Gait variability and symmetry assessment with inertial sensors for quantitative discrimination of Trendelenburg sign in total hip arthroplasty patients: a pilot study based on convenience sampling

Abstract Introduction The aim of this pilot study based on convenience sampling was to analyze the feasibility to quantitatively discriminate Trendelenburg sign (TS), a characteristic drop in pelvic position during gait in hip disfunctions, in patients with total hip arthroplasty (THA), by assessing gait variability and symmetry using inertial sensors. Methods Thirteen patients with right THA, divided into two groups with (GTS, n=4) and without TS (GnTS, n=9) assessed by experienced physician, were enrolled in the study. Harris Hip Score was applied for specific evaluation of THA. The protocol consisted in walking on a level treadmill during 3 minutes with two inertial sensors attached at anterior superior iliac spine of both sides. For each left and right step, features were extracted from the Y-axis gyroscope signals: peak value, mean absolute value, standard deviation and range. For each feature, a symmetry ratio was calculated as the ratio between left and right side. Results No significant differences were found in Harris Hip Score between groups. The variability assessed by standard deviation for left step, contralateral to the replaced side, was significantly larger for GTS group (p<0.001). Significant differences in the symmetry ratios were found between GTS and GnTS for all features extracted from gyroscopes Y-axis (W=144, p<0.001). The symmetry ratios for GnTS group were approximately equal one (except for range), whereas for the GTS group they exceed the 10% criterion. Conclusion The variability and symmetry ratios of gait features extracted from inertial sensors were successful to discriminate TS in THA patients.