Darbepoetin Alfa: Impact on Treatment for Chemotherapy-Induced Anemia and Considerations in Special Populations

October 1, 2002
Johan Vansteenkiste, MD, PhD

Erik Poulsen, MS

Gregory Rossi, PhD

John Glaspy, MD, MPH

Oncology, ONCOLOGY Vol 16 No 10, Volume 16, Issue 10

Our objective was to evaluate the effects of darbepoetin alfa (Aranesp) on hemoglobin and transfusions in anemic patients with cancer undergoing chemotherapy, and the impact of age, sex, baseline hemoglobin, chemotherapy

ABSTRACT: Our objective was to evaluate the effects of darbepoetin alfa (Aranesp)on hemoglobin and transfusions in anemic patients with cancer undergoingchemotherapy, and the impact of age, sex, baseline hemoglobin, chemotherapytype, and tumor type. Patients were randomized to one of three darbepoetin alfagroups based on average weekly dose (< 1.5 µg/kg, 1.5 to 2.25 µg/kg, and> 2.25 µg/kg) or to placebo. Dose response was evaluated for change inhemoglobin, hemoglobin and hematopoietic responses, and red blood celltransfusion rates. Hazard ratios for the incidence of hemoglobin response andtransfusions were calculated. Adverse events and antibody formation wereassessed. Treatment effects were observed for all hemoglobin end points andincidence of transfusion. The incidence of hematopoietic response among thedarbepoetin alfa dose groups ranged from 46% (95% confidence interval [CI] = 33%-60%)to 74% (95% CI = 66%-81%) and increased with higher darbepoetin alfa dose.Patients receiving darbepoetin alfa were more likely to exhibit a hemoglobinresponse and less likely to require a transfusion, compared with placebo,irrespective of the patient characteristics examined. No increased risk ofadverse events and no development of neutralizing antibodies were observed withdarbepoetin alfa use. Darbepoetin alfa increased the likelihood of a hemoglobinresponse and decreased the need for transfusions in cancer patients withchemotherapy-induced anemia. [ONCOLOGY 16(Suppl 11):37-44, 2002]

Cancer patients commonlyexhibit anemia that has a multi-factorial etiology and involves a variety of symptoms, including fatigue,dizziness, dyspnea, headache, and lethargy.[1-5] This anemia can have an adverseimpact on these patients, potentially resulting in increased morbidity anddecreased quality of life.[2,5] The frequency and severity of anemia aredependent upon many factors, including the type, duration, and extent of themalignancy, as well as the type and intensity of the chemotherapy.[5] Severeanemia or symptomatic mild-to-moderate anemia requires medical management,possibly through the use of red blood cell (RBC) transfusions, which involvespotential risks including acute transfusion reactions or transmission ofinfectious agents.[6,7] Such risks may prove to be unacceptable, as frequent RBCtransfusions with allogeneic blood may adversely affect the immune system ofthese patients, thereby increasing the likelihood of infections, decreasingtimes to relapse, or shortening survival times.[7] Even mild (10 to 12 g/dL) andmoderate (8 to 10 g/dL) anemia may be poorly tolerated by some patients becauseof the resultant increased fatigue and decreased quality of life; however, RBCtransfusions are no longer considered routine medical management in thesepatients because of the associated risks.[5,8,9]

The availability of recombinant human erythropoietin (rHuEPO)for the safe and effective treatment of chemotherapy-induced anemia in cancerpatients has provided an alternative, or supplement, to blood transfusions inthis population.[10] Furthermore, some studies have shown rHuEPO to beassociated with improvements in health-related quality of life.[8,9] Manystudies have demonstrated generally similar effects of rHuEPO on the reductionof the requirement for RBC transfusions and the ability to increase hemoglobinconcentration in patients across a broad range of nonmyeloid malignancies andchemotherapy agents, although a range of response rates has beenobserved.[8,9,11-15]

As observed by Seidenfeld et al,[16] comparisons between studiesare difficult because different studies used varying definitions of response,different analysis populations (eg, intent-to-treat vs efficacy evaluablesubsets), and an assortment of treatment period durations. Abels andcolleagues[11] reported 48% and 58% response rates (³ 6 hematocrit points) forpatients treated with cisplatin and noncisplatin chemotherapy, respectively,during a 12-week treatment period in which rHuEPO dose increases were permitted,while Glaspy et al[8] reported a 53% hemoglobin response rate (³ 2.0-g/dLincrease in hemoglobin) after a 16-week treatment period.

