Interventions for managing medication‐related osteonecrosis of the jaw

Abstract

Background

Medication‐related osteonecrosis of the jaw (MRONJ) is a severe adverse reaction experienced by some individuals to certain medicines commonly used in the treatment of cancer and osteoporosis (e.g. bisphosphonates, denosumab, and antiangiogenic agents), and involves the progressive destruction of bone in the mandible or maxilla. Depending on the drug, its dosage, and the duration of exposure, this adverse drug reaction may occur rarely (e.g. following the oral administration of bisphosphonate or denosumab treatments for osteoporosis, or antiangiogenic agent‐targeted cancer treatment), or commonly (e.g. following intravenous bisphosphonate for cancer treatment). MRONJ is associated with significant morbidity, adversely affects quality of life (QoL), and is challenging to treat. This is an update of our review first published in 2017.

Objectives

To assess the effects of interventions versus no treatment, placebo, or an active control for the prophylaxis of MRONJ in people exposed to antiresorptive or antiangiogenic drugs.

To assess the effects of non‐surgical or surgical interventions (either singly or in combination) versus no treatment, placebo, or an active control for the treatment of people with manifest MRONJ.

Search methods

Cochrane Oral Health’s Information Specialist searched four bibliographic databases up to 16 June 2021 and used additional search methods to identify published, unpublished, and ongoing studies. 

Selection criteria

We included randomised controlled trials (RCTs) comparing one modality of intervention with another for the prevention or treatment of MRONJ. For 'prophylaxis of MRONJ', the primary outcome of interest was the incidence of MRONJ; secondary outcomes were QoL, time‐to‐event, and rate of complications and side effects of the intervention. For 'treatment of established MRONJ', the primary outcome of interest was healing of MRONJ; secondary outcomes were QoL, recurrence, and rate of complications and side effects of the intervention.

Data collection and analysis

Two review authors independently screened the search results, extracted the data, and assessed the risk of bias in the included studies. For dichotomous outcomes, we reported the risk ratio (RR) (or rate ratio) and 95% confidence intervals (CIs).

Main results

We included 13 RCTs (1668 participants) in this updated review, of which eight were new additions. The studies were clinically diverse and examined very different interventions, so meta‐analyses could not be performed.

We have low or very low certainty about available evidence on interventions for the prophylaxis or treatment of MRONJ.

Prophylaxis of MRONJ

Five RCTs examined different interventions to prevent the occurrence of MRONJ.

One RCT compared standard care with regular dental examinations at three‐month intervals and preventive treatments (including antibiotics before dental extractions and the use of techniques for wound closure that avoid exposure and contamination of bone) in men with metastatic prostate cancer treated with zoledronic acid. The intervention seemed to lower the risk of MRONJ (RR 0.10, 95% CI 0.02 to 0.39, 253 participants). Secondary outcomes were not evaluated.

Dentoalveolar surgery is considered a common predisposing event for developing MRONJ and five RCTs tested various preventive measures to reduce the risk of postoperative MRONJ. The studies evaluated plasma rich in growth factors inserted into the postextraction alveolus in addition to standardised medical and surgical care versus standardised medical and surgical care alone (RR 0.08, 95% CI 0.00 to 1.51, 176 participants); delicate surgery and closure by primary intention versus non‐traumatic tooth avulsion and closure by secondary intention (no case of postoperative MRONJ in either group); primary closure of the extraction socket with a mucoperiosteal flap versus application of platelet‐rich fibrin without primary wound closure (no case of postoperative MRONJ in either group); and subperiosteal wound closure versus epiperiosteal wound closure (RR 0.09, 95% CI 0.00 to 1.56, 132 participants). 

Treatment of MRONJ

Eight RCTs examined different interventions for the treatment of established MRONJ; that is, the effect on MRONJ cure rates. 

One RCT analysed hyperbaric oxygen (HBO) treatment used in addition to standard care (antiseptic rinses, antibiotics, and surgery) compared with standard care alone (at last follow‐up: RR 1.56, 95% CI 0.77 to 3.18, 46 participants). 

Healing rates from MRONJ were not significantly different between autofluorescence‐guided bone surgery and conventional bone surgery (RR 1.08, 95% CI 0.85 to 1.37, 30 participants). Another RCT that compared autofluorescence‐ with tetracycline fluorescence‐guided sequestrectomy for the surgical treatment of MRONJ found no significant difference (at one‐year follow‐up: RR 1.05, 95% CI 0.86 to 1.30, 34 participants). 

Three RCTs investigated the effect of growth factors and autologous platelet concentrates on healing rates of MRONJ: platelet‐rich fibrin after bone surgery versus surgery alone (RR 1.05, 95% CI 0.90 to 1.22, 47 participants), bone morphogenetic protein‐2 together with platelet‐rich fibrin versus platelet‐rich fibrin alone (RR 1.10, 95% CI 0.94 to 1.29, 55 participants), and concentrated growth factor and primary wound closure versus primary wound closure only (RR 1.38, 95% CI 0.81 to 2.34, 28 participants).  

Two RCTs focused on pharmacological treatment with teriparatide: teriparatide 20 μg daily versus placebo in addition to standard care (RR 0.96, 95% CI 0.31 to 2.95, 33 participants) and teriparatide 56.5 μg weekly versus teriparatide 20 μg daily in addition to standard care (RR 1.60, 95% CI 0.25 to 1.44, 12 participants). 

Authors' conclusions

Prophylaxis of medication‐related osteonecrosis of the jaw

One open‐label RCT provided some evidence that dental examinations at three‐month intervals and preventive treatments may be more effective than standard care for reducing the incidence of medication‐related osteonecrosis of the jaw (MRONJ) in individuals taking intravenous bisphosphonates for advanced cancer. We assessed the certainty of the evidence to be very low.

There is insufficient evidence to either claim or refute a benefit of the interventions tested for prophylaxis of MRONJ in patients with antiresorptive therapy undergoing dentoalveolar surgery. Although some interventions suggested a potential large effect, the studies were underpowered to show statistical significance, and replication of the results in larger studies is pending.

Treatment of medication‐related osteonecrosis of the jaw

The available evidence is insufficient to either claim or refute a benefit, in addition to standard care, of any of the interventions studied for the treatment of MRONJ. 

Plain language summary

Interventions for managing medication‐related osteonecrosis (severe bone damage) of the jaw

Review question

What are the effects of different interventions to either prevent or treat medication‐related osteonecrosis of the jaw compared with each other or compared with no treatment or an inactive intervention ('placebo')?

Background

Medication‐related osteonecrosis of the jaw (MRONJ) is severe bone damage in the jaw bone that occurs in some people as a reaction to certain medicines commonly used in the treatment of cancer and osteoporosis (a disease that makes bones fragile). It is a painful condition that can be difficult to treat. MRONJ occurs rarely in people taking some medicines for osteoporosis. However, in people receiving these drugs at higher doses for cancer‐related conditions, the risk of MRONJ may be higher and has been reported to occur in up to five in 100 individuals. It is important to identify effective preventive measures to reduce the risk of MRONJ, and better treatments for those who have it.

This is an update of our review first published in 2017. It is based on a search for articles that was conducted most recently in June 2021.

Studycharacteristics

Working with Cochrane Oral Health, we searched for studies that had been published up to June 2021. We found five studies that focused on the prevention of MRONJ and eight studies that tested treatments for MRONJ. The studies involved 1668 adults, with the smallest study having 13 participants and the largest study having 700 participants. Most study participants were women, but one study was of men with prostate cancer receiving bisphosphonate infusions (given by drip into a vein). All but two studies included only participants treated with bisphosphonates (used to support treatment and reduce risk of fracture and bone pain), although several other drugs are also known to induce MRONJ. Two trials also included patients treated with bisphosphonates or denosumab.

Keyresults

One study provided very low‐certainty evidence that dental examinations at three‐month intervals and preventive treatments (antibiotics before dental extractions and the use of techniques for wound closure that avoid exposure and contamination of bone) are more effective than standard care for reducing the number of cases with MRONJ in a group of people receiving intravenous bisphosphonates for cancer‐related conditions. In the experimental group, which received preventive care consisting of antibiotics and specific wound closure, fewer people developed MRONJ: two participants per 100 who underwent close monitoring developed MRONJ compared to 23 participants per 100 in the control group (standard care).

There was insufficient evidence to conclude that the use of the other interventions investigated would reduce the risk of MRONJ or would improve healing of MRONJ.

Certainty of evidence

The certainty of evidence was low or very low. This was due to limitations in how the studies were designed and run. For example, some participants changed groups during the study, some participants did not finish the study, and the outcomes were measured at different follow‐up times. In addition, most of the studies had only a small number of participants.

Summary of findings

Summary of findings 1. Dental examinations at three‐month intervals and preventive treatments (experimental) compared to standard care (control) for prophylaxis of MRONJ.

Dental examinations at three‐month intervals and preventive treatments (experimental) compared to standard care (control) for prophylaxis of MRONJ
Population: people at risk of MRONJ Setting: hospital Intervention: dental examinations at three‐month intervals and preventive treatments (experimental) Comparison: standard care (control)
Outcomes	Anticipated absolute effects* (95% CI)		Relative effect (95% CI)	Number of participants (studies)	Certainty of the evidence (GRADE)	Comments
	Risk with standard care (control)	Risk with dental examinations at 3‐month intervals and preventive treatments (experimental)
MRONJ (incidence proportion) Diagnostic criteria for MRONJ: non‐healing exposed bone in mandible or maxilla for longer than 8 weeks without any change of the stage of disease   Follow‐up: mean 32 months	233 per 1000	23 per 1000 (5 to 91)	RR 0.10 (0.02 to 0.39)	253 (1 RCT)	⊕⊝⊝⊝ VERY LOW1	Participants: high‐risk (i.e. individuals with cancer exposed to intravenous zoledronic acid). The outcome MRONJ was also reported as number of cases per patient‐year (incidence rate): rate ratio 0.18 (95% CI 0.04 to 0.74).
*The risk in the intervention group (and its 95% CI) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: confidence interval; MRONJ: medication‐related osteonecrosis of the jaw;RCT: randomised controlled trial; RR: risk ratio
GRADE Working Group grades of evidence High certainty: we are very confident that the true effect lies close to that of the estimate of the effect. Moderate certainty: we are moderately confident in the effect estimate; the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low certainty: our confidence in the effect estimate is limited; the true effect may be substantially different from the estimate of the effect. Very low certainty: we have very little confidence in the effect estimate; the true effect is likely to be substantially different from the estimate of effect.

1. We downgraded the certainty of the evidence by three levels due to very serious risk of bias (high and unbalanced rate of crossovers after randomisation; high dropout rates due to high mortality; failure to adhere to the intention‐to‐treat principle; mean follow‐up differed between experimental and control groups) and very serious limitation of indirectness (all male and high‐risk patients).

Summary of findings 2. Dental extraction protocol with plasma rich in growth factors (PRGF) (experimental) compared to standard dental extraction protocol without PRGF (control) for prophylaxis of MRONJ in people treated with IV bisphosphonates who need dental extractions.

Dental extraction protocol with plasma rich in growth factors (PRGF) (experimental) compared to a standard dental extraction protocol without PRGF (control) for prophylaxis of MRONJ in people treated with IV bisphosphonates who need dental extractions
Population: people treated with IV bisphosphonates who need dental extractions Setting: hospital Intervention: dental extraction protocol with PRGF (experimental) Comparison: standard dental extraction protocol without PRGF (control)
Outcomes	Anticipated absolute effects* (95% CI)		Relative effect (95% CI)	Number of participants (studies)	Certainty of the evidence (GRADE)	Comments
	Risk with standard dental extraction protocol without PRGF (control)	Risk with dental extraction protocol with PRGF (experimental)
MRONJ (incidence proportion) Diagnostic criteria of MRONJ: pain, swelling, and non‐healing exposed necrotic bone or fistulae, or both, with connection to the bone   Follow‐up: 24‐60 months	59 per 1000	5 per 1000 (0 to 89)	RR 0.08 (0.00 to 1.51)	176 (1 RCT)	⊕⊝⊝⊝ VERY LOW1	Participants: high risk, i.e. individuals with cancer exposed to IV zoledronic acid
*The risk in the intervention group (and its 95% CI) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: confidence interval; IV: intravenous: MRONJ: medication‐related osteonecrosis of the jaw;RCT: randomised controlled trial; RR: risk ratio
GRADE Working Group grades of evidence High certainty: we are very confident that the true effect lies close to that of the estimate of the effect. Moderate certainty: we are moderately confident in the effect estimate; the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low certainty: our confidence in the effect estimate is limited; the true effect may be substantially different from the estimate of the effect. Very low certainty: we have very little confidence in the effect estimate; the true effect is likely to be substantially different from the estimate of effect.

1. We downgraded the certainty of evidence by three levels due to imprecision and very serious risk of bias (high or unclear risk of selection bias, performance bias, detection bias, and attrition bias).

Summary of findings 3. Subperiosteal wound closure versus epiperiosteal wound closure after tooth extraction for prevention of MRONJ in patients on antiresorptive treatment.

Subperiosteal wound closure versus epiperiosteal wound closure after tooth extraction for prevention of MRONJ in patients on antiresorptive treatment
Population: people on antiresorptive treatment Setting: hospital Intervention: subperiosteal wound closure Comparison: epiperiosteal wound closure after tooth extraction
Outcomes	Anticipated absolute effects* (95% CI)		Relative effect (95% CI)	Number of participants (studies)	Certainty of the evidence (GRADE)	Comments
	Risk with epiperiosteal wound closure after tooth extraction	Risk with subperiosteal wound closure after tooth extraction
MRONJ after tooth extraction Assessed with: absence of complete mucosal integrity   Follow‐up: 6 months	77 per 1000	7 per 1000 (0 to 120)	RR 0.09 (0.00 to 1.56)	132 (1 RCT)	⊕⊕⊝⊝ LOW1	8 patients changed intervention from epiperiosteal wound closure to subperiosteal wound closure.
*The risk in the intervention group (and its 95% CI) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: confidence interval; MRONJ: medication‐related osteonecrosis of the jaw;RCT: randomised controlled trial; RR: risk ratio
GRADE Working Group grades of evidence High certainty: we are very confident that the true effect lies close to that of the estimate of the effect. Moderate certainty: we are moderately confident in the effect estimate; the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low certainty: our confidence in the effect estimate is limited; the true effect may be substantially different from the estimate of the effect. Very low certainty: we have very little confidence in the effect estimate; the true effect is likely to be substantially different from the estimate of effect.

1. We downgraded the certainty of the evidence by two levels due to imprecision and serious risk of bias (unclear selection bias, detection bias, high risk of performance bias, attrition bias and reporting bias).

Summary of findings 4. Hyperbaric oxygen therapy (HBO) as an adjunct to conventional therapy (experimental) compared to conventional therapy (control) for treatment of MRONJ.

HBO as an adjunct to conventional therapy (experimental) compared to conventional therapy (control) for treatment of MRONJ
Population: people with MRONJ Setting: hospital Intervention: HBO as an adjunct to conventional therapy (experimental) Comparison: conventional therapy (control)
Outcomes	Anticipated absolute effects* (95% CI)		Relative effect (95% CI)	Number of participants (studies)	Certainty of the evidence (GRADE)	Comments
	Risk with conventional therapy (control)	Risk with HBO therapy as an adjunct to conventional therapy (experimental)
Healing of MRONJ Diagnostic criteria for healing of MRONJ: gingival coverage with no exposed bone   Follow‐up: up to 24 months (outcome was measured at last follow‐up)	333 per 1000	520 per 1000 (257 to 1000)	RR 1.56 (0.77 to 3.18)	46  (1 RCT)	⊕⊝⊝⊝ VERY LOW1
*The risk in the intervention group (and its 95% CI) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: confidence interval; HBO: hyperbaric oxygen therapy; MRONJ: medication‐related osteonecrosis of the jaw;RCT: randomised controlled trial; RR: risk ratio
GRADE Working Group grades of evidence High certainty: we are very confident that the true effect lies close to that of the estimate of the effect. Moderate certainty: we are moderately confident in the effect estimate; the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low certainty: our confidence in the effect estimate is limited; the true effect may be substantially different from the estimate of the effect. Very low certainty: we have very little confidence in the effect estimate; the true effect is likely to be substantially different from the estimate of effect.

1. We downgraded the certainty of the evidence by three levels due to imprecision and very serious risk of bias (unclear and high risk of selection bias, performance bias, detection bias, and attrition bias; failure to adhere to the intention‐to‐treat principle).

Summary of findings 5. Autofluorescence‐guided bone surgery (experimental) compared to tetracycline fluorescence‐guided bone surgery (control) for treatment of MRONJ.

Autofluorescence‐guided bone surgery (experimental) compared to tetracycline fluorescence‐guided bone surgery (control) for treatment of MRONJ
Population: people with MRONJ Setting: hospital Intervention: autofluorescence‐guided bone surgery (experimental) Comparison: tetracycline fluorescence‐guided bone surgery (control)
Outcomes	Anticipated absolute effects* (95% CI)		Relative effect (95% CI)	Number of participants (studies)	Certainty of the evidence (GRADE)	Comments
	Risk with tetracycline fluorescence‐guided bone surgery (control)	Risk with autofluorescence‐guided bone surgery (experimental)
Healing of MRONJ Criteria for healing of MRONJ: mucosal integrity   Follow‐up: 1 year	889 per 1000	933 per 1000 (764 to 1000)	RR 1.05 (0.86 to 1.30)	34  (1 RCT)	⊕⊝⊝⊝ VERY LOW1
*The risk in the intervention group (and its 95% CI) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: confidence interval; MRONJ: medication‐related osteonecrosis of the jaw;RCT: randomised controlled trial; RR: risk ratio
GRADE Working Group grades of evidence High certainty: we are very confident that the true effect lies close to that of the estimate of the effect. Moderate certainty: we are moderately confident in the effect estimate; the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low certainty: our confidence in the effect estimate is limited; the true effect may be substantially different from the estimate of the effect. Very low certainty: we have very little confidence in the effect estimate; the true effect is likely to be substantially different from the estimate of effect.

1. We downgraded the certainty of the evidence by three levels due to imprecision and very serious risk of bias (unclear and high risk of selection bias, performance bias, and detection bias).

Summary of findings 6. Bone morphogenetic protein‐2 together with platelet‐rich fibrin (experimental) compared to platelet‐rich fibrin alone (control) for treatment of MRONJ.

Bone morphogenetic protein‐2 together with platelet‐rich fibrin (experimental) compared to platelet‐rich fibrin alone (control) for treatment of MRONJ
Population: people with MRONJ Settings: hospital Intervention: bone morphogenetic protein‐2 together with platelet‐rich fibrin (experimental) Comparison: platelet‐rich fibrin only (control)
Outcomes	Anticipated absolute risks* (95% CI)		Relative effect (95% CI)	Number of participants (studies)	Certainty of the evidence (GRADE)	Comments
	Risk with platelet‐rich fibrin only (control)	Risk with bone morphogenetic protein‐2 adjacent to platelet‐rich fibrin (experimental)
Healing of MRONJ  Defined in the study as full coverage with absence of exposed bone, mucosal swelling and erythema, purulent drainage, intra‐ and extra oral fistula and/or any pain or discomfort   Follow‐up: 16 weeks post surgery	880 per 1000	968 per 1000 (831 to 1000)	RR 1.10 (0.94 to 1.29)	55  (1 RCT)	⊕⊝⊝⊝ VERY LOW1
*The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: confidence interval; MRONJ: medication‐related osteonecrosis of the jaw;RCT: randomised controlled trial; RR: risk ratio
GRADE Working Group grades of evidence High certainty: we are very confident that the true effect lies close to that of the estimate of the effect. Moderate certainty: we are moderately confident in the effect estimate; the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low certainty: our confidence in the effect estimate is limited; the true effect may be substantially different from the estimate of the effect. Very low certainty: we have very little confidence in the effect estimate; the true effect is likely to be substantially different from the estimate of effect.

1. We downgraded the certainty of the evidence by three levels due to very serious limitations of imprecision and very serious risk of bias (unclear risk of selection bias, detection bias, high risk of performance bias, attrition bias, reporting bias).

Summary of findings 7. Autofluorescence‐guided surgery (experimental) compared to conventional surgery (control) for treatment of MRONJ.

Autofluorescence‐guided surgery (experimental) compared to conventional surgery (control) for treatment of MRONJ
Population: people with MRONJ Settings: hospital Intervention: autofluorescence guided surgery (experimental) Comparison: conventional surgery (control)
Outcomes	Anticipated absolute risks* (95% CI)		Relative effect (95% CI)	Number of Participants (studies)	Certainty of the evidence (GRADE)	Comments
	Risk with conventional surgery (control)	Risk with autofluorescence guided surgery (experimental)
Healing of MRONJ  Criteria for healing: absence of bone exposure   Follow‐up: 1 year (at last follow‐up)	867 per 1000	933 per 1000 (734 to 1000)	RR 1.08 (0.85 to 1.37)	30 (1 RCT)	⊕⊝⊝⊝ VERY LOW1	High drop‐out rate. 6 patients were excluded due to mortality and no show at follow‐up appointments.
*The risk in the intervention group (and its 95% CI) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: confidence interval; MRONJ: medication‐related osteonecrosis of the jaw;RCT: randomised controlled trial; RR: risk ratio
GRADE Working Group grades of evidence High certainty: we are very confident that the true effect lies close to that of the estimate of the effect. Moderate certainty: we are moderately confident in the effect estimate; the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low certainty: our confidence in the effect estimate is limited; the true effect may be substantially different from the estimate of the effect. Very low certainty: we have very little confidence in the effect estimate; the true effect is likely to be substantially different from the estimate of effect.

1. We downgraded the certainty  of the evidence by three levels due to very serious imprecision and serious risk of bias (high selection bias, performance bias, attrition bias, unclear risk of detection bias).

Summary of findings 8. Platelet‐rich fibrin after bone surgery (experimental) compared to surgery alone (control) for treatment of MRONJ.

Platelet‐rich fibrin after bone surgery (experimental) compared to surgery alone (control) for treatment of MRONJ
Population: people with MRONJ Settings: hospital Intervention: platelet‐rich fibrin (experimental) Comparison: conventional (control)
Outcomes	Anticipated absolute risks* (95% CI)		Relative effect (95% CI)	Number of participants (studies)	Certainty of the evidence (GRADE)	Comments
	Risk with surgery alone (control)	Risk with platelet‐rich fibrin after bone surgery (experimental)
Healing of MRONJ Defined as absence of infection and mucosal integrity without fistula   Follow‐up: 1 year	913 per 1000	958 per 1000 (824 to 1000)	RR 1.05 (0.90 to 1.22)	47 (1 RCT)	⊕⊝⊝⊝ VERY LOW1	The outcome healing of MRONJ was also reported as absence of infection, mucosal integrity without fistula, no need for re‐intervention: rate ratio 1.60 (95% CI 1.04 to 2.46). Follow‐up: 1 year See Analysis 8.2.
*The risk in the intervention group (and its 95% CI) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: confidence interval; MRONJ: medication‐related osteonecrosis of the jaw;RCT: randomised controlled trial; RR: risk ratio
GRADE Working Group grades of evidence High certainty: we are very confident that the true effect lies close to that of the estimate of the effect. Moderate certainty: we are moderately confident in the effect estimate; the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low certainty: our confidence in the effect estimate is limited; the true effect may be substantially different from the estimate of the effect. Very low certainty: we have very little confidence in the effect estimate; the true effect is likely to be substantially different from the estimate of effect.

1. We downgraded the certainty of the evidence by three levels due to very serious limitation of imprecision and serious risk of bias (unclear selection bias, detection bias, high risk of performance bias, reporting bias).

