- Email: [email protected]
Estimation of benchmark doses as threshold levels of urinary cadmium, based on excretion of β2-microglobulin in cadmium-polluted and non-polluted regions in Japan
Toxicology Letters 179 (2008) 108–112
Contents lists available at ScienceDirect
Toxicology Letters journal homepage: www.elsevier.com/locate/toxlet
Estimation of benchmark doses as threshold levels of urinary cadmium, based on excretion of 2 -microglobulin in cadmium-polluted and non-polluted regions in Japan Etsuko Kobayashi a,∗ , Yasushi Suwazono a , Mirei Dochi a , Ryumon Honda b , Muneko Nishijo c , Teruhiko Kido d , Hideaki Nakagawa c a Department of Occupational and Environmental Medicine (A2), Graduate School of Medicine, Chiba University, 1-8-1 Inohana, Chuohku, Chiba 260-8670 Japan b Department of Nursing, Kanazawa Medical University, Japan c Department of Public Health, Kanazawa Medical University, Kanazawa, Japan d Department of Community Health Nursing, Kanazawa University School of Health Sciences, Japan
a r t i c l e
i n f o
Article history: Received 17 March 2008 Received in revised form 24 April 2008 Accepted 24 April 2008 Available online 3 May 2008 Keywords: Cadmium Environment Threshold level Renal effect 2 -microglobuline-uria
a b s t r a c t The threshold level of urinary cadmium (Cd) for 2 -microglobulin (MG)-uria was calculated using a benchmark dose (BMD) approach. Total number of subjects was 3103 for Cd-polluted areas and 2929 for non-polluted areas. Multiple logistic regression analysis was employed to fit the dose–response model taking into consideration an age effect. Cut-off values for urinary 2 -MG were defined as those corresponding to the 84th and 97.5th percentile of 2 -MG levels in the controls, and 1000 g/g creatinine (cr). The BMD low (BMDL) was calculated using the profile likelihood method. When the benchmark response was 5%, the BMD/BMDL of Cd for the 84th percentile of 2 -MG for mean age, 55, 65, and 75 years was 3.0/2.7, 4.6/4.2, 2.8/2.6, and 1.8/1.6 g/g cr in men and 3.4/3.2, 5.8/5.5, 3.2/3.1, and 1.8/1.7 g/g cr in women, respectively. The value for the 97.5th percentile for each age was 4.9/4.5, 7.6/7.0, 4.6/4.3, and 2.6/2.4 g/g cr in men and 5.9/5.6, 9.7/9.2, 5.6/5.3, and 2.8/2.6 g/g cr in women. Namely it became clear that the margin between the threshold level and average excretion level of urinary Cd was small in the older population in Japan. To prevent the adverse health effects caused by exposure to Cd, it is important to establish the threshold level of Cd exposure at each age. © 2008 Elsevier Ireland Ltd. All rights reserved.
1. Introduction It is well known that long-term exposure to cadmium (Cd) in the general environment causes renal dysfunction (Nogawa, 1981; Nomiyama, 1986). To protect people from the adverse effects of Cd exposure, it is important to determine threshold exposure levels for Cd using different indices of Cd exposure and health effects. Physiologic indicators of exposure may include Cd concentrations in tissues, blood or urine. In western countries, urinary Cd concentration is often used as an index of Cd body burden, and a dose–response relationship between renal dysfunction and urinary Cd excretion level has been established (Jakubowski et al., ¨ 1987, 1992; Buchet et al., 1990; Bernard et al., 1992, 1995; Jarup et al., 2000). Similarly, we previously established a dose–response relationship between urinary Cd concentrations and urinary 2 microglobulin (2 -MG) as an indicator of renal dysfunction and
∗ Corresponding author. Tel.: +81 43 226 2065; fax: +81 43 226 2066. E-mail address: [email protected] (E. Kobayashi). 0378-4274/$ – see front matter © 2008 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.toxlet.2008.04.013
the threshold level of urinary Cd in Cd-exposed subjects of the Kakehashi River basin (Ishizaki et al., 1989; Hayano et al., 1996). Moreover, we investigated the association between Cd concentration in either blood or urine and indicators of renal dysfunction (urinary total protein, 2 -MG, and N-acetyl--d-glucosaminidase (NAG)) for 2778 inhabitants ≥50 years of age (1114 men and 1664 women) in three different Cd non-polluted areas in Japan. As a result, we showed the presence of renal effects induced by Cd exposure not only in Cd-polluted areas but also in the general environment without any known Cd pollution (Yamanaka et al., 1998; Oo et al., 2000; Suwazono et al., 2000). In health risk assessment of environmental contaminants, a benchmark dose (BMD) approach as defined by Crump (1984, 1995) has been used to estimate the threshold levels of toxic substances (Gaylor et al., 1998). The BMD can be defined as the exposure that corresponds to a certain increase in the probability of an adverse response compared to the background at zero exposure. The benchmark dose low (BMDL) is defined as the value corresponding to the lower 95% confidence interval of the BMD and can be used in evaluating dose–response relationships as a replacement for the
E. Kobayashi et al. / Toxicology Letters 179 (2008) 108–112
no observed adverse effect level (NOAEL) or the lowest observable adverse effect level (LOAEL) (Gaylor et al., 1998). We estimated the BMDL from the threshold levels of urinary Cd using prior data from the Kakehashi River basin (Shimizu et al., 2006) or three different Cd non-polluted areas in Japan (Kobayashi et al., 2006). The values were estimated at 2.9 g/g creatinine (cr) for men and 1.5 g/g cr for women in the Kakehashi River basin and 2.4 g/g cr for men and 3.3 g/g cr for women in the Cd non-polluted areas. By using a multiple logistic model (Budtz-Jørgensen et al., 2001) it is possible to take into account the effects of other possible covariates. We were unable to evaluate the effect of age in the estimation of the threshold level of urinary Cd in the previous Shimizu study (2006) and Kobayashi study (2006). Therefore in the present study we calculated the threshold level of urinary Cd for 2 -MG-uria in Cd-polluted and non-polluted areas (the Kakehashi River basin along with the three different Cd non-polluted areas) by using a multiple logistic model.
Table 1 Means of age, urinary Cd and 2 -microglobulin concentrations according to sex Men (N = 2578)
Age (years)a Cd (g/g cr) 2 -MG (g/g cr)
2.1. Study population The study population for the Cd-polluted area was drawn from the population evaluated in the 1981 and 1982 health survey conducted among the entire population (>50 years of age) residing in the Kakehashi River basin (Shimizu et al., 2006). Among these, subjects who were known from the questionnaire survey to ingest household rice were selected as the target population of the present study. The groups consisted of 3103 participants inhabiting Cd-polluted hamlets (1397 men and 1706 women) and 289 participants (130 men and 159 women) inhabiting non-polluted hamlets. Moreover, 2640 inhabitants (1051 men and 1589 women, ≥50 years of age) of the three different Cd non-polluted areas described in a previous study (Kobayashi et al., 2006) were included in the study population. Thus, the number of subjects for the non-polluted areas totaled 2929 (1181 men and 1748 women). 2.2. Collection of samples and analytical methods Morning urine specimens were collected from all participants and were kept frozen at −20 ◦ C until analysis. The specimens for 2 -MG determination were adjusted to a pH above 6.0, while the specimens for Cd determination were acidified with nitric acid, and then kept frozen. However, when the pH of urine samples obtained from non-polluted areas was <5.4, the 2 -MG concentrations were not measured. Urinary Cd was measured by flameless atomic absorption spectrometry using a Hitachi Model Z180-80 after digestion with nitric acid (Kido et al., 1984). Urinary 2 -MG was determined by radioimmunoassay method (Pharmacia 2 -micro RIA, Pharmacia Diagnostics AB, Sweden) and urinary creatinine concentration was determined by a modified method of the Jaffe reaction (Bonsnes and Taussky, 1945). 2.3. Statistical analysis Multiple logistic regression analysis was used to fit the dose–response model to the data. The dependent variable was a positive finding of 2 -MG-uria. The independent variables were urinary Cd concentration and age, as continuous variables. The concentrations of urinary Cd and 2 -MG were expressed in corrected creatinine units (g cr−1 ). Cut-off values for urinary 2 -MG were defined as those corresponding to the 84th percentile and 97.5th percentile of the 2 -MG level in subjects from nonpolluted areas. Those values were 492 g/g cr and 965 g/g cr in men and 407 g/g cr and 798 g/g cr in women, respectively. Additionally, 1000 g/g cr of urinary 2 -MG was also used as a cut-off value for 2 -MG-uria.