Recently, in a report of a meta-analysis of the use of rHuEPO totreat chemotherapy-induced anemia, Seidenfeld et al indicated that evaluation ofa treatment effect for rHuEPO was impeded because there were a limited number ofstudies with large populations, and few studies with consistent designs.[16]However, based on this analysis, Seidenfeld et al[16] concluded that there wasno evidence that earlier initiation of rHuEPO therapy spared more patients fromtransfusions or resulted in greater quality-of-life improvements than waitingfor hemoglobin concentrations to decline to nearly 10 g/dL.

Darbepoetin alfa (Aranesp) is a next-generation erythropoieticprotein that has been shown to be safe and effective in treatingchemotherapy-induced anemia in cancer patients.[17-22] Compared with epoetinalfa, darbepoetin alfa has an increased sialic acid content, which contributesto its extended serum half-life (approximately threefold longer) and increasedbiologic activity.[23,24] Darbepoetin alfa and rHuEPO appear to act by similarmechanisms, essentially by binding to the erythropoietin receptor andstimulating erythropoiesis, thus causing RBC precursors to proliferate andenhancing their survival. Darbepoetin alfa, however, exhibits an early andsustained erythropoietic effect, allowing darbepoetin alfa to be administeredless frequently than rHuEPO while maintaining efficacy.

At the time of initial licensure of a new agent, many questionsregarding its clinical utility are still not yet fully reported in theliterature. There exists a need to elaborate the various characteristics of thenew agents to permit effective clinical use. These characteristics include thedose response, the impact of important covariates (ie, age or sex) on thetreatment effect, the nature of any specific recommendations that may be neededif a differential treatment effect is observed in specific subpopulations, andany specific safety precautions that should be undertaken with regard tohigh-risk groups, eg, those with cardiovascular comorbidity. Frequently, asingle study cannot provide the answers to all these questions, and acomprehensive evaluation of all the experience gained in clinical studies isrequired to draw meaningful conclusions. While the answers to these questionsare unlikely to differ significantly between darbepoetin alfa and rHuEPO due tothe identical mechanism of action, comprehensive reviews addressing theseimportant data are limited even for rHuEPO, despite 10 years of clinical use.

We pooled data from four completed, similarly designed studiesof darbepoetin alfa[17-19] to perform an integrated analysis of the relationshipof darbepoetin alfa dose to a number of efficacy variables. Additionally,results are presented that evaluate the treatment effect of darbepoetin alfaamong a number of subgroups based on important patient characteristics,including age, sex, baseline hemoglobin, tumor type, and chemotherapy type.


Study Population

The main differences between the eligibility criteria in thefour studies included in this analysis were tumor type (lung cancer vs solidtumors vs lymphoproliferative malignancies) and chemotherapy restriction(platinum-containing only vs no restrictions). Common eligibility criteriaincluded the following: Eligible patients had to be of legal age with adiagnosis of cancer (solid tumors or lymphoproliferative malignancies) andanemia (hemoglobin concentration £ 11.0 g/dL) primarily due to cytotoxicchemotherapy or cancer, and had to be receiving concomitant chemotherapy for theduration of the study treatment period (12 weeks). Patients also were requiredto have adequate renal function (serum creatinine level of ≤ 2.0 mg/dL) and anEastern Cooperative Oncology Group (ECOG) performance status score of 0 to 2(three studies) or 0 to 3 (one study).

Patients who were anemic for reasons other than cancer or itstreatment and patients who had received RBC transfusions immediately beforestudy entry (two RBC transfusions within 4 weeks or any RBC transfusions within2 weeks) were excluded. To ensure that patients were not anemic due to nutritionor iron deficiency, patients were required to have a vitamin B12 level of ³ 200pg/mL, a folate level of ³ 2.0 ng/mL, and ferritin and transferrin saturationlevels of at least 10 mg/dL and 15%, respectively. Patients with primary ormetastatic cancer involving the central nervous system, and patients withseizures, unstable angina, or uncontrolled hypertension, were also excluded fromall studies.

The institutional review boards or ethics committees of theparticipating centers approved the protocol. Informed consent was obtained fromall patients before any study-related procedure was performed.

Study Design and Conduct

Four multicenter, randomized studies were conducted in theUnited States, Europe, Australia, and Canada. One study was open-label in designand used rHuEPO as an active control. The remaining three studies weredouble-blind and placebo-controlled.

