Real World Treatment Outcomes and Cost Among Newly-Diagnosed Multiple Myeloma Patients Treated with Bortezomib and Dexamethasone in Combination with Cyclophosphamide or Lenalidomide

Background: Despite an increasing incidence of multiple myeloma (MM) with advancing age and life expectancy, there are few real-world claims-based analyses describing treatment patterns and healthcare costs associated with use of novel treatments.^1,2 This study aimed to assess treatment patterns and healthcare costs among newly-diagnosed MM patients using the US Medicare database.

Methods: This retrospective study identified adult patients with ≥2 claims for MM (International Classification of Diseases, 9th Revision, Clinical Modification code: 203.0x) 30 days apart and ≥1 treatment during the identification period (01JAN2011-30JUN2014) from the 100% Medicare dataset. Medicare dataset contains medical and pharmacy claims submitted by healthcare providers, facilities and pharmacy. It includes comprehensive demographic information for beneficiaries and a longitudinal picture of their healthcare utilizations and costs .The initial course of therapy (COT1) date was the index date and included all treatments prescribed within 60 days of this date. Patients were required to have continuous enrollment for 12 months pre- and ≥6 months post-index date unless the patient died in <6 months (follow-up period), ≥1 full cycle of therapy with a valid COT1 regimen, no evidence of prior MM diagnosis or treatment (including autologous stem cell transplant [ASCT]), and no evidence of ASCT in the follow-up period. COT2 was defined as the earliest occurrence of: addition of a new drug or switch in regimen after the first 60 days, restart of a previous regimen after >180-day gap, or dose increase from maintenance to relapse therapy. Steroids (dexamethasone/prednisone [d]) were assumed to be included regardless of whether or not they were observed during the study period; this did not impact the ongoing COT. Treatment patterns and healthcare costs during the follow-up period were compared among those initiating lenalidomide (R) with bortezomib (V) ± steroids (RVd) and cyclophosphamide (Cy) with bortezomib (bor) ± steroids (CyBorD). Time-to-next treatment (TTNT) was defined as the duration from initiation of COT1 plus any treatment gaps until the initiation of COT2. Kaplan Meier (KM), Cox regression analyses and a generalized linear model (GLM) were performed to evaluate TTNT, assess the impact of various predictors on TTNT, and estimate the 12-month per patient per month (PPPM) total healthcare costs respectively among patients initiating RVd and CyBorD.

Results:After accounting for the patient selection criteria, 9.9% (n=345) of patients initiated RVd and 5.0% (n=175) initiated CyBorD as COT1. CyBorD-treated patients were significantly older (76.1 vs. 74.2 years, p=0.0009) with a higher age-adjusted Charlson Comorbidity Index score (9.5 vs 8.8, p=0.0119). The overall mean duration of COT1 was significantly longer among patients treated with RVd vs CyBorD (13.2 vs 8.5 months, p<0.0001). Among patients who completed COT1, the mean duration of COT1 was longer for patients treated with RVd vs. CyBorD (12.8 vs 6.7 months, p<0.0001). A higher percentage of patients treated with CyBorD progressed to COT2 (27.4%, vs 21.7% p=0.1491) versus RVd, however no significant difference was observed. Among patients who progressed to COT2, TTNT was significantly shorter among those treated with CyBorD vs RVd (Mean: 7.9 vs 15.9 months, p<0.0001). KM analysis suggested that patients initiating CyBorD progressed much faster than patients receiving RVd. After adjusting for baseline characteristics using Cox regression, TTNT remained significantly shorter for CyBorD vs. RVd treated patients (hazard ratio: 2.2, 95% confidence interval: 1.5-3.4, p=0.0002). Results from GLM analysis suggested that adjusted total PPPM cost during 12 months follow up was higher among patients treated with RVd vs. CyBorD ($13,941 vs $9,340, p=0.0001), and the majority of the extra cost are due to higher pharmacy costs for patients treated with RVd.

Conclusion: Patients on RVd incurred higher costs, however, they progressed significantly slower and their TTNT was almost twice as long as for CyBorD patients. The difference remained significant after controlling for baseline characteristics including markers for higher disease severity among patients on CyBorD.

¹Song X, et al. Curr Med Res Opin 2015;32(1):95-103

²Teitelbaum A, et al. Oncologist 2013;18:37-45