Survival After Rituximab-Based Chemoimmunotherapy In Older Patients With Follicular, Nodal Marginal Zone and Small Lymphocytic Lymphoma: Analysis Of SEER-Medicare Data

Abstract

BackgroundIndolent B-cell lymphomas differ in clinical features yet treatments for the nodular marginal zone (NMZL) and small lymphocytic (SLL) subtypes often follow paradigms for follicular lymphoma (FL). In predominantly younger FL patients, the FOLL05 randomized trial (Federico et al., J Clin Oncol 2013;31;1506) demonstrated differences between different rituximab-based regimens in progression-free but not overall survival (OS). The objective of our study was to evaluate heterogeneity of survival outcomes after initial chemotherapy in older patients with FL, NMZL and SLL using a population-wide database. MethodsWe extracted clinicopathologic, treatment and survival data on FL, NMZL and (nodal) SLL cases diagnosed between 1996 and 2009 from the linked Surveillance, Epidemiology and End-Results (SEER)-Medicare database. We excluded patients diagnosed at age <65 years (y) or with incomplete Medicare claims. Standard chemotherapy regimens were classified based on specific drug administration as: rituximab alone (R), R with cyclophosphamide/vincristine (RCVP), with additional doxorubicin (RCHOP) and R with fludarabine (RF, including combinations with cyclophosphamide, RFC, and mitoxantrone, RFM). Overall survival (OS) was compared in multivariate proportional hazard models. Rates of adverse events were compared in Poisson models using inpatient claims and diagnoses within 90 days of treatment initiation. ResultsWe identified 24,714 patients with a median age of 75y of whom 71% had complete Medicare claims. Median survival was 6.7y in FL, 6.1y in NMZL and 4.9y in SLL. Rituximab-based regimens were administered in 6,993 patients at median 2 months since diagnosis. The most common therapy was single-agent R (40%) followed by RCHOP (26%). The median number of cycles was: 6 for RCHOP/RCVP, 5 for RFM, and 3 for RF/RFC. After adjustment for multiple clinical and demographic confounders, initial treatment with fludarabine-containing regimens, compared with RCHOP, was associated with significantly worse OS in FL and NMZL (Table). Conversely, in SLL there were no significant differences, even with R alone. The OS advantage of RCHOP in FL/NMZL persisted in alternative, propensity score-based models.The hospitalization rate after RF (27%) was lower than after RCHOP (32%) with a relative risk, RR, of 0.71 (95%CI, 0.61-0.82, P<.0001) in a multivariate model. RF was also associated with a lower risk of infection (RR .61, P<.0001), neutropenic infection (RR .53, P=.003), transfusion (RR .68, P=0.0003) or admission to a critical care unit (RR .68, P=.009), but with a similar risk of cardiac event (RR 1.03, P=.84). All adverse events were significantly less common after R alone (P<0.001). Prophylactic granulocyte growth factor was administered in 42% of first cycles of multi-agent regimens. ConclusionsInitial treatment with fludarabine-based regimens is associated with inferior OS compared with RCHOP in older patients with FL and NMZL despite higher rates of complications during RCHOP therapy. In SLL, different chemoimmunotherapy regimens are associated with similar survival outcomes. Optimal initial treatment choices for indolent B-cell lymphomas in this age group should be individualized, balancing the OS benefits, life expectancy and risks of serious toxicity. Disclosures:No relevant conflicts of interest to declare.TableHazard ratios (HR) with 95% confidence intervals (CI) for OS in multivariate Cox models for each lymphoma subtype.RegimenFL (N=4,807)NMZL (N=843)SLL (N=1,343)NHR (95%CI)PNHR (95%CI)PNHR (95%CI)PRCHOP1,6081 (reference)11711181R1,7191.22 (1.08-1.37).0014771.08 (0.78-1.49).665891.03 (0.78-1.36).82RCVP1,1591.09 (0.96-1.24).181661.15 (0.80-1.67).452380.99 (0.73-1.34).77RF3211.53 (1.29-1.82)<.0001831.88 (1.27-2.79).0023981.09 (0.82-1.45).54