OUP user menu

Randomized, double-blind, placebo-controlled study to evaluate the effect of two dosing regimens of darbepoetin alfa in patients with heart failure and anaemia

Dirk J. van Veldhuisen, Kenneth Dickstein, Alain Cohen-Solal, Dirk J.A. Lok, Scott M. Wasserman, Nigel Baker, Dylan Rosser, John G.F. Cleland, Piotr Ponikowski
DOI: http://dx.doi.org/10.1093/eurheartj/ehm328 2208-2216 First published online: 6 August 2007


Aims Anaemia is common in chronic heart failure (CHF) and associated with worse outcome. This randomized, double-blind, placebo-controlled study evaluated the effect of two darbepoetin alfa dosing regimens on haemoglobin (Hb) rate of rise and clinical effects in patients with CHF and anaemia.

Methods and results Patients with CHF (≥3 months), left ventricular ejection fraction (LVEF) ≤ 40%, and Hb 9.0 to 12.5 g/dL received darbepoetin alfa subcutaneously every 2 weeks for 26 weeks at a starting weight-adjusted dose of 0.75 mcg/kg (n = 56) or a fixed dose of 50 mcg (n = 54), or placebo (n = 55), to gradually achieve and maintain a target Hb of 14.0 ± 1.0 g/dL. Endpoints included rate of Hb rise per week during titration, safety, and changes in 6 min walk distance, New York Heart Association (NYHA) class, LVEF, and quality of life. Most subjects were NYHA class II–III. Mean (SD) age was 71 (11) years, LVEF was 28 (9), and Hb 11.5 (0.7) g/dL. Rate of Hb rise was equivalent between darbepoetin alfa weight-based (+1.87 ± 1.36 g/dL) and fixed dosing (+1.64 ± 0.98 g/dL) groups, vs. + 0.07 ± 1.08 g/dL in the placebo group. Mean Hb concentrations by week 27 were 13.4 and 13.2 g/dL, in the weight-based and fixed dosing groups, respectively. There were non-significant improvements in the combined darbepoetin alfa group vs. placebo for 6 min walk distance (P = 0.074) and Patient’s Global Assessment score (P = 0.057). There was a significant improvement in Kansas City Cardiomyopathy Questionnaire total symptom score (8.2 vs. 1.5 points; P = 0.027) but no change in NYHA class, LVEF, and Minnesota Living With Heart Failure Questionnaire score. Six treatment-unrelated deaths occurred in the 110 darbepoetin alfa treated patients, and none in the 55 placebo treated patients. Other adverse events were similar between groups.

Conclusion In this study of patients with CHF and anaemia, treatment with darbepoetin alfa raised Hb using different dosing regimens. Darbepoetin alfa improved some quality of life indices, but its safety requires further exploration. Larger trials are needed to determine the effects on long-term morbidity and mortality.

  • Anaemia
  • Heart failure
  • Exercise
  • Haemoglobin
  • Trials
See page 2185 for the editorial comment on this article (doi:10.1093/eurheartj/ehm331)


Anaemia is common in patients with chronic heart failure (CHF) and it plays a role in the pathophysiology of the disease.14 Its prevalence increases with the severity of CHF. Using the most commonly employed definition of the World Health Organization (Haemoglobin [Hb] < 12 g/dL in women and < 13 g/dL in men), anaemia is present in ≥ 15–30% of CHF patients. Its pathogenesis in CHF is complex and renal dysfunction, inflammation, changes in plasma volume, hematinic deficiencies, and drug treatment are common causes.15 Anaemia in CHF is independently associated with increased morbidity and mortality.

