OSTEONECROSIS

Because of the limited biological response of different
organ systems, various etiologic factors can lead to similar pathologic
changes. In most cases of osteonecrosis, a specific etiologic factor or
factors can be identified. In traumatic cases such as hip dislocation
and femoral neck fracture, there is a clear cause-and-effect
relationship between the insult, mechanical damage to the vessels that
supply the femoral head, and the resulting osteonecrosis. A number of
etiologic factors can be identified in patients with nontraumatic
osteonecrosis. These include intraosseous marrow displacement disorders
that lead to increased pressure in the femoral head and neck, such as
Gaucher’s disease, leukemia, and myeloproliferative disorders (52,58,85,144).
Other conditions, such as radiation and chemotherapy, lead to bone
necrosis by direct cell toxicity. Mechanical blockage of vessels may be
caused by emboli composed of abnormal red blood cells, such as in
sickle cell disease and thalassemia; or nitrogen bubble emboli, such as
in caisson disease, or dysbarism (59,60).

In many cases of osteonecrosis, the etiology is less
clear. In these cases, one can often identify specific factors known to
be frequently associated with osteonecrosis, although the exact
mechanisms by which they act have not been completely delineated (Table 125.1).
For example, high doses of corticosteroids and excessive alcohol intake
have been identified in nearly 80% of cases of nontraumatic
osteonecrosis (30,52,53,61,84,85,87,144).
However, a one-to-one relationship between these factors and this
disorder has not been established, because most patients exposed to
steroids or alcohol do not develop osteonecrosis (30,84,85).
Although the entity of idiopathic osteonecrosis most likely does exist,
as we learn more about the etiology of this condition, the number of
patients who will be relegated to this category will diminish (61).

There have been many mechanisms proposed to explain the
pathogenesis of osteonecrosis. They can be categorized into six groups:
(a) direct cellular toxicity, (b) extraosseous arterial, (c)
extraosseous venous, (d) intraosseous extravascular, (e) intraosseous
intravascular, and (f) multifactorial (Table 125.2).

Although several different etiologic factors can lead to
vascular impairment in osteonecrosis, the sequence of events that
follow the initial insult or insults is similar. The resultant hypoxia
rapidly leads to increased cell membrane permeability, which allows
fluid and electrolytes to enter the cell, causing it to swell.
Intracellular lysosomal enzymes are released, resulting in
autodigestion or coagulation necrosis and cell rupture. Vascular injury
leads to tissue edema and hemorrhage. An inflammatory response ensues,
marked by the appearance of neutrophils and macrophages.

Death of bone is more difficult to diagnose than marrow
necrosis and is predicated on the disappearance of osteocytes from
within their lacunae, a process that takes several days to develop (Fig. 125.1).
The center of the necrotic lesion remains avascular and repair is not
possible. Bone and soft-tissue detritus accumulate. As a result of
repeated stresses, dead trabeculae undergo microfractures that cannot
be repaired.

Toward the periphery of the necrotic area where
sufficient vascularity is present, the transition zone, an active
process of repair begins. Macrophages and osteoclasts remove dead
marrow elements and bone. Granulation and fibrous tissue are formed.
Osteoblasts form new bone, much of which is laid down directly on
remnants of dead trabeculae (Fig. 125.2). The
resulting trabeculae are much thicker than normal and are responsible
for the sclerotic margin that surrounds the lesion and that is one of
the radiographic hallmarks of osteonecrosis (Fig. 125.3).

If the lesion is small, and particularly if it is not in
a major weight-bearing region of the femoral head, it may undergo
revascularization and be completely replaced with viable bone.
Alternatively, it may remain as a small sequestrum surrounded by a wall
of new bone. In either case, femoral head collapse does not occur, the
hip continues to function normally, and the patient remains
asymptomatic.

Larger lesions, particularly those in a region of major
weight bearing, have a poor prognosis. The attempt at repair is only
partially effective. Bone resorption is more rapid than formation.
There is a gradual collapse of cancellous bone beneath the bony
endplate of the weight-bearing

aspect
of the femoral head. If the contour of the articular surface remains
intact, a fluid-filled space beneath the cortical subchondral bone
develops, which gives the appearance of a crescent sign on radiographs (Fig. 125.4).
With flattening and collapse of the articular surface, this space is
obliterated and the crescent disappears. Gross flattening of the
femoral head is now apparent (Fig. 125.5).

