Can pelvic tilt be restored by spinal osteotomy in ankylosing spondylitis patients with thoracolumbar kyphosis? A minimum follow-up of 2 years

BACKGROUND: Defining the postoperative pelvic tilt (PT) individually can help to reconstruct sagittal balance. However, the postoperative actual PT is hardly restored to theoretical value. Some cases with theoretical postoperative PT was overcorrected and still did not have normal horizontal visual field after surgery. The objective of this study is to describe the pelvic tilt change after spinal osteotomy in ankylosing spondylitis (AS) kyphotic deformity and evaluate the effect on clinical outcomes. METHODS: Twenty-three AS patients including 21 men and two women with thoracolumbar kyphosis, who underwent spinal osteotomy from 2013 to 2015 in our center, were retrospectively reviewed. A series of parameters including sacral slop (SS), pelvic incidence (PI), PT, and sagittal vertical axis (SVA) measured on preoperative and postoperative standing radiographs were analyzed. The theoretical postoperative PT (tPT) was calculated by the formula tPT = 0.37 × PI − 7. The radiographic measurements were compared before surgery, 2 weeks and at least 2 years postoperatively. Clinical outcomes were performed with the Oswestry disability index and Scoliosis Research Society-22 surveys. RESULTS: Mean age of the patients (2 women, 21 men) was 39.8 ± 9.1 years. Mean follow-up was 27.4 ± 3.8 months, at least 24 months. After spinal osteotomy, SS and SVA were corrected from 11.9° ± 11.2° and 18.0 ± 7.6 mm preoperatively to 25.8° ± 8.1° and 9.6 ± 6.3 mm postoperatively, respectively (p < 0.001). PT reduced from 37.6° ± 12.1° to 21.8° ± 9.8° postoperatively (p < 0.001). The tPT was different from postoperative actual PT significantly (p < 0.001). The clinical evaluations were not correlated with postoperative PT. CONCLUSION: The abnormal PT is corrected by spinal osteotomy but is hard to restore to theoretical normal value. PT is a helpful parameter in making surgery plan. But pursuing postoperative PT being totally equal to tPT is undesirable and even may cause for overcorrection.