Short follow-up evaluation of proximal femoral varus osteotomy for treatment of Legg–Calvé–Perthes disease

Legg–Calvé–Perthes disease (LCPD) is defined as an idiopathic osteonecrosis of the femoral head which leads to variable complications with resultant deformity of the femoral head and, later, osteoarthritis [26, 28]. Unlike normal, healthy bone, the bone of the avascular epiphysis is not capable of withstanding the stresses on the epiphysis of the femoral head in cases of LCPD. The aim of treating Perthes disease is to reduce the risk of later osteoarthritis by preventing femoral head deformity, which may occur if adequate containment is not achieved [11, 18, 19]. To achieve containment, the femoral head is centered within the acetabulum during the fragmentation and reossification phase. This allows the acetabulum to act as a mold during the healing or revascularization phase when the biologically plastic femoral head is at risk of subluxation, hinged abduction, and permanent femoral head deformation. At skeletal maturity, severe femoral head deformity and joint incongruity increase the risk of loss of function, leading to osteoarthritis later on [8, 17, 18].

There are many treatment methods for Legg–Calvé–Perthes disease, and the appropriate method to use depends on the grade of the disease. Such methods include spica cast immobilization, bed rest, traction, and walking with a weight-relieving caliper [3]. Surgical methods are also employed in young patients with Legg–Calvé–Perthes disease. Some authors have recommended nonoperative means such as bracing and cast immobilization, and have reported satisfactory outcomes for most patients [4, 9, 14]. Many other authors have reported good results with operative techniques such as femoral varus or valgus osteotomy, as well as other types of pelvic osteotomies such as innominate (Salter) pelvic osteotomy, lateral shelf osteotomy, and triple osteotomy [4‐6, 8, 10, 14, 20, 35, 36, 39‐41]. The study reported in the present paper was designed to investigate the short-term outcome results of proximal femoral varus osteotomy in the treatment of Catteral grade III (Herring groups B, BC) and IV (Herring group C) Perthes disease according to various classifications and grading schemes, and to compare the results to those achieved using other methods of treating Perthes disease reported in the literature.

Between July 2005 and December 2011, 23 patients with unilateral Legg–Calvé–Perthes disease were treated using a proximal femoral varus osteotomy procedure at Zagazig University Hospitals. The right side was affected in 15 patients and the left side in the remaining 8 patients. The mean age of the patients was 7.8 years (range: 6–11.5 years). All of them were males. The average follow-up was 36.2 months (range: 29–48 months). The patients’ main complaints were hip pain with limping and accompanying knee pain. The clinical complaints and their onset dates were noted, and the flexion, extension, abduction, adduction, and internal and external rotation ranges of the hips and limbs were also recorded. Clinical measurements and scanograms were used to measure and detect leg length discrepancies. Anteroposterior pelvic X-rays were obtained in neutral, abduction, and abduction–internal rotation positions as well as frog-leg lateral (Lauenstein) views for all of the patients. Epiphyseal or femoral head involvement was graded according to the classifications of Catterall [7] and Herring et al. [16]. Both the extrusion index [12] and Wiberg’s CE angle [40] were measured and graded, and the risk factors of the patients were also identified.

Based on the Catterall and Herring classifications, the 23 patients were categorized as follows: 6 patients were Catterall group III (Herring group B), another 10 patients were Catterall group III (Herring group BC), and the other 7 patients were Catterall group IV (Herring group C). The preoperative and postoperative mean epiphyseal extrusion indices were as follows: group III (B), 10.88 % and 7.22 %, P = 0.027; group III (BC), 15.81 and 8.93 %, P = 0.005; group IV (C), 72.64 and 39.44 %, P = 0.018. The preoperative and postoperative mean Wiberg’s CE angle were as follows: group III (B), 26.88° and 37.81°, P = 0.028; group III (BC), 24.4° and 32.2°, P = 0.005; group IV (C), 20.89° and 28.41°, P = 0.018. The changes in the femoral neck–shaft angle were as follows: group III (B), 137.07° to 117.02°, P = 0.028; group III (BC), 137.0° to 117.3°, P = 0.005; group IV (C), 137.67° to 115.64°, P = 0.017. The preoperative and the postoperative mean Iowa clinical hip score were as follows: group III (B), 54.8 and 92.33, P = 0.027; group III (BC), 47.3 and 87.8, P = 0.005; group IV (C), 34.43 and 68.29, P = 0.017. Mean limb length discrepancy was 0.9 cm (range: 0.0–2 cm) of shortening at the operated side compared to the normal side at the last follow-up after plate removal. We did not see any progressive change in this parameter during the follow-up period, especially after hardware removal and in the younger boys. Limping related to limb length discrepancy and the gluteal weakness was reported for all patients, but this improved gradually over the course of 8 months after the operation. All of the osteotomies united within 3 months without loss of fixation (see Table 1; Fig. 1a–d).

