Chemoprophylaxis for Heterotopic Ossification After Hip Arthroscopy: A Systematic Review and Meta-analysis

Methods

Study protocol for this systematic review was performed following PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines. A meta-analysis was performed on a subset of the included articles.

Study Eligibility

The inclusion criteria were English-language studies reporting the incidence of HO after primary hip arthroscopy. All included studies were required to report if patients did or did not receive chemical prophylaxis; and, when prophylaxis was prescribed, what the prophylaxis regimen was, as well as what the agent used was. Studies were excluded if they included nonhip arthroscopy data, included revision hip arthroscopy, failed to state if chemical prophylaxis was or was not used, and failed to specify which prophylaxis was used. Studies were included if they contained data with all patients receiving prophylaxis. Systematic reviews, review articles, editorials or commentaries, instructional courses, case reports, and publications in languages other than English were also excluded. Two independent reviewers assessed the eligibility of each article (R.M. and D.M.). Reviewer disagreements were discussed, and a third independent reviewer (R.S.D.) was used to achieve consensus. The initial search identified 410 articles, of which 44 were included in the final analysis (Figure 1).

OPEN IN VIEWER

Figure 1.

PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) flow diagram demonstrating selection and screening for included papers in study.

Data Extraction and Statistical Analysis

The 44 included studies were evaluated, and data relating to study design, methodology, patient demographics, surgical procedure and indication, prophylactic agent utilized, use, dosing and length of treatment of postoperative prophylaxis, presence of HO adjacent to the operative hip, and length of follow-up. The Brooker classification system was used to determine the degree of HO from postoperative anteroposterior radiographs of the operative hip (Figure 2). ⁷ Data relating to symptoms from HO, revision surgery, and complications from prophylactic treatments were collected. Patients were grouped in 2 groups: prophylaxis group and no prophylaxis group.

OPEN IN VIEWER

Figure 2.

Brooker classification of hip heterotopic ossification. (A) Grade I is isolate bone island, (B) Grade II is bone spurs from acetabulum or greater trochanter with greater than 1 cm distance between ends, (C) Grade III is less than 1 cm between ends, and (D) Grade IV is bridging of the hip joint. Reproduced with permission from Bedi et al. ⁷

The risk of bias was assessed via the methodological index for non-randomized studies criteria. Higher scores are associated with a lower risk of bias, with a maximum score of 24 for comparative studies and 16 for noncomparative studies.

Descriptive data were listed as mean ± SD when appropriate, and continuous variables were compared using independent t test when normally distributed as determined by Anderson-Darling testing; otherwise, the Mann-Whitney U test was performed. Categorical data were compared using chi-square tests. Meta-analysis was performed on the subset of studies that evaluated HO in both non-prophylaxis and prophylaxis cohorts. The DerSimonian and Laird method of random-effects meta-analysis was used for all extrapolated and pooled outcome data. Heterogeneity between studies was quantified using I², with values greater than 75% indicating high heterogeneity. Two-tailed P-values < .05 were determined to be statistically significant. All analysis and creation of tables and plots were performed using open-source R software Version 4.4.1. Because included studies recorded both the development of HO after surgery and patient follow-up, the incidence of HO was determined from the combined data.

Results

From the 44 studies, a total of 32,603 patients were included in the combined analysis. Of those, 7971 patients across 19 studies received NSAID prophylaxis postoperatively, whereas 24,632 patients across 30 studies did not receive prophylaxis of any kind.

Demographic factors relating to the combined data set are listed in Table 1. There were significantly more male patients in the prophylaxis group (53.5% vs 41.2%; P < .001). Age, body mass index, and length of follow-up were not statistically significantly different.

OPEN IN VIEWER

Table 1

Patient Demographics of Included Studies, Compared Between Patients Receiving HO Prophylaxis and Those Who Did Not Receive HO Prophylaxis ^a

	Prophylaxis	No Prophylaxis	P
Patients, n	7971	24,632
Age, y	34.4 ± 6.59	34.9 ± 8.28	.80
Sex, male, %	53.5	41.2	<.001
BMI	25.4 ± 1.50	26.0 ± 1.46	.17
Follow-up time, mo	19.0 ± 8.70	23.1 ± 7.51	.09

a

Data are presented as mean ± SD unless otherwise indicated. Bold P value <.05 indicates statistical significance. BMI, body mass index.

