- Email: [email protected]
Validity of Cadmium Concentration in Rice as the “Dose” of the Dose–Response Relationship between Cadmium Intake and Renal Dysfunction
Environmental Research Section A 84, 275}281 (2000) doi:10.1006/enrs.2000.4093, available online at http://www.idealibrary.com on
Validity of Cadmium Concentration in Rice as the ‘‘Dose’’ of the DoseIResponse Relationship between Cadmium Intake and Renal Dysfunction1 Takashi Izuno,* Minoru Sugita,* Seizaburou Arita,- Yumi Otahara,* Ikuo Nasu,? Kenzaburo Tsuchiya,A and Yuzo HayashiB *Department of Environmental and Occupational Health, Toho University School of Medicine, Tokyo 143-8540, Japan; -Department of Mathematics, Kansai Medical College; ?Department of Public Dental Health, School of Dentistry at Matsudo, Nihon University; ASchool of Medicine, Keio University; and BSchool of Pharmaceutical Sciences, Kitasato University Received October 4, 1999
difAcult to clarify precisely the confounding factors from the available data, it is concluded that deriving a safety level for foodstuffs using only the Cd-R level as a reference is not appropriate. 2000
It is well known that cadmium (Cd) causes renal dysfunction such as increase of b2-microglobulin excretion into urine. Although Cd in rice seems to be one of the largest sources of total Cd intake in Japan, there are very few studies that have epidemiologically clariAed the relationship between Cd concentration in rice (Cd-R) and renal dysfunction, because such studies are basically ecological studies, in which confounding factors are difAcult to take into consideration. To derive safety levels for foodstuff from Cd-R, it is essential to evaluate the effect of confounding factors. Thus, we investigated the dose}response relationship between renal dysfunction and not only Cd-R but also confounding factors, and we tried to determine whether Cd-R is an adequate indicator of ‘‘dose’’ in the dose}response relationship between Cd intake and renal dysfunction. In 1971, Cd-R data were obtained from rice samples collected by the Environment Agency, Government of Japan in the Fuchu area of Toyama Prefecture, which is known as a place where many itai-itai disease patients were found, and medical data were collected during 1979}1984 by Toyama Prefecture. First, the dose}response relationship between Cd-R and renal dysfunction was analyzed using the data from the Fuchu area. Second, to investigate the effect of confounding factors, analysis using the data from both the Fuchu area and an unpolluted area with environmental factors different from those of the Fuchu area was performed. The results showed that the cause of renal dysfunction could not be explained by Cd-R alone, and confounding factors were not negligible. Although it is
Academic Press
Key Words: epidemiology; dose}response relationship; renal dysfunction; cadmium concentration in rice.
INTRODUCTION
Many epidemiologic studies on cadmium (Cd) in the environment (Tsuchiya, 1978) and a series of related experimental studies using animals (Nomiyama, 1981, 1986; Nomiyama and Nomiyama, 1988, 1992; Nomiyama et al., 1979b, 1981; Ellis et al., 1984) have been reported. There appears to be no room for doubt that Cd causes health disorders, especially renal dysfunction, such as increase of b2microglobulin (b2-MG) excretion into urine. Cd intake from rice is thought to be one of the most important sources of Cd intake in Japan (Tsuchiya, and Iwao, 1978), but there are very few studies that have epidemiologically clari7ed the relationship between Cd concentration in rice (Cd-R) and renal dysfunction (Nogawa et al., 1989; Kido and Nogawa, 1993; Kido et al., 1993). Moreover, consideration of confounding factors has not been suf7cient. On the other hand, year to year variation in Cd-R was reported to be very large even in the same rice 7eld (Masui, et al., 1971). In such a situation, it is not appropriate to derive safety levels of Cd in foodstuffs by referring only to the Cd-R level (World Health Organization, 1992). Therefore, further studies are thought to be necessary to evaluate the effect of
1 This study was conducted in accordance with national and institutional guidelines for the protection of human subjects.
275 0013-9351/00 $35.00 Copyright 2000 by Academic Press All rights of reproduction in any form reserved.
276
IZUNO ET AL.
confounding factors and con7rm the dose}response relationship between Cd intake based on Cd-R and renal dysfunction. The aim of this study was to evaluate the adequacy of Cd-R as an indicator of ‘‘dose’’ in the dose}response relationship between Cd intake and renal dysfunction. MATERIALS AND METHODS
Cd-R was measured by the Environment Agency, Government of Japan, in 1971 in the Fuchu area where many itai-itai disease patients had been found. Rice was sampled for Cd measurement at 310 points every 200 m in a lattice pattern. The Environment Agency made two lists of data, one list containing the code numbers of the Cd measurement points and measured Cd values and the other list containing the code numbers of the Cd measurement points and the names of the farmers who tilled the measurement point. By linking the code number of the Cd measurement point, the list, which contains Cd-Rs and the names of the farmers who tilled the points, was obtained. When the number of Cd measurement points of a farmer was more than one, the mean value of the Cd-Rs was assigned as his Cd-R. The main householder’s Cd-R was assigned as the Cd-R of his family members. Medical examination was performed by Toyama Prefecture during 1979}1984 in Fuchu and areas not polluted by Cd in Toyama Prefecture. Concerning renal function, urinary protein was measured qualitatively, and b2-MG was measured if urinary protein was positive. Renal dysfunction was de7ned as positive when b2-MG was equal to or greater than 0.2 mg/100 ml and as negative when renal protein was negative or b2-MG was less than 0.2 mg/100 ml. The dose of Cd intake was calculated as in a preceding study (Nogawa et al., 1989) as Cd intake (lg) "(Cd-R lg/g;333.5 g/day#34 lg/day) ;365.25 days/year;residence in years in the Fuchu area #50 lg/day;365.25 days/year ;(age!residence in years in the Fuchu area), where Cd-R is Cd concentration of rice, 333.5 g/day is the average rice intake per person per day, 34 lg/day is Cd intake per day from foods other than rice (Ishikawa Prefecture, 1976), and 50 lg/day is the average Cd intake per day in an unpolluted area of Japan (Yamagata, 1978). In the unpolluted area,
