Failure of nonoperative care in adult symptomatic lumbar scoliosis: incidence, timing, and risk factors for conversion from nonoperative to operative treatment

OBJECTIVE The Adult Symptomatic Lumbar Scoliosis (ASLS) study is a prospective multicenter trial with randomized and observational cohorts comparing operative and nonoperative treatment for ASLS. The objective of the present study was to perform a post hoc analysis of the ASLS trial to examine factors related to failure of nonoperative treatment in ASLS. METHODS Patients from the ASLS trial who initially received at least 6 months of nonoperative treatment were followed for up to 8 years after trial enrollment. Baseline patient-reported outcome measures (Scoliosis Research Society-22 [SRS-22] questionnaire and Oswestry Disability Index), radiographic data, and other clinical characteristics were compared between patients who did and did not convert to operative treatment during follow-up. The incidence of operative treatment was calculated and independent predictors of operative treatment were identified using multivariate regression. RESULTS Of 135 nonoperative patients, 42 (31%) crossed over to operative treatment after 6 months and 93 (69%) received only nonoperative treatment. In the observational cohort, 23 (22%) of 106 nonoperative patients crossed over to surgery. In the randomized cohort, 19 (66%) of 29 patients randomized to nonoperative treatment crossed over to surgery. The most impactful factors associated with crossover from nonoperative to operative treatment were enrollment in the randomized cohort and baseline SRS-22 subscore < 3.0 at the 2-year follow-up, closer to 3.4 at 8 years. In addition, baseline lumbar lordosis (LL) < 50° was associated with crossover to operative treatment. Each 1-point decrease in baseline SRS-22 subscore was associated with a 233% higher risk of conversion to surgery (hazard ratio [HR] 2.33, 95% confidence interval [CI] 1.14-4.76, p = 0.0212). Each 10° decrease in LL was associated with a 24% increased risk of conversion to operative treatment (HR 1.24, 95% CI 1.03-1.49, p = 0.0232). Enrollment in the randomized cohort was associated with a 337% higher probability of proceeding with operative treatment (HR 3.37, 95% CI 1.54-7.35, p = 0.0024). CONCLUSIONS Enrollment in the randomized cohort, a lower baseline SRS-22 subscore, and lower LL were associated with conversion from nonoperative treatment to surgery in patients (observational and randomized) who were initially managed nonoperatively in the ASLS trial.