Neurogenic Heterotopic Ossification Heterotopic ossification (HO) following neurological injury is defined as soft tissue bone formation which is most often seen after spinal cord or head injury and reported to occur after numerous other neurological insults. It can result in ankylosis of associated joints restricting patient mobility. Lesions present with signs of local inflammation. Evidence of ossification is noted on Technetium 99 bone scan and then on X-ray. Serum alkaline phosphatase may be elevated. Although it resembles the histopathology of myositis ossificans traumatica, its pathogenesis remains unclear. Current research favors exercise of joints to attenuate loss in range of motion. Ethane-l-hydroxy-l, I-diphosphonic acid, a diphosphonate analogue of pyrophosphate, is the only medical treatment specifically indicated in heterotopic ossification. It has not been as effective as first hoped. Nonsteroidal anti-inflammatory drugs, warfarin, and X-radiation therapy hold promise as treatments. Surgical excision to remove ankylosing ectopic bone Ims been plagued by reoccurrence. Studies to better elucidate the etiology of HO, accelerate diagnosis of the condition, and accurately determine its risk of reoccurrence are needed. Further controlled studies of the various treatments are also warranted.

The authors would like to thank a number of professionals including Drs. John E. Eisele, Ze' ev Grosswasser, Roy Weiss, Solomon Weiss, and E. Ungar-Paul, orR for their inspiration and encouragement in preparing this review.

Scott M. Paul, MD Children's Rehabilitation, Upper Valley Medical Centers Miami County, OH, and, Wright State University Dayton,OH

Janet R. Barlow, BS Wright State UniveTSity School of Medicine Dayton,OH

Heterotopic ossification is bone formation in soft tissues. Four major clinical entities have been observed. Myositis ossificans progressiva is a genetic disorder in which microdactyly is associated with ossification in voluntary muscles which may progress to ankylosis of related joints. Myositis ossificans traumatica occurs in hematoma resulting from local i~jury. Heterotopic ossification also occurs after orthopedic surgical procedures such as total hip arthroplasty. Heterotopic ossification following neurological injury (HO) is bone formation related to neurological insult. HO has been described most commonly in spinal cord iI~ury (SCI) and traumatic brain injury (TEl) but has been reported as a sequelae of many other disorders (Table 1). The nomenclature ascribed to the lesion in varied (Table 2). This attests to the present lack of understanding of its etiology. HO may be found as an incidental finding on roentgenological examination. However, it may limit an associated joint's range of motion (ROM) or it may progress to ankylosis ofthe joint. It, therefore, has serious rehabilitation implications. The limitations it may impose on ROM can impede posture, alignment, and mobility. These effects can then place the patient at higher risk of developing secondary complications such as decubitus ulcers. There may NeuroRelwbil 1993; 3(3):66-78 Copyright © 1993 by Andover Medical.

Neurogenic Heterotopic Ossification

Table 1. Neuropathologies in Which Heterotopic Ossifications Have Been Reported. 7,IO,IOB

Traumatic Spinal Cord Traumatic Brain Injury Toxic Anoxia Carbon Monoxide Poisoning Malignant Neurolepsis Infections Tetanus Poliomyelitis Syphilis Meningitis Myelitis Encephalomyelitis Intracranial Abscess Epidural Abscess Polyneuritis

Spinal Cord Anomalies Myelodysplasia Syringomyelia Degenerative Diseases Multiple Sclerosis Vascular CVA Spinal Cord Thrombosis Neoplastic Cerebral Tumors Spinal Cord Tumors

Table 2. Terms Used to Describe Neurogenic

Heterotopic Ossification in the Literature.

Ectopic Bone Ectopic Ossification Heterotopic Bone Heterotopic Ossification Heterotopic Ossification in Paraplegia Metaplastic Ossification Myositis Ossificans in Paraplegia Myositis Ossificans Neurotica

Neurogenic Ossifying Fibromyopathy Neurogenic Ossifying Fibromyositis Paraosteoarthropathy Periarticular New Bone Formation Pathological Ossification Soft Tissue Calcification in Paraplegia

be long-term risk of sarcomatous transformation of the lesion, as has been reported in rare cases of myositis ossificans traumatica and progressiva. I,2

CLINICAL PRESENTATION Foci of heterotopic ossification first present with signs of inflammation. Discovered is an area of soft tissue swelling which may be warm, erythematous, and painful (in patients with sensory sparing). Researchers have also observed a systemic febrile reaction,3-5 swollen joint,4,6,7 and, several days after onset, a circumscribed firm mass within the

67

edema. 5,B The general signs of inflammation may be seen with much less frequency in the pediatric population. 9 In adult SCI patients these signs have been observed as early as 19 days and as late as 7 months postinjury,IO averaging about 60 days postinjury.ll,12 The onset appears to be delayed in children with SCI. The average onset in this group is reported as 14 months after injury, with a range from 3 to 36 months. 9 Similar to adults with SCI, the onset of signs in adult TBI patients has been from 1 to 7 months postinjury-mean onset 100 days. 13 The differential diagnosis of soft tissue swelling in patients with neurological insults must also consider thrombophlebitis, cellulitis, septic arthritis, periostitis, local trauma, hematoma, and fracture. Some have proposed serum levels of alkaline phosphatase (AlkP) as a discriminating test and, indeed, increased enzyme levels have been consistently observed early in heterotopic ossification. 4,7,12,14-16 A recent study of SCI patients 17 has shown a significant increase in AlkP in SCI patients with HO compared to those without HO six weeks after the onset of injury. Garland et a1. 7 point out, however, that AlkP will be raised in any patient with healing fractures, and Szabo et al. lB note that elevated AlkP can be a natural result of adolescent growth. These facts diminish the reliability of this enzyme level, especially in an adolescent patient with associated bony trauma. X-ray studies also have aided little to early diagnosis. Early in the progression of the lesion, X-ray will only reveal signs of soft tissue swelling. 4,B,12 However, Technetium 99 bone scanning is a reliable parameter in early diagnosis.1 2,19,20 The most sensitive method of bone scanning appears to be the three-phased Technetium 99 scan, 14 which detects increased vascularity in the area of the lesion 2 to 4 weeks before static bone scan is positive. As the lesion of HO matures, a progressive decrease in the passive ROM of the neighboring joint may be noted. (This may be the only sign in children with SCI. 9) Signs of swelling will diminish, patchy areas of calcification will be detected by X-ray, and Technetium scan will show increased uptake in the area of the lesion (suggestive of the diagnosis of tumor). Further X-ray studies reveal a

