Residual dysplasia or subluxation of the hip in childhood is reported to be a leading cause of premature arthritis of the hip in young adults, particularly women1-6. It has been postulated that early surgical correction of residual dysplasia or subluxation to achieve a normal radiographic result by the time that the patient reaches school age will delay or prevent the onset of premature arthritis7,8.
Dysplasia is a broad term that refers, after the neonatal period, to inadequate development of the acetabulum, femoral head, or both, as documented on plain radiographs1,6,7,9. With hip dysplasia, Shenton's line is intact1,4,5,10. The acetabulum may be abnormally steep when the acetabular index is >20°, in patients who are more than two years old, or when the Sharp angle is >42°11-13. In addition, acetabular changes may include a short acetabular roof, a shallow joint surface, and/or thickening of the lateral aspect of the sourcil (acetabular subchondral bone)5,6,13,14. On the femoral side, there may be slight flattening of the femoral head with or without varus or valgus angular deformity and excessive anteversion. Angular deformities of the proximal part of the femur are characterized by a neck-shaft angle of <117° (varus) or >132° (valgus)11. Femoral anteversion of >60° is considered excessive15.
A subluxated hip is defined as one with a clear break in Shenton's line, as noted on the anteroposterior radiograph, indicating that the femoral head is superiorly and/or laterally displaced from the medial wall of the acetabulum. In most cases, severe dysplasia is also present. Thus, a dysplastic hip that is also subluxated is a more serious problem and requires more vigorous treatment.
During the past few decades, there has been a trend toward early treatment of residual dysplasia and/or subluxation of the hip7,8,16. The purpose of the present study was to analyze the results of operative correction of dysplasia or subluxation with a femoral and/or pelvic osteotomy in patients who were between two and eight years old. We compared the results in patients who were managed with early operative treatment (when they were between two and eight years old), and those in patients who were managed with an operation when they were between eight and eighteen years old to determine whether treatment at an earlier age improved the outcome.
Materials and Methods
Group I included all patients with residual hip dysplasia or subluxation who were between two and eight years old and were treated with either a femoral and/or pelvic osteotomy at our institution between 1976 and 1997. Patients with a specific syndrome or a neuromuscular condition were excluded. Patients who had a minimum follow-up of two years and for whom essential radiographs were available for review were included in the present study. Dislocated hips that required an osteotomy as part of the treatment at the time of closed or open reduction were excluded.
Group II included patients with residual dysplasia and/or subluxation of the hip who were treated with a pelvic and/or femoral osteotomy at an older age (between eight and eighteen years old) during the same time-period. The results in these patients were compared with those in Group I.
Data on previous treatment, age at the time of surgery, surgical procedure performed, complications, need for subsequent operations, and clinical and radiographic outcome at the latest follow-up examination were obtained from the hospital records and a review of the plain radiographs.
Anteroposterior radiographs of the pelvis were made in a reproducible fashion with the patient standing and the feet facing forward. Measurements were made preoperatively, two to four months postoperatively, yearly (when such radiographs were available), and at the time of the latest follow-up. In Group I, the patients had an adequate number of radiographs to allow data to be recorded from those made preoperatively, two to four months postoperatively, one year postoperatively, and at the time of the latest follow-up. In Group II, the available radiographs from which it was possible to record data accurately included those made preoperatively and at the time of the latest follow-up only.
The measurements included the acetabular index, the Sharp angle (the acetabular index in an adult), the center-edge angle of Wiberg, Shenton's line, and the presence or absence of avascular necrosis, according to the classification system of Bucholz and Ogden6,11-13,17,18. Because the serial radiographs of any patient in Group I may have been made before and after closure of the triradiate cartilage, both the acetabular index and the Sharp angle were recorded on each radiograph. In Group-II patients, only the Sharp angle was measured. Despite its limitations in the analysis of the ossifying femoral head, a Mose concentric-circle template was used to document the center of the femoral head for measurements of the center-edge angle in Group I. The congruity of Shenton's line was assessed with use of Mose concentric circles fitted to the inferior surface of the proximal femoral metaphysis and the superior pubic ramus12.
