Introduction: Bone destruction is common multiple myeloma. At diagnosis 79% of multiple myeloma (MM) patients has radiological bone disease and 59% has bone pain. This usually increases further during the course of their disease.

Risk of progressive bone disease can be minimized with monthly ZOL treatment and the original bisphosphonate studies had a median treatment period of 21 and 24 months respectively. We have showed earlier that monthly treatment beyond two years offers significant protection from progressive bone disease in newly diagnosed MM patients treated with zoledronic acid (ZOL) for two years and then randomized to either continued ZOL or observation. But is this true for all patients regardless of depth of response, as other groups have shown the bone protection of ZOL to be most pronounced in patients with a suboptimal response? To answer this question we made a post-hoc analysis of the Magnolia trial looking at patients who had achieved at least VGPR at latest time point prior to randomization.

Methods:

All patients in study were classified according to International Myeloma Working Group (IMWG) Uniform Response Criteria achieved at latest time point before randomization after two years of ZOL treatment. All patients who had not achieved at least VGPR was excluded. We similarly analyzed patients achieving at least CR. Analysis was performed as a cox regression with PBD as the failure state. This was done to include patients risk time in study. Criteria for PBD were unchanged from earlier investigation and required ≥25% increase in size of existing osteolytic lesions or new osteolytic lesions (both at least 10 mm increase/diameter), new fractures, or lesions needing irradiation therapy or surgery.

Results: Of 193 patients in the main study 150 patients had achieved at least VGPR at latest time point before randomization after two years of ZOL treatment. 77 remained in monthly ZOL and 73 moved to observation. 25 cases of PBD were found (observation n =18, ZOL n = 7). The hazard ratio favored continued ZOL treatment with hazard ratio for treatment group of 0.37 and 95% CI (0.15-0.89) p = 0.027. For patients achieving at least CR the hazard was still in favor of ZOL 0.47, but it was not significant due to lack of statistical power; number of patients (n = 69), number of PBD cases 9 (observation n =6, ZOL n = 3) CI (0.12-1.84) p = 0.278.

Conclusions: Continued treatment with ZOL offers decreased risk of PBD in myeloma patients who had achieved at least VGPR at latest treatment after two years initial ZOL treatment compared to observation.