Introduction

Neurogenic heterotopic ossification (NHO) is a frequent complication in spinal cord injury (SCI). It is characterized by the formation of new extra osseous (ectopic) bone in soft tissue surrounding peripheral joints in patients with neurologic disorders. In SCI patients, the incidence ranges from 10–53% depending on the study design, the methods of detection (radiologically or by clinical symptoms), and the diagnostic criteria used.1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28 Moreover, in the early literature, no attempt was made to separate cases with ectopic bone formation associated with trauma or decubitus from true ectopic ossification not associated with a history of (local) trauma or infection. In non-traumatic myelopathies the incidence of NHO seems less compared to traumatic SCI ranging somewhere between 6–15%.6,14,19,29,30 NHO is less common in children than in adults with an incidence generally reported between 3–10%.31,32 The clinical symptoms in children do not differ from those reported in adults, but are less pronounced compared to adult patients. Moreover, spontaneous regression of NHO is more often reported in children and young adults than in older adults.31,32

The clinical spectrum of NHO ranges from an incidental finding on X-rays to severe limitation of the range of motion and even complete ankylosis of peripheral joints. In the majority of cases, the extent of NHO is minimal and only diagnosed as a radiographic finding. However, in 20–30% of the SCI patients clinically significant NHO is present with a reduction of joint range of motion, whereas in 3–8% of the SCI patients ankylosis develops.7,8,10,12,19,20,22,23,24,27,29,33 NHO always occurs below the level of the SCI, most commonly at the hip (70–97%).4,5,7,8,10,12,14,15,18,19,22,24,26,28,34,35,36,37 Other body segments including the knee, elbow, shoulder, hand and spine (in decreasing incidence) may be involved.7,10,12,14,22,23,33,36,38,39,40,41,42 Incidentally, NHO has been reported after soft tissue surgery in SCI patients.43,44

Although NHO may develop even several years after SCI, it is generally diagnosed between 1 to 6 months post-injury with a peak incidence at 2 months7,8,10,12,19,22,45,46,47,48,49,50 Although NHO may begin well before the clinical signs become evident, the initial signs are often seen within the first 3 weeks after SCI.3,4,7,9,39,51 The most common clinical findings are a decreased joint range of motion and a peri-articular swelling due to interstitial oedema of the soft tissues.1,7,8,10,30,36,39,45,52,53 In patients with sensory sparing, the first symptom may be pain in the affected area.10,54 Peri-articular erythema and warmth may also occur, sometimes accompanied by a low-grade fever.55 Spasticity may increase secondary to the NHO development.53 Reduction of hip joint movement and spasticity may lead to loss of an adequate sitting position,51 pressure sores, and related pain complaints52 and may also compromise transfers and activities of daily living. Although not commonly reported in SCI patients, ectopic ossification has been associated with compression of vascular structures and nearby peripheral nerves.56,57,58,59,60,61 The relatively low incidence of neuropathies found in SCI patients may be due to the sensory loss in (complete) SCI patients, so that they will not experience painful paraesthesias associated with nerve compression. Moreover, progression of NHO sufficient to compress nerves and nearby vascular structures and thereby causing clinical symptoms may result in ankylosis prior to possible neurovascular compression.

Pathophysiology

NHO originates in the connective tissues and may be contiguous with the skeleton, but does not involve the periosteum. When near a joint, it leaves the joint space and capsule preserved.34,35,36,48,62 Muscle fibres are not primarily involved in the process, but can be incorporated in or compressed by the fibrosing and calcifying soft tissues leading to local muscle necrosis.8,34,48

NHO begins as an area of oedematous, inflammatory reaction coinciding with an increased bloodflow in the affected soft-tissues.48 First, an exsudative cellular infiltration is seen, then fibroblastic proliferation occurs followed by osteoid formation, and subsequent deposition of bone matrix. Primitive osteoid is deposited as small masses within an area of fibroblastic mesenchymal reaction early within the first 2 weeks at first in the periphery. Osteoblasts produce tropocollagen, which polymerizes to form collagen, and secrete alkaline phosphatase (AP). The AP lyses pyrophosphate, a compound that prevents calcium deposition. Thus, by inactivating the pyrophosphate near the developing ectopic bone matrix, AP allows calcium to precipitate and the bone matrix to mineralize.63,64,65

The mineralization process of the soft tissues consists of an amorphous calcium phosphate phase, which is gradually replaced by enlarging hydroxyapatite crystals.48 The centripetal pattern of maturation that is seen in the following weeks is the basis of the zone phenomenon described by Ackerman.66 The zone phenomenon is characterized by a thin outer zone in the surrounding muscle that encloses a broader intermediate zone. In the intermediate zone, areas of immature bone are lined by osteoblasts, while in the outer margins of this zone mature bone forming a well-demarcated outer trabecular rim is already present. The intermediate zone surrounds the central zone that consists of an undifferentiated highly cellular proliferation of fibroblasts with haemorrhage and muscle necrosis. As the lesion matures, the peripheral rim of the intermediate zone becomes radiographically opaque due to progressive mineralization.67,68,69 The entire sequence of bone maturation is usually completed within 6–18 months.10,35,52,70 Mature NHO resembles normal bone, both histologically and radiologically.71,72 and consists of cancellous bone with Haversian canals, cortex, blood vessels and bone marrow, although with a minor amount of hematopoesis.34,36,39,48,73

