台灣HIV感染者合併結核病感染之流行病學與防治策略研究

A study in epidemiology and preventive strategy of Taiwan HIV-infected persons coinfected with tuberculosis

楊靖慧(Chin-Hui Yang)

指導教授 ： 邱弘毅

臺北醫學大學/公共衛生暨營養學院/公共衛生學系暨研究所/博士(2014年)

摘要

研究目的
結核病(TB)與人類免疫不全病毒感染(HIV) 列名為全球最重要的三大傳染病之一。在2012年，全球新增110萬個HIV與TB合併感染者(佔TB新感染人數的13%)，而有30萬人因此而死亡(佔TB死亡人數的23%)。因此，世界衛生組織建議針對此二疾病，要合併防治策略，加強對HIV感染者的TB感染進行監測、治療與預防策。台灣自2006年起推動「結核病十年減半全民動員計畫」，結核病新增個案數每年以7%的幅度下降，屬於TB中度盛行率的國家。而我國之新通報與累積存活的HIV感染人數則呈現逐年增加的趨勢，雖然仍屬於HIV低度盛行率，但是HIV與TB合併感染的風險也隨之增加。台灣此種特殊的流行疫情模式，對HIV與TB合併感染的影響，有待進一步評估。與非感染者相比，HIV感染者有較高的TB死亡率，其相對危險值依不同國家及不同研究統計約為4-10倍。而且HIV感染者開始結核病治療後，15-45%的個案可能會出現免疫重建發炎症候群(immune reconstitution inflammatory syndrome, IRIS)，嚴重時甚至可能造成死亡。在TB治療期間越早開始HAART越容易發生IRIS。但是，一些觀察性研究也發現，在TB治療期間，若延遲使用HAART，可能會增加HIV-TB合併感染者的死亡率。所以，HAART在TB治療期間的開始時機，是目前世界衛生組織認定需要探討的重要課題之一。除了HAART外，另一個可以降低HIV感染者TB發生率的策略是潛伏性結核感染(latent tuberculosis infection, LTBI)的主動篩檢暨治療。但是，台灣因為卡介苗(BCG)的高接種率，使得利用皮膚結核菌素檢查(TST)來診斷LTBI的特異性較不可靠。因此，針對臺灣的HIV感染者，何種LTBI篩檢流程較為合適仍待進一步研究。
本研究的第一個目是要比較台灣HIV感染者與非感染者的TB治療預後，以及與TB死亡率相關的因子。第二個目的是希望了解HIV感染者發生TB後，HAART使用時機與TB死亡率的相關性。第三個目的是要比較TST與T-SPOT.TB(一種IGRA檢驗)在台灣用以診斷HIV感染者LTBI的可行性，及其對未來TB發病的預測性。並且對有LTBI的HIV感染者進行IPT，觀察此治療的安全性與有效性。本研究的重要性是希望提供一個完整的台灣HIV與TB合併感染的流行病學調查，並且評估LTBI主動篩檢暨治療策略的效益，以提供公衛部門制定政策的的參考。

