Ischemic Stroke Subtypes in a Japanese Population
Takashima Stroke Registry, 1988–2004
Background and Purpose—Population-based information on the epidemiology of ischemic stroke (IS) subtypes is scant. In this study, we characterized IS subtypes in terms of incidence, time trend, and risk factor profiles in a community-based population.
Methods—We obtained data from the Takashima Stroke Registry on approximately 55 000 residents of Takashima County in central Japan and calculated age-adjusted stroke incidence rates for different IS subtypes. We determined the incidence time trend by calculating the average annual change across years and also compared risk factors between subtypes.
Results—There were 1389 first-ever ischemic strokes registered during 1988 to 2004. Lacunar infarction was the most frequent etiology (54.1%) followed by cardioembolic infarction (22.9%). Age-adjusted incidence rates for different IS subtypes were lacunar, 77.1; cardioembolic, 31.5; and nonlacunar, 29.7/105 person-years. The average annual change was not significant for the IS subtypes except for nonlacunar infarction, which showed a decreasing trend. Risk factor analysis showed that patients with lacunar infarctions were younger and less likely to have a history of transient ischemic attack or a drinking habit than patients with nonlacunar infarctions.
Conclusion—Lacunar infarct was the most common IS subtype in our population. We found no significant change in the incidence of subtypes during the study, except a decrease in nonlacunar infarction.
Similar to other developed countries, ischemic stroke (IS) is the most common type of stroke in Japan.1–3 However, population-based information on IS subtypes in Japan is scarce. Because of differences in pathogenesis, prognosis, and treatment among IS subtypes, it is important to explore the population-based incidence and its trends for the subtypes. Previous studies on IS subtypes in Japan were either based on small cohorts,4,5 or on series of patients recruited consecutively from hospital admissions.6,7 On the contrary, to comprehensively investigate the incidence and trends of IS subtypes in a population, the ideal study population would include all patients with an incident event in a particular geographical area. A registry data following an entire community population continuously for a long time period would be very appropriate to determine the incidence of IS subtypes as well as any trends over time.
Very few cardiovascular disease registries covering an entire community exist in Japan to define the incidence and the secular trends of IS subtypes over a prolonged time period. In accordance with the World Health Organization–Monitoring Trends and Determinants in Cardiovascular Disease (WHO-MONICA) Study, we have monitored the incidence of stroke in Takashima County, Shiga Prefecture, in Japan for a number of years. In the present report, we estimated the incidence of IS by subtypes, analyzed their risk factors, and examined the secular time trends in incidence using an entire community covering population-based stroke registry in Japan.
Population and Methods
The Takashima Registry
The Takashima Stroke Registry is an integrated part of the Takashima Cardiocerebrovascular Disease Registration System, which was established in 1988 in Takashima County, Shiga, Japan.8,9 The population of Takashima County remained stable over the 17-year study period and was 55 451 (men: 49.2%; women: 50.7%) in the Year 2000.10 Among the population, 22.3% were ≥65 years old, which is higher than the all Japan proportion of 17.4%.
Case Finding and Registration Process
The registered patients included all suspected stroke cases among the residents of Takashima County. Details of the case finding, registration process, diagnostic criteria, and items registered have been described elsewhere.8,11 In brief, we used multiple case-finding sources for case ascertainment that included hospital records, emergency ambulance service records as well as death certificates. For possible inclusion in the registry, all suspected stroke events, both hospitalized and out-of-hospital cases, were identified, evaluated, and confirmed using several sources of information and were validated by diagnostic and registration criteria established by the Monitoring System for Cardiovascular Disease commissioned by the Ministry of Health and Welfare, Japan.12 These criteria are in accord with the WHO-MONICA project.13 For hospitalized cases, hospital admission, discharge records, or death records were routinely collected for all suspected stroke events. For nonhospitalized cases (ambulance service usage, emergency room visits, deaths in emergency rooms, deaths on or during arrival at the hospital, or community deaths), the event information was collected from ambulance service records, emergency room records, or death records. For information on prehospital fatal cases, ambulance service records and death certificates at the local government were screened for acute events.
