Seasonal Variation in Mortality from Ischemic Heart Disease and Cerebrovascular Disease in Okinawa and Osaka: the Possible Role of Air Temperature

Okinawa is located in a subtropical area and is well-known for low mortality due to ischemic heart disease (IHD) and cerebrovascular disease (CVD). However, the factors that contribute to these low mortality rates remain unclear. We examined the seasonal variation in the mortality due to IHD and CVD among Okinawa and Osaka residents, aged 45 to 84 years, between 1992 and 1996. In addition, we studied if there was a relationship between the monthly mortality rate from IHD or CVD and the monthly mean daily air temperature in Naha City and Osaka City. Data on the monthly mean daily air temperature was obtained from the meteorological stations in Naha City and Osaka City. Our results showed that there were inverse correlations between the monthly mean daily temperature in a city and each of the monthly mortality from IHD in Okinawa (r=-0.794, p<0.01), the monthly mortality from CVD in Okinawa (r=-0.837, p<0.001), and the monthly mortality from CVD in Osaka (r=-0.954, p<0.001). In Osaka, the monthly mortality rate from IHD was at or near its minimum value when the mean daily temperature was approximately 25 (in September), and it increased in a linear fashion as the mean monthly temperature fell (r=-0.975, p<0.001). The difference between the monthly mortality from IHD or CVD among the Okinawa and the Osaka residents increased in the winter season in comparison with that in the other seasons, with the exception for IHD in July and in August. These findings indicate that the lower mortality from IHD and CVD in Okinawa is affected, at least in part, by Okinawa's warm winter.


INTRODUCTION
Okinawa is located in a subtropical area between the East China Sea and the Pacific Ocean, and is approximately 1000 kilometers from Honshu Island.Okinawa is well-known for its low mortality from ischemic heart disease (IHD) and cerebrovascular diseases (CVD).In 1995, the age-standardized mortality (per 105, based on the standard population in Japan in 1985) from IHD in Okinawa was 50 among males and 23 among females, which is 12% and 24% lower, respectively, thanthose among.the Japanese population 1).The mortality from CVD in Okinawa was 87 among males and 57 among females, which are 30% and 38% lower, respectively, than those among the Japanese population 1).However, the reasons for the lower mortalities from IHD and CVD in Okinawa remain unclear.
According to recent nutrition surveys performed in Japan, the proportion of people with either hypertension or borderline hypertension, hypercholesterolemia, or who smoke in Okinawa does not differ from the respective proportion among the Japanese people as a whole 2,3).
Seasonal variation in mortality due to IHD and/or CVD has been noted in both the northern4-7) and southern 8,9) hemispheres, with higher death rates during the winter than in the summer.Twenty years ago, the seasonal pattern of CVD

METHODS
Okinawa Prefecture (1995 population: 1.3 million) consists of the Ryukyu-archipelago including Okinawa Island, which is located between latitudes 26 and 27 degrees north.Approximately 92% of the Okinawa people reside on Okinawa Island.Osaka Prefecture (1995 population: 8.8 million) lies between latitudes 34 and 35 degrees north.The metropolitan areas on Honshu Island are located around this latitude.We obtained data from the annual reports of the Meteorological Agency of Japan, on the monthly mean 24-hour air temperature in each month of the years between 1992 and 1996 that was recorded by the meteorological station in Naha City located on Okinawa Island, and the meteorological station in Osaka City.The monthly values in each year between 1992 and 1996 were averaged to obtain the mean daily temperature in each month of the year in each prefecture.
Cause-specific mortality figures were obtained from the annual statistics prepared by Okinawa and Osaka Prefectures, and accumulated by 5-year age and month of death between 1992 and 1996.The underlying cause of death was defined according to the International Classification of Diseases (ICD), 9th revision for 1992-94, and 10th revision for 1995-96, according to the coding rule prepared by the World Health Organization 17).We defined IHD as 410-414 in the ICD 9th revision and as 120-125 in the ICD 10th revision.We defined CVD as 430-438 in the ICD 9th revision, and as 160-167 and 169 in the ICD 10th revision.Consistency was maintained between the 9th and 10th ICD series by excluding ICD 168 from the 10th series, which is the code for cerebrovascular disorders in diseases classified elsewhere after 1995.The agestandardized monthly mortality rate due to IHD or CVD was   (1.94), using the rate in Okinawa as a reference (Figure 4).The monthly mortality from CVD was significantly higher in Osaka than in Okinawa throughout the year, with a rate ratio ranging between 1.20 and 1.40 (Figure 5).During the year, the monthly mortality rate ratios from CVD were higher from October to February (1.35 to 1.40)

DISCUSSION
Our results demonstrate that in the 45-84-year age group, the monthly mortalities from IHD and CVD in Okinawa were inversely correlated with the monthly daily temperature.These findings are consistent with reports of seasonal variation in cardiovascular mortality in regions with a warm winter such as Athens 7), Hawaii 18), and Hong Kong 19).The data suggest that even a relatively small seasonal change in temperature has a significant effect on mortality from coronary artery disease and cerebrovascular disease among middle-aged and elderly people.Donaldson and Keatinge 20) demonstrated in the population of southeast England that the mortalities lagged a temperature change by 3 days for IHD and 12 days for respiratory diseases.This finding and those of the seasonal analysis indicate that cold exposure can cause rapid IHD death by inducing an increase in arterial pressure and thrombosis possibly due to haemoconcentration 11,12).Further IHD deaths can be expected secondary to respiratory diseases, since the incidence of respiratory disease increases in cold weather 7,21), and since influenza 22,23) and other respiratory infections u 24, 25) are reported to be accompanied by increased IHD mortality.
Other seasonal factors such as sunlight exposure 18,19) Jan Feb March April May June July Aug Sep Oct Nov Dec Figure 5.The age-standardized rate ratios of monthly CVD mortality (Osaka/Okinawa) among 45-to-84year-old individuals between 1992-1996.
the winter season, with the exception of IHD in July and in August.This finding, together with the correlation analysis, indicates that the lower mortality from IHD and CVD in Okinawa may be affected by Okinawa's warm winter.However, our data also showed that the monthly mortality from both IHD and CVD in Okinawa was always lower than that in Osaka in the same temperature range.This finding suggests the existence of other substantial causal factors inducing the different mortalities we observed.Further epidemiological studies are required to clarify the factors that contribute to the lower IHD and CVD mortalities in Okinawa, focusing on life styles at the individual level.

Figure 2 .
Figure 2. Relationship between the monthly age-standardized mortality rate from IHD in individuals between 45-84 years and the average monthly mean daily air temperature between 1992-1996 in Okinawa and Osaka.Correlation analysis in Osaka was limited to between September and June because of its linearity.The age-standardized mortality rate from IHD was inversely correlated with the monthly mean daily air temperature in Osaka and in Okinawa.
, humidity 10), and maximum and minimum temperatures 26), that have been assumed to affect resulting in the occurrence of RID.Reports from New Orleans 27) and Kuwait 28) support this hypothesis, in which they observed a higher mortality rate from coronary artery disease during hot weather than during cool weather.Our results demonstrate that the difference between the mortality from 1HD or CVD in Okinawa and Osaka increased in Cerebrovascular Disease