HETEROTOPIC OSSIFICATION AND CHARCOT NEUROARTHROPATHY

Formation of mature lamellar bone in nonosseous tissue
is termed heterotopic ossification (HO) or heterotopic bone.
Histologically, it differs from soft-tissue calcification in terms of
its osteoblastic activity. Myositis ossificans is histologically
identical to HO but, by definition, forms in inflammatory muscle.
Periarticular calcification does not manifest the radiologic features
of organized bone and is usually limited to distinct structures, such
as the collateral ligaments. HO is metabolically active and
histologically similar to native bone, with increased numbers of
osteoblasts and osteoclasts.

The pathogenesis of HO is not fully understood, although
it is clearly associated with certain clinical situations. HO can
result in severe limitation of motion and ankylosis. Chalmers et al. (19)
proposed three conditions necessary for HO formation: an osteogenic
precursor cell, an inducing agent, and an environment conducive to
osteogenesis. Mesenchymal cells, which have the ability to
differentiate to osteogenic stem cells, are found in the soft tissues
surrounding joints. Osteoinductive substances are probably released as
a result of the insult, and they may cause a proliferation of
mesenchymal cells. Growth factors, such as bone morphogenetic proteins
(BMPs), are potential inducers of undifferentiated mesenchymal cells to
proliferate and differentiate into cartilage and bone and may play a
role in the formation of HO (125,126).

The preoperative identification of patients who are at
risk has been investigated with bone scans and with laboratory tests,
such as erythrocyte sedimentation rates (ESRs) and alkaline phosphatase
levels. The results of these tests do not always correlate well with
the development of the condition Theoretically, an immediate
postoperative change

in
the serum level of substances related to osteoblast activity, such as
osteocalcin, circulating growth factors, or the bone isoenzyme of
alkaline phosphatase, might aid in the selection of high-risk patients
needing prophylaxis. Most investigators have observed an increase in
the serum alkaline phosphatase up to 3.5 times normal, beginning in the
first month of HO and peaking at about 3 months (84).
However, by the time these levels are noted to be elevated, the process
is well established and prophylaxis is unlikely to be useful.

There has not been any conclusive evidence in the
literature with regard to which test is the most accurate predictor of
HO formation and maturation. Despite this, most surgeons usually allow
a triple-phase bone scan and alkaline phosphatase to return to baseline
before resecting heterotopic bone. Computerized tomography is often
useful for operative planning.

The most common classification scheme used clinically is
the Brooker classification, which is based on the extent of HO seen on
the AP radiograph of the hip (Table 124.1). Wright et al. (131)
tested the reliability and validity of this grading system. The
intraobserver reliability between two surgeons was 86% and 77%, and the
interobserver reliability was 68%. Although it is useful in determining
the radiographic appearance of the HO, the Brooker classification does
not address its functional effect.

HO prevention is based on preventing bone induction. The
mainstays of this prevention are nonsteroidal anti-inflammatory agents
(NSAIDs), ionizing radiation, and meticulous handling of surgical
tissue.

The inhibitory action of NSAIDs on HO is probably
related to their inhibition of the activity of prostaglandin
synthetase. The resulting decrease in local concentrations of
osteoactive prostaglandins may be key to the inhibition of formation of
heterotopic bone. The change in environment tends to inhibit the
proliferation and maturation of pluripotential osteogenic precursors
and limits the development of heterotopic bone. Indomethacin has been
shown in an animal model to be effective only if administered in the
initial phase of bone induction (81). In
studies of animals, NSAIDs have been shown to delay the healing of bone
after local trauma. Rats that had been fed an aspirin-rich diet
revealed a dose-dependent delay in fracture healing (4). Dahl (24)
first reported on the use of indomethacin after total hip arthroplasty.
Its use as an analgesic showed reduction of HO compared with the use of
a placebo. Since that time, numerous studies have been conducted to
compare the different anti-inflammatory agents with varying dosing and
length of treatment. Because of the potential side effects of
dyspepsia, bleeding, and inhibited bone healing, much emphasis has
focused on finding the lowest dose and shortest length of treatment,
without compromising the results of suppression of HO formation.
Coumadin has been proven to be ineffective (103).

Numerous authors have investigated the use of radiation
therapy. Most of their work centers around the use of such therapy
after total hip replacement (23,25,43,47,48,51,59,71,74,78,86,106,107,127).