The studies included doses of darbepoetin alfa from 0.5 to 8.0µg/kg once weekly and 3.0 to 9.0 µg/kg every 2 weeks. In addition to similareligibility criteria, all of the studies had identical, planned treatmentlengths (12 weeks), identical dose adjustment rules upon reachingpredefined hemoglobin thresholds, and identical recommendations regardingtransfusions. Two studies allowed dose increases of darbepoetin alfa if thepatient’s hemoglobin had not increased by at least 1 g/dL after approximately1 month of therapy (week 5 or 6).

Rationale for Pooling Data

The information from each of the selected studies was integratedto obtain a larger patient sample in which to evaluate the dose-responserelationship for a nonmyeloid malignancy population as a whole, as well as toevaluate the impact of important covariates such as age, sex, baselinehemoglobin, chemotherapy type, and tumor type on the treatment effect ofdarbepoetin alfa. As the individual studies included in this analysis involvedspecific tumor types, pooling of the data from these studies resulted in apatient population with a broader spectrum of tumor types and chemotherapyregimens consistent with the previously published data on rHuEPO.

The criteria for pooling were mainly clinical—the studies hadbeen designed with common characteristics in terms of treatment length,intervention strategies, and assessments, as well as patient eligibilitycriteria. Importantly, both the hemoglobin- and transfusion-based results fromthe four studies were similar where common average weekly doses existed for sucha comparison. For example, in the average weekly dose range 1.5 to 2.25 µg/kg,the difference between darbepoetin alfa and placebo with respect to change inhemoglobin over a 12-week treatment phase was between 1.3 and 1.6 g/dL for thethree placebo-controlled studies. In the active controlled study, similarchanges from baseline were observed.

Statistical Analysis

The primary comparison in this analysis is between darbepoetinalfa and placebo, both to examine the treatment effect of darbepoetin alfa invarious subgroups as well as to examine the relationship of dose and response.Patients who received rHuEPO in the active control study were not included inthis analysis as the sample size was relatively small compared with the othergroups, thereby prohibiting meaningful comparisons.

Patients who received darbepoetin alfa were divided into threedose groups based on the average weekly dose: low dose, medium dose, and highdose. The low-dose group (< 1.5 µg/kg/wk darbepoetin alfa) combineddata from patients who received darbepoetin alfa dosages of 0.5 and 1.0 µg/kgevery week. The medium-dose group (1.5 to 2.25 µg/kg/wk darbepoetin alfa)included patients who received darbepoetin alfa dosages of 1.5 and2.25 µg/kg every week or 3.0 µg/kg every 2 weeks. The high-dose group(> 2.25 µg/kg/wk darbepoetin alfa) included patients who receiveddarbepoetin alfa doses of 4.5, 6.0, and 8.0 µg/kg every week or 5.0, 7.0, and9.0 µg/kg every 2 weeks. Due to the relatively small sample size in the low- andhigh-dose groups, the subgroup analyses and safety results are presented for themedium-dose group only.

Baseline demographic and clinical characteristics weresummarized by the mean (standard deviation [SD]) for continuous measures andnumber (percentage) for categorical measures. Common definitions and methods todeal with missing data were used for hemoglobin, transfusion, and safety endpoints across all studies. Change in hemoglobin concentration after12 weeks of treatment was summarized by the mean value (95% confidenceinterval [CI]) using the last available value not within 28 days of a RBCtransfusion to calculate the change. Hematopoietic response was defined aseither a ³ 2.0-g/dL increase in hemoglobin from baseline or a hemoglobinconcentration ³ 12.0 g/dL in the absence of a RBC transfusion in thepreceding 28 days. Hemoglobin response was defined as a ³ 2.0-g/dLincrease in hemoglobin from baseline in the absence of a RBC transfusion in thepreceding 28 days. The incidence of RBC transfusions was calculated based on allRBC transfusions given for any reason, regardless of hemoglobin concentrationover two time periods: first, the period beginning in week 5 (study day 29) andcontinuing through the end of treatment, and second, the entire treatmentperiod.

The Kaplan-Meier method was used to estimate hematopoieticresponse, hemoglobin response, and RBC transfusion percentages by subtractingthe Kaplan-Meier estimate of the survivor function from 1 for each end point.Point estimates and 95% confidence intervals, calculated using Greenwood’sestimate of variance, are provided for hemoglobin response, hematopoieticresponse, and RBC transfusion percentages.[25] The time to hematopoieticresponse was calculated as the number of weeks from study day 1 to the firstpostbaseline hemoglobin concentration that met the criteria for hematopoieticresponse.

