Incidence of heterotopic ossification following hip arthroscopy is low: considerations for routine prophylaxis

Purpose This scoping review aims to map and summarise the available literature on heterotopic ossification (HO) following hip arthroscopy, with particular focus on incidence, distribution as per Brooker classification, efficacy of prophylactic measures and factors that may influence the likelihood of production of HO. Methods A computer-based search was performed on PubMed, Embase, Emcare, Cinahl, ISI web of science and Scopus using the terms ‘heterotopic ossification’ and ‘hip arthroscopy’. Articles reporting heterotopic ossification following hip arthroscopy for any condition were included after two-stage title/abstract and full-text screening. Results Of the 663 articles retrieved, 45 studies were included. The proportion of patients with HO ranged from 0 to 44%. The majority of the cases were either Brooker grade I or II. Of the six studies investigating the effect of NSAID prophylaxis, five reported a significantly lower incidence of heterotopic ossification associated with its use. Weak evidence suggests that an outside-in arthroscopic approach, no capsular closure, male sex and mixed cam and pincer resection may be associated with an increased risk of HO. Conclusion Although there is a large variation in rates of HO following hip arthroscopy in the current literature, the majority of studies report a low incidence. Evidence exists advocating the administration of post-operative NSAIDs to reduce the incidence of HO following hip arthroscopy. This, combined with the low risk of complications, means there is a favourable risk–benefit ratio for prophylactic NSAID used in HA. Future research should work to identify patient clinical and demographic factors which may increase the risk of development of HO, allowing clinicians to risk stratify and select only specific patients who would benefit from receiving NSAID prophylaxis. Supplementary Information The online version contains supplementary material available at 10.1007/s00264-022-05402-4.

HO is presumed to be caused following soft tissue injury and a subsequent inflammatory cascade. The resultant inflammatory environment leads to angiogenesis, progenitor cell differentiation and ectopic bone formation from Level of Evidence: IV.
* Vikas Khanduja vk279@cam.ac.uk differentiated cell types [22,23]. Whilst Brooker grades I and II HO may be asymptomatic, grades III and IV may present with stiffness, pain and reduced range of motion [24,25]. Some evidence exists highlighting the role of nonsteroidal anti-inflammatory drugs (NSAIDs) in the reduction of HO following hip arthroscopy [22,24,26]. This effect is thought to occur through modulation of osteoprogenitor cells and interference of cell signalling pathways [27,28]. However, the debate on routine prophylaxis for HO following HA continues. This scoping review aims to map and summarise the available literature on HO following hip arthroscopy, with particular focus on incidence, distribution as per Brooker classification, efficacy of prophylactic measures and factors that may influence the likelihood of production of HO. By doing so, we aim to enhance the reader's understanding of this common complication, inform clinical decision-making with regard to the use of prophylactic measures and critically identify areas for future research. We hypothesise that whilst the incidence of HO may be increasing, the overall incidence remains low and that prophylaxis may not be required for all patients.

Materials and methods
The methodological framework for scoping reviews was first outlined by Arksey and O'Malley in 2005 and more recently updated by Levac et al. and The Joanna Briggs Institute [29][30][31]. These articles outline five common steps in a scoping review, all of which are outlined below.

Identifying the research question
The following review questions were developed: 1) What is the incidence of HO following hip arthroscopy? 2) What prophylactic measures are reported and what is their efficacy? 3) In which locations does HO occur? 4) What is the distribution of severity in terms of the Brooker classification? 5) What patient-and treatment-related factors are associated with the development of HO?

Identification of relevant studies
Free text and medical subject heading (MeSH) terms including 'heterotopic ossification', 'hip arthroscopy' and 'femoroacetabular impingement' were used, with the Boolean operators 'and', 'or' used to combine search terms as appropriate. The full search strategy is available in the Appendix. The search was initially performed in PubMed, before being adapted for and used in five other databases including, OVID Embase, OVID Emcare, Scopus, ISI Web of Science and CINAHL. All searches were computer-based and performed on 17th January 2021. Reference list checking was also performed using review articles identified by the above search process.

Study selection
Selected manuscripts were imported into Rayyan systematic reviews web application (Qatar Computing Research Institute, Doha, Qatar) for screening and selection [32]. A two-stage title/abstract and full-text screening was performed by two authors independently, using the outlined selection criteria: Studies describing treatment of heterotopic ossification were also included as were those that reported no complications following hip arthroscopy, with the assumption that no patients showed signs of HO. Studies which reported results of any hip arthroscopy procedure yet did not specifically describe the formation of HO were excluded. Studies reporting a mini-open approach or the use of both an arthroscopic and open treatment were also excluded. Non-English articles were excluded at the fulltext screening stage rather than through imposing limits on the database search. This allowed the display of these potentially relevant foreign language articles in an appendix, to ensure transparency.
Differences in opinion regarding the inclusion or exclusion of articles were first resolved by discussion between the two authors and, failing this, by consultation with a third author.

