【作者】 郑荣强；

【导师】 王予彬；

【作者基本信息】 上海体育学院 ， 运动人体科学， 2009， 博士

【摘要】 目的:探讨关节镜下髌外侧支持带松解、清理术及术后康复对膝关节骨性关节炎的治疗作用;探讨膝关节骨性关节炎股四头肌和腘绳肌肌肉功能特点;探讨关节镜下髌外侧支持带松解、清理术和术后康复对股四头肌各肌束及腘绳肌肌肉功能的影响;基于BP神经网络建立膝关节骨性关节炎关节镜治疗临床疗效的预测模型。方法:膝骨性关节炎患者共50例,随机分为A、B两组,A组为“单纯关节镜下清理术”,B组为“关节镜下清理术+髌外侧减压术”,A组采用关节镜下骨性关节炎清理术,术中刨削明显增生、变性的滑膜及滑膜皱襞,根据半月板损伤、磨损情况,分别行半月板部分切除、大部分切除和全切除;搔刮表面毛糙、软化或者大部剥脱的软骨、刨削修整病损的软骨面;取出关节内游离体和剥脱的软骨。B组在A组手术的基础上行髌外侧支持带松解,使髌骨外侧倾斜可达30°以上,髌骨向内侧移动范围明显增加。两组术后按照相同的康复治疗计划进行康复训练,早期消肿、止痛、促进组织愈合;术后2天至2周保护修复区域,关节活动度训练;术后3周至3个月获得全范围关节活动度,强化肌力训练;晚期使肌力、关节活动度、本体感觉达到运动功能水平。两组患者均于术前、术后3个月和术后6个月接受膝关节功能评估、等速肌力测试和表面肌电测试。应用Matlab6.5神经网络工具箱建立BP神经网络模型,以膝关节OA患者术前病史、症状、体征、辅助检查及术中所见的关节内改变等变量作为输入端,以术后疗效情况作为输出端。50例接受关节镜治疗并进行随访的膝关节骨性关节炎患者分为两组,35例进行神经网络的训练,15例进行神经网络的测试。结果:1临床结果:A组术后3个月Lysholm量表评分、疼痛评分与术前相比明显改善,有统计学意义;术后6个月Lysholm量表评分、疼痛评分与术后3个月比较差异无统计学意义。B组术后3个月Lysholm量表评分、疼痛评分优于术前,差异有统计学意义;术后6个月Lysholm量表评分、疼痛评分优于术后3个月,差异有统计学意义。术后三个月B组膝关节Lysholm量表评分较A组下降,具有显著性差异。术后六个月B组患者膝关节Lysholm量表评分较A组患者增高,有显著性差异。两组患者术后3个月总疗效A组好于B组,差异具有统计学意义。两组患者术后6个月总疗效B组好于A组,差异具有统计学意义。2等速肌力测试结果:等速肌力测试的结果显示,膝关节OA患者健侧股四头肌在60°/s、120°/s时峰力矩较患侧高,有显著性差异,在180°/s无显著性差异。健侧股四头肌在60°/s、120°/s、180°/s相对峰力矩高于患侧,有显著差异。健侧股四头肌在60°/s、120°/s、180°/s时单次最佳做功高于患侧,有显著性差异。健侧股四头肌在60°/s、120°/s时平均功率高于患侧,有显著性差异。双侧腘绳肌峰力矩、最佳单次做功、平均功率均无显著性差异。健、患侧H/Q比值无显著性差异。股四头肌峰力矩和平均功率与骨性关节炎的疼痛指数呈中度正相关和Lysholom量表呈中度负相关。A组术后三个月股四头肌峰力矩较术前60°/s、120°/s提高,有显著差异,术后六个月较术后三个月无明显差异;B组术后三个月较术前无明显差异,术后六个月较术后三个月三个角速度都有明显差异;术后六个月B组在60°/s、120°/s、180°/s明显大于A组,有显著性差异。A组术后三个月股四头肌平均功率较术前60°/s、120°/s、180°/s提高,有显著差异,术后六个月较术后三个月无明显差异;B组术后三个月较术前无明显差异,术后六个月较术后三个月在60°/s、120°/s提高,有显著差异。术后六个月B组股四头肌平均功率在60°/s大于A组,有显著性差异。A组术后三个月股四头肌最佳单次做功较术前60°/s、120°/s、180°/s提高,有显著差异,术后六个月较术后三个月无明显差异;B组术后三个月较术前无明显差异,术后六个月在60°/s高于术后三个月,有显著性差异。术后六个月B组在60°/s、120°/s高于A组,有显著差异。A、B组腘绳肌峰力矩、单次最佳做功和平均功率术前、术后三个月、术后六个月均无显著差异。A、B两组组H/Q比值术后三个月较术前无显著差异,术后六个月较术后三个月无显著差异。3表面肌电分析结果膝关节OA患者健、患侧的表面肌电研究显示患侧VL在90~100°、100~110°、120~130°时小于健侧;RF在110~120°、120~130°时小于健侧;VMO在110~120°、140~150°和150~160°时小于健侧;BF在150~160°时大于健侧,均有统计学意义。健侧VMO/VL比值在60°/s、180°/s高于患侧,有统计学意义;在120°/s健、患侧无显著性差异。健侧VMO比VL早激活约5.5ms左右,而患侧VMO与VL早激活1.4ms左右,两者有显著性差异。A、B两组最大等长收缩时表面肌电平均振幅在术前、术后三个月和术后六个月都没有显著变化。在等速运动状态下表面肌电经标准化处理后术后三个月A组VL在60°/s时较B组高,有显著差异;RF在60°/s、120°/s较B组高,有显著差异;A、B两组VMO、ST、BF相比均无明显差异。术后六个月两组无明显差异。A组术前、术后三个月、术后六个月VMO/VL比值无明显变化;B组术后三个月VMO/VL比值较术前60°/s、120°/s、180°/s提高,有显著性差异,术后六个月较术后三个月无明显差别。术后三个月、术后六个月VMO、VL激发时间B组VMO较VL首先激活较A组有显著差异。4基于BP神经网络建立膝关节骨性关节炎关节镜治疗临床疗效的预测模型。50例样本随机分为两部分,35例为训练样本,15例为测试样本。用训练好的网络对测试集的15例样本进行预测,实际值和预测值间的差别无统计学意义表明网络有较好的预测效果,网络训练没有发生过度拟合现象,具有较好的泛化性能。结论:①关节镜下清理+髌外侧减压术对于入选的膝关节骨性关节炎患者具有良好的临床治疗效果,为OA患者康复训练、肌力恢复与膝关节功能改善提供了有利的条件;②关节镜下清理+髌外侧减压术后患者的肌力恢复经历了代偿期和功能恢复期;③等速肌力测试结果表明:膝关节OA患者关节镜清理术后经康复治疗伸、屈肌峰力矩、单次最佳做功及平均功率明显提高,尤其以股四头肌提高最为显著,术前H/Q比值基本处于正常范围,术后H/Q值比术前降低,应考虑加强屈膝肌力的训练,使屈伸肌力量比进一步协调;④表面肌电测试结果表明VMO/VL比值和VMO、VL激发时间髌外侧支持带减压患者较未减压的患者有明显变化,说明髌外侧支持带减压后经康复训练明显改善VMO和VL的协调性,进而改善髌骨运动轨迹和髌股关节的力学分布情况;⑤术后康复治疗对提高股四头肌和腘绳肌肌力和膝关节功能恢复具有重要的临床意义;⑥BP神经网络具有强大的模拟与预测功能,可以从复杂的信息中提取系统的输出与输入间的相互关系,本研究利用BP神经网络方法建立了膝关节OA患者关节镜治疗临床疗效的预测模型,为膝关节OA关节镜治疗的预后判断提供了一种有效方法。更多还原

【Abstract】 Objective: To study the mechanism of knee osteoarthritis treated with arthroscopic debridement and release of lateral patellar retinaculum; Study the musculus quadriceps fexoris and hamstring functional characteristics of knee osteoarthritis;To study the effect of arthroscopic debridement and release of lateral patellar retinaculum on the musculus quadriceps fexoris and hamstring functional; Based on the BP neural network to establish the model of curative effect estimation of arthroscopic debridement for knee osteoarthritis.Methods: 50 patients with knee