Development of a risk scoring system for predicting post-extraction medication-related osteonecrosis of jaws: a derivation and validation cohort study

Authors

• Ling-Ying Wei, Division of Oral and Maxillofacial Surgery, Department of Dentistry, National Taiwan University Hospital, Taipei, Taiwan
• Ching-Ming Chiu, Department of Family Medicine, National Taiwan University Hospital, Taipei, Taiwan
• Sang-Heng Kok, Division of Oral and Maxillofacial Surgery, Department of Dentistry, National Taiwan University Hospital, Taipei, Taiwan
• Hung-Ying Lin, Division of Oral and Maxillofacial Surgery, Department of Dentistry, National Taiwan University Hospital, Taipei, Taiwan
• Wei-Yih Chiu, Division of Endocrinology and Metabolism, Department of Internal Medicine, National Taiwan University Hospital, Taipei, Taiwan
• Hao-Hong Chang, Division of Oral and Maxillofacial Surgery, Department of Dentistry, National Taiwan University Hospital, Taipei, Taiwan
• Shih-Jung Cheng, Division of Oral and Maxillofacial Surgery, Department of Dentistry, National Taiwan University Hospital, Taipei, Taiwan
• Jang-Jaer Lee, Division of Oral and Maxillofacial Surgery, Department of Dentistry, National Taiwan University Hospital, Taipei, TaiwanFollow

First Page

1654

Last Page

1661

Abstract

Background/purpose: Although risk factors for medication-related osteonecrosis of the jaw (MRONJ) are known, clinical assessment prior to dental extraction remains largely subjective. We aimed to develop and externally validate a simple clinical risk scoring system to predict post-extraction MRONJ in osteoporotic patients receiving antiresorptive therapy (ART) and support evidence-based clinical decision-making.

Materials and methods: This retrospective cohort study used a derivation cohort (N = 1,067 extractions, 2003‒2022) to develop a 5- to 16-point risk score system based on five predictors (age ≥ 75, ART duration ≥ 24 months, bisphosphonate use, drug interruption < 3 months, and extraction site). The model was validated in an independent cohort (N = 928 extractions, 2022–2024) using a pre-defined ≥12 cut-off. Performance metrics included the area under the curve (AUC), sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV).

Results: MRONJ prevalence was lower in the validation cohort (2.2% vs. 24.5%; P < 0.001), reflecting stricter drug interruption adherence (90.0% vs. 71.5%). The AUC was 0.86 and 0.75 in derivation and validation cohorts, respectively. At the ≥ 12 cut-off, sensitivity was 65.0%, specificity 67.5%, PPV 4.2%, and NPV 98.9%.

Conclusion: The model demonstrates good discrimination and excellent "rule-out" utility in a real-world, low-prevalence setting, identifying low-risk patients (< 12 points) for safe dental extraction, thereby facilitating evidence-based clinical decision-making.

Recommended Citation

Wei, Ling-Ying; Chiu, Ching-Ming; Kok, Sang-Heng; Lin, Hung-Ying; Chiu, Wei-Yih; Chang, Hao-Hong; Cheng, Shih-Jung; and Lee, Jang-Jaer (2026) "Development of a risk scoring system for predicting post-extraction medication-related osteonecrosis of jaws: a derivation and validation cohort study," Journal of Dental Sciences: Vol. 21: Iss. 3, Article 33.
Available at: https://jds.ads.org.tw/journal/vol21/iss3/33

Publication Date

2026