Weather and Stroke in a Subtropical Area: Ilan, Taiwan
Background and Purpose The aim of this study was to clarify the association of weather and stroke occurrence in a subtropical area in Ilan, Taiwan.
Methods We studied 517 patients with stroke (date of onset, January 1 to December 31, 1991) in Poh-Ai Hospital and St Mary’s Hospital: 316 patients (61.1%) with cerebral infarction, 170 (32.9%) with intracerebral hemorrhage, and 31 (6.0%) with other types of stroke. The daily occurrence of cerebral infarction and intracerebral hemorrhage was analyzed and correlated with three major meteorologic factors: air temperature, air pressure, and relative humidity.
Results The occurrence of cerebral infarction was rather uniform in all kinds of weather. The occurrence of intracerebral hemorrhage was approximately twice as great on cold days (0.71 case per day) and high-pressure days (0.73) as on warm days (0.31) and low-pressure days (0.39) (P<.005 by χ2 test). When a linear regression model was used to test whether air temperature or air pressure had more influence on intracerebral hemorrhage, only air temperature showed a significant effect. Regarding intracerebral hemorrhage, the relative risks of cooler and median temperature days versus warmer days were 18.5 and 5.1, respectively.
Conclusions Intracerebral hemorrhage but not cerebral infarction occurs more frequently on cooler days, with a dose-response relationship, in Ilan, Taiwan.
The influence of weather on stroke is very impressive to many clinicians. It has been discussed since ancient times in both oriental and western traditional medicine1 2 but is rarely mentioned in modern textbooks on stroke.3 In recent decades, although many quantitative studies on the relationship between meteorologic factors and stroke have been performed, their conclusions seem rather inconsistent.4 5 6 7 8 9 10 11 12 13 14 15 16 Despite these inconsistent conclusions, many clinicians have the impression that stroke occurs in groups, ie, on some days many patients with stroke are admitted and on other days there are no stroke admissions. We conjecture that this occurrence of stroke in groups might be influenced by weather. To verify our conjecture, we analyzed daily stroke occurrence in a regional area to clarify which weather factors affect stroke and which type of stroke is affected most by weather.
Subjects and Methods
Ilan County, located in northeast Taiwan, is a relatively isolated area surrounded by high mountains and the Pacific Ocean. Although it has an area of approximately 2137 km2, most of its 450 000 inhabitants live in a plain of approximately 500 km2 with rather uniform weather. Ilan plain (latitude, 24°40′ north; longitude, 121°40′ east) has a subtropical monsoon climate and an average annual rainfall of more than 250 cm. In 1991 the mean annual temperature (±SD) was 22.3±5.0°C, mean annual relative humidity was 84±7.2%, and mean annual air pressure was 1015.7±6.9 millibars. To clarify how weather affects stroke, we compared daily stroke occurrences under different meteorologic conditions, based on the date of stroke onset. Three major meteorologic factors—daily air temperature, air pressure, and relative humidity—were used as variables. According to the daily mean values of temperature, pressure, and relative humidity, all 365 days in a year were divided into three groups by annual mean values and 1 SD, as follows: temperature, 62 cooler days (daily mean temperature, <17.3°C), 104 warmer days (daily mean temperature, >27.3°C), and 199 days of median temperature (daily mean temperature, 17.3°C to 27.3°C); air pressure, 67 higher-pressure days (daily mean pressure, >1022.6 millibars), 61 lower-pressure days (daily mean pressure, <1008.8 millibars), and 237 days of median pressure (daily mean pressure, 1008.8 to 1022.6 millibars); and relative humidity, 56 drier days (daily mean humidity, <76.6%), 62 wetter days (daily mean humidity, >91.1%), and 247 days of median humidity (daily mean humidity, 76.6% to 91.1%). The meteorologic data were recorded at Ilan Weather Observatory Station, which is located at the center of Ilan plain.
Lotung St Mary’s Hospital and Poh-Ai Hospital are the only two general hospitals and referral hospitals for stroke patients in Ilan County. Because of the region’s geographic isolation, stroke patients who live in Ilan are rarely admitted to hospitals outside Ilan County. Most stroke patients are admitted to these two hospitals in a timely manner; in this study 74% of patients were admitted within 24 hours. The estimation of the onset of stroke was reliable and precise.
Stroke was defined as rapidly developed clinical signs of local or global disturbance of cerebral function lasting for more than 24 hours or leading to death before 24 hours.17 Causes other than vascular lesions and cases of transient ischemic attack were excluded by clinical and head computed tomography (CT) findings. Strokes were further classified into cerebral infarction (CI), intracerebral hemorrhage (ICH), subarachnoid hemorrhage (SAH), and other types. Head CT was performed in 99% of all cases, which was helpful for the differential diagnosis of type of stroke.
