• 027071

    Classification of Cervical Incomplete Spinal Cord Injury based on Trajectory of Recovery: An Analysis of 652 Patients

     

    Christopher Witiw, Michael Fehlings, Jefferson Wilson, Jetan Badhiwala

     

    Introduction: There is substantial heterogeneity in the outcomes of cervical spinal cord injury (SCI).

    Objective: Using a novel approach, we sought to dissociate subgroups of patients with distinct longitudinal trajectories of upper limb motor recovery following cervical motor incomplete SCI.

    Methods: Patients with cervical motor incomplete SCI (AIS C-D; C1-C8) were identified from four prospective multi-center datasets. A group-based trajectory model was fit to upper extremity motor scores out to 2 years. A maximum of three trajectory groups was assumed. A grid search strategy was implemented to derive the optimal statistical model, minimizing the Bayesian information criterion. To determine baseline features that characterize groups, multinomial logistic regression was performed adjusting for AIS grade and neurological level.

    Results: In total, 652 patients were eligible. The final model had three groups. Group 1 (‘severe injury phenotype’) was characterized by severe initial injury with minimal recovery over time. Patients in Group 3 (‘mild injury phenotype’) had relatively preserved motor function and quickly achieved (near)-normal status. Patients in Group 2 (‘recovery phenotype’) had a moderately-severe initial injury with significant subsequent recovery. After adjustment for AIS and neurological level, older age and fall mechanism were associated with membership to less favorable trajectory group. Presence of central cord syndrome and absence of fracture/dislocation also correlated with poorer trajectories. Early surgery and steroids were associated with more favorable trajectory groups.

    Conclusions: Patients with cervical motor incomplete SCI demonstrate distinct profiles of recovery in upper limb motor function. The trajectory a patient is likely to follow may be predicted by baseline characteristics.

We use cookies to improve the performance of our site, to analyze the traffic to our site, and to personalize your experience of the site. You can control cookies through your browser settings. Please find more information on the cookies used on our site here. Privacy Policy