台南市和高雄市登革大流行後的環境、病媒蚊密度與防治策略等因子對病毒感染率及病例發生率之影響，2008-2011

The impact of environmental factors, mosquito densities, and public health intervention strategies on the incidence rates of dengue virus infection and cases after the large-scale epidemics in Tainan and Kaohsiung Cities, 2008-2011

李崢嶸(Zheng-Rong Tiger Li)

指導教授 ： 金傳春

國立臺灣大學/公共衛生學院/流行病學與預防醫學研究所/碩士(2012年)

https://doi.org/10.6342/NTU.2012.01115

摘要

登革自 1988 在南台灣流行後，近年發生頻率與病例數逐年增加。台南市歷經 2007 年 6 月至隔年 1 月第一型登革病毒流行，造成 1,389 例本土確定病例，主要發生在北區和東區。高雄市亦於 2009 年中至 2011 年初爆發兩波登革流行，第一波為期 29 週（2009 年 7 月至 2010 年 2 月），主要為第三型登革病毒；而第二波為期 42 週（2010 年 3 月至 2011 年 2 月），且分離出四型登革病毒。高雄市衛生局在 2010 年至 2011 年的第二波流行中首度全市使用噴霧罐。本研究的主要目的即以血清流行病學測量兩市居民和學童的抗登革病毒抗體血清盛行率及發生率，整合病媒蚊密度、人口密度、環境因子和防治措施，並探查其對登革病例發生率的影響。
本研究包含三部分：（一）登革病毒的血清流行病學：於 2008 至 2010 年間，在台南市和高雄市各進行一社區居民及學童的抗登革病毒抗體血清流行病學探究。研究區為 2007 年台南市第一型登革病毒流行的高發區及高、低發區的學校，與 2009 至 2010 年高雄市第一波第三型登革病毒流行的高發區及高、低發區的小學。實驗法採用台灣疾病管制局發展的抗登革病毒第一型非結構性蛋白（non-structural protein 1，NS1）人類免疫球蛋白 G（immunoglobulin G，IgG）間接酵素連結免疫呈色法（indirect enzyme-linked immunosorbent assay，indirect ELISA）及抗登革病毒人類免疫球蛋白 M（IgM）捕捉酵素連結免疫呈色法（capture ELISA）。所有陽性血清以血清型專一（serotype-specific）的抗登革病毒 NS1 IgG indirect ELISA 和中和試驗（neutralization test）分辨血清型別、判定抗體效價；並由問卷收集人口學、臨床病徵、流行病學等資訊，所有陽性感染者再進行流行病學分析。（二）以時序指標描述高雄市 2009 至 2011 年流行的登革病例的時序聚集特徵，和斯皮爾曼等級相關係數（Spearman's rank correlation coefficient）及因素分析（factor analysis）檢視高雄市登革病例發生率和解釋變項的相關性。（三）病媒蚊、環境因子和防治措施對登革病例出現與否和發生率的影響：以欄模型（hurdle model）複合一對數連結截切負二項分配（truncated negative binomial with log link）和風險函數連結二項分配（binomial with logit link），推論解釋變項對登革病例出現與否和發生率的影響，以逐步模式選擇赤池資訊量指標（Akaike information criterion，AIC）最低之模式。