8.2. Analysis.

Comparison 8: Platelet‐rich fibrin after bone surgery (experimental) compared to surgery alone (control) for treatment of MRONJ, Outcome 2: Healing of MRONJ (defined as absence of infection, mucosal integrity without fistula,  and no need for re‐intervention) at 1 year

Summary of findings 9. Concentrated growth factor and primary wound closure (experimental) versus primary wound closure only (control) for treatment of MRONJ.

Concentrated growth factor and primary wound closure compared with primary wound closure only for treatment of MRONJ
Population: people with MRONJ Setting: surgical treatment Intervention: concentrated growth factor and primary wound closure Comparison: primary wound closure only
Outcomes	Anticipated absolute effects* (95% CI)		Relative effect (95% CI)	Number of participants (studies)	Certainty of the evidence (GRADE)	Comments
	Risk with primary wound closure only (control)	Risk with concentrated growth factor and primary wound closure (experimental)
Healing of MRONJ Defined as soft tissue healing Assessed with: mucosal integrity (without flap dehiscence or infection) Follow‐up: 6 months	521 per 1000	286 per 1000 (89 to 922)	RR 1.38 (0.81 to 2.34)	28 (1 RCT)	⊕⊝⊝⊝ VERY LOW1
*The risk in the intervention group (and its 95% CI) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: confidence interval; MRONJ: medication‐related osteonecrosis of the jaw;RCT: randomised controlled trial; RR: risk ratio
GRADE Working Group grades of evidence High certainty: we are very confident that the true effect lies close to that of the estimate of the effect. Moderate certainty: we are moderately confident in the effect estimate; the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low certainty: our confidence in the effect estimate is limited; the true effect may be substantially different from the estimate of the effect. Very low certainty: we have very little confidence in the effect estimate; the true effect is likely to be substantially different from the estimate of effect.

1. We downgraded the certainty of the evidence by three levels because of serious risk of bias (unclear selection bias, high risk of performance bias), very serious limitation of indirectness (only female participants with osteoporosis) and very serious limitation of imprecision (few participants).

Summary of findings 10. Teriparatide 20 μg daily (experimental) versus placebo (control) in addition to standard care for treatment of MRONJ.

Teriparatide 20 μg daily compared with placebo for treatment of MRONJ
Population: people  with MRONJ Settings: outpatient and inpatient treatment Intervention: teriparatide 20 μg daily Comparison: placebo
Outcomes	Illustrative comparative risks* (95% CI)		Relative effect (95% CI)	Number of Participants (studies)	Quality of the evidence (GRADE)	Comments
	Risk with placebo (control)	Risk with teriparatide 20 µg daily (experimental)
Healing of MRONJ Primary outcomes were the clinical and radiologic resolution of MRONJ lesions, as evaluated by oral examination and CBCT imaging; secondary outcomes included improvement in MRONJ stage, change in MRONJ lesion size, quality of life, bone mineral density, and evidence of osteoblastic response measured biochemically using P1NP and radiologically using 18F‐fluoride PET‐CT imaging   Last follow‐up: 52 weeks	278 per 1000	267 per 1000 (87 to 819)	RR 0.96  (0.31 to 2.95)	33 (1 RCT)	⊕⊕⊝⊝ LOW1
*The risk in the intervention group (and its 95% CI) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: confidence interval; MRONJ: medication‐related osteonecrosis of the jaw;RCT: randomised controlled trial; RR: risk ratio
GRADE Working Group grades of evidence High certainty: we are very confident that the true effect lies close to that of the estimate of the effect. Moderate certainty: we are moderately confident in the effect estimate; the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low certainty: our confidence in the effect estimate is limited; the true effect may be substantially different from the estimate of the effect. Very low certainty: we have very little confidence in the effect estimate; the true effect is likely to be substantially different from the estimate of effect.

1. We downgraded the certainty of the evidence by two levels due to imprecision and serious risk of bias (unclear selection bias, high risk of attrition bias and reporting bias).

Summary of findings 11. Teriparatide 56.5 μg weekly (experimental) versus teriparatide 20 μg daily (control) in addition to standard care for treatment of MRONJ.

Teriparatide 56.5 μg weekly (experimental) compared with teriparatide 20 μg daily (control) in addition to standard care for treatment of MRONJ
Population: people with MRONJ Settings: outpatient and inpatient treatment Intervention: teriparatide 56.5 μg weekly in addition to standard care Comparison: teriparatide 20 μg daily in addition to standard care
Outcomes	Illustrative comparative risks* (95% CI)		Relative effect (95% CI)	Number of participants (studies)	Certainty of the evidence (GRADE)	Comments
	Risk with teriparatide 20 μg daily (control)	Risk with teriparatide 56.5 µg weekly (experimental)
Healing of MRONJ Measured changes in MRONJ clinical stage at 6 months after the start of the treatment as clinical course, changes in bone metabolism (using bone scintigraphy), percentage of bone formation on the osteolysis of MRONJ, and measurement of bone turnover markers)   Follow‐up: 6 months after start of treatment	500 per 1000	300 per 1000 (125 to 721)	RR 0.60 (0.25 to 1.44)	12 (1 RCT)	⊕⊝⊝⊝ VERY LOW1
*The risk in the intervention group (and its 95% CI) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: confidence interval; MRONJ: medication‐related osteonecrosis of the jaw;RCT: randomised controlled trial; RR: risk ratio
GRADE Working Group grades of evidence High certainty: we are very confident that the true effect lies close to that of the estimate of the effect. Moderate certainty: we are moderately confident in the effect estimate; the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low certainty: our confidence in the effect estimate is limited; the true effect may be substantially different from the estimate of the effect. Very low certainty: we have very little confidence in the effect estimate; the true effect is likely to be substantially different from the estimate of effect.

1. We downgraded the certainty of the evidence by three levels due to imprecision and very serious risk of bias (unclear selection bias and detection bias, high risk of performance bias, attrition bias and reporting bias).

Background

Description of the condition

Medication‐related osteonecrosis of the jaw (MRONJ) is a severe adverse reaction experienced by some individuals to certain medicines commonly used in the treatment of cancer and osteoporosis (e.g. bisphosphonates, denosumab and antiangiogenic agents), and involves the progressive destruction of bone in the mandible or maxilla.

Osteonecrosis of the jaw (ONJ) associated with bisphosphonate treatment was first reported in 2003 (Marx 2003; Migliorati 2003; Ruggiero 2007; Sigua‐Rodriguez 2014). Subsequently, ONJ was observed in individuals who took denosumab, an antiresorptive medication unrelated to the bisphosphonate class (Bone 2017). A growing number of case reports currently suggest that ONJ is also associated with antiangiogenic agents such as bevacizumab, aflibercept, sunitinib, temsirolimus, and everolimus (Ruggiero 2014; Zhang 2016). The condition formerly referred to as 'bisphosphonate‐related ONJ' has been renamed 'medication‐related ONJ' due to the growing number of ONJ cases associated with non‐bisphosphonate treatments (Ruggiero 2014).

The exact mechanisms underlying MRONJ remain unknown. Interestingly, MRONJ is primarily limited to the maxillofacial region. In contrast to other skeletal bones, jaw bones (the alveolar process and periodontium) have relatively high vascularity, bone turnover, and remodelling because of continuous mechanical stress, which may make them vulnerable to the adverse effects of drugs. Proposed hypotheses that attempt to explain the localisation of MRONJ exclusively to the jaws include altered bone remodelling, angiogenesis inhibition, constant microtrauma, suppression of innate or acquired immunity, and possible effects of inflammation or infection (Ruggiero 2014).

According to the case definition provided by the American Society for Bone and Mineral Research and the American Association of Oral and Maxillofacial Surgeons, people may be considered to have MRONJ if all of the following characteristics are present: (i) current or previous treatment with antiresorptive or antiangiogenic agents, (ii) exposed or necrotic bone in the maxillofacial region that did not heal (by primary or secondary intent) within eight weeks after identification by a healthcare provider, (iii) no history of radiation therapy to the jaws, and (iv) no evidence of metastatic disease to the jaws (Ruggiero 2007; Sigua‐Rodriguez 2014). MRONJ has been divided into four stages based on clinical symptoms. Stage 0 describes individuals with prodromal disease (unexposed variant). Bone exposure is common in individuals with stage 1 to 3 MRONJ without infection (stage 1), with infection (stage 2), or with infection as well as a pathological fracture or fistula, or evidence of osteolysis extending to the inferior border of the mandible or sinus floor (stage 3) (Table 12) (Ruggiero 2007; Ruggiero 2014; Sigua‐Rodriguez 2014; Vescovi 2012a).

1. Clinical staging of MRONJ.

MRONJ stage	Description
AT RISK	No apparent necrotic bone in patients who have been treated with oral or intravenous bisphosphonates
STAGE 0	No clinical evidence of necrotic bone but nonspecific clinical findings, radiographic changes, and symptoms
STAGE 1	Exposed and necrotic bone or fistulas that probes to bone in patients who are asymptomatic and have no evidence of infection
STAGE 2	Exposed and necrotic bone or fistulas that probes to bone associated with infection as evidenced by pain and erythema in the region of exposed bone with or without purulent drainage
STAGE 3	Exposed and necrotic bone or a fistula that probes to bone in patients with pain, infection, and ≥ 1 of the following: exposed and necrotic bone extending beyond the region of alveolar bone (i.e. inferior border and ramus in mandible, maxillary sinus, and zygoma in maxilla) resulting in pathologic fracture, extraoral fistula, oral antral, or oral nasal communication, or osteolysis extending to inferior border of the mandible or sinus floor

From the American Association of Oral and Maxillofacial Surgeons position paper on medication‐related osteonecrosis of the jaw‐‐2014 update (Ruggiero 2014)

The frequency of MRONJ is highly variable and ranges from very rare (less than 1/10,000) to common (1/100 or more), depending on the drug, treatment indication (cancer versus osteoporosis), dose, and duration of treatment (Dodson 2015). For example, in randomised controlled trials (RCTs) and a meta‐analysis, the incidence of MRONJ in individuals with cancer exposed to intravenous (IV) zoledronic acid was between 0.3% and 5% (Coleman 2011; Lopez‐Olivo 2012; Mauri 2009; Morgan 2010). The reported risk of MRONJ in individuals with cancer treated with denosumab ranged from 0.7% to 1.9% (Boquete‐Castro 2016; Qui 2014; Ruggiero 2014). A meta‐analysis that compared the safety of denosumab and zoledronic acid in individuals with bone metastases did not reveal a significant difference in the risk of MRONJ between the denosumab and zoledronic acid groups (Chen 2016).

Among individuals with osteoporosis, who receive substantially lower doses of bisphosphonates or denosumab than those with cancer, MRONJ is rare and the incidence may not be substantially greater than the natural background incidence of the condition. In people receiving bisphosphonates to treat osteoporosis, incidence estimates range from less than 0.1 to 0.7 cases per 10,000 patient years of exposure (Chamizo Carmona 2013; Grbic 2010). In a recent report studying people exposed to denosumab for treatment of osteoporosis, the incidence of MRONJ was 5.2 per 10,000 patient‐years (Bone 2017). The risk for MRONJ among people with osteoporosis treated with bisphosphonates or denosumab approximates the risk for MRONJ that is observed in placebo groups (Bone 2017; Grbic 2010). The risk of MRONJ among people exposed to antiresorptive medications for the treatment of osteoporosis is approximately 100‐fold smaller than the risk in people with cancer (Ruggiero 2014).

Evidence supporting the association of antiangiogenic medications with the development of MRONJ is primarily based on case reports. The frequency of MRONJ in people receiving antiangiogenic agents is not known accurately and reliably. Analysis of the United States Food and Drug Administration’s Adverse Event Reporting System database showed that the intravenous BPs were associated with the highest risk for MRONJ, denosumab was associated with risk comparable to bisphosphonates used for osteoporosis, and the antiangiogenic agents were associated with the lowest risk for MRONJ (Zhang 2016). In a combined analysis of three phase III trials the incidence of MRONJ in people exposed to the angiogenesis inhibitor bevacizumab was 0.2% (Guarneri 2010). The incidence was substantially higher in those exposed to both zoledronic acid and bevacizumab (Guarneri 2010).

The treatment of MRONJ is challenging, and an effective and appropriate therapy that substantially improves the outcome remains to be identified. The median time to resolution of osteonecrosis symptoms may be up to 12 months and depends on the specific therapeutic intervention (Hinson 2015). Additional information on the natural history of MRONJ comes from a report of individuals with multiple myeloma who were prospectively observed for a minimum of 3.2 years following diagnosis (Badros 2008). MRONJ resolved in 62% of cases, resolved and then recurred in 12%, and did not heal in 26%.

Antiresorptive medications associated with MRONJ

Bisphosphonates are osteotropic agents with antiresorptive activity that are used in a wide spectrum of indications such as the treatment and prevention of osteoporosis, as well as the treatment of Paget's disease, multiple myeloma, and malignancy‐associated hypercalcaemia. Bisphosphonates bind to bone hydroxyapatite and specifically inhibit the activity of osteoclasts, the bone‐resorbing cells. Bone turnover is thereby reduced, which results in an increase in the mineral density of the bone and a reduction in serum calcium (Chestnut 2001; Guyatt 2002; Ruggiero 2007; Sigua‐Rodriguez 2014). Bisphosphonates have a long retention time in bone, and effects may persist for some time after treatment has been stopped. There are two major risk categories for bisphosphonate‐related ONJ: (i) low risk in individuals without cancer treated with oral bisphosphonates (e.g. alendronic acid, clodronic acid, etidronic acid, ibandronic acid, and risedronic acid), or intravenous bisphosphonates (e.g. ibandronic acid and zoledronic acid) for osteoporosis, Paget’s disease, osteopenia, and osteogenesis imperfecta; and (ii) high risk in individuals with cancer treated with intravenous bisphosphonates (e.g. zoledronic acid, pamidronic acid, and ibandronic acid) for multiple myeloma and bone metastases (Bagan 2009; Ruggiero 2014; Vescovi 2012a). Additional parameters affecting the development of bisphosphonate‐related ONJ include the duration of bisphosphonate exposure, age, comedication, comorbidity, smoking, and oral health/oral hygiene (Bamias 2005; Dimopoulos 2006; Katsarelis 2015; Ruggiero 2014; Sigua‐Rodriguez 2014).

Denosumab, a potent antiresorptive agent, is used to treat osteoporosis in postmenopausal women and in men who have an increased risk of fracture. The recommended dose is 60 mg administered as a single subcutaneous injection once every six months. Denosumab is also used to prevent bone complications in adults with bone metastases from solid tumours and to treat a type of bone cancer called giant cell tumour of bone. The recommended maintenance dose for the latter indications is much higher, 120 mg every four weeks. Denosumab is a monoclonal antibody, which has been designed to attach to an antigen called RANK ligand (RANKL). By attaching to and blocking RANKL, denosumab reduces the formation and activity of osteoclasts, the cells in the body that are involved in breaking down bone tissue (Katsarelis 2015; Pageau 2009; Ruggiero 2014; Xu 2013). The exact pathophysiological mechanisms of denosumab‐related ONJ are currently unknown.

Antiangiogenic medications associated with MRONJ

Antiangiogenic agents are increasingly used as anticancer drugs for the treatment of renal cell carcinomas, gastrointestinal tumours, and other solid tumours. The drugs interfere with the formation of new blood vessels by inhibiting angiogenesis signalling cascades, such as vascular endothelial growth factor signalling (bevacizumab and aflibercept), mechanistic target of rapamycin signalling (temsirolimus and everolimus), or receptor tyrosine kinase signalling (sunitinib). MRONJ is a known, rare side effect of these agents, possibly resulting from their interaction with wound healing or osteoclast differentiation and survival (Patel 2015; Ruggiero 2014). Drug approval authorities (US Food and Drug Administration, European Medicines Agency) have included drug safety warnings in the drug labels of bevacizumab, aflibercept, and sunitinib regarding the risk of MRONJ.

Description of the intervention

Interventions for the prevention of MRONJ in at‐risk individuals or the management of MRONJ in individuals with manifest disease may include the following.

Prophylaxis of MRONJ

A range of dental prophylactic measures may be used alone or in combination. A primary means of prevention is the completion of all dental treatment (such as restorative therapy, root canal treatment, periodontitis therapy, or tooth extraction) before the commencement of antiresorptive or antiangiogenic therapy or as soon as possible following the commencement of antiresorptive or antiangiogenic therapy to ensure that treatment is completed within the specified ‘time frame’ for the intended agent. Antibiotic prophylaxis or antiseptic mouthwash (e.g. chlorhexidine) may be used. Individuals may take part in a preventive recall programme, or be provided with information regarding antiresorptive or antiangiogenic therapy risks, professional teeth cleaning, effective oral hygiene, and the importance of limiting or ceasing oral health risk behaviours (such as smoking and drug and alcohol use), or both. Surgical interventions may use a non‐traumatic surgical technique (i.e. surgical treatment designed to minimise tissue damage). The use of plasma rich in growth factors (PRGF) may promote bone and adjacent soft tissue regeneration in post‐extraction defects, thereby reducing the risk of MRONJ. To minimise wound exposure to bacteria, reconstructive surgical techniques for wound closure can be used. Some specific dental extraction methods recommend the discontinuation of antiresorptive or antiangiogenic agents before dentoalveolar surgery.

Treatment of MRONJ

For individuals with established MRONJ, the objective is to control infection, minimise necrosis progression, and promote tissue healing (Bagan 2009; Rollason 2016; Ruggiero 2014; Sigua‐Rodriguez 2014; Vescovi 2006; Vescovi 2012a). The standard medical care of MRONJ is currently anti‐infective treatment with systemic antibiotics or oral antiseptic rinses (e.g. chlorhexidine), or both, and surgical debridement or resection (Ruggiero 2014).

Non‐surgical treatment options

Healing may be stimulated by oral pentoxifylline and α‐tocopherol (vitamin E) in addition to antimicrobial therapy. Other options are adjunct hyperbaric oxygen (HBO) therapy, which involves breathing pure oxygen in a pressurised room or tube, or topical ozone therapy (OT) to improve healing. Low‐level laser therapy (LLLT) is also considered a promising adjunctive treatment method for MRONJ. The lasers most commonly used for biomodulation in bone are argon, carbon dioxide, helium/neon, and neodymium‐doped yttrium‐aluminium‐garnet. The use of (autologous) platelet‐rich plasma (PRP) has been suggested to enhance postsurgical wound healing. PRP is commonly used in a gel formulation, which is formed by mixing PRP (derived from the centrifugation of autologous whole blood) with thrombin and calcium chloride. PRP gel contains higher amounts of fibrinogen, platelets, and growth factors than whole blood. Moreover, bone may be restored by teriparatide, a recombinant form of parathyroid hormone. Teriparatide is approved for the treatment of osteoporosis but is used off‐label for other indications such as fracture healing, dental stability, and ONJ. Recombinant human bone morphogenetic proteins (rhBMPs), which also have the ability to induce osteogenesis, are another treatment option to enhance bone healing in MRONJ. After sequestrectomy, a carrier/scaffold (absorbable collagen sponge) that contains rhBMP is placed into the defect.

Surgical treatment options

Surgical treatments include sequestrectomy, debridement, resection, immediate reconstruction. Surgical treatment may also include extraction of teeth within exposed necrotic bone.

How the intervention might work

Prophylaxis of MRONJ

Controlling risk factors for MRONJ may represent an effective prophylaxis for MRONJ. MRONJ is a complication that can develop spontaneously after dentoalveolar surgery in combination with antiresorptive agents. Therefore, the completion of necessary elective dentoalveolar surgery before the start of this therapy may help reduce the risk of MRONJ (Ruggiero 2007; Ruggiero 2014). Another known risk factor is infection (Katsarelis 2015; Ruggiero 2014). Dental prophylaxis, caries control, and conservative restorative dentistry are expected to minimise the number of bacteria and eliminate the ports of entry for bacteria, thereby reducing the risk of infection. Regular dental evaluations during antiresorptive or antiangiogenic therapy may help to recognise significant risks at an early stage and enable prompt measures to be taken to counter them (Ruggiero 2014; SDCEP 2017). If surgery is necessary, for example, during bisphosphonate therapy, wound exposure to bacteria may be controlled by antibiotic prophylaxis, antiseptic mouthwash, or both. Choosing of surgical procedures that help minimise bone exposure or trauma to the jaws may reduce the risk of MRONJ. Platelet‐derived growth‐factor preparations, such as PRP and PRGF, applied at the surgical site may accelerate wound healing and reduce the time of increased infection risk. Stopping antiresorptive drugs before an invasive dental procedure (drug holiday) could be useful for the prevention of MRONJ. Due to the pharmacokinetics, the antiresorptive effect of bisphosphonates and denosumab is maintained for several weeks or months. This would require cessation of antiresorptive therapy for at least two months to significantly reduce the risk of MRONJ during invasive dental procedures (Damm 2013; Ruggiero 2014).

Treatment of MRONJ

Treatment objectives for people with a defined diagnosis of MRONJ are to control infection of the soft and hard tissues, and minimise the progression or occurrence of bone necrosis to optimise wound healing. Stage‐dependent strategies to treat MRONJ have been proposed (Ruggiero 2014), which can be classified into non‐surgical and surgical treatment.

Non‐surgical treatment options

Non‐surgical management includes, for example, drug treatment with teriparatide, which is a recombinant form of parathyroid hormone that stimulates osteoblasts to increase bone density when used intermittently. Alternative options are treatment with pentoxifylline and α‐tocopherol in combination with anti‐microbial therapy, OT, HBO, and LLLT (Vescovi 2012a). Pentoxifylline and α‐tocopherol have been used to treat osteoradionecrosis for many years. Pentoxifylline, a methylxanthine derivative and phosphodiesterase inhibitor, improves blood flow by increasing erythrocyte flexibility and vasodilatation, and modulates immunological activity; α‐tocopherol has antioxidant properties (Epstein 2010); pentoxifylline and α‐tocopherol may play a role in encouraging wound healing and reducing scarring; ozone has antimicrobial and wound‐healing properties, and OT as an adjunct treatment has been hypothesised to induce the repair of tissues by cleansing osteonecrotic lesions, which leads to mucosal healing (Petrucci 2007; Ripamonti 2011). HBO has been shown to be effective in addition to conventional therapies to treat osteoradionecrosis (Bennett 2016). HBO has been proven to stimulate new blood vessel growth within the damaged tissues and to improve the availability of oxygen for wound healing. Thus, HBO has been hypothesised to be a useful adjunctive treatment for MRONJ (Freiberger 2009). Phototherapy with a low‐intensity laser is used as an adjunctive therapy for the treatment of various diseases including wounds. The laser light used with LLLT is in the red visible and near infrared wavelength ranges and promotes biological effects, such as inflammation and angiogenesis. It also increases the inorganic matrix, which may support wound healing (Martins 2012; Vescovi 2006). Platelet‐derived growth‐factor preparations, such as PRP and PRGF, are applied at the surgical site as an adjuvant to stimulate regeneration of osseous and epithelial tissues, thereby accelerating wound healing. Platelet‐derived growth factors are proposed to support angiogenesis and to improve bone formation by enhancing osteoblast formation and activity (Lee 2007; Lopez‐Jornet 2016). rhBMP is used in surgical procedures to improve bone formation and remodelling during bone healing by enhancing the effects of osteoblast formation and activity (Gerard 2014).

Surgical treatment options

Surgical treatments may include a more conservative approach, such as sequestrectomy and surgical debridement or aggressive therapies, such as resections of affected bone with reconstruction. One of the advantages of a more conservative surgical approach such as sequestrectomy is that a better healing can be expected because the periosteum and unaffected bone are conserved (Eckardt 2011; Stanton 2009; Comas‐Calonge 2017).

