- Email: [email protected]
Cadmium and nutritional intake in pregnant Japanese women
Toxicology Letters 148 (2004) 171–176
Cadmium and nutritional intake in pregnant Japanese women Muneko Nishijo a,∗ , Kenji Tawara a , Ryumon Honda b , Jun-ichi Kuriwaki a , Hideaki Nakagawa a , Kyoko Tanebe c , Shigeru Saito c a
c
Department of Public Health, Kanazawa Medical University, 1-1 Daigaku, Uchinada, Ishikawa 920-0293, Japan b Department of Hygiene, Kanazawa Medical University, 1-1 Daigaku, Uchinada, Ishikawa 920-0293, Japan Department of Obstructive and Gynecology, Faculty of Medicine, Toyama Medical Pharmaceutical University, Toyama, Japan Received 15 August 2003; received in revised form 20 September 2003; accepted 26 September 2003
Abstract A study to clarify the food composition and nutritional factors that contribute to the levels of blood and urinary cadmium (Cd) was conducted on 50 pregnant Japanese women with mean age of 29 years. The mean iron (Fe) intake of subjects was 9.2 mg, which is much lower than the recommended level of 20 mg for pregnant women. Cd in urine samples collected at 30–32 weeks of gestation were correlated (r = 0.354), but urinary Cd was related to age more than blood Cd. Urinary Cd and blood Cd levels were inversely related to total energy (rpartial = −0.325, and − 0.334, respectively) and fat intake (rpartial = −0.419, and − 0.379, respectively), even after adjustment for age. Blood Cd was also correlated to protein and iron intake (rpartial = −0.299, and − 0.353, respectively). These results indicate that Cd exposure levels of pregnant women with low energy intake, especially less fat intake, were higher than those of women with more energy and fat intake. In particular, blood Cd may be affected by protein and iron intake in pregnant women with increased these nutrients demand. © 2003 Elsevier Ireland Ltd. All rights reserved. Keywords: Nutritional intake; Cadmium; Pregnant women
1. Introduction Since the description of itai-itai disease in Japan, cadmium (Cd) exposure levels have been extensively investigated in that country. The relationships between biomarkers of Cd exposure and Cd concentration in rice have been described in inhabitants living in cadmium-polluted areas in Japan (Nogawa et al., 1978; Kido et al., 1992). In addition, a survey ∗ Corresponding author. Tel.: +81-76-286-2211; fax: +81-76-286-3728. E-mail address: [email protected] (M. Nishijo).
to clarify urinary and blood Cd levels which reflect Cd body burden and recent exposure (IPCS, 1992) in the Japanese general population aged more than 50 years living in a non-polluted area was conducted recently: a higher level of Cd exposure than those shown in European studies was reported even for inhabitants living in non-polluted areas (Swazono et al., 2000). One reason for this higher Cd exposure level in Japanese is suspected to be higher Cd levels in rice, which is a staple food for Japanese. Recently, Shimbo et al. (2000a) conducted surveys to clarify the Cd level of younger women in their 20s in Japan. The study reported that their blood and urinary Cd
0378-4274/$ – see front matter © 2003 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.toxlet.2003.09.016
172
M. Nishijo et al. / Toxicology Letters 148 (2004) 171–176
concentrations were lower, and that their intake of rice was much less than that of middle aged women. However, they also suggested that rice was still a greater Cd source than other food, contributing 40% of their Cd intake (Watanabe et al., 2000). Not only amounts of food with high Cd content, but also food components and nutritional factors may influence the increase of blood and urinary Cd. Berglund et al. (1994) reported that dietary fiber intake was related to blood Cd levels in non-pregnant women under 50, including vegetarians and women on mixed diets. Another Swedish study suggested that blood Cd of pregnant women is higher than that of non-pregnant women at the same age because of the increased Cd absorption engendered by depletion of body iron (Fe) stores during pregnancy (Järup et al., 1998). However, very few studies have considered the importance of blood and urinary Cd in Japanese mothers. In particular, information on the influence of nutritional factors and food on Cd levels is limited for pregnant women, whose nutritional demands are increased. The objective of this study was to investigate Cd exposure levels and to determine food composition and nutritional factors affecting blood and urinary Cd in young Japanese pregnant women.
2. Materials and methods 2.1. Subjects We recruited 50 pregnant women (30–32 weeks of gestation), who were interested in their Cd exposure levels, and cooperated in providing their blood and urine samples and recording their diet. They had all visited Toyama Medical Pharmaceutical University Hospital for maternity health checks and infant delivery, and were not selected for prenatal medical problems. They were all Japanese citizens living in an area near the hospital. Three women lived in the cadmium-polluted Jinzu River basin located in this area. More than 70% were housewives whose most common occupation before marriage was ‘office worker’; they had no occupational history of Cd exposure. No significantly different characteristics related to socioeconomic status were found among all subjects. Informed consent for this study was obtained from these subjects in an appropriate manner.
