Background and Objectives: Ventral stabilization of thoracolumbar spine fractures can be achieved using different interbody reconstruction techniques, including titanium cages, vertebral body replacements (VBR), and autologous pelvic bone grafts (APBG). Although all approaches aim to restore anterior column stability and alignment, comparative data on long-term patient-reported outcomes remain limited. The objective of this study was to compare long-term patient-reported wellbeing following ventral stabilization using these three techniques. Materials and Methods: A retrospective, non-randomized single-center cohort study with prospective follow-up was analyzed. Treatment allocation was indication-based. Ninety-one patients treated between 2008 and 2018 underwent ventral stabilization using cage implantation (n = 12), vertebral body replacement (n = 45), or autologous pelvic bone grafting (n = 34). Clinical outcome was assessed at least 12 months postoperatively using a modified Visual Analog Scale Spine Score (VAS-Spine). Statistical analysis included linear and ordinal regression adjusted for age and sex. Potential baseline differences between groups were considered in the interpretation of the results. Results: Sixty-three patients (mean age 52 ± 15 years; 41% female) completed follow-up. The mean VAS-Spine score was lowest after cage implantation (2.7 ± 3.6), followed by VBR (3.9 ± 2.8) and APBG (4.9 ± 1.8; p* = 0.021). The observed difference between cage and APBG approached the minimal clinically important difference reported for VAS-based measures. Patients treated with cage implantation reported less pain during rest and activity and fewer limitations in daily life. No significant differences were observed regarding age or sex. Conclusions: In this observational cohort, cage implantation was associated with more favorable patient-reported outcomes compared with VBR and APBG. Autologous pelvic bone grafting was associated with worse patient-reported outcomes, potentially related to donor-site morbidity. Given the non-randomized design and potential confounding, these findings should be interpreted as associative and hypothesis-generating.
Jäckle et al. (Wed,) studied this question.