accumulated Cd intake was calculated using zero numbers of years of residence in a Cd-polluted area. Only when both the medical data and the Cd-R data for inhabitants were available were they used in the present study. The dose of Cd intake calculated from Cd-R was de7ned as an independent variable, and the result for renal function (renal dysfunction positive or negative) was de7ned as a dependent variable. First, the distribution of Cd intake in the Fuchu area was analyzed to determine whether the data of this study were 7t to examine the dose}response relationship between Cd intake and renal dysfunction. Second, logistic regression analysis was performed using the data from the Fuchu area to examine the dose}response relationship between Cd intake and renal dysfunction. Third, to investigate the effect of confounding factors, another logistic regression analysis was done using data from both Fuchu and unpolluted areas. In this analysis, the variable ‘‘area,’’ which is a binary variable for the Fuchu area or an unpolluted area, was introduced into this logistic model. This variable was considered to represent a confounding factor. To evaluate the dose}response relationship between Cd intake and renal dysfunction, the mean and standard deviation of the incidence of renal dysfunction directly adjusted by age and the theoretical curve were drawn using the results of the logistic regression analysis by gender. RESULTS
Table 1 shows the number, age, years of residence, measured Cd-R, and accumulated Cd intake calculated from the Cd-R of the subjects in the Fuchu area by gender. In the Fuchu area, the mean and standard deviation of Cd-R and accumulated Cd intake were 0.76$0.38 lg/g and 5.15$3.09 g (0.90}17.24 g), respectively, in males and 0.75$ 0.35 lg/g and 4.30$2.53 g (1.03}13.55 g), respectively, in females. The range of Cd-R in Fuchu was 0.21}2.16 lg/g. Table 2 shows the results of renal function in medical examinations in the Fuchu area by gender. The b2-MG of subjects whose renal protein was negative was not measured, because b2-MG has a strong relationship with renal protein (Friberg et al., 1974). Using 70.2 mg/100 ml b2-MG as an indication of renal damage, 23.2% of the male subjects and 30.3% of the female subjects were positive for nenal dysfunction. Table 3 shows the results of renal function in medical examinations in the unpolluted area by gender. These data were obtained from a government report by Toyama Prefecture. The
277
VALIDITY OF Cd IN RICE IN DOSE}RESPONSE RELATIONSHIP
TABLE 1 The Ages, Duration of Residence, Cd-R, and Accumulated Cd Intake of the Subjects in the Fuchu Area (meanⴞSD)
Male Female
Number of subjects
Age (years)
Residence duration (years)
Cd-R (lg/g)
Accumulated Cd intake (g)
82 56
66.5$12.9 65.1$11.3
59.7$18.7 45.5$12.0
0.76$0.38 0.75$0.35
5.15$3.09 4.30$2.53
data of the government report consisted of the numbers of subjects for the results of renal protein and b2-MG by age group and gender. Because only the range of age for each age group was shown in the government data by Toyama Prefecture, the median age of each age group was assigned as the age for all members of the age group in the unpolluted area. The means and SD of the assigned age for males and females in the unpolluted area were 63.1$8.6 and 64.3$9.5 years, respectively, and the means and SD of the accumulated Cd intake were 1.15$0.16 and 1.17$0.17 g for males and females, respectively. Only 0.3% of the male subjects and 0.4% of the female subjects in the unpolluted area were positive for renal dysfunction. Table 4 shows the results of logistic regression analysis for the dose}response relationship between accumulated Cd intake based on Cd-R and renal dysfunction in the Fuchu area. In the logistic regression analysis, the dependent variable is renal dysfunction and the independent variables are age (years) and accumulated Cd intake (lg). Because b2-MG is strongly related to age, age was entered into this logistic regression model as a confounding factor for accumulated Cd intake. The results show that the odds ratio of accumulated Cd intake in males is greater than 1.0 and statistically signi7cant, whereas that in females is less than 1.0 and not signi7cant. The odds ratios of age were signi7cant in both males and females. The odds ratio of accumulated Cd intake shows the effect of the independent variable by an increase of 1 lg, and the odds ratio of age shows the effect by an increase of one year. Figure 1 shows the dose-response relationship between the incidence of renal dysfunction and the
accumulated Cd intake directly adjusted by age. In each gender group, the subjects were divided into three groups (low, medium, high) by dose of accumulated Cd intake. The position of each group on the X axis of Fig. 1 is the mean value of accumulated Cd intake for the group. The numbers of subjects in the low, medium, and high groups were 27, 27, and 28 in males and 18, 19, and 19 in females, respectively. The means of accumulated Cd intake for the three groups were 2.44, 4.61, and 8.58 g in males and 1.99, 3.72, and 7.07 g in females, respectively. In this 7gure, mean$SD shows the incidence of renal dysfunction and the curve shows the theoretical curve calculated from the result of the logistic regression analysis. Table 5 shows the results of logistic regression analysis using the combined data from Fuchu and the unpolluted area. The dependent variables of this analysis were the same as those for the previous logistic regression analysis. These results show that the odds ratio of accumulated Cd intake is signi7cant in both males and females, contrary to the results in Table 4, whereas the odds ratio of age is signi7cant in both males and females in accordance with the results of Table 4. Table 6 and Fig. 2 show the results of logistic regression analysis using data from both the Fuchu area and the unpolluted area, adding the variable ‘‘area,’’ where ‘‘area’’ is a binary variable for Fuchu or the unpolluted area. In Fig. 2, to understand at a glance, the unit of age was set at 10 years, which means effect by increase of 10 years. The odds ratios of the variable ‘‘area’’ were much larger than 1.0. The odds ratios of age were larger than 1.0 in both males and females. The odds ratios of accumulated Cd were greater than 1.0 in males, whereas they were less than 1.0 in females. This
TABLE 2 Results of Renal Function of Medical Examinations in the Fuchu Area b2-MG
Male Female
Number of subjects
Subjects with no urinary protein
:0.2 mg/100 ml
70.2 mg/100 ml
82 56
44(53.7%) 29(51.8%)
19(23.2%) 10(17.9%)
19(23.2%) 17(30.3%)
278
IZUNO ET AL.
