Purpose: Previously reported association of anemia with shorter survival in newly diagnosed metastatic prostate cancer may simply reflect extent of disease. The impact of anemia on response to androgen deprivation is not known. We examined the prognostic value of anemia in a multivariate analysis that included disease extent and other tumor and demographic covariates in 957 patients starting hormonal therapy for metastatic prostate cancer as part of Southwest Oncology Group study 8894.
Materials and methods: The multivariate associations of disease and patient measures with anemia (hemoglobin less than 12 gm/dl) were evaluated with a logistic regression model. The associations between hemoglobin and survival and progression-free survival (PFS) were evaluated using a proportional hazards model, and included indicators for quartiles of hemoglobin and baseline covariates. Prostate specific antigen (PSA) normalization (PSAN, or PSA of 4 ng/ml or less) was evaluated with a logistic model.
Results: Quartiles of hemoglobin were 10.1 or less, 10.2 to 12.0, 12.1 to 13.4 and greater than 13.4 gm/dl. In a multivariate model anemia was significantly associated (p <0.02) with being black, performance status 2 to 3 (vs 0 to 1), extensive disease and higher PSA. Anemia was inversely associated with prior therapy with curative intent and with Gleason score 6 to 7 (vs 5 or less), and was not associated with age or bone pain. After adjusting for potential confounders, lower hemoglobin was associated with shorter survival and PFS, and lower likelihood of PSAN with hormonal therapy (p <0.01).
Conclusions: In this newly diagnosed metastatic prostate cancer sample, anemia was common and was associated with shorter survival, shorter PFS, and lower likelihood of PSAN with hormonal therapy after adjustment for disease status and other covariates.