Risk factors for cage migration and cage retropulsion following transforaminal lumbar interbody fusion

Abstract

Transforaminal lumbar interbody fusion (TLIF) is a widely accepted surgical procedure, but cage migration (CM) and cage retropulsion (CR) are associated with poor outcomes. This study seeks to identify risk factors associated with these serious events. A prospective observational longitudinal study. Over a 5-year period, 881 lumbar levels in 784 patients were treated using TLIF at three spinal surgery centers. We evaluated the odds ratio of the risk factors for CM with and without subsidence and CR in multivariate analysis. Our study classified CM into two subgroups: CM without subsidence and CM with subsidence. Cases of spinal canal and/or foramen intrusion of the cage was defined separately as CR. Patient records, operative notes, and radiographs were analyzed for factors potentially related to CM with subsidence, CM without subsidence, and CR. Of 881 lumbar levels treated with TLIFs, CM without subsidence was observed in 20 (2.3%) and CM with subsidence was observed in 36 (4.1%) patients. Among the CM cases, CR was observed in 17 (17/56, 30.4%). The risk factors of CM without subsidence were osteoporosis (OR 8.73, p < .001) and use of a unilateral single cage (OR 3.57, p < .001). Osteoporosis (OR 5.77, p < .001) and endplate injury (OR 26.87, p < .001) were found to be significant risk factors for CM with subsidence. Risk factors of CR were osteoporosis (OR 7.86, p < .001), pear-shaped disc (OR 8.28, p = .001), endplate injury (OR 18.70, p < .001), unilateral single cage use (OR 4.40, p = .03), and posterior cage position (OR 6.45, p = .04). A difference in overall fusion rates was identified, with a rate of 97.1% (801 of 825) for no CM, 55.0% (11 of 20) for CM without subsidence, 41.7% (15 of 36) for CM with subsidence, and 17.6% (3 of 17) for CR at 1.5 years postoperatively. Our results suggest that osteoporosis is a significant risk factor for both CM and CR. In addition, a pear-shaped disc, posterior positioning of the cage, the presence of endplate injury and the use of a single cage were correlated with the CM with and without subsidence and CR.