Final Report Summary - SEMI-NUC (Prospective cohort study of residents near the Semipalatinsk nuclear test site - feasibility)
Executive Summary:
TThe Semipalatinsk nuclear test site (SNTS) covers 18,500 km2 in the north east of Kazakhstan. From 1949–1989 the Soviet Union conducted nuclear tests at the SNTS, including detonations above ground and in the atmosphere that produced radioactive contamination in the soil and air. Some of these releases caused radioactive plumes that exposed people and territories adjacent to the site. In the past two decades, groups of researchers have initiated two cohort studies to assess the health effects of exposure to fallout from Soviet nuclear weapons testing among residents living near the SNTS. These studies used different control groups and different dosimetric methods and produced conflicting results that are of limited use for evaluating the true effects of the exposure.
The SEMI-NUC project was a support action that brought together scientists from Europe, Kazakhstan, and Japan with a common interest in the health effects from chronic exposure to low doses of ionizing radiation.
One aim of the project was to test record linkage possibilities between the two previously-established cohorts (so-called “new” and “historical”) to assess the feasibility of creating a larger, unified cohort for future long-term studies. It was postulated that he resulting large population could be used for a long-term, prospective follow-up study of the health effects from exposure to fall-out in the Semipalatinsk area.
During the course of the project it was established that although record linkage between the two databases that contain the information on the new and historical cohorts is possible, it would be very labour intensive due to the different structures of the databases. It was also discovered during a site visit that one of the partners, the Scientific Research Institute of Radiation Medicine and Ecology (NIIRME), possesses a registry of exposed population around the SNTS that is larger than both the previously established cohorts combined. All information stored in the NIIRME database is verified to avoid duplication and technical errors and it is kept updated continuously. It was concluded that a cohort formed from the individuals whose information is stored in the NIIRME database would be the most suitable for a long-term, prospective follow-up study of the health effects from exposure to fall-out in the Semipalatinsk area.
The ultimate aim of this project was to produce a feasibility report to determine if a long term, prospective follow-up study of the residents near Semipalatinsk can be developed. The feasibility report contains the background and evidence that was collected during the course of the project that led the project partners to the conclusion that proves that this is indeed the case and recommends using a cohort selected from the NIIRME registry for the study. The report also includes information that will be the foundation for the follow-up study, including descriptions of follow-up mechanisms; assessment of dosimetry; description of biological material to study, mechanisms that underpin the effects of low-dose radiation and an outline of the health outcomes to be studied.
Project Context and Objectives:
• Introduction
The health effects of exposure to fallout from Soviet nuclear weapon’s testing among residents living near to the Semipalatinsk nuclear test site in Kazakhstan have not been thoroughly investigated in the past. Two previous studies that were conducted on two independent cohorts, the so-called “historical” and “new” cohorts produced contradicting and non-conclusive results. These studies also involved differing control groups and different dosimetric methods.
The SEMI-NUC project was developed in accordance with the Strategic Research Agenda of MELODI, a European Platform dedicated to low dose radiation risk research. The project was designed as a feasibility study to investigate the possibility of merging the previously-established historical and new cohorts into a single, large cohort for future studies of the health effects of long-term exposure to low-dose ionizing radiation in these populations. The merging of the two existing cohorts into a single, larger cohort would have the advantage of avoiding unnecessary duplication of effort and resources in future studies plus the benefit of increased statistical power. In addition, there are not many populations worldwide that have been exposed to protracted low-dose radiation and that are suitable for studies on the morbidity of cardiovascular diseases. The large unique cohort of individuals from around the SNTS has the potential to prove that prospective epidemiological research on the association between radiation and cardiovascular diseases is feasible.
Other populations exposed to low to moderate dose radiation, such as in Fukushima or elsewhere, can benefit from the outcomes of studying the unique Semipalatinsk cohort and the results will contribute to a better understanding and quantification of radiation risks for low to moderate chronic doses.
• Creation of a unique cohort for future prospective studies
To create a single combined cohort from the two existing cohorts, it was necessary to know and understand the data that is available for the historical and new cohort members in the respective databases. Site visits were organized to the two partner institutes (NIIRME and NNC) where the data for each cohort is kept to thoroughly investigate the structure and composition of the different databases and the mechanisms used for identification and follow-up of cohort members.
Evaluation of the databases at NIIRME and NNC revealed that each has a different structure and contains different kinds of information about individuals. In order to conduct record linkage of the two databases, it was necessary to create another version of the NIIRME database with just one record per person.
To test a record linkage procedure an initial linkage of a random sample of 3,000 records selected from each of the two databases was performed using RecLink3 software. Linkage variables included first name, last name, patronymic name, gender, and date of birth. It was concluded that despite their different structures, a complete linkage of both databases would be possible, but very labour intensive. It was estimated that one person-year would be needed for a complete linkage.
However, it was decided during the course of the project that the linkage of the two databases would not be needed for the study. This decision resulted from the discovery made during one of the site visits that NIIRME possesses a registry of exposed population around the SNTS that is larger than both the previously established cohorts combined. Moreover, all information stored in the NIIRME database is verified to avoid duplication and technical errors and it is kept updated continuously.