Collating, summarising and reporting results
The number of studies retrieved and removed following each screening stage are shown in the PRISMSA flow diagram ( Fig. 1) [33]. Study characteristics, including first author, type of study, number of hips and patients, patient age and sex and follow-up period are shown in Table 1. Forest plots generated using R studio are used to display the incidence of HO in included studies (Fig. 2). The distribution of severity of HO according to the Brooker classification is displayed in Table 2. A qualitative thematic approach was used throughout, with results reported according to the key themes described, including incidence of HO, Brooker classification, prophylactic measure used and their effects and factors associated with the development of HO. This approach is commonly used in scoping reviews and aids in the identification and mapping of key themes within a broad topic [31,34].

Quality assessment and risk of bias
Although the assessment of quality of the studies and risk of bias forms a key part of systematic reviews, these steps are not required in a scoping review [29][30][31].

Results
The search strategy outlined, resulted in the identification of 663 unique articles, of which 45 (6.8%) were finally included. Most articles (33/45, 73.3%) were level IV evidence, whilst ten level III studies (22.2%) and two level I randomised control trials (4.4%) were also included ( Table 1). The total number of hips was 12,613 (12,538 patients) and the mean pooled age was 34.4 years. A forest plot displaying the overall incidence of HO is shown (Fig. 2). Rates of HO vary widely from 0 to 44%. Of the included studies, 14 (31.1%) report an incidence of HO of 1% or less, whilst a total of 30 (66.7%) studies report an incidence of under 5% and 36 (80%) studies less than 10%. Unfortunately, it was not possible to pool individual study results to derive an overall incidence due to low level of evidence included and inherent risk of bias and heterogeneity associated with such study designs [77,78].
Of the included studies, 19 reported a severity stratification of HO cases in terms of the Brooker classification (

Prophylactic measures
A total of 16 included studies (34.8%) reported on the prophylactic use of NSAIDs (naproxen, celecoxib, indomethacin or aspirin). Six studies directly compared the development of HO in those receiving NSAID prophylaxis to controls receiving no prophylaxis. A summary of these studies is shown in Table 3. Of these, five reported a significantly reduced incidence of HO in those receiving NSAIDs compared to those receiving no treatment. Two studies, Bedi et al. (2012) and Randelli et al. (2010), compare the efficacy of two different NSAID regimens [24,61].

Factors influencing the development of heterotopic ossification
A total of six studies described other factors which may influence the development of HO. A description of these factors is provided in Table 4.

