A Systematic Review of Oxygen Therapy for the Management of Medication-Related Osteonecrosis of the Jaw ( MRONJ )

Background: Medication-related osteonecrosis of the jaw (MRONJ) can be a life changing iatrogenic complication of antiresorptive and antiangiogenic drug therapy. It is most often associated with high doses of these medications that are used to prevent skeletal-related events in patients with cancer and bone pathologies. Unfortunately, managing MRONJ lesions has proven difficult and remains a major challenge for clinicians. Due to the lack of efficacy in treating MRONJ by surgical modalities (local debridement and free flap reconstruction), the nonsurgical management of MRONJ is still advocated to aid healing or avoid disease progression. The aim of this systematic review is to identify, analyse and understand the published evidence related to the success of oxygen therapies such as ozone (OT) and hyperbaric oxygen (HBO) in treating MRONJ. Material and methods: A multi-database (PubMed, MEDLINE, EMBASE, CINAHL and Cochrane CENTRAL) systematic search was performed by three authors. The identified articles were independently assessed for their risk of bias. Any type of study evaluating humans treated with antiresorptive and antiangiogenic drugs were considered. The aim is primarily to evaluate the success of OT and HBO in resolving MRONJ and secondarily to identify any improvements in quality of life (QoL), rate of complications, time-to-event and severity of side effects related to these treatments. Results: In total, just 13 studies were eligible for analysis. A pooled total of 313 patients (HBO group n = 82; OT group n = 231) described in these studies have shown good tolerance for oxygen therapies. Complete resolution of MRONJ was reported in 44.58% of OT patients but only 5.17% of the HBO group. Progression of MRONJ was reported only in the HBO studies in 10.34% of cases (6 patients). The quality of evidence was low or very low in all studies. This was due to limitations in how the studies were designed, run and reported. Conclusions: Based on the limited data available, it is difficult to suggest OT is better or worse than HBO or whether it is better than a placebo. As the level of evidence available is low, this necessitates larger well-designed trials to justify these interventions for patients affected by MRONJ. Appl. Sci. 2019, 9, 1026; doi:10.3390/app9051026 www.mdpi.com/journal/applsci Appl. Sci. 2019, 9, 1026 2 of 23


