Early Postoperative Neurological Changes After Surgical Treatment of Traumatic Spinal Injury: A Retrospective Cohort Study

Introduction: Early neurological changes after surgical treatment of traumatic spinal injury are heterogeneous and may be influenced by baseline neurological status, injury characteristics, and the timing of assessment.This study aimed to describe early postoperative changes in American Spinal Injury Association Impairment Scale (AIS) grade and explore factors associated with neurological improvement.Methods: This retrospective observational cohort study included adults who underwent surgery for documented traumatic spinal injury at a tertiary spine center in Mexico City, Mexico, between February 2025 and May 2026.The primary outcome was an increase of at least one AIS grade between the preoperative examination and the latest documented postoperative assessment.Associations between anatomical region and neurological improvement were examined, and an exploratory multivariable logistic regression model was fitted.Bootstrap resampling with 1,000 requested samples was used to examine coefficient uncertainty; the available bias-corrected and accelerated (BCa) 95% confidence interval was reported for the relevant coefficient estimate.Results: Fifty-two patients were included (mean age, 39.9 ± 12.2 years; 43 (82.7%)male).Baseline AIS grades were A in 18 patients (34.6%),B in two (3.8%),C in one (1.9%),D in 29 (55.8%), and E in two (3.8%).At the latest documented postoperative assessment, 24 patients (46.2%) had improved by at least one AIS grade, 27 (51.9%)remained unchanged, and one (1.9%)experienced neurological deterioration.Improvement rates ranged from 30.0% in cervical/cervicothoracic injuries to 53.8% in lumbar injuries; the association between anatomical region and improvement was not statistically significant (Pearson χ² = 1.418, df = 3, p = 0.701).The exploratory multivariable model, which included 50 patients, was not statistically significant overall (χ² = 5.930, df = 6, p = 0.431; Nagelkerke R² = 0.149), and no individual predictor was statistically significant.Bootstrap analysis showed substantial uncertainty in some coefficient estimates.Four recorded transitions from AIS A to AIS E could not be independently validated from all required examination components.Median postoperative follow-up was 33 days (interquartile range, 21-55.5 days).Conclusions: Nearly half of the cohort had a recorded increase of at least one AIS grade during postoperative follow-up.However, the short and variable assessment interval, uncertainty surrounding several AIS grade transitions that could not be confirmed through the available clinical documentation, and instability of exploratory multivariable estimates limit interpretation.These findings describe early postoperative neurological changes and do not establish long-term recovery, independent prognostic associations, or a causal effect of surgery.