It was concluded that a cohort formed from the individuals whose information is stored in the NIIRME database would be the most suitable for a long-term, prospective follow-up study of the health effects from exposure to fall-out in the Semipalatinsk area and hence it was no longer thought necessary to merge the two previously-established cohorts.
• Identification of potential health outcomes
A second major objective of the project was to identify the potential health outcomes (cancer and non-cancer diseases) that can be studied in a future prospective study using the unified cohort described previously. NRPA in collaboration with NIIRME and NNC worked on the evaluation of mechanisms of data collection and the availability and quality of the various information sources of mortality and morbidity data to identify potential health outcomes for future prospective studies of radiation-related risks. Mechanisms of data collection on death cases and causes, and information sources on mortality data were identified, followed by an evaluation of mechanisms of data collection, information sources and data quality for both incidence and mortality data.
A quality assurance protocol was developed that included procedures for case ascertainment and abstraction of information on both death and incident cases of cancer and non-cancer diseases. The protocol describes the procedures for data collection, data entry and data management. The data quality was checked including consistency of diagnostic criteria within the same disease and correctness of disease coding using International Classification of Disease (ICD) and ICD for Oncology (ICD-O).
Potential health outcomes for future studies were identified based on a detailed inventory of all available information sources for different types of health outcomes (cancer and non-cancer, incidence and mortality), followed by assessment of data quality and completeness including a review of case ascertainment procedures, diagnostic criteria, morphological verification and cross-checking the information from the different sources. After reviewing sources of information and data availability on cancer incidence, we consider feasible a future study of cancer incidence in relation to early life exposure. Because information on adulthood cancer risks following in utero and early life exposure is limited (mainly based on studies of A-bomb survivors and offspring of Mayak female workers and Techa River population), we consider as a priority setting up a prospective cohort study of radiation-related cancer risk in a sub-cohort of people exposed in utero or early life. We also recommend a study on cardiovascular diseases based on the national screening examination data. It is feasible to study stroke and CVD outcomes in the Semipalatinsk population. Ongoing national screening programs on CVD will ensure uniformity of diagnostic criteria and compliance with international approaches resulting in reliable quality of CVD diagnoses. Because data on behavioural and other risk factors are collected as a part of the screening examination procedure, it will allow inclusion in the study additional information on non-radiation risk factors. A follow-up of the NCI ultrasound-detected thyroid nodules prevalence study was also considered as the most feasible and informative study. Because of systematic follow-up of population in the NIIRME registry and regular collection of information on death cases and causes by NIIRME researchers, it is feasible to carry out studies both on cancer and non-cancer mortality risks in a cohort of people exposed around the SNTS. The detailed report can be read in Deliverable 3.1.
• Evaluation of follow-up mechanisms
Possible follow-up mechanisms for a future prospective study were reviewed and evaluated. The characteristics of the different follow-up mechanisms used in the previous studies using the historical and new cohorts were summarized based on existing publications and reports and other relevant information. Alternative sources of information for follow-up were evaluated and discussed. After completing these tasks, follow-up mechanisms were suggested for a future prospective study. The mechanism of follow-up was established and proved to be efficient and not discriminative upon disease and exposure status. We consider availability of IIN for each Kazakh citizen as an advantage for follow-up that facilitates the follow-up process and allows expansion of the study follow-up area to the territory of entire Kazakhstan.
After reviewing the information, we confirmed that ZAGS (population registry) records should serve the main official source for mortality follow-up, i.e. for ascertaining date and cause of death in exposed population. The results of our comparison between three different datasets (historical cohort, new cohort, and NIIRME registry) showed that the NIIRME’s registry of exposed population is a comprehensive source of information that can be used not only for establishing a cohort but also for obtaining vital status data for a future prospective study. A report on identified and verified follow-up mechanisms for a prospective study was produced (Deliverable 4.1). In addition, a comparison of information on follow-up results on selected samples from the historical cohort, the new cohort, and the NIIRME Registry of exposed population was carried out.
• Identification of key pathways and mechanisms for dose formation
The main pathways for external and internal exposure to the public living in areas of radioactive fallout due to nuclear weapons testing were identified for both external and internal exposure taking into account the geographical location of the population and lifestyle factors such as diet, occupation, and materials used for housing.
• Assessment of the approaches used previously for individual external and internal dose reconstruction, including evaluation of the parameters used for assessing cumulative doses
The advantages and weaknesses of the two methodologies (called the Russian methodology and U.S. methodology) used in the previous studies were assessed in detail based on existing reports and publications.
A third methodology - the Joint U.S.-Russian Methodology – has been developed jointly by U.S. (U.S. National Cancer Institute, U.S. Department of Energy and University of Utah) and Russian (Institute of Biophysics) specialists on fallout dosimetry taking into account the advantages and weaknesses of both the U.S. and Russian methodologies. It is currently the best method available for reconstruction of individual external and internal doses hence it is this method that is recommended for use in the future study.
• Development of an approach for the assessment of individual cumulative doses to the exposed population
The objective was to prepare a proposal for a unified individual dose reconstruction approach for the public exposed to fallout from nuclear weapons testing at the SNTS that can be used in the future epidemiological study.