TABLE 3 Results of Renal Function in Medical Examinations in the Unpolluted Area b2-MG
Male Female
Number of subjects
Subjects with no urinary protein
:0.2 mg/100 ml
70.2 mg/100 ml
1089 1290
910(83.6%) 1153(89.4%)
175(16.1%) 131(10.2%)
4(0.3%) 6(0.4%)
implies that the variable ‘‘area’’ is an important factor in the analyses. DISCUSSION
The hypothesis that Cd intake causes health disorders is assumed to be strongly supported by many studies. The relation between Cd intake and renal dysfunction such as increase of b2-microglobulin excretion into urine is especially well known. Cd in rice seems to be one of the most important source of Cd intake in Japan, but there have been very few studies that have epidemiologically clari7ed the relationship between Cd-R and renal dysfunction. Moreover, confounding factors have not been taken into due consideration. Thus, deriving safety levels of Cd in foodstuffs employing the Cd-R level alone is ill-advised. In this study, we attempted to clarify the validity of Cd-R as the dose of the dose}response relationship between Cd intake and renal dysfunction, taking into consideration confounding factors. Many studies have revealed that b2-MG excretion into urine increases with age (Kido and Nogawa, 1993; Kido et al., 1993; Kitamura and Koizumi, 1975; Nomiyama et al., 1978; Tsuchiya, et al., 1979). Age was entered into all logistic models in this study as a confounding factor for accumulated Cd intake. The odds ratio of the relationship between age and renal dysfunction judged by the b2-MG level was also larger than 1.0 and statistically signi7cant. After adjusting for the effect of age, in the Fuchu area the odds ratio between renal dysfunction and accumulated Cd intake based on Cd-R was found to be different by gender. The odds ratio of males was
larger than 1.0, whereas that of females was smaller than 1.0. According to Fig. 1 for the dose}response relationship between incidence of renal dysfunction and accumulated Cd intake calculated from Cd-R drawn with the direct age adjustment method, the incidence in males rises with increase of accumulated Cd intake, whereas that in females does not increase. The theoretical curve drawn using the results of logistic regression analysis shows this relation more clearly. Although cases of itai-itai disease were found mainly in females in the Fuchu area, the odds ratio of females was found to be smaller than 1.0 in this study. The lower limit of Cd-R in the Fuchu area was almost the same as that in the unpolluted areas (Tsuchiya, 1978; Masui et al., 1971), whereas the upper limit of Cd-R in the Fuchu area was quite higher than that in the unpolluted areas. As the CdR values in the Fuchu area varied from low to very high, Fuchu was thought to be suitable for a study of the dose}response relationship for some health disorders caused by Cd intake calculated from Cd-R. Therefore, if yearly Cd-R represents total Cd exposure over a long period of years, signi7cant results of the relationship between accumulated Cd intake and renal dysfunction should have been detected in females of the Fuchu area. In this study, however, such a relationship was not detected among females in the Fuchu area. Thus, Cd-R should not be used alone as an indicator of the lifetime Cd intake of the inhabitants. In the 7rst logistic regression analysis in this study, confounding factors for accumulated Cd intake were not taken into consideration. Such factors
TABLE 4 The Results of Logistic Regression Analysis in the Fuchu Area Male
Accumulated Cd intake Age
Female
Odds ratio
95% C.I.
P-value
Odds ratio
95% C.I.
P value
1.205 1.088
1.011}1.437 1.035}1.143
0.0375 0.0009
0.877 1.070
0.668}1.151 1.011}1.133
0.3424 0.0196
279
VALIDITY OF Cd IN RICE IN DOSE}RESPONSE RELATIONSHIP
FIG. 1. Dose}response relationship between incidence of renal dysfunction and accumulative Cd intake by gender directly adjusted by age. Mean$SD: the incidence data Curve: the theoretical curve calculated from the results of the logistic regression analysis.
were virtually considered to be almost the same for subjects who lived in the same area. The second logistic regression analysis, using combined data from Fuchu and the unpolluted areas, showed that the odds ratio of accumulated Cd intake was greater than 1.0. To consider this inconsistency, the variable ‘‘area,’’ representing a confounding factor, was introduced in the third logistic regression analysis. The result showed that the odds ratios of the ‘‘area’’ were very large. In females the odds ratio of accumulated Cd intake was less than 1.0 and not statistically signi7cant. Therefore, a dose}response relationship between Cd intake calculated from Cd-R and renal dysfunction could not be derived from the results of the analysis of Fuchu and the unpolluted areas when a confounding factor was taken into consideration. According to the above logistic regression analyses, the difference between the incidences of renal dysfunction in the two areas could be explained by both the Cd-R and the variable ‘‘area.’’ The proportion of the effect due to Cd-R and that due to ‘‘area’’ is, however, unknown from the results of the present
study. It may therefore be impossible to calculate a safety level from Cd-R. The incidence of renal dysfunction in the Fuchu area was higher than that in the unpolluted area. Although in the Fuchu area the level of Cd-R differs from the level in the control (unpolluted) area to a large extent, the relationship between Cd intake and renal dysfunction is not clear in the Fuchu area. If there is a clear dose}response relationship and the range of the dose extends from the control level at which there are few responses to the high level at which there are many responses, the dose}response relationship is easily detected epidemiologically. However, the difference in the incidence of renal dysfunction could not be explained by Cd-R alone in this study. Nevertheless, the difference in the incidence of renal dysfunction between the Fuchu area and the control area is thought to be mostly explainable by the total pollution due to Cd. In the logistic analysis adding the variable ‘‘area,’’ the odds ratio of the variable ‘‘area’’ was found to be very large. That shows that a confounding factor might have caused
TABLE 5 The Results of Logistic Analysis Using the Data of Both the Fuchu Area and the Unpolluted Area Male
Accumulated Cd intake Age
Female
Odds ratio
95% C.I.
P-value
Odds ratio
95% C.I.
P value
1.671 1.111
1.442}1.938 1.058}1.166
:0.0001 :0.0001
1.801 1.061
1.465}2.214 1.017}1.108
:0.0001 0.0065
280
IZUNO ET AL.
TABLE 6 The Results of Logistic Regression Analysis Using the Data of Both the Fuchu Area and the Unpolluted Area, Adding the Variable ‘‘Area’’ Male Odds ratio Accumulated Cd intake Age Area
1.204 1.086 26.949
95% C.I. 1.011}1.433 1.039}1.135 6.433}112.894
renal dysfunction in the Fuchu area. One of these factors may be the drinking water, but it is very dif7cult to pinpoint the true factor because the available data are very limited. On the other hand, the year to year variation in Cd-R was revealed to be very large (Masui et al., 1971) and Cd-R could not represent the total Cd intake throughout life, so the dose}response relationship between the accumulated Cd intake calculated from Cd-R and the renal dysfunction have not been detected in this study. Earlier studies were basically ecological studies, and confounding factors are dif7cult to take into consideration in such studies (Rothman, 1986). In this study, the only genuine ecological observation is the simple comparison of the positive rates of b2-MG in two areas, as shown in Table 5. Table 6 shows the results of analysis adding the confounding factor of ‘‘area.’’ The results show
Female P-value 0.0374 0.0003 :0.0001
Odds ratio 0.868 1.085 159.048
95% C.I. 0.657}1.146 1.034}1.138 42.254}598.664
P value 0.3169 0.0009 :0.0001
that this confounding factor is not negligible. Therefore, it is not possible to determine a safety level from this study without taking into consideration confounding factors. We conclude that the hypothesis that Cd intake based on Cd-R is a reasonable ‘‘dose’’ for determining the dose}response relationship between Cd intake and renal dysfunction is not valid. CONCLUSIONS
We analyzed the dose}response relationship between Cd intake calculated from Cd-R and renal dysfunction using Cd-R data measured by the Environment Agency and medical examination data collected by Toyama Prefecture. The results showed that Cd-R is not an adequate indicator for calculating the Cd safety level of environmental pollution.
