Majer Istvan, van de Wetering Gijs, Polanyi Zoltan, Krishna Arun, Gray Elisabeth, Roy Anuja
Pharmerit International, Health Economics and Outcomes Research, Marten Meesweg 107, 3068 AV, Rotterdam, The Netherlands.
Novartis Pharmaceuticals UK, Surrey, UK.
Appl Health Econ Health Policy. 2017 Feb;15(1):45-55. doi: 10.1007/s40258-016-0271-0.
In the UK, the standard of care for patients with multiple myeloma who received ≥2 prior treatments is lenalidomide plus dexamethasone (LEN + DEX) and pomalidomide plus DEX (POM + DEX) (in Wales only). Recently, panobinostat plus bortezomib and DEX (PAN + BTZ + DEX) was licensed in this setting. The current study assessed the progression-free survival (PFS) and overall survival (OS) outcomes with PAN + BTZ + DEX versus LEN + DEX (primary comparator) and POM + DEX (exploratory comparator).
Since an anchor-based indirect treatment comparison was not feasible, the matching-adjusted indirect treatment comparison approach was used. To compare the survival outcomes, patient-level data were generated for the comparators utilizing published Kaplan-Meier survival estimates. The use of approximated patient-level data and matched data for PAN + BTZ + DEX allowed the use of Cox proportional hazards models and the assessment of the proportional hazards assumption. In cases where there was evidence that the proportional hazards assumption was violated, time-dependent hazard ratios (HRs) were estimated. Median and mean values for PFS and OS were predicted.
For both PFS and OS, the proportional hazards assumption was not satisfied, therefore time-dependent HRs were estimated. Using time-dependent HRs, the mean PFS was estimated to be 11.83 months for PAN + BTZ + DEX and 10.96 months for LEN + DEX. The corresponding mean OS estimates were 30.73 and 27.76 months, respectively. Comparisons with POM + DEX were affected by large uncertainty and did not allow making robust inferences.
To our knowledge, this is the first study that combined matching-adjusted indirect treatment comparison with time-dependent HRs to address changing patterns in the HR. The results suggest that treatment with PAN + BTZ + DEX and LEN + DEX are associated with similar mean PFS and OS in the third-line treatment setting of multiple myeloma.
在英国，接受过≥2次先前治疗的多发性骨髓瘤患者的标准治疗方案是来那度胺联合地塞米松（LEN + DEX）以及泊马度胺联合地塞米松（POM + DEX）（仅在威尔士）。最近，帕比司他联合硼替佐米和地塞米松（PAN + BTZ + DEX）在此种情况下获得了许可。本研究评估了PAN + BTZ + DEX与LEN + DEX（主要对照）以及POM + DEX（探索性对照）相比的无进展生存期（PFS）和总生存期（OS）结果。
由于基于锚定的间接治疗比较不可行，因此采用了匹配调整间接治疗比较方法。为了比较生存结果，利用已发表的Kaplan-Meier生存估计值为对照生成患者水平数据。使用近似的患者水平数据和PAN + BTZ + DEX的匹配数据允许使用Cox比例风险模型并评估比例风险假设。在有证据表明比例风险假设被违反的情况下，估计时间依赖性风险比（HRs）。预测了PFS和OS的中位数和平均值。
对于PFS和OS，比例风险假设均未得到满足，因此估计了时间依赖性HRs。使用时间依赖性HRs，PAN + BTZ + DEX的平均PFS估计为11.83个月，LEN + DEX为10.96个月。相应的平均OS估计分别为30.73和27.76个月。与POM + DEX的比较受到较大不确定性的影响，无法进行可靠推断。
据我们所知，这是第一项将匹配调整间接治疗比较与时间依赖性HRs相结合以解决HR变化模式的研究。结果表明，在多发性骨髓瘤的三线治疗中，PAN + BTZ + DEX和LEN + DEX治疗的平均PFS和OS相似。
