Epidemiology in a Focus of Motor Neuron Disease on the Kii Peninsula of Japan : A Death Rate Survey

In the Hohara area of Nansei-cho, one of the foci of motor neuron disease (MND) on the Kii peninsula, the death rates of MND and of leading causes of death were surveyed to obtain a clue to the etiology of MND in the Western Pacific. The crude annual death rate of MND in the Hohara area was 47.6 per 100,000 population (for 1961-1990), and the standardized mortality ratio of MND was 5403. The death rate survey showed that the mortality of nonischemic heart disease was higher and that of diabetes mellitus was lower in Hohara as compared with the whole Japanese population . The mortality of MND has a positive correlation with the mortality of cerebrovascular diseases, and also with that of nephritis and nephrotic syndrome in Nansei-cho. These death rate data may suggest that etiological factors of MND in this focus could also influence the mortality of some other diseases. J Epidemiol, 1994; 4 : 31-35.


ORIGINAL CONTRIBUTION
Epidemiology in a Focus of Motor Neuron Disease on the Kii Peninsula of Japan : A Death Rate Survey Okujou Iwamil, Yoshiko Niki2*, Takao Watanabe3, and Masayuki Ikeda1 In the Hohara area of Nansei-cho, one of the foci of motor neuron disease (MND) on the Kii peninsula, the death rates of MND and of leading causes of death were surveyed to obtain a clue to the etiology of MND in the Western Pacific.The crude annual death rate of MND in the Hohara area was 47.6 per 100,000 population (for , and the standardized mortality ratio of MND was 5403.The death rate survey showed that the mortality of nonischemic heart disease was higher and that of diabetes mellitus was lower in Hohara as compared with the whole Japanese population .The mortality of MND has a positive correlation with the mortality of cerebrovascular diseases, and also with that of nephritis and nephrotic syndrome in Nansei-cho.
These death rate data may suggest that etiological factors of MND in this focus could also influence the mortality of some other diseases.
motor neuron disease, Kii Peninsula, underlying cause of death Three high incidence foci of motor neuron disease (MND) have been reported in the Western Pacific area, i.e., the Kii peninsula in Japan (the Hohara** and Kozagawa areas), Guam, and West New Guinea1-8).Parkinsonismdementia complex has been also reported to occur in high incidence in these areas2.3.5.7).The etiology of MND in the foci is not yet established, but it has been suggested that environmental factors play an important role in the high incidence in the foci 2.5,8).
Whatever etiological factors exist in the foci, it may be expectable that these factors influence incidences of disease other than MND and parkinsonism-dementia complex.Death rates from MND and 15 leading causes in the Hohara area were surveyed in comparison with the rates in the neighboring population and also with the general Japanese population.
The aim of this death rate survey is to ascertain whether the Hohara area is specific only for the high incidence of MND or whether other causes of death are also unusual.
A possible association among the death rates may offer a clue to the etiology of MND.

Description of study area
Hohara is located on the Kii peninsula of Japan, facing the Western Pacific Ocean (Figure 1).The Hohara area is one of five administrative areas of Nansei-cho (109 km2 in space) in the Mie prefecture (Figure 2).The population of the five areas, as of 1988, was 2731 in the Nankai area, 1662 in the Hohara area, 3793 in the Gokasho area, 949 in the Kamihara area, and 3337 in the Shukutaso area, and 12,472 for Nansei-cho as a whole.

Death rate analysis
The Ise Health Center is in charge of collecting abstracts of death certificates from Nansei-cho.Each of 3541 mortality cases in the five areas in Nansei-cho in 1961-90 was reviewed, and the underlying causes of death were re-coded according to the 9th revision of International Classification of Disease (ICD)9).The standardized mortal-O.
Iwami et al,  Nansei-cho is divided into 5 administrative areas as A) the Nankai area, B) the Hohara area, C) the Gokasho area, D) the Kamihara area, and E) the Shukutaso area y ity ratios (SMRs) were calculated for the deaths in which underlying cause of death was MND or another 15 leading causes.
A whole Japanese population of the same period