Careful handling of tissue during any surgical
procedure, especially around the hip and elbow, is also of importance
to minimize any trauma and subsequent inflammation. The role of proper
surgical technique in the formation of heterotopic bone is difficult to
quantitate.

Several unproved methods of prophylaxis have been studied. Riska and Michelsson (92)
had encouraging results with the transplantation of fat after excision
of HO, theorizing that filling the gap created in the soft tissues
would prevent recurrence. Biphosphonates have been employed because
they inhibit the transformation of amorphous calcium phosphate into
hydroxyapatite and thereby the mineralization of osteoid matrix.
Unfortunately, mineralization occurs when the medicine is discontinued.
Thomas and Amstutz (120) found no difference
when they compared biphosphonates to a placebo 2 years after the
operation. Furthermore, the medication is expensive and can adversely
affect calcium and phosphate metabolism, leading to conditions that
include osteomalacia.

HO is a frequent occurrence after a total hip arthroplasty (Fig. 124.1).
Certain patients are at a higher risk for developing clinically
significant HO: those with ankylosing spondylitis, osteoarthritis with
large acetabular osteophytes,

Forestier’s
disease, Paget’s disease, bilateral disease, extensive operative
trauma, formation of a hematoma, and those who have developed HO after
a previous total hip arthroplasty. Men with osteoarthritis are at
higher risk for HO than women (93). Ahrengart and Lindgren (3)
concluded that men with osteoarthritis, men with fractures, women older
than 65 with osteoarthritis, and women with hypertrophic osteoarthritis
should be considered for prophylactic treatment.

The roentgenographic incidence of HO after cemented
total hip arthroplasty without prophylactic treatment approximates 30%,
with a reported range from 5% (21) to 90% (99) (Table 124.2).
Of those who develop HO, 3% to 10% have pain and decreased range of
motion as a direct result of the HO. Maloney and Krushell (72)
found a significant increase in the severity of HO with uncemented
total hip arthroplasty as compared with hybrid arthroplasty. The rate
of clinically significant HO in cementless total hip arthroplasty is
approximately 30% (62).

HO usually becomes visible on radiographs 3 to 4 weeks
after operation and matures by 3 to 6 months. Histologically, HO is of
the membranous type and has a woven structure that is later transformed
into a more organized trabecular bone. Six months to 1 year after the
operation, the rate of remodeling decreases, and the tissue tends to
stabilize (Fig. 124.2).

HO is common after open reduction and internal fixation
of acetabular fractures. It is seen in only about 5% of patients
treated conservatively (87) but in as many as 90% after extensile approaches to acetabular fractures (66). Severe HO has been reported in approximately 24% of patients (56).
Certain types of exposures, including the Kocher-Langenbeck, extended
iliofemoral, and triradiate, have resulted in a high occurrence of HO,
whereas the ilioinguinal exposure has not. The extended iliofemoral
approach has the highest prevalence of HO (66).
Postoperative HO prophylaxis is not required for fractures treated
through the ilioinguinal exposure. Prophylaxis with either NSAIDs,
single-dose local field radiation, or a combination should be used
routinely early after open reduction and internal fixation with
approaches other than the ilioinguinal approach. Moed and Letournel (78)
reviewed 53 patients who underwent the posterior or extended
iliofemoral approach and were treated with indomethacin and
irradiation. Indomethacin was given for 3 weeks, and radiation therapy
was given, either in divided or a single dose. Forty-four fractures
showed no HO, and 10 had grade I fractures. There was no difference
between radiation doses.

From studies that have mostly focused on the hip,
information about the incidence, pathophysiology, and treatment has
been applied to the upper extremity. HO around the elbow is associated
with local injury, electrical and thermal burns, and neurologic injury
to the brain and spinal cord. The most frequent cause of heterotopic
bone about the elbow is direct trauma (Fig. 124.3) (46).
There appears to be a direct correlation between the frequency of HO
and the magnitude of the injury. HO is usually seen in the collateral
ligaments when it is associated with trauma.

HO develops in approximately 1% of patients who have
sustained burns, 3% of patients with a local injury of the elbow, and
11% of patients with a head injury (36,88,121).
The elbow is the most frequent site after a burn. When a fracture
combined with a dislocation occurs, the incidence rises to 15% to 20%.
The incidence of HO about the forearm is approximately 2% to 4%, with
high-energy and open fractures most at risk. HO around the elbow,
associated with head injury, has been investigated by Garland and
O’Hollaren (38). The incidence of HO increases fifteenfold when the neurotraumatized patient has an associated elbow injury.