Articular cartilage derives its nourishment from
synovial fluid within the joint rather than from vessels in the
cancellous bone, so it remains viable while the attached subchondral
bone dies. The mechanical stresses on the collapsed and irregular
articular surface eventually result in damage to and death of
chondrocytes (Fig. 125.6). These abnormal
stresses are transferred to the otherwise normal cartilage of the
acetabulum, which undergoes secondary degenerative changes. Joint line
narrowing becomes obvious on radiographs. As the processes of
degeneration continue, the underlying acetabular bone becomes affected.
Typical changes of degenerative joint disease appear and include
sclerosis, cyst formation, and marginal osteophytes. End-stage
arthritis of the hip eventually ensues (18,28,55,56,58,76,113,114).

Osteonecrosis is quite different histologically from transient osteoporosis (or osteopenia) of the hip (TOH) (39,42),
although these two conditions can be confused. In TOH, there are no
findings of bone infarction or repair, which are the hallmarks of
osteonecrosis. The pathologic picture is primarily one of marrow edema,
hence TOH is also referred to as bone marrow edema syndrome (BMES).
Some of the pathologic features of TOH, such as the presence of edema
fluid, have also been found in early stages of osteonecrosis, and this
has led some authors to conclude that these two entities are at
different ends of the spectrum of the same disorder (49).

Clinically, transient osteoporosis also behaves
differently from osteonecrosis. The onset of pain is usually more
sudden and it is often more severe. To minimize the chance of fracture,
place patients on crutches until there is clinical, radiographic, and
MRI evidence of resolution. This may require 4–6 months. This condition
affects men more often than women, and these patients have no
associated risk factors as in osteonecrosis. In women, it classically
develops during the third trimester of pregnancy, and the incidence of
fracture is greater than in men. The disease rarely involves both hips
at the same time. Occasionally, the opposite hip is affected months or
years later. It is difficult to make a definitive diagnosis on the
basis of standard radiographs because the only abnormality is mild
osteopenia of the femoral head and neck. However, the diagnosis can be
made readily based on MRI in most cases.

It is essential to distinguish transient osteoporosis
from osteonecrosis, because the natural histories, prognoses, and
treatments of these two conditions are quite different. Because TOH is
usually self-limited, it is treated with protected weight bearing to
prevent fracture. Infrequently, core decompression may be indicated if
a patient has an inordinate amount of pain or if the diagnosis is in
doubt. Osteonecrosis, on the other hand, is a progressive disorder if
untreated, and in most instances early diagnosis and treatment are
required (18,48,49).

The signs and symptoms of osteonecrosis are nonspecific.
For weeks and perhaps even months after the initial vascular insult,
the involved area may be entirely asymptomatic. When symptoms develop,
they usually do so gradually. This may be due to a buildup of
intraosseous pressure initially and later perhaps to microfractures of
affected trabeculae. The pain is usually localized to the inguinal
region but may involve the buttock or the upper thigh. Rarely does it
radiate as far as the knee. It may be present at rest but is often
exacerbated by activity. Later, a limp and a slight decrease in ROM,
associated with pain, may develop.

Although both hips are affected in more than 50% of the
cases, the involvement is usually not simultaneous. Therefore, the
patient will normally present initially with symptoms on only one side.
If the contralateral hip is afflicted, symptoms will usually develop
within 3 to 6 months. Approximately 80% of asymptomatic hips with
MRI-proven osteonecrosis eventually become symptomatic. The incidence
is much less in the case of small lesions. When the asymptomatic
contralateral hip of a patient with osteonecrosis on the opposite side
appears normal on the initial MRI, there is less than a 10% incidence
that this hip will develop osteonecrosis at a later date (32).
In a small number of cases of advanced osteonecrosis, the patient may
develop a positive Trendelenburg sign and have shortening of the limb,
due both to collapse of the femoral head and to a limited range of hip
motion resulting in a functional shortening.

In approximately 15% of cases of hip involvement, other
areas will also develop symptomatic osteonecrosis. These include knees,
shoulders, wrists, feet, and, rarely, elbows and facial bones.
Accordingly, these areas should be evaluated by history and physical
examination. If there is any question of involvement, plain radiographs
and perhaps a bone scan or an MRI should be obtained. Recently, MRI
studies have indicated that the incidence might be considerably higher
than expected.

Take a careful history, looking for possible etiologic
or associated factors. High doses of corticosteroids and prolonged,
excessive alcohol intake are most often implicated, but a number of
other factors should also be considered, as discussed earlier. The
quantities of alcohol or steroid required to cause osteonecrosis have
not been exactly determined, and it is clear that there is a tremendous
variation in the sensitivity of patients to these agents. For example,
it has been shown that patients on corticosteroids who develop
Cushingoid features are much more likely to develop osteonecrosis than
individuals taking the same dose of steroid who do not become
Cushingoid (120).

