|Home | About | Journals | Submit | Contact Us | Français|
Combined with clinical examination and MRI, radiographs have been mainstays in the management femoroacetabular impingement (FAI). Because hip morphology often portends intraoperative damage, radiographic features should inform surgical management.
We determined (1) the radiographic features of the various hip morphologies; (2) the prevalence of radiographic coxa profunda in each group; (3) the radiographic differences between hips with and without coxa profunda; and (4) its sensitivity and specificity as a measure of global acetabular overcoverage.
We reviewed preoperative radiographs and operative notes of 144 hips that underwent surgical dislocation and correction for FAI between August 2002 and February 2011. Hips were divided into four FAI subtypes by radiographic analysis (cam, global overcoverage, retroversion, and combined) and three subtypes (cam, pincer, or combined) by intraarticular pathology. Standard radiographic measurements were performed, and we introduce a novel measurement that assesses femoral head coverage.
We found differences in median Angle of Sharp, femoral head-neck angle, and median roof length (and its subset) among the FAI morphologies. The prevalence of radiographic coxa profunda was 48% in cam hips, 85% in global overcoverage hips, 66% in retroverted hips, and 32% in combined hips. The sensitivity and specificity of radiographic coxa profunda as a measure of global overcoverage was 75% (95% CI, 0.62–0.85) and 62% (95% CI, 0.51–0.73), respectively.
We found major differences in radiographic measurements between FAI morphologies. Radiographic coxa profunda was poorly specific for global overcoverage. Measurement of roof length and ratio should be used to determine the morphology of the impinging hip.
Level II, prognostic study. See Guidelines for Authors for a complete description of levels of evidence.
Femoroacetabular impingement (FAI) is a mechanism that leads to hip pain and has been associated with osteoarthritis [7, 8, 12, 16, 18]. The morphologic features associated with the deformity have helped define the type of impingement and as such have been divided into cam impingement when the deformity is on the femoral side; pincer impingement when there is sectorial or global acetabular overcoverage; and combined impingement when there is a combination of both acetabular and femoral deformity. Both cam and pincer impingement have been associated with intraarticular cartilage and labral damage, but the type and degree of damage vary among impingement types [2, 7, 8, 11, 12, 16, 18]. Typically, cam impingement leads to debonding and sheering of the articular cartilage, initially with preservation of the labrum, whereas pincer-type morphologies lead to rim chondromalacia with labral thinning and metaplasia. It is through a careful radiographic examination that the surgeon will determine degrees of intraarticular damage and choose the best surgical approach to treat the impinging hip.
Pincer-type impingement is often hard to diagnose in the presence of seemingly normal radiographs. In recent years, the radiographic feature coxa profunda has been associated with global overcoverage but this relationship has been poorly defined. Radiographically, it is defined as an acetabular wall at or medial to the ischioilial line of Kohler’s irrespective of other morphologic characteristics . This radiographic feature is not an uncommon finding in the female pelvis  and does not seem to be associated in all hips with acetabular overcoverage because it can also be seen in some dysplastic hips. More recently, a lateral center-edge angle greater than 40° has also been associated with acetabular overcoverage  but this again is unlikely to be the sole indicator of a pincer-type morphology. A comprehensive analysis of the radiographic features associated with coxa profunda and its prevalence in the different hip morphologies is warranted to determine its usefulness in diagnosing FAI. Radiographs are the mainstay in diagnosing FAI and as such the radiographic features associated with differing hip morphologies require further study.
The aims of this study are therefore (1) to delineate the femoral and acetabular radiographic features associated with differing hip morphologies; (2) to report the prevalence of radiographic coxa profunda in each of these groups; (3) to report the differences in radiographic features between hips with and without radiographic coxa profunda; and (4) to report the sensitivity and specificity of coxa profunda as a measure associated with global acetabular overcoverage or pincer FAI.
This is an institutional review board-approved retrospective case series. We reviewed all 155 patients (178 hips) who underwent surgical hip dislocation (SHD) and correction for FAI between 2002 and February 2011. Cases were identified using a computerized database compiled by the Mayo Clinic Surgical Index. The indication for SHD in the presence of FAI was symptomatic impingement in patients with a structural anatomic abnormality leading to impingement. The contraindications for SHD included infection and patients with lesser degrees of structural abnormalities that could be treated arthroscopically. Only patients with primary FAI were included, excluding FAI occurring from slipped capital femoral epiphysis and Perthes disease. In addition, we excluded hips that had concomitant osteotomies about the femur and pelvis, leaving 144 hips that underwent a detailed radiographic analysis. There were 59 females (41%) and 85 males (59%) in this group with an average age of 29.5 years (range, 17–50 years). Four hips were lost to followup.