68

NEUROREHABILITATION / SUMMER 1993

progression to immature ossification with no distinct margins and an absence of linear trabeculae and, finally, to an area resembling mature bone. 8 Nicholas 16 suggested four stages in the maturation of Ho. First observed is soft tissue swelling, decreased joint ROM, and increased serum AlkP. In the second stage, these signs remain and X-ray examination begins to show signs of ectopic bone. In stage three, signs of swelling recede leaving a palpable bony mass, AlkP remains elevated, and X-ray signs progress. In stage four only the decreased ROM and X-ray signs remain. HO is usually thought to be mature by 12 to 18 months postinjury.21 Blood chemistries (including AlkP) and serial X-ray scans have not been useful in judging maturity ofHO. 4 However, serial Technetium 99 bone scans have been helpful in establishing maturation. 22 ,23 There may also be a role for magnetic resonance imaging. One recent study of ossification of the posterior longitudinal ligament of the spinal cord 24 found progression of ossification in this unrelated condition to be more frequently related to low-, iso-, or high-MRI signal intensity than to no signal, as in cases where progression was complete. The incidence of HO in adult SCI patients has varied in the literature from 8%10 to 53%.25 The incidence in pediatric SCI patients has been reported in one study to be much lower, 3.3%.10 In TBI patients, the reported incidence has also varied, from 11%7 to 76.8%.1 3 As in SCI, the incidence in children appears to be lower, from 3% to 22.5%.26,27 Of SCI patients with HO, between 20%8 and 46%3 have been reported to have had lesions which significantly reduce ROM in or entirely ankylosed one or more joints. In TBI patients, 16% of joints with HO detected on X-ray had significant decrease in ROM or ankylosis. 7 The ossifications in SCI patients have all presented in sites below the level of neurological lesion. The joint most commonly affected is the hip.21,28,29 Interestingly, TBI patients have been reported to present with lesions both ipsilateral and contralateral to areas of paralysis. 3o In TBI patients the hip 7 and the shoulder30 have been varyingly reported to be most frequently affected. Ossifications have also been found around the

elbow, knee, the spine, 8 and the proximal interphalangeal joints ofthe hand. 31 ,32 Reports of involvement of the wrist, ankle, feet, or other joints of the hands could not be found in the literature of neurogenic Ho. There have been increasing numbers of reports of entrapment of surrounding anatomic structures as a secondary complication in patients with BO. It is unclear whether this is related to increased clinical vigilance or an evolution in the presentation ofHo. These complications have left clinicians with difficult questions not yet resolved. The most troubling of these complications has been related to the venous system. There have been at least five reports of extrinsic compression of the deep venous system, mimicking deep venous thrombosis (DVT).1,22,33-36 One additional report 37 was unable to definitively differentiate between extrinsic compression and DVT. HO is related to increased vascularity and may worsen with bleeding. This may result in increased risk to HO patients from the use of anticoagulants. At the same time, this population is also at risk for DVT and this risk may be exacerbated by the decrease in venous outflow related to extrinsic compression of the veins by HO. To date, there is no clear choice for treatment. At least one group34 has used a Greenfield vena cava filter to minimize the risk of pulmonary embolism, rather than risk anticoagulation. Other structures which have been reported as compressed or entrapped by HO include peripheral nerves 35,38 and lymphatic channels. 35 Computerized tomography has been useful in most of the above studies in establishing a definitive diagnosis and describing the extent of involvement. One37 of the studies also suggested that radionuclide venogram be combined in one study with triple-phase Technetium 99 bone scan (by introducing the Technetium into a pedal vein) to differentiate between extrinsic venous compression and DVT.

PATHOGENESIS Studies of the histopathology of HO have been undertaken to help elucidate the etiology of the process. Specimens of recent onset (5 days to 8

Neurogenic Heterotopic Ossification

weeks after diagnosis) disease studied by Rossier et al. 4 described a firm lesion arising at the muscle periphery along with an area of edematous connective tissue between muscle and lesion which matured into a cleavage plane. The sections showed vascular regions of edematous, fibroblastic, osteogenic connective tissue with few lymphocytes. Areas of newly formed woven bone undergoing osteoclastic erosion were seen, along with areas of active osteogenesis and accompanying active chondrogenesis. Patches of randomly oriented hydroxyapatite crystals were occasionally found. The periphery consisted of edematous, nonossified connective tissue in which some atrophic muscle fibers could be found. Long-standing (7 to 46 months after clinical onset) disease specimens 4 showed the adjacent muscle mass to be well defined and not atrophied, sometimes firmly attached to the mass and sometimes traveling through grooves or tunnels in the mass. Atrophic muscle fibers were incorporated into the mass as in the early sections. The mass was composed of a 1 mm thick cortex, with a tightly latticed spongiosa without well-defined borders. Bone was lamellar, harder than normal, with occasional Haversian systems. There were a rarity of osteoclasts and osteoblasts, but still more than in normal bone. De novo adipose marrow was present, with little or no hematopoiesis. The joint capsule was not broached, but where the mass ankylosed a joint there was disruption of the normal bone cortex with loss of defined limits between pathological and normal bone. Amorphous calcium phosphate predominated in the analysis of the mineral content of the early lesions. Later, there were more hydroxyapatite crystals and, after approximately 30 months, the mineral content approached that of normal bone. The average crystal size was less than that of normal bone. Based on these observations, Rossier et al 4 suggested that ossification proceeds from edematous connective tissue to such tissue with foci of calcification, to maturation of calcification, to ossification. This is similar to the histopathology of myositis ossificans traumatica, although HO need not arise in muscle, arising also in aponeurotic tissue and the tissue separating muscle mass.