Introduction
The term 'medication-related osteonecrosis of the jaws' (MRONJ) refers to a potentially serious iatrogenic complication of treatment with medications, such as antiangiogenic or antiresorptive drugs.These drug families are used primarily for the treatment of malignancies (e.g., multiple myeloma or bone metastases) but also have important roles in the management of osteoporosis, Paget's disease and hypercalcemia.The bone targeting agents (BTAs) can reduce the risk of skeletal-related adverse events to protect the skeleton in patients with primary or secondary bone pathology [1,2].
Since the first reports of bisphosphonate-related osteonecrosis of the jaw in 2003, an increasing number of reports have been published showing similar clinical complications associated with other drugs.New evidence has shown that along with bisphosphonates (BPs), other BTAs such as denosumab also cause osteonecrosis of the jaw bones.In addition, monoclonal antibodies able to bind and selectively inhibit vascular endothelial growth factor-A (VEGF-A), specifically mammalian target of rapamycin (mTOR) inhibitors, can also cause similar lesions [3][4][5].For this reason, the term MRONJ was adopted in the 2014 position paper of the American Association of Oral and Maxillofacial Surgeons (AAOMS) [3].The medications currently reported to be associated with MRONJ are listed in Table 1 (antiresorptive drugs) and Table 2 (antiangiogenic agents) [6,7].
The AAOMS position paper states that "patients may be considered to have MRONJ if all the following characteristics are present: current or previous treatment with antiresorptive or antiangiogenic agents; exposed bone or bone that can be probed through an intraoral or extraoral fistula in the maxillofacial region that has persisted for longer than 8 weeks; and no history of radiation therapy to the jaws or obvious metastatic disease to the jaws".However, it is accepted that some patients may present with non-specific symptoms and may not have evidence of exposed bone such as those classified as Stage 0 by the AAOMS staging system [8].The term 'stage 0' was first used by Mawardi et al. to gather suspected MRONJ cases presenting with clinical and radiological signs of disease other than intraoral bone exposure [9].The AAOMS classification and staging system is based entirely on the intraoral presentation, which they propose should also guide potential treatment.This has been criticised in studies where it was highlighted that radiological findings are an important part of the clinical picture necessary for accurate MRONJ staging (Tables 3 and 4) [6,8,10].
The major risk factor for the development of MRONJ is dento-alveolar surgery, with a history of tooth extraction or oral surgery procedure (apicectomy or cystectomy) reported in 52% to 80% of patients with MRONJ [11][12][13].The overall risk of developing MRONJ after dento-alveolar surgery, in patients on IV bisphosphonates ranges from 1.6% to 14.8% in comparison to 0.5% for patients taking oral bisphosphonates [6].
To date, there is no current standard for the treatment of MRONJ associated with antiresorptive or antiangiogenic drug therapy.Several treatment options have been described, with the earlier stages of MRONJ reportedly responding well to conservative management such as topical or systemic antibiotics, or limited bone debridement, although this still remains controversial [14,15].
Ozone therapy (OT) and hyperbaric oxygen therapy (HBO) have been reported as effective adjunctive therapies in situations where normal bony wound healing is impaired, such as osteoradionecrosis and chronic osteomyelitis of the jaw [16][17][18].HBO increases local concentrations of reactive nitrogen species (RNS) and reactive oxygen species (ROS) by providing substrates (oxygen and L-arginine) for nitric oxide synthase, as well as by the generation of superoxide [19,20].ROS and RNS influence osteoclast (OC) differentiation and activity and participate in the regulation of various aspects of bone metabolism [21][22][23].Nitric oxide (NO) is constitutively synthesised by both OCs and osteoblasts (OBs) and has contrasting biphasic effects.At lower levels, NO decreases bone resorption and stimulates its turnover; whereas at higher concentrations, NO promotes inflammatory processes and inhibits bone formation [24].Indeed, ROS stimulate the expression of receptor activator of nuclear factor kappa-B ligand (RANKL), changing the RANKL/osteoprotegerin ratio and favouring OC differentiation, and avoiding osteopetrosis in animal models [22,25].Authors have also suggested that HBO-generated ROS could induce suppression of OC activity and promote bone healing [21,23].Recent research has shown that HBO-generated ROS and RNS induce stem cell mobilisation, vasculogenesis, mitochondrial biogenesis and preconditioning [26][27][28][29].
Ozone dissolves physically in biological water (physiological saline, plasma, lymph, urine).All these reagents act as donor electrons and are oxidisable, and participate in the ozonation process and the consequent formation of ROSs and lipid oxidation products (LOPs).These molecules are responsible for the biochemical actions of ozone and function as biochemical regulators of inflammation at distinct times and physiological concentrations [30,31].Through interactions with cellular components and depending on the concentration of ozone in the tissue, these molecules can trigger biological effects that are either therapeutic or detrimental to health [30,31].
During normal metabolism, osteoclasts, platelets, lymphocytes, neutrophils, monocytes and fibroblasts can induce the formation of ROS [32,33].When in excess, these ROSs can trigger damage to cellular constituents, extracellular components and affect the metabolism of the cells responsible for extracellular matrix synthesis-fundamental in tissue repair-which leads to apoptosis and cellular necrosis [34,35].
The aim of this review is to analyse all available evidence and evaluate the reported outcome of oxygen therapy as treatment for patients affected by MRONJ.

Material and Methods
This systematic review was performed according to Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines [36].The following databases were interrogated: PubMed, MEDLINE, EMBASE, CINAHL and the Cochrane Central Register of Controlled Trials (CENTRAL).A three-stage screening approach was used to ensure precision and the quality of the search.The screening of titles and abstracts was carried out independently by three authors (RS, RL and SO) to eliminate any irrelevant material (i.e., reviews, animal studies, non-clinical studies and studies that did not report patients undergoing oxygen therapy treatments).Disagreements were resolved by discussion until a consensus was reached.
A data screening and abstraction form was used to: (1) Verify the study eligibility derived from the inclusion/exclusion criteria.
(2) Carry out the methodological quality assessment.
(3) Extract data on study characteristics and outcomes for the included studies.
The authors of any studies eligible for inclusion in the review, unless without sufficient information, were contacted directly (Figure 1).
The aim of this review is to analyse all available evidence and evaluate the reported outcome of oxygen therapy as treatment for patients affected by MRONJ.

