Conklin Sarah J, Cavalcanti Helen Meira, Almeida Lorena Rosa S, Mishra Virendra, Oliveira-Filho Jamary, Mari Zoltan, Landers Merrill R, Longhurst Jason K
Program in Physical Therapy, Doisy College of Health Sciences, Saint Louis University, 3437 Caroline Street, St. Louis, MO 63104, USA.
Department of Physical Therapy, Northeast Adventist University Center for Education, BR-101, km 197, Estrada de Capoeiruçu, s/n, P.O. Box 18, Capoeiruçu, Cachoeira, Bahia 44300-000, Brazil; Postgraduate Program in Health Sciences, School of Medicine, Federal University of Bahia, Av. Milton Santos, s/nº - Ondina, Salvador, Bahia 40170-110, Brazil.
Gait Posture. 2025 Jul 18;122:225-231. doi: 10.1016/j.gaitpost.2025.07.323.
BACKGROUND & AIM: Freezing of gait (FOG) is common in Parkinson's disease (PD) and increases fall risk. While medication may reduce FOG, many with PD still show abnormalities in unfrozen gait. This study aimed to identify gait characteristics in both ON- and OFF-PD medication states that are associated with FOG severity and those that can detect FOG presence.
Data were collected from 36 people with PD (15 with FOG, 21 without FOG), including demographics (age, sex), PD severity (MDS-UPDRS part III), Levodopa-Equivalent Daily Dose (LEDD), observed FOG severity (FOG Score), and temporal-spatial gait measures from an instrumented walkway. Gait was assessed in the OFF-state (>12-hour medication withdrawal) and ON-state (∼45 min after medication). Ridge regression was used to explore the relationship between OFF-state FOG severity and gait characteristics (eGVI, gait velocity, stride length (mean and coefficient of variation (CV)), cadence, stride time (mean and CV), single support time (mean and CV) in both states. Lasso regression identified characteristics most sensitive for detecting FOG.
In the ON-state, only single support time CV (β=0.90, p < 0.001) predicted FOG severity (R2 =0.80). In the OFF-state, MDS-UPDRS III (β=0.28, p = 0.136) and stride time CV (β=0.69, p = 0.002) predicted FOG severity (R2 =0.70). ON-state eGVI (β=0.12, p = 0.06), OFF-state stride length mean (β=-0.018, p = 0.77), and OFF-state velocity (β=-0.021, p = 0.70) distinguished between freezers and non-freezers with 88 % accuracy.
DISCUSSION/CONCLUSIONS: These results suggest that gait variability during unfrozen gait may reflect a subtle manifestation of FOG. Longitudinal studies should explore the development of FOG and associated gait changes over time.
冻结步态（FOG）在帕金森病（PD）中很常见，会增加跌倒风险。虽然药物治疗可能会减少冻结步态，但许多帕金森病患者在非冻结步态时仍表现出异常。本研究旨在确定帕金森病患者在服药和未服药状态下与冻结步态严重程度相关的步态特征以及能够检测到冻结步态存在的特征。
收集了36例帕金森病患者（15例有冻结步态，21例无冻结步态）的数据，包括人口统计学信息（年龄、性别）、帕金森病严重程度（MDS-UPDRS第三部分）、左旋多巴等效日剂量（LEDD）、观察到的冻结步态严重程度（冻结步态评分）以及来自仪器化步道的时空步态测量数据。在未服药状态（停药超过12小时）和服药状态（服药后约45分钟）下评估步态。采用岭回归探索未服药状态下冻结步态严重程度与步态特征（有效步态速度指数（eGVI）、步态速度、步长（平均值和变异系数（CV））、步频、步幅时间（平均值和CV）、单支撑时间（平均值和CV））之间的关系。套索回归确定了检测冻结步态最敏感的特征。
在服药状态下，只有单支撑时间变异系数（β=0.90，p<0.001）可预测冻结步态严重程度（R2=0.80）。在未服药状态下，MDS-UPDRS第三部分（β=0.28，p=0.136）和步幅时间变异系数（β=0.69，p=0.002）可预测冻结步态严重程度（R2=0.70）。服药状态下的有效步态速度指数（β=0.12，p=0.06）、未服药状态下的步长平均值（β=-0.018，p=0.77）和未服药状态下的速度（β=-0.021，p=0.70）区分冻结步态患者和非冻结步态患者的准确率为88%。
讨论/结论：这些结果表明，非冻结步态期间的步态变异性可能反映了冻结步态的一种细微表现。纵向研究应探索冻结步态的发展以及随时间推移相关的步态变化。
