Disease Characteristics, Patterns of Care, and Survival in Very Elderly Patients with Diffuse Large B-Cell Lymphoma

Context: Despite having the highest incidence of diffuse large B-cell lymphoma (DLBCL), individuals older than 80 years are rarely included in DLBCL clinical trials or epidemiological studies. We sought to better characterize DLBCL presentation, treatment, and survival patterns for this age group.

Objective: We investigated demographic and clinical characteristics at diagnosis, treatment selection factors, and the impact of treatment regimen on overall survival (OS) and lymphoma-related survival (LRS) for DLBCL patients >80 years. We hypothesized that patients >80 years were more likely to undergo observation and less likely to receive standard-of-care rituximab, cyclophosphamide, doxorubicin, vincristine, and prednisone (R-CHOP). We also hypothesized that patients >80 years who received R-CHOP would have superior OS and LRS, even after controlling for demographic and clinical factors.

Methods: We used the Surveillance, Epidemiology, and End Results (SEER)-Medicare linked database to examine DLBCL patients diagnosed from 1999-2009 and followed through 2010. Our population-based cohort contained 5,924 DLBCL patients aged ≥66 years; 1,422 were >80 years. Only patients treated within 6 months of diagnosis with R-CHOP; cyclophosphamide, doxorubicin, vincristine, and prednisone (CHOP); cyclophosphamide, vincristine, and prednisone (CVP); rituximab plus CVP (R-CVP); or patients undergoing observation were included in order to examine factors associated with the use of anthracyclines. Chi-squared tests compared characteristics and initial treatments of DLBCL patients >80 years and 66-80 years. Multivariable logistic regression models examined treatment selection factors in patients >80 years. Standard and propensity score-adjusted multivariable Cox proportional hazards models adjusted for patient demographics, clinical characteristics, comorbidities, performance status, and year of diagnosis examined relationships between treatment regimen, treatment duration, and survival.

Results: Among patients >80 years, 58% were female, 91% were Caucasian, 36% had stage III/IV disease, 39% had extranodal involvement, 7% had B-symptoms, 28% had poor performance status, and 14% had ≥2 comorbidities. Patients >80 years were less likely to receive R-CHOP (43% vs. 61%) and more likely to be observed (30% vs. 15%) or receive R-CVP (12% vs. 4%); all p<0.0001.

Sex, marital status, area-level poverty, year of diagnosis, performance status, and disease site were associated with R-CHOP treatment in patients >80 years. The initial receipt of R-CHOP was more commonly associated with female sex (odds ratio (OR) 1.31, 95% confidence interval 1.01-1.71), being married (OR 1.69, 1.07-2.66) and a diagnosis after 2001 (OR for 2002 11.71, 6.32-21.70; persistently increased ORs thereafter). The initial receipt of R-CHOP was less commonly associated with extranodal disease (OR 0.71, 0.55-0.91), poor performance status (OR 0.57, 0.44-0.75), and residence in a census tract with >12% of residents living in poverty (OR 0.69, 0.50-0.96). Initial observation was more commonly associated with the same factors that were less commonly associated with R-CHOP use and was less commonly associated with stage III/IV disease (OR 0.66, 0.50-0.87).

Kaplan-Meier survival curves revealed that in patients >80 years, R-CHOP was associated with the best OS and LRS. Multivariable Cox proportional hazards models revealed that R-CHOP for >4 cycles was associated with the best OS in patients >80 years of all stages (hazard ratio (HR) 0.48, 0.37-0.62). Among stage III/IV patients, R-CHOP for >4 cycles (HR 0.48, 0.31-0.72) and R-CVP for >4 cycles (HR 0.40, 0.21-0.76) demonstrated significantly longer OS.

Conclusions:Although DLBCL patients >80 years were less likely to receive R-CHOP, this regimen conferred the best survival. The failure of very elderly DLBCL patients to receive R-CHOP may occur due to clinical factors such as poor performance status, but commonly varies across demographic factors such as area-level poverty, which may reflect bias in the under-utilization of R-CHOP in very elderly patients that is not based on clinical parameters. Further studies are needed to characterize the impact of DLBCL treatment on quality of life in very elderly patients, and algorithms should be developed to help guide therapy in this population.