Serial changes in urinary cadmium concentrations and degree of renal tubular injury after soil replacement in cadmium-polluted rice paddies.
ABSTRACT Targeting persons requiring observation in the Cd-polluted Kakehashi River basin a serial observation study was conducted. Namely we followed the serial changes in 50 subjects who ingested household rice for 10 years after replacement of Cd-polluted soil in rice paddies. The serial changes in urinary substance levels in individuals were determined adjusting for the potential effect of age using a general linear mixed model. Cd excretions decreased with increasing number of years elapsed, with the partial regression coefficients of the number of years elapsed statistically significant in the women. The ratio of the Cd excretion theoretical values at the completion of soil replacement and 10 years later was 0.60 and the reduction rate was calculated as 5.0% per year in women. However, it was surmised that in practice a decrease to the level of inhabitants of non-polluted districts would not be achievable. The indices of renal tubular injury (beta(2)-microglobulin, retinol binding protein (RBP), total protein, amino-N and glucose) with the exception of amino-N in men showed increased excretion in both sexes with increasing number of years elapsed with statistically significant differences in RBP and total protein in both sexes and glucose in men. In this study using a general linear mixed model, which is an appropriate statistical method to perform a follow-up study, Cd concentrations in rice and urine were found to decrease after Cd exposure was reduced, but the degree of renal tubular injury was not found to improve, leading to the conclusion that the renal tubular injury induced by environmental Cd exposure is irreversible.
Serial changes in urinary cadmium concentrations and degree of renal
tubular injury after soil replacement in cadmium-polluted rice paddies
Etsuko Kobayashia,∗, Yasushi Suwazonoa, Ryumon Hondab, Mirei Dochia,
Muneko Nishijoc, Teruhiko Kidod, Hideaki Nakagawac
aDepartment of Occupational and Environmental Medicine (A2), Graduate School of Medicine, Chiba University,
1-8-1 Inohana, Chuohku, Chiba 260-8670, Japan
bDepartment of Nursing, Kanazawa Medical University, 1-1 Daigaku, Uchinada, Ishikawa 920-0293, Japan
cDepartment of Public Health, Kanazawa Medical University, 1-1 Daigaku, Uchinada, Ishikawa 920-0293, Japan
dDepartment of Community Health Nursing, Kanazawa University School of Health Sciences, 5-11-80 Kodatsuno, Kanazawa 920-0942, Japan
Received 23 August 2007; received in revised form 24 October 2007; accepted 24 October 2007
Available online 1 November 2007
the serial changes in 50 subjects who ingested household rice for 10 years after replacement of Cd-polluted soil in rice paddies. The serial changes
in urinary substance levels in individuals were determined adjusting for the potential effect of age using a general linear mixed model.
Cd excretions decreased with increasing number of years elapsed, with the partial regression coefficients of the number of years elapsed
statistically significant in the women. The ratio of the Cd excretion theoretical values at the completion of soil replacement and 10 years later
was 0.60 and the reduction rate was calculated as 5.0% per year in women. However, it was surmised that in practice a decrease to the level of
inhabitants of non-polluted districts would not be achievable.
The indices of renal tubular injury (?2-microglobulin, retinol binding protein (RBP), total protein, amino-N and glucose) with the exception of
amino-N in men showed increased excretion in both sexes with increasing number of years elapsed with statistically significant differences in RBP
and total protein in both sexes and glucose in men.
In this study using a general linear mixed model, which is an appropriate statistical method to perform a follow-up study, Cd concentrations in
rice and urine were found to decrease after Cd exposure was reduced, but the degree of renal tubular injury was not found to improve, leading to
the conclusion that the renal tubular injury induced by environmental Cd exposure is irreversible.
© 2007 Elsevier Ireland Ltd. All rights reserved.