The Brooker classification is insufficiently sensitive to describe the functional loss of motion in the elbow (74). Hastings and Graham (46) divided HO in the elbow and forearm after all types of trauma into three classes (Table 124.6). A treatment plan should be based on severity of HO (Table 124.7).

Early range of motion is essential to prevent stiffness.
Early after injury, there is no completely effective prophylaxis. The
difficulty with the elbow is its smaller size, compared with the hip;
therefore, a small amount of bone results in more functional
limitations. Bridging bone in one plane is amenable to excision with
good results, whereas synostosis in two or more planes has poorer
results. Most surgeons wait 12 to 18 months before excision, with
recurrence demonstrated frequently in patients (Table 124.7) (37,88,97).
There is an advantage to early excision, particularly in those with
neurologic injury. This allows for early rehabilitation, enhancing
function and independence. In a recent review of eight patients who had
excision of HO at an average of 7 months followed by radiation therapy,
there was no recurrence and significant improvement in arc of motion (74). This compares with previous reports of recurrence rates up to 30% (37).
Because the preoperative testing for maturation of HO is not reliable
for recurrence, individualize the decision for excision.

The most common site for HO about the elbow is the
posterolateral aspect of the elbow, from the olecranon to the lateral
humeral condyle. The elbow can be approached by various incisions
dictated by the location of HO (see Chapter 1).
Postoperatively, early motion is imperative, with the continuous
passive motion device used commonly. Initiate irradiation and
anti-inflammatory medications within 72 hours of surgery.

The complication of HO around the joints of patients was first described in patients with spinal cord injuries in the

late 1800s, whereas its association with traumatic brain injury was first reported by Roberts in 1968 (98). The incidence of HO in adults who have had an injury to the brain ranges from 11% to 76% (100). Garland et al. (34,35,36 and 37)
have studied the traumatic brain injury patient in depth. The frequency
of HO in the upper extremity parallels that in the lower extremity. The
incidence of HO in the shoulder is similar to that in the elbow (Fig. 124.4).
Involvement of the upper extremities is more frequent in patients who
have had injuries to the brain than in those with spinal cord injuries (34).

The incidence of HO in patients with spinal cord injury is reported to be 20% to 30% (34). Limitation of joint motion occurs in 18% to 37% of patients who develop HO (11,52). HO develops more often in patients with complete paralysis (102).
Patients with spinal cord injuries usually present with HO anteromedial
to the hip in the vicinity of the lesser trochanter, whereas in
traumatic brain injury patients, the location is more variable. New
bone formation in the head-injured patient is usually periarticular,
whereas patients with spinal cord injury develop HO at a distance from
the joint.

Whereas the hip is the most common location for HO in
neurologically injured patients, the elbow most often becomes
ankylosed. New bone formation in the hip and elbow correlate with the
position of the extremity. It is thought that diffuse axonal injury
often produces spasticity, which is known to be associated with HO (1).
Patients with spastic quadriplegia and those with low levels of
recovery are particularly prone to the formation of HO and recurrence
after resection.

The exact timing for resection is better defined when
the lesion occurs after a spinal cord injury. Resection is usually
performed when serial alkaline phosphatase and bone scan have returned
to baseline. Brain-injured patients have a high variability of
neurologic involvement and severity, as well as variability in anatomic
location of HO. In a review of 23 patients with resection of HO in 37
joints, a recurrence of HO was found in 56%. The recurrence was closely
linked to the severity of cognitive deficit and physical disability (37).

Postoperatively, patients are irradiated or treated with
NSAIDs. Biphosphonates are usually selected for prophylaxis if there is
a contraindication to NSAIDs. The size of the initial HO mass may be
the most important factor in predicting postoperative recurrence (116).

The site of the HO at the elbow or hip dictates the
approach for resection. The shoulder is usually amenable to
manipulation and only uncommonly requires resection. Most resections
are carried out through a deltopectoral approach. The hip has three
areas of possible development of new bone: anterior, inferomedial to
the hip joint and distal to the lesser trochanter (associated with
adductor spasticity), and posterior to the femoral head (associated
with flexion contractures). The hip may be approached anteriorly or
posteriorly.

