糖尿病神经病变

　　【摘要】 目的 探讨5项筛查方法对糖尿病周围神经病变(DPN)的筛查效果，寻找早期临床诊断最佳筛查组合。 方法 选取糖尿病患者1214例，以神经传导速度(NCV)检查作为参照，分别接受5项筛查方法及多伦多临床评分系统(TCSS)，分析5项筛查方法和TCSS的敏感性、特异性、阳性预测值、阴性预测值和约登指数。绘制受试者工作特征曲线，比较不同组合的筛查效果，评价各组合与NCV的拟合度。 结果 踝反射、振动觉、压力觉、温度觉和刺痛觉的敏感性和特异性分别为35.21%、76.87%、29.41%、36.26%、34.54%和69.57%、62.78%、92.26%、89.25%、85.69%。5项筛查方法和TCSS切点≥6时的敏感性、特异性、阳性预测值、阴性预测值和约登指数分别为89.94%、67.65%、82.19%、79.28%、0.58 和78.21%、84.11%、89.33%、69.40%、0.62。5项筛查方法与TCSS比较差异无统计学意义。5项筛查方法中选取1、2、3、4、5项作为组合的曲线下面积(AUC)最大值分别为0.71、0.75、0.77、0.78、0.79。AUC值拐点出现在5项选取3项的组合中，其中踝反射+振动觉+温度觉组合AUC值最高。 结论 踝反射+振动觉+温度觉为最佳筛查组合，与NCV有较好拟合度，可全面、准确筛查DPN。

　　【关键词】糖尿病周围神经病变;神经传导速度;筛查;受试者工作特征曲线

　　【Abstract】 Objective To investigate the effect of five screening methods for diabetic prripheral neuropathy (DPN), and search the best combination of DPN screening methods. Methods Five screening methods and Toronto clinical scoring system (TCSS) were performed in 1214 diabetic patients respectively. The nerve conduction velocity (NCV) as a gold standard was used to analyze five screening methods and TCSS and to calculate the sensitivity, specificity, positive predictive value, negative predictive value and Youden index. The effectiveness of various combination were assessed using receiver operating characteristic curve and its AUC, and at last the best screening combination was selected. Results The sensitivity and specificity of ankle reflex, vibration perception, feeling the pressure, temperature sensation, pinprick were 35.21%, 76.87%, 29.41%, 36.26%, 34.54% and 69.57%, 62.78%, 92.26%, 89.25%, 85.69% respectively. Sensitivity, specificity, positive predictive value, negative predictive value and Youden index of five screening methods and TCSS≥6 were 89.94%, 67.65%, 82.19%, 79.28%, 0.58 and 78.21%, 84.11%, 89.33%, 69.40%, 0.62 respectively. There was no significant difference between five screening methods and TCSS. Five screening methods in the combinations of one, two, three, four, five of maximum area under curve (AUC) were 0.71, 0.75, 0.77, 0.78, 0.79. When taking three kinds among five kinds of examination, combination of ankle reflex plus vibration perception plus temperature sensation showed highest AUC value. Conclusion The combination of ankle reflex, vibration perception and temperature sensation has a good compliance with NCV. It is an optimal screening combination and can be used as a simple and effective method to screen DPN.

　　【Key words】Diabetic peripheral neuropathy (DPN);Nerve conduction velocity (NCV);Screening;Receiver operating characteristic curve