Maintenance rituximab revisited: is it worth the cost?

As the overall survival of patients with follicular lymphoma (FL) edges closer to 12 or 15 years [1,2], an analysis of therapeutic interventions must be considered in many contexts, including disease control and survival, toxicity, quality of life and healthcare costs. This latter point is complex, since the timing of treatment initiation, choice of frontline therapy and clinical heterogeneity of patients with FL all confound the true cost of therapy. Nevertheless, in this issue of Leukemia and Lymphoma, Hornberger and colleagues from Cedar Associates LLC, in collaboration with Genentech, have performed a provocative cost-effectiveness analysis of maintenance rituximab (MR) using a model based on patients enrolled in the PRIMA (Primary RItuximab and MAintenance) trial [3].

Considering that the PRIMA trial provided the model patients, it is worth briefly reviewing several details of the original trial [4]. First, all patients had high tumor burden FL (grades 1–3a) requiring treatment as per Groupe d’Etude des Lymphomes Folliculaires (GELF) criteria. Induction therapy consisted of one of three chemoimmunotherapy combinations: R-CHOP (rituximab, cyclophosphamide, doxorubicin, vincristine and prednisone) R-CVP (rituximab, cyclophosphamide, vincristine and prednisone) and R-FCM (rituximab, fludarabine, cyclophosphamide and mitoxantrone). Following induction, only responding patients were randomized 1:1 between rituximab 375 mg/m² every 8 weeks versus observation. The duration of maintenance rituximab was capped at 2 years for an intended total of 12 infusions. In total, 1018 patients were randomized to either observation (n = 513) or MR (n = 505). With a median follow-up of 36 months, progression-free survival (PFS) was significantly improved in the MR arm (75% vs. 58%, p < 0.001) and provided benefit across subgroups stratified by age, gender, Follicular Lymphoma International Prognostic Index (FLIPI) score and type of response to induction. There was no difference in overall survival, which exceeded 95% in both groups. Importantly, quality of life questionnaires were included, and did not show a difference by treatment arm. The PRIMA trial was the first prospective randomized trial to show a PFS benefit in patients receiving rituximab as part of chemotherapy induction, and has likely profoundly impacted practice patterns.

The intent of the current analysis was to determine whether MR following chemoimmunotherapy induction, as delivered in the PRIMA trial, was cost-effective with respect to PFS and overall survival (OS), and focusing on cost per life-year gained (LYG) and cost per quality adjusted life-year (QALY) gained. In creating their model, the authors extrapolate from the PRIMA population and further apply several assumptions, including that patients will not receive all intended doses of MR, only one computed tomography (CT) scan is performed annually, quality of life declines with disease progression, the difference in PFS curves in the PRIMA persists for 6 years, and that progression to a more advanced stage has higher mortality. With these assumptions, the model projects an improved PFS of 1.5 years and improved OS of 1.21 years, both in favor of MR. In addition, the authors find a decreased QALY in observation patients due to an assumed decline in quality of life following first progression. The main conclusion is that patients on MR for 2 years incur an increased cost of $US38 545, which is offset by substantial gains in LYG and QALY, and that MR is thus cost-effective in this setting.

Although the findings are intriguing, several items are worth consideration. First, all patients had symptomatic, high tumor burden FL treated with one of three chemoimmunotherapy regimens already mentioned. Improvement in the induction regimen could attenuate the impact of MR, and newer regimens such as bendamustine plus rituximab [5] might diminish the PFS and OS advantages of MR calculated in this model. Second, multiple maintenance schedules exist for which the cost-effectiveness may be different, including one dose every 3 months, four weekly doses repeated every 6 months, and “extended induction” regimens [6–10]. To date, there are no comparative data showing the efficacy of any one MR regimen, and differing PFS outcomes would likely affect the cost-effectiveness as determined in this report. Compounding the choice of MR regimens is the uncertainty of how much maintenance is enough, and it is important to note that the current cost analysis assumed that MR ceased at 2 years. The continued cost-effectiveness beyond 2 years remains unknown. Another aspect is the clinical cost of toxicity associated with MR for patients. Only grade 3 and 4 toxicities with a greater than 2% difference were included in the current QALY model, whereas, for many patients, a persistent and unrelenting grade 1 or 2 infection or fatigue is likely to impact both quality of life and healthcare expenditures.

An important caveat is that most attempts to dissect the relative benefit of MR versus retreatment have not shown a survival advantage to MR [8,11], and none of the investigations thus far have been in patients with high tumor burden. Many have taken the lack of a survival benefit in the face of PFS differences to imply that salvage strategies (i.e. rituximab at relapse) are effective and can compensate for early progression in observation patients. To date, there is no evidence that MR following frontline therapy improves overall survival, although a recent meta-analysis shows a significant survival advantage to MR in patients with relapsed FL [12]. The authors’ calculated prediction of improved overall survival is intriguing, but not yet proven in a clinical setting.

Finally, the cost-effectiveness of this analysis is, by intent, limited to frontline patients with FL treated per specific criteria and per a limited time of 6 years of follow-up. It may be worthwhile to take a step back and consider the entire patient lifetime in which he/she is likely to receive repeated rounds of rituximab-containing therapy with or without additional years of maintenance rituximab, all extending over at least 10 years. At some point, most FL will acquire a resistant phenotype, and we are likely at least a decade away from being able to absorb the full impact of frontline treatment choices on subsequent therapies from a population cost perspective.

In summary, the study by Hornberger and colleagues provides a unique analysis supporting what many practitioners already implement. Their modeled calculations find that 2 years of MR following chemoimmunotherapy induction is cost-effective and improves outcomes. With the cautions discussed above, this report is nevertheless timely, and supports that MR is both clinically and economically a viable option for frontline patients with FL.

Potential conflict of interest

Disclosure forms provided by the authors are available with the full text of this article at www.informahealthcare.com/lal.