In most instances, routine laboratory tests are within
normal limits. In selected cases, serologic testing can be used to rule
out or diagnose other possible causes of hip pain. Although these tests
do not diagnose osteonecrosis per se, they may identify certain risk
factors, thus helping to support the diagnosis. These include sickle
cell disease, systemic lupus erythematosis, hyperuricemia, abnormal
amounts of circulating lipids, and subtle coagulopathies.

Abnormalities in coagulation factors have recently been
identified in more than 70% of patients with previously diagnosed
“idiopathic” osteonecrosis (45,46).
They have also been identified in a large number of patients with
osteonecrosis in whom other inciting factors, such as alcohol or
steroids, are present. It is thus felt that either alone or in
combination with other agents, these coagulation abnormalities may
predispose to or cause osteonecrosis.

These abnormalities cannot be detected on standard
measurements of coagulation, such as prothrombin time, partial
thromboplastin time, and bleeding time, which are usually within normal
limits. A more complete coagulation profile is therefore required to
identify these factors. Thrombophilia may be associated with decreased
levels of protein C, protein S, and antithrombin III (AT III);
increased resistance to activated protein C (RAP-C); and elevated
antiphospholipid antibodies. Hypofibrinolysis may be associated with
increased plasminogen activator inhibitor activity (PAI-1), increased
lipoproteins, and decreased stimulated plasminogen activator activity.
Unfortunately, at the present time these tests cannot be performed in
all facilities, and they are expensive.

The earlier literature recommended that certain invasive
tests, referred to as “functional exploration of bone,” be performed in
patients suspected of having osteonecrosis despite normal radiographs (12,38,39,52).
These included intraosseous venography, intraosseous pressure
measurements, and histologic examination of core biopsy specimens taken
from the femoral head. Although these tests can often lead to the
diagnosis of osteonecrosis prior to the development of radiographic
changes, they are seldom indicated if MRI is available. Rarely,
histologic evaluation may be useful in distinguishing between
osteonecrosis and certain other conditions affecting the femoral head,
such as TOH (18,39,45,48,49).

Even more important than the treatment of a disease
state is its prevention. In osteonecrosis, certain specific risk
factors have been identified and these should be eliminated or
minimized to the extent possible. This applies particularly to alcohol
ingestion and steroid administration, the two leading causes of
osteonecrosis, as well as to smoking. In regard to steroid use, altered
dosage schedules and the substitution of other agents have already led
to a decrease in the incidence of osteonecrosis under circumstances in
which the use of corticosteroids has not yet been completely
eliminated. Established guidelines for divers and those working under
hyperbaric conditions must be followed. When they become generally
available, medications shown to be effective in combating some of the
known causes of osteonecrosis should be appropriately utilized in those
patients particularly at risk.

Certain systemic disorders have been associated with an
increased incidence of osteonecrosis. These include coagulopathies,
hyperlipidemias, and the presence of specific antibodies. Screen
patients with osteonecrosis for the presence of these disorders, with
coagulation studies, liver function tests, lipid profiles, and tests
for anticardiolipin antibodies and lupus anticoagulant. If definite
abnormalities are found, decide whether to treat them and which agents
to use. Limited studies have shown a possible role for stanozolol, an
anabolic steroid that alters lipoproteins and suppresses clotting
factors; nifedipine, a vasodilator; agents that lower circulating
lipids; and long-term anticoagulation to treat coagulopathies. A
limited number of patients in this latter category have been treated to
date. Anticoagulation is not a benign form of treatment, and the
appropriate dose and duration of treatment has not been established. In
general, neither the effectiveness nor the safety of these agents has
been determined, nor do we have appropriate guidelines for their use at
present. We can hope they will prove to be effective and will be
available for the prevention and treatment of osteonecrosis in the not
too distant future (46).

When a patient is diagnosed with osteonecrosis of a
weight-bearing joint, there is a natural desire on the part of the
treating physician to protect the involved area from excessive stress
by using some form of limited weight bearing. Thus, canes or even
crutches are frequently prescribed. Although these may decrease the
degree of discomfort in patients who are symptomatic, they have not
been shown to alter the natural course of this disorder.

Protected weight bearing may be used empirically as an
alternative to surgical management in cases with a good prognosis.
These cases include small, asymptomatic lesions and those in regions of
relatively low weight bearing, such as the medial aspect of the femoral
head. However, it has not been established that limited weight bearing
improves the already favorable course in such instances. Symptomatic
management, including protected weight bearing, may be the treatment of
choice when the patient’s age, general prognosis, and associated
medical conditions, or the patient’s own wishes, contraindicate
surgical intervention.

