Seasonal Variation of Stroke Incidence in Japan for Stroke Patients in the Japanese Standard Stroke Registry,

Transcription

1 Seasonal Variation of Stroke Incidence in Japan for Stroke Patients in the Japanese Standard Stroke Registry, Shunya Takizawa, MD, PhD,* Takeo Shibata, PhD, Shigeharu Takagi, MD, PhD,* and Shotai Kobayashi, MD, PhD for the Japan Standard Stroke Registry Study Group Background: Seasonal variation of stroke incidence has been reported in many countries. The present study was designed to elucidate seasonal and monthly variations in the incidence of subtypes of acute ischemic stroke and hypertensive hemorrhagic stroke using the Japanese Standard Stroke Registry Study (JSSRS) database, which is currently the world s largest hospital-based stroke database, accumulating records from 163 Japanese institutions. Methods: Among 47,782 patients with acute stroke registered with JSSRS between 1998 and 2007, we selected 35,631 for analysis (patients with ischemic or hemorrhagic stroke of unknown etiology were excluded). A simple moving average was used to examine monthly variation of stroke incidence. We also examined seasonal variation of ischemic stroke subtypes. Results and Conclusions: Monthly variation in incidence of all ischemic stroke was significant (P,.001). Noncardioembolic ischemic stroke was more frequent in summer than in winter (P,.001). Lacunar stroke showed higher incidence in summer than in winter (P,.001), although the increase did not reach significance for atherothrombotic stroke (P 5.057). In contrast, cardioembolic stroke (P,.001) and hemorrhagic stroke (P,.001) occurred more frequently in winter than in summer. Hemorrhagic stroke showed a regional difference of incidence between northern and southern Japan. There is a temporal variation of stroke incidence in Japan, with different patterns of variation depending on stroke subtype. These findings may help in developing strategies for preventing stroke. Key Words: Atherothrombolic stroke cardioembolic stroke hemorrhagic stroke lacunar stroke seasonal variation stroke. Ó 2013 by National Stroke Association Seasonal differences of hospital- or community-based stroke incidence have been examined in many countries, 1-15 including Japan However, the results show considerable variability. Many studies have concluded that stroke occurs most frequently in winter. 1-3,5,8-13 In Japan, Suzuki et al 11 and Shinkawa et al 12 found the From the *Department of Neurology; Department of Medical Informatics, Tokai University School of Medicine, Kanagawa; and Director of Shimane University Hospital, Shimane, Japan. Received April 8, 2011; accepted June 1, Address correspondence to Shunya Takizawa, MD, PhD, Department of Neurology, Tokai University School of Medicine, 143 Shimokasuya, Isehara, Kanagawa, shun@is.icc. u-tokai.ac.jp /$ - see front matter Ó 2013 by National Stroke Association doi: /j.jstrokecerebrovasdis same tendency, but a recent report from Takashima Stroke Registry 14 found that stroke occurs most frequently in spring, followed closely by winter. On the other hand, the Japan Multicenter Stroke Investigators Collaboration 15 reported that stroke occurs least frequently in spring, followed by winter, fall, and summer. These discrepancies may be partly related to differences in ambient temperature, 2 atmospheric pressure, 16 and platelet aggregation. 17 Here, we describe seasonal and monthly variations of stroke using the Japanese Standard Stroke Registry Study (JSSRS) database. 18 Between December 1998 and November 2007, JSSRS accumulated 47,782 patients with stroke in 163 institutions (see Appendix A) throughout Japan; as the largest hospital-based stroke registration database in the world at present, it offers a unique opportunity to obtain reliable data on seasonal variations in stroke 36 Journal of Stroke and Cerebrovascular Diseases, Vol. 22, No. 1 (January), 2013: pp 36-41