Women (N = 3454)
GM
Min.
Max.
GM
Min.
Max.
64.0 3.0 157
50 0.01 1
91 49.6 107922
64.1 4.2 195
50 0.02 2
95 57.6 186668
N: number of subjects examined; GM: geometric mean; Min.: minimum value; Max.: maximum value; a Arithmetic mean of age was calculated.
The BMD of urinary Cd excretion was calculated using parameters obtained by multiple logistic regression analysis. The benchmark response (BMR) corresponding to the BMD is defined as an additional pre-specified increase in the probability of a positive finding. Assuming that age is fixed to certain values, the log odds of probability of a positive finding, P(di ) at dose di of urinary Cd concentration, is given by ln
2. Materials and methods
109
P(0) P(di ) + ˇ × di = ln 1 − P(0) 1 − P(di )
where P(0) = estimated probability of a “hypothetical” control distribution at urinary Cd = 0 and a certain age, and ˇ = slope for dose-log odds relationship. P(BMD) is given by P(BMD) = P(0) + BMR The log odds of P(BMD) is given by ln
P(d0 ) P(BMD) = ln + ˇ × BMD 1 − P(BMD) 1 − P(d0 )
(1)
BMD is obtained by combining the equation for P(BMD) with that for the dose–response model (1): BMD =
1 [P(d0 ) + BMR] × [1 − P(d0 )] ln ˇ [1 − P(d0 ) − BMR] × [P(d0 )]
In this study, BMDL was calculated using the profile likelihood method (Crump, 1984; Filipsson et al., 2003). BMD/BMDL with BMR at 5% or 10% was calculated for each cut-off value of urinary 2 -MG concentration. The analyses were performed with SPSS, version 12.0J (multiple logistic regression, SPSS Japan Inc., Tokyo, Japan), and Microsoft Excel 2003 (BMDL calculation, Microsoft Corporation, Redmond, WA, USA). P values < 0.05 were considered to be statistically significant.
3. Results Mean values of age, urinary Cd and 2 -MG concentrations of the study population according to sex are shown in Table 1. Prevalences of 2 -MG-uria using three different cut-off values for urinary 2 -MG according to sex and age class (50–59, 60–69, ≥70 years and total) are shown in Table 2. The prevalence calculated for each cut-off value increased with increasing age class in both men and women. Table 3 lists results of multiple logistic regressions for 2 -MG-uria. There were significant positive associations between 2 -MG-uria and urinary Cd concentrations in both men and women. Age also associated positively with 2 -MG-uria in both sexes.
Table 2 Prevalence (%) of 2 -microglobulin-uria using three different cut-off values by sex and age class Age class
Men N
50–59 years 60–69 years 70 years or more Total
902 963 713 2578
Women Cut-off value
N
84th percentile
97.5th percentile
1000 g/g cr
8.6 16.9 31.3 18.0
3.9 9.2 20.5 10.5
3.9 8.8 20.1 10.2
1174 1312 968 3454
Cut-off value 84th percentile
97.5th percentile
1000 g/g cr
10.7 19.4 37.6 28.9
4.6 10.7 24.7 16.8
3.2 9.7 22.3 14.8
N: number of subjects examined. The 84th percentile and 97.5th percentile of the 2 -MG level were calculated in subjects from non-polluted areas (492 g/g cr and 965 g/g cr in men and 407 g/g cr and 798 g/g cr in women).
110
E. Kobayashi et al. / Toxicology Letters 179 (2008) 108–112
Table 3 Results of multiple logistic regression analyses for 2 -microglobulin-uria Cut-off value
Men, OR (95%CI)
Women, OR (95%CI)
84th percentile Cd (g/g cr) Age (years)
1.23 (1.19–1.26) 1.08 (1.07–1.09)
1.20 (1.18–1.23) 1.10 (1.09–1.11)
97.5th percentile Cd (g/g cr) Age (years)
1.24 (1.20–1.29) 1.10 (1.08–1.11)
1.25 (1.22–1.28) 1.12 (1.11–1.14)
1000 g/g cr Cd (g/g cr) Age (years)
1.25 (1.21–1.29) 1.10 (1.08–1.11)
1.26 (1.23–1.30) 1.13 (1.12–1.15)
OR: odds ratio = 95% CI: 95% confidence intervals.
The BMD/BMDL for each of the cut-off values was calculated using a multiple logistic model, setting BMR at 5% or 10% (Table 4). When BMR was 5%, the BMD/BMDL of urinary Cd for the 84th percentile of urinary 2 -MG for mean age, 55, 65, and 75 years was 3.0/2.7, 4.6/4.2, 2.8/2.6, and 1.8/1.6 g/g cr in men and 3.4/3.2, 5.8/5.5, 3.2/3.1, and 1.8/1.7 g/g cr in women, respectively. When an abnormal value of 5% was employed, the BMD/BMDL of urinary Cd for the 97.5th percentile for mean age, 55 years, 65 years, and 75 years was 4.9/4.5, 7.6/7.0, 4.6/4.3, and 2.6/2.4 g/g cr in men and 5.9/5.6, 9.7/9.2, 5.6/5.3, and 2.8/2.6 g/g cr in women, respectively. When an abnormal value of 5% was employed, the BMD/BMDL of urinary Cd for 1000 g/g cr of urinary 2 -MG for mean age, 55, 65, and 75 years was 5.0/4.6, 7.7/7.1, 4.7/4.3, and 2.7/2.4 g/g cr in men and 6.7/6.3, 10.9/10.4, 6.3/5.9, and 3.1/2.6 g/g cr in women, respectively. 4. Discussion In the present study, we calculated the threshold level of urinary Cd for 2 -MG-uria in Cd-polluted and non-polluted areas in Japan, including the Kakehashi River basin and three different Cd nonpolluted areas, using a multiple logistic model. To our knowledge, this is the first estimation of BMD/BMDL from a multiple logistic model for the threshold level of urinary Cd taking into account an age effect. As shown in Table 2, it is clear that the prevalence of 2 MG-uria using any cut-off value is increased with increasing age in both sexes. In our previous studies (Shimizu et al., 2006; Kobayashi et al., 2006) we were not able to consider the effects of age on the estimation of BMD, due to limitations of the software used (Bench-
mark Dose Software; US EPA). Therefore the threshold values of urinary Cd were estimated without taking the confounding effect of age on 2 -MG-uria into consideration in those previous studies. On the other hand, we could estimate the threshold level of urinary Cd at each age in this study by using a multiple logistic regression analysis including age into the model. The obtained BMDL gradually decreased with increasing age, for example the value for 84th percentile for 55, 65, and 75 years was 4.2, 2.6, and 1.6 g/g cr in men and 5.5, 3.1, and 1.7 g/g cr in women (BMR was 5%) as shown Table 4. As shown in Tables 3 and 4, age was positively related to the 2 -MG-uria, yielding increased probability of 2 -MG-uria at 0 exposure of cadmium (P(0)). Therefore, significant increase in probability of 2 -MG-uria compared to 0 exposure occurs at lower urinary Cd level in older population than in younger population. Thus, the present study reveals how much the threshold amount of urinary Cd is affected by individual age. From the epidemiological standpoint these results indicated that the margin between the threshold level and average excretion level of urinary Cd was small in the older population in Japan. To prevent the adverse health effects caused by exposure to Cd, it is thus important to establish the threshold level of Cd exposure at each age. Moreover, the age range of the target population should be as wide as possible for more accurate adjustment for the confounding effect of age. In terms of statistical power, the investigated cohort (N = 6032) was much larger than in Shimizu’s (2006, N = 3392) and Kobayashi’s (2006, N = 2778) studies. Furthermore, the