Protected weight bearing is mandated following certain
types of surgical procedures, such as core decompression, grafting, and
osteotomies, where it is used as an adjunct to the basic operative
approach. Here it serves to protect the weakened regions from fracture,
and perhaps protects the femoral head as well, until the healing
processes have progressed satisfactorily. Perhaps the most important
role for protected weight bearing is in the patient with relatively
advanced stages of osteonecrosis, when it is felt that prophylactic
procedures are of little avail. Here the use of a cane or perhaps
crutches can diminish symptoms and improve function considerably until
such time as a reconstructive procedure is indicated (3,81,85,94).

Various types of electrical stimulation have
demonstrated an ability to enhance bone formation and fracture healing.
It was thus natural for electrical stimulation to be applied to
patients with osteonecrosis, either alone or as an adjunct

to
other surgical procedures, in the hope that it would stimulate healing
of the necrotic regions. Three specific signals have been used: direct
current (DC), capacitive coupling (CC), and pulsing electromagnetic
fields (PEMF). The latter two are transmitted to the involved bone by
means of surface electrodes or coils placed on the skin over the
involved region. Thus they can be used alone or in conjunction with
surgical procedures. DC stimulation requires the insertion of an
electrode directly into the bony region to be stimulated, and therefore
it has been used as an adjunct to core decompression rather than as an
independent treatment modality.

Only limited studies have been reported using these techniques (18).
Capacitive coupling, with the specific signal utilized, failed to show
any enhancement of the effects of core decompression and bone grafting
alone. DC stimulation, despite indications of an early response, did
not show a long-term effect. Results with PEMFs, both in individual
studies and in a multicenter study, were promising (1,4).
They were shown to be more effective than symptomatic management in
precollapsed and minimally collapsed hips; as effective as core
decompression in precollapsed lesions; and more effective in hips with
early collapse, both as determined by radiographic progression and
delaying the need for arthroplasty. A special signal, different from
that used to treat fractures, was used to treat osteonecrosis. At
present, this is not generally available for routine use in the United
States. If further evaluation by other investigators confirms these
early positive findings, PEMF may be added to the treatment modalities
available for this condition (1,4,119).

During the early 1960s, Arlet and Ficat (12,38,39)
used core decompression to examine the pathologic changes in femoral
heads of patients who were suspected of having AVN. This procedure
caused a decrease in the abnormally elevated intraosseous marrow
pressures and frequently produced immediate relief of preoperative
pain. Subsequently, this approach was used as a therapeutic rather than
a diagnostic procedure. It is presumed to work by decreasing the
abnormally high intraosseous pressures present in osteonecrotic
lesions, by opening up channels for vascular ingrowth, and by
stimulating the natural processes of repair. This technique can be
utilized as initially described or it can be supplemented with loosely
fitted cancellous or cortical grafts, electrical stimulation, or
various agents with the potential to stimulate bone healing, such as
bone morphogenetic protein (BMP) and demineralized bone matrix (DBM) (2).

Investigators do not completely agree as to the safety or the effectiveness of this procedure. In 1985, Ficat and Arlet (38)
reported on 133 hips in stages I and II treated with core
decompression. They noted “good and very good results” in 90% of hips
clinically and in 79% of hips radiographically. In 1986, Camp and
Colwell (26) retrospectively reviewed 40 core
decompressions performed by 13 surgeons in their area. Sixty percent of
hips treated before collapse failed either radiographically or
clinically, and all hips treated after collapse were considered
clinical failures. Four patients sustained postoperative fractures.

Although opinion remains somewhat divided on the
efficacy of core decompression, most studies report a low incidence of
complications and results significantly better than with symptomatic
management. In an extensive review of the literature published in 1996,
Mont et al. (81) reviewed 42 reports of 2,025
hips, of which 1,206 were treated by core decompression and 819 by
nonoperative means. Satisfactory results were found in 64% of hips
treated by core decompression, compared with only 23% of hips treated
nonoperatively. In the hips treated before collapse, good results were
obtained in 71% with core decompression and 35% treated nonoperatively.
The best results were achieved at centers doing the greatest numbers of
procedures (37,81,85,110,124).

Various types of osteotomies have been utilized in the treatment of osteonecrosis of the femoral head over the past 30 years (13,33,65,72,76,83,106,133).
These are technically demanding procedures that are not frequently
employed in North America. Results vary considerably, and subsequent
arthroplasty may be compromised. The most commonly cited rationale for
their efficacy is the biomechanical effect of moving the necrotic
segment of the femoral head from the principal weight-bearing area of
the hip to an area that bears less weight, and replacing

it
with relatively normal bone and cartilage. Others have suggested that
the beneficial effects of osteotomies are secondary to the procedure
effecting a reduction in venous hypertension and intramedullary
pressure (13).