Why it is important to do this review

Cochrane Oral Health undertook an extensive prioritisation exercise in 2014 to identify a core portfolio of titles that were the most clinically important to maintain on the Cochrane Library (Worthington 2015). This review was identified as a new priority title by the oral and maxillofacial surgery expert panel (Cochrane Oral Health priority review portfolio).

Among the drugs associated with MRONJ, bisphosphonates are by far the most widely used for a wide range of clinical indications. For example, bisphosphonates can be used in breast cancer and prostate cancer, which have the highest sex‐related incidence rates worldwide. Osteoporosis, another common indication for bisphosphonates, is estimated to affect 200 million women worldwide: approximately one‐tenth of women aged 60 years, one‐fifth of women aged 70 years, two‐fifths of women aged 80 years, and two‐thirds of women aged 90 years (Kanis 2007). Moreover, several other drugs (denosumab, antiangiogenic medications) have recently been associated with MRONJ. MRONJ may occur as a common side effect, particularly in individuals with cancer, depending on the drug and the dosage used. Therefore, the population at risk for MRONJ is large and expanding, and the public health implications may be substantial.

MRONJ significantly affects the quality of life (QoL), and the decline in QoL correlates with MRONJ stage (Kyrgidis 2012; Miksad 2011). The following factors contribute to impairment of QoL: (i) infected and painful necrotic jaw bone; (ii) ulcerated, painful, and swollen oral mucosa; (iii) chronic sinus tracts and facial disfigurement; (iv) impaired speech, swallowing, and eating; and (v) frequent medical and dental evaluations and treatments (Migliorati 2010). Rehabilitation after a complete cure of MRONJ is often protracted. A further aggravating circumstance is a high risk of recurrence, which is higher than in other diseases of the jaw bone. Thus, it is important to develop strategies to prevent or manage MRONJ. Preventative dentistry may be shown to decrease the incidence of MRONJ, in which case the implementation of preventive strategies will become an important consideration for individuals, clinicians, and policy makers (Dimopoulos 2006; Ripamonti 2009). Epidemiological studies have shown that the risk of MRONJ increases with a longer duration of treatment and with higher drug doses. Effective measures to prevent and treat MRONJ may significantly improve the risk‐benefit balance, in particular for people requiring long‐term or high‐dose therapy.

However, there is uncertainty regarding how to prevent MRONJ before and during bisphosphonate therapy and how to manage manifest MRONJ (Lopez‐Jornet 2010). As a consequence, current recommendations are contradictory in certain respects (Ruggiero 2014; SDCEP 2017). This review complements and extends the previous Cochrane Review by Rollason 2016, which focused on interventions for treating ONJ associated with bisphosphonate drugs.

This review was first published in 2017 evaluating the prophylaxis and treatment of MRONJ (Beth‐Tasdogan 2017). We completed this update in 2022.

Objectives

To assess the effects of interventions versus no treatment, placebo, or an active control for the prophylaxis of medication‐related osteonecrosis of the jaw (MRONJ) in people exposed to antiresorptive or antiangiogenic drugs.

To assess the effects of non‐surgical or surgical interventions (either singly or in combination) versus no treatment, placebo, or an active control for the treatment of people with manifest MRONJ.

Methods

Criteria for considering studies for this review

Types of studies

We included randomised controlled trials (RCTs) comparing one modality of intervention with another for the prevention or treatment of MRONJ. We excluded quasi‐randomised and non‐RCTs, as well as case studies, case series (or those of case series design), and cross‐sectional studies. We did not exclude studies on the basis of language, publication status or date of publication.

Types of participants

To assess preventive strategies, we included participants who were treated with known risk medications and who had not yet developed MRONJ before assignment to the experimental or control group.

To assess interventions to treat MRONJ, we included people who had developed clinically apparent MRONJ. Case definition included exposure to risk drug and the presence of necrotic bone or fistulas that probe to bone.

We applied no restrictions regarding participant sex, age, initial health status, and pre‐existing conditions, or type of ONJ‐related drug (e.g. alendronic acid, clodronic acid, etidronic acid, ibandronic acid, incadronic acid, olpadronic acid, pamidronic acid, risedronic acid, tiludronic acid, zoledronic acid, denosumab, bevacizumab, aflibercept, sunitinib, temsirolimus, or everolimus), dose, or duration of therapy. To comply with the MRONJ case definition (Ruggiero 2014), we did not include participants with a history of head and neck radiation therapy.

Types of interventions

For prophylaxis of MRONJ

Any intervention (before or after commencement of antiresorptive or antiangiogenic drug therapy) that aims at prevention of MRONJ. Examples of interventions discussed in the literature include the following.

• Completion of all necessary dental treatment before the commencement of antiresorptive or antiangiogenic agents or as soon as possible following commencement of antiresorptive or antiangiogenic agents

• Antibiotic prophylaxis or antiseptic mouthwash

• Preventive recall programme and provision of information for patients

• Non‐traumatic surgery (i.e. surgical treatment designed to minimise tissue damage), reconstructive techniques for wound closure to minimise wound exposure to bacteria, and specific dental extraction protocols

• Supportive measures to accelerate wound healing after surgery, such as platelet‐rich plasma (PRP) and plasma rich in growth factors (PRGF)

• Cessation of therapy with antiresorptive or antiangiogenic agents (‘drug holiday’) before invasive dental procedures

For treatment of MRONJ

Any intervention (non‐surgical, surgical, or a combination of both) that aims to treat clinically manifest MRONJ. Examples of interventions discussed in the literature include the following.

• Non‐surgical

  • Antiseptic mouthwashes

  • Antibiotic and antifungal therapy

  • Parathyroid hormone and teriparatide

  • Pentoxifylline and α‐tocopherol

  • Ozone therapy (OT)

  • Hyperbaric oxygen therapy (HBO)

  • Laser therapy (low‐level laser therapy (LLLT))

  • Platelet‐derived growth‐factor preparations, such as PRP and PRGF

  • Recombinant human bone morphogenetic proteins (rhBMPs)

• Surgical

  • Surgical debridement, sequestrectomy

  • Jaw bone resection

  • Extraction of teeth within exposed necrotic bone

Comparisons: any single or combined experimental intervention versus control. The control arm consisted of participants receiving no treatment, placebo, or an active control (e.g. standard care).

Types of outcome measures

Primary outcomes

Prophylaxis of MRONJ

Incidence of MRONJ

Two related measures are often used to describe the incidence of MRONJ: incidence proportion (cumulative incidence) and incidence rate of MRONJ. As the incidence rate of MRONJ peaks after two to four years of exposure to bisphosphonates or denosumab in individuals with cancer (Henry 2011; Nakamura 2015; Saad 2012), we had originally planned to include only trials with a follow‐up period of at least three years for the primary outcome. However, we found that the three‐year follow‐up threshold was not applicable as a strict selection criterion for the following reasons: a large proportion of individuals with metastatic cancer (i.e. those most likely to be affected by MRONJ) may die before reaching a three‐year follow‐up. Moreover, follow‐up periods were reported inconsistently between studies (mean follow‐up versus range, follow‐up period of the total study population versus that for each study arm separately, follow‐up per protocol versus follow‐up period as observed).

Treatment of MRONJ

Healing of MRONJ

There is no standardised scale for the assessment of MRONJ healing. Healing of MRONJ may be defined based on clinical examination, imaging findings, or both. Wound healing may be defined as absolute area healed per day, percentage of initial area healed per day, and advance of the wound margin towards the wound centre per day. Wound healing may also be defined as the time taken for mucosa to completely cover necrotic tissue and exposed bone (‘cure period’). The number of participants with resolution of MRONJ (defined as mucosal healing with coverage of exposed bone) within a given time period (e.g. six months) may also be used to describe the healing of MRONJ. Follow‐up time should be at least six months for this primary outcome.

Secondary outcomes

Prophylaxis of MRONJ

• Quality of life (QoL)

• Time‐to‐event

• Rate of complications and side effects of the intervention

Treatment of MRONJ

• QoL

• Recurrence

• Rate of complications and side effects 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.

For the outcome 'complications', if the intervention involved interruption/delay of antiresorptive or antiangiogenic treatment or progression of the underlying disease (e.g. fracture in osteoporosis or disease progression in cancer) were considered complications of the intervention.

Search methods for identification of studies

Electronic searches

Cochrane Oral Health’s Information Specialist conducted systematic searches in the following databases for randomised controlled trials and controlled clinical trials. Due to the Cochrane Centralised Search project to identify all clinical trials on the database and add them to CENTRAL, we only searched recent years of the Embase database. Please see the searching page on the Cochrane Oral Health website for more information. We did not place any other restrictions on the language or date of publication when searching the electronic databases:

• Cochrane Oral Health’s Trials Register (searched 16 June 2021) (see Appendix 1);

• The Cochrane Central Register of Controlled Trials (CENTRAL; 2021, Issue 5) in the Cochrane Library (searched 16 June 2021) (see Appendix 2);

• MEDLINE Ovid (1946 to 16 June 2021) (see Appendix 3);

• Embase Ovid (23 May 2016 to 16 June 2021) (see Appendix 4).

Subject strategies were modelled on the search strategy designed for MEDLINE Ovid. Where appropriate, they were combined with subject strategy adaptations of the highly sensitive search strategies designed by Cochrane for identifying randomised controlled trials and controlled clinical trials (as described in the Cochrane Handbook for Systematic Reviews of Interventions, Version 6.3, (Lefebvre 2022)).

Searching other resources

The following trial registries were searched for ongoing studies:

• US National Institutes of Health Ongoing Trials Register (ClinicalTrials.gov; searched 16 June 2021) (see Appendix 5);

• World Health Organization International Clinical Trials Registry Platform (apps.who.int/trialsearch;  16 June 2021) (see Appendix 6).

We asked experts in the field to help identify unpublished literature and searched the reference lists of potential clinical trials in an attempt to identify any study not found by the other searches.

We searched the reference lists of included studies and relevant systematic reviews for further studies.

We checked that none of the included studies in this review were retracted due to error or fraud.

We did not perform a separate search for adverse effects of interventions used, we considered adverse effects described in included studies only.

Data collection and analysis

Selection of studies

Two review authors (NBT (for the original and updated review) and JP) independently assessed the titles and abstracts of each paper identified by the review search strategy. We excluded only clearly irrelevant records at this stage. The search was designed to be sensitive and include controlled clinical trials; these were filtered out early in the selection process if they were not randomised. Following this, we obtained the full text of potentially relevant studies and assessed these for eligibility based on the inclusion criteria as outlined above. In the event that the two review authors could not reach a consensus, another review author (OZ (for the original and updated review)) acted as arbiter. We maintained a detailed log of study eligibility and reasons for exclusion, and recorded these in 'Characteristics of excluded studies' tables.

Data extraction and management

Two review authors (NBT (for the original and updated review) and JP) independently collected details from the included trials using a structured form. If necessary, a third review author (OZ (for the original and updated review)) was consulted to resolve inconsistencies. We extracted the following details and entered them into 'Characteristics of included studies' tables in Review Manager 5 (RevMan 5) (RevMan 2014).

• Methods

  • Trial design

  • Duration of study

  • Sample size calculation

  • Country of origin

  • Year of publication

  • Language of the original publication

  • Category (i.e. prophylaxis or treatment of MRONJ)

  • Funding

  • Registration in a public trials registry

• Participants

  • Number of participants

  • Age

  • Sex

  • Condition treated with antiresorptive or antiangiogenic agents

  • Inclusion criteria

  • Exclusion criteria

• Interventions (i.e. the type of intervention and procedural information)

• Outcomes

  • Primary outcomes

  • Secondary outcomes

We planned to contact study authors to ask for further information or clarification of their data if necessary.

Assessment of risk of bias in included studies

Two review authors (NB and JP) independently assessed the risk of bias in the included studies according to guidelines in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011).

We assessed the included trials for risk of bias (high, low, or unclear) in the following key domains:

• random sequence generation (allocation bias);

• allocation concealment (allocation bias);

• blinding of participants and personnel (performance bias);

• blinding of outcome assessors (detection bias);

• incomplete outcome data (attrition bias);

• selective reporting (reporting bias).

'Unclear’ indicates either a lack of information or uncertainty over the potential for bias. We completed a risk of bias table for each study and presented the results graphically by study and by domain for all studies. If the risk of bias was not clear because of a lack of detail in the studies, we planned to contact the study authors to request further information.

We categorised overall risk of bias by outcome as shown in the table below.

Risk of bias	Interpretation	Within a study	Across studies
Low risk of bias	Plausible bias unlikely to seriously alter the results	Low risk of bias for all key domains	Most information is from studies at low risk of bias
Unclear risk of bias	Plausible bias that raises some doubt about the results	Unclear risk of bias for one or more key domains	Most information is from studies at low or unclear risk of bias
High risk of bias	Plausible bias that seriously weakens confidence in the results	High risk of bias for one or more key domains	The proportion of information from studies at high risk of bias is sufficient to affect the interpretation of results

Measures of treatment effect

We used RevMan 5 (RevMan 2014) to perform the analyses.

For continuous data, we planned to calculate the mean differences (MDS)and 95% confidence intervals (CIs). We planned to report continuous outcomes as means and standard deviations (SDs). When studies used different instruments to measure the same construct, we planned to use the standardised difference i(SMD) n means in the analysis to combine the data.

For dichotomous outcomes, we calculated risk ratios (RRs) along with 95% CIs from cumulative incidence data. In cases of reported incidence rates, the rate ratio was the effect measure of choice. In cases when no events are observed in one or both groups, RevMan automatically checks for problematic zero counts, and add a fixed value to all cells of a table where the problems occur. Studies with no events in either arm are excluded from the meta‐analysis of risk ratios (Higgins 2011).

To summarise time‐to‐event data, we planned to use methods of survival analysis and we planned to express the intervention effect as a hazard ratio (HR), along with 95% CI.

Where insufficient information was reported to enable effect measures to be calculated, we provided a narrative report of the summary measures.

Unit of analysis issues

The individual participant was the unit of analysis.

If there was a choice of time points for a primary outcome, we selected the time point closest to three years for prophylaxis and one year for treatment. We avoided multiple testing of the effect at each of the time points.

Dealing with missing data

We attempted, where feasible, to contact authors from the primary studies to obtain missing data. We used the methods outlined in the Cochrane Handbook for Systematic Reviews of Interventions to estimate the missing standard error of the log rate ratio (Higgins 2011).

Assessment of heterogeneity

To identify and measure the statistical heterogeneity of the data, we planned to use the I² statistic (Higgins 2003). This value (percentage) defines the variability in effect estimates between studies that is beyond what would be expected by chance. The I² value can be categorised as not important heterogeneity (0% to 40%), moderate heterogeneity (30% to 60%), substantial heterogeneity (50% to 90%), and very substantial heterogeneity (75% to 100%) (Higgins 2003). We also planned to use graphical displays, such as Galbraith plots, if appropriate. Galbraith plots enable the display of several estimates of the same quantity having different standard errors; this is why they provide a useful way of checking for the presence of heterogeneity (Anzures‐Cabrera 2010; Copas 2009). Clinical diversity (i.e. variability in the participants, interventions, and outcomes studied) may contribute to statistical heterogeneity. If a sufficient number of studies was included, we planned to explore heterogeneity by conducting subgroup analyses. If there was substantial evidence for between‐study heterogeneity, we planned to use a random‐effects meta‐analysis.

Assessment of reporting biases

If there had been sufficient studies, we would have assessed publication bias using methods based on a funnel plot, such as Egger's test (Egger 1997). However, all publication bias methods were characterised by a relatively low power and could not be assumed to prove or exclude publication bias (Higgins 2011).

Data synthesis

We followed the guidelines in the Cochrane Handbook for Systematic Reviews of Interventions for the statistical analysis of results (Higgins 2011). If the studies had been sufficiently similar with respect to the participants included, interventions compared, and outcomes and time points reported, we would have conducted meta‐analyses. We would have used a random‐effects or fixed‐effect meta‐analysis as appropriate to combine quantitative data. For comparisons in which a meta‐analysis could not be carried out, we have provided a narrative report of the summary measures and treatment effects.

Subgroup analysis and investigation of heterogeneity

Clinical heterogeneity (i.e. differences associated with the participants, interventions, or outcomes across the included studies) may contribute to statistical heterogeneity (i.e. differences in the effects of interventions). If a sufficient number of studies were included, we planned to explore heterogeneity by conducting subgroup analyses in any case (i.e. whether statistical heterogeneity was present or not). To assess the effect of particular aspects of the studies on the primary and secondary outcome variables, we had planned to conduct the following subgroup analyses: medication dose or dose intensity (i.e. unit dose of medication administered per unit time); medication type (e.g. nitrogenous or non‐nitrogenous bisphosphonate) or compound; stage and type of disease (e.g. cancer or non‐cancer); and risk factors (e.g. multimorbidity, age, smoker). If we had included at least 10 studies, we would have investigated these effects using a meta‐regression analysis.

In the case of significant statistical heterogeneity, we would have attempted to identify the source of the heterogeneity with subgroup analyses.

Sensitivity analysis

If there had been sufficient RCTs for meta‐analyses, we would have performed a sensitivity analysis to check the robustness of results when omitting studies with high or unclear risk of bias or to investigate whether the meta‐analysis result was heavily determined by outlier studies. We would have used the Galbraith plot to detect potential outliers.

Summary of findings and assessment of the certainty of the evidence

We have developed a summary of findings table for each comparison, and have presented summary information for the primary outcomes.

Following GRADE methods and using GRADEPro software (GRADEPro 2014), two review authors (NB and OZ) assessed the certainty of evidence with reference to the overall risk of bias of the included studies, directness of the evidence, consistency of the results, precision of the estimates, and risk of publication bias. Factors that may lead to downgrading of evidence in the GRADE approach are: (a) risk of bias, (b) inconsistency between studies, (c) indirectness, (d) imprecision, and (e) likely publication bias. We assessed the certainty of the body of evidence for each comparison and outcome as high, moderate, low, or very low.

Results

Description of studies

See Characteristics of included studies; Characteristics of excluded studies; Characteristics of ongoing studies.

Results of the search

We brought forward five included studies (eight references), 11 excluded studies and four ongoing studies from the previous version of the review. For details of the study selection process for the previous version of the review, see Table 13.

2. Previous searches.

The initial electronic search of 2016 retrieved 1105 references after de‐duplication. After screening the titles and abstracts, we excluded all but 23 references from further evaluation. We examined the full text of the remaining 23 articles and found that eight references relating to five studies met the prespecified inclusion criteria and were therefore included in this review. We identified four additional studies that are ongoing and listed these under Characteristics of ongoing studies. We excluded 11 full‐text articles for reasons noted in the Characteristics of excluded studies table.

Our searches for the update identified a total of 1047 references. After de‐duplication, two review authors NBT (for the original and updated review) and JP screened the remaining 728 references as titles and abstracts, most of which did not fulfil the inclusion criteria. Twenty‐nine articles were selected for further evaluation in full text. We excluded nine ineligible articles (see Characteristics of excluded studies), and there were nine ongoing studies (see Characteristics of ongoing studies). We identified eight new RCTs (11 references) for inclusion (see Included studies). 

The flow diagram summarises the study selection process for the whole review (Figure 1). 

1.

Included studies

We included 13 studies in the review: five studies from the previous review (Freiberger 2012; Mozzati 2012; Mozzati 2013; Mücke 2016; Ristow 2016) and eight additional RCTs identified for this update (Giudice 2018a; Giudice 2018b; Ohbayashi 2020; Park 2017; Poxleitner 2020; Ristow 2020; Sim 2020; Sim 2020). Five studies focused on the prophylaxis of MRONJ (Mozzati 2012; Mozzati 2013; Mücke 2016; Poxleitner 2020; Ristow 2020), and eight trials investigated options for the treatment of MRONJ (Freiberger 2012; Giudice 2018a; Giudice 2018b; Ohbayashi 2020; Park 2017; Ristow 2016; Sim 2020; Yüce 2021). The trials varied in sample size between 13 (Ohbayashi 2020) and 700 participants (Mozzati 2013). In total, 1668 participants were included in this review. More women than men took part in the studies, with the exception of one study (Mücke 2016), which recruited only men with prostate cancer. Full details of the studies are in the Characteristics of included studies tables. 

Prophylaxis of MRONJ

Characteristics of the trial designs and settings 

Of the five studies concerning prophylaxis of MRONJ, three were conducted in Germany (Mücke 2016; Poxleitner 2020; Ristow 2020), and two in Italy (Mozzati 2012; Mozzati 2013). All of the five studies used a parallel‐group design and had two treatment arms.

Characteristics of the participants

The five studies focusing on prophylaxis of MRONJ involved a total of 1366 participants, ranging from 77 to 700 participants per study. All participants were scheduled for tooth extraction while receiving intravenous bisphosphonate for the treatment of cancer (Mücke 2016; Mozzati 2012), or oral bisphosphonate or subcutaneous denosumab for the treatment of osteoporosis (Mozzati 2013; Poxleitner 2020; Ristow 2020), cancer (Ristow 2020), or rheumatoid arthritis and Paget's disease (Mozzati 2013).

Characteristics of the interventions

The following interventions were used in the included studies.

• Dental examinations at three‐month intervals and preventive treatments versus standard care (Mücke 2016).

• Dental extraction protocol with plasma rich in growth factors (PRGF), which was inserted into the postextraction alveolus versus standard dental extraction protocol without PRGF. All participants had a professional oral hygiene session one week before surgery and antibiotics for six days starting the evening before surgery. Surgical care included anaesthesia by alveolar nerve block, no intraligamentous or intrapapillary infiltrations, mucosal flap, and suturing to enable healing via primary intention (Mozzati 2012).

• Surgical extractions carried out via an intrasulcular incision and mobilisation of a mucoperiosteal flap versus extractions carried out without detachment of full‐thickness flaps, and sockets were filled with absorbable haemostatic gelatine sponges (Mozzati 2013).

• Primary closure of the extraction socket with a mucoperiosteal flap versus application of platelet‐rich fibrin (PRF) without subsequent primary closure. Bisphosphonate or denosumab drug holiday prior to the tooth extraction was not required. All patients received perioperative intravenous antibiotic therapy, initiated one day before surgery and continued until one day after surgery. Prior to the extraction, all patients rinsed their mouths with chlorhexidine solution. Postoperatively, patients were instructed to consume a soft diet, to apply daily mouth rinses with chlorhexidine solution, and to refrain from wearing dentures until complete mucosal healing was achieved (Poxleitner 2020).

• Subperiosteal wound closure versus epiperiosteal wound closure after tooth extraction. All patients were pre‐treated with oral antibiotics from the week before surgery until one week after surgery and used an antimicrobial mouth rinse with chlorhexidine three times daily, starting two days before surgery and for at least five days after surgery. If a drug holiday was possible from an oncologic/osteologic perspective, antiresorptive therapy was discontinued one month before surgery until one month after surgery (Ristow 2020).

Characteristics of the outcomes 

Primary outcome 

The primary outcome was the development of MRONJ. The major diagnostic criterion of MRONJ was non‐healing exposed bone in the mandible or maxilla for longer than eight weeks (according to the AAOMS criteria) after surgery (Mücke 2016; Ristow 2020). Mozzati 2013 defined the success rate as the proportion of participants without clinical signs of postoperative MRONJ, such as pain, swelling, and non‐healing exposed necrotic bone or fistulas, or both, with connection to the bone. Poxleitner 2020) defined the primary outcome as complete mucosal coverage at the extraction site at the follow‐up examination 90 days postoperatively. Mücke 2016 determined the incidence of MRONJ calculated as the incidence rate (i.e. the number of people developing MRONJ per patient‐years) and incidence proportion (i.e. the number of people developing MRONJ relative to the number of people in the study group.