According to Chalmers et al74 three conditions must be met for the formation of ectopic ossification: the presence of osteogenic percursor cells, an inducing agent, and a permissive environment. Although the precise causal mechanism for NHO is still unknown, humoral, neural and local factors probably all play a role in its pathophysiology (Figure 1). There is either a migration of distant mesenchymal cells to the area involved, with subsequent transformation of these cells into osteoblasts, or a transformation of the local mesenchymal cells directly into osteoblasts.52,63,75,76,77 Whether these cells migrate at random or in response to some chemotactic factor is still not known, but the importance of several factors has been suggested in the transformation of mesenchymal cells into osteoblasts.77,78

Figure 1
figure 1

Visual approach to the pathophysiology of NHO in SCI patients. OAC: oral anti-coagulant. DVT: deep venous thrombosis

Full size image

Humoral factors

Recent work with the serum of traumatic brain injured (TBI) and SCI patients seems to indicate that humoral mechanisms may be involved.79,80 In the study of Binder,79 the serum of a TBI patient increased the osteoblast growth factor activity in foetal rats. Kurer et al80 incubated sera from SCI patients with and without NHO 4–7 months post injury and sera from healthy control subjects with human osteoblasts in tissue culture. An increase in osteoblast stimulating factors was found only in the SCI groups and the activity was more pronounced in the group of SCI patients with NHO compared to patients without NHO. Renfree et al26 incubated sera from SCI and TBI patients throughout the first 12 weeks post-injury with osteoblasts from foetal rats. They observed a significant rise in serum mitogenic activity during the period after the injury in both patient groups. However, no significant differences were seen when patients who developed NHO were compared to other patients and healthy control subjects. Because their findings did not support the existence of a humoral factor that directly stimulates osteoblast proliferation within the first 12 weeks post injury, Renfree et al26 hypothesized that the rise in mitogenic activity seen in their patients may indirectly play a role in the bone inductive process.

Although humoral factors may play a role in the osteo-inductive process their origin and their biological characteristics have not actually been identified in SCI patients. Yet, in experimental studies concerning the in vitro induction of ectopic bone, an osteo-inductive protein released from demineralized bone tissue could be identified.81 This factor was called bone morphogenic protein (BMP).77,82,83,84 It is, therefore, possible that bone resorption and collagen degradation in acute SCI patients may well release (still unidentified) osteo-inductive factors.85

Neuro-immunological factors

The neural influence on NHO development cannot be disregarded taking into account its high incidence in neurologic disorders and the confinement of the NHO to the body regions with neurologic deficits. Dejerine1 was one of the first to suggest that damage to the intermedio-lateral sympathetic columns of the traumatized spinal cord might predispose to NHO through autonomic dysregulation. Secondary to an altered balance within the autonomic nervous system, a diversity of metabolic and vascular changes may occur.1,3,36,46,48,73,86 Indeed, several studies have shown that the initial stage of NHO is characterized by local microvascular alterations such as an increased vascularity, venous haemostasis and arterio-venous shunting in the involved tissues.36,48,87 These modifications in the blood perfusion and oxygen levels of the soft tissues might play a critical role in the formation of NHO,36,87,88 although it remains unclear whether these changes are secondary to or a causal factor in the NHO. Whether a disruption of the neuro-immunological pathways could result in an abnormal balance between osteoblast and osteoclasts and the subsequent induction of NHO is also still hypothetical.89 Yet the presence of interstitial oedema by itself, whether due to autonomic dysregulation,1 hypersensitivity, or hypoproteinemia6 may add to a permissive environment by facilitating pathological calcification of the osteoid.

Local factors

The local factors that may predispose to NHO are venous thrombosis or haemostasis, (local) infection, decubitus ulcers and (micro) trauma. These factors may lead to tissue damage and subsequent inflammatory reactions causing oedema and tissue hypoxia12 and may predispose to ectopic bone formation either by providing a permissive environment or by releasing humoral factors through the inflammatory process. In animal studies it has been suggested that eg prostaglandin E2 (PGE2) and interleukin-1 can induce a dose dependent increase in the subperiostal lamellar bone formation90 and that subcutaneous injections of PGE2 in growing rats can induce heterotopic bone formation.91,92 Measurement of 24-h PGE in SCI patients with NHO showed abnormal 24-h PGE2 urinary excretion as long as the bone-scan had not stabilized.27 However, the mechanism by which prostaglandins may induce ectopic bone formation in SCI patients remains unclear and there are well-known difficulties in the interpretation of PGE2 urinary excretion measurements in men presenting with a positive urinary sperm count or in patients with lower urinary tract infections. Both groups account for a substantial amount of the SCI population.

