Definition and estimation of osteonecrosis of jaw (ONJ), and optimal duration of antiresorptive treatment in bone metastatic cancer patients: supplementary data from the denosumab extension study?

Abstract

To the Editor, In their recent article, Stopeck et al. [1] concluded that denosumab confirms its known safety profile even after longterm exposure, or after switching to it from zoledronic acid, and that osteonecrosis of jaws (ONJ) rates increased with increasing exposure to antiresorptives, consistent with previous reports. This is based on the open label extension phase of two phase 3 studies in patients with breast and prostate cancer with bone metastases who were randomized to receive denosumab or zoledronic acid (ZA) [2, 3]. The patients were offered open-label denosumab for up to an additional 2 years after the results of the primary analysis, favorable for denosumab on several aspects. Patients initially randomized to denosumab (denosumab/ denosumab group) continued to receive denosumab at 120 mg Q4W whereas patients on ZA were switched to denosumab in the open-label phase (ZA/denosumab group) at 120 mg Q4W starting 4 weeks from their last ZA dose. Patients who declined further therapy in the open-label extension phase, or who did not complete the blinded treatment phase, continued follow-up for survival every 12weeks (Q12W) for up to 2 years after their last dose. We collected data from the text and the tables of the paper published by Stopeck et al. [1] and summarized them in a new table (Table 1). Although authors’ conclusions are quite reassuring both in terms of denosumab safety and efficacy, it is noteworthy that the median exposure of patients to denosumab in the extension phase study is lower than expected, even in the presence of a longer range. The final median exposure of the 318 denosumab/denosumab breast cancer patients is 19.1 months (range 0.1–59.8), not much longer than that registered for the whole cohort of the 1019 breast cancer patient population enrolled in the blinded phase, that was 17.6 months (range 0– 23.7). As far as prostate cancer patients are concerned, the median denosumab exposure was 12.0 months (range 0.1– 67.2) for the 147 denosumab/denosumab patients from the extension study versus 12.0 months (range 0.1–23.3) for the 942 patients enrolled in the original randomized trial. We could not work out from the paper the median denosumab exposure for 318 breast and 147 prostate cancer patients of the extension study denosumab/denosumab population in the previous blinded phase. This appears to us a weakness of the article. Interestingly, the frequency of ONJ cases in the open label extension study appears substantially higher than what is found in the initial blinded phase (ONJ frequency ranging from 1 to 2 %) [2, 3] despite median denosumab exposure was not significantly longer in the former. The crude ONJ figures increased both in denosumab/denosumab groups and in ZA/denosumab populations: ONJ cases were respectively 20/318 (6.3 %) in breast patients and 12/147 (8.2 %) in prostate patients in the denosumab/denosumab group, whereas they were 18/334 (5.4 %) breast patients and 7/118 (5.9 %) prostate patients in the ZA/denosumab group. Such an increase in ONJ frequency highlights the need for longer patients’monitoring and the adoption of nonparametric actuarial estimation (Kaplan-Meier), as done in other studies * Vittorio Fusco vittoriofusco1@tin.it