Incidence and risk factors for the progression of proximal junctional kyphosis in degenerative lumbar scoliosis following long instrumented posterior spinal fusion

The aim of this study was to identify the prevalence of proximal junctional kyphosis (PJK) in degenerative lumbar scoliosis (DLS) following long instrumented posterior spinal fusion, and to search for predictable risk factors for the progression of junctional kyphosis.In total 98 DLS patients with a minimum 2-year follow-up were reviewed prospectively. According to the occurrence of PJK at the last follow-up, patients were divided into 2 groups: PJK group and non-PJK group. To investigate risk values for the progression of PJK, 3 categorized factors were analyzed statistically: patient characteristics-preoperative data of age, sex, body mass index (BMI), bone mineral density (BMD) were investigated; surgical variables-the most proximal and distal levels of the instrumentation, the number of instrumented levels; pre- and postoperative radiographic parameters include the scoliotic angle, sagittal vertical axis, thoracic kyphosis, thoracolumbar junctional angle, lumbar lordosis, pelvic incidence, pelvic tilt, and sacral slope.PJK was developed in 17 of 98 patients (17.3%) until to the final follow-up and were enrolled as the PJK group, and 81 patients without PJK at final follow-up were enrolled as the non-PJK group. There was no statistically significant difference between the 2 groups in age at operation (P = 0.900). The patient's sex was excluded in statistical analysis because of the predominance of female patients. There were statistically significant difference between the 2 groups in BMI ([25.5 ± 1.7] kg/m in the PJK group and [23.6 ± 1.9] kg/m in the non-PJK group, P < 0.001) and BMD ([-1.4 ± 0.8] g/cm in the PJK group and [-0.7 ± 0.3] g/cm in the non-PJK group, P < 0.001). No specific surgery-related variables were found to be associated with an increased risk of developing PJK, except when the most proximal instrumented vertebrae stopped at thoracolumbar junction (T11-L1). The upper instrumentation vertebrae (UIV) at thoracolumbar junction was more common in the PJK group than that in the non-PJK group (P = 0.007). No preoperative and early postoperative variable did reveal a statistically significant difference between the 2 groups. When included in a multivariate logistic regression model, BMI>25 kg/m, osteoporosis, and UIV at thoracolumbar junction were independently associated with PJK.In conclusion, osteoporosis, obesity, and UIV at thoracolumbar junction are risk factors for the development and progression of PJK in DLS patients following long instrumented posterior spinal fusion. Antiosteoporosis treatment extends the fusion level above the thoracolumbar region and controlling body weight before and after surgery could provide opportunities to reduce the rate of PJK and to improve therapeutic outcomes.