exposure range was broader than in the previous studies because we combined data from non-polluted and Cd-polluted areas. For example, only 4.0% of subjects had urinary cadmium ≥7.0 g/g cr in the previous study conducted in non-polluted areas (Kobayashi et al., 2006), whereas that value was observed more than 30% of the time in the other study conducted in Cd-polluted areas (Shimizu et al., 2006). The estimated threshold values of urinary Cd for 2 -MG-uria (BMR = 5%, with a cut-off value at the 84th percentile of urinary 2 -MG) were estimated at 2.9 g/g cr in men and 1.5 g/g cr in women of the Kakehashi River basin (Shimizu et al., 2006), and 2.4 g/g cr in men and 3.3 g/g cr in women in Cd non-polluted areas (Kobayashi et al., 2006). Since there was little difference between the prevalence of 2 -MG-uria in the Cd-polluted areas (21.0%) and in the non-polluted areas (14.2%) in men, the threshold values (for the mean age) were similar. On the there hand, since the prevalence of 2 -MG-uria was 30.4% and 12.9% in each study, respectively. The threshold values of urinary Cd were different. In this study the prevalence of 2 -MG-uria was 18.0% in men and 28.9% in
Table 4 BMD/BMDL of urinary cadmium for 2 -microglobulin-uria according to sex Cut-off value
84th percentile
97.5th percentile
1000 g/g cr
Age
Mean age 55 years 65 years 75 years Mean age 55 years 65 years 75 years Mean age 55 years 65 years 75 years
Men
Women
BMD/BMDL of U-Cd (g/g cr)
BMD/BMDL of U-Cd (g/g cr)
BMR = 5%
BMR = 10%
P(0) (%)
BMR = 5%
BMR = 10%
3.0/2.7 4.6/4.2 2.8/2.6 1.8/1.6 4.9/4.5 7.6/7.0 4.6/4.3 2.6/2.4 5.0/4.6 7.7/7.1 4.7/4.3 2.7/2.4
5.0/4.6 7.1/6.6 4.8/4.4 3.2/2.9 7.4/6.8 10.5/9.6 7.1/6.5 4.4/4.1 7.5/6.9 10.6/9.7 7.2/6.6 4.5/4.1
6.8 3.5 7.3 14.4 2.8 1.3 3.1 7.4 2.7 1.2 3.0 7.1
3.4/3.2 5.8/5.5 3.2/3.1 1.8/1.7 5.9/5.6 9.7/9.2 5.6/5.3 2.8/2.6 6.7/6.3 10.9/10.4 6.3/5.9 3.1/2.6
5.7/5.4 8.8/8.3 5.4/5.1 3.3/3.1 8.6/8.1 12.7/12.1 8.2/7.8 4.6/4.4 9.4/8.9 13.9/13.4 9.0/8.5 5.0/4.5
P(0) (%) 6.4 2.8 7.0 16.3 2.0 0.7 2.2 6.7 1.4 0.5 1.6 5.3
BMR: benchmark response; P (0): probability of positive finding at zero exposure = BMD/BMDL are calculated assuming mean age (men: 64.0; women: 64.1), and 55, 65, and 75 years of age.
E. Kobayashi et al. / Toxicology Letters 179 (2008) 108–112
women, and the threshold values of urinary Cd for the mean age was 2.7 g/g cr in men and 3.2 g/g cr in women, respectively. Thus, we feel that the present study improved the accuracy of BMD/BMDL determination by adopting an improved methodology and a greater quality and quantity of subjects. Urinary Cd threshold values have been estimated using different methodologies from that in the present study. Hayano et al. (1996) applied a logistic regression model to data from Cd-polluted areas and employed a cut-off value for 2 -MG of 1000 g/g cr. The estimated Cd threshold values ranged from 1.6 to 4.6 g/g cr in different sex and age groups. In the Hayano study, the abnormality rate of urinary 2 -MG in the controls was substituted into each regression formula calculated with the urinary Cd concentration and the abnormality rate of 2 -MG in Cd-exposed subjects, and threshold levels of urinary Cd were obtained. However, those control groups had in fact been exposed to Cd because Cd was detected in their urine. Thus the control groups did not strictly correspond to a population without any exposure to Cd. Unlike in experimental animals, a perfect control with no Cd exposure is physically impossible in humans. In contrast, the BMD method does not require abnormality rates of urinary findings in the control subjects. The probability of adverse response at zero exposure was estimated based on the whole dose–response curve. Then, the exposure level was calculated where a certain change in response compared to the background was expected. In this sense, the BMD/BMDL calculations obtained in the present study are more accurate and provide more information than the threshold levels obtained in the Hayano study. In this study we used the 84th percentile and 97.5th percentile of the non-exposed subjects and 1000 g/g cr of 2 -MG as cut-off values. Kido et al. (1988) investigated the reversibility of 2 -MG-uria in 74 inhabitants (32 men and 42 women) of Cd-polluted regions of the Kakehashi River basin. The subjects participated in two examinations, one conducted just after the cessation of Cd exposure and the other 5 years later (when Cd exposure had ended). In cases who had ≥1000 g/g cr of 2 -MG in the first investigation, almost all individuals showed deterioration of 2 -MG-uria. Accordingly the threshold level for irreversible renal damage was determined to be 1000 g/g cr of urinary 2 -MG excretion. Other researchers in Japan have also obtained similar results (Iwata et al., 1993; Cai et al., 2001). Nakagawa et al. (1993) conducted a 9-year follow-up study of 3178 subjects (aged ≥50 years) in the Cd-polluted Kakehashi River basin. When compared with a group with <300 g/g cr of urinary 2 -MG at entry, mortality ratios were found to increase as urinary 2 MG excretion increased in both sexes. Thus, the greater the tubular dysfunction, the less favorable the life prognosis, and even slightly increased urinary excretion of 2 -MG (300 to <1000 g/g cr) was associated with increased mortality. In the present study, the 84th percentile for 2 -MG was 492 g/g cr for men and 407 g/g cr for women, higher than the 300 g/g cr that was reported to be the borderline value indicative of a poorer life prognosis. Taken together, 1000 g/g cr for 2 -MG is thought to be an extremely robust cut-off which reflects both irreversible renal dysfunction and poor life prognosis. The lowest cut-off level in the present study, the 84th percentile, is also thought to be valid because of the negative impact on life prognosis according to the Nakagawa study (1993). Recently, Nakagawa et al. (2006) reported the results of a 15-year follow-up survey of the same population as mentioned above (Nakagawa et al., 1993) to investigate excess mortality and the dose–response relationship between mortality and urinary Cd excretion. The mortality risk was increased among the subjects, especially in women, with urinary Cd >3 g/g cr after adjustment for age. In our study, the threshold level of urinary Cd for the 84th percentile of 2 -MG among the controls was estimated at
111
2.7 g/g cr in men and 3.2 g/g cr in women for the mean age (about 64 years in both sexes). Thus the present study and their study together suggest that subjects with a urinary Cd of greater than about 3 g/g cr will have renal dysfunction and an increased risk of mortality. In conclusion, the threshold level of urinary Cd was estimated in persons living in the general environment with and without Cd pollution in Japan. The values were calculated more exactly than previously using a multiple logistic model taking into consideration the effects of increasing age.