All but 2^7,23 of the included studies reported their indications for hip arthroscopy. The most common indications for hip arthroscopy among the included studies was femoroacetabular impingement (FAI): 7 studies included only those patients who underwent hip arthroscopy for FAI.^{3,4,9,19,20,22,25} Each of the remaining studies included cohorts made up primarily of FAI patients but also included pathologies such as hip dysplasia, ¹⁶ labral lesions, ⁸ chondromatosis, ²⁸ and snapping hips. ²⁸ Additionally, the most common surgeries performed were labral repair, labral debridement, femoral osteoplasty, acetabular osteoplasty, synovectomy, chondroplasty, microfracture, iliopsoas release.

Table 2 lists the number of patients in each group who developed radiographic HO after hip arthroscopy. Among the included patients, 8.0% of those who did not receive prophylaxis developed HO compared with 2.4% of patients with prophylaxis who developed HO; 22 studies included data relating to the severity of HO using the Brooker classification with a total of 163 patients from the prophylaxis group and 127 patients in the no prophylaxis group. The breakdown by Brooker classification is provided in Table 2. The prophylaxis group had a higher distribution of patients with Brooker 1 HO compared with the no prophylaxis group; this difference was deemed to be statistically significant (χ² = 8.656; P = .03).

OPEN IN VIEWER

Table 2

Development of HO Between Patients Who Received HO Prophylaxis and Patients Who Did Not (Severity of HO Subdivided Based on Brooker Classification) ^a

	Prophylaxis	No Prophylaxis	P
Patients, n	7971	24,632
HO	190 (2.4)	1973 (8.0)	<0.001
Patients, adjusted n b	163	127
Severity of HO			0.03 c
Brooker 1	120 (73.6)	80 (63.0)
Brooker 2	25 (15.3)	38 (29.9)
Brooker 3	17 (10.4)	10 (7.9)
Brooker 4	1 (0.6)	0 (0)

a

Data are presented as n (%) unless otherwise indicated. Bold P values indicate statistical significance. HO, heterotopic ossification.

b

Totals shown are the adjusted number of patients with HO in each group. Percentages are based only on studies that included data on Brooker classification.

c

χ² = 8.656, P = .03.

For patients who did not receive prophylaxis, the incidence was 80.1 cases per 1,000 patients per year, which was statistically significantly higher than the incidence of 15.3 cases per 1000 patients per year in the patients who received chemoprophylaxis (P = .01).

Table 3 lists the different prophylactic regimens and the number of patients to receive each treatment. All included studies that used chemoprophylaxis used NSAIDs as the agent; however, there was wide variability in the specific medication used, dosing, and length of prophylaxis with 14 different protocols. The most used protocol overall was naproxen 500 mg twice daily for 3 weeks, which was used on 2760 patients between 5 different studies.^{5,11,21,28,33} Other NSAIDs used included aspirin, ³¹ diclofenac, ³⁴ celecoxib,^13,15 etodolac, ³⁰ indomethacin,^14,21 and etoricoxib. ²⁷ The percentage of patients on each prophylactic regimen who eventually developed HO ranged from 0% to 36%.

OPEN IN VIEWER

Table 3

Different HO Prophylactic Treatment Medication and Dosing Schedules ^a

Medication	Dose	Duration, weeks	Patients, n	% With HO
Naproxen	500 mg 1-2x per day	2-6	4384	92 (2.1)
Aspirin	325 mg BID	6	2379	19 (0.8)
Diclofenac	50 mg TID	2	700	15 (2.1)
Celecoxib	200-400 mg qD	2-6	292	1 (0.3)
Etodolac	600 mg qD	2	100	36 (36.0)
Indomethacin	75 mg qD	3	62	0 (0)
Etoricoxib	60 mg qD	2	54	0 (0)

a

χ² = 588.487; P < .001. BID, twice a day; HO, heterotopic ossification; qD, once a day; TID, 3 times per day.

Six of the studies in the systematic review included both prophylaxis and no prophylaxis group. These data were pooled and underwent meta-analysis. The odds ratio for developing HO in the prophylaxis groups compared with those without prophylaxis was 0.10 (95% CI, 0.03-0.35; P < .001), indicating a statistically significant decrease in HO among those receiving prophylaxis (Figure 3).

OPEN IN VIEWER

Figure 3.

Forest plot showing the individual odds ratios of the 6 studies to compare prophylaxis and no prophylaxis patients. Combined odds ratio with 95% CI as well as heterogeneity are included.

