Epidemiological Transition of Stroke in China
Twenty-One–Year Observational Study From the Sino-MONICA-Beijing Project
Background and Purpose— Epidemiological patterns of stroke may change with economic development. It is important to understand these changes for making timely strategies for stroke prevention. The aim of this study was to examine the changes in trends of stroke epidemiology during a period of fast economic development in China, based on data of the Sino-MONICA-Beijing project.
Methods— Acute stroke events were registered in a large defined population aged 25 through 74 years from 1984 to 2004. The age standardized incidence rates and case fatality rates of total stroke, ischemic stroke and hemorrhagic stroke were calculated. The trends were analyzed by a regression model.
Results— There was a total of 14 584 stroke events registered in the study population from 1984 through 2004. The incidence rate of hemorrhagic stroke declined by 1.7% and the incidence rate of ischemic stroke increased by 8.7% annually on average. The case fatality rates significantly reduced in both types of stroke. The mean onset age of stroke was delayed by 2.7 years in men and 3.6 years in women. The proportion of deaths of cerebrovascular disease out of total deaths decreased and the proportion of ischemic heart disease increased during the study period.
Conclusions— Characteristics of stroke transition were found during a period of economic development in China. The changes in patterns of stroke have raised new challenges and the need for priority adjustment for stroke prevention in China.
Several epidemiological characteristics of stroke are used as a symbol of poverty, which include more hemorrhagic stroke, a higher case fatality rate, more stroke deaths in total deaths than that of ischemic heart disease (IHD) and a younger onset age. These characteristics of stroke should change in the opposite direction during a period of economic transition, which is referred to as the epidemiological transition of stroke.1–2 However, few published studies have demonstrated stroke transition prospectively in a population from a low economic level to a higher economic level by a long-term observational study with data of incidence rates. China has been experiencing fast economic development in the last 2 decades, which has changed people’s lifestyle and health care greatly as a major driving force. It is important for China and other developing countries to recognize and understand the characteristics of epidemiological transition of stroke for making timely strategies of stroke prevention. This study aimed to examine the trends of epidemiology of stroke events in the last 2 decades in China and to describe the changes of trend of stroke quantitatively, based on the data of the Sino-MONICA (Monitoring Trends and Determinants in Cardiovascular Disease)-Beijing project from 1984 to 2004. The Sino-MONICA-Beijing project was a long-term surveillance study for incidences and mortality rates of both acute stroke events and acute IHD events in Beijing, China. This was a collaboration study of the WHO MONICA project from 1984 to 1993 and was continually conducted from 1994 to 2004 in part of the original Sino-MONICA-Beijing population with the same study protocol.3–7
Materials and Methods
The Sino-MONICA-Beijing population was defined by the method of nonrandom cluster sampling. The study population was scattered in 7 of 18 administrative districts (including counties) in Beijing and covered all residents with permanent census registration in a number of defined communities in each of the 7 administrative districts. The study population accounted for approximately 5% to 7% of the total population in Beijing from 1984 through 1993 and approximately 1% to 1.5% of the total population from 1994 to 2004, with a similar composition in age groups, gender, and urban and rural areas as that in the total population of Beijing. Stroke events were registered in the study population aged 25 through 74 years.
The population size by 5-year age strata for men and women was obtained from censuses. The intercensal estimation of the mid-year population was based on available data in local police substations for births, deaths, immigration and emigration in defined study areas annually or biennially. The population size in each year was used as the denominator for calculation of incidence rates.
Definition of Stroke Events
The definition of acute stroke events was based on the criteria developed by the WHO MONICA project.3 Stroke events were defined as rapidly developing signs of focal (or global) disturbance of cerebral function lasting >24 hours (unless interrupted by surgery or death) with no apparent nonvascular causes. The definition included patients presenting with clinical signs and symptoms of subarachnoid hemorrhage, intracerebral hemorrhage, thrombosis and embolism. Hemorrhagic stroke was defined as stroke events with diagnosis of subarachnoid hemorrhage or intracerebral hemorrhage and ischemic stroke was defined as stroke events with diagnosis of thrombosis or embolism. Transient ischemic attack (TIA) and chronic cerebral vascular disease were not included. On the basis of the survival status within 28 days of onset, stroke events were subdivided into fatal and nonfatal events.
