No abstract

No abstract

Objective. This study investigates the trend of perinatal mortality rates in Fukushima Prefecture and four neighboring prefectures (Miyagi, Gunma, Tochigi, and Ibaraki) after the disaster at the Fukushima Daiichi nuclear power plant in March 2011. Material and methods. Japanese monthly perinatal mortality data on a prefecture level are available on a website of the Japanese government. A combined regression of perinatal mortality rates from the study region and the rest of Japan (the control region) is conducted. The regression model allows for an asymptotic lower limit and a level change of perinatal mortality rates in 2012–2017 in the study region relative to the predicted trend. Results. In 2012–2017, perinatal mortality in the study region shows a significant 10.6% increase relative to the trend in preceding years (p = 0.006). The excess mortality translates to 195 (95% CI: 28, 462) excess perinatal deaths. The increase is three times greater in Fukushima Prefecture than in the four neighboring prefectures and the difference in excess rates is statistically significant (p = 0.010). Periodic peaks of perinatal mortality are found in 2012–2017 with maxima around April. Conclusion. We find an increase in perinatal mortality in Fukushima and four neighboring prefectures after the Fukushima nuclear accident. The results agree with similar observations in Germany and Ukraine after the Chernobyl disaster. Due to its ecological design, the study cannot prove a causal link between radiation exposure and perinatal mortality. Continued observation of the trend of perinatal mortality in contaminated regions of Japan is recommended.

The authors investigated a possible association between background gamma radiation (BGR) and cancer and infant mortality rates. In an in-country ecological study, they performed a population-weighted linear regression of cancer (infant) mortality rates on BGR, adjusted for unemployment rate and population density. Crude cancer rates showed a highly significant increase with BGR: 38 excess cases per 100,000 person-years per millisievert/year (p < .0001). After adjusting for unemployment rate and population density, the authors found that the excess absolute risk reduced to 23.6 cases per 100,000 person-years per mSv/year (p = .0014). The corresponding excess relative risk was 10.2% (95% confidence interval = 3.9-16.7) per mSv/year. The excess relative risk for infant mortality rates was 24% (95% confidence interval = 9-42) per mSv/year. The cancer risk derived from this ecological study is 0.24/Sv, which compares with an International Commission on Radiological Protection value of 0.05/Sv. However, because they are based on highly aggregated data, the results should be interpreted with caution.

Using trend analysis, the author sought a possible association between perinatal mortality rates in West Germany, 1955-1993, and the fallout from atmospheric nuclear weapons testing in the years 1952-1993. The regression model used a continuously falling trend and a superimposed extra term that reflects the average strontium content in pregnant women. Mortality rates show an upward deviation that peaked in 1970. The model attributes more than 100,000 excess perinatal deaths to strontium in the fallout. The dose-response curve is curvilinear with a power of dose of 1.81 +/- 0.23. In addition, using a combined regression model, the author analyzed the two data subsets of perinatal mortality (i.e., stillbirth rate and early neonatal mortality). The strontium effect is 3.4 times greater on early infant deaths than on stillbirths. According to the prevailing wisdom, the fetus is protected against damage from ionizing radiation by a threshold dose of 50-200 mSv, but the doses from strontium in the fallout were well below 1 mSv/yr in Germany. The results reported here seem to contradict the existence of a threshold dose for perinatal mortality at low doses.