The joint U.S.-Russian methodology was selected for retrospective reconstruction of individual external and internal doses to the population permanently living in the vicinity of the SNTS. In the framework of the current pilot project, it was possible only to identify and describe the main sources of uncertainty in the dose estimates. A strategy for estimation and inclusion of uncertainty in the dose estimates is in a progress. It is proposed to use the recently developed two-dimensional Monte-Carlo procedure for future long-term study to enable separation of shared and unshared errors. The instrumental methods of retrospective dosimetry (Retrospective Luminescence Dosimetry- RLD, and Electron Spin Resonance Dosimetry - ESR) were selected and tested for validation of calculated external doses. It is proposed to use these instrumental methods for validation of calculated doses in the future extended study.
Project Results:
By demonstrating the feasibility of using a unique cohort from the data held by NIIRME, the SEMI-NUC project has prepared all the main aspects for conducting a full epidemiological study on residents near the Semipalatinsk test site. The results of such a study would greatly contribute to the quantification of health risks from protracted low and moderate doses of radiation.
The Semipalatinsk site is particularly interesting in this sense because of an initial high exposure during the nuclear testing and a subsequent protracted exposure due to the radioactive fallout. The dose range for members of the cohort vary from very low to moderate doses and the outcomes that could be studied include both cancer and non-cancer diseases. In particular, there is increasing interest in studies of the risk of cardiovascular diseases following exposure to low-dose protracted radiation as this may have important implications for radiation protection of general public and patients exposed to radiation.
The review of the dosimetric tools considered will be valuable for other researches involved in dose-effect studies of other population groups, since good dosimetry is essential for epidemiological studies.
Proposals for follow-up research
• Follow-up of ultrasound-detected thyroid nodules
The most feasible and near-term study would be a follow-up of ultrasound-detected thyroid nodules prevalence study. A first round of screening of about 3,000 people was performed by the National Cancer Institute (NCI), USA, in collaboration with Kazakh scientists in 1998, and since then no attempts of organized systematic follow-up of the persons diagnosed with thyroid nodules has been made. For each screening participant there is also a study questionnaire recording their residential history, medical history and dietary habits with special attention to their consumption of milk and milk products. Individual thyroid doses due to external and internal exposures were reconstructed for each study participant.
Given gaps in current understanding of both the natural history of development of thyroid nodules, their potential to transform into neoplasms, and mechanisms of radiation-induced thyroid carcinogenesis, it is important to initiate a follow-up of the people with detected thyroid nodules and also those who were nodule-free at the time of the initial screening.
To perform a longitudinal study assessing the development of thyroid nodules the following steps are required:
a) Reach a data access agreement with the NCI to have access to the list of study participants;
b) Link the list of study participants
i. with the NIIRME registry of exposed people to ascertain their vital status and residence history;
ii. with the population-based cancer registry in the Semipalatinsk regional oncology centre to identify cases of thyroid cancer among them; and also
iii. with the database of the Semipalatinsk clinical-diagnostic centre to identify those among them with thyroid nodules.
The linkage result will serve as a basis for the risk evaluation of neoplastic transformation of benign nodules into thyroid cancer.
After the first phase, depending on the percentage of deceased, migrated or lost-to-follow-up study subjects, a screening exercise can be planned to assess the current thyroid status including the presence of thyroid nodules in this cohort.
• Pilot study on cardiovascular diseases
We recommend a pilot study on cardiovascular diseases based on national screening examination data. It is feasible to study stroke and CVD outcomes in the Semipalatinsk population. To carry out such study, it will be necessary to:
a) perform a linkage of the NIIRME registry data with the database on the 26,093 exposed individuals examined at the clinical-diagnostic centre in the framework of the established collaboration.
b) collect information on risk factors and examination results of exposed individuals who are patients of the clinical-diagnostic centre.
c) collect information on CVD screening results, including information on diet, lifestyle and other risk factors, from the local medical institutions in the catchment area, i.e. East-Kazakhstan region, through linkage between the NIIRME registry on exposed people with the database on screened patients.
The proposed study will have a prospective nature because data on CVD incidence are available from 2011 onward. Ongoing national screening programs on CVD will ensure uniformity of diagnostic criteria and compliance with international approaches resulting in reliable quality of CVD diagnoses. Because data on behavioural and other risk factors are collected as a part of the screening examination procedure, it will allow inclusion in the study additional information on non-radiation risk factors. Availability of IIN will ensure precise and effective linkage collecting this information on the members of NIIRME registry who were screened for CVDs.
• Cohort study on cancer risk following early life exposure
We suggest a cohort study on cancer risk following early life exposure including exposure in utero. After reviewing sources of information and data availability on cancer incidence, we conclude that a future study of cancer incidence in relation to early life exposure is feasible. However, taking into account existing limitations in cancer incidence data, for the successful implementation of such a study the following would be needed:
a) Identify potential members of a cohort with early life exposure using the NIIRME registry of exposed population;
b) Collect data on cohort members’ residence history and vital status using the NIIRME registry of exposed population;
c) Establish collaboration with the Semey regional oncology centre to identify cancer cases in the NIIRME registry members through linkage with the population-based cancer registry. Cancer follow-up is nearly complete from 2003 onward, but information in the period from 1980–2002 is of questionable quality. No cancer incidence data available before 1980;
d) Reconstruct individualised doses due to external exposure using all available sources of information including historical data on exposure rate measurements, individual residential history, time spent outdoors, etc.;
e) Perform a dose-response analysis for solid cancer and leukaemia cases using radiation epidemiology statistical tools and software.
