Lipidomic Predictors of Diabetic Kidney Disease Progression in Patients with Type-1 Diabetes

摘要 糖尿病是美国终末期肾病的主要原因。目前还没有生物标记物 在保留GFR的情况下，确定糖尿病肾病(DKD)进展的高风险患者(&gt；90 Ml/min)，尿白蛋白排泄量在正常范围内。在这项研究中，我们旨在测试预测能力 血浆C16-C24游离脂肪酸S和C40-C46三酰甘油S预测糖尿病肾病进展 早期估计肾小球滤过率(EGFR)>90毫升/分钟，尿白蛋白/肌酐比值(ACR) <30 mg/g。这项研究将是病例对照观察，病例组定义为 DKD的纵向随访进展情况。患者群体为1型糖尿病患者。学习 样本选自包括Steno糖尿病在内的4个已建立的1型糖尿病患者队列 哥本哈根中心，芬兰糖尿病肾病(FinnDiane)，科罗拉多州冠状动脉钙化 在1型糖尿病(CACT1)和匹兹堡糖尿病并发症流行病学(EDC)中。抽样是 根据纳入和排除标准的适用情况。入选标准为18岁或以上。 在样本选择时，表皮生长因子受体≥为90ml/分钟，≥3纵向测量表皮生长因子受体，并随访超过 四年了。排除标准为18岁。病例组被定义为患有1型糖尿病的患者。 随访期间EGFR丢失ml/min/年。对照组定义为1型糖尿病患者，没有 或在随访期间EGFR损失小于1毫升/分钟/年，根据年龄、性别、种族和EGFR匹配的频率在 与病例组的基线。总体而言，350名患者包括无进步和进步的患者，比例为2：1 都被选中。选择后，患者将被随机分成两组进行培训(57名进展者和117名非进步者 进步者)和验证队列。结果是DKD的进展，定义为EGFR中3毫升/分钟/年的损失 在后续访问期间。基线就诊时的临床数据和血浆样本(匹配病例的对应日期 和控件)可用。有针对性的脂体学研究(基于我们的初步数据)将应用于量化 用AB Sciex三重四极杆/QTRAP在多反应监测(MRM)模式下建议的脂类 6500质谱仪。对于分析，我们将使用带有错误发现率校正的t-检验 使用逐个化合物的比较来预测DKD进展的能力。另外， 我们将使用主成分进行数据约简，并将合并重要的脂类以及 在调整后的Logistic回归模型中分离主成分检验独立预测 DKD进展的建议标记物。我们将计算c统计量，并将其与EGFR和ACR的结果进行比较 评估分类能力提升情况。我们将在验证子集中复制分析 共175例，包括无进展者和进展者，比例为2：1。总的来说，我们期待着 确定一个能准确预测早期DKD进展的定量预后脂质小组。

ABSTRACT Diabetes is the leading cause of end stage kidney disease in the United States. Currently there is no biomarker to identify patients at high risk of progression of diabetic kidney disease (DKD) when GFR is preserved (>90 mL/min) and urine albumin excretion is within normal limit. In this study we aim to test the predictive power of C16-C24 free fatty acids (FFA)s and C40-C46 triacylglycerols (TAG)s in plasma to predict progression of DKD at early stage when estimated GFR (eGFR) is greater than 90 mL/min and urine albumin-creatinine ratio (ACR) is less than 30 mg/g. This study will be a case control observation in which the case group is defined as progression of DKD in longitudinal follow up visits. Patient population is the patients with type-1 diabetes. Study samples are selected from 4 established cohorts of patients with type-1 diabetes including the Steno Diabetes Center Copenhagen, the Finnish Diabetic Nephropathy (FinnDiane), the Colorado Coronary Artery Calcification in Type 1 diabetes (CACT1), and the Pittsburgh Epidemiology of Diabetic Complications (EDC). Sampling is based on the application of the inclusion and exclusion criteria. The inclusion criteria are age of 18 years or older at the time of sample selection, eGFR ≥90 ml/min, ≥3 longitudinal measure of eGFR, and follow up of more than 4 years. Exclusion criterion is age<18 years. Case group is defined as patients with type-1 diabetes who had >3 ml/min/year loss in eGFR during follow up. Control group is defined as patients with type-1 diabetes who had no or less than 1 mL/min/year loss in eGFR during follow up, frequency matched by age, sex, race, and eGFR at baseline with the case group. Overall, 350 patients including non-progressors and progressors with a 2:1 ratio are selected. After selection, patients will be randomly split to the training (57 progressors and 117 non- progressors) and validation cohorts. Outcome is progression of DKD defined as >3 mL/min/year loss in eGFR during follow up visits. Clinical data and plasma samples at baseline visit (corresponding date of matching cases and controls) are available. Targeted lipidomic studies (based on our preliminary data) will be applied to quantify the proposed lipids in multiple reaction monitoring (MRM) mode using an AB Sciex Triple Quadrupole/QTRAP 6500+ mass spectrometer. For analysis, we will apply t-test with false discovery rate correction for multiple comparisons using a compound by compound comparison for ability to predict DKD progression. Additionally, we will use principal component for data reduction, and will incorporate the significant lipids as well as the principal components separately in adjusted logistic regression models to test the independent prediction of proposed markers on DKD progression. We will calculate c-statistics and compare it to that of eGFR and ACR to assess the improvement of classification power. We will replicate the analysis in the validation subset consisting of 175 patients including non-progressors and progressors with 2:1 ratio. Collectively, we anticipate identifying a quantitative prognostic lipid panel that accurately predicts early DKD progression.