Mortality among Fukushima nuclear emergency workers, 2012–2021 Authors: Huan Hu, Kotaro Ozasa, Shinji Yoshinaga, Tetsuya Mizoue, Toshiteru Okubo Published: 2026 ABSTRACT AbstractObjectivesThe 2011 Fukushima Daiichi Nuclear Power Plant accident involved approximately 20 000 emergency workers. Given the scale and nature of the accident, concerns have been raised regarding the long-term health of the emergency workers. This study presents the first 10 year overview of mortality among these workers.MethodsWe analysed the mortality data for 19 358 male emergency workers (mean age in 2012: 45.5 years; SD: 11.3) from 2012 to 2021. Mortality data were monitored through linkage to national vital statistics, including the dates and causes of death. Standardised mortality ratios (SMRs) with 95% confidence intervals (CIs) were calculated using the general male population in Japan as a reference, standardised by calendar year and attained age (5 year age groups).ResultsDuring 190 700 person-years of follow-up, 708 deaths were identified. The overall SMR was 0.86 (95% CI 0.80 to 0.93). SMRs for all solid cancers and radiation-associated solid cancers and leukaemia were below 1.0 during 2012–2014, approached 1.0 during 2015–2017 and exceeded 1.0 during 2018–2021, suggesting marginal excess mortality (both P values for trend <0.05). In contrast, the increasing trends were less evident for neoplasms excluding radiation-associated solid cancers and leukaemia. No elevated SMR for suicide was observed.ConclusionsThe overall mortality among Fukushima emergency workers was lower than expected based on general population rates, likely reflecting the healthy-worker effect. However, marginal elevations in SMRs for all solid cancers and for radiation-associated solid cancers and leukaemia emerged during later follow-up, highlighting the importance of continued health monitoring in this population. WHAT IS ALREADY KNOWN ON THIS TOPIC Following major nuclear accidents, concerns have been raised about potential long-term health effects among emergency workers, including cancer and suicide. Evidence from the Chornobyl accident suggested lower overall mortality in early follow-up but increased risks for certain outcomes, including suicide. However, mortality data for Fukushima emergency workers have not previously been reported. Following major nuclear accidents, concerns have been raised about potential long-term health effects among emergency workers, including cancer and suicide. Evidence from the Chornobyl accident suggested lower overall mortality in early follow-up but increased risks for certain outcomes, including suicide. However, mortality data for Fukushima emergency workers have not previously been reported. WHAT THIS STUDY ADDS In this first 10-year mortality overview of Fukushima emergency workers, overall mortality was lower than expected, and no increase was observed for suicide. Standardised mortality ratios for all solid cancers and for radiation-associated solid cancers and leukaemia increased over time, from below 1.0 during 2012–2014 to above 1.0 during 2018–2021, suggesting marginal excess mortality in later years. In this first 10-year mortality overview of Fukushima emergency workers, overall mortality was lower than expected, and no increase was observed for suicide. Standardised mortality ratios for all solid cancers and for radiation-associated solid cancers and leukaemia increased over time, from below 1.0 during 2012–2014 to above 1.0 during 2018–2021, suggesting marginal excess mortality in later years. HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY These findings provide early evidence on mortality patterns among Fukushima emergency workers and support the need for continued long-term health monitoring, including future evaluation of dose–response relationships. These findings provide early evidence on mortality patterns among Fukushima emergency workers and support the need for continued long-term health monitoring, including future evaluation of dose–response relationships. Background The Fukushima Daiichi Nuclear Power Plant accident, classified as level 7 (major accident) on the International Nuclear Event Scale, resulted in the release of large quantities of radionuclides into the environment, prompting extensive emergency response efforts.1 Approximately 20 000 workers participated in emergency tasks between 14 March and 16 December 2011, including cooling nuclear reactors, removing fuel rods and constructing tanks to store contaminated water.2 These workers included both nuclear plant employees and individuals recruited from various industries across Japan for emergency response activities. To enable these activities, the radiation dose limit for emergency workers was temporarily set at a cumulative dose of 250 mSv received during the emergency work period from 14 March to 16 December 2011, independently of the usual occupational exposure limit of 100 mSv over 5 years.3 The scale and nature of the accident raised significant concerns regarding the potential health consequences for emergency