2.2. Analysis of urine and blood samples Spot urine and heparinized whole blood samples were collected on the day when they visited the hospital in the morning for maternity health check-ups at 30–32 gestational weeks, and were kept frozen at −20 ◦ C until analysis. At that time, the subjects underwent physical examination including measurement of their height, weight, blood count, and serum Fe and ferritin. Serum Fe was measured by colorimetric assay and serum ferritin was analyzed by latex immuno agglutination method conducted with an automated system. Dietary diaries for three consecutive days were distributed. A nutritional survey was conducted during the next visit in 32–34 gestational weeks by collecting dietary diary records. The record contents were checked through a direct interview conducted by a dietitian using food models, eating utensils, and photos of food for subjects to accurately monitor the amount of food eaten for three days. The food and nutrition intake was calculated for each individual using these results based on the Standard Tables for Food Composition in Japan, Fourth Revised Edition (Resources Council, science and Technology Agency, Japan, 1985). Then the daily average intake was calculated. The Cd concentration was measured using a flameless atomic absorption spectrophotometer (AAS) (Model 180-80; Hitachi Co, Japan). Before Cd analysis, blood samples were diluted with 10% HNO3 and urinary samples were digested by wet ashing in HNO3 /H2 SO4 /HClO4 and extraction with ammonium pyrrolidine dithiocarbamate- methyl isobutyl ketone (APDC-MIBK) (Honda et al., 1979). Urine reference material No. 2670 (The National Bureau of Standards, Washington, D.C.) and control blood (Behring Institut, Lot No. 620302) were used to test the accuracy and precision of the analytical method of determining Cd. The certified value (confidence range) of these controls were 88 (94–82) g/l for urine and 5.0 (4.2–5.8) g/l for blood. Our measurement of Cd was 87.3 (83.7–90.9) g/l for urinary samples and 5.1 (4.3–5.7) g/l for blood samples. Urinary Cd concentration was corrected for the urinary creatinine concentration determined by using the reaction method of Jaffe.
M. Nishijo et al. / Toxicology Letters 148 (2004) 171–176
2.3. Statistical analysis
2.5
The geometrical mean of urinary Cd concentration was 1.05 g/g Cd (geometrical standard deviation = 1.64), and the mean of blood Cd was 0.93 g/l (standard deviation = 0.46) (Table 1). The mean age was 29.3 years for the range of 22–39 years. Correlation between urinary Cd and age was positive and significant with a correlation coefficient of 0.354 (P = 0.019). However, the blood Cd highly correlated to urinary Cd (Fig. 1), although it did not show the significant correlation with age (r = 0.207, P = 0.150). Primipara subjects were 50% of the subjects, Table 1 Population characteristics
b
1.5 1 0.5 0 0.1
1
10
Urinary Cd (µg/gCr) Fig. 1. The relationship between blood and urinary cadmium (logtransformed value). Correlation coefficient (Spearman) = 0.594 (P < 0.001).
3. Results
a
2
Blood Cd (µg/L)
The log-transformed values of urinary Cd were used for statistical analysis, and geometric mean and geometric standard deviation of urinary Cd were calculated. Simple correlations between indices of Cd exposure and amount of nutrients and food intake were tested by Spearman’s rho; then multiple regression analysis was used to compute partial correlation coefficients between them after adjusting for age. These statistical analyses were all performed using the SPSS (version 10.0) package.
Urinary Cd (g/g Cr)a Blood Cd (g/l) Age (year) Height (cm) Weight before pregnancy (kg) BMI before pregnancy Weight (kg) BMI RBC (×104 /mm3 ) Hb (g) Ht (%) Serum Feb (mg/dl) Serum ferritinb <5 ng/ml, n (%) Primipara, n (%) Smoking habit before pregnancy, n (%) During pregnancy, n (%)
173
Mean
S.D.
1.05 0.93 29.3 157.4 52.9 21.3 60.9 42.6 367.8 10.8 33.2 52.8 20 25 9 2
1.64 0.46 4.3 4.6 8.5 3.1 7.9 2.8 25.7 0.9 2.4 33.6 41.7 50 18 4
Geometrical mean (geometrical standard deviation). The number of the subjects examined was 48.
but those subjects showed no significant difference of urinary and blood Cd, as compared with multipara. The number of subjects who smoked before pregnancy was 9 (18%): their urinary and blood Cd concentrations were not different from those of non-smoking women. Table 1 shows physical measurements before pregnancy and 30–32 weeks’ gestation and indices of anemia including serum Fe and serum ferritin at 30–32 weeks’ gestation. Height and weight are typical for Japanese women; moreover, the body mass indexes (BMI) before pregnancy of the subjects were within normal limits except for one subject. The mean body weight gain during pregnancy was 8 kg, indicating adequate progress of pregnancy. These subjects were slightly anemic, with their Hb below 12 g that is the lower range of Hb for healthy women (Table 1). In addition, mean of serum Fe was 52.8 mg/dl closed to 50 mg/dl which is the lower range of reference value, and 41.7% of these subjects had depletion of serum ferritin (<5 ng/ml), as usual in the third trimester of pregnancy (Table 1). However, no significant relationships between these anemia indices and Cd exposure indices were observed in these subjects. Their nutritional intake including macronutrients, calcium and iron is shown in Table 2. The total energy intake was 1799 kcal, and this figure conforms to optimal values for 30-year-old women, but was lower than optimal for pregnant women listed in the Recommended Dietary Allowance for Japanese, Fifth Version (1994). The means of protein, fat and carbohydrate intake converted to a percentage of total
174
M. Nishijo et al. / Toxicology Letters 148 (2004) 171–176
Table 2 Nutritional intake and biomarkers of Cd exposure Total Energy intake (kcal) Mean S.D. Max Min
Protein intake (g)
1799 321 2642 1282
Fat intake (g)
Carbonate intake (g)
Protein energy ratio
Fat energy ratio
Carbonate energy ratio
644 227 1296 288
Iron intake (mg)
67.2 15.3 109 44.6
59.6 15.8 97.5 34.8
244.9 50.8 385.1 158.6
14.9 2.1 19.6 10.4
29.7 5.2 41 20.1
−0.273 −0.156
−0.431∗∗ −0.45∗∗
−0.097 −0.064
−0.094 0.011
−0.257 −0.362∗
0.272 0.338∗
−0.152 −0.029
−0.346∗ −0.209
Partial correlation coefficient controlling for age −0.299∗ −0.379∗∗ Blood Cd −0.334∗ Urinary Cd −0.325∗ −0.226 −0.419∗∗
−0.187 −0.151
−0.081 0.072
−0.232 −0.298∗
0.236 0.241
−0.239 −0.132
−0.353∗ −0.28
Correlation coefficient Blood Cd −0.325∗ Urinary Cd −0.291∗
55.4 5.8 67.6 43.6
Calcium intake (mg)
9.2 2.4 6 15.7
Correlation coefficients were tested by Spearman’s rho. ∗ P < 0.05. ∗∗ P < 0.01.