FIG. 2. The odds ratios and their 95% con7dence intervals of the analysis using the data of both the Fuchu area and the unpolluted area, adding the variable ‘‘area.’’
VALIDITY OF Cd IN RICE IN DOSE}RESPONSE RELATIONSHIP
Investigating the dose}response relationship between Cd intake based on Cd-R and renal dysfunction in Cd-polluted and -unpolluted areas is not easy from a practical point of view because it is very dif7cult to collect valid exposure data throughout 50 years of life. REFERENCES Ellis, K. J., Yuen, J., Yasumura, S., and Cohn, S. H. (1984). Dose}response analysis of cadmium in man: Body burden vs kidney dysfunction. Environ. Res. 33, 216}226. Friberg, L., Piscator, M., Nordberg, G. F., and Kjollstom, T. (1974). ‘‘Cadmium in the Environment,’’ pp. 101}130. CRC Press, Cleveland. Kido, T., and Nogawa, K. (1993). Dose-response relationship between total cadmium intake and b2-microglobulinuria using logistic regression analysis. Toxicol. Lett. 69, 113}120. Kido, T., Shaikh, Z. A., Kito, H., Honda, R., and Nogawa, K. (1993). Dose-response relationship between total cadmium intake and metallotioneinuria using logistic regression analysis. Toxicology 80, 207}215. Kitamura, M., and Koizumi, N. (1975). Results of urine tests for bedridden old persons in Hyogo Prefecture (Report 2). Kankyo Hoken Report (Japanese Public Health Association) 36, 135}137. [In Japanese] Masui, M., Kanamaru, N., Takesako, H., Miyakoda, H., Nanga, I., and Takahashi, H. (1971). Annual survey on correlation between the degree of cadmium contamination of paddy 7eld rice grain and the number of dry-paddy 7eld days in the cadmium contaminated area in the Tama region of Tokyo. Bull. Tokyo Agric. Exp. Station 5, 1}5. [In Japanese] Nogawa, K., Honda, R., Kido, T., Tsuritani, I., Yamada, Y., Ishizaki, M., and Yamaya, H. (1989). A dose}response analysis of cadmium in the general environment with special reference to total cadmium intake limit. Environ. Res. 48, 7}16. Nomiyama, K. (1981). Renal effects of cadmium. In ‘‘Cadmium in the Environment’’ (J. O. Nriagu, Ed.), pp. 643}689. Wiley, New York. Nomiyama, K. (1986). The chronic toxicity of cadmium: In8uence of environmental and other variables. In ‘‘Handbook of Experimental Pharmacology,’’ (E. C. Foulkes, Ed.), Vol. 80, pp. 101}133. Springer-Verlag, Berlin.
281
Nomiyama, K., and Nomiyama, H. (1988). Health effects of six years of dietary cadmium (cadmium-contaminated rice) in monkeys. In. ‘‘Essential and Toxic Trace Elements in Human Health and Disease’’ (A. S. Prasad, Ed.), pp. 589}609. A. R. Liss, New York. Nomiyama, K., and Nomiyama H. (1992). Invalid assessment of allowable daily intake of cadmium by critical concentration or critical total dose of cadmium, in population exposed to cadmium of low dose. In ‘‘Proceedings of the International Symposium on Bio-Clinical Signi7cance of Urinary b2-Microglobulin’’ (K. Nomiyama, Ed.). Kankyo Hoken Report 59, 112}116. [In Japanese] Nomiyama, K., Nomiyama H, and Masui, K. (1979a). Chemical form and intestinal absorption of cadmium in rice. Jpn. J. Hyg. 34, 103. Nomiyama. K., Nomiyama, H., Akabori, F., and Masaoka, T. (1981). Further studies on dietary cadmium on rhesus monkeys (10) Tissue cadmium, zinc, and copper. In ‘‘Recent Studies on Health Effects of Cadmium in Japan,’’ pp. 154}188. Japan Environmental Agency, Tokyo. Nomiyama, K., Nomiyama, H., Yotoriyama, M., and Taguchi, T. (1978). Some recent studies on the renal effects of cadmium. In ‘‘Cadmium 77,’’ pp. 186}194. Metal Bull Ltd., San Francisco. Nomiyama, K., Nomiyama, H., Nomura, Y., Taguchi, T., Masui, K., Yotoriyama, M., Akabori, F., Iwao, S., Koizumi, N., Masaoka, T., Kitamura, S., Tsuchiya, K., Suzuki, T., and Kabayashi, K. (1979b). Effects of dietary cadmium on rhesus monkeys. Environ. Health Perspect. 28, 223}243. Rothman K. J. (1986). ‘‘Modern Epidemiology.’’ Little, Brown and Co., Boston. Tsuchiya, K. (1978). ‘‘Cadmium Studies in Japan: A Review.’’ Elsevier/North-Holland, Amsterdam. Tsuchiya, K., and Iwao, S. (1978). Results and evaluation on cadmium intake of Cd-exposed inhabitants in Akita, Ishikawa, and Nagasaki Prefectures. Kankyo Hoken Report 44, 86}115. [In Japanese] Tsuchiya, K., Iwao, S., Sugita, M., and Sakurai, H. (1979). Increased urinary b2-Microglobulin in cadmium exposure: Dose-effect relationship and biological signi7cance of b2Microglobulin. Environ. Health Perspect. 28, 147}153. Watanabe, H., and Murayama, H. (Eds.) (1974). ‘‘Proceedings of the International Symposium on Recent Advances in the Assessment of the Health Effects of Environmental Pollution,’’ Vol. 1. World Health Organization. (1992). ‘‘Environmental Health Criteria 134: Cadmium.’’ World Health Organization, Geneva.