Risk factors

Several risk factors have been associated with NHO. In general, no association was found between NHO and race or gender,13,19 although in some studies a slightly higher incidence has been suggested for (young) male patients.10,13,14,17,24,47 A genetic predisposition for NHO based on human lymphocyte antigen (HLA)-typing has also been suggested. The human leukocyte antigen system (HLA) consists of a series of glycoprotein molecules that appear on the surface of virtually all nucleated cells. The major histocompatibility complex (MHC) is located on the short arm of the sixth chromosome and regulates a large and complex portion of the body's immune system and possibly other aspects of cell proliferation. The MHC contains a number of loci, more specifically the classic HLA-antigens (HLA-A, HLA-B, and HLA-C).93 Some authors have suggested an association between HLA B18 and HLA B27 on the one hand and NHO on the other hand,94,95,96 whereas others could not corroborate this association.93,97,98,99

The degree of completeness of SCI seems to be more important than the level of the lesion. Although Catz et al37 did not find a relationship between radiologically diagnosed NHO and the severity of the motor deficits, other authors have reported that complete transverse SCI is more commonly associated with NHO than incomplete SCI with relative risks (RR) reported between 2.0–4.2.1,13,14,19,21,22,24,88 Also, NHO is less frequently reported (<5%) in patients with lumbosacral or conus-cauda lesions, which regain or retain ambulation.88

Other clinical factors associated with NHO are the presence of pressure sores,6,21,22,24,30,47 urinary tract infections or renal stones,3,4,6,24,100 deep venous thrombosis (DVT),24 severe spasticity,22 and (micro) trauma. An area of soft tissue damage due to pressure ulceration with subsequent oedema may predispose to the development of ectopic ossification.101 On the other hand, pressure sores may occur secondary to NHO due to a decrease in eg hip range of motion that affects sitting position and alters pressure patterns. As a result, body weight is unequally distributed among the tubera and pressure ulceration may evolve, mostly contralateral to the NHO affected side.11

An infected urinary tract could serve as a source of antigenic material precipitating an immune response that triggers subsequent NHO.89 On the other hand, it may be that demineralization of bone accompanied by a loss of calcium and a subsequent loss of collagen as is seen in the acute phase of SCI102,103 is associated with an increased risk of developing both urinary tract stones, osteoporosis and NHO.3,4,17,100 As yet, the conclusion seems justified that the relatively high incidence of urinary tract infection and stone formation in SCI patients with NHO is not well understood from a pathophysiological perspective.21,22

Patients in the acute phase of traumatic SCI are known to be hypercoagulable and at risk of developing thromboembolic complications. In several studies an association was found between DVT and NHO, with RR reported between 1.8 and 2.0 for patients with DVT compared to patients without DVT. Colachis et al25 found a 5.3% co-incidence of DVT and NHO. Although the authors routinely screened all SCI patients on admission for DVT, they did not routinely screen for NHO at the same time and hence, there may have been an underestimation of the real incidence of NHO in this study. In all cases the DVT was diagnosed prior to the NHO and the NHO activity was present at the same side of the body. Although these findings suggest that DVT may be a possible risk factor for NHO, Perkash et al104 emphasized in their report of three chronic SCI patients with acute NHO the possibility of a secondary hypercoagulable state. They reported that the NHO activity correlated well with the increased coagulation parameters. As the NHO activity decreased, the coagulation parameters returned to near baseline values suggesting that NHO may alter blood coagulation and thereby predispose to DVT. Another possible mechanism by which NHO may predispose to DVT is a compression of vascular structures by local oedema as well as by the expanding ectopic bone mass.56,57,59,61,105,106 As a consequence, it can also be concluded that the precise pathophysiological relationship between DVT and NHO has yet to be established.

Controversy also exists regarding the possible association between NHO and spasticity. In some studies NHO is more commonly seen in SCI patients with spasticity and more extensively in those with severe spasticity (RR 0.17–2.0).21,22,30 Moreover, ectopic ossification seems to be rare in flaccid limbs or in limbs not affected by spasticity.17,21,52 Although such observations may suggest that spasticity predisposes to the development of ectopic ossification, the direction of this putative causality may easily be reversed such that a developing ectopic mass leads to an increase in spasticity.6,7,10,14

As for the role of (micro) trauma, mechanical stress to the musculotendinous apparatus may arise either from vigorous passive exercises9,17,46 or from loss of mobility and muscle imbalance causing peak pressure on soft tissue areas. Mechanical stress causes local micro-trauma that may induce ossification either indirectly through an inflammatory response or directly by releasing osteoblast-stimulating factors. Results from several case series revealed that the time-interval between the SCI and the beginning of passive movement exercises plays a vital role in the risk of developing NHO.6,9,51,107 In the study of Daud,107 clinically apparent NHO only occurred in SCI patients in whom the start of passive motion exercises was delayed until 7 or more days after the SCI. These results were supported by animal studies.108,109 Studies performed by Izumi110 and Michelson108,109 demonstrated that heterotopic bone could be produced in rabbits by forced passive movements in paralytic limbs that had been immobilized for a certain period. It is, therefore, tempting to speculate that (forced) passive movements following a period of immobilization may easily result in shear and tear of soft tissues leading to an increased risk of developing NHO.111,112

In conclusion, apart from the existence and completeness of SCI and the probable role of (micro) trauma, the reports of other possible risk factors and their presumed causal relationship with NHO are still inconclusive.