Treatment of anaemia with erythropoiesis-stimulating proteins (ESP) in patients with CHF would thus seem attractive, and has been explored in a few relatively small studies.69 Silverberg et al.6 were the first to report beneficial effects in two early studies: one open-label study in 26 patients and a subsequent open-label controlled study in which they compared 16 patients who received rHuEPO with 16 untreated patients. Mancini et al.7 studied 26 patients, of whom 15 received subcutaneous rHuEPO. After 3 months, use of rHuEPO was associated with an increased exercise capacity. In two more recent placebo-controlled studies of around 40 patients, beneficial clinical effects were also reported by Palazzuoli et al.8 and Ponikowski et al.9 Although these data seem promising, safety cannot be assessed in small studies. This may be relevant since studies in patients with cancer10 and severe kidney disease or on haemodialysis,11 have associated treatment with rHuEPO with an increase in vascular events.

The aim of this phase 2 study was to evaluate the effects of treatment with darbepoetin alfa, administered subcutaneously in two regimens (weight-based and fixed dosing) vs. placebo, in patients with CHF and anaemia (baseline Hb 9.0–12.5 g/dL) who were iron replete. Darbepoetin alfa stimulates erythropoiesis in the same manner as endogenous erythropoietin, but has greater in vivo biologic activity, allowing for extended dosing intervals.12,13 The primary objective was to examine the effect of two darbepoetin alfa dosing regimens on Hb response. Secondary objectives included the effect on exercise tolerance (6 min walk), left ventricular ejection fraction (LVEF), New York Heart Association (NYHA) functional class, and quality of life. In addition, the safety profile was examined.


Patient population

Patients were eligible if they had symptomatic CHF for ≥ 3 months and a LVEF ≤ 40% as measured by echocardiography, radionuclide ventriculography, or magnetic resonance imaging within 12 weeks prior to screening. They were on optimal medication for CHF, which included an angiotensin converting enzyme (ACE) inhibitor and/or an angiotensin receptor blocker (ARB) and a β-blocker unless intolerant, for ≥ 8 weeks and without dose change for ≥ 4 weeks. Patients’ Hb concentration had to be ≥ 9.0 g/dL and ≤ 12.5 g/dL (mean of 2 screening period measurements) and transferrin saturation (TSAT) ≥ 15.0% at screening. Serum vitamin B12 and folate levels were greater than or equal to the lower limit of the normal range.

Patients were excluded from the study if they had: uncontrolled hypertension (resting blood pressure > 160/100 mmHg), recent (< 3 months) acute coronary syndrome or cerebrovascular accident, hypertrophic (obstructive) cardiomyopathy, active myocarditis, constrictive pericarditis, or severe uncorrected valvular disease. Patients who were potential recipients of an organ transplant or renal replacement therapy < 6 months were excluded, as well as those with chemotherapy and/or radiation therapy within the preceding 12 weeks, or current malignancy, those with active, systemic haematologic disease, and those with anaemia due to acute or chronic bleeding and infection with the human immunodeficiency virus (HIV). Further, uncontrolled atrial fibrillation or flutter, or symptomatic brady- or tachyarrhythmias within 1 month, and significant renal dysfunction (serum creatinine level > 3.0 mg/dL) were not allowed. We also excluded patients who had received ESP therapy or whole blood or red blood cell transfusion within 12 weeks of intended randomization, those with a known contraindication for iron therapy or hypersensitivity to any of the products to be administered, and those who were currently receiving, or had recently received (< 30 days), another investigational agent(s). All study participants gave written informed consent and the study was approved by the Institutional Review Board at each participating centre.

Study protocol

This was a double-blind, randomized, placebo-controlled, multicentre study. Following a 2-week screening period, eligible subjects were randomly assigned in a 1:1:1 ratio to receive a weight-based dose of subcutaneous (SC) darbepoetin alfa (starting dose 0.75 mcg/kg), a fixed dose of darbepoetin alfa (starting dose 50 mcg), or placebo every 2 weeks (Q2W) for 25 weeks. At the end of treatment, there was a 4-week follow-up period.

