Introduction

Heterotopic ossification (HO) or ectopic ossification is associated with several medical conditions, including spinal cord injuries (SCIs), traumatic brain injuries and hip arthroplasties. It is commonly reported after SCI in which it involves the formation of mature lamellar bone in soft tissues, usually para-articularly.1 Reported incidence varies greatly in the SCI population, ranging from 10–53%.2, 3 HO begins to develop most frequently within the first 2–3 weeks after SCI and occurs below the level of paralysis, usually at the hip (70–97%) and followed by the knee.2, 3 In patients with SCI with clinically significant HO, 20–30% present with a reduction in joint range of motion, whereas only 3–8% develop ankylosis.2

Initial clinical features of HO may include joint and muscle pain, decreased range of motion, tissue swelling, redness and heat in the involved region, as well as a low-grade fever.2 Diagnostic testing can be helpful; in the early phase of HO, three-phase nuclear bone scanning is positive with increased uptake of osteotropic radionucleotides. Nuclear bone scans have proven to be more sensitive than plain radiography in detecting early HO with neurogenic HO becoming evident in the first and second phase of the three-phase bone scan 2–6 weeks before diagnosis on plain radiographs.4, 5 Computed tomography scanning may be a useful tool in the later stages when considering surgery as it allows better visualization of the heterotopic bone.6 The use of sonography has also been assessed to diagnose HO before radiographic evidence.7 Some studies have looked into diagnosing HO through elevations of biochemical markers such as alkaline phosphatase8, 9 and creatine phosphokinase.8, 10, 11 The predictive value of alkaline phosphatase has not been validated,8, 10, 11 although there is conflicting evidence of an association of HO with increased serum creatine phosphokinase levels.8, 10 Schurch et al.12 studied individuals with acute SCI and found increases in the 24-h prostaglandin E2 urinary excretion a valid indicator of early HO formation.

Various clinical factors have been associated with neurogenic HO, including degree of completeness of the SCI and the presence of pressure sores, urinary tract infections, deep venous thrombosis, severe spasticity, and trauma.13, 14, 15, 16 A causal relationship between these factors and the development of HO after SCI has yet to be determined.

HO may be a significant complication for individuals who have suffered an SCI; despite this, no critical review of HO interventions in SCI has been conducted. The purpose of this systematic review was to examine the effectiveness of treatments that used to prevent and treat HO in the SCI population. This review was conducted as part of the Spinal Cord Injury Rehabilitation Evidence project (http://www.icord.org/scire),17 an evidence-based review of the literature assessing rehabilitation interventions in patients with SCI.

Methods

Literature search strategy

A systematic review of all relevant literature, published from 1980 to April 2009, was conducted using multiple databases (MEDLINE, CINAHL, EMBASE and PsycINFO). Keywords included heterotopic ossification, HO, ectopic ossification; excision surgery, etidronate, pharmacological, non-pharmacological, medication, radiation, treatment and intervention. All references were scanned for relevant citations and assessed against inclusion criteria. References of selected articles were reviewed for pertinent articles.

Study selection

The literature search and secondary hand search resulted in a total of 194 studies. Studies were selected on the basis of previously established Spinal Cord Injury Rehabilitation Evidence methodology.18 Studies were only included for analysis if at least 50% of subjects had an SCI, there were at least three SCI subjects and if there was a definable intervention being studied. Only studies published in English language were included. A total of 13 articles were identified as research interventions for HO after SCI. Excluded articles included: (1) review articles, (2) noninterventional studies, (3) pediatric studies, (4) non-SCI populations (hip arthoplasty, brain injury, multiple sclerosis and Guillain–Barre syndrome).

This review included randomized controlled trials (RCTs) and nonrandomized controlled trials (nonRCTs), including prospective controlled trial, cohort, case–control, pre–post, post-test and case series.

Study appraisal

A quality assessment was conducted for each RCT study by two reviewers, using the Physiotherapy Evidence Database (PEDro) scoring19 system. Discrepancy in scoring was resolved by a third-blind reviewer. The PEDro tool19 consists of 11 questions with a maximum score of 10; higher score indicating a better methodological study with scores from 10 to 6 rating excellent to good and <5 fair to poor. Downs and Black (D&B) tool20 was used in the assessment of nonRCTs. The D&B tool20 contains 27 items with a maximum score of 28; again higher scores reflect a higher methodological quality of the rated study.

