Patient and procedural risk factors for decline in lower-extremity motor scores following adult spinal deformity surgery.

Abstract

The purpose of this study was to discern factors that differentiate patients who experience postoperative lower-extremity motor function decline in the early postoperative period. Adult spinal deformity (ASD) patients who were enrolled in a multicenter, observational, and prospectively collected study from 2018 to 2021 at 18 spinal deformity centers in North America were queried. Eligible participants met at least one of the following radiographic and/or procedural inclusion criteria: pelvic incidence minus lumbar lordosis (PI-LL) ≥ 25°, T1 pelvic angle (T1PA) ≥ 30°, sagittal vertical axis (SVA) ≥ 15 cm, thoracic scoliosis ≥ 70°, thoracolumbar scoliosis ≥ 50°, global coronal malalignment ≥ 7 cm, 3-column osteotomy, spinal fusion ≥ 12 levels, and/or age ≥ 65 years with ≥ 7 levels of instrumentation. Patients with an inflammatory or autoimmune disease and those who were incarcerated or pregnant were excluded, as were non-English speakers. Only patients with baseline and 6-week postoperative lower-extremity motor score (LEMS) were analyzed. Patient information, including demographic data, operative data, patient-reported outcomes, and radiographic parameters, were collected. Univariate and multivariable logistic regression models were built to quantify the degree to which a patient's postoperative LEMS decline was related to demographic and clinical characteristics. In total, 205 patients (mean age 61.5 years, mean total instrumented levels 12.6, 67.3% female, 54.2% primary cases, 79.5% with pelvic fixation) were evaluated. Of these 205 patients, 32 (15.5%) experienced LEMS decline in the perioperative period. These patients were older (p = 0.0014) and had greater BMI (p = 0.0176), higher frailty scores (p = 0.047), longer operating room times (p = 0.033), and greater estimated blood loss (p < 0.0001), and they were more frequently observed to have intraoperative neurophysiological monitoring (IONM) changes (p = 0.018). The deteriorated cohort had greater C7SVA at baseline (p = 0.0028) but were comparable in terms of all other radiographic parameters. No radiographic differences were seen between the groups at the 6-week visit; however, the deteriorated cohort experienced greater change in PI-LL (p < 0.0001), lumbar lordosis (p = 0.0461), C7SVA (p = 0.0004), and T1PA (p < 0.0001). Multivariate logistic regression demonstrated that the presence of IONM changes and each degree of negative change in T1PA conferred 3.71 (95% CI 1.01-13.42) and 1.09 (1.01-1.19) greater odds of postoperative LEMS deterioration, respectively. In this study, 15.6% of ASD patients incurred LEMS decline in the perioperative period. The magnitude of change in global sagittal alignment, specifically T1PA, was the strongest independent predictor of LEMS decline, which has implications for surgical planning, patient counseling, and clinical research.