workers, particularly radiation-induced illnesses and mortality. To address these concerns and provide evidence-based insights for future preparedness, the Epidemiological Study of Health Effects in Fukushima Nuclear Emergency Workers (NEWS) was initiated in 2016. Nuclear emergency workers may develop radiation-related health risks, including certain cancers and non-radiation effects, such as psychosocial or psychiatric impairments; both have been observed among Chornobyl cleanup workers.4–7 A study conducted approximately 10 years after the Chornobyl accident revealed that the overall mortality among emergency workers was lower than that of the general population.8 However, a significantly higher risk of suicide was observed among these workers, which was attributed to the psychological and emotional distress arising from the acute stress and challenging conditions of emergency work.8 9 A previous study of Chornobyl emergency workers suggested that increased risks of radiation-induced leukaemia were observed primarily during the first 10 years after the accident.4 Despite a similar period having passed since the Fukushima accident, no mortality data have been reported for Fukushima emergency workers. To fill this information gap, the present study used data from the NEWS cohort to provide the first overview of mortality among NEWS during the first 10 years (2012–2021) following their involvement in the accident response. Methods The NEWS study is an ongoing, lifetime cohort study of Fukushima nuclear emergency workers. The source population comprised 19 812 company-employed emergency workers registered in the healthcare system established by the Ministry of Health, Labour and Welfare (MHLW), Japan. For mortality and cancer surveillance, all eligible workers were included unless they declined data linkage under an opt-out procedure. Further details of the NEWS study design have been described elsewhere.10 11 Radiation dose data during the emergency work period were obtained from the MHLW and were based on individual monitoring records, including external exposures measured by personal dosimeters and internal exposures assessed by whole-body counters when applicable. As previously reported,10 the mean dose received during emergency work was 15.4 mSv (range: 0.0–645.5 mSv). However, a comprehensive evaluation of individual radiation exposure requires validated dose information covering periods before and after the emergency work in addition to the recorded emergency-work doses. Verification and supplementary collection of these dose data are ongoing; therefore, complete and validated individual exposure profiles were not available for analytical use in the present study. Ethical approval for this study was obtained from the ethics committees of the Radiation Effects Research Foundation (RP 6–15) and the National Institute of Occupational Safety and Health (2024N28), Japan. Study participants From the 19 812 registered emergency workers, women (n=25), individuals who declined personal data linkage (n=412), those with missing dates of birth (n=7) and individuals who died in 2011 (n=10) were excluded, resulting in a final cohort of 19 358 participants. As 2011 corresponded to the emergency response period during which workers were actively engaged in emergency work while accumulating radiation exposure, follow-up was initiated on 1 January 2012 to evaluate post-emergency-work mortality. Mortality Mortality data, including the dates and causes of death, were ascertained through periodic linkage to the national vital statistics system. This system registers all deaths nationwide and is widely used for epidemiological research. The linkage was performed using personal information, including name, sex, date of birth and place of residence. The causes of death were classified according to the International Classification of Diseases, 10th Revision (ICD-10). The study focused on radiation-related cancers, including all solid cancers combined, cancers of the salivary gland, oesophagus, stomach, colon, liver, lung, bone and connective tissue, non-melanoma skin, breast, bladder, brain and central nervous system and thyroid, as well as leukaemia, excluding chronic lymphocytic leukaemia.12 Suicide was also evaluated as a potential outcome associated with disaster-related psychological distress. Statistical analysis Person-years at risk were calculated from 1 January 2012, until either the date of death or 31 December 2021, whichever occurred first. To evaluate the mortality risk among emergency workers compared with the general male population of Japan, standardised mortality ratios (SMRs) with 95% confidence intervals (CIs) were calculated. Overall and cause-specific SMRs were determined as the ratios of observed to expected deaths among emergency workers. The expected numbers of deaths were calculated for each calendar year by multiplying the person-years at risk by the corresponding annual mortality rates for the Japanese male population, stratified by attained age in 5-year age groups (15–19, 20–24, 25–29, …, 80–84 and ≥85). The 95% CIs were computed using OpenEpi.13 Descriptive analyses and Poisson regression analyses of trends in SMRs were performed using SAS V.9.4 (SAS Institute, Cary, NC, USA). Results The analysis comprised 19 358 male emergency workers, with a mean age of 45.5 (SD: 11.3) years in 2012 and 54.2 (11.2) years in 2021. During 190 700 person-years of follow-up, 708 deaths were identified, resulting in a mortality rate of approximately 3.7 deaths per 1000 person-years. The leading causes of death were neoplasms (43.1%), cardiovascular diseases (23.7%) and external causes (14.7%). As shown in table 1, the overall SMR among emergency workers during 2012–2021 was 0.86 (95% CI 0.80 to 0.93), corresponding to a 14% lower mortality than expected based on general male population rates in Japan. Cause-specific analyses showed no evidence of increased SMRs for all neoplasms (0.96; 95% CI 0.86 to 1.08), all solid cancers combined (1.00; 95% CI 0.89 to 1.12), radiation-associated solid cancers (1.01; 95% CI 0.87 to 1.16), leukaemia (0.70; 95% CI 0.25 to 1.50) or neoplasms other than radiation-associated solid cancers and leukaemia (0.92; 95% CI 0.76 to 1.10). Additionally, no elevated SMR was observed for suicide (0.87; 95% CI 0.65 to 1.14). Lower SMRs were identified for diseases of the circulatory system (0.84, 95% CI 0.72 to 0.98), respiratory system (0.52, 95% CI 0.35 to 0.75), nervous system (0.11, 95% CI 0.01 to 0.41) and genitourinary system (0.18, 95% CI 0.02 to 0.66). Table 1Observed deaths and standardised mortality ratios (SMRs) among 19 358 male Fukushima emergency workers, 2012–2021Cause of deathICD-10ObservedExpectedSMR (95% CI)All causesA01-U99708820.70.86 (0.80 to 0.93)Certain infectious and parasitic diseasesA01-B99612.80.47 (0.17 to 1.02)NeoplasmsC00-D48305316.50.96 (0.86 to 1.08) All solid cancersC00-C80, C97289289.51.00 (0.89 to 1.12) Radiation-associated solid cancers*See footnote189188.11.01 (0.87 to 1.16) Neoplasms excluding radiation-associated solid cancers and leukaemia†See footnote110119.60.92 (0.76 to 1.10) Malignant neoplasm of salivary glandsC07, C080NA– Malignant neoplasm of oesophagusC151517.20.87 (0.51 to 1.41) Malignant neoplasm of stomachC163839.90.95 (0.68 to 1.29) Malignant neoplasm of colonC183125.51.22 (0.84 to 1.70) Malignant neoplasm of liver and intrahepatic bile ductsC222726.21.03 (0.69 to 1.48) Malignant neoplasm of bronchus and lungC33, C346969.01.00 (0.78 to 1.26) Malignant neoplasm of bone, articular cartilage, and connective tissueC40, C41, C490NA– Malignant neoplasm of skin (non-melanoma skin cancer)C4401.2– Malignant neoplasm of breastC5000.1– Malignant neoplasm of bladderC67<54.90.82 (0.26 to 1.97) Malignant neoplasm of brain and central nervous systemC70, C7253.91.28 (0.47 to 2.84) Malignant neoplasm of thyroid glandC730NA– Malignant neoplasms of lymphoid, hematopoietic and related tissuesC81-C961120.20.54 (0.29 to 0.95) Malignant neoplasms of lymphoid, hematopoietic and related tissues (excluding leukaemia)C81-C90, C96511.50.43 (0.14 to 1.02)Leukaemia‡C91-C9568.70.70 (0.25 to 1.50) Myelodysplastic syndromesD46<52.01.50 (0.31 to 4.39)Endocrine, nutritional, and metabolic diseasesE00-E881416.40.85 (0.49 to 1.40)Mental and behavioural disordersF00-F99<54.01.00 (0.27 to 2.56)Nervous systemG00-G98<517.60.11 (0.01 to 0.41)Circulatory systemI00-I99168198.90.84 (0.72 to 0.98)Respiratory systemJ00-J982650.20.52 (0.35 to 0.75)Digestive systemK00-K924549.30.91 (0.67 to 1.21)Skin and subcutaneous tissueL00-L98<50.91.11 (0.03 to 6.19)Genitourinary systemN00-N99<511.00.18 (0.02 to 0.66)Symptoms, signs, and abnormal clinical and laboratory findings, not elsewhere classifiedR00-R992934.20.85 (0.58 to 1.20)Injury, poisoning and certain other consequences of external causesS00-T98104100.41.04 (0.85 to 1.25) Transport accidentsV01-V98139.21.41 (0.75 to 2.42) Other external causes of accidental injuryW00-X593225.41.26 (0.88 to 1.76) Intentional self-harm (suicide)X60-X845057.60.87 (0.65 to 1.14) Other external causesX85-Y9898.21.10 (0.50 to 2.08)Special purposes (COVID-19)U00-U99<5NA–Expected numbers could not be calculated for some rare cancer sites because they were not separately classified in the annual cause-specific mortality reference data.*Radiation-associated solid cancers included C15, C16, C18, C22, C33–C34, C44, C50, C67, and C70–C72. Cancers of the salivary gland (C07–C08), bone and connective tissue (C40–C41, C49), and thyroid (C73) were excluded because they were not separately classified in the annual cause-specific mortality reference data.†Neoplasms excluding radiation-associated solid cancers and leukaemia were defined as neoplasms (C00–D48) excluding radiation-associated solid cancers (C15, C16, C18, C22, C33–C34, C44, C50, C67, and C70–C72) and leukaemia (C91–C95).