References Bernard, A., Roels, H., Thielemans, N., Lierde, V.M., Lauwerys, R., 1992. Assessment of the causality of cadmium-protein relationships in the urine of the general population with reference to the Cadmibel study. IARC Sci. Pub. 118, 341– 346. Bernard, A., Thielemans, N., Roels, H., Lauwerys, R., 1995. Association between NAGB and cadmium in urine with no evidence of a threshold. Occup. Environ. Med. 52, 177–180. Bonsnes, R.W., Taussky, H.H., 1945. On the calorimetric determination of creatinine by the Jaffe reaction. J. Biol. Chem. 15, 581–591. Buchet, J.P., Lauwerys, R.R., Roles, H., Bernard, A., Claeys, F., Ducoffer, G., Plaen, P.D., Staessen, J., Amery, A., Lijnen, P., Thijs, L., Rondia, D., Sartor, F., Remy, A.S., Nick, L., 1990. Renal effects of cadmium body burden of the general population. Lancet 336, 699–702. Budtz-Jørgensen, E., Keiding, N., Grandjean, P., 2001. Benchmark dose calculation from epidemiological data. Biometrics 57, 698–706. Cai, Y., Aoshima, K., Katoh, T., Teranishi, H., Kasuya, M., 2001. Renal tubular dysfunction in male inhabitants of a cadmium-polluted area in Toyama, Japan—an eleven-year follow-up study. J. Epidemiol. 11, 180–189. Crump, K.S., 1984. A new method for determining allowable daily intakes. Fundam. Appl. Toxicol. 4, 854–871. Crump, K.S., 1995. Calculation of benchmark dose from continuous data. Risk Anal 15, 79–89. Filipsson, A.F., Sand, S., Nilsson, J., Victorin, K., 2003. The benchmark dose methods—Review of available models, and recommendations for application in health risk assessment. Crit. Rev. Tocicol. 33, 505–542. Gaylor, D., Ryan, L., Krewski, D., Zhu, Y., 1998. Procedures for calculation benchmark doses for health risk assessment. Reg. Toxicol. Pharmacol. 28, 150–164. Hayano, M., Nogawa, K., Kido, T., Kobayashi, E., Honda, R., Tsuritani, I., 1996. Dose–response relationship between urinary cadmium concentration and 2 microglobulinuria using logistic regression analysis. Arch. Environ. Health 51, 162–167. Ishizaki, M., Kido, T., Honda, R., Tsuritani, I., Yamada, Y., Nakagawa, H., Nogawa, K., 1989. Dose–response relationship between urinary cadmium and 2 microglobulin in a Japanese environmentally cadmium exposed population. Toxicology 58, 121–131. Iwata, K., Saito, H., Moriyama, M., Nakano, A., 1993. Renal tubular function after reduction of environmental cadmium exposure: a ten-year follow-up. Arch. Environ. Health 48, 157–163. Jakubowski, M., Trojanowska, B., Kowalska, G., Gendek, E., Stajewska, B., Krajewska, B., Jajte, J., 1987. Occupational exposure to cadmium and kidney dysfunction. Int. Occup. Environ. Health 59, 567–577. Jakubowski, M., Razniewska, G., Halatek, T., Trzcinka, O.M., 1992. Integrated of occupational exposure to cadmium as a predictor of kidney dysfunction. IARC Sci. Pub. 118, 319–324. ¨ Jarup, L., Hellstrom, L., Alfven, T., Carlsson, M.D., Grubb, A., Persson, B., Pettersson, C., Spang, G., Schut, A., Elinder, C.G., 2000. Low level exposure to cadmium and early kidney damage: the OSCAR study. Occup. Environ. Med. 57, 668– 672. Kido, T., Tsuritani, I., Honda, R., Ishizaki, M., Yamada, Y., Nogawa, K., 1984. A direct determination of urinary cadmium by graphite-furnace atomic absorption spectrometry using the zeeman effect. J. Kanazawa Med. Univ. 9, 70–75 (in Japanese). Kido, T., Honda, R., Tsuritani, I., Yamaya, H., Ishizaki, M., Nogawa, K., 1988. Progress of renal dysfunction in inhabitants environmentally exposed to cadmium. Arch. Environ. Health 43, 213–221. Kobayashi, E., Suwazono, Y., Uetani, M., Inaba, T., Oishi, M., Kido, T., Nishijo, M., Nakagawa, H., Nogawa, K., 2006. Estimation of benchmark dose as the threshold levels of urinary cadmium, based on excretion of total protein, beta2-microglobulin, and N-acetyl-beta-d-glucosaminidase in cadmium non-polluted regions in Japan. Environ. Res. 101, 401–406. Nakagawa, H., Nishijo, M., Morikawa, Y., Senma, M., Kawano, S., Ishizaki, M., Sugita, N., Nishi, M., Kido, T., Nogawa, K., 1993. Urinary 2 -microglobulin concentration and mortality in a cadmium-polluted area. Arch. Environ. Health 48, 428– 435. Nakagawa, H., Nishijo, M., Morikawa, Y., Miura, K., Tawara, K., Kuriwaki, J., Kido, T., Ikawa, A., Kobayashi, E., Nogawa, K., 2006. Urinary cadmium and mortality among inhabitants of a cadmium-polluted area in Japan. Environ. Res. 16, 273–279.
112
E. Kobayashi et al. / Toxicology Letters 179 (2008) 108–112
Nogawa, K., 1981. Itai-Itai disease and follow-up studies. In Nriagu, J.O. (Ed.), Cadmium in the Environment. Part 2. Health Effects. Wiley Interscience Publication, New York, Chichester, Brisbane, Toronto, Singapore, pp. 1–37. Nomiyama, K., 1986. The chronic toxicity of cadmium: influence of environment and other variables. In: Foulkes, E.C. (Ed.), Handbook of Environmental Pharmacology. Springer-Verlag, Berlin/Heidelberg/New York/Tokyo, pp. 101– 133. Oo, Y.K., Kobayashi, E., Nogawa, K., Okubo, Y., Suwazono, Y., Kido, T., Nakagawa, H., 2000. Renal effects of cadmium intake of a Japanese general population in two areas unpolluted by cadmium. Arch. Environ. Health 55, 98–103.
Shimizu, A., Kobayashi, E., Suwazono, Y., Uetani, M., Oishi, M., Inaba, T., Kido, T., Nogawa, K., 2006. Estimation of benchmark doses for urinary cadmium based on 2 -microglobulinuria excretion in cadmium-polluted regions of the Kakehashi River basin, Japan. Int. J. Environ. Health Res. 16, 329–337. Suwazono, Y., Kobayashi, E., Nogawa, K., Okubo, Y., Kido, T., Nakagawa, H., 2000. Renal effects of cadmium exposure in cadmium nonpolluted areas in Japan. Environ. Res. 84, 44–55. Yamanaka, O., Kobayashi, E., Nogawa, K., Suwazono, Y., Sakurada, I., Kido, T., 1998. Association between renal effects and cadmium exposure in cadmiumnonpolluted area in Japan. Environ. Res. 77, 1–8.