There were 519 patients with stroke onset admitted to these two hospitals in 1991: 318 (61.27%) with CI, 170 (32.76%) with ICH, 23 (4.43%) with SAH, and 8 (1.54%) with other types. The details of case ascertainment are included in our other report.18 SAH and other types of stroke were excluded from further analysis because the number of patients was too small. Two patients with CI with uncertain date of onset were also excluded.
We used a χ2 test to compare whether the number of daily stroke occurrences differed significantly under various weather conditions. Meteorologic factors shown to be significant by χ2 test were placed into a linear regression model to further test their independent effect on stroke. For calculating relative risk, the logarithmic daily occurrence rate was used as a dependent variable. Risk factors of stroke such as age, sex, hypertension, diabetes mellitus, and hyperlipidemia were compared on days with different temperatures by Student’s t test and χ2 test. Probability values were calculated by a two-tailed test; P<.05 was considered statistically significant.
The number of patients, mean age, and sex ratio of patients with CI and ICH are shown in Table 1⇓.
We recorded 62 cooler days, 104 warmer days, and 199 days of median temperature. As shown in Table 2⇓, ICH occurred more frequently than expected on cooler days. This difference was significant by χ2 test (P<.005). The average daily occurrence of ICH was 0.71 (case per day) on cooler days and only 0.31 on warmer days.
We recorded 67 higher-pressure days, 61 lower-pressure days, and 237 days of median pressure. The daily occurrence of ICH significantly increased on the higher-pressure days (0.73 case per day) compared with lower-pressure days (0.39), as shown in Table 3⇓ (P<.005 by χ2 test). CI increased slightly but not significantly on lower-pressure days.
We recorded 56 drier days, 62 wetter days, and 247 days of median humidity. There was no significant difference in type of stroke occurrence in regard to different levels of humidity, as shown in Table 4⇓.
Independent Effect of Air Temperature and Air Pressure
In Ilan, air temperature and pressure have a strong negative correlation (correlation coefficient, −.89). A linear regression model, with logarithmic daily stroke occurrence rate as the dependent variable and the three categories of air temperature and pressure as independent variables, was used to clarify whether air temperature or air pressure exerts more of an influence on ICH. The relative risks of daily occurrence of ICH are shown in Table 5⇓. The relative risks for cooler and median temperatures compared with warmer temperatures in regard to the occurrence of ICH were 18.5 and 5.1, respectively. Air pressure showed no significant effect on the occurrence of ICH.
Other Risk Factors for Stroke
Mean age, sex, and presence of hypertension, diabetes mellitus, hyperlipidemia, and heart disease showed no significant differences in patients with ICH on days with different temperatures, as shown in Table 6⇓.
This study reveals that only ICH is significantly affected by air temperature. The occurrence of CI is rather uniform under various weather conditions. This study has two distinguishing features. First, the relationship between stroke onset and weather change was clear because the study was conducted in a small and rather isolated area with uniform weather, and most patients were admitted within 24 hours. Several meteorologic studies use stroke mortality and not stroke onset as a variable.7 9 11 12 Because most patients with stroke die several days or months after stroke onset, usually of extracranial causes,19 the conclusions drawn from these studies regarding the association between weather and stroke are rather vague. Second, the proportion of ICH is two to four times higher in Taiwan than in European and North American countries.20 21 22 This provides us with an opportunity to determine the effect of weather on different types of stroke and to infer the pathogenesis from our observations.
However, generalizations drawn from our results also have two potential sources of error. First, our data are based on hospitalized patients and not on the general population. Some stroke patients might not have been admitted to the hospitals because they died before admission or because their symptoms were minor.18 Second, geographic, ethnic, and social factors may affect the influence of weather on people. When we reviewed previous reports on the association between stroke and weather, we noted many inconsistent results. When the type of stroke was not specified, increases in stroke incidence in cold weather were reported in Sarajevo (Yugoslavia),4 Nottingham (England),8 and Melbourne (Australia).9 However, in the Negev Desert (Israel),11 the incidence of stroke increased on warmer days. When the type of stroke was taken into account, an increased incidence of ICH was found in Minnesota,6 Brussels (Belgium),13 Hisayama (Japan),14 and Iowa.15 However, a report from the Lehigh Valley (Pennsylvania)16 did not find a seasonal variation in incidence of ICH. An increased incidence of CI in the winter was reported in Hisayama, but in Iowa CI increases in the warmer months, and in the Lehigh Valley the peak months of CI were February through April.
These discrepancies may be due to the peculiar physiological responses of different people acclimating to different geographic environments. Using the hypothesis that people will acclimate to the median weather condition in their living environment, we divided the three meteorologic factors into median, lower, and higher categories. The higher and lower values will be a source of stress for people with potential cardiovascular disorders. The higher or lower value is relative. For example, cooler days in Ilan are not as cold as in Japan or North America, but the effect of stress may be similar. We also noted a skew of the temperature distribution toward the warmer days in Ilan, with 102 warmer days and only 62 cooler days in 1991. The fewer cooler days would therefore be more stressful. In addition, living habits may also affect the influence of weather on stroke. In Ilan, almost all houses have electric fan and/or cooler equipment for hot days, but most houses do not have heaters for cold weather. This may also account for the fact that cooler days were more stressful to our patients than warmer days.
