Usami T, Koyama K, Takeuchi O, Morozumi K, Kimura G. Regional Variations in the Incidence of End-Stage Renal Failure in Japan. JAMA. 2000;284(20):2622-2624. doi:10.1001/jama.284.20.2622
Author Affiliations: Department of Internal Medicine and Pathophysiology, Nagoya City University Medical School, Nagoya, Japan.
Context Despite recent medical advances, the number of patients beginning dialysis
annually is increasing in both the United States and Japan. The ethnically
homogeneous population of Japan presents an opportunity to study the presence
of factors other than race/ethnicity that might contribute to incidence of
end-stage renal disease (ESRD).
Objective To determine if and where regional differences exist in ESRD in Japan.
Design, Setting, and Subjects Analysis of data reported by the Japanese Society for Dialysis Therapy
based on the annual number of patients with ESRD beginning maintenance dialysis
therapy in all 47 prefectures of Japan from 1982 to 1998.
Main Outcome Measures Mean annual ESRD incidence and increasing rate of ESRD in each of 11
predefined areas making up the entire country.
Results Incidence of ESRD increased approximately 3-fold in Japan during the
study years, from 81.3 per 1 million in 1982 to 237.6 per 1 million in 1998.
Significant regional differences were found in both measures. The mean (SEM)
annual ESRD incidence (P<.01) and increasing rate
of ESRD (P<.01), respectively, were significantly
different across Japan. Koshinetsu (140  per 1 million and 9.1 [0.6] per
1 million/y) and Hokuriku (141  per 1 million and 9.7 [0.5] per 1 million/y)
were the areas with the lowest incidence and increasing rate of incidence,
while Okinawa (188  per 1 million and 13.4 [0.6] per 1 million/y) and
Kyushu (179  per 1 million and 12.0 [0.6] per 1 million/y) were the areas
with the highest incidence and increasing rate of incidence.
Conclusions We found definite and significant regional differences in incidence
and increasing rate of incidence of ESRD in Japan. Further analyses are needed
to identify factors that contribute to these regional differences and thereby
improve strategies for treatment of renal disease.
Despite recent advances in nephrology and dramatic decreases in the
incidence of cardiovascular diseases,1 the
number of patients beginning dialysis therapy annually is increasing in both
the United States2 and Japan.3
In the United States, racial differences in the incidence of end-stage renal
disease (ESRD) have been noted.4,5
On the other hand, Japan has a relatively homogeneous racial composition.
Therefore, we constructed maps to compare the annual incidence of ESRD and
the increasing rate of ESRD incidence among different areas in Japan. Regional
differences in Japan may suggest the presence of factors other than race that
contribute to differences in incidence and increasing rate of incidence and
that may be controllable by treatment.
Based on the numbers of patients with ESRD beginning maintenance dialysis
therapy (both hemodialysis and peritoneal dialysis) annually in the 17-year
period from 1982 to 1998 reported as an overview of regular dialysis treatment
in Japan by the Japanese Society for Dialysis Therapy,3
we calculated the mean annual ESRD incidence and the increasing rate of ESRD
incidence in each area for 11 areas comprising the entire country to construct
maps on renal failure in Japan.
The 47 prefectures of Japan were organized into 11 areas that have internally
homogeneous cultural and socioeconomic activities that differ between regions.
Incorporation of data based on prefecture into larger units based on area
should decrease errors due to the influx and efflux of the population that
occur across the prefecture lines but that is mostly restricted to within
a given area. Division into 11 areas (Hokkaido, Tohoku, Kanto, Koshinetsu,
Hokuriku, Tokai, Kinki, Chyugoku, Shikoku, Kyushu, and Okinawa) was based
on conventionally used and widely accepted definitions.
The mean annual incidence of ESRD in each area (per population of 1
million) was calculated as the average in the area for the 17-year period
from 1982 to 1998 of the number of patients with ESRD annually beginning dialysis
therapy in prefectures3 and corrected for population
in prefectures in each year. The increasing rate of ESRD incidence in each
area (per population of 1 million per year) was calculated as the slope of
regression lines between the annual incidence of ESRD in the area corrected
for population and the year during 17 years.
We examined factors that might affect regional differences in ESRD dynamics.
These were estimated from data based on prefectures for each area and included
the average age of patients with ESRD entering dialysis therapy in 1998,3 the average percentage of people older than 65 years
in 1998,6 the cost of medical care per 1 person
in 1995,7 the number of nephrologists per general
population in 1997,8 and the number of hospital
beds available for dialysis therapy in 1997.9
One-way repeated-measures analysis of variance (ANOVA) was used to compare
the mean annual incidence of ESRD among areas, while 1-way ANOVA was used
for the increasing rate of ESRD incidence, followed by the Newman-Keuls multiple
comparison test. Pearson product moment correlation was used to examine the
correlation among areas. Numeric data were expressed as the mean (SEM), and P<.05 was considered statistically significant.