Validity of Cadmium Concentration in Rice as the ‘‘Dose’’ of the DoseIResponse Relationship between Cadmium Intake and Renal Dysfunction1 Takashi Izuno,* Minoru Sugita,* Seizaburou Arita,- Yumi Otahara,* Ikuo Nasu,? Kenzaburo Tsuchiya,A and Yuzo HayashiB *Department of Environmental and Occupational Health, Toho University School of Medicine, Tokyo 143-8540, Japan; -Department of Mathematics, Kansai Medical College; ?Department of Public Dental Health, School of Dentistry at Matsudo, Nihon University; ASchool of Medicine, Keio University; and BSchool of Pharmaceutical Sciences, Kitasato University Received October 4, 1999
difAcult to clarify precisely the confounding factors from the available data, it is concluded that deriving a safety level for foodstuffs using only the Cd-R level as a reference is not appropriate. 2000
It is well known that cadmium (Cd) causes renal dysfunction such as increase of b2-microglobulin excretion into urine. Although Cd in rice seems to be one of the largest sources of total Cd intake in Japan, there are very few studies that have epidemiologically clariAed the relationship between Cd concentration in rice (Cd-R) and renal dysfunction, because such studies are basically ecological studies, in which confounding factors are difAcult to take into consideration. To derive safety levels for foodstuff from Cd-R, it is essential to evaluate the effect of confounding factors. Thus, we investigated the dose}response relationship between renal dysfunction and not only Cd-R but also confounding factors, and we tried to determine whether Cd-R is an adequate indicator of ‘‘dose’’ in the dose}response relationship between Cd intake and renal dysfunction. In 1971, Cd-R data were obtained from rice samples collected by the Environment Agency, Government of Japan in the Fuchu area of Toyama Prefecture, which is known as a place where many itai-itai disease patients were found, and medical data were collected during 1979}1984 by Toyama Prefecture. First, the dose}response relationship between Cd-R and renal dysfunction was analyzed using the data from the Fuchu area. Second, to investigate the effect of confounding factors, analysis using the data from both the Fuchu area and an unpolluted area with environmental factors different from those of the Fuchu area was performed. The results showed that the cause of renal dysfunction could not be explained by Cd-R alone, and confounding factors were not negligible. Although it is
Academic Press
Key Words: epidemiology; dose}response relationship; renal dysfunction; cadmium concentration in rice.
INTRODUCTION
Many epidemiologic studies on cadmium (Cd) in the environment (Tsuchiya, 1978) and a series of related experimental studies using animals (Nomiyama, 1981, 1986; Nomiyama and Nomiyama, 1988, 1992; Nomiyama et al., 1979b, 1981; Ellis et al., 1984) have been reported. There appears to be no room for doubt that Cd causes health disorders, especially renal dysfunction, such as increase of b2microglobulin (b2-MG) excretion into urine. Cd intake from rice is thought to be one of the most important sources of Cd intake in Japan (Tsuchiya, and Iwao, 1978), but there are very few studies that have epidemiologically clari7ed the relationship between Cd concentration in rice (Cd-R) and renal dysfunction (Nogawa et al., 1989; Kido and Nogawa, 1993; Kido et al., 1993). Moreover, consideration of confounding factors has not been suf7cient. On the other hand, year to year variation in Cd-R was reported to be very large even in the same rice 7eld (Masui, et al., 1971). In such a situation, it is not appropriate to derive safety levels of Cd in foodstuffs by referring only to the Cd-R level (World Health Organization, 1992). Therefore, further studies are thought to be necessary to evaluate the effect of
1 This study was conducted in accordance with national and institutional guidelines for the protection of human subjects.
275 0013-9351/00 $35.00 Copyright 2000 by Academic Press All rights of reproduction in any form reserved.
276
IZUNO ET AL.
confounding factors and con7rm the dose}response relationship between Cd intake based on Cd-R and renal dysfunction. The aim of this study was to evaluate the adequacy of Cd-R as an indicator of ‘‘dose’’ in the dose}response relationship between Cd intake and renal dysfunction. MATERIALS AND METHODS
Cd-R was measured by the Environment Agency, Government of Japan, in 1971 in the Fuchu area where many itai-itai disease patients had been found. Rice was sampled for Cd measurement at 310 points every 200 m in a lattice pattern. The Environment Agency made two lists of data, one list containing the code numbers of the Cd measurement points and measured Cd values and the other list containing the code numbers of the Cd measurement points and the names of the farmers who tilled the measurement point. By linking the code number of the Cd measurement point, the list, which contains Cd-Rs and the names of the farmers who tilled the points, was obtained. When the number of Cd measurement points of a farmer was more than one, the mean value of the Cd-Rs was assigned as his Cd-R. The main householder’s Cd-R was assigned as the Cd-R of his family members. Medical examination was performed by Toyama Prefecture during 1979}1984 in Fuchu and areas not polluted by Cd in Toyama Prefecture. Concerning renal function, urinary protein was measured qualitatively, and b2-MG was measured if urinary protein was positive. Renal dysfunction was de7ned as positive when b2-MG was equal to or greater than 0.2 mg/100 ml and as negative when renal protein was negative or b2-MG was less than 0.2 mg/100 ml. The dose of Cd intake was calculated as in a preceding study (Nogawa et al., 1989) as Cd intake (lg) "(Cd-R lg/g;333.5 g/day#34 lg/day) ;365.25 days/year;residence in years in the Fuchu area #50 lg/day;365.25 days/year ;(age!residence in years in the Fuchu area), where Cd-R is Cd concentration of rice, 333.5 g/day is the average rice intake per person per day, 34 lg/day is Cd intake per day from foods other than rice (Ishikawa Prefecture, 1976), and 50 lg/day is the average Cd intake per day in an unpolluted area of Japan (Yamagata, 1978). In the unpolluted area,
