Vertebral Bone Quality Score as a Predictor of Pedicle Screw Loosening Following Surgery for Degenerative Lumbar Disease

Abstract

Study design: Retrospective study.

Objective: To investigate whether magnetic resonance imaging-based vertebral bone quality (VBQ) score can predict pedicle screw loosening in patients who underwent pedicle screw fixation, and to compare, which measurement, the VBQ score or the Hounsfield unit (HU) value, is more predictive of pedicle screw loosening.

Summary of background data: In clinical work, we found that patients with screw loosening had higher VBQ scores. In addition, some studies have found a correlation between VBQ scores and osteoporosis.

Patients and methods: Patients who were treated with lumbar pedicle screw fixation were reviewed. The VBQ score was measured using magnetic resonance imaging scans. The HU value for L1 to L4 lumbar bone mineral density was measured with computed tomography scans. Logistic regression analysis was used to identify factors associated with pedicle screw loosening. Receiver-operating characteristic curve analysis was used to evaluate the value of VBQ scores in predicting pedicle screw loosening.

Results: A total of 156 patients were included in the final analysis. The pedicle screw loosening rate was 35% (55 of 156 patients). The postoperative low-back pain visual analog scale score was higher in the loosening group (3.0 ± 2.0 vs . 2.4 ± 1.8; P < 0.05). The VBQ score was higher in the loosening group than in the nonloosening group (3.28 ± 0.58 vs . 2.82 ± 0.50; P < 0.01). In multivariable analysis, nonsingle segment fixation [odds ratio (OR): 3.992; 95% CI: 1.643-9.701; P = 0.002], lowest instrumented vertebrae at S1 (OR: 3.378; 95% CI: 1.387-8.226; P = 0.007), HU value (OR: 0.988; 95% CI: 0.976-1.000; P = 0.047), and VBQ score (OR: 3.908; 95% CI: 1.624-9.405; P = 0.002) were factors associated with screw loosening. The areas under the curve for using the VBQ score and HU value to predict pedicle screw loosening were 0.720 and 0.702, respectively. The optimal VBQ score threshold was 3.05 for predicting pedicle screw loosening (sensitivity: 0.655; specificity: 0.713).

Conclusions: The VBQ score was an influential factor associated with lumbar pedicle screw loosening, and a higher VBQ score was significantly correlated with a higher risk of screw loosening. The VBQ score was a better predictor of pedicle screw loosening than the HU value in patients who underwent pedicle screw fixation for degenerative lumbar disease.

Figure 1

Flowchart highlighting the inclusion and exclusion criteria. CT indicates computed tomography; LIV, lowest instrumented vertebrae; MRI, magnetic resonance imaging; PLIF, posterior lumbar interbody fusion.

Figure 2

Noncontrast-enhanced T1-weighted MRI image shows the determination of SI of L1 to L4 and SI of CSF using ROIs. CSF indicates cerebrospinal fluid; MRI, magnetic resonance imaging; ROI, regions of interest; SI, signal intensity.

Figure 3

CT images show HU values determined using the ROI. A, Shows the axial plane of the ROI in the middle of L1 to L4. B, Shows that PACS software automatically calculates the HU value of the ROI. CT indicates computed tomography; HU, Hounsfield unit; ROI, region of interest.

Figure 4

Radiograph of patients with postoperative screw loosening.

Figure 5

The scatter plot shows the distribution of patients in the loosening and nonloosening groups based on the mean HU values of the L1 to L4 and VBQ scores. HU indicates Hounsfield unit; VBQ, vertebral bone quality.

Figure 6

Receiver operating characteristic curve analysis was used to evaluate the mean HU values of the L1 to L4 and VBQ scores in predicting pedicle screw loosening. HU indicates Hounsfield unit; VBQ, vertebral bone quality.