Incidence and Predictors of Hardware Failure After Instrumentation for Spine Metastasis: A Single-Institutional Series

Abstract

We report a retrospective analysis of hardware failure in patients requiring instrumentation for spinal metastasis. In a retrospective study at a single institution, we identified 58 patients who underwent spinal instrumentation for metastasis from 2012 to 2018. Hardware failure was defined as screw pullout/loosening, cage migration, progressive kyphosis, or an otherwise-noticeable instrumentation deficit detectable on imaging. Risk factors for hardware failure with a P < 0.05 in in univariate were included in multivariate logistic regression models controlled for age, sex, and previously identified risk factors for hardware failure. In total, 58 patients required instrumentation for metastatic spine disease. Median age was 60.2 years (interquartile range 49.0-66.3), 38 patients (65.5%) were male, and median follow-up was 8.1 months (interquartile range 3.1-20.7). Eight patients (13.8%) developed signs of hardware failure during follow-up, of whom 2 patients (3.4%) underwent operative revision. In univariate analysis, Eastern Cooperative Oncology Group performance status >2 (P= 0.049) and multiple myeloma lesions (P= 0.010) were significant predictors of failure. Both factors maintained significance in a multivariate logistic regression model controlled for age, sex, history of spine radiation, and number of fused levels with P= 0.047; odds ratio 12.7 (95% confidence interval 1.03-156.4) for Eastern Cooperative Oncology Group performance status over 2 and P= 0.012; odds ratio 31.5 (95% confidence interval 2.2-460.0) for multiple myeloma lesions. The rate of hardware failure in this cohort was 13.8%, although operative revision rate was 3.4%. Spinal instrumentation in patients with poor preoperative functional status or multiple myeloma may be more likely to develop instrumentation failure.