders, six of these were identified as non-preferable coding in the register and three were identified as non-preferable coding during the validation. One case who died in 1950 was not classifiable without extra information and three cases for which no agreement was reached as to what code to use since both classifications appeared to be valid. Among discrepancies in the three digit categories, two main sources were the coding of multiple cancers and specific vs unspecific coding of cerebrovascular diseases (Deliverable 2.3.1).

SOLO sub-project 3: Pooled Analysis of Plutonium Worker Cohorts
The aim of the subproject was to perform a feasibility study of the potential to carry out analyses of risks from lung cancer, leukaemia and circulatory disease in a pooled cohort of plutonium workers from Mayak PA and the Sellafield plant in the UK (part of the BNFL cohort). Two main areas of compatibility for pooling were addressed, considering the feasibility of generating both harmonised dosimetry and health data across the two cohorts. If the analyses were deemed feasible, then the data would be gathered and a preliminary analysis of the disease risks carried out.
In addition during the course of the project an opportunity arose to complete and publish a paper on FISH analysis of plutonium exposed Mayak works that was started by Westlakes Scientific Consulting (WSC), part of the University of Central Lancashire, but that was not completed due to the closure of WSC.
WP 3.1: Feasibility study: harmonising dosimetry data
The aim of this work package was to develop the methodology and approaches to be used for dose reconstruction for the pooled epidemiological analysis of the Mayak and Sellafield worker cohorts and to ensure that the overall approach for dose estimation was harmonised as far as possible between these cohorts (Deliverable 3.1.5).
The feasibility of using early (1953-1970) plutonium urinalysis monitoring data for Mayak workers was investigated. It was concluded that these data are unreliable and that it would be preferable to conduct further monitoring of earlier workers and a strategy for this was developed. Modelling of plutonium excretion after administration of chelating agents was reviewed. A report summarising this work was produced (Deliverable 3.1.1).
An independent expert review of external dosimetry for the Mayak and Sellafield cohorts was conducted. The report of this review concluded that while there were some differences in ways in which estimates of external doses were produced they were not significant. In order to improve compatibility, further work was undertaken to develop factors that would permit existing whole-body external dose records to be used to estimate organ-specific doses for the Sellafield worker cohort. This was the first time organ-specific doses resulting from external exposures have been calculated and used within epidemiological analyses for the Sellafield cohort (Deliverable 3.1.2).
Current approaches to the reconstruction of doses to the systemic organs and to the lung from plutonium exposure were reviewed, developed and harmonised across the cohorts. Notable developments included the adoption of a revised structure for the ICRP Human Respiratory Tract Model (which will also form the basis of the forthcoming improved ICRP model) and work indicating that chemical binding of material in the lung had, at best, only a limited role to play in assessment of lung doses. A key issue was that of the lung solubility parameters for plutonium nitrate. Evidence with respect to plutonium nitrate solubility differs between Mayak and Sellafield. It was not possible to resolve this dichotomy with the time and resources available to the project, thus a pragmatic view was taken: two sets of plutonium nitrate absorption parameters were taken forward, one set based on the evidence from Mayak and the other set based on evidence from Sellafield and elsewhere. Both sets of assumptions will be used to produce two sets of plutonium doses for analysis. The importance of these parameters to the final risk estimates can then be assessed (Deliverable 3.1.3).
Methodology was developed to evaluate the uncertainties associated with estimates of dose, based on refinements to the Bayesian methodology that had previously been employed in the Sixth Framework Programme ALPHA-RISK project. Prior distributions were derived for all key internal dosimetry parameters. The final methodology for estimation of doses and associated uncertainties has been comprehensively documented in the internal dosimetry protocol for the study and will be published in due course (Deliverable 3.1.4).
WP 3.2: Feasibility study: harmonising health data
Key steps in this WP were the definition of a Research Study Protocol (RSP) (Deliverable 3.2.1) the assessment of health data comparability (Deliverable 3.2.2) and obtaining appropriate consents and permissions from stakeholders to undertake the study (Deliverable 3.2.3).
The comparability of the coding of causes of death was addressed in WP2.3 where a re-coding and validation exercise was undertaken on a sample of workers death certificates from the Ozyorsk cause of death registry. This study found good agreement with international coding standards and thus it was concluded that there was sufficient agreement in the coding of deaths between the cohorts to support the undertaking of the pooled analysis. A RSP was developed and agreed in October 2012. A long delay then occurred in gaining access to the Sellafield data due to the closure of Westlakes Scientific Consulting, which managed the cohort database. PHE took over management of the database but reinstating access to the data took considerable time and effort. A feasibility report on the viability of the pooled analyses was produced and reviewed by Dr Dale Preston (a member of the project Scientific Advisory Board) in October 2014. Dr Preston endorsed with some minor caveats the conclusion that the pooling was feasible. An ethical review of the study was carried out which had a positive outcome. As a result of this and the positive review of the feasibility report, permission was given to use the Sellafield data. Following the reinstatement of access to the database, work to develop datasets for analysis was completed by mid-2014, enabling the analysis work to commence thereafter.
