2型糖尿病胫神经病变超声特征及其危险因素

目的分析2型糖尿病胫神经病变超声特征。探讨2型糖尿病胫神经病变出现前后的神经形态学改变的危险因素。方法选取超声检查胫神经的452例2型糖尿病患者,以密歇根神经筛查量表(MNSI)体格检查2分作为诊断截断值分为有神经病变组和非神经病变组。以性别、年龄、体重指数、小腿周长条件近似1:1配对,共配对成功40对。搜集胫神经超声形态学特征包括宽度、厚度、横截面积、胫神经“蜂窝样”结构是否显示,建立条件Logistic回归模型分析周围神经病变发生的危险因素,并使用ROC曲线对模型进行验证。结果胫神经厚度(OR=6.250,P=0.044)和横截面积(OR=2.259,P=0.049)是2型糖尿病周围神经病变发生的危险因素,而胫神经宽度和“蜂窝样”结构是否可清晰显示不是其危险因素(P均>0.05)。条件Logistic回归模型曲线下面积0.91,敏感度、特异度、阳性预测值及阴性预测值为85.71%、95.24%、94.70%、87.00%。胫神经横截面积对糖尿病胫神经病变诊断的ROC曲线下面积0.794,敏感度、特异度、阳性预测值及阴性预测值为71.43%、90.48%、88.20%、76.00%;厚度的曲线下面积0.726,敏感度、特异度、阳性预测值及阴性预测值为70.27%、85.71%、83.30%、75.00%。结论胫神经的厚度及横截面积是2型糖尿病周围神经病变的危险因素。胫神经厚度与横截面积对诊断周围神经病变的敏感度接近,横截面积的特异度高于厚度。

The ultrasound features and risk factors of tibial neuropathy in Type 2 diabetic peripheral neuropathy

Objective To analysis the ultrasound characteristics of tibial neuropathy in type 2 diabetes and to study the morphological changes before and after tibial lesions in diabetic tibial neuropathy.Methods A retrospective analysis of 452 patients with type 2 diabetes who underwent tibial nerve ultrasound examination was performed.We divided subjects into two groups using the Michigan Neuropathy Instrumen(MNSI)as the diagnostic cut-off value of"2 points".The two groups were"neuropathy group"and"non-neuropathy group".According to gender,age,BMI and calf circumference,40 pairs were matched successfully.A conditional Logistic regression model was established based on the ultrasound morphological characteristics of the tibial nerve:width,thickness,cross-sectional area,and whether the"honeycomb"structure of the tibial nerve was shown as independent variables.ROC curve was used to verify the conditional Logistic regression model.Results Results of conditional Logistic regression model analysis showed atibial nerve thickness(OR=6.250,P=0.044)and cross-sectional area(OR=2.259,P=0.049)were risk factors for type 2 diabetic peripheral neuropathy.The tibial nerve width(P=0.355)and the"honeycomb"structure were not statistically significant(P=0.291).The AUC of the conditional Logistic regression model was 0.91,and the sensitivity,specificity,positive predictive value and negative predictive value of the model were 85.71%,95.24%,94.70%and 87.00%.The AUC of tibial nerve cross-sectional area was 0.794,sensitivity,specificity,positive predictive value and negative predictive value were 71.43%,90.48%,88.20%and 76.00%.The AUC of tibial nerve thickness was 0.726,sensitivity,specificity,positive predictive value and negative predictive value were 70.27%,85.71%,83.30%and 75.00%.Conclusion The thickness and cross-sectional area of tibial nerve are high risk factors for peripheral lesions in type 2 diabetic peripheral neuropathy and also important indicators for ultrasound observation.The sensitivity of the thickness of the tibial nerve is similar to that of the cross-sectional area and the specificity of the cross-sectional area is higher than that of the thickness.