Natural history of scoliosis in nonambulatory spastic tetraplegic cerebral palsy

Objective: To analyze the development and progression of scoliosis in children and adolescents with nonambulatory spastic tetraplegic cerebral palsy.

Design: Retrospective longitudinal review.

Setting: Pediatric nursing home.

Participants: A total of 110 children and adolescents <18 years of age, with scoliosis.

Interventions: N/A.

Main outcome measures: Cobb angle, age, weight, height, history of hip dislocation, tracheostomy.

Results: The Cobb angle increased with age, weight, and height in a nonlinear fashion. A square root transformation of the Cobb angle was chosen to model the nonlinear relationship between the Cobb angle and predictors. Age, height, and weight were significant univariate predictors of the square root of the Cobb angle (slopes = 0.377, 0.067, and 0.06, respectively). In the multivariate mixed model, age remained a significant predictor of the Cobb angle (slope = 0.456), but height and weight did not. If the Cobb angle was > 40° by age 12 years, scoliosis was more likely to progress than if the Cobb angle was ≤ 40°. The effect of age was stronger for those with history of tracheostomy (age slope = 0.631 vs 0.281) than those without. The relationship of age and Cobb angle did not differ significantly between hip dislocated and non-hip-dislocated groups.

Conclusions: Age was found to be the most significant predictor of Cobb angle, and the effect of age was greater in the tracheostomy group than in the nontracheostomy group. After adjustment for age, the weight and height were not significant predictors of Cobb angle. Cobb angles of > 40° by the age of 12 years were associated with greater increases in Cobb angle with age.