Abstract
Background
The oldest old constitute a large proportion of the total patient (pt) population with FL. Therapeutic decision making in this group is limited by comorbidities, adverse disease and pts' characteristics, potential treatment toxicity, and limited life expectancy. Further, randomized clinical trials have rarely included this pt population. Whether current practice patterns for these pts affect their outcome remains unanswered. Therefore, we aimed to determine treatment selections, patterns of care, prognostic factors, and survival outcomes of first-line management strategies in a large United States (US) based cohort of the oldest old (pts aged > 80 years at diagnosis).
Methods
We used the linked Surveillance, Epidemiology, and End Results -Medicare database to identify 1,878 FL cases in pts > 80 years diagnosed between 1995 and 2009 and focused on the period when rituximab (R) claims occurred. We ascertained first-line management strategies from Medicare claims made within 90 days of diagnosis. We used multiple variable logistic regression models to evaluate the relationship between pt characteristics and the use of two common first-line management strategies—observation (obs) and treatment with R, cyclophosphamide, doxorubicin, vincristine, and prednisone (R-CHOP). We used Kaplan-Meier estimators stratified by stage to evaluate survival functions for first-line management strategies and Cox proportional hazards models adjusted for pt demographics, comorbidity index, disease characteristics, and year of diagnosis to compare the impact of first-line management strategies on survival.
Results
Of the 1,878 oldest adult pts, 63% were female, 95% were white, 2% were African American, 52% had stage III/IV FL, 17% had grade 3 FL, 5 % had B-symptoms, 35% had extranodal involvement, and 14% had a comorbidity index ≥ 2. Common first-line management strategies were: obs, 46%; R, 17%; chemotherapy (chemo) plus R, 11%; chemo, 11%; and radiotherapy (XRT), 11%.
In the cohort of pts diagnosed between 1995 and 2009, obs was more commonly associated with urban pts (ref. less urban/rural pts; OR 1.91; 95% CI 1.15-3.18), and comorbidity index of ≥ 1 (ref. index=0; OR 1.28; 95% CI 1.00-1.64). Obs was less commonly associated with stage III/IV FL (ref. stage I/II; OR 0.67; 95% CI 0.54-0.84), grade 3 FL (ref. grade 1/2; OR 0.35; 95% CI 0.26-0.47), and year of diagnosis (ref. year 1995; OR for 1997 0.23; 95% CI 0.07-0.75; steady decrease thereafter). In the cohort of pts diagnosed between 1999 and 2009, the use of R-CHOP was associated with grade3 FL (ref. grade 1/2; OR 8.20; 95% CI 3.83-17.55) and presence of B-symptoms (ref. absent; OR 4.18; 95% CI 1.81-9.62). R-CHOP use did not vary with year of diagnosis.
The table displays median survival and hazard ratios (HRs) for first-line management strategies. Most favorable outcomes were associated with first-line R-Chemo. Among stage III/IV cases, the least favorable outcomes were observed in the group that received chemo without R. The HRs did not vary with more recent years of diagnosis.
Conclusion
In this largest retrospective analysis of the oldest old US-based FL pts, we demonstrate that first-line R-Chemo is associated with improved survival. Confirmatory prospective studies specifically designed for this pt population are warranted.
CVP-cyclophosphamide, vincristine, prednisone; CHOP- cyclophosphamide, doxorubicin, vincristine, prednisone; R-CVP- rituximab, cyclophosphamide, vincristine, prednisone.
Disclosures:
No relevant conflicts of interest to declare.
First-line management of canine status epilepticus at home and in hospital-opportunities and limitations of the various administration routes of benzodiazepines