energy intake, confirmed with the recommended value. The mean Fe intake was 9.2 mg, which is much lower than the recommended 20 mg for pregnant women. Urinary Cd was inversely related to total energy and fat intake, and fat energy ratio; it was positively related to the carbohydrate energy ratio (Table 2). In addition, multiple regression analysis indicated partial correlation between urinary Cd and nutritional intake because fat intake and urinary Cd are related to age. Table 2 shows significant inverse relationships between urinary Cd and total energy intake, fat intake, and fat intake ratio after adjustment for age. These results indicate that urinary Cd is higher in women who
consumed diets with low fat content than those taking high fat diets. Blood Cd was also inversely correlated with total energy and fat intake even after controlling for age, but it is not related to fat energy ratio (Table 2). These results indicate that there is the relationship between blood Cd and lower food intakes, especially foods rich in fat. Blood Cd also correlated inversely to Fe and protein intake, even after adjustment for age (Table 2). These results suggest that blood Cd levels of women with lower iron intake are higher than those of women of the same age with higher iron intake because protein intake was highly related to Fe intake in these subjects.
Table 3 Food intake and biomarkers of cadmium exposure Cereals Mean (g) S.D. Max (g) Min (g)
394.4 113.7 691 183
Correlation coefficient Blood Cd 0.133 Urinary Cd 0.071
Potato
Oil
Green vegetable
41.1 36.7 123.3 0
17.2 7.8 36 1.7
108.1 67.9 330.7 2.5
−0.126 −0.089
−0.268 −0.239
−0.191 −0.007
Partial correlation coefficient controlling for age Blood Cd 0.117 −0.252 −0.232 Urinary Cd 0.011 −0.212 −0.221
−0.299∗ −0.017
Correlation coefficients were tested by Spearman’s rho. ∗ P < 0.05. ∗∗ P < 0.01.
Other vegetable 132.5 75.1 445 33
Fish
Meat
Egg
Milk
55.7 34.6 221.3 2.3
59.3 30.1 164 11.7
35.1 21.1 84 0
252.1 142.1 815 0
0.058 0.067
−0.081 −0.051
−0.364∗∗ −0.334∗
−0.065 −0.187
−0.09 0.03
−0.118 −0.112
−0.174 −0.155
−0.236 −0.212
0.044 −0.03
−0.167 0.037
M. Nishijo et al. / Toxicology Letters 148 (2004) 171–176
Next, correlations between indices of Cd exposure and intake by food groups were investigated to clarify that the food component contributed to higher Cd exposure level. Those results are shown in Table 3. Both blood and urinary Cd were inversely correlated to meat intake. However, partial coefficients between blood Cd and meat controlling for age were not significant because meat intake was related to age in these subjects. Blood Cd was also inversely related to iron-rich green vegetables like spinach, but no significant correlation was found among other food groups and indices of Cd exposure after elimination of age effects.
4. Discussion Food is the main source of Cd exposure among the non-smoking, non-occupationally exposed general population (IPCS, 1992; Järup et al., 1998), but nutrition may also represent an influential factor for gastrointestinal Cd absorption. Iron status influences the Cd exposure level because of the increase of Cd absorption accompanying the increase of Fe absorption resulting from the depletion of body Fe stores. Berglund et al. (1994) reported that blood Cd was inversely correlated with serum ferritin, which is a marker of body, Fe status in Swedish women, including subjects with low body iron stores because of their dietary habits. Jacobsson Lagerkvist et al. (1993) reported that higher blood Cd level during pregnancy in nonsmoking and ex-smoking women, as compared with the level of non-pregnant women without smoking in Sweden. Åkesson et al. (2002) investigated the relationship between Fe status markers and blood and urinary Cd in pregnancy and lactation. That study found increased blood and urinary Cd in lactation of women with tissue iron deficiency in late pregnancy. These reports suggest that iron deficiency is an important factor regulating Cd body burden. Women in late pregnancy with exhausted body iron stores are at high risk for Cd exposure. The present study detected no significant relationships between serum ferritin and blood and urinary Cd measured at the same time. However, lower Fe intake correlated with higher blood Cd. Moreover, lower intake of green vegetables, an important source of Fe in Japan, was related to higher blood Cd. These results
175
suggested that blood Cd in late pregnancy might be affected by Fe intake because almost all pregnant women have lower Fe status as a consequence of fetal growth. We found that protein intake was inversely correlated with blood Cd. This result also suggested the influence of Fe intake on blood Cd because protein is a macronutrient whose intake showed the highest correlation with Fe intake (r = 0.805, P < 0.0001) in these subjects. Even though blood and urinary Cd were highly correlated, the relations of Fe, protein, and green vegetables intakes to urinary Cd were not significant. One explanation for these differences is that blood Cd reflects not only the Cd burden on the body, but also current exposure through food during pregnancy. Fat intake was inversely related to both urinary and blood Cd in these subjects. In recent years, Japanese, especially young people, have adopted a diet characterized by increased fat intake and decreased rice consumption. In addition, the mean of fat intake energy ratio among the present subjects was the upper limit of the recommended value for Japanese; the food groups contributing to fat intake were oils (r = 0.669, P < 0.0001) and meats (r = 0.409, P < 0.01). Meat intake was inversely related to blood and urinary Cd, although this relationship was suspected to be biased by age. Therefore, women who prefer a low fat diet achieved through low meat and oil intake might have a greater Cd body burden. An inverse relation between total energy intake and these indices of Cd exposure was also observed, even after adjustment for age. Among present subjects, the total energy intake was related to intakes of eggs, milk, oils, and meats. This result also suggested that lower energy intake as a result of lower animal food intake is related to higher Cd body burden in Japanese women. A Swedish study (Berglund et al., 1994) of nonpregnant women showed that non-smoking Swedish women eating diets rich in cereals and root vegetables (i.e., a high fiber diet) or a shell fish diet had a higher Cd intake than women eating a mixed diet. However, intake of cereals and vegetables, excluding green vegetables, was not related to biomarkers of Cd exposure in the present study. Also, the carbohydrate energy ratio was positively related to urinary Cd, but its relation was not significant after adjustment for age because both of carbohydrate intake and urinary Cd were correlated with the mothers’ age.