osteoarthritis were randomly divided into groups A (29 cases) and B (21 cases) to underwent arthroscopic debridement and arthroscopic debridement combined with lateral patellar decompression respectively. All the operations were performed by a same surgeon. Postoperative rehabilitation therapies were carried out by one group of therapist. The Lysholm, PRI, PPI scores , isokinetic muscle strength test and surface electromyogram test were evaluated before the operation . The examinations were repeated in 3 and 6 months respectively after surgery.Based on the BP neural network to establish the model of curative effect estimation of arthroscopic debridement for knee osteoarthritis. Input end is preoperative case history , complaints, physical sign, auxiliary examination and the changes in the knee joint during operation of the knee OA patients. Output end is the postoperative curative effect. 50 cases of knee OA patient treated with arthroscopic debridement were divided into two groups, 35 cases were used to train the artificial neural network, 15 cases were to test the network. Results:1 Clinical results:In group A , the scores of Lysholm and PPI determined at postoperative 3 months were significantly higher than those before the operative, but not changed markedly at month 6. Whereas, in group B, the scores kept being increased after the surgery. At month 3, the scores in group A were higher than those in group B, however, at months 6, the Lysholm scores in group B were higher than those in group A. At month 3, the curatively effective rate of group A was higher than that of group B, however, at month 6, the rate of group B was higher than that of group A.2 Isokinetic muscle strength test resultsThe peak torque of the unaffected side were higher than affected side at angular speed of 120°/s and 180°/s, but there was no difference at angular speed of 180°/s. The hamstring’s peak torque was no significant difference between the two sides. The quadriceps’s relative peak torque of the unaffected side was higher than that of affected at angular speed of 60°/s、120°/s、180°/s, however the hamstring’s relative peak torque was no significant difference between the two sides. The quadriceps’s average power of the unaffected side was higher than that of the affected side at angular speed of 60°/s、120°/s.The hamstring’s average power was no significant difference between the two sides. The quadriceps’s total work of unaffected side was higher than that of the affected side at angular speed of 60°/s、120°/s、180°/s. The H/Q rate of the affected side was higher than that of the affected side.In group A the peak torque of postoperative month 3 was higher than that of preoperative at angular speed of 60°/s、120°/s, but there was no difference between month 3 and month 6. In group B there was no difference between preoperative and postoperative month 3, but month 6 was higher than that of month 3. At postoperative month 6, the peak torque of group B was higher than that of group A at all three angular speed. In group A the average power of postoperative month 3 was higher than that of preoperative at all three angular speed, but there was no difference between between month 3 and month 6. In group B there was no difference between preoperative and postoperative month 3, but month 6 was higher than that of month 3. In group A the total work of postoperative month 3 was higher than that of preoperative at all three angular speed, but there was no difference between between month 3 and month 6. In group B there was no difference between