Clinical significance of stride length variability in patients with degenerative cervical myelopathy

Gait impairment is a major manifestation of degenerative cervical myelopathy (DCM). Most previous gait studies have focused on mean spatiotemporal parameters, whereas gait variability may provide a more sensitive indicator of gait instability and fall risk. However, the clinical significance of stride length variability in patients with DCM remains unclear. This study aimed to investigate the clinical significance of stride length variability in patients with DCM by examining its associations with objective and patient-reported clinical measures, as well as other gait parameters. Sixty-three consecutive patients with DCM who completed all assessments were included. Stride length variability was quantified using the coefficient of variation (CV) of stride length during treadmill walking using the KinemaTracer three-dimensional gait analysis system. Associations between stride length variability and objective clinical measures, gait parameters, and the domains of the Japanese Orthopaedic Association Cervical Myelopathy Evaluation Questionnaire (JOACMEQ) were evaluated. Receiver operating characteristic analysis was performed to determine the optimal gait speed cutoff for identifying high stride length variability (CV ≥ 4%). Individual JOACMEQ lower extremity function items were examined in separate age- and sex-adjusted linear regression models, followed by additional adjustment for overground gait speed and quadriceps strength. The mean age was 68.9 ± 11.1 years, and 46 of 63 patients (73.0%) were men. Stride length variability demonstrated a moderate negative correlation with overground gait speed (ρ = −0.647). The optimal gait speed cutoff for identifying high stride length variability was 1.08 m/s (AUC, 0.797; 95% CI, 0.685–0.908; sensitivity, 84.0%; specificity, 65.6%). Among the JOACMEQ domains, lower extremity function demonstrated the strongest correlation with stride length variability (ρ = −0.441). In age- and sex-adjusted models, difficulty standing on one leg (standardized β = −0.271, p = 0.042) and walking for more than 15 min (β = −0.261, p = 0.036) were associated with higher stride length CV. Neither association remained statistically significant after this additional adjustment. In patients with DCM, stride length variability was associated with gait speed and patient-reported lower-extremity function. These findings suggest that stride length variability may help characterize gait impairment in this population. Not applicable.