Diagnosis

NHO is primarily diagnosed based on clinical signs and likelihood. The early symptoms must be differentiated from arthritis,113 thrombophlebitis, DVT, cellulitis, soft-tissue haematoma, complex regional pain syndrome (CRPS),114 and soft tissue tumour.8,35,36,105,115,116,117,118 Elevated serum alkaline phosphatase (SAP) levels may be of value in differentiating early NHO from other inflammations, as SAP may be markedly elevated during active osteogenesis.16 Ultrasonography can be used for the early identification of clinically suspected NHO and for differentiating NHO from DVT, a developing pressure sore, infection, or tumor.60,65,69,112,119 The use of phlebography to differentiate between NHO and DVT can be misleading. A large mass of ectopic bone can distort and compress vascular structures, causing venographic findings that mimic venous thrombosis.60,61,106 In order to differentiate between NHO and arthritis due to rheumatic disease, synovial fluid analyses can be performed. In contrast with arthritis due to rheumatic disease, synovial fluid analysis in NHO reveals a low white blood cell count, high protein levels, low viscosity and no crystals.39,48,113,120

Laboratory examinations

It has been shown that routine blood chemistry, eg the determination of isolated urinary and serum calcium, is of little value in the diagnosis and monitoring of NHO.16,48,65,72,100,121,122,123 Also an elevated erythrocyte sedimentation rate may reflect the initial ‘inflammatory’ phase of NHO, but is very aspecific.

The SAP consists of a series of iso-enzymes which are sensitive, but non-specific indicators of heterotopic ossification. They are found in several human tissues including the skeleton, the liver, the intestinal mucosa, and the placenta.64 The various iso-enzymes can be separated by electrophoresis. Electrophoretical studies performed with the sera of NHO patients revealed that elevated SAP levels reflect the activity of the ossification process.16,116,124 When new bone is actively deposited, the SAP levels are elevated. As soon as the ossification has stopped, the enzyme levels return to normal.16,48,70,72,125 In general, SAP levels start to rise on average 7 weeks before the first clinical signs of NHO become apparent, exceed normal levels 1 week later and reach peak levels 3 weeks after the appearance of the first clinical signs.65,70 Thereafter, the SAP levels gradually decline to reach normal values at about 5 months post onset. However, extensive bone formation may lead to prolonged elevation of SAP levels,39,48 whereas a minor amount of bone formation may not lead to SAP elevation at all. Normalization of the SAP levels does not constitute adequate proof of stabilization of the osteogenic process45,48,49,52,62 nor does the height of the SAP levels accurately correlate with either the peak activity in the bone formation or the number or amount of NHO lesions.100 Therefore, SAP levels are generally considered of limited value for judging NHO maturity, as well as for detecting possible recurrence or reactivation. Nonetheless, SAP levels may play a role in diagnosing NHO, since the persistent elevation of both inorganic phosphate and SAP levels increases the likelihood of active heterotopic bone formation.122

The determination of urinary collagen metabolite excretion has also been advocated in SCI patients with NHO. Normally, the urinary excretion of collagen related glycosides is an indication of the tissue origin of the collagen being currently degraded. A degradation of bone collagen will produce a large increase in the excretion of gal-hydroxylysine and a smaller increase in the glu-gal hydroxylysine excretion. The reverse will be true in the case of skin collagen degradation as seen in decubitus.100 In SCI patients with NHO the gal-hydroxylysine excretion in urine was found to be higher than in SCI patients with NHO, and did not return to control values until the bone turnover had stabilized.

Also the urinary concentration of hydroxyproline, which is another collagen metabolite, was found to parallel the level of SAP in SCI patients with NHO.72,48 The peak in the hydroxyproline excretion is related to a specific polypeptide fraction that reflects the collagen degradation.48,122 Yet, the urinary hydroxyproline levels appear to be elevated in nearly all tetra en paraplegic SCI patients, which makes it difficult to differentiate between the general influence of SCI and the specific influence of bone formation in patients with early NHO development. Moreover, although urine hydroxyproline levels are associated with SAP levels, they are unreliable parameters in determining NHO maturation.48

Longitudinal studies are necessary to determine the precise temporal relationship between an increase in the characteristic urinary glycoside excretion and the appearance of clinical signs of bone formation. Until now, measurements of collagen metabolites derived from bone, connective tissue, or muscles have been abandoned, because of their unspecificity and methodological difficulty.36

Radiological examination

In the early stages of NHO, the bone formation mainly consists of osteoid that shows a high uptake of osteotropic radionuclides through which it is readily detectable by three-phase 99m Technetium bone scanning.15,16,28,39,70,126,127,128,129,130,131,132 The first phase of the three-phase bone scan is the period immediately after the intravenous injection of the radionuclides and detects areas of increased blood blow, which is an early indicator of the inflammation process (the dynamic blood flow phase). The second phase (the static blood pool phase), identifies areas of increased blood pooling several minutes after the injection. The third phase (the static bone phase) determines the degree of osseous uptake of the labelled radionuclides several hours after the injection. The blood flow and blood pool phases of the bone scan are able to detect NHO as early as 2.5 weeks after SCI and are followed by a positive static bone phase 1–4 weeks later.28,129 Bone scans return to normal as the NHO matures usually within 6–18 months after its first clinical signs.70 Compared to plain radiography, bone scanning is a more sensitive diagnostic test for early NHO, but radiography is more specific. A disadvantage of the three-phase bone scan is its low specificity leading to potential difficulty in discriminating NHO from other inflammatory, traumatic, or degenerative processes of the skeleton, eg fracture, bone tumor, metastasis, or osteomyelitis which all show increased osteoblastic activity and thereby increased uptake of osteotropic radionuclides.126,127,132,133,134 Moreover, it requires expensive specialist equipment and the use of ionising radiation. Yet bone scanning may reach a fair specificity once clinical signs and likelihood have given rise to reasonable suspicion of NHO.28 In monitoring NHO maturation, the three-phase bone scan seems the most sensitive tool and currently the ‘golden standard’.49,126,129,132 Nonetheless, the criteria for maturity rating on the bone scans remains a topic of discussion. Whether a qualitative assessment eg a continuous decrease in the uptake or a steady state in the radionuclide accumulation, or a quantitative assessment such as specific uptake ratios should be considered as the best maturity index remains unclear.