This systematic review was performed according to Preferred Reporting Items for Systematic
Reviews and Meta-Analyses (PRISMA) guidelines [36].The following databases were interrogated: PubMed, MEDLINE, EMBASE, CINAHL and the Cochrane Central Register of Controlled Trials (CENTRAL).A three-stage screening approach was used to ensure precision and the quality of the search.The screening of titles and abstracts was carried out independently by three authors (RS, RL and SO) to eliminate any irrelevant material (i.e., reviews, animal studies, non-clinical studies and studies that did not report patients undergoing oxygen therapy treatments).Disagreements were resolved by discussion until a consensus was reached.
A data screening and abstraction form was used to: (1) Verify the study eligibility derived from the inclusion/exclusion criteria.
(2) Carry out the methodological quality assessment.
(3) Extract data on study characteristics and outcomes for the included studies.
The authors of any studies eligible for inclusion in the review, unless without sufficient information, were contacted directly (Figure 1).

Types of Studies
The types of studies included in the research strategy were published or unpublished randomised controlled trials, case-controlled trials, case series, retrospective studies and case reports.Papers were obtained from January 2003 to September 2018.Animal studies, reviews and those studies including patients with a previous history of radiation therapy to the head and neck regions were excluded.No language restrictions were imposed to the search.

Types of Participants
The review considered studies involving patients who developed MRONJ and subsequently underwent OT and/or HBO treatment.No restriction of age, gender or ethnic origin was applied.There was also no restriction on the minimum number of patients included in the studies.

Types of Interventions
Patients affected by MRONJ who underwent OT and/or HBO as either standalone or adjuvant treatment were considered.

Objectives
The objectives was to assess the therapeutic effects of HBO and OT in patients exposed to antiresorptive or antiangiogenic drugs and affected by MRONJ.Moreover, it was to assess the effects of OT and HBO therapy as standalone or adjuvant treatments (either singly or in combination to other treatments) in people with manifest MRONJ.

Types of Outcome Measures
Primary outcomes

•
Healing of MRONJ as indicated by one or more of the following indicators: Improvement in the clinical grade of the lesions according to the AAOMS staging of MRONJ (Table 3).Wound healing (yes or no).Plain film radiological examination (improvement of sclerotic changes, mottling and bone fragmentation, improvement of formed sequestrum or persistent extraction sockets), computed tomography (CT) scan, magnetic resonance imaging (MRI) (surface area of the bone disease, localisation, evidence of bone marrow disease), positron emission tomography (PET)/CT imaging (decreased abnormal focal uptake) [37].
Healing of sinus tract or deep periodontal pockets.

Secondary outcomes
• Quality of life (QoL).

•
Rate of complications and side effects of the intervention.
For the 'complications' outcome measure, interventions involving an interruption or delay of antiresorptive or antiangiogenic treatments, or progression of the underlying disease (e.g., fracture in osteoporosis or disease progression in cancer), were considered to be complications of the intervention.
For QoL measures, we reported whether validated scales were used.Non-validated scales were not excluded a priori.QoL had to have been measured at baseline and at least once during follow-up.

Data Extracted
Data extracted from the studies included number of patients; patient gender and age; predisposing factors for and localisation of MRONJ; type of antiangiogenic or antiresorptive drugs and their cumulative dose; clinical indications for the drug or combined therapy; type of intervention; complications; follow up time; MRONJ evolution and MRONJ recurrence.
All selected papers were carefully read to identify author(s); year of publication; study design; population and treatment characteristics.
In the case of missing information, we contacted the authors and allowed six weeks for a reply.If the information was still missing, we then indicated the missing data as 'Not Reported (NR)' in the text and tables.

