Introduction

Osteonecrosis of the femoral head, which is also known as avascular necrosis, describes the clinical picture that is observed after the death of bone marrow and osteocytes. Resorption of the dead bone marrow and subchondral tissue leads to the generation of weaker tissue. The newly formed bone is prone to both fracture and collapse, and this may cause pain and decreased function of the hip joint. Without intervention, the disease process usually leads to articular cartilage destruction and resulting osteoarthritis. Osteonecrosis is most often seen in individuals who are 40 years old or younger, with approximately 10,000 to 20,000 new cases identified in the United States each year.

Trauma is the most common cause of osteonecrosis, but it can also be classified as atraumatic, in which the cause is not well defined. Traumatic osteonecrosis is seen when the blood supply of the femoral head and neck is disrupted as a result of trauma to the joint. This may occur after femoral head and neck fracture, hip dislocation, or both. Some atraumatic conditions (e.g., Caisson disease, sickle cell disease, myeloproliferative diseases, coagulation disorders) may directly cause osteonecrosis via the impairment of blood supply to the bone. However, most cases of atraumatic osteonecrosis have unknown causes, although they have numerous associated risk factors, including corticosteroid use, alcohol abuse, smoking, systemic lupus erythematosus, chronic renal disease, inflammatory bowel disease, human immunodeficiency virus infection, and hypertension. Some individuals may also have a genetic predisposition for osteonecrosis.

Various treatment methods have been used in an attempt to alleviate the symptoms or to slow the progression of the disease, including nonoperative (e.g., limited weight bearing, medication) and operative modalities (e.g., core decompression, bone grafting, osteotomies, hip arthroplasty). Nonvascularized bone grafting is a surgical technique that attempts to remove necrotic bone, to increase the amount of viable bone present in the femoral head for remodeling, and to provide support to decrease damage to the articular cartilage. In this chapter, we will describe the basic science of osteonecrosis as well as the indications, diagnostic methods, surgical techniques, rehabilitation, results, and complications of nonvascularized bone grafting for the treatment of osteonecrosis of the femoral head.

Basic science

In patients with traumatic osteonecrosis, the initial insult is a decrease in the blood supply that is caused by the disruption of the blood vessel. Bone marrow death follows within 12 hours of vascular compromise. The event that precipitates atraumatic osteonecrosis is less well defined, in part because of the lack of knowledge of the causes of this disease. Some causes are believed to incorporate the combined effects of metabolic factors, mechanical stresses, local factors that affect blood supply, and increased intraosseous pressure. Corticosteroid use and excessive alcohol intake are associated with the majority of cases.

The pathologic disease course that occurs after the initial insult is similar for both traumatic and atraumatic osteonecrosis. Reperfusion initiates a cascade of local metabolic factors that cause demineralization and trabecular thinning. Despite attempts to replace dead bone and repair weak bone, subchondral fractures occur. The resulting lack of mechanical support for the articular cartilage leads to altered joint mechanics and the degeneration of the hip joint.

Indications

As with other treatment modalities, the successful employment of bone grafting is most dependent on the stage of the disease. Several classification systems, including the Ficat and Steinberg systems, have been used to define various stages of tissue involvement, as shown in Table 30-1. The goal of bone grafting is to preserve the structure of the bone and the articular cartilage, so the procedure is most useful during precollapse stages (Ficat and Steinberg stages I and II), especially when less than 30% of the femoral head is involved. Bone grafting may also be used for a limited number of patients who have smaller Ficat stage III lesions if the articular cartilage is mostly intact.

Table 30–1 Descriptions of the stages of two widely used systems for classifying osteonecrosis of the femoral head

* The University of Pennsylvania stages I-V are further subclassified into three grades: Grade A = mild, involving less than 15% of the femoral head. Grade B = moderate, involving 15 to 30% of the femoral head. Grade C = severe, involving greater than 30% of the femoral head. In Stage V, the grade is determined by averaging the extent of involvement of the femoral head and the acetabulum.

