Thoracic vertebrectomy and spinal reconstruction via anterior, posterior, or combined approaches: clinical outcomes in 91 consecutive patients with metastatic spinal tumors

Abstract

Adequate decompression of the thoracic spinal cord often requires a complete vertebrectomy. Such procedures can be performed from an anterior/transthoracic, posterior, or combined approach. In this study, the authors sought to compare the clinical outcomes of patients with spinal metastatic tumors undergoing anterior, posterior, and combined thoracic vertebrectomies to determine the efficacy and operative morbidity of such approaches. A retrospective review was conducted of all patients undergoing thoracic vertebrectomies at a single institution over the past 7 years. Characteristics of patients and operative procedures were documented. Neurological status, perioperative variables, and complications were assessed and associations with each approach were analyzed. Ninety-one patients (mean age 55.5 +/- 13.7 years) underwent vertebrectomies via an anterior (22 patients, 24.2%), posterior (45 patients, 49.4%), or combined anterior-posterior approach (24 patients, 26.4%) for metastatic spinal tumors. The patients did not differ significantly preoperatively in terms of neurological assessments on the Nurick and American Spinal Injury Association Impairment scales, ambulatory ability, or other comorbidities. Anterior approaches were associated with less blood loss than posterior approaches (1172 +/- 1984 vs 2486 +/- 1645 ml, respectively; p = 0.03) or combined approaches (1172 +/- 1984 vs 2826 +/- 2703 ml, respectively; p = 0.05) but were associated with a similar length of stay compared with the other treatment cohorts (11.5 +/- 9.3 [anterior] vs 11.3 +/- 8.6 [posterior] vs 14.3 +/- 6.7 [combined] days; p = 0.35). The posterior approach was associated with a higher incidence of wound infection compared with the anterior approach cohort (26.7 vs 4.5%, respectively; p = 0.03), and patients in the posterior approach group experienced the highest rates of deep vein thrombosis (15.6% [posterior] vs 0% [other 2 groups]; p = 0.02). However, the posterior approach demonstrated the lowest incidence of pneumothorax (4.4%; p < 0.0001) compared with the other 2 cohorts. Duration of chest tube use was greater in the combined patient group compared with the anterior approach cohort (8.8 +/- 6.2 vs 4.7 +/- 2.3 days, respectively; p = 0.01), and the combined group also experienced the highest rates of radiographic pleural effusion (83.3%; p = 0.01). Postoperatively, all groups improved neurologically, although functional outcome in patients undergoing the combined approach improved the most compared with the other 2 groups on both the Nurick (p = 0.04) and American Spinal Injury Association Impairment scales (p = 0.03). Decisions regarding the approach to thoracic vertebrectomy may be complex. This study found that although anterior approaches to the thoracic vertebrae have been historically associated with significant pulmonary complications, in our experience these rates are nevertheless quite comparable to that encountered via a posterior or combined approach. In fact, the posterior approach was found to be associated with a higher risk for some perioperative complications such as wound infection and deep vein thromboses. Finally, the combined anteriorposterior approach may provide greater ambulatory and neurological improvements in properly selected patients.