• OBJECTIVE
    • To evaluate predictors of fusion across the disc space and posterolaterally following transforaminal lumbar interbody fusion (TLIF).
  • METHODS
    • Patients who underwent 1- or 2-level TLIF for degenerative pathologies were examined for fusion across the disc space, posterolateral fusion, and circumferential radiographic fusion. Multivariable logistic regression was used to identify independent predictors of the primary and secondary outcomes.
  • RESULTS
    • 278 unique patients (328 unique levels) were included (median age 66.2 yr; 58.3 % female) 25.9 % had complete circumferential fusion, 69.6 % had fusion across the disc space and 71 % had unilateral posterolateral fusion. Fusion across the disc space was less likely with a minimally invasive [versus open] approach (OR 0.39; 95 % CI [0.16, 0.95]; p = 0.04) or bullet [versus banana] cage (OR 0.18; [0.09, 0.38]; p < 0.001). Posterolateral fusion was also less likely with an MIS approach (OR 0.10; [0.03, 0.30]; p < 0.001) or bullet cage (OR 0.14; [0.06, 0.30]; p < 0.001) but was significantly more likely with BMP use (OR 2.97; [1.14, 7.74]; p = 0.026). Circumferential fusion was predicted by BMP use (OR 3.29; [1.48; 7.35]; p = 0.004), use of a bullet cage (OR 0.19; [0.04, 0.87] p = 0.033), and use of a longer interbody device (OR 1.15; [1.01, 1.30]; p = 0.032). Cage material was not predictive in any of the multivariable analyses.
  • CONCLUSION
    • The minority of patients experience circumferential fusion following TLIF. Use of a minimally invasive approach or a bullet-type cage lowers the odds of successful fusion both posterolaterally and across the disc space. BMP increases fusion odds due to improved odds of posterolateral fusion.