Long-term clinical and radiological outcome in patients with severe Legg-Calvé-Perthes disease after Chiari pelvic osteotomy: a mean of 14 years follow-up

Introduction: The Chiari pelvic osteotomy (CPO) has been recommended as a salvage procedure to improve head coverage in case of hip joint incongruence in paediatric hip disease. In this study, we aimed to assess the long-term results of CPO for severe Legg-Calvé-Perthes disease (LCPD). Methods: A total of 39 patients who underwent a CPO at our department between 1995 and 2010 were prospectively followed both radiologically (Stulberg classification) and clinically (Harris Hip Score [HHS], conversion into total hip arthroplasty). In this study, we retrospectively reviewed the cases of 12 hips (12 patients, 3 girls, 8 left hips) treated by CPO for severe LCPD (Catterall grade 3 or 4) with hip joint incongruence. Mean follow-up was 14.0 (range 7.6–21.3) years. Results: Mean age at surgery was 10.2 (range 8.2–17.8) years. Additional femoral osteotomy was performed in 8 patients. A good radiological result (Stulberg I or II) was achieved in 2 patients, a fair result (Stulberg III) in 4 patients, and a poor outcome (Stulberg IV or V) in 6 patients. Mean postoperative HHS averaged 93 (range 65–100) points. An excellent functional outcome (HHS 90–100 points) was achieved in 9 patients. No patient underwent total hip arthroplasty during follow-up. Postoperative limb-length discrepancy was found in 3 patients. Conclusions: CPO for severe LCPD with hip joint incongruence resulted in good long-term clinical outcome in about ⅔ of our patients after a mean of 14 years. Our results suggest that CPO can still be considered as a salvage joint-conserving procedure in this selected group of younger patients.


Introduction
There is great variability in the clinical and radiographic expression of Legg-Calvé-Perthes disease (LCPD). 1,2 LCPD results in a temporary softening of the femoral head, which may further lead to structural abnormities, hip joint incongruence, and secondary hip osteoarthritis. Conservative and operative treatments aim to minimise femoral head deformity to prevent premature osteoarthritis of the hip. 3 Main goals of LCPD treatment are femoral head containment and maintenance of hip motion. 3 A variety of different surgical treatment options have been described in the literature and can be classified as proximal femoral procedures, 4,5 pelvic procedures, [5][6][7][8][9] or combined procedures. 10 From a biomechanical view, correction of the acetabulum is the preferred procedure. 11 Already in 1952 Prof. Karl Chiari performed the first Chiari pelvic osteotomy (CPO) and reported encouraging short-term results. 12,13 Initially, Chiari limited the indications to subluxated and dislocated hips and claimed that his method provides stable coverage of the femoral head and improved biomechanics of the treated hip. 14 In the current literature CPO and shelf-acetabuloplasty are mainly considered salvage procedures for the treatment of LCPD with hip joint incongruence in adolescents and young adults. [15][16][17][18][19][20][21] In this study, we reviewed the results of a selected group of patients with incongruent hips resulting from severe LCPD and treated by CPO at our department. We aimed to assess long-term radiographic and clinical outcome of patients with severe LCPD after CPO.