accumulated Cd intake was calculated using zero numbers of years of residence in a Cd-polluted area. Only when both the medical data and the Cd-R data for inhabitants were available were they used in the present study. The dose of Cd intake calculated from Cd-R was de7ned as an independent variable, and the result for renal function (renal dysfunction positive or negative) was de7ned as a dependent variable. First, the distribution of Cd intake in the Fuchu area was analyzed to determine whether the data of this study were 7t to examine the dose}response relationship between Cd intake and renal dysfunction. Second, logistic regression analysis was performed using the data from the Fuchu area to examine the dose}response relationship between Cd intake and renal dysfunction. Third, to investigate the effect of confounding factors, another logistic regression analysis was done using data from both Fuchu and unpolluted areas. In this analysis, the variable ‘‘area,’’ which is a binary variable for the Fuchu area or an unpolluted area, was introduced into this logistic model. This variable was considered to represent a confounding factor. To evaluate the dose}response relationship between Cd intake and renal dysfunction, the mean and standard deviation of the incidence of renal dysfunction directly adjusted by age and the theoretical curve were drawn using the results of the logistic regression analysis by gender. RESULTS
Table 1 shows the number, age, years of residence, measured Cd-R, and accumulated Cd intake calculated from the Cd-R of the subjects in the Fuchu area by gender. In the Fuchu area, the mean and standard deviation of Cd-R and accumulated Cd intake were 0.76$0.38 lg/g and 5.15$3.09 g (0.90}17.24 g), respectively, in males and 0.75$ 0.35 lg/g and 4.30$2.53 g (1.03}13.55 g), respectively, in females. The range of Cd-R in Fuchu was 0.21}2.16 lg/g. Table 2 shows the results of renal function in medical examinations in the Fuchu area by gender. The b2-MG of subjects whose renal protein was negative was not measured, because b2-MG has a strong relationship with renal protein (Friberg et al., 1974). Using 70.2 mg/100 ml b2-MG as an indication of renal damage, 23.2% of the male subjects and 30.3% of the female subjects were positive for nenal dysfunction. Table 3 shows the results of renal function in medical examinations in the unpolluted area by gender. These data were obtained from a government report by Toyama Prefecture. The
277
VALIDITY OF Cd IN RICE IN DOSE}RESPONSE RELATIONSHIP
TABLE 1 The Ages, Duration of Residence, Cd-R, and Accumulated Cd Intake of the Subjects in the Fuchu Area (meanⴞSD)
Male Female
Number of subjects
Age (years)
Residence duration (years)
Cd-R (lg/g)
Accumulated Cd intake (g)
82 56
66.5$12.9 65.1$11.3
59.7$18.7 45.5$12.0
0.76$0.38 0.75$0.35
5.15$3.09 4.30$2.53
data of the government report consisted of the numbers of subjects for the results of renal protein and b2-MG by age group and gender. Because only the range of age for each age group was shown in the government data by Toyama Prefecture, the median age of each age group was assigned as the age for all members of the age group in the unpolluted area. The means and SD of the assigned age for males and females in the unpolluted area were 63.1$8.6 and 64.3$9.5 years, respectively, and the means and SD of the accumulated Cd intake were 1.15$0.16 and 1.17$0.17 g for males and females, respectively. Only 0.3% of the male subjects and 0.4% of the female subjects in the unpolluted area were positive for renal dysfunction. Table 4 shows the results of logistic regression analysis for the dose}response relationship between accumulated Cd intake based on Cd-R and renal dysfunction in the Fuchu area. In the logistic regression analysis, the dependent variable is renal dysfunction and the independent variables are age (years) and accumulated Cd intake (lg). Because b2-MG is strongly related to age, age was entered into this logistic regression model as a confounding factor for accumulated Cd intake. The results show that the odds ratio of accumulated Cd intake in males is greater than 1.0 and statistically signi7cant, whereas that in females is less than 1.0 and not signi7cant. The odds ratios of age were signi7cant in both males and females. The odds ratio of accumulated Cd intake shows the effect of the independent variable by an increase of 1 lg, and the odds ratio of age shows the effect by an increase of one year. Figure 1 shows the dose-response relationship between the incidence of renal dysfunction and the
accumulated Cd intake directly adjusted by age. In each gender group, the subjects were divided into three groups (low, medium, high) by dose of accumulated Cd intake. The position of each group on the X axis of Fig. 1 is the mean value of accumulated Cd intake for the group. The numbers of subjects in the low, medium, and high groups were 27, 27, and 28 in males and 18, 19, and 19 in females, respectively. The means of accumulated Cd intake for the three groups were 2.44, 4.61, and 8.58 g in males and 1.99, 3.72, and 7.07 g in females, respectively. In this 7gure, mean$SD shows the incidence of renal dysfunction and the curve shows the theoretical curve calculated from the result of the logistic regression analysis. Table 5 shows the results of logistic regression analysis using the combined data from Fuchu and the unpolluted area. The dependent variables of this analysis were the same as those for the previous logistic regression analysis. These results show that the odds ratio of accumulated Cd intake is signi7cant in both males and females, contrary to the results in Table 4, whereas the odds ratio of age is signi7cant in both males and females in accordance with the results of Table 4. Table 6 and Fig. 2 show the results of logistic regression analysis using data from both the Fuchu area and the unpolluted area, adding the variable ‘‘area,’’ where ‘‘area’’ is a binary variable for Fuchu or the unpolluted area. In Fig. 2, to understand at a glance, the unit of age was set at 10 years, which means effect by increase of 10 years. The odds ratios of the variable ‘‘area’’ were much larger than 1.0. The odds ratios of age were larger than 1.0 in both males and females. The odds ratios of accumulated Cd were greater than 1.0 in males, whereas they were less than 1.0 in females. This
TABLE 2 Results of Renal Function of Medical Examinations in the Fuchu Area b2-MG
Male Female
Number of subjects
Subjects with no urinary protein
:0.2 mg/100 ml
70.2 mg/100 ml
82 56
44(53.7%) 29(51.8%)
19(23.2%) 10(17.9%)
19(23.2%) 17(30.3%)
278
IZUNO ET AL.