2 SEASONAL VARIATION OF STROKE IN JAPAN 37 incidence. Further, because stroke subtype is available for most JSSRS-registered patients, we also evaluated seasonal variation in lacunar, atherothrombotic, and cardioembolic stroke. Third, because the Japanese islands extend over 3000 kilometers, we also evaluated the difference of stroke incidence between northern and southern Japan. Methods To examine seasonal and monthly variation of stroke, we analyzed data for 47,782 patients diagnosed with cerebral infarction (atherothrombotic, lacunar, cardioembolic stroke, and others), transient ischemic attack, cerebral hemorrhage, subarachnoid hemorrhage, and others who were hospitalized within 7 days after symptom onset who were registered in JSSRS from December 1998 to November Atherothrombotic, lacunar, or cardioembolic subtype was determined from the patients neurologic, radiologic, cardiologic, and hematologic profiles, principally according to the classification of cerebrovascular diseases III by the National Institute of Neurological Disorders and Stroke. 19 To clarify the seasonal incidence of each subtype of ischemic stroke or hypertensive hemorrhagic stroke, we excluded ischemic or hemorrhagic stroke of unknown cause or etiology and subarachnoid hemorrhage. The monthly number of patients was corrected for a standard month of 30 days (for months with 31 days, the actual number was multiplied by 30/31; for February, it was multiplied by a factor of [30 3 9]/[{28 3 7} 1 {29 3 2}] to take account of the 2 leap years). A simple moving average, calculated by averaging the number of patients in the month of interest and the previous and following months, was used to smooth random variation. To examine seasonal variation in the incidence of ischemic and nonischemic stroke patients, months were grouped into 4 seasons: spring (March-May), summer (June-August), fall (September-November), and winter (December-February). To analyze regional difference in noncardioembolic and hypertensive hemorrhagic stroke incidence, we divided Japan into northern and southern regions, taking the Kanto district as the dividing line. Because the numbers of patients in the northern and southern regions differed, we used the moving average of monthly patients calculated as a percentage of total annual incidence of noncardioembolic or hemorrhagic stroke. To examine the relation between environmental temperature and stroke incidence, we obtained regional monthly temperature data for 1998 to 2007 from the Japan Meteorological Agency. The Chi-square test was used to analyze the seasonal and monthly variations of stroke incidence. SPSS software (version 16; SPSS, Inc, Chicago, IL) was used for statistical analysis, and P,.05 was considered statistically significant. Kruskal Wallis tests were used to evaluate the significance of age in each subgroup. Results Table 1 shows the characteristics of registered stroke patients. Among the 47,782 patients registered with JSSRS between 1998 and 2007, we excluded those with ischemic stroke of unknown cause (2426 patients) and hemorrhagic stroke of unknown etiology (1272 patients). We enrolled 35,631 patients ( years of age; 21,302 men and 14,329 women) comprised of 29,238 ischemic stroke patients ( years of age) and 6393 hypertensive hemorrhagic stroke patients ( years of age). Ischemic stroke was classified into 3 subtypes: atherothrombotic (9227 patients; 29.1% of ischemic stroke), lacunar (10,828; 34.2%) and cardioembolic (9183; 29.0%). A significant difference of age was observed among subgroups. Monthly and Seasonal Variation in All Ischemic Strokes Fig 1A shows the moving average of monthly incidence of all ischemic strokes. The monthly variation was significant (P,.001), and the incidence was highest in May. The seasonal difference in incidence was not statistically significant (Fig 1, B; P 5.065), but the incidence in summer was significantly higher than in winter (incidence ratio 1.035; P,.05). Table 1. Characteristics of registered stroke patients Type of stroke Lacunar Atheromatous Cardioembolic Hemorrhagic No. of patients No. of men No. of women Age, y (mean 6 SD) * * * *,75 y y *Significant difference of age among subgroups (P,.001; Kruskal Wallis test).