Methods
The local ethics committee approved the study protocol and written informed consent was obtained from all patients.
We identified all patients at our institution who had undergone CPO from March 1995 to June 2010. CPO was performed in selected cases of LCPD with inadequate femoral head coverage. In this context, CPO was preferred over varus femoral osteotomy in cases with limited hip abduction (<30°). However, according to our previous institutional practice, we did not restrict CPO to cases with hip joint incongruence at that time. Additional femoral valgus extension osteotomy was performed in case of LCPD with hinge abduction (impingement of the extruded superolateral portion of the femoral head against the lateral acetabulum on arthrography).
In order to determine the value of CPO for severe cases of LCPD (Catterall grade 3 or 4), we retrospectively reviewed the cases of 12 hips (12 patients, 3 girls, 8 left hips) with hip joint incongruence in this study. In addition to our routine follow-up examinations, all patients were invited by telephone to participate in a separate study visit at our institution. This study visit included completion of Harris Hip Score (HHS) questionnaires, a physical examination, and a review of the latest x-rays. 1 patient was unable to follow our invitation for the follow up-examination in our hospital. Thus, 11 patients completed the follow-up study examination. Mean follow-up was 14.0 (range 7.6-21.3) years. Preoperative x-rays were not available for retrospective analysis in 3 patients. Patients' characteristics are shown in Table 1.
The operative technique of the Chiari pelvic osteotomy is described elsewhere. 12 The basic elements comprise a semi-circular osteotomy in the sagittal plane and an ascending straight osteotomy in the frontal plane of the ilium. 12,13,18 Preoperative hip arthrography was routinely performed. A single leg-hip spica cast was used for postoperative immobilisation in younger patients for 4-6 weeks followed by full weight-bearing and physical therapy.
Follow-up examinations were performed by 3 authors. The HHS was utilised to assess clinical outcome. 22 Furthermore, all patients were asked about their actual pain level using the numeric rating scale (NRS; 0-10; 0 = no pain; 10 = pain as intense as you can imagine). Hip range of motion was measured using a goniometer. Leglength discrepancy was assessed using standing blocks.
For radiographic evaluation we used anteroposterior films of the pelvis taken before and after osteotomy and at  follow-up examination. Acetabular Index (AI) and Wiberg's Centre-edge Angle (CEA) measurements were taken as references for the acetabulum shape. 23 The AI was measured as the angle formed between the lateral margin of the acetabular roof and the inferior aspect of the pelvic teardrop and a horizontal line between the inferior aspect of both pelvic teardrops. The CEA was formed by a vertical line from the centre of the femoral head and a line from the centre of the femoral head to the lateral edge of the acetabulum. The femoral head cover percentage (FHC) and Reimer's Index (RI) were calculated as the horizontal distance of the lateral femoral head covered (A)/uncovered (A') by the acetabulum divided by the total horizontal width of the femoral head (B). Both indices were expressed as a percentage (Figures 1 and 3). 24 Catterall and Herring classifications were used to preoperatively assess the stage of femoral head involvement. 2, 25 Stulberg classification was applied to evaluate the femoral head condition in radiographic outcomes and summarised as good (I/II), fair (III) or poor (IV/V) postoperatively. 26,27 Statistical analysis The statistical analysis was conducted in a purely descriptive manner with IBM SPSS Statistics 22 (IBM Corporation, Armonk, NY, USA). Variables on a nominal scale are reported as absolute values and percentages, variables on an ordinal scale as median and minimum/maximum, variables on an interval scale with mean and standard deviation. The non-parametric Shapiro-Wilk test was used to test for normality.

Results
CPO was performed in 12 patients (12 hips, nine boys and three girls). Mean age at presentation at our department was 9.1 (range 5.9-17.0) years and mean age at operation was (range 8.2-17.8) 10.2 years, respectively. In 3 cases, a concurrent femoral valgus osteotomy was applied via a lateral approach to obtain concentric reduction of the femoral head within the acetabulum. In 5 cases, femoral valgus (3 hips) or varus (2 hips) osteotomy was performed at an average of 1.5 years (range 0.3-2.3) before CPO. Screw removal was carried out routinely at 0.9 (range 0.3-1.6) years after CPO. 2 of the 3 females had normal spontaneous vaginal delivery after CPO.

Radiological findings
According to the preoperative Herring and Catterall classifications, 2 hips were graded Herring B, 10 hips Herring C, 7 hips Catterall III, and 5 hips Catterall IV. A good   Table 1.

Clinical findings
Mean postoperative Harris Hip Score averaged 93 (range 65-100) points. According to the HHS, an excellent functional outcome (90-100 points) was achieved in 9 patients, a good functional outcome in 1 patient, and a poor functional outcome (HHS <70 points) in 1 patient. Mean NRS for pain at the time of latest follow-up was 1 (0-5). No patient underwent total hip arthroplasty during follow-up.
Postoperative limb-length discrepancy was found in 3 patients and averaged 1.3 cm (range 1.0-1.8). All 3 limb-length discrepancies were caused by lengthening of the operated limb. 9 patients showed no limb-length discrepancy.