preoperative and postoperative month 3, but month 6 was higher than that of month at angular speed of 60°/s and 120°/s. There was no difference between group A and group B on the hamstring’s peak torque,total work and average power, as well as the month 3 and month 6.There was no difference between group A and group B on H/Q rate preoperative, as well as month 3 and month 6.3 Surface electromyogram test resultsThe average sEMG of the unaffected side VL was higher than that of the affected side at range of 90-100°,100-110°and 120-130°. The average sEMG of the unaffected side RF was higher than that of affected side at range of 110~120°、120~130°, and the unaffected side VMO is higher than affected side at range of 110~120°、140~150°、150-160°.The unaffected side BF is higher than affected side at range of 150-160°.The VMO/VL rate of unaffected side was higher than that of affected side at angular speed of 60°/s and 180°/s. There was no significant difference at angular speed 120°/s. VMO of unaffected side was activated more early 6 second than that of affected side.The MVC average sEMG of group A and group B didn’t changed significantly preoperative, month 3 and month 6. Normalized the sEMG of isokinetic eccentric exercise, the VL average sEMG of group A was higher than that of group B at angular speed of 60°/s, as RF at angular speed of 60°/s and 120°/s. There was no difference on VMO,ST and BF. The VMO/VL rate of group A changed no significantly on preoperative, postoperative month 3 and month 6. The preoperative VMO/VL rate of group B was higher than month 3 at angular speed of 60°/s、120°/s、180°/s.And there was no significant difference between month 3 and month 6. In group B VMO was activated more early than VL on month 3 and month 6.4 Based on the BP neural network to establish the model of curative effect estimation of arthroscopic debridement for knee osteoarthritis. 86 cases of knee OA patient treated with arthroscopic debridement were divided into two groups, 66 cases were used to train the artificial neural network, 20 cases were to test the network. There was no difference between the real value and predictive value. The network had a good predictive function. There was no excessive curve fitting during network training.Conclusion:①The patients of OA treated with arthroscopic debridement and release of lateral patellar retinaculum had a good curative effect, which provide a good condition to the rehabilitation.②The patients of OA treated with arthroscopic debridement and release of lateral patellar retinaculum experience a quadriceps compensation period and recovery period.③The extensional and flexional muscle peak torque,total work and average power of the patients of OA increased significantly after rehabilitation, specially quadriceps.The H/Q rate was in normal range preoperative, but decreased after operation.④The VMO/VL rate and VMO and VL activation time changed remarkably after operation in group B, which showed that release of lateral patellar retinaculum can improve the coordination of VMO and VL.⑤Rehabilitation after operation play an important role in increasing the muscle strength and the knee joint function.⑥BP network had a strong simulated and predictive ability, which can extract the relation between input and output from the distortion and defect information. This study establish the curative effective estimation model of knee OA patients treated with arthroscopic debridment, which provide a effective method to judge the prognosis of knee OA treated with arthroscopic debridment.更多还原