Comparative efficacy of carfilzomib, lenalidomide, and dexamethasone (KRd) versus bortezomib, lenalidomide, and dexamethasone (VRd) in newly‐diagnosed multiple myeloma: A systematic review and meta‐analysis

In view of the increasing data evaluating carfilzomib-based induction for newly-diagnosed multiple myeloma (NDMM), we conducted a systematic review and meta-analysis comparing the efficacy of carfilzomib/lenalidomide/dexamethasone (KRd) versus bortezomib/lenalidomide/dexamethasone (VRd). Three studies totaling 1597 patients (50% KRd-treated, 50% VRd-treated) were included. Despite similar survival outcomes and overall response rate compared with the VRd arm, KRd-treated subjects showed higher odds of achieving complete responses and measurable residual disease negativity. Among patients with high-risk cytogenetics (n = 348), KRd was associated with significant improvement in progression-free survival (HR = 0.70; 95% CI = 0.50–0.97; p = .03; I2 = 0%), suggesting carfilzomib-based induction may be preferable in this NDMM subpopulation. The treatment paradigm of newly-diagnosed multiple myeloma (NDMM) has undergone a profound change over the past few decades, with induction regimens evolving beyond the traditional steroid/alkylator backbones to highly effective combinations incorporating immunomodulatory drugs, proteasome inhibitors, and/or monoclonal antibodies.1 As a result, recent large-scale studies of patients treated in resource-rich settings have reported an impressive overall survival (OS) of around 126 months.2 However, given the numerous options of primary therapy currently endorsed by consensus guidelines, there is ongoing discussion about the relative safety and efficacy of the combinations available. While ENDURANCE—a multicenter, open-label, phase 3, randomized controlled trial (RCT)—found no significant difference in progression-free survival (PFS) between previously-untreated patients without intention for immediate autologous stem cell transplantation receiving carfilzomib/lenalidomide/dexamethasone (KRd) versus bortezomib/lenalidomide/dexamethasone (VRd), subsequent real-world results from NDMM cohorts have varied.3-5 Moreover, the practical use of carfilzomib-based induction in frontline settings continues to rise. Considering this dilemma, we conducted a systematic review and meta-analysis (SRMA) comparing efficacy outcomes between patients treated with KRd versus VRd as first-line induction for NDMM. This study adhered to the Cochrane Handbook for Systematic Reviews of Interventions and the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) 2020 statement. Our protocol was registered in the International Prospective Register of Systematic Reviews (PROSPERO; ID number: CRD42024504804). We systematically searched PubMed, Embase, and Cochrane Library from inception to January 2024 using the following subject headings and keywords with various synonyms: "carfilzomib," "bortezomib," "lenalidomide," "dexamethasone," "induction therapy," "multiple myeloma," and "newly diagnosed." Two investigators independently screened search records using the Rayyan application. We included RCTs and observational studies comparing efficacy outcomes between individuals who received VRd or KRd as induction therapy for NDMM. Studies meeting any of the following criteria were excluded: (1) enrollment of patients with relapsed/refractory multiple myeloma, other plasma cell dyscrasias, or non-plasma cell malignancies; (2) absence of a VRd arm and/or a KRd arm; (3) non-reporting of comparative data for outcomes of interest; (4) overlapping population with another eligible study, where we prioritized the most comprehensive publication in which outcome measures were adjusted for potential confounding factors. Following prespecified search criteria and quality assessment, two investigators independently extracted relevant data from full-text publications. Differences in study screening/selection or data extraction were resolved after discussion with the senior author. Two investigators also independently assessed the eligible studies for bias. While RCTs were appraised using Cochrane's risk-of-bias tool for randomized trials version 2 (RoB 2), observational studies were appraised using Cochrane's risk of bias in non-randomized studies of interventions (ROBINS-I). The main outcome of interest was PFS (time from induction randomization/initiation until the earlier of disease progression or death from any cause). Secondary endpoints included OS (time from induction randomization/initiation to either death from any cause or last