Application of hyperbaric oxygen therapy in femoral head necrosis: a systematic review and meta-analysis

Abstract

Purpose

• To evaluate the effectiveness of hyperbaric oxygen therapy (HBOT) in the treatment of osteonecrosis of the femoral head (ONFH).

Methods

• Search for term systems related to ONFH and HBOT in PubMed, Cochrane Library, Embase and Web of Science databases. The risk ratio (RR) was used as the effective index for dichotomous variables while mean difference (MD) and 95% confidence interval (CI) were used as the effective index for continuous variables, with a two-sided P < 0.05 considered statistically significant. I² statistic and Q test were used to evaluate the statistical heterogeneity of the included studies.

Results

• Ten studies were included, involving 568 participants. Pre-post meta-analyses to show the visual analog scale after HBOT (MD = −2.94, 95% CI: −4.27, −1.60, P < 0.0001), short form 12 physical component summary scale (SF12 PCS) (MD = 17.28, 95% CI: 8.26, 26.31, P = 0.0002), short form 12 mental component summary scale (SF12 MCS) (MD = 4.26, 95% CI: 2.56, 5.95, P < 0.00001), Harris hip score (HHS) (MD = 26.91, 95% CI: 0.35, 53.46, P = 0.05), modified Harris hip score (mHHS) (MD = 44.31, 95% CI: 13.75, 74.87, P = 0.004) were significantly different from those before treatment. The results of SF12 PCS (MD = −0.39, 95% CI: −7.85, 7.06, P = 0.92) and SF12 MCS (MD = 0.76, 95% CI: −7.02, 8.52, P = 0.85), patients’ improved events (RR = 1.83, 95% CI: 0.37, 9.09, P = 0.46) showed no significant difference between the HBO and non-HBO groups.

Conclusion

• The results of this study indicate that HBOT cannot be regarded as an effective measure for the treatment of early-stage necrosis of the femoral head and more large-scale randomized controlled trials are needed for further verification.

Introduction

Osteonecrosis of the femoral head (ONFH) is a disease caused by blood circulation disorders in bone joints under the stimulation of various harmful factors such as physical trauma, long-term exposure to steroids, alcohol abuse, chemotherapy and so on (1). The clinical manifestations of ONFH include femoral head collapse, hip pain and dysfunction (2). Microfracture of bone trabeculae and local osteocyte necrosis of the femoral head were observed under microscope (3). It is estimated that there are about 7.5–10 million patients with ONFH in China and about 30 million worldwide (2). ONFH mainly occurs among people aged 20–60 years, with an average age of 38 years (4). The mechanism of ONFH is still not fully understood but it is thought to be related to pathophysiological processes including apoptosis and autophagy, endothelial cell injury and coagulation dysfunction, lipid metabolism disorder and oxidative stress (5). At present, the treatment of ONFH remains a challenge.

Hyperbaric oxygen therapy (HBOT), in which the patient’s body is placed in a pressurized environment containing 100% oxygen, two to three times atmospheric pressure at sea level, can cause oxygen tension in the arteries to exceed 2,000 mmHg and oxygen tension in the tissues to almost 400 mmHg (6). HBOT is mainly used in clinical diseases such as carbon monoxide poisoning, decompression sickness and arterial gas embolism (6). Compared to the more invasive interventions currently used to treat ONFH such as core decompression (CD) or total hip replacement, HBOT is a new attempt by researchers to explore conservative treatments (7). According to the Tenth European Consensus Conference on Hyperbaric Medicine, HBOT is an effective biophysical treatment for early-stage femoral head necrosis which is yet unapproved globally (8). Bosco et al. found that HBOT achieved therapeutic effect by regulating the level of IL-1β, TNF-α and IL-6 to modulate inflammation and oxidative stress (7). Wu et al. found that the expression of S100 calcium binding protein A9 decreased in the femoral head under HBOT which can promote the angiogenesis of the femoral head (9). Although two previous systematic reviews (10, 11) have addressed this topic, their inclusion of non-English literature and the emergence of recent high-quality studies necessitate an updated meta-analysis to consolidate current evidence.

