Aims The outcome of treatment for Perthes’ disease has historically been assessed using the Stulberg classification, which is categorical and subjective, as reported in a large prospective multicentre study by the original Perthes’ Study Group in 2004. The aim of this study was to reanalyze these outcomes and prognostic determinants using a quantitative outcome measure, the sphericity deviation score. Methods We reanalyzed the demographic and radiological data, which included 319 patients followed up for a mean of 8.7 years (2.2 to 15.3), until skeletal maturity. The sphericity deviation scores of all 319 hips were measured by two independent observers. The forms of treatment included no treatment or range of motion (ROM) exercises, bracing, proximal femoral varus osteotomy, and Salter innominate osteotomy. We compared differences in the outcomes of treatment after stratification by significant patient characteristics determined by a multivariate linear regression. The patients were also analyzed using a regression tree model. Results Multivariate analysis identified age at the time of onset of the condition, sex, and the form of treatment as significant prognostic factors, with bone age at onset as the most significant factor, followed by chronological age and the form of treatment. In patients with a bone age of ≤ six years at the time of onset, the form of treatment did not affect the outcome. In patients with a bone age of > six years, femoral osteotomy or Salter osteotomy produced significantly better sphericity scores than no treatment or ROM exercises. The regression tree analysis corroborated these findings and found that those who had a femoral osteotomy had significantly better outcomes than those who did not, in patients with a chronological age of > 7.6 years. Conclusion Analysis based on scores of the sphericity of the femoral head showed that age at the onset of Perthes’ disease, sex, and the form of treatment were significant prognostic factors of outcome. Multivariate analysis with dichotomization of the bone age (≤ 6 yrs or > 6 yrs) revealed significantly better sphericity scores after proximal femoral osteotomy or Salter osteotomy than after no treatment or ROM exercises in patients with a bone age of > six years at the time of onset. The regression tree analysis without prior dichotomization showed that the patients who were treated with a femoral osteotomy had significantly better outcomes than those who did not, in patients with a chronological age of > 7.6 years. Cite this article: Bone Joint J 2026;108-B(2):251–258.
Kang et al. (2026) studied this question.