Two general types of osteotomies have been described:
varus or valgus osteotomies, usually combined with flexion or
extension, and transtrochanteric rotational osteotomies. The varus or
valgus osteotomies have been associated with variable rates of success
after short-term follow-up of approximately 5 years (13,72).
Recent reports also include a prospective study of valgus-extension
osteotomies combined with curettage and bone-grafting of the necrotic
segment in a group of non-corticosteroid-associated hips (106).
These authors reported a good or excellent clinical result in 36 of 45
hips (80%) at a mean follow-up of 65 months (range, 32 to 126 months).
In a recent study of varus osteotomies combined with flexion or
extension (83), there were 28 of 37 good and
excellent outcomes (76%) at a mean follow-up of 11.5 years (range, 5 to
18 years). In patients without a corticosteroid association in this
study, there were 17 of 20 successful clinical outcomes (85%).

The indications for varus or valgus osteotomies depend
on the location and size of the lesion. Osteotomies may be utilized for
both pre- and postcollapse lesions, but they should usually not be
performed if there is acetabular involvement. Osteotomies work best
when the lesions are small or medium sized with a combined necrotic
angle of less than 200° (Fig. 125.20) (65)
or with less than 30% of femoral head involvement. For varus
osteotomies, there should be at least 20° of the superolateral femoral
head not involved with disease, because this area of cartilage will be
shifted into weight bearing after the osteotomy.

Likewise,
a valgus osteotomy requires normal bone and cartilage in the central or
medial aspect of the head. Guidelines for when to add a flexion or
extension component to the osteotomy are similarly based on the
location of the lesion. Extension can be added when the necrotic
segment is posterior and flexion can be added if the lesion is anterior.

Bone-grafting procedures can be divided into two general categories: nonvascularized (19,20,23,35,41,56,76,82,98,102,109,114,142) and vascularized (16,22,57,77,134,135,143).
Most of these procedures are designed to provide some degree of
structural support to prevent collapse of the articular cartilage
surface. There are several different types of nonvascularized
bone-grafting procedures, and they differ considerably. The graft may
be introduced through the femoral head, the neck, or the
intertrochanteric area. Either cancellous or cortical bone can be used,
and this can be obtained from the patient’s own iliac crest, proximal
femur, tibia, or fibula, or from the bone bank.

There are also various techniques for performing
vascularized bone grafting. The graft may be taken from the ilium,
fibula, or greater trochanter. It may involve a muscle pedicle, which
contains its own vascular supply, or a microvascular anastomosis can be
performed using regional vessels. These procedures have been used both
before and after femoral head collapse but work best for precollapse
lesions. They should not be utilized when there are obvious
degenerative changes in the acetabulum.

As mentioned earlier, protected weight bearing does not
alter the natural progression of this condition. Thus, it should not be
considered an alternative method of prophylactic management. In the
later stages, however, it is a useful adjunct for symptomatic
management of a patient who is not yet ready for a definitive
arthroplasty. Intermittent use of a single cane or perhaps even two
crutches at the discretion of the patient, together with a reasonable
limitation of activities and the use of mild analgesics, will often
decrease pain and enhance function.

Prior to the development of THR, cup arthroplasty was
often considered the standard reconstructive procedure for patients
with late stages of osteonecrosis, degenerative arthritis, and other
arthritides. The postoperative regimen was lengthy and complicated, the
results were inconsistent, and even relatively good results were
generally not completely satisfactory by today’s standards.
Accordingly, conventional cup arthroplasty was essentially abandoned
with the advent of THR.

Because osteonecrosis initially affects the femoral head
and involves the acetabulum only much later, a number of surgeons have
advocated performing arthroplasty on the femoral side alone, provided
that the acetabulum is still relatively normal. This can be
accomplished by using a modified cup arthroplasty or a femoral
endoprosthetic replacement arthroplasty (25,53,67,93,108,116,132).

Investigators in France (65)
modified the standard cup arthroplasty and referred to it as an
adjusted cup arthroplasty. In this procedure, the femoral head is
reamed and a snugly fitting metallic cup is impacted over it. The
articular cartilage of the acetabulum is left intact. Kerboul et al. (65)
reported on 80 hips treated by this operation and followed for 1–6
years. At follow-up, 66% were felt to be successful, 16% somewhat
improved but still experiencing pain and a limp, and 16% failures. The
failures were related to rotation of the cup on the femoral head,
femoral head resorption and collapse, and acetabular erosion. The
authors felt that with more rigid indications and improvements in
techniques, the results could be improved.