Contents lists available at ScienceDirect
Toxicology Letters journal homepage: www.elsevier.com/locate/toxlet
Estimation of benchmark doses as threshold levels of urinary cadmium, based on excretion of 2 -microglobulin in cadmium-polluted and non-polluted regions in Japan Etsuko Kobayashi a,∗ , Yasushi Suwazono a , Mirei Dochi a , Ryumon Honda b , Muneko Nishijo c , Teruhiko Kido d , Hideaki Nakagawa c a Department of Occupational and Environmental Medicine (A2), Graduate School of Medicine, Chiba University, 1-8-1 Inohana, Chuohku, Chiba 260-8670 Japan b Department of Nursing, Kanazawa Medical University, Japan c Department of Public Health, Kanazawa Medical University, Kanazawa, Japan d Department of Community Health Nursing, Kanazawa University School of Health Sciences, Japan
a r t i c l e
i n f o
Article history: Received 17 March 2008 Received in revised form 24 April 2008 Accepted 24 April 2008 Available online 3 May 2008 Keywords: Cadmium Environment Threshold level Renal effect 2 -microglobuline-uria
a b s t r a c t The threshold level of urinary cadmium (Cd) for 2 -microglobulin (MG)-uria was calculated using a benchmark dose (BMD) approach. Total number of subjects was 3103 for Cd-polluted areas and 2929 for non-polluted areas. Multiple logistic regression analysis was employed to fit the dose–response model taking into consideration an age effect. Cut-off values for urinary 2 -MG were defined as those corresponding to the 84th and 97.5th percentile of 2 -MG levels in the controls, and 1000 g/g creatinine (cr). The BMD low (BMDL) was calculated using the profile likelihood method. When the benchmark response was 5%, the BMD/BMDL of Cd for the 84th percentile of 2 -MG for mean age, 55, 65, and 75 years was 3.0/2.7, 4.6/4.2, 2.8/2.6, and 1.8/1.6 g/g cr in men and 3.4/3.2, 5.8/5.5, 3.2/3.1, and 1.8/1.7 g/g cr in women, respectively. The value for the 97.5th percentile for each age was 4.9/4.5, 7.6/7.0, 4.6/4.3, and 2.6/2.4 g/g cr in men and 5.9/5.6, 9.7/9.2, 5.6/5.3, and 2.8/2.6 g/g cr in women. Namely it became clear that the margin between the threshold level and average excretion level of urinary Cd was small in the older population in Japan. To prevent the adverse health effects caused by exposure to Cd, it is important to establish the threshold level of Cd exposure at each age. © 2008 Elsevier Ireland Ltd. All rights reserved.
1. Introduction It is well known that long-term exposure to cadmium (Cd) in the general environment causes renal dysfunction (Nogawa, 1981; Nomiyama, 1986). To protect people from the adverse effects of Cd exposure, it is important to determine threshold exposure levels for Cd using different indices of Cd exposure and health effects. Physiologic indicators of exposure may include Cd concentrations in tissues, blood or urine. In western countries, urinary Cd concentration is often used as an index of Cd body burden, and a dose–response relationship between renal dysfunction and urinary Cd excretion level has been established (Jakubowski et al., ¨ 1987, 1992; Buchet et al., 1990; Bernard et al., 1992, 1995; Jarup et al., 2000). Similarly, we previously established a dose–response relationship between urinary Cd concentrations and urinary 2 microglobulin (2 -MG) as an indicator of renal dysfunction and
∗ Corresponding author. Tel.: +81 43 226 2065; fax: +81 43 226 2066. E-mail address: [email protected] (E. Kobayashi). 0378-4274/$ – see front matter © 2008 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.toxlet.2008.04.013
the threshold level of urinary Cd in Cd-exposed subjects of the Kakehashi River basin (Ishizaki et al., 1989; Hayano et al., 1996). Moreover, we investigated the association between Cd concentration in either blood or urine and indicators of renal dysfunction (urinary total protein, 2 -MG, and N-acetyl--d-glucosaminidase (NAG)) for 2778 inhabitants ≥50 years of age (1114 men and 1664 women) in three different Cd non-polluted areas in Japan. As a result, we showed the presence of renal effects induced by Cd exposure not only in Cd-polluted areas but also in the general environment without any known Cd pollution (Yamanaka et al., 1998; Oo et al., 2000; Suwazono et al., 2000). In health risk assessment of environmental contaminants, a benchmark dose (BMD) approach as defined by Crump (1984, 1995) has been used to estimate the threshold levels of toxic substances (Gaylor et al., 1998). The BMD can be defined as the exposure that corresponds to a certain increase in the probability of an adverse response compared to the background at zero exposure. The benchmark dose low (BMDL) is defined as the value corresponding to the lower 95% confidence interval of the BMD and can be used in evaluating dose–response relationships as a replacement for the
E. Kobayashi et al. / Toxicology Letters 179 (2008) 108–112
no observed adverse effect level (NOAEL) or the lowest observable adverse effect level (LOAEL) (Gaylor et al., 1998). We estimated the BMDL from the threshold levels of urinary Cd using prior data from the Kakehashi River basin (Shimizu et al., 2006) or three different Cd non-polluted areas in Japan (Kobayashi et al., 2006). The values were estimated at 2.9 g/g creatinine (cr) for men and 1.5 g/g cr for women in the Kakehashi River basin and 2.4 g/g cr for men and 3.3 g/g cr for women in the Cd non-polluted areas. By using a multiple logistic model (Budtz-Jørgensen et al., 2001) it is possible to take into account the effects of other possible covariates. We were unable to evaluate the effect of age in the estimation of the threshold level of urinary Cd in the previous Shimizu study (2006) and Kobayashi study (2006). Therefore in the present study we calculated the threshold level of urinary Cd for 2 -MG-uria in Cd-polluted and non-polluted areas (the Kakehashi River basin along with the three different Cd non-polluted areas) by using a multiple logistic model.
Table 1 Means of age, urinary Cd and 2 -microglobulin concentrations according to sex Men (N = 2578)
Age (years)a Cd (g/g cr) 2 -MG (g/g cr)
2.1. Study population The study population for the Cd-polluted area was drawn from the population evaluated in the 1981 and 1982 health survey conducted among the entire population (>50 years of age) residing in the Kakehashi River basin (Shimizu et al., 2006). Among these, subjects who were known from the questionnaire survey to ingest household rice were selected as the target population of the present study. The groups consisted of 3103 participants inhabiting Cd-polluted hamlets (1397 men and 1706 women) and 289 participants (130 men and 159 women) inhabiting non-polluted hamlets. Moreover, 2640 inhabitants (1051 men and 1589 women, ≥50 years of age) of the three different Cd non-polluted areas described in a previous study (Kobayashi et al., 2006) were included in the study population. Thus, the number of subjects for the non-polluted areas totaled 2929 (1181 men and 1748 women). 2.2. Collection of samples and analytical methods Morning urine specimens were collected from all participants and were kept frozen at −20 ◦ C until analysis. The specimens for 2 -MG determination were adjusted to a pH above 6.0, while the specimens for Cd determination were acidified with nitric acid, and then kept frozen. However, when the pH of urine samples obtained from non-polluted areas was <5.4, the 2 -MG concentrations were not measured. Urinary Cd was measured by flameless atomic absorption spectrometry using a Hitachi Model Z180-80 after digestion with nitric acid (Kido et al., 1984). Urinary 2 -MG was determined by radioimmunoassay method (Pharmacia 2 -micro RIA, Pharmacia Diagnostics AB, Sweden) and urinary creatinine concentration was determined by a modified method of the Jaffe reaction (Bonsnes and Taussky, 1945). 2.3. Statistical analysis Multiple logistic regression analysis was used to fit the dose–response model to the data. The dependent variable was a positive finding of 2 -MG-uria. The independent variables were urinary Cd concentration and age, as continuous variables. The concentrations of urinary Cd and 2 -MG were expressed in corrected creatinine units (g cr−1 ). Cut-off values for urinary 2 -MG were defined as those corresponding to the 84th percentile and 97.5th percentile of the 2 -MG level in subjects from nonpolluted areas. Those values were 492 g/g cr and 965 g/g cr in men and 407 g/g cr and 798 g/g cr in women, respectively. Additionally, 1000 g/g cr of urinary 2 -MG was also used as a cut-off value for 2 -MG-uria.