One of us (GB) reviewed all preoperative radiographs to determine radiographic criteria consistent with the deformity and to classify hips into different hip morphologies according to accepted criteria. For the AP pelvic radiograph, the patient lies supine with the legs 15° internally rotated to compensate for femoral antetorsion and provide better visualization of the lateral femoral head-neck junction . The central beam is directed at the midpoint between a line connecting both anterosuperior iliac spines and the superior border of the symphysis . The crosstable lateral view of the proximal femur is taken with the leg internally rotated and with the central beam directed to the inguinal fold . Measurements were obtained from standardized preoperative AP radiographs and included the lateral center-edge angle of Wiberg (LCE), Tönnis angle, and Angle of Sharp. We also obtained two novel measurements of the acetabular roof length, one lateral (Roof 1) and one medial (Roof 2) (Fig. 1). From these values, a sum (the total roof distance) and a ratio (total roof/Roof 2) were calculated. On the femoral side, we determined the sphericity of the femoral head and measured femoral head neck angle on the AP pelvic radiograph and the alpha angle on available lateral radiographs. According to accepted criteria of radiographic morphologies, hips were classified as cam, global overcoverage (pincer), retroverted , and combined (Table 1).
Of the 144 hips, 29 hips had cam morphology, 41 hips had pincer or global overcoverage morphology, 17 hips had a retroverted morphology, and 57 had a combined morphology.
Subsequently, hips were reclassified into three impingement morphologies with regard to their intraarticular cartilage and labral damage. These were classified by RJS into cam, pincer, and combined hip morphologies after review of surgical notes and intraoperative pictures. Cam morphology hips produced outside-in sheering and debonding of the affected cartilage with tearing of the chondrolabral junction. In these, hip rim resection was most often performed for labral reattachment or chondroplasty with or without microfracture of the remaining subchondral bone. In pincer-type morphologies, the damage to the labrum leads to a thinned, ossified labrum, whereas the cartilage damage is typically located in the periphery with little extension into the depth of the acetabulum. Combined morphologies characteristically have a combination of cartilage and labral findings. Using these criteria, 92 hips were classified as cam, 38 as pincer, and 14 as combined hip morphologies. Of note, in some instances, it was difficult to discern whether a hip should be placed into the combined group or the cam group, because in some cases, the intraarticular lesions had a predominance of cam-type damage (eg, debonding of cartilage).
The presence or not of radiographic coxa profunda was determined from AP pelvic radiographs. Of the 144 hips, 78 hips were identified as having a radiographic coxa profunda defined as the medial wall of the acetabulum lying medial to Kohler’s line. Radiographic measurements were then compared between hips with radiographic coxa profunda and those that did not to determine the difference between the two groups.
Statistical analysis focused primarily on comparisons between patient subgroups based on impingement morphology (a four-level group based on radiographic measurements and a three-level group based on intraarticular cartilage and labral damage). Within each grouping scheme, study outcomes including radiographic and operative findings were compared between patients with different impingement morphologies using analysis of variance for variables measured on a continuous scale and chi square tests for binary or categorical data. When the global tests from the analysis of variance models were statistically significant, further analysis was performed using a multiple comparisons procedure to identify which morphology subgroups were different from each other while controlling the overall Type I error rate. Significant global results for the categorical data were further analyzed by performing pairwise chi-square tests, and the resulting p values were adjusted using the Benjamini-Hochberg  procedure to control the false discovery rate. Using logistic regression, we attempted to find a cutoff value of roof length or ratio that could be used to predict global overcoverage using pincer impingement as stratified by intraoperative findings. Comparisons between patients classified as coxa profunda and those not classified as such were performed using two-sample t-tests for continuous variables and chi square tests for categorical data. The sensitivity and specificity of coxa profunda with respect to global acetabular coverage were reported with 95% binomial confidence intervals. All statistical tests were two-sided. All analysis was performed using SAS version 9.2 (SAS Institute Inc, Cary, NC) and R version 2.14.0 (http://www.R-project.org) .