176
M. Nishijo et al. / Toxicology Letters 148 (2004) 171–176
Fish intake in these subjects also showed no relation to blood and urinary Cd. This difference from the Swedish report may arise from the different types of diets in Sweden and Japan, and methods of the food intake survey. However, this time we did not ask about the frequency of eating shellfish, mollusks, and crustaceans, which accumulate Cd well, but are often consumed in this area. In the future, the survey of frequency of non-fish seafood with high Cd content will be necessary to clarify their contribution to Cd exposure. In Japanese cadmium-polluted areas, rice has been considered to be the main source of Cd. Recent studies targeted to middle-aged farmers’ wives showed that total Cd intake from rice contributed 40% of total Cd intake (Watanabe et al., 2000). This study analyzed correlations between daily intake of boiled rice and indices of Cd exposure, but could not detect significant correlation. Probably, this result may be caused by less rice contribution to Cd exposure in our subjects, who were younger and consumed less boiled rice than those reported by Watanabe et al. (2000). However, Shimbo et al. (2000b) reported that there was regional difference of Cd content in rice: the rice collected in our area showed the highest Cd content among samples from various regions in Japan. Future larger scale surveys targeted to young people in this area are necessary to clarify the relation between rice intake and Cd exposure.
Acknowledgements We thank Makiko Ebie, Mutumi Katsuo, Aki Ohta in Toyama Medical Pharmaceutical University Hospital who corrected data, and Katushi Yoshita who cooperated with nutritional calculation. This work was supported by grant for Project Research from High-Technology Center of Kanazawa Medical University (H2001-4).
References Åkesson, A., Berglund, M., Schutz, A., Bjellerup, P., Bremme, K., Vahter, M., 2002. Cadmium exposure in pregnancy and
lactation in relation to iron status. Am. J. Public Health 321, 177–181. Berglund, M., Akesson, A., Nermell, B., Vahter, M., 1994. Intestinal absorption of dietary cadmium in women depends on body iron stores and fiber intake. Environ. Health Perspect. 102, 1058–1066. Health promotion and nutrition devision, Health service Bureau, Ministry of Health and welfare of Japan, 1994. Recommended Dietary Allowance for Japanese, fifth revision, Daiichi Shuppan, Tokyo, pp. 26, 67–72. Honda, R., Nogawa, K., Kobayashi, E., Ishizaki, A., 1979. A simplified determination of urinary cadmium by flameless atomic absorption spectrophotometry using a heated graphite atomizer. Hokuriku J. Public Health (in Japanese) 6, 13–19. IPCS, 1992. International Programme on Chmical Safety, Environmental Health Criteria 134, Cadmium, WHO, Geneva. Jacobsson Lagerkvist, B., Soderberg, H.A., Nordberg, G.F., Ekesrydh, S., Englyst, V., 1993. Biological monitoring of arsenic, lead and cadmium in occupationally and environmentally exposed pregnant women. Scand. J. Work Environ. Health 19 (Suppl 1), 50–53. Järup, L., Berglund, M., Elinder, C.G., Vahter, M., 1998. Health effects of cadmium exposure—a review of the literature and a risk estimate. Scand. J. Work Environ. Health 24 (Suppl 1), 7–51. Kido, T., Nogawa, K., Ohmichi, M., Honda, R., Tsuritani, I., Ishizaki, M., Yamada, Y., 1992. Significance of urinary cadmium concenration in a Japanese population environmentally exposed to cadmium. Arch. Environ. Health 47, 196–202. Nogawa, K., Ishizaki, A., Kawano, S., 1978. Statistical observation of the dose–response relationship of cadmium based on epidemiological studies in the Kakehashi River basin. Environ. Res. 15, 185–198. Resources Council, Science and Technology Agency, Japan, 1985. Standard Tables for Food Composition in Japan, fourth revised ed. (1982), Daiichi Shuppan, Tokyo, pp. 18–196. Shimbo, S., Zhang, Z.W., Moon, C.S., Watanabe, T., Nakatsuka, H., Matsuda-Inoguchi, N., Hogashikawa, K., Ikeda, M., 2000a. Correlation between urine and blood concentrations and dietary intake of cadmium and lead among women in the general population of Japan. Int. Arch. Occup. Environ. Health 73, 163–170. Shimbo, S., Zhang, Z.W., Watanabe, T., Matsuda-Inoguchi, N., Nakatsuka, H., 2000b. Survey on dietary exposure of Japanese population to cadmium. Kankyo Hoken Report 67, 186–191 (in Japanese). Swazono, Y., Kobayashi, E., Okubo, Y., Nogawa, K., Kido, T., Nakagawa, H., 2000. Renal effects of cadmium exposure in cadmium non-polluted areas in Japan. Environ. Res. 84, 44– 55. Watanabe, T., Zhang, Z.W., Moon, C.S., Shimbo, S., Nakatsuka, H., Matsuda-Inoguchi, N., Hogashikawa, K., Ikeda, M., 2000. Cadmium exposure of women in general populations in Japan during 1991–1997 compared with 1997–1981. Int. Arch. Occup. Environ. Health 73, 26–34.