Sedel et al. (108) reported
satisfactory short-term results with the adjusted cup arthroplasty but
later abandoned this procedure in favor of ceramic THR arthroplasty. In
the United States, Nelson and Walz (93)
modified the cup arthroplasty by cementing the cup onto the femoral
head, leaving the relatively normal acetabulum intact. They reported
satisfactory intermediate-term results (17,65).

Surface replacement arthroplasty (SRA) was popularized
in the 1970s as an alternative to conventional THR. It was thought to
be especially suited for the young adult. Several different designs
were available, and a number of procedures were performed in patients
with osteonecrosis. Despite the theoretical advantages of this
approach, the clinical results were unacceptable, with a high 5-year
failure rate. The procedure was therefore virtually abandoned. However,
McMinn et al. (75), Amstutz (5,9),
and others have continued their work to eliminate the problems inherent
in the original designs, which used thin polyethylene acetabular
components cemented into place. Newer devices incorporate
metal-on-metal articulations with acetabular components adapted for
both cemented and cementless fixation. To date, results have been
variable and we must await further reports with larger numbers and
longer follow-ups to determine the roles that these components will
eventually play in the treatment of patients with osteonecrosis (5,9,75,107,140).

Hemisurface replacement, also known as cup
hemiarthroplasty, is a direct descendant of cup arthroplasty, which was
originally conceived by Smith-Petersen (111).
Because it involves resurfacing the femoral head only, it is a viable
treatment option for patients with osteonecrosis of the femoral head
when bone quality is adequate and when acetabular cartilage is
relatively normal. Because the femoral canal is not violated, nor is
the femoral intramedullary diaphyseal bone compromised, as it is with
THA, hemisurface replacement preserves bone stock, affording potential
for easy revision. With normal transfer of stress to the proximal
femur, proximal bone loss due to stress shielding may be mitigated.
Cartilage durability is enhanced

by
optimizing contact of the hemispherical bearing to the acetabular
cartilage. An added advantage of this technique is that the larger ball
size ensures greater joint stability than with THA (7).

The rationale for using a femoral endoprosthesis rather
than a THR for osteonecrosis once again stems from the fact that
initial changes are in the femoral head and not the acetabulum. It was
thus felt by some investigators that replacing only the femoral head
would be more conservative than THR (25,116). The early procedures used solid

press-fit femoral components that employed neither cement nor bone
ingrowth for fixation of the stem into the femoral canal. The heads
were sized rather crudely to the acetabulum because the importance of
an exact fit was not realized at that time. The long-term results with
these prostheses when used in younger patients with osteonecrosis were
generally poor (Fig. 125.37), although some good to excellent results did occur (Fig. 125.38).

More recently, endoprostheses have been rigidly fixed
into the femoral canal either by cement or by biological ingrowth.
Modular components came into use and both unipolar and bipolar devices
were employed. Initially it was felt that the use of a bipolar
component would yield better results than a unipolar in that there
would be less frictional wear on the acetabulum and thus a lower
incidence of acetabular failure. Unfortunately, other problems with the
bipolar prostheses developed, and there is currently little evidence of
the superiority of the bipolar over the unipolar device (Fig. 125.39).

Although a number of surgeons continue to use femoral
endoprostheses in patients with osteonecrosis, the current trend is
away from these devices and toward THR arthroplasty or other procedures.

When the results of endoprosthetic replacement, using
current techniques and components, are compared to THR, most reports
indicate better results with THR. Because the stem design and the
fixation of endoprostheses and total hips are now similar, it is
probable that the difference in results can be attributed to problems
within the acetabulum. A metallic femoral head places abnormal stress
on the acetabular cartilage and creates a situation that is far from
physiologic. This is not consistently well tolerated, even by normal
articular cartilage, and it has been noted that the acetabular
cartilage in osteonecrosis is no longer normal by the time arthroplasty
is needed (24,25,31,67,130).

Although THR is by far the most commonly used procedure
for the treatment of patients with advanced stages of arthritis and
other disorders of the hips, there has been a certain reluctance to
advise this procedure for patients with AVN. This is because these
patients are much younger than patients with degenerative joint disease
at

the
time that arthroplasty is required, and it is generally accepted that
the survivorship for THR in young patients is less than in older
patients. In addition, patients with osteonecrosis seem to have a
higher complication rate and earlier incidence of failure than patients
of a similar age but with other disorders of the hip (96).
In five reports published between 1981 and 1984 on THR arthroplasty in
younger patients, the mean incidence of revision was 13%, and results
were rated as good to excellent in 81% with a 5-year follow-up (3). Salvati and Cornell (105)
reviewed their experience with THR in patients with osteonecrosis from
1972 through 1975. They found that 11 of 28 arthroplasties failed at a
mean of 8 years (range, 5–10 years), for an overall failure rate of
37%. Of the remaining 18 THRs, 16 had good to excellent ratings. It was
their opinion that the failure rate in osteonecrosis was four times
greater than in osteoarthritis.