Radiographic analysis of the neck-shaft angle is often inaccurate because of the lack of standardization in positioning of the patient12. As a result, determination of the true neck-shaft angle was made with use of intraoperative fluoroscopy. Once the patient was anesthetized and positioned supine on a radiolucent table, an intraoperative assessment of the neck-shaft angle and femoral anteversion was made. With the pelvis level, the hip was abducted and internally rotated until the maximum width of the femoral neck was visualized on the image intensifier, allowing accurate measurement of the true neck-shaft angle. The normal value for the neck-shaft angle is 125° ± 7°11. The femoral anterversion was determined with use of the technique developed by Ryder and Crane19.
In addition, each hip was rated with respect to the most common indicators of dysplasia to provide a simple summary of femoral head coverage. A hip was considered to have normal findings when the acetabular index was <20° (or the Sharp angle was <42°), the center-edge angle was >20°, and Shenton's line was intact.
Thirty-six patients (fifty hips) in the juvenile age-group (two to eight years old) met the inclusion criteria of the study. There were thirty girls and six boys. Fourteen patients had bilateral involvement. The average age at the time of surgery was 3.7 years (range, 2.0 to 7.1 years), and the mean duration of follow-up was 4.3 years (range, 2.0 to 10.7 years).
Several patients had been treated during infancy because of acetabular dysplasia and/or instability of the hip. Twenty-two patients (thirty-four hips) had a previous diagnosis of acetabular dysplasia without instability in infancy. Twelve hips had had no previous intervention, whereas twenty-two hips had been managed with bracing. The type of brace used was age-dependent. Infants who were less than ten months old were treated with a Pavlik harness. Infants and children who were more than ten months old were managed with a polypropylene abduction brace.
Fourteen patients (sixteen hips) had a previous diagnosis of hip instability with or without acetabular dysplasia in infancy. Three hips were treated with an open reduction only. The remaining thirteen hips had a combination of prior treatments, including bracing (six hips), traction (ten), closed reduction (six), adductor tenotomy (six), and open reduction (five). The open reduction was performed from an anterior approach alone in five hips and through a medial (Ludloff) approach in two hips. The remaining hip had both an initial medial open reduction and a later anterior open reduction and capsulorrhaphy because the stability of the hip could not be maintained in the cast following the medial open reduction.
At the time of the index osteotomy, twenty-two hips had residual dysplasia (Group IA) and twenty-eight hips had residual subluxation (Group IB).
A neck-shaft angle of >140° (mean and two standard deviations), as measured by intraoperative fluoroscopy, was considered excessive and, in such cases, a varus osteotomy was performed. Femoral anteversion of >60° was considered a relative indication for derotation osteotomy15. A femoral derotation osteotomy was rarely done because of isolated, increased femoral anteversion without an associated elevation in the neck-shaft angle. The indications for a pelvic osteotomy were an acetabular index of >20° or a Sharp angle of >42°, and/or a center-edge angle of <20° on the preoperative plain radiographs.
The twenty-two hips in Group IA (residual dysplasia) were managed with a femoral osteotomy only (two), a pelvic osteotomy only (seventeen), or combined femoral and pelvic osteotomies (three). The two hips that had a femoral osteotomy only were both treated with a varus derotation osteotomy. In the subgroup that had a pelvic osteotomy only, sixteen hips underwent a Pemberton osteotomy ( Figs. 1-A , 1-B , and 1-C ) and one, a Sutherland double innominate osteotomy. In the subgroup that had combined femoral and pelvic osteotomies, the femoral osteotomy consisted of a derotation osteotomy (one hip) or a varus derotation osteotomy (two hips). The pelvic procedures were all Pemberton osteotomies. Additional procedures performed at the time of the index operation included an arthrogram in four hips.