The early goal of treatment is to prevent head deformation by weight-related forces during remodeling and ossification, so containment is the widely accepted treatment principle [28]. The main indication for operative containment treatment of Perthes disease is age >6 years along with lateral subluxation and advanced femoral head involvement [3, 14]. The most commonly reported surgical method for the treatment of Legg–Calvé–Perthes disease is proximal femoral varus osteotomy, which was first introduced in 1965 and has since become a popular surgical treatment for LCPD. Femoral varus osteotomy improves the intraosseous circulation, the mechanics around the proximal femoral head, and subsequently the degree of femoral head sphericity after healing, and it permits the regeneration of the necrotic tissues of the femoral head. It also prevents the subluxation of the femoral head, covering it with the acetabulum. It restores joint congruity and reduces femoroacetabular impingement [3‐5, 13, 24, 28, 38].

On the other hand, the main aims of femoral valgus osteotomy are to reduce hinged abduction during remodeling and to improve the symptoms and the range of motion [20]. In severely deformed femoral heads treated with femoral valgus osteotomy, greater congruency is obtained in adduction rather than in abduction. Kim et al. evaluated the effectiveness of valgus osteotomy based on femoral head roundness, femoral head subluxation, and function. They found that this technique helped to keep the deformed femoral head inside the acetabulum during the fragmentation phase so that it could be remodeled to fit neatly inside the acetabulum [20]. Besides this, valgus osteotomy is valuable for relieving hinged abduction after skeletal maturity has been reached [41].

Recently, many authors have argued against nonoperative treatment, especially in children >6 years old with lateral pillar type B, B/C, or C LCPD. This group benefits more from varus or innominate osteotomy than nonoperative treatment because both pain and hip dysfunction are common in them. The clinical signs of femoroacetabular impingement and the radiographic signs of hip osteoarthritis were also found to be correlated with pain in nonoperatively treated patients [13, 22, 24, 33]. The main advantage of Salter or innominate osteotomy is its effect on femoral head remodeling during the remaining growth. Patients who are indicated for this osteotomy alone are usually younger children with a recent clinical onset of LCPD and no femoral head deformity or subluxation [34, 39].

One of the surgical methods used when other treatment options are contraindicated is arthrodiastasis of the hip joint with soft-tissue release. The advantages of this procedure are that it improves the range of motion, reduces superior and lateral subluxation, and provides better radiographic sphericity of the femoral head. This treatment can even be performed with distraction in stiff hips and deformed hips [1, 2, 21, 27]. In a comparative study by Voplon in which he used arthrodistraction as a primary treatment for active forms of LCP disease and prospectively compared the results with those obtained using Salter innominate osteotomy, although the methods gave similar final radiological results, morbidity was higher with arthrodistraction than with innominate osteotomy. Consequently, the author does not recommend arthrodistraction as a primary treatment for the early stages of Legg–Calvé–Perthes disease [39].

Although shelf acetabuloplasty leads to improved femoral head coverage, the available literature does not support the use of this procedure to prevent late osteoarthritis and improve function long term [17]. Recently, a new labral support technique has been reported. This shelf arthroplasty technique includes a minimal-incision variant of labral support shelf arthroplasty, arthroscopic visualization, and an allograft buttress on the shelf support that has been proposed to maintain containment. This minimal-incision technique yielded similar results to those obtained using a Petrie cast, a femoral varus osteotomy, or an innominate (Salter) osteotomy. The authors concluded that the labral support shelf arthroplasty technique is simple to perform and does not induce a permanent deformity in the proximal femur or acetabulum [6]. An advantage compared to femoral varus osteotomy is that there is additional lateral growth of the true acetabulum to generate more coverage following surgery. In this way, the labral support technique can stimulate lateral acetabular growth, restore the shelf after femoral epiphyseal reossification, and prevent subluxation [10].

Triple innominate osteotomy is considered one of the most efficient techniques for femoral head containment in any LCPD case. The main disadvantage of this technique is overcoverage of the femoral head, which can lead to pincer impingement. To prevent this complication, correction of the center-edge angle beyond 44° should be avoided [16].

