Early Versus Late Surgical Decompression for Central Cord Syndrome: A Propensity Score-Matched Analysis

Abstract

Abstract INTRODUCTION The optimal timing of surgical decompression for central cord syndrome (CCS) remains unclear. We sought to perform a head-to-head comparison of the neurological and functional outcomes of early (<24 h) vs late (≥24 h) surgical decompression in patients with CCS. METHODS Patients who underwent surgery for CCS (AIS grade C or D; LEMS UEMS ≥ 5) were identified from 3 prospective, multi-center SCI datasets (NACTN SCI Registry; STASCIS; NASCIS III). Propensity scores were calculated as the probability of undergoing early (<24 h) versus late (≥24 h) surgery using the logit method with age, injury mechanism, data source, and initial ASIA motor score, AIS grade, and neurological level as covariates. Propensity score matching (nearest neighbor) was performed in a 1:1 ratio. The primary outcome was ASIA motor score at 1 yr. Secondary outcomes included Functional Independence Measure (FIM) motor score at 1 yr. Outcomes were compared between study groups by t-test. Effect sizes for each outcome measure were summarized by mean differences (MDs). RESULTS A total of 300 patients were eligible. Propensity score matching produced a final cohort of 186 patients (N = 93 early surgery; N = 93 late surgery). Patient demographic, injury, and treatment characteristics at baseline were balanced between matched early and late surgery groups. ASIA motor score at 1 yr was significantly higher in the early surgery (mean: 91.8) compared to late surgery (mean: 87.1) group (MD 4.64, 95% CI 0.19 to 9.09, P = .04). Similarly, patients who underwent early (mean: 84.2) as opposed to late (mean: 77.2) surgery achieved better functional status at 1 yr, as measured by the FIM motor score (MD 7.04, 95% CI 0.59 to 13.49, P = .03). CONCLUSION In patients with CCS, we found early surgery, within 24 h of injury, to result in improved neurological and functional outcomes at 1 yr, as compared to late surgery, after adjusting for key potential confounders by a propensity score matching technique.