Because information on adulthood cancer risks following in utero and early life exposure is limited (mainly based on studies of A-bomb survivors and offspring of Mayak female workers and the Techa River population), we consider setting up a prospective cohort study of radiation-related cancer risk in a sub-cohort of people exposed in utero or early life as a priority. Nested case-control studies for specific cancer sites could be considered in the future.
• Prospective molecular epidemiology study of cancer risks following early life exposure
The currently available biobank data on exposed individuals are not sufficient to carry out a molecular epidemiology study. We therefore recommend establishing a sustained collaboration with the cancer oncology centre and the clinic-diagnostic centre to collect blood and tumour tissue samples for members of the NIIRME registry of exposed population. Expanding the collection of biological samples will make it possible to conduct a prospective study of cancer risks following early life exposure as described above.
The proposed action will also open up new collaborations outside existing European networks of low-dose radiation research, having built up a very close collaboration with the scientists from Kazakhstan, Japan and also the US. It has also improved the capacity for conducting high-quality radiation research in Kazakhstan through the involvement of local scientists who benefitted from participation in the support action to better exploit their competences.
Through clarifying the Intellectual Property Rights for the database shared jointly between the Institute of Radiation Safety and Ecology of the National Nuclear Centre of Kazakhstan (NNC) and the Radiation Effects Association (REA) of Japan, it was possible to reach agreement for this data to be made available to researchers for a future study. Kazakhstan partners NNC also agreed to make their version of this data available to a wider scientific population for future research.
Potential Impact:
In the long term, this project and a subsequent full study, has a potential improve our understanding of the health risks associated with low and moderate chronic doses of radiation. A better quantification of risks from low and protracted doses is valuable for regulating authorities so that exposure due to planned, existing or emergency exposure situations, such as Chernobyl or Fukushima, can be better regulated. It will also serve as a valuable reference for making public health decisions.
The dissemination of the knowledge arising from the project was achieved through presentations at scientific conferences, a dedicated project dissemination meeting, publications and other forms of communication such as websites. The table in Annex 1 presents a list of the webpages containing news and information about the project.
• Webpages
Webpages containing news and information about the project are attached following the table in Annex 1.
• Publications
Grosche B., Zhunussova T., Apsalikov K., Kesminiene A. Studies of Health Effects from Nuclear Testing near the Semipalatinsk Nuclear Test Site, Kazakhstan. Central Asian Journal of Public Health, Vol 4(1), 2015. http://cajgh.pitt.edu/ojs/index.php/cajgh/article/view/127/203
Zharmukhambetova A.K. Lipikhina A.V. Mansarina A.E. Zhakupova Sh.B. Bright Yu.Yu. Integration of data from the medical registry of Kazakhstan population exposed to low doses of ionizing radiation into the European platform MELODI. Abstracts of XI International scientific conference «Ecology. Radiation. Health», August 28-29, 2015, Semey state medical university, 2015, p. 20. http://ssmu.kz/ru/pages/default/view?id=sci-downl
World Hypertension League (WHL) Newsletter, No. 147, March 2015, p.5.
http://www.thaihypertension.org/files/314.Newsletter%20No.%20147,%20March%202015.pdf
• Project meetings, dates and venues
06-08/05/2013, IARC, Lyon, France: Project kick-off meeting: Prospective cohort study of residents near the Semipalatinsk nuclear test site – feasibility assessment and health risk from exposure to radioactive fallout from nuclear testing in Kazakhstan at the Semipalatinsk nuclear test site. This meeting was organized together with the meeting co-sponsored by IARC and the National Cancer Institute to overview the previous studies conducted on health effects in Semipalatinsk. It helped to identify methods and approaches used in the previous studies
17-20/09/2013, Semey, Kazakhstan: Site visit; visit to Astana to the Ministry of Health of Kazakhstan – to present the project and discuss accessibility of existing data and registries to European investigators
09-10/06/2014, BfS, Munich, Germany: SEMI-NUC Mid-Term project meeting, External Advisory Board meeting.
30/03/2015-02/04/2015, Site visit to Semey, Kazakhstan: visit to NIIRME, Semey and NNC, Kurchatov to learn more about the data in the registries.
11-12/05/2015, External Advisory Board (EAB) meeting, IARC, Lyon, France: The Work Package leaders presented the work achieved and the External Advisory Board made comments and suggestions as to how to resolve problems and about the best way to proceed.
26/05/2015, Kyoto, Japan: the study coordinator and Dr Grosche attended a Meeting with the president of REA to discuss the IPR for the new cohort data.
09-10/11/2015, BfS, Munich, Germany: Project dissemination meeting and follow-up meeting.
09/03/2016, Fukushima, Japan: meeting at the Fukushima medical university to discuss future longitudinal study of thyroid nodules in populations around the SNTS
14-15/03/2016, NRPA, Oslo, Norway: Final project meeting with partners and EAB.
The reports from these meetings are attached in Annex 2.
