Clinical Informatics to Advance Epidemiology and Pharmacogenetics of Serious Cutaneous Adverse Drug Reactions

项目摘要 严重的皮肤不良反应(疤痕)是对药物的病态免疫反应，这些药物赋予 死亡率为10%-50%。在过去的十年里，对预测和预防的重大承诺来自于 发现许多疤痕与人类白细胞抗原I类等位基因的变异有关。对于HLA-B*15：02，这是 导致在许多东南亚国家对卡马西平进行常规处方前筛查，并对 卡马西平SJS/TEN病例的减少。尽管取得了这些进展，但人们对基因和基因的研究知之甚少 与抗生素等常用药物有关的疤痕的流行病学危险因素。也有有限的 关于美国不同人群中疤痕的人类白细胞抗原风险的信息。此外， 临床表型的不精确和缺乏标准化的编码导致了寻找疤痕病例的挑战 在电子健康记录(EHR)中。我们建议的研究旨在解决我们的关键挑战和差距 关于抗生素疤痕的知识。 在目标1中，我们将利用先进的信息学和纵向EHR数据，从 自20世纪80年代以来一直与医疗保健系统合作，以确定疤痕病例。我们将创建、优化和规范 用于发现疤痕病例和验证疤痕患者队列的可重复性方法。这个迭代过程 将被用来提炼和传播电子表型，以供跨机构验证。 在目标2中，我们将分析疤痕患病率并进行病例对照研究，以确定药物特异性和 抗生素相关疤痕的特定患者风险因素。我们将比较临床后遗症、生活质量和 通过有效的调查工具比较疤痕患者与对照组的依从性。 在目标3中，我们将从使用有效抗生素的患者中确定候选的人类白细胞抗原和基因关联。 相关的伤疤。我们将检查少数民族和健康差距人群的遗传风险差异，并 预测我们将有能力建立万古霉素服装的人类白细胞抗原关联(即，药物反应与 嗜酸性粒细胞增多和全身症状)以及磺胺类抗菌药物和β-内酰胺类疤痕。单独使用HL A，或者 结合临床危险因素，可以改善疤痕预防和早期诊断。我们会 建立一个数据共享平台，以在线电子表型和患者登记的形式，可以 用于扩大SCAR队列，用于未来的大规模基因组研究。 我们开发的路线图将转化为针对严重不良反应的电子表型的开发 促进基因发现的药物反应。所获得的知识将对基因数据的翻译至关重要 转化为临床决策。这与美国国立卫生研究院加速遗传基因的研究任务密切相关 发现医源性和可预防的药物引起的疾病，将转化为预防，更早 诊断和增强机制的理解，这可能导致有针对性的治疗方法。

Project Summary Severe cutaneous adverse reactions (SCARs) are morbid immunologic reactions to drugs that confer a mortality of 10-50%. Over the last decade, significant promise for prediction and prevention has come from the discovery that many SCARs are associated with variation within HLA class I alleles. For HLA-B*15:02, this has led to routine pre-prescription screening for carbamazepine in many Southeast Asian countries and a significant reduction in cases of carbamazepine SJS/TEN. Despite this progress, there is little known about genetic and epidemiological risk factors for SCARs related to commonly used drugs such as antibiotics. There is also limited information about HLA risk for SCARs across the diverse populations present in the United States. Furthermore, imprecision of clinical phenotyping and lack of standardized coding has led to challenges in finding SCAR cases in the electronic health record (EHR). Our proposed study aims to address critical challenges and gaps in our knowledge of antibiotic SCARs. In Aim 1, we will leverage advanced informatics and longitudinal EHR data for over 11 million patients from Partners HealthCare System since the 1980s to identify SCAR cases. We will create, optimize and standardize reproducible methods for finding SCAR cases and validating a cohort of SCAR patients. This iterative process will be used to refine and disseminate an electronic phenotype to be validated cross-institutionally. In Aim 2, we will analyze SCAR prevalence and conduct a case-control study to identify drug-specific and patient-specific risk factors for antibiotic-associated SCARs. We will compare clinical sequelae, quality of life and adherence of SCAR patients compared to controls through validated survey instruments. In Aim 3, we will identify candidate HLA and genetic associations from patients with validated antibiotic- associated SCARs. We will examine difference in genetic risk in minority and health disparity populations and predict that we will be powered to establish HLA associations for vancomycin DRESS (i.e., drug reaction with eosinophilia and systemic symptoms) and sulfonamide antimicrobial and beta-lactam SCAR. HLA alone, or in combination with clinical risk factors, can lead to improved SCAR prevention and early diagnosis. We will establish a data sharing platform, in the form of an online electronic phenotype and patient registry, that can be used to enlarge SCAR cohorts for future large-scale genomics studies. The roadmap we develop will translate into the development of electronic phenotypes for serious adverse drug reactions that facilitate genetic discovery. Knowledge gained will be crucial to the translation of genetic data into clinical decision making. This is in close alignment with NIH’s research mission to accelerate genetic discovery for iatrogenic and preventable drug-induced diseases that will translate into prevention, earlier diagnosis and an enhanced mechanistic understanding that may lead to targeted therapeutic approaches.