Abstract
Background: The principal complications that follow the treatment of developmental dysplasia of the hip are redislocation and growth disturbance of the femoral head and neck as a result of osteonecrosis of the femoral epiphysis. Growth disturbance secondary to osteonecrosis is difficult to determine until long after the treatment episode has passed. Consequently, the treating surgeon has little early feedback regarding the long-term consequences of management interventions. We therefore sought to devise a quantitative method to identify early evidence of growth disturbance related to osteonecrosis.
Methods: The width and height of the epiphyses were measured on anteroposterior radiographs of the pelvis made twelve to eighteen months after successful closed reduction and on the latest available radiograph for each patient (mean age, 8.6 years). The epiphyseal index was calculated by dividing the height by the width. The radiographs were also scored for osteonecrosis with use of the Kalamchi and MacEwen classification system and were also assessed for sphericity with use of Mose rings.
Results: Forty-seven patients with late-presenting developmental dysplasia of the hip who subsequently underwent successful closed reduction were included. An index of <0.357 on the twelve to eighteen-month post-treatment radiograph strongly predicted the development of a nonspherical femoral head on the latest radiograph (sensitivity, 0.83; specificity, 0.95; positive predictive value, 0.55; and negative predictive value, 0.99).
Conclusions: The height-to-width index appears to be a simple and quantifiable measurement of the severity of growth disturbance as a consequence of osteonecrosis following treatment for developmental dysplasia of the hip. It is predictive of asphericity at the time of intermediate-term follow-up and appears likely to predict asphericity at maturity, but this must be confirmed with follow-up to maturity. Unlike the currently used methods of assessing osteonecrosis, the index allows for the quantifiable evaluation of growth disturbance within a few years after the corrective procedure.
Level of Evidence: Diagnostic Level III. See Instructions to Authors for a complete description of levels of evidence.
Growth disturbance of the capital femoral epiphysis as a result of osteonecrosis is one of the main complications of treatment of developmental dysplasia of the hip1-10. Most authors have agreed that an alteration of the blood supply to the femoral head causes this complication6,11-15. The reported frequency of osteonecrosis has varied enormously, from 0% to 73%8,16-18. Osteonecrosis can lead to deformities of the femoral head, neck, and greater trochanter as well as poor acetabular development. Cooperman et al.3 showed that loss of femoral head sphericity, persistent lateral and proximal subluxation, irregularity of the medial part of the femoral head, and acetabular dysplasia are related to premature painful degenerative osteoarthritis. Several other studies have corroborated these findings3-6,8,10,19-22.
Several classifications of osteonecrosis have been described. In 1980, Kalamchi and MacEwen7 developed a classification system emphasizing the growth disturbances associated with various degrees of physeal arrest. In this system, Group 1 includes changes affecting the ossific nucleus, following which the head usually regains its spherical shape. Group 2 involves lateral physeal arrest, which might not be evident on early radiographs. This type of involvement can take up to 12.5 years (mean, nine years) to become evident radiographically and leads to coxa valga and coxa breva, progressive uncovering of the femoral head, and relative trochanteric overgrowth. Group 3 is characterized by central physeal damage, which leads to a short femoral neck and a functional coxa vara. In Group 4, there is total damage to the head and physis, leading to femoral head irregularity, flattening, and coxa magna. Using this system, Kalamchi and MacEwen were unable to classify twenty-five of the 119 hips in their original report.
Bucholz and Ogden2 described a classification system similar to that of Kalamchi and MacEwen. Types 1 and 2 are similar to Groups 1 and 2 of the Kalamchi and MacEwen system, Type 3 involves the entire proximal part of the femur, and Type 4 affects the medial part of the epiphysis and the subjacent metaphysis. Thomas et al.23 later classified Type 2 into two subtypes, with Type 2B having additional central involvement of the femoral head. O'Brien et al.24,25 discussed the importance of growth disturbance lines in predicting future femoral development. However, Kim et al.26 noted that hips that demonstrated normal growth lines may still experience lateral physeal growth disturbance. The intensity of these lines in their study was variable and was potentially obscured by the degree of femoral rotation on the radiograph.
The wide variation in the reported frequency of osteonecrosis following the treatment of developmental dysplasia of the hip could be related in part to different treatment methods and the age of the patients at the time of treatment. In addition, Westin et al.9 believed that it also could be due to the lack of definition of terms. Thomas et al.22 concluded that there was some association between the reported frequency in a given series and the rigor with which the diagnosis had been sought. In addition, interobserver variability for the various classification systems in use has to be taken into account when comparing outcomes. Omeroglu et al.27 found that the interobserver agreement was only fair for the classification of osteonecrosis type with use of the Kalamchi and MacEwen system (kappa = 0.66).