WP 3.3: Calculation of dose estimates and consideration of uncertainties
Validated organ-specific dose estimates from plutonium exposure, and associated uncertainties, were calculated for the Mayak and Sellafield worker cohorts, based on assessment systems that employed the methodology specified in the common internal dosimetry protocol developed under WP3.1. The computer codes developed/adapted for this work, PuMA and IMBA UA (Uncertainty Analyser) for the Sellafield cohort and Pandora for the Mayak cohort, were extensively validated both in isolation and against each other. This demonstrated that the dose and uncertainty assessment systems were working as expected, and implemented the agreed common methodology (Deliverable 3.3.1).
Two sets of doses were produced for each of the 11,603 Sellafield and 7,499 Mayak plutonium workers using the two different lung solubility parameter assumptions for Mayak and Sellafield (see WP3.1). The effect of the lung solubility parameters was most pronounced on the lung dose estimates with median doses for Mayak PA of 29mGy and 19mGy and for Sellafield of 0.85mGy and 0.22mGy based on the Mayak and Sellafield assumptions respectively. These results indicate that attempting to improve understanding of solubility in the lung should be a major aim of future dosimetry research in this area. With doses for 19,102 individuals, this is by far the largest plutonium dose data set ever produced for epidemiological analysis (Deliverable 3.3.2).
Work Package 3.4: Preliminary pooled analyses
Preliminary epidemiological evaluations of potential associations of both internal (plutonium) and external (gamma) radiation exposure with lung cancer and leukaemia incidence and mortality, and with cardiovascular disease mortality, for the pooled Sellafield-Mayak worker cohort was performed.
The pooled cohort comprises 45,817 workers from the two enterprises; 23,443 radiation workers first employed in 1946-2002 from the Sellafield Workers Cohort (SWC) and 22,774 radiation workers from the Mayak Worker Cohort (MWC) first employed at the main plants in 1948-1982. The period of follow-up is terminated at the end of 2008 for Mayak workers who were Ozyorsk city residents, and at the end of 2005 for Sellafield workers and Mayak workers who had emigrated from Ozyorsk.
The epidemiological analyses of lung cancer incidence and mortality was carried out in relation to external and plutonium doses to the lung while that for leukaemia incidence and mortality was performed in relation to doses to the red bone marrow (RBM) and for circulatory disease mortality doses to the liver were used. The datasets analysed were aggregated tables containing the number of study events, number of person-years, and mean values of studied factors tabulated by categories of radiation and non-radiation factors.
Leukaemia analysis
There were 94 cases of leukaemia excluding CLL (60 MWC and 34 SWC) and 116 deaths (87 MWC and 29 SWC) in the pooled cohort dataset. Plutonium exposure dose estimates were only available for 59 workers with leukemia (excluding CLL) incidence (25 MWC and 34 SWC) and 65 deaths (36 MWC and 29 SWC) . The study of background rates of leukaemia incidence and mortality demonstrated comparability of age-related trends between the cohorts and variation in background by year of birth and gender.
The excess relative risk associated with radiation dose, of both external gamma radiation and internal alpha-radiation from Pu, was estimated taking into account all of the available non-radiation factors: those affecting background rates were sex, attained age and birth cohort, while the only non-radiation factor affecting the radiation risk estimates was attained age.
Radiation risk estimates for external gamma exposure were comparable between the MWC and SWC both in separate dose intervals and across the whole dose range. The average linear excess relative risk estimate was approximately 3.0 per Gy of accumulated dose to red bone marrow, which is similar to the results of past investigations of Mayak and Sellafield workers, and of other groups of people exposed to low LET radiation. The final model for leukaemia incidence and mortality in the pooled cohort contained a quadratic dose response with attained age effect modification terms which showed a decrease in the risk estimate with increasing attained age. There was no evidence of heterogeneity of the risk estimates between the MWC and SWC and analysis showed that there was a statistically significant risk of leukaemia excluding CLL incidence across a wider dose range in the pooled cohort than that in each cohort separately.
There was no evidence of any relationship between leukaemia risks and accumulated red bone marrow dose from internal alpha-exposure due to incorporated Pu-239. However, the number of leukaemia cases in the pooled cohort with positive plutonium doses was low (34 deaths and 35 incidences) and the range of doses among these cases was relatively small (especially in the SWC) so the power to detect a significant dose response effect was low while the power to detect a difference in risk between the two cohorts was extremely low (Deliverable 3.4.1).