Women (N = 3454)
GM
Min.
Max.
GM
Min.
Max.
64.0 3.0 157
50 0.01 1
91 49.6 107922
64.1 4.2 195
50 0.02 2
95 57.6 186668
N: number of subjects examined; GM: geometric mean; Min.: minimum value; Max.: maximum value; a Arithmetic mean of age was calculated.
The BMD of urinary Cd excretion was calculated using parameters obtained by multiple logistic regression analysis. The benchmark response (BMR) corresponding to the BMD is defined as an additional pre-specified increase in the probability of a positive finding. Assuming that age is fixed to certain values, the log odds of probability of a positive finding, P(di ) at dose di of urinary Cd concentration, is given by ln
2. Materials and methods
109
P(0) P(di ) + ˇ × di = ln 1 − P(0) 1 − P(di )
where P(0) = estimated probability of a “hypothetical” control distribution at urinary Cd = 0 and a certain age, and ˇ = slope for dose-log odds relationship. P(BMD) is given by P(BMD) = P(0) + BMR The log odds of P(BMD) is given by ln
P(d0 ) P(BMD) = ln + ˇ × BMD 1 − P(BMD) 1 − P(d0 )
(1)
BMD is obtained by combining the equation for P(BMD) with that for the dose–response model (1): BMD =
1 [P(d0 ) + BMR] × [1 − P(d0 )] ln ˇ [1 − P(d0 ) − BMR] × [P(d0 )]
In this study, BMDL was calculated using the profile likelihood method (Crump, 1984; Filipsson et al., 2003). BMD/BMDL with BMR at 5% or 10% was calculated for each cut-off value of urinary 2 -MG concentration. The analyses were performed with SPSS, version 12.0J (multiple logistic regression, SPSS Japan Inc., Tokyo, Japan), and Microsoft Excel 2003 (BMDL calculation, Microsoft Corporation, Redmond, WA, USA). P values < 0.05 were considered to be statistically significant.
3. Results Mean values of age, urinary Cd and 2 -MG concentrations of the study population according to sex are shown in Table 1. Prevalences of 2 -MG-uria using three different cut-off values for urinary 2 -MG according to sex and age class (50–59, 60–69, ≥70 years and total) are shown in Table 2. The prevalence calculated for each cut-off value increased with increasing age class in both men and women. Table 3 lists results of multiple logistic regressions for 2 -MG-uria. There were significant positive associations between 2 -MG-uria and urinary Cd concentrations in both men and women. Age also associated positively with 2 -MG-uria in both sexes.
Table 2 Prevalence (%) of 2 -microglobulin-uria using three different cut-off values by sex and age class Age class
Men N
50–59 years 60–69 years 70 years or more Total
902 963 713 2578
Women Cut-off value
N
84th percentile
97.5th percentile
1000 g/g cr
8.6 16.9 31.3 18.0
3.9 9.2 20.5 10.5
3.9 8.8 20.1 10.2
1174 1312 968 3454
Cut-off value 84th percentile
97.5th percentile
1000 g/g cr
10.7 19.4 37.6 28.9
4.6 10.7 24.7 16.8
3.2 9.7 22.3 14.8
N: number of subjects examined. The 84th percentile and 97.5th percentile of the 2 -MG level were calculated in subjects from non-polluted areas (492 g/g cr and 965 g/g cr in men and 407 g/g cr and 798 g/g cr in women).
110
E. Kobayashi et al. / Toxicology Letters 179 (2008) 108–112
Table 3 Results of multiple logistic regression analyses for 2 -microglobulin-uria Cut-off value
Men, OR (95%CI)
Women, OR (95%CI)
84th percentile Cd (g/g cr) Age (years)
1.23 (1.19–1.26) 1.08 (1.07–1.09)
1.20 (1.18–1.23) 1.10 (1.09–1.11)
97.5th percentile Cd (g/g cr) Age (years)
1.24 (1.20–1.29) 1.10 (1.08–1.11)
1.25 (1.22–1.28) 1.12 (1.11–1.14)
1000 g/g cr Cd (g/g cr) Age (years)
1.25 (1.21–1.29) 1.10 (1.08–1.11)
1.26 (1.23–1.30) 1.13 (1.12–1.15)
OR: odds ratio = 95% CI: 95% confidence intervals.
The BMD/BMDL for each of the cut-off values was calculated using a multiple logistic model, setting BMR at 5% or 10% (Table 4). When BMR was 5%, the BMD/BMDL of urinary Cd for the 84th percentile of urinary 2 -MG for mean age, 55, 65, and 75 years was 3.0/2.7, 4.6/4.2, 2.8/2.6, and 1.8/1.6 g/g cr in men and 3.4/3.2, 5.8/5.5, 3.2/3.1, and 1.8/1.7 g/g cr in women, respectively. When an abnormal value of 5% was employed, the BMD/BMDL of urinary Cd for the 97.5th percentile for mean age, 55 years, 65 years, and 75 years was 4.9/4.5, 7.6/7.0, 4.6/4.3, and 2.6/2.4 g/g cr in men and 5.9/5.6, 9.7/9.2, 5.6/5.3, and 2.8/2.6 g/g cr in women, respectively. When an abnormal value of 5% was employed, the BMD/BMDL of urinary Cd for 1000 g/g cr of urinary 2 -MG for mean age, 55, 65, and 75 years was 5.0/4.6, 7.7/7.1, 4.7/4.3, and 2.7/2.4 g/g cr in men and 6.7/6.3, 10.9/10.4, 6.3/5.9, and 3.1/2.6 g/g cr in women, respectively. 4. Discussion In the present study, we calculated the threshold level of urinary Cd for 2 -MG-uria in Cd-polluted and non-polluted areas in Japan, including the Kakehashi River basin and three different Cd nonpolluted areas, using a multiple logistic model. To our knowledge, this is the first estimation of BMD/BMDL from a multiple logistic model for the threshold level of urinary Cd taking into account an age effect. As shown in Table 2, it is clear that the prevalence of 2 MG-uria using any cut-off value is increased with increasing age in both sexes. In our previous studies (Shimizu et al., 2006; Kobayashi et al., 2006) we were not able to consider the effects of age on the estimation of BMD, due to limitations of the software used (Bench-
mark Dose Software; US EPA). Therefore the threshold values of urinary Cd were estimated without taking the confounding effect of age on 2 -MG-uria into consideration in those previous studies. On the other hand, we could estimate the threshold level of urinary Cd at each age in this study by using a multiple logistic regression analysis including age into the model. The obtained BMDL gradually decreased with increasing age, for example the value for 84th percentile for 55, 65, and 75 years was 4.2, 2.6, and 1.6 g/g cr in men and 5.5, 3.1, and 1.7 g/g cr in women (BMR was 5%) as shown Table 4. As shown in Tables 3 and 4, age was positively related to the 2 -MG-uria, yielding increased probability of 2 -MG-uria at 0 exposure of cadmium (P(0)). Therefore, significant increase in probability of 2 -MG-uria compared to 0 exposure occurs at lower urinary Cd level in older population than in younger population. Thus, the present study reveals how much the threshold amount of urinary Cd is affected by individual age. From the epidemiological standpoint these results indicated that the margin between the threshold level and average excretion level of urinary Cd was small in the older population in Japan. To prevent the adverse health effects caused by exposure to Cd, it is thus important to establish the threshold level of Cd exposure at each age. Moreover, the age range of the target population should be as wide as possible for more accurate adjustment for the confounding effect of age. In terms of statistical power, the investigated cohort (N = 6032) was much larger than in Shimizu’s (2006, N = 3392) and Kobayashi’s (2006, N = 2778) studies. Furthermore, the