Secondary outcome 

Intraoperative and postoperative complications were reported as secondary outcomes (Mozzati 2013). Quality of life was not investigated, except for one study that used the QoL EQ‐5D score sheet as one of the secondary study outcomes (Ristow 2020).

Treatment of MRONJ

Characteristics of the trial designs and settings 

The eight included studies concerning treatment and healing of MRONJ were conducted in the USA (Freiberger 2012), in Germany (Ristow 2016), in Korea (Park 2017), two in Italy (Giudice 2018a; Giudice 2018b), in Japan (Ohbayashi 2020), in Australia (Sim 2020), and in Turkey (Yüce 2021).

All of the eight studies used a parallel‐group design and had two treatment arms.

Characteristics of the participants

The eight studies focusing on treatment of MRONJ involved a total of 302 participants ranging from 13 to 55 participants per study. All had MRONJ related to the use of antiresorptive drugs (bisphosphonates or denosumab), mainly for the treatment of cancer or osteoporosis.

Characteristics of the interventions

The interventions used in the included studies included the following.

• Standard care with hyperbaric oxygen (HBO) versus standard care without HBO. Treatment for MRONJ included surgical debridement at the discretion of the referring surgeon and antibiotics for any sign of local infection. Participants in the HBO group received 40 HBO sessions at 2 atmospheres of pressure for two hours each over four weeks (Freiberger 2012).

• Preoperative treatment with doxycycline, which is incorporated into viable bone and is visualised with a certified medical lamp intraoperatively versus treatment with ampicillin/sulbactam (or clindamycin in case of hypersensitivity to penicillin or a penicillin allergy) preoperatively without doxycycline labelling. Autofluorescence of vital bone, which was induced with a special fluorescence lamp (provided for the study by the manufacturer), was used to visualise vital bone intraoperatively (Ristow 2016).

• Single application of leukocyte rich and platelet‐rich fibrin (L‐PRF group) versus combined application of L‐PRF and recombinant human bone morphogenetic protein 2 (rhBMP‐2) (PRF plus BMP group) after surgical removal of bone sequestra and before primary wound closure. All patients received antibiotics, analgesics, 0.12% chlorhexidine for daily rinsing, and professional dental prophylaxis in the week before surgery. Postoperatively, all patients were treated with an antibacterial mouth rinse until complete healing and with intravenous antibiotics for one week and oral antibiotics for another two weeks. (Park 2017).

• Autofluorescence‐guided surgery (AF) versus the traditionally treated group (non‐AF). In the AF group resection margins were visualised using the VELscope which reveals natural autofluorescence of healthy bone by emitting high‐intensity polarised ultraviolet light. In the non‐AF group, bone resection margins of patients were removed by the surgeon according his surgical experience. Absorbable sutures were used in all patients for tension‐free wound closure, and antibiotics started three days preoperatively were continued until seven days postoperatively (Giudice 2018a).

• Removal of necrotic bone and treatment with local application of platelet‐rich fibrin (PRF) at the surgical site before primary wound closure (tension‐free with absorbable suture material) versus removal of necrotic bone and primary wound closure without treatment with PRF. Starting three days before surgery, each patient was treated with antibiotics for 10 days. All patients also received a professional oral hygiene session and a mouth rinse with 0.2% chlorhexidine one week before surgery (Giudice 2018b).

• Removal of necrotic bone and treatment with local application of concentrated growth factor (CGF) at the surgical site before primary wound closure versus removal of necrotic bone and primary wound closure without treatment with CGF (Yüce 2021).

• Eight weeks of subcutaneous teriparatide (TPTD) (20 mg/day) injections versus placebo injections. All participants received calcium and vitamin D supplementation and standard clinical care (Sim 2020).

• Injections of 1×/week 56.5‐μg TPTD for six months versus 20‐μg TPTD injections daily for six months. Participants in both groups received conventional therapy and intensive antibiotic therapy as needed (Ohbayashi 2020).

Characteristics of the outcomes 

Primary outcome 

The primary outcome was success rate, defined as the absence of a MRONJ site after surgery i.e. full mucosal coverage or the absence of exposed necrotic bone with full coverage at eight weeks after surgery (according to the AAOMS criteria) (Ristow 2016). Giudice and colleagues defined primary outcome as mucosal integrity at six months postoperatively (Giudice 2018a; Giudice 2018b). For the primary outcome, oral lesions were scored by size and number, and a change in lesion scores compared with the baseline condition was used to grade the primary outcome. Possible outcome categories were healed (defined as gingival coverage with no exposed bone), improved, unchanged, or worse (Freiberger 2012).  Park 2017 categorised healing outcomes as complete healing, delayed healing or no resolution at four and 16 weeks post‐surgery. Ohbayashi 2020 defined the outcome as change in MRONJ stage, changes in bone metabolism (bone scintigraphy), the percentage of bone formation (axial and coronal computed tomography (CT) images of the maximum osteolysis area), and bone turnover markers (osteocalcin, procollagen type I N‐terminal propeptide, bone‐specific alkaline phosphatases, urinary cross‐linked N‐telopeptide of type I collagen, urinary deoxypiridinolin, and tartrate‐resistant acid phosphatase‐5b).

Secondary outcomes 

Secondary outcomes were quality of life (QoL), laboratory measures of bone turnover, and molecular indicators of osteoclast activation, such as RANK, RANKL, OPG, and pAKt (Freiberger 2012), or osteoblastic responses as measured biochemically and radiologically (Sim 2020).  Park 2017 analysed serum concentration of C‐terminal cross‐linked telopeptide of type I collagen (sCTX) of each patient with MRONJ additionally to the primary outcomes. Two studies reported mucosal integrity or full coverage at the remaining measurement time points, no signs of infection (Ristow 2016), loss of sensitivity (numbness) of the alveolar nerve, and subjective pain as secondary outcomes (Giudice 2018a).

Excluded studies

We excluded 20 studies (11 in the last version of the review and nine from the updated searches). They were all excluded because they were not RCTs (Asaka 2017; Bonacina 2011; Bramati 2015; Calvani 2018; Colapinto 2018; Coviello 2012; DE Iuliis 2014; Dimopoulos 2009; Garcia‐Martinez 2017; Giovannacci 2016; Jung 2017; Lee 2014; Montebugnoli 2007; Nica 2021; Pelaz 2014; Rodriguez 2019; Szentpeteri 2020; Vescovi 2010; Vescovi 2012; Watanabe 2021). See the Characteristics of excluded studies.

Ongoing studies

There are 12 RCTs currently in process that may be included in future iterations of this review (ChiCTR1900027382; DRKS00012888; jRCTs071200006; NCT01526915; NCT02198001; NCT03040778; NCT03269214; NCT03390777; NCT04512638; NCT04531800; UMIN000009132; UMIN000042862). See Characteristics of ongoing studies.

Risk of bias in included studies

See below and the Characteristics of included studies tables for more details about our risk of bias assessments. A graphical overview is provided in Figure 2. 

2.

Risk of bias summary: review authors' judgements about each risk of bias item for each included study

Allocation

In all 13 trials, participants were randomly divided into two groups. The method of random sequence generation was noted in only two studies: participants were assigned by a computer randomisation programme, and we judged the risk of bias to be low (Mozzati 2013; Ristow 2020). The authors of 10 trials did not mention the generation of the randomisation sequence and we therefore rated the level of risk as unclear (Freiberger 2012; Giudice 2018b; Mozzati 2012; Mücke 2016; Ohbayashi 2020; Park 2017; Poxleitner 2020; Ristow 2016; Sim 2020; Yüce 2021). In Giudice 2018a, participants were assigned to one of two groups by an independent resident senior, which we considered to be high risk of bias. 

We considered the use of sealed envelopes in three studies put them at low risk of bias due to inadequate allocation concealment (Freiberger 2012; Ristow 2020; Yüce 2021). Allocation concealment was not reported for 10 studies, so we categorised the risk level as unclear (Giudice 2018a; Giudice 2018b; Mozzati 2012; Mozzati 2013; Mücke 2016; Ohbayashi 2020; Park 2017; Poxleitner 2020; Ristow 2016; Sim 2020). 

Combining these possible sources of selection bias, we therefore judged only one study to be at low risk of selection bias (Ristow 2020), with one at high risk (Giudice 2018a), and the rest unclear. 

Blinding

Performance bias

We considered the level of risk for performance bias to be high in all but one of study (Sim 2020) because personnel were not blinded, which in many studies was due to the nature of the intervention. 

Detection bias

Detection bias was rated low for three studies where outcome assessors were blinded (Ristow 2020; Sim 2020; Yüce 2021). Therefore, we considered the level of risk for performance and detection bias to be high for all but one study (Sim 2020).

Incomplete outcome data

We judged attrition bias to be high in six studies (Freiberger 2012; Giudice 2018a; Mücke 2016; Ohbayashi 2020; Ristow 2020; Sim 2020). Although a description of losses and withdrawals was given, data analysis was performed as‐treated and not by intention‐to‐treat, or numbers of participants excluded were not balanced across intervention groups. None of the studies reported data in a format that would have enabled us to recalculate effects on an intent‐to‐treat basis.

We judged attrition bias to be low in six studies: follow‐up was complete in five studies, with all participants included in analyses (Giudice 2018b; Park 2017; Poxleitner 2020; Ristow 2016; Yüce 2021), and in Ristow 2016, although some participants were lost for the assessment of secondary endpoints; all participants were included in the assessment of the primary endpoint.

The level of risk of attrition bias was unclear in the other two studies because completeness or loss to follow‐up was not reported (Mozzati 2012; Mozzati 2013).

Selective reporting

We judged four studies to be at high risk of reporting bias (Giudice 2018b; Ohbayashi 2020; Park 2017; Sim 2020). In the report by Giudice 2018b, preoperative pain levels in control and intervention groups were not reported. (Ohbayashi 2020 presented in their publication P values of statistical analyses but did not report effect estimates of outcomes such as improvement of MRONJ. Park 2017 did not report results for all follow‐up time points. Sim 2020 did not report all outcomes; for example, quality of life survey data were missing. All of the other studies presented data for their planned outcomes and were therefore judged to be at low risk of reporting bias.

Other potential sources of bias

We identified no other sources of bias. 

Effects of interventions

See: Table 1; Table 2; Table 3; Table 4; Table 5; Table 6; Table 7; Table 8; Table 9; Table 10; Table 11

Prophylaxis of MRONJ

Regular dental examinations at three‐month intervals and preventive treatments versus standard care for the prophylaxis of MRONJ in men with metastatic prostate cancer and intravenous zoledronic acid

We identified one study with 253 participants that explored the preventive effect of a prophylactic treatment to reduce MRONJ in men with metastatic prostate cancer treated with zoledronic acid (Mücke 2016). The study compared regular dental examinations at three‐month intervals and preventive treatments (including antibiotics before dental extractions, and the use of techniques for wound closure that avoid exposure and contamination of bone) versus standard care (i.e. monitoring and treatment if necessary at the discretion of the participant's dentist).

Incidence of MRONJ

Our primary outcome, incidence of MRONJ, was reported as incidence rate per year and incidence proportion. MRONJ was defined as the non‐healing of exposed bone in the mandible or maxilla for longer than eight weeks without any change in the stage of disease. The mean follow‐up time was 28.8 months. Regular dental examinations at three‐month intervals and preventive treatments showed a lower risk ratio (RR) for MRONJ (0.10; 95% CI 0.02 to 0.39) compared to standard care when dental extractions were performed. There was also a clinically relevant  difference in the number of MRONJ cases per patient‐years (rate ratio 0.18; 95% CI 0.04 to 0.74). We rated the certainty of the evidence for the primary outcome to be very low due to very serious risk of bias (high and unbalanced rate of cross‐overs after randomisation; high dropout rates due to high mortality; failure to adhere to the intention‐to‐treat principle; mean follow‐up differed between experimental and control groups), and very serious limitation of indirectness (all male and high‐risk patients). See Table 1, Analysis 1.1, and Analysis 1.2.

1.1. Analysis.

Comparison 1: Dental examinations at three‐month intervals and preventive treatments (experimental) versus standard care (control) for prophylaxis of MRONJ, Outcome 1: MRONJ (incidence proportion)

1.2. Analysis.

Comparison 1: Dental examinations at three‐month intervals and preventive treatments (experimental) versus standard care (control) for prophylaxis of MRONJ, Outcome 2: MRONJ (incidence rate: MRONJ cases per patient‐year)

Plasma rich in growth factors inserted into the postextraction alveolus in addition to standardised medical and surgical care (experimental) versus standardised medical and surgical care alone (control) for prophylaxis of MRONJ

One RCT reported the effect of PRGF for preventing MRONJ in 176 participants with cancer undergoing dental extractions (Mozzati 2012).

Incidence of MRONJ

The diagnosis of MRONJ was based on clinical examination and radiographic examinations. Clinical signs of MRONJ were pain, swelling, and non‐healing exposed necrotic bone or fistulas, or both, with connection to the bone. The study group had a total follow‐up period of 24 to 60 months. At the last contact, no participants in the PRGF group (N = 91) but five participants in the control group (N = 85) developed MRONJ. The RR was 0.08 (95% CI 0.00 to 1.51). We rated the certainty of the evidence for the primary outcome to be very low due to imprecision and very serious risk of bias (high or unclear risk of selection bias, performance bias, detection bias, and attrition bias). See Table 2 and Analysis 2.1.

2.1. Analysis.

Comparison 2: Plasma rich in growth factors inserted into the postextraction alveolus in addition to standardised medical and surgical care (experimental) versus standardised medical and surgical care alone (control) for prophylaxis of MRONJ, Outcome 1: MRONJ (incidence proportion)

Rate of complications and side effects of the intervention

No intraoperative complications were observed in either of the groups.

Delicate surgery and closure by primary intention (experimental) versus non‐traumatic tooth avulsion and closure by secondary intention (control) for the prophylaxis of MRONJ 

One RCT with 700 participants compared wound closure by primary intention with wound closure by secondary intention after dental extractions in individuals treated with oral bisphosphonates (Mozzati 2013).

Incidence of MRONJ

The participants were regularly monitored for clinical signs of MRONJ: pain, swelling, and non‐healing exposed necrotic bone or fistulas, or both, with connection to the bone. In both study arms, no case of postoperative MRONJ was observed.

Rate of complications and side effects of the intervention

No intraoperative complications were observed in either of the two groups.

Primary closure of the extraction socket with a mucoperiosteal flap (experimental) versus application of platelet‐rich fibrin without primary closure (control) for the prophylaxis of MRONJ 

One RCT, which randomised 77 participants, reported on the effects of a PRF clot inserted into the extraction socket without primary closure versus primary closure of the extraction socket with a mucoperiosteal flap after tooth extraction (Poxleitner 2020). 

Incidence of MRONJ

All participants had complete mucosal coverage without any signs of MRONJ at the final control examination 90 days postoperatively.

Rate of complications and side effects of the intervention

No intraoperative complications occurred in either of the groups. Postoperative complications were found in six  participants (15.4%) in the mucoperiosteal flap group and in one participant (2.6%) in the PRF group. Postoperative complications were defined as any deviation from the normal postoperative course such as delayed wound healing, inflammation at the site of extraction, postoperative hematoma/haemorrhage, exposed bone/sequestrum. 

Subperiosteal prepared (SPP) mucoperiosteal flap versus epiperiosteal prepared (EPP) mucosa flap for the prevention of MRONJ in individuals treated with antiresorptive therapy for cancer or osteoporosis who underwent dental extractions

One RCT with 160 participants compared two methods of alveoplasty, namely the epiperiosteal prepared (EPP) mucosal flap and the subperiosteal prepared (SPP) mucoperiosteal flap after tooth extraction (Ristow 2020). 

Incidence of MRONJ

Six months after surgery, no participant in the EPP group (N = 67) but 5 participants (7.7%) in the SPP group (N = 65) had MRONJ. Diagnosis of MRONJ was based on the absence of full mucosal integrity. The RR was 0.09 (95% CI 0.00 to 1.56). We rated the certainty of the evidence for the primary outcome to be low due to imprecision and serious risk of bias (unclear selection bias, detection bias, high risk of performance bias, attrition bias and reporting bias). See Table 3 and Analysis 3.1. 

3.1. Analysis.

Comparison 3: Subperiosteal wound closure versus epiperiosteal wound closure after tooth extraction for prevention of MRONJ in patients on antiresorptive treatment, Outcome 1: MRONJ after tooth extraction (assessed with: absence of complete mucosal integrity) at 6 months

Treatment of MRONJ

Hyperbaric oxygen as an adjunct to conventional therapy (experimental) versus conventional therapy (control) for treatment of MRONJ

One RCT with 49 participants analysed the healing of MRONJ using HBO treatment in addition to standard care (antiseptic rinses, antibiotics, surgery) (Freiberger 2012). All participants terminated bisphosphonate administration before or at the time of consent, except one who continued bisphosphonate administration for one month after the initial examination.

Healing of MRONJ

Oral lesions were graded by size and number, and staged by clinical severity. The last contact was intended to be 24 months after consent; however, only 18 participants completed the full 24‐month observation period. Healing was defined as gingival coverage with no exposed bone. HBO in addition to standard care did not improve healing from MRONJ at any of the investigated time points (at last follow‐up: RR 1.56; 95% CI 0.77 to 3.18; n = 46). We rated the certainty of the evidence for the primary outcome to be very low due to imprecision and very serious risk of bias (unclear and high risk of selection bias, performance bias, detection bias, and attrition bias; failure to adhere to the intention‐to‐treat principle). See Table 4 and Analysis 4.1.

4.1. Analysis.

Comparison 4: Hyperbaric oxygen as an adjunct to conventional therapy (experimental) versus conventional therapy (control) for treatment of MRONJ, Outcome 1: Healing of MRONJ at last contact

Quality of life

Quality of life (QoL) was measured using the Duke Health Profile, a 17‐question generic self‐reporting instrument with six health domains (physical, mental, social, general, perceived health, and self‐esteem) and four dysfunction measurements (anxiety, depression, pain, and disability) (Freiberger 2012). QoL assessments were recorded at the time of the initial interview and after six months. Only within‐group comparisons for each domain were provided based on a dichotomous classification (‘improved’, ‘no change, or worse’). Because no score values were provided, we were unable to make a between‐group analysis.

Autofluorescence‐guided bone surgery (experimental) versus tetracycline fluorescence‐guided bone surgery (control) in individuals with MRONJ referred for surgical treatment

One RCT with 40 participants compared autofluorescence‐guided and tetracycline fluorescence‐guided bone surgery for the treatment of MRONJ (Ristow 2016).

Healing of MRONJ

The primary endpoint reported by Ristow 2016 was success rate. Success was defined as the absence of a MRONJ site after surgery, specified as the maintenance of full mucosal coverage (mucosal integrity) after surgery at the time of the evaluation. All measurements were acquired at five specific time points: preoperatively, and 10 days, eight weeks, six months, and one year after surgery. There was no relevant difference between the autofluorescence‐ and the tetracycline fluorescence‐guided groups at any of the time points (at one‐year follow‐up: RR 1.05; 95% CI 0.86 to 1.30; n = 34). We rated the certainty of the evidence for the primary outcome to be very low due to imprecision and very serious risk of bias (unclear and high risk of selection bias, performance bias, and detection bias). See Table 5 and Analysis 5.1.

5.1. Analysis.

Comparison 5: Autofluorescence‐guided bone surgery (experimental) versus tetracycline fluorescence‐guided bone surgery (control) for treatment of MRONJ, Outcome 1: Healing of MRONJ (defined as mucosal integrity) at 1 year

Bone morphogenetic protein‐2 together with platelet‐rich fibrin (experimental) versus platelet‐rich fibrin alone (control) for treatment of MRONJ

One RCT compared the healing results of the combined use of bone morphogenetic protein‐2 (BMP‐2) and platelet‐rich fibrin (PRF) with the use of PRF alone in patients with MRONJ and bony lesions requiring surgical debridement (Park 2017). Fifty‐five patients were included, of whom 30 patients were assigned to the experimental group (PRF and rhBMP‐2) and 25 participants to the PRF control group. 

Healing of MRONJ

Healing of MRONJ was defined as full mucosal coverage without clinical or radiographical evidence of MRONJ. At 16 weeks postoperatively, 29 participants (96.7%) of the PRF plus BMP group and 22 participants (88.0%) of the PRF group showed complete resolution (RR 1.10, 95% CI 0.94 to 1.29). We rated the certainty of the evidence for the primary outcome to be very low due to very serious limitations of imprecision and very serious risk of bias (unclear risk of selection bias, detection bias, high risk of performance bias, attrition bias, reporting bias). See Table 6 and Analysis 6.1. 

6.1. Analysis.

Comparison 6: Bone morphogenetic protein‐2 together with platelet‐rich fibrin (experimental) compared to platelet‐rich fibrin alone (control) for treatment of MRONJ., Outcome 1: Healing of MRONJ (defined as full mucosal coverage without clinical or radiographical evidence of MRONJ) after 16 weeks

Autofluorescence‐guided bone surgery (experimental) versus conventional bone surgery (control) for treatment of MRONJ

An RCT with 36 participants investigated whether autofluorescence‐guided delineation of resection margins for necrotic bone results in improved MRNOJ cure rates compared with conventional bone surgery based on surgeon experience alone (Giudice 2018a).

Healing of MRONJ

Healing of MRONJ was defined as the absence of exposed necrotic bone with full mucosal coverage and no signs of residual infection (pain, purulent discharge, or numbness) at one week, one month, six months, and one year after surgery. The primary outcome as reported by Giudice 2018a was mucosal integrity six months after surgery. MRONJ cure rates were not different between both treatment arms at one‐year follow‐up (RR 1.08; 95% CI 0.85 to 1.37; n = 30) or any other time points. We rated the certainty of the evidence for the primary outcome to be very low due to very serious imprecision and serious risk of bias (high selection bias, performance bias, attrition bias, unclear risk of detection bias). See Table 7 and Analysis 7.1.

7.1. Analysis.

Comparison 7: Autofluorescence‐guided surgery (experimental) compared to conventional surgery (control) for treatment of MRONJ, Outcome 1: Healing of MRONJ (defined as full mucosal coverage and no signs of residual infection) at 1 year

Platelet‐rich fibrin after bone surgery (experimental) versus surgery alone (control) for treatment of MRONJ

Giudice 2018b recruited 47 participants with the diagnosis of stage II or III of MRONJ to compare the efficacy of bone surgery either with or without  PRF for mucosal healing in an RCT. 

Healing of MRONJ

Giudice 2018b reported mucosal integrity, absence of residual infection or cutaneous fistulas, and no re‐intervention necessary to healing as study outcomes, evaluated at one month, six months, and one year after surgery. We used a composite endpoint requiring mucosal integrity and  absence of residual infection or cutaneous fistulas to define successful healing of MRONJ at 1‐year follow‐up. The RR was 1.05 (95% CI 0.90 to 1.22). We rated the certainty of the evidence for the primary outcome to be very low due to very serious limitation of imprecision and serious risk of bias (unclear selection bias, detection bias, high risk of performance bias, reporting bias).  See Table 8 and Analysis 8.1.

8.1. Analysis.

Comparison 8: Platelet‐rich fibrin after bone surgery (experimental) compared to surgery alone (control) for treatment of MRONJ, Outcome 1: Healing of MRONJ (defined as absence of infection and mucosal integrity without fistula) at 1 year

The need for re‐intervention may be considered a treatment failure, i.e. the endpoint of cure would likely not have been achieved without re‐intervention. Therefore, as part of a sensitivity analysis, we additionally considered the combined endpoint of mucosal integrity, absence of residual infection or cutaneous fistulas, and no re‐intervention necessary. The RR was 1.60 (95% CI 1.04 to 2.46). We rated the certainty of the evidence for the primary outcome to be very low due to very serious limitation of imprecision and serious risk of bias (unclear selection bias, detection bias, high risk of performance bias, reporting bias). See Table 8 and Analysis 8.2.

