Abstract
Introduction: WM is a rare B-cell lymphoma affecting older patients, and its management relies on limited evidence. Efficacy of new agents (rituximab, R, bortezomib, BOR, bendamustine, BEND) has been reported in phase II clinical trials, but data on treatment patterns and outcomes in the population are lacking. Our objective was to analyze trends in the management and outcomes of WM using the linked Medicare and Surveillance, Epidemiology and End Results (SEER-Medicare) data.
Methods: We selected patients aged ≥65 years, diagnosed with WM (including lymphoplasmacytic lymphoma) between 1994 and 2011, who had complete Medicare records up to December 2013. Using claims for related health services, we determined indicators of WM severity at diagnosis (anemia, neuropathy, etc.), date and type of the initial chemotherapy. Regimens were classified as purine analogue-(PUR), alkylator-based (ALK), BOR or BEND, with or without R. Trends in proportions were studied by log-binomial regression, factors associated with treatment type by mixed-effects logistic models, and overall survival (OS) by proportional hazard models, reporting, respectively, annual percent change (APC), odds ratios (OR) or hazard ratios (HR), with 95% confidence intervals (CI). Mean Medicare payments were estimated using inverse probability of censoring, and were inflation-adjusted to 2013 dollars.
Results: Among the 2,666 WM patients, median age was 78 years, and 57% of cases were male. Prevalence of recorded WM complications at diagnosis was as follows: anemia, 46%; transfusions, 6%; autoimmune hemolytic anemia (AIHA), 3%; neuropathy 5%; plasmapheresis, 2.5%; amyloidosis, <1%. Between 1994 and 2011, the prevalence of baseline anemia (APC, +1.5%, P=.001), transfusions (APC, +7.7%, P<.0001)and neuropathy (APC, +4.2%, P=.022) significantly increased, as did the proportion of patients receiving chemotherapy within 1 year of WM diagnosis (from 39% in 1994/99, to 56% in 2009/13, APC, +2.7%, P<.0001), although mean age did not change (P=.62). Within 10 years from diagnosis, 67% of patients started chemotherapy and 28% died without treatment. Median time to chemotherapy was 9.5 months (CI, 7.8-11.4), decreasing from 29 months in 1994/99 to 4 months in 2009/13.
In the pre-rituximab era (1994/99), 53% of first-line regimens were PUR-based, and 29% were ALK-based (Fig. A). By 2005/08, 54% of patients received R alone, and further 32% R with chemotherapy. Since 2009, 56% have received R alone, 10% BOR (+/-R) and 5% BEND (+/-R), while only 8% got PUR (+/-R) and 18% ALK (+/-R). Single-agent R was more often used in patients >80 years old (OR, 1.97, CI, 1.30-2.97), women (OR, 1.33, CI, 1.01-1.74), those with neuropathy (OR, 1.87, CI, 1.18-2.96) or AIHA (OR, 2.13, CI, 1.01-4.49), and was dependent on physician's preference (intraclass correlation, 18%, CI, 8-35%, P<.0001).
Median OS was 4.6 years (CI, 4.4-4.9) from diagnosis. Adjusting for baseline risk factors, OS progressively improved, with HR of 0.76 (CI, 0.67-0.87) for patients diagnosed in 2000/04 (compared with 1994/99), 0.65 (CI, 0.56-0.76) for 2005/08, and 0.48 (CI, 0.40-0.58) for 2009/2011. Presence of anemia, transfusions, plasmapheresis or amyloidosis at diagnosis were associated with worse OS. After adjusting for immortal-time bias, patients treated within 1 year from diagnosis had a worse OS than those who were not (HR, 1.17, CI, 1.05-1.30). Among patients treated in 2005-2013, administration of R was associated with OS advantage (HR, 0.68, CI, 0.52-0.88, Fig. B). Mean Medicare costs for all health care claims from WM diagnosis were $164,095 per patient, and mean costs for chemotherapy were $44,286. Mean Medicare chemotherapy-related costs accrued during the first year of treatment increased from $7,538 in 1994/99 to $29,469 in 2009/13.
Conclusions: The increasing proportion of WM patients receiving chemotherapy within 1 year of diagnosis may reflect changes in the diagnostic definition of WM or in the management paradigms. OS improvement over time occurred despite increasing prevalence of anemia, neuropathy and early treatment, and it may be related to better, albeit substantially more expensive therapy. After a dramatic shift since 2000, a majority of patients now receive R alone, while the PUR-based therapy predominating in the 1990's is rarely used. Further research may elucidate how much the introduction of novel agents (R, BOR, BEND) contributed to improved OS.
Olszewski:Genentech, Inc.: Research Funding; Bristol-Myers Squibb, Inc.: Consultancy. Off Label Use: Rituximab, bortezomib, bendamustine-for treatment of Waldenstrom macroglobulinemia.