Review Quality Assessment Data
Two review authors (RS, AL) appraised the risk of bias in the included study with the tool recommended by the Cochrane Handbook for Systematic Reviews of Interventions as appropriate for randomised control trials (RCTs) [38].Moreover, the authors used the CARE Checklist for case report and the Modified Delphi Checklist for the case series studies [39,40].We referred instances of disagreement in risk of bias assessments to one of the other members of the review team (SO) and resolved them by discussion.

List of Excluded Studies
Currently the treatment of MRONJ is controversial, but many researchers agree that intravenous drug administration or longer period of drug intake contribute to high risk of developing MRONJ [54][55][56].
The most common site for MRONJ was the mandible (18.29%) followed by the maxilla (3.65%).In 2.43% of patients, MRONJ lesions were reported in both (Table 7).However, in 75.60% the MRONJ site was not reported (NR).The drug most commonly responsible of MRONJ was Zoledronate (12.19%), but only 33 patients out of 82 had this detail reported (40.24%).
HBO was most commonly used as a neoadjuvant and/or adjuvant therapy, followed by surgery in 4 studies out of 6.In one study the HBO therapy was given as standalone treatment (Table 9).
The patients were followed for a period of time ranging from 1 to 32 months.At the end of the follow up, MRONJ was seen as completely resolved in 5.17% of the cases (n = 3), while the majority (48.27%) of the patients (n = 28) were reported to have some benefit due to stability or improvement of the disease presentation.In just 10.34% of patients (n = 6), the disease progressed, but for 13 patients (22.41%), data were not available (Table 10).
The most common site for MRONJ was the mandible (43.72%), followed by the maxilla (22.94%).In 8.65% of cases, MRONJ was reported in both, and the site was not reported in 24.67% of cases (Table 12).
The drug most commonly responsible for MRONJ was Zoledronate (14.71%).However, only 37 patients out of 231 had details of the specific MRONJ-associated drug reported (16.01%).OT was most commonly used as a neoadjuvant and/or adjuvant therapy, followed by surgery, in 6 studies out of 7 (Table 14).
The patients were followed for a period of time ranging from 7 to 36 months.At the end of the follow up, MRONJ was completely resolved in 44.58% of patients (n = 103), whilst 22.94% showed some improvement or remission of the disease (n = 53).No progression of the disease has been reported in any studies (Table 15); however, for 30 patients (12.98%), this data was not available.

Risk of Bias and Review Quality Assessment
In all three case report studies, we identified a lack of clarity in many of the thirteen domains of the CARE Checklist, with missing information.We found that the lack of clarity was predominantly on follow-up and diagnostic procedures at the time of follow-up.Hence, we concluded the level of bias to be high for all the included case reports.In the nine case series studies, we reported a consistent lack of clarity in some of the seven domains of the Modified Delphi Checklist.These were predominantly regarding the outcome measurement methods, hence we considered the level of bias to be high for all case series studies.The only eligible RCT was an evaluation of adjuvant HBO therapy for people undergoing surgery.The authors of the study did not mention the generation of randomisation sequences but reported the concealment of allocation using a series of opaque envelopes containing the assignment, and we therefore rated the level of risk as unclear.The personnel involved in the study were not blinded because this was deemed to be impractical.The loss of patients to follow-up was substantial, and although a clear description of losses and withdrawals was given, data analysis was performed as-treated.Moreover, the study had a very high and unbalanced rate of crossovers between study arms.Therefore, we considered the level of risk of bias to be high across the study [41][42][43][44][45][46][47][48][49][50][51][52][53].