exposure range was broader than in the previous studies because we combined data from non-polluted and Cd-polluted areas. For example, only 4.0% of subjects had urinary cadmium ≥7.0 g/g cr in the previous study conducted in non-polluted areas (Kobayashi et al., 2006), whereas that value was observed more than 30% of the time in the other study conducted in Cd-polluted areas (Shimizu et al., 2006). The estimated threshold values of urinary Cd for 2 -MG-uria (BMR = 5%, with a cut-off value at the 84th percentile of urinary 2 -MG) were estimated at 2.9 g/g cr in men and 1.5 g/g cr in women of the Kakehashi River basin (Shimizu et al., 2006), and 2.4 g/g cr in men and 3.3 g/g cr in women in Cd non-polluted areas (Kobayashi et al., 2006). Since there was little difference between the prevalence of 2 -MG-uria in the Cd-polluted areas (21.0%) and in the non-polluted areas (14.2%) in men, the threshold values (for the mean age) were similar. On the there hand, since the prevalence of 2 -MG-uria was 30.4% and 12.9% in each study, respectively. The threshold values of urinary Cd were different. In this study the prevalence of 2 -MG-uria was 18.0% in men and 28.9% in
Table 4 BMD/BMDL of urinary cadmium for 2 -microglobulin-uria according to sex Cut-off value
84th percentile
97.5th percentile
1000 g/g cr
Age
Mean age 55 years 65 years 75 years Mean age 55 years 65 years 75 years Mean age 55 years 65 years 75 years
Men
Women
BMD/BMDL of U-Cd (g/g cr)
BMD/BMDL of U-Cd (g/g cr)
BMR = 5%
BMR = 10%
P(0) (%)
BMR = 5%
BMR = 10%
3.0/2.7 4.6/4.2 2.8/2.6 1.8/1.6 4.9/4.5 7.6/7.0 4.6/4.3 2.6/2.4 5.0/4.6 7.7/7.1 4.7/4.3 2.7/2.4
5.0/4.6 7.1/6.6 4.8/4.4 3.2/2.9 7.4/6.8 10.5/9.6 7.1/6.5 4.4/4.1 7.5/6.9 10.6/9.7 7.2/6.6 4.5/4.1
6.8 3.5 7.3 14.4 2.8 1.3 3.1 7.4 2.7 1.2 3.0 7.1
3.4/3.2 5.8/5.5 3.2/3.1 1.8/1.7 5.9/5.6 9.7/9.2 5.6/5.3 2.8/2.6 6.7/6.3 10.9/10.4 6.3/5.9 3.1/2.6
5.7/5.4 8.8/8.3 5.4/5.1 3.3/3.1 8.6/8.1 12.7/12.1 8.2/7.8 4.6/4.4 9.4/8.9 13.9/13.4 9.0/8.5 5.0/4.5
P(0) (%) 6.4 2.8 7.0 16.3 2.0 0.7 2.2 6.7 1.4 0.5 1.6 5.3
BMR: benchmark response; P (0): probability of positive finding at zero exposure = BMD/BMDL are calculated assuming mean age (men: 64.0; women: 64.1), and 55, 65, and 75 years of age.
E. Kobayashi et al. / Toxicology Letters 179 (2008) 108–112
women, and the threshold values of urinary Cd for the mean age was 2.7 g/g cr in men and 3.2 g/g cr in women, respectively. Thus, we feel that the present study improved the accuracy of BMD/BMDL determination by adopting an improved methodology and a greater quality and quantity of subjects. Urinary Cd threshold values have been estimated using different methodologies from that in the present study. Hayano et al. (1996) applied a logistic regression model to data from Cd-polluted areas and employed a cut-off value for 2 -MG of 1000 g/g cr. The estimated Cd threshold values ranged from 1.6 to 4.6 g/g cr in different sex and age groups. In the Hayano study, the abnormality rate of urinary 2 -MG in the controls was substituted into each regression formula calculated with the urinary Cd concentration and the abnormality rate of 2 -MG in Cd-exposed subjects, and threshold levels of urinary Cd were obtained. However, those control groups had in fact been exposed to Cd because Cd was detected in their urine. Thus the control groups did not strictly correspond to a population without any exposure to Cd. Unlike in experimental animals, a perfect control with no Cd exposure is physically impossible in humans. In contrast, the BMD method does not require abnormality rates of urinary findings in the control subjects. The probability of adverse response at zero exposure was estimated based on the whole dose–response curve. Then, the exposure level was calculated where a certain change in response compared to the background was expected. In this sense, the BMD/BMDL calculations obtained in the present study are more accurate and provide more information than the threshold levels obtained in the Hayano study. In this study we used the 84th percentile and 97.5th percentile of the non-exposed subjects and 1000 g/g cr of 2 -MG as cut-off values. Kido et al. (1988) investigated the reversibility of 2 -MG-uria in 74 inhabitants (32 men and 42 women) of Cd-polluted regions of the Kakehashi River basin. The subjects participated in two examinations, one conducted just after the cessation of Cd exposure and the other 5 years later (when Cd exposure had ended). In cases who had ≥1000 g/g cr of 2 -MG in the first investigation, almost all individuals showed deterioration of 2 -MG-uria. Accordingly the threshold level for irreversible renal damage was determined to be 1000 g/g cr of urinary 2 -MG excretion. Other researchers in Japan have also obtained similar results (Iwata et al., 1993; Cai et al., 2001). Nakagawa et al. (1993) conducted a 9-year follow-up study of 3178 subjects (aged ≥50 years) in the Cd-polluted Kakehashi River basin. When compared with a group with <300 g/g cr of urinary 2 -MG at entry, mortality ratios were found to increase as urinary 2 MG excretion increased in both sexes. Thus, the greater the tubular dysfunction, the less favorable the life prognosis, and even slightly increased urinary excretion of 2 -MG (300 to <1000 g/g cr) was associated with increased mortality. In the present study, the 84th percentile for 2 -MG was 492 g/g cr for men and 407 g/g cr for women, higher than the 300 g/g cr that was reported to be the borderline value indicative of a poorer life prognosis. Taken together, 1000 g/g cr for 2 -MG is thought to be an extremely robust cut-off which reflects both irreversible renal dysfunction and poor life prognosis. The lowest cut-off level in the present study, the 84th percentile, is also thought to be valid because of the negative impact on life prognosis according to the Nakagawa study (1993). Recently, Nakagawa et al. (2006) reported the results of a 15-year follow-up survey of the same population as mentioned above (Nakagawa et al., 1993) to investigate excess mortality and the dose–response relationship between mortality and urinary Cd excretion. The mortality risk was increased among the subjects, especially in women, with urinary Cd >3 g/g cr after adjustment for age. In our study, the threshold level of urinary Cd for the 84th percentile of 2 -MG among the controls was estimated at
111
2.7 g/g cr in men and 3.2 g/g cr in women for the mean age (about 64 years in both sexes). Thus the present study and their study together suggest that subjects with a urinary Cd of greater than about 3 g/g cr will have renal dysfunction and an increased risk of mortality. In conclusion, the threshold level of urinary Cd was estimated in persons living in the general environment with and without Cd pollution in Japan. The values were calculated more exactly than previously using a multiple logistic model taking into consideration the effects of increasing age.