A 3-level network was developed at the beginning of the study to form a case report system with close collaboration among coordinating centers (first level), 42 research teams in 42 collaborating hospitals (second level) and 412 resident committees or health stations (third level or grassroots level) over the whole study area. Suspected acute stroke events occurring in defined areas were registered by filling out a standardized event form. There were clear definitions and explanations for each item on the standardized form in the WHO MONICA protocol, which was translated into Chinese as a manual of operation. A combination of “hot pursuit” (screening of hospital or emergency room admission) and “cold pursuit” (identifying and extracting medical records after discharge) was used for hospitalized cases or cases dead in emergency rooms. Official routine death statistics were also available in each defined area, which was used as a channel of identification of fatal stroke cases. For a few nonhospitalized cases, all efforts were made to collect relevant medical and social information used for diagnosis. The case ascertainment and validation procedures have been described in detail previously.4–6
Total Cause of Death
All causes of death based on the official routine death register in the defined area were obtained from a local policy substation. The cause of death was coded using the 9th version of the International Classification of Diseases (ICD) by trained research staffs. The deaths with codes of cerebral vascular death (ICD, 9th, 430.0 to 438.0) and deaths with codes of IHD (ICD 9th, 410.0 to 414.9) were validated by the WHO MONICA criteria for fatal events.
Quality Assurance of Data Collection
The approaches of quality control and standardization of event registration and validation have been reported in detail previously.4–7 It was also an important quality assurance for the long-term study that certain key persons in our research team continuously worked for the long-term project. The urbanization and city construction undoubtedly increased immigration and emigration in some defined study areas in the late 1990s. However, the majority of acute stroke events in these defined areas were admitted to collaborating hospitals of the Sino-MONICA project. We also maintained collaborations with the resident committees, health stations (third level or grassroots level) and reimbursement agencies to assure that some suspected stroke events admitted to nonlocal hospitals were reported.
The incidence rates and case fatality rates of total acute stroke events (first-ever and recurrent events) and first-ever events for 2 subtypes of stroke were calculated separately for men and women. The rates of first-ever stroke events served as validity purpose. Age-standardized incidence and age-standardized case fatality rates were calculated by the direct method with 5-year age-group rates in the range of 25 to 74 years. For incidence rates, the weights of age groups were derived from the age distribution of the latest China national census in 2000. For case fatality rate, the weights of age groups were derived from the age distribution of all the stroke cases in 1984 to 2004.
Trends of stroke events were calculated from age-standardized annual rates by using the regression model log rt=a+bt, where log denotes the natural logarithm and t is the year. The trend b was estimated from the ordinary regression, and P<0.05 was used as the level of statistical significance; 100b represents the estimated annual percentage changes in the rate.5 The trend of case fatality rates were calculated using log-linear regression, assuming a binomial distribution within age groups. The trends of stroke were also calculated separately for the 2 study periods from 1984 to 1993 and from 1994 to 2004 with different sizes of study population.
SPSS version 13.0 was used for statistics calculations.
There was a total of 14 584 acute stroke events registered in the Sino-MONICA population from 1984 to 2004 (except for 2000 and 2003). Nonfatal stroke events accounted for 72.3% of total acute stroke events. Among the nonfatal stroke events, 87% were admitted to hospital after the onset and 13% were treated in nursing home or home; 82% were examined by neurologist and 18% were examined by physician. Among fatal stroke events, 61% were admitted to hospital and 39% either died in emergency room or outside of hospital; 67% of the fatal stroke events were examined by neurologists and 33% were examined by physicians. The utility rates of CT examination was 4.4% at 1984, 60.1% in 1988 and 97.5% in 2004. Table 1 lists the size of the population and numbers of stroke events in each year during the study period. The stroke events aged 65 to 74 years accounted for 41.2% of the total stroke events.