Our hypothesis is that HBOT can improve the functional score of patients with ONFH and contribute to the improvement of their condition. This study conducted a comprehensive review of the relevant literature to determine whether HBOT is truly safe and effective in the treatment of ONFH. In conclusion, this paper provides more evidence-based medical evidence for clinical selection of better treatment methods.

Methods

We conducted this systematic review and meta-analysis based on Methodological Guidelines of the Cochrane Handbook of Systematic Reviews (12) and the PRISMA (Preferred Reporting Program for Systematic Reviews and Meta-Analyses) (13) (Supplementary Material S1 (see section on Supplementary materials given at the end of the article)). We registered this systematic review on the PROSPERO website on March 10th, 2024 (CRD42024509483).

Search strategy

The search was conducted in four databases, PubMed/MEDLINE, Cochrane Library, Embase and Web of Science, and up until October 2023. The specific retrieval process and details can be referred to Supplementary Material S2.

Eligibility criteria

Population, intervention, comparison, outcomes and study (PICOs) questions are the basis for eligibility criteria for systematic reviews and meta-analyses (14). P: Osteonecrosis of the femoral head; I: hyperbaric oxygen therapy; C: blank or use other physical therapy (such as extracorporeal shockwave therapy and CD); O: function score and favorable events; S: randomized controlled trials and cohort studies.

Inclusion criteria: i) necrosis of the femoral head; ii) HBO treatment or treatment involving HBO; iii) there are no restrictions on the patient’s race, gender or nationality.

Exclusion criteria: i) incomplete data and unavailability of full texts; ii) non-English literature; iii) osteonecrosis without specific data on ONFH; iv) review, comment, case series studies and case reports; v) mechanism research.

Data extraction and quality assessment

Two researchers (YC and PT) extracted the data and assessed its quality and if there were any differences, a third researcher (YL) helped evaluate the data and reach a consensus. The author, year, country, number of patients, age, follow-up time and other information were extracted. The main outcome indicators extracted in this study included: i) visual analog scale (VAS); ii) short form 12 physical component summary scale (SF12 PCS); iii) short form 12 mental component summary scale (SF12 MCS); iv) Harris hip score (HHS); v) modified Harris hip score (mHHS); vi) patients’ improved events. Using the Cochrane offset risk assessment tool and the Newcastle-Ottawa Scale (NOS); two reviewers (YC and PT) independently assessed the methodological quality of included studies. The Cochrane offset risk assessment tool, mainly used to evaluate randomized controlled trials, consists of six domains, each of which can be classified into three grades: low, medium and high, and an overall grade evaluation is made. The NOS for cohort studies contains seven items and each item needs to be rated as ‘yes’ or ‘no’ which means complete or no report respectively.

Statistical analysis

Review manager 5.4 was used for meta-analysis of the included literature. The dichotomous outcomes were represented by risk ratio (RR) and the continuous outcomes were represented by mean difference (MD). Effect sizes were assessed with 95% confidence intervals (95% CI) with a P-value less than 0.05 considered significant. Heterogeneity of included studies was assessed using Cochran’s Q statistics and I² statistics. When the I² of the study indicators was <50%, the fixed-effects model was used for meta-analysis and when the I² of the study indicators was >50%, the random-effects model was used for meta-analysis. Significant clinical heterogeneity was addressed using methods such as sensitivity analysis or only descriptive analysis.

Results

Search results

Initially, 164 articles were obtained by using the search strategy, 62 duplicated articles were eliminated, and 18 articles were obtained after reading the title and abstract strictly according to the inclusion and exclusion criteria. Finally, eight studies were excluded by reading the full text (Supplementary material S3), and ten studies were included. Figure 1 depicts the literature screening process.

