Early results of surgery for femoroacetabular impingement in patients with osteonecrosis of femoral head

Purpose: Femoroacetabular impingement and its surgical treatment have not been described before in osteonecrosis of femoral head. We present here outcomes of 15 patients with femoroacetabular impingement secondary to osteonecrosis of femoral head. This results from partial collapse of femoral head, particularly in the anterosuperior region, secondary to osteonecrosis. With subsequent remodelling, periphery of the femoral head flattens and osteophytes form in this area. All these patients were managed with open/arthroscopic osteochondroplasty of femoral head. Methods: These patients were symptomatic for hip impingement. Cam deformity was studied using computed tomography and magnetic resonance imaging. In six patients open osteochondroplasty was carried out using surgical hip dislocation. In nine patients arthroscopic femoral head osteochondroplasty was done. All the patients were followed up for hip pain (VAS), Harris hip score (HHS), Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC), and hip range of motion. Results: A statistically significant improvement in the VAS for pain, HHS, and WOMAC score was noted. Average HHS improved from 71.3 (SD, 13) to 89.7 (SD, 14.5), p-value 0.0079. Average WOMAC improved from 73.6 (SD, 15.4) to 92.4 (SD, 16), p-value 0.0154. Impingement test became negative in all the patients. A significant improvement in hip ROM was noted. There was no conversion to total hip arthroplasty. All patients could sit on the floor cross-legged and squat. Conclusion: Some patients with partial collapse of femoral head due to osteonecrosis present chiefly with symptoms of femoroacetabular impingement. They should be identified as osteochondroplasty gives successful results in these patients. Level of evidence − IV


Introduction
Cam-type femoroacetabular impingement (FAI) is due to a femoral headÀneck deformity. A number of causes have been identified for such cam-type deformity, such as developmental abnormality in shape of femoral head, sequelae of slipped capital femoral epiphysis (SCFE), Legg-Calvé-Perthes disease (LCPD), post-infectious sequelae, and post-traumatic remodelling. But in majority of cases of cam-type deformity in adults, no antecedent cause can be identified [1].
Osteonecrosis of the femoral head (ONFH) is a result of inadequate blood supply leading to death of the osteocytes. Uneventful healing of ONFH is uncommon and variable amount of collapse occurs depending upon involvement of the femoral head [2,3]. In adults remodelling potential of the femoral head is limited and collapse due to ONFH leads to irreversible changes in the shape of femoral head. Size and site of necrosis in the femoral head have been identified as reliable prognostic indicators of femoral head collapse [4][5][6]. But we have noticed that some patients with a large-size lesion in the weight-bearing area do not show extensive collapse of the femoral head. Contrary to the expectation, these patients do not progress to advanced degeneration of the hip joint. Rather they presented with FAI due to deformity at femur head-neck junction.
FAI as a result of healed ONFH has not been described in literature before. We are describing this entity where partial collapse of femoral head occurs, particularly in the anterosuperior region, secondary to ONFH (Figure 1). A step deformity develops at the junction of this collapsed segment with normal bone in femur head-neck region ( Figure 2). With subsequent remodelling, periphery of the femoral head flattens and osteophytes form in this area. We have encountered 15 such patients who presented to us with hip pain and classical signs and symptoms of FAI in whom we could confirm that the antecedent cause was ONFH using previous MRI scans. All these patients were managed with open or arthroscopic osteochondroplasty and follow-up results of these patients are presented in this article.

