Radiographic Predictors for Mechanical Failure After Adult Spinal Deformity Surgery: A Retrospective Cohort Study in 138 Patients.

STUDY

DESIGN

Retrospective cohort study at a single institution.

OBJECTIVE

We aimed at estimating the rate of revision procedures and identify radiographic predictors of mechanical failure after adult spinal deformity surgery.

SUMMARY OF BACKGROUND DATA: Mechanical failure rates after adult spinal deformity surgery range 12% to 37% in literature. Although the importance of spinal and spino-pelvic alignment is well documented for surgical outcome and ideal alignment has been proposed as sagittal vertical axis (SVA) < 5 cm, pelvic tilt < 20° and lumbar lordosis (LL) = pelvic incidence ± 9°, the role of radiographic sagittal spine parameters and alignment targets as predictors for mechanical failure remains uncertain.

METHODS

A consecutive cohort of adult spinal deformity patients who underwent corrective surgery with at least 5 levels of instrumentation between January 2008 and December 2012 at a single tertiary spine unit were followed for at least 2 years. Time to death or failure was recorded and cause-specific Cox regressions were applied to evaluate predictors for mechanical failure or death.

RESULTS

A total of 138 patients with median age of 61 years were included for analysis. Follow up ranged 2.1 to 6.8 years. In total 47% had revision and estimated failure rates were 16% at 1 year increasing to 56% at 5 years. A multivariate analysis adjusting for age at surgery showed increased hazard of failure from LL change > 30°, postoperative TK > 50°, and SS ≤30°. LL change was mostly because of 3-column osteotomy and ending the instrumentation at L5 or S1 increased the hazard of failure more than 6 fold compared with more cranial lumbar levels.

CONCLUSION

Mechanical failure rate was 47% after adult spinal deformity corrective surgery. LL change > 30°, postoperative TK > 50°, and postoperative SS ≤30° were independent radiographic predictors associated with increased hazard of failure.

LEVEL OF EVIDENCE: 4.