Effectiveness of Immediate vs. Delayed Use of Metformin in New-Onset T2 Diabetes

描述（由申请人提供）：本申请涉及广泛的挑战领域（05）比较有效性和特定挑战主题05-DK-101* 为新发现的2型糖尿病患者选择最佳初始治疗方案。2007年对2型糖尿病初始药物治疗的建议进行了修改-从磺脲类药物改为二甲双胍作为初始单药治疗，从建议进行药物营养治疗（MNT）和生活方式改变的初始试验改为建议在糖尿病初始诊断时立即开始药物治疗沿着MNT。来自UKPDS和两项观察性研究的结果表明，二甲双胍单药治疗相对于磺脲类药物可能降低心血管并发症的风险;来自ADOPT研究的结果表明，二甲双胍单药治疗的有效性持久性长于磺脲类药物格列本脲。二甲双胍似乎对LDL-胆固醇水平也有适度的相对益处，通常不会引起低血糖。来自糖尿病预防计划的证据表明，二甲双胍可有效预防葡萄糖耐量受损进展为糖尿病，这表明一旦超过2型糖尿病的诊断阈值，如果立即开始，它也可能有效减缓胰腺功能的恶化。最近对ACORD，ADVANCE和VADT试验的建议也表明，早期血糖控制可能比后期更有效地预防2型糖尿病的微血管和大血管并发症。这些指南的作者承认这些指南是“基于共识的”，并明确指出需要更好的证据，特别是在长期结果方面。它们尚未在临床实践中被广泛采用，目前尚不清楚它们是否会影响不同的真实患者人群的结局。毫无疑问，在不同的医生、系统和时间中，二甲双胍的采用存在差异，这为研究早期开始二甲双胍治疗是否上级既往治疗延迟糖尿病恶化和并发症发作提供了“自然实验”机会。我们建议检查立即开始二甲双胍单药治疗与延迟开始二甲双胍或早期/晚期开始磺脲类药物单药治疗对各种结局的比较有效性。使用来自HMO研究网络糖尿病联盟的4个大型卫生系统成员的电子病历数据，我们将创建一个新诊断的2型糖尿病患者的大型队列，平均随访3.5年。在纵向分析中，我们将检查与其他策略相比，及时开始二甲双胍治疗（最早发现糖尿病后6个月内）是否与以下因素相关：1.单药治疗血糖控制的持久性更长（从最初检测到糖尿病到单药治疗失败）2.初次诊断后1年和2年体重增加较少3。严重低血糖发生率的差异4.更好的LDL-胆固醇控制或较低强度的降脂治疗需要5。血压控制差异或降压治疗强度较低需要6。心血管疾病终点发生率的差异（急性心肌梗死、卒中或冠状动脉、颈动脉或外周血运重建的联合终点）7.微血管并发症（新发糖尿病视网膜病变、微量白蛋白尿或估计肾小球滤过率下降率）发生率的差异。我们认识到观察方法固有的混淆风险。我们可以从EMR记录中获得许多临床和人口统计学混杂因素，并将使用倾向评分对这些变量进行调整，这些变量在首次检测糖尿病时或之前测量。作为这些分析的补充，我们将进行工具变量分析，使用医生处方实践作为工具变量。最近的研究表明，二甲双胍是糖尿病初始治疗的最佳口服药物，并且在首次检测到糖尿病时立即开始二甲双胍，而不是等到血糖达到更高的目标水平，可以延缓糖尿病的进展并预防并发症，如心脏病，中风和肾脏疾病。这项研究在来自4个大型医疗保健系统的近40，000名新诊断的糖尿病患者中比较了二甲双胍的早期治疗与更传统的方法。如果早期使用二甲双胍显示出优势，该研究的发现可能会极大地改变美国2型糖尿病的治疗方式，并减少这种非常常见的并发症的负担。

DESCRIPTION (provided by applicant): This application addresses broad Challenge Area (05) Comparative Effectiveness and Specific Challenge Topic 05-DK-101* Selecting the Optimal Initial Treatment Regimen for Patients With Newly Discovered Type 2 Diabetes. Recommendations for initial pharmacotherapy in type 2 diabetes were modified in 2007 - to change from sulfonylureas to metformin as initial monotherapy, and from recommending an initial trial of medical nutrition therapy (MNT) and lifestyle change to recommending prompt initiation of pharmacotherapy along with MNT at initial diagnosis of diabetes. Findings from UKPDS and from two observational studies suggest that metformin monotherapy may lower risk for cardiovascular complications relative to sulfonylureas; findings from the ADOPT Study indicate that durability of monotherapy effectiveness is longer with metformin than with the sulfonylurea glyburide. Metformin also appears to have a modest relative benefit on LDL- cholesterol levels and usually does not cause hypoglycemia. Evidence from the Diabetes Prevention Program demonstrates that metformin is effective in preventing progression from impaired glucose tolerance to diabetes, suggesting that it may also be effective in slowing deterioration of pancreatic function once the diagnostic threshold for type 2 diabetes has been crossed if started immediately. Recent recommendations in response to the ACORD, ADVANCE, and VADT trials also suggest that early glycemic control may be more effective than later in preventing both microvascular and macrovascular complications of type 2 diabetes. These guidelines are acknowledged by their authors to be "consensus-based" and the need for better evidence, especially