Twenty-nine hips were classified per radiographic criteria as having cam impingement, 41 hips as global overcoverage, 17 hips as retroverted, and 57 as combined. We identified differences in multiple radiographic features among the groups (Table 2). Likewise, when subclassified according to intraoperative damage, we found differences in radiographic features among the three groups (Table 3). Hip morphology associated with global overcoverage or pincer impingement included a more valgus femoral neck angle, a shorter total roof length, and a shorter Roof 2 (distance from Kohler’s line to medial aspect of roof or sourcil) with no difference in Roof 1 (true sourcil). Ratios that included Roof 2 demonstrated differences among impingement types (Table 3). As a result of the overlap of ratios within each group, it was not possible to determine a specific cutoff value for the measurement of roof length and ratios. In addition, despite a linear increase in probability, the confidence intervals were wide, not allowing to draw any conclusive data to its use (Figs. 2,, 3 3).
The prevalence of radiographic coxa profunda was 48% in hips with cam morphology, 85% in hips with global overcoverage, 66% in hips with retroversion, and 32% in hips with combined morphology.
In total, 78 of 144 hips (55%) had radiographic features consistent with coxa profunda. Among these hips, the mean lateral center-edge angle was greater (p = 0.01) in hips without coxa profunda than in those with: 36.2° versus 32.6°, respectively. The mean alpha angle in patients with coxa profunda was lower (p = 0.002) when compared with noncoxa profunda hips (53.5 versus 62.7). The mean Tönnis angle was less (p = 0.001) in the group with coxa profunda (2.8° versus 5.7°), whereas the mean femoral neck angle was higher (p = 0.004) in patients with coxa profunda (130.2° versus 128.0°). Compared with the noncoxa profunda hip, the coxa profunda group had a lower mean total roof length (47.1 mm versus 55.7 mm, p = 0.001) and lower roof ratio (1.3 versus 1.9, p = 0.001) (Table 4).
The sensitivity and specificity of radiographic coxa profunda as a measure of global acetabular overcoverage was 75% (95% CI, 0.62–0.85) and 62% (95% CI, 0.51–0.73), respectively, when using radiographic criteria and 68% (95% CI, 0.51–0.82) and 50% (95% CI, 0.41–0.6) when using intraoperative findings.
The diagnosis of FAI is based on a clinical examination in symptomatic patients combined with radiographic findings supporting a structural deformity that leads to hip damage. Understanding radiographic features consistent with impingement morphologies will help the impingement surgeon determine the best surgical approach to treat the impinging hip. We therefore (1) delineated the femoral and acetabular radiographic features associated with differing hip morphologies; (2) determined the prevalence of radiographic coxa profunda in each of these groups; (3) identified differences in radiographic features between hips with and without radiographic coxa profunda; and (4) determined the sensitivity and specificity of coxa profunda as a measure associated with global acetabular overcoverage or pincer FAI.
There are limitations to the present study. First, there may be some selection bias present as a result of not including patients who did not undergo SHD for FAI. However, this bias is somewhat reduced including all patients undergoing SHD within the given timeframe. Second, radiographic measurements may vary with patient position and could be subject to interpretation error . The radiographic technique is standardized to accommodate for these errors. All measurements were performed digitally on an in-house templating program that adjusts for technique differences. Third, these measurements are clearly subject to error, and we did not assess intraobserver or interobserver reliability. Fourth, we defined the morphologies according to established radiographic criteria, but these criteria have not been universally accepted. Finally, we classified hips according to intraarticular chondrolabral damage as noted on operative reports and intraoperative pictures. This was especially challenging for hips with combined morphology, because cam lesions typically dominate the damage picture. This might contribute to the differences between these groups.
Our analysis shows there are clear differences in pelvic and femoral parameters among impingement morphologies. We found differences, for example, in Angle of Sharp and femoral neck angles in between morphologic subgroups. Because the Angle of Sharp is the degree of lateral opening, a larger angle was seen in retroverted sockets when compared with all other morphologic subtypes. Retroverted sockets do not appear to be characterized by lateral overcoverage. Hips with global overcoverage or pincer morphology had considerably more valgus femoral neck angles compared with all other morphologies. One could theorize that the valgus of the femoral neck may be adaptive in nature, helping with clearance of the trochanter with ROM because the structure is further away from the pelvis.