List of Websites:
http://semi-nuc.iarc.fr/
TThe Semipalatinsk nuclear test site (SNTS) covers 18,500 km2 in the north east of Kazakhstan. From 1949–1989 the Soviet Union conducted nuclear tests at the SNTS, including detonations above ground and in the atmosphere that produced radioactive contamination in the soil and air. Some of these releases caused radioactive plumes that exposed people and territories adjacent to the site. In the past two decades, groups of researchers have initiated two cohort studies to assess the health effects of exposure to fallout from Soviet nuclear weapons testing among residents living near the SNTS. These studies used different control groups and different dosimetric methods and produced conflicting results that are of limited use for evaluating the true effects of the exposure.
The SEMI-NUC project was a support action that brought together scientists from Europe, Kazakhstan, and Japan with a common interest in the health effects from chronic exposure to low doses of ionizing radiation.
One aim of the project was to test record linkage possibilities between the two previously-established cohorts (so-called “new” and “historical”) to assess the feasibility of creating a larger, unified cohort for future long-term studies. It was postulated that he resulting large population could be used for a long-term, prospective follow-up study of the health effects from exposure to fall-out in the Semipalatinsk area.
During the course of the project it was established that although record linkage between the two databases that contain the information on the new and historical cohorts is possible, it would be very labour intensive due to the different structures of the databases. It was also discovered during a site visit that one of the partners, the Scientific Research Institute of Radiation Medicine and Ecology (NIIRME), possesses a registry of exposed population around the SNTS that is larger than both the previously established cohorts combined. All information stored in the NIIRME database is verified to avoid duplication and technical errors and it is kept updated continuously. It was concluded that a cohort formed from the individuals whose information is stored in the NIIRME database would be the most suitable for a long-term, prospective follow-up study of the health effects from exposure to fall-out in the Semipalatinsk area.
The ultimate aim of this project was to produce a feasibility report to determine if a long term, prospective follow-up study of the residents near Semipalatinsk can be developed. The feasibility report contains the background and evidence that was collected during the course of the project that led the project partners to the conclusion that proves that this is indeed the case and recommends using a cohort selected from the NIIRME registry for the study. The report also includes information that will be the foundation for the follow-up study, including descriptions of follow-up mechanisms; assessment of dosimetry; description of biological material to study, mechanisms that underpin the effects of low-dose radiation and an outline of the health outcomes to be studied.
Project Context and Objectives:
• Introduction
The health effects of exposure to fallout from Soviet nuclear weapon’s testing among residents living near to the Semipalatinsk nuclear test site in Kazakhstan have not been thoroughly investigated in the past. Two previous studies that were conducted on two independent cohorts, the so-called “historical” and “new” cohorts produced contradicting and non-conclusive results. These studies also involved differing control groups and different dosimetric methods.
The SEMI-NUC project was developed in accordance with the Strategic Research Agenda of MELODI, a European Platform dedicated to low dose radiation risk research. The project was designed as a feasibility study to investigate the possibility of merging the previously-established historical and new cohorts into a single, large cohort for future studies of the health effects of long-term exposure to low-dose ionizing radiation in these populations. The merging of the two existing cohorts into a single, larger cohort would have the advantage of avoiding unnecessary duplication of effort and resources in future studies plus the benefit of increased statistical power. In addition, there are not many populations worldwide that have been exposed to protracted low-dose radiation and that are suitable for studies on the morbidity of cardiovascular diseases. The large unique cohort of individuals from around the SNTS has the potential to prove that prospective epidemiological research on the association between radiation and cardiovascular diseases is feasible.
Other populations exposed to low to moderate dose radiation, such as in Fukushima or elsewhere, can benefit from the outcomes of studying the unique Semipalatinsk cohort and the results will contribute to a better understanding and quantification of radiation risks for low to moderate chronic doses.
• Creation of a unique cohort for future prospective studies
To create a single combined cohort from the two existing cohorts, it was necessary to know and understand the data that is available for the historical and new cohort members in the respective databases. Site visits were organized to the two partner institutes (NIIRME and NNC) where the data for each cohort is kept to thoroughly investigate the structure and composition of the different databases and the mechanisms used for identification and follow-up of cohort members.
Evaluation of the databases at NIIRME and NNC revealed that each has a different structure and contains different kinds of information about individuals. In order to conduct record linkage of the two databases, it was necessary to create another version of the NIIRME database with just one record per person.
To test a record linkage procedure an initial linkage of a random sample of 3,000 records selected from each of the two databases was performed using RecLink3 software. Linkage variables included first name, last name, patronymic name, gender, and date of birth. It was concluded that despite their different structures, a complete linkage of both databases would be possible, but very labour intensive. It was estimated that one person-year would be needed for a complete linkage.
However, it was decided during the course of the project that the linkage of the two databases would not be needed for the study. This decision resulted from the discovery made during one of the site visits that NIIRME possesses a registry of exposed population around the SNTS that is larger than both the previously established cohorts combined. Moreover, all information stored in the NIIRME database is verified to avoid duplication and technical errors and it is kept updated continuously.