Data synthesis

Data from each of the studies were extracted and put into a table; investigations involving similar interventions were grouped. Data extracted included the type of study, a brief summary of intervention outcomes, study results and methodological quality (PEDro19 or Downs and Black score20). A modified Sackett scale rated the strength of evidence for each category of intervention21 (see Table 1).

Table 1 Levels of evidence modified from Sackett scale
Full size table

Results

There were 13 studies that met the inclusion criteria; this included four prophylaxis1, 3, 22, 23 and nine treatment24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35 studies of HO. Three of these intervention studies had Level 1,3, 22, 23 two Level 2,24, 25 seven Level 426, 27, 28, 29, 32, 33, 34, 35 and one Level 51 evidence.

Prophylaxis of HO

Interventions studied included primary prevention of HO shortly after SCI onset and later secondary prevention after surgical removal of the heterotopic bone. These include pharmacological and electromagnetic field treatment (Table 2).

Table 2 Prevention of heterotopic ossification after SCI
Full size table

NSAIDs drugs as a prophylaxis

There were two studies that investigated nonsteroidal anti-inflammatory drugs (NSAIDs) in the prevention of HO early after SCI. Banovac et al.22 randomized 76 patients early after SCI into two groups; a treatment group, in which subjects received 25 mg rofecoxib daily for 4 weeks, and a nontreatment control group. After 1 month, there was a significantly lower likelihood of developing clinical and radiographic evidence of HO in the rofecoxib group (13.4%) when compared with the placebo group (33.3%) (P<0.05). The most common adverse effect of rofecoxib was upper gastrointestinal symptoms.22 Owing to a reported increased risk of cardiovascular side effects of rofecoxib (Vioxx) in other populations, it is no longer available as a treatment.

Banovac et al.3 randomized 33 patients with an SCI ∼3 weeks after injury into prophylactic treatment with either slow-release indomethacin (75 mg daily) or a placebo control for a total of 3 weeks. Patients were followed up carefully with regular clinical follow-up and nuclear bone scans. After 3 weeks, there was a significantly higher incidence of HO (diagnosed through clinical signs and nuclear bone scan and/or radiographs) in the placebo group (64.7%) when compared with the group taking indomethacin (25.0%) (P<0.001). Patients receiving the indomethacin also experienced HO symptoms later than those in the placebo group (31.7 vs 19.2 days (P<0.048)). Upper abdominal discomfort was the most common symptom in both groups; however, this did not result in any patient discontinuing the medication.

In conclusion, there was Level 1 evidence (on the basis of two positive RCTs)3, 22 that NSAIDs, rofexocib and indomethacin, reduced the incidence of HO when administered early (3 weeks) after SCI.

Warfarin as a prophylaxis

Warfarin is a well-known anticoagulant that has been observed to reduce the development of HO after SCI. A single observational retrospective study examined the association between warfarin use and HO after SCI.1 Buschbacher et al.1 reviewed 227 patients with SCI. Diagnosis of HO was made through bone scans, x-rays and clinical signs. None of the 33 patients treated with warfarin (for an average of 5.4 weeks after SCI) were diagnosed with HO, whereas 34 of the remaining 194 patients (17.5%) not treated with warfarin were diagnosed with HO on an average of 12.5 weeks after injury.

In conclusion, there was only Level 5 evidence that warfarin is associated with a reduction in the incidence of HO after SCI.1

PLIMF

Pulse low-intensity electromagnetic field (PLIMF) therapy uses magnetic fields to increase oxygen levels and decrease toxic by-products of inflammation by increasing local blood flow into the area.23 Durovic et al.23 conducted an RCT examining the effect of PLIMF therapy as a prophylaxis for HO in individuals with SCI. Diagnosis of HO was through plain radiography and the Brooker's grading system. Both the control and treatment groups received a range of motion and exercise therapy. Individuals in the treatment group (n=14) also received 4 weeks of PLIMF therapy at a mean of 7 weeks after injury. The study showed significant differences in the incidence of HO between the two groups (P=0.04). None of the individuals in the treatment group developed HO, although HO progressed in 33% of individuals in the control group as measured by Brooker grades and radiographs.

