We retrospectively reviewed the clinical records and radiographs of 51 selected patients with a diagnosis of femoroacetabular impingement who had failed nonoperative management and were treated with an open surgical dislocation of the hip [6
] from May 2000 to September 2003. Five patients had inadequate preoperative or postoperative data for inclusion in the study group. This left 46 patients for evaluation. Two of these 46 patients had bilateral hip procedures, creating a study group of 48 hips. The indications for surgical intervention included skeletally mature patients with consistent clinical and radiographic findings of femoroacetabular impingement who had failed at least 1 month of conservative management including activity modification and NSAIDs. Exclusion criteria included skeletally immature patients and patients with advanced hip osteoarthritis evidenced by Tönnis grade 3 findings [22
]. The average age of the patients was 33 years (range, 18–51 years). There were 21 women and 25 men. Twenty-one left hips and 27 right hips underwent surgery. The minimum followup time for a postoperative questionnaire (Appendix 1) was 6 months (mean, 38 months; range, 6–67 months); all patients in the study group completed the questionnaire. The minimum radiographic followup was 1 month (mean, 12 months; range, 1–60 months); this was compromised by 16 patients having followup visits in the patient’s state of residence. Prior to the review, the study received institutional ethics review board approval.
Onset of pain was initially activity-related, but in 42 patients became constant. All patients localized pain to the groin area and had a positive impingement test [12
]. The average duration of symptoms between onset and surgery was 42 months (range, 1–132 months). Thirty-five of 46 were taking NSAIDs for pain control. Eleven of the 46 were taking narcotics. Three of the 46 had previous arthroscopic hip surgery. Eleven of the remaining had other previous hip procedures (open relocation, irrigation and débridement, ORIF of femoral neck fracture, iliopsoas tenotomy, IT band tenotomy, trochanteric bursectomy, benign tumor resection, periacetabular osteotomy, intertrochanteric osteotomy, and removal of a dynamic hip screw from a previous femoral neck fracture).
Radiographic evaluation included an AP pelvis, AP hip, faux-profile view, and cross-table lateral view of the involved hip. AP pelvis radiographic adequacy was evaluated based on previous cadaveric radiographic positioning studies [21
]. To clarify, the pelvic tilt was documented by the relative position of the sacrococcygeal joint and the pubic symphysis compared to average population values. To ensure we appropriately assessed retroversion, we documented and ensured proper pelvic rotation by comparing the position of the coccyx with respect to the symphysis, as well as the symmetry of the obturator foramina and iliac wings. One of us not associated with the surgery (MG) documented from the AP pelvis the following preoperative findings: acetabular retroversion, coxa profunda, protrusio, pistol grip deformity, aspherical femoral head, coxa vara, coxa valga, herniation pits, Tönnis OA classification [22
], lateral center-edge angle, acetabular index, anterior center-edge angle, and the amount of cartilage space of the superior joint measured in millimeters above the center of the femoral head. The acetabulum was defined as retroverted if the anterior border crossed the posterior border medial to the lateral edge of the acetabulum [20
]. Coxa profunda was defined as the floor of the fossa acetabuli contacting the ilioischial line, when associated with a center edge angle of greater than 35 degrees [2
]. Protrusio was defined as the medial aspect of the femoral head contacting the ilioischial line with an associated center-edge angle of greater than 40°. A pistol grip deformity of the femoral head was defined as an extension of the lateral epiphysis down to the base of the femoral neck [9
]. An aspherical head was defined as an extension of the lateral epiphysis onto the cephalad neck creating a nonspherical head. This was corroborated by evaluating the cross-table lateral view of the hip [4
]. Coxa vara was defined by a neck-shaft angle less than 125° in association with the tip of the trochanter being greater than or equal to 5 mm above the center of the femoral head. Coxa valga was defined by a neck-shaft angle greater than 135° in association with the tip of the trochanter being greater than or equal to 5 mm below the center of the femoral head.
All hips on plain radiographs had insufficient femoral head-neck offset. The offset was considered inadequate if a the epiphysis extended outside of a circle drawn around the femoral head [4
]. This was often notable on both the cross-table lateral and the anteroposterior hip view. Thirty-six of 48 hips had pure cam impingement, defined by primary proximal femoral pathology [10
] (Fig. A–B). The remaining 12 had a mixed deformity with signs of primary proximal femoral and acetabular pathology [8
]. There were no pure pincer impingement cases in this series (Table ).
