**4. Discussion**

In the present study, we collected food items from local markets in two Cd-polluted areas in Akita, Japan, measured their Cd concentrations, and attempted to accurately assess Cd intake levels in local female farmers who had been exposed to high Cd levels by calculating individual Cd intake based on data obtained from diet surveys performed in health examinations. The present results demonstrated that Cd intake was approximately 2.5-fold higher in our subjects than in the general population. Furthermore, weekly Cd intake per body weight, which was confirmed by the Monte Carlo simulation, were approximately the same as the PTWI of JECFA or TWI of Japan, while 38.0–51.7% of subjects had weekly Cd intake per body weight that was above it.

The high Cd intake levels observed among our subjects may be attributed to high Cd concentrations in local agricultural products, particularly rice. Although Cd concentrations in rice in these areas (GMs of 0.158 and 0.109 mg/kg in areas A and B, respectively) were lower than the safety standard (0.4 mg/kg), they were markedly higher than the median rice Cd concentrations in Japan (0.06 mg/kg in 1997 and 1998 and 0.05 mg/kg in 2009 and 2010) [16], and there were rice that had Cd concentrations above the safety standard (8.2 and 5.8% in areas A and B, respectively). Although the intake of Cd was higher in area A than in area B, the actual accumulated levels of Cd in subjects were higher in area B than in area A, as demonstrated by their blood and urinary Cd levels (for example, urinary Cd levels in 70–79-year-old subjects were 4.90 and 9.34 μg/g cr. in areas A and B, respectively, and 2.99 μg/g cr. in control subjects) [5]. Since the biological half-life of Cd in humans is very long (10–30 years), blood and urinary Cd levels are stable indicators of the accumulation of Cd in the human body, particularly when the level of Cd that accumulates is high [1]. This discrepancy may have been due to the timing of the initiation of measures in rice farming to lower Cd absorption from the soil in 2002, namely, the flooding of paddy fields before and after heading during August [5,17]. Actually, the rates of rice with Cd concentration above the safety standard in these areas were decreased after that according to the results of intensive inspections targeted for Cd-polluted rice fields by Akita prefecture (Akita Prefect. Department of of Agriculture, Forestry, and Fisheries) (Supplemental Table S1). More recently, our own investigations on farmers in the areas show that the median of Cd concentration in their self-harvested rice from 2010 to 2018 was 0.096 mg/kg and 0.7% of them were above the safety standard (*n* = 599) (unpublished data). These results indicate the flooding of paddy fields have effectively lowered Cd concentrations in rice from 2002 in these areas. The health examination in area A was performed in 2001–2002, at the initiation of the flooding of paddy fields, while that in area B was performed in 2002–2003, just after the flooding of paddy fields had started. Therefore, Cd intake levels may have been higher in area B than in area A before the start of the flooding of paddy fields.

Cd intake levels from food items other than rice and rice products, which remained constant independent of the flooding of rice paddy fields, were also higher in these areas in Akita than the average in Japan (Figure 2) [15]. Although total Cd intake other than rice and rice products in these areas was nearly twice the average in Japan in 2007; those from other food items varied [15]. Cd intake from cereals, tubers, and roots was 2.3–2.4 μg/day in these areas, which is similar to the average in Japan of 2.7 μg/day. On the other hand, Cd intake from soybeans and soybean products was 3.4 μg/day in these areas, which was approximately 3-fold higher than the average in Japan of 1.1 μg/day. Cd intake from vegetables and seafood in the general Japanese population was reported in the classification as "brightly colored vegetables", "vegetables, seaweeds", and "seafood" by the Ministry of Agriculture, Forestry, and Fisheries. Therefore, we compared total Cd intake from vegetables and seafood in areas A and B with total Cd intake from "brightly colored vegetables", "vegetables, seaweeds", and "seafood" in the general Japanese population; the former, 12.3–13.1 μg/day, was higher than the latter, 8.2 μg/day. In addition, Cd intake from mushrooms in these areas, 2.6–2.8 μg/day, which was absent from the report by the Ministry of Agriculture, Forestry, and Fisheries of Japan, significantly contributed to total Cd intake. These comparisons revealed that agricultural products in these areas contained high Cd concentrations due to Cd contamination in the local farmland, and the consumption of these products was a significant contributor to Cd overexposure among local farmers.

Among agricultural products, Cd concentrations were high in sesame seeds, taro (satoimo in Japanese), yams (yamaimo in Japanese), soybeans, carrot, spinach, burdock (gobo in Japanese), garland chrysanthemum (shungiku in Japanese), Japanese mustard spinach (komatsuna in Japanese), garlic, belvedere fruit (tomburi in Japanese), and shiitake mushroom. Among seafood, many types of seaweed (such as wakame, kombu, and hijiki in Japanese), salted squid guts, scallops (hotate in Japanese) and oysters (kaki in Japanese) with innards, Japanese littleneck clam (asari in Japanese), and freshwater clam (shijimi in Japanese) showed very high Cd concentrations. On the other hand, Cd concentrations were generally low in livestock food and fruit. These results are consistent with previous findings showing that Cd concentrations were high in tubers, soybeans, brightly colored vegetables, and seafood, particularly seaweed, and the innards of squid or shellfish in Japan [10,15].