The twenty-eight hips in Group IB (residual subluxation) were managed with a femoral osteotomy only (eight hips), a pelvic osteotomy only (three), or combined femoral and pelvic osteotomies (seventeen). In the subgroup that had a femoral osteotomy only, seven hips underwent a varus derotation osteotomy and one hip, a pure varus osteotomy. In the subgroup that had a pelvic osteotomy only, all three hips underwent a Pemberton osteotomy. In the subgroup treated with combined femoral and pelvic osteotomies, the femoral osteotomy consisted of a varus derotation osteotomy (thirteen hips) or a pure varus osteotomy (four hips), and the pelvic procedures included fifteen Pemberton osteotomies and two Salter innominate osteotomies. Additional procedures performed at the time of the index operation included an arthrogram in two hips and capsulorrhaphy in three hips.
In the comparison group, the corrective osteotomies in fourteen patients (eighteen hips) were delayed. There were twelve girls and two boys, and four of them had a bilateral procedure. The mean age at the time of the surgery was 13.7 years (range, 9.1 to 17.2 years), and the mean duration of follow-up was 4.9 years (range, 2.0 to 12.5 years).
Thirteen hips had a previous diagnosis of acetabular dysplasia without instability that was made in the neonatal period or during infancy. None of these hips had prior treatment. In the remaining five hips, instability with or without acetabular dysplasia was diagnosed in the neonatal period or during infancy. Two of them had a previous closed reduction only, and three had a combination of prior treatments, including traction (two hips), closed reduction and application of a cast (three), and adductor tenotomy (one).
At the time of the index osteotomy, four hips had residual dysplasia (Group IIA) and fourteen hips had residual subluxation (Group IIB).
All four Group-IIA hips were treated with a pelvic osteotomy only. The pelvic procedures included a Pemberton osteotomy (one hip), a double innominate osteotomy (two), and a triple innominate osteotomy (one).
The fourteen hips in Group IIB were managed with a pelvic osteotomy only (six) and combined femoral and pelvic osteotomies (eight). In the subgroup treated with pelvic osteotomy only, the pelvic procedures included double innominate osteotomy (four hips) and triple innominate osteotomy (two). In the subgroup treated with combined femoral and pelvic osteotomies, the femoral procedure consisted of varus derotation osteotomy (five hips) or pure varus osteotomy (three), and the pelvic procedures included a double innominate osteotomy (five), a triple innominate osteotomy (two) ( Figs. 2-A , 2-B , and 2-C ), and a Pemberton osteotomy (one).
Severin Radiographic and Functional Classifications
The radiographic and functional results at the time of the latest follow-up in both age-groups were analyzed according to a modification of the classification systems of Severin6,20,21. Although juvenile patients are, for a number of reasons, far less likely to exhibit signs and symptoms of hip dysplasia than are adolescents, who are heavier and less resilient, functional data were collected without the intent of comparing the two groups.
All radiographic results were analyzed with respect to age-group (two to eight years old or eight to eighteen years old) and the presence of residual hip dysplasia or subluxation at the time of the index osteotomy ( Table I ).
The data for Group I were sufficient to allow each measurement and all measurements as a whole to be analyzed with respect to the duration of time after surgery. In all of the subgroups, there was a general trend toward normalization of the radiographic results, starting in the early postoperative period, which was defined as two to four months following surgery. This trend continued during the first year and seemed to plateau around the one-year mark. Insufficient numbers of radiographs were available at two and three years following surgery to allow us to comment accurately on the findings. At the time of the latest follow-up visit, at a mean of 4.3 years (range, 2.0 to 10.7 years), the correction gained in the first year after surgery had been maintained for the most part, with minimal loss of correction of the various radiographic indices.