Figure 1.

The preferred reporting items for systematic reviews and meta-analysis (PRISMA) flow diagram to show study selection.

Study characteristics

A total of ten articles were included in the meta-analysis, published between 2000 and 2021. It included three randomized controlled trials (15, 16, 17) and seven cohort studies (18, 19, 20, 21, 22, 23, 24). Table 1 lists the basic information for all included studies. In the included studies, eight studies studied necrosis of the femoral head, and the other two studies studied transient osteoporosis of the hip (20) and bone marrow edema respectively (18). All studies were treated with hyperbaric oxygen. All studies included 586 participants in total, all using magnetic resonance imaging (MRI). Nine of the studies had populations with a mean age of 35 years or older, and one study focused on children (mean age 8 years) (18). Eight studies had follow-up of more than 3 months, one study had no follow-up (19), and one study had an unknown follow-up duration (18). Seven studies provided the staging of femoral head necrosis. As for Journal Citation Reports (JCR) ranking, one study is in Q1 (15), three are in Q2 (16, 17, 22), four are in Q3 (18, 21, 23) and two are in Q4 (19, 20, 24). All ten studies reported respective HBO treatment schemes. We also separately summarized the conclusions of the included studies. Camporesi et al. demonstrated that HBOT can effectively treat patients with ONFH without the need for hip arthroplasty surgery (15). Bozkurt et al. agreed with Camporesi et al. that HBOT reduced pain and improved functional scores in stage II patients (16). Hsu et al. did not observe a synergistic effect of extracorporeal shockwave therapy (ESWT), HBO and alendronate over ESWT alone and called for long-term experiments (17). Reis et al. concluded that hyperbaric oxygen is effective in the treatment of stage I femoral head necrosis (23). Moghamis et al. concluded HBO is as promising and effective as CD in the treatment of ONFH before non-traumatic collapse of the femoral head and recommended HBO as a non-invasive treatment option (22). Koren et al. showed that HBOT was effective for the preservation of hip joints with stage I and II ONFH (21). Vezzani et al. stated that HBO appears to be a viable treatment for Ficat I, II and even III ONFH patients (24). Chandrinou et al. showed that HBOT did not cause a reduction in quality of life and was accepted by patients with ONFH (19). For transient osteoporosis of the hip, Guler et al. reported that HBOT significantly improved the patient’s condition without complications (20). Bernbeck et al. demonstrated that HBO had not shown a beneficial effect in children over 10 years of age with lymphoblastic leukemia and non-Hodgkin’s lymphoma, who were at risk for bone marrow edema and aseptic osteonecrosis, particularly in knee and hip lesions (18). All studies included in this meta-analysis had NOS scores between 6 and 8. The GRADE evaluation results indicated that only one outcome indicator was of high quality of evidence, with the remaining outcome indicators being of very low quality (Supplementary Material S4).

Table 1.

Baseline characteristics of the included literature.

