Timing of epiphysiodesis to correct leg-length discrepancy: A comparison of prediction methods

Background: The purpose of this study was to compare the accuracy of different methods used to predict ultimate leg lengths and residual leg-length discrepancy in a group of patients treated with epiphysiodesis at our institution. Methods: Seventy-seven patients with adequate preoperative radiographs, no postoperative complications, and follow-up to skeletal maturity composed the study group. We compared the predicted lengths of both legs and residual leg-length discrepancy at maturity with actual outcomes using the White-Menelaus, Anderson-Green, Moseley, and multiplier methods. Results: Skeletal age varied >1 year from chronological age in 61 (26%) of 231 observations, including 19 patients (25%) whose average skeletal age from 3 determinations differed by >1 year from chronological age. The prediction accuracy of each method was improved by using skeletal, rather than chronological, age. Error in prediction of the length of the short leg varied from a mean (and standard deviation) of 1.8 ± 1.2 cm for the straight-line graph to 2.5 ± 2.0 cm for the multiplier method. Prediction error for the long leg (after epiphysiodesis) varied from a mean of 1.2 ± 1.1 cm for the straight-line graph to 1.7 ± 1.5 cm for the multiplier method. Leg-length-discrepancy prediction error ranged from a mean of 0.7 ± 0.6 cm for the White-Menelaus method incorporating a growth inhibition factor to 1.1 ± 0.9 cm for the multiplier method. The multiplier method was the least accurate of all. All differences were significant. Conclusions: Determination of skeletal age is clinically relevant in patients being considered for epiphysiodesis to manage leg-length inequality. The multiplier method was the least accurate of the prediction methods in this patient population, which may have implications in calculating the appropriate timing of epiphysiodesis.