研究方法：
第一部分是回溯性世代研究，利用TB與HIV通報資料庫與健保資料庫勾稽，擷取2002/1/1至2007/12/31新通報之結核病個案為研究對象，進一步與健保資料檔中有進行HIV檢查醫令者及HIV通報資料庫勾稽，確認TB個案的HIV狀態，分成HIV染者、非感染者與HIV未知者三組，比較不同HIV狀態個案之一年TB死亡率以及相關因子。第二部分亦是回溯性世代研究，勾稽2007年以前的TB與HIV通報個案資料庫以確認HIV與TB合併感染個案，經由調閱HIV與TB合併感染個案的病歷，進行臨床資料與病程的收集，比較與個案一年TB死亡率的相關因子，包括HAART使用時機等。利用Cox Proportional Hazard model分析不同變項(HIV診斷年齡、性別、HIV危險因子、HAART與否、HAART使用情形、HIV診斷時的CD4+淋巴細胞計數等)在TB死亡率的相對風險(Hazzard ratio, HR)。先進行單變項分析，以alpha = 0.1決定該變項是否納入多變項分析。利用Kaplan-Meier 存活分析與log-rank test來比較各組別死亡率。所有的統計分析均為雙尾， P值小於0.05視為顯著。
第三個部分是追蹤研究，招募HIV感染者，同時進行TST及T-SPOT TB檢查。招募時間自2008年1月1日至2010年11月30日止。參與者需簽署書面知情同意書，填寫標準化問卷(包括社會人口學背景和臨床資料)，HIV相關資料由審查病歷取得。個案須經由醫師診察排除為活動性結核病者始納入。TST 檢查是以 PPD 2TU皮內注射於左前臂掌側中段，48至72小時之間判讀結果，硬節橫徑超過5公厘視為陽性。TST陽性者會給予免費每天INH 300mg共九個月的治療(IPT)。利用多變項Logistic 回歸模式來評估與TST陽性，T -SPOT.TB和雙陽性結果相關的因素。主要的觀察結果是活動性結核病的發生率，所有參與者(含未IPT者)會定期追蹤至2012年4月30日(觀察截止日)或個案死亡或個案發生活動性結核，以最早發生的日期為結束日期。利用Poisson distribution來計算未接受IPT個案發生活動性結核病的相對危險性，用以評估TST和T-SPOT.TB的表現，並評估IPT的保護效果及安全性。