69

The studies of Dejerine, Cellier, and Dejerine 39 were perhaps the earliest to clearly document the pathology of HO and differentiate it from lesions attributable to sepsis. They hypothesized a modification of both inter- and intramuscular fibrous tissue in paralyzed limbs, followed by microscopic hemorrhages, whose occurrence was not related to the level ofthe cord lesions but to their severity. They postulated as initiating and propagating the ossification an undefined systemic factor which altered bone metabolism in paraplegia and a number of local factors in the affected limb, including: a pool of available calcium from the adjacent skeleton, soft tissue edema, circulatory alterations, pH changes, and the presence of a mesenchymally derived cell capable of osteoblastic activity. Over 80 years later, the etiology of HO is still not clearly defined. However, groups have studied a number of the factors Dejerine and Cellier proposed, as described below. Rossier et al. 4 studied circulatory alterations by angiography and comparisons of arterial and venous blood gases. They found signs of increased vascularity and early venous return, 2 to 3 months post-SCI, suggesting the presence of arteriovenous shunts. The signs seemed to be proportional to the extent of injury, so they postulated that the angiographic changes were related to the degree of damage to the sympathetic system. They concluded, however, that these changes were secondary and not an etiological factor. Conversely, Major et al. 40 theorized that stasis in the paravertebral venous plexus of Batson (combined with increased local calcium due to disuse osteoporosis) may be the primary factor precipitating calcium salts and initiating growth of hydroxyapatite crystals. The marked vascular blush and blood pool observed around the hips in Technetium scan by Freed et al. 14 three weeks before bone scan was positive, would seem to support this theory. Nechwatal 15 found that SCI patients admitted for rehabilitation had evidence of acidosis. He suggested that the correction of the acidosis and the concurrent hypercalcemia resulted in calcium precipitation in the form of Ho. Researchers have reviewed the HLA system to find evidence of a genetic predisposition which might suggest the systemic factor proposed above.

70

NEUROREHABILITATION / SUMMER 1993

Markers HLA B-18 41 and HLA B-2742 have been reported to be associated with Ho. However, other investigators 43 - 45 have studied multiple HLA loci, including B-18 and B-27, without finding any significant difference between controls and HO patients. Sazbon et al. 46 performed serum growth hormone (GH) stimulation tests on a group of comatose patients. They found an increase in GH in response to stimulation only in patients who developed HO, which would suggest that the proposed systemic factor may be neurohormonal. There has been a recent explosion in research on human tissue growth factors. This has lead researchers to investigate whether any of these factors may be the circulating factor postulated by Dejerine, Cellier, and Dejerine. 39 A number of growth factors have been demonstrated to influence osteoblast activity (see Table 3). These factors include fibroblast growth factor (FGF), transforming growth factor I3(TGF-I3), and the bone morphogenic proteins (types 2A, 2B, and 3 of which are related to TGF-I3).47 Levels of FGF and TGF-13 in bone excised from patients with heterotopic ossification after total hip replacement have been found to be increased significantly, compared to age-matched controls. 48 A recent study by Bidner et al. 49 indirectly demonstrated evidence of increased activity of growth factors in serum of patients who sustained TBIs. However, the study did not identify any specific factor and, as pointed out by Garland,5o it did not establish a causal relationship between this evidence and Ho. As research on cellular and biochemical factors continues, other groups have attempted to retrospectively identify clinical factors which may be related to the formation or progression of Ho. The demographic factors of age greater than 30 Table 3.

Human Growth Factors with Influence on Osteoblast Activity.

Osteogenin Osteoinductive factor Platelet derived growth factor Fibroblast growth factor Epidermal growth factor

Prostaglandins Insulinlike growth factors Transforming growth factor 13 Bone morphogenic proteins

years and completeness oflesion were found in the study of Lal et al. 51 to be significantly related in adult SCI patients with Ho. In another study of TBI patients, 13 no relationship was found between age or length of coma and Ho. A recent study in pediatric TBI patients by Hurvitz et al. 27 found age greater than 11 years or length of coma greater than seven days to be significant. Data by Citta-Pietrolungo, et al. 26 seems to support the relationship with coma. Spasticity has been implicated as a significant factor in some groups of adult SCI and pediatric TBI patients. 26 ,51 Hurvitz's group27 found no relationship to spasticity in their study. Bouchard and D'Astous 52 found postoperative HO to be more common in children with meningomyelocele than in those with cerebral palsy, which could argue against spasticity as a significant factor. The presence of decubitus ulcers was found to be significant by Lal's group.51 However, as first pointed out by Dejerine, Cellier, and Dejerine,39 this may represent a different pathological process with a septic origin. Physical therapy has been implicated as an etiological factor in HO. As in myositis ossificans traumatica, it has been thought that the lesion might result from local trauma. Soule proposed 53 that passive ROM exercises against spastic resistance induce Ho. However, all of the patients in this study had decubiti at some time, so the ossifications might have had septic origins. Hossack and King,54 studying four cases, suggested that ossification occurred only after forceful ranging by overzealous therapists. In a case study, Caughey and Gray55 suggested that strenuous activity in the pool and gym may be etiologically significant. Supporting the previous observations of Hardy and Dickson, 10 Wharton and Morgan,21 studying 90 patients, concluded that there was no relationship between spasticity and the development of ossification. It was noted that hip ankylosis occurs in extension rather than in flexion, as would be expected if spasticity was a factor. In addition, it was indicated that failure to continue passive manipulation of the joints may have been the dominating factor leading to ankylosis in some patients. This has been supported by recent studies, as will be discussed.

Neurogenic Heterotopic Ossification

TREATMENT Physical 1. Exercise Physical therapy in the treatment of 110 has been fraught with controversy. Stover et al. 29 reviewed 70 articles and found 10 different therapeutic techniques with some authors for and against each. Some textbooks 56,57 have implicated range of motion exercise as causitive or have recommended joint immobilization as part of the treatment protocol. However, Wharton and Morgan,21 Stover et al. 29 and Wharton 8 have all shown that passive ROM exercises reduced loss of joint motion. Wharton and Morgan 21 additionally suggested that forceful joint manipulation under anesthesia may reduce ankylosis, a theory which is supported by other studies. 57 ,58 2. Physical Modalities Evaluations have revealed that diathermy and ultrasound have no beneficial results. 21 There have been numerous studies 6o..64 in recent years demonstrating the efficacy of X-radiation in preventing heterotopic ossification after total hip arthroplasty. It has been postulated 65 that by interfering with DNA synthesis, radiation therapy blocks the differentiation of pluripotential mesenchymal cells to osteoblasts. Single-dose postoperative regimens with radiation dosages as low as 700 centigray 64 have been found to be effective in patients after total arthroplasty. One animal model study66 has revealed findings which suggest that preoperative radiation may be effective. The use of X-radiation in neurological HO has not been as extensively studied. To date, its use has been advocated only in TEl patients with HO requiring surgical resection. Garland 67 recommends a total dose of 1000 centigray given in fractions of 200 centigray beginning on the day of the surgery. Increased risk of sarcomatous transformation has been a concern in the use of therapeutic radiation. Long-term studies 68 ,69 seem to indicate that this is unlikely at the dosages and schedules recommended above.