Cadmium and nutritional intake in pregnant Japanese women Muneko Nishijo a,∗ , Kenji Tawara a , Ryumon Honda b , Jun-ichi Kuriwaki a , Hideaki Nakagawa a , Kyoko Tanebe c , Shigeru Saito c a
c
Department of Public Health, Kanazawa Medical University, 1-1 Daigaku, Uchinada, Ishikawa 920-0293, Japan b Department of Hygiene, Kanazawa Medical University, 1-1 Daigaku, Uchinada, Ishikawa 920-0293, Japan Department of Obstructive and Gynecology, Faculty of Medicine, Toyama Medical Pharmaceutical University, Toyama, Japan Received 15 August 2003; received in revised form 20 September 2003; accepted 26 September 2003
Abstract A study to clarify the food composition and nutritional factors that contribute to the levels of blood and urinary cadmium (Cd) was conducted on 50 pregnant Japanese women with mean age of 29 years. The mean iron (Fe) intake of subjects was 9.2 mg, which is much lower than the recommended level of 20 mg for pregnant women. Cd in urine samples collected at 30–32 weeks of gestation were correlated (r = 0.354), but urinary Cd was related to age more than blood Cd. Urinary Cd and blood Cd levels were inversely related to total energy (rpartial = −0.325, and − 0.334, respectively) and fat intake (rpartial = −0.419, and − 0.379, respectively), even after adjustment for age. Blood Cd was also correlated to protein and iron intake (rpartial = −0.299, and − 0.353, respectively). These results indicate that Cd exposure levels of pregnant women with low energy intake, especially less fat intake, were higher than those of women with more energy and fat intake. In particular, blood Cd may be affected by protein and iron intake in pregnant women with increased these nutrients demand. © 2003 Elsevier Ireland Ltd. All rights reserved. Keywords: Nutritional intake; Cadmium; Pregnant women
1. Introduction Since the description of itai-itai disease in Japan, cadmium (Cd) exposure levels have been extensively investigated in that country. The relationships between biomarkers of Cd exposure and Cd concentration in rice have been described in inhabitants living in cadmium-polluted areas in Japan (Nogawa et al., 1978; Kido et al., 1992). In addition, a survey ∗ Corresponding author. Tel.: +81-76-286-2211; fax: +81-76-286-3728. E-mail address: [email protected] (M. Nishijo).
to clarify urinary and blood Cd levels which reflect Cd body burden and recent exposure (IPCS, 1992) in the Japanese general population aged more than 50 years living in a non-polluted area was conducted recently: a higher level of Cd exposure than those shown in European studies was reported even for inhabitants living in non-polluted areas (Swazono et al., 2000). One reason for this higher Cd exposure level in Japanese is suspected to be higher Cd levels in rice, which is a staple food for Japanese. Recently, Shimbo et al. (2000a) conducted surveys to clarify the Cd level of younger women in their 20s in Japan. The study reported that their blood and urinary Cd
0378-4274/$ – see front matter © 2003 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.toxlet.2003.09.016
172
M. Nishijo et al. / Toxicology Letters 148 (2004) 171–176
concentrations were lower, and that their intake of rice was much less than that of middle aged women. However, they also suggested that rice was still a greater Cd source than other food, contributing 40% of their Cd intake (Watanabe et al., 2000). Not only amounts of food with high Cd content, but also food components and nutritional factors may influence the increase of blood and urinary Cd. Berglund et al. (1994) reported that dietary fiber intake was related to blood Cd levels in non-pregnant women under 50, including vegetarians and women on mixed diets. Another Swedish study suggested that blood Cd of pregnant women is higher than that of non-pregnant women at the same age because of the increased Cd absorption engendered by depletion of body iron (Fe) stores during pregnancy (Järup et al., 1998). However, very few studies have considered the importance of blood and urinary Cd in Japanese mothers. In particular, information on the influence of nutritional factors and food on Cd levels is limited for pregnant women, whose nutritional demands are increased. The objective of this study was to investigate Cd exposure levels and to determine food composition and nutritional factors affecting blood and urinary Cd in young Japanese pregnant women.
2. Materials and methods 2.1. Subjects We recruited 50 pregnant women (30–32 weeks of gestation), who were interested in their Cd exposure levels, and cooperated in providing their blood and urine samples and recording their diet. They had all visited Toyama Medical Pharmaceutical University Hospital for maternity health checks and infant delivery, and were not selected for prenatal medical problems. They were all Japanese citizens living in an area near the hospital. Three women lived in the cadmium-polluted Jinzu River basin located in this area. More than 70% were housewives whose most common occupation before marriage was ‘office worker’; they had no occupational history of Cd exposure. No significantly different characteristics related to socioeconomic status were found among all subjects. Informed consent for this study was obtained from these subjects in an appropriate manner.