Discussion
Our study found that the overall incidence of HO after arthroscopy of the hip is low. Although individual studies report a HO incidence between 0 and 44%, two thirds of included studies describe an incidence of 5% or under and one third report an incidence under 1%. Although no formal meta-analysis was possible, majority of cases were Brooker grade 1 or 2 cases, with more severe cases rarely found in 11 studies. Given the rise in popularity of hip arthroscopy and corresponding increase in publications in its field, it is not surprising that the majority of included articles were published in the last decade [80].
Although there is a large amount of literature concerning the topic of HO following hip arthroscopy, with 45 articles included in this review, almost 75% are level IV evidence, with limited higher levels of evidence as exemplified by the availability of only two randomised control studies. Due to the high degree of heterogeneity present in level III and IV studies, pooling of rates of HO in these studies was not possible, and hence a statistical comparison against the 30 Table 3 Results of those studies comparing heterotopic ossification occurrence in patients given prophylaxis to those receiving no prophylaxis or comparing heterotopic ossification development after the use of two different NSAID regimes. PO by mouth, BD twice a day, QD once daily, RR risk ratio  [56] Celecoxib 200 mg QD, 3 weeks (122) No prophylaxis (240) 0% 3/240 (1.3%) No significant association between prophylaxis sand HO incidence Randelli [61] Etoricoxib 90 mg daily, 3 weeks (15), Naproxen 500 mg PO BD, 3 weeks (248), Others-aceclofenac,i ndomethacin,ketopro fen, 3 weeks (22) No prophylaxis (15 studies not reporting any use of prophylactic measures was not performed. A large variation in reported rates of HO is seen. The majority (36/45, 80%) of studies report rates of less than 10%, with six articles reporting no cases of HO in their cohorts. However, rates as high as 36% and 44% are reported in the studies of Amar et al.(2015) and Rath et al.(2013), respectively [21,35]. One reason for this large variation may be the use of prophylactic NSAID therapy to prevent HO. Such a strategy has been described in 16 (34.8%) of the included studies. Another explanation for this variation may be because radiographs are not obtained routinely, six months or a year post-HA in all included studies, which may produce an underestimate of HO incidence.
Of the six studies directly comparing the occurrence of HO in those receiving prophylactic NSAIDs with controls, five report significantly decreased rates in the former group [26,36,42,56,61,62]. There is therefore robust evidence suggesting that post-operative prophylactic NSAID administration is effective in reducing the incidence of HO following HA. Nevertheless, questions remain regarding its routine use in all patients undergoing HA, and it is therefore important to carefully consider the risk-benefit ratio. Firstly, the use of NSAID prophylaxis does not completely eliminate the risk of HO, with rates as high as 22.  [24,42]. Furthermore, the results of this review suggest that the large majority of HO cases may be classified as Brooker Grade I or II, which typically does not present with symptoms and is instead detected on radiographic imaging [24]. There is also currently no evidence to suggest that prophylactic NSAID therapy may decrease the severity of HO according to the Brooker classification, with Beckmann et al.(2015) describing no significant difference in Brooker classification between patients receiving prophylactic NSAIDs and controls receiving no prophylaxis [36]. Furthermore, there is no evidence to suggest that the development of HO affects patient outcomes following hip arthroscopy. Dow et al.(2020) report no significant difference in iHOT33 values between patients with and without HO, whilst no significant difference was also found between the prophylaxis and control group in terms of the proportion of patients achieving MCID at 6 months, 1 year or 2 years post-operatively [42]. Similarly, Zheng et al. (2020) found significant post-operative improvements in mHHHS VAS, HOS-ADL and HOS-SS in those with asymptomatic HO [76].
Whilst NSAID use is cost effective and generally considered safe, complications may occur in some, particularly those with pre-existing cardiovascular or renal risk [81][82][83][84][85][86][87]. However, the majority of studies included in this review did not routinely collect or report on side effects or complications associations with NSAID use. Of the two studies that reported this information, one reported a case of acute renal failure, one haematochezia following acute colitis and three cases of gastritis [36], whilst the other reported relatively minor side effects such as headache, weight gain and gastrointestinal upset [26]. It is likely that the patients Table 4 Summary of factors which may affect the incidence of heterotopic ossification following hip arthroscopy