研究結果
第一個部分：2002年至2007年間共計有94,919 位TB新通報個案，其中僅有17.3%的個案曾做過HIV篩檢，確認HIV感染者有568位(0.6%)，非HIV感染者有15,848位(16.7%)。HIV感染者的年齡較低(30-49歲佔六成)，但在非感染者與未知者，年齡超過50歲者占六成以上。其中共計17,541人(18.5%)於TB診斷後一年內死亡，在不同HIV狀態分組間沒有顯著差異。但是若分年齡層來看，HIV感染者的每個年齡層的TB死亡率均顯著高於非感染者和HIV未知者，尤其是在小於50歲的個案。HIV感染者TB治療成功的比例顯著較低，而失落率則正好相反，顯示HIV感染者的TB治療預後顯著較差(p<0.05)。在多變項分析中，TB死亡率會隨著年齡增長而增加。在調整年齡、性別以及TB臨床表現等變項後，HIV感染者比非感染者有較高的風險(2.3至3.7倍)在TB診斷後一年內死亡。即使HIV感染者有使用HAART，仍是有顯著較高的死亡風險。在年齡小於50歲的TB個案中，HIV感染對死亡率有極大的影響力，aHR達3.52(95% CI 2.70-4.59)。無論有否使用HAART，其預後均較非感染者/HIV未知者顯著較差。但在年齡大於50歲者，影響就不顯著。
第二個部分：至2006年底累計新通報HIV感染者13,103人中有473人(3.6%，498人次)發生TB感染。臨床表現上，肺外結核十分常見，佔52.8%。HIV感染者在TB診斷後12個月的治療結果如下：治療成功率為69.5%，死亡率為26.9%。所有死亡個案中有64.5%被判定為結核病相關死亡。多變量分析發現年齡增加與CD4+細胞計數減少與TB死亡率呈現顯著相關。TB診斷時已經使用HAART者，比起未使用HAART者在單變項分析時有顯著較好的存活率，但調整其他變項後，未達統計顯著差異。至於在TB診斷時未使用HAART之HIV感染者，除了年齡增長和較低的CD4+淋巴細胞計數外，在抗結核治療期間沒有開始HAART與TB死亡率有顯著相關。不同HAART使用時機之TB年死亡率依高低順序排列如下：在未使用HAART者、TB治療60天後、 15天內與16到30天內，31到60天內無人死亡。合併HCV感染者比起HCV陰性或未知者，有較高的機會無法在TB治療期間開始HAART (p<0.01)，但HBV合併感染則無此現象。使用HAART個案中有31%發生IRIS，發生IRIS者並未伴隨較高的死亡率。30天內開始HAART的個案比30天後才開始者有顯著較高的機會發生IRIS，36.8% 對 10.5%， p<0.05。
第三個部分：總共有909個HIV感染者同時完成TST 與 T-SPOT.TB兩種檢查且結果可判讀。LTBI篩檢結果如下列：TST檢查陽性率(TST≥5公厘)為25％， T-SPOT.TB陽性率為15％，雙陽性率為9%。CD4 +淋巴細胞計數增加與結核病接觸史兩項因素，是TST、T-SPOT.TB與同時檢驗陽性的獨立預測因子。年齡增加與男性與T-SPOT.TB檢驗陽性有關，但是與TST則無相關。監獄停留超過6個月與TST陽性反應有關，但是與T-SPOT.TB則無關。兩項測試的kappa統計量數為0.32，一致性為78.1%，表現只能算尚可。IPT治療結果如下，82人(92.1%)完成服藥，有11人因副作用中斷服藥，有10人(11.2%)發生急性肝炎(肝功能超過正常值5倍)，但僅有1人(1/89, 1.1%)因肝功能過高而永久停藥。所有參加者追蹤2.97年後，有5人(0.55%)發生活動性結核病，TB發生率為0.17例/100人年(95％CI 0.003-0.29/100 PY)，此5人之TST和T-SPOT.TB檢查均為陽性結果。根據Poisson regression分析，以陰性結果對照，TST陽性、T-SPOT.TB陽性和雙陽性結果的受試者，發生結核病的相對危險性各為41倍、74倍和227倍。假設IPT可提供完全的保護力，要預防一個後續的TB個案，需要IPT治療的TST陽性、T-SPOT.TB陽性和雙陽性個案人數分別為35，22和8人。而在TST與T-SPOT.TB檢查雙陽性的參與者中，未接受IPT的32人中有4人發生活動性結核病，而接受IPT的49人中僅有1人發病，(log-rank檢驗 p <0.05)。由此估算TST與T-SPOT.TB檢查雙陽性者，未接受IPT的個案比接受IPT者增加7.8倍活動性結核病的發生風險。
結論
根據我們的研究結果發現，即使在已經有HAART治療的時代，HIV感染者仍比非感染者有較高的結核病死亡風險。要降低HIV感染者的TB死亡率，一方面要從加強HIV感染防治計畫著手，包括積極篩檢、個案管理等策略，使感染者可以適時開始HAART治療，如此不但能降低TB的發生率，亦可降低HIV與TB合併感染的死亡率。另一方面，應該加強TB病人的HIV篩檢率，使未及時開始HAART治療的HIV-TB合併感染者能盡快開始治療，才能有效降低死亡率。至於不同HAART治療時機(TB治療開始0-15天、16-30天、31-60天與60天後)與TB一年內死亡率則沒有顯著的關係。但是早期開始HAART(30天內)會伴隨著增加IRIS發生的風險。因此在TB治療後30到60天之間在開始HAART，可以平衡因治療引起的風險與利益，應是較可行的方式。此外，降低TB發生率的另一個策略為LTBI篩檢與IPT，採用TST和T-SPOT.TB陽性結果來篩檢HIV感染者的LTBI是可行的，若使用雙陽性結果作為IPT治療標準，可以降低需要IPT治療的人數，減少因而產生的副作用處理，亦能有效降低未來的TB發病率，進一步阻斷新的傳染鏈。因此，我們建議針對所有的HIV感染者，發生TB時要適時使用HAART，可以有效降低死亡率；而未發生TB時，進行LTBI主動篩檢並且給予IPT治療，可以有效降低TB發生率。此二策略並行，應可有效降低台灣HIV與TB合併感染的人數與相關死亡。