History and physical examination

Patients who have osteonecrosis of the femoral head often present in a similar manner to patients who have other hip diseases (e.g., osteoarthritis), except that patients with osteonecrosis are frequently 40 years old or younger. The most common symptom of osteonecrosis is groin pain, which often occurs with weight bearing or other activity. Patients may describe the feeling of a groin pull (i.e., a sudden loss of hip stability) or groin fullness. Some patients who have an advanced stage of osteonecrosis may experience a sudden change in their ambulatory status. The physician should understand the risk factors for this disease. Patients may have a history of corticosteroid or alcohol use. Corticosteroid doses of more than 2 g of prednisone in 3 months or analogous doses of other steroids are generally associated with osteonecrosis. Alcohol exhibits a clear relationship with osteonecrosis, with higher doses of alcohol being associated with an increased risk of osteonecrosis. Other important risk factors include chronic conditions (e.g., systemic lupus erythematosus, renal disease) and immunocompromising conditions (e.g., human immunodeficiency virus, previous organ transplantation).

During the physical examination, pain is reproduced with passive as well as active internal rotation of the leg. The patient will point to the groin as the area of maximum pain, but he or she occasionally may point to the lateral trochanteric region, which may signify referred discomfort. Patients may be unable to fully bear weight on the affected extremity because of the pain. It is important to routinely perform a straight-leg raise and contralateral straight-leg raise to ensure that the cause of the pain is not the lumbar region.

Imaging and diagnostic studies

When evaluating a patient who has hip pain, several imaging studies may facilitate the diagnosis of osteonecrosis, including plain radiographs, bone scans, and magnetic resonance imaging (MRI). Currently, the standard diagnostic studies include plain radiographs followed by MRI. Plain radiographs should include an anteroposterior pelvic view to include the contralateral hip as a comparison and a cross-table lateral view of the affected hip. The presence of sclerosis, cystic changes, a crescent sign, or collapse on the plain radiographs may indicate the presence of osteonecrosis. The MRI is the current gold standard for the diagnosis of osteonecrosis; it is considered to be 98% to 99% sensitive and specific for the condition, and it also allows for the quantification of the size of the necrotic segment on both the sagittal and coronal axes.