The sonographic diagnosis of NHO depends on the age of the lesion, the rate of bone formation and the degree of mineralization.133 The first finding on ultra-sound (US) examination of a focal, elongated hypoechoic mass is unspecific and is also seen in muscle tears, abscess, and soft tissue tumor. The muscle fibres and soft tissues surrounding the lesion may appear compressed, but the interface is smooth and there is no evidence of invasion. Thereafter, the development of a centripetal pattern of maturation can be seen. Initially, the intermediate zone contains foci of echodense islands and will not be uniformly reflective. Later, the foci of echodense islands rapidly become confluent68,69,135 and the complete zone phenomena can be seen sonographically. As the NHO matures, the peripheral rim of the intermediate zone becomes more reflective on US due to increased mineralization67,68,69 and 4–6 weeks after the first clinical signs the NHO behaves like corticated bone. At that time, the US beam is totally reflected and the NHO can also be seen radiographically.133,69 Mechanical forces can modify the appearance of the zone phenomena. Therefore, the sensitivity and specificity of US for diagnosing NHO strongly depend on the experience of the radiologist. Serial sonography allows differentiation from muscle tears, soft tissue haematoma, abscess, thrombosis and soft tissue tumor.67,68,119,135,136 Sonography has the advantage of the possibility of bedside application, is relatively cheap, and requires no radiation.

The earliest radiographic sign of NHO is an increased density of the peri-articular soft tissues due to oedema.39,47,48 Gradually, the image shows flocculent densities as calcium precipitates. The maturation process proceeds rapidly with increased delineation of the soft tissue mass and the formation of bony cortex and trabeculations. The skeleton underlying the NHO may show variable degrees of demineralization, but the underlying bone configuration and the joint surfaces remain unaffected.18 NHO will become evident on plain radiography upon accumulation of minerals about 2–6 weeks after the three-phase bone scan first becomes positive.70,129,137 In several studies radiographic signs of NHO were found on average 1–10 weeks after the first clinical signs.19,36,47 Plain radiography may be positive even in asymptomatic patients as early as 4.5 weeks after the SCI. For monitoring NHO activity, plain radiography may be valuable as long as the X-rays show modification from one examination to the next.48 However, with a large block of ectopic bone, interpretation of the X-rays is difficult because immature elements may be obscured by mature bone.47 Therefore, radiography is generally considered of little value for judging NHO maturity,47 as well as for detecting possible recurrence or reactivation.49

CT scanning allows a better visualization of the ectopic bone in relation to the soft tissues.118,138 The CT scan may aid in the development of a surgical plan,139 especially three-dimensional tomography, eg in order to avoid areas of immature bone.140 MRI scanning is the best technique to define the extent of the soft tissue oedema, but has no role in the early diagnosis of NHO. The appearance of soft tissue swelling is aspecific and the lack of signal from calcification prevents differentiation from other inflammatory processes.118

The vascular changes seen immediately after SCI and more frequently in the case of concomitant NHO can be visualized by angiography. They consist of opening of the arteriovenous shunts, early venous return, and locally increased vascularity, which can easily be determined by angiography.48,141 However, the persistence or absence of venous return is not a valid indication of the degree of maturity. As ectopic bone matures, the pathological angiographic findings regress, in particular the vasodilatation and blood pooling without necessarily returning to normal.48,141 Complications of angiography are haematoma and subsequent scar tissue formation. As yet, angiography has no role in diagnosing or monitoring NHO.

Treatment

Disodium etidronate (EHDP)

Once NHO has been diagnosed, patients can be treated with sodium etidronate, which is a disphosphonate. Diphosphonates are structural analogues of the naturally occurring inorganic pyrophosphates, which play an important role in the calcium-phosphate metabolism. Unlike the natural pyrophosphates, diphosphonates are almost completely stable to biochemical degradation. EHDP has been most widely used in patients with ectopic bone formation. In vitro EHDP strongly binds with hydroxyapatite, the main inorganic constituent of bone, thereby blocking the transformation of amorphous calcium phosphate into hydroxyapatite crystals, without inhibiting the formation of bone matrix.142,143,144,145 In animal studies and experimental models of ectopic ossification, it has also been demonstrated that EHDP can reduce the number of bone forming cells and alter their cell morphology.146,147 EDHP might also have an anti-inflammatory effect probably by affecting the production of interleukin-1.148