TABLE 3 Results of Renal Function in Medical Examinations in the Unpolluted Area b2-MG
Male Female
Number of subjects
Subjects with no urinary protein
:0.2 mg/100 ml
70.2 mg/100 ml
1089 1290
910(83.6%) 1153(89.4%)
175(16.1%) 131(10.2%)
4(0.3%) 6(0.4%)
implies that the variable ‘‘area’’ is an important factor in the analyses. DISCUSSION
The hypothesis that Cd intake causes health disorders is assumed to be strongly supported by many studies. The relation between Cd intake and renal dysfunction such as increase of b2-microglobulin excretion into urine is especially well known. Cd in rice seems to be one of the most important source of Cd intake in Japan, but there have been very few studies that have epidemiologically clari7ed the relationship between Cd-R and renal dysfunction. Moreover, confounding factors have not been taken into due consideration. Thus, deriving safety levels of Cd in foodstuffs employing the Cd-R level alone is ill-advised. In this study, we attempted to clarify the validity of Cd-R as the dose of the dose}response relationship between Cd intake and renal dysfunction, taking into consideration confounding factors. Many studies have revealed that b2-MG excretion into urine increases with age (Kido and Nogawa, 1993; Kido et al., 1993; Kitamura and Koizumi, 1975; Nomiyama et al., 1978; Tsuchiya, et al., 1979). Age was entered into all logistic models in this study as a confounding factor for accumulated Cd intake. The odds ratio of the relationship between age and renal dysfunction judged by the b2-MG level was also larger than 1.0 and statistically signi7cant. After adjusting for the effect of age, in the Fuchu area the odds ratio between renal dysfunction and accumulated Cd intake based on Cd-R was found to be different by gender. The odds ratio of males was
larger than 1.0, whereas that of females was smaller than 1.0. According to Fig. 1 for the dose}response relationship between incidence of renal dysfunction and accumulated Cd intake calculated from Cd-R drawn with the direct age adjustment method, the incidence in males rises with increase of accumulated Cd intake, whereas that in females does not increase. The theoretical curve drawn using the results of logistic regression analysis shows this relation more clearly. Although cases of itai-itai disease were found mainly in females in the Fuchu area, the odds ratio of females was found to be smaller than 1.0 in this study. The lower limit of Cd-R in the Fuchu area was almost the same as that in the unpolluted areas (Tsuchiya, 1978; Masui et al., 1971), whereas the upper limit of Cd-R in the Fuchu area was quite higher than that in the unpolluted areas. As the CdR values in the Fuchu area varied from low to very high, Fuchu was thought to be suitable for a study of the dose}response relationship for some health disorders caused by Cd intake calculated from Cd-R. Therefore, if yearly Cd-R represents total Cd exposure over a long period of years, signi7cant results of the relationship between accumulated Cd intake and renal dysfunction should have been detected in females of the Fuchu area. In this study, however, such a relationship was not detected among females in the Fuchu area. Thus, Cd-R should not be used alone as an indicator of the lifetime Cd intake of the inhabitants. In the 7rst logistic regression analysis in this study, confounding factors for accumulated Cd intake were not taken into consideration. Such factors
TABLE 4 The Results of Logistic Regression Analysis in the Fuchu Area Male
Accumulated Cd intake Age
Female
Odds ratio
95% C.I.
P-value
Odds ratio
95% C.I.
P value
1.205 1.088
1.011}1.437 1.035}1.143
0.0375 0.0009
0.877 1.070
0.668}1.151 1.011}1.133
0.3424 0.0196
279
VALIDITY OF Cd IN RICE IN DOSE}RESPONSE RELATIONSHIP
FIG. 1. Dose}response relationship between incidence of renal dysfunction and accumulative Cd intake by gender directly adjusted by age. Mean$SD: the incidence data Curve: the theoretical curve calculated from the results of the logistic regression analysis.
were virtually considered to be almost the same for subjects who lived in the same area. The second logistic regression analysis, using combined data from Fuchu and the unpolluted areas, showed that the odds ratio of accumulated Cd intake was greater than 1.0. To consider this inconsistency, the variable ‘‘area,’’ representing a confounding factor, was introduced in the third logistic regression analysis. The result showed that the odds ratios of the ‘‘area’’ were very large. In females the odds ratio of accumulated Cd intake was less than 1.0 and not statistically signi7cant. Therefore, a dose}response relationship between Cd intake calculated from Cd-R and renal dysfunction could not be derived from the results of the analysis of Fuchu and the unpolluted areas when a confounding factor was taken into consideration. According to the above logistic regression analyses, the difference between the incidences of renal dysfunction in the two areas could be explained by both the Cd-R and the variable ‘‘area.’’ The proportion of the effect due to Cd-R and that due to ‘‘area’’ is, however, unknown from the results of the present
study. It may therefore be impossible to calculate a safety level from Cd-R. The incidence of renal dysfunction in the Fuchu area was higher than that in the unpolluted area. Although in the Fuchu area the level of Cd-R differs from the level in the control (unpolluted) area to a large extent, the relationship between Cd intake and renal dysfunction is not clear in the Fuchu area. If there is a clear dose}response relationship and the range of the dose extends from the control level at which there are few responses to the high level at which there are many responses, the dose}response relationship is easily detected epidemiologically. However, the difference in the incidence of renal dysfunction could not be explained by Cd-R alone in this study. Nevertheless, the difference in the incidence of renal dysfunction between the Fuchu area and the control area is thought to be mostly explainable by the total pollution due to Cd. In the logistic analysis adding the variable ‘‘area,’’ the odds ratio of the variable ‘‘area’’ was found to be very large. That shows that a confounding factor might have caused
TABLE 5 The Results of Logistic Analysis Using the Data of Both the Fuchu Area and the Unpolluted Area Male
Accumulated Cd intake Age
Female
Odds ratio
95% C.I.
P-value
Odds ratio
95% C.I.
P value
1.671 1.111
1.442}1.938 1.058}1.166
:0.0001 :0.0001
1.801 1.061
1.465}2.214 1.017}1.108
:0.0001 0.0065
280
IZUNO ET AL.
TABLE 6 The Results of Logistic Regression Analysis Using the Data of Both the Fuchu Area and the Unpolluted Area, Adding the Variable ‘‘Area’’ Male Odds ratio Accumulated Cd intake Age Area
1.204 1.086 26.949
95% C.I. 1.011}1.433 1.039}1.135 6.433}112.894
renal dysfunction in the Fuchu area. One of these factors may be the drinking water, but it is very dif7cult to pinpoint the true factor because the available data are very limited. On the other hand, the year to year variation in Cd-R was revealed to be very large (Masui et al., 1971) and Cd-R could not represent the total Cd intake throughout life, so the dose}response relationship between the accumulated Cd intake calculated from Cd-R and the renal dysfunction have not been detected in this study. Earlier studies were basically ecological studies, and confounding factors are dif7cult to take into consideration in such studies (Rothman, 1986). In this study, the only genuine ecological observation is the simple comparison of the positive rates of b2-MG in two areas, as shown in Table 5. Table 6 shows the results of analysis adding the confounding factor of ‘‘area.’’ The results show
Female P-value 0.0374 0.0003 :0.0001
Odds ratio 0.868 1.085 159.048
95% C.I. 0.657}1.146 1.034}1.138 42.254}598.664
P value 0.3169 0.0009 :0.0001
that this confounding factor is not negligible. Therefore, it is not possible to determine a safety level from this study without taking into consideration confounding factors. We conclude that the hypothesis that Cd intake based on Cd-R is a reasonable ‘‘dose’’ for determining the dose}response relationship between Cd intake and renal dysfunction is not valid. CONCLUSIONS
We analyzed the dose}response relationship between Cd intake calculated from Cd-R and renal dysfunction using Cd-R data measured by the Environment Agency and medical examination data collected by Toyama Prefecture. The results showed that Cd-R is not an adequate indicator for calculating the Cd safety level of environmental pollution.