disease evaluation), overall response rate (ORR; proportion of patients achieving partial response or better at induction completion), complete response rate (CRR; proportion of patients achieving complete or stringent complete response [CR/sCR] at induction completion), and measurable residual disease (MRD) negativity rate by flow cytometry (proportion of patients with <0.001% neoplastic plasma cells on post-induction bone marrow samples). Disease response/progression was assessed according to the International Myeloma Working Group (IMWG) uniform criteria.1 Pre-planned subgroup analyses for PFS and OS were conducted in patients who had detailed cytogenetic testing, with high-risk NDMM being defined by one or more of the following features: del(17p), t(4;14), t(14;16), t(14;20), and 1q21 gain/amplification.1 After extracting PFS and OS hazard ratios (HRs) and 95% confidence intervals (CIs) from each study, we pooled log-HRs according to the intent-to-treat principle. For MRD negativity and response rates, we pooled the absolute number of recorded events to estimate odds ratios (ORs) and 95% CIs for binary endpoints. The DerSimonian–Laird random-effects model was consistently employed for a conservative estimation of the overall effect size, with two-sided p-values <.05 deemed statistically significant. To assess between-study heterogeneity, we employed Cochran's Q test (p-values <.10 deemed significant) and I-squared statistics (I2 values <25%, 25%–50%, and >50% deemed low, moderate, and high heterogeneity, respectively). We checked for small study effects (publication bias) via visual funnel plot inference. Review Manager 5.4.1 (The Cochrane Collaboration, Copenhagen, Denmark) was used for all statistical analyses. Our initial search identified 494 studies upon deduplication. Based on titles/abstracts and full-text reviews, three studies (1 RCT and 2 retrospective cohort studies) conformed to the predefined criteria (Figure S1). The total pooled population consisted of 1597 individuals. Among the 798 patients (50%) who underwent VRd induction, 783 were response evaluable and 696 were MRD evaluable. Among the 799 patients (50%) who underwent KRd induction, 780 were response evaluable and 760 were MRD evaluable. Table S1 summarizes study baseline characteristics.3-5 In the pooled analyses for the overall population, KRd-treated and VRd-treated subjects demonstrated no significant differences in PFS (HR = 0.84; 95% CI = 0.61–1.15; p = .27; I2 = 55%; Figure 1A), OS (HR = 0.77; 95% CI = 0.42–1.44; p = .41; I2 = 61%; Figure 1B), or ORR (OR, 0.32; 95% CI = 0.07–1.43; p = .14; I2 = 75%; Figure 1C). In comparison to the bortezomib arm, the carfilzomib arm showed 36% higher odds of achieving CR/sCR (OR, 0.64; 95% CI = 0.44–0.92; p = .02; I2 = 43%; Figure 1D) and 38% higher odds of achieving MRD negativity (OR, 0.62; 95% CI = 0.45–0.86; p = .003; I2 = 0%; Figure 1E). Among the 1137 individuals with available cytogenetic data, 348 (31%) were found to have high-risk NDMM. In this subgroup, KRd induction was ultimately associated with a significant improvement in PFS when compared with VRd induction (HR = 0.70; 95% CI = 0.50–0.97; p = .03; I2 = 0%; Figure 1F). Although a trend toward improved OS in the KRd arm was also observed, it did not reach statistical significance (HR = 0.42; 95% CI = 0.17–1.02; p = .06; I2 = 55%; Figure 1G). Quality appraisal of individual studies is available in Figure S2. While neither observational study was considered to have a serious risk of bias, the lack of blinding of outcome assessors in the RCT could potentially lead to observer bias. In the funnel plot, studies with similar weights were symmetrically distributed against their standard errors, indicating an absence of publication bias (Figure S3). The small number of included studies (n < 10) hindered the application of Egger's regression test to quantitatively assess funnel plot asymmetry. By pooling data from a large multicenter RCT and well-designed retrospective cohorts, this comparative SRMA demonstrated superior rates of CR/sCR and MRD negativity in patients receiving KRd instead of VRd for NDMM management.3-5 While such efficacy measures are valuable, they serve as surrogate/proxy endpoints for OS and should not be the sole basis for determining therapeutic decisions. Given the lack of PFS/OS benefit with KRd in the overall population analyzed, our study suggests that VRd-based induction remains a standard-of-care option for previously-untreated individuals without high-risk cytogenetic features, especially due to the higher incidence of