Association between instrumental stability measures of gait and clinical rating scales in stroke patients

Main topics: Analysis of gait and motor disorders; Movement deviation indexes Introduction and aim: Subjects who suffered a stroke often experience alterations in the normal pattern of walking and a higher fall risk [1]. Clinical rating scales represent the standard approach used by clinicians for motor function assessment of post-stroke patients; this approach is however highly dependent on the clinician's subjective judgement [2]. Instrumental measurements of trunk accelerations during gait resulted promising in the assessment of gait stability and fall risk in healthy elderly subjects [3]. A more reliable and objective quantification of motor function could be obtained from the integration of stability measures and clinical rating scales. In order to evaluate the potential clinical benefits derivable from this approach, the association between stability measures and clinical scales has to be assessed. The aim of the present study is the assessment of the relationship between instrumental gait stability measures of trunk accelerations and some of the most used clinical rating scales in a sample of post-acute stroke patients. Patients/materials and methods: Thirty-three subjects post-acute stroke patients with hemiparesis (56 ± 14 years, 71 ± 14 kg) participated in the study. Seventeen subjects were not able to walk without a cane. Data were collected at Sol et Salus Hospital (Rimini, Italy). Motricity Index (MI), Trunk Control Test (TCT), Functional Ambulation Category (FAC), Walking Handicap Scale (WHS), Rivermead Mobility Index (RMI), Cumulative Illness Rating Scale (SI and CI), Timed-Up and Go Test (TUG) and Two Minute Walk Test (2MWT) were administered to subjects by the same operators. The 2MWT was administered with the addition of an IMU located on the trunk, at the height of the fifth lumbar vertebra. Four gait stability measures, namely Harmonic Ratio (HR), Index of Harmonicity (IH), Multiscale Entropy (MSE), and Recurrence Quantification Analysis (RQA), were calculated on trunk acceleration signal during gait obtained from 2MWT. Each measure was calculated for anterior-posterior (AP), mediolateral (ML) and vertical (V) acceleration directions. Log transformed measures were then used as inputs for linear regression models. Results: Subjects who used a cane showed a high correlation between MSE, RQA and TCT, RMI. More sparse but significant correlation has been found between MSE and WHS, CIRS. IH showed no correlation with clinical parameters. Subjects who were able to walk without a cane showed high correlation values between MSE, RQA and TUG, 2MWT and WHS. TUG, 2MWT and SI correlated in a lesser extent with HR. IH showed high correlation values with TUG and RMI; RMI correlated with MSE and RQA also. Discussion and conclusions: In subjects who walked with a cane, correlations between TCT and MSE were particularly consistent, highlighting an affinity between the instrumental measure of complexity of trunk acceleration and the clinical assessment of trunk control. For subjects who were able to walk without a cane TUG, 2MWT showed high correlations with stability measures, in particular with MSE and RQA (mainly in the ML direction), highlighting the importance of the medio-lateral control of the trunk during gait. In conclusion, gait stability measures based on trunk accelerations (in particular MSE and RQA) showed promising correlation with clinical scales in stroke patients, and could complement the standard clinical scores in the assessment of locomotor performance of subjects with stroke, helping clinicians and physical therapists in the patient's rehabilitation process. Moreover, the use of instrumental measurements could lead, in the future, to a more reliable quantification of locomotor features, allowing avoiding inter-operator differences.