血清學結果顯示於台南市四個 2007 年流行高發里的抗登革病毒 NS1 IgG 血清盛行率範圍在 6.1%（3/49）至 28.1%（16/57）之間，且 70.8%（17/24）的陽性血清來自 65 歲以上居民。此外，50%（12/24）的 IgG 陽性參加者為第一型登革病毒感染，41.7%（10/24）為多血清型感染，且多為 65 歲以上居民（8/10）。四所位於台南市 2007 年登革病例高發區的學校，其學童抗登革病毒 NS1 IgG 血清盛行率較低，介於 0.31%（2/656）至 2.34%（3/132）之間，且 88.9%（8/9）的學童為第一型登革病毒感染。高雄市港墘與草衙二里各有 14.0%（12/86）和 21.4%（12/56）的參加者曾感染登革病毒，且多為第一型登革病毒；此二里的登革病毒近期感染發生率各為 11.6%（10/86）和 1.8%（1/56）。高雄市五校的學童抗登革病毒 NS1 IgG 血清盛行率介於 0.5%（4/789）至 2.1%（4/188）之間，而登革病毒感染發生率介於 0.16%（1/625）到 2.6%（5/188）之間。本研究在小港和前鎮兩區仍發現 7 位學童為第四型登革病毒感染。
高雄市的兩波流行分別持續 29 周（2009/7/27 至 2010/2/12）及 49 周（2010/3/20 至 2011/2/16），造成 656 例及 866 例第三型登革病毒感染確診病例，其中有 9 例和 7 例為登革出血熱；時序指數、性別、年齡分布均無顯著差異。在高雄市第二波流行中，時序指標和登革出血熱病例數有顯著正相關，且各里的登革病例數和疾病嚴重度亦與雌埃及斑蚊（Aedes aegypti）比例正相關。由斯皮爾曼等級相關係數檢視，發現和登革病例發生率相關的變項中，以戶內容器指數為甚。家戶指數、容器指數、布氏指數、戶外容器指數彼此間有強烈正相關。因素分析得以上四指數的變異可以同一潛在因素描述，故其後選擇容器指數進入後續統計模式探究。欄模式（hurdle model）得出兩群指數，分別影響里的登革病例出現與否和發生率：在高雄市第一波登革流行中，成蚊指數和容器級數對登革病例出現與否有正向影響，而一旦登革病例出現，容器密度和成蚊指數即和登革病例發生率正相關；在第二波流行中 ，人口密度和容器級數對登革病例出現與否有正相關，而其後是大型孳生源密度和容器密度兩因素對病例發生率有正相關。此外，使用噴霧罐年度第一個發生病例的區，其流行時間（40 週）較使用熱煙霧機年度第一個發生病例的區（25 週）長。
綜言之，台南市和高雄市的 8 至 12 歲學童之抗登革病毒抗體血清盛行率和發生率甚低，推斷台灣尚未成為地方性流行（endemic）區域。本血清流行病學研究發現不顯性第四型登革病毒感染，且各里的抗登革病毒抗體血清盛行率差異甚大，反映潛在易感染人口的不同，在未來建模中需特別考慮。另本研究發現兩群不同因素分別影響登革病例出現與否和登革病例發生率，在公共衛生防治上有其重要意義，即減少容器數和清管大型孳生源，雨後應立即清理所有積水容器，減少其在人群聚集處扮演傳播角色；而在流行時民眾應著重避免病媒蚊叮咬，發燒者應有自主健康管理。基於上述重要發現，本研究嘗試提出一整合登革病毒感染與發病、病媒蚊密度、環境因子的方案，可作為後續建置南台灣登革整合偵測系統之參考。

