Patient-related and radiographic predictors of inferior health-related quality-of-life measures in adult patients with nonoperative spinal deformity

OBJECTIVE Patients with nonoperative (N-Op) adult spinal deformity (ASD) have inferior long-term spinopelvic alignment and clinical outcomes. Predictors of lower quality-of-life measures in N-Op populations have yet to be sufficiently investigated. The aim of this study was to identify patient-related factors and radiographic parameters associated with inferior health-related quality-of-life (HRQOL) scores in N-Op ASD patients. METHODS N-Op ASD patients with complete radiographic and outcome data at baseline and 2 years were included. N-Op patients and operative (Op) patients were propensity score matched for baseline disability and deformity. Patientrelated factors and radiographic alignment parameters (pelvic tilt [PT], sagittal vertical axis [SVA], pelvic incidence [PI]- lumbar lordosis [LL] mismatch, mismatch between cervical lordosis and T1 segment slope [TS-CL], cervical-thoracic pelvic angle [PA], and others) at baseline and 2 years were analyzed as predictors for moderate to severe 2-year Oswestry Disability Index (ODI > 20) and failing to meet the minimal clinically importance difference (MCID) for 2-year Scoliosis Research Society Outcomes Questionnaire (SRS) scores (< 0.4 increase from baseline). Conditional inference decision trees identified predictors of each HRQOL measure and established cutoffs at which factors have a global effect. Random forest analysis (RFA) generated 5000 conditional inference trees to compute a variable importance table for top predictors of inferior HRQOL. Statistical significance was set at p < 0.05. RESULTS Six hundred sixty-two patients with ASD (331 Op patients and 331 N-Op patients) with complete radiographic and HRQOL data at their 2-year follow-up were included. There were no differences in demographics, ODI, and Schwab deformity modifiers between groups at baseline (all p > 0.05). N-Op patients had higher 2-year ODI scores (27.9 vs 20.3, p < 0.001), higher rates of moderate to severe disability (29.3% vs 22.4%, p = 0.05), lower SRS total scores (3.47 vs 3.91, p < 0.001), and higher rates of failure to reach SRS MCID (35.3% vs 15.7%, p < 0.001) than Op patients at 2 years. RFA ranked the top overall predictors for moderate to severe ODI at 2 years for N-Op patients as follows: 1) frailty index > 2.8, 2) BMI > 35 kg/m2, T4PA > 28°, and 4) Charlson Comorbidity Index > 1. Top radiographic predictors were T4PA > 28° and C2-S1 SVA > 93 mm. RFA also ranked the top overall predictors for failure to reach 2-year SRS MCID for N-Op patients, as follows: 1) T12-S1 lordosis > 53°, 2) cervical SVA (cSVA) > 28 mm, 3) C2-S1 angle > 14.5°, 4) TS-CL > 12°, and 5) PT > 23°. The top radiographic predictors were T12-S1 Cobb angle, cSVA, C2-S1 angle, and TS-CL. CONCLUSIONS When controlling for baseline deformity in N-Op versus Op patients, subsequent deterioration in frailty, BMI, and radiographic progression over a 2-year follow-up were found to drive suboptimal patient-reported outcome measures in N-Op cohorts as measured by validated ODI and SRS clinical instruments.