1499P - Effectiveness of darbepoetin alfa (DA) for chemotherapy-induced anaemia (CIA) when initiated at haemoglobin (Hb) ≤10 g/dL

Date 28 September 2014
Event ESMO 2014
Session Poster Display session
Topics Complications of Treatment
Supportive Care
Presenter Robert Pirker
Citation Annals of Oncology (2014) 25 (suppl_4): iv517-iv541. 10.1093/annonc/mdu356
Authors R. Pirker1, J.F. Vansteenkiste2, M. Hedenus3, E. Hernandez4, L. Belton5, J. Terwey6
  • 1Department Of Medicine, Medical University of Vienna, AT-1090 - Vienna/AT
  • 2Respiratory Oncology Unit (pulmonology), University Hospital KU Leuven, 3000 - Leuven/BE
  • 3Department Of Medicine, Sundsvall Hospital, Sundsvall/SE
  • 4Department Of Obstetrics, Gynecology And Reproductive Sciences, Temple University Hospital, Philadelphia/US
  • 5Biostatistics, LB Biostatistics, London/GB
  • 6Research And Development, Amgen (Europe) GmbH, Zug/CH



In 2008, the label for DA was updated, decreasing Hb treatment initiation and discontinuation thresholds to ≤10 and >12 g/dL, respectively. Few published placebo-controlled data are available on the effectiveness of DA initiated at Hb ≤10 g/dL. Pooled and meta-analyses were conducted to assess DA effectiveness when initiated as per current labelling.


Data for patients with cancer and CIA who initiated DA at Hb ≤10 g/dL were extracted from a database of Amgen-sponsored trials. A comparative analysis was limited to randomised, controlled trials in patients treated with DA or control (placebo/best supportive care). Data for the DA arm(s) of randomised, multiple-arm or prospective, single-arm trials were also extracted (DA-only analysis; non-front loaded studies only). Outcomes included Hb increase ≥1g/dL or ≥2g/dL during the first 12 weeks of treatment. Crude and Kaplan–Meier (KM) proportions of subjects who experienced each outcome, and time (days) to each outcome were summarised by treatment arm. Meta-analysis (fixed effect inverse-variance method) was performed to compare outcomes for DA versus control.


The comparative analysis included four studies (two in lung cancer, one in lymphoproliferative disease and one in non-myeloid malignancy): DA, n = 261; control, n = 273. DA-only analysis included 15 studies (n = 3768). Results of the crude and KM analyses are shown in the Table; data are presented as percentages of patients (95% confidence interval [CI]). Hazard ratios (95% CI) for patients with an Hb increase of ≥1 g/dL and ≥2 g/dL for DA vs. control were 2.07 (1.62–2.63) and 2.91 (2.09–4.06), respectively. Median (95% CI) time to a ≥1 g/dL increase was 43 (37,50) days for DA and was not achieved for placebo. First quartile time to a ≥2 g/dL increase was 43 (40–50) days and 83 (76–not evaluable) days, respectively.

Hb change % (95% CI) DA (n = 261) Placebo (n = 273) DA only (n = 3768)
ΔHb ≥1 g/dL
Crude 66 (60–71) 40 (35–46) 67 (65–68)
KM 76 (69–83) 48 (40–57) 82 (80–85)
ΔHb ≥2 g/dL
Crude 45 (39–51) 19 (15–24) 43 (42–45)
KM 58 (49–68) 27 (18–37) 60 (57–63)


Our study suggests that DA effectively increases Hb levels and reduces time to Hb increase in patients for whom treatment was initiated at Hb ≤10 g/dL, as per current product labelling.


M. Hedenus: Minor honoraria for temporary advisory boards for Takeda, Pharmacosmos,and Vifor; L. Belton: Contractor funded by Amgen; J. Terwey: Employee of Amgen (Europe) GmbH. All other authors have declared no conflicts of interest.