Introduction: There have been consistently opposing reports regarding selection of uppermost instrumented vertebra (UIV) at thoracolumbar junction in long spinal fusion. However, some recent literatures claimed thoracolumbar junction could be considered selectively as UIV for long spinal fusion without certain potential risk factors. There have been a paucity of study investigating the risk factors associated with proximal junctional kyphosis (PJK) following long instrumented spinal fusion from L1 to the sacrum. Therefore, we aimed to investigate the incidence and risk factors of PJK following spinal instrumented fusion from L1 to sacrum in patients with mild to moderate sagittal imbalance.

Methods: This retrospective study recruited consecutive patients undergone instrumented fusion from L1 to the sacrum for degenerative lumbar disease between June 2006 and November 2019 in single institution. Inclusion criteria were as follows: 1) etiology of spinal stenosis, spondylolisthesis, or adjacent segment disease following previous spinal fusion; 2) completion of a long-segment spinal instrumented fusion surgery from L1 vertebra to the sacrum; 3) with a minimum follow-up period of 2 years. Exclusion criteria were as follows: 1) C7 sagittal vertical axis (SVA) more than 15cm; 2) history of pedicle subtraction osteotomy or other equivalent procedures; 3) early postoperative (within 1 year) complications requiring revision for index surgery. The patients’ preoperative clinical data, muscle status at T12-L1 on magnetic resonance images, and sagittal spinopelvic parameters on radiographs at preoperative, immediate postoperative, regular postoperative visits were collected and analyzed. PJK was defined if postoperative proximal junctional angle (made by T11 upper endplate and L1 lower endplate) was >20 degrees or increased more than 10 degrees compared to the baseline. Univariate analysis was used to compare clinical and radiographic data between PJK and non-PJK patients. Logistic regression analysis was used to investigate the independent risk factors for PJK.

Results: A total of 41 patients were included as study cohort. The mean age at surgery was 67.3 years and mean follow-up period was 37.3 months. Seven were male and 34 were female. PJK developed in 17 (41.5%) out of 41; of these patients, 15 (88.2%) developed PJK within postoperative 1 year and 5 (29.4%) of them were diagnosed with proximal junctional failure. On univariate analysis between PJK and non-PJK patients, PJK group showed more frequent osteoporosis, lower body mass index, smaller cross-sectional area (CSA) and more fat infiltration (FI) in erector spinae muscle at T12-L1, larger preoperative TLK and PT with statistical significance (p<0.05). On logistic regression analysis, severe (>50%) FI in erector spinae muscle (OR=43.60, CI 4.10-463.06, R²_{N =}0.730, p=0.002) and osteoporosis (OR=20.49, CI 1.58-264.99, R²_{N =}0.730, p=0.021) were identified as significant risk factors for PJK.

Discussion: In a review of minimum 2 years, most (88.2%) of PJK developed within postoperative 1 year. Preexisting severe (>50%) fat infiltration in erector spinae muscle and osteoporosis were significant independent risk factors associated with PJK following instrumented fusion from L1 to the sacrum.