We question the validity of the indirect comparison of everolimus and sorafenib by Di Lorenzo et al. Citation[1] because it used data that are inconsistent with other reports on the same patients and an unproven analytic approach.
Di Lorenzo et al. used patient-level data on everolimus from RECORD-1, a Phase III trial comparing everolimus and placebo Citation[2], and patient-level data on sorafenib from a single-arm Phase II trial Citation[3]. For the entire everolimus arm of RECORD-1 and the subset of sunitinib-failure patients (prior to the matching for the indirect comparison), their Table 2 reports estimates of median overall survival (OS) and median progression-free survival (PFS). The estimate of median OS in the everolimus arm of RECORD-1 (n = 277), 64 weeks (or 14.8 months), is consistent with the same statistic reported by Motzer et al. Citation[2]. However, the estimate of median PFS for the everolimus arm (n = 277), 24.1 weeks (5.6 months), is 14% longer than the 21.2 weeks (4.9 months) reported by Motzer et al. Citation[2] and cited in the prescribing information for everolimus and in the Introduction of the article itself. The estimate of median PFS for sunitinib-failure everolimus patients, also 24.1 weeks, is 43% longer than the 16.9 weeks (3.9 months) reported by Motzer et al. Citation[2] using what should be the same data.
Di Lorenzo et al. report 127 patients in that subset, whereas reports from RECORD-1 show only 124 Citation[2,4]. In their Table 1, the percentages of pre-matched sunitinib-failure RECORD-1 everolimus patients in the categories of MSKCC risk score (28.1, 54.7, and 17.2%) are not compatible with any subset of 127 (or 124).
Di Lorenzo et al. have not established the validity of their matching-adjusted approach (selecting subsets of patients, matching on a single variable, and using repeated samples to estimate a confidence interval). Their references 10 and 23 – 25 do not support it. They discussed and analyzed differences between the everolimus patients and the sorafenib patients that remained after matching, but they provide no evidence that their approach compensated for the absence of a comparator common to the everolimus trial and the sorafenib trial. In established practice, indirect comparisons require randomized trials with a common comparator Citation[5].
The concerns that we have expressed also have implications for the economic analysis by Casciano, Chulikavit, and Di Lorenzo Citation[6], which used the results of this study as a basis.
Declaration of interest
DC Hoaglin is an independent consultant. He worked on this letter as a consultant to United BioSource Corporation, in connection with a project supported by Bayer HealthCare Pharmaceuticals. PR Cislo is an employee of Bayer HealthCare Pharmaceuticals.
Bibliography
- Di Lorenzo G, Casciano R, Malangone E, An adjusted indirect comparison of everolimus and sorafenib therapy in sunitinib-refractory metastatic renal cell carcinoma patients using repeated matched samples. Expert Opin Pharmacother 2011;12(10):1491-7
- Motzer RJ, Escudier B, Oudard S, Phase 3 trial of everolimus for metastatic renal cell carcinoma: final results and analysis of prognostic factors. Cancer 2010;116(18):4256-65
- Di Lorenzo G, Carteni G, Autorino R, Phase II study of sorafenib in patients with sunitinib-refractory metastatic renal cell cancer. J Clin Oncol 2009;27(27):4469-74
- Motzer RJ, Escudier B, Oudard S, Efficacy of everolimus in advanced renal cell carcinoma: a double-blind, randomised, placebo-controlled phase III trial. Lancet 2008;372(9637):449-56
- Jansen JP, Fleurence R, Devine B, Interpreting indirect treatment comparisons and network meta-analysis for health-care decision making: report of the ISPOR task force on indirect treatment comparisons good research practices: Part 1. Value Health 2011;14:417-28
- Casciano R, Chulikavit M, Di Lorenzo G, Economic evaluation of everolimus versus sorafenib for the treatment of metastatic renal cell carcinoma after failure of first-line sunitinib. Value Health 2011;14(6):846-51