目標：登革出血熱致病機轉迄今仍未完全了解。本研究目標為釐清登革熱併發登革出血熱 的危險因子，以及併發登革出血熱後死亡的預測因子。方法：本研究之研究對象為2003-2013 年間所有確診登革熱或登革出血熱個案。以病例對照研究探討登革熱併發登革出血熱之危險 因子，以世代追蹤研究分析登革出血熱個案死亡的預測因子，以邏輯斯迴歸進行統計分析。 結果：病例對照研究納入142例登革出血熱病例，並從同時期9,787例未併發登革出血熱之登 革熱病例中，以隨機抽樣方式選出568位依發病時間及居住縣市1:4配對之對照。60歲以上長 者(OR=4.53, p<0.0001)與第二型登革病毒感染(OR=5.39, p=0.0081)是發生登革出血熱的獨立危 險因子。世代追蹤研究納入142例登革出血熱病例，其中24例死亡。60歲以上長者(OR=3.21, p<0.027)及糖尿病患者(OR=3.08, p<0.005)是登革出血熱死亡的獨立預測因子。結論：本研究結 果可提供民眾對登革出血熱危險因子及預後因子的正確認知，加強醫事人員臨床警覺、早期診 斷及治療，以降低登革出血熱死亡率。(台灣衛誌 2015；34(4)：437-446)

Objectives: The pathogenesis of dengue hemorrhagic fever is not fully understood. The aims of this study were to examine the risk factors for developing dengue hemorrhagic fever after acquiring dengue infection and to identify the predictors of mortality after developing dengue hemorrhagic fever. Methods: All confirmed dengue and dengue hemorrhagic fever cases during 2003-2013 in Taiwan were included in this study. In this case-control study, we compared cases of dengue hemorrhagic fever with controls who were randomly selected from identified patients with dengue but without hemorrhagic fever, individually matched by month of onset and residential county in a 1:4 ratio. We tracked the outcomes of patients with dengue hemorrhagic fever, and analyzed the predictors of mortality. Multivariate logistic regression was used to analyze the data. Results: The study included 142 dengue hemorrhagic fever cases and 568 controls. Age over 60 (OR = 4.53, p <0.0001) and infection by the type 2 dengue virus (OR = 5.39, p = 0.0081) were two independent risk factors for dengue hemorrhagic fever. There were 24 deaths among the 142 dengue hemorrhagic fever cases. Age over 60 (OR = 3.21, p <0.027) and diabetes mellitus (OR = 3.08, p <0.005) were the two independent predictors of mortality. Conclusions: Knowledge about risk factors and prognostic factors will enable communities and health care professionals to better coordinate the effort to make early diagnoses, initiate treatment for dengue hemorrhagic fever, and minimize dengue-associated morbidity and mortality. (Taiwan J Public Health. 2015;34(4):437-446)