In SCI patients EDHP has mainly been used to block ectopic bone formation in the early phase after the clinical diagnosis of NHO. Stover62 treated patients on average 56 days post-SCI with EHDP 20 mg/kg/day for the first 2 weeks, followed by 10 mg/kg/day for the next 10 weeks. At the start of the EHDP treatment radiographic evidence of NHO was found in 25% of the placebo-group and in 22% of the EHDP group. After the treatment 41% of the placebo group and 30% of the EHDP group showed radiographic evidence of NHO. At its best EHDP may have slowed down the mineralization process. The authors suggested that for an optimal effect EHDP should be started at an earlier phase of NHO before significant amounts of ectopic bone have been formed. Finerman et al29 treated patients with clinical signs but without radiographic evidence of NHO according to Stover's62 treatment protocol on average 112 days post-SCI. At the end of the treatment period of 12 weeks, 6% of the EHDP group demonstrated NHO on X-rays, of which only 2% was clinically significant. In the placebo group, 27% demonstrated radiographic evidence of NHO, of which 13% was clinically significant. However, after cessation of treatment, the incidence of NHO increased in the EHDP group, although the lesions were less extensive than in the placebo group. Since the majority of NHO is radiographically diagnosed within the first 6 months after SCI,52,149,70 Garland et al20 stated that EHDP should be administered for at least a 6-month period. Although EHDP seems to block the mineralization process to a certain extent, it does not influence the formation of the osteoid matrix. After cessation of EHDP administration, the bone matrix already formed may undergo an uninhibited mineralization, the so-called ‘rebound ossification’.52,65

Some of the differences seen in therapeutic response to EHDP between patients may be due to variations in the bioavailability of EHDP after absorption by the gut. Because of the low solubility of EHDP in water, absorption in the intestinal tract can fluctuate from 1–10%. To avoid such variations and to improve the bioavailability of EHDP, Banovac et al137 administered EHDP intravenously in SCI patients with early clinical signs of NHO and a positive I and II phase on the bone scan on average 28 days post-SCI using high dosages over a prolonged time period: 300 mg for 3 days, followed by 20 mg/kg/day orally for 6 months. A prompt reduction in the peri-articular swelling was seen during the first 48 h. However, after the treatment period, no radiographic differences with regard to extent of NHO were found between the EHDP patients and a control group treated with 20 mg/kg/day orally for 2 weeks followed by 10 mg/kg/day for 6 months. In the subgroup without radiographic evidence of NHO (N=12), only two patients (15%) developed minimal radiographic evidence of NHO at the end of the follow-up period on average 11 months post onset of therapy. Banovac28,150 repeated his study in asymptomatic patients with a positive third phase on the three-phase bone scan. Only one patient developed NHO to a minimal extent.

Side effects of EHDP include hyperphosphatemia. In animal studies EHDP induced osteomalacia and spontaneous fractures have been reported. However, these complications have not been found in patients after total hip arthroplasty (THA) nor in SCI patients. Moreover, animal studies suggest that three to five times higher dosages are needed to induce such complications compared to the dosages currently used in SCI patients.151 The main reasons for patients to discontinue EHDP treatment are gastro-intestinal symptoms such as nausea, diarrhoea, and abdominal distress.

Non-steroidal anti-inflammatory drugs (NSAID)

The basic knowledge about the prophylactic treatment of ectopic bone formation by NSAID comes from animal studies and studies in patients undergoing THA.152,153,154,155,156,157 The effect of NSAID has been attributed to the inhibition of the inflammation process and the suppression of mesenchymal cell proliferation.52 Anti-inflammatory drugs inhibit the release of prostaglandins and related substances, thereby reducing the stimulation of the formative and resorptive phases of the bone remodelling.152,153,154,155,156,157 Moreover, NSAID are known to inhibit the differentiation of the mesenchymal cells into osteogenic cells and to reduce ectopic ossification in soft tissues in experimental conditions.

In the only randomised clinical trial (RCT) performed in SCI patients, Banovac et al158 allocated patients to either indomethacin 75 mg/day or placebo for 3 weeks from admission to a rehabilitation centre on average 20 days after their injury. There was a significantly higher incidence of NHO diagnosed by bone scan in the placebo group (65%) compared to the indomethacin treated group (25%). In the indomethacin treated patients with NHO the onset of symptoms was delayed on average 31.7 days compared to the placebo treated patients with NHO on average 19.2 days. Also, the inflammatory symptoms, ie swelling, erythema, and fever were less pronounced.

No side effects were reported, although gastro-intestinal complications eg dyspepsia and gastric ulceration or bleeding, and delayed fracture healing have been reported in other studies.152,153 Moreover indomethacin is also known to have an anti-thrombotic effect and may alter blood coagulation especially with the concomitant use of other anticoagulants to prevent deep vein thrombosis (DVT).

Other types of medication

Two other types of medication have been used in the treatment of NHO after SCI. Calcitonin has been used in an uncontrolled study with inconclusive results.159

Warfarin, which is an anti-coagulant that inhibits vitamin-K-dependent synthesis of calcium binding proteins such as osteocalcin has been used in a historic cohort of 227 SCI patients, of whom 33 had been treated with a low dose (prothrombin time 1.5–2) for the prevention of DVT.23 The authors did not observe clinically significant NHO as long as 10 years after SCI in the Warfarin treated group, whereas in the other patients clinically significant NHO developed in 15%. Although methodologically non-optimal, this study suggests a possible inhibiting effect of Warfarin on NHO, yet the nature of this effect remains to be elucidated.

Glucocorticoids, such as prednisolone can decrease the formation of bone and scar tissue in vitro. Animal studies showed that prednisolone decreased the inflammatory changes and fibrosis in rabbits caused by forceful manipulation after prolonged immobilization and therefore, also decreased ectopic bone formation.154 However, until now, glucocorticoids have not been used in the treatment of NHO in SCI patients.

