Correction of anemia has been linked to enhanced benefit from radiotherapy in patients with cancer, but the potential benefit of erythropoietin on cancer control has not been established. Henke and colleagues studied the effect of erythropoietin on cancer control and survival in more than 300 patients undergoing radiotherapy for head and neck cancer.

The double-blind, randomized, placebo-controlled, multicenter trial included 351 adult patients who received definitive or postoperative radiotherapy for squamous cell carcinoma of the oral cavity, oropharynx, hypopharynx, or larynx at 23 European medical centers between 1997 and 2001. The patients all had advanced disease (T3, T4, or nodal involvement), hemoglobin levels below 12 g per dL (120 g per L) in women or 13 g per dL (130 g per L) in men, and Karnofsky scores of at least 60. Exclusion criteria included thrombocytosis, treatment-resistant hypertension, epilepsy, and treatment with certain drugs. Before randomization, patients were stratified according to tumor resection status.

Subcutaneous epoetin beta (300 IU per kg) or placebo was administered subcutaneously three times weekly beginning 10 to 14 days before radiotherapy. Treatment was continued throughout radiotherapy or until target hemoglobin concentrations were achieved (at least 14 g per dL [140 g per L] in women and 15 g per dL [150 g per L] in men). Patients who had transferrin saturations below 25 percent also received oral or intravenous iron.

The 171 patients assigned to placebo were similar to the 180 patients assigned to epoetin-beta therapy, except that there were more smokers and patients with relapsed cancer in the epoetin-beta treatment group. Patients were assessed six weeks after completing radiotherapy and then every three months thereafter. The primary outcomes monitored were locoregional progression-free survival and death. Adverse events, changes in hematologic values, and time to change in tumor status also were monitored.

Target hemoglobin concentrations were achieved in 148 (82 percent) of the patients treated with epoetin beta, compared with 26 (15 percent) of the patients who received placebo. In patients treated with epoetin beta, mean hemoglobin concentrations increased for up to six weeks and then remained stable. Overall, 208 patients (59 percent) experienced locoregional tumor progression or died during follow-up. The adjusted median locoregional progression-free survival was 745 days in the placebo group, but 406 days in the epoetin-beta–treated group.

Multivariate analysis to adjust for tumor stage, treatment stratum, smoking status, tumor site, hemoglobin concentration, relapse status, and days between starting drug therapy and beginning radiotherapy concluded that the relative risk of unfavorable outcome was increased for epoetin-beta therapy compared with placebo (relative risk: 1.2). Subgroup analysis found that poorer outcome with epoetin-beta therapy was significant only in patients younger than 60 years, patients with a baseline hemoglobin level of at least 11 g per dL (110 g per L), and patients with advanced cancer or cancer of the hypopharynx.

The rates of cancer-related and noncancer-related adverse events were almost identical in the treatment and placebo groups. Overall, 52 percent of placebo-treated patients and 61 percent of epoetin-beta–treated patients died, with 34 percent of the deaths attributed to cancer. Treatment-related adverse events occurred in 6 percent of placebo-treated patients and 8 percent of epoetin-beta–treated patients.

The authors found that although epoetin-beta therapy improved hemoglobin concentrations, patients did not benefit in terms of disease progression or survival. They stress that erythropoietin may affect specific cancers or patient subgroups in different ways. Erythropoietin is known to promote angiogenesis and growth of certain breast cancers. Conversely, low hemoglobin concentrations reduce tumor oxygenation and increase tumor hypoxia, but might decrease radiosensitivity. Previous studies on erythropoietin focused on quality of life in patients receiving palliative cancer therapy. These studies had methodologic problems and inconsistent results.

The authors conclude that erythropoietin may impair cancer control and adversely affect survival in some patients undergoing radio-therapy. They call for further trials to confirm their findings and to explain the underlying mechanism of the hemoglobin effect.

editor's note: This study raises concerns and choices for patients, families, physicians, and even drug companies. The authors are careful to present their results objectively, and all (sponsor, journal, and authors) are to be commended for publishing what could be interpreted as “negative” results in a drug company–sponsored clinical trial.

The poorer outcomes in the treated patients appear counterintuitive. Yet if some cancers have erythropoietin receptors, erythropoietin therapy could directly benefit cancer cells, as well as indirectly benefit all active cells through rising hemoglobin levels. In selected cases, could patients be asked to trade feeling worse because of anemia for improved long-term survival? Clearly this would be an individual decision, driven by the values of the patient and those closest to him or her.

A clear distinction has to be made between patients receiving palliative therapy, for whom the quality of remaining life is of paramount consideration, and patients aiming for curative treatment or improved long-term survival. In this complex and emotional situation, one wonders if advertisements for erythropoietin need to be modified to incorporate the results and potential implications of this study.—a.d.w.