The hips that were considered to have a normal result according to the radiographic criteria (an acetabular index of <20° [or a Sharp angle of <42°], a center-edge angle of >20°, and an intact Shenton's line) were categorized with respect to the presence of residual hip dysplasia or subluxation at the time of the operation and the type of procedure. In Group IA (residual dysplasia), both hips that had a femoral osteotomy only, sixteen of seventeen hips that had a pelvic osteotomy only, and all three hips that had combined femoral and pelvic osteotomies met the criteria. In Group IB (residual subluxation), four of the eight hips treated with a femoral osteotomy, two of the three hips treated with a pelvic osteotomy only, and fifteen of the seventeen hips that had combined femoral and pelvic osteotomies met the criteria.
In Group IIA (residual dysplasia), three of the four hips that were treated with a pelvic osteotomy only were considered to have normal radiographic findings on the basis of our definition. In Group IIB (residual subluxation), three of the six hips that were treated with the pelvic osteotomy only and four of the eight hips that were treated with the combined femoral and pelvic osteotomies had normal radiographic findings.
The radiographic measurements were also used to determine the Severin radiographic classification for each hip in both age-groups ( Table II ).
Group I had remarkably few complications, with no osteonecrosis, wound infections, implant failure, or loss of internal fixation.
In Group II, none of the fourteen patients (eighteen hips) had development of osteonecrosis or deep infection. Pseudarthrosis developed in one hip in a thirteen-year-old girl with residual subluxation who had had a left triple innominate osteotomy. A nonunion that developed at the site of the iliac osteotomy healed after it was treated with an iliac crest bone graft and screw fixation. In addition, two patients (two hips) in Group II had a complication related to fixation. A sixteen-year-old girl with residual subluxation of the hip had undergone a right double innominate osteotomy with distal transfer of the greater trochanter. The iliac pins broke after she started walking, against medical advice. No additional treatment was required. The other patient, a fifteen-year-old girl with residual hip dysplasia who underwent a left double innominate osteotomy, had bending of the iliac pins with medial displacement of the osteotomy in the early postoperative period. No additional treatment was undertaken.
Need for Subsequent Surgery to Correct Dysplasia
Two (4%) of the fifty hips in Group I and four of the eighteen hips in Group II needed a subsequent revision hip operation. The radiographic data in Table I represent the measurements made before the revision operations. In Group I, one of the revisions was performed in a girl with residual subluxation of the hip who had had a right proximal femoral varus osteotomy when she was six years old. Five years postoperatively, she had poor acetabular development and a progressive increase in the neck-shaft angle. At the age of 10.9 years, she had a repeat proximal femoral osteotomy combined with a Salter innominate osteotomy. The other revision was performed in a girl with residual subluxation of the hip who had had a bilateral Pemberton pelvic osteotomy when she was two years old. She had a persistent slight elevation in both neck-shaft angles at eleven months postoperatively. At the age of 2.9 years, she underwent a bilateral proximal femoral varus derotation osteotomy.
The four patients who needed revision surgery in Group II included a sixteen-year-old girl with residual subluxation who had had a double innominate osteotomy. She required subsequent distal transfer of the greater trochanter and proximal femoral varus osteotomy one year after the index operation. A thirteen-year-old girl with residual subluxation who had been managed with a triple innominate osteotomy required an iliac crest bone graft and screw fixation nine months postoperatively because of a nonunion at the site of the iliac osteotomy. The third patient, a girl with residual subluxation that was treated with a left double innominate osteotomy at the age of eleven years, subsequently required a capsulorrhaphy and shelf arthroplasty three years later. The fourth patient was a twelve-year-old boy with residual dysplasia who had had a left proximal femoral varus derotation osteotomy and a double innominate osteotomy. Four years postoperatively, he required a shelf arthroplasty. The procedure was complicated by a foot drop, which was attributed to a retraction injury to the peroneal branch within the sciatic nerve and was documented by electromyographic and nerve-conduction studies.
Severin Functional Classification
The clinical outcome was recorded at the time of the final follow-up with use of the Severin functional classification system ( Table II ). Most (88%) of the fifty hips in Group I were rated as class A (complete freedom from signs and symptoms and without limp or pain). In Group II, a higher proportion of hips were rated as class B (patient has no limp but feels the hip on any great exertion) or class C (a continual limp but no other symptoms).