Low-field irradiation

The exact mechanism by which irradiation affects NHO formation remains unknown, although it is suggested that irradiation may disrupt the differentiation process of pluripotent mesenchymal cells into osteoblasts.160,161,162,163,164 Irradiation may also decrease pain perception associated with the tissue inflammation surrounding the NHO, or it may induce the ablation of pain receptors.164 Sautter-Bihl et al165 treated 20 SCI patients with low field irradiation with 10 Gy in single fractions of 2 to 2.5 Gy. In 15 patients irradiation was given as a primary treatment to prevent NHO, whereas in seven patients it was used after operative NHO resection to prevent NHO recurrence. Four of the seven operated SCI patients, received a THA. During the follow-up period of 12 weeks, none of the 22 patients showed progression of their NHO. In the five patients followed for 44 months, the hip joint range of motion and sitting position did not deteriorate. In a second study, the same author166 treated 49 SCI (70 joints) with low field irradiation with 8 to 10 Gy in 2–25 single dose fractions. In 58 joints, irradiation was the primary treatment after the first clinical signs of NHO, whereas in 12 joints irradiation was performed after surgical resection of the NHO. In 71% no progression of the incipient ossification or post-operative recurrence of NHO was seen during a follow-up period of on average 11 months.

Complications of radiation therapy include delayed wound and bone healing, osteonecrosis,167 as well as the risk of developing radiation-induced sarcoma. Although these complications are not mentioned in relation to dosages less than 30 Gy delivered within a 3-week period,168,169 they certainly limit the use of radiation therapy as the primary prevention of NHO, when a relatively long life expectancy and the frequent involvement of multiple joints in SCI patients are taken into account.170

Surgical resection

The indications for surgical resection of NHO are improvement of joint motion, eg to achieve an adequate sitting or standing position,49,116,171 reduction of spasms, and prevention of pressure sores.172,173,174,175 However, operative resection of NHO has been associated with severe complications and poor outcome.171 These complications include deep (2–5%)174,175 or superficial infection (7–38%),45,49,174,175 post-operative haemorrhage (5–38%),45,173 and excessive intra-operative bleeding requiring (multiple) blood transfusions (9–83%).49,174,175,176 Moreover, intra- and post-operative fracturation of bone may occur in the case of severe osteoporosis. Stover174 has reported intra-operative fracturation of the femoral head in 5% and post-operative (spontaneous) fracturation in 3–16% of the SCI patients operated for NHO around the hip joint. Since other authors mention a higher percentage (29%)175 of SCI patients with femoral heads too brittle to preserve during operation, the real incidence of this problem may be even higher. NHO resection is also associated with a poor outcome due to a high recurrence rate. The post-operative recurrence of NHO depends on the definition used and the follow-up period. Radiographic signs of NHO recurrence have been mentioned in 82–100% of the operated SCI patients,45,49,173,174 although clinically significant NHO would develop in only 17–58% of the cases.10,45,49,173,174,176 It has been suggested that the recurrence after resection can be reduced if NHO is removed only once the first and second phase of the bone scan have normalized.10,52,115,129,132 However, such a maturity rating does not guarantee that NHO recurrence is unlikely. Garland et al49 reported that still 36% recurrence of clinically significant NHO was found in SCI patients with mature NHO on the bone scan. It was not so much the maturity that was associated with the risk of NHO recurrence, but rather the extent of the NHO preoperatively. There was also no correlation between NHO recurrence and the time post SCI or NHO onset.49,174,177 Recently McAuliffe et al170 and Freebourn177 reported three patients that had successfully undergone early resection of NHO, that is within 7–11 months post SCI. Earlier surgery may be warranted to prevent fibrous ankylosis, muscle contractures, and severe disuse osteoporosis to decrease the intra-operative fracturation rate. Hence, although many experts advise a time period of 12–18 months between NHO diagnosis and resection, no adequate studies are available to corroborate this advice or to give an indication of the most effective timing for the surgery.

Different prophylactic measures have been advocated to reduced the NHO recurrence post surgery. As for THA, both irradiation and indomethacin seem to be effective in reducing the occurrence of heterotropic ossification after THA,152,153,160,161,162,163 whereas EHDP has been abandoned due to its ineffectiveness on bone matrix formation and the high risk of ‘rebound’ ossification.142,146 In SCI patients EHDP,174 NSAID, irradiation,163,165,166,170,175 or a combination of treatments177,178,179 have all been recommended to reduce the NHO recurrence after surgery, however, no controlled studies are available.163,165,166,170,174,175,176,177,178,179

Primary prevention

Since the pathophysiology of NHO is poorly understood, the only preventive treatment possible is the early identification and adequate treatment of the putative risk factors. Through adequate nursing management the incidence of urinary tract infections, decubitus ulcers, and DVT may be reduced, and thus, the risk of developing NHO. Although in the early literature aggressive passive physiotherapy has been recommended to improve joint mobility and to counteract ankylosis in the case of contractures, it is now generally accepted that SCI patients profit from early, regular, and cautious joint mobilization to prevent more rigorous exercises with the risk of (micro) trauma to the peri-articular tissues becoming necessary. When gentle passive movements of the large peripheral joints are started and maintained from the day of the injury, the joint capsules are kept as supple as possible, muscles will not easily shorten and contractures will not readily develop, so that NHO may be prevented. This insight takes into account the highly ‘vascular’ state of the paralysed area and the concomitant use of anticoagulants (to prevent DVT), that may predispose to haematoma and secondary NHO, particularly during the rehabilitation phase after SCI.19

Various, more specific prophylactic measures have been proposed to prevent NHO, including the use of diphosphonates, NSAID, and more recently irradiation. The latter two methods have been successful to some degree in preventing heterotopic ossification eg after THA,142,152,153,160,161,162,163,180 but no controlled studies are available in SCI patients.

