Introduction: Management of adult spinal deformity (ASD) has increasingly favored operative intervention, however the incidence of complications and reoperations is high and patients may fail to reach optimal realignment parameters. While non-operative interventions have varying rates of efficacy, these patients presumably approximate the natural history of ASD. The objective of this study is to compare health-related quality of life (HRQL) metrics of patients with suboptimal surgical outcomes to those undergoing non-operative management (i.e. natural history).

Methods: ASD patients > 18 considered for operative intervention with full baseline (BL) and up to 1-year (1Y) follow-up HRQL and radiographic data were included. All patients ≥1 of the following BL radiographic parameters for ASD: pelvic tilt (PT) >25°, Cobb angle >20°, sagittal vertical axis (SVA) > 5cm, and thoracic kyphosis (TK) >60°. While all patients were offered operative intervention, for those in the non-operative group, the decision to forego operative management was ultimately made by the patient. Operative patients were selected for “suboptimal outcomes,” defined as any reoperation, major complication, or ≥2 severe (i.e. ++) SRS-Schwab modifiers at follow-up. Patients were separated into two groups: suboptimal operative (SOp) vs. natural history, i.e. non-operative (NH). SOp and NH patients were then propensity score matched (PSM) by BL age, deformity, Oswestry Disability Index (ODI), and Charlson Comorbidity Index (CCI). Means comparison tests analyzed differences between groups in outcomes of interest, including ODI and Scoliosis Research Society-22 (SRS-22) domains. 

Results: 370 patients met inclusion criteria, 284-SOp and 86-NH. After PSM, the resulting groups included 67 “SOp” and 67 “NH” patients. Cohort and group demographics, HRQLs, and radiographic parameters can be found in Table 1. After PSM the SOp and NH groups were not significantly different in demographics, HRQLs, or deformity (pelvic tilt (PT), SVA, pelvic-incidence-minus-lumbar-lordosis (PI-LL)). However, NH group had significantly higher baseline SRS-activity and SRS-satisfaction (p<.05 for both). Additionally, there were no significant differences in percent of patients with severe (i.e. ++) BL SRS-Schwab modifiers (p<.05).  At 1Y, SOp group had a mean improvement in ODI from BL (-9.16) vs. NH group (-0.51), p<.002 and significantly more SOp patients (39.7%) gained ≥1 MCID in ODI compared to NH patients (2.9%), p<.001. At 1Y, the SOp group had a significantly higher SRS-22 satisfaction domain score and demonstrated significantly greater improvement in SRS-Activity, SRS-Pain, SRS-Satisfaction, and SRS-Total (all p-values <.05). Finally, significantly more SOp patients gained ≥1 MCID in SRS-Activity (40.7%) and SRS-Pain (57.6%) compared to the NH patients, 17.1% and 31.4% respectively, p<.02 for both values.

Discussion: When compared to the natural history of non-operative ASD patients, operative patients with suboptimal outcomes still experience significantly greater improvements in HRQLs after surgical intervention. Such divergent outcomes seen at 1Y, highlight the stagnation and deterioration in outcomes associated with the natural history of ASD. While additional investigation should be conducted to explore these differences, spine surgeons and patients should consider the outcomes associated with the non-operative course when weighting risk and benefits of operative intervention for ASD.