關鍵字

愛滋病 ； 結核病 ； 死亡率 ； 潛伏結核感染

並列摘要

Purpose
Tuberculosis (TB) and human immunodeficiency virus (HIV) are both the most important infectious diseases worldwide. According to the estimation by World Health Organization (WHO), there were 1.1 million HIV-positive new TB cases globally in 2012 and 320,000 people died of HIV-associated TB. Therefore, the WHO recommends adopting collaborative HIV/TB activities and strengthens the interventions needed to prevent, diagnose and treat TB in people living with HIV. After implementation of the “Mobilization Plan to Halve Tuberculosis Incidence in Ten Years” (national TB plan) since 2006, the number of new TB cases declined 7% annually, although Taiwan still belong to moderate TB burden country. On the contrary, the new and cumulative HIV-infected cases showed a rising trend, although still a low prevalence of HIV, but the risk of HIV and TB co-infection is increasing. What is the impact of such unique epidemiology on the HIV and TB co-infection needs further analysis. Persons co-infected with TB and HIV are 29.6 times (27.1 – 32.1) more likely to develop active TB disease than persons without HIV, and the risk of death are also higher among co-infected persons , 4-10 times, than persons without HIV. About 15 to 45% of co-infected patients would develop immune reconstitution inflammatory syndrome (IRIS) and some might resulted in death, make the treatment of TB complex in HIV-infected persons. Early initiation of HAART during TB treatment is strongly associated with the occurrence of IRIS. However, several observational studies found that deferral of HAART in TB treatments are associated with higher mortality. Therefore, the timing of HAART initiation during TB treatment is one of the important issues currently WHO identified the need to explore. Another important strategy to reduce the incidence of HIV-infected TB is screening and treatment of latent tuberculosis infection (LTBI). The tuberculosis skin test (TST) is the traditional method for the diagnosis of LTBI, but the test has low specificity due to cross reactions with the bacillus Calmette-Gue’rin (BCG) vaccine. Due to the low specificity, Tuberculin skin test (TST) may not be a good tool to screen LTBI in Taiwan due to high BCG vaccination coverage rate. Therefore, the appropriate LTBI screening protocol for HIV-infected persons in Taiwan need further study.
The first aim of this study is to compare the treatment outcome and fatality rate of TB cases in Taiwan, stratified by HIV sero-status, and further to elucidate risk factors related to death. The second is to understand the epidemiology of HIV and TB co-infection in Taiwan and hope to contribute to the understanding of the appropriate timing to initiate HAART in co-infected patients. The third is to evaluate the T-SPOT.TB (an IGRA) and the TST, for the diagnosis of LTBI and the predictive value of the development of subsequent active tuberculosis. We also offer isoniazid preventive therapy (IPT) for LTBI patients and observe the safety of the treatment. We tried to determine the clinical and epidemiological risk factors associated with positive results for the TST and the T-SPOT.TB test in HIV-infected individuals and recommend an appropriate protocol to better select HIV-infected individuals for IPT.

Methods:
The first part is retrieved from national TB registry dataset, collected the data of reported new TB cases from 2002/1/1 to 2007/12/31 for analysis. We further linked with HIV testing data from national health insurance dataset and national HIV registry data to classify the HIV status of TB patients as HIV-infected, HIV-uninfected and unknown HIV status. The second part was a retrospective study, through linkage of the HIV and TB registry in Taiwan CDC before 2007 to identify HIV/TB co-infected patients. All the medical and microbiologic records were reviewed by physician to obtain clinical information which included date of diagnosis; type of TB; antiretroviral drugs used before and after the diagnosis of TB; and the CD4+ lymphocyte and viral load closest to the date of TB diagnosis. The outcome of interest studied was all-cause mortality within 1 year following TB diagnosis and treatment. Cox Proportional Hazard model was used to evaluate the hazard ratio of variables (e.g. age, gender, HAART initiation timing, etc.) associated with one-year TB mortality between the groups. Variables associated with a p value < 0.15 were retained in the final multivariate model and use backward stepwise selection method. Results are expressed as hazard ratios (HRs) with 95% confidence intervals (CIs). All comparisons were two-tailed and p value <0.05 was considered statistically significant.
The third part is to recruit HIV-infected individuals without clinical suspicion of active TB or a past history of TB from 1 January 2008 to 30 November 2010. Both T-SPOT.TB test and TST were offered to the participants whom were followed up prospectively until April 30, 2012 for development of TB. Multivariate logistic regression model was used to determine the risk factors for positive results. The concordance between the TST and T-SPOT.TB results was assessed using Cohen’s kappa (κ) coefficient. IPT was offered to recruited participants with positive TST reactions with free isoniazid, 300 mg daily for 9 months. All enrolled subjects were routinely followed up every 3-6 months in the HIV clinics until April 30, 2012; death; or the development of active TB, whichever was earliest. We used the Poisson distribution to calculate the relative risks (RRs) of incident TB between different tests and IPT or not.

