Freezing of gait: Problem analysis and rehabilitation strategies

Abstract

Freezing remains a disabling problem with gait in patients with Parkinson's disease (PD). We conducted a series of investigations into gait events, prior to freezing, to understand the aetiology of this complex phenomenon. Using a 3D gait laboratory, with an integrated EMG-system, we compared gait cycles before freezing, with normal gait within the same population of freezers. Gait analysis was performed in 14 patients with Parkinson's disease, in the off-phase of the medication cycle with a mean age of 64.5 years (± 4.7) and mean Hoehn and Yahr stage of 3.7. After several trials of normal walking, a distracting cognitive task and/or obstacles on the walkway were used to invoke freezing. Firstly, we analysed the spatiotemporal pattern of gait leading up to freezing or festination [Nieuwboer A, Dom R, De Weerdt W, Desloovere K, Vangheluwe S. Electromyographic profiles of gait at the onset of freezing in Parkinson's disease. Mov Disord 2001;16(6):1066–75]. The duration of the pre-freezing gait cycles were reduced by 35% in comparison with normal (off) gait. Within the gait cycle, the swing phase was significantly reduced by 5.2% before freezing. In addition, a progressive decrease in stride length and an excessively increased cadence prior to freezing was observed. These findings suggest that freezing is caused by a combination of an increasing inability to generate stride length superimposed on a dyscontrol in the cadence of walking. To further investigate this hypothesis, we investigated whether PD patients demonstrate disturbed co-ordination of the gastrocnemius (GS) and tibialis anterior (TA) muscles prior to freezing, affecting both the timing and magnitude of EMG-activity [Nieuwboer A, Dom R, De Weerdt W, Desloovere K, Janssens L, Vangheluwe S. Electromyographic profiles of gait prior to freezing episodes in patients with Parkinson's disease. Brain 2004;127:650–1660]. In 11 of the 14 previous patients (those who displayed actual freezing episodes), we analysed the rectified and smoothed EMG-signals, expressed as a percentage of the gait cycle. EMG-onset was determined using a preset threshold, corrected following visual inspection. The magnitude of EMG was calculated by integrating EMG-signals over time. In addition, iEMG was normalised to allow for the altered timing of activity. We showed that significantly abnormal timing occurred in the TA and GS muscles with overall preserved reciprocity. TA activity occurred prematurely during stance. GS activity was also premature during stance and prolonged in swing, indicating hastened muscle activity. The magnitude of EMG activity generated in both lower limb muscles was reduced when compared with normal walking. However, this reduction was mostly due to the shortened time in which the muscles were active. In TA, we noticed an increased level of peak and normalised EMG activity, indicating that patients attempt to compensate for the tendency to freeze by pulling the leg into swing. Eventually, a breakdown of movement may occur as patients find it difficult to generate sufficient swing activity quickly. As PD patients with freezing apparently have a central disturbance of gait cycle timing, we hypothesised that freezers would be prone to error in synchronising their stepping to auditory cues, and, as a result, would benefit less from cueing. Therefore, we compared two groups of 10 PD patients, with and without freezing, with 10 control subjects of a similar age [Nieuwboer A, Willems A, Chavret F, Kwakkel G, Jones D, Dom R. Synchronising walking to various auditory cueing frequencies: differences between freezers and non-freezers with Parkinson's disease. Mov Disord 2004;19(Suppl.9):S156]. Freezers and non-freezers were comparable for UPDRS scores, Hoehn and Yahr stage and baseline gait speed. Five cueing frequencies were presented in a randomized order using a metronome with a baseline frequency of 10% and 20% above and 10% and 20% below baseline. Surprisingly, freezers demonstrated a greater ability than non-freezers to synchronise their gait to the imposed cueing frequencies and possessed a normal direction of error. Non-freezers reached their optimal speed and stride length at the +10% cueing rhythm, similar to controls. However, freezers showed an inability to improve their gait speed beyond baseline levels in response to auditory cues, demonstrating a marked reduction of stride length at the +10% cueing rhythm.