Rituximab (Rituxan), when combined with CHOP (cyclophosphamide [Cytoxan, Neosar], doxorubicin HCl, vincristine [Oncovin], prednisone) chemotherapy (R-CHOP) in the treatment of patients with CD20-positive diffuse large B-cell lymphoma (DLCL), significantly prolongs event-free and overall survival (GELA [Groupe d’Etude des Lymphomes de l’Adulte] LNH 98-5 study). Our objective was to estimate the cost-effectiveness of R-CHOP based on the evidence currently available.
Patients enrolled in GELA were required to be between 60 and 80 years old, to have stage II-IV DLCL, and to have an Eastern Cooperative Oncology Group (ECOG) performance status of 0-2; they were randomized to receive R-CHOP or CHOP. A protocol-scheduled interim analysis was conducted after at least 200 patients had been recruited and 100 were followed for at least 1 year.
To assess the cost-effectiveness of R-CHOP vs CHOP from a French health system perspective, we estimated years of life gained, costs during treatment (every 3 weeks for eight cycles), costs of long-term follow-up for survivors, and cost per year of life gained over the trial period (where time of last follow-up equaled 2.4 years) and over 10 years. A 10-year model was developed because approximately 30% of low-risk patients in this aged cohort are expected to survive at least 10 years. Costs and survival were discounted at 3%, and then used to calculate the cost-effectiveness ratio (costs per discounted life-year gained).
To estimate expected survival patterns for 10 years of follow-up, and thus project years of life gained, trial-based survival was stratified according to the age-adjusted International Prognostic Index and matched to the published long-term survival data on DLCL patients (N Engl J Med 329:987, 1993). To estimate costs, French diagnosis-related group (DRG) payments were applied to trial data on hospital use and on treatments for adverse events. French drug prices and drug administration costs were used to estimate the costs of R-CHOP and CHOP regimens. Average annual costs of cancer surveillance after treatment—estimated at 500 euros—were based on published data.
R-CHOP increased survival by 10%, from 59.4% to 69.4% (P < .01) at 22 months (time of last event). The mean survival for the 328 patients who met the interim analysis criteria was 724 days for R-CHOP and 645 days for CHOP, resulting in an increase in mean survival of 0.21 years. If the survival advantage seen more than 1 year after completion of eight cycles of treatment is found to persist to 10 years, R-CHOP is predicted to increase discounted mean survival by 0.8 life-years. Estimated therapy-related costs during the trial period were 14,840 euros higher with R-CHOP, including costs for drug acquisition and administration and for treatment of adverse events.
Projecting over 10 years, the total added cost per patient is 15,190 euros, and the estimated cost per year of life gained is 22,150 euros—a ratio that is generally perceived as favorable by health authorities. Sensitivity analyses indicate that this base-case result depends on the durability of the survival benefit. For example, a short-term estimate based on the trial period alone would produce an upper boundary of cost per life-year gained equal to 68,580 euros.
CONCLUSION: Evidence from this trial supports the assumption that R-CHOP has resulted in a higher number of sustained cures compared with CHOP. Additional follow-up of all 399 randomized patients will provide more definitive evidence on the durability of these findings. The cost-effectiveness of R-CHOP compares favorably with other oncology treatments in widespread use.