The studies included in this systematic review were inconsistent in reporting complications related to both side effects of NSAIDs as well as reoperations because of symptomatic HO after the initial procedure. Two studies report a total of 25 adverse events in 244 patients receiving HO chemoprophyaxis.^5,6 This included 7 with gastrointestinal upset, 2 with heartburn, 3 reporting rash, 1 with acute renal failure, and 1 with acute colitis resulting in a total of 8 patients discontinuing NSAID prophylaxis because of side effects. Seven studies reported results on reoperation for symptomatic HO.^{6,7,13,14,16,31,36} A total of 228 patients among these studies developed HO with 59 undergoing resection. Among 6 of these studies, 17% of patients who received HO chemoprophylaxis and developed HO required resection (16/92), and 24% of patients who did not receive HO chemoprophylaxis and developed HO required resection (40/167). However, data collection was not sufficient for comparative analysis to determine if prophylactic NSAID use either increases risk related to side effects or decreases revision surgery for symptomatic HO.

Discussion

The primary findings of this study demonstrate that the use of NSAIDs postoperatively significantly decreases the incidence and severity of HO after hip arthroscopy. Despite the variety of NSAID regimens utilized, these medications helped mitigate the risk of developing radiographic HO. While most patients who develop HO will be asymptomatic, some will develop significant pain and stiffness to the point of requiring reoperation. ¹⁶ These findings provide surgeons with a pooled reference to justify NSAID prophylaxis for HO after hip arthroscopy.

The incidence of HO in patients without prophylaxis reported in this analysis was 8%, which is similar to previous studies.^1,2,26 Comparatively, the incidence with the use of chemoprophylaxis was found to be significantly lower (2.4%). Results of the meta-analysis suggested up to a 10-fold decrease in HO after hip arthroscopy with chemoprophylaxis. NSAIDs have also been used recently in place of radiation for prophylaxis of HO in patients undergoing total hip arthroplasty, and while several meta-analyses show higher rates of HO in hip arthroplasty (9.6%-52.7%) as compared with hip arthroscopy, there is a similar ratio of decrease in incidence within patients who received prophylaxis.^18,24,35 Isolated use of NSAIDs for surgical treatment of traumatic fractures, however, has not been shown to significantly decrease the incidence of HO when compared with no prophylaxis.^3,22

Among patients who received prophylaxis, there was an additional benefit seen in the severity of HO. Based on Brooker classification, patients with chemoprophylaxis were likely to have lower grade HO. Interestingly, 73.6% of patients who received prophylaxis and subsequently developed HO were classified as Brooker 1, which is associated with no or minimal clinical symptoms. Conversely, 37.0% of patients without prophylaxis had a higher grade of radiographic HO according to the Brooker classification (Table 2). This suggests that NSAIDs prophylaxis not only decreased the incidence of HO formation, but also the severity of extra bone formation in patients who did end up developing HO. This is supported by the predicted role of NSAIDs in interfering with signaling pathways and altering osteoprogenitor cells, which modify the surrounding tissue to create an environment conducive to bone formation.^4,9,19,25 In effect, while prophylactic NSAIDs may not entirely prevent HO in some patients, the blunted effect limits the extent of bone formation and potentially reduces clinical symptoms.

Our results indicated that NSAIDs were effective in reducing HO postoperatively, with a wide variability between the specifics of postoperative chemoprophylaxis. The most used prophylaxis method was Naproxen 500 mg twice daily for 3 weeks.^{5,11,21,28,33} Aspirin 325 mg twice daily for 6 weeks was used for a large number of patients; however, these were all from a large single study. ³¹ As a result, while NSAIDs overall showed a positive effect, it is difficult to discern whether one NSAID is better than another or if certain dosing regimens are more effective. Six of the included studies used COX-2 inhibitors for prophylaxis.^{7,10,15,27,30,34} COX-2 inhibitors are favored over other NSAIDs because of decreased side effects. ⁸ A recent meta-analysis by Zhu et al ³⁷ demonstrated no difference in HO after total hip arthroplasty when comparing selective COX-2 inhibitors with nonselective NSAIDs. This suggests that selective COX-2 inhibitors can be considered when there is a contraindication to nonselective NSAIDs.

Conclusion

HO is a common complication after hip arthroscopy. Postoperative use of NSAIDs is effective in reducing the incidence and severity of HO in the surgical hip. A meta-analysis of matched-cohort studies estimated a 10-fold decrease in incidence. Further evaluation of NSAID dosage and time is necessary to maximize the benefits of prophylaxis while limiting harm from side effects.