severe cardio-pulmonary–renal toxicities (16% vs. 4.7%) and treatment-related deaths (2.1% vs. 0.4%) observed with KRd in the ENDURANCE trial.3 Alternatively, the present SRMA is the first to demonstrate a significant PFS advantage when using KRd over VRd in treating high-risk NDMM. Our findings add up to results from the FORTE trial in supporting the first-line use of KRd induction for patients with high-risk cytogenetic abnormalities, going further to suggest KRd-based combinations could be preferable than VRd-based combinations in this subpopulation.6 A notable exception may apply for individuals with underlying cardiovascular comorbidities, among whom carfilzomib tolerability can be an important concern.3, 4 In such cases, the increased risk of carfilzomib-related adverse events might outweigh the drug's potential benefits. This SRMA has limitations. There was a small number of eligible studies, culminating in a modest total sample size. Second, the level of statistical heterogeneity observed in some of the pooled analyses might affect our findings' generalizability. As our protocol focused on comparing efficacy outcomes between treatment groups, we did not conduct analyses on patient-reported outcomes or safety profile, key factors for clinical decision-making in myeloma care. Although a carfilzomib dosing schedule of 70 mg/m2 once weekly is more commonly used in current practice due to similar safety/efficacy and increased convenience, a majority of patients in the KRd arm received a carfilzomib dosing schedule of 36 mg/m2 twice weekly. Finally, efficacy data on VRd-based and KRd-based quadruplets containing an anti-CD38 monoclonal antibody (daratumumab or isatuximab) were not included, given the absence of prospective/retrospective head-to-head comparisons of these regimens in current literature. Considering the above, results from ongoing RCTs comparing these induction regimens in frontline settings, such as COBRA (NCT03729804), 219TiP (CTRI/2022/05/042759), and ADVANCE (NCT04268498), will be pivotal in refining and personalizing therapeutic strategies for NDMM. In addition, future multinational studies focused on previously-untreated patients with high-risk cytogenetics are essential to further explore VRd-based and KRd-based combinations and validate our findings in broader patient populations. Bruno Almeida Costa: Conceptualization; methodology; investigation; data curation; formal analysis; and writing (original draft). Thomaz Alexandre Costa: Investigation; data curation; formal analysis; visualization; and writing (original draft). Kevin Pak: Investigation; data curation; validation; writing (original draft). Aesha Patel: Investigation; data curation; validation; writing (original draft). Nicole Felix: Software; formal analysis; visualization; writing (original draft). Tarek H. Mouhieddine: Investigation; validation; supervision; and writing (review and editing). Joshua Richter: Conceptualization; methodology; supervision; project administration; and writing (review and editing). Bruno Almeida Costa gratefully acknowledges the support provided by the American Society of Hematology (ASH) through the "Hematology Opportunities for the Next Generation of Research Scientists" (HONORS) Award. Cytogenetics-stratified data for ENDURANCE trial participants were collected from a post-hoc analysis (Kapoor P, Schmidt T, Jacobus S, et al. Impact of chromosome 1 abnormalities on newly diagnosed multiple myeloma treated with proteasome inhibitor, immunomodulatory drug, and dexamethasone: Analysis from the ENDURANCE ECOG-ACRIN, E1A11 trial) presented at the 18th International Myeloma Workshop (Vienna, Austria; September 8–11, 2021). Full slides for this oral presentation were made available by the International Myeloma Society (IMS) via their official website (https://www.myelomasociety.org/wp-content/uploads/2021/09/Kapoor.pdf). Tarek H. Mouhieddine reports advisory board fees from Legend Biotech and Sanofi. Joshua Richter is a consultant/adviser for Janssen, BMS, Pfizer, Karyopharm, Sanofi, Takeda, and AbbVie, and is a member of the speaker's bureau for Janssen, BMS, Sanofi, and Adaptive Biotechnologies. All other authors have no conflicts of interest to disclose. This meta-analysis incorporates data derived from research previously published and available in the public domain. Researchers interested in obtaining patient-level data from the studies included should directly contact the corresponding authors of each respective publication. Data S1. Supporting Information. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.