Comparing Options for Heterotopic Ossification Prophylaxis following Elbow Trauma: A Systematic Review and Meta-Analysis

Abstract

Introduction  Heterotopic ossification (HO) can be a potentially serious and devastating complication following traumatic injury to the elbow. HO prophylaxis options include nonsteroidal anti-inflammatory drugs (NSAIDs) and radiation therapy (RT) but neither has been proven more effective. The purpose of this review is to compare effectiveness and outcomes between NSAID and RT prophylaxis for HO about the elbow following a traumatic injury.

Materials and Methods  We performed a systematic review of PubMed and Cochrane Library for cases of HO prophylaxis following elbow trauma utilizing PRISMA guidelines to determine the most effective form of prophylaxis. Outcomes of interest included recurrence of HO, range of motion (ROM), and Mayo elbow performance index (MEPI). A total of 36 articles and 826 elbows of which 203 received RT and 623 received NSAID were identified and included in the final analysis.

Results  Rates of HO formation or recurrence following elbow trauma were similar between radiation and NSAID prophylaxis (15.6% vs. 22.2%, respectively p = 0.457). ROM was similar in flexion and extension arc (109.0 degrees in radiation vs. 112.8 in NSAIDs, p = 0.459) and in pronation and supination arc (118.9 degrees radiation vs. 134.7 degrees NSAIDs, p = 0.322). MEPI scores were 79.19 in the radiation group and 88.82 in the NSAIDs group at the final follow-up.

Conclusion  There is no statistical difference in HO development, recurrence, or final ROM between NSAIDs and RT prophylaxis following trauma to the elbow. We recommend the choice of modality based on patient characteristics, cost, and surgeon preference.

Level of Evidence  Level III.

Introduction

Heterotopic ossification (HO) can be a potentially serious and devastating complication following traumatic injury to the elbow. Rates of HO following elbow trauma can be as high as 50% in some series. 1 2 Despite proper management at the time of injury, HO is a late complication that can cause pain, restricted motion, and lower quality of life. 3 4

HO prophylaxis options include nonsteroidal anti-inflammatory drugs (NSAIDs) and single-dose radiation therapy (RT). The exact mechanisms by which NSAIDs and RT inhibit HO formation are unknown. Osteoblastic mesenchymal progenitor cells have high-mitotic rates and thus are thought to be sensitive to radiation, which is similar to mechanisms by which RT induces apoptosis of metastatic cancers. 5 NSAIDs may inhibit HO formation through inhibition of bone morphogenic protein-related signaling. 6

While both modalities have been studied extensively, there exists little consensus on the most effective form of HO prophylaxis in the elbow following trauma. One systematic review examined the effectiveness of RT on HO prophylaxis about the elbow, but no studies have compared modalities, nor have they limited comparisons to HO prophylaxis in the setting of traumatic injury to the elbow. 7 The purpose of this review is to compare the effectiveness and outcomes between NSAIDs and RT prophylaxis for HO about the elbow following traumatic injury.

Materials and Methods

We performed a systematic review of PubMed and Cochrane Library for cases of heterotopic ossification prophylaxis following elbow trauma utilizing Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines ( Fig. 1 ). Search terms for studies included “Elbow OR Medial condyle OR lateral condyle OR distal humerus OR radial head OR coranoid OR olecranon OR terrible triad AND fracture AND radiation AND prophylaxis” and “Elbow OR Medial condyle OR lateral condyle OR distal humerus OR radial head OR coranoid OR olecranon OR terrible triad AND fracture AND NSAID OR indomethacin OR celecoxib AND prophylaxis.”

Fig. 1.

Preferred reporting items for systematic reviews and meta-analyses ( PRISMA ) flow diagram.

HO prophylaxis was primary (before the development of HO) or secondary (after excision of previously developed HO). Case series, retrospective studies, observational cohort studies, and randomized controlled trials were included. Articles were excluded for HO etiology other than fracture about the elbow or subjects that received both prophylactic interventions. Outcomes of interest included occurance or recurrence of HO, range of motion (ROM) in both flexion and extension as well in pronation and supination at the elbow, and Mayo elbow performance index (MEPI) scores.

A total of 35 articles were included in the final analysis ( Table 1 ). Separate analyses were subsequently performed for ROM at the final follow-up in flexion and extension arcs ( n = 20) as well as pronation and supination arcs ( n = 10). MEPI means were calculated but not compared due to underreporting in studies. Continuous variables were examined using weighted means and Wald’s test for p values. Categorical variables were examined via proportions testing. All statistical analyses were done using R Studio (Version 3.6.3, Vienna, Austria).