FIG. 2. The odds ratios and their 95% con7dence intervals of the analysis using the data of both the Fuchu area and the unpolluted area, adding the variable ‘‘area.’’
VALIDITY OF Cd IN RICE IN DOSE}RESPONSE RELATIONSHIP
Investigating the dose}response relationship between Cd intake based on Cd-R and renal dysfunction in Cd-polluted and -unpolluted areas is not easy from a practical point of view because it is very dif7cult to collect valid exposure data throughout 50 years of life. REFERENCES Ellis, K. J., Yuen, J., Yasumura, S., and Cohn, S. H. (1984). Dose}response analysis of cadmium in man: Body burden vs kidney dysfunction. Environ. Res. 33, 216}226. Friberg, L., Piscator, M., Nordberg, G. F., and Kjollstom, T. (1974). ‘‘Cadmium in the Environment,’’ pp. 101}130. CRC Press, Cleveland. Kido, T., and Nogawa, K. (1993). Dose-response relationship between total cadmium intake and b2-microglobulinuria using logistic regression analysis. Toxicol. Lett. 69, 113}120. Kido, T., Shaikh, Z. A., Kito, H., Honda, R., and Nogawa, K. (1993). Dose-response relationship between total cadmium intake and metallotioneinuria using logistic regression analysis. Toxicology 80, 207}215. Kitamura, M., and Koizumi, N. (1975). Results of urine tests for bedridden old persons in Hyogo Prefecture (Report 2). Kankyo Hoken Report (Japanese Public Health Association) 36, 135}137. [In Japanese] Masui, M., Kanamaru, N., Takesako, H., Miyakoda, H., Nanga, I., and Takahashi, H. (1971). Annual survey on correlation between the degree of cadmium contamination of paddy 7eld rice grain and the number of dry-paddy 7eld days in the cadmium contaminated area in the Tama region of Tokyo. Bull. Tokyo Agric. Exp. Station 5, 1}5. [In Japanese] Nogawa, K., Honda, R., Kido, T., Tsuritani, I., Yamada, Y., Ishizaki, M., and Yamaya, H. (1989). A dose}response analysis of cadmium in the general environment with special reference to total cadmium intake limit. Environ. Res. 48, 7}16. Nomiyama, K. (1981). Renal effects of cadmium. In ‘‘Cadmium in the Environment’’ (J. O. Nriagu, Ed.), pp. 643}689. Wiley, New York. Nomiyama, K. (1986). The chronic toxicity of cadmium: In8uence of environmental and other variables. In ‘‘Handbook of Experimental Pharmacology,’’ (E. C. Foulkes, Ed.), Vol. 80, pp. 101}133. Springer-Verlag, Berlin.
281
Nomiyama, K., and Nomiyama, H. (1988). Health effects of six years of dietary cadmium (cadmium-contaminated rice) in monkeys. In. ‘‘Essential and Toxic Trace Elements in Human Health and Disease’’ (A. S. Prasad, Ed.), pp. 589}609. A. R. Liss, New York. Nomiyama, K., and Nomiyama H. (1992). Invalid assessment of allowable daily intake of cadmium by critical concentration or critical total dose of cadmium, in population exposed to cadmium of low dose. In ‘‘Proceedings of the International Symposium on Bio-Clinical Signi7cance of Urinary b2-Microglobulin’’ (K. Nomiyama, Ed.). Kankyo Hoken Report 59, 112}116. [In Japanese] Nomiyama, K., Nomiyama H, and Masui, K. (1979a). Chemical form and intestinal absorption of cadmium in rice. Jpn. J. Hyg. 34, 103. Nomiyama. K., Nomiyama, H., Akabori, F., and Masaoka, T. (1981). Further studies on dietary cadmium on rhesus monkeys (10) Tissue cadmium, zinc, and copper. In ‘‘Recent Studies on Health Effects of Cadmium in Japan,’’ pp. 154}188. Japan Environmental Agency, Tokyo. Nomiyama, K., Nomiyama, H., Yotoriyama, M., and Taguchi, T. (1978). Some recent studies on the renal effects of cadmium. In ‘‘Cadmium 77,’’ pp. 186}194. Metal Bull Ltd., San Francisco. Nomiyama, K., Nomiyama, H., Nomura, Y., Taguchi, T., Masui, K., Yotoriyama, M., Akabori, F., Iwao, S., Koizumi, N., Masaoka, T., Kitamura, S., Tsuchiya, K., Suzuki, T., and Kabayashi, K. (1979b). Effects of dietary cadmium on rhesus monkeys. Environ. Health Perspect. 28, 223}243. Rothman K. J. (1986). ‘‘Modern Epidemiology.’’ Little, Brown and Co., Boston. Tsuchiya, K. (1978). ‘‘Cadmium Studies in Japan: A Review.’’ Elsevier/North-Holland, Amsterdam. Tsuchiya, K., and Iwao, S. (1978). Results and evaluation on cadmium intake of Cd-exposed inhabitants in Akita, Ishikawa, and Nagasaki Prefectures. Kankyo Hoken Report 44, 86}115. [In Japanese] Tsuchiya, K., Iwao, S., Sugita, M., and Sakurai, H. (1979). Increased urinary b2-Microglobulin in cadmium exposure: Dose-effect relationship and biological signi7cance of b2Microglobulin. Environ. Health Perspect. 28, 147}153. Watanabe, H., and Murayama, H. (Eds.) (1974). ‘‘Proceedings of the International Symposium on Recent Advances in the Assessment of the Health Effects of Environmental Pollution,’’ Vol. 1. World Health Organization. (1992). ‘‘Environmental Health Criteria 134: Cadmium.’’ World Health Organization, Geneva.