Study	Year	Country	M/F	Age	Hips	Intervention		Disease stages (hips)	HBO treatment scheme				Follow-up	JCR ranking
						EXP	CON		Pressure	STD, min	Treatment times	Weekly frequency
Moghamis et al. (22)	2021	Qatar	9/10	35.2 ± 9.8	EXP: 11; CON: 12	HBO	CD	II	2.4ATA	90 min	40	3–4	34.2 ± 18.4 mo	Q2
Camporesi et al. (15)	2010	America	12/7	48.9	NA	HBO	HBA	II	2.5ATA	82 min	30	5	7 y	Q1
Bozkurt et al. (16)	2021	Turkey	NA	39.6 ± 9.2	EXP: 46; CON: 34	HBO	HBO + CD	IIa, IIb	2.4ATA	120 min	30	6	41.2 mo	Q3
Hsu et al. (17)	2010	China	45/18	39.4 ± 12.1	EXP: 50; CON: 48	Cocktail therapy	ESWT	NA	2.5ATA	90 min	20	5	24 mo	Q2
Reis et al. (23)	2003	Israel	20/7	46.2	NA	HBO	Blank	I	2–2.4ATA	90 min	100	6	4 mo	Q3
Koren et al. (21)	2015	Israel	47/21	43.3 ± 11.7	78	HBO	Blank	I, II	2–2.4ATA	90 min	80*	6	11.1 ± 5.1 y	Q3
Vezzani et al. (24)	2018	Italy	144/73	53.8	NA	HBO	Blank	I, II, III	2.5ATA	82 min	90	5	4 y	Q2
Chandrinou et al. (19)	2020	Greece	49/24	40.43 ± 9.99	91	HBO	Blank	I, II	2ATA	120 min	40	5	-	Q4
Guler et al. (20)	2015	Turkey	12/6	38.6	NA	Medicine + HBO	Blank	NA	2.5ATA	120 min	30	7	17.8 ± 5 mo	Q4
Bernbeck et al. (18)	2004	German	6/13	8 ± 5	NA	HBO	Blank	NA	2.4ATA	90 min	45*	5	NA	Q3

^*Mean value.

HBO, hyperbaric oxygen; M/F, male/female; CD, core decompression; HBA, hyperbaric air; EXP, experimental group; CON, control group; ESWT, extracorporeal shock wave treatment; ATA, atmospheres absolute; NA, not access; STD, single treatment duration; min, minutes; mo, months; y, years.

Results of meta-analysis

VAS pain score

The pre-port meta-analysis included three studies with 75 patients (17, 18, 19). A random-effects model was used to analyze the preoperative and postoperative effects, and the results of econometric analysis showed a substantial relationship between HBOT and the improvement of VAS pain score (MD = −2.94, 95% CI: −4.27, −1.60, P < 0.0001, I² = 76%) (Fig. 2).

Figure 2.

Forest plots of VAS pain score.

Short form 12 physical component summary scale

The pre-post meta-analysis included two studies with 104 hips (17, 21). The quantitative analysis of preoperative and postoperative effects using a random-effects model showed a substantial relationship between HBOT and SF12 PCS improvement (MD = 17.28, 95% CI: 8.26, 26.31, P = 0.0002, I² = 93%) (Fig. 3A). In addition, two studies involving 121 hips reported the effects of HBOT on SF12 PCS (17, 22). Analysis using a random-effects model showed no discernible difference between the HBO and non-HBO groups (MD = −0.39, 95% CI: −7.85, 7.06, P = 0.92, I² = 69%) (Fig. 3B).

Figure 3.

Forest plots of short form 12 physical component summary scale.

Short form 12 mental component summary scale

The pre-post meta-analysis included two studies with 104 hips (17, 21). A fixed-effects model was used to analyze the preoperative and postoperative effects, and the results of econometric analysis showed that HBOT improved SF12 MCS significantly (MD = 4.26, 95% CI: 2.56, 5.95, P < 0.00001, I² = 0%) (Fig. 4A). In addition, two studies involving 121 hips reported the effects of HBOT on SF12 MCS (17, 22). Analysis using a random-effects model showed no discernible difference between the HBO and non-HBO groups (MD = 0.76, 95% CI: −7.02, 8.52, P = 0.85, I² = 62%) (Fig. 4B).

Figure 4.

Forest plots of short form 12 mental component summary scale.

Harris hip score

The pre-post meta-analysis included two studies with 46 patients (17, 20). The random-effects model was used to analyze the preoperative and postoperative effects, and the econometric analysis results showed a substantial relationship between HBOT and HHS improvement (MD = 26.91, 95% CI: 0.35, 53.46, P = 0.05, I² = 99%) (Fig. 5).

Figure 5.

Forest plots of Harris hip score.