Table 1. Data extracted from included articles.

Author	Year	# Elbows	Age	# Male	Follow-up (months)	Radiation dose (Gy)	Fractions	Prophylaxis timing	HO development/recurrence	Preop ROM flexion extension arc	Preop ROM pronation supination arc	Postop ROM flexion extension arc	Postop ROM pronation supination arc	Preop MEPI	Postop MEPI
Cullen, et al 24	1994	1	40	1	24	200	5	Secondary	0	–	0	–	120	–	–
Robinson et al 25	2010	31	42	27	8.7	700	1	Secondary	3	35	47.5	101	132.5	–	–
Koh et al 26	2013	48	–	–	–	700	1	Secondary	6	–	–	116	–	–	–
McAuliffe et al 27	1997	1	71	0	52	200	5	Secondary	0	20	–	125	–	–	–
McAuliffe et al	1997	1	27	1	45	200	5	Secondary	0	10	–	115	–	–	–
McAuliffe et al	1997	1	34	1	25	200	5	Secondary	0	0	–	90	–	–	–
Hamid et al 2	2010	21	44	14	7.5	700	1	Primary	7	–	–	87	141	–	69
Mishra et al 11	2011	21	–	–	–	700	1	Secondary	2	–	–	–	–	–	–
Stein et al 28	2003	11	51	3	12	700	1	Primary	3	–	–	114.5	173	–	–
Lo et al 29	1996	1	63	0	1	600	1	Primary	1	–	–	–	–	–	–
Lo et al	1996	1	71	1	8	700	1	Secondary	0	–	–	–	–	–	–
Heyd al 30	2009	20	47.4	13	43.3	700	1	Primary/Secondary	2	–	–	–	–	–	95.75
Abrams et al 31	1993	1	34	1	21	700	1	Secondary	0	–	0	–	140		–
Abrams et al	1993	1	67	1	43	1000	4	Secondary	0	–	0	–	160		–
Yang et al 32	2003	4	41	6	17	700	1	Secondary	0	–	–	116	–	–	–
Henket et al 33	2007	13	25	8	42	700	1	Secondary	4	74	0	98	102	–	77
Haglin al 34	2018	20	–	–	–	700	1	Secondary	3	–	–	–	–	–	–
Poggi et al 35	1999	1	20	1	10.5	700	1	Secondary	0	–	–	80	20	–	–
Poggi et al	1999	1	22	1	10.5	700	1	Secondary	0	–	–	105	–	–	–
Poggi et al	1999	1	36	0	10.5	800	1	Secondary	0	–	–	97	–	–	–
Lindenhovius et al 36	2007	2	–	–	–	700	1	Secondary	0	–	–	–	–	–	–
Kamineni et al 37	2002	1	28	1	144	–	–	Secondary	0	10	0	25	115	50	75
Sun et al 38	2015	77	43.6	40	9	200	Celecoxib	Secondary	8	38	98	122	142	–	–
Li et al 39	2019	17	46.5	10	8.7	200	Celecoxib	Secondary	17	–	–	–	–	–	–
Li et al	2019	20	43.1	11	9.2	200	Celecoxib	Primary	2	–	–	–	–	–	–
Bimmel et al 40	2006	1	52	0	9	25	Indomethacin	Primary	1	–	–	130	0		–
Viola et al 41	1999	12	28	14	24	25	Indomethacin	Secondary	0	43	79	120	152	–	–
Bell et al 42	1999	1	33	–	12	30	Indomethacin	Secondary	0	70	0	95	150
