Gait analysis on chronic Low Back Pain clusters before rehabilitation program: Kinematic and kinetic approaches -

Abstract

Introduction Chronic Low Back Pain (cLBP) can have major impacts on life, with psycho-social consequences. An initial analysis based on psycho-social subgrouping scheme was performed last year on a heterogeneous population (Delpierre et al. 2017). Using several psycho- social scales and questionnaires, two psychosocial clusters (cluster 2 with higher scores) were described before a multidisciplinary reconditioning program and compared in relation to pain intensity, muscular activity on erector spinae during flexionextension, and Global Trunk Mobility (GTM). This global mobility is based on an index of trunk mobility computed from five trunk segments during flexionextension, bending and rotation (Delpierre et al. 2017). Such global mobility could be correlated with lower mobility during walking. Gait analysis allowed quantifying kinematics and kinetics with Ground Reaction Forces (GRF). Several authors have described differences in back kinematics between cLBP and asymptomatic subjects (Lee et al. 2007; Kuai et al. 2017). These patients presented lower pelvic mobility, a higher standard deviation in the case of cLBP (Kuai et al. 2017) and used strategies to attenuate the amount of force imposed on their body (Lee et al. 2007). Such variations seem to confirm the existence of several sub-groups or clusters. Moreover, pelvic disequilibrium may be associated with a difference in GRF (Kuai et al. 2017). Lateral GRF present a classically low amplitude with considerable variations, thus more attention is given to anteroposterior and vertical forces. These forces are associated with two maximal/minimal amplitudes in reference to a braking phase and an impulsion phase. Therefore, in relation to the clusters mentioned previously, it could be interesting to compare anteroposterior and vertical GRF. Do patients in cluster2 minimize GRF and pelvic mobility in reference to patients in cluster1? The aims of this study were to: (1) compare pelvic mobility and reaction forces during gait cycles between these clusters, and (2) evaluate correlation between the index of trunk mobility and specific gait parameters. 2. Methods 2.1. Population Patients (n=18; 11 women and 7 men; 37.22 years (6.10)) with cLBP previously enrolled in a clinical trial comparing behavioral physical therapy interventions. They were classified on two clusters from previously described psycho-social scales. They were analyzed before rehabilitation program. Nine patients of each cluster were included in this study. 2.2. Material and method Gait analysis were realized with a 3D motion capture system (ViconT10, 100 Hz) which recorded data for 34 passive markers (14mm) in relation to plug-in gait model. Only lower limbs and pelvic were considered in this study. From GRF (AMTI force plate, 1000 Hz), minima/maxima in anteroposterior axis (defined braking and propulsion values) and vertical axis (defined reception and impulsion values) were computed (Matlab). Ranges of Motion (RoM) on pelvis in each plane were computed too. Only correlations between left side parameters and index of trunk mobility were studied. Per subject, five strides were averaged. Analyses were performed with Statistica (V13, Dell, USA). Results were presented as: mean (standard deviation). Mann-Whitney tests and Spearman's rang correlation were applied (-rfpy1≤0.05). 3. Results and discussion 3.1. Gait analysis with two clusters Cluster1 walked at 1.11 (0.17) m s-1 with a cadence of 98.13 (7.65) steps min-1. This cluster presented pain evaluated at EVA = 6.5 (0.5). Cluster2 walked at 1.08 (0.20) m s-1 with a cadence of 91.54 (10.67) steps min-1. This second cluster presented pain evaluated at EVA = 8 (2). The measurement of speed and cadence revealed no difference between these two clusters. Hence, pelvic mobility and GRF were compared independently of free speed. Table 1 reveals differences in pelvic mobility and GRF. No difference was noticed for pelvic mobility. Regarding GRF, anteroposterior forces presented similar maxima/minima values. The observation of vertical forces revealed differences in reception value: cluster 2 (which was previously associated with higher psycho-social scores) had lower reception values on each side. Thus, these two clusters were associated with differences in vertical forces. Cluster2 minimized the reception phase, which could correspond to a strategy described earlier to reduce the impact of heel strike (Lee et al. 2007). 3.2. Correlations between index of trunk mobility and gait parameters Regarding the index of trunk mobility, the correlations between this index and the gait parameters were: -0.35 (RoM of pelvis in the sagittal plane), •0.07 (RoM pelvis in the frontal plane), -0.01 (RoM pelvis in the transversal plane), -0.12 (GRFAnteroposterior Impulsion), 0.51° (GRFAnteroposterior Braking), 0.48° (GRF- Vertical Reception), -0.04 (GRF- Vertical Impulsion). Two correlations appeared significant but moderate. Therefore, the reduction in trunk mobility and braking value seems moderately correlated. This could suggest specific strategies (Lee et al. 2007). Pain distribution seems to be influenced by trunk mobility and GRF. Although preferred speed was tested, the patients adopted a gait that was less painful or aggravating (Lee et al. 2007). Several limitations should be mentioned. Firstly, only nine patients for each cluster were studied. Secondly, the incidences of the rehabilitation program were not defined. Finally, only preferred speed was evaluated and only linear correlation was tested. 4. Conclusions With cLBP shown for two psycho-social clusters, this study determined specific differences in the reception value observed for GRF. It could be interesting to evaluate the magnitude of kinesiophobia for each cluster (Thomas et al. 2010). (Figure Presented).