A stroke was classified as IS when brain imaging revealed acute infarction or showed no evidence of hemorrhage. All suspected IS cases in the population during the study period were identified and evaluated. IS was further classified into lacunar, cardioembolic, nonlacunar infarctions, and unclassifiable infarction.11,14 Lacunar infarction characterizes patients who present with focal neurological symptoms and signs, brain image evidence of infarct size ≤2 cm with typical location or unstated size or when the imaging result was negative, and no source of cardiac embolism could be identified. Cardioembolic infarction required the same criteria of IS plus evidence of a possible source of embolus such as valvular heart disease, atrial fibrillation, or history of acute myocardial infarction. Nonlacunar infarction characterizes patients with IS with brain imaging evidence of infarct size >2 cm, no definite cardioembolic source, moderate to severe arterial stenosis, or infarction of other determined causes. The infarcts that did not meet any of these criteria or may have had inadequate evaluation such that reasonable diagnostic classification was difficult were categorized as unclassifiable infarction. Items recorded at registration of a stroke were the date and time of stroke onset, the situation and symptoms at the event onset, the extent of neurological symptoms and clinical observations at the event, history, and so on. This study was approved by the Institutional Review Boards of the Shiga University of Medical Science.
Our analysis included all patients from the Takashima Registry who had their first-ever IS during the period from January 1, 1988, to December 31, 2004. Average annual incidence rates for subtypes of IS were calculated with denominators interpolated from the yearly population demographic data derived from the routine census and vital statistics system for Takashima County for the study period. We also divided the total observation period of 17 years into 3 year groups: 1988 to 1993, 1994 to 1998, and 1999 to 2004; and incidence rates across the year groups were calculated. The rates were adjusted applying the direct method using the Japanese population from the 2000 census as the standard population. For all of the incidence rates, 95% CIs were calculated by Byar’s method.15 To examine the incidence of stroke across time, the number of stroke events within a single year was modeled as an overdispersed integer response following a Poisson distribution.16 Generalized linear models were used with year as an independent variable. This method allows for estimation of trends across individual calendar years to obtain average annual percentage changes. The Poisson regression procedure fits a model of the following form: “log (incidence)=a+β (year).” The exponential of the estimate of regression coefficient “β (year)” is the relative risk per passing of each calendar year. Based on the corresponding estimated regression coefficient and SE, a yearly change in stroke can be derived along with 95% CI. The regression coefficient of year multiplied ×100 gives the average annual change of incidence in percents.9 The 95% CI was calculated as exp (β±1.96×SE) with “SE” serving as the SE of “β.”
Distributions of patient characteristics and risk factors among the IS subtypes were compared. Categorical variables were analyzed using the χ2 test. Comparisons of means were made using the analysis of variance. In multivariate models, for the baseline variables’ association with IS subtypes, each subtype was compared against the other subtypes. We estimated the odds of the presence of a risk factor in a particular IS subtype to the odds of its presence in other subtypes (lacunar versus cardioembolic, lacunar versus nonlacunar, and cardioembolic versus nonlacunar). Results were expressed in ORs and the 95% CIs were calculated.
All statistical analysis was performed using SAS Version 9.1 (SAS Institute, Cary, NC).
During the years 1988 to 2004, we registered a total of 1389 cases (781 men and 608 women) of first-ever IS incidence in Takashima County. The average age for men and women was 73.7 (SD: 11.0) years and 76.8 (SD: 10.5) years, respectively. The majority of IS was of the lacunar subtype (54.1%) followed by cardioembolic infarction (22.9%) and nonlacunar infarction (21.0%). Tables 1 and 2⇓ show the patient characteristics and prevalence of risk factors across IS subtypes, respectively.
Among the IS subtypes, age adjusted incidence rate (per 105 person-years) of lacunar infarction was 77.1 (95% CI: 71.6 to 82.6) followed by cardioembolic infarction, 31.5 (95% CI: 28.0 to 34.9), and nonlacunar infarction, 29.7 (95% CI: 26.3 to 33.2). Table 3 shows age-specific and age-adjusted incidence rates of first-ever IS and its subtype (per 105 person-years) for men and women. The incidence rates for both genders increased with advancing age.
