Abstract
Introduction: ESAs are associated with an increased and quantifiable risk of venous thromboembolism (VTE); however, the treatment alternative, RBC transfusion, has only a transient effect and the associated risks are often more numerous and less predictable. Repeated transfusions also diminish patients’ personal independence. Transfusion risks include mistransfusions, hemolytic reactions, exposure to infectious agents, immune-mediated reactions, and iron and circulatory overload. Previous studies have suggested that in order to maximally reduce transfusion requirements, ESAs should be initiated when anemia is mild. We conducted a retrospective analysis to assess the effect of hemoglobin (Hb) level at ESA initiation on the risk of transfusions and safety outcomes.
Methods: A combined analysis of patient-level data from randomized, blinded, placebo-controlled studies in which the ESA darbepoetin alfa (DA) was administered subcutaneously in patients with nonmyeloid tumors and CIA (hemoglobin ≤11 g/dL at study screening) was performed. Patients had received at least 1 prior cycle of chemotherapy and had further chemotherapy planned with or without radiotherapy. Endpoints were analyzed by baseline Hb level and included transfusions, embolic/thrombotic events, death, and death or disease progression. Hazards ratios (HR, with 95% confidence interval limits) were used to describe the risk of these endpoints in the DA group relative to placebo.
Results: Patients (n=2112) were mostly white, with a mean age of 62 years. Nearly half were ≥65 years. Baseline characteristics were similar, with the most frequent tumor types in both treatment groups being lung and hematologic cancers, and most with stage III or higher disease. For both groups, patients with lower baseline Hb levels had a higher probability of transfusions; yet the reduced risk of transfusions in the DA group relative to placebo was apparent at each baseline Hb level. The known increased risk of VTEs in the DA group did not appear to be associated with baseline Hb level. There was no increased risk of death or progression at any baseline Hb level.
Conclusions: Results from this patient-level meta-analysis in CIA patients suggest that baseline Hb is a predictor of transfusion risk, and that earlier ESA intervention lowers the risk of future transfusions. Initiating DA at Hb ≥ 9 to < 10 g/dL resulted in a rate of transfusion of 32%; this rate was substantially lowered when DA was initiated at Hb ≥ 10 to < 11 g/dL. There did not appear to be an increased risk of death or VTEs as a result of initiating DA therapy at the higher Hb levels. Therefore, delaying the start of therapy may only increase the transfusion requirements, resulting in increased risk to the patient, and additional strain on the national blood supply.
