2型糖尿病正常白蛋白尿且eGFR保留者的隐匿性肾小管表型
本研究纳入4046例2型糖尿病成人，均尿白蛋白肌酐比（UACR）<30 mg/g且eGFR≥60 mL/min/1.73 m²，检测晨尿α1-微球蛋白和β2-微球蛋白。结果显示19.7%存在低分子量蛋白定义的肾小管表型，UACR单独无法识别该表型。独立人肾数据集证实存在近端小管损伤相关程序及显著细胞和空间异质性。
要点
- 近五分之一尿白蛋白正常且eGFR保留的2型糖尿病患者存在隐匿性肾小管损伤表型
- UACR单独不能识别该肾小管表型，需补充肾小管标志物评估
- 肾小管表型可作为2型糖尿病肾脏评估的补充手段
核心结论
- 在UACR正常且eGFR保留的2型糖尿病成人中，低分子量蛋白定义的肾小管表型发生率近五分之一，且UACR难以识别该表型。
- 独立人肾数据集显示该表型伴随近端小管损伤相关程序，具有显著细胞和空间异质性。
研究设计与样本
- 纳入4046例2型糖尿病成人，UACR<30 mg/g，eGFR≥60 mL/min/1.73 m²，检测配对晨尿α1-微球蛋白和β2-微球蛋白。
- 肾小管表型定义为α1-微球蛋白>12 mg/L且β2-微球蛋白>0.20 mg/L。
- 进行药物、入院背景、高eGFR和肝功能敏感性分析，并分析独立显微分离肾小管、单细胞/单核及CosMx数据集。
主要发现
- 797例（19.7%）存在肾小管表型。
- UACR单独区分该表型的AUC为0.466（95% CI：0.443–0.489）。
- 临床校正后ln(UACR+1)与肾小管表型无关联（OR：0.873，95% CI：0.710–1.075；p=0.201）。
- 药物校正后估计相似（OR：0.871，95% CI：0.708–1.072；p=0.191），其他敏感性分析结果一致。
- 显微分离肾小管中DKD患者10基因损伤评分更高（p<0.001）；单细胞图谱中DKD样本近端小管损伤比例最高（p=0.029）；空间分析识别出5个损伤近端小管亚群。
关键数值与推荐
- 肾小管表型定义：晨尿α1-微球蛋白>12 mg/L且β2-微球蛋白>0.20 mg/L。
- 研究入选标准：UACR<30 mg/g且eGFR≥60 mL/min/1.73 m²。
临床怎么用
- 对2型糖尿病患者，即使UACR正常且eGFR保留，也应考虑检测尿低分子量蛋白以识别隐匿性肾小管损伤。
- 肾小管表型评估可作为UACR之外肾脏评估的补充手段。
以上解读依据论文摘要（全文未开放获取或许可不允许翻译转载）。
原文
ABSTRACT Aims To determine the prevalence of a low-molecular-weight protein (LMWP)-defined tubular phenotype in adults with Type 2 diabetes, normoalbuminuria and preserved estimated glomerular filtration rate (eGFR), its relation to UACR, and its human kidney context. Materials and Methods We studied 4046 adults with Type 2 diabetes, UACR < 30 mg/g, eGFR ≥ 60 mL/min/1.73 m2, and paired morning urinary α1-microglobulin (α1MG) and β2-microglobulin (β2MG). The phenotype required α1MG > 12 mg/L and β2MG > 0.20 mg/L. We tested associations with UACR and performed medication, admission-context, high-eGFR and liver-function sensitivity analyses. Independent microdissected tubular, single-cell/single-nucleus and CosMx datasets were analysed. Results The phenotype was present in 797 patients (19.7%). UACR alone did not discriminate the phenotype (AUC 0.466, 95% CI: 0.443–0.489), and ln(UACR+1) was not associated with it after clinical adjustment (OR: 0.873, 95% CI: 0.710–1.075; p = 0.201). Medication adjustment produced a similar estimate (OR: 0.871, 95% CI: 0.708–1.072; p = 0.191), with consistent findings in the other sensitivity analyses. In microdissected tubules, a 10-gene injury score was higher in DKD (p < 0.001); sample-level injured proximal-tubular fractions were highest in DKD in the single-cell atlas (p = 0.029), and spatial profiling resolved five injured proximal-tubular subclusters. Conclusions An LMWP-defined tubular phenotype was present in nearly one in five adults with Type 2 diabetes despite normoalbuminuria and preserved eGFR, and was poorly captured by UACR. Independent human kidney datasets showed related proximal-tubular injury-associated programmes with marked cellular and spatial heterogeneity. These findings support tubular phenotyping as a complementary approach to kidney assessment in Type 2 diabetes and warrant prospective validation.
引用
- Diabetes, Obesity and Metabolism：A Clinically Occult Tubular Phenotype in Type 2 Diabetes Despite Normoalbuminuria and Preserved eGFR: Multiscale Human Kidney Evidence
供执业医师工作参考，不替代临床判断；数值与分级以指南原文为准。