European Spine Journal - 2026-10-14 - Journal Article
Multifactorial prediction of mechanical complications after lumbar fusion using bone quality and turnover markers.
Ogura Y, Kato T, Nishiyama Y, Varthi A, Suzuki Y
Topics
Key Takeaway
A multivariate model combining TRACP-5b, fusion levels, and bone density parameters achieved AUC 0.830 for predicting mechanical complications after lumbar fusion, versus AUC ≤0.688 for any single bone quality metric alone.
Summary Depth
Choose how much analysis to show on this article page.
Summary
This study asked whether preoperative bone health indices—DEXA-derived BMD/T-score, CT Hounsfield units, and serum bone turnover markers (P1NP, TRACP-5b)—predict mechanical complications after lumbar fusion with pedicle screw instrumentation. In 120 patients, mechanical complications occurred in 40 (33%); femoral T-score (AUC 0.688) and BMD (AUC 0.681) showed limited individual discrimination, while multivariable logistic regression identified TRACP-5b (OR 1.02, p=0.042) and number of fused levels (OR 5.24, p=0.001) as independent predictors. The combined multivariate model reached AUC 0.830, demonstrating that integrating metabolic bone markers with surgical variables substantially outperforms any single bone quality parameter.
Key Limitation
Mechanical complications were not stratified by type (screw loosening, cage subsidence, rod fracture, pseudarthrosis), preventing determination of which specific failure mode TRACP-5b most strongly predicts.
Original Abstract
PURPOSE
To evaluate the predictive value of preoperative bone health indices including T-score, bone mineral density (BMD), Hounsfield units (HU), and serum bone turnover markers for postoperative mechanical complications following lumbar fusion surgery.
METHODS
This retrospective study analyzed prospectively collected data from 120 patients who underwent lumbar fusion with pedicle screw instrumentation between August 2021 and June 2024. Preoperative assessments included dual-energy X-ray absorptiometry (DEXA)-derived BMD and T-scores, computed tomography-derived HU values, and serum bone turnover markers including procollagen type I N-terminal propeptide (P1NP) and tartrate-resistant acid phosphatase type 5b (TRACP-5b). Mechanical complications were evaluated radiographically at ≥ 12 months postoperatively. Receiver operating characteristic (ROC) curve analyses were performed for individual parameters, and multivariable logistic regression was used to identify independent predictors.
RESULTS
Mechanical complications occurred in 40 patients (33%). In univariate analysis, patients with complications were significantly older and had lower femoral BMD and T-scores and higher P1NP levels (all p < 0.05). ROC analysis demonstrated modest predictive ability for femoral T-score (AUC = 0.688) and BMD (AUC = 0.681). In multivariable logistic regression, TRACP-5b (OR = 1.02, p = 0.042) and the number of fused levels (OR = 5.24, p = 0.001) were associated with complications. A multivariate model combining significant variables yielded an AUC of 0.830, indicating good overall discrimination.
CONCLUSIONS
Preoperative T-score, BMD, and HU values alone showed limited predictive accuracy for mechanical complications after lumbar fusion. However, integrating serum bone turnover markers and surgical factors substantially improved predictive performance. A multifactorial preoperative assessment incorporating bone density, metabolic, and mechanical parameters may better stratify risk and guide surgical planning in patients with poor bone quality.