Giannicola et al 10	2020	12	46	6	20.5	100	Indomethacin	Secondary	1	85	0	116	123	65	89
Kamrani et al 15	2014	15	29	–	31	25	Indomethacin	Secondary	4	78	0	81	101	57	91
Sonderegger et al 12	2012	1	54	–	26	28	Indomethacin	Secondary	0	70	25	80	140	–	–
Sonderegger et al	2012	1	48	–	29	29	Indomethacin	Secondary	0	55	0	80	150	–	–
Sonderegger et al	2012	1	45	–	17	27	Indomethacin	Secondary	0	105	0	120	155	–	–
Sonderegger et al	2012	1	61	–	54	25	Indomethacin	Secondary	0	130	0	130	135	–	–
Sonderegger et al	2012	1	24	–	15	26	Indomethacin	Secondary	0	160	0	160	155	–	–
He et al 43	2018	25	36	17	33	25	Indomethacin	Secondary	6	35	–	93	–	47	85
He et al	2018	17	37	13	42	25	Indomethacin	Secondary	4	16	–	98	–	37	91
Haglin et al 34	2018	11	–	–	–	–	Indomethacin	Secondary	1	–	–	–	–	–	–
Ahmed et al 44	2013	1	–	–	5	–	Indomethacin	Secondary	0	0	–	60	–		–
Ahmed et al	2013	1	–	–	7.9	–	Indomethacin	Secondary	0	20	–	60	–		–
Ahmed et al	2013	1	–	–	4.4	–	Indomethacin	Secondary	0	–	30	–	160		–
Sun et al 3	2019	49	36	38	69	200	Celecoxib	Secondary	2	27	115	131	145	54	95
Tarallo et al 45	2017	13	–	–	–	200	Celecoxib	Primary	2	–	–	–	–	–	–
Wang et al 46	2019	15	44.3	5	32.5	25	Indomethacin	Primary	1	–	–	112.7	160.4	–	89
Kayalar et al 17	2008	18	27	13	47	–	Indomethacin	Secondary	0	25	–	92	–	–	–
Chen et al 13	2015	52	38.3	35	15	25	Indomethacin	Secondary	15	38	102	122	122	55	92
Chen et al	2015	112	35.9	64	13	25	Indomethacin	Secondary	30	37	111	124	102	57	91
Zhang et al 16	2014	21	38.4	17	32	25	Indomethacin	Primary	2	–	–	126	139.1	–	95.2
Sreenivas et al 47	2013	1	42	1	18	75	Indomethacin	Secondary	0	0	0	95	173	–	–
Sandeep et al 48	2017	9	38.7	7	18.1	25	Indomethacin	Secondary	1	20	20	100	144	–	81
Lindenhovius et al 36	2007	3	–	–	–	25	Indomethacin	Secondary	0	–	–	–	–	–	–
Zhou al 14	2017	38	37	21	31	25	Indomethacin	Secondary	1	27	148	126	153	68	96
Chen al 49	2015	12	33.6	8	15.5	25	Indomethacin	Primary	3	–	–	105	126		–
Durakbasa et al 18	2013	15	36	5	50	25	Indomethacin	Primary	7	–	–	–	–	–	83.3
Rex et al 50	2008	47	30.8	36	33.9	25	Indomethacin	Secondary	0	33.9	–	105	–	66.59	93.82
Kamineni et al 37	2002	1	58	1	24	–	Indomethacin	Secondary	0	95	140	110	175	55	85
Kamineni et al	2002	1	56	0	60	–	Indomethacin	Secondary	0	60	30	120	100	50	75