Result
The first part: Among the 94,919 reported new TB cases between 2002 and 2007, 17.3% among them ever received 568(0.6%) were HIV-infected and 15,878(16.7%) were HIV-infected. Around 60% of HIV-uninfected and unknown HIV status TB patients were aged over 50 years old, but HIV-infected TB patients were significantly younger and over 60% were less than 50 years old. A total of 17,541 people died within one year after TB diagnosis, the overall TB mortality are 18.5% and no significant differences between the three HIV sero-status groups. However, HIV-infected individuals had significantly higher mortality rate when compared with HIV-uninfected and unknown HIV status in each age groups. HIV-infected cases had lower successfully treated and higher default rate, p<0.05. In multivariate analysis after adjustment with age, gender and sputum smear and culture status, HIV infection had a significantly higher risk of death than HIV-uninfected and HIV unknown status and the aHR was 3.70 (95% CI 3.01-4.51). Even those who have HAART, there is still a significantly higher risk of death and aHR was 3.15 (95% CI 2.48-4.00). In TB cases less than 50 years old, HIV infection was the most important factor on one-year TB mortality, aHR of 3.52 (95% CI 2.70-4.59), though the impact in aged over 50 years old was not significant.
The second part: Among the 13,013 reported HIV-infected persons in Taiwan by the end of 2006, 473 patients (3.6%, 489 episodes) developed TB infection. Extra-pulmonary involvement was common and accounted for 52.8%. Twelve month TB treatment outcomes of HIV infected individuals included 69.5% was successfully treated and 26.9% died. There are 64.5% among deaths were interpreted as tuberculosis-related deaths. On multivariate analysis, only increasing age at TB diagnosis and decreasing CD4+ lymphocyte counts was associated with TB mortality. Already on HAART at TB diagnosis showed better survival than those not on HAART, though not significant. In cases not on HAART during TB treatment, in addition to decreasing age and higher CD4 lymphocyte count, ever start HAART during TB treatment was significantly associated with better survival (adjusted Hazard Ratio was 0.11, 95% CI 0.06-0.21). As to the timing of HAART initiation, the TB mortality rate was highest among cases initiating HAART after 60 days of TB treatment, followed by within 15 days, 16-30 days and 31-60 days. Cases with HCV co-infection had significantly lower chance to initiate HAART during TB treatment than cases without or had unknown HCV status (p < 0.01); HBV co-infection did not show such findings. One third of the patients developed IRIS after initiation of HAART. Cases initiated HAART after 30 days had lower risk to develop IRIS than cases initiate HAART earlier. Cases with IRIS had significant higher rate of re-hospitalization (49% vs. 4%, p<0.001) and prolonged hospitalization (28 days vs. 18.5 days, p<0.01).
The third part: Among the 909 participants, 25% had positive TST reactions with cut-off point of 5 mm and 15% had positive TSPOT.TB and 9% had dual positive results. Increasing CD4+ lymphocyte counts and a history of contact with TB case were independently associated with a positive response in both tests. Increasing age and male gender independently predicted a positive T-SPOT.TB response, but not TST. Prison stay for over 6 months was associated with positive TST response. The overall agreement between the two tests was only fair (78.1%, κ = 0.32). There were 82 participants (92.1%) completed IPT, 11 persons ever interrupted due to adverse events and 10 persons (11.2%) experienced acute hepatitis, but only one participants (1.1%) discontinued IPT permanently. After a median follow-up of 2.97 years, there were 5 cases developed culture-confirmed active TB (all had dual positive TST and T-SPOT.TB results), and the incidence was 0.17 per 100 person-years. The relative risks (RRs) for subsequent active TB in HIV-infected individuals with positive TST results, positive T-SPOT.TB results and dual positive results compared with the risk for individuals with negative results were 40.6 (95% CI 2.1–767.9), 73.9 (95% CI 3.9–1397.7) and 226.5 (95% CI 12.0–4284), respectively. Among patients with dual positive results, there were 4 active TB cases among 32 untreated patients, compared with 1 TB case among 49 patients taking IPT (p<0.05 by the log-rank test), Dual positive TST and T-SPOT.TB results were associated with a 7.8-fold increased risk of active TB in patients who did not receive IPT.