Conclusion

NHO is a frequent complication in SCI. Although the precise pathophysiology of NHO is still unknown, humoral,26,79,80,81,82,83,84,85 neural,1,3,6,36,46,48,73,86,87,88,89 and local factors12,27,90,91,92 probably all play a role in the formation of ectopic bone. Apart from the completeness of SCI1,13,14,19,21,22,24,88 and the possible role of (micro) trauma,6,9,17,46,51,107,108,109,110,111,112 reports of possible risk factors like DVT, pressure ulceration, infection, and spasticity6,7,10,14,21,22,24,30,47,89 and their presumed causal relationship with NHO are inconclusive. Studies on experimental bone induction have emphasized the role of neuro-immunological and humoral factors. The role of these factors is currently also studied in the pathophysiology of CRPS and shoulder–hand syndrome and seems to be a promising area of further research in the pathophysiology of NHO. Until now, most studies on experimental bone induction have been performed in animal models without SCI. To enhance our knowledge of the pathophysiology of NHO in SCI, research needs to be done in SCI patients as well as in animal models with SCI.

The diagnosis of NHO is primarily based on clinical signs and likelihood. It is the concurrence of persistent elevated SAP levels with characteristic clinical signs and, ultimately, positive radiographic findings that is essential for definitively diagnosing NHO.122 Sonography may also play a role in the early diagnosis of NHO, however its sensitivity and specificity greatly depend on the experience of the radiologist.67,68,69,133 The three-phase bone scanning is the most sensitive technique for diagnosing NHO and, although generally aspecific, may reach a fair degree of specificity once clinical signs and likelihood have given rise to reasonable clinical suspicion.28,49,70,126,129,132 For monitoring NHO activity, plain radiographs and elevated SAP levels may be valuable as long as they show modification from one examination to the next, but currently the three-phase bone scan seems to be the ‘golden standard’. CT and MRI scanning are mainly used to aid in the development of a surgical plan.118,138,140

Several studies have documented that early diagnosis is of crucial importance in the treatment of NHO.29,65 Waiting for late clinical or radiographic symptoms of NHO may allow a significant amount of ectopic bone to be deposited before treatment is initiated. Although EHDP inhibits the mineralization of bone, it does not influence the formation of osteoid.142,143,144,145 Hence, after cessation of EHDP, the bone matrix already formed may still undergo mineralization as soon as the drug has been cleared from the tissues, especially in the first 6 months after SCI, the so-called ‘rebound’ ossification.52,65 Therefore, in the treatment with diphosphonates prolonged administration up to 6 months seems to be of crucial importance.20,28,137,150 Yet, there is also some evidence that with the late application of EHDP, the severity of the lesions remains less compared to placebo treatment.29,62 More recently, NSAID and irradiation have been used in SCI to prevent NHO occurrence. The effects of both NSAID and irradiation have been attributed to the suppression of the differentiation process of the mesenchymal cells into osteoblasts.52,152,153,154,155,156,157,160,161,162,163,164 Irradiation may also decrease pain perception association with the soft tissue inflammatory reaction surrounding the NHO, or it may result in the ablation of pain receptors.164 NSAID may also exert an effect through the inhibition of the inflammatory response.152,153,154,155,156,157 However, the exact mechanisms by which NSAID or irradiation prevent NHO development remain unknown. Both treatments, however, do not influence the ectopic bone already formed, and are, therefore, effective only in the early stages of NHO.158,165,166 More controlled studies are needed to determine the effectiveness and complication rates of different (prophylactic) treatments for NHO in SCI patients.

The indication for surgery is to improve joint range of motion to achieve adequate posture, reduction in spasms, and prevention of pressure sores.49,116,171,175 Operative resection of NHO is generally associated with severe complications and poor outcome.45,49,71,173,174,175,176 It has been suggested that the recurrence of NHO after resection can be reduced, if the NHO is removed during a radionuclide steady stage, however, this policy is not supported by unequivocal clinical evidence.10,52,115,129,132 Although most authors advise a delay of 12 to 18 months between NHO diagnosis and surgical resection, until now no adequate controlled studies are available to corroborate such advice. Hence, there is no consensus on the most effective timing for surgery.

To date there is no satisfactory prevention of NHO and its prophylaxis is mainly based on the early identification and adequate treatment of putative risk factors, such as DVT, pressure ulcers, urinary tract infection and spasticity. Yet, there is ample empirical evidence that regular and cautious mobilization of the large peripheral joints should be recommended, from the day of the injury, to keep the joint capsules as supple as possible and to maintain adequate muscle length. With such an approach contractures will not readily develop and NHO related to traumatizing mobilization might be prevented.