Can pelvic tilt be predicated by the sacrofemoral-pubic angel in patients with thoracolumbar kyphosis secondary to ankylosing spondylitis?

STUDY DESIGN: A retrospective radiographical study.

OBJECTIVE: To construct a predictive model for pelvic tilt (PT) based on the sacrofemoral-pubic (SFP) angle in patients with thoracolumbar kyphosis secondary to ankylosing spondylitis (or AS).

SUMMARY OF BACKGROUND DATA: PT is a key pelvic parameter in the regulation of spine sagittal alignment that can be used to plan the appropriate osteotomy angle in patients with AS with thoracolumbar kyphosis. However, it could be difficult to measure PT in patients with femoral heads poorly visualized on lateral radiographs. Previous studies showed that the SFP angle could be used to evaluate PT in adult patients with scoliosis. However, this method has not been validated in patients with AS.

METHODS: A total of 115 patients with AS with thoracolumbar kyphosis were included. Full-length anteroposterior and lateral spine radiographs were all available, with spinal and pelvic anatomical landmarks clearly identified. PT, SFP angle, and global kyphosis were measured. The patients were randomly divided into group A (n=65) and group B (n=50). In group A, the predictive model for PT was constructed by the results of the linear regression analysis. In group B, the predictive ability and accuracy of the predictive model were investigated.

RESULTS: In group A, the Pearson correlation analysis revealed a strong correlation between the SFP angle and PT (r=0.852; P<0.001). The predictive model for PT was constructed as PT=72.3-0.82×(SFP angle). In group B, PT was predicted by the model with a mean error of 4.6° (SD=4.5°) with a predictive value of 78%.

CONCLUSION: PT can be accurately predicted by the SFP angle using the current model: PT=72.3-0.82×(SFP angle), when the femur heads are poorly visualized on lateral radiographs in patients with AS with thoracolumbar kyphosis.

LEVEL OF EVIDENCE: 4.