Author Factors
Beckmann [36] In those patients receiving femoral osteoplasty, the degree of resection was significantly higher in those who went on to develop HO (18.9°) compared to those who did not (12.3°), P = 0.036. Of those given prophylactic NSAID therapy, 3 patients who underwent capsular repair developed HO, whilst 8 who did not undergo capsular repair developed HO, although this difference was not significant. No significant difference in HO development was found between the first and second 46 cases of the performing surgeon. In a multivariate logistic regression model, mixed type FAI resection was associated with an increased risk of HO compared to CAM only resection (Odds ratio 52.5, P < 0.011). No significant associated observed with regards to age or sex Bedi [24] Of the 29 cases of HO, 7 occurred in patients undergoing cam femoral osteoplasty, 2 in those receiving isolated acetabular resection for pincer impingement and 20 in those receiving mixed resection. However, no significant association was found between type of procedure and development of HO, likely due to the small numbers involved. Most cases of HO occurred in male patients receiving osteoplasty for FAI, during which the capsule was cut. Multivariate logistic regression found no association between type of procedure and HO development Dow [42] OF the 92 cases of HO, significantly more (P < 0.001) occurred in males (69/92, 75%) compared to females. This significant difference was also seen specifically in both the treatment and control groups where 75.8% and 73.3% of patients who developed HO were male Randelli [61] Significantly lower HO incidence was seen in the NSAID prophylaxis group. However, no significant difference in age, sex, weight or type of procedure performed (pincer rim trimming or cam head neck junction osteoplasty) seen between the treatment group receiving NSAID prophylaxis and controls Rath [62] Significantly lower HO incidence was seen in those receiving prophylaxis compared to controls. The latter group also had a significantly longer mean surgery time of 121.9 min, compared to 106.2 min in the control group Rath [21] Bivariate logistic backward stepwise regression analysis showed no significant association between sex, diagnosis, procedure performed, anchor use and surgery time and HO development undergoing hip arthroscopy are in general a healthy population with few concomitant cardiovascular or renal diseases. This combined with the relatively short period of NSAID administration (Table 3) and the lack of evidence suggesting a high risk of serious side effects indicates that the safety profile of NSAIDs should not be a major concern in this clinical context. Despite our results suggesting the majority of HO are Brooker grade I or II, this does not provide an adequate rationale against routine NSAID prophylaxis as it does not necessarily reflect the proportion of patients requiring revision excision. For example, 7/29 patients in the cohort of Bedi et al.(2012) required revision excision of HO, all of which were either grade I or II cases [24]. Furthermore, reoperation for excision rates of between 9.8% and 37.5% are reported in the included studies. It is thus clear that even a small area of HO, with a Brooker grade of I or II may cause significant symptoms requiring excision [24]. This effect may be explained through consideration of the location of ectopic ossification. For example, small areas of ossification located in the joint capsule, iliopsoas or rectus femoris may cause impingement on movement and hence require excision due to functional deficit and/or pain [24].
One way in the need to prevent HO may be balanced against the potential complications and may be through alterations in dosage. For example, Mortensen et al.(2020) found no significant difference in the rate of HO between those receiving two weeks and three weeks of Naproxen 500 mg PO BD [54]. Careful patient stratification and selection using risk factors for the development of HO may provide another suitable option.
Some studies have suggested factors such as male sex, arthroscopic approach, lack of capsular closure, length of operation and type of procedure (cam resection, pincer resection or mixed resection) may influence the development of HO [24,35,36,42,62,69]. However, these factors are derived from small-scale individual studies with low-quality study designs and therefore cannot be used to guide clinical decision-making. This is a key area of further research, with more high-quality comparative studies required in determining patient or treatment-related risk factors for the development of HO. This will not only enable the stratification and selection of appropriate patients for NSAID prophylaxis but also the development of an evidence-based protocol for the prevention of HO development following HA. However, until the influence of these factors can be determined, it is more appropriate to administer prophylactic NSAIDs to all patients undergoing HA, owing to their efficacy at preventing HO and the subsequent need for surgical excision and favourable safety profile. To aid the development of a HO reduction protocol, further research should also investigate other prophylactic techniques, such as radiation therapy. Research suggests this technique is effective in reducing the risk of HO following open total hip arthroplasty; however its use in HA has not been investigated [88,89].

Limitations
It is important to be aware of potential differences in radiographic follow-up between studies, which may also partly account for the large variation in rates of HO. Some studies included in this review focus specifically on the incidence of HO after hip arthroscopy, providing patients with regular radiographic imaging, whilst others have a more general aim of reporting outcomes/complications after hip arthroscopy. Our results suggest that the majority of cases of HO are likely to be Brooker grade I or II that is often asymptomatic and detected only through radiographic imaging. It could therefore be the case that the latter group of studies, which often have long periods of clinical follow-up, but might not provide patients with such thorough radiographic follow-up, may provide an underestimate of HO rates. Dow et al. (2020) describe 67 cases of HO after six months radiographic follow-up, 85 after one year and 92 after two years [42]. Although the majority of cases occur in the first six months, an increase is seen across a two year period, despite the loss to follow up of 150 patients. This highlights the effect a potential lack of radiographic follow-up may have on the underestimation of HO rates.

Conclusion
Although a large variation in rates of HO following HA is seen in the current literature, the majority of studies report a low incidence and most cases described are typically asymptomatic Brooker Grade I or II presentations. Whilst evidence suggests that post-operative NSAIDs do reduce the incidence of heterotopic ossification following hip arthroscopy, their routine administration in all patients undergoing HA needs to be considered in terms of the risk-benefit ratio, specifically given our findings suggesting a low incidence of symptomatic HO. Identification of patient-and treatmentrelated risk factors associated with the production of heterotopic ossification may allow clinicians to risk stratify patients, thereby aiding in patient selection and decisionmaking with respect to the use of NSAIDs for prophylaxis. However, as such a strategy cannot yet be implemented in a safe manner, prophylactic NSAIDs should be administered to all patients undergoing HA.

Data availability Not applicable.
Code availability Not applicable.

Declarations
Ethics approval Ethical approval was not required as this study only contains data extracted from previous studies which are freely available online.
Consent to participate Not applicable for a systematic review.

Consent for publication
Not applicable for a systematic review.

Competing interests
The authors declare no competing interests. VK is a Consultant for Smith and Nephew & Arthrex Limited.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visithttp:// creat iveco mmons. org/ licen ses/ by/4. 0/.