Trend of Incidence Rates of Ischemic Stroke and Hemorrhagic Stroke
Table 2 shows age standardized incidence rates of total stroke events, ischemic stroke events and hemorrhagic stroke events from 1984 to 2004 with trend b and its confidence interval (CI). The trend estimations were also shown separately from men and women and by 2 time periods from 1984 to 1993 and from 1994 to 2004 because of the higher utility rate of CT examination from 1994 to 2004. The age standardized incidence rate of hemorrhagic stroke declined by 1.7% (CI: −2.7, −1.2), P<0.05) annually from 1984 to 2004. The declining trend started from the first decade with a 1.8% (CI: −2.7, −0.8, P<0.05) decrease annually and it then declined more significantly in the second decade to 3.2% (CI: −4.4, −1.9, P<0.05) annually. The age standardized incidence rates of ischemic stroke increased by 8.7% (CI: 4.3, 8.9, P<0.05) annually from 1984 to 2004. The increasing trend of incidence rate of ischemic stroke event started from the first decade with 3.5% (CI: 2.5, 4.5, P<0.05) annual percentage changes and then it stayed at a 4.3% increase in the second decade, but this was not statistically significant. The incidence rate of total strokes increased by 6.7% (CI: 4.3, 8.9, P<0.05) annually from 1984 to 2004. Ischemic stroke increased more significantly in the first 10 years and hemorrhagic stroke decreased more significantly in the second time period of the study. No remarkable differences in the trends were found between men and women.
The trends of age standardized incidence rates of first-ever ischemic stroke events and age standardized incidence rates of first-ever hemorrhagic stroke events from 1984 to 2004 are presented in Table 3. The declining trend in the incidence of first-ever hemorrhagic stroke events and the increasing trend of incidence rates of first-ever ischemic stroke events were almost identical with the trends of incidence rates of stroke events in Table 2.
Trend of Case Fatality Rates of Stroke
Age standardized case fatality rates were decreased in hemorrhagic stroke, ischemic stroke and total stroke in both men and women (Table 4). There was a 1.7% decline in the case fatality rate of hemorrhagic stroke and a 0.5% decline in that of ischemic stroke annually. The declining trends of case fatality rates in both subtypes of stroke were more significant in the second decade of the study.
Changes in Proportional Mortality Rates of Stroke and Ischemic Heart Disease
Table 5 shows the proportion of deaths from cerebral vascular disease (ICD 430.0 to 438.0) and the proportion of deaths from IHD (ICD 410.0 to 414.9) in men and women from 1984 to 2004. The proportion of deaths from IHD increased from 9.7% in men and 7.0% in women in 1984 to 17.5% in men and 13.5% in women in 2004. The proportion of deaths from IHD was much lower than those from cerebral vascular disease in both men and women in 1984, but they were almost equal in 2004.
Changes in Onset Age of Stroke
The mean onset ages of total stroke was 2.7 years older in men and 3.6 years older in women in 2004 compared with the mean onset age in 1984 (men: 63.9±8.4 versus 61.2±7.9 years, P<0.01; women: 64.9±7.0 versus 61.3±8.8 years, P<0.01).
Stroke, as a largely preventable disease, remains the second or third leading cause of death in China, even with dramatic economic development in recent years.8–10 From a global vision, stroke in developing countries accounted for nearly 70% of global stroke deaths and 40% of stroke deaths in developing countries were in China.11 It is critically important for timely policy making of stroke prevention and control in China to recognize and understand the epidemiological trends during a period of economic transition based not only on mortality data but also on the data of incidence and case fatality.
Four characteristics of epidemiological transition of stroke were actually observed by the 21-year community-based surveillance study in China, which included declining incidence rates of hemorrhagic stroke, reduced case fatality rates, increased onset age, and an expanded proportion of IHD deaths and a shrunken proportion of stroke deaths in total deaths in the study population. This study also found a significant increase in the incidence rate of ischemic stroke, which actually led to a significant increase in total stroke burden in spite of remarkable decline in hemorrhagic stroke. The significant increase in incidence rates of ischemic stroke is also an important feature of epidemiological transition of stroke in China.
There may be several explanations for the 2 opposite directions in trends of incidence rates of subtypes of stroke in this study. Except for the common risk factors associated with hemorrhagic stroke, such as hypertension and smoking, ischemic stroke is associated more closely with atherosclerotic risk factors, such as an increase in serum LDL cholesterol, which is mainly due to an unhealthy lifestyle. The total fat intake in the study area increased from 88.1 g/d in 1983 to 97.4 g/d in 2002. Cholesterol intake increased from 334.5 g/d to 488.4 g/d in urban areas and from 124.8 g/d to 350.7 g/d in rural areas from 1983 to 2002.12–13 Mean serum cholesterol levels increased by 24% from 1984 to 1999, which explained 77% of the increase in IHD mortality in the same population.14 The prevalence rate of diabetes increased by 97% from 1994 to 2002.10 The prevalence rate of overweight or obesity increased by 13% in urban areas and by 85% in rural areas and in this region.14,15 The rapid increase in atherosclerotic risk factors in the study population is the most logical explanation for trends of ischemic stroke.