More recent literature, however, indicates improved
results at short- and medium-term follow-up times with more modern
components and techniques (21,25,40,43,54,62,71,92,99). In 1994, Piston et al. (100)
reported on 35 hips with AVN treated with porous-coated THR
arthroplasty. Follow-up ranged from 5 to 10 years with a mean of 7.5
years. They encountered one femoral and two acetabular revisions, for
an overall revision rate of 6%. There was, however, significant
remodeling in six hips and osteolytic reactions in six others. In 1997,
Garino and Steinberg (43) reported on 123 THAs
in patients with osteonecrosis. The follow-up was 2–10 years with a
mean of 4.6 years. All femoral stems and 71 sockets were fixed with
acrylic cement. Fifty-two sockets were designed for bone ingrowth. The
average Harris hip score improved from 45 points preoperatively to 92
points at the time of last follow-up. Of the 246 components used, 6
acetabular and 4 femoral prostheses (4%) were revised. The incidence of
revision for aseptic loosening was 2.5%. None of the noncemented
acetabular components were either radiographically loose or revised.
Fifty-two hips had a minimum follow-up of 5 years (mean, 6.6 years). In
this group, five components (5%) were revised and four (4%) were
radiographically loose. The revision rate for hybrid hips with a
minimum 5-year follow-up was 2%. These results seem significantly
better than reports of older series. The authors concluded that using
modern techniques and components in THR arthroplasty can give excellent
results over a period of 2–10 years, even in young patients with
osteonecrosis. They felt that this procedure was the treatment of
choice when reconstructive surgery was required (Fig. 125.40 and Fig. 125.41).

Resection arthroplasty, also referred to as a
Girdlestone pseudarthrosis, was used as a primary arthroplasty for both
septic and nonseptic hips prior to the advent of THR. Today it is used
most frequently as a salvage for the failed and usually infected THR or
endoprosthesis. Under certain selected circumstances, however, it can
be used as a primary arthroplasty. In the patient with end-stage
osteonecrosis whose pain and disability require some type of operative
intervention, but in whom prosthetic replacement or THR is
contraindicated, the resection arthroplasty may be a reasonable
alternative. This procedure may be considered to treat the
osteonecrotic hip complicated by sepsis, and the patient in whom THR
arthroplasty has a high potential for complications and failure.
Patients with chronic recurrent septicemia, those on renal dialysis,
and individuals with active sickle cell disease may fit into this
category. Resection arthroplasty might also be considered in the
individual who is felt to be too unreliable to take proper precautions
following THR, such as the chronic alcoholic, drug abuser, and mentally
retarded. It also might be used for the individual who is confined to a
bed/chair existence due to neurologic or medical problems and who
simply needs pain relief and satisfactory motion rather than stability
and the ability to ambulate.

Resection arthroplasty must be considered a compromise.
If this is clear from the outset to both the patient and physician, and
if there is an understanding of the advantages and disadvantages of
this procedure, then most patients will be satisfied with the outcome.
A resection arthroplasty will eliminate sepsis if present, provide pain
relief, give a good ROM, and not deteriorate with the passage of time.
At the same time, the joint will be unstable, the limb will be an
average of 1–1.5 inches short, and most patients will require both a
shoe lift and the use of at least one cane or possibly two crutches for
ambulation (Fig. 125.42) (27,121).
Resection arthroplasty may be used as a definitive procedure or as an
intermediate step, such as in the case of infection, until a THR can be
safely performed.

In the patient with significant pain and disability, who
is no longer responsive to conservative management, and in whom
prosthetic replacement or THR is contraindicated but stability of the
hip rather than motion is required, arthrodesis rather than resection
arthroplasty might be considered. Most of the indications described for
resection

arthroplasty
also apply to hip fusion. This procedure might also be considered in
the large, young, active man who desires a single “permanent” procedure
rather than a THR with the need for more than a single revision during
his lifetime. Before embarking on this procedure, the individual must
be fully aware of the possibility that the attempt at fusion may not be
successful, and of the permanent limitations that a fusion will impose.
At times the use of a pantaloon spica as a trial will help the patient
make an informed decision. The usual contraindications for hip fusion
apply here as well and include pathology in the knees, lumbar spine, or
contralateral hip. Because of the known high incidence of bilaterality
of osteonecrosis, it is essential that involvement of the opposite hip
be ruled out prior to embarking on a fusion. If, despite these
precautions, osteonecrosis does develop in the contralateral hip, THR
can be performed with a satisfactory outcome. The incidence of nonunion
in osteonecrosis is somewhat higher than in patients with other
disorders because of the necrotic bone present in the femoral head.