Discussion
Our study findings show good long-term clinical outcome at 14 years after CPO for severe LCPD with hip joint incongruence in about ⅔ of our patients. HHS averaged 93 points, showing only minor impairments for activities of daily living. An excellent functional clinical outcome (HHS 90-100 points) was achieved in ¾ (9 of 12) of patients. These clinical findings are in contrast to the radiographic outcomes, as only 2 patients who were graded as Herring B (Caterall 3) and underwent surgery before the age of 10 years exhibited a good postoperative result according to the Stulberg classification.
Long-term results following CPO have already been reported in the literature. [28][29][30] Most of those reports deal with the outcome in patients with dysplastic hips and poor joint congruency. 31 The best treatment for patients with LCPD and hinge abduction continues to be controversial and the role of CPO is still under debate. To our best knowledge only few authors have investigated outcome following CPO in patients with severe LCPD. 3,[32][33][34][35] From the given literature it is seen that the most important factors for successful surgical treatment in severe LCPD are congruency of the hip joint and femoral head sphericity. 32,35-37 Different types of surgical procedures for the treatment of LCPD are suggested in the literature. CPO is generally considered to be a salvage procedure for the treatment of LCPD when other osteotomies seem inappropriate. 3 Bennett et al. 15 reported in their study with severe cases of LCPD and incongruent hip joints that most pelvic osteotomies like the Salter-Innominate osteotomy are contraindicated. Klisic et al. 32 concluded that CPO is needed in hips where neither non-operative treatment nor femoral or innominate redirectional osteotomy could be expected to be successful. Reinker 38 mentioned that hinge abduction should be considered a contraindication to containment by redirectional pelvic or femoral varus osteotomy. For patients in whom hinge abduction is suspected, a shelf arthroplasty or CPO is recommended, but if a reasonably sized shelf would not adequately cover the femoral head, a Chiari osteotomy would be a valuable alternative. 38 We do not perform shelf arthroplasty and therefore have no experience with this procedure at our institution. Lack et al. 35 stated that CPO might delay the clinical effects of hip osteoarthritis for years in patients with LCPD. This finding was also supported by other studies that indicated that the degenerative process might be slowed as a result of the favourable biomechanical properties of CPO. 18,[35][36][37] This might be explained by reduction of the joint pressure due to increased weight-bearing surface, shortening of the lever arm of the body weight, and centralisation of the femoral head within its muscle cuff. [35][36][37] We cannot support the view of Kamegaya et al. 39 that CPO leads to a medialisation of the hip centre. According to Stulberg et al. 26 , the most important factors for successful surgical treatment of severe LCPD are congruency of the hip joint and femoral head sphericity. However, for patients with an incongruent hip joint and/or hinge abduction phenomenon CPO could be a valuable alternative. In our study, we demonstrate a significant improvement in femoral head coverage and a good clinical outcome in about ⅔ of such patients.
CPO has been recommended mainly for adolescents and young adults. 40 Long-term survival of the hip joint was related to age at operation for hip dysplasia. 28,29,41,42 Our results indicate that children and adolescents with LCPD aged between 8 and 18 years undergoing CPO can achieve good functional results. This might be explained by the improved coverage of the femoral head after CPO, which hypothetically contributes to the epiphyseal growth, and promotes remodelling of the femoral head. Reddy and Morin 3 reported good results after CPO in children with an average age at surgery of 8.5 years. Crutcher and Staheli 43 found a 50% improvement in femoral head sphericity in young children classified as Caterall group III and IV cases after combined Salter innominate and varus femoral osteotomies. Consequently, Canale et al. 6 report poor remodelling in all cases operated at late stage LCPD. Also Windhager et al. 29 found a poorer outcome with increasing age at the time of operation. In contrast, in a study by Cahuzac et al. 16 , younger patients were seen to have a poorer outcome. In summary, the question as to the right age for an LCPD patient with an incongruent hip joint to undergo CPO remains unanswered.
We acknowledge the following limitations of our study. Foremost, the relatively small number of patients may inhibit adequate analysis of significant risk and prognostic factors of CPO in LCPD. Furthermore, the complete dataset was not available for all patients and patient specifications were inhomogeneous at the time of surgery. This, however, is mostly the case in paediatric orthopaedic surgery with rare diseases and heterogeneous approaches. Furthermore, the Harris hips score used to assess functional outcome was developed for adults with hip osteoarthritis. Therefore, it may not be sufficiently sensitive in a population of young adults with the deformity prior to the onset of osteoarthritis.
In conclusion, the best treatment for LCPD patients with hinge abduction continues to be controversial. The Chiari pelvic osteotomy is a salvage procedure, effective for improving hip coverage and diminishing joint pressure. We therefore believe that the Chiari pelvic osteotomy can be considered a salvage procedure for children with severe cases of LCPD with incongruent hip joints and hinge abduction, when containment procedures are contraindicated.