RESULTS
The SMRs are shown in Table 1.There was no death certificate that described MND as other than underlying cause of death.The crude annual death rate of MND was 47.6 per 100,000 population in the Hohara area for 1961-1990.The SMR of MND in the Hohara area was 5403, and the probability of this SMR value was as small as less than 10-30.The SMR of MND in the Gokasho area was 294, but the probability of this SMR value was insignificant (p>0.05).Among leading causes of death, SMR for non-ischemic heart diseases (191) and diabetes mellitus (0) were significantly (p<0.05)different in the Hohara area.The death rates for non-ischemic heart diseases were generally high and those for diabetes mellitus were generally low in Nansei-cho.The SMRs of some other causes, e.g.hypertensive disease, were also significant in the whole of Nansei-cho, but only the SMRs of these three diseases were significant (p<0.05) in the Hohara area.
Correlation between SMRs of the 16 causes of death were examined in five areas in Nansei-cho.The coefficients of correlation between MND and the other 15 leading causes are shown in Table 2.At the 5% level, four correlations were significant, i.e., gastroenteritis versus pneumonia and bronchitis (r=0.96,p=0.010), ulcer of stomach and duodenum versus cerebrovascular disease (r=0.88,p=0.047),MND versus nephritis and nephrotic syndrome (r=0.90, p=0.036), and MND versus cerebrovascular disease (r=0.91,p=0.033).
Parkinsonism (code 332) was also surveyed in the abstracts of death certificates regardless of underlying cause of death.Two patients with parkinson disease and one patient with secondary parkinsonism were found in death certificates in the Hohara area for 1961-1990.No patient with parkinsonism-dementia complex or presenile dementia was found in the death certificates of the Hohara area.Thus, the annual incidence of parkinson disease was 3.7 per 100,000 population in the Hohara area.