71

Surgical In cases ofHO which result in significantly diminished ROM or ankylosis, excision of the ectopic bone has been the only hope for the restoration of joint mobility. Many protocols 4,8,7o have advised incomplete wedge resection. Garland 67 has recently advocated resection of the entire bridge of bone since reankylosis could occur earlier with wedge resection if HO were to reoccur. Conscientious hemostasis and postoperative suction drainage of the surgical wound is common to both methods. Surgical complications include hematoma formation, wound breakdown, infection, sinus formation, osteomyelitis, and, most frustratingly, reoccurrence and reankylosis. Reoccurrence has been thought to be related to HO maturity. For this reason, an accurate measure of maturity of the lesion (i.e., Technetium 99 bone scan) has been important in scheduling operation. Although the presence of increased activity on bone scan has been related to reoccurrence,12,19 the converse has not been borne out. Bone scans without increased activity have not guaranteed the absence of reoccurrence. 71 A major thrust, then, of postoperative treatment is prevention of reoccurence. Passive range of motion of the effected joint is part ofthe postoperative management in each of the studies cited above. The postoperative prophylactic use of radiation therapy, disodium etidronate, and nonsteroidal antiinflammatory drugs is discussed in other sections of this review.

Medical 1. Diphosphonates The search for a specific medical therapy which might aid in the treatment of ectopic calcifications, including HO, derives from the development of water softeners in the 1930s. Low concentrations of condensed phosphates inhibited the deposition of calcium carbonate in hard water. In the 1960s, Fleish and Neuman 72 demonstrated that collagen promotes crystallization of dissolved calcium phosphate and that pyrophosphate was an inhibitor of this precipitation. Alkaline phosphatase, they postulated, regulates calcification by acting on pyrophosphate. The knowledge that there is

72

NEUROREHABILITATION 1 SUMMER 1993

increased pyrophosphate and a failure of calcification in hypophosphatasia supported the hypothesis. This promulgated research to find a substance analogous to pyrophosphate which would be absorbed intact from the gut and not destroyed by AlkP. Ethane-l-hydroxy-l, I-diphosphonic acid (EHDP; disodium etidronate) is a diphosphonate analogue of pyrophosphate. No enzyme has been found which can catalyze its hydrolysis. EHDP absorption in the gut is affected by stomach contents, especially calcium. One-half of the absorbed compound is excreted in urine, one-half goes to bone. The T1/2 of bone retention is approximately 4 weeks in rats. 73 Figure 1 shows the similarity between EHDP, pyrophosphate, and the condensed phosphate water softeners. Initial studies ofEHDp72-75 revealed a number of in vitro and in vivo (animal) actions including: (1) inhibition of hydroxyapatite dissolution in vitro and bone resorption in tissue culture, (2) inhibition of calcium phosphate crystal formation in vitro and aortic calcification in vivo, (3) in vitro block of the conversion of amorphous calcium phosphate into hydroxyapatite crystals, and, in high doses, (4) conversion of hydroxyapatite into a colloid state and formation of polynuclear complexes in the presence of calcium. It was postulated that these effects resulted from EHDP's strong affinity to both calcium ions and calcium phosphate. Further animal studies have suggested both intracellular and extracellular actions of EHDP. Woltgens et al. 76 found the same affinity for both pyrophosphate and EHDP in hamster pyrophosphatase. This would suggest that EHDP acts extracellularly as a competitive inhibitor. Plasmans et al.,77 however, found atypical cells which displayed characteristics of both osteoclasts and osteoblasts suggestive of an intracellular effect. Rath et al. 78 compared EHDP and polyvinyl phosphate (PVP), a phosphonate whose actions are confined to the extracellular space by its high-molecular weight. Although PVP did not effect induced bone formation, EHDP did inhibit bone formation, strongly suggestive of an intracellular action. More recent studies have suggested that EHDP has a direct cytotoxic effect. 79

o II

OH

I o=pI

o-p

I

OH

OH

OH

I

-o-p=o

I

OH

n

Condensed Phosphate (Polyphosphate)

OH

OH

OH

OH

I I o=p-o-p=o I Pyrophosphate

OH

CH3

OH

OH

OH

OH

I I I o=p-c-p=o I Ethane-l-hydroxy-l , l-diphosphonic acid

Figure 1. The chemical structures of condensed phosphate, pyrophosphate, and ethane-I-hydroxyI, I-diphosphonic acid. Uncertainty is not limited to the pharmacology of EHDP. Efficacy of the drug even in animal models is unclear. Plasmans et al. 77 and H u et al. 79 have found reversal of EHDP's inhibition of mineralization after the drug was discontinued. McClure 80 found that EHDP had no effect on chondrification nor later ossification in his model. In spite of uncertainty concerning the mode of action and efficacy in animal studies, EHDP has been approved for treatment in Paget's disease, HO associated with SCI (20mglkglday for two weeks followed by lOmg/kg/day for ten additional weeks) and after total hip arthroplasty. Studies have also evaluated its use in myositis

Neurogenic Heterotopic Ossification

ossificans progressiva 73,81 and immobilization hypercalcemia. 82,83 Stover et al. 5 presented the first double-blind study of EHDP in SCI patients. They reported a significantly lower incidence of HO with a significantly lower X-ray grade in the EHDP-treated group as compared to the placebo group. Subsequent studies 3. 11 ,14,84-86 (summarized in Table 4) have demonstrated mixed results. Most of these studies have focused on SCI and only a few studies have looked at patients with TBI (with dosing schedules similar to those used in SCI). A number of these researchers 3,84,86 have found that the desired results are not sustained after EHDP is discontinued, similar to the findings of the Plasmans et al. 77 and Hu et al.7 9 animal studies. Side effects reported in EHDP treatment have included an increase in serum phosphate (which diminishes posttreatment),3,5 abdominal distress, nausea, vomiting, and diarrhea. 3 ,5,85 In the treatment of Paget's patients, more serious side effects have been reported, including an increase in bone pain 73 and traumatic fractures of the radii. 87 Safety of EHDP in children has not been established. There has been a case report of rickets in a pediatric TBI patient 88 and inhibition of ossification and weakness in a child with myositis ossificans progressiva. 89 Table 4. Ho.

Results of EHDP Therapy in Studies of

Ret No.