Table 4 shows the age-adjusted incidence rates (per 105 person-years) of IS subtypes for men and women among the 3 observation periods and the average annual change in the incidence of IS during the 17-year study period adjusted for age and gender. Although we observed a significant decrease of lacunar infarction incidence between 1994 to 1998 and 1999 to 2004, when we examined the incidence across time, the overall yearly change was insignificant across the study years of 1988 to 2004 (average annual change of 0.54%; 95% CI: −1.17 to 2.24). The age- and gender-adjusted average annual change was not significant for all the IS subtypes except for nonlacunar infarction, which showed a decreasing trend.
The associations between risk factors and IS subtypes are shown in Table 5. Lacunar infarction was negatively associated with age, previous transient ischemic attack, and the presence of a drinking habit compared with cardioembolic infarctions. Lacunar infarction was negatively associated with age and the presence of a drinking habit compared with nonlacunar infarctions. Cardioembolic infarction was positively associated with age and the presence of a previous transient ischemic attack compared with nonlacunar infarction. No difference in a history of hypertension, diabetes, or dyslipidemia was observed across IS subtypes.
Our results supports previous observations that lacunar stroke is the dominant type of IS in Japan,2,4–6 whereas in Western countries, large thromboembolic stroke prevails.2,14,17 In our study, the proportion of lacunar stroke was >50% of the IS subtypes, which is similar to the other reported studies from Japan2,4–6 but is much greater than that reported in American, Australian, or German populations.17–19 In the present study, we observed a level trend in the age-adjusted incidence rate of lacunar infarction in Japanese population during the 17-year periods from 1988 through 2004. Nonlacunar infarction showed a decreasing trend. The age-adjusted proportions of IS subtypes did not significantly differ from 1992 to 2002 for a study conducted in the urban setting of Osaka.6 On the contrary, it was reported that lacunar stroke decreased by approximately 50% from 1961 to 2000 in the Hisayama study.5
Analysis of the distribution of risk factors in our study showed that patients with lacunar infarction were younger, had a lower prevalence of a history of transient ischemic attack, and were less likely to have a drinking habit than patients with other infarction subtypes. Although we did not observe any difference in the prevalence of hypertension across IS subtypes, some previous studies have reported that patients with lacunar infarction tend to have a higher prevalence of hypertension compared with other subtypes.4,19 However, a recent systematic review concluded that the apparent excess of hypertension in lacunar stroke compared with nonlacunar stroke was mainly due to bias because many of the studies included hypertension in the definition of IS subtypes.20 Diabetes has been reported to be more closely associated with lacunar infarction than with other subtypes.4,14,19 Diabetes is associated with elevated coagulation factors and hyperinsulinemia, which might contribute to the development of lacunar infarction.21,22 However, we did not find any significant difference in the prevalence of diabetes among IS subtypes in our study. A meta-analysis of studies on risk factors in patients with lacunar versus nonlacunar infarction indicated there was no clear association between smoking, excess alcohol consumption, previous transient ischemic attack, or elevated cholesterol and lacunar versus nonlacunar infarctions.20
The comprehensiveness of the registration system for cardiocerebrovascular diseases such as stroke is essential in determining the incidence and trend in a particular area. A system to capture all patients in the study area together with accurate diagnosis is required to ensure the comprehensiveness of the registration. Factors that reduce the comprehensiveness of a register are missing cases or cases lacking confirmed diagnosis, patients being admitted to hospitals outside the registration area, and nonregistration. The quality of our registration system was assured by its completeness. Our registry system was planned to capture all the cases in the study area by covering all the hospitals of the county. It has been estimated that >98% of all hospital admissions of Takashima County are seen in these institutions.9 To ensure that eligible patients hospitalized outside the county were not omitted, registration procedures were also conducted at 4 high-level medical facilities outside the county. In