Materials and methods
This case series included 15 patients (15 hips) who underwent surgery in our centre between January 2014 and January 2016. Clearance from the Institutional Ethics Committee was taken for this study. Informed consent was obtained from all the patients for inclusion in the study.
Inclusion criteria :  patients with little or no hip pain on standing and walking, which did not interfere with routine activities or necessitated use of walking aid; patients symptomatic mostly on hip motion particularly flexion and hip rotations; near normal joint space on plain radiographs, or more than 50% joint space preserved laterally compared to opposite hip; -MRI showing a continuous lining of cartilage on acetabular and femoral head; duration of symptoms more than 2 years to ensure that the hip has undergone sufficient remodelling and further collapse due to ONFH is not expected; presence of a cam-type lesion on femoral head-neck junction on radiographs/CT scans; presence of at least 50% of the normal arc of hip motion in coronal and sagittal planes with no fixed deformities.
In six patients open osteochondroplasty was carried out using surgical hip dislocation as described by Ganz et al. [7]. In nine patients arthroscopic femoral head osteochondroplasty was carried out. Technique used for hip was arthroscopy, which was similar to that described by Dienst et al. [8] where peripheral compartment arthroscopy is carried out first to address the femoral head-neck lesion. Indication for choosing open procedure over hip arthroscopy was more severe involvement of the femoral head with flattening of a larger circumference of the head-neck junction.
Preoperatively plain radiographs, magnetic resonance imaging (MRI), and computed tomography (CT) with 3D reconstruction were performed in all the patients for delineation of the lesion. Diagnosis of ONFH in all the patients was based on previous magnetic MRI records. Preoperative MRI scans of these patients were also evaluated for percentage area of femoral head involvement using method described by Kim et al. [9].

Details of arthroscopic procedure
Hip arthroscopy was carried out in supine position on a fracture table with a perineal post. An unscrubbed assistant could manipulate the operated limb in positions of flexion/extension, rotations, and abduction/adduction on the limb of the fracture table. No traction was used for the procedure. If at all required, traction was used at the end of the osteochondroplasty procedure to visualise the articular cartilage. 1 mg/L of epinephrine was added to the irrigation fluid.
Standard anterolateral portal was used to insert the arthroscope. Image intensifier was used to confirm the position of the cannulated guide pin in the femoral headneck area. A Nitinol wire was passed through the pin and the tract is dilated using dilators. Blunt trocar and a flow-sheath of arthroscope were introduced to feel bony contact in the anterior part of femoral neck to confirm its intra-articular placement. Arthroscope was then introduced. Alternatively, an outside-in capsulotomy could be 1 year follow-up shows partial collapse in both femoral heads, relatively preserved joint space, relative flattening at the lateral periphery of the femoral heads, and formation of impinging cam-type lesion. carried out as described by Thaunat et al. [10]. Standard anterior portal was used for instrumentation. Attempt was made to identify the normal articular cartilage just peripheral to the labrum in neutral position of the limb. This would help to identify the extent of the deformed head-neck portion (cam deformity). Extension of the hip joint would help to identify the anterior cam deformity, and internal rotation helped to identify the lateral cam deformity.
It is important to understand that the goal of the surgery is mainly to remove only those part of the cam deformity which are causing impingement in flexion and rotations. In patients with more circumferential lesions, there is a possibility that some parts of this deformity located posteriorly (posterolateral and posteromedial) may be left after the procedure if they are not found to be contributing to impingement.
Hip range of motion was carried out perioperatively for dynamic assessment of femoroacetabular contact and this was also confirmed fluoroscopically. All the patients were followed up prospectively for hip pain (VAS for pain), Harris hip score (HHS), Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC), hip movements in flexion, abduction, adduction, and arc of rotation. End point for success of this procedure was total hip replacement. Preoperative and post-operative functional outcome scores and range of motion were compared for statistical significance using Wilcoxon signed rank test.  Hip range of motion exercises were started as soon as the pain allowed, generally in first 1-2 days. Postoperative rehabilitation protocol included 6 weeks of protected weight bearing for patients undergoing surgical dislocation of hip. Immediate weight bearing was started for patients undergoing arthroscopic osteochondroplasty.