How cold weather increases hemorrhagic stroke is not clear. A physiological study of cold stress in healthy subjects showed that 6 hours of mild surface cooling in moving air at 24°C increased platelet and red cell count, blood viscosity, and arterial pressure.23 The elevated blood pressure may account for the increased ICH on cooler days. Higher arterial blood pressure in the winter or spring has been reported, but the causal relationship was not clear.24 25
In summary, when comparing previous reports on the association of weather and stroke, the influence of weather on stroke is not universal in different geographic areas. In our study, conducted in a subtropical rainy area, only ICH occurs more frequently on cooler days, and it does so with a dose-response relationship. Elevated blood pressure may account for the increased incidence of ICH in cold weather. Further epidemiological and physiological studies are needed to clarify the influence of weather on stroke.
- Received September 13, 1994.
- Revision received December 15, 1994.
- Accepted January 3, 1995.
- Copyright © 1995 by American Heart Association
Shon LP. Pathophysiology of Chinese Medicine [in Chinese]. Taipei, Taiwan: Chi-Ye; 1988:89-99.
Frederick S. Hippocratic Heritage: A History of Ideas About Weather and Human Health. New York, NY: Pergamon Press, Inc; 1982.
Marquardsen J. Epidemiology of stroke in Europe. In: Barnett HJM, Stein BM, Mohr JP, Yatsu FM, eds. Stroke: Pathophysiology, Diagnosis, and Management. New York, NY: Churchill Livingstone, Inc; 1986:36.
Bokonjic R, Zec N. Stroke and the weather: a quantitative statistical study. J Neurol Sci. 1968:483-491.
Lejeune J-P, Vinchon M, Amouyel P, Escartin T, Escartin D, Christaens J-L. Association of occurrence of aneurysmal bleeding with meteorologic variations in the north of France. Stroke. 1994;25:338-341.
Barer D, Ebrahim S, Smith C. Factors affecting day to day incidence of stroke in Nottingham. Br Med J. 1984;289:662.
Christie D. Stroke in Melbourne, Australia: an epidemiological study. Stroke. 1981;12:467-469.
Berginer VM, Goldsmith J, Batz U, Vardi H, Sapiro Y. Clustering of stroke in association with meteorologic factors in Negev Desert of Israel: 1981-1983. Stroke. 1989;20:65-69.
Knox EG. Meteorological associations of cerebrovascular disease mortality in England and Wales. J Epidemiol Community Health. 1981;35:220-223.
Capon A, Demeurisse G, Zheng L. Seasonal variation of cerebral hemorrhage in 236 consecutive cases in Brussels. Stroke. 1992;23:24-27.
Shinkawa A, Ueda K, Hasuo Y, Kiyohara Y, Fujishima M. Seasonal variation in stroke incidence in Hisayama, Japan. Stroke. 1990;21:1262-1267.
Sobel E, Zhang ZX, Alter M, Lai SM, Davanipour Z, Friday G, McCoy R, Isack T, Levitt L. Stroke in the Lehigh Valley: seasonal variation in incidence rates. Stroke. 1987;18:38-42.
World Health Organization. Cerebrovascular diseases: prevention, treatment, and rehabilitation: report of a WHO meeting. World Health Organ Tech Rep Ser No. 469, 1971.
Chang SF, Su CL, Chen ZY, Hung TP. Stroke incidence in Ilan, Taiwan: a community hospital-based registry. J Formos Med Assoc. In press.
Bounds JV, Wiebers DO, Whisnant JP, Okazaki H. Mechanisms and timing of deaths from cerebral infarction. Stroke. 1981;12:474-477.
Jernotorp P, Berglund G. Stroke registry in Malmo, Sweden. Stroke. 1992;23:357-361.
Wolf PA, Kannel WB, McGee DL. Epidemiology of strokes in North America. In: Barnett HJM, Stein BM, Mohr JP, Yatsu FM, eds. Stroke: Pathophysiology, Diagnosis, and Management. New York, NY: Churchill Livingstone, Inc; 1986:19-29.
Keatinge WR, Coleshaw SRK, Cotter F, Mattock MB, Murphy M, Chelliah R. Increases in platelet and red cell counts, blood viscosity, and arterial pressure during mild surface cooling: factors in mortality from coronary and cerebral thrombosis in winter. Br Med J. 1984;289:1405-1408.
Brennan PJ, Greenberg G, Miall WE, Thompson SG. Seasonal variation in arterial blood pressure. Br Med J. 1982;285:919-924.