Keywords: Cadmium; Soil replacement; Follow-up study; General linear mixed model; Renal tubular injury
The agricultural district of the Kakehashi River basin in
Ishikawa Prefecture has suffered from a continuous decrease
in production from around 1900 due to cadmium (Cd) pollution
originating from an upstream mine. The mine provided finan-
cial compensation until 1965. In 1973 heavily Cd-polluted rice
was detected in this region, prompting the undertaking of two
∗Corresponding author. Tel.: +81 43 226 2065;
fax: +81 43 226 2066.
E-mail address: email@example.com (E. Kobayashi).
large-scale health investigations in 1974–1975 and 1981–1983
check-ups up. Our group has retained the results of urinary sub-
stances (Cd, ?2-microglobulin (?2-mg), retinol binding protein
(RBP), total protein, amino-N, glucose, and others) of the per-
sons requiring observation from the Kakehashi River basin for
34 years since 1974.
To reduce Cd concentrations in homegrown rice the most
efficacious method is to remove the Cd-polluted soil present in
rice paddies and to replace it with non-polluted soil. In 1977,
in the scattered Cd-polluted districts in Japan (35 districts in 19
prefectures) rice paddies measuring a total of 4297.2ha were
0378-4274/$ – see front matter © 2007 Elsevier Ireland Ltd. All rights reserved.
judged to require such treatment. In the Kakehashi River basin,
459.7ha (28 hamlets) were thus designated, with soil replace-
ment undertaken during the 12-year period from 1977 to 1988.
With regard to the influence of such soil replacement on renal
dysfunction, in the Kakehashi River basin Kido et al. (1988), in
Tsushima Island, Nagasaki Prefecture Iwata et al. (1993), and
in the Jinzu River basin, Toyama Prefecture, Cai et al. (2001)
reported their respective findings. All three groups noted that
soil replacement, the excretion of indices of renal tubular injury
such as ?2-mg increased. However, since they performed the
comparison with the excretion values only at every two time
points of the observation in their studies, the serial changes
throughout the observational period could not be shown.
with missing data.
2. Materials and methods
2.1. Replacement of polluted soil and changes in hamlet rice Cd
In this study, to quantitatively determine changes in urinary substance con-
centrations after replacement of polluted soil, seven hamlets in which soil
replacement had been completed approximately 1 year were selected for anal-
ysis. In four of these seven hamlets (hamlets A–C, F) the soil replacement
had been completed within 1 year in 1978 or 1979. In the other three hamlets
(hamlets D, E, G) this had been completed within a 2-year period: 1979–1980,
1980–1981, or 1986–1987. The changes in hamlet rice Cd concentrations are
shown in Table 1 (Ishikawa Prefecture, Department of Environment, necessary
data from 1974–1995 extracted).
Of the inhabitants of the seven hamlets who were recognized as being
persons requiring observation based on the results of the large-scale screen-
ing check-ups conducted in 1974–1975 and 1981–1983, 50 subjects (20 men,
30 women) who were known from the questionnaire survey to ingest house-
hold rice were selected for the purposes of this study. Age (mean age and
range) at the completion of polluted soil replacement was 72.4 years (52–89
years) and 70.1 years (51–90 years) in the men and women respectively.
Fifteen subjects (8 men, 7 women) had died by the end of observational
2.3. Analysis of urinary substance concentrations
The urinary Cd concentration was measured by graphite-furnace atomic
absorption spectrometry after wet ashing in HNO3/H2SO4/HClO4and extrac-
tion with ammonium pyrrolidine dithiocarbamate-metyl isobutyl ketone
(APDC-MIBK). Urinary ?2-mg concentration was measured by a radioim-
munoassay method (Pharmacia ?2-micro RIA, Pharmacia Diagnostics AB,
Sweden). RBP was analyzed with a kit using a single immuno-diffusion method
(Nippon Boehringer Ingelheim Co. Ltd., Japan). Urinary total protein was
determined by the modified method of Kingsbury-Clark. Amino-N in urine
was determined with the TNBS (2,4,6-trinitorobenzene sulfonic acid) method
(Fukushima and Kobayashi, 1975). Urinary glucose was analyzed with a kit
using o-toluidine-boric acid (Glucose Test Wako, Wako Pharmacy, Japan). Uri-
nary creatinine was determined by the Jaffe reaction method (Bonsnes and
2.4. Calculation methods of urinary substance concentrations
some subjects were missing for the year that soil replacement was completed.
