Cranial sagittal vertical axis to the hip as the best sagittal alignment predictor of patient-reported outcomes at 2 years postoperatively in adult spinal deformity surgery.

OBJECTIVE

The objective was to discern whether the cranial sagittal vertical axis (CrSVA) can best predict the trajectory of patient-reported outcome measures (PROMs) at 2 years postoperatively.

METHODS

This was a retrospective cohort study of prospectively collected adult spinal deformity patient data. CrSVA relative to the sacrum, hip (CrSVA-H), knee, and ankle was measured as the horizontal distance to the vertical plumb line from the nasion-inion midpoint, with positive values indicating an anterior cranium. Standard sagittal alignment parameters were also collected. Outcome variables were PROMs as measured by Scoliosis Research Society-22r questionnaire (SRS-22r) total and subdomain scores and the Oswestry Disability Index. Pearson's correlation coefficients and univariate regressions were performed to investigate associations between predictors and PROMs. Two conceptual multivariable linear regression models for each 2-year outcome measure were built after adjusting for the impact of preoperative SRS-22r scores. Model 1 assessed pre- and postoperative alignment only relative to C2 and C7, while model 2 assessed alignment relative to C2 and C7 as well as the cranium.

RESULTS

There was a total of 363 patients with 2 years of radiographic and PROM follow-up (68.0% female, mean [standard error of the mean] age 60.8 [0.78] years, BMI 27.5 [0.29], and total number of instrumented levels 12.8 [0.22]). CrSVA measures were significantly associated with the 2-year SRS-22r total and subdomain scores. In univariate regression, revision surgery, number of prior surgeries, frailty, BMI, total number of osteotomies, and lower baseline total SRS-22r score as well as postoperative sagittal alignment were significantly associated with worse 2-year SRS-22r scores. In multivariable regression, after adjusting for baseline SRS-22r scores, greater preoperative C2 to sacrum sagittal vertical axis (SVA) and C7 SVA were found to be the only independent predictors of 2-year total SRS-22r score (β = -0.011 [p = 0.0026] and β = 0.009 [p = 0.0211], respectively) when alignment was considered only relative to C2. However, in the subsequent model, CrSVA-H replaced C7 SVA as the independent factor driving postoperative SRS-22r total scores (β = -0.006, p < 0.0001). That is, when the model included alignment relative to the cranium, C2, and C7, greater or more anterior CrSVA-H resulted in worse SRS-22r scores, while smaller or more posterior CrSVA-H resulted in better scores. Similar models for subdomains again found CrSVA-H to be the best predictor of function (β = -0.0095, p < 0.0001), pain (β = -0.0091, p < 0.0001), self-image (β = -0.0084, p = 0.0004), and mental health (β = -0.0059, p = 0.0026).

CONCLUSIONS

In multivariable regression, C7 SVA was supplanted by CrSVA-H alignment as a significant, independent predictor of 2-year SRS-22r scores in patients with adult spinal deformity and should be considered as one of the standard postoperative sagittal alignment target goals.