Results
Mean age of the patients was 27.5 years (range 18-34 years). Mean duration of hip symptoms was 2.9 years (2.1-4.2 years). Hip joint space was normal or only slightly  reduced in all the patients. Preoperative MRI scans of all these patients showed findings consistent with ONFH in all the patients. Average percentage area of femoral head involvement in these patients was 37% (range 28-50%).
At the time of surgery, articular cartilage in the weight-bearing area of the hip joint was found intact in all patients. Degenerative changes in the labrum at the site of impingement were seen in all the patients. Debridement of these tears was carried out. Labral repair was not required in any of these patients. A statistically significant improvement in the VAS for pain, HHS, and WOMAC score was noted in all the patients (Table 1). Average HHS improved from 71.3 (SD, 13) to 89.7 (SD, 14.5), p-value 0.0079. Average WOMAC improved from 73.6 (SD, 15.4) to 92.4 (SD, 16), p-value 0.0154. Impingement test became negative in all the patients after the surgery. A significant improvement in hip ROM was noted (Table 1). There was no conversion to total hip arthroplasty at the final follow-up. All patients could sit on the floor cross-legged and squat. There were no instances of osteonecrosis, superficial or deep infection, implant failure, or non-union of the trochanteric osteotomy site after the surgery. Average follow-up duration was 2 years (range 1.5-2.5 years). Preoperative and post-operative follow-up of representative cases is shown in Figures 3-7.

Discussion
We propose that the cause of anterosuperior femoral head-neck deformity is partial collapse of the femoral head due to osteonecrosis. It is not easy to understand why behaviour of these hips differed from the more usual event of advanced collapse in ONFH. ONFH most commonly involves the anterosuperolateral part of the femoral head. Involvement is generally in shape of a cone with base towards the articular surface. Biomechanically the cortical shell of the femoral head is stiffer than the underlying cancellous bone [10,11]. In ONFH the support of the cortical shell from underlying cancellous bone weakens due to necrosis of bone. If the cortical shell and underlying subchondral bone is strong, the entire conical area of the involved part of the head may depress as a single piece with buckling of the central metaphyseal bone. This collapse is generally small if good-quality bone is present in subchondral region. This anterosuperior part of the femoral head becomes relatively aspherical and a step appears at its junction with the normal bone laterally. With remodelling, peripheral osteophytes start appearing at this periphery.
In all our patients, a larger part of the femoral head was involved, including weight-bearing area and a large conical part of underlying femoral metaphysis, reaching up to the centre of the femoral head. This entire part sequestered together during collapse forming a step deformity at the periphery of the head or the head-neck junction. With revascularisation, repair of this lesion had occurred, as evidenced by this segment being well fixed and normally bleeding during the surgery.
Surgical options for hip preservation surgery in ONFH include core decompression or bone grafting in precollapse stage and redirectional osteotomies once collapse ensues. Hip arthroplasty is the standard treatment for hips involved with osteonecrosis. As these patients are young, most of the patients undergoing arthroplasty can expect more surgeries of their hip joints in future. Hip preservation should be attempted in these patients if possible. Osteochondroplasty of the femoral head-neck junction has not been described as an option for hip preservation in ONFH before. These patients with partial collapse of the head predominantly have symptoms of FAI. They have little pain while weight bearing and the sphericity of femoral head in weight-bearing region is relatively maintained. A significant pain relief and restoration of function can be expected in these patients after osteochondroplasty of the femoral head to remove the cam deformity and restoration of the femoral head-neck contour.
Natural history of ONFH is still incompletely understood. These findings also bring to our notice deficiencies in current understanding of changes occurring in the subchondral and metaphyseal bone in ONFH and mechanisms of head collapse and healing. One important concern in these patients with ONFH is possibility of progression of the collapse of the femoral head. All these patients were symptomatic for more than 2 years, which indicates that the vascular insult to the hip was old and remodelling has already occurred in the head. No further collapse was noted in any of these patients 2 years after osteochondroplasty. As these patients presented with a diagnosis of ONFH, presence of hip symptoms, and radiological changes in the femoral head, they had been advised hip arthroplasties before they were seen in our hospital.

Conclusion
Some patients with partial collapse of femoral head due to osteonecrosis present chiefly with symptoms of femoroacetabular impingement. They should be identified as femoral head-neck osteochondroplasty gives successful results in these patients. This article adds a new dimension to femoral head preservation surgery in carefully selected patients with osteonecrosis of femoral head. This study presents only a short-term follow-up. A longer follow-up of these patients is awaited and should be available in future.