However, we were able to confirm retrospectively that values for the previous
the influence of increasing age into account the observation was concluded at
10 years taking the successive years after the replacement of polluted soil was
Changes in cadmium concentrations (ppm) in rice associated with soil replacement in target hamlets
Values after soil replacement were shown in parenthesis. * Soil replacement uncompleted. ** Rice samples could not be collected. When cadmium concentrations in
rice after soil replacement were repeatedly lower than 0.4ppm, the value provisionally set by the Japanese government at which the distribution of rice is forbidden,
subsequently measured values were not made public.
completed as year 1, year 2, etc. Accordingly, in the hamlets in which soil
until 1997. The soil replacement in Hamlets D, E, and G reached 81.8% in
1980, 91.7% in 1980, and 80.7% in 1987 respectively. Accordingly, these years
were regarded as the respective years of completion of soil replacement. Since
the methodology used to measure urinary ?2-mg concentrations in 1974–1982
differed, these values were excluded from the present analysis. The urinary
substance concentrations were used after correction for creatinine values and
common logarithmic transformation.
2.5. Statistical analysis methods
The data of the analyzed period were calculated according to sex using a
general linear mixed model and expressed as serial estimated values (Ishimura
and Nejima, 2004). In order to avoid multicollineality among the num-
ber of observation years and the years of age at annual measurement, the
age of each subject at the completion of the replacement of polluted soil
was inserted into the model as a covariate to adjust for the age effect.
SPSS 12.0J (SPSS Japan Inc., Japan) was used for statistical analysis of the
Fig. 1. Serial changes in urinary substance concentrations after the completion of soil replacement according to sex and age group.
Fig. 1. (Continued).
Fig. 1 shows the changes in the urinary substance concen-
trations of the analyzed subjects listed according to sex and
age group (50–59, 60–69, 70–79, ≥80 years) with the year of
completion of soil replacement considered as year 0, and the
subsequent years as years 1–10. Urinary Cd concentrations in
men showed an increasingly higher excretion level in increas-
ing age groups. Within the respective age groups, Cd excretion
appeared to decrease as more time elapsed from the completion
of the soil replacement in both sexes. Excretion levels of ?2-
mg, RBP, total protein, amino-N, and glucose were increased in
both sexes in increasing age groups. Within the respective age
appeared to increase with increasing number of years elapsed.
Table 2 shows the regression coefficients for urinary sub-
linear mixed model. Urinary Cd concentrations increased with
increasing age in men, while in contrast decreasing in women.
In both sexes Cd excretions decreased with increasing number
Regression coefficients for urinary substance concentrations according to sex
calculated from a general linear mixed model
Cd (?g/g cr)
?2-mg (?g/g cr)
RBP (mg/g cr)
Total protein (mg/g cr)
Amino-N (mg/g cr)
Glucose (mg/g cr)
Age: age at the beginning of the observation; year: year after the completion of
of years elapsed, with the partial regression coefficients of the
number of years elapsed statistically significant in the women.
The ratio of the Cd excretion theoretical values at the comple-
tion of soil replacement and 10 years later was 0.60 in women.
Indices of renal tubular injury with the exception of amino-N in
men showed increased excretion in both sexes with increasing
number of years elapsed. Partial regression coefficients of the
number of years elapsed were statistically significant for RBP
and total protein in both sexes and for glucose in men. From the
regression formulas the respective ratios of the excretion theo-
retical values at the end of soil replacement and 10 years later
were calculated as 3.36, 2.67, and 2.41 for RBP, total protein,
and glucose in men, and 2.81 and 4.20 for RBP and total protein
in women. Fig. 2 shows the serial changes in urinary substances
in the 55, 65, 75 and 85 years old groups.
