Correlation between change in graft height and change in segmental angle following central corpectomy for cervical spondylotic myelopathy

Abstract

Object: This study was undertaken to examine the correlation between change in graft height and change in angulation across grafted segments (segmental angle) in patients undergoing central corpectomy (CC) with autologous bone reconstruction for cervical spondylotic myelopathy (CSM). Methods: The authors performed a retrospective analysis of 70 cases in which patients with CSM underwent uninstrumented single- or multilevel CC and had evidence of osseous fusion of their grafts at follow-up. The segmental angles and heights of the grafted segments on preoperative, postoperative, and follow-up radiographs were compared. Results: The mean change in graft height (± standard deviation) was -7.3 ± 3.8 mm (mean duration of follow-up 19.7 ± 5.4 months, range 13-53 months). There was a mean kyphotic change in segmental angle of -7.3 ± 3.8° (p < 0.001). In patients who had a straight or kyphotic cervical spine (28 patients) or a straight or kyphotic segment (32 patients) preoperatively, there was a significant linear correlation between changes in graft height and changes in segmental angle (Pearson correlation, r = 0.40, p = 0.03; r = 0.40, p = 0.02, respectively). Such a correlation was not seen in the patients who had a lordotic cervical spine (42 patients) or a lordotic segment (38 patients) preoperatively (Pearson correlation, r = -0.04, p = 0.81; r = 0.08, p = 0.62, respectively). The change in segmental angle did not influence improvement in Nurick grade (p = 0.8). The degree of agreement between the 2 observers was almost perfect for measurement of graft height (postoperative intraclass correlation coefficient [ICC] = 0.94, follow-up ICC = 0.90) but was significantly lower for measurement of segmental angles (postoperative ICC = 0.71, follow-up ICC = 0.67). Conclusions: Among patients undergoing uninstrumented CC for CSM, there is a significant correlation between postoperative settling and kyphotic change across fused segments in those who had straight or kyphotic cervical spines or segments preoperatively but not in those who had lordotic cervical spines or segments preoperatively. A more vigorous surgical correction of the segmental kyphosis than achieved in this study might have caused the kyphotic segments to behave like the lordotic segments. Paraspinal muscles and ligaments may play a role in determining the segmental angle as graft settling in patients with lordotic spines or segments is not linearly correlated with angular change.