‡Observed leukaemia cases did not include chronic lymphocytic leukaemia (C91.1); however, the expected numbers may have included C91.1 because leukaemia was reported as a single category in the annual cause-specific mortality reference data.ICD-10, International Classification of Diseases, 10th Revision; NA, not available. Observed deaths and standardised mortality ratios (SMRs) among 19 358 male Fukushima emergency workers, 2012–2021 Expected numbers could not be calculated for some rare cancer sites because they were not separately classified in the annual cause-specific mortality reference data. Radiation-associated solid cancers included C15, C16, C18, C22, C33–C34, C44, C50, C67, and C70–C72. Cancers of the salivary gland (C07–C08), bone and connective tissue (C40–C41, C49), and thyroid (C73) were excluded because they were not separately classified in the annual cause-specific mortality reference data. Neoplasms excluding radiation-associated solid cancers and leukaemia were defined as neoplasms (C00–D48) excluding radiation-associated solid cancers (C15, C16, C18, C22, C33–C34, C44, C50, C67, and C70–C72) and leukaemia (C91–C95). Observed leukaemia cases did not include chronic lymphocytic leukaemia (C91.1); however, the expected numbers may have included C91.1 because leukaemia was reported as a single category in the annual cause-specific mortality reference data. ICD-10, International Classification of Diseases, 10th Revision; NA, not available. Figure 1 illustrates trends in SMRs over time for all-cause and major cause-specific mortality. The overall SMR during 2012–2014 was 0.63 (95% CI 0.53 to 0.75), corresponding to a 37% lower mortality than expected. Thereafter, the SMR gradually approached 1.0, reaching 0.91 (95% CI 0.79 to 1.04) for 2015–2017 and 0.95 (95% CI 0.86 to 1.05) for 2018–2021 (figure 1 and online supplemental table S1). For all neoplasms, all solid cancers, and radiation-associated solid cancers and leukaemia combined (owing to the small number of leukaemia deaths), the SMRs were below 1.0 during 2012–2014, approached 1.0 during 2015–2017 and exceeded 1.0 during 2018–2021. For neoplasms excluding radiation-associated solid cancers and leukaemia, the SMR was below 1.0 during 2012–2014 and approached 1.0 thereafter. These temporal trends were quantitatively evaluated using both the entire and restricted follow-up periods (online supplemental table S2). Increasing trends in SMRs were statistically significant for several cancer outcomes over the entire follow-up period; however, the trends were attenuated after exclusion of the early follow-up period (2012–2014) and were no longer statistically significant during 2015–2021 (online supplemental table S2). Nevertheless, the attenuation was less marked for radiation-associated solid cancers and leukaemia. For cardiovascular disease, the SMR was below 1.0 during 2012–2014 and subsequently approached 1.0 over time, although the increasing trend was not statistically significant (P for trend=0.15–0.68). In contrast, SMRs for suicide remained below or close to 1.0 throughout the study period (P for trend=0.31–0.93). For other causes excluding neoplasms, cardiovascular disease and suicide, the SMR remained below 1.0 throughout the study period, although a slight increase over time was observed (P for trend=0.17–0.85). Standardised mortality ratios (SMRs) by cause of death and calendar period among 19 358 male Fukushima emergency workers, 2012–2021. Radiation-associated solid cancers included C15, C16, C18, C22, C33–C34, C44, C50, C67 and C70–C72. Expected numbers could not be calculated for some rare cancer sites (salivary glands, bone, articular cartilage and connective tissue, thyroid gland) because they were not separately classified in the annual cause-specific mortality reference data; therefore, SMRs could not be calculated for these cancers, and these sites were excluded from the ‘Radiation-associated solid cancers’ group. Observed leukaemia cases did not include chronic lymphocytic leukaemia (C91.1); however, the expected numbers may have included C91.1 because leukaemia was reported as a single category in the annual cause-specific mortality reference data. Due to the small number of leukaemia deaths, leukaemia was combined with radiation-associated solid cancers for analysis. Neoplasms excluding radiation-associated solid cancers and leukaemia were defined as neoplasms (C00–D48) excluding radiation-associated solid cancers (C15, C16, C18, C22, C33–C34, C44, C50, C67 and C70–C72) and leukaemia (C91–C95). Discussion Among Fukushima nuclear emergency workers, overall mortality during the first 10 years after emergency work was lower than expected based on rates in the general male population of Japan. Increasing trends in SMRs were observed for several cancer outcomes over the entire follow-up period; however, these trends were substantially attenuated after exclusion of the early follow-up period, particularly for neoplasms excluding radiation-associated solid cancers and leukaemia. In contrast, the increasing trends for all solid cancers and radiation-associated solid cancers and leukaemia were less attenuated. No elevated SMR for suicide was observed in any period. The lower overall SMR in the early years of follow-up likely reflects the healthy-worker effect, as individuals engaged in emergency response activities are generally required