with respect to longer-term outcomes is clearly stated. They have not been widely adopted in clinical practice, and it is unknown whether they are impacting outcomes in diverse, real-world, patient populations. Undoubtedly, there has been variation in adoption across physicians, systems, and time, providing this "natural experiment" opportunity to study whether early initiation of metformin is superior to prior practice for delaying deterioration of diabetes and the onset of complications. We propose to examine the comparative effectiveness of immediate initiation of metformin monotherapy vs. delayed initiation of metformin or early/late initiation of sulfonylurea monotherapy for a variety of outcomes. Using electronic medical record data from 4 large health system members of the HMO Research Network Diabetes Consortium, we will create a large cohort of newly diagnosed type 2 diabetic patients with average follow-up of 3.5 years. In longitudinal analyses, we will examine whether prompt initiation of metformin (within 6 months of the earliest detection of diabetes) compared with other strategies is associated with: 1. Longer durability of glycemic control on monotherapy (from initial detection of diabetes to failure of monotherapy) 2. Less weight gain at 1 and 2 years post initial diagnosis 3. Differences in the incidence of severe hypoglycemia 4. Better LDL-cholesterol control or lower intensity of lipid-lowering therapy required 5. Differences in blood pressure control or lower intensity of anti-hypertensive therapy required 6. Differences in the incidence of cardiovascular disease endpoints (combined endpoint of acute myocardial infarction, stroke, or coronary, carotid or peripheral revascularization) 7. Differences in incidence of microvascular complications (new onset diabetic retinopathy, microalbuminuria, or rate of decline in estimated glomerular filtration rate). We recognize the risks of confounding inherent in observational approaches. Many clinical and demographic confounders are available to us from EMR records and all comparisons will be adjusted, using propensity scores, for these variables measured at or before initial detection of diabetes. As a complement to these analyses, we will conduct instrumental variable analyses, using physician prescribing practices as the instrumental variable. Recent studies suggest that metformin is the best oral medication for initial treatment of diabetes and that starting metformin immediately when diabetes is first detected, rather than waiting until blood sugars reach a somewhat higher target level, could delay progression of diabetes and prevent complications such as heart disease, stroke and kidney disease. This study compares very early treatment with metformin to more traditional approaches in a population of nearly 40,000 newly diagnosed diabetic patients from 4 large health care systems. If an advantage is shown for early use of metformin, the study's findings could dramatically change the way in which type 2 diabetes is treated in the U.S. and reduce the burden of complications from this very common condition.