Fig. 1A–B (A) This preoperative AP pelvis of a patient in the series reveals radiographic signs of bilateral cam impingement. (B) This postoperative AP pelvis was taken 14 months after surgery on the right side. Bilateral surgical hardware is noted, reflecting (more ...)
We performed a preoperative gadolinium-enhanced arthrographic MRI on all but one hip (47 of 48). Almost all were completed and read by the same musculoskeletal radiologist (TL) with specialized training in arthrographic MRI. Documented pathology included loss of head/neck offset, the type and location of labral damage (all labral damage located between 10 o’clock and 2 o’clock except ossification that often extended posteroinferiorly), the type of acetabular and femoral head cartilage damage, and the presence of herniation pits (Table ).
The common goal of all procedures was restoring impingement-free motion. All patients underwent a surgical dislocation through a modification of the Gibson approach and osteochondroplasty of the head-neck junction [8
]. The other procedures were varied based on preoperative evaluation and intraoperative findings. Acetabuloplasty (trimming of the acetabular rim) was performed to decrease overcoverage. The labrum was never excised in this series. The portion of the labrum that was damaged was débrided, with the remainder undergoing repair to the acetabular rim. Rim fractures were excised or repaired based on the degree of coverage, with the goal to prevent undercoverage. Abrasion chondroplasty was performed after cartilage flap débridement when necessary in an attempt to initiate fibrocartilage formation. Additional procedures in the series included relative femoral neck lengthening, intertrochanteric osteotomy for coxa vara, lateralization of the greater trochanter, osteochondral grafting of femoral head cartilage defects, curtain osteophyte resection, sciatic neurolysis, and loose body removal (Table ).
Labral and cartilage damage were localized by translating left and right hips into a right hip clock face system with 12 o’clock located superiorly, 6 o’clock located inferiorly, 3 o’clock located anteriorly, and 9 o’clock located posteriorly. Intraoperative data correlated with preoperative MRI findings.
A first-generation cephalosporin was given preoperatively and continued until the drain was removed. Drains were removed once less than 30 mL was noted over a 24-hour period. Thromboembolic prophylaxis consisted of a low-molecular-weight heparin while in the hospital. Upon discharge, a buffered aspirin was given daily for 6 weeks.
Postoperative physical therapy consisted of a continuous passive motion machine while in the hospital. This was initially set at 30° and advanced 5° per day until discharge. Active assisted range of motion began immediately and advanced to resistance exercises once the greater trochanter showed radiographic signs of union. Postoperative weight-bearing protocol included weight-of-leg weight bearing for 6 weeks using crutches or a walker. If radiographs revealed greater trochanteric healing at 6 weeks, single crutch progressive weight bearing began and continued for 2 weeks. Patients were allowed full weight bearing without an assistive device after this time.
Patient clinical outcome was assessed using the Merle D’Aubigné-Postel score [14
]. As noted previously, 16 patients had followup out of state. In addition to receiving the followup clinic notes from the local orthopaedic surgeons, we (MG) contacted these patients and/or their doctors by phone or mailed them questionnaires (Appendix 1). This combination allowed us to get Merle D’Aubigné-Postel scores on the out-of-state patients. The standard followup protocol was 2 weeks, 6 weeks, 3 months, 6 months, 1 year, and yearly thereafter.
Radiographs were taken at 6 weeks, 12 weeks, and yearly thereafter. An attempt was made to obtain radiographs from all out-of-state patients who were followed by other orthopaedic surgeons. We were able to locate followup radiographs on 38 of the 46 patients. The same observer evaluated the followup radiographs for the same preoperative variables as well as the Brooker classification of heterotopic ossification [3
]. We assessed offset by noting whether or not there was a clear difference in the head-neck junction (ie, an extended lateral epiphysis).
Outcome scores were first evaluated for the normality of distribution of the data. We did not find a normal distribution with the Merle d’Aubigné-Postel scores, and therefore used the Wilcoxon signed rank test (SPSS Incorporated, Chicago, IL).