It was concluded that a cohort formed from the individuals whose information is stored in the NIIRME database would be the most suitable for a long-term, prospective follow-up study of the health effects from exposure to fall-out in the Semipalatinsk area and hence it was no longer thought necessary to merge the two previously-established cohorts.
• Identification of potential health outcomes
A second major objective of the project was to identify the potential health outcomes (cancer and non-cancer diseases) that can be studied in a future prospective study using the unified cohort described previously. NRPA in collaboration with NIIRME and NNC worked on the evaluation of mechanisms of data collection and the availability and quality of the various information sources of mortality and morbidity data to identify potential health outcomes for future prospective studies of radiation-related risks. Mechanisms of data collection on death cases and causes, and information sources on mortality data were identified, followed by an evaluation of mechanisms of data collection, information sources and data quality for both incidence and mortality data.
A quality assurance protocol was developed that included procedures for case ascertainment and abstraction of information on both death and incident cases of cancer and non-cancer diseases. The protocol describes the procedures for data collection, data entry and data management. The data quality was checked including consistency of diagnostic criteria within the same disease and correctness of disease coding using International Classification of Disease (ICD) and ICD for Oncology (ICD-O).
Potential health outcomes for future studies were identified based on a detailed inventory of all available information sources for different types of health outcomes (cancer and non-cancer, incidence and mortality), followed by assessment of data quality and completeness including a review of case ascertainment procedures, diagnostic criteria, morphological verification and cross-checking the information from the different sources. After reviewing sources of information and data availability on cancer incidence, we consider feasible a future study of cancer incidence in relation to early life exposure. Because information on adulthood cancer risks following in utero and early life exposure is limited (mainly based on studies of A-bomb survivors and offspring of Mayak female workers and Techa River population), we consider as a priority setting up a prospective cohort study of radiation-related cancer risk in a sub-cohort of people exposed in utero or early life. We also recommend a study on cardiovascular diseases based on the national screening examination data. It is feasible to study stroke and CVD outcomes in the Semipalatinsk population. Ongoing national screening programs on CVD will ensure uniformity of diagnostic criteria and compliance with international approaches resulting in reliable quality of CVD diagnoses. Because data on behavioural and other risk factors are collected as a part of the screening examination procedure, it will allow inclusion in the study additional information on non-radiation risk factors. A follow-up of the NCI ultrasound-detected thyroid nodules prevalence study was also considered as the most feasible and informative study. Because of systematic follow-up of population in the NIIRME registry and regular collection of information on death cases and causes by NIIRME researchers, it is feasible to carry out studies both on cancer and non-cancer mortality risks in a cohort of people exposed around the SNTS. The detailed report can be read in Deliverable 3.1.
• Evaluation of follow-up mechanisms
Possible follow-up mechanisms for a future prospective study were reviewed and evaluated. The characteristics of the different follow-up mechanisms used in the previous studies using the historical and new cohorts were summarized based on existing publications and reports and other relevant information. Alternative sources of information for follow-up were evaluated and discussed. After completing these tasks, follow-up mechanisms were suggested for a future prospective study. The mechanism of follow-up was established and proved to be efficient and not discriminative upon disease and exposure status. We consider availability of IIN for each Kazakh citizen as an advantage for follow-up that facilitates the follow-up process and allows expansion of the study follow-up area to the territory of entire Kazakhstan.
After reviewing the information, we confirmed that ZAGS (population registry) records should serve the main official source for mortality follow-up, i.e. for ascertaining date and cause of death in exposed population. The results of our comparison between three different datasets (historical cohort, new cohort, and NIIRME registry) showed that the NIIRME’s registry of exposed population is a comprehensive source of information that can be used not only for establishing a cohort but also for obtaining vital status data for a future prospective study. A report on identified and verified follow-up mechanisms for a prospective study was produced (Deliverable 4.1). In addition, a comparison of information on follow-up results on selected samples from the historical cohort, the new cohort, and the NIIRME Registry of exposed population was carried out.
• Identification of key pathways and mechanisms for dose formation
The main pathways for external and internal exposure to the public living in areas of radioactive fallout due to nuclear weapons testing were identified for both external and internal exposure taking into account the geographical location of the population and lifestyle factors such as diet, occupation, and materials used for housing.
• Assessment of the approaches used previously for individual external and internal dose reconstruction, including evaluation of the parameters used for assessing cumulative doses
The advantages and weaknesses of the two methodologies (called the Russian methodology and U.S. methodology) used in the previous studies were assessed in detail based on existing reports and publications.
A third methodology - the Joint U.S.-Russian Methodology – has been developed jointly by U.S. (U.S. National Cancer Institute, U.S. Department of Energy and University of Utah) and Russian (Institute of Biophysics) specialists on fallout dosimetry taking into account the advantages and weaknesses of both the U.S. and Russian methodologies. It is currently the best method available for reconstruction of individual external and internal doses hence it is this method that is recommended for use in the future study.
• Development of an approach for the assessment of individual cumulative doses to the exposed population
The objective was to prepare a proposal for a unified individual dose reconstruction approach for the public exposed to fallout from nuclear weapons testing at the SNTS that can be used in the future epidemiological study.