Lung cancer analysis
There were 893 cases of lung cancer (509 MWC; 384 SWC) and 1,195 deaths (789 MWC; 406 SWC) in the pooled cohort dataset. Among those monitored for plutonium exposure, there were 747 incident cases of lung cancer and 817 deaths. The study of background morbidity and mortality lung cancer rates demonstrated comparability of age-related trends between the cohorts, but a difference in variation in background by year of birth or the calendar period. However, information about smoking status was not available in the framework of this project and the differing temporal trends in smoking rates in the UK and the Russian Federation may well explain these differences in birth cohort effects.
The excess relative risk associated with radiation dose, of both external gamma radiation and internal alpha-radiation from plutonium, was estimated taking into account all of the available non-radiation factors: those affecting background rates were cohort (i.e. MWC and SWC), sex, attained age and birth cohort, while those affecting the radiation risk estimates were sex and attained age. The pooled radiation risk analysis, in terms of the cumulative internal alpha-radiation dose to the lung from plutonium, revealed compatible risk estimates to those obtained for the two cohorts separately.
Uncertainty surrounding the best estimate for the value of the parameter representing the rate of slow solubility for plutonium nitrate in the two cohorts meant that two sets of plutonium lung doses were generated for use in the analysis. One set based on an estimate of the slow solubility parameter derived from observations on Mayak workers (ss_mayak) and the other based on investigations of volunteers’ by PHE (ss_sel). Examination of the plutonium dose-response for lung cancer incidence found significantly increased risks at relatively low dose for the SWC, 2-5 and 5-10 mGy using ss_mayak, and 1-2, 10-20 and 20+ mGy using ss_sel. In the MWC, an increased risk was only observed at relatively high doses (200-500 mGy using ss_mayak and 50-100 mGy using ss_sel). As in previous MWC studies, the plutonium dose-response was found to be linear across the whole dose range and the point estimates of risk were consistent down to a relatively low dose range when restricting the range of plutonium dose included in the analysis. For example, in the incidence analysis a significant pooled ERR/Gy estimate was detectable at 0.2 Gy when using ss_mayak and 0.1 Gy when using ss_sel, with the risk estimates (ERR/Gy) down to 0.05 Gy positive and consistent with the overall estimate.
The study of potential effect modifiers on the plutonium ERR/Gy such as, attained age, gender and age at first plutonium exposure was hampered by the lack of power in the SWC. For MWC, and, as a consequence, for the pooled cohort, it was revealed that gender and attained age were significant factors affecting the value of the risk estimate. Gender was found to significantly modify the plutonium risk estimate in the MWC with the Mayak females having risks 4 times higher than Mayak males for incidence and 2-3 times higher for mortality. The SWC male risk estimate was compatible with that of the MWC males but the number of female lung cancers was very low (10 deaths, 8 incidences) in the SWC and models that allowed the risk estimate to vary by gender within the SWC converged poorly. In relation to attained age, a declining pattern in the plutonium risk estimate with increasing attained age was observed in both cohorts, although the power to detect this effect in the SWC was lower (p=0.10 in the SWC for mortality), the scale of this effect was very similar in both cohorts (e.g. for lung cancer incidence using ss_sel the age effect was Exp(-3.04*log(age/60) for MWC and Exp(-5.85*log(age/60) for the SWC).
The final pooled risk estimates for the cohort are within the range 5-8 per Gy for males at age 60 for both mortality and incidence and using both solubility assumptions. The risk estimates are very similar in magnitude to those obtained in previous studies of the Mayak PA workers. The ERR/Gy estimates in relation to external gamma-radiation were in the range 0.2-0.4 per Gy of gamma-dose to the lung, which is similar to the results of past investigations of Mayak and Sellafield workers, and of other groups of persons exposed to low LET radiation.
The estimates of risk, for both external and plutonium exposure obtained in this study show comparability of risks between the cohorts which suggest that the pooling of cohorts is sensible and leads to increased statistical significance and allows the study of a wider range of doses. The risk values obtained for both types of radiation allow an estimation of relative biological effectiveness (RBE); a point estimate from the results of this investigation is within the range of 10-30, which can be considered, in view of uncertainties in dose and risk estimates, as a broad confirmation of the value of 20 currently adopted in radiological protection as the radiation weighting factor for alpha-particles.
Circulatory diseases analysis
The radiation-induced risk of circulatory disease was assessed on the basis of mortality modelling only as incident cases were not available for the SWC. Analyses were conducted separately for the main circulatory disease categories, ischemic heart disease (IHD) and cerebrovascular disease (CeVD), as well as cardiovascular disease (CVD) as a whole. For the plutonium analyses, dose to the liver was used in the absence of reliable knowledge of the relevant target tissue for CVD induced by low-level radiation exposure. In the absence of reliable knowledge on what was an appropriate latent period for CVD following low-level exposure to radiation, lag periods of 5 and 10 years were adopted for the purposes of analysis. For a 5-year lag period, the overall number of CVD deaths among radiation workers was 7445 (MWC: 5123; SWC: 2322), with 4465 (MWC: 2905; SWC: 1560) deaths from IHD and 2048 (MWC: 1610; SWC: 438) deaths from CeVD. For workers monitored for exposure to plutonium (and for a 5-year lag period), the overall number of CVD deaths is 2787 (MWC: 1627; SWC: 1160), with 1744 (MWC: 963; SWC: 781) deaths from IHD and 748 (MWC: 533; SWC: 215) deaths from CeVD.
The analysis was complicated by a number of factors, including the different background rates of CVD mortality relevant to the two worker cohorts, in particular the high rates of CVD mortality experienced in Russia during the 1990s. Also, for the pooled analysis, no information on important background risk factors (such as smoking) was available.
Four basic dose-response models to assess ERR/Gy were employed. Each of these was stratified by sex, attained age, calendar period and (for Mayak) Ozyorsk migration status, and individual models then further adjusted for plutonium exposure (in the gamma analyses) or gamma exposure (in the plutonium analyses), year of first employment, or both of these as appropriate Dose groups of 0-50, 50-200, 200-750 and 750+ mGy for gamma dose, and 0-2, 2-20, 20-100 and 100+ mGy for plutonium dose, were used, with the lowest dose groups used as reference categories. Additional analyses were performed with doses truncated at 4 Gy for gamma dose and at 1 Gy for plutonium dose. Finally, in addition to analyses with plutonium workers followed up from first plutonium monitoring, analyses were conducted with the first two years following first plutonium monitoring treated as unmonitored. Sets of results were generated for both plutonium lung solubility assumptions: Mayak and Sellafield.
For the MWC, the ERR/Gy for gamma exposure tended to be positive and of marginal statistical significance for CVD and IHD, and of variable sign and non-significant for CeVD, whatever the lag period, although associations tended to be stronger for IHD. The ERR/Gy for the liver dose from plutonium were generally non-significantly positive.
For the SWC, the ERR/Gy was significantly positive for gamma exposure for CVD and IHD, and tended to be non-significantly positive for CeVD, whatever the lag period. The ERR/Gy for the liver dose from plutonium were largely positive, but convergence problems meant that the statistical significance of these trends could not be assessed.
The statistical power of the Mayak data in comparison to the Sellafield data produced results for the pooled dataset that were broadly a reflection of the findings for Mayak alone. However, the Sellafield ERR/Gy values were notably greater than those for Mayak for CVD and IHD, although convergence problems for plutonium doses for the Sellafield cohort made an assessment of these exposures difficult. The lung solubility assumption did not affect the overall pattern of results.

