| 摘要: | 目的：本研究主要想監控HIV-1在台灣地區的傳播並找出防治的方法。因此研究分為兩個部份，(1)進行HIV-1亞型(subtype)監測，以了解目前HIV-1亞型分佈狀況，並進一步推估病毒株演化情形與其起始年代，並且探討HIV-1亞型與危險族群的關係；(2)透過問卷調查調查影響注射藥癮者感染HIV-1的危險因子，以建立注射藥癮者感染HIV-1的風險預測模式，用來預測高危險族群感染HIV-1的風險，並提高其自覺感染HIV-1的危險性，以逹到防治的效果。
方法：研究對象收集來自2004年起已建立的HIV-1監測網，此監測網包括每1~2年收集台灣各地監所高危險群 – 注射藥癮者及中國醫藥大學附設醫院愛滋病特別門診內病患所做的HIV/AIDS知識、態度和行為之危險因子調查問卷和HIV-1陽性個案血液樣本。研究(1)，在HIV-1亞型監測方面，以2007~2009年間監測網內所收集來自監所和愛滋特別門診內HIV-1陽性病患之血液樣本，抽取DNA並針對HIV-1 env基因定序(Sequencing)做系統發生學分析(Phylogenetic analysis)以及多重巢式聚合酶連鎖反應(Nested multiplex PCR)來偵測HIV-1病毒株之env基因亞型，並應用溯祖理論(coalescent theory)、分子鐘 (Molecular Clock)原理推估其演化情形及起源年代。最後，利用統計分析了解不同亞型感染者之感染HIV-1危險因子特性。研究(2)，注射藥癮者感染HIV-1風險預測模式是使用監測網內2004~2005年間危險因子資料庫內注射藥癮者問卷資料，以病例對照研究法(case-control study)探討其感染HIV-1的危險因子，並使用邏輯斯迴歸(logistic regression)建立風險預測模式，再利用ROC curve (Receiver operating characteristic curve) 分析進而得知注射藥癮者其感染HIV-1風險閾值(Threshold)。驗證此預測模式是以2007~2008年間危險因子資料庫內注射藥癮者問卷資料作為驗證。
結果：(1)HIV-1亞型監測：經過系統發生學分析及多重巢式聚合酶連鎖反應區別HIV-1 亞型後，包括三種HIV-1亞型，分別為CRF07_BC、subtype B及CRF01_AE，以CRF07_BC占大多數(68.8%)。HIV-1基因序列演化分析方面主要是以CRF07_BC env序列做分析。根據在不同的substitution model與coalescent model下推估後得知，中國CRF07_BC出現的年代約在
1995.5~1996.1年間，台灣南部的CRF07_BC約為2001.1~2001.9 年，台灣中部的CRF07_BC約為2003.4~2003.7年，台灣北部的CRF07_BC約為2003.5~2003.7年。而HIV-1亞型與其相關危險因子的分析中，subtype B感染者多為同性戀者，CRF01_AE感染者大多為異性戀者，CRF07_BC感染者大多為異性戀且有注射毒品相關的危險行為。(2)HIV-1危險因子監測：男性注射藥癮者感染HIV-1的風險預測模式中的顯著危險因子共有4個(「教育程度」、「常得到性病的人比較容易得到愛滋病 (知識題)」、「共用針頭」、「共用毒品稀釋之溶液」)。進一步使用ROC curve挑選總危險對比值(Odds ratio，OR)為7.435做為最適當的風險閾值(敏感度為95.6%，特異度為52.2%)。驗證此預測模式結果發現，其敏感度為72.2%，特異度為60.5%。
Objectived：The main purposes of this study were to surveillance the HIV transmission and try to prevent it in Taiwan. (1) For the subtype surveillance purpose, this study was try to investigate the distribution of HIV-1 subtypes among different high-risk groups、spread pathways and evolution time period in Taiwan. (2)For the prevention purpose, this study wish to develop the HIV-1 infection risk prediction model for IDUs by using the questionnaire of risk factor investigation. This risk prediction model would aim to be a tool for predict their probable risk of HIV infection and increase their self-awareness HIV-1 infection.
Methods：The study population came from a Taiwan HIV-1 surveillance network which builded from 2004. This network established a database which included HIV-1 infection risk factor questionnaire collection and HIV-1 seropositive blood samples from IDUs of jails and HIV/AIDS Outpatient department of China medical university hospital in Taiwan annually. (1) About HIV-1subtype surveillance, we collected 2007-2009 HIV-1 seropositive blood samples from HIV-1 surveillance network. Their subtypes were examined by using DNA sequencing and phylogenetic analysis, and applyed coalescent theory, molecular clock analysis to estimated evolution rate and time of the most recent common ancestors (TMRCA). And, the risk factor characteristics of different subtypes were calculated by the biostatistic method. (2) Using 2004 to 2005 risk factor questionnaire database in HIV-1 surveillance network for building up HIV-1 infection risk prediction model for IDUs, we investigated risk factors of HIV-1 infection by case-control study. The risk prediction model was established by logistic regression, and odds ratios (ORs) for each risk factor were summed. Then, we used ROC curve (Receiver operating characteristic curve) to know HIV-1 risk threshold of IDUs. Next, we collected 2007 -2008 data of risk factor questionnaire database to validate the accuracy of this risk prediction model.
Result：(1) HIV-1 subtype surveillance : HIV-1 subtype determined by phylogenetic analysis and nested multiplex PCR. After subtyping, we found three HIV-1 subtypes in our cases, including CRF07_BC, subtype B and CRF01_AE, most of these is CRF07_BC. CRF07_BC originated in 1993.6-1997.3 in China. TMRCA of CRF07_BC from southern, middle and northern were dated to 1998.3-2003.3, 2003.4-2004.4 and 2002.7-2004.3. In HIV-1 subtype and related HIV-1risk factor analysis showed, demographic and risk behavior factos among HIV-1subtype were differenct. (2) For risk factor prevention purpose: In the risk prediction model of IDUs, the significant variables among men were education, syringe sharing, heroin diluention sharing, and knowledge of AIDS- persons often had the venereal diseases, they easier had AIDS, and among women was syringe sharing. The risk threshold value of ORs in men is 7.435 (sensitivity: 0.96; specificity: 0.52). Because only one significant variable among women, the risk threshold value of ORs was not calculated. Finally, we compared with the sum of total ORs for each risk factor and the risk threshold of ORs, then, we could identify the risk of HIV infection. Furthermore, we collected questionnaire in 2007 ~ 2008, and tested this prediction model. The test result is : sensitivity – 72.2% ; specificity – 60.5%. |