The classification systems that have been described to date are observations and descriptions of the anatomical sequelae of osteonecrosis at maturity. None of the classification systems can give an early prediction of the outcome before maturity, and, because of the descriptive nature of the parameters used, they are subject to substantial interobserver variation. The objective of the present study was to identify a quantitative method of predicting the consequences of osteonecrosis soon after an index treatment episode so that surgeons can gain useful early feedback regarding the rate of osteonecrosis following treatment.
The purposes of the present study were (1) to develop an objective quantitative index that correlates with femoral head growth disturbance and loss of sphericity secondary to osteonecrosis, (2) to measure the intraobserver and interobserver variability of this index, and (3) to determine if the index on early radiographs can predict the eventual Kalamchi and MacEwen grade.
The medical records and radiographs of children managed with closed reduction for the treatment of developmental dysplasia of the hip at our institution between August 1997 and June 2001 were reviewed. Forty-seven patients were included in the present study. All patients were managed with arthrography, closed reduction, and spica immobilization with the hips in 100° to 110° of flexion, 40° to 60° of abduction (within the safe zone of Ramsey), and neutral rotation. A computed tomographic scan was performed to ensure that the reduction was maintained after the spica cast had been applied.
Twenty other patients who were managed during the same period were excluded from the study as they subsequently required open reduction or pelvic osteotomy following the initial closed reduction, and one patient was excluded because of loss to follow-up at four years. None of the patients had neuromuscular disease, arthrogryposis, or any teratological conditions. Seven patients were male, and forty were female. The left hip was affected in twenty-nine patients, the right hip was affected in fifteen, and both hips were affected in three. The patients underwent arthrography from August 1997 to June 2001, at which time the mean age was forty-two weeks. The patients were managed with a hip spica for an average of four months (range, three to six months). Thirty-two hips underwent percutaneous adductor tenotomy at the time of closed reduction. Following successful treatment with the hip spica, the patients were followed both clinically and radiographically. The average age at the time of the latest follow-up was 8.6 years (range, six to twelve years). Both the affected and the unaffected hip were used in the analysis, so that the number of hips at risk in the present study was ninety-four.
The predictive measure of osteonecrosis used in the present study is the epiphyseal index, which is a ratio of the maximum height to the maximum width of the ossified epiphysis. This index was originally described by Eyre-Brooke in 193628. Tönnis29 described a similar index whereby the maximum width of the ossified epiphysis is measured and the height is measured along a perpendicular line drawn from the midpoint of the transverse width. This method seemed more difficult to use in clinical practice, and hence the Eyre-Brooke method was adopted for the present study.
The epiphyseal index was measured on two anteroposterior radiographs of the pelvis that were made twelve to eighteen months after the closed reduction and on the latest follow-up radiograph. The height-to-width index was calculated by dividing the maximum height by the maximum width, both measured in millimeters. The radiographs were measured by two of the authors, a pediatric orthopaedic fellow and a senior resident (A.F. and J.A.C.), and the interobserver agreement was calculated. In addition, the radiographs were measured twice by one observer (J.A.C.) for the determination of intraobserver agreement.
The acetabular index of Hilgenreiner29,30 and the center-edge angle of Wiberg31 were measured to assess acetabular development. The femoral head was assessed for sphericity with use of Mose rings4. The radiographs from the twelve to eighteen-month period following the arthrogram and the latest radiographs were scored for osteonecrosis with use of the Kalamchi and MacEwen system by the two senior authors (C.E.B. and A.B.), and the interobserver agreement was calculated.
Index Measurement
For determination of the index, a line was first drawn in the orientation of the physis (Fig. 1). When the physis was undulating as seen on radiographs made when the patient was more mature, the physeal line was drawn from the medial to the lateral point of the physis. The width was measured between the most medial and most lateral points of the ossified epiphysis along a line parallel to the physeal line. The height was measured between the most superior and inferior points of the ossified epiphysis along a line perpendicular to the physeal line. The inferior margin of the epiphysis was usually easy to identify, but occasionally it cast two radiographic shadows when the physis was not in the plane of the radiograph beam. On these occasions, the inferior margin was measured from the point between the two radiographic shadows.
