Variability of gait analysis in children with Cerebral Palsy across different conditions

Three-dimensional gait analysis (3DGA) is a standard part of clinical follow-up in children with cerebral palsy (CP). It is the ‘gold standard’ for measurement and description of gait [1]. However, individual variability can be observed and should be considered when interpreting the results. Variations in 3DGA-data can arise naturally (intrinsic) or from experimental errors (extrinsic) [2]. This study focused on the inter-trial (intrinsic) and inter-session (extrinsic) variability in children with CP. Previous studies [3] only investigated the variability in the barefoot condition, and mainly during overground walking, however most of this population receives ankle-foot orthoses (AFO) to improve their gait and gait analysis may also be formed on a treadmill. What is the inter-trial & inter-session standard error of 3DGA of children with CP and how is it different across conditions (barefoot, ankle-foot orthoses, overground, treadmill)? Seven children with CP (2 M; 5 F; 9.6 y ± 3.6 y; GMFCS level I-III; uni- (N=2) and bilateral (N=5) CP) performed repeated gait analyses (1–14 days interval) assessed by the same clinician at their preferred walking speed, in four conditions (barefoot overground, barefoot treadmill, AFO overground and AFO treadmill). Kinematic & kinetic data in the sagittal plane were collected (VICON, plug-in-gait markerset) from 10 steps overground & 15 steps on the treadmill. From the 2 CP children with CMFCS-level III, no overground kinetic data were included. The patient’s most involved side, based on the highest Modified Ashworth Scale score of gastrocnemius was used for further analysis. Inter- and inter-session timepoint-by-timepoint standard error was calculated from the continues waveforms in the sagittal plane, according to the methods of Schwartz et al. [2]. The inter-trial error (Fig. 1) served as a baseline indication of intrinsic variability. The ankle joint angle was the most repeatable (errors: 0.342°-0.575°), except for the AFO treadmill condition. The hip had the highest errors for joint moments (0.14–0.56 Nm/kg) and power (0.020–0.134 W/kg). Same conclusions can be made for the inter-session errors (Fig. 1), except for ankle power, which showed the highest errors (0.120–0.314 W/kg). In overground condition, the inter-trial and -session error for the kinematic parameters were smaller compared to the treadmill condition. For the kinetic parameters, the treadmill-condition had the smallest inter-trial and -session errors. Fig. 1Download : Download high-res image (120KB)Download : Download full-size image The observed inter-trial error of the joint angles is in line with the findings of Schwartz et al. (healthy adults: hip=1.17°; knee=0.60°; ankle≈0.65°). The decreased standard-error of the ankle angle in the AFO condition can be explained by the fixation of the ankle in the orthosis, reducing the range of motion. For kinetics, the treadmill condition resulted in smaller errors in comparison to the overground condition, which could be explained by the constant walking speed.