The effects of age (< 65 years vs ³ 65 years), sex(men vs women), tumor type (solid vs lymphoproliferative), chemotherapy (ever vsnever received platinum-containing chemotherapy), and baseline hemoglobin(< 10 g/dL vs ³ 10 g/dL) on hemoglobin response and RBCtransfusion end points were investigated using hazard ratios with 95% confidenceintervals.

The incidence of adverse events was analyzed, and the oddsratios with 95% confidence intervals are presented for the adverse events ofhypertension and thrombosis. Although not identified as a potential safetyconcern in patients with cancer, adverse events were evaluated relative to highhemoglobin concentrations and hemoglobin rates of rise because of potentialsafety concerns that have been identified in patients with chronic renal failurereceiving epoetin alfa. A patient was considered to have a rapid rate of rise ofhemoglobin if the maximum increase in hemoglobin during any 28-day period afterthe first dose of study drug in the absence of RBC transfusion was ³ 2.0g/dL.


Overall, 729 patients with cancer received darbepoetin alfa and339 received placebo. Demographic and baseline characteristics are presented in Table 1. Patients ranged in age from 18 to 91 years, with a mean of 63.1 years(standard deviation, 11.7 years). Nearly equal numbers of men and women wereenrolled overall and within each dose group, although a slight imbalance infavor of women was evident in the low-dose group. Most patients (93% overall)were white. Most patients had an ECOG performance status of ≤ 1 (79% darbepoetinalfa, 76% placebo). Across all darbepoetin alfa dose groups, mean baselinehemoglobin concentration was approximately 9.8 g/dL.

Effect of Darbepoetin Alfa Dose on Hemoglobin End Points

The effect of erythropoietic agents on hemoglobin change hasbeen measured using a variety of end points in previously reported studies. Weevaluated three of the most commonly used: change in hemoglobin from baseline,hematopoietic response (³ 2.0-g/dL increase from baseline or apostbaseline hemoglobin concentration ³ 12.0 g/dL), and hemoglobinresponse (³ 2.0-g/dL increase from baseline). Because erythropoieticagents are used to reduce the risk of transfusions and alleviate signs andsymptoms of anemia in patients being treated with additional myelotoxicchemotherapy, the time to elicit a response is an important characteristic.Therefore, we also evaluated the relationship of darbepoetin alfa dose to timeto response.

Substantial increases in hemoglobin (mean, 1.1 g/dL [95% CI= 0.7-1.6 g/dL] to 2.1 g/dL [95% CI = 1.9-2.4 g/dL]) were observedafter 12 weeks of treatment for all three of the darbepoetin alfa dose groupscompared with placebo, with increasing response apparent with increasing dose (Figure1). The mean change in hemoglobin for the medium-dose group was1.8 g/dL (95% CI = 1.6-2.0 g/dL). Similarly, a dose response was alsoobserved for hematopoietic response (Figure2). The proportion of patientsachieving a hematopoietic response ranged from 46% (95% CI = 33%-60%) to 74%(95% CI = 66%-81%). As observed with change in hemoglobin, even the low-dosegroup had a greater proportion of patients with a hematopoietic responsecompared with placebo. Furthermore, the medium- and high-dose groups exhibitedgreater incidences of hematopoietic response than the low-dose group, as theconfidence intervals do not overlap (Figure2). A dose-response relationshipwith respect to time to hematopoietic response was also observed; higher dosesof darbepoetin alfa were associated with a decreased time to hematopoieticresponse and a higher percentage of patients responding at a given time point (Figure3). The median times to hematopoietic response were 10 weeks (95% CI = 8-11weeks) for the medium-dose group and 8 weeks (95% CI = 7-9 weeks) for thehigh-dose group. Similar findings were observed for hemoglobin response (Figure4).

Effect of Darbepoetin Alfa Dose on RBC Transfusions

Results published regarding rHuEPO in this patient population[11] have shown that the effect on RBC transfusions is not apparent until after the first month of treatment. Therefore, in this analysis, the impact of darbepoetin alfa dose on RBC transfusions was evaluated after the first 4 weeks of treatment.[26,27] A positive treatment effect was observed with darbepoetin alfa with respect to the incidence of RBC transfusions from week 5 to the end of the treatment phase. All darbepoetin alfa dose groups had lower RBC transfusion requirements (18% [95% CI = 8%-28%] to 31% [95% CI = 24%-38%]) compared with the placebo group (50% [95% CI = 44%-55%]) (Figure 5); however, no apparent dose-response effect was observed.

