Categorization of gait patterns in adults with cerebral palsy: A clustering approach

Abstract Gait patterns in adults with cerebral palsy have, to our knowledge, never been assessed. This contrasts with the large number of studies which have attempted to categorize gait patterns in children with cerebral palsy. Several methodological approaches have been developed to objectively classify gait patterns in patients with central nervous system lesions. These methods enable the identification of groups of patients with common underlying clinical problems. One method is cluster analysis, a multivariate statistical method which is used to classify an entire data set into homogeneous groups or “clusters”. The aim of this study was to determine, using cluster analysis, the principal gait patterns which can be found in adults with cerebral palsy. Data from 3D motion analyses of 44 adults with cerebral palsy were included. A hierarchical cluster analysis was used to subgroup the different gait patterns based on spatiotemporal and kinematic parameters in the sagittal and frontal planes. Five clusters were identified (C1–C5) among which, 3 subgroups were determined, based on spontaneous gait speed (C1/C2: slow, C3/C4: moderate and C5: almost normal). The different clusters were related to specific kinematic parameters that can be assessed in routine clinical practice. These 5 classifications can be used to follow changes in gait patterns throughout growth and aging as well to assess the effects of different treatments (physiotherapy, surgery, botulinum toxin, etc.) on gait patterns in adults with cerebral palsy.