The decrease in hemorrhagic stroke should be closely related with the changes in health care and health access during economic development. The improvement in hypertension control is the most likely explanation because 50% of acute hemorrhagic strokes can be attributed to hypertension in Chinese.16,17 The control rate of hypertension was 3.4% in 1984 and 6.33% in 1996 and significantly increased in 1999 (17.5%) in urban areas. For rural areas, the control rate of hypertension was 1.1% in 1984 and didn’t change until 1999 (6.9%).15 The changing pattern of control rates of hypertension appeared to be related to a more significant decline of hemorrhagic stroke in the second time period of the study. The results imply that economic development probably has a much earlier impact on lifestyle and levels of risk factors but has a relatively later impact on the level of treatment and control of risk factors.
The change in diagnostic approach for subtypes of stroke is another possible contributor of the opposite trends of ischemic and hemorrhagic stroke. There was a dramatic increase in use of CT imaging technology during the first decade from 4.4% in 1984 to 84% in 1993 and kept in a higher use rate from 1994 to 2004.18 The increasing use of CT imaging actually represented the change in health care and health access in study population. It was possible the changes in health care and health access created certain misclassification of subtype stroke. But it is unlikely to be the major explanation for the trend because the declining trend of hemorrhagic stroke and increasing trend of ischemic stroke remained to exist in both periods with lower CT utility rate from 1984 to 1993 and higher CT utility rate from 1994 to 2004.
Case fatality rates from both ischemic and hemorrhagic stroke were reduced during the study period. The decrease in case fatalities of total stroke was due to the significant reduction in the proportion of hemorrhagic stroke events in total stroke event. Improvement in medical care for acute stroke events should also be the explanation for the remarkable reduction in case fatality rate of stroke.
This study showed a mean 2.7 year increase in onset age for male stroke patients and a 3.6 year increase in female stroke patients from 1984 to 2004. The increased onset age may reflect population aging. The proportion of the population aged over 65 years increased from 7% in 1984 to 15% in 2004 in the Sino-MONICA population.
One limitation of this study is that the diagnosis of subtype of stroke was based on clinical diagnostic criteria for subtypes of stroke. The diagnosis of subtype of stroke was mainly dependent on clinical symptoms, signs and examination results of lumbar puncture in the earlier stage of the study. We believe that there was some degree of misclassification for subtypes of stroke in this study. Another limitation is that more immigration and emigration of population in defined study areas as the result of city construction from 1994 to 2004 may increase the missing rate of register of stroke event.
The characteristics of epidemiological transition of stroke were found by a long-term surveillance study during a period of fast economic development in Beijing, China. The new challenges raised by the epidemiological transition of stroke are: (1) the risk factor control for stroke prevention should address not only hypertension but also multiple risk factors, such as elevated cholesterol, and diabetes and obesity, which are major risk factors for ischemic stroke and IHD; (2) secondary prevention of stroke has increasing importance because more stroke patients survived from their first ever events and they have composed of a large high risk group for recurrent stroke, and (3) there is increased demand for rehabilitation for stroke patients.
Substantial contributions to the intellectual content from each author: Dong Zhao: conception and design, acquisition of data, analysis and interpretation of data, drafting and revision of the manuscript, and administrative duties; Jing Liu: concept and design and critical revision of the manuscript for important intellectual content, and acquisition of data; Wei Wang: concept and design and critical revision of the manuscript for important intellectual content, and acquisition of data; Zhechun Zeng: concept and design and critical revision of the manuscript for important intellectual content, and acquisition of data; Jun Cheng: acquisition of data, statistical analysis, analysis and interpretation of data, and technical support; Jun Liu: acquisition of data, statistical analysis and technical support; Jiayi Sun: acquisition of data, statistical analysis, and technical support; Zhaosu Wu: concept and design and critical revision of the manuscript for important intellectual content, and obtaining funding.