Patient Type

5 86

SCI SCI

3

SCI

14

SCI

11

SCI

85 84

TBI Mixed; neurological and nonneurological

Results

Decreased incidence Reocurrence after drug withdrawn Decreased incidence but some reoccurrence after drug withdrawn Equivocal No effect on progression or new appearance but no ankyloses Decreased incidence No reossification while under treatment; reossification after drug withdrawn

73

2. Nonsteroidal Anti-Inflammatory Drugs HO initially presents with signs of inflammation, both clinically and on a microscopic level. A diverse group of drugs, including aspirin, indomethacin, and ibuprofen, are known as nonsteroidal anti-inflammatory drugs (NSAIDs). Extensive studies have shown them all to act on the prostaglandin pathways and decrease or limit inflammation throughout the body. It is not surprising, then, that researchers have been interested in the effects of NSAIDs on HO. There have been a number of studies investigating the effects of indomethacin on animal models of bone formation. The drug has been shown to inhibit cyclo-oxygenase and, therefore, reduce the production of prostaglandin E 2.90 Other studies 91 have shown that prostaglandin E2 is a mediator ofHo. Indomethacin has been shown to also have a number of effects at the microscopic level in addition (or perhaps in relation) to these findings. It has been shown in animal models to inhibit local proliferation of preosteoblasts, 92,93 inhibit differentiation of preosteoblasts into osteoblasts,94 decrease the rate of fracture repair,95,96 and inhibit Haversian remodeling. It has also been hypothesized 47 that the NSAIDs may decrease bone formation indirectly by inhibiting angiogenesis. The earliest clinical study of the effect of NSAIDs on 1I0 is attributed to Dahl 97 in 1974. He observed decreased heterotopic ossification after total hip arthroplasty with treatment with indomethacin. His findings have been confirmed in the postarthroplasty population in double-blind, placebo controlled studies. Indomethacin at a dose of 25 mg, TID for six weeks, postoperative, has been shown to significantly decrease incidence and severity of postarthroplasty heterotopic ossification. 98 These effects have also been shown to persist on follow-up of three years or more. 99 Other studies have demonstrated similar results with the use of ibuprofen and naproxen ("standard dosage" for 1 to 3 weeks),IOO diclofenac (one preop dose of 75 mg, followed by 50 mg TID for six weeks),101 and flurbiprofen (200 mg QD for three weeks).102 Indomethacin has also been shown to be effective in prophylaxis for heterotopic ossification after acetabular fracture on the same

74

NEUROREHABILITATION / SUMMER 1993

dosage schedule (see Table 5).103 There have been little complications of treatment reported in these studies. Side effects seem to be limited to nausea, vomiting, and stomach discomfort, without report of significant gastrointestinal bleeding. Unfortunately, the literature on the use of NSAIDs in HO after neurological injuries has been limited to case reports and very small nonblinded studies. The reports are promising, however. Two groups31,104 have each reported a single case where indomethacin was used after excision of HO in adults with TBI at the dosage used in the postarthroplasty studies. 98 ,99 One case lO4 is somewhat confounded by the concurrent administration of corticosteroids. There has also been one study on aspirin, at a dose of 60 mg/kg for six weeks. 105 They reported minimal or no recurrence after excision of HO in eight pediatric TBI patients and no progression after use in two additional children with HO after TBI. No reports in the SCI population or reports of prophylactic NSAID use before the diagnosis of HO could be found on this review. 3. Warfarin The anticoagulant, warfarin, functions as a Vitamin K inhibitor. Vitamin K is known to catalyze the production of the osteocalcin, a bone protein thought to be involved in bone matrix mineralization. 106 One retrospective studyl07 of patients with SCI found that no patients receiving warfarin for treatment of DVT developed clinically detected HO, despite an overall 15% incidence ofHO in the study group.

Table 5. NSAIDs Shown to Decrease the Incidence of Postarthroplasty Heterotopic Ossification. Drug Indomethacin Ibuprofen Naproxen Diclofenac Flurbiprofen

Dosage 25 mg, TID "standard dosage" "standard dosage" 75 mg preop., then 50 mgTID 200 mgQD

Duration of Treatment 6 weeks 1-3 weeks 1-3 weeks 6 weeks 3 weeks

CONCLUSIONS The aforementioned studies indicate that further research into the pathological mechanisms of HO are necessary in order to effectively develop better early diagnosis, measures of risk of reoccurrence and more rational treatment. An animal model of neurological injury which reliably produced HO after injury would be an especially valuable aid. Further study should include search for increased levels of bone growth factors in patients with HO, perhaps by the use of tagged antibodies for these factors. The efficacy of ROM exercise after diagnosis has not been fully established. Neither has its causative role. However, the alternative, immobilization, places patients at greater risk of ankylosis from HO. EHDP, although promising in early studies, seems to delay but not prevent HO. However, this ability to delay the progression of HO may make EHDP valuable when other treatments are contraindicated due to other patient complications. Further study, including attention to the possibility of different responses in patients with SCI versus those with TBI, is warranted. Based on the studies of the related problem of heterotopic ossification after arthroplasty, NSAIDs and radiation therapy both hold promise as treatments in neurological Ho. The short-term use of NSAIDs appears to be effective and safe. Well structured, double-blind, placebo-controlled trials of the prophylactic use of NSAIDs to prevent HO must be developed for children and adults with TBI and SCI. Warfarin may also have a role in prophylaxis. However, further animal and human study is first required, especially given the risks of bleeding while taking warfarin. Although radiation therapy is not attractive as a preventive measure, studies of its use in conjunction with surgical excision of HO should also be considered. The prognosis of surgery can be unpredictable. The literature is still unclear about proper timing and extent of excision required. It is clear that surgery should be followed by a careful postoperative program. This program should at least include range of motion exercise. There is probably benefit to including the use of radiation therapy and NSAIDs. Until further research is available,

Neurogenic Heterotopic Ossification

the choice and timing of surgical excision (like most interventions in patients with neurological i~ury) is probably best guided by its impact on the patient's functional outcome versus the risks of

75

surgery and/or recurrence and should be entertained only ifits goal is to significantly improve the patient's independence or minimize secondary complications.