Statistical Analysis
Means were used for the index, and the highest Kalamchi and MacEwen grade was used whenever there was disagreement between the two observers. Interobserver and intraobserver agreement was analyzed with use of the kappa coefficient for categorical data and the intraclass correlation coefficient for continuous data. An intraclass correlation coefficient value of >0.75 was considered to be adequate. Receiver operating characteristic curves measured the predictive value of the index and were used to determine the optimum cutoff values. The cutoff values for the index were chosen to produce the highest value for sensitivity and specificity. Continuous data were tested for normality of distribution with use of the Shapiro-Wilk W test before use of the unpaired t test, with the level of significance set at p = 0.05.
Source of Funding
No external funding was received for the present study.
The interobserver agreement when the Kalamchi and MacEwen classification system was used to grade the hips on the pelvic radiograph made twelve to eighteen months after arthrography and the latest available radiograph was assessed with use of the kappa coefficient. The agreement is summarized in Table I.
Construct Validity
The relationship of the index to the presence or absence of osteonecrosis was an observation over the years by a senior author (C.E.B.). To assess construct validity, epiphyseal index data from the side affected by developmental dysplasia of the hip for all patients were used. The index was correlated with the presence or absence of osteonecrosis in each hip as subjectively agreed on by the two senior authors (C.E.B. and A.B.). The data are shown in Figure 2. All of the hips affected by osteonecrosis had an index of <0.5, with minimum overlap with the hips unaffected by osteonecrosis.
Repeatability
The interobserver and intraobserver agreement values for width, height, and the index are shown in Table II. The results indicate good intraobserver and interobserver repeatability.
Five femoral heads were nonspherical on the latest follow-up radiograph. A nonspherical femoral head was defined as one with an ovoid, mushroom, or umbrella shape that fell outside a 2-mm concentric circle on the anteroposterior radiograph. There was no significant difference between the center-edge angle of the nonspherical heads and the spherical heads (mean, 21.4° compared with 22.3°; p = 0.77). There was also no significant difference between the acetabular index of the nonspherical heads and that of the spherical heads (mean, 20.2 compared with 18.2; p = 0.41).
The receiver operating characteristic curves were drawn to establish cutoff points for the measured epiphyseal index. This was done postoperatively from initial twelve to eighteen-month radiographs to predict the final osteonecrosis outcome for the individual patient (Table III). Similarly, this was also done retrospectively on the basis of the latest available radiograph to quantify the development of early osteonecrosis according to the final epiphyseal index (Table IV). A number of the receiver operating characteristic curves that were produced to establish these epiphyseal index cutoff values are presented in the Appendix.
All but one of the hips that were not affected by osteonecrosis had an index value (as measured on the latest radiograph) of >0.38. Hence, an index of >0.38 had a negative predictive value (indicating the probability that a hip with an epiphyseal index above the given cutoff may not have development of the condition) of 0.98 for a history of osteonecrosis. Similarly, an index cutoff of 0.41 could reliably distinguish between the hips that had development of osteonecrosis at some stage following treatment and those that did not. An index of >0.41 had a negative predictive value of 0.98. Even hips that had recovered from osteonecrosis, as judged by the senior authors (according to the Kalamchi and MacEwen criteria) on the latest radiograph, had a persistently low index on the latest radiograph. The index (as measured on the latest radiograph) for hips that had osteonecrosis that subsequently recovered was compared with the index for those that never had evidence of osteonecrosis. The difference was significant (p = 0.0003, 95% confidence interval for the difference between means = 0.03 to 0.11).
Effect of Time
The mean age of the patients at the time of the latest follow-up was 8.6 years (range, six to twelve years). To analyze the effect of time on the hips, patients with more than six years of follow-up were further studied. The radiographs that were made when the patients were six years of age were screened for sphericity of the femoral head. Subsequent radiographs were analyzed similarly. All of the femoral heads had become aspherical by the age of six years and remained aspherical throughout the rest of the follow-up period. No new hips became aspherical after the age of six years.
Stability of the Index
The index was found to be stable over time. The mean fractional difference between the twelve to eighteen-month ratio and the latest follow-up ratio (at a mean age of 8.6 years) was 0.12 ± 0.036.
