探討血脂異常是否增加活動性肺結核風險:世代追蹤研究

Association between Serum Lipid Profiles and Risk of Active Tuberculosis: A Prospective Cohort Study

劉澄杰(Cheng-Chieh Liu)

指導教授 ： 林先和

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(2019年)

https://doi.org/10.6342/NTU201903904

摘要

背景
血脂異常對於整個結核病的發病與進展仍不清楚;過去的動物模式和一篇流病研究都顯示膳食來源的膽固醇會增加肺結核的易感受性及發生率，實際上膳食來源的膽固醇大約只佔血清膽固醇的1/3，無法全面評估血脂異常對於整個活動性結核病的發病率影響。另一方面，史他汀一類血脂用藥在過去文獻中對於肺結核發病率具有保護的效果，但無法釐清整體的保護的效果是來自於史他汀一類血脂用藥或是血脂的保護。因此,本研究的目的為 (a) 探討四種常見血脂參數對於結核病的發病率影響 (b) 探討在結核病發病族群中四種常見血脂參數對於結核病的嚴重度影響(包含CXR 開洞與否以及塗片陽性的強弱) (c) 釐清血脂參數、史他汀一類血脂用藥和活動性結核病的相關性
方法
研究對象來自新北市社區成人健康篩檢資料 (n=118,097) , 血脂異常的切點參考ATP III之血脂分類標準以及成人代謝症候群診斷標準分為理想濃度、邊際危險濃度、高危險濃度。新發病的活動性結核病的個案定義是細菌學確認並通報至疾病管制署之結核病通報系統。第一，我們利用Kaplan-Meier 存活曲線觀察正常濃度、邊際濃度、高濃度血脂狀態對於結核病發生率的差異;Cox比例風險模式探討邊際和高濃度血脂狀態相對於正常濃度血脂狀態結核病發生率的風險比值，最後以spline 迴歸模式去觀察四種常見血脂參數與結核病風險的劑量效應關係。第二，我們使用單變項邏輯式迴歸去評估在確診結核病族群中，血脂參數與結核病的嚴重度的相關。第三 ,我們使用時間相依共變數調整追蹤期間史他汀用藥紀錄

結果
經過平均8年的追蹤，總共有429名結核病新發個案,發現高LDL-c
(>160mg/dl)和高總膽固醇(>200mg/dl)與結核病發生率存在著反向相關，在Cox比例風險模式中，在調整了其他共變異項後，相較於正常血脂狀態,高LDL-c 族群的結核病風險比值為0.587 (95% CI: 0.417-0.828)，高總膽固醇族群的結核病風險比值為0.623(95% CI: 0.450-0.861) 。我們更進一步針對排除條件進行敏感度分析以及在模式中調整追蹤期間史他汀用藥紀錄，高LDL-c (>160mg/dl)和高總膽固醇與結核病的反向相關仍無改變，呈現一致性
結果。

結論
高LDL-c和高總膽固醇的受試者罹患結核病的風險分別低於正常血脂42%和38%，但由於血脂狀態與結核病之間的相互作用是複雜的，需要更多研究進一步去了解血脂如何調節宿主對抗結核病。

關鍵字

結核病 ； 血脂異常 ； 史他汀 ； 世代追蹤研究 ； 劑量效應關係 ； 時間相依共變數

並列摘要

Introduction
The impact of dyslipidemia on tuberculosis incidence and progression remains unclear and has potential implications contributed to TB control strategy. Previous studies from animal studies and one observational study revealed that dietary cholesterol may increase TB susceptibility and incidence rate. However, dietary cholesterol could not fully represent the association between serum lipid status and TB.
On the other hand, the cholesterol-lowering drugs, statins, showed a protective effect on reducing TB incidence without differentiating whether the protective effect was derived from drugs or lipid profiles. Therefore, evaluating the serum lipid effect on TB incidence beyond the independence of statin effect would be essential.
Methods
We conducted a cohort study from a community-based health screening program in northern Taiwan from 2005 to 2008, including a total of 118,097 participants. Serum lipid profiles including triglycerides, high-density lipoprotein cholesterol (HDL-c), low-density lipoprotein cholesterol (LDL-c) and total cholesterol were ascertained at baseline. The occurrence of active tuberculosis was confirmed from the National Tuberculosis Registry. First, Kaplan-Meier curves by different lipid profiles were produced to compare the time to TB. Next, Cox proportional hazards regression analyses to estimate the hazard ratios (HRs) between lipid profiles and active TB. Later, spline regression was utilized to investigate the dose-response relationship. Second, the crude analysis was conducted to assess the association between lipid profiles and severity of TB. Eventually, we adjusted follow-up statin as a time-varying covariate in cox-model.

Results
After a median follow-up of 8 years, 429 cases of confirmed TB occurred. An inverse association was observed between LDL-c, total cholesterol, and incidence of TB (Table 2). After adjusting for other potential covariates, high LDL-c group (>160mg/dl) compared with normal LDL-c group (<100mg/dl) was associated with a lower risk of active TB (adjusted HR: 0.587, 95% CI: 0.417-0.828 ). Similarly, High total cholesterol group (>240mg/dl) also had a lower risk of active TB, with adjusted HR of 0.623 ( 95% CI: 0.450-0.861 ). On the other hands, serum triglycerides and HDL-c were not associated with risk of active TB. With further sensitivity analysis on exclusion criteria and adjustment of the follow-up statin usage, high LDL-c (>160mg/dl) and high total cholesterol remain inverse association with TB, consistently.
Conclusion
In this community-based cohort study, elevated serum cholesterol level and high LDL-c were associated with a decreased risk of active tuberculosis. Limitations of the research included a single measurement of serum lipid at baseline and lack of further information on latent TB. Further analyses should aim to examine the interplay of statin treatment, lipid profiles, and risk of TB.

並列關鍵字

tuberculosis ； dyslipidemia ； dose-response relationship ； time-varying covariate

參考文獻

1.World Health Oraniganization. Global tuberculosis report 2018 Geneva: World Health Organization; 2018 [cited 2018 Sep 18]. Available from: https://apps.who.int/iris/bitstream/handle/10665/274453/9789241565646-eng.pdf.

2.World Health Oraniganization. End TB brochure Geneva: World Health Organization; 2015. Available from: https://www.who.int/tb/End_TB_brochure.pdf?ua=1.

3.Narasimhan P, Wood J, Macintyre CR, Mathai D. Risk factors for tuberculosis. Pulmonary medicine. 2013;2013:828939-. Epub 2013/02/12. doi: 10.1155/2013/828939. PubMed PMID: 23476764.

4.Lonnroth K, Williams BG, Cegielski P, Dye C. A consistent log-linear relationship between tuberculosis incidence and body mass index. Int J Epidemiol. 2010;39(1):149-55. Epub 2009/10/13. doi: 10.1093/ije/dyp308. PubMed PMID: 19820104.

5.Lin HH, Wu CY, Wang CH, Fu H, Lonnroth K, Chang YC, et al. Association of Obesity, Diabetes, and Risk of Tuberculosis: Two Population-Based Cohorts. Clin Infect Dis. 2018;66(5):699-705. Epub 2017/10/14. doi: 10.1093/cid/cix852. PubMed PMID: 29029077; PubMed Central PMCID: PMCPMC5850624.