Osteonecrosis can involve both distal femoral condyles or the proximal tibia (85,89,91). This is probably the second most common site of involvement after the femoral head, with one report (91)
finding an incidence of knee involvement in 9% of patients with
osteonecrosis. Two entirely separate entities can affect this region:
One is known as spontaneous osteonecrosis of the knee (or SPONK), and
the other, which will be referred to as avascular necrosis, is the
disorder that can also affect the hip and other areas of the skeleton (Fig. 125.43).
SPONK occurs in patients over 55 years of age, and it typically
involves one femoral condyle or tibial plateau without other joint
involvement. The risk factors commonly found in osteonecrosis of the
hip have not been identified. AVN of the knee, on the other hand, is
typically found in patients younger than 45 years of age, may involve
several sites about the knee itself, is associated with hip involvement
90% of the time, has a different appearance on radiographs and MRI, and
is often associated with corticosteroid use or alcohol intake.

Involvement of the humeral head has been reported in 5% to 12% of patients with osteonecrosis (29,69,86).
Involvement is bilateral in 50% of the patients when associated with
corticosteroids. The initial treatment of this entity is symptomatic
and includes gentle exercises to preserve strength and ROM and the
avoidance of excessive stresses on the joint. There have been
relatively few reports of core decompression; however, the best results
have been obtained when the procedure was performed prior to humeral
head collapse (88). A recent study (69)
reported good to excellent results in 47 of 62 shoulders (76%) with a
mean follow-up of 8.5 years (range, 2–18 years). Some shoulders may
continue to function reasonably well even after humeral head collapse
has occurred. In some, however, the pain and disability is sufficient
to warrant either hemiarthroplasty or total shoulder replacement. These
are described in only a few reports, and there is mixed opinion as to
which of the two procedures should be utilized, even if radiographs do
not demonstrate changes in the glenoid (63).

Although osteonecrosis of the talus is most often seen
following fracture, atraumatic osteonecrosis is occasionally seen in
association with involvement of the hip and other regions (34,86,90,136).
Talar involvement is seen in approximately 2% of patients with
osteonecrosis. Eighty percent of the patients are women and the
disorder is noted bilaterally 50% of the time (Fig. 125.44).
Other bones of the foot and ankle are also involved occasionally. These
include the distal tibia, calcaneus, medial cuneiform, metatarsal
heads, and distal fibula (Fig. 125.45).
Treatment options include analgesics, bracing, or casting. Core
decompression of the talus has been performed in a small number of
cases. A recent report (90) noted satisfactory
results in 29 of 32 cases. If these measures fail and the talus goes on
to collapse, significant pain and disability may result, requiring
tibiotalar arthrodesis or possibly pantalar arthrodesis (86). Fusion can be difficult to obtain. In one report (86), there were 8 of 9 nonunions (89%) following the first attempt. In another report (136),
9 of 11 ankles (82%) eventually went on to fusion, but multiple
procedures were often required; these included autogenous bone
grafting, crossed cancellous screws, tibial struts, the insertion of a
blade plate, and prolonged casting.

Osteonecrosis of the adult elbow is encountered infrequently (70). One recent report (68) described eight adult patients with atraumatic osteonecrosis of the elbow. These

included four men and four women with a mean age of 35. Four patients
had an associated autoimmune disease, all had involvement of other
joints, and all were on corticosteroid therapy. Four cases involved the
capitellum, three the trochlea, and three the lateral condyle. Four
elbows presented in the precollapsed stage radiographically and
responded well to nonoperative management. Two patients with collapse
failed nonoperative treatment and underwent core decompression for pain
relief. One elbow with advanced disease required a total elbow
arthroplasty. In contrast to the pediatric form of osteonecrosis,
Panner’s disease, osteonecrosis of the adult elbow can lead to endstage
arthritis.

Osteonecroses involving various bones in children and
adolescents were previously grouped together with a variety of skeletal
disorders collectively known as the osteochondroses. These include
Perthes disease, Freiberg’s disease, Koöhler’s disease, and Kienboöck’s
disease. With the exception of Perthes disease, these conditions are
rarely encountered (120).