Concentrated growth factor and primary wound closure (experimental) versus primary wound closure only (control) for treatment of MRONJ

One RCT examined the effect of concentrated growth factor (CGF) applied to the surgical site compared with a control group without CGF (Yüce 2021). Twenty‐eight patients with stage 2 or 3 MRONJ and inadequate improvement with conservative treatment who underwent surgical debridement were randomised 1:1.  

Healing of MRONJ 

Healing was defined as soft tissue coverage without signs of infection or necrotic bone exposure. Healing of MRONJ was observed in 11 participants (78.6%) in the CGF group and 8 participants (57.1%) in the control group at six months after surgery (RR 1.38, 95% CI 0.81 to 2.34). We rated the certainty of the evidence for the primary outcome to be very low due to serious risk of bias (unclear selection bias, high risk of performance bias), very serious limitation of indirectness (only female participants with osteoporosis) and very serious limitation of imprecision (few participants). See Table 9 and Analysis 9.1.  

9.1. Analysis.

Comparison 9: Concentrated growth factor and primary wound closure (experimental) versus primary wound closure only (control) for treatment of MRONJ, Outcome 1: Healing of MRONJ (assessed with: mucosal integrity) at 6 months

Teriparatide 20 μg daily (experimental) versus placebo (control), in addition to standard care, for treatment of MRONJ

One RCT evaluated the efficacy of eight weeks of subcutaneous teriparatide (20 mg/day) versus placebo injections in addition to calcium and vitamin D supplementation and standard clinical care in 34 participants with established MRONJ (Sim 2020). Participants were observed for 12 months, with primary outcomes that included the clinical and radiologic resolution of MRONJ lesions.

Healing of MRONJ

The proportion of participants without any unresolved lesion at 52 weeks was 27.8% in the placebo group versus 26.7% in the teriparatide group (RR 0.96, 95% CI 0.31 to 2.95). We rated the certainty of the evidence for the primary outcome to be low due to imprecision and serious risk of bias (unclear selection bias, high risk of attrition bias and reporting bias). We downgraded the certainty of the evidence by two levels due to imprecision and serious risk of bias (unclear selection bias, high risk of attrition bias and reporting bias). See Table 10 and Analysis 10.1.

10.1. Analysis.

Comparison 10: Teriparatide 20 μg daily (experimental) versus placebo (control), in addition to standard care, for treatment of MRONJ, Outcome 1: Healing of MRONJ (defined as absence of any unresolved lesion) at 1 year

Teriparatide 56.5 μg weekly (experimental) versus teriparatide 20 μg daily (control) in addition to conventional treatment of MRONJ

Ohbayashi 2020 examined the clinical outcomes of teriparatide (TPTD) treatment in 13 patients with stage II‐III bisphosphonate‐induced ONJ and compared two TPTD dosing regimens, namely 20 μg TPTD injections daily for six months versus 1×/week 56.5 μg TPTD injections for six months, in an RCT. After one dropout, 12 participants, six in each group, were analysed. The last follow‐up time point was six months after the start of teriparatide treatment. 

Healing of MRONJ

The authors reported MRONJ remission as an outcome measure based on the clinical staging of MRONJ. However, the report aggregated the rates of partial remission and complete remission, leaving the cure rate, i.e. the rate of complete remission, unclear. At the end of six months of TPTD treatment, partial remission or complete remission was achieved in three of six participants in the experimental group and in five of six participants in the control group (RR 0.60, 95% CI 0.25 to 1.44). We rated the certainty of the evidence for the primary outcome to be very low due to imprecision and very serious risk of bias (unclear selection bias and detection bias, high risk of performance bias, attrition bias and reporting bias). See Table 11 and Analysis 11.1.

11.1. Analysis.

Comparison 11: Teriparatide 56.5 μg weekly (experimental) versus teriparatide 20 μg daily (control), in addition to standard care, for treatment of MRONJ, Outcome 1: Healing of MRONJ (defined as partial or complete remission) at 6 months

Discussion

At present, the mechanisms of medication‐related osteonecrosis of the jaw (MRONJ) are not well known, and the prevention and treatment of MRONJ remains challenging. Thus, it is important to identify effective strategies for managing this well‐known complication of antiresorptive medication.

Summary of main results

For the prevention of MRONJ, we identified five randomised controlled trials (RCTs), each evaluating different interventions (Mozzati 2012; Mozzati 2013; Mücke 2016; Poxleitner 2020; Ristow 2020). 

There is very low‐certainty evidence that dental examinations at three‐month intervals and preventive treatments are more effective than standard care in reducing the incidence proportion and the incidence rate of MRONJ in men taking intravenous bisphosphonates for advanced cancer and bone metastases. 

Tooth extractions are considered one of the main triggers for the development of MRONJ and various preventive measures have been tested (Mozzati 2012; Mozzati 2013; Poxleitner 2020; Ristow 2020). After evaluating the available evidence, it is not possible to assert or refute a benefit of any of the interventions reviewed, such as plasma rich in growth factors (PRGF) or platelet‐rich fibrin (PRF) inserted into the postextraction alveolus or a specific surgical strategy (i.e. wound closure by primary or secondary intention) or a specific alveoplasty technique (i.e. epiperiosteal  or subperiosteal wound closure) for prevention of MRONJ. 

We identified eight RCTs that evaluated specific methods to improve the healing of MRONJ, namely hyperbaric oxygen (HBO) therapy (Freiberger 2012), fluorescence‐guided bone surgery (Ristow 2016; Giudice 2018a), growth factors such as  PRF, concentrated growth factor or bone morphogenic protein 2 (Giudice 2018b; Yüce 2021; Park 2017), and teriparatide (Ohbayashi 2020; Sim 2020). There was insufficient evidence to either claim or refute a benefit of any of these therapies for improved healing of MRONJ. There was also insufficient evidence to support either autofluorescence‐guided bone surgery or tetracycline fluorescence‐guided bone surgery for improved healing of MRONJ. The small sample size may have contributed to a lack of measurable effect. 

Overall completeness and applicability of evidence

The types of interventions evaluated in the included RCTs varied widely and so we were not able to combine data from different studies.

Most RCTs included only people with bisphosphonates for the treatment of cancer or osteoporosis, or both. Two studies included patients treated with both denosumab and bisphosphonates (Poxleitner 2020; Ristow 2020). Although another study also allowed individuals treated with denosumab to participate, only individuals treated with bisphosphonates were included (Ristow 2016). None of the trials investigated the association between MRONJ and antiangiogenic medications. Thus, the included RCTs do not cover the entire spectrum of medications associated with MRONJ.

One trial recruited a highly selective group of participants (i.e. men with prostate cancer receiving zoledronic acid for the treatment of bone metastases) (Mücke 2016). The applicability of the results of this study to other populations is unclear.

A subgroup of individuals, namely those with a history of head and neck radiation therapy, were generally excluded from the RCTs due to the widely accepted case definition of MRONJ. Although exclusion of these individuals may be useful in reducing the heterogeneity of the study populations and in controlling for an important influencing variable, this may have impaired the overall completeness of evidence.

Quality of the evidence

We judged the overall certainty of the evidence to be low or very low, meaning that we are uncertain about the estimates of effect.

All included studies were assessed to have a high risk of bias overall, as they had at least one domain rated at high risk. All trials but one (Sim 2020) were open‐label. Due to their nature, some interventions could not be blinded to participants or surgeons. In only a few studies were outcome assessors blinded (Ristow 2020; Sim 2020; Yüce 2021). Altogether, a lack of blinding confers a high risk of bias. In one trial, the length of follow‐up differed between comparison groups, which may have biased the results of the study (Mücke 2016). A high and unbalanced rate of cross‐overs after randomisation between the comparison groups in two trials may also have conferred a high risk of bias (Freiberger 2012; Mücke 2016). We also downgraded the certainty of the evidence due to imprecision. This is because most studies included relatively few participants (Freiberger 2012; Ristow 2016; Giudice 2018a; Ohbayashi 2020; Poxleitner 2020; Sim 2020; Yüce 2021) or had few or no events  (Mozzati 2012; Mozzati 2013; Poxleitner 2020), making it impossible to determine the effect estimates, or the effect estimates had wide 95% confidence intervals. In two comparisons, the certainty of evidence was downgraded due to indirectness (based on Mücke 2016 and Yüce 2021); only one sex (either males or females) was included with one particular disease or the participants were high‐risk patients.

Potential biases in the review process

The methods we used in the review were established and documented in advance of the review being undertaken. We were not influenced by prior knowledge of the study results when making judgements regarding study eligibility. We made no subsequent changes to the types of studies and types of participants to be included in the review as specified in the protocol, with one exception. For trials investigating the effects of interventions for the prophylaxis of MRONJ, we originally required a follow‐up period of at least three years. The three‐year follow‐up threshold, however, turned out not to be a feasible selection criterion (see Primary outcomes). We consider this change to the inclusion criteria to be well justified, however, and we do not believe that we have introduced a relevant selection bias.

Cochrane Oral Health Information Specialist (Anne Littlewood) conducted comprehensive searches of journal and conference databases to ensure that all published and unpublished trials were identified. We did not limit the searches to a particular language. Two review authors independently extracted the trials that met the inclusion criteria. The study authors were contacted where necessary to ascertain if any newer data were available following publication.

Agreements and disagreements with other studies or reviews

In the last two years, several systematic reviews have been published addressing the prophylaxis and treatment of MRONJ (Cabras 2021; Di Fede 2021; Dos Santos Ferreira 2021; Fortunato 2020; Goker 2021; Govaerts 2020; de Souza Tolentino 2019). Unlike our work, which included only RCTs, these reviews additionally included non‐RCTs, such as case reports, case series, retrospective studies, and prospective cohort studies. The reviews investigated the efficacy of hyperbaric oxygen, low‐intensity laser, autologous platelet concentrates (platelet‐rich plasma, plasma‐rich growth factors, and platelet‐rich fibrin), teriparatide, ozone applications, or fluorescence‐guided surgery in the treatment of MRONJ (de Souza Tolentino 2019; Di Fede 2021; Dos Santos Ferreira 2021;  Fortunato 2020; Goker 2021; Govaerts 2020; de Souza Tolentino 2019). Another focus of the reviews was prophylaxis of MRONJ by periprocedural administration of systemic antibiotics (Cabras 2021) or autologous platelet concentrates (Fortunato 2020) in patients under antiresorptive and/or biologic agents undergoing tooth extraction or implant surgery. All these reviews agree that high‐quality research, i.e. randomised controlled trials with large samples are required before conclusive statements can be made regarding strategies for the prevention or treatment of MRONJ.

Authors' conclusions

Implications for practice.

Prophylaxis of MRONJ

We identified five randomised controlled trials (RCTs) that evaluated various interventions for the prophylaxis of medication‐related osteonecrosis of the jaw (MRONJ). 

One open‐label RCT provided very low‐certainty evidence that dental examinations at three‐month intervals plus preventive treatments are more effective than standard care in reducing the incidence of MRONJ in men taking intravenous bisphosphonates for advanced cancer and bone metastases (Mücke 2016). Our conclusion from the study is that people receiving intravenous bisphosphonates for advanced cancer and bone metastases should be placed on a regular recall schedule. Recall visits should include a check of oral hygiene, periodontal diseases, cavities, and effective infection control. Of note, 29% of participants randomly allocated to the experimental arm later declined to have frequent dental check‐up visits (Mücke 2016). Given that the adherence rates observed in clinical trials generally exceed those observed in a real‐life setting, the limited acceptance of a dental monitoring programme among such individuals may limit the success of this preventive intervention. Thus, the motivation of these individuals is very important. The applicability of the study results to populations other than individuals with cancer taking intravenous bisphosphonates is unclear.

With dentoalveolar surgery being considered a major risk factor for developing MRONJ, five RCTs evaluated interventions that were proposed to reduce the incidence of MRONJ in individuals undergoing dental extractions. There was insufficient evidence to either claim or refute a benefit of any of the tested interventions for the prophylaxis of MRONJ. The small sample size relative to the low event rate of MRONJ may have contributed to the lack of a measurable effect. As we are unable to draw definitive conclusions about these interventions and techniques, we cannot make any clinical recommendations regarding them.

Treatment of MRONJ

We found eight RCTs that compared regimens for treating MRONJ. There was insufficient evidence to either claim or refute a benefit of any of the tested interventions for the treatment of MRONJ. The small sample size may have contributed to a lack of measurable effect. Moreover, methodological constraints of the trials were associated with a high risk of bias, contributing to uncertainty about any estimates of effect.

Implications for research.

Prophylaxis of MRONJ

Incidence rates for MRONJ depend on the specific drug, its dose, and the duration of treatment, and range from 0.004% to 6.7% (Ruggiero 2014). Thus, depending on the population under investigation and the specific at‐risk drug therapy, studies may require several hundred to several thousand participants to provide sufficient statistical power to detect meaningful effects of preventive measures on the incidence of MRONJ. Although dentoalveolar surgery is considered a common predisposing event for developing MRONJ (Ruggiero 2014), well‐designed RCTs are lacking to identify effective preventive strategies in individuals at risk undergoing dentoalveolar surgery. Importantly, the concept of 'a drug holiday' (stopping the antiresorptive medication) in individuals receiving oral bisphosphonates or denosumab who require tooth extractions is a matter of debate and requires future research (Damm 2013; Ruggiero 2014).

Treatment of MRONJ

Future RCTs should address important practice‐related research questions, namely the comparison of surgical versus non‐surgical protocols or conservative versus aggressive surgical protocols for the stage‐specific treatment of MRONJ. Moreover, the evaluation of add‐on effects for adjunct treatments such as HBO, α‐tocopherol, pentoxifylline, ozone therapy, or low‐level laser therapy, is important. Blinding of participants and clinicians (surgeons) may not be possible because of the nature of most interventions, but efforts should be made to ensure the blinding of outcome assessors (data collectors), which is crucial to ensure unbiased outcome assessment. One important limitation of existing RCTs was the small sample size. The sample size of future trials should be appropriate to allow meaningful conclusions to be drawn. In order to deal with the rare event rates of MRONJ, future trials should preferably follow a multicentric design and include sufficient participating centres. This will facilitate reaching a large number of cases.

What's new

Date	Event	Description
16 June 2021	New citation required but conclusions have not changed	The addition of new evidence did not change our conclusions.
16 June 2021	New search has been performed	Search updated. Eight new trials included

History

Protocol first published: Issue 11, 2016
Review first published: Issue 10, 2017

Notes

This is an update of a review published in 2017 (Beth‐Tasdogan 2017).

Appendices

Appendix 1. Cochrane Oral Health’s Trials Register search strategy

Cochrane Oral Health’s Trials Register is available via the Cochrane Register of Studies. For information on how the register is compiled, see https://oralhealth.cochrane.org/trials 

1 ((((medication or bisphosphonate or drug) and (osteonecrosis or necrosis) and jaw*)):ti,ab) AND (INREGISTER)
2 (MRONJ or BRONJ or BONJ:ti,ab) AND (INREGISTER)
3 (#1 or #2) AND (INREGISTER)
4 ((osteonecrosis or "bone necrosis"):ti,ab) AND (INREGISTER)
5 (osteochemonecro*:ti,ab) AND (INREGISTER)
6 (#4 or #5) AND (INREGISTER)
7 ((jaw* or jawbone* or mandib* or maxill* or (alveolar and bone*)):ti,ab) AND (INREGISTER)
8 ((diphosphonate* or bisphosphonate* or aminobisphosphonate* or alendronate or risedronate or pamidronate or "zoledronic acid" or ibandronate or "alendronic acid" or bevacizumab or denosumab or "etidronate disodium" or "ibandronic acid" or sirolimus or "sodium clodronate" or sorafenib or sunitinib or "tiludronic acid" or zoledronate or didronel or "clodronate disodium" or tiludronate or "risedronic acid" or "clodronic acid"):ti,ab) AND (INREGISTER)
9 ((Fosamax or Fosavance or Actonel or Aclasta or Zometa or Reclast or Didronel or Skelid or Bondronat or Bonviva or Aredia or Bonefos or Nexavar or Avastin or Prolia or Xgeva or Boniva or Atelvia or Rapamune or Rapamycin or Sutent or Zometa):ti,ab) AND (INREGISTER)
10 ((denosumab or prolia or ranmark or xgeva):ti,ab) AND (INREGISTER)
11 ((antivegf of avastin or bevacizumab):ti,ab) AND (INREGISTER)
12 ((aflibercept or eylea or "vegf trap" or zaltrap):ti,ab) AND (INREGISTER)
13 (("su 11248" or sunitinib or sunitinibum or sutent):ti,ab) AND (INREGISTER)
14 (("bms 907351" or bms907351 or cabozantinib or cometriq or "xl 184" or "xl 184" or xl184):ti,ab) AND (INREGISTER)
15 ((temsirolimus or torisel):ti,ab) AND (INREGISTER)
16 ((afinitor or certican or everolimus or everolimus or rad001 or "sdz rad" or votubia or zortress):ti,ab) AND (INREGISTER)
17 (#8 or #9 or #10 or #11 or #12 or #13 or #14 or #15 or #16) AND (INREGISTER)
18 (#6 and #7 and #17) AND (INREGISTER)
19 (#3 or #18) AND (INREGISTER)

Appendix 2. Cochrane Central Register of Controlled Clinical Trials (CENTRAL) search strategy

#1 [mh ^"Bisphosphonate‐associated osteonecrosis of the jaw"]
#2 ((medication or bisphosphonate or drug) near/4 (osteonecrosis or necrosis) near/3 jaw*)
#3 (MRONJ or BRONJ or BONJ):ti,ab
#4 {or #1‐#3}
#5 [mh ^Osteonecrosis]
#6 (osteonecro* or "bone necrosis"):ti,ab
#7 osteochemonecro*:ti,ab
#8 {or #5‐#7}
#9 [mh jaw]
#10 [mh ^"alveolar bone loss"]
#11 [mh ^"jaw diseases"]
#12 (jaw* or jawbone* or mandib* or maxill* or (alveolar near/4 bone*)):ti,ab
#13 {or #9‐#12}
#14 [mh diphosphonates]
#15 (diphosphonate* or bisphosphonate* or aminobisphosphonate* or alendronate or risedronate or pamidronate or "zoledronic acid" or ibandronate or "alendronic acid" or bevacizumab or denosumab or "etidronate disodium" or "ibandronic acid" or sirolimus or "sodium clodronate" or sorafenib or sunitinib or "tiludronic acid" or zoledronate or didronel or "clodronate disodium" or tiludronate or "risedronic acid" or "clodronic acid"):ti,ab
#16 (Fosamax or Fosavance or Actonel or Aclasta or Zometa or Reclast or Didronel or Skelid or Bondronat or Bonviva or Aredia or Bonefos or Nexavar or Avastin or Prolia or Xgeva or Boniva or Atelvia or Rapamune or Rapamycin or Sutent or Zometa):ti,ab
#17 [mh ^Denosumab]
#18 (denosumab or prolia or ranmark or xgeva):ti,ab
#19 [mh ^Bevacizumab]
#20 (antivegf of avastin or bevacizumab):ti,ab
#21 (aflibercept or eylea or "vegf trap" or zaltrap):ti,ab
#22 ("su 11248" or sunitinib or sunitinibum or sutent):ti,ab
#23 ("bms 907351" or bms907351 or cabozantinib or cometriq or "xl 184" or "xl 184" or xl184):ti,ab
#24 (temsirolimus or torisel):ti,ab
#25 [mh ^Everolimus]
#26 (afinitor or certican or everolimus or everolimus or rad001 or "sdz rad" or votubia or zortress):ti,ab
#27 {or #14‐#26}
#28 #8 and #13 and #27
#29 #4 or #28

Appendix 3. MEDLINE Ovid search strategy

1. Bisphosphonate‐associated osteonecrosis of the jaw/
2. ((medication or bisphosphonate or drug) adj4 (osteonecrosis or necrosis) adj3 jaw$).ti,ab.
3. (MRONJ or BRONJ or BONJ).ti,ab.
4. or/1‐3
5. Osteonecrosis/
6. (osteonecro$ or "bone necrosis").ti,ab.
7. osteochemonecro$.ti,ab.
8. or/5‐7
9. exp Jaw/
10. Alveolar bone loss/ci
11. Jaw diseases/ci
12. (jaw or jawbone$ or mandibl$ or maxill$ or (alveolar adj4 bone$)).ti,ab.
13. or/9‐12
14. exp Diphosphonates/
15. (diphosphonate$ or bisphosphonate$ or aminobisphosphonate$ or alendronate or risedronate or pamidronate or "zoledronic acid" or ibandronate or "alendronic acid" or bevacizumab or denosumab or "etidronate disodium" or "ibandronic acid" or sirolimus or "sodium clodronate" or sorafenib or sunitinib or "tiludronic acid" or zoledronate or didronel or "clodronate disodium" or tiludronate or "risedronic acid" or "clodronic acid").ti,ab
16. (Fosamax or Fosavance or Actonel or Aclasta or Zometa or Reclast or Didronel or Skelid or Bondronat or Bonviva or Aredia or Bonefos or Nexavar or Avastin or Prolia or Xgeva or Boniva or Atelvia or Rapamune or Rapamycin or Sutent or Zometa).ti,ab.
17. Denosumab/
18. (denosumab or prolia or ranmark or xgeva).ti,ab.
19. Bevacizumab/
20. (antivegf or avastin or bevacizumab).ti,ab.
21. (aflibercept or eylea or "vegf trap" or zaltrap).ti,ab.
22. ("su 11248" or sunitinib or sunitinibum or sutent).ti,ab.
23. ("bms 907351" or bms907351 or cabozantinib or cometriq or "xl 184" or "xl 184" or xl184).ti,ab.
24. (temsirolimus or torisel).ti,ab.
25. Everolimus/
26. (afinitor or certican or everolimus or everolimus or rad001 or "sdz rad" or votubia or zortress).ti,ab.
27. or/14‐26
28. 8 and 13 and 27
29. 4 or 28

The above subject search was linked with the highly sensitive search strategy designed by Cochrane for identifying randomised controlled trials and controlled clinical trials in MEDLINE (as described in Lefebvre 2022, box 3c).