The most notable feature of the present study was that the
serial observation conducted over long periods, and the serial
mined adjusting for the potential effect of age using a general
linear mixed model. More recently, one of the analysis meth-
ods used in lieu of classical multivariable regression models
is the general linear mixed model (Harville, 1977; Laird and
Ware, 1982). In this model are included parameters taking into
account individual variables, thereby facilitating more accurate
calculation of changes in repeatedly measured data. For exam-
changes in prostate specific antigen (PSA) levels. However, in
this general linear mixed model have yet to be conducted. As
can also be appreciated from Fig. 1, the fact that excretion lev-
els of urinary substances differ according to age group makes it
difficult to discern trends in excretion levels. Our study is thus
thought to represent the first study in which the age factor was
taken into consideration. As shown in Fig. 2 the serial changes
in levels of urinary substances in the 55, 65, 75 and 85 years old
groups in the both sexes were obtained.
With regard to the health impact of Cd exposure attenua-
tion, a number of studies in Western countries have focused
on factory workers exposed to Cd (Roels et al., 1982; Piscator,
1983; Elinder et al., 1985). In Japan, polluted soil replacement
has been undertaken to reduce the amount of Cd exposure of
inhabitants of Cd-polluted districts, and the influence of such
soil replacement on renal dysfunction has been investigated in
three studies. Kido et al. (1988) focused on the same Kakehashi
River basin in Ishikawa Prefecture studied in the present inves-
tigation. Namely, in one heavily polluted hamlet in which soil
was replaced during the period 1977–1980, urinary data in 74
5 years later (1986). Comparing the geometric mean values, the
?2-mg, glucose, and amino-acid concentrations in 1986 were
increased by 46%, 60%, and 13% respectively as compared to
the values noted in 1981. No significant changes in urinary Cd
were found during the 5-year period. Iwata et al. (1993) inves-
tigated renal tubular function from 1983, which was the year
of completion of soil replacement, in a Cd-polluted district in
Tsushima Island, Nagasaki Prefecture using health screening
data obtained six times (1979, 1982, 1986, 1987, 1988, 1989)
during the 10-year period 1979–1989. They found that the geo-
metric mean of the urinary ?2-mg concentration for 28 subjects
aged ≥40 years in 1979 increased from 1135.8?g/g cr in 1979
in 16 subjects with urinary ?2-mg concentrations greater than
1000?g/g cr in 1979. Moreover, they reported that in 48 per-
sons examined in 1982, 1986 and 1989, the geometric mean
of urinary Cd concentrations decreased from 8.49?g/g cr in
1982 to 6.03?g/g cr in 1989. Cai et al. (2001) reached the same
conclusion as Kido et al. (1988) and Iwata et al. (1993) in an
investigation conducted twice (1985–86, 1996–97) in a total of
Prefecture, where numerous Itai-itai disease patients have been
lar dysfunction (RTD), which was defined as urinary ?2-mg
cr, increased from 18% and 23% respectively in the 1985–1986
study to 25% and 32% respectively in the 1996–1997 study, and
a total of 12 new cases (12%) of RTD were found. The mean Cd
levels in urine were significantly reduced from 7.9 and 9.5?g/g
cr in the initial study to 6.9 and 6.8?g/g cr in the follow-up
one. In each of these three studies despite the fact that Cd con-
Fig. 2. Serial changes in urinary substance concentrations in 55, 65, 75 and 85 years old according to sex calculated from estimated values in a general linear mixed
model. (?) 55 years old; (?) 65 years old; (?) 75 years old; (×) 85 years old; N, number of subjects examined; NS, not significant.
centrations in food, rice and urine decreased after Cd exposure
was reduced, the degree of renal tubular injury was not found to
improve, leading to the conclusion that the renal tubular injury
induced by environmental Cd exposure is irreversible.
