Spine 2026;51:1197–1208
Choosing between reconstruction and nonoperative care for adult spinal deformity demands a realistic estimate of benefit rather than radiographic enthusiasm alone. This multicenter analysis uses propensity matching to compare patients with similar measured baseline characteristics. At 2 years, surgery produced substantially greater improvement in disability and scoliosis-specific quality of life. Clinically meaningful benefit was achieved by a much larger proportion of surgical patients, although the cohort was relatively young and only mildly to moderately impaired at baseline. The results strengthen comparative-effectiveness evidence but do not replace individualized discussion of complications and residual confounding.
Objective
To compare 2-year patient-reported outcomes after surgical and nonoperative management of adult spinal deformity while reducing measured treatment-selection bias through propensity-score matching.
Methods
The retrospective comparison used prospectively collected data from an adult spinal-deformity registry at six centers. Of 580 eligible patients, 338 underwent surgery and 242 received nonoperative care. Propensity scores incorporated demographic variables, radiographic alignment measures and baseline Oswestry Disability Index (ODI) and Scoliosis Research Society–22 (SRS-22) scores. One-to-one matching produced 160 balanced pairs.
Main results
At 2 years, mean ODI improvement was −19.4 ± 14.2 points after surgery and −4.2 ± 12.3 with nonoperative care (p<0.001). A clinically meaningful ODI improvement of at least 15 points was achieved by 72% of surgical patients versus 29% of nonoperative patients.
The mean 2-year SRS-22 total score was 3.95 ± 0.67 after surgery and 3.46 ± 0.75 after nonoperative care (p<0.0001). The minimum clinically important difference for the SRS-22 total score was reached by 81.3% and 36.9%, respectively.
Interpretation
Among matched patients, reconstruction was associated with a large and clinically visible improvement beyond that observed with nonoperative management. Propensity matching makes the groups more comparable but cannot recreate randomization. The magnitude of benefit should therefore be presented alongside the patient’s operative risk, deformity phenotype, frailty, goals and probability of complications or revision.
Limitations
Treatment was not randomized, and unmeasured factors such as surgeon judgment, symptoms, patient preference and socioeconomic context may still influence both selection and outcome. The matched cohort had a mean age of about 45 years and mild-to-moderate baseline impairment, limiting extrapolation to older or frailer patients. Two-year follow-up does not capture late mechanical failure, revision burden or durability.
Clinical takeaway
For an appropriately selected adult with symptomatic deformity, surgery offers a substantially greater chance of meaningful 2-year improvement than continued nonoperative care. Use those probabilities in shared decision-making, but pair them with center-specific complication and revision estimates rather than presenting benefit in isolation.