1. randomized controlled trial.pt.
2. controlled clinical trial.pt.
3. randomized.ab.
4. placebo.ab.
5. drug therapy.fs.
6. randomly.ab.
7. trial.ab.
8. groups.ab.
9. or/1‐8
10. exp animals/ not humans.sh.
11. 9 not 10

Appendix 4. Embase Ovid search strategy

1. ((medication or bisphosphonate or drug) adj4 (osteonecrosis or necrosis) adj3 jaw$).ti,ab.
2. (MRONJ or BRONJ or BONJ).ti,ab.
3. 1 or 2
4. "Bone necrosis"/
5. "Jaw osteonecrosis"/
6. (osteonecro$ or "bone necrosis").ti,ab.
7. osteochemonecro$.ti,ab.
8. or/4‐6
9. exp Jaw/
10. Alveolar bone loss/
11. Jaw disease/
12. (jaw or jawbone$ or mandibl$ or maxill$ or (alveolar adj4 bone$)).ti,ab.
13. or/9‐12
14. exp Bisphosphonic acid derivative/
15. (diphosphonate$ or bisphosphonate$ or aminobisphosphonate$ or alendronate or risedronate or pamidronate or "zoledronic acid" or ibandronate or "alendronic acid" or bevacizumab or denosumab or "etidronate disodium" or "ibandronic acid" or sirolimus or "sodium clodronate" or sorafenib or sunitinib or "tiludronic acid" or zoledronate ordidronel or "clodronate disodium" or tiludronate or "risedronic acid" or "clodronic acid").ti,ab.
16. (Fosamax or Fosavance or Actonel or Aclasta or Zometa or Reclast or Didronel or Skelid or Bondronat or Bonviva or Aredia or Bonefos or Nexavar or Avastin or Prolia or Xgeva or Boniva or Atelvia or Rapamune or Rapamycin or Sutent or Zometa).ti,ab.
17. Denosumab/
18. (denosumab or prolia or ranmark or xgeva).ti,ab.
19. Bevacizumab/
20. (antivegf or avastin or bevacizumab).ti,ab.
21. (aflibercept or eylea or "vegf trap" or zaltrap).ti,ab.
22. ("su 11248" or sunitinib or sunitinibum or sutent).ti,ab.
23. ("bms 907351" or bms907351 or cabozantinib or cometriq or "xl 184" or "xl 184" or xl184).ti,ab.
24. (temsirolimus or torisel).ti,ab.
25. Everolimus/
26. (afinitor or certican or everolimus or everolimus or rad001 or "sdz rad" or votubia or zortress).ti,ab.
27. or/14‐26
28. 8 and 13 and 27
29. 3 or 28

The above subject search was linked with the highly sensitive search strategy designed by Cochrane for identifying randomised controlled trials and controlled clinical trials in Embase (as described in Lefebvre 2022, box 3e). 

1. Randomized controlled trial/

2. Controlled clinical study/

3. random$.ti,ab.

4. randomization/

5. intermethod comparison/

6. placebo.ti,ab.

7. (compare or compared or comparison).ti.

8. ((evaluated or evaluate or evaluating or assessed or assess) and (compare or compared or comparing or comparison)).ab.

9. (open adj label).ti,ab.

10. ((double or single or doubly or singly) adj (blind or blinded or blindly)).ti,ab.

11. double blind procedure/

12. parallel group$1.ti,ab.

13. (crossover or cross over).ti,ab.

14. ((assign$ or match or matched or allocation) adj5 (alternate or group$1 or intervention$1 or patient$1 or subject$1 or participant$1)).ti,ab.

15. (assigned  or allocated).ti,ab.

16. (controlled adj7 (study or design or trial)).ti,ab.

17. (volunteer or volunteers).ti,ab.

18. human experiment/

19. trial.ti.

20. or/1‐19

21. random$ adj sampl$ adj7 ("cross section$" or questionnaire$1 or survey$ or database$1)).ti,ab. not (comparative study/ or controlled study/ or randomi?ed controlled.ti,ab. or randomly assigned.ti,ab.)

22. Cross‐sectional study/ not (randomized controlled trial/ or controlled clinical study/ or controlled study/ or randomi?ed controlled.ti,ab. or control group$1.ti,ab.)

23. (((case adj control$) and random$) not randomi?ed controlled).ti,ab.

24. (Systematic review not (trial or study)).ti.

25. (nonrandom$ not random$).ti,ab.

26. "Random field$".ti,ab.

27. (random cluster adj3 sampl$).ti,ab.

28. (review.ab. and review.pt.) not trial.ti.

29. "we searched".ab. and (review.ti. or review.pt.)

30. "update review".ab.

31. (databases adj4 searched).ab.

32. (rat or rats or mouse or mice or swine or porcine or murine or sheep or lambs or pigs or piglets or rabbit or rabbits or cat or cats or dog or dogs or cattle or bovine or monkey or monkeys or trout or marmoset$1).ti. and animal experiment/

33. Animal experiment/ not (human experiment/ or human/)

34. or/21‐33

35. 20 not 34

 

Appendix 5. US National Institutes of Health Ongoing Trials Register (ClinicalTrials.gov) search strategy

bisphosphonates and jaw
osteonecrosis and jaw
necrosis and jaw

Appendix 6. World Health Organization International Clinical Trials Registry Platform search strategy

bisphosphonates and jaw

osteonecrosis and jaw or necrosis and jaw

Data and analyses

Comparison 1. Dental examinations at three‐month intervals and preventive treatments (experimental) versus standard care (control) for prophylaxis of MRONJ.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1.1 MRONJ (incidence proportion)	1	253	Risk Ratio (M‐H, Fixed, 95% CI)	0.10 [0.02, 0.39]
1.2 MRONJ (incidence rate: MRONJ cases per patient‐year)	1		Rate Ratio (IV, Fixed, 95% CI)	0.18 [0.04, 0.74]

Comparison 2. Plasma rich in growth factors inserted into the postextraction alveolus in addition to standardised medical and surgical care (experimental) versus standardised medical and surgical care alone (control) for prophylaxis of MRONJ.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
2.1 MRONJ (incidence proportion)	1	176	Risk Ratio (M‐H, Random, 95% CI)	0.08 [0.00, 1.51]

Comparison 3. Subperiosteal wound closure versus epiperiosteal wound closure after tooth extraction for prevention of MRONJ in patients on antiresorptive treatment.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
3.1 MRONJ after tooth extraction (assessed with: absence of complete mucosal integrity) at 6 months	1	132	Risk Ratio (M‐H, Fixed, 95% CI)	0.09 [0.00, 1.56]

Comparison 4. Hyperbaric oxygen as an adjunct to conventional therapy (experimental) versus conventional therapy (control) for treatment of MRONJ.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
4.1 Healing of MRONJ at last contact	1	46	Risk Ratio (M‐H, Fixed, 95% CI)	1.56 [0.77, 3.18]

Comparison 5. Autofluorescence‐guided bone surgery (experimental) versus tetracycline fluorescence‐guided bone surgery (control) for treatment of MRONJ.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
5.1 Healing of MRONJ (defined as mucosal integrity) at 1 year	1	34	Risk Ratio (M‐H, Fixed, 95% CI)	1.05 [0.86, 1.30]

Comparison 6. Bone morphogenetic protein‐2 together with platelet‐rich fibrin (experimental) compared to platelet‐rich fibrin alone (control) for treatment of MRONJ.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
6.1 Healing of MRONJ (defined as full mucosal coverage without clinical or radiographical evidence of MRONJ) after 16 weeks	1	55	Risk Ratio (M‐H, Fixed, 95% CI)	1.10 [0.94, 1.29]

Comparison 7. Autofluorescence‐guided surgery (experimental) compared to conventional surgery (control) for treatment of MRONJ.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
7.1 Healing of MRONJ (defined as full mucosal coverage and no signs of residual infection) at 1 year	1	30	Risk Ratio (M‐H, Fixed, 95% CI)	1.08 [0.85, 1.37]

Comparison 8. Platelet‐rich fibrin after bone surgery (experimental) compared to surgery alone (control) for treatment of MRONJ.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
8.1 Healing of MRONJ (defined as absence of infection and mucosal integrity without fistula) at 1 year	1	47	Risk Ratio (M‐H, Fixed, 95% CI)	1.05 [0.90, 1.22]
8.2 Healing of MRONJ (defined as absence of infection, mucosal integrity without fistula,  and no need for re‐intervention) at 1 year	1	47	Risk Ratio (M‐H, Fixed, 95% CI)	1.60 [1.04, 2.46]

Comparison 9. Concentrated growth factor and primary wound closure (experimental) versus primary wound closure only (control) for treatment of MRONJ.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
9.1 Healing of MRONJ (assessed with: mucosal integrity) at 6 months	1	28	Risk Ratio (M‐H, Fixed, 95% CI)	1.38 [0.81, 2.34]

Comparison 10. Teriparatide 20 μg daily (experimental) versus placebo (control), in addition to standard care, for treatment of MRONJ.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
10.1 Healing of MRONJ (defined as absence of any unresolved lesion) at 1 year	1	33	Risk Ratio (M‐H, Fixed, 95% CI)	0.96 [0.31, 2.95]

Comparison 11. Teriparatide 56.5 μg weekly (experimental) versus teriparatide 20 μg daily (control), in addition to standard care, for treatment of MRONJ.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
11.1 Healing of MRONJ (defined as partial or complete remission) at 6 months	1	12	Risk Ratio (M‐H, Fixed, 95% CI)	0.60 [0.25, 1.44]

Characteristics of studies

Characteristics of included studies [ordered by study ID]

Freiberger 2012.

Study characteristics
Methods	Trial design: single‐centre, interventional, prospective, unblinded RCT Duration of study: enrolment period from July 2006 to December 2010 Follow‐up: per protocol 2 years Sample size calculation: quote: "The target study sample size was calculated based on the primary outcome variable (change in oral lesion size and number) and indicated a requirement for 33 to 37 subjects with MRONJ per group. This assumed a spontaneous remission rate of 5% to 10% for the study to detect at least a 25% difference in cure rates between HBO‐treated patients and non‐treated controls. The authors used alpha value equal to 0.05 and a power equal to 0.80 for these calculations." Country of origin: USA Year of publication: 2012 Language of the original publication: English Category: treatment of MRONJ, non‐surgical Funding: quote: "This work was supported by a grant from Novartis Healthcare." Registration in a public trials registry: NCT00462098
Participants	49 participants with MRONJ randomised into 2 groups: 27 control (standard care), 22 experimental (standard care + hyperbaric oxygen (HBO)) Mean age: control 66 yr, HBO 66 yr Sex: control 56% female, HBO 59% female Condition treated with bisphosphonates: osteoporosis (15% of total sample), cancer and other indications (85% of total sample) Inclusion criteria Able to consent Has taken bisphosphonates Presence of exposed bone in the maxillofacial area with no evidence of healing after 6 weeks of appropriate evaluation and dental care No radiation history of the affected area Exclusion criteria Unable to consent Ineligible for HBO Taking protease inhibitors for HIV Any past history of radiation to the jaw Metastatic or recurrent malignant disease of the jaw or oropharynx Life expectancy less than 12 months Tobacco use Pregnancy
Interventions	Control (n = 22): standard care (antiseptic rinses, antibiotics, and surgery, if indicated by the participant's individual conditions) Experimental (n = 27): standard care plus 40 sessions of 100% oxygen at 2 atmospheres of pressure for 2 hours each, twice a day
Outcomes	Primary Change from baseline in oral lesion size and number Secondary Pain (0‐ to 10‐point Likert scale) Quality of life (Duke Health Profile, a 17‐question generic self‐reporting instrument) Serum measurements of bone turnover (data collected but not reported) Molecular measures of osteoclast signalling
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Method of sequence generation not reported
Allocation concealment (selection bias)	Low risk	Quote: "The randomization of patients with MRONJ to treatment groups was performed after informed consent, but before the initial staging examination using a series of 70 opaque envelopes containing the assignment."
Blinding of participants and personnel (performance bias) All outcomes	High risk	Quote: "The subjects and staff were not blinded to therapy because of the impracticality of providing sham HBO; however, the oral‐maxillofacial surgeon was not told the subjects’ assignments before the initial staging examination."
Blinding of outcome assessment (detection bias) All outcomes	High risk	Quote: "Lesion scores at the time of last contact were assigned by the study team, including the oral‐maxillofacial surgeon." No blinding of outcome assessment
Incomplete outcome data (attrition bias) All outcomes	High risk	High attrition rate: at the 12‐ and 18‐month evaluations 50% and 63%, respectively, of participants were lost to follow‐up. High and unbalanced rate of cross‐overs: after randomisation 5 participants switched from the control to the HBO group; 1 participant assigned to the HBO group declined HBO treatment and was switched to the control group. Data analysis: as‐treated, not by intention‐to‐treat
Selective reporting (reporting bias)	Low risk	All outcome variables listed in the Methods and the study protocol were reported.

Giudice 2018a.

Study characteristics
Methods	Trial design: monocentric, prospective RCT Duration of study: enrolment period from April 2015 to May 2016 Follow‐up: per protocol 1 year Sample size calculation: not calculated Country of origin: Italy Year of publication: 2017 Language of the original publication: English Category: treatment of MRONJ, surgical Funding: quote: "not provided" Registration in a public trials registry: not stated
Participants	36 participants with MRONJ randomised into 2 groups: 18 control (standard care (non‐AF)), 18 experimental (autofluorescence‐guided bone surgery (AF)) Mean age: non‐AF group 72.14 yr, AF‐group 72.3 yr Sex: non‐AF group 52% female, AF‐group 48% female Condition treated with bisphosphonates: osteoporosis (36% of total sample), cancer and other indications (64% of total sample) Inclusion criteria Exposed necrotic bone for longer than 8 weeks with history of antiresorptive drug treatment according to AAOMS Lesions resistant to drug therapy Exclusion criteria Radiation therapy in neck and head area Metastatic bone disease in jaws Contraindication to surgery
Interventions	Control (n = 18): standardised medical and surgical care. This included antibiotic therapy starting 3 days before surgery, professional oral hygiene, and antiseptic mouth rinse. Intraoperatively, the necrotic bone was resected according to the surgeon's assessment and his experience in recognising vital bone and resection margins. Tension‐free wound closure was achieved with absorbable sutures, and antibiotic treatment was continued postoperatively. Experimental (n = 18): standardised medical and surgical care as in the control group, but with autofluorescence‐guided surgery, i.e. natural autofluorescence of the healthy bone visualised with VELscope device to detect resection margins
Outcomes	Primary: Mucosal integrity 6 months (T3) after surgery Secondary: Mucosal integrity at 1 week (T1), 1 month (T2) and 1 year (T4) after surgery Absence of residual infection at 1 week (T1), 1 month (T2), 6 months (T3) and 1 year (T4) after surgery
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	High risk	Quote: "Patients of each stage were randomly assigned to 1 of 2 surgical groups by an independent resident senior."
Allocation concealment (selection bias)	Unclear risk	Allocation concealment not reported
Blinding of participants and personnel (performance bias) All outcomes	High risk	Blinding of participants and personnel not reported. Due to the nature of the intervention, surgeons were not blinded.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Blinding of outcome assessment not reported.
Incomplete outcome data (attrition bias) All outcomes	High risk	Six participants excluded: 2 died during follow‐up and 4 did not attend the 6‐month follow‐up visit.
Selective reporting (reporting bias)	Low risk	All outcomes reported

Giudice 2018b.

Study characteristics
Methods	Trial design: monocentric, prospective, single‐blind RCT Duration of study: enrolment period from November 2015 to December 2016 Follow‐up: per protocol 1 year Sample size calculation: not performed Country of origin: Italy Year of publication: 2018 Language of the original publication: English Category: treatment of MRONJ, surgical Funding: quote: "No financial support was received for this study." Registration in a public trials registry: not stated
Participants	47 participants with MRONJ randomised into 2 groups: 23 control (standard care (non‐PRF)), 24 experimental (standard care + platelet‐rich fibrin (PRF)) Mean age: non‐PRF group 73.9 yr, PRF group 74.7 yr Sex: non‐PRF group 58% female, PRF group 42% female Condition treated with bisphosphonates: osteoporosis (26% of total sample), cancer and other indications (74% of total sample) Inclusion criteria Exposed necrotic bone for longer than 8 weeks with a history of antiresorptive drug treatment according to the AAOMS Exclusion criteria Radiation therapy in neck and head area Metastatic bone disease in jaws Contraindication to surgery
Interventions	A combined antibiotic therapy with amoxicillin (1 g every 12 hours) and metronidazole (250 mg every 8 hours) for 10 days, beginning 3 days before surgery, a professional oral hygiene session and chlorhexidine 0.2% as a mouth rinse was performed in both groups. In case of hypersensitivity or allergy, clindamycin (600 mg 3 times a day) was received by participants. Control (n = 23): surgical removal of the necrotic bone Experimental (n = 24): PRF treatment in addition to surgical necrotic bone removal All participants received a tension‐free wound closure with resorbable sutures and obtained routine postoperative instructions and the same postoperative drug therapy as before surgery. To avoid mucosal damage, no prosthesis should be used.
Outcomes	Mucosal integrity: no presence of necrotic bone exposure Absence of residual infection: no presence of purulent exudate and reduction of swelling Presence of cutaneous fistulas Re‐intervention necessary to healing Reduction of pain: VAS score evaluation Outcomes evaluated preoperatively (T0) and at 1 month (T1), 6 months (T2), and 1 year (T3) after treatment Primary outcome Mucosal integrity 6 months (T2) after surgery
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote: "They were allocated to 2 groups according to their medication history: the first group included patients with bone metabolic disease being treated with a low‐dose (L) antiresorptive therapy (oral bisphosphonates, denosumab 60 mg/6 months) while the other group included patients with malignancy being treated with a high‐dose (H) antiresorptive therapy (intravenous bisphosphonates, denosumab 120 mg/4 weeks). Each group was divided in 2 subgroups by the same Senior Consultant (H2, L2 and H3, L3) according to the MRONJ stage diagnosed (stage 2 or 3)." "To create a homogeneous study population, an independent senior resident randomly assigned patients of each subgroup (H2, L2, H3, L3) to a specific surgical study group (PRF or non‐PRF)." Method of sequence generation not clearly described
Allocation concealment (selection bias)	Unclear risk	Allocation concealment not reported
Blinding of participants and personnel (performance bias) All outcomes	High risk	Quote: "The study was designed as a prospective randomized, single‐blind, monocentric clinical trial."
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Blinding of outcome assessors not reported
Incomplete outcome data (attrition bias) All outcomes	Low risk	Follow‐up completed for all participants
Selective reporting (reporting bias)	High risk	Quote: "Data and clinical outcomes were collected by an independent resident preoperatively (T0) and at T1, T2, and T3." However, preoperative pain VAS scores were not reported.

Mozzati 2012.

Study characteristics
Methods	Trial design: single‐centre, interventional, prospective, unblinded RCT Duration of study: January 2005 to December 2009 Follow‐up: between 24 and 60 months Sample size calculation: not provided Country of origin: Italy Year of publication: 2012 Language of the original publication: English Category: prophylaxis of MRONJ Funding: not reported Registration in a public trials registry: not stated
Participants	176 participants randomised into 2 groups: 85 control, 91 experimental (plasma rich in growth factors (PRGF)) Age 44‐60 yr: control 27, PRGF 22 Age 60‐70 yr: control 36, PRGF 43 Age 70‐83 yr: control 22, PRGF 26 Sex: control 54% female, PRGF 60% female Condition treated with intravenous (IV) bisphosphonate (zoledronic acid): breast cancer, prostate cancer, ovarian cancer, lung cancer, and multiple myeloma Inclusion criteria Current IV bisphosphonate therapy The necessity for removal of strongly compromised dental elements Exclusion criteria Previous history of irradiation to maxillofacial area Dental extractions before study period
Interventions	Control: no PRGF Experimental: PRGF fraction inserted into the postextraction alveolus All participants: professional oral hygiene session 1 week before surgery; antibiotics for 6 days starting the evening before surgery, anaesthesia by alveolar nerve block, no intraligamentous or intrapapillary infiltrations; mucosal flap and suturing to enable healing via primary intention
Outcomes	Postoperative bisphosphonate‐associated osteonecrosis (clinical signs of MRONJ: pain, swelling, non‐healing, exposed necrotic bone, and/or fistulas with connection to the bone) Follow‐up examinations: mucosal healing was monitored at 3, 7, and 14 days postoperatively; monitoring for MRONJ was continued at 21, 30, 60, 90, and 120 days, and 6 months, followed by visits every 6 months
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote: "The study cohort was divided  into two similar groups: 91 patients were treated with PRGF (study group) and 85 patients were not treated with the growth factor preparation (control group). The randomized group distribution was set up specifically to obtain groups that were homogenous for gender, age, smoking habits, systemic pathology based on the computerized clinical file we used in the first visit." Generation of randomisation sequence not reported
Allocation concealment (selection bias)	Unclear risk	Concealment of allocation not reported
Blinding of participants and personnel (performance bias) All outcomes	High risk	Due to the nature of the intervention, the personnel were not blinded. Because an extra 15 mL blood sample was obtained from the participants in the PRGF group, the participants were most likely not blinded.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	It is not reported whether outcome was monitored by an independent and blinded outcome assessor.
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Completeness or loss to follow‐up was not reported.
Selective reporting (reporting bias)	Low risk	The outcomes mentioned in the Methods were all reported.

Mozzati 2013.

Study characteristics
Methods	Trial design: single‐centre, interventional, prospective, unblinded, 2‐arm RCT Duration of study: January 2005 to April 2011 Follow‐up: between 12 and 72 months Sample size calculation: not provided Country of origin: Italy Year of publication: 2013 Language of the original publication: English Category: prophylaxis of MRONJ Funding: not reported Registration in a public trials registry: not stated
Participants	700 participants receiving oral bisphosphonates: tooth extractions were performed in 334 participants with protocol A (delicate surgery and closure by primary intention), in 366 participants with protocol B (non‐traumatic avulsion and closure by secondary intention) Age 50‐60 yr: protocol A 85, protocol B 93. Age 60‐70 yr: protocol A 185, protocol B 179. Age 70‐80 yr: protocol A 64, protocol B 94. Sex: protocol A 96% female, protocol B 98% female. Condition treated with bisphosphonates: osteoporosis, rheumatoid arthritis, Paget's disease Inclusion criteria Current oral bisphosphonate therapy Treatment with oral bisphosphonates for more than 24 months The necessity for the removal of compromised dental elements Exclusion criteria Any previous history of irradiation to the maxillofacial area Dental extractions before the study period
Interventions	All participants: professional oral hygiene session 1 week before surgery; antibiotics for 6 days starting the evening before surgery Protocol A (n = 334): surgical extractions carried out by intrasulcular incisions and detachment of full thickness flaps to allow wound healing via primary intention Protocol B (n = 366): extractions carried out without detachment of full thickness flaps; sockets filled with absorbable gelatin sponge haemostatic to allow wound healing via secondary intention
Outcomes	Intraoperative complications Success rate (absence of postoperative MRONJ) Follow‐up examinations: mucosal healing monitored at 3, 7, and 14 days postoperatively; monitoring for MRONJ continued at 21, 30, 60, and 90 days, and 6 months, followed by visits every 6 months
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "Each patient was assigned by a computer‐randomization program to one of two groups."
Allocation concealment (selection bias)	Unclear risk	Concealment of allocation not reported
Blinding of participants and personnel (performance bias) All outcomes	High risk	Due to the nature of the intervention, the personnel and participants were not blinded.
Blinding of outcome assessment (detection bias) All outcomes	High risk	At least during the mucosal healing period, due to the nature of the intervention, blinding of outcome assessors is not possible.
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Completeness or loss to follow‐up was not reported.
Selective reporting (reporting bias)	Low risk	The outcomes mentioned in the Methods were all reported.

Mücke 2016.