Japan, almost 100% of the residents are covered by health insurance under the control of the Ministry of Health and Welfare.9,10 Individuals with mild stroke who visited general physicians in the community are almost always referred to secondary- or tertiary-level hospitals for more extensive investigations. In addition, 24-hour emergency ambulance service is available for residents without charge. The usual practice in Japan is to take patients with any acute conditions to the emergency facilities. Furthermore, for all deaths including out-of-hospital deaths, registration of the death events and their causes is mandatory under the Family Registration Law in Japan and is implemented robustly throughout the country. It is very unlikely that individuals with suspected stroke, having acute symptoms lasting >24 hours, were treated out-of-hospital because of several unique characteristics of the Japanese healthcare system: insurance coverage for all the residents, general tendency toward hospitalization for any acute condition, and free of charge and round-the-clock emergency ambulance service. Nevertheless, there is always a possibility that nonfatal nonhospitalized strokes might be left out; however, we believe that the number of these events, if any, was negligible. Japan has the most MRI units (40 U per million population) and CT scanners (93 U per million population) per capita among the developed countries.23 Only 2% of our IS stroke cases had to be classified as unclassifiable infarction due to lack of clinicoimaging results. Therefore, we believe that in our registration system, identification of stroke cases in the study area was almost complete and stroke categorization was quite accurate. The number of unclassified strokes changed across the study years (yearly decrease by 12.38%), but presumably the total number of unclassifiable events was too small (n=28) to account for the stable incidence of some of the IS categories over time.
Potential limitations of this study warrant consideration. Our subclassifications were made by a review of medical records and neuroimaging reports rather than direct examination of patients or images. The information regarding the timing of scans for neuroimaging after stroke onset was not available. Thus, there is a chance that some cases that were categorized based on early scan results may have been misclassified. Finally, the Takashima Registry covers a rural population in Japan, which may be different from the metropolitan population. Thus generalization of our study results needs caution.
In conclusion, our study showed that the incidence of lacunar infarction is much higher than the other subtypes for both men and women. For lacunar and cardioembolic infarction, we observed no significant change in incidence during the study period, although nonlacunar showed a decreasing trend in our study population.
Sources of Funding
Supported by the Research on Cardiovascular Disease (3A-1, 6A-5, 7A-2), The Comprehensive Research on Cardiovascular and Life Style Related Diseases (H18-CVD-Ippan-029) of the Ministry of Health and Welfare, and Grants-in-Aid Scientific Research (B-17390186, B-20390184, C-213670361, No.17015018) of Ministry of Education, Culture, Sports, Science and Technology. T.C.T. and N.R. are supported by fellowships and Research Grants-In-Aid (P-20.08124 and P-21.09139) from the Japan Society for the Promotion of Science.
- Received February 3, 2010.
- Revision received March 17, 2010.
- Accepted April 7, 2010.
Sudlow CL, Warlow CP. Comparable studies of the incidence of stroke and its pathological types: results from an international collaboration. International Stroke Incidence Collaboration. Stroke. 1997; 28: 491–499.
Ueshima H, Sekikawa A, Miura K, Turin TC, Takashima N, Kita Y, Watanabe M, Kadota A, Okuda N, Kadowaki T, Nakamura Y, Okamura T. Cardiovascular disease and risk factors in Asia: a selected review. Circulation. 2008; 118: 2702–2709.
Tanizaki Y, Kiyohara Y, Kato I, Iwamoto H, Nakayama K, Shinohara N, Arima H, Tanaka K, Ibayashi S, Fujishima M. Incidence and risk factors for subtypes of cerebral infarction in a general population: the Hisayama Study. Stroke. 2000; 31: 2616–2622.
Kubo M, Kiyohara Y, Ninomiya T, Tanizaki Y, Yonemoto K, Doi Y, Hata J, Oishi Y, Shikata K, Iida M. Decreasing incidence of lacunar vs other types of cerebral infarction in a Japanese population. Neurology. 2006; 66: 1539–1544.
Kitamura A, Nakagawa Y, Sato M, Iso H, Sato S, Imano H, Kiyama M, Okada T, Iida M, Shimamoto T. Proportions of stroke subtypes among men and women > or =40 years of age in an urban Japanese city in 1992, 1997, and 2002. Stroke. 2006; 37: 1374–1378.
