There is no role for ESAs in cancer patients receiving chemotherapy with a normal Hgb level, as data have shown adverse responses in terms of OS when ESAs are used in non-anaemic patients.³² The Breast Cancer Erythropoietin Survival Trial (BEST) randomised non-anaemic patients with metastatic breast cancer receiving chemotherapy to epoetin alpha or placebo, targeting Hgb levels of 12–14g/dl. The study was terminated prematurely as early analysis demonstrated a trend of declining one-year survival in the epoetin alpha group (70%) compared with the placebo group (76%) (p=0.012).³²

In the Danish Head and Neck Cancer Group (DAHANCA) 10 study, 522 patients were randomly assigned to receive radiotherapy plus darbepoetin alpha or radiotherapy alone in head and neck cancer. The target Hgb was 14–15g/dl. The study was terminated early because preliminary analysis demonstrated a significantly decreased five-year locoregional tumour control in the darbepoetin arm compared with the control group (RR 1.44; p=0.02). OS was lower in the ESAs group, but did not reach statistical significance (RR 1.28; p=0.08).³⁸

Table 1: Guidelines for ESAs

The Erythropoietin in Head and Neck Cancer (ENHANCE) trial randomised non-anaemic patients with head and neck cancer receiving radiotherapy to either placebo or pegzerepoetin alpha, aiming for an Hgb level of 14–15g/dl. The pegzerepoetin alpha group had a substantially shorter locoregional progression-free survival (PFS) (hazard ratio [HR] 1.62; p=0.0008) and shorter time to locoregional progression (HR 1.69; p=0.007), as well as decreased OS, compared with the placebo group (HR 1.39; p=0.02).³³

The Preoperative Epirubicin Paclitaxel Aranesp Study (PREPARE) trial randomised 733 patients with breast cancer receiving neoadjuvant chemotherapy to placebo or darbepoetin alfa. Hgb was maintained between 12.5 and 13g/dl. Patients who received ESAs were found to have shorter three-year OS (86 versus 90%) and relapse-free survival rate (72 versus 78%) than those on placebo. Moreover, tumour progression was faster in the patients treated with darbepoetin alfa.⁵¹

In another randomised trial, by the National Cancer Institute (NCI) Gynecologic Oncology Group (GOG), 114 patients with advanced cervical cancer receiving concurrent cisplatin and radiotherapy were randomly assigned to receive ESAs or transfusion when the Hgb concentration was ≤10mg/dl. The study was terminated early when a data analysis showed lower three-year OS (59 versus 62%) and PFS (61 versus 71%) in patients treated with epoetin alpha compared with the standard of care.³⁹

Despite the uncertain effect of ESA therapy on survival rates that were reported in earlier meta-analyses,²⁴ the result of a more recent large metaanalysis, which was obtained from 51 phase III trials conducted between 1985 and 2008, has established a statistically significant increased mortality in patients receiving ESA therapy (HR 1.10, 95% CI 1.01–1.20; p=0.03).³⁵