The epiphyseal height-to-width index gives a continuous outcome measurement of the deformity of the femoral head affected by osteonecrosis following the treatment of developmental dysplasia of the hip. This index is a reproducible, reliable measurement that seems to reflect the severity of growth disturbance following osteonecrosis. In the present study, hips that had an abnormal epiphyseal index and were aspherical at the time of early evaluation remained so until the latest follow-up at a mean age of 8.6 years. The index is therefore a predictor of intermediate-term asphericity, and it seems likely that it will predict longer-term asphericity and consequently a poor long-term outcome. The index should minimize the influence of interobserver variability arising from the use of present classification systems for the measurement of the frequency and severity of osteonecrosis following the treatment of developmental dysplasia of the hip. It should enable a more objective comparison of the frequency and severity of osteonecrosis between series of patients managed with different means and at different centers. The index also might make it possible to establish a prognosis for the individual patient who is affected by osteonecrosis at an early stage. The index therefore may be a useful research tool for the assessment of treatment in the early stages of osteonecrosis in addition to allowing surgeons to assess the magnitude of complications in patients under their care and to instigate changes in their practice to explore ways to minimize this risk.
The height-to-width index is a two-dimensional measurement of the secondary ossific nucleus of the proximal part of the femur. The index could have been more accurate if three-dimensional measurements from magnetic resonance imaging or computed tomography were possible. Several authors have relied on radiographic changes of the femoral head to predict subsequent deformity2,7,32. Salter et al.14, Gage and Winter6, and Berkeley et al.17 discussed the signs of total and partial osteonecrosis seen initially in the femoral head but did not correlate the early changes with the future development of the affected hip. Others7,25,33 have argued that the later development of the proximal part of the femur is influenced mostly by damage to the proximal femoral epiphysis. This is particularly true for Kalamchi and MacEwen Group-2 and Bucholz and Ogden Type-2 osteonecrosis, in which femoral deformities frequently become apparent only toward maturity. It has been hypothesized that these changes are a consequence of the damage to the lateral part of the physis, which may not be manifest before the adolescent growth spurt26; however, this notion remains mostly an observation and has never been scientifically proven, to our knowledge.
To our knowledge, the present study is the first study that has focused mainly on the accurate measurement of the dimensions of the ossific nucleus rather than its shape or pattern of fragmentation. A recent similar study of the deformity index of the epiphysis in patients with Legg-Calvé-Perthes disease demonstrated good validity, repeatability, and discriminatory ability34. The frequently used nonobjective descriptions such as "woolly" or "crescentic" epiphysis, medial bowing, lateral tilt, and premature physeal closure are prone to different interpretations of presence and severity. In contrast, by measuring the size of the ossific nucleus, interobserver variability is substantially diminished. Disruption of physeal growth seems to explain the various deformities of the proximal part of the femur, although the often subtle to nonapparent changes to the physis cannot be relied on to predict outcome at an early stage. This index appears to reflect the severity of the ischemic insult sustained by the proximal part of the femur, whether this affects mostly the ossific epiphysis or the physis.
The present study showed that the index is an accurate early predictor of subsequent femoral head asphericity. The resultant asphericity of the femoral heads at the time of the latest follow-up in the present study almost certainly was secondary to osteonecrosis rather than inadequate acetabular cover, given that there was no difference in the acetabular index and the center-edge angle between the nonspherical heads and the rest. Asphericity is probably the worst outcome following osteonecrosis as it is not amenable to surgical correction in the same way that other deformities such as trochanteric overgrowth, coxa vara, and coxa valga are, and it is associated with the early onset of osteoarthritis. Thus, it can be argued that the index is a potential predictor of the worst outcome of osteonecrosis.
It is interesting to see that the measured index of hips that had early osteonecrosis and subsequently recovered was still lower than that of hips that never had evidence of osteonecrosis. This could make it possible to retrospectively identify hips that had osteonecrosis during the first years following treatment.
The limitations of the present study arose from the duration of follow-up, given that not all patients had reached skeletal maturity by the time of the latest follow-up. Several authors have pointed out that Type-2 changes may not develop until the age of nine or ten years7,22,35-37. Consequently, there could be hips in this group in which Grade-2 changes have yet to develop as the mean age at the time of the latest follow-up was 8.6 years.
Other secondary parameters of the femoral shape were measured in the present study, including neck-shaft angle, neck width, and length and articulotrochanteric distance. There was a concern that there was substantial error in these measured parameters arising from variable femoral rotation on the projected radiographic image. The data that were gathered were found to be unsuitable for any additional statistical analysis. Computed tomography or magnetic resonance imaging of the hips would have enabled more accurate measurements of these parameters, but these are not routinely available or indicated.
