Sixty-two children with unilateral Perthes disease who underwent trochanteric epiphyseodesis combined with varus osteotomy of the femur during the active stage of the disease, (mean age at surgery: 8.4 years) and twenty controls were followed up untill skeletal maturity.

The following measurements were taken on radiographs taken at skeletal maturity: the articulo-trochanteric distance (ATD), the center-trochanteric distance (CTD), the length of the abductor lever arm, the neck-shaft angle, the radius of the femoral head and the Reimer's migration index of normal and affected hips. The shape of the femoral head was assessed according to the criteria of Mose. The range of hip motion, the strength of hip abduction and limb lengths were measured and the Trendelenburg sign was elicited.

The mean values of ATD and CTD were greater and the frequency of a positive Trendelenburg sign was less in children who had undergone trochanteric epiphyseodesis in 60% of operated children. The procedure was not effective in 30% and there was over-correction in in 10% of children. Logistic regression analysis showed that the size of the femoral head and the age at surgery were variables that significantly influenced the effectiveness of trochanteric growth arrest.

At skeletal maturity, the mean shortening of the affected limb in operated children was 0.44 cm (SD 0.68 cm), while that of non-operated children was 0.86 cm (SD 0.78 cm) (p: 0.023). The range of motion of the hip was excellent and there were no significant differences in the range of motion between children with optimal correction, under-correction and over-correction.

A probability curve plotted on the basis of the of a logistic regression model suggests that effective trochanteric arrest may be achieved in a high proportion of children operated at, or before, 8.5 years of age, and in half the children operated between the age of 8.5 years and 10 years.