Does dual-task walking reveal gait differences between patients with cervical spinal stenosis with and without degenerative cervical myelopathy and healthy controls?

Background Degenerative cervical myelopathy (DCM) causes gait impairments in patients with cervical spinal stenosis (CSS), yet early differentiation of mild DCM from CSS without myelopathy remains difficult. Dual-task (DT) walking can reveal subtle gait deficits in other neurological conditions. This study investigated whether DT walking reveals subtle gait differences between CSS patients with and without DCM, and healthy controls, that are not apparent during single-task (ST) walking. Methods Gait analysis was performed in 31 patients with CSS (14 with DCM and 17 without DCM) and 30 healthy controls in a 20 m hallway using seven inertial sensors (RehaGait, Germany) placed on the pelvis, thighs, shanks and feet. Two ST trials at self-selected walking speed were recorded, followed by two DT trials while reciting the months of the year backwards. Spatiotemporal parameters, sagittal joint range of motion (ROM) and their coefficients of variation (CoV) were analyzed using linear mixed-effects models with group, condition and their interaction as fixed effects, age as covariate and participant as random intercept. In patients, the association between the DT effect and disease severity (modified Japanese Orthopaedic Association score, mJOA) was tested using the condition x mJOA interaction in a separate linear mixed-effects model. Results Most patients with DCM had mild disease (median mJOA 16), and the mJOA did not differ significantly from patients with CSS without DCM (median 17, p = 0.068). Across groups, the DT reduced walking speed and cadence (both p < 0.001) and increased the CoV of stride time (p = 0.010). Cadence decreased in all groups (DCM −9.9 steps/min, p = 0.001; CSS without DCM −7.6 steps/min, p = 0.005; healthy controls −7.0 steps/min, p < 0.001), whereas walking speed decreased significantly in healthy controls (−0.08 m/s, p = 0.007) and patients with CSS without DCM (−0.10 m/s, p = 0.020), but not in patients with DCM (−0.09 m/s, p = 0.052). The DT effect did not differ between groups for any parameter (group × condition interaction p ≥ 0.705 for primary outcomes). Irrespective of the task, both patient groups walked slower than healthy controls (p = 0.003) but did not differ from each other. The DT effect was not associated with the mJOA (p ≥ 0.18). Conclusions Walking while reciting the months backwards elicited similar gait changes in patients with CSS with and without DCM and in healthy controls. In this cohort with predominantly mild DCM, this DT paradigm did not provide additional value for distinguishing the presence of DCM in patients with CSS.