Racial disparity in breast cancer survival: the impact of pre-treatment hematologic variables.

PURPOSE: A survival disparity of black versus white breast cancer patients has been extensively documented but not adequately explained. Blacks and whites also have significant differences in hematologic traits including hemoglobin (HGB). However, a link between survival disparity and hematologic differences has not been reported. We aimed to explore the effect of pre-treatment hematologic variables on this survival disparity. METHODS: We sequentially matched 443 black patients, using a minimum distance approach, to four different sets of 443 whites on demographics (age, year of diagnosis, smoking, and drinking status), tumor presentation (all demographic variables plus tumor stage, grade, and hormone receptor status), treatment (all presentation variables plus surgery, chemotherapy, radiation therapy, and hormone therapy), and presentation plus pre-treatment hematologic variables. Racial survival for each matched dataset was analyzed by Cox proportional hazards model. RESULTS: We found that white patients matched on demographic characteristics had more favorable survival than blacks [hazard ratio (HR) 0.57, 95 % confidence interval (CI) 0.42-0.77, p log-rank = 0.0002]. Presentation match diminished this disparity [HR 0.72 (0.54-0.95), p log-rank = 0.0199], which was not further reduced in treatment match [HR 0.73 (0.55-0.96), p log-rank = 0.0249]. However, the survival disparity was largely reduced when pre-treatment level of HGB or red blood cell distribution width was further matched in addition to presentation match [HR 0.83 (0.64-1.09), p log-rank = 0.1819 and HR 0.83 (0.64-1.09), p log-rank = 0.1760, respectively]. CONCLUSIONS: We found that in our patient population, differences in tumor presentation and certain pre-treatment hematologic traits, but not treatment, were associated with the survival disparity between black and white breast cancer patients.

Duke Scholars

Author Terry Hyslop Biostatistics & Bioinformatics, Division of Translational Bi ...