The Relationship Between Socioeconomic Status and Clinical Outcomes in African Americans with Chronic Lymphocytic Leukemia

Epidemiologic data compiled by the Survey Epidemiology and End Results (SEER) identifies important differences in incidence and survival for African Americans with chronic lymphocytic leukemia (CLL). Although the incidence of CLL is lower among African Americans (AA) than Caucasians (Cauc) (4.4 and 6.1 per 100,000 men, respectively), age adjusted survival is inferior. The overall 5-year relative survival for 1999–2006 from 17 SEER geographic areas by race and sex was: 77.0% for Cauc men, 81.1% for Cauc women, 62.4% for AA men, and 68.3% for AA women; hence, AA CLL patients have an approximate 15% greater absolute risk of CLL-related death in the first 5 years after diagnosis. We have undertaken biologic and genetic investigations to identify factors that contribute to poorer outcomes among AA patients with CLL. A central unanswered question is whether the differences in biology and genetics are the primary cause of inferior overall survival, or whether differences in socioeconomic status (SES) also play a significant role in clinical outcomes.

The Duke University CLL database contains information on greater than 550 CLL patients at Duke University Hospital and the Durham Veterans Affairs Medical Center. We performed a detailed retrospective analysis of this cohort, using average income by ZIP code as a surrogate marker for SES. This was compared with survival data abstracted from the Social Security Outcomes Database. The primary outcome measure was overall survival, and secondary outcome measure was time to treatment (TTT). These outcome measures were compared against average income by ZIP code in addition to other standard validated clinical risk parameters such as race, gender, age, and Rai stage. Statistical analyses were performed using JMP software, version 9.0.1 (SAS, Cary, NC). Average income by ZIP code is publicly available information via IRS tax returns; we used the average adjusted gross income from 2007.

Mean household income, as estimated by ZIP code, was $43,255/year for AA (n= 37) and $55,353/year for Cauc (n= 363) (p<0.0001). For univariate analyses, Kaplan-Meier survival curves were generated for age, race, gender, and average income by ZIP code. Mean survival was 8.5 years for AA, and 17.9 years for Cauc (p<0.0001, Log-Rank). Mean survival was 18.5 years for those under the age of 60, and 14.8 years for those 60 and above (p<0.0001, Log-Rank). Mean survival was 18.6 years for females, and 15.9 years for males (p= 0.25, Log-Rank). For univariate analysis, average income by ZIP code was divided into quintiles, and mean survival was 16.0, 13.4, 16.2, 18.0, and 13.9 years, in order of ascending incomes (p= 0.73, Log-Rank). Mean TTT was 2.7 years for AA, and 4.5 years for Cauc (p= 0.0004, t-test). A multivariable analysis using a Cox proportional hazards model showed that the following prognostic variables (followed by respective hazard ratios (HR) and p value) were independent predictors of shorter overall survival: AA race (versus Cauc race; HR= 3.6, p= 0.0005), age (examined as a continuous variable; HR= 1.07, per unit change in regressor, HR= 41.1 per change in regressor over entire range, p= <0.0001), Rai stage (HR varies depending on stages under comparison, p= 0.003), and male gender (versus female; HR= 1.9, p= 0.015). Mean income by ZIP code was not an independent predictor of shorter overall survival (examined as a continuous variable; HR= 0.99, per unit change in regressor, HR= 0.53 per change in regressor over entire range, p= 0.39).

AA patients with CLL have poorer overall survival than Cauc. We have examined whether SES plays a role in these inferior outcomes. While the AA patients in our cohort have lower average household incomes, this was not an independent predictor of survival in a multivariable model. We conclude that clinical and biologic, rather than socioeconomic, factors are primary determinants of inferior clinical outcomes. However, SES is intrinsically a difficult construct to measure, and it is possible that using average incomes by ZIP code does not accurately approximate socioeconomic status.

No relevant conflicts of interest to declare.