Impingement and instability with intra- or extra-articular deformities of the hip can lead to joint damage and premature osteoarthritis of the hip. Surgical dislocation of the hip with trochanteric advancement faciliates lengthening of the femoral neck and the reduction of femoral head deformities. Leuing and Ganz reported 14 patients with surgical dislocation of the hip and trochanteric advancement with a minimum follow-up of 3 years. They found that pain, hip mobility, and gait improved greatly in this group of patients, with no major complications observed [25]. They noted that surgical dislocation of the hip yielded promising results in the treatment of femoral head deformities following LCPD [25, 31]. The authors reported transtrochanteric rotational osteotomy as a new technique for patients in whom the onset of LCPD occurs after 9 years of age. They concluded that this technique is an effective procedure for salvaging late-onset LCPD in affected hips, and that the amount of head involvement and the lateral pillar influence the surgical outcome [29, 36].

Herring et al. [14], in a prospective multicenter study, found a strong correlation between the lateral pillar classification, age at disease onset, and outcome in patients with Legg–Calvé–Perthes disease. Patients >8.0 years old at disease onset who had a hip categorized as lateral pillar B or B/C according to the Herring classification were found to have better outcomes following surgical treatment with either femoral varus osteotomy or innominate osteotomy than they did with nonoperative treatment. Group B hips in children <8.0 years of age at disease onset have very favorable outcomes that are unrelated to choice of treatment, whereas group C hips in children of all ages frequently have poor outcomes, which again appear to be unrelated to the choice of treatment.

In contrast to Herring, in another multicenter Norwegian prospective study on Legg–Calvé–Perthes disease that was published by Terjesen et al. [37], 70 patients who were diagnosed at >6 years old with unilateral LCPD and had femoral head necrosis of more than 50 % were treated with femoral varus osteotomy. In that study, both the Catterall and Herring classifications of necrosis were applied, and the results were compared with a control group of 51 similar children who received physiotherapy treatment. They concluded that, in children aged 6.0–10.0 years in whom the whole femoral head is affected, femoral head sphericity at 5 years follow-up after femoral varus osteotomy was better than that achieved with physiotherapy. Those results are in good accord with the results of the present study because the three unsuccessfully treated patients in the present study were classified as having Catterall group IV (Herring type C) LCPD and were over 10 years old. These findings support the efficacy of early surgery for appropriately selected patients due to the remarkable decrease in the ability to remodel after 5–6 years of age, meaning that patients over 6 years of age should be the main candidates for operative treatment [3, 18, 19, 37].

Rather than Salter’s osteotomy, the author of the present study preferred proximal femoral varus osteotomy because this method achieves decompression, enables dynamic treatment, does not increase intra-articular pressure, and does not cause a frozen joint postoperatively with good coverage of the femoral head in the hip joint [3, 12, 15, 30, 37]. It is also worth noting that femoral varus osteotomy is said to have certain disadvantages or complications, such as femoral shortening, limping, excessive varus, nonunion, and overgrowth and elevation of the greater trochanter [3, 12, 15, 30, 37]. The most important predictor of leg length discrepancy (LLD) is the extent of lateral pillar involvement, and no other factor (including age, sex, and treatment modality) is correlated with LLD at skeletal maturity [32]. In the present study, the author decided to use an open-wedge osteotomy, as persistent limb shortening tends to be greater after a closed-wedge osteotomy in the older child. We did not see any progressive change in this parameter during the follow-up period, especially after hardware removal and in the younger boys, but it may decrease with time as the varus angulation of the subtrochanteric osteotomy gradually changes.

Limping related to limb length discrepancy or gluteal weakness or both have generally been reported by other authors after a proximal femoral varus osteotomy. In the present study, the author encountered limping, but the limping gradually improved within 8 months after the operation, in agreement with observations reported by other authors. No complications such as delayed union, nonunion, overgrowth, or elevation of the greater trochanter were encountered in the present study. However, the author did face common problems and limitations associated with studies in this field, including the variable nature of Perthes disease (which makes the condition difficult to study) and the use of different classification systems and outcome measures (which leads to confusion). Analysis of surgical procedures is hampered by the use of small subject groups, the infrequent use of a control group, the unmatched selection of patients of varying ages, and the varying severity of the disease process.

In conclusion, proximal femoral varus osteotomy gives good results in children aged 6–10 years who do not exhibit any femoral head deformity or flattening, especially those with good containment in abduction. Treatment failure is not usually because of the treatment method; it is due to technical errors, inappropriate patient selection, and delayed treatment. All recently reported techniques aim to reshape the femoral head in both congruency and size to match the acetabulum and sequentially decrease the impingement, as well as to restore the normal cartilage in the weight-bearing zone of the head.