Surgical technique

When performing a nonvascularized bone graft, the authors prefer to have the patient placed in the lateral decubitus position (Figures 30-1 through 30-6). The patient is secured to the operating table, pads and cushions are placed on the extremities, and the airway is controlled by the anesthesiologist. The direct lateral approach is the surgical approach of choice for the author, because the vasculature to the femoral head is more easily preserved. The posterior approach is also used, but greater care needs to be taken to avoid injury to the lateral epiphyseal vessel, which is the branch of the medial obturator vessel that supplies the femoral head. After the patient is appropriately positioned, prepared, and draped, an 8-cm to 14-cm incision is made overlying the center of the greater trochanter and brought proximally and distally. Sharp dissection is performed with the use of a No. 10 scalpel down to the level of the iliotibial band. The iliotibial band is then incised and retracted posteriorly and anteriorly. A blunt Hohmann retractor is placed posteriorly, again to avoid injury to the neurovascular structures located posterior to the greater trochanter. The bursa can be excised for the better identification of the vastus ridge, the greater trochanter, and the insertion site of the gluteus medius muscle. With the use of an electrocautery device, the anterior 40% of the gluteus medius muscle is elevated from the anterior neck of the proximal femur in conjunction with the gluteus minimus muscle. The bursa between the iliopsoas muscle and the hip capsule is explored and retracted anteriorly with the use of a Cobb elevator; this allows for the preservation of the muscle and avoids injury to the neurovascular structures anteriorly. Care is taken to avoid the excessive stripping of the hip capsule. The gluteus medius and minimus tendons are taken proximally and held in place with the use of a Taylor retractor. Care is taken to avoid injury to the neurovascular structures that innervate the gluteus medius and gluteus minimus tendons. Excessive dissection is not necessary posteriorly. If additional exposure is necessary, then a Cobb elevator is used to elevate some of the muscular attachments of the gluteus muscle onto the capsule. We prefer to place a Hohmann retractor anteriorly and a Taylor retractor superiorly, and this is followed by the use of a blunt Hohmann retractor posteriorly. These retractors allow for excellent exposure to the hip capsule. The hip capsule is incised from the femoral neck and extending to the acetabulum. It is important to avoid injury to the labrum of the acetabulum. The capsule incision is then extended from the base of the neck anteriorly and then posteriorly approximately 180 degrees around the circumference of the femoral neck. The posterior capsule is not violated. The capsule is also peeled off on the acetabular side approximately 180 degrees, thereby exposing the whole anterior lateral labrum. Next, the capsule incision is shaped like an “H,” which allows for the visualization of the neck–cartilage junction. The anterior neck is outlined with the use of a surgical marker. A 1.5-cm by 1.5-cm square area of bone is identified. The surgeon should take great care to avoid extending this window too far medially and thus compromising the neurovascular structures or too far laterally and thus creating a potential stress riser for a femoral neck fracture. An oscillating saw is used to remove the area of bone. The authors prefer to bevel the bony cuts to allow for the replacement of the bony window after the bone grafting procedure is completed. A high-speed burr is used to remove all of the necrotic bone from the femoral head. This procedure can successfully be performed without dislocation, but if dislocation is necessary, then we recommend partial dislocation to obtain better visualization of the femoral head. A full dislocation increases the risk of complete femoral head death. After the necrotic segment within the femoral head is removed, a light is placed into the femoral head to ensure that the entire necrotic segment has been removed, and the remaining bone is observed to ensure that bleeding is still present. The authors prefer to use fresh- frozen bone graft that has been milled and then mixed with bone morphogenic protein to maintain the osteoinductive and conductive properties of the bone. Autograft is also used, when available. The autograft can be harvested from small cores within the greater trochanter or from a secondary incision along the iliac crest. The femoral head is packed tightly with the substrate. If a large volume of the head has been denuded of bone, then small cortical strips can be added to the substrate to provide structural integrity. The cortical window is then placed back onto the femoral neck and fixed with the use of three resorbable pins.

Figure 30–1 The patient is in the direct lateral position, and the incision is centered over the trochanter.

Figure 30–2 At the completion of the bone grafting, the cortical window is replaced and fixed with the use of three bioabsorbable pins.

Figure 30–3 With this approach, the femoral head does not need to be dislocated to allow for access to the anterior cortex or to preserve the posterior blood supply.

Figure 30–4 The femoral head is delivered into the wound, thereby minimizing internal rotation and maximizing the protection of the posterior blood supply.

Figure 30–5 A light can be used to evaluate the amount of bone that has been removed from the femoral head.

Figure 30–6 Cortical cancellous chips are packed tightly into the femoral head.

Closure is accomplished in the following manner. The capsule is reduced to its normal anatomic position and held in place with the use of bioabsorbable suture. The gluteus minimus tendon is separated from the gluteus medius tendon. The authors prefer to use No. 5 nonabsorbable polyester sutures to secure both the gluteus minimus and gluteus medius tendons back to bony tunnels onto the greater trochanter; this promotes tendon–bone healing, and the authors believe that this helps to reduce the incidence of postoperative limp. A drain is then placed deep into the wound, and closure is continued. The iliotibial band and the subcutaneous tissue are reapproximated, and the skin is often closed with staples. A sterile dressing is applied, and an abduction pillow is placed between the patient’s legs.

