The quality-of-life impact of osteonecrosis of the jaw: Implications for bisphosphonate use in metastatic breast cancer

6620 Background: Osteonecrosis of the jaw (ONJ) has been linked to bisphosphonates used to prevent skeletal related events (SREs) in metastatic breast cancer (MBC). The primary aim of this decision-analysis study is to determine the preference threshold at which the quality of life (QOL) impact of ONJ may change bisphosphonate treatment decisions. Methods: We developed a Markov decision- analysis model of bisphosphonate use in MBC that includes the risk of ONJ. For the base case we estimated the QOL impact of ONJ by evaluating published ONJ reports with the Oral Health Impact Profile (OHIP). OHIP scores were transformed to EQ5D utilities and adjusted for MBC (published utility for MBC=0.63). We used published utility values for SRE: 0.46 for the month in which SRE occurs. Based on published data, we estimated that bisphosphonates reduce the incidence of SREs by 41% and that the incidence of SRE rises with increased bisphosphonate exposure: year 1=0.004/month; year 2=0.022/month; year 3=0.034/month. We inspected 2 treatment strategies: treat all patients with bisphosphonates (treat all) and treat no patient with bisphosphonates (treat none). Results: 18 published cases were adequate for evaluation. The mean OHIP score=27 (possible range 14–70), s.d.=1.8. We calculated that patients with MBC and ONJ have a utility=0.53 (s.d.=0.04) for the base-case. The model predicted a mean survival of 22 months for both strategies. In the treat all strategy each patient received a mean of 19 months of bisphosphonates and suffered 2.4 SREs. In the treat none strategy each patient suffered 4.0 SREs. In the base case, the treat all strategy maximized net quality-adjusted life, although by less than 1/2 month per patient. The treat all strategy was optimal for only 33% of patients. The treat all strategy does not maximize net quality-adjusted life if the risk of ONJ is 4.5 times higher than the base case or the ratio of the utility for ONJ to the utility for SRE is less than 0.4 (base-case ratio=1.152). Conclusions: The QOL impact of ONJ alters the decision to use bisphosphonates when 1) the incidence of ONJ is 4.5 times higher than published estimates; or 2) the long-term preference for ONJ is 60% lower than the short-term preference for SRE. Further QOL research may refine these estimates. No significant financial relationships to disclose.