to be in good health to perform demanding tasks under challenging conditions. Furthermore, the healthy-worker survivor effect may also play a role, considering that some workers had long employment histories at the nuclear power plant prior to the accident. In our cohort, this lower mortality in the early years of follow-up was driven mainly by fewer deaths due to neoplasms and circulatory diseases during 2012–2014. A similar mortality deficit has been reported among Chornobyl cleanup workers, with an overall SMR reported as 0.85 between 1991 and 1998.8 Overall mortality gradually approached that of the general population over time, a trend also reported among Chornobyl cleanup workers.8 This suggests that the healthy-worker effect diminishes over time, as noted in previous research.14 Although overall cancer mortality during the entire follow-up period (2012–2021) did not substantially exceed the expected level based on general population rates, SMRs for all solid cancers and for radiation-associated solid cancers and leukaemia increased over time, with some indication of marginal excess mortality during 2018–2021. In contrast, SMRs for neoplasms excluding radiation-associated solid cancers and leukaemia were below 1.0 during the early follow-up period and subsequently approached 1.0 in 2015–2017, with little further increase thereafter. The lower mortality observed during the early years of follow-up likely reflects the healthy-worker effect. In a cohort of 65 905 Chornobyl emergency workers who started work during 1986–1990, overall cancer mortality remained below 1.0 during 1991–1998.8 However, a later study of 67 568 emergency workers who worked in the Chornobyl exclusion zone during 1986–1987 reported an 18% excess incidence of solid cancers during follow-up from 1992 to 2009.15 No increase in the SMR for suicide was observed during the study period. In contrast, studies of Chornobyl cleanup workers reported an elevated suicide risk, attributed to sudden and forced recruitment, uncertainty about radiation exposure, poor communication from authorities, highly challenging working conditions and other disaster-related stressors.9 16 Differences in radiation exposure—with substantially higher doses among Chornobyl cleanup workers (mean dose: approximately 100 mSv) and much lower doses among emergency workers (mean dose: approximately 15 mSv)—may also have contributed to the contrasting suicide outcomes observed between the two groups. In addition to these dose differences, multiple protective measures were implemented at the emergency worksite to safeguard the workers’ health, including individual radiation dose monitoring, improvements to the working environment, medical examinations and mental health support services.17 18 Furthermore, all workers were registered in the MHLW healthcare system to ensure ongoing health follow-up, including access to mental healthcare after their emergency work.17 18 These measures may help explain the absence of excess suicide mortality observed in this cohort. Strengths and limitations The major strengths of this study include the inclusion of almost the entire population of emergency workers and access to national death registries, which enabled individual-level linkage using personal identification data. This study has several limitations. First, the relatively short follow-up period limits the ability to detect long-latency outcomes, particularly solid cancers. Second, mortality data do not capture disease incidence, limiting the evaluation of cancers with low fatality, such as thyroid and prostate cancers. Third, data on major potential confounders—such as smoking, alcohol consumption, diet and pre-accident and post-accident radiation exposure—were not available for SMR calculations. The lack of these data may influence the interpretation of cause-specific mortality patterns. Fourth, data on specific emergency work tasks were not available for all workers, precluding an analysis of the effects of task-specific characteristics on mortality. Fifth, the cohort was based on the MHLW registry of company-employed emergency workers and did not include government personnel (eg, firefighters and police officers); therefore, the findings may not be generalisable to all individuals involved in the Fukushima accident response. Finally, the healthy-worker effect may have contributed to the low mortality observed during the early follow-up period. Conclusions In this first 10-year follow-up study of Fukushima nuclear emergency workers, overall mortality was lower than that observed in the general male population of Japan, likely reflecting the healthy-worker effect. In later years of follow-up (2018–2021), marginal elevations in SMRs for all solid cancers and for radiation-associated solid cancers and leukaemia were observed, underscoring the need for continued monitoring. Further studies are needed to investigate the potential associations between radiation exposure and mortality among emergency workers once the detailed radiation dose data become available. Supplementary material