Abstract
Abstract 2891
Hairy cell leukemia (HCL) is a rare B cell malignancy with a variable clinical course. Therapies have changed dramatically in the last 3 decades. Specifically, approval of recombinant alpha interferon (IFN-α) in 1984, followed by the development of purine nucleoside analogues (pentostatin in 1986 and cladribine in 1990), represent significant therapeutic advances, affording high response rates and (using the latter agents) disease control lasting years. Monoclonal antibodies including rituximab -first approved in 1999- have also shown promise, to date mainly in relapsed HCL. However, given the rarity of HCL, large population analyses of survival of affected patients, and survival trends over time, are lacking. Therefore, using the Surveillance Epidemiology and End Results (SEER-17) database, we evaluated overall survival, and factors associated with mortality trends, in HCL patients in the general US population over the past 35 years.
The SEER-17 database was interrogated for HCL patients presenting over the last 35 years, from 1973–2008. 3904 patients were identified; after excluding patients whose survival status was unavailable, 3480 were analyzed. Available demographic information, including age, gender, race, and year of diagnosis, were analyzed. Age was analyzed as a continuous variable. We utilized Cox proportional hazards models to assess prognostic factors. Additional models were formed to evaluate a possible step change in survival in the years following key therapeutic advances (1984, 1990, and 1999). Specifically, the population was separated into four cohorts (pre-1984, 1984–1990, 1991–1999, and 2000–2008) and the entire cohort was included as an additional predictor in the model. Kaplan Meier curves were constructed to estimate overall survival and statistically significant differences between the curves was estimated using the 2 tailed log rank test. The survival curves for the 3 groups post 1984 were plotted along with patients from 1973–1983 and are shown below in Figure 1. The three predictors associated with survival were identified as age, race and year of diagnosis.
The mean age was 56 years (95% CI 33–86) for the entire study population. There was no difference in the mean age before 1984 (56.5 yrs.) compared to the three cohorts after 1984 (mean ages 58 yrs., 57 yrs. and 57 years respectively). The risk of death increased by 5.8% (95% CI 5.14– 6.55) per year with older age at diagnosis. However, after adjusting for age, survival has improved over the study period, as the risk of death in an unselected patient of the same age with HCL in the US has decreased by 90.5%. African-American patients were noted to have a significantly higher mortality than Caucasians (OR 1.942, 95% CI 1.254–3.008). Survival of patients appears to have improved over the years with the risk of mortality decreasing by 6.5% per year (95% CI 5.6%-7.4%) with a later year of diagnosis. With additional models, the risk of death for an individual of a given age decreased by 5.3 % every year with a further reduction of 27.5% occurring in 1984 (p=0.04), consistent with a dramatic reduction in mortality after 1984. Models derived using other landmark years did not show any further step wise reduction in mortality (1990 p=0.441; 1999 p=0.535) at these specific time points. However, the reduction in mortality conferred by diagnosis at a later year has resulted in significantly improved survival being different for all groups before and after 1984 (Figure 1).
- Overall Survival of HCL Cohorts in the US
Survival of modern cohorts of patients with HCL in the US has improved significantly. While the diagnosis of HCL has employed classic morphologic, cytochemical, and (more recently) immunohistologic features, effective antineoplastic therapy has evolved significantly in recent decades. Thus, modern therapies may have played a significant role in the survival improvements of modern HCL patients. African- American ethnicity is associated with higher mortality, a disparity requiring further analysis in the context of patient factors and an improving understanding of HCL pathobiology.
No relevant conflicts of interest to declare.
