THA was described as “The operation of the century” [13]. Despite success of the procedure for patients with idiopathic or traumatic degenerative arthritis, hip replacement for congenital dislocation or subluxation was controversial as late as the 1970s [3, 6, 8, 13]. Different anatomic characteristics observed on AP radiographs of the pelvis historically were used to assess the degree of acetabular dysplasia. These included acetabular angle of Sharp, center edge angle of Wiberg, acetabular index of depth to width as described by Heyman and Herndon [11], acetabular roof obliquity described by Massie and Howorth [14], femoral head extrusion index, lateral subluxation, and peak to edge distance [15]. Crowe et al. described a relatively simple method to determine degree of hip dysplasia in 1979 [4].
The controversy surrounding the indications and techniques for hip arthroplasty in adult patients with developmental dysplasia of the hip highlighted the need for a system to gauge the degree of dysplasia. A fair comparison among the outcomes for different operative techniques was possible only with a classification system for preoperative assessment of the degree of dysplasia.
An ideal staging system should be practical, reproducible, and provide prognostic estimation. As the normal anatomy of the hip is distorted in a patient with developmental hip dysplasia, a practical staging system should be based on readily identifiable landmarks that will be affected minimally by the disorder to ensure measurements can be made in a consistent and reliable manner for patients with diseases across the spectrum.
Crowe’s classification is based on three easily identifiable anatomic landmarks: (1) the height of the pelvis; (2) the medial head-neck junction in the affected hip; and (3) the inferior margin of the acetabulum (the teardrop).
The measurements are made on AP radiographs of the pelvis. The reference line is drawn joining the inferior margins of each teardrop. The medial head-neck junction is identified and its distance from the reference line is noted. In the absence of subluxation this distance is close to zero. This distance was described as a measure of the degree of dysplasia by Crowe et al. [4]. According to the original investigation, the height of the pelvis is measured as the vertical distance from the highest point on the iliac crest to the inferior margin of the ischial tuberosity. The normal ratio of vertical diameter of the femoral head to the height of the pelvis was approximately 1:5 [4]. A hip was considered subluxated 50% or greater only if the medial head-neck junction was situated above the reference line by at least 10% of the measured height of the pelvis.
|
Group |
Description |
|---|---|
|
I |
Subluxation < 50% or proximal dislocation < 0.1% of the pelvic height |
|
II |
Subluxation 50%–75% or proximal dislocation of 0.1% to 0.15% of pelvic height |
|
III |
Subluxation 75%–100% or proximal dislocation of 0.15% to 0.20% of pelvic height |
|
IV |
Subluxation > 100% or proximal dislocation of > 0.20% of pelvic height |
|
Type |
Description |
Anatomic correlation as verified during surgery |
|---|---|---|
|
Dysplasia |
The femoral head is not dislocated out of the acetabulum despite the degree of subluxation |
Segmental deficiency of the superior wall, secondary shallowness attributable to fossa-covering osteophytes |
|
Low dislocation |
The femoral head is dislocated and articulates with a false acetabulum which partially covers the true acetabulum to a varying degree |
Complete absence of the superior, anterior, and posterior segmental deficiency, narrow opening, and inadequate depth of the true acetabulum |
|
High dislocation |
The femoral head is dislocated and has migrated superiorly and posteriorly with no articulation with any part of the true acetabulum |
Segmental deficiency of the entire acetabulum with narrow opening, inadequate depth, excessive anteversion, abnormal distribution of bone stock, mainly located anteroposteriorly in relation to the true acetabulum |
In addition to these two commonly used classification systems, other classification systems have been proposed [7, 12]. Eftekhar classified dysplastic hips in four types: A through D [7]. Type A is a slightly elongated dysplastic acetabulum accommodating a flattened mushroom-shaped femoral head, Type B is intermediate dislocation, Type C is high dislocation, and Type D is an old, unreduced dislocation. Kerboul et al. [12] classified dysplastic hips in three categories: Type A, anterior dislocation; Type B, intermediate dislocation; and Type C, posterior dislocation.
High levels of interobserver and intraobserver reliability have been reported using Crowe’s classification. Yiannakopoulos et al. reported kappa coefficients as high as 0.92 and 0.95 for interobserver and intraobserver reliabilities, respectively [16], and Decking et al. reported kappa coefficients of 0.82 and 0.86 for interobserver and intraobserver reliabilities, respectively [5]. These two investigations compared the classifications of Crowe et al. and Hartofilakidis et al. for dysplasia for reliability. Although Crowe’s classification had higher kappa coefficients for reliability in both investigations, the difference was not clinically meaningful. Both investigations concluded that both classification systems are reliable.
Others have examined the reliability of the classification systems of Eftekhar and Kerboul et al. [1]. Although both classification systems showed sufficient reliability, they were less reliable when compared with Crowe’s classification system. Brunner et al. recommended using the classification systems of Crowe et al. or Hartofilakidis et al. to assess the severity of hip dysplasia [1].
Validity for classification systems refers to the inherent ability of the proposed system to measure what it is supposed to measure. Validity has not been formally investigated for Crowe’s classification, as it measures radiographic findings pathognomonic and unique to dysplastic hips. However, the influence of Crowe’s rating on the outcome of THA for patients with congenital dysplasia was investigated by Cameron et al. [2]. They concluded that the complication rate increased with an increase in Crowe’s rating, validating its prognostic significance.
The main limitation of Crowe’s classification system is the lack of correlation of the rating system with anatomy as seen during surgical exposure. This arises because Crowe’s classification is a 2-D assessment of a 3-D problem. This limitation was addressed by Hartofilakidis et al. with their classification system (Table 2) [10]. They used 3-D CT to investigate four parameters of acetabular anatomy: (1) segmental deficiencies in the rim of the acetabulum superiorly, anteriorly, and posteriorly; (2) the amount of anteversion and the depth and opening, ie, the distance between the anterior and posterior rim of the true acetabulum; (3) the amount of acetabular bone stock superiorly, anteriorly, and posteriorly; and (4) the presence of osteophytes in the area of the true and false acetabula. Anatomic correlations are shown in Table 2.
Crowe’s classification system for dysplastic hips is reliable, reproducible, and related to prognosis of patients with THA who have developmental hip dysplasia [2]. It is well accepted and widely used for classification by orthopaedic adult reconstruction surgeons worldwide. However, the classification system of Hartofilakidis et al. is similarly reliable and a combination of both classification systems probably would be most useful for clinical decision-making.