3 38 S. TAKIZAWA ET AL. Figure 1. Moving average of monthly (A) and seasonal (B) incidence of all ischemic stroke. Monthly variation in the incidence of all ischemic stroke was significant (P,.001). Monthly and Seasonal Variation in Noncardioembolic Strokes Monthly variation in noncardioembolic strokes, including atherothrombotic and lacunar stroke, was significant (Fig 2, A; P,.001). Seasonal variation in noncardioembolic stroke was also significant (Fig 2, B; P,.001); the incidence in summer was significantly higher than in winter (incidence ratio 1.113; P,.001). Monthly variations in noncardioembolic stroke were significant for both sexes (P,.01), but a seasonal difference was seen only in men (P,.001) and not in women (P 5.313; data not shown). Seasonal Variation in Subtype of Ischemic Stroke Fig 3, A shows the distribution of ischemic stroke across the 4 seasons by subtype. Seasonal variation in lacunar stroke was significant (P,.001); the incidence in summer was significantly higher than in winter (incidence ratio 1.187; P,.001). Atherothrombotic stroke showed a similar distribution, but the difference was not significant (P 5.057; incidence ratio [summer/winter] 1.083; P,.01). In contrast, seasonal difference in cardioembolic stroke was significant (P,.001); the incidence in winter was significantly higher than in summer (incidence ratio 1.128; P,.001). Next, we examined the relation between onset time and seasonal incidence by subtype (Fig 3, B-D). All subtypes of ischemic stroke occurred most frequently at 7 to 12, followed by 13 to 18, 19 to 24, and 1 to 6 o clock. There was no difference in onset pattern among subtypes, although cardioembolic stroke at 7 to 12 o clock was more frequent in winter than in summer, but without significance (Fig 3, D; P 5.084). Monthly and Seasonal Variations in Hypertensive Hemorrhagic Stroke Fig 4, A shows the moving average of monthly incidence of hypertensive hemorrhagic stroke. Monthly variation was significant (P,.001); the highest number Figure 2. Moving average of monthly (A) and seasonal (B) incidence of noncardioembolic stroke. Monthly variation and seasonal difference of noncardioembolic stroke were significant (P,.001).

4 SEASONAL VARIATION OF STROKE IN JAPAN 39 Figure 3. Distribution of ischemic stroke across the 4 seasons according to subtype. (A) For lacunar stroke, the incidence was highest in summer and lowest in winter (P,.001). Atherothrombotic stroke showed a similar distribution, but without significance (P 5.057). In cardioembolic stroke, the incidence was highest in winter and lowest in summer (P,.001). There was no difference in the pattern of onset time among subtypes (B-D). of patients occurred in February and the lowest in August. Seasonal difference in hemorrhagic stroke patients was also significant (Fig 4, B; P,.001); the incidence in summer was significantly lower than in others (P,.001). Monthly and seasonal variations in hypertensive hemorrhagic stroke were seen in both sexes (P,.001). Interestingly, the incidence of hemorrhagic stroke in men was higher during activity, compared with at rest or during sleep (P,.05; data not shown). Regional Differences in Monthly Incidence of Noncardioembolic and Hemorrhagic Stroke Fig 5, A and B show the moving average of monthly patients divided by the annual incidence (%), in northern and southern Japan, for noncardioembolic and hemorrhagic stroke, respectively. As shown in Fig 5, A, there was a significant difference in monthly incidence of noncardioembolic ischemic stroke within the northern region (P,.001) but not within the southern region (P 5.422). Figure 4. Moving average of monthly (A) and seasonal (B) incidence of hypertensive hemorrhagic stroke. Monthly variation in the incidence of hemorrhagic stroke was significant (P,.001). Seasonal variation was also significant (P,.001).