Darbepoetin alfa also exhibited a positive treatment effect whenRBC transfusions were analyzed across the entire treatment period. Fewerpatients receiving darbepoetin alfa required RBC transfusions during the entiretreatment period (28% [95% CI = 17%-40%] to 38% [95% CI = 31%-45%]) comparedwith patients receiving placebo (58% [95% CI = 53%-64%]).

Impact of Patient/Treatment Characteristics on Efficacy of Darbepoetin Alfa

We examined the impact of various patient characteristics onhemoglobin and RBC transfusion end points for patients within the medium-dosegroup compared with patients receiving placebo, as this dose group had thehighest number of patients, thereby allowing a more robust comparison. Theresults are displayed as hazard ratios as this allows for simple presentation oftreatment effect across the various characteristics examined. We chosehemoglobin response to be representative of the hemoglobin variables for tworeasons. First, as hemoglobin response is a simple yes/no variable, it is easilypresented as a hazard ratio. Second, we wanted to evaluate the impact ofbaseline hemoglobin on response. Hemoglobin response was used to analyze theimpact of patient characteristics on hemoglobin end points because hematopoieticresponse, which counts patients as responders if they achieve either ahemoglobin concentration ³ 12.0 g/dL or a ³ 2.0-g/dL increase inhemoglobin from baseline, would have a higher degree of bias in favor of thepatients with higher hemoglobin levels at baseline compared with hemoglobinresponse alone. For hemoglobin response, a hazard ratio above 1 indicates ahigher likelihood of achieving a response on darbepoetin alfa compared withplacebo, and for RBC transfusions a hazard ratio below 1 indicates a lowerlikelihood of receiving a transfusion on darbepoetin alfa treatment.

Darbepoetin alfa patients had an increased likelihood of ahemoglobin response (Figure 6) and a decreased risk of transfusions (Figure7) compared with placebo patients, regardless of baseline hemoglobin category(< 10 vs ³ 10 g/dL), tumor type (lymphoproliferative vs solid),and chemotherapy type (ever vs never received platinum-containing chemotherapy).

For hemoglobin response, the hazard ratio for each subgroupanalysis was approximately 4 or greater indicating a fourfold greater chance ofa hemoglobin response on active treatment (hazard ratio ranges: 3.4 to 6.2) (Figure6). No difference was observed between patients who did or did not everreceive platinum therapy or between patients with solid and withlymphoproliferative malignancies. Regardless of baseline hemoglobin category,all patients receiving darbepoetin alfa had a higher likelihood of hemoglobinresponse compared with those receiving placebo. While no evidence of adifference between the two baseline hemoglobin groups was detected, a slighttrend toward increased response for patients initiating therapy at a baselinehemoglobin above 10 g/dL was observed.

Similar treatment effects were observed when hemoglobin responsewas analyzed by either age (< 65 vs ³ 65 years) or sex (men vswomen). The hazard ratios of darbepoetin alfa vs placebo for the impact of agewere 4.3 (95% CI = 2.7-6.6) for < 65 years old and 4.0 (95% CI =2.6-6.0) for ³ 65 years old. The hazard ratios of darbepoetin alfa vsplacebo for the impact of sex were 3.7 (95% CI = 2.5-5.4) for men and 4.9 (95%CI = 3.0-8.1) for women.

A treatment effect vs placebo was also seen in all subgroupswith respect to transfusions, with darbepoetin alfa patients being approximatelyhalf as likely to receive a RBC transfusion as patients receiving placebo (rangeof hazard ratios: 0.4 to 0.7) (Figure 7). As observed for hemoglobin response,no major differences were observed between patients receiving platinum therapyand those on nonplatinum therapy, or between patients with solid and those withlymphoproliferative malignancies. However, a trend to slightly greater treatmenteffect was again observed in patients with a baseline hemoglobin concentration ³ 10 g/dL, although no statistical difference was observed despite arelatively large sample size.