On the basis of a single RCT, there was Level 1 evidence, supporting the efficacy of PLIMF in prophylaxis of HO after SCI.23

Treatment of HO

Bisphosphates

Six studies reported on the efficacy of bisphosphonates in the treatment of HO, once diagnosed (Table 3). Banovac et al.24 studied 46 patients with SCI with HO treated for 3 days with intravenous disodium etidronate, followed by oral etidronate for 6 months. Patients were subsequently categorized as being in one of the two groups. Group 1 (n=33) had positive bone scintigraphy but negative radiographic findings for HO; of these, five discontinued treatment and experienced a gradual progression of their HO. Of the remaining 28 patients, 22 continued to have no radiographic evidence of HO. Group 2 (n=13) tested positive on both bone scintigraphy and radiographs. Within this group, there was no further progression of soft-tissue ossification with etidronate in six patients, whereas the remaining seven exhibited progression of HO despite treatment.24 The authors concluded that once the bone scan was positive, discontinuation of treatment, regardless of radiographic findings, resulted in progression of HO; furthermore, once radiographic changes were present, there was a high probability that HO would progress regardless of whether treatment was continued or not.

Table 3 Treatment of heterotopic ossification after SCI
Full size table

Banovac et al.25 treated 27 patients diagnosed with HO based on radiographs and the three-phase nuclear bone scan after SCI, with intravenous etidronate for 3–5 days, followed by oral etidronate for 6 months. These 27 subjects were compared with 11 patients treated with oral etidronate alone for a period of 6 months. After the initial intravenous etidronate therapy, 20 of the 27 patients (74.1%) showed a reduction in swelling over the first 48 h, whereas 7 patients (25.9%) had either no change or increased swelling. Overall, it was determined that treatment reduced swelling from baseline (P<0.01). No significant difference was noted in the development of HO between those who had received etidronate both intravenously and orally and those treated with the oral drug alone.25

Banovac26 again studied patients with SCI with HO (n=40), diagnosed early (3–6 weeks after SCI) with positive bone scans and negative radiographs, who were treated with etidronate (intravenous for 3 days and then orally for 6 months). These patients were then followed over an extended period of time (6 years). In all, 11 of the 40 patients (27.5%) developed radiographic evidence of HO between 1.5 and 6 years after the initiation of therapy.

Garland et al.27 assessed 14 patients with SCI with clinical signs of HO and found no evidence of improvement in patients who were administered etidronate. In one patient, ossification appeared to plateau at the end of the treatment, whereas ossification continued to increase in the remaining patients to varying degrees.

In two small case series, each involving five individuals with an SCI, no recurrence of HO was found following surgical removal and subsequent treatment. Subbarao et al.28 used disodium edridonate before and after surgical hip wedge resection, whereas Schuetz et al.29 used pamidronate before and after HO surgical removal.

In conclusion, there was Level 2 evidence that etidronate can halt the progression of HO once the diagnosis is made, if initiated earlier (3–6 weeks).25 It seems to be most effective if the treatment is initiated earlier when the nuclear bone scan is positive and radiographs are normal.26 Once radiographs were positive, there was Level 4 evidence that bisphosphonates are not as effective. There was Level 4 (at a case series level) evidence that bisphosphonates effectively halt secondary HO progression after surgical resection of HO.28, 29

Radiotherapy

Radiotherapy involves the irradiation of pluripotential mesencymal cells, which may be responsible for the formation of heterotopic bone.31 Sautter-Bihl et al.32 reported on 52 patients with SCI, some of whom already had HO, who were treated with radiotherapy with single doses between 2 to 10 Gy. The study found neither a progression nor recurrence of the resected bone in 71% of joints. Regression in the Brooker grading score was not seen in any of the patients. Two joints increased in Brooker grades; one patient's hip went from a Brooker grade I to III and another patient's knee went from grade 0 to II. However, neither of them developed any associated functional impairments.

Sautter-Bihl et al.33 studied 36 patients with SCI diagnosed with HO, 3 of whom were subsequently lost to follow-up. In 11 of 33 patients (13 joints), radiotherapy treatment was used 24–36 h after surgical resection of heterotopic bone. Two of the 33 patients received radiotherapy both before and after surgery. The mean duration of follow-up was 23.6 months. In all, 30 of the 33 irradiated patients (90.9%) experienced no progression of HO and had normal mobilization and rehabilitation. In three patients, ossification after radiotherapy resulted in a moderate decrease in joint mobility with progression in the Brooker grade. No relevant adverse effects were seen due to irradiation.