Median weekly Cd intake levels in our subjects were 7.2 and 6.0 μg/kg BW/week in areas A and B, respectively, which were approximately the PTWI of JECFA at that time or the current TWI in Japan of 7 μg/kg BW/week. In our previous study on area A, we estimated weekly Cd intake levels to be between 5.70–6.72 μg/kg BW/week, which was consistent with the present results, based on 2 assumptions: all food items other than rice may be contaminated by Cd at the same percentage contribution as rice (50%) or at a constant level (15.0 μg/day) [4]. In addition, the percentage of subjects with weekly Cd intake levels higher than 7 μg/kg BW/week in area A, 51.7%, in the present study was similar to previously estimated percentages, 33–51%. These results indicate that previously estimated Cd intake levels, based only on Cd concentrations in rice and miso, were not inaccurate, and that local farmers in Cd-polluted areas in Akita exposed to Cd were at risk of adverse effects. Weekly Cd intake per body weight was above the PTWI or TWI in older subjects in area A, suggesting a higher risk of developing renal tubular dysfunction among the elderly. There were no age-classified subgroups above the PTWI or TWI in area B, but actually the older subjects might had been exposed to much higher levels of Cd like in area A before the start of the flooding of paddy fields.

Two methods are generally employed in exposure assessments of chemicals in food: a total diet study (TDS) and individual food analysis. There are two approaches in TDS: a market basket method

and duplicate portion study [5,18]. In the market basket method, which assesses chemical intake in divided food groups, the average intake of a chemical in a certain population may be calculated; however, the concentration of this chemical in individual food items remains unknown. On the other hand, the duplicate portion study, which measures chemical concentrations in whole meals, provides information on actual chemical intake by individuals, but does not give stable results on chemical intake. We simultaneously performed a duplicate portion study in area A on 17 female farmers for three days to support the present results, and obtained a smaller median Cd intake of 19.0 μg/day that ranged between 9.8–63.1 μg/day (unpublished data). In contrast, the individual food analysis, adopted in the present study, allowed us to identify the source foods of chemical intake and calculate the average intake of a chemical in a population, similar to the market basket method. The individual food analysis generally cannot exclude uncertainties due to processing and cooking foodstu ffs. However, Cd itself does not increase or diminish in foodstu ffs by processing and cooking. Furthermore, mass changes in individual foods were included in the calculation of Cd intake based on the tables of mass changes in individual food, from the Standard Tables of Food Compositions in Japan (7th edition, Japanese Ministry of Education, Culture, Sports, Science, and Technology). Furthermore, a probabilistic assessment, the Monte Carlo simulation, may be performed using individual Cd concentrations in food items.

The National Institute of Health Sciences previously investigated Cd intake by the general Japanese population using the market basket method [15] and an assessment of Cd intake based on the individual food analysis, similar to the present study, was not previously performed in Japan. Cd intake in the Jinzu River basin of Toyama, the heaviest Cd-polluted area in Japan, was reported to be as high as 600 μg/day in 1968 using the market basket method [19]. On the other hand, in Kosaka town in Akita, the intake of Cd from rice by local residents, assessed from Cd concentrations in rice, was 100.6 μg/day in 1974–1976 [20], 92 μg/day in 1978, and 55 μg/day in 1999–2000 using the duplicate portion study [19]; the latter was similar to the present results. These findings indicate that Cd intake in Cd-polluted areas in Akita has been decreasing for decades, and are consistent with the present results. Cd intake in Cd-polluted areas in Akita is still higher than those by the general Japanese population and Western counties. Recent Cd intake by general populations assessed using the market basket method was 4.63 μg/day or 0.54 μg/kg body weight/week in U.S.A. [21], 0.16 μg/kg body weight/week in France [22], 0.77 μg/kg body weight/week in Spain [23], 1 μg/kg body weight/week in Sweden [24], 5.00 μg/day in Italy [25], 0.85μg/kg body weight/week in Belgium [26], and 13.5 μg/day in Denmark [27]. Although Cd intake in Eastern Asian countries, where rice is consumed as a staple food, is generally high, such as 32.7 μg/day in China [28] and 22.0 μg/day in Korea [29], Cd intake levels remain higher in Akita.

There are some limitations in the present study. The data used in this study were obtained in 2001–2003. Nevertheless, the results obtained remain important, even after approximately 20 years, because few Cd-polluted areas remain in Japan. Furthermore, data were collected at the start of the flooding of paddy fields, which has successfully decreased Cd exposure levels in farmers in these areas; therefore, these data will never be obtained again in the future.

Furthermore, the subjects examined were females and therefore the Cd intake by male farmers remains unknown. Cd intake levels by males may be higher than those by females based on di fferences in the amount of food consumed. However, since females are generally more vulnerable to Cd toxicity than males, such as higher intestinal Cd absorbability in females than in males, it is not inappropriate to use the results of the present study for risk assessments of Cd in the general population. In area B, the results of the Cd exposure assessment did not reflect previous Cd intake because of the flooding of paddy fields. However, the Cd intake level in area B was still higher than that by the general population. In DHQ, which assesses general nutritional intake, some food items that showed markedly higher Cd concentrations were absent, such as kelp and hijiki. Therefore, we were unable to include Cd intake from these foods in total Cd intake, which may have led to the underestimation of Cd exposure. The Japanese governmen<sup>t</sup> provides Cd intake values by the general population as averages. Although median values in Cd-polluted areas cannot be statistically compared to these averages, the di fferences are large enough to identify areas at risk of health problems.
