NSAIDs
The prophylactic effectiveness of indomethacin following elective lower limb surgery is well-accepted, and has been shown to help prevent HO, lessen the extent of the development of HO and assist in preventing inflammation associated with acute HO; it is currently part of a widely-employed prophylaxis protocol[
1,
5,
24‐
27]. The drug is a simple and low cost option for prophylaxis. It is important to note that several trials have shown that aspirin does not appear to have the same beneficial effects as indomethacin[
27,
28]. The drug works by inhibiting prostaglandin-mediated (specifically PGE-2) bone remodeling and also by directly inhibiting the differentiation of osteoprogenitor cells[
1,
24,
25]. A Cochrane Review was conducted in 2004, comprising 16 randomized controlled trials analyzing the effectiveness of prevention of HO with the use of NSAIDs[
28]. The use of NSAIDs peri-operatively was shown to reduce the risk of developing heterotopic bone formation by 59% over placebo. Another systematic review by Neal, et al. demonstrated a 57% reduction in the risk of HO when NSAIDs were used as prophylaxis[
27]. The statistical data from this study leads the authors to predict that for every 100,000 THAs performed in the US each year, perioperative NSAID use has the potential to prevent anywhere from 10 to 20,000 cases of heterotopic bone formation. Further, the review showed no significant difference in the incidence of gastro-intestinal complications between those treated with NSAIDs and the control. Though both review studies demonstrated a significant reduction in incidence of HO, neither addressed clinical outcomes of pain and physical function among treatments.
However not all trials have seen such a benign side-effect profile with the use of NSAIDs. Karunakar et al[
26]. conducted a randomized, prospective double-blind placebo-controlled clinical trial on the effect of indomethacin after the operative treatment of fractures of the acetabulum. Though as the authors point out the study is lacking in power, it is significant in that it was determined that there was no significant difference in the incidence of HO between the indomethacin and placebo treatment groups. However, it was noted that one of the indomethacin patients developed gastrointestinal hemorrhage and one had a perforated ulcer, while 13 total patients withdrew from the study due to side effects of the medication, compared with only 1 withdrawing in the placebo treatment. Others such as Banovac[
24] have used misoprostol to aid in the prevention of gastrointestinal complications. Another problem with non-selective (ie those actively inhibiting cyclo-oxygenase 1 and 2, or COX-1 and COX-2) NSAIDs such as indomethacin is increased perioperative bleeding secondary to inhibition of COX-1, leading to reduced production of thromboxane A2, which is essential to platelet aggregation[
29]. Fransen et al. followed up her 2004 Cochrane review with a randomized controlled trial comparing postoperative pain and physical function in patients taking either ibuprofen or placebo following total hip arthroplasty and revision total hip arthroplasty surgery[
30]. Though the overall risk of HO was reduced by 31% (as assessed radiographically) with use of ibuprofen, the study demonstrated no clinically significant difference in either pain or physical function 6 to 12 months postoperatively. Further, though the risk reduction is impressive, the raw data (ie number of patients affected) does not suggest as large of an impact on the actual incidence of Brooker Grade 3 and 4 HO. Among those patients in the ibuprofen treatment set, 2.5% developed Brooker Grade 3 or 4 HO, whereas the placebo treatment demonstrated an incidence of 6%[
30]. Further, there was a significant increase in major bleeding complications in the ibuprofen treatment group, again serving to emphasize untoward side effects of this treatment modality[
30].
One method of circumventing the pitfalls of NSAIDs use is by employing COX-2 selective NSAIDs, such as meloxicam. Weber et al. have shown that use of this selective NSAID following total hip arthroplasty was associated with a 17% reduction in intraoperative and postoperative (first 24 hours following surgery) blood loss[
29]. Furthermore, fewer gastrointestinal side-effects have been noted as a result of preferential use of COX-2 selective NSAIDs [
31‐
35]. As noted in additional File
1, more recent studies have highlighted the both these advantages as well as efficacy comparable to that of traditional, non-selective NSAIDs[
32,
34‐
37]. However, it is notable that the practice of prescribing COX-2 specific NSAIDs is not without consequences, as several trials have shown an increased risk of cardiovascular events associated with their use [
38‐
40]. Despite this, a few trials have shown either no increased risk for cardiovascular events when compared to non-slective NSAID use[
41,
42] (albeit with lower rates of GI side effects) or a decreased risk (in the case of celecoxib[
43]) as shown in one study. Due to the current lack of evidence for safety of routine use of COX-2 selective NSAIDs as prophylaxis for postoperative HO, indomethacin remains the gold standard of treatment when employing NSAIDs therapy[
44].
A serious problem with the use of high-dose indomethacin and other NSAIDS for HO prophylaxis is that while new heterotopic bone may be prevented from forming, the formation of bone for healing the fracture site may also be impaired. Thus, the large number of reports of nonunion or malunion as well as poor ligament healing[
5,
25,
45]. One such study by Burd et al. noted 29% incidence of nonunion of long bone fractures following indomethacin prophylaxis, whereas in the radiation arm the incidence was just 7%[
25]. Of note, there were no instances of acetabular nonunion. In a study by Persson et al., 142 patients were followed for formation of HO following THA. Of the 11 that underwent a revision procedure secondary to aseptic loosening, 10 belonged to the indomethacin group[
46].