Investigators used a central interactive voice-response system (IVRS) to obtain a unique subject number for each subject. Randomization was done centrally by a computer-generated scheme and stratified according to body weight: < 60 kg, 60 to 90 kg, or > 90 kg. Investigational product was provided in single-dose vials containing either placebo or darbepoetin alfa (Aranesp®; Amgen Inc., 1000 Oaks, CA, USA) 200 or 500 mcg/mL. The appropriate starting dose for each subject at randomization was calculated by the IVRS. Darbepoetin alfa doses were titrated to reach and maintain Hb within the target range of 14.0 ± 1.0 g/dL according to predefined dosing rules based on Hb concentration and rate of rise. Once subjects’ Hb concentration had reached the target range and had increased by ≥ 1.5 g/dL from baseline, the dose of darbepoetin alfa was reduced by 25%. Dose adjustments to maintain Hb within the target range of 14.0 ± 1.0 g/dL were made as in the titration phase. Patients and personnel at the study sites were blinded to investigational product. All patients were administered approximately 1 mL of investigational product, which was supplemented with 200 mg/day of elemental oral iron, unless ferritin was > 800 mcg/L.

The primary endpoint of the study was the rate of rise of Hb concentration (rise/week) during the titration phase. Secondary endpoints were changed from baseline at week 27 in: LVEF (echocardiography), exercise tolerance (6 min walk), NYHA functional classification, and quality of life measurements (see below). Safety endpoints were incidence of adverse events, laboratory parameters and vital signs, and seroreactivity to darbepoetin alfa and other ESPs.


Study visits occurred weekly during the screening period (weeks –2 and –1) and treatment period (weeks 1 to 25). An end-of-study assessment was performed at week 27, or at the time of withdrawal. A safety follow-up was performed at week 29. Haemoglobin concentrations were measured weekly during screening and treatment periods, and then at week 26 and the end-of-study visit (week 27 or on withdrawal). Haematology, serum chemistry, and TSAT were assessed at screening, weeks 9 and 17, and at the end-of-study visit. Anti-erythropoietic seroreactivity assays were performed prior to administration of the first dose of darbepoetin alfa and at the end-of-study visit. Assessment of NYHA functional class was performed at weeks 1, 9, 13, 17, and at the end-of-study visit. Adverse events were collected throughout the study. A 6 min walk test was performed during screening, and at weeks 13 and 27. Health-related quality of life was assessed using the Patient’s Global Assessment (PGA) score, which consists of a simple judgement indicating improvement, no change, or worsening since the start of the study,14 was performed at weeks 9, 13, 17 and at the end-of study. The Minnesota Living with Heart Failure Questionnaire (MLHFQ)15 and the Kansas City Cardiomyopathy Questionnaire (KCCQ)16 were performed at weeks 1, 13 and at the end of study. With the MLHFQ a higher score means more impairment or worsening.15 The KCCQ identifies seven individual domains (symptom frequency, symptom burden, symptom stability, physical limitation, social limitation, quality of life, and self-efficacy). In addition, there are three composite scores: total symptom score, which is the mean of symptom frequency and symptom burden score; clinical (summary) score, which is the mean of the total symptom score and the physical limitation score; and the overall summary score, which is the mean of four scores: total symptom score, physical limitation score, social limitation score, and quality of life score. In general, a higher score means improvement or less functional impairment and fewer symptoms.16

Statistical analysis

A target sample size of 150, with 50 subjects per treatment group, was determined to give 90% power to detect a difference of 0.18 g/dL/week between the rates of rise of Hb concentration in the darbepoetin alfa weight-based dosing and placebo group with a 5% significance level. This sample size would also provide 90% power to conclude equivalence, within ± 0.5 g/dL/week, between the darbepoetin alfa groups. A drop-out rate of 10% was anticipated.

Analyses of the primary endpoint were based on the per-protocol analysis set.17 This set included all subjects who were randomly assigned and received the treatment to which they were assigned until the end of the titration phase, and had ≥ 6 Hb concentration measurements and received ≥ 8 weeks’ exposure to study drug. Analyses of secondary endpoints included data from all randomized subjects (intent-to-treat analysis set). Safety analyses included data from all subjects who were randomly assigned to treatment and received ≥ 1 dose of study drug.