Results

A combined 203 elbows received RT and 623 received NSAIDs prophylaxis ( Table 2 ). The average age of combined subjects is 41.6 in RT and 36.9 in the NSAIDs group ( p = 0.220). As much as 68.5% of those who received RT are male and 64.7% of those who received NSAIDs are male ( p = 0.693). Follow-up duration is 20.7 months on average in the RT group and 29.4 months in the NSAIDs group ( p = 0.109) Rates of HO formation or recurrence following elbow trauma are similar between RT and NSAID prophylaxis (15.6% vs. 22.2%, respectively p = 0.457). ROM is similar in flexion and extension arc (109.0 degrees in RT vs. 112.8 in NSAIDs, p = 0.459) and in pronation and supination arc (118.9 degrees RT vs. 134.7 degrees NSAIDs, p = 0.322). Average MEPI scores are 79.19 in the RT group and 88.82 in the NSAID group at the final follow-up.

Table 2. Heterotopic ossification and range of motion in RT versus NSAIDs prophylaxis.

Variable	Radiation	NSAIDs	p -Value
Abbreviations: MEPI, Mayo elbow performance index; NSAIDs, nonsteroidal anti-inflammatory drugs; RT, radiation therapy. a No comparison due to insufficient reporting in included studies.
Number of subjects	203	623	–
Age	41.67 (34.47–48.86)	36.92 (34.50–39.33)	0.220
Male	68.50% (49.81%–82.65%)	64.74% (57.55%–71.31%)	0.693
Follow-up time	20.72 (12.89–28.54)	29.44 (22.18–36.70)	0.109
HO development/recurrence	15.68% (10.79% - 22.22%)	11.77% (5.87%–22.22%)	0.457
Flexion/extension arc	109.02 (100.87–117.18)	112.87 (106.81–118.93)	0.459
Pronation/supination arc	118.92 (89.21–148.63)	134.75 (124.84–144.65)	0.322
MEPI	88.82 ( n = 16)	79.19 ( n = 4)	a

Discussion

Relative to other joints, the elbow is especially prone to HO formation. The exact etiology of HO formation at the elbow is unclear but likely is multifactorial. 8 HO at the elbow can form in the setting of genetic predisposition, upper extremity neurologic injury, traumatic brain injury, previous surgery (iatrogenic), or trauma. 8 9 10 Following elbow trauma, there is no clear consensus on the best prophylactic HO prevention strategy.

This meta-analysis compares demographics, rates of HO formation, and ROM and MEPI scores between NSAID and RT HO prophylactic management strategies. No differences were found in any combined outcome measures across 35 articles and 826 total subjects with elbow HO due to trauma. Both RT and NSAIDs are viable options to prevent HO formation in this setting. Patient and surgeon preferences, patient-specific factors, clinical judgment, and cost and available resources should guide decision-making when deciding on optimal HO management strategy in at-risk patients.

RT can be given as either a preoperative or postoperative dose, usually of 7 Gy in one to five fractions, depending on the protocol. A previous systematic review examined the safety and efficacy of RT prophylaxis for HO at the elbow but did not compare RT to other treatment strategies. They found RT to be safe with no differences in primary or secondary prophylaxis as well as the timing of RT with regard to surgical procedures. 7 One study in both this systematic review as well as our meta-analysis presented here randomized patients to RT or no therapy and found an unacceptably high rate of nonunion in the RT group. 2 The same study also reported two cases of infection in the RT group. No other studies reported on nonunion or infection as a complication. Only one other complication that occured was triceps rupture in one subject. 11 Many factors may contribute to complications in RT including timing, dose, fractions, and predisposing patient factors, however, the overall complication rate remains low.

Various NSAID medications can be used for HO prophylaxis. Indomethacin 25 mg three times daily and celecoxib 200 mg two times daily were the most frequently used dose and frequencies for HO prophylaxis. 12 13 14 Timing varied from 2 to 6 weeks postoperatively. Various complications were reported in the NSAIDs cohort including infection, nerve palsy, hardware failure, and nonunion. 3 15 16 17 18 Rates of complications reported in studies were very low and comparable to those in the RT therapy group.

This is the first known meta-analysis to compare HO prophylaxis options in the elbow. A similar study compared NSAID to RT for HO prophylaxis at the hip in a meta-analysis of randomized controlled trials (RCTs). 19 They found no differences in effectiveness or complications between either treatment strategy, while also noting the very heterogeneous nature of studies included with regard to treatment protocols. Other joints have been studied as well, including the knee 20 21 and ankle. 22 23 Results of these studies can be extrapolated to the elbow, however, the elbow is unique in that it is more prone to HO formation relative to other joints. More research is needed to determine precise pathophysiologic pathways that contribute to elbow HO formation to develop effective treatment options.

There are limitations to this meta-analysis. Compiled articles were not limited to randomized controlled trials, and no attempt was made to compare study quality. As such, the data represented here is only as good as the quality of studies included. To increase the quality of our meta-analysis, we only included studies that reported outcomes of interest and limited the number of outcomes to increase the number of studies and subjects that met inclusion criteria. Patient populations across studies are inherently heterogeneous. HO prophylaxis was either primary (before the formation of HO) or secondary (after HO formation or after attempted excision of HO). We included patients with only traumatic etiology and excluded neurologic injury to increase the homogeneity of our cohort. Few studies included patient-reported outcome measures. Therefore, we were not able to make meaningful comparisons in this domain. Recurrence of HO and ROM may not be the best indicators of clinical success. A RCT is needed to make the most accurate assessments with regard to the quality and outcomes of different HO management strategies.

Conclusion

No differences were found in demographic risk factors, HO occurrence, and ROM and MEPI scores between NSAIDs and RT prophylaxis for HO of the elbow caused by trauma. Situational factors should be considered when deciding on a proper prophylactic treatment strategy. RCTs are needed to better delineate clinical differences between different HO prophylaxis strategies.