The number of patients with ESRD newly beginning maintenance dialysis
in Japan increased from 4652 in 1982 to 30,051 in 1998, an approximately 3-fold
increase in incidence from 81.3 to 237.6 per 1 million population. The average
age of patients was 50.8 years in 1982 and 62.7 years in 1998. The major causes
of ESRD were chronic glomerulonephritis (59.6% and 35.0% in 1982 and 1998,
respectively) and diabetic nephropathy (13.7% in 1982 and 35.7% in 1998).
The mean annual ESRD incidence and the increasing rate of ESRD incidence
in each area between 1982 and 1998 are shown in Table 1. Differences in the 2 measures among 11 areas were significant
(P<.01) based on ANOVA.
To construct the maps (Figure 1)
of renal failure in Japan, areas were classified into 3 categories: the lowest
3, the highest 3, and other intermediate areas, for incidence and increasing
rate of ESRD, respectively. The maps illustrate that the mean annual incidence
and the increasing rate of incidence were both low in Koshinetsu and Hokuriku,
but were high in Okinawa and Kyushu. Figure
2 shows the comparison of the annual incidence of ESRD from 1982
to 1998 among these 4 areas. There were significant differences between the
2 regions with the lowest incidence and increasing rates (Koshinetsu and Hokuriku)
and 2 regions with the highest (Okinawa and Kyushu) for both the mean annual
incidence (P<.001) and the increasing rate (P<.05).
The number of hospital beds for dialysis therapy was significantly related
to both the mean annual incidence (r = 0.92; P<.001) and the increasing rate (r = 0.67; P = .02). None of the other factors
examined, such as average age of patients, percentage of the population older
than 65 years, cost of medical care, or the number of nephrologists was correlated
with the mean annual incidence and increasing rate of incidence.
Our data indicate the presence of clear regional differences within
Japan in both the annual incidence of patients with ESRD beginning maintenance
dialysis therapy and the increasing rate of ESRD incidence from 1982 to 1998.
For example, the mean annual incidence and the increasing rate were both lowest
in Koshinetsu, with annual incidence 1.3 times lower than that in Okinawa
and the increasing rate 1.5 times lower than in Hokkaido. The Japanese Society
for Dialysis Therapy3 has reported the absolute
number of patients beginning dialysis therapy in each prefecture annually
for 17 years without correcting for population. Therefore, the regional differences
presented here have not been previously noted.
To our knowledge, this is the first study demonstrating the regional
differences in ESRD dynamics within a relatively homogenous national population.
Although studies have shown that hypertension and renal failure are more common
among blacks than whites in the United States,4,5
no regional differences have been reported within a relatively homogeneous
national population. Incidence of hypertension10- 13
and stroke13- 15
are known to vary by region in Japan, both being high in northern Japan and
low in the southern part of the country. Association between hypertension
and the amount of salt intake, proven in Japan,10- 13
is well known.12,13,16,17
Once regional differences in ESRD dynamics have been established, the
factors affecting the differences must be elucidated. If we can identify such
factors, it may be possible to improve strategies to prevent renal failure.
We analyzed several factors to determine whether they were related to the
regional differences in ESRD dynamics and found that only the number of beds
available for dialysis therapy was correlated with regional differences. However,
this is probably the result of increased numbers of patients with ESRD, rather
than a causative factor. The dynamics of ESRD are determined by both the incidence
of nephropathy and the progression rate. Lack of clear evidence concerning
regional differences in the incidence of glomerulonephritis and diabetic nephropathy
within Japan may suggest that the progression rate of nephropathy differs
among geographic areas, but the current evidence does not allow us to reach
Recently, it has been noted that despite dramatic decreases in the incidences
of stroke and ischemic heart disease, the incidence of ESRD increased by almost
3-fold in the 14-year period between 1982 and 1995 in the United States.1,2 Our data show a similar trend in Japan,
with the incidence of ESRD increasing approximately 3-fold in the 17-year
period between 1982 and 1998. Therefore, it is extremely important to determine
the risk factors, especially modifiable risk factors, leading to ESRD. Angiotensin-converting
enzyme inhibitor is now considered 1 of the most promising interventions for
arresting renal failure.18- 20
If the factors contributing to the regional differences in ESRD dynamics can
be identified, new strategies for treatment of renal disease should become
available. Thus, further studies of regional differences in ESRD dynamics
are needed. We hope that the present study will stimulate such analyses because
the regional differences in ESRD dynamics appear definite and significant.