The primary analysis assessed Hb rate of rise during the titration phase with a two-stage hypothesis testing procedure using analysis of covariance (ANCOVA) to adjust for three body weight categories (<60, 60 to 90, and > 90 kg). Weight-based darbepoetin alfa dosing was compared with placebo and if significantly different (P < 0.05), the equivalence of weight-based and fixed dosing of darbepoetin alfa was tested. These two dosing regimens were considered equivalent if the lower limit of the two-sided 95% confidence interval (CI) for the difference in mean rate of rise of Hb between the two regimens was above –0.5 g/dL/week, and the upper limit was below 0.5 g/dL/week.

For the 6 min walk test, changes from baseline were evaluated using ANCOVA, with the last value carried forward method used to impute missing data. For the quality-of-life parameters, the analysis was based on available data without imputation. Logistic regression was used to compare efficacy between treatment groups for NYHA classification. A χ2 test was used to evaluate the proportion of subjects with an improvement in PGA. Post hoc analyses were performed in which the two darbepoetin alfa groups were combined after confirming that they were equivalent with respect to Hb rate of rise.


Patient characteristics

Patients were enrolled at 44 study centres in 15 countries in Europe and the USA. Enrollment started in July 2003; the last subject completed the study in May 2005. Of the 317 patients who were screened for the study, 165 were eligible and randomized (Figure 1), while 152 were ineligible due to: Hb concentration outside the prespecified range (n = 65), TSAT < 15% (n = 29), vitamin B12 and/or folate levels lesser than the lower limit of the normal range (n = 19), no LVEF ≤ 40% within 12 weeks prior to screening (n = 13) or other exclusion criteria (n = 26). Of the 165 patients who were randomly allocated to treatment and were included in the intent-to-treat analysis set, 160 were eligible for inclusion in per-protocol analyses. At baseline, the three treatment groups were well-balanced, except for a slight imbalance in presence of diabetes (Table 1). The body weight of most patients (62%) was between 60 and 90 kg, while 21% weighed < 60 kg and 16% weighed > 90 kg. Blood pressure [mean 118 (17)/67 (9) mmHg] and heart rate [mean 70 (11)] were similar between groups.

Figure 1

Study design and subject disposition.

View this table:
Table 1

Demographics and baseline characteristics

Placebo (n = 55)Darbepoetin alfa
Weight-based dosing (n = 56)Fixed dosing (n = 54)All (n = 110)
Sex, n (%)
 Men34 (62)31 (55)31 (57)62 (56)
Ethnicity, n (%)
 Caucasian49 (89)50 (89)52 (96)102 (93)
 African American4 (7)3 (5)2 (4)5 (5)
 Asian2 (4)3 (5)03 (3)
Mean (SD) age (years)71 (9)73 (9)70 (12)71 (11)
Mean (SD) weight (kg)74 (16)74 (15)75 (19)74 (17)
Mean (SD) body mass index (kg/m2)26 (5)26 (5)26 (5)26 (5)
NYHA classification, n (%)
 Class I1 (2)1 (2)2 (4)3 (3)
 Class II24 (44)18 (32)22 (41)40 (36)
 Class III29 (53)35 (63)30 (56)65 (59)
 Class IV1 (2)2 (4)02 (2)
Mean (SD) LVEF (%)27 (10)28 (8)30 (9)29 (9)
Mean (SD) 6-min walk distance (m)304 (111)268 (117)307 (129)287 (124)
Mean (SD) duration of CHF (years)5 (6)7 (7)4 (4)6 (6)
Primary etiology CHF, n (%)
 Ischaemic heart disease35 (64)36 (64)40 (74)76 (69)
 Acquired cardiomyopathy10 (18)7 (13)11 (20)18 (16)
 Hypertension5 (9)7 (13)1 (2)8 (7)
 Valvular heart disease3 (5)2 (4)02 (2)
Medical history, n (%)
 Diabetes15 (27)27 (48)22 (41)49 (45)
 Ischaemic heart disease40 (73)39 (70)42 (78)81 (74)
 Hypertension33 (60)36 (64)30 (56)66 (60)
 Myocardial infarction31 (56)31 (55)38 (70)69 (63)
 Valvular heart disease15 (27)12 (21)14 (26)26 (24)
Heart failure medication, n (%)
 β-blockers50 (91)44 (79)45 (83)89 (81)
 ACEI and/or ARB50 (91)53 (95)50 (93)103 (94)
 ACE inhibitors41 (75)44 (79)41 (76)85 (77)
 ARB11 (20)10 (18)10 (19)20 (18)
 Glycosides14 (25)12 (21)11 (20)23 (21)
 Diuretics51 (93)54 (96)52 (96)106 (96)
 Anticoagulants12 (22)11 (20)5 (9)16 (15)
 Antiplatelet medications24 (44)27 (48)35 (65)62 (56)
Mean (SD) laboratory values
 Haemoglobin (g/dL)11.4 (0.9)11.5 (0.7)11.6 (0.7)11.5 (0.7)
 Ferritin (μg/L)200 (224)179 (175)220 (280)198 (232)
 Fe (μmol/L)13 (4)13 (4)16 (7)15 (6)
 TSAT (%)25 (8)24 (7)28 (11)26 (9)
 Creatinine (mg/dL)1.5 (0.6)1.3 (0.6)1.4 (0.5)1.4 (0.5)
 eGFR (mL/min/1.73 m2)53.5 (23.5)57.6 (21.9)55.3 (21.6)56.5 (21.7)
  • eGFR, estimated glomerular filtration rate.

