郭柏秀Yong-Sheng Wong翁咏聖2024-09-262024-09-262024https://ntu.primo.exlibrisgroup.com/permalink/886NTU_INST/14poklj/alma991039261578604786https://scholars.lib.ntu.edu.tw/handle/123456789/721569指導教授：郭柏秀Research has suggested that cannabis use is a modifiable behavioral factor associated with anxiety, which in turn may influence cannabis use. However, the causality between cannabis use and anxiety is yet to be elucidated. We conducted a bidirectional, two-sample Mendelian randomization investigation leveraging the latest publicly available genome-wide association study summary statistics to explore potential causal relationships between lifetime cannabis use (LCU), cannabis use disorder (CUD), and anxiety disorders among individuals of European ancestry. Using the inverse-variance weighted (IVW) multiplicative random-effects model, with the median- and mode-based estimates, MR-Egger, MR-PRESSO, and other supplementary analyses, positive causal effects were found in several tests: CUD on anxiety disorders (ORIVW = 1.088, 95% CI: 1.018-1.163, p-value = 0.013), CUD on social phobias (ORIVW = 1.290, 95% CI: 1.062-1.567, p-value = 0.010), and anxiety disorders on LCU (ORIVW = 1.166, 95% CI: 1.007-1.351, p-value = 0.040). Our findings align with existing epidemiological evidence, but on the whole, this study alone may not offer compelling evidence to confirm the causal associations between cannabis use and anxiety. Discrepant results with the literature may arise from confounding or reverse causation in observational studies, as well as differences between genetic instrumental variable analysis and clinical trials or community interventions. To boost the power to detect causality and improve the generalizability of research findings, future efforts should be made for the richness and diversity of omics data. This includes identifying more specific genetic markers linked to cannabis use or anxiety with biomedical implications. Most importantly, evidence synthesis and triangulation, along with critical appraisal of causal fallacies, will provide better guidance for substance use issues and mental health practices.研究指出大麻使用是與焦慮有關的可改變行為因素，焦慮也可能反過來影響大麻的使用，但兩者之間的因果性仍有待商榷。我們利用最新公佈且可直接取得的全基因體關聯性研究匯總數據，對歐裔血統人群終生大麻使用、大麻使用障礙，以及焦慮之間的潛在因果關係執行了一項雙向、雙樣本孟德爾隨機化調查。藉由變異數倒數加權的相乘隨機效應模型，輔以基於中位數和眾數的估計、MR-Egger、MR-RAPS、MR-PRESSO等穩健統計方法，我們在大麻使用障礙對焦慮症（ORIVW = 1.088，95%信賴區間：1.018-1.163，p值= 0.013）、大麻使用障礙對社交恐懼症（ORIVW = 1.290，95%信賴區間：1.062-1.567，p值= 0.010），以及焦慮症對終生大麻使用（ORIVW = 1.166，95%信賴區間：1.007-1.351，p值= 0.040）的幾種測試中找到顯著的正向因果效應。我們的發現與既有流行病學證據相符，但整體而言，僅憑本研究似乎無法提供強而有力的證據來確認大麻使用對焦慮，或是焦慮對大麻使用之因果相關。部分結果與文獻不一致，可能源自於觀察性研究的干擾作用或反向因果，以及遺傳工具變項分析和臨床試驗或社區介入的差異。未來的工作需要擴大體學資料的豐富度及多樣性，探尋與大麻使用或焦慮在生物醫學意義上更為具體的位點，以提昇檢驗因果關聯的能力，並增進研究成果類推的適用性。最重要的是，對證據的混合比對，加上對因果謬誤的思辨，將更好地為物質使用議題及心理健康實踐指引方向。Mendelian randomization孟德爾隨機化cannabis大麻anxiety焦慮causal inference因果推論genetic instrumental variables遺傳工具變項thesis