Treatment of Diffuse Large B-Cell Lymphoma (DLBCL) Patients (pts) Age 80 and Older: Analysis of the Veterans Health Administration (VHA) National Database

The incidence of DLBCL increases with age and the population ≥80 is the fastest growing segment of the elderly population. Optimal treatment of these very elderly DLBCL pts remains unclear. While chemo-immunotherapy can be curative, benefits of treatment must be balanced against toxicities in older, frail pts. Previous studies have suggested that doxorubicin treatment is associated with better survival in the very elderly population, though only on univariate analyses. To better understand the relationship between treatment and survival in pts ≥80, we performed a retrospective analysis of veterans to examine the relationship between treatment and overall survival (OS).

We identified 523 DLBCL pts ≥80 diagnosed between 1998 and 2008 in the VHA Central Cancer Registry. We excluded pts with primary cutaneous or central nervous system DLBCL and pts treated with unknown agents outside the VHA. Data on stage, LDH, co-morbidities, date of death, treatment and supportive care medications were obtained for each pt. Treatment related mortality (TRM) was defined as death in the 30 days following any cycle of therapy. OS was defined as time from diagnosis to death. ECOG Performance Status (PS) at diagnosis was either obtained from pt charts or assessed from physician, nursing, and physical therapy notes available for each patient.

Of the 476 pts that met inclusion criteria, 193 received at least one dose of doxorubicin therapy (DT), 92 were treated without doxorubicin (non-DT) and 191 received no system therapy (NST). Median OS times were 30.2 months, 12.3 months, and 1.9 months in the DT, non-DT, and NST groups, respectively (Log-rank p<0.001). Baseline characteristics of 285 treated patients presented in Table 1. Among the 273 pts in whom treatment dates were available, 48 (18%) experienced TRM. Of these 48 deaths, 32 (67%) were associated with the first treatment cycle. On multivariate analysis, rituximab significantly reduced the risk of death (HR .62, CI .44 - .87), while GCSF use demonstrated a trend towards decreased death (HR .76, CI .57 – 1.03). Comorbidities (HR 1.1, CI 1.02 – 1.19) and PS were associated with increased risk of death. PS of 2–4 was associated with a marked increased risk of death within the first 45 days after diagnosis (HR 19.7, CI 2.6– 146) and a lesser risk after 45 days (HR 1.9, CI 1.4 –2.6). After controlling for other variables, doxorubicin use was no longer significantly associated with OS (HR .87, CI .64 – 1.17), suggesting the OS difference noted between the DT and non-DT groups was primarily due to differences in baseline patient characteristics.

While we acknowledge the limitations of analyses of observational data, the observed 18% rate of TRM (compared to historical TRM rates of 4–6% seen in trials of R-CHOP) suggests that standard treatment paradigms are too toxic for many patients in this age group. The lack of association between doxorubicin and OS calls into question the importance of this drug in this patient population. In the absence of better risk stratification, treatment with less toxic regimens- either through attenuated doses or different drugs- may be appropriate in this population.

Carson:Genentech: Consultancy, Honoraria, Speakers Bureau. Nabhan:Genentech: Research Funding, Speakers Bureau.