The joint U.S.-Russian methodology was selected for retrospective reconstruction of individual external and internal doses to the population permanently living in the vicinity of the SNTS. In the framework of the current pilot project, it was possible only to identify and describe the main sources of uncertainty in the dose estimates. A strategy for estimation and inclusion of uncertainty in the dose estimates is in a progress. It is proposed to use the recently developed two-dimensional Monte-Carlo procedure for future long-term study to enable separation of shared and unshared errors. The instrumental methods of retrospective dosimetry (Retrospective Luminescence Dosimetry- RLD, and Electron Spin Resonance Dosimetry - ESR) were selected and tested for validation of calculated external doses. It is proposed to use these instrumental methods for validation of calculated doses in the future extended study.
Project Results:
By demonstrating the feasibility of using a unique cohort from the data held by NIIRME, the SEMI-NUC project has prepared all the main aspects for conducting a full epidemiological study on residents near the Semipalatinsk test site. The results of such a study would greatly contribute to the quantification of health risks from protracted low and moderate doses of radiation.
The Semipalatinsk site is particularly interesting in this sense because of an initial high exposure during the nuclear testing and a subsequent protracted exposure due to the radioactive fallout. The dose range for members of the cohort vary from very low to moderate doses and the outcomes that could be studied include both cancer and non-cancer diseases. In particular, there is increasing interest in studies of the risk of cardiovascular diseases following exposure to low-dose protracted radiation as this may have important implications for radiation protection of general public and patients exposed to radiation.
The review of the dosimetric tools considered will be valuable for other researches involved in dose-effect studies of other population groups, since good dosimetry is essential for epidemiological studies.
Proposals for follow-up research
• Follow-up of ultrasound-detected thyroid nodules
The most feasible and near-term study would be a follow-up of ultrasound-detected thyroid nodules prevalence study. A first round of screening of about 3,000 people was performed by the National Cancer Institute (NCI), USA, in collaboration with Kazakh scientists in 1998, and since then no attempts of organized systematic follow-up of the persons diagnosed with thyroid nodules has been made. For each screening participant there is also a study questionnaire recording their residential history, medical history and dietary habits with special attention to their consumption of milk and milk products. Individual thyroid doses due to external and internal exposures were reconstructed for each study participant.
Given gaps in current understanding of both the natural history of development of thyroid nodules, their potential to transform into neoplasms, and mechanisms of radiation-induced thyroid carcinogenesis, it is important to initiate a follow-up of the people with detected thyroid nodules and also those who were nodule-free at the time of the initial screening.
To perform a longitudinal study assessing the development of thyroid nodules the following steps are required:
a) Reach a data access agreement with the NCI to have access to the list of study participants;
b) Link the list of study participants
i. with the NIIRME registry of exposed people to ascertain their vital status and residence history;
ii. with the population-based cancer registry in the Semipalatinsk regional oncology centre to identify cases of thyroid cancer among them; and also
iii. with the database of the Semipalatinsk clinical-diagnostic centre to identify those among them with thyroid nodules.
The linkage result will serve as a basis for the risk evaluation of neoplastic transformation of benign nodules into thyroid cancer.
After the first phase, depending on the percentage of deceased, migrated or lost-to-follow-up study subjects, a screening exercise can be planned to assess the current thyroid status including the presence of thyroid nodules in this cohort.
• Pilot study on cardiovascular diseases
We recommend a pilot study on cardiovascular diseases based on national screening examination data. It is feasible to study stroke and CVD outcomes in the Semipalatinsk population. To carry out such study, it will be necessary to:
a) perform a linkage of the NIIRME registry data with the database on the 26,093 exposed individuals examined at the clinical-diagnostic centre in the framework of the established collaboration.
b) collect information on risk factors and examination results of exposed individuals who are patients of the clinical-diagnostic centre.
c) collect information on CVD screening results, including information on diet, lifestyle and other risk factors, from the local medical institutions in the catchment area, i.e. East-Kazakhstan region, through linkage between the NIIRME registry on exposed people with the database on screened patients.
The proposed study will have a prospective nature because data on CVD incidence are available from 2011 onward. Ongoing national screening programs on CVD will ensure uniformity of diagnostic criteria and compliance with international approaches resulting in reliable quality of CVD diagnoses. Because data on behavioural and other risk factors are collected as a part of the screening examination procedure, it will allow inclusion in the study additional information on non-radiation risk factors. Availability of IIN will ensure precise and effective linkage collecting this information on the members of NIIRME registry who were screened for CVDs.
• Cohort study on cancer risk following early life exposure
We suggest a cohort study on cancer risk following early life exposure including exposure in utero. After reviewing sources of information and data availability on cancer incidence, we conclude that a future study of cancer incidence in relation to early life exposure is feasible. However, taking into account existing limitations in cancer incidence data, for the successful implementation of such a study the following would be needed:
a) Identify potential members of a cohort with early life exposure using the NIIRME registry of exposed population;
b) Collect data on cohort members’ residence history and vital status using the NIIRME registry of exposed population;
c) Establish collaboration with the Semey regional oncology centre to identify cancer cases in the NIIRME registry members through linkage with the population-based cancer registry. Cancer follow-up is nearly complete from 2003 onward, but information in the period from 1980–2002 is of questionable quality. No cancer incidence data available before 1980;
d) Reconstruct individualised doses due to external exposure using all available sources of information including historical data on exposure rate measurements, individual residential history, time spent outdoors, etc.;
e) Perform a dose-response analysis for solid cancer and leukaemia cases using radiation epidemiology statistical tools and software.