Additional deliverable D3.4.4: FISH analysis of Mayak workers

The aim of this work was to determine if translocation yield has the potential to be a useful tool in the validation of red bone marrow doses resulting from mixed exposure to external and internal radiation. This study was completed and published and its results suggest that translocation yield does have the potential to validate red bone marrow doses. This work complements that of SP1 and provides additional evidence that the doses to red bone marrow from both the external and internal exposures used in the analysis of the leukaemia risk in the pooled Sellafield and Mayak are valid for use in epidemiological analyses.

SP4:
The aim of this subproject was to assess the risk of cancer following in utero irradiation in a pooled cohort of subjects born to female Mayak workers and to the females in the Techa River cohort. Studies of the effects of prenatal X-ray exposures have found raised risks of childhood cancer following in utero exposure (Monson R.R. MacMahon B., 1984; Gilman E. et al., 1988) but studies of atomic bomb survivors exposed in utero have been less clear about possible raised risks of cancer in childhood, although they do show a raised risk of solid cancer in adulthood (Preston D. L. et al., 2008). However there is little information on the risk of cancer following protracted external and/or internal exposure in utero, particularly many years after exposure.

The work was planned as four work packages. WP 4.1 comprised an analysis of the cancer morbidity in the Techa River In utero exposed offspring (TRIUC), as well as a feasibility study to check the viability of the proposed pooling project. WP 4.2 comprised an analysis of the cancer morbidity of in-utero exposed offspring of female Mayak workers (MWOC) (an additional task not originally planned but a necessary precursor to the pooled analysis), and conditional on a favorable result of the feasibility study, an analysis of the cancer risk in the pooled cohorts of TRIUC and MWOC. Both WP4.1 and WP4.2 used the outputs of WP4.3 in which fetal doses to the study subjects were calculated. Postnatal doses were already available from previous research. In WP4.4 the vital status of the members of the (TRIUC) was updated with an additional four years of follow-up.

WP 4.1 Analysis of the cancer risk in the Techa River in utero cohort and feasibility of the Pooled analysis of Mayak and Techa River in utero cohorts

The Techa River in utero cohort (TRIUC) was defined and created. It included all offspring born between Jan 2, 1950 and September 30, 1961 to women residing at any point in time in 41 villages along the Techa River between January 1, 1950 and December 31, 1960, after their mothers’ exposure to the radioactive discharged from Mayak PA into the river. As of August 2013, the TRIUC numbered 11,574 members (50.6% men; 49.4% women), the mean age of cohort members who were alive was 57 years. Tartars and Bashkirs comprised 34%, and Slavs 66% of the cohort. About 63% of the cohort members were Chelyabinsk Oblast residents, 17% were Kurgan Oblast residents while about 20 % of the cohort members had migrated outside the mortality catchment area (Chelyabinsk and Kurgan Oblasts) by the end of 2009. An extended cohort was also defined (ETRIUC) which included all offspring born to women residing at any point in time in the 41 villages of the Techa River between January 1, 1950 and December 31, 1960, after their mothers’ exposure to the releases. This expansion increased the cohort to 20,045 children. The mean age of live cohort members in this cohort was 49 years with 64% still alive and 18% known to be dead while 19% had migrated from the catchment area. Catchment areas and periods were different depending on cancer type (hematological malignancies or solid tumors) and morbidity or mortality analyses.
The analysis of the incidence of solid cancer and haematological cancers in the TRIUC was based on 26 cases of haematological cancer (with 426,618 person years) and 242 cases (with 381,948 person years) of solid cancer. Estimates of excess relative risk were derived based on models that included separate independent variables for in utero and postnatal doses. No statistically significant excess risks were found in relation to either dose. The point estimates of the risks were all close to zero (deliverables 4.1.1 4.1.3).
The Mayak female Worker Offspring Cohort (MWOC) comprised all offspring born in Ozyorsk between 1948 and 1988 to women who were members of the Mayak worker 5-plants cohort, which includes all persons who worked for any period of time between 1948 and 1982 at one of the 5 main plant types of Mayak PA (reactors, radiochemical, plutonium production, water treatment and mechanical repair plants). The cohort comprised 8466 individuals of which 4361 were male. Around 60% of this group was born between 1948 and 1960 - the period of highest exposures at Mayak. Vital status of the MWOC participants was known for 97% of the cohort; 46% were alive and 11% had died while living in Ozyorsk by the end of 2009. Around 35% had migrated from Ozyorsk and were thus lost to follow-up.

The feasibility of the pooling of the TRIUC with the MWOC was evaluated. As both cohorts were from the same time period and the same area, with very similar methods for follow-up of the subjects, pooling was feasible. The predicted statistical power to detect excess risks was sufficiently high for high risk scenarios to justify the completion of the study. It was substantially higher when the cohorts were combined and the power to detect associations was higher in the cancer incidence analyses compared to cancer mortality analyses for haematological cancers as well as for solid cancers. However, there was little benefit from the inclusion of the ETRIUC over and above the TRIUC in the analyses due to the low numbers of death in this population (deliverable 4.1.2).