Conclusion
Even in the era of HAART, HIV-infected individuals still experienced higher one-year TB mortality. In order to reduce the TB mortality rate of HIV-infected individuals is implementing HIV prevention program which include aggressive HIV screening program and case management program, thus enable HIV–infected persons can get HAART in time. As to persons not able to initiate HAART at TB diagnosis, we recommended that HAART should initiate during TB treatment. Early initiation of HAART during TB treatment demonstrated a non-significant better one-year survival than late initiation (60 days after TB treatment). Nevertheless, early initiation of HAART within 30 days appeared to increase the risk of IRIS. Deferring HAART to 31–60 days of TB treatment might be optimal after considering the risks and benefits. In addition, adopting positive results of the TST and T-SPOT.TB to screen LTBI among BCG-vaccinated HIV-infected individuals might be feasible. Number needed to treat for isoniazid therapy could be reduced significantly by using the dual positive strategy, thus can reduced adverse events and still can lower the future incident TB cases, further blocking the transmit ion. Therefore, we recommended all HIV-infected persons to initiate HAART timely when developed TB, can effectively reduce mortality; and screening for LTBI and offer IPT for cases without TB to reduce the incidence of TB. These two strategies in parallel could reduce the number of HIV and TB co-infection and related deaths in Taiwan.

並列關鍵字

HIV ； TB ； Mortality ； Latent tuberculosis infection

參考文獻

Abdool Karim, S. S., Naidoo, K., Grobler, A., Padayatchi, N., Baxter, C., Gray, A. L., et al. (2011). Integration of antiretroviral therapy with tuberculosis treatment. N Engl J Med, 365(16), 1492-1501.

Aichelburg, M. C., Rieger, A., Breitenecker, F., Pfistershammer, K., Tittes, J., Eltz, S., et al. (2009). Detection and prediction of active tuberculosis disease by a whole-blood interferon-gamma release assay in HIV-1-infected individuals. Clin Infect Dis, 48(7), 954-962.

Akksilp, S., Karnkawinpong, O., Wattanaamornkiat, W., Viriyakitja, D., Monkongdee, P., Sitti, W., et al. (2007). Antiretroviral therapy during tuberculosis treatment and marked reduction in death rate of HIV-infected patients, Thailand. Emerg.Infect Dis, 13(7), 1001-1007.

Akolo, C., Adetifa, I., Shepperd, S., & Volmink, J. (2010). Treatment of latent tuberculosis infection in HIV infected persons. Cochrane.Database.Syst.Rev.(1), CD000171.

Allen, S., Batungwanayo, J., Kerlikowske, K., Lifson, A. R., Wolf, W., Granich, R., et al. (1992). Two-year incidence of tuberculosis in cohorts of HIV-infected and uninfected urban Rwandan women. Am Rev.Respir Dis, 146(6), 1439-1444.