2.2. Analysis of urine and blood samples Spot urine and heparinized whole blood samples were collected on the day when they visited the hospital in the morning for maternity health check-ups at 30–32 gestational weeks, and were kept frozen at −20 ◦ C until analysis. At that time, the subjects underwent physical examination including measurement of their height, weight, blood count, and serum Fe and ferritin. Serum Fe was measured by colorimetric assay and serum ferritin was analyzed by latex immuno agglutination method conducted with an automated system. Dietary diaries for three consecutive days were distributed. A nutritional survey was conducted during the next visit in 32–34 gestational weeks by collecting dietary diary records. The record contents were checked through a direct interview conducted by a dietitian using food models, eating utensils, and photos of food for subjects to accurately monitor the amount of food eaten for three days. The food and nutrition intake was calculated for each individual using these results based on the Standard Tables for Food Composition in Japan, Fourth Revised Edition (Resources Council, science and Technology Agency, Japan, 1985). Then the daily average intake was calculated. The Cd concentration was measured using a flameless atomic absorption spectrophotometer (AAS) (Model 180-80; Hitachi Co, Japan). Before Cd analysis, blood samples were diluted with 10% HNO3 and urinary samples were digested by wet ashing in HNO3 /H2 SO4 /HClO4 and extraction with ammonium pyrrolidine dithiocarbamate- methyl isobutyl ketone (APDC-MIBK) (Honda et al., 1979). Urine reference material No. 2670 (The National Bureau of Standards, Washington, D.C.) and control blood (Behring Institut, Lot No. 620302) were used to test the accuracy and precision of the analytical method of determining Cd. The certified value (confidence range) of these controls were 88 (94–82) g/l for urine and 5.0 (4.2–5.8) g/l for blood. Our measurement of Cd was 87.3 (83.7–90.9) g/l for urinary samples and 5.1 (4.3–5.7) g/l for blood samples. Urinary Cd concentration was corrected for the urinary creatinine concentration determined by using the reaction method of Jaffe.
M. Nishijo et al. / Toxicology Letters 148 (2004) 171–176
2.3. Statistical analysis
2.5
The geometrical mean of urinary Cd concentration was 1.05 g/g Cd (geometrical standard deviation = 1.64), and the mean of blood Cd was 0.93 g/l (standard deviation = 0.46) (Table 1). The mean age was 29.3 years for the range of 22–39 years. Correlation between urinary Cd and age was positive and significant with a correlation coefficient of 0.354 (P = 0.019). However, the blood Cd highly correlated to urinary Cd (Fig. 1), although it did not show the significant correlation with age (r = 0.207, P = 0.150). Primipara subjects were 50% of the subjects, Table 1 Population characteristics
b
1.5 1 0.5 0 0.1
1
10
Urinary Cd (µg/gCr) Fig. 1. The relationship between blood and urinary cadmium (logtransformed value). Correlation coefficient (Spearman) = 0.594 (P < 0.001).
3. Results
a
2
Blood Cd (µg/L)
The log-transformed values of urinary Cd were used for statistical analysis, and geometric mean and geometric standard deviation of urinary Cd were calculated. Simple correlations between indices of Cd exposure and amount of nutrients and food intake were tested by Spearman’s rho; then multiple regression analysis was used to compute partial correlation coefficients between them after adjusting for age. These statistical analyses were all performed using the SPSS (version 10.0) package.
Urinary Cd (g/g Cr)a Blood Cd (g/l) Age (year) Height (cm) Weight before pregnancy (kg) BMI before pregnancy Weight (kg) BMI RBC (×104 /mm3 ) Hb (g) Ht (%) Serum Feb (mg/dl) Serum ferritinb <5 ng/ml, n (%) Primipara, n (%) Smoking habit before pregnancy, n (%) During pregnancy, n (%)
173
Mean
S.D.
1.05 0.93 29.3 157.4 52.9 21.3 60.9 42.6 367.8 10.8 33.2 52.8 20 25 9 2
1.64 0.46 4.3 4.6 8.5 3.1 7.9 2.8 25.7 0.9 2.4 33.6 41.7 50 18 4
Geometrical mean (geometrical standard deviation). The number of the subjects examined was 48.
but those subjects showed no significant difference of urinary and blood Cd, as compared with multipara. The number of subjects who smoked before pregnancy was 9 (18%): their urinary and blood Cd concentrations were not different from those of non-smoking women. Table 1 shows physical measurements before pregnancy and 30–32 weeks’ gestation and indices of anemia including serum Fe and serum ferritin at 30–32 weeks’ gestation. Height and weight are typical for Japanese women; moreover, the body mass indexes (BMI) before pregnancy of the subjects were within normal limits except for one subject. The mean body weight gain during pregnancy was 8 kg, indicating adequate progress of pregnancy. These subjects were slightly anemic, with their Hb below 12 g that is the lower range of Hb for healthy women (Table 1). In addition, mean of serum Fe was 52.8 mg/dl closed to 50 mg/dl which is the lower range of reference value, and 41.7% of these subjects had depletion of serum ferritin (<5 ng/ml), as usual in the third trimester of pregnancy (Table 1). However, no significant relationships between these anemia indices and Cd exposure indices were observed in these subjects. Their nutritional intake including macronutrients, calcium and iron is shown in Table 2. The total energy intake was 1799 kcal, and this figure conforms to optimal values for 30-year-old women, but was lower than optimal for pregnant women listed in the Recommended Dietary Allowance for Japanese, Fifth Version (1994). The means of protein, fat and carbohydrate intake converted to a percentage of total
174
M. Nishijo et al. / Toxicology Letters 148 (2004) 171–176
Table 2 Nutritional intake and biomarkers of Cd exposure Total Energy intake (kcal) Mean S.D. Max Min
Protein intake (g)
1799 321 2642 1282
Fat intake (g)
Carbonate intake (g)
Protein energy ratio
Fat energy ratio
Carbonate energy ratio
644 227 1296 288
Iron intake (mg)
67.2 15.3 109 44.6
59.6 15.8 97.5 34.8
244.9 50.8 385.1 158.6
14.9 2.1 19.6 10.4
29.7 5.2 41 20.1
−0.273 −0.156
−0.431∗∗ −0.45∗∗
−0.097 −0.064
−0.094 0.011
−0.257 −0.362∗
0.272 0.338∗
−0.152 −0.029
−0.346∗ −0.209
Partial correlation coefficient controlling for age −0.299∗ −0.379∗∗ Blood Cd −0.334∗ Urinary Cd −0.325∗ −0.226 −0.419∗∗
−0.187 −0.151
−0.081 0.072
−0.232 −0.298∗
0.236 0.241
−0.239 −0.132
−0.353∗ −0.28
Correlation coefficient Blood Cd −0.325∗ Urinary Cd −0.291∗
55.4 5.8 67.6 43.6
Calcium intake (mg)
9.2 2.4 6 15.7
Correlation coefficients were tested by Spearman’s rho. ∗ P < 0.05. ∗∗ P < 0.01.
