Proximal Junctional Kyphosis in Patients Following Surgical Treatment for Scheuermann’s Kyphosis: What Are the Risk Factors?

Purpose: To analyze the risk factors associated with proximal junctional kyphosis (PJK) in patients undergoing long instrumented spinal fusion for Scheuermann’s kyphosis with different types of fixation constructs.

Background: Previously reported risk factors for PJK include inappropriate end vertebrate selection, curve correction > 50%, or excessive junctional soft tissue dissection. Proximal junctional angle is defined as the cranial endplate of the upper instrumented vertebrate to the cranial endplate two vertebrate above. Abnormal PJK was defined by proximal junctional angle > 10 degrees and at least 10 degrees greater than the corresponding preoperative measurement.

Materials and Methods: Clinical and radiographic data on 45 patients treated with instrumented fusion for Scheuermann’s kyphosis were reviewed. All patients had complete radiographic data with minimum 5 year follow-up (mean 72 months). Ten patients (22%) with flexible curves (hyperextension <50 degrees) were treated with posterior only while others (78%) combined anterior and procedures. Ten patients (22%) had associated scoliosis> 20 degrees. The instrumentations used include Harrington compression rods (6), Luque sublaminar wires (6), segmental hook system (26), and hybrid constructs consisting of hooks proximally and pedicle screw anchors distally (6). Radiographic measurements include Cobb angle, C7 plumb line, and proximal junctional angle on preoperative, immediate post-operative, and final follow-up radiographs. Due to the difficulty in visualizing the upper thoracic vertebrae on some standing long cassette radiographs, a novel method with previously confirmed intra- and inter-observer reliability was used to select the proximal end vertebrae. This method relies on identification of the anterior and posterior cortex in the visualized portion of lower cervical and upper thoracic vertebral bodies, drawing a “best fit line” between these points, and using the perpendicular to the “best fit line” to estimate the proximal end vertebrae.

Results: Incidence of PJK as defined is seen in 12 patients (27%). The development of PJK is associated with failure to incorporate the proximal end vertebra (9 patients) and disruption of junctional ligamentum flavum (3 patients). The ligamentum flavum disruption was due to the use of supralaminar hooks in 2 patients and sublaminar wire in one patient. The development of PJK does not appear to be associated with magnitude of the curve or the amount of correction achieved (range 20-68%). The most common cause of inappropriate end vertebrae selection was poor visualization of upper thoracic vertebrae and this error can be minimized by using the “best fit line” technique. Four patients required additional surgery for instrumentation removal (2), repair of pseudoarthrosis (1) and repair of symptomatic PJK.

Conclusion: Minimum 5 year follow-up of patients who underwent surgical treatment for Scheuermann’s kyphosis revealed high incidence of PJK although the majority are asymptomatic. The incidence of PJK can be minimized by the appropriate selection of upper end vertebrae and avoiding disruption of junctional ligamentum flavum. Large amount of curve correction (>50%) does not correlate with the development of PJK.