Radiation Therapy
Studies by Cooley and Goss[
47] in 1958 and later those by Craven and Urist[
48] in 1971 demonstrated the effects of irradiation therapy on bone growth and repair. After demonstrating the inhibiting influence of the radiation on bone repair of rat bone and seeing that the effects were more pronounced when the treatment was initiated closer to the time of the fracture, the authors hypothesized that the early osteoprogenitor cells involved in bone repair were more radiosensitive than the more mature cells seen later[
4]. In 1981, Coventry et al. established the utility of radiotherapy (RT) as prophylactic treatment for HO by irradiating the hips of 42 patients who had undergone hip surgery[
49] 48 hips, each designated as high risk for the formation of HO, were treated with a 20 Gy dose of radiation. Of those treated, 19% developed ectopic bone, noting that patients treated earlier enjoyed lower rates of HO, though data specifics were not included. Today, RT is used prophylactically (albeit in much smaller doses) both pre- and postoperatively for the prevention of HO following bone fracture or manipulation secondary to trauma or operative treatment[
3,
4,
50]. Recently, Childs et al. noted that based on a retrospective cohort study of 263 patients having experienced traumatic acetabular fracture, HO was discovered in 5.3% of patients receiving RT, while 60% of patients who did not receive treatment developed some degree of ectopic bone[
51]. Further, a study by Chao et al. has demonstrated RT to be capable of preventing HO in high-risk patients, specifically those in whom a history of HO exists[
52].
Several studies comparing preoperative to postoperative radiation therapy (RT) were reviewed by Balboni et al. in a critical review of RT for HO prophylaxis[
4]. Studies cited by the authors suggest that there is not a statistically significant difference between employing preoperative (<4 hours preoperatively) or postoperative (<72 hours postoperatively) RT. In particular the authors cite Gregoritch et al[
50] and Seegenschmiedt et al[
53] In Gregoritch's study 122 patients undergoing THA and deemed at a high risk of HO were treated with either preoperative or postoperative RT. The authors reported no significant difference among the treatment arms, noting a 28% incidence in the postoperative treatment and 26% incidence in the preoperative treatment (
p =
1.0). They calculated a 5% incidence of clinically significant (Brooker Grade 3 or 4) HO in the postoperative group and a 2% incidence in the preoperative group (
p =
0.58). In Seegenschmiedt's study, 161 patients were randomized to receive either preoperative or postoperative RT. Of those treated, 4 (5%) failures were noted in the postoperative group, compared with 11 (19%) in the preoperative group (
p <
0.05). However, the authors noted that the majority of failures in the preoperative group occurred in patients with preexisting Brooker Grade 3 or 4 HO which had not been removed prior to treatment. Thus, the authors conclude that there both pre- and postoperative treatment are equally effective.
Childs et al. had similar results regarding timing of treatment in a study evaluating postoperative RT for HO prophylaxis following traumatic acetabular fractures[
51]. Of 152 patients studied, 58 received radiation within 24 hours of surgery, 41 within 48 hours, 53 within 72 hours, 13 within 4 days and 4 were delayed longer than 4 days. The authors noted no increase in HO when prophylactic RT was initiated anywhere from one to four days postoperatively.
There are several potential side effects of RT, the most concerning of which is the theoretical effect of carcinogenesis; however there has yet to be a documented case of a radiation-induced tumor after RT for HO prophylaxis[
4]. This positive outcome is thought to be the effect of both low doses of radiation as well as an older patient population – as the latency period for induction of malignancy following RT is from 15 to 24 years, it is a possibility that there are too few patients that survive long enough after treatment for the carcinogenic effects to be realized[
4,
50]. Therefore, as this treatment option is employed for younger patients, this concern is worth considering.
Another possible complication of RT (and one that also plagues indomethacin treatment, as mentioned above) is the risk of bony nonunion, as has been demonstrated when trochanteric osteotomy is necessary to remove the prosthesis during a revision procedure[
4,
49,
54]. Rates of nonunion range from 12–30% after RT[
4], with the highest in Lo et al., in which the authors noted that of the 6 patients in whom osteotomy was necessary, 2 developed a nonunion[
54]. In non-irradiated trochanteric osteotomy, the rate of nonunion is diminished, occurring only 2–15% of the time[
4]. Therefore some authors have advocated the use of shielding to prevent nonunion[
9,
50].
Finally, radiation dose to the testis is also a concern with the use of radiation prophylaxis. Animal studies have shown that reversible oligospermia can be induced with doses as low as 20 to 70 cGy, and doses of 120 cGy have been shown to cause permanent azoospermia[
4]. The same study also measured effective testicular dose, showing an average of 25.1 cGy, with a 54% reduction to 11.3 cGy using a testicular shield. For this reason it is recommended that testicular shields always be used during this treatment and that patients be made aware of the risks involved with this treatment modality.
Regarding which of these treatments (ie NSAIDs or RT) is most effective, several trials have found both are effective means of prophylaxis following arthroplasty and trauma. They point out that clinical decisions should not be made based on efficacy but rather on factors such as availability, side effects and cost [
55‐
57]. Strauss et al. (2008) recommend that, when accounting for efficacy and all costs associated with the use of NSAIDs versus RT (including patient disability due to HO) for prophylaxis, RT has the advantage due to a lower incidence of serious side effects[
58]. This attitude is not supported universally, though several other trials have found an advantage in efficacy using RT rather than NSAID therapy[
9,
59]. Kölbl et al[
59] conducted a trial in which 301 patients were randomized to either an NSAID treatment arm, a single 5 Gy fraction of RT or a 7 Gy fraction of RT. 113 patients were randomized to the NSAID arm, 95 to the 5 Gy arm and 93 to 7 Gy. The data from the study supported the 7 Gy therapy as being the most effective postoperative treatment schedule in prevention of clinically significant (Brooker 3 or 4) heterotopic ossification. A study by Pakos et al[
9] had similar results, demonstrating a difference in effective HO prophylaxis for Brooker grade 3 and 4, albeit only a 1.2% absolute risk reduction.