energy intake, confirmed with the recommended value. The mean Fe intake was 9.2 mg, which is much lower than the recommended 20 mg for pregnant women. Urinary Cd was inversely related to total energy and fat intake, and fat energy ratio; it was positively related to the carbohydrate energy ratio (Table 2). In addition, multiple regression analysis indicated partial correlation between urinary Cd and nutritional intake because fat intake and urinary Cd are related to age. Table 2 shows significant inverse relationships between urinary Cd and total energy intake, fat intake, and fat intake ratio after adjustment for age. These results indicate that urinary Cd is higher in women who
consumed diets with low fat content than those taking high fat diets. Blood Cd was also inversely correlated with total energy and fat intake even after controlling for age, but it is not related to fat energy ratio (Table 2). These results indicate that there is the relationship between blood Cd and lower food intakes, especially foods rich in fat. Blood Cd also correlated inversely to Fe and protein intake, even after adjustment for age (Table 2). These results suggest that blood Cd levels of women with lower iron intake are higher than those of women of the same age with higher iron intake because protein intake was highly related to Fe intake in these subjects.
Table 3 Food intake and biomarkers of cadmium exposure Cereals Mean (g) S.D. Max (g) Min (g)
394.4 113.7 691 183
Correlation coefficient Blood Cd 0.133 Urinary Cd 0.071
Potato
Oil
Green vegetable
41.1 36.7 123.3 0
17.2 7.8 36 1.7
108.1 67.9 330.7 2.5
−0.126 −0.089
−0.268 −0.239
−0.191 −0.007
Partial correlation coefficient controlling for age Blood Cd 0.117 −0.252 −0.232 Urinary Cd 0.011 −0.212 −0.221
−0.299∗ −0.017
Correlation coefficients were tested by Spearman’s rho. ∗ P < 0.05. ∗∗ P < 0.01.
Other vegetable 132.5 75.1 445 33
Fish
Meat
Egg
Milk
55.7 34.6 221.3 2.3
59.3 30.1 164 11.7
35.1 21.1 84 0
252.1 142.1 815 0
0.058 0.067
−0.081 −0.051
−0.364∗∗ −0.334∗
−0.065 −0.187
−0.09 0.03
−0.118 −0.112
−0.174 −0.155
−0.236 −0.212
0.044 −0.03
−0.167 0.037
M. Nishijo et al. / Toxicology Letters 148 (2004) 171–176
Next, correlations between indices of Cd exposure and intake by food groups were investigated to clarify that the food component contributed to higher Cd exposure level. Those results are shown in Table 3. Both blood and urinary Cd were inversely correlated to meat intake. However, partial coefficients between blood Cd and meat controlling for age were not significant because meat intake was related to age in these subjects. Blood Cd was also inversely related to iron-rich green vegetables like spinach, but no significant correlation was found among other food groups and indices of Cd exposure after elimination of age effects.
4. Discussion Food is the main source of Cd exposure among the non-smoking, non-occupationally exposed general population (IPCS, 1992; Järup et al., 1998), but nutrition may also represent an influential factor for gastrointestinal Cd absorption. Iron status influences the Cd exposure level because of the increase of Cd absorption accompanying the increase of Fe absorption resulting from the depletion of body Fe stores. Berglund et al. (1994) reported that blood Cd was inversely correlated with serum ferritin, which is a marker of body, Fe status in Swedish women, including subjects with low body iron stores because of their dietary habits. Jacobsson Lagerkvist et al. (1993) reported that higher blood Cd level during pregnancy in nonsmoking and ex-smoking women, as compared with the level of non-pregnant women without smoking in Sweden. Åkesson et al. (2002) investigated the relationship between Fe status markers and blood and urinary Cd in pregnancy and lactation. That study found increased blood and urinary Cd in lactation of women with tissue iron deficiency in late pregnancy. These reports suggest that iron deficiency is an important factor regulating Cd body burden. Women in late pregnancy with exhausted body iron stores are at high risk for Cd exposure. The present study detected no significant relationships between serum ferritin and blood and urinary Cd measured at the same time. However, lower Fe intake correlated with higher blood Cd. Moreover, lower intake of green vegetables, an important source of Fe in Japan, was related to higher blood Cd. These results
175
suggested that blood Cd in late pregnancy might be affected by Fe intake because almost all pregnant women have lower Fe status as a consequence of fetal growth. We found that protein intake was inversely correlated with blood Cd. This result also suggested the influence of Fe intake on blood Cd because protein is a macronutrient whose intake showed the highest correlation with Fe intake (r = 0.805, P < 0.0001) in these subjects. Even though blood and urinary Cd were highly correlated, the relations of Fe, protein, and green vegetables intakes to urinary Cd were not significant. One explanation for these differences is that blood Cd reflects not only the Cd burden on the body, but also current exposure through food during pregnancy. Fat intake was inversely related to both urinary and blood Cd in these subjects. In recent years, Japanese, especially young people, have adopted a diet characterized by increased fat intake and decreased rice consumption. In addition, the mean of fat intake energy ratio among the present subjects was the upper limit of the recommended value for Japanese; the food groups contributing to fat intake were oils (r = 0.669, P < 0.0001) and meats (r = 0.409, P < 0.01). Meat intake was inversely related to blood and urinary Cd, although this relationship was suspected to be biased by age. Therefore, women who prefer a low fat diet achieved through low meat and oil intake might have a greater Cd body burden. An inverse relation between total energy intake and these indices of Cd exposure was also observed, even after adjustment for age. Among present subjects, the total energy intake was related to intakes of eggs, milk, oils, and meats. This result also suggested that lower energy intake as a result of lower animal food intake is related to higher Cd body burden in Japanese women. A Swedish study (Berglund et al., 1994) of nonpregnant women showed that non-smoking Swedish women eating diets rich in cereals and root vegetables (i.e., a high fiber diet) or a shell fish diet had a higher Cd intake than women eating a mixed diet. However, intake of cereals and vegetables, excluding green vegetables, was not related to biomarkers of Cd exposure in the present study. Also, the carbohydrate energy ratio was positively related to urinary Cd, but its relation was not significant after adjustment for age because both of carbohydrate intake and urinary Cd were correlated with the mothers’ age.