Primary endpoint

There was a significant difference between the darbepoetin alfa weight-based dosing group and the placebo group in the rate of rise of Hb concentration of 0.2 g/dL/week (95% CI: 0.16, 0.24; P < 0.001). Although the rate of Hb rise in the weight-based group was statistically greater than in the fixed dosing group [0.05 g/dL/week (95% CI: 0.01, 0.09)], the two darbepoetin alfa dosing regimens met the pre-specified criteria for equivalence.

At baseline, mean Hb was similar in the three groups (Table 1). At week 27, mean (SD) Hb had increased by 1.87 (1.36) g/dL and 1.64 (0.98) g/dL in the weight-based and fixed dose darbepoetin alfa groups, respectively, vs. 0.07 (1.08) g/dL in the placebo group (Figure 2). The least-squares mean (95% CI) treatment difference between the darbepoetin alfa weight-based dose and placebo groups was 0.20 (0.16, 0.24) g/dL/week (P < 0.001), and the darbepoetin alfa weight-based and fixed dose groups were considered equivalent according to the protocol. Mean (SD) haemoglobin concentrations at week 27 were 13.4 (1.4) g/dL in the weight-based and 13.2 (1.1) g/dL in the fixed-dose group, and 11.4 (0.9) g/dL in the placebo group. The difference in mean change in Hb from baseline to week 27 was 1.69 (95% CI: 1.18–2.2) g/dL between the darbepoetin alfa groups combined vs. placebo (P < 0001).

Figure 2

Change in haemoglobin concentration over time (intent-to-treat analysis set). Values are mean ± standard error (SE).

Iron, ferritin, and TSAT were similar in the three groups at baseline and did not significantly change during the study (Table 2). There was a significant difference between the placebo and darbepoetin alfa groups combined, regarding the change from baseline to 27 weeks, in serum creatinine (P = 0.012) and eGFR (P = 0.024).

View this table:
Table 2

Selected laboratory values during treatment (27 weeks)

PlaceboWeight-based DAFixed DAAll DA
Mean (SD) laboratory values
 Haemoglobin (g/dL)11.4 (1.4)13.4 (1.4)13.2 (1.1)13.3 (1.2)*
 Ferritin (μg/L)219 (262)214 (211)194 (181)203 (195)
 Fe (μmol/L)13 (5)17 (7)17 (8)17 (8)
 TSAT (%)27 (10)33 (14)32 (17)33 (15)
 Creatinine (mg/dL)1.6 (0.8)1.2 (0.5)1.3 (0.5)1.3 (0.5)*
 eGFR (mL/min/1.73 m2)52.5 (24.3)61.5 (24.1)59.1 (24.4)60.2 (24.1)*
  • *Indicates P < 0.05 for difference between change after 27 weeks between placebo and darbepoetin alfa groups combined.

