Pharmacological and surgical predictors of postoperative healing in medication-related osteonecrosis of the jaw: A retrospective multicenter cohort study.
retrospective_cohort · Level III
Where this comes from
- Record sourced from PubMed, PMID 42431290.
- Also identified by DOI 10.1016/j.bone.2026.118010.
- No licence information is recorded for this record.
- Because redistribution is not established, this page shows the abstract only. Follow the links below for the full text.
Abstract
Failure to achieve postoperative healing after surgical treatment of medication-related osteonecrosis of the jaw (MRONJ) remains a clinical challenge. We aimed to investigate the independent predictors of postoperative nonhealing and to evaluate the impact of regenerative adjuncts. In a multicenter surgical cohort (2019-2025) of 531 patients with MRONJ, pharmacological, lesion level, and surgical factors were assessed using univariable testing, elastic net regularization for variable selection, and multivariable logistic regression. Postoperative nonhealing was defined as the need for additional surgical intervention or failure to achieve complete mucosal coverage, accompanied by clinical or radiographic evidence of persistent or progressive disease. Kaplan-Meier analysis was used to estimate nonhealing-free survival. The overall postoperative nonhealing rate was 18.5% (98/531). In the multivariable models, high-dose antiresorptive therapy (odds ratio [OR]: 2.94, 95% confidence interval [CI]: 1.70-5.06) and multiple-site lesions (OR: 2.29, 95% CI: 1.09-4.67) were independent risk factors for postoperative nonhealing, whereas intraoperative bone morphogenetic protein (BMP) use was associated with a lower risk of postoperative nonhealing (OR: 0.32, 95% CI: 0.20-0.52). The number needed to treat to prevent one postoperative nonhealing event was 5.3 (95% CI: 3.9-8.7) for BMP, and the number needed to harm was 4.4 (95% CI: 3.0-8.5) for high-dose therapy. Solitary maxillary lesions showed an insignificant protective effect. Antiangiogenic agents were associated with postoperative nonhealing in the univariate and elastic net analyses, but not in the final multivariable model. Neither treatment duration nor preoperative drug holiday was independently associated with postoperative nonhealing. The elastic net model showed acceptable discrimination (area under the curve [AUC]: 0.724), and the Kaplan-Meier curves were consistent with these findings. Postoperative nonhealing after surgical management of MRONJ was mainly associated with systemic pharmacological burden and lesion extent, whereas BMP application was associated with improved postoperative healing. These findings support further investigation of regenerative adjuncts and risk-adapted surgical management strategies.