In our study, Cd concentrations also decreased in both sexes
with the partial regression coefficients of the number of years
elapsed statistically significant in the women. It is well-known
that urinary Cd excretion has been shown to be related to body
burden since the half-life of Cd is very long. However, it has
a limitation in that a sharp increase may be noted after tubular
dysfunction occurs or some cancers damage organs rich in Cd,
such as stomach or liver (Nordberg, 1972; Bernard et al., 1980;
Kido et al., 1989). Previously I reported that when 541 inhab-
itants aged over 20 of the Jinzu River basin were divided into
2 groups according to urinary ?2-mg concentration, in the low
?2-mg (<4mg/L) group the average of urinary Cd concentra-
tions increased with increasing residence time in years and the
trend was similar in the controls showing a peak in 50–59 years.
On the other hand in the high ?2-mg group (≥4mg/L) aver-
age of urinary Cd concentrations showed higher levels than the
low ?2-mg group and did not relate with number of residence
years (Kobayashi, 1982). These differences between low and
high ?2-mg groups may be explained by the increased excre-
tion of Cd after renal damage has occurred. In the present study
Cd excretion levels were higher in both sexes throughout the
observational period as compared to controls whose levels were
itants (2–86 years, 157 men, 272 women) of a non-Cd polluted
district (Honda et al., 1981). On the other hand, Cd excretions
significantly decreased with increasing number of years elapsed
attributable to a reduction in Cd concentrations in homegrown
rice (Table 1). On the other hand, Ishizaki et al. (1971) reported
organ Cd values from five Itai-itai disease patients and 38 con-
in the Itai-itai disease patients than in the controls of the same
age group. In contrast, Cd contents of the kidney cortices were
lower. They ascribed the low Cd content in the kidney in the
Itai-itai disease patients to the advanced kidney damage present
study reflects not only the reduction in rice Cd concentrations
but also the decrease in renal Cd content in the subjects. But it
should be noted that the Cd excretion level was higher yet than
that of the controls. From the regression formulas in the females
it was calculated that the urinary Cd reduction rate was 5% per
year, and that 21–28 years would be required to decrease the
Cd excretion level to that of the controls. Thus, although the
urinary Cd concentration decreased after the completion of pol-
to the level of inhabitants of non-polluted districts would not be
lar injury indices with the exception of urinary glucose in both
sexes in their 50s and 60s already exceeded the respective val-
ues of a control group in Ishikawa Prefecture (Kobayashi et al.,
1981). After completion of the soil replacement, renal tubular
injury indices with the exception of amino-N in men showed
increased excretion in both sexes with increasing number of
as a percentage of the previous year’s values were 113%, 110%,
environmental Cd exposure was found to be irreversible in the
present study. Kjellstrom (1986) described the probable mecha-
nism for Cd-induced renal damage as follows; the accumulation
of Cd bound to metallothionein (MT) complex in the kidney
takes place mainly in the proximal tubular cell. If so much Cd
enters the cell that the available MT cannot bind any more Cd, a
likely to follow and Cd interference with renal tubular cell func-
tion occurs. In our study the target subjects had renal tubular
dysfunction. Therefore, it may be supposed that the reduction
of MT production capacity in the renal cells was a cause of
their irreversible renal tubular injury. Moreover, the extent of
decreases in urinary Cd concentration following the comple-
tion of polluted soil replacement was clearly different between
unclear, it should be noted that renal tubular injury in women
was demonstrated to be more severe than in men, suggesting
impairment in the ability to product MT and retain Cd in renal
In the present study using a general linear mixed model, we
conducted a 10-year follow-up investigation of the same sub-
jects, and determined the serial changes in urinary substance
levels in individuals adjusting for the potential effect of age.