The median (25th, 75th) darbepoetin alfa Q2W dose during the study was 0.80 (0.54, 1.01) mcg/kg in the weight-based group and 0.72 (0.62, 0.98) mcg/kg in the fixed-dose group.

Secondary endpoints

Six-minute walk distance

The mean (SD) 6 min walk distances at baseline in the three groups are shown in Table 1. At week 27, there were no significant differences in change from baseline among the three treatment groups. If the two darbepoetin alfa groups were combined, there was a mean (SE) increase of 34.2 (7.3) m, compared with 11.4 (10.3) m in the placebo group (P = 0.074) (Figure 3).

Figure 3

Change from baseline at week 27 in 6 min walk distance. Values are mean ± standard error (SE).

Left ventricular ejection fraction

The mean (SD) LVEFs at baseline in the three groups are provided in Table 1. At week 27, there were no differences among the groups; if the two darbepoetin alfa groups were combined, the mean (SE) changes were +1.27 (1.29)% in the placebo group vs. −0.02 (0.91)% in the combined darbepoetin alfa group (P = 0.415).

New York Heart Association classification

At baseline, NYHA functional class was similar in the three groups (Table 1). Although patients receiving placebo (−0.23 (0.08)) or darbepoetin alfa (−0.30 (0.06)) both showed minor improvements in NYHA class, there was no difference between groups (P = 0.473).

Patient Global Assessment

At week 27, 49% of patients in the placebo group reported improvement vs. 65% of patients in the combined darbepoetin alfa group (P = 0.057).

Minnesota Living with Heart Failure Questionnaire

At baseline, the mean (SD) MLHFQ total scores were similar in the three groups: 40.4 (22.7) in the placebo group, 39.0 (23.6) in the darbepoetin alfa weight-based group, and 42.1 (20.3) in the fixed dosing group. By week 27, there was a mean (SE) decrease of 7.4 (2.7) points in the placebo group vs. 10.1 (1.9) points in the combined darbepoetin alfa group (P = 0.413).

Kansas City Cardiomyopathy Questionnaire

At baseline, the mean (SD) KCCQ overall summary scores were similar in the three groups: 57.7 (21.9) points in the placebo group, 58.7 (20.3) points in the darbepoetin alfa weight-based, and 57.0 (18.1) points in the fixed dosing group. At week 27, the mean (SE) improvement in the combined darbepoetin alfa group was 8.0 vs. 4.9 points in the placebo group (P = 0.238) (Figure 4). For the KCCQ total symptom score, mean baseline values were 67.2 (22.1) points (placebo) and 65.4 (21.8) points (combined darbepoetin alfa). At week 27, there was a mean change of 1.5 points on placebo vs. 8.2 points in the combined darbepoetin alfa group (P = 0.027). At week 27, mean changes in the KCCQ clinical summary score were 3.7 (placebo) vs. 8.6 points in the combined darbepoetin alfa group (P = 0.086). Mean changes at week 27 in individual domains of the KCCQ were mostly similar between groups, apart from the KCCQ symptom frequency score: −2.7 (placebo) vs. +8.1 points in the combined darbepoetin alfa group (P = 0.003).

Figure 4

Change from baseline at week 27 in Kansas City Cardiomyopathy Questionnaire (KCCQ) scores. Values are mean ± standard error (SE).