Postoperative rehabilitation

Patients who have undergone nonvascularized bone grafting procedures usually remain nonambulatory for 6 weeks. The authors recommend a maximum of 20% weight bearing with two crutches or a walker during this period. Patients are allowed to perform gentle range-of-motion exercises, but they are instructed to avoid excessive abduction in an attempt to protect the gluteus medius and gluteus minimus muscle repair. They are also encouraged to avoid hip flexion beyond 90 degrees, excessive hip rotation, and adduction to reduce the risk of dislocation. After 6 weeks, radiographs are obtained to confirm the lack of progression of osteonecrosis and the beginning of cortical remodeling. If adequate remodeling is present, then weight bearing is increased to 50% with a cane, abductor strengthening exercises can begin, and gait training is encouraged. Radiographs are obtained 3 months, 6 months, and then yearly after the procedure. If adequate remodeling has occurred at 3 months, patients are advanced to full weight-bearing status. Strenuous activity is not allowed for a minimum of 6 months.

Results and outcomes

Multiple published studies have indicated the success of nonvascularized bone grafting procedures for the treatment of early-stage osteonecrosis. Ko and colleagues reported about the use of the “trapdoor” bone-grafting procedure in addition to a femoral or acetabular osteotomy to treat 9 teenagers (10 hips) who had severe osteonecrosis and articular surface collapse. Cancellous bone grafting was used to pack the femoral head. At an average follow up of 53 months (range, 2 to 9 years), 7 patients (8 hips) had good clinical results, and 2 patients (2 hips) had fair clinical results. Six hips were rated as good radiographically, 3 hips were rated as fair, and 1 hip was rated as poor.

Rosenwasser and colleagues examined the use of the “lightbulb” bone-grafting approach to treat 13 patients (15 hips) who had a mean age of 34 years and who had Ficat stage I, II, or III osteonecrosis of the femoral head. At a mean follow up of 12 years (range, 10 to 15 years), 11 patients (87%) remained pain free, with minimal progression of osteoarthritis. No fractures, infections, or thromboembolic events were reported.

Mont and colleagues reported about 30 trapdoor procedures that were performed on 23 patients who had Ficat stage III or early stage IV osteonecrosis of the femoral head. The mean Harris Hip Score improved from 41 points (range, 31 to 64 points) preoperatively to 92 points (range, 80 to 100 points) at the final follow up. In 2003, Mont and colleagues reported about the use of the lightbulb procedure to treat 19 patients (21 hips) who had Ficat and Arlet stage II or stage IIII osteonecrosis of the femoral head. The mean Harris Hip Score improved from 48 points (range, 25 to 62 points) preoperatively to 91 points (range, 80 to 100 points) at the final follow up.

Seyler and colleagues provided a retrospective study of 33 patients (39 hips) who had Ficat stage II or III lesions and who were treated with trapdoor procedures. At a mean follow up of 36 months (range, 24 to 50 months), 24 of the 30 hips (80%) that had small and medium-sized lesions did not require further surgery. Eighteen out of 22 hips that had stage II disease did not require further surgery. It was also noted that lateral lesions fared poorly as compared with centrally located lesions.

In summary, the lightbulb and trapdoor techniques have demonstrated excellent results in several studies of patients who have Ficat stages II and III osteonecrosis of the femoral head with small to medium-sized lesions.

Complications

Complications associated with nonvascularized bone grafting are similar to those described for core decompression. These include infection, continued pain, progression of collapse, and femoral head or neck fractures. The most common complication that has been reported in the literature is the progression of the osteonecrosis that ultimately results in a hip replacement. In one study, 2 out of 15 hips (13%) that were treated with cancellous bone grafting required total hip arthroplasties for disease progression at a mean follow up of 12 years (range, 10 to 15 years). Mont and colleagues reported a 14% rate of conversion to total hip arthroplasty at a shorter follow up of 48 months. There were no reported cases of femoral neck fractures. In our experience, there were no femoral neck fractures after nonvascularized bone-grafting procedures. Seyler and colleagues reported that the size of the lesion plays a role in the need for a subsequent hip replacement. Their study revealed 9 large lesions that required conversion to total hip arthroplasties. There have been no reported cases of venous thromboembolism after this procedure.