Legg-Calve-Perthes Disease: How Should It Be Treated?

Published on 11/03/2015 by admin

Last modified 11/03/2015

Print this page

This article have been viewed 1140 times

Chapter 40 Legg-Calve-Perthes Disease: How Should It Be Treated?

The necessity for evidenced-based investigation is nowhere better illustrated than in the case of the treatment of Legg–Calve–Perthes disease. In the United States, as well as around the world, an amazing variety of treatment methods are used for this condition. These vary from long-term relief from weight bearing, to months and months of abduction cast treatment, to combined femoral and pelvic surgical procedures, to ignoring the disease altogether. Even more striking is the degree with which orthopedists defend their chosen method.

The literature reflects this unique situation. Most studies have no control subjects and numbers too small for statistical validity. Different methods are used to determine severity and outcome, and until recently, few were subjected to validation by interobserver trials. It is abundantly clear that, in this instance, Level I and II studies are needed. Although many believe that the pressure for evidence-based knowledge is an academic exercise, I fully disagree, for it is the confused parent seeking treatment for a child who will benefit from this discipline. Time and again, orthopedists must dispel absurd treatment measures that other physicians have presented to the parent as an absolute necessity.

This chapter first reviews the investigations that reach Levels I, II, and III. It then briefly covers the information from other recent studies that have useful retrospective information.

Some years ago, I critically reviewed the literature up to 1994.¹ Although much has been written since that date, some principles of importance came out of that review that apply specifically to Legg–Perthes disease.

1. It is well known that the severity of the disorder is closely related to the age at onset. A study that does not adequately control for age at onset cannot reach useful conclusions.

2. The severity of femoral head involvement has been shown to be an even more important variable in outcome within all age groups up to 10 years of age at onset. Classification of outcome and severity of disease must be standardized by interobserver trials based on clear definitions. The Catterall classification and the associated risk factors were widely used in earlier works, but interobserver studies failed to show agreement among observers, even when well-versed in the details. The lateral pillar classification has had good interobserver reliability and has been shown to strongly correlate with outcome. The Stulberg classification of outcome has also been shown to have reliable reproducibility when detailed descriptors are used.

STUDIES THAT REACH LEVEL I OR II

In 1993, Fulford, Lunn, and Macnicol’s ² comparison of 2 groups of patients randomized by source of referral was published. The children treated at Royal Hospital for Sick Children were compared with those treated at the Princess Margaret Rose Orthopaedic hospital. The children at the first hospital were treated with an ischial-bearing patten-ended caliper. Those at the second hospital had a varus, derotational proximal femoral osteotomy. This study can be considered a controlled, prospective study. The authors identified no differences in outcome between the groups. Outcome in their study was found to correlate with femoral head shape at presentation and the age at onset. They note that of the heads with severe deformity at the initial arthrogram, only 4 of 25 hips improved to have a good result, and all were younger than 5 years at onset. Two of these patients were receiving brace therapy, and two had surgery. Compared with historical control subjects, the authors believe that both methods are effective. Because of the wide variation in reporting of historical control subjects, I do not believe that such comparisons are useful in this disorder.

Another difficulty is that the authors do not state what stage of disease was present when the patient first presented. Those who presented with severe head deformity must have been beyond the stage of increased density. Others presenting before head deformation had occurred may still go on to severe deformity, so this factor has questionable validity as a predictor.

Harrison and Bassett ³ performed a double-blind trial of pulsed electromagnetic field treatment in patients also being treated with a non–weight-bearing Birmingham orthosis. This study shows no discernible effect of the electromagnetic field treatment. The effectiveness of the orthosis was not elucidated by this study because the study did not include a control or comparative group.

Herring and colleagues ⁴ report on the results of a long-term, controlled study performed by the Legg Perthes Study Group. Thirty-nine pediatric orthopedists from North America and New Zealand agreed to a “best-effort” study in lieu of randomization. In this paradigm, each surgeon agreed to apply a single treatment method to each patient who met the study criteria. The presumption of this type of study is that so large a sample will likely override local selection factors and provide comparable treatment groups. The treatment groups were no treatment, range of motion treatment in which the patient did exercises once a day, Atlanta brace treatment, femoral varus osteotomy, and Salter innominate osteotomy. Within each treatment group, except the no treatment group, Petrie casts could be used up to two times to overcome hip stiffness.

Part one of the study provides detailed classifications that were tested in interobserver and intraobserver trials that showed good reliability ⁵. The classifications evaluated were the lateral pillar classification in the early stages of disease and the Stulberg classification at skeletal maturity. The outcomes were analyzed relative to age at onset and severity of disease. Patients were managed to skeletal maturity and assessed independently by an observer not involved in patient treatment.

This study came to the following conclusions: First, the lateral pillar classification was the strongest predictor of outcome. Those in lateral pillar groups A and B were much more likely to have a favorable outcome than those in lateral pillar group C. Those classified as B/C border had an outcome intermediate between groups B and C. The second strongest factor in outcome was age at onset. Those whose age at onset was older than 8 years had a less favorable outcome than those 8 years and younger. An identical effect was found when the outcome of those with onset at skeletal age of 6 years and younger was compared with those with onset beyond skeletal age of 6. (The mean delay of skeletal age in the boys was 2 years.)

Treatment methods were also significantly related to outcome. In the entire cohort, the patients treated surgically had better outcome compared with the nonsurgical groups (P = 0.02). The improved outcome with surgery was found in the lateral pillar groups B and B/C border, whose age at onset was greater than 8.0 years. In lateral pillar group B hips with an age at onset of more than 8.0 years, 73% of the operated hips had a Stulberg I or II result compared with 44% of the nonoperated hips (P = 0.02). The group B hips with onset at 8.0 years or younger had a favorable outcome profile, and there was no advantage demonstrated for the surgical group. This is an important finding given that 63% of the hips in the study were classified as group B. The group C hips, the most severely involved, were not shown to benefit from surgical treatment overall or in either age group when compared with the nonsurgical groups. No difference in outcome was reported between the Salter and varus osteotomy groups. Neither were there any differences among no treatment, range-of-motion treatment, and brace treatment.

The strength of this study lies in the prospective comparison among many surgeons in many centers. Longitudinal follow-up to maturity adds to the strength of the study, but the loss of follow-up of some patients before maturity weakens it. Exclusion forms were required for patients not enrolled to help reduce selection bias. A small number of hips (16) were excluded because of protocol violation. A total of 345 hips remained at final follow-up of 451 initially enrolled.

Buy Membership for Orthopaedics Category to continue reading. Learn more here