DISCUSSION
MND occurs with high incidence in the Hohara area in Nanseicho.The crude annual mortality rate was 47.6 per 100,000 population for 1961-1990, with a SMR of 5403.These values are comparable to values in other reports6-8.14).At least four cases of parkinsonism with dementia were reported in the Hohara area in 19697).Perusal of death certificate, however, showed only two cases of parkinson disease.Information on parkinsonism with dementia in Nansei-cho in a recent four year period (1987)(1988)(1989)(1990) was obtained from two neurologists and one psychiatrist in two central hospitals in Ise.There was one case of parkinsonism with dementia in the Hohara area which was not detected in this survey of death certificates probably due to incomplete certificate description.One case of parkinsonism with dementia and one case of MND in the Hohara area were alive in 1990.Accordingly, there are at least six cases (four cases in Reference 7, one dead case and one live case by medical information) of parkinsonism with dementia in the Hohara area in the period 1969 to 1990.
The death rate survey showed that not only deaths from MND but also deaths from non-ischemic heart diseases and from diabetes mellitus were significantly higher and lower, respectively, in the Hohara area than in all Japan.It was also made clear that deaths from cerebrovascular disease, and nephritis/nephrotic syndrome correlate with deaths from MND in Nansei-cho (Tables 1 and 2).
It is possible that factors which increase the incidence of non-ischemic heart diseases have some etiological relations to the factors that increase MND in the Hohara area.If this is the case, a significant correlation between the SMR for non-ischemic heart diseases and that for MND should be observed in Nansei-cho.Although the SMR for nonischemic heart diseases is highest in the Hohara area of Nansei-cho, the correlation with the SMR for MND is weak (r=0.65,p=0.237;Table 2).Thus, the factors which increase the SMR for non-ischemic heart diseases are not strong factors for MND in the Hohara area.They might be regarded as relatively weak factors, even if they had any etiological relation to MND.A similar consideration is possible for factors decreasing the SMR for diabetes mellitus.Although the SMR for diabetes mel-litus is 0 and the lowest in the Hohara area, the correlation between the SMR for MND and that for diabetes mellitus is statistically insignificant (r=-0.73,p=0.158).
As the SMRs for cerebrovascular disease and nephritis/ nephrotic syndrome correlate significantly (p<0.05) with the SMR for MND in Nansei-cho, the factors which increase deaths from cerebrovascular disease and/or nephritis/nephrotic syndrome should also increase deaths from MND in Nansei-cho.It is, however, possible that the analysis for correlations between pairs could be statistically significant just by chance.If a 5.0% significance level is employed, two or more significant correlations by chance should be observed with the possibility of 0.17 in the 15 pairs.Accordingly, the above two significant correlations might have resulted from statistical chance alone.
Some case-control studies suggest that MND is related to injuries15-17) The death rates from accidents and adverse effect were, however, not high in the Hohara area and were not related to those of MND in Nansei-cho in the present study (Tables I and 2).
In Guam, another MND focus in the Western Pacific, diabetes mellitus death rates in persons at ages of 45 years and over were two to three times greater than that in the corresponding population in the United States.Diabetesassociated deaths were highest in Chamorros who are a high-risk population of MND among the four ethnic groups in Guam18).A report19) on the comparison between Chamorros in Guam and whole Californian population described five points, i,e, 1) the rates for coronary heart disease were lower in Guam for both sexes, 2) the rates for cerebrovascular accidents were higher for Guam males, 3) the rates for hypertensive heart disease were higher in Guam for both sexes, 4) the rates for accidents and injuries were higher for Guam males, and 5) the rates for neoplastic diseases were lower in Guam for both sexes.In Guam, the incidence of MND varies depending on regions within the island1).No report is available to show the correlation of MND with other causes of death in Guam.Thus, no obvious tendency of the causes of non-MND death common to the Hohara area and Guam can be deduced at present.
Although the relation of other diseases to MND is not close enough in Nansei-cho to reach a statistically significant level (Table 2), it might be possible to postulate that some of the risk factors for non-ischemic heart disease, cerebrovascular disease, and/or nephritis/nephrotic syndrome, as well as some factors which decrease the incidence of diabetes mellitus, are also the risk factors for MND in this focus.This hypothesis might be partly explained by the metal exposure theory which postulates that low calcium, low magnesium, and high aluminum in the water and soil are MND-inducing factors in the foci5.6.20) because low calcium and low magnesium in the environment are related to higher incidence of circulatory disease21-24).Up to now, however, there are no convincing data to show that the above metals are common to the environment in the three foci and that the people there are unusually exposed to these metals.The analysis of food and water in Nansei-cho for metals is in progress in this study group, and the results are to be published elsewhere.

Figure 1 .
Figure 1.Map to show the location of the Hohara area in Japan.The Hohara area is located on the Kii peninsula of Honshu Island of Japan facing the Western Pacific Ocean.

Figure 2 .
Figure 2. The Hohara area in Nansei-cho.Nansei-cho is divided into 5 administrative areas as A) the Nankai area, B) the Hohara area, C) the Gokasho area, D) the Kamihara area, and E) the Shukutaso area was taken as the standard population, and the mortality data were obtained from Vital Statistics of Japan without mention of pychosis (797), and MND (335.2).The validity of death certificates for an epidemiological study of MND has been described in a previous study11).For the period of 1961-1979, the codes of the 7th and 8th revisions which corresponds to those of 9th were used in the standard population.Although MND is coded as 335.2 in the 9th revision, data for code 335.2 was not available in the Vital Statistics of Japan.Accordingly, anterior horn cell disease (code 335) was taken to substitute MND in the standard population, because other anterior horn cell diseases are rare as compared with MND (e.g., the rate of deaths from total anterior horn cell disease to that from MND in 1978 was 53 to 5112)).Code 356 of 7th revision in 1961-67, and code 348 and 330.1 of 8th revision in 1968-78 were used in place of MND in the standard population.Population data in Nanseicho were obtained from the Nansei Public Office.The populations in years1965, 1975, 1985were used for calculation of SMRs for 1961-1990.Data on the standard population were obtained from the Vital Statistics of Japan10) For any cause of death, the observed numbers were treated as a Poisson variable.Statistical analysis of SMRs was by the method of mathematical link between Poisson and the chi-squares distributions13).