There are two theories of pathogenesis: neurotraumatic
and neurovascular. The neurotraumatic theory proposed by Johnson in
1967 (53) involves repetitive trauma sustained
by a joint unable to sense pain. In cats, a denervated immobilized limb
does not produce a destructive arthropathy, but it can occur when the
cats are allowed to freely walk (16). This does
not account for the sensation of pain and neuroarthropathic changes
that can occur after acute fractures. It also does not explain the
gross destruction of joints that do not bear weight.

The neurovascular theory, suggested by Allman and Brower (5),
refers to a sympathetic dysfunction with persistent hyperemia and
active bone resorption by osteoclasts. As early as 1927, Leriche noted
that lesions of the sympathetic nerve led to an increase in blood flow (5). Blood flow may actually increase up to five times the normal rate in diabetic neuropathy.

Joint destruction in the neuropathic joint is probably
brought on by a combination of factors that include damage to the
nociceptors of the joint and the periarticular tissues. This injury may
initially be caused by mechanical trauma, chronic infections such as
syphilis, or metabolic diseases such as diabetes associated with loss
of pain and proprioception. This leads to stretched ligaments, and lax
and subluxed joints. The activity of peptides such as substance P,
calcium gene–related peptide, and vasoactive intestinal peptide (VIP)
could result in increased vascularity and inflammation, contributing to
further joint destruction. Substance P can enhance the cellular
synthesis of collagenase and prostaglandin-E; activate T lymphocytes,
monocytes, and neutrophils; and take an active part in inflammation.
Glycolization of the supporting ligaments in the foot due to high
glucose levels in the blood can increase the stiffness and
cross-linking of the collagen fibers, making them brittle.

The initial pathologic changes occur in the underlying
bone and cartilage. Recurrent effusions occur due to hyperplasia of the
synovium. The articular cartilage is slowly destroyed by a pannus,
which helps distinguish Charcot’s joints from other forms of
osteoarthritis. Histology of the fragments of cartilage and bone in
synovium shows deeply embedded tissue with normal metachromatic
staining. In contrast, in osteoarthritis the fragments are deposited
superficially and have staining characteristics suggestive of
degeneration (68).

Excess osteoclast activity is believed to be responsible
for the early bone changes associated with Charcot’s neuroarthropathy
in diabetes mellitus. In a recent study by Gough et al. (41)
the serum carboxyterminal telopeptide of type 1 collagen, a marker of
osteoclastic bone resorption, had significantly increased levels in the
acute Charcot foot. The lack of an associated increase in osteoblastic
activity supports the idea that excess osteoclast activity is a feature
of the early stages of Charcot’s neuroarthropathy.

High-peak plantar pressures have been identified as a
risk factor for ulceration in diabetics. In a study of 164 diabetic
patients, higher peak plantar pressures occurred in those with
Charcot’s arthropathy and those with neuropathic ulcers than in those
without. Measuring plantar pressures might be an effective means of
screening large numbers of patients (8).

The most common presentation of neuroarthropathy in a
diabetic patient is a single, painless, swollen, deformed foot and no
specific history of trauma. The joint has a large effusion, ligamentous
laxity, and increased warmth, but the peripheral white blood count
(WBC) and the ESR are normal. The swelling and warmth may suggest an
infection or pseudoseptic arthritis. The synovial fluid is clear, straw
colored, and viscous, although it can be bloody. Usually, there is
evidence of peripheral neuropathy in the presence of normal or bounding
pulses. Joint changes frequently precede the neurologic deficit, and as
many as one third of patients actually have pain at presentation (57).

The deep tendon reflexes at the knee are absent in a majority of
patients. Some patients present with nontraumatic dislocations or rapid
joint destruction without any known trauma. The spine is involved in 6%
to 21% of cases (40). Anand et al. (7)
reported on a patient with neuropathic spinal arthropathy in
Charcot-Marie-Tooth disease. Nonspiral fractures of the long bones
should heighten concern, because considerable force would normally be
necessary to cause such an injury. Rule out other disease processes
such as avascular necrosis (AVN), infection, crystal-induced arthritis,
hemophilia, and rheumatoid arthritis (RA).

Early radiographs may demonstrate changes very similar
to those observed in early osteoarthritis. Nontraumatic dislocations
may be an early sign. With time, there is radiographic evidence of
joint distention caused by fluid, hypertrophic synovitis, osteophytes,
and subluxation. The normal architecture of the joint is lost, with
dislocation, fragmentation, attempted repair by osteophytes, and
sclerosis. Various degrees of bone absorption have been observed, with
margins resembling surgical amputation coupled with periarticular
speckled calcification in the soft tissue. Infection usually produces
indistinct margins of bone as opposed to the sharp ones seen in
neuroarthropathy.