WP4.3 Internal dosimetry for those exposed in utero
Individual doses of in utero and postnatal exposure for members of TRIUC and the MWOC cohorts due to both external and internal exposures were estimated, based on the biokinetic and dosimetric models for in utero exposure to radionuclides of concern, principally isotopes of strontium and plutonium. A total of seven Urals-based models were constructed at the following foetal ages, in weeks post-conception: 12, 18, 22, 26, 30, 34 and 38. Because radionuclides are retained in body tissues of the new-born infant, it was also necessary to develop postnatal models and also consider the transfer of radionuclides in breast-milk from mother to infant. The biokinetic model for strontium transfer to breast milk for the Techa River population was improved to take into account an increased maternal intake of calcium and 90Sr with diet during lactation and maternal adaptations of bone mineral metabolism during partial breastfeeding. Doses of gamma postnatal exposure are individualized based on individual data on age, gender and history of residence in the Techa River settlements and on the EURT areas (Degteva et al, 2009) (deliverables 4.3.1 4.3.2 4.3.3).
Doses of in utero exposure for TRIUC were calculated using the TRDS-based computer code. The code allows calculation of in utero doses from two pathways: (1) maternal intakes of the radionuclides in the Techa River villages before and/or during pregnancy and (2) external exposure due to maternal proximity to contaminated Techa River shoreline and floodplain soils during pregnancy.

For internal in utero exposures, doses were calculated using data on the maternal intakes of radionuclides and dose coefficients for fetal organ doses following maternal intakes of radionuclides. Data on the intakes of radionuclides in the Techa River settlements were obtained in the framework of JCCRER Project 1.1. The dose coefficients represent cumulative in utero doses to the red bone marrow and all soft tissues from a unit (1 Bq) intake of strontium radioisotopes by the mother. The latest radionuclide ‘S’ values and a biokinetic model for strontium transfer to the fetus were used for the refinement of dose coefficients from maternal intakes of 89Sr and 90Sr. In the calculation of dose coefficients from 90Sr, the unit intake was applied to a female of different ages at the beginning of the intake and different times relative to the delivery. For 89Sr, the unit intake was considered to occur at different intervals between one month and five years before delivery and applied to an adult mother. The refined values of the dose coefficients considered irradiation of the target organs by the radioisotopes retained in foetal soft tissues and the skeleton. For 137Cs, dose coefficients were as evaluated previously using models from ICRP Publication 88 (ICRP 2002).

For external in utero exposures, the doses were calculated using data on external exposure rates in the Techa River villages obtained in the framework of JCCRER Project 1.1 (Shagina et al. 2012c), using maternal regimes on contaminated areas and conversion factors from the absorbed dose in air to the absorbed dose in the maternal uterus which was used as a surrogate for the evaluation of foetal doses.

Absorbed organ doses for members of the MWOC were also calculated, considering in utero and postnatal doses from plutonium-239, as well as in utero doses from external gamma-exposure for the members of MWOC were also calculated (deliverable 4.3.4).

WP 4.4: Cause-of-death registry for Techa River population
In the framework of the EU funded SOUL project, a cause of death registry was created for the Chelyabinsk and Kurgan Oblasts which comprise the catchment areas for the Techa River Cohort. This registry includes 148,000 deaths registered in that area from 1950 to 2005. It provided the cause of death information for the TRIUC and ETRIUC cohorts. The major aim of this WP was to update the follow-up of the register by four years to 2009. Cancer registrations were obtained from the URCRM cancer registry which covered the catchment area. For Ozyorsk residents, cause of death was obtained from the Ozyorsk registry of deaths and cancer incidences maintained by SUBI.
The work of updating the Chelyabinsk and Kurgan cause of death registry was made difficult by the enactment of new privacy laws in the Russian Federation. These laws required the development of new sources of information which was particularly challenging but an additional 12,144 deaths were identified which occurred between 2006 and the end of 2009 bringing the total deaths on the register to 159 837. An analysis of the structure of cause-of-death in the register and its dynamics over the sixty-year period was conducted. Mortality levels from all causes, as well as from the cardio-vascular and cancer in the catchment area, do not differ markedly from the average level in the Russian Federation over the period 1965-2009. However, it was observed that the number of cases of death from ill-defined causes increased over time (deliverables 4.4.1 4.4.2 4.4.3).