Study characteristics
Methods	Trial design: single‐centre, prospective, unblinded, 2‐arm RCT Duration of study: enrolment period from 2008 to 2014 Follow‐up: quote: "During this study patients of group A were evaluated 3.2 times (range 2‐4, at least 2 years) while group B was treated 6.8 times (range 4‐24, at least 1 year)" Sample size calculation: not provided Country of origin: Germany Year of publication: 2016 Language of the original publication: English Category: prophylaxis of MRONJ Funding: no extramural funding; quot e"The study was not funded" Registration in a public trials registry: not stated
Participants	253 men with prostate cancer with planned zoledronic acid for treatment of bone metastases were randomised into 2 groups: 127 in group A and 126 in group B Mean age: group A 69 yr, group B 72 yr Sex: male Condition treated with IV zoledronic acid: metastatic adenocarcinoma of the prostate with bone metastases Inclusion criteria Metastatic adenocarcinoma of the prostate with bone metastases Not yet treated with IV zoledronic acid Exclusion criteria Kidney failure (creatinine clearance < 30 mL/min)
Interventions	Group A (control) (n = 127): participants received an initial examination at the study centre and were monitored and treated where deemed necessary by the individual's dentist, and were re‐evaluated once a year Group B (experimental) (n = 126): participants received an initial examination and were treated if needed at the study centre. Participants were monitored and treated where necessary by the authors at 12‐week intervals. Extractions were performed under prophylactic antibiotic treatment and wound closure was carried out without tension on the local flap
Outcomes	Incidence rate per year for MRONJ Incidence proportion for MRONJ
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote: "All patients were prospectively examined before the start of therapy with zoledronic acid and were randomly allocated into two groups." Generation of randomisation sequence not reported
Allocation concealment (selection bias)	Unclear risk	Concealment of allocation not reported
Blinding of participants and personnel (performance bias) All outcomes	High risk	Quote: "treatment was not possible to be blinded"
Blinding of outcome assessment (detection bias) All outcomes	High risk	Not blinded
Incomplete outcome data (attrition bias) All outcomes	High risk	High and unbalanced rate of cross‐overs: quote: "36 patients, who were randomized to participate in group B did not want to be part of a close follow‐up and were then regrouped in group A." "At the end of this study, 153 (93.3%) patients of group A and 79 (87.8%) patients from group B have died." Data analysis: as‐treated, not by intention‐to‐treat
Selective reporting (reporting bias)	Low risk	The outcomes mentioned in the Methods were all reported.

Ohbayashi 2020.

Study characteristics
Methods	Trial design: monocentre, prospective, parallel‐group (2‐arm), randomised clinical pilot trial Duration of study: enrolment period from October 2011 to August 2017 Follow‐up: 6 months Sample size calculation: not reported Country of origin: Japan Year of publication: 2020 Language of the original publication: English Category: MRONJ treatment Funding: not stated Registration in a public trials registry: not stated
Participants	13 participants were randomly assigned to 2 groups: 7 to the weekly teriparatide (TPTD) group (1 dropout in this group, i.e. 6 were analysed), 6 to the daily TPTD group Median age: weekly TPTD group 82 yr (range 79‐86 yr), daily TPTD group 82 yr (range 52‐88 yr) Sex: weekly TPTD group 100% female, daily TPTD group 100% female Condition treated with bisphosphonates: osteoporosis (100% of total sample) Inclusion criteria Female, ≥ 50 years of age Not a cancer patient Required to continue osteoporosis treatment Stage ≥ 2 MRONJ according to the AAOMS criteria Provision of informed consent Exclusion criteria Hypercalcemic complication High risk of osteosarcoma Bone cancer Metabolic bone disease Pregnant or nursing Serious complications such as cancer, heart disease, liver disease, and kidney failure
Interventions	Weekly TPTD group (n = 7): weekly injections of teriparatide (56.5 μg), starting at first visit and continuing for 6 months Daily TPTD group (n = 6): daily injections of teriparatide (20 μg), starting at first visit and continuing for 6 months All participants were given MRONJ stage‐appropriate conventional treatment plus intensive antibiotic therapy as needed.
Outcomes	Change in MRONJ stage after treatment Changes in bone metabolism using bone scintigraphy The percentage of bone formation by using axial and coronal CT images of the maximum osteolysis area Bone turnover markers: osteocalcin, procollagen type I N‐terminal propeptide, bone‐specific alkaline phosphatases, urinary cross‐linked N‐telopeptide of type I collagen, urinary deoxypiridinolin, and tartrate‐resistant acid phosphatase‐5b
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote: "We randomly assigned the 13 patients" Generation of randomisation sequence not reported
Allocation concealment (selection bias)	Unclear risk	Concealment of allocation not reported
Blinding of participants and personnel (performance bias) All outcomes	High risk	Participants and personnel not blinded
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Blinding of outcome assessors not reported
Incomplete outcome data (attrition bias) All outcomes	High risk	One participant did not complete the study and was excluded from the analysis. Quote: "weekly group, n = 7 including 2 patients with RA, with 1 later dropout"
Selective reporting (reporting bias)	High risk	Outcomes were reported but with inadequate detail for the data to be included in a meta‐analysis. Authors present P values of statistical analyses but do not report effect estimates of outcomes such as improvement of MRONJ.

Park 2017.

Study characteristics
Methods	Trial design: prospective RCT Duration of study: enrolment period from 2012 to 2015 Follow‐up: 6 months Sample size calculation: not provided Country of origin: Korea Year of publication: 2016 Language of the original publication: English Category: treatment of MRONJ Funding: not reported Registration in a public trials registry: not stated
Participants	55 participants with MRONJ randomised into 2 groups: 25 control (L‐PRF alone), 30 experimental (L‐PRF and rhBMP‐2) Mean age: L‐PRF group 75.24 yr, L‐PRF + rhBMP‐2 group 75.2 yr Sex: L‐PRF group 43% female, L‐PRF + rhBMP‐2 group 57% female Condition treated with bisphosphonates: osteoporosis (87% of total sample), cancer and other indications (13% of total sample) Inclusion criteria Current or previous treatment with antiresorptive agents Exposed bone or bone or an intraoral/extraoral fistula in the maxillofacial region that has persisted for longer than 8 weeks Bony lesion with sequestra or necrotic bone in the jaw that needs to be surgically removed Exclusion criteria History of radiation therapy Metastatic disease to the jaws
Interventions	Antibiotic therapy with '1g of third‐generation cephalosporin given intravenously twice daily, analgesics, daily irrigation using 0.12% chlorhexidine, and professional dental prophylaxis during 1 week before surgery' was performed in both groups. Control (n = 25): L‐PRF treatment alone after surgical removal of the necrotic bone Experimental (n = 30): combined L‐PRF and rhBMP‐2 treatment in addition to surgical necrotic bone removal After performing intravenous sedation or general anaesthesia with local anaesthesia on patients with MRONJ, surgeons used surgical curettes to remove bony sequestra and granulation until flesh bleeding from the bone was confirmed. Sharp bony margins were smoothened and an antibiotic solution (2g of third generation cephalosporin in 1L saline) was applied to remove debris and foreign bodies and to reduce bacterial contamination. Before primary closure of the mucoperiosteal flap, L‐PRF was added to the bone surface of patients in the L‐PRF group. In the PRF plus BMP group a collagen sponge section with rhBMP‐2 was placed first and then L‐PRF was placed to the bony margins. By using a rhBMP‐2 commercial kit with a 0.5 mL rhBMP‐2 solution and hydroxylapatite, rhBMP‐2 solution was gained and applied to a collagen sponge and then used in patients with MRONJ.  All patients received intravenous antibiotics for 1 week postoperatively and oral antibiotics (third‐generation cephalosporin) for another 2 weeks.
Outcomes	Healing of MRONJ, defined as: complete: full mucosal coverage with absence of clinical or radiographical evidence of MRONJ delayed: MRONJ was present at 4 weeks but had resolved completely with full mucosal coverage by 16 weeks no resolution: persistence or progression of MRONJ at 16 weeks postoperatively Follow‐up examinations: mucosal healing was monitored weekly for the first month then monthly for 6 months.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote: "Patients were randomly assigned" Method of random sequence generation not described
Allocation concealment (selection bias)	Unclear risk	Allocation concealment not reported
Blinding of participants and personnel (performance bias) All outcomes	High risk	Blinding of participants and personnel not reported. Due to the nature of the intervention, surgeons were not blinded.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Blinding of outcome assessment not reported
Incomplete outcome data (attrition bias) All outcomes	Low risk	No dropouts or exclusions from the analysis
Selective reporting (reporting bias)	High risk	Results were not reported for all follow‐up time points.

Poxleitner 2020.

Study characteristics
Methods	Trial design: prospective, parallel‐group (2‐arm), randomised clinical trial Duration of study: enrolment period not reported Follow‐up: 90 days Sample size calculation: not reported Country of origin: Germany, Turkey Year of publication: 2020 Language of the original publication: English Category: MRONJ prevention Funding: quote: "University of Freiburg Faculty of Medicine Research Committee" Registration in a public trials registry: not stated
Participants	77 participants randomised into 2 groups: 39 group A (post‐extraction primary closure of the extraction socket with a mucoperiosteal flap), 38 group B (insertion of a PRF clot into the extraction socket without subsequent primary closure) Median age: group A 77 yr (range 44‐88 yr), group B 78 yr (range 53‐87 yr) Sex: group A 100% female, group B 97% female Condition treated with bisphosphonates or denosumab: osteoporosis (100% of total sample) Inclusion criteria Diagnosis of osteoporosis Current or previous antiresorptive therapy (bisphosphonates or denosumab) for osteoporosis Indication for extraction of one or more teeth with a hopeless prognosis Exclusion criteria History of irradiation to the maxillofacial region or neoplastic involvement/disease of the maxillofacial region
Interventions	Group A (n = 39): post‐extraction primary closure of the extraction socket with a mucoperiosteal flap Group B (n = 38): insertion of a PRF clot into the extraction socket without subsequent primary closure Bisphosphonate or denosumab drug holiday prior to the tooth extraction was not required. All patients received perioperative intravenous antibiotic therapy (penicillin 10, 000,000 IU once daily or clindamycin 600 mg three times daily in case of penicillin allergy), initiated 1 day before surgery and continued until 1 day after surgery. Prior to the extraction, all patients rinsed their mouth with chlorhexidine solution. Postoperatively, patients were instructed to consume a soft diet, to apply daily mouth rinses with chlorhexidine solution, and to refrain from wearing dentures until complete mucosal healing was achieved.
Outcomes	Primary Complete mucosal coverage at extraction site Secondary Intraoperative complications Postoperative complications (Clavien‐Dindo classification)
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote: "Assignment to one of the two groups was performed via block randomization with randomly selected block sizes." The process of selecting the blocks, such as a random number table or a computer random number generator, was not specified.
Allocation concealment (selection bias)	Unclear risk	Allocation concealment not reported
Blinding of participants and personnel (performance bias) All outcomes	High risk	Due to the nature of the intervention, blinding was not possible.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Blinding of outcome assessors was not reported.
Incomplete outcome data (attrition bias) All outcomes	Low risk	No dropouts or exclusions from the analysis
Selective reporting (reporting bias)	Low risk	No selective reporting

Ristow 2016.

Study characteristics
Methods	Trial design: single‐centre, prospective, unblinded, 2‐arm RCT Duration of study: quote: "patients [...] were recruited over a time period of 12 months and followed up for 12 months" Follow‐up: one year Sample size calculation: not provided; quote: "Because of the preliminary ‘proof‐of‐concept’ character of this study, the sample size estimation was disclaimed" Country of origin: Germany Year of publication: 2016 Language of the original publication: English Category: treatment of established MRONJ Funding: quote: "There was no source of funding for this research" Registration in a public trials registry: not stated
Participants	40 participants with MRONJ randomised into 2 groups: 20 control (tetracycline fluorescence‐guided bone surgery (TF)), 20 experimental (autofluorescence‐guided bone surgery (AF)) Mean age: control 67 yr, AF 71 yr Sex: control 69% female, AF 70% female Condition treated with antiresorptive medication: cancer (85% of total sample), osteoporosis (15% of total sample) Inclusion criteria History of antiresorptive drug treatment (bisphosphonates or denosumab, or both) in the absence of radiotherapy to head and neck region Exposed osteonecrosis of the jaw, defined as long‐standing (more than 8 weeks) transmucosal exposure of necrotic bone in the jaw Exclusion criteria History of head and neck irradiation Metastatic bone disease of maxillofacial region Contradictions to surgery under general anaesthesia
Interventions	Control (TF group) (n = 20): participants received 100 mg doxycycline twice a day for at least 7 days preoperatively. Incorporation of doxycycline into vital bone and absence of doxycycline in necrotic bone was detected by a fluorescent light source (VELscope fluorescence lamp; LED Dental, White Rock, British Columbia, Canada). Doxycycline fluorescence was used for intraoperative identification of bone resection margins and guided debridement of necrotic bone. AF group (n = 20): participants received antibiotic prophylaxis with ampicillin/sulbactam 2000 mg/1000 mg (or clindamycin 600 mg in case of hypersensitivity to penicillin or a penicillin allergy) before operation. Autofluorescence of vital bone, induced with the VELscope fluorescence lamp (LED Dental, White Rock, British Columbia, Canada) was used for intraoperative identification of bone resection margins and guided debridement of necrotic bone. In all participants, a tension‐free wound closure was achieved using mucoperiosteal flaps. All participants remained in hospital for 4 days after the operation. Participants received routine postoperative instructions and the same postoperative analgesic drug therapy. Antibiotic treatment involved the administration of ampicillin/sulbactam 2000 mg/1000 mg (or clindamycin in case of hypersensitivity to penicillin or a penicillin allergy) intravenously while in hospital and then orally for a further 6 days after discharge from the hospital.
Outcomes	Primary Success rate: absence of a MRONJ site (i.e. maintenance of full mucosal coverage) at 8 weeks (T2) after surgery Secondary Mucosal integrity at 10 days (T1), 6 months (T3), and 1 year (T4) after surgery Loss of sensitivity (numbness) of the alveolar nerve (Vincent sign) at 10 days (T1), 8 weeks (T2), 6 months (T3), and 1 year (T4) after surgery Subjective pain at 10 days (T1), 8 weeks (T2), 6 months (T3), and 1 year (T4) after surgery Signs of infection at 10 days (T1), 8 weeks (T2), 6 months (T3), and 1 year (T4) after surgery
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote: "Over a period of 12 months, the study population was prospectively referred for the treatment of MRONJ and divided randomly into two study groups" Generation of randomisation sequence not reported
Allocation concealment (selection bias)	Unclear risk	Concealment of allocation not reported
Blinding of participants and personnel (performance bias) All outcomes	High risk	Not blinded
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not reported
Incomplete outcome data (attrition bias) All outcomes	Low risk	4 participants (2 each in the TF and the AF group) died after T2 (8 weeks after the operation) and 2 participants in the AF group failed to attend the 1‐year follow‐up (T4). No participant was lost for assessment of the primary endpoint at T2 (8 weeks after the operation).
Selective reporting (reporting bias)	Low risk	The outcomes mentioned in the Methods were all reported.

Ristow 2020.

Study characteristics
Methods	Trial design: single‐centre, prospective, parallel‐group (two‐arm), single‐blind (patient/subject‐blind and assessor‐blind, randomised clinical pilot trial Duration of study: enrolment period from  April 2016 to April 2018 Follow‐up: at least 6 months Sample size calculation: quote: "Because of the preliminary “proof‐of‐concept” character of this study, no formal sample size calculation was performed." Country of origin: Germany Year of publication: 2020 Language of the original publication: English Category: MRONJ prevention Funding: institutional budget, no external funding Registration in a public trials registry: German Clinical Trials Register (DRKS00010106)
Participants	160 participants randomised into 2 groups: 78 active control (epiperiosteal wound closure), 82 experimental (subperiosteal wound closure) Median age: active control group 67.8 yr, experimental group 68.4 yr Sex: active control group 77% female, experimental group70% female Condition treated with bisphosphonates or denosumab for more than three years: malignant disease with bone metastasis or multiple myeloma (57% of total sample); osteoporosis (43% of total sample) Inclusion criteria Patients with malignant disease or osteoporosis and antiresorptive medication  who are scheduled for preventive tooth extraction Minimum age 18 years Exclusion criteria Previous history of a radiation to the head neck region Known metastasis of the jaw Exposed bone in the operation field Existing diagnosis MRONJ in the operation field
Interventions	All patients were pretreated with oral antibiotics from the week before surgery until one week after surgery. All patients used an antimicrobial mouth rinse with chlorhexidine three times daily, starting 2 days before surgery and for at least 5 days after surgery. If a "drug holiday" was possible from an oncologic/osteologic perspective, antiresorptive therapy was discontinued 1 month before surgery until 1 month after surgery. Control (n = 78): epiperiosteal wound closure after tooth extraction Experimental (n = 82): subperiosteal wound closure after tooth extraction
Outcomes	Primary Absence of an MRONJ site after tooth extractions, specified as the maintenance of full mucosal coverage Secondary Histopathologic evaluation of alveolar bone samples for the detection of early necrotic bone changes at the time of surgical tooth extraction Presurgical subjective pain (measured as pain “yes/no” and a VAS scale) Presurgical infection signs (defined as swelling and/or redness and/or purulent discharge from the tooth socket) The period of antiresorptive application at the time point of surgery Method of extraction (single extraction (one tooth), multiple extractions (>1 tooth), extractions with interruption) Mucosal integrity at the remaining measurement points Demographics and baseline characteristics for reassurance of group comparability
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "An independent statistical consultant was employed to computer generate the randomization sequence via a centralized web‐based tool (www.randomizer.at) prior to the start of the study. This generated a pseudorandom code with permuted blocks of randomly variable size."
Allocation concealment (selection bias)	Low risk	Quote: "The sequence was known only to the programmer until the database lock. No one directly involved in the project had access to the allocation codes."  Quote: "the allocation of the patients to the treatment groups was performed by the surgeons by means of sealed envelopes immediately before surgery."
Blinding of participants and personnel (performance bias) All outcomes	High risk	Quote: "patients were not informed about their allocation" Surgeons were not blinded.
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Quote: "The physicians and research nurses who carried out the postoperative follow‐ups and assessed the outcomes and the statistician were all blinded to treatment allocation during the entire study."
Incomplete outcome data (attrition bias) All outcomes	High risk	Several dropouts during follow‐up. 8 participants allocated to EPP did not receive allocated intervention but switched to SPP.
Selective reporting (reporting bias)	Low risk	The outcomes mentioned in the Methods were all reported.

Sim 2020.

Study characteristics
Methods	Trial design: multicentre, prospective, double‐blind RCT Duration of study: enrolment period from November 2012 to May 2015 Follow‐up: per protocol 1 year Sample size calculation: the trial was designed to recruit 68 participants to allow 80% power to detect a significant difference between groups. Country of origin: Australia Year of publication: 2020 Language of the original publication: English Category: treatment of MRONJ, pharmacological Funding: funding source government body Registration in a public trials registry: Australian New Zealand Clinical Trials Registry ACTRN12612000950864
Participants	34 participants with MRONJ randomised into 2 groups: 19 control (placebo), 15 experimental (teriparatide) Median age: placebo group 64 yr, teriparatide 64 yr Sex: placebo group 47% female, teriparatide group 47% female Condition treated with bisphosphonates:  malignant bone disease, myeloma, breast cancer, and prostate cancer (79% of total sample); osteoporosis (21% of total sample) Inclusion criteria Osteonecrosis of the jaw Previous/current treatment with either bisphosphonates or denosumab Minimum age 18 years Exclusion criteria Previous craniofacial radiotherapy Pregnancy Hypercalcaemia or pre‐existing primary hyperparathyroidism Severe renal impairment (eGFR < 30) Known metabolic bone disease, excluding osteoporosis or metastatic bone disease Growth hormone deficiency Secondary hyperparathyroidism with PTH greater than twofold above upper limit of reference range
Interventions	Control (n = 19): placebo saline injections, plus calcium (600 mg tablet daily) and vitamin D (1000 IU tablet daily) supplementation for 8 weeks Experimental (n = 15): subcutaneous teriparatide injections (20 micrograms daily), plus calcium (600 mg tablet daily) and vitamin D (1000 IU tablet daily) supplementation for 8 weeks
Outcomes	Primary Clinical staging of osteonecrosis of the jaw ‐ described by the American Association of Oral and Maxillofacial Surgeons position paper Radiological staging of osteonecrosis of the jaw, as assessed by cone beam CT. Staging system as described by Bianchi et al (2007) Secondary Bone formation and resorption markers (P1NP, beta‐CTX) Jaw osteoblast activity, as measured by NaF‐PET imaging Quality of life, as measured by Oral Health Impact Profile 14 questionnaire
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote: "Random assignment was performed in blocks and was stratified according to duration of MRONJ diagnosis: < 12 months or ≥ 12 months" The process of selecting the blocks, such as a random number table or a computer random number generator, was not specified.
Allocation concealment (selection bias)	Unclear risk	Concealment of allocation not reported
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Quote: "we conducted a prospective doubleblind, placebo‐controlled, randomized trial"
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Quote: "we conducted a prospective doubleblind, placebo‐controlled, randomized trial"
Incomplete outcome data (attrition bias) All outcomes	High risk	In the placebo group, 1 participant was lost to follow‐up and 1 participant died during follow‐up.  Quote: Quote: "All analyses were based on the intention‐to‐treat principle". Although 19 participants were assigned to the placebo group, only 18 placebo participants were included in analyses.
Selective reporting (reporting bias)	High risk	Not all outcomes were reported; for example, quality of life survey data were missing.

Yüce 2021.

Study characteristics
Methods	Trial design: RCT Duration of study: enrolment period from May 2016 to April 2018 Follow‐up: per protocol 6 months Sample size calculation: 8 patients per treatment group were calculated to provide 80% power. 14 patients per treatment group were included to compensate for dropouts. Country of origin: Turkey Year of publication: 2020 Language of the original publication: English Category: treatment of MRONJ, pharmacological Funding: not reported Registration in a public trials registry: ClinicalTrials.gov (NCT04531800)
Participants	28 elderly female osteoporotic participants were randomised into 2 groups: 14 in CGF group (CGF clots + primarily closure), 14 in control group (without CGF placement + primarily closure ) Median age: CGF group 73 yr; control group 73 year Sex: female Condition treated with bisphosphonates: osteoporosis Inclusion criteria Treatment with oral bisphosphonates (BPs) for osteoporosis MRONJ diagnosis with exposed bone in the jaws that had persisted for longer than 8 weeks  MRONJ stage 2 or 3 with bone destruction and sequestrum confirmed by clinical and radiographic examination Insufficient improvement with conservative treatment Exclusion criteria History of a radiation to the head neck region Metastatic bone disease of the jaws Platelet values under than 150,000 mm3 in a complete blood count
Interventions	CGF, like platelet‐rich plasma or platelet‐rich fibrin, is an autologous preparation obtained from the patient's blood immediately before surgery. All participants underwent removal of necrotic bone and primary closure of the surgical site. Control (n = 14): treatment without concentrated growth factor (CGF) + primary wound closure Experimental (n = 14): CGF clots placement + primary wound closure
Outcomes	Primary outcome Soft tissue healing 6 months postoperatively
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Method of sequence generation not reported
Allocation concealment (selection bias)	Low risk	Quote: "The sequentially numbered sealed envelopes were used by the examiner (G.I.) to randomly assign patients to the study groups. Before the surgery, these sealed envelopes were opened by the same examiner, and each patient was assigned to one of the study groups."
Blinding of participants and personnel (performance bias) All outcomes	High risk	Blinding of participants were not reported. Surgeons were not blinded.
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Quote: "Preoperative and postoperative data were recorded by the blinded examiner (E.A.), who did not participate in randomization or surgical procedures."
Incomplete outcome data (attrition bias) All outcomes	Low risk	The study was completed with all participants.
Selective reporting (reporting bias)	Low risk	The outcomes mentioned in the Methods were all reported.

