Analysis of success versus failed realignment in patients with cervical deformity: in-construct angles provide novel targets for correction

BACKGROUND CONTEXT: Correction of cervical deformity (CD) has the potential to improve patient function significantly. However, previously described radiographic parameters cannot be measured intraoperatively. The present study investigates in-construct measurements which can be used as targets for CD correction to optimize radiographic realignment and improve outcomes. PURPOSE: Provide targets for CD correction through in-construct measurements (SA) to optimize radiographic realignment and improve outcomes. STUDY DESIGN/SETTING: Retrospective analysis of a prospective, multicenter CD database. PATIENT SAMPLE: Patients with adult cervical deformity with either cervical kyphosis more than 10°, cSVA of more than 4 cm, CBVA more than 25°, and a LIV at T1 or caudal were included. Patients were categorized into the failed outcome group if they had a cSVA of more than 4 cm postoperatively. OUTCOME MEASURES: Demographic factors; surgical factors; radiographic measures including novel in-construct measures describing sagittal alignment within the fusion with a line from the C2 centroid to the the centroid of a thoracic vertebra and a line along the posterior cortex of that vertebra (C2-T1 SA, C2-T4 SA, C2-T10 SA, C2-LIV SA) and established parameters (cSVA, T1S, C2S, TS-CL, T4-T12 Cobb, TPA, SVA, PI-LL, PT, DJK angle. 1-year HRQLs: NDI, mJOA, NRS neck, EQ5D. METHODS: The in-construct measurements were used depending on the patients’ LIV. All patients had a C2-T1 SA. C2-T4 SA were compared between groups with LIV below T4, and C2-T10 SA between groups with LIV below T10. Change in C2-LIV SA described the sagittal correction within the fusion for each patient. Analyses between failed and success realigned groups were performed using t-test, chi square analysis, and multivariate regression. Linear regression analysis was used to determine the C2-T1, C2-T4, C2-10 SA measures that correspond to a cSVA=4 cm and DJK =10°. HRQL analysis was done in subset of patients with 1-year follow-up. RESULTS: A total of 143 patients with CD (mean age 63 yr, 60% female) were included with 73 having failed radiographic outcomes by high cSVA (51% Failed). There were no differences in demographics. Failed was associated with greater levels fused (10.44 vs 8.46; p=0.005) and EBL (1108.41 vs 651.69; p=0.002). Failure to correct cSVA was associated with worse baseline deformity including cSVA (64.37 mm vs 40.36), T1S (42.97° vs 32.25°) C2S (51.38° vs 34.40°) TS-CL (51.16° vs 35.78°) (all p<0.001), greater kyphotic change in DJKA (10.45° vs 6.33°; p= 0.026), and greater postoperative C2-T1 SA (-5.23° vs -12.84°; p<0.001). Multivariate regression for variables with p<0.05 revealed that the postoperative C2-T1 in-construct angle independently predicted failed realignment outcome (OR= 1.25, CI 1.11-1.41; p<0.001). Using linear regression, a cSVA measurement of 4.0cm corresponded to a C2-T1-SA of -9.55°, C2-T4 of -0.37°, C2-T10-SA of 14.67°(all r>0.38, p<0.05). Linear regression revealed that postoperative C2-T10 SA was able to predict change in DJKA, where a change of 10° yielded a C2-T10 SA of 20.67°(r>0.57, p=0.02). While there was no difference in postop HRQL between the groups, improvement in C2-LIV SA was associated with improvement in neck pain (NRS neck) at 1 year (r>0.42, p=0.036). CONCLUSIONS: Failure to restore cSVA patients was independently associated with undercorrection, as evidenced by significantly larger postoperative in-construct angles. In-construct measures can be used as alignment targets to optimize radiographic outcomes and prevent DJK thereby improving patient reported outcomes. FDA Device/Drug Status: This abstract does not discuss or include any applicable devices or drugs.

Duke Scholars

Author Christopher Ignatius Shaffrey Orthopaedic Surgery

Author Peter Passias Orthopaedic Surgery