Sources of Funding
This study was funded by the People’s Republic of China Ministry of Science and Technology and the Ministry of Health through the 7th, 8th and 10th National Five-year Key Research Plan (contracts 85-915-01-01, 85-915-01-02 and 2001BA703B01), with additional funding from the World Health Organization and a project from Beijing Municipal Science & Technology Commission (contract 953850700).
- Received August 23, 2007.
- Revision received October 23, 2007.
- Accepted November 14, 2007.
Howson CP, Reddy HK, Ryan TJ, Bale JR. Control Cardiovascular Diseases in Developing Countries. Washington D.C.: National Academy Press; 1998.
Yusuf S, Reddy KS, Ounpuu S, Anand S. Global burden of cardiovascular diseases Part I: general consideration, the epidemiology transition, risk factors and impact of urbanization. Circulation. 2001; 104: 2746–2753.
WHO MONICA project MONICA Manual. Available at: URL: http://www.ktl.fi/publications/monica. URNLNBN:fi-fe 19981147, 1990.
Wu ZS, Yao CH, Chen DY, Li N, Zhang M, Wu YY, Zhao D, Wu GX, Wu YK. The Sino-MONICA-Beijing study: report on results between 1984 and 1986. Acta Med Scand. 1988; Suppl 728: 60–66.
Wu ZS, Yao CH, Zhao D. Wu GX, Wang W, Liu J, Zeng ZC, Wu YK. A collaborative study on trends and determinants in cardiovascular disease in China, Part I: morbidity and mortality monitoring. Circulation. 2001; 103: 462–468.
Sarti C, Stegmayr B, Tolonen H, Mahonen M, Tuomilehto J, Asplund K. Are changes in mortality rate from stroke caused by changes in stroke events rates or case fatality? Stroke. 2003; 34: 1833–1841.
National center for cardiovascular diseases. 2005 report on cardiovascular diseases in China. Encyclopedia of China Publishing House. 2006(In Chinese).
Wu GX, Wu ZS, Liu J, Liu J, Zeng ZC, Wang WH, Qin LP, Zhou MR, Liu S. Trend of changes in mortality of cardiovascular diseases in some areas of Beijing during 1984 to 1998. Chinese J Preventive Med. 2001; 35: 98–1019(In Chinese).
Reddy KS, Yusuf S. Emerging epidemic of cardiovascular disease in developing countries. Circulation. 1998; 97: 596–601.
Yao CH, Hao EH, Wu YY, Dong L. Dietary survey in Beijing in 1983. Chinese Med J. 1985; 98: 439–441.
Pang XH, Jiao SF, Huang L, Duan JL, Ren ZY, Liu ZJ. Nutrition and health status of the Beijing residents. Chinese J Preventive Medicine. 2005; 39: 269–272(In Chinese).
Critchley J, Liu J, Zhao D, Wang W, Capewell S. Explaining the increasing in coronary heart disease mortality rate in Beijing between 1984 and 1999. Circulation. 2004; 110: 1236–1244.
Liu S, Zhao D, Wang W, Liu J, Qin LP, Zeng ZC, Wu ZS. The trends of cardiovascular risk factors in urban and rural areas of Beijing during 1984–1999. J Cardiovascular & Pulmonary Disease. 2006; 25: 129–134(In Chinese).
Tolomen H, Mohomen M, Asplund K, Rastenyte D, Kuulasmaa K, Vanuzzo D, Tuomilehto J. Do trends in population levels of blood pressure and other cardiolvascular risk factors explain trends in stroke event rates? Stroke. 2002; 33: 2367–2375.
Wang W, Zhao D, Sun JY, Wang WH, Cheng J, Liu J, Qin LP, Liu S, Wu ZS. Risk factors comparison in Chinese developing acute coronary event, ischemic or hemorrhagic stroke: a multi-provincial cohort study. Chinese J Cardiol. 2006; 34: 1133–1137(In Chinese).
Wu GX, Wu ZS, Zeng ZC, Qin LP, Liu J. Epidemiological characteristics of stroke in Beijing(Sino-MONICA-Beijing study). J J Cardiovascular & Pulmonary Disease. 1998; 17: 252–255(In Chinese).