In conclusion, the epiphyseal height-to-width index seems to be a simple and accurate measurement of the severity of growth disturbance following osteonecrosis of the femoral head after treatment for developmental dysplasia of the hip. It is predictive of asphericity at the time of intermediate-term outcome (mean, 8.6 years) and seems likely to predict asphericity at maturity, with the eventual development of adverse outcomes. It is a valid measurement with excellent intraobserver and interobserver repeatability. It offers the potential to compare outcomes following the treatment of developmental dysplasia of the hip before maturity is reached and may facilitate comparison between studies with a quantifiable index that is relatively free of operator interpretation and therefore subject to less intraobserver and interobserver error than are the traditional methods in current usage. It has the potential to alert practitioners to growth disturbance within a few years after treatment, which will facilitate practice audit and the early identification of growth disturbance. Only then can unusually high rates of osteonecrosis be identified and potential causes and solutions considered.
Four receiver operating characteristic curves used to establish the epiphyseal index cutoff values are available with the electronic versions of this article, on our web site at (go to the article citation and click on "Supplementary Material") and on our quarterly CD/DVD (call our subscription department, at 781-449-9780, to order the CD or DVD). 
Brougham DI, Broughton NS, Cole WG, Menelaus MB. Avascular necrosis following closed reduction of congenital dislocation of the hip. Review of influencing factors and long-term follow-up. J Bone Joint Surg Am.1990;72:557-62.72557
1990
[PubMed]
Bucholz RW, Ogden JA. Patterns of ischemic necrosis of the proximal femur in non-operatively treated congenital hip disease. In: The hip. Proceedings of the Sixth Open Scientific Meeting of the Hip Society. St. Louis: CV Mosby; 1978. p 43-63.
1978
Cooperman DR, Wallensten R, Stulberg SD. Post-reduction avascular necrosis in congenital dislocation of the hip. J Bone Joint Surg Am.1980;62:247-58.62247
1980
Cooperman DR, Wallensten R, Stulberg SD. Acetabular dysplasia in the adult. Clin Orthop Relat Res.1983;175:79-85.17579
1983
Fisher RL, Cary JM. Avascular necrosis complicating congenital dislocation of the hip. Course, prognosis, and orthopaedic management. Internat Orthop.1978;2:229-40.2229
1978
[CrossRef]
Gage JR, Winter RB. Avascular necrosis of the capital femoral epiphysis as a complication of closed reduction of congenital dislocation of the hip. A critical review of twenty years' experience at Gillette Children's Hospital. J Bone Joint Surg Am.1972;54:373-88.54373
1972
Kalamchi A, MacEwen GD. Avascular necrosis following treatment of congenital dislocation of the hip. J Bone Joint Surg Am.1980;62:876-88.62876
1980
Malvitz TA, Weinstein SL. Closed reduction for congenital dysplasia of the hip. Functional and radiographic results after an average of thirty years. J Bone Joint Surg Am.1994;76:1777-92.761777
1994
Westin GW, Ilfeld FW, Provost J. Total avascular necrosis of the capital femoral epiphysis in congenital dislocated hips. Clin Orthop Relat Res.1976;119:93-8.11993
1976
Keret D, MacEwen GD. Growth disturbance of the proximal part of the femur after treatment for congenital dislocation of the hip. J Bone Joint Surg Am.1991;73:410-23.73410
1991
Allen RP. Ischemic necrosis following treatment of hip "dysplasia". JAMA.1962;180:497-9.180497
1962
Gore DR. Iatrogenic avascular necrosis of the hip in young children. A review of six cases. J Bone Joint Surg Am.1974;56:493-502.56493
1974
Ponseti IV, Frigerio ER. Results of treatment of congenital dislocation of the hip. J Bone Joint Surg Am.1959;41:823-46.41823
1959
Salter RB, Kostuik J, Dallas S. Avascular necrosis of the femoral head as a complication of treatment for congenital dislocation of the hip in young children: a clinical and experimental investigation. Can J Surg.1969;12:44-61.1244
1969
Weiner DS, Hoyt WA Jr, O'dell HW. Congenital dislocation of the hip. The relationship of premanipulation traction and age to avascular necrosis of the femoral head. J Bone Joint Surg Am.1977;59:306-11.59306
1977
Crego CH Jr, Schwartzmann JR. Follow-up study of the early treatment of congenital displacement of the hip. J Bone Joint Surg Am.1948;30:428-42.30428
1948
Berkeley ME, Dickson JH, Cain TE, Donovan MM. Surgical therapy for congenital dislocation of the hip in patients who are twelve to thirty-six months old. J Bone Joint Surg Am.1984;66:412-20.66412
1984
Kalamchi A, Schmidt TL, MacEwen GD. Congenital dislocation of the hip. Open reduction by the medial approach. Clin Orthop Relat Res.1982;169:127-32.169127
1982
Gibson PH, Benson MK. Congenital dislocation of the hip. Review at maturity of 147 hips treated by excision of the limbus and derotation osteotomy. J Bone Joint Surg Br.1982;64:169-75.64169
1982
Harris NH, Lloyd-Roberts GC, Gallien R. Acetabular development in congenital dislocation of the hip. With special reference to the indications for acetabuloplasty and pelvic or femoral realignment osteotomy. J Bone Joint Surg Br.1975;57:46-52.5746
1975
Stulberg SD, Harris WH. Acetabular dysplasia and development of osteoarthritis of the hip. In: Harris WH, editor. The hip. Proceedings of the Second Open Scientific Meeting of the Hip Society. St. Louis: CV Mosby; 1974. p 82-93.
