This is the first study showing that denosumab treatment can significantly increase the risk of developing MRONJ, when compared to ZA, among patients who received BMA treatment for bone metastasis in a real-world clinical practice setting using a propensity score-matched analysis. Tooth extraction after starting BMA treatment was also a significant risk factor. The incidences of MRONJ in the present study were 9.6% and 4.8% in the denosumab and ZA groups, respectively. The reported incidence of MRONJ is 1–17% [25, 27, 29–42], and the incidences of MRONJ in our study were within this range for both groups. In a combined analysis of three phase III randomized control trials, the incidences of MRONJ were low in patients receiving both denosumab (1.8%; n = 52/2862) and ZA (1.3%; n = 37/2861) [25]. Another randomized control trial also reported that the incidences of MRONJ were low in patients with multiple myeloma receiving both denosumab (4.1%; n = 35/850) and ZA (2.8%; n = 24/852) [37]. In contrast, in retrospective observational studies, the reported incidences of MRONJ were found to vary from 2–17% [29–31, 38–42] and were relatively higher than those in randomized controlled trials. In randomized controlled trials, the study protocol specified that all participants underwent scheduled periodic dental examinations (e.g., at baseline and every 6 months thereafter) [25, 27, 32–34, 36, 37]. On the other hand, in the real-word clinical practice setting used in our study, although all subjects underwent dental examination before the initiation of BMA treatment, the vast majority of them did not receive scheduled periodic dental examinations after the initiation of BMA treatment. If the patients complained of dental symptoms, the attending physicians consulted the dentists. Because scheduled dental examination can reduce the risk of developing MRONJ [17–19, 31, 35, 41–43], this discordance between randomized control trials and observational studies in real-world clinical practice might affect the risk of MRONJ.
This study revealed via multivariate analysis that denosumab treatment was associated with a significantly higher risk of developing MRONJ than ZA treatment. Moreover, this result was also clearly confirmed by propensity score-matched analysis. The higher risk for developing MRONJ with denosumab than with ZA was consistent with some previous real-world data [40–43]. Previous randomized controlled trials reported that the incidence of MRONJ in patients treated with denosumab was not significantly different from that in patients treated with ZA, although it tended to be higher in the former [25, 27, 32–34, 36, 37]. One potential explanation for this discordance might be due to scheduled dental examination, which we discussed previously. However, when investigating the risk of MRONJ in retrospective observational studies, several biases, such as dental examinations before BMA treatments and dental interventions after starting BMA treatments, should be considered. As such, to reduce the potential bias of patient characteristics associated with the development of MRONJ, we limited the study participants to those examined by dentists before starting BMA treatment. In addition, we performed propensity score matching to control for and reduce selection bias in the denosumab and ZA groups.
The higher incidence of denosumab-associated osteonecrosis of the jaw seems to reflect its superior effect on bone resorption compared to that of ZA [26, 32, 33]. Recently, we reported that MRONJ caused by denosumab resolves faster than that caused by ZA [40]. We believe that diagnosing MRONJ at an earlier stage through appropriate monitoring and multidisciplinary collaborative work with healthcare providers is essential for BMA treatment [44].
We also revealed that tooth extraction after starting BMA treatment was significantly increased by multivariate analysis, which was consistent with the findings of previous reports [19, 21, 25]. In contrast, tooth extractions performed before starting BMAs tended to increase the risk of developing MRONJ in this study, although this increase was not significant. Poor oral health status is known to be a significant risk factor [17–20], and our results likely support the notion that in unavoidable cases, tooth extraction before starting BMAs is a meaningful intervention to reduce the risk of developing MRONJ. However, in this retrospective study, we could not obtain convincing data related to oral health status before starting BMAs. Further studies are required to confirm these results. Since tooth extraction before starting BMAs significantly increased the risk of developing MRONJ in the univariate analysis, early dental consultation should be considered after patients are diagnosed with cancer.
This study has some limitations. First, oral health status, such as periodontal disease and dental caries, was not fully investigated. To reduce the effect of these factors, we limited the study participants to those examined by dentists before starting BMA treatment. Second, we did not evaluate the effect of other risk factors, such as dental prosthesis and tobacco use [17–20]. Despite our best efforts to obtain clinical information, we were not able to collect all of these data within this retrospective study design. However, to our knowledge, these missing data should have similar impacts among the groups. Lastly, since the patients who complained of dental symptoms consulted a dentist following the request of the attending physicians, mild cases of MRONJ might have been underdiagnosed. Despite these limitations, this real-world observational study clearly revealed that the risk of developing MRONJ was significantly higher in advanced cancer patients treated with denosumab than in those treated with ZA.
In conclusion, the results of this study suggest that denosumab significantly increases the risk of developing MRONJ compared to ZA in cancer patients undergoing treatment for bone metastasis. Tooth extraction after starting BMA treatment is also significantly associated with developing MRONJ. Taken together, these patients require close monitoring in a clinical setting.