Lumbar lordosis restoration by minimally invasive short-segment fusion with anterior column realignment for adult spinal deformity: minimum 2-year follow-up

Abstract

OBJECTIVE The efficacy of anterior column realignment (ACR) remains relatively unclear, possibly because some safety concerns have limited its adoption and extensive evaluation. The authors aimed to study whether a minimally invasive surgery (MIS) triad consisting of ACR, lateral lumbar interbody fusion, and percutaneous pedicle screw fixation in a select group of adult spinal deformity (ASD) patients helps shorten fusion length without compromising clinical and radiographic outcomes over a minimum 2-year follow-up period. METHODS A series of 61 ASD patients (mean age 72.8 years) with pelvic incidence (PI) − lumbar lordosis (LL) (PI-LL) mismatch > 10° underwent the short-segment MIS triad (mean fusion length 3.0 levels) as a single-stage operation with a mean operative time and estimated blood loss of 157 minutes and 127 mL, respectively. Exclusion criteria were 1) thoracic scoliosis as the main deformity, 2) thoracolumbar junction kyphosis > 25°, 3) ankylosed facet joints, and 4) previous spinal fusion surgery. Seven patients, who needed fusion to be extended to S1, underwent mini-open transforaminal lumbar interbody fusion at L5–S1. RESULTS The segmental disc angle at the ACR level more than quintupled, averaging from 2.9° preoperatively to 18.9° at the latest follow-up (p < 0.0001). LL, in turn, nearly doubled from 17.0° to 32.8° (p < 0.0001) and PI-LL decreased by nearly half from 28.8° to 13.2° (p < 0.0001). At the same time, other spinopelvic deformity parameters as well as Oswestry Disability Index (ODI) scores significantly improved. Patients were divided into two groups at the latest postoperative evaluation: 36 patients whose PI-LL improved to < 10° and 25 patients who maintained a PI-LL mismatch > 10°. Binary logistic regression revealed preoperative PI-LL mismatch as the only factor that significantly influenced this dichotomous separation postoperatively. Receiver operating characteristic curve analysis identified the critical preoperative mismatch of 26.4° with 68% sensitivity and 84% specificity. Despite this different radiographic consequence, the two groups had an equally successful clinical outcome with no significant difference in ODI scores. CONCLUSIONS As long as the ASD characteristics are consistent with the authors’ exclusion criteria, the short-segment MIS triad served as an excellent surgical option in the patients with preoperative PI-LL mismatch < 26.4°, but the technique also worked well even in those with a mismatch > 26.4°, although ideal spinopelvic alignment targets were not necessarily achieved in these patients.