人類乳突病毒感染與子宮頸癌變之分子流行病學研究：人類乳突病毒基因型別、病毒量、嵌入宿主細胞與持續感染對子宮頸癌變之角色探討

Abstract

子宮頸癌為婦女重要的癌症之一，在台灣歷年來向居婦女癌症首位。人類乳突病毒(HPV)被廣泛認為是子宮頸癌的必要因子，但是過去的研究大多以橫斷式或病例對照研究的設計進行，如此一來，病毒感染狀態很難區分是病毒持續感染或暫時性感染，且與子宮頸病變的因果時序性不易辨明。惟以長期追蹤研究與重複採樣的研究設計方有助於闡明HPV感染與其他危險因子誘發子宮頸癌的相關。 本研究係以於1991年間所建立的11923名參與社區性癌症篩檢研究的婦女世代為研究對象，所有自願參加者分別於1991-3及1993-5年兩段期間分別接受邀請參加基線健康檢查及追蹤健康檢查，除了簽署同意書外，由訪視人員依標準化問卷收集社會人口學等基本資料、家族疾病史及多種危險因子的暴露資料，並接受醫師進行子宮頸抹片檢查及子宮頸細胞檢體之採集。疑似罹患子宮頸病變的個案進一步轉診輔以陰道鏡採集組織切片進行確診及例行追蹤。HPV感染之檢測以子宮頸細胞進行HPV DNA及型別檢測。病毒量及病毒嵌入宿主細胞的檢測則以同時參加兩次檢查的對象中且於基線檢查感染HPV16、18、 52及58等四型中任一型者之檢體進行檢測。在追蹤期間，與全國癌症登記系統與死亡登記系統進行資料連結，以獲得子宮頸癌的新發病例。 透過長期追蹤研究設計及HPV感染的重複測量，本研究發現在30-65婦女之HPV盛行率為24.5%，以HPV16, 18, 52, 58與11等型為最常見型別。於抹片正常婦女在平均追蹤1.4年間，以新偵感染任一型別而言，平均新偵感染率為8.4%，兩次檢查感染同一型的平均持續感染率為27.7%。已感染HPV者、初次性交年齡較早者、性交後陰道有灌洗習慣者與離婚或寡居者較易獲致新偵感染；年齡較高、高陰道產次(≥4)、曾使用子宮內避孕器者及已停經者等因素則與持續感染有高度相關。在追蹤15年後，感染HPV或高危險型別HPV的子宮頸癌發生率分別為每十萬人年171及265人，相對危險性則分別是12.8及19.6倍。與台灣婦女子宮頸癌高度相關的主要高危險型別為HPV16,18,52,58，預防這四型的感染估計可減少63%的子宮頸癌，疫苗型別(HPV16/18)則可預防51%。持續感染在子宮頸癌發生自然史中居重要關鍵地位，其危險性高達44.3倍，一旦病毒清除，則其危險性即未達統計顯著(2.4,0.6-9.2)。HPV16/18/52/58之病毒量與持續感染呈有明顯劑量效應，高病毒量(>104 copies/50ng DNA)並可預測子宮頸癌罹癌風險達3.4倍，追蹤期間病毒量降低，後續罹癌風險即明顯降低達十分之ㄧ。病毒量與持續感染及子宮頸癌之相關，在停經後婦女尤著。病毒嵌入與是否患有子宮頸病變具統計相關(p=0.0076)，卻無法於正常婦女驗證其預測子宮頸癌之發生。上述結果將有助於子宮頸癌的預防，早期偵測以接受治療，提昇婦女健康及免於重大疾病之死亡。

Cervical cancer is the second common cancer in the world and has been the leading female cancer in Taiwan over than a decade. Human papillomavirus (HPV) is well documented as the necessary cause of cervical cancer. But most previous studies were based on the cross-sectional case-control design, which can neither differentiate transient and persistent infection nor clarify the causal temporality of risk factor exposure and health outcome. The best way to examine the causation between HPV infection and various cancers should be based on a long-term follow-up study with repeated measurement of HPV infection. In this study, 11,923 women were recruited as cohort members from a community-based cancer screening project (CBCSP) since 1991. Participants received health examinations in two bi-annual cycles in 1991-1993 and 1993-1995. After giving their informed consent, cohort members were personally interviewed according to a structured questionnaire to obtain information on socio-demographical characteristics and history of exposures to various cancer risk factors. Virapaps were used to collect cervical cells. Pap smear and health examination were performed. All women with suspected squamous intraepithelial lesions were further examined by colposcopy-guided biopsy to confirm the diagnosis. They were referred to intensive follow-up examinations every four months. The cervical cell samples and Pap smear were collected at baseline and follow-up were tested for HPV DNA by polymerase chain reaction and genotyping by EasyChip. For cohort members infected with HPV types 16, 18, 52 and 58 further tests on viral load and integration into host genome will be carried out. During follow-up, cases of newly-diagnosed cervical cancers and cervical neoplasia will be ascertained through data linkage with profiles of National Cancer Registry and National Death Certification Registry. Through this long-term follow-up study in CBCSP woman cohort with repeated measurements of HPV infection, we found HPV prevalence was 24.5% and HPV16, 18, 52, 58 and 11 were common types. Among normal women with a mean follow-up duration of 1.4 years, the summarized acquisition rate was 8.4% for any type and the mean of type-specific persistence rate was 27.7%. Women had HPV infection, earlier age at initial coitus, douching use after sexual intercourse, or not currently married had higher risk to acquire newly infection. The determinants for HPV persistence were higher age, high frequency of vaginal delivery, IUD user or post-menopause. With 15-year follow-up, the incidence of cervical cancer for HPV infection and persistence were 171 and 265 per 100000 person-year, the corresponding hazard ratio were 12.8 and 19.6, respectively. HPV16, 18, 52, 58 were the major high-risk types associated with cervical cancer in Taiwan. Around 63% of cervical cancer could be attributed with these 4 types and 51% for HPV16 and/or 18, which vaccine against. In our study, HPV persistence was confirmed the pivotal role in the natural history of cervical cancer with a 44.3-fold risk, once the virus was cleared, the risk was non-significant (HR=2.4, 0.6-9.0). The viral load of HPV16, 18, 52 or 58 associated with persistence in a dose-response relationship and higher viral load (>10^4 copies/ 50ng DNA) was more likely to develop cervical cancer during follow-up; lowering viral load had a protective effect (HR=0.1, 0.03-0.3) with cancer incidence. The integration was associated with cervical neoplasia in the cross-sectional study, however, it couldn’t predict the risks of HPV persistence and cervical cancer in the longitudinal study. The findings will be useful for the primary prevention, early detection and intervention for cervical cancer.