Modified Harris hip score

The pre-post meta-analysis included two studies with 98 hips (19, 21). The random-effects model was used to perform a quantitative analysis of the preoperative and postoperative effects, and the results showed a significant relationship between HBOT and mHHS improvement (MD = 44.31, 95% CI: 13.75, 74.87, P = 0.004, I² = 96%) (Fig. 6).

Figure 6.

Forest plots of modified Harris hip score.

Patients’ improved events

Two studies involving 125 hips reported the effect of HBOT on improving outcomes in patients with ONFH (17, 23). The random-effects model analysis showed no significant difference between the HBO group and the non-HBO groups (RR = 1.83, 95% CI: 0.37, 9.09, P = 0.46, I² = 89%) (Fig. 7).

Figure 7.

Forest plots of patients’ improved events.

Discussion

The meta-analysis of ten studies showed that in the HBOT group postoperative values of VAS pain score decreased, whereas values of SF12 PCS, SF12 MCS, HHS and mHHS were significantly improved compared with the preoperative. There was no significant difference in SF12 PCS, SF12 MCS and patients’ improved events in the HBOT group compared with the non-HBOT, suggesting that HBOT could not be considered effective in treating ONFH, which is the most important finding of this paper. Sensitivity analysis of the VAS pain score suggests that Bernbeck et al.’s study may be a major source of heterogeneity. The possible reasons are i) different population characteristics. Bernbeck et al. studied children with an average age of 8 years (18), while Hsu et al. and Chandrinou et al. studied populations with an average age of over 39 years (17, 19). ii) The data of Bernbeck et al. were VAS pain score of the femoral head after five sessions of HBOT, which may not have achieved significant effect at this time (18).

ONFH is classified as traumatic and non-traumatic. The mechanism of traumatic ONFH is the change of blood supply and tamponade effect (25), while the mechanism of non-traumatic ONFH is not well understood (26). Alcohol-induced aseptic necrosis of the femoral head is due in part to an increase in circulating cortisol (27). Vezzani et al. found that HBOT can increase the level of serum osteoprotegerin (OPG), inhibit osteoclast activation and promote bone regeneration (28). Camporesi et al. concluded that when the high concentration of oxygen in HBOT is dissolved in the blood and delivered to bone tissue, oxygen is available to ischemic bone cells (29). Meanwhile, due to its vasoconstriction effect, HBOT can reduce bone marrow edema and improve venous drainage and microcirculation (30). This could partly explain the significant improvement in VAS pain score, SF12 PCS, SF12 MCS, HHS and mHHS numerical evaluation of patients under HBOT compared with those before treatment. In addition, except for SF12 MCS in the pre-post meta-analysis, other outcome indicators showed high heterogeneity. The high heterogeneity could be explained by several reasons: i) this may be due to different frequency, interval, oxygen pressure of HBOT; ii) the handling methods of randomized controlled trials’ and cohort studies’ control groups are different; iii) age is one important coadjuvant factor, the older the age, the more likely to occur ONFH (25); iv) some outcome indicators may be influenced by the personal subjective feelings and measurer’s induction. In conclusion, we are not entirely sure that HBOT is effective for ONFH.