Osteomyelitis typically is secondary to infected ulcers
or cellulitis, with direct infectious spread to and through the cortex
of the adjacent bone. Establishing the diagnosis of osteomyelitis in
the foot may be difficult (see Chapter 116). Use imaging studies to confirm the clinical suspicion of osteomyelitis. Lipman et al. (67)
studied the diagnostic efficacy of combined three-phase bone
scintigraphy and indium-111–labeled WBC scintigraphy, magnetic
resonance imaging (MRI), and conventional radiography in detecting
osteomyelitis of the neuropathic foot. MRI was found to be comparable
to conventional radiography. In the midfoot and hindfoot, three-phase
or WBC scans were more specific than either MRI or conventional
radiography. MRI was most accurate in the forefoot.

Charcot neuroarthropathy in the hip is rare. As with any
joint, the critical element in treating a patient with a neuropathic
hip joint is to establish the diagnosis (Fig. 124.8) (9).

If the joint remains painless and functional, no
intervention is required. If pain is present and function is impaired,
pursue conservative treatment with protected weight bearing as long as
possible before considering any type of surgical procedure. Treatment
with hip arthrodesis results in a high rate of nonunion; historically,
it has been as high as 100% (53). Total hip
arthroplasty also has a high failure rate, especially in patients with
significant neurologic findings or ataxia. The overwhelming evidence in
the literature is that these patients do poorly with any type of
prosthetic replacement. The number of patients treated with total hip
arthroplasty reported in the literature is small, reflecting the
relatively low incidence of neuropathic hips. They have universally
done poorly (93,114).
This limited evidence strongly indicates that prosthetic replacement of
the Charcot hip joint is contraindicated. Of the dozen or so cases
reported in the literature, only one patient was stated to have done
well with a total hip joint arthroplasty (114).
The remainder did poorly even when treated with a broad range of total
hip designs. Resorting to a resection arthroplasty may be the only
viable solution in the treatment of the painful hip in these patients.

The treatment of hip fractures in a Charcot hip joint also remains problematic. Johnson (53)
found that over 50% of fractures of the femoral neck in diabetics
developed Charcot’s joints. Many times, the neuropathic nature of this
fracture is not recognized, and hip nailing or a prosthesis is
inserted, only to result in failure. Internal fixation has been
documented to do as poorly as prosthetic replacements. Treatment with
restricted weight bearing may be the most prudent option with a delayed
resection arthroplasty if the joint subsequently becomes painful.

Complex fractures of the acetabulum in patients with diabetes are difficult to treat. Berg (10)
reported on three patients who developed Charcot’s changes after open
reduction internal fixation for fractures of the acetabulum. None had
any previous evidence of a neuropathy.

Charcot’s neuroarthropathy in the knee has previously
been reported as the most common neuropathic joint problem secondary to
syphilis and one of the most difficult to treat (53).
There is tremendous leverage, predisposing to instability, and little
soft-tissue envelope to cushion the joint. When the knee is painless,
bracing is the treatment of choice. Johnson (53)
recommends a corrective osteotomy to prevent shearing stress, followed
by bracing for severely damaged knees, particularly with bilateral
involvement. If only one knee is involved and destruction is severe,
fusion is indicated. Bracing is imperative to promote a solid union.

Fractures in diabetics are best treated by prolonged immobilization and splinting until there is definite radiographic

evidence of repair. The potential for complications with immobilization
is high. Arthrodesis, usually the treatment of choice in advanced
disease of the knee, may fail because of nonunion and breakage of
fixation devices. Drennan (26)
found a 55% clinically successful arthrodesis in his series of patients
with knee fusion. Arthroplasty is generally considered contraindicated
in Charcot’s joints. Loosening and breakage of the prosthesis is more
likely to occur secondary to lack of proprioception, although the
results of nine total knee arthroplasties were considered excellent in
eight knees after a 3-year follow-up (113).
Correct ligamentous balancing, use of long-stemmed components, and
meticulous surgical detail are imperative for good long-term results.
Despite these good reported early results, total knee arthroplasty
should be approached in these patients with great caution because the
long-term results undoubtedly will not be as favorable.