We demonstrated that the urinary Cd reduction rate was 5% per
environmental Cd exposure was found to be irreversible based
on the serial changes in urinary substance excretion.
Bernard, A., Goret, A., Buchet, J.P., Roles, H., Lauwerys, R.R., 1980. Signif-
icance of cadmium level in blood and urine during long-term exposure of
rats to cadmium. J. Toxicol. Environ. Health 6, 175–184.
Bonsnes, R.W., Taussky, H.H., 1945. On the calorimetric determination of cre-
atinine by the Yaffe reaction. J. Biol. Chem. 158, 581–591.
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.
Elinder, C.G., Edling, C., Lindberg, E., Kagedal, B., Verterberg, O., 1985. ?2-
Microglobulinuria among workers previously exposed to Cd: follow-up and
dose-response analysis. Am. J. Ind. Med. 8, 553–564.
Fukushima, M., Kobayashi, E., 1975. Studies on some urinary screening tests
for renal dysfunction induced by cadmium exposure. 4. Determination of
free amino acid in urine by 2,4,6-trinitorobenzene 1-sulfonic acid method.
J. Juzen Med. 84, 166–171.
Harville, D.A., 1977. Maximum likelihood approaches to variance component
estimation and to related problems. J. Am. Stat. Assoc. 72, 320–340.
Honda, R., Kobayashi, E., Nogawa, K., Sakamoto, M., 1981. Urinary excretion
of cadmium, copper, calcium and phosphorus in the general population in
Ishikawa Prefecture. J. Kanazawa Med. Univ. 6, 26–33 (in Japanese).
Ishikawa Prefecture, 1974–1995. Environmental Report in Ishikawa Prefecture.
Department of Environment, Ishikawa Prefecture, Kanazawa (in Japanese).
Tokyo Tosho, Tokyo (in Japanese).
Ishizaki, A., Fukushima, M., Sakamoto, M., 1971. Contents of cadmium and
zinc in organs of Itai-itai disease patients and residents of Hokuriku district.
Jpn. J. Hyg. 26, 268–273 (in Japanese).
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.
K., 1988. Progress of renal dysfunction in inhabitants environmentally
exposed to cadmium. Arch. Environ. Health 43, 213–217.
K., 1989. High urinary cadmium concentration in a case of gastric cancer.
Br. J. Ind. Med. 46, 288.
Kjellstrom, T., 1986. Renal effects. In: Friberg, L., Elinder, C.G., Kjellstrom,
T., Nordberg, G.F. (Eds.), Cadmium and Health: A Toxicological and Epi-
demiological Appraisal. CRC, Boca Raton, Florida, pp. 21–109.
Kobayashi, E., 1982. An epidemiological study on the health effects of envi-
ronmental cadmium. (Part 2) Some urinary findings by residence time in
Cd-polluted area. Jpn. J. Pub. Health 29, 201–207 (in Japanese).
Kobayashi, E., Honda, R., Tsuritani, I., Nogawa, K., 1981. Urinary excretion
of protein, glucose, amino acid, praline and ?2-microglobulin in the general
population in Ishikawa Prefecture. J. Kanazawa Med. Univ. 6, 76–84 (in
Laird, N.M., Ware, J.H., 1982. Random effects models. Ann. Stat. 17, 624–642.
Nordberg, G.F., 1972. Cadmium metabolism and toxicity. Environ. Physiol.
Biochem. 2, 7–36.
Stat. Med. 13, 587–601.
Piscator, M., 1983. The progress of renal dysfunction in Cd-exposed persons,
Cadmium 83. In: Proceedings of the Fourth International Cadmium Confer-
ence, Metal Bulletin Ltd., London, pp. 157–160.
of cadmium-induced renal dysfunction in workers removed from exposure.
Scand. J. Work Environ. Health 8, 191–200.