References Bernard, A., Roels, H., Thielemans, N., Lierde, V.M., Lauwerys, R., 1992. Assessment of the causality of cadmium-protein relationships in the urine of the general population with reference to the Cadmibel study. IARC Sci. Pub. 118, 341– 346. Bernard, A., Thielemans, N., Roels, H., Lauwerys, R., 1995. Association between NAGB and cadmium in urine with no evidence of a threshold. Occup. Environ. Med. 52, 177–180. Bonsnes, R.W., Taussky, H.H., 1945. On the calorimetric determination of creatinine by the Jaffe reaction. J. Biol. Chem. 15, 581–591. Buchet, J.P., Lauwerys, R.R., Roles, H., Bernard, A., Claeys, F., Ducoffer, G., Plaen, P.D., Staessen, J., Amery, A., Lijnen, P., Thijs, L., Rondia, D., Sartor, F., Remy, A.S., Nick, L., 1990. Renal effects of cadmium body burden of the general population. Lancet 336, 699–702. Budtz-Jørgensen, E., Keiding, N., Grandjean, P., 2001. Benchmark dose calculation from epidemiological data. Biometrics 57, 698–706. Cai, Y., Aoshima, K., Katoh, T., Teranishi, H., Kasuya, M., 2001. Renal tubular dysfunction in male inhabitants of a cadmium-polluted area in Toyama, Japan—an eleven-year follow-up study. J. Epidemiol. 11, 180–189. Crump, K.S., 1984. A new method for determining allowable daily intakes. Fundam. Appl. Toxicol. 4, 854–871. Crump, K.S., 1995. Calculation of benchmark dose from continuous data. Risk Anal 15, 79–89. Filipsson, A.F., Sand, S., Nilsson, J., Victorin, K., 2003. The benchmark dose methods—Review of available models, and recommendations for application in health risk assessment. Crit. Rev. Tocicol. 33, 505–542. Gaylor, D., Ryan, L., Krewski, D., Zhu, Y., 1998. Procedures for calculation benchmark doses for health risk assessment. Reg. Toxicol. Pharmacol. 28, 150–164. Hayano, M., Nogawa, K., Kido, T., Kobayashi, E., Honda, R., Tsuritani, I., 1996. Dose–response relationship between urinary cadmium concentration and 2 microglobulinuria using logistic regression analysis. Arch. Environ. Health 51, 162–167. Ishizaki, M., Kido, T., Honda, R., Tsuritani, I., Yamada, Y., Nakagawa, H., Nogawa, K., 1989. Dose–response relationship between urinary cadmium and 2 microglobulin in a Japanese environmentally cadmium exposed population. Toxicology 58, 121–131. Iwata, K., Saito, H., Moriyama, M., Nakano, A., 1993. Renal tubular function after reduction of environmental cadmium exposure: a ten-year follow-up. Arch. Environ. Health 48, 157–163. Jakubowski, M., Trojanowska, B., Kowalska, G., Gendek, E., Stajewska, B., Krajewska, B., Jajte, J., 1987. Occupational exposure to cadmium and kidney dysfunction. Int. Occup. Environ. Health 59, 567–577. Jakubowski, M., Razniewska, G., Halatek, T., Trzcinka, O.M., 1992. Integrated of occupational exposure to cadmium as a predictor of kidney dysfunction. IARC Sci. Pub. 118, 319–324. ¨ Jarup, L., Hellstrom, L., Alfven, T., Carlsson, M.D., Grubb, A., Persson, B., Pettersson, C., Spang, G., Schut, A., Elinder, C.G., 2000. Low level exposure to cadmium and early kidney damage: the OSCAR study. Occup. Environ. Med. 57, 668– 672. Kido, T., Tsuritani, I., Honda, R., Ishizaki, M., Yamada, Y., Nogawa, K., 1984. A direct determination of urinary cadmium by graphite-furnace atomic absorption spectrometry using the zeeman effect. J. Kanazawa Med. Univ. 9, 70–75 (in Japanese). Kido, T., Honda, R., Tsuritani, I., Yamaya, H., Ishizaki, M., Nogawa, K., 1988. Progress of renal dysfunction in inhabitants environmentally exposed to cadmium. Arch. Environ. Health 43, 213–221. Kobayashi, E., Suwazono, Y., Uetani, M., Inaba, T., Oishi, M., Kido, T., Nishijo, M., Nakagawa, H., Nogawa, K., 2006. Estimation of benchmark dose as the threshold levels of urinary cadmium, based on excretion of total protein, beta2-microglobulin, and N-acetyl-beta-d-glucosaminidase in cadmium non-polluted regions in Japan. Environ. Res. 101, 401–406. Nakagawa, H., Nishijo, M., Morikawa, Y., Senma, M., Kawano, S., Ishizaki, M., Sugita, N., Nishi, M., Kido, T., Nogawa, K., 1993. Urinary 2 -microglobulin concentration and mortality in a cadmium-polluted area. Arch. Environ. Health 48, 428– 435. Nakagawa, H., Nishijo, M., Morikawa, Y., Miura, K., Tawara, K., Kuriwaki, J., Kido, T., Ikawa, A., Kobayashi, E., Nogawa, K., 2006. Urinary cadmium and mortality among inhabitants of a cadmium-polluted area in Japan. Environ. Res. 16, 273–279.
112
E. Kobayashi et al. / Toxicology Letters 179 (2008) 108–112
Nogawa, K., 1981. Itai-Itai disease and follow-up studies. In Nriagu, J.O. (Ed.), Cadmium in the Environment. Part 2. Health Effects. Wiley Interscience Publication, New York, Chichester, Brisbane, Toronto, Singapore, pp. 1–37. Nomiyama, K., 1986. The chronic toxicity of cadmium: influence of environment and other variables. In: Foulkes, E.C. (Ed.), Handbook of Environmental Pharmacology. Springer-Verlag, Berlin/Heidelberg/New York/Tokyo, pp. 101– 133. Oo, Y.K., Kobayashi, E., Nogawa, K., Okubo, Y., Suwazono, Y., Kido, T., Nakagawa, H., 2000. Renal effects of cadmium intake of a Japanese general population in two areas unpolluted by cadmium. Arch. Environ. Health 55, 98–103.
Shimizu, A., Kobayashi, E., Suwazono, Y., Uetani, M., Oishi, M., Inaba, T., Kido, T., Nogawa, K., 2006. Estimation of benchmark doses for urinary cadmium based on 2 -microglobulinuria excretion in cadmium-polluted regions of the Kakehashi River basin, Japan. Int. J. Environ. Health Res. 16, 329–337. Suwazono, Y., Kobayashi, E., Nogawa, K., Okubo, Y., Kido, T., Nakagawa, H., 2000. Renal effects of cadmium exposure in cadmium nonpolluted areas in Japan. Environ. Res. 84, 44–55. Yamanaka, O., Kobayashi, E., Nogawa, K., Suwazono, Y., Sakurada, I., Kido, T., 1998. Association between renal effects and cadmium exposure in cadmiumnonpolluted area in Japan. Environ. Res. 77, 1–8.