Early hemoglobin response and alternative metrics of efficacy with erythropoietic agents for chemotherapy-related anemia

Abstract

Objective: To examine associations between early hemoglobin response and alternative measures of efficacy following treatment with an erythropoietic agent for chemotherapy-related anemia.

Research design and methods: Preliminary data from an ongoing randomized, multicenter, 16-week, open-label clinical trial of epoetin alfa versus darbepoetin alfa were used to dichotomize patients based on attainment of early hemoglobin response (> or = 1 g/dL increase in hemoglobin level within 4 weeks of treatment initiation). Measures of efficacy were compared between patients with early hemoglobin response and those without. Sensitivity analyses were then performed to evaluate the impact of various methods for handling censored data and hemoglobin values following blood transfusion.

Main outcome measures: Efficacy measures included: the proportion of patients with a > or = 1 g/dL increase in hemoglobin by 4 weeks or a > or = 2 g/dL increase by 8 weeks; mean hemoglobin levels at 4, 8, 12, and 16 weeks; area under the curve for change in hemoglobin level; proportion of patients who required a blood transfusion after 4 weeks; proportion of follow-up days on which patients had hemoglobin levels within the therapeutic range of 11 g/dL to 13 g/dL; and proportion of patients who never had a hemoglobin level within this range.

Results: A total of 274 patients were included (66.1% female, mean age 62.4), of whom 48.9% had an early hemoglobin response and 51.1% did not. Mean duration of follow-up was 10.1 +/- 5.05 weeks. All metrics indicated superior longer-term response among patients with early hemoglobin response compared to patients without early response. The findings were robust across sensitivity analyses. Although the analysis establishes a significant relationship between early hemoglobin response and alternative efficacy metrics, causality cannot be inferred.

Conclusions: Early hemoglobin response is significantly associated with various metrics of clinical response to erythropoietic agents and is an appropriate measure for evaluating treatment effects.