REFERENCES 1. Shanoff LB, Spira M, Hardy SB. Mysositis ossificans evolution to osteogenic sarcoma, report of a histologically verified case. Am j Surg 1967; 113:537-541. 2. Thyss A, Michiels JF, Caldani C, et al. Sarcome osteogenique des tissus mous apres myosite ossificate post-traumatique. Presse Medicale 1984; 13:1333-1334. 3. Finerman GAM, Stover SL. Heterotopic ossification following hip replacement or spinal cord injury. Two clinical studies with EHDP. Metab Bone Dis Rel Res 1981; 4 and 5:337-342. 4. Rossier AB, Bussat P, Infanti F, et al. Current facts on para-osteoarthropathy (POA). Paraplegia 1973; 11:36-78. 5. Stover SL, Hahn HR, Miller JM. Disodium etidronate in the prevention of heterotopic ossification following spinal cord injury (preliminary report). Paraplegia 1976; 14: 146-156. 6. Buschbacher R, Coplin B, Buschbacher L, et al. Noninflammatory knee joint effusions in spinal comr-injured and other paralyzed patients, four case studies. Amj Phys Med Rehabil1991; 70:309-312. 7. Garland DE, Blum CE, Waters RL. Periarticular heterotopic ossification in head injured adults. Incidence and location. j Bone joint Surg (Am) 1980; 62: 1143-1146. 8. Wharton Gw. Heterotopic ossification. Clin Orthop 1975; 112:142-149. 9. Garland DE, Shimoyama ST, Lugo C, et al. Spinal cord insults and heterotopic ossification in the pediatric population. Clin Orthop 1989; 245:303-310. 10. Hardy AG, DicksonJW. Pathological ossification in traumatic paraplegia. j Bone joint Surg (Br) 1963; 45:76-87. 11. Garland DE, Alday B, Venos KG, et al. Diphosphonate treatment for heterotopic ossification in spinal cord injury patients. Clin Orthop 1983; 176:197-200. 12. Hsu D, Sakimura I, Stauffer ES. Heterotopic ossification around the hip joint in spinal cord injured patients. Clin Orthop 1975; 112: 165-169. 13. Sazbon L, Najenson T, Tartakovsky M, et al. Widespread periarticular new-bone formation in long-term comatose patients. j Bone joint Surg (Br) 1981; 63:120-125.

14. Freed JH, Hahn H, Menter R, et al. The use of the three-phase bone scan in the early diagnosis of heterotopic ossification (HO) and in the evaluation of didronel therapy. Paraplegia 1982; 20:208-216. 15. Nechwatal E. Early recognition of heterotopic calcifications by means of alkaline phosphatase. Paraplegia 1973; 11:79-85. 16. Nicholas JJ. Ectopic bone formation in patients with spinal cord injury. Arch Phys Med Rehabil 1973; 54:354-359. 17. Wittenberg RH, Peschke U, Botel U. Heterotopic ossification after spinal cord injury. j Bone joint Surg (Br) 1992; 74:215-218. 18. Szabo Z, Ritzl F, Chittima S, Laszlo GA, et al. Bone scintigraphy of myositis ossificans in apallic syndrome. Eur j Nucl Med 1982; 7:426-428. 19. Suzuki Y, Hisada K, Takeda M. Demonstration of myositis ossificans by 99mTc pyrophosphate bone scanning. Radiology 1974; 111:633-664. 20. Tyler JL, Derbekyan V, Lisbona R. Early diagnosis of myositis ossificans with Tc-99m diphosphonite imaging. Clin Nucl Med 1984; 9:256258. 21. Wharton GW, Morgan TH. Ankylosis in the paralyzed patient. j Bone joint Surg (Am) 1970; 52:105-112. 22. Jensen LL, Halar E, Little JW, et al. Neurogenic Heterotopic Ossification. Am j Phys Med Rehabil 1988; 66:351-363. 23. Tanaka T, Rossier AB, Hussey RW, et al. Quantitative assessment of para-osteo-arthropathy and its maturation on serial radionuclide bone images. Radiology 1977; 123:217-221. 24. Sakamoto R, Ikata T, Murase M, et al. Comparative study between magnetic resonance imaging and histopathologic findings in ossification or calcification ofligaments. Spine 1991; 16:1253-1261. 25. Venier LH, Ditunno JH. Heterotopic ossification in the paraplegic patient. Arch Phys Med Rehabil 1971; 52:475-479. 26. Citta-Pietrolungo TJ, Alexander MA, Steg NL. Early detection of heterotopic ossification in young patients with traumatic brain ir~ury. Arch Phys Med Rehabill992; 73:258-262.

27. Hurvitz EA, Mandac BR, DavidoffG, et al. Risk factors for heterotopic ossification in children

76

28. 29. 30.

3l. 32. 33.

34. 35.

36. 37. 38.

NEUROREHABILITATION / SUMMER 1993

and adolescents with severe traumatic brain injury. Arch Phys Med Rehabil 1992; 73:459-462. Blane CE, Perkash I. True heterotopic bone in the paralysed patient. Skeletal Radio11981; 7:21-25. Stover SL, Hathaway CJ, Zeiger HE. Heterotopic ossification in spinal cord injured patients. Arch Phys Med Rehabil 1975; 56: 199-204. Mendelson L, Grosswasser Z, Najenson T, et al. Periarticular new bone formation in patients suffering from severe head injuries. Scan j Rehab Med 1975; 7:141-145. Asselmeir M, Light TR. Heterotopic paraarticular ossification of the proximal interphalangeal joint. j Hand Surg 1992; 17A:154-157. Spencer RE Heterotopic ossification in a finger following head injury. j Hand Surg (Br) 1991; 16:217-218. Paul SM, Hammerman A, Stoddard E. Heterotopic Ossification causing extrinsic compression of the deep venous system of the lower extremity (abstr). Arch Phys Med Rehabil 1988; 69:760. Silver K. Concurrent existance of heterotopic ossification and thrombophlebitis. Am j Med 1990; 88:318. Varghese G, Williams K, Desmet A, et al. Nonarticular complication of heterotopic ossification: a clinical review. Arch Phys Med Rehabi11991; 72:1009-1013. Yarkony GM, Lee MY, Green D, et al. Heterotopic ossification pseudophlebitis. Am j Med 1989; 87:342-344. Bradleigh LH, Perkash A, Linder SH, et al. Deep venous thrombosis associated with heterotopic ossification. Arch Phys Med Rehabil1992; 73:293-294. Brooke MM, Heard DL, deLateur BJ, et al. Heterotopic ossification and peripheral nerve entrapment: early diagnosis and excision. Arch Phys Med Rehabi11991; 72:425-429.

