Radiotherapy for the prophylaxis of heterotopic ossification: A single 7–8 Gy fraction seems optimal

Abstract

Heterotopic ossification (HO) is the pathological development of ectopic bone within soft tissues typically as a complication of trauma, surgery, neurological injury or severe burns. Most common in the hip region, it may also affect other sites including the elbow, shoulder, knee and temporo-mandibular joint. In the early stages, it is asymptomatic, evident on imaging only, but may progress to severe debility from pain and reduced range of movement, through to complete ankylosis [1]. Recently, several novel prophylactic strategies have been investigated (e.g. pulsed electromagnetic fields, free radical scavengers, bone morphogenic protein inhibitors). However, the two most established treatments are oral non-steroidal anti-inflammatory drugs (NSAIDs) and low dose radiotherapy (RT). Both are thought to act ultimately by inhibiting conversion of so-called pluripotent mesenchymal stem cells to osteoblasts in the traumatized tissue. The outcomes of these two preventative strategies have been reported in numerous retrospective studies and prospective randomized trials, including head-to-head comparisons suggesting that they have similar efficacy, with acceptable though obviously very different side effect risk profiles [2]. For a variety of reasons, it has proven challenging to come to definitive conclusions on optimal management of this condition. There are many questions regarding the RT option in particular: What is the optimal dose and fractionation? What is the optimal timing (preor post-operatively)? Are the answers different for different sites, for preventing severe vs mild HO, or for preventing progression of pre-existing vs de novo HO? Is RT truly iso-effective to NSAIDs? Which is more globally cost-effective? What about the combination? Are the newer strategies better than either? How does each option impact upon patients’ quality of life? In the current issue of the Journal, the Toronto Sunnybrook group presents a meta-analysis of randomized radiotherapy trials on prophylaxis of HO using a literature search over the period 1946 to early 2015 [3]. The same group published a broader meta-analysis using the same search terms (but including non-randomized data) in the October 2014 issue of the Journal, searching between 1946 and mid 2013 [4]. The latter included 61 studies, each with more than 5 treated sites (total 5464 sites) and the current paper is based on 1253 of those sites from randomized trials published between 1992 and 2010. The term ‘‘sites’’ is used throughout, rather than patients, because a few patients had RT to more than one site. Nearly all treated sites were hips (97.7% and 98.3% from the earlier and current papers, respectively); the remainder were elbows (1.6%, 1.7%), temporomandibular joints (0.4%, 0%) and knees (0.4%, 0%). The main outcome measure in both papers was the percentage of sites which developed HO. Two secondary outcome measures were also reported, namely percentages of sites which developed Brooker grade 1–2 HO (asymptomatic) and grade 3–4 HO (clinically significant). The actual results were not provided in the earlier paper, except for the development of HO in the whole group of patients (24.8%). In the second paper, the incidence of HO was tabulated for each trial arm, which enabled us to explore the data further. We here refer to the trials by the first author’s surname and year of publication; the full citations are given by Milakovic et al. [3]. When we examined the 12 papers (13 trials) they reviewed, we found several errors: incorrect fractionation for two trials (10 Gy/2, instead of 10 Gy/5), incorrect calculation of the Biologically Effective Dose (BED) for trials with 5 Gy/1 (12.33 Gy, instead of 13.33), and reporting the prevalence of HO in patients at 6 months, rather than the development of HO, in 8 of the 20 RT trial arms (Seegenschmiedt 1997, Pellegrini 1992 and 1996). We present the corrected data in the Table 1 and Fig. 1. We had to exclude 9 sites from one trial (Padgett 2003) because they had previous ipsilateral HO and it was unclear how much HO developed after treatment. We attempted to answer the following two questions.