調整腰臀圍比的身體質量指數和高密度脂蛋白膽固醇：雙樣本孟德爾隨機化分析

WHR-adjusted BMI and High Density Lipoprotein Cholesterol: Two-Sample Mendelian Randomization

吳亦舒(Yi-Shu Wu)

指導教授 ： 沈志陽

共同指導教授 ： 王姿乃(Tsu-Nai Wang)

高雄醫學大學/健康科學院/公共衛生學系公共衛生學碩士班/碩士(2018年)

摘要

研究背景與目的：肥胖與血脂異常是已開發國家常見的公共衛生議題。現有的證據顯示肥胖對心臟代謝特徵有不利的影響，繼而可能引發心血管疾病等相關疾病風險。公認且最常使用的肥胖定義是身體質量指數(BMI)，但脂肪分布在個體間存在很大的異質性，腹部內脂肪的含量與多項心臟代謝特徵息息相關。本研究的目的是利用遺傳變異，探討台灣漢人其調整腰臀圍比的身體質量指數(WHR-adjusted BMI)與高密度脂蛋白膽固醇(HDL-C)的因果關係。
研究方法：本研究為雙樣本孟德爾隨機化分析。使用GWAS catalog彙整發表文獻中，且以亞洲族群發現與肥胖指標相關的單核苷酸多型性(SNP），在台灣人體生物資料庫16000位參與者中進行兩次分析，包括第一次的發現分析及第二次的驗證分析。本研究以5個SNPs所計算的多基因風險評分(GRS)作為暴露的工具變項，進行以台灣人群為基礎的WHR-adjusted BMI與HDL-C因果關係探討。在研究中分別驗證孟德爾隨機化(Mendelian Randomization Analysis, 簡稱MR)的主要假設，並進行因果效應估計，因果效應的檢驗及檢驗多效性。
研究結果：台灣人體生物資料庫參與樣本經過品質控制後有14568人，595646點位進入研究分析中。經由回歸分析發現，工具變項每增加1-SD則HDL-C降低5.95 mg/dl(95%CI，-10.55~-1.35；P值: 0.0112)，並具有劑量效應關係(P值: 0.0162)；MR分析的三個方法中，都發現WHR-adjusted BMI對HDL-C為負相關。多基因風險評分作為工具變項的F估計值為241.8，符合假設一；工具變項與潛在的干擾因子皆無顯著相關(P值>0.05)，符合假設二；MR-Egger的多效性檢驗(P值: 0.076)，沒有發現多效性的證據，符合假設三。且多基因風險評分對台灣人體生物資料庫的15項疾病的相關皆未具有統計上顯著意義，為獨立於多效性的假設提供了間接的支持。
結論：WHR-adjusted BMI 的遺傳傾向與HDL-C的降低因果有關。這個結果可提供台灣人群在公共衛生上以WHR-adjusted BMI來做為調控HDL-C的證據。

關鍵字

孟德爾隨機化 ； 身體質量指數 ； 高密度脂蛋白膽固醇

並列摘要

Background and purpose: Obesity and dyslipidemia are common public health issues in developed countries. Evidence shows that obesity has an adverse effect on the cardiometabolic traits, leading to the development of cardiovascular diseases and other related diseases. The most commonly used definition of obesity is the body mass index (BMI), but fat distribution in organs is highly heterogeneous among individuals-Therefore the intra-abdominal fat is closely related to cardiometabolic traits. The purpose of this study was to use genetic variants to explore the causal relationship between WHR-adjusted BMI (waist-hip ratio) and high-density lipoprotein cholesterol (HDL-C) in Han Chinese in Taiwan.
Methods: This study is a Two-sample Mendelian Randomized analysis. We used the single nucleotide polymorphism (SNP) associated with BMI, which was selected from the GWAS catalog and was identified specifically significant in Asian populations. Based on the participants of Taiwan Biobank, we performed these analyses, which were separated into two phases, the first discovery phase and second verification phases. In this study, the polygenic risk score (GRS) calculated by 5 SNPs of BMI was used as instrumental variables (IV) to represent exposure, by which we investigated the WHR-adjusted BMI with HDL-C causality relationship. In this study, the assumption of Mendelian Randomization (MR) was examined, and the causal estimation, causal effect test and the pleiotropy effect were tested.
Results: Taiwan Biobank participants after quality control (QC)--there were 14,568 participants and 595,646 SNPs included in the analysis. The study shows that linear regression analysis demonstrated one-standard deviation (SD) increase in IV was associated with a decreased HDL-C of 5.95 mg/dl (95% CI, -10.55 to -1.35；P value: 0.0112), furthermore, there was a dose-response relationship (P value, 0.0162). Among three methods of MR analysis, WHR-adjusted BMI was found to be negatively correlated with HDL-C. That the GRS was suitable as an instrumental variable, as indicated by the F-test value, 241.8, (F>10 is considered significant), agreeing with the assumption 1; There was no significant correlation between IV with potential confounders, including smoking status, alcohol consumption, sport habit and education (P> 0.05), agreeing with the assumption 2. MR-Egger intercept showed a P value of 0.076, meaning that no evidence of pleiotropy effect was found, agreeing with the assumption 3. In addition, the association between GRS and medical history of fifteen diseases collected by the Taiwan Biobank was examined, and there was no statistically significant, indirectly supporting independent for pleiotropy effect.
Conclusions: The genetic predisposition of WHR-adjusted BMI is causally associated with reduction of HDL-C. This result provides evidence for public health implication to control WHR-adjusted BMI to regulate HDL-C.

並列關鍵字

Mendelian Randomization ； BMI ； HDL-C

參考文獻

1. worldobesity, Obesity.

2. 衛生福利部國民健康署, 肥胖防治網.

3. Ng, M., et al., Global, regional, and national prevalence of overweight and obesity in children and adults during 1980-2013: a systematic analysis for the Global Burden of Disease Study 2013. Lancet, 2014. 384(9945): p. 766-81.

4. Seidell, J.C., et al., Regional distribution of muscle and fat mass in men--new insight into the risk of abdominal obesity using computed tomography. Int J Obes, 1989. 13(3): p. 289-303.

5. Ashwell, M., et al., Obesity: new insight into the anthropometric classification of fat distribution shown by computed tomography. Br Med J (Clin Res Ed), 1985. 290(6483): p. 1692-4.