5 40 S. TAKIZAWA ET AL. Figure 5. Regional difference in monthly incidence of noncardioembolic (A) and hemorrhagic stroke (B). There was a significant difference in monthly variation of the incidence of noncardioembolic ischemic stroke in the northern part of Japan (P,.001), but not in the southern part (P 5.422). Hemorrhagic stroke showed a regional difference of incidence between northern and southern parts. Hemorrhagic stroke was also unevenly distributed throughout the 12 months in the northern and southern regions (Fig 5, B; P,.05 and P,.001, respectively). Hemorrhagic stroke in the northern region showed a significantly higher incidence in June, July, and August compared to that in southern Japan. Its incidence in northern Japan was significantly lower in January compared to that in southern Japan. There was no relation between the regional variation of incidence of either noncardioembolic ischemic or hemorrhagic stroke and the regional monthly temperatures (high, low, and daily range of temperatures in the northern and southern regions; see Appendix B). Discussion Previous studies have mostly found that stroke incidence peaks in winter, 1-3,5,8-13 possibly because a decrease in temperature increases blood pressure, which tends to increase stroke incidence. 20,21 However, recent studies in Japan 14,15 have indicated that winter is not necessarily the peak season of stroke. Although it is unclear why the seasonal pattern of stroke occurrence would have changed during the past 2 decades, global warming is a possible reason. According to the 2007 Fourth Assessment Report by the Intergovernmental Panel on Climate Change, global surface temperature increased by C during the 20th century. In Japan, the annual mean temperature increased by around 1.13 C during the 20th century, and the increase accelerated after the 1990s (data from Japan Meteorological Agency). Hot summers may cause dehydration, which may lead to hyperviscosity of blood and enhanced platelet aggregation. 22,23 In the present large-scale study, we also analyzed seasonal variation in the incidence of subtypes of ischemic stroke in Japan; the only previous report on this topic is that of Ogata et al. 15 We found that lacunar stroke occurred more frequently in summer than in winter (P,.001). Atherothrombotic stroke showed a similar tendency (P 5.057). Atherothrombotic stroke is reported to be associated with enhanced platelet aggregation and blood hyperviscosity, 17,23,24 and hyperviscosity is also involved in the pathogenesis of lacunar stroke. 25 On the other hand, cardioembolic stroke occurred more frequently in winter (P,.001), in agreement with previous studies. 4,10,26 We also found that cardioembolic stroke occurred rather frequently at 7 to 12 o clock in winter. Excitation of the sympathetic system in the morning in winter may result in atrial fibrillation 26 and high blood pressure. 4,20 In addition, infection may promote the formation of intracardiac thrombi through a hypercoagulable state, 27,28 particularly in elderly patients. 29 This is consistent with the fact that patients with cardioembolic stroke were older than those with other subtypes in this study. The incidence of hypertensive hemorrhagic stroke showed significant monthly variation (P,.001), being highest in February and lowest in August, in agreement with reports that it occurs most frequently in winter. 3,5,6,11-13 Therefore, elevated blood pressure in winter, 20 particularly during activity, may trigger hemorrhagic stroke, whereas dehydration in summer may play a role in ischemic stroke. The mechanism underlying the observed difference between northern and southern Japan is unclear. We found no weather-related association, as shown in Appendix B. Other factors, including alcohol consumption, salt intake, and blood pressure also showed no significant correlation with regional variation of stroke incidence (data not shown). There are some limitations in our study. Because the 163 facilities contributing data to JSSRS are not equally distributed across Japan, it is unclear whether the division