The effectiveness of HBO in the treatment of ONFH has been controversial. Salameh et al. reported a case series of hyperbaric oxygen therapy for Steinberg I and II ONFH, showing that most patients achieved satisfactory results without further surgical treatment (31). Li et al. treated a patient with steroid-associated ONFH with hyperbaric oxygen at 1.6ATA pressure. After 20 treatments, the VAS decreased from 7 to 2, and after 50 treatments, the patient’s symptoms almost completely disappeared (32). For transient osteoporosis of the hip, Yagishita et al.’s case series study did not find that HBOT significantly accelerated the recovery of patients (33). But due to the small number of cases in both studies, the conclusions are difficult to convince. A randomized controlled trial conducted by Camporesi et al. showed significant improvements in extension, adduction, abduction and intensity of pain in 17 patients after 30 sessions of HBOT (15). A combination of alendronate, extracorporeal shock and HBO can delay or stop the development of ONFH in patients with post-severe acute respiratory syndrome (34). However, in a randomized controlled study of 98 patients conducted by Hsu et al. no theoretical synergic effect between HBOT and ESWT was observed in cocktail therapy over ESWT alone (17). For the studies of Camporesi et al. and Hsu et al., the reason for their different conclusions may be the different stages of femoral head necrosis included in the two studies. ONFH is often caused by intense corticosteroid treatment in children and adolescents with acute lymphoblastic leukemia (35). MRI may help to improve staging, imaging investigation of hidden collapse, assessment of prognosis and evaluation of treatment of ONFH (36). Scherer et al. found no significant changes in MRI morphology during late-stage chemotherapy associated with AON when treated with HBO compared to a control group treated only by reducing the affected weight-bearing structure (37). In addition to the uncertainty of efficacy, Klumpp et al. also summarize other disadvantages of hyperbaric oxygen therapy, such as high treatment costs, high treatment duration and frequency, and limited service structure, which may be detrimental to patient compliance (38).

Li et al. and Paderno et al. both conducted systematic reviews and meta-analyses of the clinical efficacy of HBOT in the treatment of ONFH, and came to the same conclusion that HBOT can significantly improve the clinical efficacy of patients with femoral head necrosis (10, 11). However, both systematic reviews contain many non-English literature and do not include hip bone marrow edema and transient osteoporosis. However, according to Iida et al., bone marrow edema is considered a marker of possible progression to advanced osteonecrosis, and osteonecrosis needs to be carefully examined when bone marrow edema is present (39). Transient osteoporosis of the hip is not a unique disease, but an early reversible subtype of non-traumatic ONFH (40). So both deserve to be included. In addition, the outcome indicator of previous systematic reviews and meta-analyses is only improved event, and continuous variables have not been examined as outcome indicators before.

This study still has the following limitations, which need to be further improved: i) the sample size of some studies is relatively small; ii) the follow-up time of included studies may affect the results of data measurement; iii) the outcome indicators of multiple studies are not uniform, which makes it difficult to evaluate the effect; iv) inconsistent comparison between multiple studies; v) by comparing different articles, we found that HBOT seemed to be effective for avascular necrosis before femoral head collapse, but larger randomized controlled studies are needed for future validation.

Conclusion

Based on the current evidence, HBOT cannot be regarded as an effective measure for the treatment of early-stage necrosis of the femoral head. Although many studies suggest the effectiveness of HBOT in its contribution to the treatment of femoral head necrosis, the heterogeneity on study methods and treatments cannot lead us to clear and safe conclusions, and more large-scale randomized controlled trials are needed for further verification.

ICMJE Statement of Interest

The authors declare that there is no conflict of interest that could be perceived as prejudicing the impartiality of the work reported.

Funding Statement

This work was supported by National Key Research and Development Program of China (No. 2023YFB4606705), National Natural Science Foundation of China (No. 82272611, 82472522, 82072506, 92268115), Hunan Provincial Science Fund for Distinguished Young Scholars (No. 2024JJ2089), Science and Technology Innovation Program of Hunan Province (No. 2023SK2024), Natural Science Foundation of Hunan Province (2023JJ30949), Shaanxi Province Technology Committee Project (No. 2021JM576).

Author contribution statement

Y Cao, P Tang and Y Li developed the idea and designed the study and had full access to all data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Y Cao, W Xiao and T Wen ran the search strategy. Y Cao and P Tang selected articles and extracted data. Y Cao evaluated the quality of the literature. Y Cao, P Tang, W Xiao, T Wen, Y Li, J Zhang, S Liu, Y Zhu, H Chai, W Ma, Y Cao, B Lin and H Tan wrote the manuscript. All listed authors reviewed and approved the final manuscript.