Florey Neuroscience Institutes, University of Melbourne, Parkville, Victoria, Australia.
J Parkinsons Dis. 2012;2(1):47-55. doi: 10.3233/JPD-2012-11071.
There is a need for objective measures of dyskinesia and bradykinesia of Parkinson's disease (PD) that are continuous throughout the day and related to levodopa dosing. The output of an algorithm that calculates dyskinesia and bradykinesia scores every two minutes over 10 days (PKG: Global Kinetics Corporation) was compared with conventional rating scales for PD in PD subjects. The algorithm recognises bradykinesia as movements made with lower acceleration and amplitude and with longer intervals between movement. Similarly the algorithm recognises dyskinesia as having movements of normal amplitude and acceleration but with shorter periods without movement. The distribution of the bradykinesia and dyskinesia scores from PD subjects differed from that of normal subjects. The algorithm predicted the clinical dyskinesia rating scale AIMS with a 95% margin of error of 3.2 units compared with the inter-rater 95% limits of agreement from 3 neurologists of -3.4 to +4.3 units. Similarly the algorithm predicted the UPDRS III score with a margin of error similar to the inter-rater limits of agreement. Improvement in scores in response to changes in medication could be assessed statistically in individual patients. This algorithm provides objective, continuous and automated assessment of the clinical features of bradykinesia and dyskinesia in PD.
需要一种客观的帕金森病（PD）运动障碍和运动迟缓的测量方法，该方法应能在一天内持续测量，并与左旋多巴剂量相关。我们比较了一种算法的输出结果，该算法每隔两分钟计算一次运动障碍和运动迟缓评分，共计算 10 天（PKG：全球动力学公司），并将其与 PD 患者的常规 PD 评定量表进行比较。该算法将运动迟缓定义为运动时加速度和幅度较低，运动之间间隔较长的运动。同样，该算法将运动障碍定义为运动幅度和加速度正常，但运动之间无运动的时间较短。PD 患者的运动迟缓与运动障碍评分分布与正常受试者不同。该算法预测临床运动障碍评定量表 AIMS 的误差范围为 95%，为 3.2 个单位，而 3 位神经科医生的组内相关系数 95%的一致性界限为-3.4 至+4.3 个单位。同样，该算法预测 UPDRS III 评分的误差范围与组内相关系数的一致性界限相似。可以对个体患者的药物治疗反应进行统计学评估。该算法可对 PD 患者的运动迟缓与运动障碍的临床特征进行客观、连续和自动评估。
