The Association Between Scoliosis Research Society-22 Scores and Scoliosis Severity Changes at a Clinically Relevant Threshold

• Published in: Spine (Philadelphia, Pa. 1976) Pa. 1976), 2010-02, Vol.35 (3), p.315-322
• Main Authors: PARENT, Eric C, WONG, Daniel, HILL, Doug, MAHOOD, James, MOREAU, Marc, RASO, V. James, LOU, Edmond
• Format: Article
• Language: English
• Subjects: Adolescent; Alberta; Biological and medical sciences; Cerebrospinal fluid. Meninges. Spinal cord; Cross-Sectional Studies; Female; Humans; Medical sciences; Nervous system (semeiology, syndromes); Nervous system as a whole; Neurology; Quality of Life - psychology; Scoliosis - classification; Scoliosis - diagnosis; Scoliosis - psychology; Severity of Illness Index; Societies, Medical - standards; Surveys and Questionnaires - standards
• ISSN: 0362-2436; 1528-1159
• DOI: 10.1097/brs.0b013e3181cabe75

Cross-sectional correlation study. To determine the threshold in spinal deformity severity measurements beyond which there is a progressive decline in health-related quality-of-life (HRQOL). The associations between HRQOL and scoliosis deformity measures are at best moderate when assessed using linear regressions. This may be because HRQOL is not affected until a severity threshold is reached. Identifying the thresholds in deformity beyond which HRQOL deteriorates could assist in treatment recommendations. The Scoliosis Research Society-22 (SRS-22) questionnaire was completed by 101 females with adolescent idiopathic scoliosis (age, 15.0 +/- 1.8; largest Cobb angle, 36.9 degrees +/- 14.6 degrees). Radiographs and surface topography were used to quantify the severity of the internal (largest Cobb angle) and external deformity (cosmetic score, decompensation, trunk twist), respectively. Segmented linear regression models were estimated to determine the association between SRS-22 domains and spinal deformity measures. This analysis also identifies deformity thresholds beyond which HRQOL is more affected. The percentage of variance explained (R2) by linear and segmented models were compared (alpha = 0.05) to identify the best models. Cobb angle predicted significantly more variance in all SRS-22 domains except mental health using segmented models (R2: 0.09-0.30) than linear models (R2: 0.02-0.21). Segmented models with a single threshold estimated at a Cobb angle between 43 degrees and 48 degrees predicted between 3% and 11% more variance compared to corresponding linear model using the same variables. Surface topography parameters were not strongly associated with SRS-22 variables with linear and segmented models explaining less than 10% of the variance. Deterioration in SRS-22 scores is mildly associated with increases in the severity of the internal deformity. HRQOL is stable until the curve reaches a maximal Cobb angle threshold at approximately 45 degrees where HRQOL declines linearly with increasing internal deformity. The association between HRQOL and scoliosis severity is low, but is better explained by segmented rather than linear models.