39. Dejerine M, Cellier A, Dejerine Y. Para-osteoarthropathies des paraplegiques par lesion medullaire. Etude anatomique et histologique. Rev Neurol 1919; 26:399-407. 40. Major P, Resnick D, Greenway G. Heterotopic ossification in paraplegia: a possible disturbance of the paravertebral venous plexus. Radiology 1980; 36:797-799. 41. Minaire P, Betuel H, Girard R, et al. Neurologic injuries, paraosteoarthropathies, and human leukocyte antigens. Arch Phys Med Rehabil 1980; 61:214-215. 42. Larson J, Michalski J, Collacott E, et al. Increased prevalance of HLA-B27 in patients with ectopic ossification following traumatic spinal cord injury. Rheumatol Rehabil 1981; 20:4. 43. Garland DE, Alday B, Venos KG. Heterotopic ossification and HLA antigens. Arch Phys Med Rehabill984; 65:531-532.

44. Hunter T, Dubo HIC, Hildahl CR, et al. Histocompatability antigens in patients with spinal cord i~jury or cerebral damage complicated by heterotopic ossification. Rheumatol Rehabil 1980; 19:97-99. 45. Weiss S, Grosswasser Z, Ohri A, et al. Histocompatibility (HLA) antigens in heterotopic ossification associated with neurological injury. j Rheumato11979; 6:88-9l. 46. Sazbon L, Sack J, Najenson T, et al. Growth hormone and periarticular new bone formation-a casual relationship? Scand j Rehab Med 1982; 15:43-46. 47. Sawyer JR, Myers MA, Rosier RN, et al. Heterotopic ossification: clinical and cellular aspects. CalcifTissue Int 1991; 49:208-215. 48. Thies RS, Bauduy M, McQuaid D. Bone morphogenic protein alters W-20 stromal cell differentiation in vitro. j Bone Mineral Res 1990; 5:S150. 49. Bidner SM, Rubins 1M, Desjardins JV, et al. Evidence for a hunoral mechanism for enhanced osteogenesis after head i~jury. j Bone joint Surg (Am) 1990; 72:1144-1149. 50. Garland DE. Letter to the editor. j Bone joint Surg (Am) 1992; 74:152-153. 5l. Lal S, Hamilton BB, Heinemann A, et al. Risk factors for heterotopic ossification in spinal cord injury. Arch Phys Med Rehabil 1989; 70:387-390. 52. Bouchard J, D'Astous J. Postoperative heterotopic ossification in children: a comparison of children with spina bifida and with cerebral palsy. JCC 1991; 34:454-456. 53. Soule AB. Neurogenic ossifYing fibromyopathies: a preliminary report. j Neurosurg 1945; 2:485-497. 54. Hossack DW, King A. Neurogenic heterotopic ossification. Medj Australia 1967; 1:326-328. 55. Caughey JE, Gray P. Heterotopic calcification complicating infectious polyneuritis. NZ Med j 1992; 95:38-39. 56. Solter RB. Textbook of DilOrders and Injuries of the Musculoskeletal System, 2nd ed. Baltimore: William and Wilkins, 1983. 57. Trombly CA, Scott Ao. Occupational Therapy for Physical Dysfunction, Baltimore: William and Wilkens, 1983. 58. Garland DE, Razza BE, Walters RL. Forceful joint manipulation in head injured adults with heterotopic ossification. Clin Orthop 1982; 169: 133-138. 59. Wenner SM. Heterotopic ossification of the shoulder following head injury. Clin Orthop 1986; 212:231-236. 60. Anthony P, Keys H, McCollister-Evarts C, et al. Prevention of heterotopic bone formation with

Neurogenic Heterotopic Ossification

61.

62. 63.

64.

65.

66.

67. 68. 69. 70.

71.

72. 73. 74.

75.

early post operative irradiation in high risk patients undergoing total hip arthroplasty: comparison of 10.00 Gy vs. 20.00 Gy schedules. Intj Rad One Biol Phys 1987; 13:365-369. Ayers DC, Pellegrini VD, McCollister-Evarts C. Prevention of heterotopic ossification in highrisk patients by radiation therapy. Clin Orthop 1991; 263:87-93. Coventry MB, Scanlon pw. The use of radiation to discourage ectopic bone. j Bone joint Surg (Am) 1981; 63:201-208. Pellegrini VD, Konski AA, Gastel JA, et al. Prevention of heterotopic ossification with radiation after total hip arthroplasty. j Bone joint Surg (Am) 1992; 74:186-199. Warren SB, Brooker AE Excision of heterotopic bone followed by irradiation after total hip arthroplasty. j Bone joint Surg (Am) 1992; 74:201-210. Evarts CM, Ayers DC, Puzas JE. Prevention of heterotopic bone formation in high-risk patients by postoperative irradiation. In: Brand R, ed. The hip: proceedings ofthe 14th open scientific meeting of the Hip Society. St. Louis, MO: CV Mosby, 1986. Kantorowitz DA, Miller GJ, FerraraJA, et al. Preoperative versus postoperative irradiation in the prophylaxis of heterotopic bone formation in rats. Int j Rad One Biol Phys 1990; 19: 1431-1438. Garland DE. Surgical approaches for resection of heterotopic ossification in traumatic braininjured adults. Clin Orthop 1991; 263:60-70. Brady LW. Radiation-induced sarcomas of bone. Skeletal Radiol 1979; 4:72. KimJH, Chu FC, Woodard HQ, et al. Radiationinduced soft-tissue and bone sarcoma. Radiology 1978; 129:501. Stover SL, Niemann KMW, Tulloss JR. Experience with surgical resection of heterotopic bone in spinal cord injured patients. Clin Orthop 1991; 263:71-77. Garland DE, Hanscom DA, Keenan MA, et al. Resection of heterotopic ossification in the adult with head trauma. j Bone joint Surg (Am) 1985; 67:1261. Fleish H, Neuman WE Mechanisms of calcification: role of collagen, polyphosphates, and phosphatase. Amj Physioll961; 200:1296-1300. Russel RGG, Smith R. Diphosphonates. Experimental and clinical aspects. j Bone joint Surg (Br) 1973; 55:66-86. Fleish H, Russell RGG, Francis MD. Diphosphonates inhibit hydroxyapatite dissolution in vitro and bone resorption in tissue culture and in vivo. Science 1969; 165:1262-1264. Francis MD, Russell RGG, Fleish H. Diphosphonates inhibit formation of calcium phosphate crystals in vitro and pathological calcification in

