The Association of an Early Hemoglobin Response with Alternative Metrics to Evaluate Treatment Efficacy of Erythropoietic Agents.

Abstract Background: Various metrics have been used to evaluate the efficacy of erythropoietic agents for the treatment of chemotherapy-related anemia (CRA) including the proportion of patients attaining a ≥1 or ≥2g/dL increase in hemoglobin (Hb), Hb change from baseline at weeks 4, 8, 12 and 16, area under the change in Hb curve (Hb AUC) and the proportion of patients achieving a target Hb level. Many trial protocols and national treatment guidelines recommend dose escalation of these agents after the first 4 or 6 weeks of treatment if the patient fails to obtain a ≥1g/dL rise in Hb. Studies have also shown that an early Hb response is associated with improved clinical outcomes, i.e. lower transfusion requirements and greater gains in quality of life; but, it is not known how well early Hb response correlates with alternative efficacy metrics. Objective: To compare several alternative measures of treatment efficacy between patients who attained an early Hb response to those who did not while receiving an erythropoietic agent for treatment of CRA. Methods: Using preliminary data from an ongoing randomized, multi-center, 16-week, open-label trial of epoetin alfa versus darbepoetin alfa, we dichotomized patients based on early Hb response (≥1 g/dL Hb rise during the first 4 weeks of treatment) and calculated several alternative measures of efficacy. Only observed, patient-level data were analyzed in the base-case analyses. Sensitivity analyses were performed to evaluate the impact of employing various methods to handle missing data, excluding Hb values within 28 days following a transfusion, limiting analyses to patients with ≥8 or 12 weeks of follow-up, and limiting analyses to Hb values obtained after 4 weeks of treatment. Comparisons were performed using chi-square and Wilcoxon rank-sum tests for categorical and continuous metrics, respectively. Results: 274 patients were analyzed; 134 had an early Hb response, 140 did not. The mean duration of follow-up was 10.1 weeks. All metrics indicated superior longer-term response among patients who had an early Hb response compared to those who did not. These findings were robust across sensitivity analyses. Comparisons of Patients Achieving and Not Achieving Early Hemoglobin Response Comparisons of Patients Achieving and Not Achieving Early Hemoglobin Response Metric Early Response (n=134) No Early Response (n=140) p-value Mean Hb at baseline 10.05 g/dL 10.18 g/dL 0.7178 Mean Hb at 4 weeks 11.58 g/dL 9.89 g/dL <0.0001 Mean Hb at 8 weeks 11.91 g/dL 10.78 g/dL <0.0001 Mean Hb at 12 weeks 12.11 g/dL 11.51 g/dL 0.0105 % with ≥2g/dL rise in Hb by 8 weeks 73.1% 13.6% <0.0001 % requiring transfusion after 4 weeks 6.0% 18.6% 0.0016 Hb AUC4 (baseline to 4 weeks) 3.15 g/dL -0.96 g/dL <0.0001 Hb AUC8 (baseline to 8 weeks) 8.20 g/dL 0.07 g/dL <0.0001 Hb AUC12 (baseline to 12 weeks) 12.25 g/dL 1.99 g/dL <0.0001 Mean % of days within therapeutic range (Hb 11-13 g/dL) 45.4% 24.6% <0.0001 % of patients never within therapeutic range (Hb 11-13g/dL) 10.4% 42.9% 0.0005 Conclusion: An early Hb response is significantly associated with alternative metrics of clinical response to erythropoietic agents and is an appropriate measure to evaluate early treatment effects.