WP 4.2: Analysis of cancer incidence in the Mayak worker offspring cohort and pooled analysis of Mayak and Techa River in utero cohorts
During the course of the project, it was found that while a separate analysis of mortality in the in-utero exposed offspring of female Mayak workers had been published, a comparable morbidity analysis had not. Information from such an analysis would provide useful information regarding the feasibility of the pooled analysis. Thus, this analysis was also undertaken (deliverable 4.2.3).
The analysis of the incidence of solid cancer and haematological cancers in the MWOC was based on 177 cases of solid cancer and 32 haematological cancers (13 leukaemia cases) during the 277,002 person-years at risk among 8466 subjects. In 2009, 52% of the cohort members were still alive and resident in Ozyorsk. The mean gamma ray dose among the 3470 in utero exposed offspring was 58 mGy. Among the 1912 offspring with positive plutonium in-utero exposure, the average in utero dose to the red bone marrow was 0.9mGy and on average they accumulated 2.5 mGy post-natally.
For haematological cancers, the analysis of the data suggested an increased risk associated with gamma rays in dose categories above 20mGy. The linear estimate of excess relative risk (ERR) was 1.11 (95% confidence interval (CI) 0.11-3.41) per 100mGy of gamma dose. The in utero red bone marrow dose from plutonium was not associated with increased risk of these cancers, neither were postnatal doses. For solid tumours, the analysis of the data did not show any statistically significant excess risk with either in utero dose or post-natal external dose to the stomach.
The results of the dosimetry calculations (WP4.3) and the ascertainment of follow-up (WP4.4) and the results of the individual cohort analyses were used to assess the statistical power (WP 4.1) of a pooled cohort analysis to detect mortality and incidence risks from solid cancers and from hematological diseases. The results of the analysis of the pooled cohorts were as follows.
For the hematological diseases, due to the relatively small number of events, the two cohorts were pooled without restrictions leading to a total of 19,536 individuals with 700,504 person years follow-up of which 9,701 (50%) were still alive at the end of follow-up. Overall 26% were lost to follow-up due to migration from the study area. There were 2762 deaths and 508 incident cancers.
The analysis of the 58 hematological cancers was based on estimated doses to red bone marrow with no adjustment for plutonium exposure. Post-natal doses were lagged by one year up to age 15 and two years subsequently. Estimates of in utero and postnatal dose were less than 1mGy for 25 and 30 of the 58 cases respectively. Estimates of excess relative risk (ERR) were calculated based on models that took into account confounding factors of ethnicity, period of birth, sex and age. For hematological cancer incidence, the estimate of linear ERR (per 10mGy) from in utero exposure was 0.08 (95% CI 0.00 0.26) and for postnatal exposure 0.02 (95%CI <0, 0.11). Results for all leukaemia incidence (based on 28 cases) gave a smaller risk from in utero exposure 0.04 (95% Ci <0, 0.23) and zero risk 0.00 (<0, 0.06) from postnatal exposure. The estimate of mortality risk from in utero exposure of hematological cancers was similar to that for incidence at 0.02 (<0, 0.12) but for postnatal exposure it was smaller at 0.01 (<0, 0.07).
For the solid tumors analyses, only the participants of the two cohorts who were born between 1950 and 1961 were pooled to improve the homogeneity of resulting dataset leading to a total of 15,813 individuals with 554,411 person years follow-up in the incidence analyses and slightly more in the mortality analyses. Of these 7,093 (45%) participants were still alive at the end of follow-up. Overall 33% were lost to follow-up due to migration from the study area. There were 2130 deaths and 425 incident cancers.
The analysis of 369 solid cancers was based on estimated dose to the stomach with no adjustment for plutonium exposure. Postnatal doses were lagged by one year up to age 15 and five years subsequently. Estimates of in utero and postnatal dose were less than 1mGy for 195 and 161 of the 369 cases respectively. Estimates of excess relative risk (ERR) were calculated based on models that took account of confounding factors of ethnicity, period of birth, sex and age. The categorical analysis showed a pattern of decreasing risk with increasing dose of in utero exposure, with a negative estimate of the excess relative risk of -0.25 (95%CI -0.61 to 0.32) in the category above 80mGy, although no relative risks were significantly different from one. The analysis of postnatal exposure showed a pattern of increasing relative risk with dose compared to the baseline category of <1mGy for postnatal exposures (the relative risk in the 80+mSv category was significantly high (RR=1.72 95%CI 1.12 2.57). Considering the 92 digestive cancer cases as a group it was possible to estimate linear ERR values per 10mGy of 0.02 (95%Ci <0, 0.13) and 0.10 (95%CI 0.02 0.26) for in utero and postnatal exposures respectively in relation to stomach dose. For the group of 68 respiratory cancers, the estimate of linear ERR was zero for in utero exposure and no trend was observed in the categorical risks analysis. The group of 49 breast cancer did not show any significant association but it was based on small numbers.
In summary, for hematological cancers the small number of cases observed prevent firm conclusions being drawn. There was a suggestion of excess risk with in utero exposure in incidence analyses but not the mortality data. No associations with postnatal dose were identified. The analysis of leukaemia did not reveal differences in risk compared to lymphoma.
For solid cancers no association with in utero dose was identified. Analyses of individual cancer types (digestive, respiratory and breast) revealed indications of heterogeneity between these types but the evidence was not strong.


Potential Impact:
Epidemiological studies of radiation risks in exposed populations have provided the quantitative information that forms the basis for protection standards applied worldwide, principally as recommended by the International Commission on Radiological Protection (ICRP). Epidemiological studies continue to provide the best approach to gain an increased understanding of risk under different circumstances of exposure of workers and members of the public to ensure that protection is adequate and commensurate. Funding agencies have an obligation to prioritise support for such proven approaches. If risks are underestimated, protection will be insufficient to adequately safeguard health, while overestimation of risks will lead to the potential for substantial unwarranted financial burdens on industry and other operations using radiation and radioactive materials. Aspects of particular importance addressed in this project are risks associated with internally incorporated radionuclides compared with external radiation, risks of non-cancer diseases (circulatory and respiratory) as well as cancer, and risks from in utero exposures.

Key developments and their impact are:
• Establishment of a joint cohort of Russian and UK plutonium workers
• Establishment of a joint cohort of individuals exposed in utero as a result of their mothers’ work at the Rusisan Mayak plant or residence along the nearby Techa river
• Evidence of similar lung cancer risks relating to plutonium exposures of Russian and UK workers, with values consistent with estimates based on studies of the effects of external radiation and the use of a value of 20 for the relative biological effectiveness of alpha particles compared with gamma rays
• Evidence that while external radiation can cause leukaemia, plutonium-239 is not an effective leukaemogen despite irradiation of the peripheral red bone marrow
• Evidence that risks of protracted radiation exposures of workers may include circulatory and respiratory disease and that protection standards may requires adjustment accordingly
• Preliminary evidence that the risks of in utero irradiation have not been substantially underestimated by current assumptions based on limited data
• Substantial improvements in dose estimates for external and internal exposures and associated improvements in the reliability of risk estimates, including:
o improvements in the reliability of EPR measurements on tooth samples as a dose assessment methodology
o insights into the use of FISH chromosome aberration assays on blood lymphocytes to measure external and internal exposures
o development of improved models for the behaviour and dosimetry of radionuclides in body tissues, notably for assessing doses from inhaled plutonium-239 in the respiratory tract, and from strontium-90 in fetal tissues following intakes by the mother during pregnancy.


Examination of the work of the project

The key outputs of this project were designed to provide improved information about the risks of cancer and non-cancer health effects in workforces exposed to protracted radiation and for cancer following in utero exposure, thereby provide a more robust scientific basis for underpinning radiation protection standards.