In Groups IA, IB, and IIA, no patient had a limb-length inequality. In Group IIB, one of five patients (six hips) who had a pelvic osteotomy alone had a limb-length inequality, with the longer limb on the involved side. This patient had a 2.5-cm discrepancy and wore an insert in the shoe to equalize the limb lengths. No patient who had combined femoral and pelvic osteotomies in Group IIB had a limb-length inequality.
In the past, the natural history of residual dysplasia and/or subluxation of the hip secondary to developmental dysplasia or dislocation was less well understood. As a result, only symptomatic patients with dysplasia or subluxation of the hip and radiographic findings of early osteoarthritis were treated, usually when they were teenagers or young adults.
Over the last twenty years, long-term follow-up studies have further demonstrated that patients with residual hip dysplasia or subluxation have a high risk for the development of premature osteoarthritis5. It has been estimated that degenerative disease of the hip in adults is secondary to subluxation or acetabular dysplasia that is a residual of childhood hip dysplasia in 20% to 50% of adults with the disease5. In the classic study by Stulberg and Harris, 48% of 130 patients with degenerative arthritis of the hip were found to have underlying acetabular dysplasia3. They also reported that premature joint disease and clinical disability predictably develop in patients with a subluxated hip that is not treated. The age at the onset of symptoms has been found to be closely associated with the severity of the subluxation3,6,14. Patients with the most severe subluxation begin to have pain in the second decade of life, whereas those with less severe subluxation may reach the fifth decade prior to noting symptoms14. The effect of hip dysplasia without subluxation is less predictable. Patients are frequently asymptomatic, and osteoarthritis tends to develop later and less predictably than it does in those who have subluxation1.
Initially, this new understanding was applied mainly to adolescents with hip dysplasia or subluxation. However, the surgical results were not uniformly predictable in this age-group, especially if the goal was to produce a completely normal hip that would be durable and functional, even in an athletic individual. In our evaluation of the patients in the present study who had surgical correction of subluxation when they were between eight and eighteen years old, we found that only about half had an excellent result. On the basis of this finding, our current policy is to perform surgical correction of residual dysplasia or subluxation when our patients are in early childhood. We found the results to be more predictable, with less extensive surgery, easier healing, and fewer complications.
A pelvic osteotomy alone should be sufficient to correct residual dysplasia. Sixteen of the seventeen hips with residual dysplasia that had a pelvic osteotomy alone in Group I had normal radiographic findings at the time of the latest follow-up compared with three of four such hips in Group II. In hips with residual subluxation, a combination of femoral and pelvic osteotomies produced the best results. Fifteen of the seventeen hips with residual subluxation that had combined femoral and pelvic osteotomies in Group I had normal radiographic findings compared with four of eight such hips in Group II.
The results after an intermediate duration of follow-up in the present study demonstrate the safety and efficacy of our approach. By intervening earlier (when the patient is between two and eight years old), we can more predictably achieve normal radiographic results. The surgery is usually less extensive and better tolerated, with a more rapid functional recovery. In addition, the surgery appears to be associated with fewer complications, including a decreased need for revision surgery. We are optimistic that early operative treatment will lead to prolonged normal function of the hip; however, long-term follow-up studies are needed to document the efficacy of early osteotomy on the delay or prevention of premature onset of arthritis of the hip.
Note: The authors thank Dr. D. Sutherland, Dr. S. Mubarak, Dr. P. Newton, Dr. H. Chambers, and Dr. D. Wallace for allowing us to include their patients in the study.
Investigation performed at the Division of Orthopedic Surgery, Children's Hospital San Diego, and the University of California at San Diego, San Diego, California
No benefits in any form have been received or will be received from a commercial party related directly or indirectly to the subject of this article. No funds were received in support of this study.