176
M. Nishijo et al. / Toxicology Letters 148 (2004) 171–176
Fish intake in these subjects also showed no relation to blood and urinary Cd. This difference from the Swedish report may arise from the different types of diets in Sweden and Japan, and methods of the food intake survey. However, this time we did not ask about the frequency of eating shellfish, mollusks, and crustaceans, which accumulate Cd well, but are often consumed in this area. In the future, the survey of frequency of non-fish seafood with high Cd content will be necessary to clarify their contribution to Cd exposure. In Japanese cadmium-polluted areas, rice has been considered to be the main source of Cd. Recent studies targeted to middle-aged farmers’ wives showed that total Cd intake from rice contributed 40% of total Cd intake (Watanabe et al., 2000). This study analyzed correlations between daily intake of boiled rice and indices of Cd exposure, but could not detect significant correlation. Probably, this result may be caused by less rice contribution to Cd exposure in our subjects, who were younger and consumed less boiled rice than those reported by Watanabe et al. (2000). However, Shimbo et al. (2000b) reported that there was regional difference of Cd content in rice: the rice collected in our area showed the highest Cd content among samples from various regions in Japan. Future larger scale surveys targeted to young people in this area are necessary to clarify the relation between rice intake and Cd exposure.
Acknowledgements We thank Makiko Ebie, Mutumi Katsuo, Aki Ohta in Toyama Medical Pharmaceutical University Hospital who corrected data, and Katushi Yoshita who cooperated with nutritional calculation. This work was supported by grant for Project Research from High-Technology Center of Kanazawa Medical University (H2001-4).
References Åkesson, A., Berglund, M., Schutz, A., Bjellerup, P., Bremme, K., Vahter, M., 2002. Cadmium exposure in pregnancy and
lactation in relation to iron status. Am. J. Public Health 321, 177–181. Berglund, M., Akesson, A., Nermell, B., Vahter, M., 1994. Intestinal absorption of dietary cadmium in women depends on body iron stores and fiber intake. Environ. Health Perspect. 102, 1058–1066. Health promotion and nutrition devision, Health service Bureau, Ministry of Health and welfare of Japan, 1994. Recommended Dietary Allowance for Japanese, fifth revision, Daiichi Shuppan, Tokyo, pp. 26, 67–72. Honda, R., Nogawa, K., Kobayashi, E., Ishizaki, A., 1979. A simplified determination of urinary cadmium by flameless atomic absorption spectrophotometry using a heated graphite atomizer. Hokuriku J. Public Health (in Japanese) 6, 13–19. IPCS, 1992. International Programme on Chmical Safety, Environmental Health Criteria 134, Cadmium, WHO, Geneva. Jacobsson Lagerkvist, B., Soderberg, H.A., Nordberg, G.F., Ekesrydh, S., Englyst, V., 1993. Biological monitoring of arsenic, lead and cadmium in occupationally and environmentally exposed pregnant women. Scand. J. Work Environ. Health 19 (Suppl 1), 50–53. Järup, L., Berglund, M., Elinder, C.G., Vahter, M., 1998. Health effects of cadmium exposure—a review of the literature and a risk estimate. Scand. J. Work Environ. Health 24 (Suppl 1), 7–51. Kido, T., Nogawa, K., Ohmichi, M., Honda, R., Tsuritani, I., Ishizaki, M., Yamada, Y., 1992. Significance of urinary cadmium concenration in a Japanese population environmentally exposed to cadmium. Arch. Environ. Health 47, 196–202. Nogawa, K., Ishizaki, A., Kawano, S., 1978. Statistical observation of the dose–response relationship of cadmium based on epidemiological studies in the Kakehashi River basin. Environ. Res. 15, 185–198. Resources Council, Science and Technology Agency, Japan, 1985. Standard Tables for Food Composition in Japan, fourth revised ed. (1982), Daiichi Shuppan, Tokyo, pp. 18–196. Shimbo, S., Zhang, Z.W., Moon, C.S., Watanabe, T., Nakatsuka, H., Matsuda-Inoguchi, N., Hogashikawa, K., Ikeda, M., 2000a. Correlation between urine and blood concentrations and dietary intake of cadmium and lead among women in the general population of Japan. Int. Arch. Occup. Environ. Health 73, 163–170. Shimbo, S., Zhang, Z.W., Watanabe, T., Matsuda-Inoguchi, N., Nakatsuka, H., 2000b. Survey on dietary exposure of Japanese population to cadmium. Kankyo Hoken Report 67, 186–191 (in Japanese). Swazono, Y., Kobayashi, E., Okubo, Y., Nogawa, K., Kido, T., Nakagawa, H., 2000. Renal effects of cadmium exposure in cadmium non-polluted areas in Japan. Environ. Res. 84, 44– 55. Watanabe, T., Zhang, Z.W., Moon, C.S., Shimbo, S., Nakatsuka, H., Matsuda-Inoguchi, N., Hogashikawa, K., Ikeda, M., 2000. Cadmium exposure of women in general populations in Japan during 1991–1997 compared with 1997–1981. Int. Arch. Occup. Environ. Health 73, 26–34.