AAOMS: Association of Oral and Maxillofacial Surgeons; CT: computed tomography;eGFR: estimated glomerular filtration rate; EPP: epiperiosteal prepared; g: gram;IU: international unit; IV: intravenous; mg: milligram;MRONJ: medication‐related osteonecrosis of the jaw;PET: positron emission tomography; PRF: platelet‐rich fibrin;PRGF: plasma rich in growth factors; PTH: parathyroid hormone; RCT: randomised controlled trial; SPP: subperiosteal prepared; VAS: visual analogue scale; yr: year(s)

Characteristics of excluded studies [author‐defined order]

Study	Reason for exclusion
Asaka 2017	Not a RCT
Bonacina 2011	Not a RCT
Bramati 2015	Not a RCT
Coviello 2012	Not a RCT
DE Iuliis 2014	Not a RCT
Dimopoulos 2009	Not a RCT
Lee 2014	Not a RCT
Montebugnoli 2007	Not a RCT
Pelaz 2014	Not a RCT
Vescovi 2010	Not a RCT
Vescovi 2012	Not a RCT
Garcia‐Martinez 2017	Not a RCT
Jung 2017	Not a RCT
Rodriguez 2019	Not a RCT
Calvani 2018	Not a RCT
Colapinto 2018	Not a RCT
Giovannacci 2016	Not a RCT
Szentpeteri 2020	Not a RCT
Watanabe 2021	Not a RCT
Nica 2021	Not a RCT

RCT: randomised clinical trial

Characteristics of ongoing studies [ordered by study ID]

ChiCTR1900027382.

Study name	Clinical study of modified curettage for medication‐related osteonecrosis of the jaw
Methods	Interventional, randomised, parallel trial
Participants	Target sample size: 50 participants Inclusion criteria According to the MRONJ diagnosis standard; stratified into stage II General body condition can tolerate surgery Agree to receive curettage surgery Sign informed consent Exclusion criteria Preoperative diagnosis as MRONJ, but the postoperative specimen was diagnosed as metastatic tumour Follow‐up of the patient was lost after surgery
Interventions	Control group: curettage Exprimental group: curettage plus bone tunnel preparation
Outcomes	Primary outcome Status evaluation of bone exposure or mucosa/skin fistula  Bony wound healing  Secondary outcome Quality of life (University of Washington Quality of Life Questionnaire UW‐QOL  V4.0)
Starting date
Contact information	Peking University Hospital of Stomotology, 22 Zhongguancun South Street, Haidian District, Beijing, China gladiater1984@163.com
Notes

DRKS00012888.

Study name	Treatment strategies for medication‐related osteonecrosis of the jaw ‐ a prospective, partially randomized patient preference trial
Methods	Interventional, randomised, open, monocentre trial
Participants	Target sample size: 90 participants Both sexes eligible for study, 18 years and older Inclusion criteria Not infected medication‐related osteonecrosis of the jaw (stadium I corresponding to AAOMS classification) Patients who received or are receiving antiresorptive therapy (bisphosphonates or RANKL antibody) Planned for treatment at clinic for craniofacial surgery of university hospital in Heidelberg Exclusion criteria Previous history of a radiation to the head neck region Known metastasis of the jaw
Interventions	Group 1 Surgical treatment of medication‐related osteonecrosis of the jaw Careful but complete fluorescence‐guided resection of necrotic bone Smoothing of sharp bone ridges after the removal of the necrotic bone Secure and sufficiently vascularized plastic coverage (optionally with the mylohyoid muscle or the Bichat’s fat pad) Systemic intravenous antibiotic treatment Change of diet type: liquid/ strained food or temporary oral bypassing by using gastric feeding tubes Group 2 Non‐surgical treatment of medication‐related osteonecrosis of the jaw Close follow‐ups Regular cleaning of the necrotic area with chlorhexidine‐gluconate (0.2 %) containing mouth rinsing solution and topical gel Oral antibiotic treatment (Unacid PD 375 mg 1‐0‐1 or alternatively Moxifloxacin 400 mg 1‐0‐0) in case of appearing infection signs until the inflammation has subsided (usually within a period of 10 days)
Outcomes	Primary outcome Mucosal healing 12 weeks after beginning of intervention: evaluated clinically (visually and probing with special periodontal probe)) Secondary outcome Mucosal integrity after treatment with/without infection signs (redness, swelling, pus) Health‐related quality of life using EORTC QLQ‐C30 and disease‐specific modules for bone metastases as well as OHIP‐G 14 questionnaire for oral health‐related quality of life Pain assessment Period till the resumption of oncological and/or osteological antiresorptive therapy Period till prosthetic rehabilitation Radiologic signs for medication‐related osteonecrosis of the jaw All secondary outcomes measured at 10‐14 days (T1), 30 days (T2), 8 weeks (T3), 3 months (T4), 6 months (T5), 1 year (T6) postoperatively
Starting date	Date of first enrolment:  28 May 2018
Contact information	oliver.ristow@med.uni‐heidelberg.de
Notes

jRCTs071200006.

Study name	Investigation of the preventive effects of antimicrobial penetrated collagen plug (TERUPLUG) for post‐extraction tooth socket for osteonecrosis of the jaw after tooth extraction in patients using high‐dose antiresorptive agent
Methods	Recruitment status: recruiting Health condition(s) or problem(s) studied: patients receiving high‐dose antiresorptive agent Countries of recruitment: Japan Study type: interventional
Participants	Target sample size: 30 Inclusion criteria Age minimum: 20 years Age maximum: not applicable Sex: both Patients receiving ZOMETA or RANMARK therapy who are scheduled to undergo tooth extraction Exclusion criteria Patients with "problem in judgement"
Interventions	Intervention group: insert of antimicrobial (MINOCYCLINE) penetrated collagen plug (TERUPLUG) for post‐extraction tooth socket, close the wound, and remove sutures 1 week later Control group: insert of collagen plug (TERUPLUG) for post‐extraction tooth socket, close the wound, and remove sutures 1 week later
Outcomes	Primary outcome  Incidence of MRONJ 2 months after tooth extraction Secondary outcome  Association of investigation factors and onset of MRONJ
Starting date	Date of first enrolment 2 July 2020
Contact information	Name  Sakiko  Soutome Address  1‐7‐1, Sakamoto, Nagasaki Nagasaki Japan 852‐8511 Telephone  +81‐95‐819‐7698 E‐mail  sakiko@nagasaki‐u.ac.jp Affiliation  Nagasaki University Hospital
Notes

NCT01526915.

Study name	Assessment of platelet rich fibrin efficiency on healing delay and on jawbone osteochemonecrosis provoked by bisphosphonates (OCN/PRF)
Methods	Interventional, randomised, parallel assignment, open‐label
Participants	270 participants are required to validate the expected objectives in this study All sexes eligible for study, 18 years and older Inclusion criteria Adults (male or female) Documented indication at the initial visit at day 0 (JO) for a maximum extraction of 3 teeth Treatment with nitrogenous or non‐nitrogenous bisphosphonate (BP) by intravenous injection or oral administration whatever the reason for this drug prescription: ongoing BP treatment individual having received a previous treatment with BPs (irrespective of the duration and withdrawal date of this treatment Individual having received the specific information letter regarding the study and having signed the clarified consent form Exclusion criteria Individual having a maxillary or mandibulary OCN3 at day 0 (JO) Positive HIV serology at Day 0 (for participants belonging to the platelet‐rich fibrin (PRF) group) Previous history of maxillo‐cervico‐facial radiotherapy Individual with estimated survival expectancy shorter than one year Lack of social security cover Inability of the individual to respect the study follow‐up Individual having reached his/her majority and under tutelage, trusteeship or protection of the court Individual whose diagnosis could not be revealed to him/her (especially when the individual or the family expressed this wish)
Interventions	Experimental: bone curettage + PRF insertion Control: bone curettage alone without PRF insertion
Outcomes	Primary outcome measures Delay in cicatrisation4 at week 8 The appearance of osteochemonecrosis during the follow‐up period Secondary outcome measures The characteristics of the received BP treatment: starting date of ongoing treatment, accumulated dose, type of BP, administration route The precise location of the extraction site according to the tooth classification number
Starting date	September 2011
Contact information	e.gerard@chr‐metz‐thionville.fr
Notes

NCT02198001.

Study name	Prospective randomized study: assessment of PRF efficacy in prevention of jaw osteonecrosis after tooth extraction (PRF)
Methods	Interventional; randomised; parallel assignment, blinded
Participants	Cohort of 100 participants: control group 50 participants and experimental 50 participants All sexes eligible for study; 50 years and older Inclusion criteria Individuals taking bisphosphonates whatever the indication, the type, the administration and the duration of treatment (we include those taking or having taken bisphosphonates, even several years ago) Individuals who need tooth extraction (not recoverable in conservative dentistry and symptomatic tooth: dental and periodontal infections, symptomatic traumatic tooth fracture Exclusion criteria Pregnant women Younger than 50 years old Jaw radiotherapy History of jaw osteonecrosis Jaw metastasis from another cancer
Interventions	Experimental: tooth extraction and insertion of PRF (non‐traumatic tooth extraction with antibiotics (amoxicillin clavulanate combination). Insertion of PRF membrane in tooth‐extraction site) Control: no PRF (non‐traumatic extraction with antibiotic without PRF insertion)
Outcomes	Number of participants with jaw osteonecrosis after tooth extraction
Starting date	January 2014
Contact information	dorothee.deneubourg@uclouvain.be, michele.magremanne@uclouvain.be
Notes

NCT03040778.

Study name	Pentoxifylline and Tocopherol (PENTO) in the treatment of Medication‐related Osteonecrosis of the Jaw (MRONJ)
Methods	Interventional, randomised, parallel, blinded
Participants	Target sample size: 100 participants All sexes eligible for study, 18 years and older Inclusion criteria Stage 1, 2, or 3 MRONJ as defined by the AAOMS Position Paper on Medication‐Related Osteonecrosis of the Jaw‐2014 Update (Ruggiero 2014) History of exposure to antiresorptive medications such as bisphosphonates or RANK‐L inhibitors Absence of tumour in the jaw at the time of recruitment Patients with the capacity to give informed consent Exclusion criteria Patients with history of external radiation therapy to the jaws Patients who underwent any surgical intervention for MRONJ in the past 4 months Patients with past microvascular reconstruction of the head and neck Patients with an expected survival less than 1 year Patients with allergy or hypersensitivity to pentoxifylline, xanthines, or tocopherol Patients with planned invasive dental procedure in the next year Patients taking oral anticoagulants Patients with known hemorrhagic and coagulation disorder Patients with a vitamin K deficiency due to any cause Female patients who are pregnant or lactating Patients with history of serious bleeding or extensive retinal haemorrhage Patients with ischaemic heart diseases, including, but not limiting, recent myocardial infarction Patients with serious cardiac arrhythmia Patient with history of prostate cancer Patients with severe liver disease Patients with severe renal failure (Creatinine clearance <30 mL/min) Patients with diagnosed hypotension Patients taking CYP1A2 inhibitors (e.g. ciprofloxacin, fluvoxamine) Diagnosis of MRONJ with no exposed bone Patient cannot tolerate impressions of exposed bone in a clinical setting, if needed There is a change in the patient's clinical presentation (tooth extraction, sequestrectomy) from alginate impression, if impression is indicated Any other situation or condition that, in the opinion of the INVESTIGATOR, may interfere with optimal PARTICIPATION in the study A patient who has taken both bisphosphonate and Denosumab
Interventions	Experimental: Standard of care + PENTO Control: Standard of care + placebo (placebo tablets)
Outcomes	Primary outcome: Change in bone exposure area (mm2) Secondary outcome: Change in MRONJ stage Change in pain Change in osseous anterior‐posterior linear dimension on orthopantomogram Change in osseous superior‐inferior linear dimension on orthopantomogram Change in osseous area on orthopantomogram Time frame for all primary and secondary outcomes are 0 months, 1 month, 3 months, 6 months, 9 months, 12 months
Starting date	April 1, 2018
Contact information
Notes	Jasjit Dillon, University of Washington

NCT03269214.

Study name	Use of topical phenytoin in bisphosphonate‐related osteonecrosis of the mandible
Methods	Interventional, randomised, parallel, blinded
Participants	Target sample size: 20 participants All sexes eligible for study, 18 years and older Inclusion criteria Criteria of bisphosphonate‐related osteonecrosis in stage II Need debridement and surgical intervention Exclusion criteria Malignancy in the area History of chemotherapy Diabetic mellitus HIV Odontogenic infection or undergo dialysis
Interventions	Experimental Debridement of necrotic bone + topical phenytoin 5% + tetracycline before final primary closure Control Debridement of necrotic bone and primary closure
Outcomes	Primary outcome Soft tissue healing after one month, six months and 12 months Pain after one month and six months Secondray outcomes Infection after one month, six months and 12 months
Starting date	September 1, 2012
Contact information	Reza Tabrizi, Shiraz University of Medical Sciences
Notes

NCT03390777.

Study name	Plasma rich in growth factors for treatment of medication related osteonecrosis of the jaw
Methods	Interventional, randomised, parallel, blinded
Participants	Target sample size: 150 participants All sexes eligible for study, 18 years or older Inclusion criteria Candidates for surgical treatment for MRONJ at stage 2 or 3 Age >18 and < 80 years Have signed a informed consent prior to randomisation and must agree to return to scheduled follow‐up visits Exclusion criteria Inability to understand and cooperate with the study procedures or provide informed consent Bleeding diathesis or coagulapathy, or will refuse autologous blood sampling Cardiovascular event in the past 30 days Any condition that limits anticipated survival to less than 3 months
Interventions	Experimental Surgery (debridement/removal of affected tissue) + PRGF Control Surgery only (debridement/removal of affected tissue)
Outcomes	Primary outcome Recurrence of disease (time frame: 12 months) Secondary outcome Morbidity (time frame: 12 months) Post‐ and peri‐operative pain (time frame: 1 week) Quality of life (QoL) (time frame: 12 months)
Starting date	September 2018
Contact information	Oreste Iocca, University of Roma La Sapienza
Notes

NCT04512638.

Study name	Best Treatment Choice for Osteonecrosis of the jaw (BETCON)
Methods	Multicentre, interventional, phase 4, randomised, parallel, open‐label trial
Participants	Target sample size: 125 participants All sexes eligible for study, 18 years or older Inclusion criteria >18 years of age Provision of signed informed consent A history of at least one administration of, or an ongoing treatment with, a bone modifying agent in dose registered for the prevention of skeletal related events in bone metastatic disease or multiple myeloma Diagnosis of stage I‐II MRONJ according to AAOMS 2014 criteria not more than 8 weeks prior to the date of screening Exclusion criteria Any prior treatment for MRONJ other than local antiseptic rinses, systemic antibiotics, or analgesics Prior radiotherapy to the head and neck region Medical contraindication to receive any of the possible study treatments Stage III MRONJ characterised by very extensive bone necrosis, pathological fracture, or fistulas to the skin or sinuses Multiple MRONJ lesions that cannot be closed in a single surgical procedure
Interventions	Active comparator: conservative treatment   Amoxicillin‐based antibiotics and chlorhexidine oral rinse. Minor debridement. Primary wound closure is not part of this treatment strategy.   Interventions        Drug: antibiotics        Drug: chlorhexidine mouthwash Experimental: minimally invasive approach + LPRF amoxicillin‐based antibiotics and chlorhexidine oral rinse. Minimally‐invasive surgical treatment, including sequestrectomy, debridement of soft tissue, and application of LPRF membranes before tension‐free wound closure is obtained. Marginal resection of all necrotic bone is not part of this treatment strategy. Interventions:        Drug: antibiotics        Drug: chlorhexidine mouthwash Procedure: minimally invasive surgery with LPRF  Experimental: primary surgical management.   Amoxicillin‐based antibiotics and chlorhexidine oral rinse. Removal of the necrotic bone without excessive resection of healthy bone. Buccal mucoperiosteal flaps will be used to achieve a tension‐free mucosal coverage.   Interventions:        Drug: antibiotics        Drug: chlorhexidine mouthwash Procedure: surgical resection
Outcomes	Primary outcome Time to confirmed mucosal healing Secondary outcome Mucosal closure  Time to MRONJ healing  Relapse rate of MRONJ  Antibiotics use  Evolution of cancer health‐related quality‐of‐life (EORTC QLQ‐C30 questionnaire). Evolution of general health status (changes over time as measured with the EUROQOL 5D questionnaire) Evolution of oral health‐specific quality‐of‐life (measured with the Oral Health Impacts Profile  and the SWOG0702 Oral Health and Oral Health‐related Quality of Life questionnaires)
Starting date	October 1, 2020
Contact information	Tim Van den Wyngaert, University Hospital, Antwerp
Notes

NCT04531800.

Study name	Surgical management of medication related osteonecrosis of the jaws with concentrated growth factor
Methods	Interventional, randomised
Participants	28 female participants; 65 to 85 years Inclusion criteria Treatment with oral bisphosphonates (BPs) for osteoporosis MRONJ diagnosis with exposed bone in the jaws that had persisted for longer than 8 weeks according to 2014 recommendations of the Association of Oral and Maxillofacial Surgeons (AAOMS) MRONJ stage 2 or 3 with bone destruction and sequestrum confirmed by clinical and radiographic examination Insufficient improvement with conservative treatment Exclusion criteria A history of head and neck radiation therapy Metastatic bone disease of the jaws Platelet values under than 150.000 mm3 in a complete blood count
Interventions	Concentrated Growth Factor (CGF) + primarily closure vs no CGF + primarily closure
Outcomes	Soft tissue healing 6 months postoperatively
Starting date	May 1, 2016
Contact information	Principal Investigator: Gözde Işık, Lecturer, Ege University, Izmir, Turkey
Notes

UMIN000009132.

Study name	Study to the effect of teriparatide formulation Forteo versus Teribone on bisphosphonate‐related osteonecrosis of the jaw in osteoporosis patients
Methods	Interventional, parallel, randomised, open study
Participants	Target sample size: 15 female participants >= 20 years of age Inclusion criteria Individuals who require continued treatment for osteoporosis Females with bisphosphonate‐related osteonecrosis of the jaw Bisphosphonate‐related osteonecrosis of the jaw stage 2 or more Outpatients Signed informed consent forms obtained Exclusion criteria Hypercalcaemic disorders Potential risk of osteosarcoma Individuals with Paget's disease of bone Unexplained elevations of alkaline phosphatase Young adults with open epiphyses Individuals with prior external beam or implant radiation involving the skeleton Individuals with bone metastases, history of skeletal malignancies Metabolic bone diseases other than osteoporosis Pregnancy or women with suspected pregnancy Individuals with hypersensitivity to teriparatide or to any of its excipients Serious cardiac disease, serious hepatic disorder, renal disease Use of active vitamin D3 or digoxin Individuals who could not provide informed consent Unsuitability for the trial based on clinical judgement
Interventions	Forteo (teriparatide) vs Teribone (teriparatide)
Outcomes	Pain Bone formation
Starting date	August 2012
Contact information	yumiko@med.kagawa‐u.ac.jp
Notes

UMIN000042862.

Study name	The clinical effect of teriparatide on BRONJ (bisphosphonate‐related osteonecrosis of the jaw)
Methods	Interventional, parallel, randomised, open study
Participants	Target sample size: 20 participants All sexes eligible for study, 20 years or older Inclusion criteria BRONJ Exclusion criteria Patients with hypercalcaemia Patients who are considered to be at high risk of developing osteosarcoma as follows Patients with Paget's disease of bone Patients with unexplained high alkaline phosphatase levels Young patients whose epiphyseal line has not yet closed Patients who have received radiation therapy that may affect bone in the past. Patients with primary malignant bone tumours or metastatic bone tumours Patients with metabolic bone diseases other than osteoporosis (hyperparathyroidism, etc.) Pregnant women, women who may become pregnant, and lactating women Patients with a history of hypersensitivity to teriparatide components or teriparatide acetate Patients with serious complications such as cancer, cardiac disease, liver disease, and renal disorder Patients currently using an activated vitamin D preparation or a digitalis preparation Patients who are unable to give informed consent Patients who are judged by the physician to be unsuitable for teriparatide treatment
Interventions	teriparatide (daily) teriparatide (weekly)
Outcomes	Primary Outcome Improvement of necrotic bone exposure bone formation
Starting date	2012/01/01
Contact information	Daigo Yoshiga, Kyshu Dental University Oral Medicine r11yoshiga@fa.kyu‐dent.ac.jp
Notes

AAOMS: Association of Oral and Maxillofacial Surgeons; LPRF: leukocyte and platelet‐rich fibrin; MRONJ: medication‐related osteonecrosis of the jaw;OCN: osteochemonecrosis; PRF: platelet‐rich fibrin.

1. P1NP, N‐terminal propeptide of type 1 collagen Procollagen I Intact N‐Terminal

2. Beta‐CTX, Beta‐carboxy‐terminal telopeptide of type 1 collagen (beta‐CrossLaps)

3. Cicatrisation: formation of scar tissue at a wound site by fibroblasts

Differences between protocol and review

The main difference between the protocol and this 2022 version of the review is that we changed the follow‐up time inclusion criterion. We reduced the required follow‐up period from one year to six months in order to include more studies. We did this because the longer the follow‐up period, the more participants drop out ‐ mainly due to progressive cancer ‐ and thus impair the significance of the analysis. Although Hinson 2015 reported delayed healing of MRONJ with a median time of 10 months in participants who received antiresorptive treatment and three to six months in patients who discontinued antiresorptive therapy, Park 2017 reported complete healing of MRONJ after four weeks or delayed healing after 16 weeks in patients receiving antiresorptive therapy. Finally, three treatment studies reported a follow‐up period of only six months (Ohbayashi 2020; Park 2017; Yüce 2021), and five treatment studies reported a longer follow‐up period of at least one year, but most of these also report primary or secondary outcomes at six months (Freiberger 2012; Giudice 2018a; Giudice 2018b; Ristow 2016; Sim 2020). Regarding prophylaxis studies, we can consider prophylaxis successful if there are no signs of MRONJ within eight weeks after surgery according to AAOMS criteria.

Contributions of authors

Drafted the protocol: NB, OZ
Wrote the protocol: NB, OZ
Developed the search strategy: NB, OZ, and Anne Littlewood (Information Specialist from Cochrane Oral Health)
Selected trials: NB, HH, OZ
Extracted data: NB, JP, HH, OZ
Assessed trial for risk of bias: NB, HH, OZ
Assessed certainty of the evidence: NB, OZ
Contacted authors of ongoing RCTs: OZ
Performed statistical analysis: NB, BM, OZ
Wrote the review: NB, JP, OZ
Produced 'Summary of findings' table: NB, JP, OZ

Sources of support

Internal sources

• Division of Dentistry, The University of Manchester; Manchester Academic Health Sciences Centre (MAHSC); and the NIHR Manchester Biomedical Research Centre, UK, UK

  Support to Cochrane Oral Health

External sources

• National Institute for Health Research (NIHR), UK

  This project was supported by the NIHR, via Cochrane Infrastructure funding to Cochrane Oral Health. The views and opinions expressed therein are those of the authors and do not necessarily reflect those of the Systematic Reviews Programme, NIHR, NHS or the Department of Health.

• Cochrane Oral Health Global Alliance, Other

  The production of Cochrane Oral Health reviews has been supported financially by our Global Alliance since 2011 (oralhealth.cochrane.org/partnerships-alliances). Contributors in recent years have been the American Association of Public Health Dentistry, USA; AS‐Akademie, Germany; the British Association for the Study of Community Dentistry, UK; the British Society of Paediatric Dentistry, UK; the Canadian Dental Hygienists Association, Canada; the Centre for Dental Education and Research at All India Institute of Medical Sciences, India; the National Center for Dental Hygiene Research & Practice, USA; New York University College of Dentistry, USA; and Swiss Society of Endodontology, Switzerland.

Declarations of interest

There are no financial conflicts of interest and the review authors declare that they do not have any associations with any parties who may have vested interests in the results of this review.

• Natalie H Beth‐Tasdogan: no interests to declare

• Benjamin Mayer: no interests to declare

• Heba Hussein: no interests to declare

• Jens‐Uwe Peter: no interests to declare

• Oliver Zolk: no interests to declare

New search for studies and content updated (no change to conclusions)