Leveraging electronic medical records to perform large-scale diabetes pharmacogenomics among ancestrally diverse patient populations

摘要 糖尿病是当今的祸害，在全球范围内影响着越来越多的人， 包括目前的2600万美国人。此外，2型糖尿病(T2D)对 历史上处于不利地位的美国少数群体，从更高的发病率和更多的 非洲裔美国人的严重并发症。尽管有多种治疗类别 口服药物可用于治疗T2D，目前推荐将二甲双胍作为一线治疗。 二甲双胍通过减少肝脏糖异生、改善骨骼肌胰岛素而降低血糖水平 敏感性和限制肠道葡萄糖摄取。它也被证明是一种有效的治疗方法 预防突发糖尿病。尽管是全球最常用的处方药之一，但很少有 已知的生物机制(S)，通过其介导其作用的二甲双胍。这一知识将会 对更好地了解和预测治疗反应具有治疗价值。推而广之，甚至更少 已知的二甲双胍在非裔美国人中的活性，因为很少有研究包括 相当数量的非欧洲人口群体。此应用程序将帮助纠正现有的知识 通过使用T2D研究大量不同的患者群体来发现差距。具体地说，我们将利用电子 大规模糖尿病药物基因组学的病历(EMR)数据。这些数据的优势在于 能够解释随时间推移的药物使用和药物暴露；提供大量的 用于联合分析和群体特定分析的个人；以及评估临床终点 回溯性和前瞻性。在本申请中，我们提出了以下研究目标：1)评估 自我报告的种族、民族和遗传因素对二甲双胍治疗的反应是否存在差异 祖先；2)使用新颖的、基于基因的关联方法来识别共享群体和种群群体 影响二甲双胍对血糖(即HbA1c水平)影响的特定遗传变异；以及3) 在单独的一组患者中重复我们的发现，并包括其他探索性分析以评估 已识别的基因变异是否会影响糖尿病相关的微血管事件、大血管事件、 以及药物不良反应。通过这项研究获得的知识将直接解决健康的目标 人民2020--“实现健康公平，消除差距，改善所有群体的健康”。

ABSTRACT Diabetes mellitus is a modern day scourge, affecting an ever increasing proportion of individuals worldwide, including 26 million Americans currently. Moreover, type-2 diabetes (T2D) disproportionately affects historically disadvantaged U.S. minority groups, as evidenced by the much higher rates of disease and more severe complications among African American individuals. Although there are multiple therapeutic classes of oral medication available for treating T2D, metformin is currently recommended as the first-line therapy. Metformin lowers blood glucose levels by reducing hepatic gluconeogenesis, improving skeletal muscle insulin sensitivity, and limiting intestinal glucose uptake. It has also been shown to be an effective therapy for preventing incident diabetes. Despite being one of the most frequently prescribed drugs worldwide, very little is known about the biologic mechanism(s) through which metformin mediates its effect. This knowledge would be of value therapeutically to better understand and predict treatment response. By extension, even less is known about the activity of metformin among African American individuals, as few studies have included substantial numbers of non-European population groups. This application will help rectify existing knowledge gaps by studying a large and diverse patient population with T2D. Specifically, we will utilize electronic medical record (EMR) data for large-scale diabetes pharmacogenomics. These data have the advantage of being able to account for medication use and drug exposure over time; to provide substantial numbers of individuals for combined and population group specific analyses; and to assess clinical end-points both retrospectively and prospectively. In this application, we propose the following study aims: 1) To assess whether there are differences in metformin treatment response by self-reported race-ethnicity and genetic ancestry; 2) To use novel, gene-based association approaches to identify both shared and population group specific genetic variants influencing metformin's effect on blood glycemia (i.e., HbA1c levels); and 3) To replicate our findings in a separate group of patients and to include additional exploratory analyses to assess whether the identified genetic variants influence diabetes-related microvascular events, macrovascular events, and adverse drug reactions. The knowledge gained through this study will directly address the goals of Health People 2020 – “achieve health equity, eliminate disparities, and improve the health of all groups.”