Reviewed by Dr. Jennifer Wagner
Shapiro, CL et al. J Clin Oncol. 2017 Dec 10;35(35):3949-3955
Women with metastatic breast cancer develop bone metastases in 65 – 75% of cases. These bone metastases cause pain, increase likelihood of pathologic fracture, spinal cord compression, and hypercalcemia; treatment of these conditions comes with significant cost. To prevent complications from bone metastases (termed “skeletal related events” or SREs), both zoledronic acid (ZA) and denosumab are commonly prescribed. This study uses models to assess cost-effectiveness of monthly ZA, every three months ZA, or monthly denosumab. The modeling was based off multiple previous studies which showed that ZA every three months is non-inferior to monthly ZA in patients with skeletal metastases (including breast cancer). ZA overall reduces SREs up to 40%. Denosumab is more effective than ZA, as it delays the onset of first and subsequent SREs. Based on this background knowledge, this study was designed to determine if denosumab is cost-effective compared to monthly or every 3 month ZA, a generic medication.
They designed a Markov model with a hypothetical cohort of 10,000 women in each treatment arm for SRE prevention. A Markov model is used in predictive modeling and assumes that a patient is always in one of a finite number of discrete health states, and events are transitions between states. The different SRE states analyzed in this model were a pathologic bone fracture, radiation to bone, surgery to bone, and spinal cord compression. They based rates of pathological fractures for each medication off previous outside studies. Their model was set to analyze a two-year time course. Costs included the medication itself, administration cost, and treatments of fractures, radiation, surgeries, etc. Every three months ZA was the least expensive and had slightly fewer SREs than monthly ZA, so was the dominant option. Quality-adjusted life-years were assessed and were similar across the three treatment groups. When they assess various probabilities of SRE events on denosumab of 50%, 70%, and 90% lower than on ZA, denosumab was syill not cost effective. When comparing monthly ZA to denosumab in these scenarios, the mean incremental costs per mean SRE avoided for denosumab ranged from $137,905 to $283,109. When they changed the assumptions to SRE probabilities of an increase of 50% or 100% with ZA rather than denosumab, mean costs became more comparable. The mean incremental costs per SRE avoided for denosumab now ranged from $6,072 to $41,432. However, this relied on an assumption that SREs with ZA were much greater than denosumab, which is not currently the case. Based on current studies, denosumab has 23% fewer SREs than monthly ZA, which makes the analysis most like when SRE probabilities of denosumab were assumed to be 75% of those of ZA every 3 months, and in this model the cost of denosumab per SRE avoided was over $200,000.
Overall, every three month ZA was found to be superior in cost effectiveness to both monthly ZA and monthly denosumab. The study is important because it is the first study of cost-effectiveness for these drugs that was not sponsored by a drug company. The study is limited by being a predictive model based on SRE probabilities rather than actual cost data from a cohort. Second, the model was only assessing a two-years time course, and as treatment generally extends longer than this, long term cost-effectiveness could be different; however the authors feel that the cost difference is so substantial that the general results would not vary much. Costs of side effects of the medications (osteonecrosis of the jaw or atypical femoral fractures) were not included, as the rates of these complications are similar between monthly ZA and monthly denosumab. The authors conclude that despite improved outcomes with monthly denosumab, for the highest value care, ZA every three months is a viable alternative.