Efficacy of Treatments in Relapsed or Relapsed and Refractory Multiple Myeloma: An Indirect Treatment Comparison

Background: In the phase 3, randomized, international, double-blind PANORAMA 1 trial, panobinostat in combination with bortezomib and dexamethasone (PAN/BTZ/DEX, n=387) demonstrated superior progression-free survival (PFS), time to progression (TTP) and near complete response plus complete response rate (nCR+CR) over placebo/BTZ/DEX (n=381) in patients with relapsed or relapsed and refractory multiple myeloma (RR MM). These results support the addition of panobinostat to BTZ/DEX and demonstrate that this regimen is a valuable addition to the current treatment options for RR MM. However, evidence on the efficacy of PAN/BTZ/DEX relative to other relevant treatments is lacking. The aim of this study was to perform an indirect treatment comparison (ITC) to estimate the relative efficacy of PAN/BTZ/DEX compared to commonly used treatment regimens.

Methods: A systematic literature review was conducted. MEDLINE, MEDLINE In-Process, EMBASE, the Cochrane Library, and conference proceedings were searched covering clinical studies dating from January 2003 to April 2014. Studies included phase 2-4 clinical trials specifically focusing on RR MM, reporting results in English for the most commonly used agents: intravenous BTZ, lenalidomide (LEN), thalidomide (THAL), and doxorubicin (DOX). Data were extracted on study details, study duration, and efficacy outcomes including the hazard ratio (HR) of PFS, the HR of TTP, and the number of patients who achieved nCR+CR.

Multi-arm randomized controlled trials were selected for inclusion in the evidence base of the ITC. Included studies were similar in terms of trial design (e.g. patient selection criteria), patient characteristics (e.g. age, time since diagnosis, number of prior lines of therapy), and included treatments that could be linked via common comparators. Data were analysed by applying fixed effects models to estimate HRs of PFS, HRs of TTP, and odds ratios (OR) of nCR+CR. PAN/BTZ/DEX was chosen as the reference treatment for ease of comparability and interpretability of the results.

The models were conducted using Markov Chain Monte Carlo simulation methods implemented in WinBUGS 1.4. The mean of the posterior distribution was taken as the point estimate and 95% credible intervals (CI), i.e. range of values containing the true mean with a probability of 95%, were calculated.

Results: The analysis screened 637 studies of which 83 were assessed by full text. Five relevant trials were identified (PANORAMA 1, MM-009, MM-010, APEX, and DOXIL-MMY-3001) that included PAN/BTZ/DEX, BTZ/DEX, BTZ, DEX, LEN/DEX, and BTZ/DOX. Since no trial was conducted to assess the efficacy of BTZ/DEX versus DEX, results of a matched-pairs comparison analysis (using propensity score matching) were utilized in the evidence network. The patient populations across the studies were similar in terms of median age (61-64 years), disease duration (24-48 months at baseline), and proportions of patients with one prior line of therapy (32-46%), except the matched pairs analysis, where only patients with one prior line of therapy were considered. The number of patients included in the studies totaled 4,109.

Results of the comparison between PAN/BTZ/DEX and other treatments in terms of PFS indicated that PAN/BTZ/DEX was associated with the lowest risk of progression or death. Similarly, the HR of time to progression was estimated to be higher for other drugs than for PAN/BTZ/DEX. The odds ratio of nCR+CR indicated higher response rates for PAN/BTZ/DEX than for other treatments. Coefficients and CIs are reported in Table 1.

Discussion: The analyses implicitly assume that covariates acting as potential relative treatment effect modifiers (e.g. patients with relapsed and refractory MM, ß2-microglobulin level) are similar across trials. The numerical trends of the analyses suggest that PAN/BTZ/DEX is at least as efficacious as other treatments in RR MM with significant advantage against most comparators.