Radiation protection standards are based in the most part on information derived from the Life Span Study (LSS) cohort comprising survivors of the atomic bombs that were dropped on Hiroshima and Nagasaki at the end of the Second World War. This cohort was exposed to a single relatively large and acute exposure to external gamma radiation but appropriate safety standards need to protect people exposed to low and or protracted exposures and also to exposure from radionuclides taken into the body, for example, by inhalation. Further, although some atomic bomb survivors were exposed while in utero the information provided by the cohort on the subsequent risks to these people from cancer as a result of their exposures is limited.

Thus, while the LSS cohort has provided a great deal of data on radiation risk it is deficient in information in a number of areas that are key to ensuring the protection of people who are exposed to radiation occupationally and to the general public who may be exposed to either natural sources of radiation or planned exposure as a result of, for example, power generation or unplanned exposures as a result of accidents, for example, the Fukushima incident.

Occupational and population radiation protection standards in Europe are based on the recommendations of the International Commission on Radiological Protection (ICRP). For low dose and low dose rate exposures, protection standards are based on risk estimates extrapolated from high dose and/or and high dose rate exposures, and for protracted exposures they are extrapolated from acute exposures. However, it remains uncertain if these extrapolations are appropriate, the potential health risks associated with protracted exposure to low doses and low dose rates could be underestimated, and therefore under-regulated, but evidence has been building that risk from chronic exposure at low doses does not differ much from that of acute exposure. These issues are particularly important since protracted low dose exposures are the most frequent form of exposure to radiation workers and to the general public.

Furthermore, the recommendations of ICRP concerning risks from internal radiation exposures from alpha radiation are not based on direct evidence but are estimated using those from external gamma radiation with the use of a conversion factor to take account of the difference in the effects between the two types of radiation. Current knowledge on health risks due to in utero exposures is also limited and guidelines for both occupational and medical exposures to expectant mothers and the fetus would benefit from additional evidence.

The health impact from exposure to radiation has been recognized to include cancer risks but in recent years concern has been raised about whether radiation exposure might also increase the risks of non-cancer diseases, in particular circulatory and respiratory disease. Evidence for these effects is sparse, particularly as many of these diseases are chronic in nature and incidence studies are more informative.

The SOLO project aimed to provide direct evidence concerning the issues described above that standard setting bodies could use to improve international radiation protection standards. The Project aimed to do this using two radiation exposed cohorts from the Southern Urals of the Russian federation; one of workers first employed at the Mayak Production Association facility between 1948 and 1982 and the second of residents of villages along the Techa River between 1948 and 1960 and who as a result were exposed to radiation discharges into the river from the Mayak plant. In addition, data from a cohort of UK radiation workers from the Sellafield plant were also used.


Assessing the reliability of the dosimetry systems used to estimate radiation exposure in the Mayak and Techa River cohorts

The reliability of the risk estimates provided by epidemiological studies is highly dependent the quality and quantity of the data on which they are based. Among the most important factors for radiation effects studies are the accuracy of the estimates of the radiation dose and of the follow-up information from which the risk estimates are derived. The current dosimetry systems used to estimate external dose in the Mayak and Techa river cohorts are known as Mayak Worker Dosimetry System 2008 (MWDS-2008) and Techa River Dosimetry System 2009 (TRDS-2009). These systems have evolved over many years to be as accurate as possible but important issues remain. SOLO continued work begun under a previous project to verify the external dose estimates provided by these systems using two methods of biological dose reconstruction; Electron Paramagnetic Resonance (EPR) measurements of teeth and Fluorescence In Situ Hybridisation (FISH) measurements of peripheral lymphocytes in blood. In addition, SOLO undertook a programme of environmental measurements at one of the Tech River villages, Metlino, to estimate the historic anthropogenic exposures and thereby improve estimates of doses to residents.

The outcome of this work was that it was possible to verify the individual external doses reconstructed by EPR, FISH and workplace doses based on MWDS-2008 (film badges) to within factor 2 with 95% confidence for around 80% the Mayak workers. For the remaining 20% it was determined that the set of assumptions made in the dosimetry system about how they were exposed in the work place did not fully take account of their occupational exposures. For example, they may have been exposed to radionuclides that were not recorded or exposed in situations that were not taken account of in the standard exposure scenarios.

For the TRDS-2009 dosimetry system the FISH and EPR based measurements were generally in agreement with the estimates of external exposure from TRDS-2009 within uncertainty bounds when the additional internal exposure from 137Cs to the residents was taken into consideration. The maximum differences in the estimates were found for villages around 50km from the plant. The estimation of historic exposures from the environmental measurements at Metlino were also found to broadly agree with those of TRDS-2009 and it was possible to test whether residents had been exposed to noble gas releases from the plant.

The impact of this work was to provide evidence that the external doses, derived using the two dosimetry systems, as used in the epidemiological studies of the effects of radiation on the risks of both cancer and non-cancer diseases in these populations carried out in this project are reliable. It also provides reassurance that future studies will be based on sound dosimetry calculations.

Assessing the risks of cancer and non-cancer diseases in the Mayak cohort

Epidemiological studies were carried out to estimate both incidence and mortality risks from cancer and selected circulatory and respiratory diseases in the Mayak worker population. These studies provided direct evidence of the effects of protracted exposures, from both external exposure and internal exposure from plutonium.
Furthermore the ability to estimate circulatory and respiratory disease incidence risks is unique in such a population and very useful since these diseases are mainly chronic in nature and while they have high underlying incidence rates they have relatively low mortality rates. Additionally this cohort has information on important confounding factors such as smoking, alcohol consumption, hypertension and body mass index which are known risk factors for these diseases and are important to take into account when estimating radiation risks.