Safety assessments

The overall incidence of adverse events during the study was similar between the three treatment groups (Table 3). The majority of patients experienced at least one adverse event, but there were no differences or even trends between groups. There were no differences in serious adverse events, or treatment-related events (as considered by the investigator). The incidence of adverse events of specific interest was also not different between groups. Deep vein thrombosis, cerebrovascular disorders, or seizures were not observed. There was no difference in the CHF hospitalization rates. All bioassay tests for neutralizing antibodies to darbepoetin alfa were negative. No significant changes in laboratory parameters (except erythrocyte-related parameters), renal function (Table 2), or vital signs were observed.

View this table:
Table 3

Summary of adverse events

Adverse event (n, %)Placebo (n = 55) n (%)Darbepoetin alfa
Weight-based dosing (n = 55) n (%)Fixed dosing (n = 53) n (%)All (n = 108) n (%)
Any adverse event42 (76)45 (82)41 (77)86 (80)
Serious adverse events15 (27)13 (24)12 (23)25 (23)
Treatment-related adverse events4 (7)5 (9)7 (13)12 (11)
 Related serious adverse events0 (0)0 (0)1 (2)1 (1)
 Related fatal adverse events0 (0)0 (0)0 (0)0 (0)
Adverse events of specific interest
 Congestive heart failure9 (16)9 (16)3 (6)12 (11)
 Hypertension2 (4)2 (4)0 (0)2 (2)
 Myocardial infarction0 (0)1 (2)1 (2)2 (2)
 Cerebrovascular disorder0 (0)0 (0)0 (0)0 (0)
 Deep vein thrombosis0 (0)0 (0)0 (0)0 (0)
 Pulmonary emboli1 (2)0 (0)0 (0)0 (0)
 Seizure0 (0)0 (0)0 (0)0 (0)
Discontinuation due to adverse events1 (2)0 (0)0 (0)0 (0)
Deaths on studya0 (0)5 (9)1 (2)6 (6)
  • aIncludes deaths occurring within 30 days of study drug administration, or until end of study, whichever is later.

Six subjects died during the study or within 30 days of the last dose: five in the darbepoetin alfa weight-based dosing group (three heart failure, one ischaemic coronary artery disorder, and one sudden death) and one in the fixed dose group (pneumonia). None of the events leading to discontinuation or death was considered to be treatment-related by the attending physicians. Close examination of the six deaths did not reveal any patterns. Only one patient who died had an Hb concentration > 15.0 g/dL (15.3 g/dL); all others had Hb concentrations < 13.5 g/dL. One patient (not the patient with Hb > 15.0 g/dL) died suddenly 7 days after the day of the first administration of darbepoetin alfa; the other five patients had participated in the study for at least 2 months. Of the six patients who died, three were taking antiplatelet medications at baseline and at the time of death. None of these six was taking anticoagulants at baseline. Darbepoetin alfa had no apparent effect on blood pressure or heart rate. Other than the expected effects of darbepoetin alfa on Hb (and other erythrocyte-related parameters), no clinically relevant changes occurred in clinical chemistry, (other) haematological markers, or other parameters (data not shown). Baseline characteristics of the subjects who died were unremarkable and none of the six deaths had atrial fibrillation.


The present multicentre, double-blind, randomized controlled trial showed the feasibility of rising and maintaining Hb concentrations within target levels, using darbepoetin alfa, in patients with CHF and anaemia. The primary study endpoint of showing similar effectiveness of weight-based and fixed starting dose regimens of darbepoetin alfa was achieved. The effects on quality of life and exercise capacity were modest. Six deaths occurred in the darbepoetin alfa groups while none occurred in the placebo group; incidences of all other adverse events were similar between treatment groups. There is a need for further research to understand the benefits and limitations of this treatment strategy.

In recent years, many studies in CHF have shown that anaemia is associated with an adverse outcome, suggesting that treatment for anaemia should benefit patients. So far, a number of small studies69 have shown promising results. The present study adds importance to this database and is the first to evaluate two dosing regimens in a substantially larger population. While the study was not powered to evaluate clinical outcome, the results are consistent with the modest benefits on symptoms and exercise capacity previously observed.