Kimura K, Kazui S, Minematsu K, Ymaguchi T; for the Japan Multicenter Stroke Investigator’s Collaboration. Analysis of 16 922 patients with acute ischemic stroke and transient ischemic attack in Japan: a hospital-based prospective registration study. Cerebrovasc Dis. 2004; 18: 47–56.
Turin TC, Kita Y, Murakami Y, Rumana N, Sugihara H, Morita Y, Tomioka N, Okayama A, Nakamura Y, Abbott RD, Ueshima H. Higher stroke incidence in the spring season regardless of conventional risk factors: Takashima Stroke Registry, Japan: 1988–2001. Stroke. 2008; 39: 745–752.
Rumana N, Kita Y, Turin TC, Murakami Y, Sugihara H, Morita Y, Tomioka N, Okayama A, Nakamura Y, Abbott RD, Ueshima H. Trend of increase in the incidence of acute myocardial infarction in a Japanese population: Takashima AMI Registry, 1990–2001. Am J Epidemiol. 2008; 167: 1358–1364.
Ministry of Public Management, Home Affairs, Posts, and Telecommunications of Japan. 2000 Population Census of Japan [in Japanese]. Tokyo: Statistical Bureau, Ministry of Public Management, Home Affairs, Posts, and Telecommunications, Government of Japan; 2001.
Turin TC, Kita Y, Rumana N, Takashima N, Ichikawa M, Sugihara H, Morita Y, Hirose K, Murakami Y, Miura K, Okayama A, Nakamura Y, Abbott RD, Ueshima H. Morning surge in circadian periodicity of ischemic stroke is independent of conventional risk factor status: findings from the Takashima Stroke Registry 1990–2003. Eur J Neurol. 2009; 16: 843–851.
Ministry of Health and Welfare. Study Project of Monitoring System for Cardiovascular Disease Commissioned by the Ministry of Health and Welfare. Manual for the Registry and Follow-Up of Stroke [in Japanese]. Osaka: National Cardiovascular Center; 1998.
World Health Organization MONICA Project. Event Registration Data Component, MONICA Manual Version 1.1. Document for meeting of MONICA Principal Investigators. Geneva: WHO, 1986: S-4: 9–11.
Ohira T, Shahar E, Chambless LE, Rosamond WD, Mosley TH, Folsom AR. Risk factors for ischemic stroke subtypes: the Atherosclerosis Risk in Communities Study. Stroke. 2006; 37: 2493–2498.
Rothman KJ. Epidemiology: An Introduction. New York: Oxford University Press; 2002.
McCullagh P, Nelder JA. Generalized Linear Models, II ed. London: Chapman and Hall; 1989: 198–200.
White H, Boden-Albala B, Wang C, Elkind MSV, Rundek T, Wright CB, Sacco RL. Ischemic stroke subtype incidence among whites, blacks, and Hispanics: the Northern Manhattan Study. Circulation. 2005; 111: 1327–1331.
Grau AJ, Weimar C, Buggle F, Heinrich A, Goertler M, Neumaier S, Glahn J, Brandt T, Hacke W, Diener HC. Risk factors, outcome, and treatment in subtypes of ischemic stroke: the German stroke data bank. Stroke. 2001; 32: 2559–2566.
Jackson C, Sudlow C. Are lacunar strokes really different? A systematic review of differences in risk factor profiles between lacunar and nonlacunar infarcts. Stroke. 2005; 36: 891–901.
Matsuda T, Morishita E, Jokaji H, Asakura H, Saito M, Yoshida T, Takemoto K. Mechanism on disorders of coagulation and fibrinolysis in diabetes. Diabetes. 1996; 45: S109–S110.
Zunker P, Schick A, Buschmann HC, Georgiadis D, Nabavi DG, Edelmann M, Ringelstein EB. Hyperinsulinism and cerebral microangiopathy. Stroke. 1996; 27: 219–223.
Organization for Economic Co-operation Development. OECD Health Data 2009: Statistics and Indicators for 30 Countries. Available at: www.oecd.org/health/healthdata/. Accessed December 14, 2009.