A Meta-analysis on the Risk Factors of Cerebrovascular Disorders in Koreans

Park, Jong Ku; Kang, Myung Keun; Kim, Chun Bae; Kim, Ki Soon; Jee, Sun Ha

Abstract

This study was conducted to integrate the results of studies which evaluated or identified the risk factors of CVD(cerebrovascular disorders) for Koreans. We retrieved the literature published in Korean by manual search and the English literature by Medline database to identify studies on the relationship between reported risk factors and CVD conducted for the Korean from 1980 to August, 1997. Hypertension and total serum cholestrol were selected as subjects of quantitative meta-analysis as risk factors of CVD in Koreans. The overall effect sizes of the risk of CVD due to hypertension and total serum cholesterol were calculated by common odds ratio(OR) and average standardized mean difference, retrospectively. Before the integration of each effect sizes into common effect sizes, the heterogeneity tests were conducted. Also, sensitivity tests were conducted for the estimated common effect sizes. Regarding hypertension and CVD, a total of 9 epidemiologic studies were identified with a total of 2,271 cases of CVD. The overall OR of hypertension associated CVD was 4.10(95% confidence interval[CI] 3.56 to 4.71). The OR of hypertension associated with hemorrhagic CVD and ischemic CVD were 6.56(95% CI : 4.92 to 8.80) and 3.28(95% CI : 2.77 to 3.90), retrospectively. The OR of hypertension in relation to hemorrhagic CVD was significantly higher than that of hypertension in relation to overall CVD or ischemic CVD. Regarding serum total cholesterol and ischemic CVD, total 9 epidemiologic studies were identified with a total of 843 cases of ischemic CVD. Average mean difference as an effect size was 0.76, which was judged an important value according to Cohen's criteria. Our data suggested that hypertension was an important risk factor of overall CVD and its subtypes, and that the total serum cholesterol was associated with ischemic CVD in Koreans. For the lack of reliable prospective studies, however, we concluded that further research designed longitudinally would be required in this area.