6 SEASONAL VARIATION OF STROKE IN JAPAN 41 of Japan into northern and southern regions was appropriate. Also, we did not evaluate regional differences of risk factors, such as diabetes, dyslipidemia, smoking, myocardial infarction, C-reactive protein levels, etc. Finally, we used registry data but not data from a community-based surveillance system, which results in various limitations in exploring the possible mechanisms or causes of the seasonal variation of stroke. In conclusion, we have demonstrated distinct patterns of seasonal variation in the incidence of stroke subtypes in Japan, using the world s largest hospital-based stroke registration database. Our findings suggest that it is particularly important to prevent dehydration during summer in elderly patients. Acknowledgment: JSSRS was developed with support ( 21st Century Type Promoting Development of Clinical Research Fund ) from the Japanese Ministry of Welfare during 1999 to Supplementary Data Supplementary data (includes Appendix A and B) associated with this article can be found in the online version at doi: /j.jstrokecerebrovasdis References 1. Sobel E, Zhang ZX, Alter M, et al. Stroke in the Lehigh Valley: Seasonal variation in incidence rates. Stroke 1987;18: Giroud M, Beuriat P, Vion P, et al. Stroke in a French prospective population study. Neuroepidemiology 1989; 8: Azevedo E, Ribeiro JA, Lopes F, et al. Cold: A risk factor for stroke? J Neurol 1995;242: Kelly-Hayes M, Wolf PA, Kase CS, et al. Temporal patterns of stroke onset. The Framingham Study. Stroke 1995;26: Jakovljevic D, Salomaa V, Sivenius J, et al. Seasonal variation in the occurrence of stroke in a Finnish adult population: The FINMONICA Stroke Register: Finnish Monitoring Trends and Determinants in Cardiovascular Disease. Stroke 1996;27: Rothwell PM, Wroe SJ, Slattery J, et al. Is stroke incidence related to season or temperature? The Oxfordshire Community Stroke Project. Lancet 1996;347: Oberg AL, Ferguson JA, McIntyre LM, et al. Incidence of stroke and season of the year: Evidence of an association. Am J Epidemiol 2000;152: Wang Y, Levi CR, Attia JR, et al. Seasonal variation in stroke in the Hunter Region, Australia: A 5-year hospital-based study, Stroke 2003;34: Khan FA, Engstrom G, Jerntorp I, et al. Seasonal patterns of incidence and case fatality of stroke in Malm o, Sweden: The STROMA Study. Neuroepidemiology 2005;24: Spengos K, Vemmos K, Tsivgoulis G, et al. Diurnal and seasonal variation of stroke incidence in patients with cardioembolic stroke due to atrial fibrillation. Neuroepidemiology 2003;22: Suzuki K, Kutsuzawa T, Takita K, et al. Clinicoepidemiologic study of stroke in Akita, Japan. Stroke 1987;18: Shinkawa A, Ueda K, Hasuo Y, et al. Seasonal variation in stroke incidence in Hisayama, Japan. Stroke 1990;21: Wang H, Sekine M, Chen X, et al. A study of weekly and seasonal variation of stroke onset. Int J Biometeorol 2002; 47: Turin TC, Kita Y, Murakami Y, et al. Higher stroke incidence in the spring season regardless of conventional risk factors: Takashima Stroke Registry, Japan, Stroke 2008;39: Ogata T, Kimura K, Minematsu K, et al. Variation in ischemic stroke frequency in Japan by season and by other variables. J Neurol Sci 2004;225: Jimenez-Conde J, Ois A, Gomis M, et al. Weather as a trigger of stroke. Daily meteorological factors and incidence of stroke subtypes. Cerebrovasc Dis 2008;26: Crawford VL, McNerlan SE, Stout RW. Seasonal changes in platelets, fibrinogen and factor VII in elderly people. Age Ageing 2003;32: Kobayashi S, Japan Stroke Scale Registry Study Group. International Experience in Stroke Registry: Japanese Stroke Databank. Am J Prev Med 2006;31(6 Suppl. 2): S240-S Special report from the National Institute of Neurological Disorders and Stroke. Classification of cerebrovascular diseases III. Stroke 1990;21: Brennan PJ, Greenberg G, Miall WE, et al. Seasonal variation in arterial blood pressure. Br Med J 1982;285: Woodhouse PR, Khaw KT, Plummer M. Seasonal variation of blood pressure and its relationship to ambient temperature in an elderly population. J Hypertens 1993; 11: Schafer AI. The hypercoagulable states. Ann Intern Med 1985;102: Keatinge WR, Coleshaw SR, Easton JC, et al. Increased platelet and red cell counts, blood viscosity, and plasma cholesterol levels during heat stress, and mortality from coronary and cerebral thrombosis. Am J Med 1986; 81: Tanizaki Y, Kiyohara Y, Kato I, et al. Incidence and risk factors for subtypes of cerebral infarction in a general population: The Hisayama study. Stroke 2000; 31: Schneider R, Ringelstein EB, Zeumer H, et al. The role of plasma hyperviscosity in subcortical arteriosclerotic encephalopathy (Binswanger s disease). J Neurol 1987; 234: Frost L, Vukelic Andersen L, Mortensen LS, et al. Seasonal variation in stroke and stroke-associated mortality in patients with a hospital diagnosis of nonvalvular atrial fibrillation or flutter. A population-based study in Denmark. Neuroepidemiology 2006;26: Woodhouse PR, Khaw KT, Plummer M, et al. Seasonal variations of plasma fibrinogen and factor VII activity in the elderly: Winter infections and death from cardiovascular disease. Lancet 1994;343: Macko RF, Ameriso SF, Gruber A, et al. Impairments of the protein C system and fibrinolysis in infectionassociated stroke. Stroke 1996;27: Stout RW, Crawford V. Seasonal variations in fibrinogen concentrations among elderly people. Lancet 1991; 338:9-13.