Meng Ling-Rui, Chen Hua, Chen Wen-Qian, Gao Yi, Li Zi-Wei, Ye Zi, Li Zhao-Hui
Senior Department of Ophthalmology, the Third Medical Center, Chinese PLA General Hospital, Beijing 100039, China.
Medical School of Chinese PLA, Beijing 100853, China.
Int J Ophthalmol. 2024 Nov 18;17(11):2060-2066. doi: 10.18240/ijo.2024.11.12. eCollection 2024.
To observe the changes in corneal subepithelial nerve fibers (CNFs) and Langerhans cells (LCs) in patients with type 2 diabetes using corneal laser confocal microscopy (CLCM).
A total of 60 patients (64 eyes), including 40 patients with type 2 diabetes (DM group) and 20 subjects without diabetes (control group) were included with CLCM. Neuron J plugin of Image J software were used for quantitative analysis of CNF length (CNFL), CNF density (CNFD), corneal nerve branch fiber density (CNBD), main branch length density, branch length density, corneal nerve fiber tortuosity (NT) score, and LCs density. An independent samples -test to analyze the variability between the two groups was performed, and Pearson correlation analysis was used to analyze the relationships between CNF and multiple biochemical indicators in the DM group. The predictive power of CNF for type 2 diabetes was assessed using the receiver operating characteristic (ROC) curve.
There were significant differences in the CNFL, CNFD, and main branch length density between two groups. The results of Pearson correlation analysis showed a significant negative correlation between CNFD and the duration of diabetes as well as triglyceride levels and total cholesterol, and a significant positive correlation between CNFD and serum albumin. In addition, the NT score showed a positive correlation and urea nitrogen, similar to the positive correlation observed between LC density and glycosylated hemoglobin (HbA1c) levels. CNFD showed the highest area under the curve (AUC of ROC) value, followed by main branch length density and CNFL. The AUC of the ROC curve under the logistic regression model also demonstrated good predictive values. The cut-off values of CNFD, CNFL, and main branch length density for diabetes showed 31.25, 18.85, and 12.56, respectively.
In patients with type 2 diabetes, there is a notable reduction in both CNFL and CNFD. These measurements can be influenced by various blood biochemical factors. However, the compromised nerve fibers can serve as valuable indicators for predicting the onset of type 2 diabetes and also as biomarkers for detecting diabetic neuropathy and its related complications.
采用角膜激光共聚焦显微镜(CLCM)观察2型糖尿病患者角膜上皮下神经纤维(CNFs)和朗格汉斯细胞(LCs)的变化。
纳入60例患者(64只眼),其中2型糖尿病患者40例(糖尿病组),无糖尿病受试者20例(对照组),均行CLCM检查。使用Image J软件的Neuron J插件对CNF长度(CNFL)、CNF密度(CNFD)、角膜神经分支纤维密度(CNBD)、主支长度密度、分支长度密度、角膜神经纤维曲折度(NT)评分及LCs密度进行定量分析。采用独立样本t检验分析两组间的差异,Pearson相关分析用于分析糖尿病组CNF与多种生化指标之间的关系。使用受试者工作特征(ROC)曲线评估CNF对2型糖尿病的预测能力。
两组间CNFL、CNFD及主支长度密度存在显著差异。Pearson相关分析结果显示,CNFD与糖尿病病程、甘油三酯水平及总胆固醇呈显著负相关,与血清白蛋白呈显著正相关。此外,NT评分与尿素氮呈正相关,与LC密度和糖化血红蛋白(HbA1c)水平呈正相关情况相似。CNFD的曲线下面积(ROC曲线的AUC)值最高,其次为主支长度密度和CNFL。逻辑回归模型下ROC曲线的AUC也显示出良好的预测价值。糖尿病的CNFD、CNFL及主支长度密度的截断值分别为31.25、18.85和12.56。
2型糖尿病患者的CNFL和CNFD均显著降低。这些测量结果会受到多种血液生化因素的影响。然而,受损的神经纤维可作为预测2型糖尿病发病的有价值指标,也可作为检测糖尿病神经病变及其相关并发症的生物标志物。