77

vivo. Science 1969; 165:1264-1266. 76. Woltgens JHM, Bonting SL, Bijvoet OLM. Influence of sodium ethane-l-hydroxy-l, I-diphosphonate and MG2+ on the inorganic pyrophosphatase in calcifying hamster-molars. Isr j Med Sci 1971; 7 :406. 77. Plasmans CMT, Kuypers W, Sloof TJJH. The effect of ethane-l-hydroxy-l, I-diphosphonic acid (EHDP) on matrix induced ectopic bone formation. Clin Orthop 1978; 132:233-243. 78. Rath NC, Dimitrevich S, Anbar M. Effect of polyvinyl phosphonates and ethane hydroxy diphosphonate on mineralization of ectopic bone. Chem Biol Interactions 1984; 48:339-347. 79. Hu HP, Kuijpers W, SlooffTJJH, et al. The effect ofbiphosphonate on induced heterotopic bone. Clin Orthop 1991; 272:259-266. 80. McClure J. The effect of diphosphonates on heterotopic ossification in regenerating Achilles tendon of the mouse.j Path 1980; 139:419-430. 81. Geho WB, Whiteside JA Experience with disodium etidronate in diseases of ectopic calcification. In: Frame B, Parfitt AM, Duncan H, eds. Clinical Aspects of Metabolic Bone Disease. Amsterdam: Exerpta Medica, 1973: pp. 506-511. 82. Hagg E, Eklund M, Torring 0. Disodium etidronate in hypercalcemia due to immobilization. Br Medj 1984; 288:607-608. 83. Merli GJ, McElwaine GE, Adler AG, et al. Immobilization hypercalcemia in acute spinal cord injury treated with etidronate. Arch Intern Med 1984; 144:1286-1287. 84. Nollen AJG. Effects of ethylhydroxydiphosphonate (EHDP) on heterotopic ossification. Acta Orthop Scand 1986; 67:358-361. 85. Spielman G, Gennarelli TA, Rogers CR. Disodium etidronate: its role in preventing heterotopic ossification in severe head injury. Arch Phys Med Rehabill983; 64:539-542. 86. Stover SL, Niemann KMW, Miller JM. Disodium etidronate in the prevention of post operative recurrance of heterotopic ossification in spinalcord injury patients.j Bone joint Surg (Am) 1976; 58:683-688. 87. Siris ES, Canfield RE,Jacobs TP, et al. Long-term therapy of Paget's disease of bone with EHDP. Arthritis Rheum 1980; 23:1177-1184. 88. Chiodo AE, Nelson V. Rickets associated with etidronate use in a pediatric head injured patient. Arch Phys Med Rehabil 1983; 64:539-542. 89. RogersJG, DorstJp, Geho WB. Use and complication of high-dose disodium etidronate therapy in fibrodysplasia ossificans progressiva. j Pediatr 1977; 91:1001-1014. 90. Ho SSw, Stem PJ, Bruno LP, et al. Pharmacological inhibition of prostaglandin E2 in bone and its effect on pathological new bone formation in a rat

78

NEUROREHABILITATION / SUMMER 1993

burn model. Trans OrtMp Res Soc 1988; 13:536. 91. Stern PJ, Bruno LP, Hopson CN. Skeletal deformities after burn injuries. An animal model. Trans Orthop Res Soc 1985; 10:175. 92. Nilsson OS, Bauer lICF, Brosjo 0, et al. Influence of indomethacin on induced heterotopic bone formation in rats. Importance oflength oftreatment and of age. Clin Orthop 1986; 207:239-245. 93. Tornkvist H, Nilsson OS, Bauer FCB, et al. Experimentally induced heterotopic ossification in rats influenced by anti-inflammatory drugs. Scan j Rheumatol1983; 12:177-180. 94. Rosenoer LML, Gonsalves MR, Roberts WE. Indomethacin inhibition of preosteoblast differentiation associated with mechanically induced osteogenesis. Trans Orthop Res Soc 1989; 14:64. 95. Allen HL, Wase A, Bear WT. Indomethacin and aspirin: effect of non-steroidal anti-inflammatory agents on the rate of fracture repair in the rat. Acta Orthop Scan 1980; 51 :595-600. 96. RO J, Sudmann E, Marton PF. Effect on indomethacin on fracture healing in rats. Acta Orthop Scan 1976; 47:588-599. 97. Dahl K. Kliniske Observasjoner. In: Symposium about hip arthrosis, Blindern, Norway, MSD, 1974. 98. Schmidt SA, Kj;crsgaard-Andersen P, Pedersen NW, et al. The use of indomethacin to prevent the formation of heterotopic bone after total hip replacement. j Bone joint Surg (Am) 1988; 70:834-838. 99. Schmidt SA, Kj;crsgaard-Andersen P. Indomethacin inhibits the recurrence of excised ectopic ossification after hip arthroplasty. Acta Orthop Scan 1988; 59:593.

100. Sodemann B, Persson P-E, Nilsson OS. Nonsteroid anti-inflammatory drugs prevent the recurrence of heterotopic ossification after excision. Arch Orthop Trauma Surg 1990; 109:5356. 101. Wahlstrom 0, Risto 0, Djerf K, et al. Heterotopic bone formation prevented by diclofenac. Acta Orthop Scarul1991; 62:419-421. 102. Hoikka V, Lindholm TS, Eskola A. Flurbiprofen inhibits heterotopic bone formation in total hip arthroplasty. Acta Orthop Trauma Surg 1990; 109:224-226. 103. McLaren AC. Prophylaxis with indomethacin for heterotopic bone. j Bone joint Surg (Am) 1990; 72:245-247. 104. Ritter MA. Indomethacin: an adjunct to surgical excision of immature heterotopic bone formation in a patient with a severe head injury. Clin Orthop 1987; 10:1379. 105. Mital MA, Garber JE, StinsomJT. Ectopic bone formation in children and adolescents with head injuries: its management. j Ped Orthop 1987; 7:83-90. 106. Gallop P, Lian J, Hauschka P. Carboxylated calcium binding proteins and vitamin K. N Engl j Med 1980; 302:1460-1466. 107. Buschbacher R, McKinley W, Buschbacher L, et al. Warfarin in prevention of heterotopic ossification. Amj Phys Med Rehabil1992; 71:86-91. 108. Peylan J, Goldberg I, Retter J, et al. Articular ossification after malignant neurolepsis. Acta Orthop Scarul1987; 58:284-286.