In conclusion, there was Level 4 evidence that radiotherapy stops primary and secondary progression of HO.32, 33

Excision

Meiners et al.34 reported on a case series of 10 quadriplegics and 19 paraplegics who underwent HO resection at the hip followed by irradiation and passive range of motion exercises. Mean hip range of motion increased from 21.95° preoperatively to 94.51° intraoperatively and 82.68° at 4-year (mean) follow-up. Garland and Orwin35 in a case series, examined the effect of HO excision to improve range of motion in 19 individuals with SCI. The study found that the largest gain of function occurred intraoperatively followed by a large loss of function within the first 6 months. At final follow-up (6 years after surgery) of the 24 hips excised, 3 had similar or less motion compared with preoperative motion, 15 improved between 10 and 39° and 6 showed >40° improvement.

In two case series (mentioned previously in the bisphosphonate section), excision and bisphosphonate treatment (etidronate,28 pamidronate29) was effective in preventing secondary HO formation.

In conclusion, there was Level 4 evidence that resection of HO about the hip after SCI can improve restricted hip range of motion.34, 35 There was Level 4 evidence that resection combined with bisphosphonate treatment halts the recurrence of HO.28, 29

Discussion

Our review focused on interventions designed to either prevent or treat primary or secondary HO development after SCI. Figure 1 outlines the known mechanisms of action for those interventions involved in the treatment of HO after SCI. Only three of the interventional studies reviewed were rated as high-quality RCTs.3, 22, 23 Two of these RCTs evaluated NSAIDs,3, 22 whereas one showed the effectiveness of PLIMF therapy in the prevention of HO.23 Treatment studies tended to involve smaller numbers of patients and were of lower quality when compared with prevention studies.

Figure 1
figure 1

Mechanism of HO intervention action after SCI.

Full size image

There was strong evidence that NSAIDs (rofecoxib and indomethacin) prevented the development of HO when provided early after SCI. Inflammation seems to have an important role in neurogenic HO and NSAIDs have been shown to prevent HO by reducing inflammation and blocking prostaglandin synthesis.22 Macfarlane et al.36 reported that the systemic inhibition of prostaglandins, especially with the use of indomethacin, reduced the incidence of HO significantly following hip and acetabular surgery in non-SCI populations. Prostaglandins regulate mesenchymal cell differentiation into osteoblastic cells, which are important for new bone formation.37, 38 These prostaglandins may be indirectly responsible for bone morphogenic protein expression in soft tissue, which is another factor in HO development. NSAIDs, by inhibiting COX enzymes, block prostaglandin synthesis which in turn is thought to inhibit heterotopic bone formation.3, 22

Both of the NSAID studies examining prophylaxis3, 22 were Level 1 studies, used the same diagnostic tools for HO (three-phase scintigraphy and radiography), and initiated treatment around 20–25 days after SCI; both reduced the incidence of HO when compared with placebo. However, patients treated with rofecoxib, a selective COX-2 inhibitor, suffered less gastrointestinal side effects than those treated with indomethacin. Indomethacin is a nonselective COX inhibitor22 and has long been associated with a greater side-effect profile. Furthermore, rofecoxib is no longer used in clinical practice because of its cardiovascular adverse effects. Another NSAID used in HO prevention, celecoxib, has been shown in other populations to be as effective an anti-inflammatory medication as indomethacin with fewer gastrointestinal adverse effects.39 Hence, it is reasonable to assume that it may be as effective in patients with SCI and should be further studied in this population.

Radiotherapy for prevention of HO after SCI also acts through the inhibition of mesenchymal cell differentiation. Given radiotherapy has been successful in the treatment of HO after hip arthroplasty,40, 41 it has been suggested that it may also be an appropriate treatment for other high-risk populations such as traumatic individuals with SCI; however, there is only limited published research evidence on the use of radiation in the prevention of HO in the SCI population.32, 33 Although there have been studies suggesting that it was effective in preventing or treating HO progression, all were of low methodological quality and more research needs to be conducted to be more confident regarding its effectiveness in the treatment and prevention of HO after SCI.

PLIMF therapy was supported by a small single Level 1 study.23 It seems to help prevent HO by increasing blood flow and oxygenation to the treated area. This may in turn remove the neurogenic stimulus that activates the HO pathway, thereby preventing HO formation. The study found a significant difference in the incidence of HO between the treatment and control groups at the end of the study period. However, bone formation may take longer than 2–3 months to form; therefore, further follow-up results are required to assess its true effectiveness in preventing HO after SCI. Research involving larger populations and longer follow-ups are needed to evaluate its clinical relevance.