關鍵字

登革 ； 血清流行病學 ； 病媒蚊密度/指數 ； 環境因子 ； 欄模式 ； 防治措施 ； 公共衛生政策 ； 台灣

並列摘要

Southern Taiwan has experienced several epidemics of dengue since 1981. After 1988, the largest-scale dengue virus serotype 1 (DENV-1) epidemic occurred in Tainan City from June 2007 to January 2008, resulted in 1,398 laboratory-confirmed indigenous dengue cases. Kaohsiung City had a two-wave epidemic of dengue/dengue hemorrhagic fever (DHF) from July 2009 to February 2011. The first wave lasted for 29 weeks (July 2009-Feb 2010) with DENV-3 as the major serotype, whereas the second wave lasted for 49 weeks (Mar 2010-Feb 2011) and all the four serotypes of DENV were detected. The objectives of this study were to measure the baseline immunity against DENV in both community and schoolchildren in these two cities, and to provide information for integrated dengue surveillance with mosquito indices, environmental factors and intervention strategies.
Two seroepidemiological studies were conducted in Tainan and Kaohsiung Cities during 2008-2010, involving a community-based study targeting at the Lis (the minimal administrative unit in Taiwan) with high incidence of dengue cases, and a school-based study including schools located in Lis with high or low incidence of dengue cases during the studied epidemics. Anti-dengue NS1 immunoglobulin G (IgG) indirect ELISA, anti-DENV IgM capture ELISA, and neutralization tests were used to screen the serum and determine the infected serotype(s). The epidemiological characteristics and temporal clustering features in the first and second waves of the 2009-2011 epidemic in Kaohsiung City were compared by temporal indices. Hurdle model (truncated negative binomial with log link and binomial with logit link) and stepwise selection by Akaike information criterion (AIC) were used to assess the effect of mosquito indices and environmental factors on the occurrence and incidence of dengue cases.
The anti-dengue NS1 IgG seroprevalence in 4 Lis of Tainan City ranged from 6.1% (3/49) to 28.1% (16/57), and about 70% (17/24) of seropositive subjects were from residents aged beyond 65 years of age. Besides, half of seropositive participants were infected with DENV-1 (12/24), and 41.7% (10/24) had responses against multiple serotypes. Schoolchildren in the 4 schools located in the high incidence areas had relatively low anti-dengue NS1 IgG seroprevalence [0.31% (2/656)-2.34% (3/132)], and 88.9% (8/9) of seropositive schoolchildren were infected with DENV-1. On the other hand, two Lis of Kaohsiung City had 14.0% (12/86) and 21.4% (12/56) participants with past infection(s), and 11.6% (10/86) and 1.8% (1/56) had recent infection during the study period. Schoolchildren also showed low anti-dengue NS1 IgG seroprevalence [0.5% (4/789)-2.1% (4/188)] and DENV seroincidence [0.16% (1/625)-2.6% (5/188)]. Seven schoolchildren in Siaogang and Cianjhen Districts were asymptomatically infected with DENV-4.
The first and second waves of the 2009-2011 epidemic of dengue in Kaohsiung City caused total of 656 and 866 dengue cases and 9 and 7 dengue hemorrhagic fever (DHF) cases, respectively, without differences in age, gender and temporal indices. Lis with DHF cases were significantly associated with high temporal indices during the 2010-2011 epidemic in Kaohsiung City. Hurdle models identified two groups of factors affecting the occurrence and incidence of dengue cases. In the first wave, adult mosquito index (AMI) and rank of container index were positively associated with the occurrence of dengue cases, while density of containers and AMI were positively associated with the incidence of dengue cases. In the second wave, population density and rank of container index had positive effect on the occurrence of dengue cases, whereas density of large breeding sites and density of containers had positive effect on the incidence of dengue cases. For the first district affected by epidemic of dengue, use of aerosol cans in the second wave lasted the epidemic longer than using thermal fog in the first wave (40 versus 25 weeks).
In summary, residents and schoolchildren had low seroprevalence and seroincidence of DENV in Tainan and Kaohsiung Cities, implying that southern Taiwan has not become dengue-endemic. In addition, herd immunity varied against different serotypes of DENV across the studied Lis, reflecting the different proportion of at-risk population for future outbreaks. This serological surveillance detected asymptomatic infection of DENV-4 that most clinical surveillance had missed. Two groups of factors associated with occurrence and/or incidence were valuable for the public health policies. For Lis have not been affected, public health professionals must focus on routine clearance of containers and avoiding people gathering; For Lis have been affected, avoiding mosquito bites and self-management for feverish cases are important. In this study, we proposed the frame of integration of mosquito densities, environmental, intervention strategies and other factors for future works to establish a more effective integrated dengue surveillance system.

並列關鍵字

Dengue ； Seroepidemiology ； Mosquito density/indices ； Environmental factors ； Hurdle model ； Prevention and control strategies ； Public Health Policies ； Taiwan

參考文獻

37. 黃智雄等. 台灣南部地區登革熱血清流行病學調查研究. 台灣疾病管制局 (2005)

1. Katherine L. Anders et al. Epidemiological factors associated with dengue shock syndrome and mortality in hospitalized dengue patients in Ho Chi Minh City, Vietnam. American Journal of Tropical Medicine and Hygiene 84(1): 127-134 (2011)

2. Angel Balmaseda et al. High seroprevalence of antibodies against dengue virus in a prospective study of schoolchildren in Managua, Nicaragua. Tropical Medicine and International Health 11(6): 935-942 (2006)

3. Angel Balmaseda et al. Trends in patterns of dengue transmission over 4 years in a pediatric cohort study in Nicaragua. The Journal of Infectious Diseases 201: 5-14 (2011)

4. L. M. Bartley et al. Assessment of the factors associated with flavivirus seroprevalence in a population in southern Vietnam. Epidemiology and Infection 128(2): 213-220 (2002)