In patients with CHF, anaemia has been associated with decreased exercise capacity and impaired quality of life in patients.14 There are several possible explanations for these associations. Oxygen supply to tissues depends on cardiac output, arterial Hb concentration, oxygen saturation, and the dissociation curve. In CHF, cardiac output is decreased and when Hb also decreases, adequate oxygen supply to tissues may become insufficient. In CHF, there is a mismatch between oxygen demand and supply of the myocardium.18 While oxygen demand is increased as a result of increased wall stress, oxygen supply is impaired which relates to several factors such as endothelial dysfunction and microvascular insufficiency. This oxygen demand-supply mismatch is particularly critical during anaemia. Raising Hb may thus improve oxygen supply to the tissues, both in peripheral muscles but also in the myocardium, which might explain the observed clinical changes. Independent of raising Hb, erythropoietin may also have other effects, in particular on angiogenesis,1,3,19 which might be clinically relevant.

In the study presented here, darbepoetin alfa was generally well tolerated. The number and incidences of non-fatal adverse events were similar in patients receiving darbepoetin alfa or placebo. However, 6 of the 110 patients in the darbepoetin alfa group died, compared with none of 55 patients in the placebo group which may be a possible cause for concern. Although this study was not powered to assess clinical outcomes, this finding prompted a detailed analysis. None of the deaths were attributed to the (double-blind) treatment, according to the investigators, and only one death occurred in a patient with an Hb concentration above the target range (Amgen, data on file). Use of ESPs has been associated with increased thrombotic risk and increased vascular resistance.1 Use of antiplatelet medications and/or anticoagulants could therefore be important, but three of the six patients who died were on antiplatelet medications. We did not study vascular resistance, but blood pressure was unaffected, and in one earlier study in CHF in which this was specifically measured by forearm plethysmography, no changes were observed after ESP administration.7 Also, all five cardiac deaths occurred in the weight-based dosing group, while none were observed in the fixed dosing group, which would suggest play of chance. Recently, data have become available from a similar study in patients with anaemia and CHF who received darbepoetin alfa (n = 162) or placebo (n = 157) Q2W for 1 year (target Hb 14.0 g/dL).20 In this study, fewer on-study deaths occurred in the darbepoetin alfa group (n = 11) than in the placebo group (n = 18). More importantly, when the clinical endpoints of that study and the present one were combined in a prespecified, pooled analysis, the data suggest that darbepoetin alfa (n = 266 patients), compared with placebo (n = 209 patients), favourably affected the risk of the composite endpoint of CHF-related hospitalization and all-cause mortality (HR 0.67, 95% CI 0.44–1.03; P = 0.06; for all-cause mortality alone: HR 0.76; 95% CI 0.39–1.48; P = 0.42).21 These data clearly need to be interpreted with caution, since neither study was powered primarily to investigate clinical outcomes. However, they support the importance of a recently launched large phase 3 study: Reduction of Events with Darbepoetin Alfa in Heart Failure (RED-HF) Trial, which will evaluate the effect of treatment of anaemia with this drug on morbidity and mortality in patients with symptomatic CHF.22

In conclusion, the current study shows a potentially beneficial effect of treatment with darbepoetin alfa on exercise tolerance and quality of life in patients with symptomatic CHF and anaemia. More deaths occurred in the darbepoetin alfa treated patients than in those on placebo, and the safety of this treatment requires further exploration. Therefore, the observed effect on clinical events, together with other recent data21 underscore the need for a larger phase 3 study, to understand the efficacy and safety of this treatment strategy.


This study was supported by Amgen, Thousand Oaks, CA, USA. The authors wish to thank Beate D Quednau, PhD (Amgen Inc.) for expert editorial assistance and Claire Chapman for providing assistance with writing and graphics.

Conflict of interest: S.M.W., N.B., and D.R. are employees of Amgen. D.J.v.V., A.C.-S., D.J.A.L., J.G.F.C., and P.P. have received research grants and/or consultancy fees from Amgen.


View Abstract