Dietary intake of radiocesium in adult residents in Fukushima prefecture and neighboring regions after the Fukushima nuclear power plant accident: 24-h food-duplicate survey in December 2011.

Since the nuclear power plant accident in Fukushima in March 2011, the Japanese government has conducted screening and removal of contaminated foods from the market that exceed provisional regulation limits for radionuclides. This study aimed to provide an urgent estimate of the dietary exposure of adult residents recruited from three areas in Japan to cesium 134 ((134)Cs), cesium 137 ((137)Cs), and, for comparison, natural potassium 40 ((40)K) on December 4, 2011. Fifty-three sets of 24-h food-duplicate samples were collected in Fukushima Prefecture and neighboring regions. The (134)Cs, (137)Cs, and (40)K levels in the samples were measured using a germanium detector. Items in the food-duplicate samples were recorded and analyzed for radiocesium intake. Radiocesium was detected in 25 of 26 samples from Fukushima. The median dietary intake of radiocesium was 4.0 Bq/day (range <0.26-17 Bq/day). The estimated annual dose from radiocesium was calculated assuming that the daily intake of radiocesium was constant throughout the year. The median estimated dose level was 23 μSv/year (range <2.6-99 μSv/year). The estimated dose level of radiocesium was significantly higher in Fukushima than in the Kanto region and western Japan. Stepwise multiple linear regression analyses demonstrated that the intake of fruits and mushrooms produced in Fukushima were significant factors for the dietary intake of (137)Cs in the 26 participants from Fukushima. The average radioactivity (±SD) of locally produced persimmons and apples (n = 16) were 23 ± 28 and 30 ± 35 Bq/kg for (134)Cs and (137)Cs, respectively. The preliminary estimated dietary dose levels among Fukushima residents were much lower than the maximum permissible dose 1 mSv/year, based on new Japanese standard limits for radiocesium in foods (100 Bq/kg for general foods). In future studies, the exposure estimates should be refined by probability sampling to eliminate biases.