Because information on adulthood cancer risks following in utero and early life exposure is limited (mainly based on studies of A-bomb survivors and offspring of Mayak female workers and the Techa River population), we consider setting up a prospective cohort study of radiation-related cancer risk in a sub-cohort of people exposed in utero or early life as a priority. Nested case-control studies for specific cancer sites could be considered in the future.
• Prospective molecular epidemiology study of cancer risks following early life exposure
The currently available biobank data on exposed individuals are not sufficient to carry out a molecular epidemiology study. We therefore recommend establishing a sustained collaboration with the cancer oncology centre and the clinic-diagnostic centre to collect blood and tumour tissue samples for members of the NIIRME registry of exposed population. Expanding the collection of biological samples will make it possible to conduct a prospective study of cancer risks following early life exposure as described above.
The proposed action will also open up new collaborations outside existing European networks of low-dose radiation research, having built up a very close collaboration with the scientists from Kazakhstan, Japan and also the US. It has also improved the capacity for conducting high-quality radiation research in Kazakhstan through the involvement of local scientists who benefitted from participation in the support action to better exploit their competences.
Through clarifying the Intellectual Property Rights for the database shared jointly between the Institute of Radiation Safety and Ecology of the National Nuclear Centre of Kazakhstan (NNC) and the Radiation Effects Association (REA) of Japan, it was possible to reach agreement for this data to be made available to researchers for a future study. Kazakhstan partners NNC also agreed to make their version of this data available to a wider scientific population for future research.
Potential Impact:
In the long term, this project and a subsequent full study, has a potential improve our understanding of the health risks associated with low and moderate chronic doses of radiation. A better quantification of risks from low and protracted doses is valuable for regulating authorities so that exposure due to planned, existing or emergency exposure situations, such as Chernobyl or Fukushima, can be better regulated. It will also serve as a valuable reference for making public health decisions.
The dissemination of the knowledge arising from the project was achieved through presentations at scientific conferences, a dedicated project dissemination meeting, publications and other forms of communication such as websites. The table in Annex 1 presents a list of the webpages containing news and information about the project.
• Webpages
Webpages containing news and information about the project are attached following the table in Annex 1.
• Publications
Grosche B., Zhunussova T., Apsalikov K., Kesminiene A. Studies of Health Effects from Nuclear Testing near the Semipalatinsk Nuclear Test Site, Kazakhstan. Central Asian Journal of Public Health, Vol 4(1), 2015. http://cajgh.pitt.edu/ojs/index.php/cajgh/article/view/127/203
Zharmukhambetova A.K. Lipikhina A.V. Mansarina A.E. Zhakupova Sh.B. Bright Yu.Yu. Integration of data from the medical registry of Kazakhstan population exposed to low doses of ionizing radiation into the European platform MELODI. Abstracts of XI International scientific conference «Ecology. Radiation. Health», August 28-29, 2015, Semey state medical university, 2015, p. 20. http://ssmu.kz/ru/pages/default/view?id=sci-downl
World Hypertension League (WHL) Newsletter, No. 147, March 2015, p.5.
http://www.thaihypertension.org/files/314.Newsletter%20No.%20147,%20March%202015.pdf
• Project meetings, dates and venues
06-08/05/2013, IARC, Lyon, France: Project kick-off meeting: Prospective cohort study of residents near the Semipalatinsk nuclear test site – feasibility assessment and health risk from exposure to radioactive fallout from nuclear testing in Kazakhstan at the Semipalatinsk nuclear test site. This meeting was organized together with the meeting co-sponsored by IARC and the National Cancer Institute to overview the previous studies conducted on health effects in Semipalatinsk. It helped to identify methods and approaches used in the previous studies
17-20/09/2013, Semey, Kazakhstan: Site visit; visit to Astana to the Ministry of Health of Kazakhstan – to present the project and discuss accessibility of existing data and registries to European investigators
09-10/06/2014, BfS, Munich, Germany: SEMI-NUC Mid-Term project meeting, External Advisory Board meeting.
30/03/2015-02/04/2015, Site visit to Semey, Kazakhstan: visit to NIIRME, Semey and NNC, Kurchatov to learn more about the data in the registries.
11-12/05/2015, External Advisory Board (EAB) meeting, IARC, Lyon, France: The Work Package leaders presented the work achieved and the External Advisory Board made comments and suggestions as to how to resolve problems and about the best way to proceed.
26/05/2015, Kyoto, Japan: the study coordinator and Dr Grosche attended a Meeting with the president of REA to discuss the IPR for the new cohort data.
09-10/11/2015, BfS, Munich, Germany: Project dissemination meeting and follow-up meeting.
09/03/2016, Fukushima, Japan: meeting at the Fukushima medical university to discuss future longitudinal study of thyroid nodules in populations around the SNTS
14-15/03/2016, NRPA, Oslo, Norway: Final project meeting with partners and EAB.
The reports from these meetings are attached in Annex 2.
List of Websites:
http://semi-nuc.iarc.fr/