Hounsfield unit of screw trajectory as a predictor of pedicle screw loosening after single level lumbar interbody fusion

Abstract

ObjectThis study aims to clarify the clinical potential of Hounsfield unit (HU), measured on computed tomography (CT) images, as a predictor of pedicle screw (PS) loosening, compared to bone mineral density (BMD). MethodsA total of 206 screws in 52 patients (21 men and 31 women; mean age 68.2 years) were analyzed retrospectively. The screws were classified into two groups depending on their screw loosening status on 3-month follow-up CT (loosening screw group vs. non-loosening screw group). Preoperative HU of the trajectory was evaluated by superimposing preoperative and postoperative CT images using three-dimensional image analysis software. Age, sex, body mass index, screw size, BMD of lumbar, and HU of screw trajectory were analyzed in association with screw loosening. Multivariate logistic regression analysis was performed, and the thresholds for PS loosening risk factors were evaluated using a continuous numerical variable and receiver operating characteristic (ROC) curve analyses. The area under the curve (AUC) was used to determine the diagnostic performance, and values > 0.75 were considered to represent good performance. ResultsThe loosening screw group contained 24 screws (12%). Multivariate analysis revealed that the significant independent risk factors were not BMD but male sex [P = 0.028; odds ratio (OR) 2.852, 95% confidence interval (CI) 1.120–7.258] and HU of screw trajectory (P = 0.006; OR 0.989, 95% CI 0.980–0.997). ROC curve analysis demonstrated that the AUC for HU of screw trajectory for women was 0.880 (95% CI 0.798–0.961). The cutoff value was 153.5. AUC for men was 0.635 (95% CI 0.449–0.821), which was not considered to be a good performance. ConclusionsLow HU of screw trajectories was identified as a risk factor of PS loosening for women. For female patients with low HU, additional augmentation is recommended to prevent PS loosening.