Is the T9, T11, or L1 the more reliable proximal level after adult lumbar or lumbosacral instrumented fusion to L5 or S1?

Study Design. A retrospective comparison study. Objective. To compare the postoperative proximal junctional change and revision prevalence as influenced by 3 different proximal levels after adult lumbar deformity instrumented fusion from the distal thoracic/upper lumbar spine (T9-L2) to L5 or S1. Summary of Background Data. Few comparative studies on postoperative sagittal plane change and revision prevalence as influenced by 3 different proximal levels after adult lumbar deformity instrumented fusion from the distal thoracic/upper lumbar spine ( T9-L2) to L5 or S1 have been published. Many surgeons have hypothesized that stopping proximally in the upper lumbar spine (L1 or L2) or the thoracolumbar junction (T11 or T12) would lead to a high percentage of rapid proximal degeneration, kyphosis, and decompensation because of the concentration of stress on these relatively mobile segments. Therein, many surgeons have felt it is unsafe to stop at these regions of the spine and it is better to always stop proximally at T9 or T10. Methods. A clinical and radiographic assessment in addition to revision prevalence of 125 adult lumbar deformity patients ( average age 57.1 year) who underwent long ( average 7.1 vertebrae) segmental posterior spinal instrumented fusion from the distal thoracic/upper lumbar spine ( T9-L2) to L5 or S1 with a minimum 2-year follow-up (2-19.8 year follow-up) were compared as influenced by T9-T10 (group1, n = 37), T11-T12 ( group 2, n = 49), and L1-L2 ( group 3, n = 39) proximal fusion levels. The revision prevalence and sagittal Cobb angle change at the proximal junction after surgery were compared. Results. Three groups demonstrated nonsignificant differences in the prevalence of proximal junctional kyphosis ( group 1 51% vs. group 2 55% vs. group 3 36%, P = 0.20) and revision ( group 1 24% vs. group 2 24% vs. group 3 26%, P = 0.99) at the ultimate follow-up. Subsequent proximal junctional angle and sagittal vertical axis changes between the ultimate follow-up and preoperative ( P = 0.10 and 0.46 respectively) were not significantly different. The SRS total and all subscale outcomes scores among the 3 groups did not demonstrate significant differences ( P > 0.50). Conclusion. Three different proximal fusion levels did not demonstrate significant radiographic and clinical outcomes or revision prevalence after surgery. Therefore the more distal proximal fusion level at a neutral and stable vertebra may be satisfactory.