1974
Thomas IH, Dunin AJ, Cole WG, Menelaus MB. Avascular necrosis after open reduction for congenital dislocation of the hip: analysis of causative factors and natural history. J Pediatr Orthop.1989;9:525-31.9525
1989
[CrossRef]
Thomas CL, Gage JR, Ogden JA. Treatment concepts for proximal femoral ischemic necrosis complicating congenital hip disease. J Bone Joint Surg Am.1982;64:817-28.64817
1982
O'Brien T. Growth-disturbance lines in congenital dislocation of the hip. J Bone Joint Surg Am.1985;67:626-32.67626
1985
O'Brien T, Millis MB, Griffin PP. The early identification and classification of growth disturbances of the proximal end of the femur. J Bone Joint Surg Am.1986;68:970-80.68970
1986
Kim HW, Morcuende JA, Dolan LA, Weinstein SL. Acetabular development in developmental dysplasia of the hip complicated by lateral growth disturbance of the capital femoral epiphysis. J Bone Joint Surg Am.2000;82:1692-700.821692
2000
Omeroglu H, Tumer Y, Bicimoglu A, Agus H. Intraobserver and interobserver reliability of Kalamchi and MacEwen's classification system for evaluation of avascular necrosis of the femoral head in developmental hip dysplasia. Bull Hosp Jt Dis.1999;58:194-6.58194
1999
Eyre-Brooke AL. Osteochondritis deformans coxae juvenilis or Perthes' Disease: the results of treatment by traction in recumbency. British J Surg.1936;24:166-82.24166
1936
[CrossRef]
Tönnis D. General radiography of the hip joint. In: Congenital dysplasia and dislocation of the hip in children and adults. New York: Springer; 1987. p 100-142.
1987
Tönnis D. Normal values of the hip joint for the evaluation of X-rays in children and adults. Clin Orthop Relat Res.1976;119:39-47.11939
1976
Wiberg G. Studies on dysplastic acetabula and congenital subluxation of the hip joint. With special reference to the complication of osteo-arthritis. Acta Chir Scand.1939;58(Suppl):5-135.585
1939
Siffert RS. Deformities caused by injuries to the growing hip. Hip.1981:246-63.246
1981
Katz JF. Legg-Calvé-Perthes disease. The role of distortion of normal growth mechanisms in the production of deformity. Clin Orthop Relat Res.1970;71:193-8.71193
1970
Nelson D, Zenios M, Ward K, Ramachandran M, Little DG. The deformity index as a predictor of final radiological outcome in Perthes' disease. J Bone Joint Surg Br.2007;89:1369-74.891369
2007
[CrossRef]
Campbell P, Tarlow SD. Lateral tethering of the proximal femoral physis complicating the treatment of congenital hip dysplasia. J Pediatr Orthop.1990;10:6-8.106
1990
Morcuende JA, Meyer MD, Dolan LA, Weinstein SL. Long-term outcome after open reduction through an anteromedial approach for congenital dislocation of the hip. J Bone Joint Surg Am.1997;79:810-7.79810
1997
Robinson HJ Jr, Shannon MA. Avascular necrosis in congenital hip dysplasia: the effect of treatment. J Pediatr Orthop.1989;9:293-303.9293
1989