We introduced a new measure, the total roof length, and its lateral and medial components and found differences in total roof length, Roof 2, and two calculated ratios (total roof/Roof 2 and Roof 1/Roof 2). Shorter total roof lengths were seen in hips with overcoverage morphology. Interestingly, the actual length of the weightbearing sourcil was no different among the groups (Roof 1), but the total roof length was actually shorter. The differences in total roof length between the groups seems to lie mainly in the differences in the distance between the ilioischial line and the medial aspect of the roof or sourcil that is substantially shorter in hips with global overcoverage. The differences in roof ratio were primarily the result of differences in Roof 2. Both total roof/Roof 2 and Roof 1/Roof 2 ratios were smaller in hips with global overcoverage. The anatomic difference in hip center of rotation without changes to true roof or sourcil length could indeed be responsible for the problems seen in hips with global overcoverage. Unfortunately, we were unable to determine a cutoff point that could help determine radiographically pincer hips despite these statistical differences. This could be related to the overlap of intraoperative criteria consistent with combined hips, erroneous classification of impinging morphology, or simply to lack of number of hips to determine these radiographic differences.
Coxa profunda, defined as the medial wall of the acetabulum at or medial to Kohler’s line, is commonly seen in patients with and without FAI and was seen in 55% of hips undergoing FAI surgery in this series [4, 10, 16]. This common finding was also noted by Armbuster et al. . In 1978, these investigators evaluated over 300 AP radiographs of the normal adult hip and determined that coxa profunda as it is currently defined was present in over 70% of female radiographs. In their study, the average distance between the medial wall of the acetabulum and Kohler’s line was in fact 2 mm lateral in males and 1 mm medial in females . Despite its common occurrence, there are clear differences in radiographic parameters between hips that have this radiographic feature compared with those that do not. These differences in part relate to pincer-type morphologic features such as greater LCE angle, smaller Tönnis angle, or shorter total roof, but because of its prevalence in all impingement morphologies, it should not be used alone but combined with other measurements to describe global overcoverage morphology. In the present series, the specificity of the radiographic feature of coxa profunda in determining global overcoverage was 62%.
In 1971, Hooper and Jones  defined protrusio acetabuli at a center-edge angle > 40º with no mention of the medial acetabular wall. Kutty et al.  recently reported on the predictability of the center-edge angle in the assessment of pincer-type FAI. They found LCE angle ≥ 40º to be 84.2% sensitive and 100% specific for pincer-type FAI . There are however many hips that have pincer-type morphology and intraarticular damage consistent with pincer FAI that have an LCE angle less than 40°. It would seem that using LCE greater than 40° as a sole measurement to determine global overcoverage is too simplistic, because the pathomechanics of FAI are more complex than one simple, two-dimensional measurement. Accordingly, we describe the major differences between the various impingement morphologies.
As our understanding of FAI improves, it is important to look for objective measurements that can conveniently and reliably inform clinical decision-making. Although the measurements are imperfect and can vary somewhat with the rotation of the radiograph, developing objective criteria for global overcoverage and other impinging morphologies and then correlating the data with intraoperative findings and patient outcome will lay the foundation for more objective and reproducible approaches to preoperative decision-making. Our study demonstrates limitations of the currently accepted parameters used to determine pincer-type morphologies and presents novel measurements that may prove to be useful with further analysis.
We acknowledge Amy Hu MS, and Dirk Larson MS, for their statistical support.
One of the authors (RJS) certifies that he has or may receive payments or benefits, in any one year, an amount in excess of $10,000 from Biomet Inc (Warsaw, IN, USA). One of the authors (RTT) certifies has or may receive payments or benefits, in any one year, an amount in excess of $100,000 from DePuy Orthopaedics (Warsaw, IN, USA) and less than $10,000 from Mako Surgical (Fort Lauderdale, FL, USA) and Wright Medical (Arlington, TN, USA). No other authors have any conflicts.
All ICMJE Conflict of Interest Forms for authors and Clinical Orthopaedics and Related Research editors and board members are on file with the publication and can be viewed on request.
Each author certifies that his or her institution approved the human protocol for this investigation, that all investigations were conducted in conformity with ethical principles of research, and that informed consent for participation in the study was obtained.