Clear evidence for a linear association between lung cancer incidence and plutonium exposure to the lung was seen for all histological types of lung cancer but with the largest for adenocarcinoma for which the risk estimate was about 11-fold higher than that for squamous-cell lung cancer. Importantly, for the justification of current radiation protection standards the overall risk for plutonium induced lung cancer was consistent with comparable risk estimates for external gamma radiation when account was taken of the current ICRP recommended value for the relative biological effectiveness of alpha particles emitted by plutonium compared to gamma radiation. No excess risk was seen for other solid cancers or for leukaemia in relation plutonium exposure.

A significant excess risk of chronic bronchitis incidence was found in relation to internal liver exposure after taking account of smoking and prior exposure to known occupational risk factors. This risk was evident at relatively low doses and is important since no other occupational based studies have been able to estimate this risk. No excess risk was identified in relation to external exposure.

Excess risks of incidence of both ischemic heart disease (IHD) and cerebrovascular disease (CeVD) were found in relation to external and internal exposure although the excess risks for IHD were restricted to males. Additionally while the risks in relation to external exposure were in very good agreement with a linear dose response relationship the form of the dose response for internal exposure was less clear.

This work provides direct evidence for ICRP and other bodies to help determine if the current radiation protection standards which are only designed to protect from excess risks of cancer are also sufficient to protect from excess risks of IHD, CeVD and chronic bronchitis incidence which was identified for the first time as being potentially inducible by exposure to external radiation.


Assessing the risks of plutonium exposure in a pooled cohort of Mayak and Sellafield radiation workers

A pooled dataset was compliled to study the effects of protracted external exposure and internal exposure to plutonium among occupationally exposed radiation workers was compiled. This dataset uniquely provided the opportunity to investigate risks across a wide range of doses and with maximum possible statistical power. It also provided the opportunity to compare risks across the two populations so the possibility that the risks seen in the Mayak cohort were confined to that particular cohort could be assessed.

Three disease groups were investigated: firstly, lung cancer which is the main cancer disease for which evidence of plutonium risk has been seen in the Mayak cohort; secondly, leukaemia which, while it has not been seen to be related to plutonium exposure in the Mayak cohort, has been found to be related to external gamma radiation exposure among the atomic bomb survivors and which is considered to be one of the cancers most sensitive to induction by external radiation; and thirdly, circulatory diseases which again have been seen as significantly related to occupational external and internal exposure in the Mayak cohort.

For lung cancer, linear estimates of risk obtained for both internal plutonium exposure and external gamma exposure were sufficiently similar to support the conclusion that risks are comparable for both cohorts. This means that the combined dataset could produce a statistically significant risk estimate over a wider dose range than for either cohort alone. Importantly for radiation protection standards, the estimates of risk for external and internal exposure allowed an estimation of the relative biological effectiveness (RBE) with respect to lung cancer to a range of 10-30, which can be considered, in view of the uncertainties both in the estimation of risk itself and of dose measurements and estimation, as a broad confirmation of the value of 20 currently used by ICRP. As large numbers of radiation workers (and other occupationally exposed groups such as hard rock miners) are exposed to alpha radiation, this provides reassurance that the current protection standards will provide an appropriate level of protection for these workers.
For leukaemia, the extra statistical power of the combined dataset did not reveal an excess risk of leukaemia incidence related to internal plutonium dose to the red bone marrow. While there was an excess risk in relation to external gamma radiation in the Mayak worker cohort, the lower statistical power for the Sellafield cohort alone did not result in a significant excess.
For circulatory diseases only risks from mortality could be estimated (as only mortality data were available for the Sellafield cohort). The statistical power of the Mayak data in comparison to the Sellafield data produced results for the pooled dataset that were broadly a reflection of the findings for Mayak alone. However, the Sellafield risk estimates were notably greater than those for Mayak for CVD and IHD.
Assessing the risks of in utero exposure in a pooled cohort of the offspring of female Mayak workers and of female members of the Techa River cohort
A unique cohort of in utero exposed offspring of female Techa River cohort members, comprising almost 11 500 people, was formed and followed-up for up to 60 years. Mortality and incidence of malignant neoplasms were analyzed, and their dependence on in utero exposure was estimated. A feasibility study showed that the combination of this cohort with the cohort of offspring of female Mayak workers generated a statistically more powerful dataset with a substantial proportion of subjects with relatively high in utero exposure.
Overall, these two cohorts provide a good future basis for obtaining reliable results on the cancer risk associated with in utero exposure to ionizing radiation. The issue preventing these cohorts having a significant impact on radiation protection standards at this time is the relatively young age of the cohort’s members and the low number of incident cases of cancer that have occurred to date.
Risk from haematological cancer and from all solid cancers as a group were evaluated, as well as for some individual types of solid cancers. However, the risk estimates that could be calculated are challenging to interpret, given the competing risk factors for many major types of solid cancers, the impact of early detection and improvement in treatment on incidence and mortality for many major cancers, and the possible differences in magnitude in risk related to ionizing radiation for the different cancer types. For haematological malignancies, results for in utero exposure are suggestive of a weak association, but confidence intervals were wide. For solid cancers, no association was seen.
An additional 10 years of follow up for these cohorts can be expected to lead to a marked increase in the number of cases as the cohort members reach an age where the underlying incidence of cancer in the population starts to rise markedly.
Overall, the generation and pooling of the two cohorts is an important development and although it has limitations, the risk analysis provides some information on the potential impact of in utero exposure to ionizing radiation in that it suggests that the risks of in utero exposure are not markedly different from those of postnatal exposure. The pooled cohorts have the potential to provide reliable results with further follow-up of subjects as they enter age groups in which cancer becomes more common.