J.A. Herring replies

Abstract

To The Editor: The paper “Legg-Calve-Perthes Disease. Part II: Prospective Multicenter Study of the Effect of Treatment on Outcome” (2004;86:2121-34), by Herring et al., is a landmark in pediatric orthopaedics. Dr. Herring et al. and the Legg-Perthes Study Group are to be congratulated on accumulating so many data in a prospective long-term study. The authors have performed an expansive analysis. Sample-size constraints led the authors to combine groups and to conclude that operative treatment was better than nonoperative treatment. However, if we compare bracing and surgical treatment (containment) with range-of-motion exercises and no treatment (noncontainment), the difference is highly significant in favor of containment (chi-square analysis; p < 0.01). When operative treatment is compared with bracing, no significant difference is observed, although the sample sizes are roughly equivalent (129 in the bracing group and 119 in the operative treatment group). The authors appear to have been selective in their comparisons. On the basis of logistic regression and Wald chi-square tests, the most important factors were classification and age, not treatment. Another way to look at the effect of surgery is to examine the number needed to treat (NNT) in order to move a patient from one Stulberg class to another. If we compare bracing and surgery in percentage terms and accept the “best case” that surgery was superior, 7% of patients moved from class IV or V to class III and 9% moved from class III to class I or II. Thus, 16% of …