Radiotherapy and the use of PLIMF therapy as a prophylaxis involves radiating or treating all patients with SCI, which makes it clinically less appealing given the costs and equipment training involved. Radiotherapy specifically may also result in other secondary complications such as carcinogenesis in treated individuals and therefore must be performed cautiously.33

The first-generation bisphosphonates (that is, etidronate) are analogs of inorganic pyrophosphate.29 Several studies have looked at etidronate in the treatment of HO (n=5);24, 25, 26, 27, 28, 29 however, the methodological quality of the research was not strong. Furthermore, many studies involved treatments for short duration and had dosage levels below normal treatment levels to effectively evaluate the effect of treatment. The lack of RCTs makes definitive statements difficult; nonetheless, research to date suggests etidronate delays or inhibits HO progression once diagnosed, and seems to be more effective when given earlier rather than later after an SCI. As etidronate inhibits the transformation of amorphous calcium phosphate into crystalline hydroxyapatite and thus blocks the mineralization of the bone matrix, it seems to act later in the HO developmental pathway than NSAIDs or radiotherapy.42 Hence, etidronate does not inhibit the production of the protein matrix but rather the mineralization of calcium phosphate on the matrix.1 Furthermore, it has been known to be difficult for patients to administer and is poorly tolerated due to adverse effects. Hence, etidronate may not be the optimal treatment choice in individuals with SCI because of the apparent need to continue it indefinitely.

The new-generation nitrogen-containing bisphosphonates such as pamidronate are more potent and result in less adverse effects. However, only one Level 4 study (with a small sample size) assessed the effectiveness of pamindronate in secondary prevention of HO.29 Further research is needed.

Warfarin as prophylaxis for HO after SCI has been suggested on the basis of the results of a single but intriguing observational study, which reported a significant association between warfarin administration and failure to develop HO. The mode of action of warfarin is likely early on through inhibition of bone matrix formation.1 Further investigation into the efficacy of warfarin in preventing HO after SCI is required before recommending its use because of concerns about the potential side effects, in particular, the increased risk of hemorrhage. Furthermore, studies involving stronger methodological quality and longer follow-up time are required to assess its clinical effectiveness.

Surgical resection of heterotopic bone was evaluated in four Level 4 studies.28, 29, 34, 35 Studies indicated that the optimal time for surgical resection of HO is after the bone matures. However, at present, there is no definitive clinical sign or test to assess the maturity of the heterotopic bone and whether the HO process is completed.28, 29, 34 Radiographs combined with clinical judgment are generally used to determine when HO is complete. Although surgical excision has been regarded as the best clinical option, all studies reviewed involved the use of additional treatment in conjunction with surgical resection.

An important limitation in assessing the effectiveness of interventions was the lack of standardization in diagnosing HO after SCI. HO was diagnosed using nuclear bone scans, radiographs, clinical symptoms and Brooker grades in various studies. To effectively compare the results of various interventions, a standardized tool for diagnosing HO must be used in future studies. Furthermore, most prophylactic studies diagnosed HO based on radiographic evidence; however, the clinical relevance was not assessed. The clinical relevance of a diagnostic tool is important to take into account the effectiveness of any prophylactic treatment.

Conclusions

HO can have a negative impact on the quality of life of individuals with SCI. For the prevention of HO after SCI, there was Level 1 evidence that NSAIDs (rofecoxib and indomethacin) and PLIMF prevented the development of HO after SCI. However, clinical application of these treatments may be complicated by adverse effects or limited by feasibility. However, these adverse effects may be small when compared with the negative impact of HO development. There was only Level 4 evidence that radiotherapy prevented the progression of HO after SCI, indicating the need for more research. One interesting observation (Level 4 evidence) was the absence of HO when warfarin was used. For the treatment of HO, once diagnosed, there was Level 2 evidence that bisphosphonates stop the progression of HO; however, it seems to have no effect on the deposition of the bone matrix and only stops bone mineralization while the patient is taking the medication. Hence, once treatment is stopped, mineralization may subsequently progress. Furthermore, dangerous adverse effects have been reported with high doses of bisphosphonates. Surgical excision of HO along with supplemental therapies was supported by four case series. Overall, owing to the generally poor quality of the studies, more research is needed into the various clinical interventions used to prevent and treat HO in patients with SCI before definitive recommendations can be made.