Discussion
Antiresorptive drugs are known to improve the quality of life for patients affected by bone metastasis, osteoporosis, osteopenia and Paget disease.Moreover, the new antiangiogenic drugs have been shown to be effective treatment modalities for a number of cancers.Unfortunately, increased use of these drugs has also increased the numbers of patients developing MRONJ.The risk appears to be highest in patients who require intravenous drug administration or an intake period greater than two years [12,54,55].Moreover, the literature suggests that local and systemic factors (such as periodontal disease and diabetes) might act as predisposing factors in developing MRONJ.Although no gold standard is currently available for the treatment of MRONJ, a number of studies debate which MRONJ stage benefits the most from surgical or conservative therapy [15,56].In general, the common opinion appears to be that for early stages of the disease (stage 0 or I), conservative management might be sufficient [6,54].
The purpose of this systematic review was to analyse the current evidence related to the treatment of MRONJ when using oxygen therapy (OT and HBO).Our findings indicate that the oxygen therapy has been used as a neoadjuvant or adjuvant therapy and may represent a viable complementary treatment or an alternative in advanced stage disease (stage II and III) where patients are unfit for aggressive surgery such as jaw resections or microvascular reconstruction.Despite the majority of studies reviewed presenting low quality evidence with a high risk of bias, there is some evidence to show the total resolution of MRONJ in 44.58% of patients with OT and 5.17% of patients with HBO, although this could equally have happened without these interventions.In addition, in a number of OT studies, the spontaneous sequestration of necrotic bone was followed by a spontaneous expulsion [47][48][49][50][51][52][53].The only reported cases of MRONJ progression were amongst patients treated with HBO (10.34%, n = 6).Unfortunately, these outcomes were not reported for 13.73% (n = 43), which could represent further cases of disease progression.
This review also found that in the majority of the MRONJ cases, antiresorptive drugs were explicitly discontinued if deemed safe from the oncological point of view [41][42][43][44][45][46][47][48][49][50][51][52][53].However, it is unclear if the discontinuation strategy leads to a better surgical outcome due to the long skeletal life of some antiresorptive drugs.Amongst all the studies, we have found several perplexities which have had an impact in the quality of the research.Indeed, we have noticed that in 11 articles out of 13, there was no mention of any specific investigations during the patients follow up.Only Lee et al. [42] and Brozoski et al. [53] have reported, during the follow up, that patients had either a CT or an orthopantomograph (OPG) to assess the radiological aspect of the MRONJ lesions.Moreover, other important data were missing

Table 5 .
List of excluded studies.

Table 6 .
Studies included in the hyperbaric oxygen (HBO) analysis, including number of patients treated and evidence level.Case series (CS); case report (CR); randomised control clinical trial (RCT); Levels of Evidence for Prognostic Studies Adapted from the American Society of Plastic Surgeons (https://www.plasticsurgery.org/Documents/medical-professionals/health-policy/evidence-practice/ASPS-Rating-Scale-March-2011.pdf).

Table 7 .
HBO Preoperative epidemiologic analysis (age, sex, predisposing factors and site of the necrosis involved).M: male; F: female; not reported (NR), Standard of Care (StC).

Table 8 .
HBO preoperative pharmacological analysis: type of drugs, indication for drug therapy, and time of drug exposure; Minodronate (MI);

Table 9 .
HBO operative analysis including type of intervention and the stage of the disease.Platelet Rich Plasma (PRP); Standard of Care (StC); Antibiotics (ABX); Mouthwash (MW); American Association of Oral and Maxillofacial Surgery (AAOMS); not applicable (N/A).

Table 10 .
[45]analysis of the MRONJ status at the end of follow-up; not reported (NR), poor patient compliance (PPC), computer tomography (CT).*RCTFreiberger et al. 2012[45]showed no statistically significant improvement in cure rate compared to placebo but there was improvement in secondary outcome.

Table 11 .
Studies included in the OT analysis, including number of patients treated and evidence level.Case series (CS); Case report (CR); Levels of Evidence for Prognostic Studies Adapted from the American Society of Plastic Surgeons (https://www.plasticsurgery.org/Documents/medicalprofessionals/health-policy/evidence-practice/ASPS-Rating-Scale-March-2011.pdf).

Table 14 .
Ozone therapy (OT) operative analysis including type of intervention and the stage of the disease.Spontaneous expulsion of necrotic bone (SENB); antibiotics (ABX); American Association of Oral and Maxillofacial Surgery (AAOMS); not reported (NR).