The individual point estimates for the incidence of RBCtransfusions, presented as medium-dose group vs placebo, were 42% (95% CI = 35%-48%)vs 57% (95% CI = 49%-64%) for patients with a baseline hemoglobin< 10 g/dL and 18% (95%CI = 13%-24%) vs 42% (95%CI = 34%-50%) for patients with a baseline hemoglobin ³ 10 g/dL.This indicates that in patients receiving darbepoetin alfa, fewer patients witha baseline hemoglobin ³ 10 g/dL required a transfusion, compared withpatients with a baseline hemoglobin < 10 g/dL. In general, fewerpatients receiving darbepoetin alfa required transfusions than patientsreceiving placebo.

Similar treatment effects were observed when transfusions wereanalyzed by either age (< 65 vs ³ 65 years) or sex (men vs women).The hazard ratios of darbepoetin alfa vs placebo for the impact of age were 0.5(95% CI = 0.4-0.8) for < 65 years old and 0.5 (95% CI = 0.4-0.7) for ³ 65 years old. The hazard ratios of darbepoetin alfa vs placebo for theimpact of sex were 0.5 (95% CI = 0.4-0.7) for men and 0.5 (95% CI = 0.4-0.8)for women.


Adverse event data are provided for patients within themedium-dose group (1.5 to 2.25 µg/kg/wk) and patients receiving placebo.Generally, the adverse event profile was similar between patients receivingdarbepoetin alfa and those receiving placebo, and consistent with that expectedfor patients with cancer receiving cytotoxic chemotherapy, with most patientsreported adverse events (95% of each treatment group), and nausea was the mostfrequently reported adverse event (Table 2). Most deaths (darbepoetin alfa 9%,placebo 7%) were attributed to cancer or cancer-related sequelae.

Identical percentages of patients in each group (5%) withdrewbecause of adverse events, most frequently due to tumor progression.Treatment-related adverse events were reported by similar proportions ofpatients in the darbepoetin alfa (11%) and placebo (7%) groups. When all adverseevents in general, and cardiovascular and thrombotic adverse events inparticular, were evaluated relative to hemoglobin concentrations, no obviouspatterns with respect to increased safety risk with high hemoglobinconcentrations and higher rates of rise of hemoglobin concentrations wereobserved.

Two adverse events were prospectively identified as being ofinterest based on specific safety concerns potentially associated with the useof rHuEPO in the nephrology setting, namely, thrombotic events and hypertension.The incidences of hypertension (4% darbepoetin alfa, 3% placebo) and thromboticevents (3% darbepoetin alfa, 2% placebo) were low and similar between thedarbepoetin alfa and placebo groups; this suggests that, in general, anemicpatients with cancer receiving darbepoetin alfa do not have a higher thanexpected risk of experiencing these events compared with those receivingplacebo. This is supported by the observation that the 95% confidence intervalsfor the odds ratios between darbepoetin alfa and placebo for hypertension orthrombotic events included 1, indicating no difference between the groups (Figure8).

Serum assays to detect antibodies to darbepoetin alfa wereperformed at multiple time points during the studies for patients treated withdarbepoetin alfa. During the 144 patient years of exposure across thestudies, no evidence of neutralizing antibody formation to darbepoetin alfa wasfound in the 729 patients who received darbepoetin alfa.


The series of analyses presented here are important for a numberof reasons. Most importantly, treating physicians need to know whether theresponse to therapy is well proven and robust. The individual studies conductedfor the darbepoetin alfa clinical program provide compelling results for adose-response relationship and an effect-dose range, whether one looks atefficacy based on hemoglobin variables or using transfusion-based endpoints.[17-19] This program has also shown the utility of weekly andonce-every-other-week dosing regimens. The pooled analysis reported herecomplements the data from the individual studies in a number of ways. Itprovides an estimate of effect for a nonmyeloid malignancy population as awhole, and by increasing the sample size to obtain a more precise estimate oftreatment effect. It also allows us to evaluate specific groups of patients,defined by either baseline characteristics or choice of concurrent treatments.

Despite many years of clinical use, the dose-responserelationship of rHuEPO is not well characterized across the general anemiccancer population. However, at the licensed dose and schedule for rHuEPO,hemoglobin response rates for rHuEPO have been previously reported to be between50% to 60%.[14] Using fixed dosing rather than weight-based dosing, and analternative measure of efficacy—hematopoietic response—rates ofapproximately 65% have been observed.[9,26] Seidenfeld and colleagues,[16] in arecent review of rHuEPO published literature, reported a wide range ofreductions in transfusion incidences (7% to 47%) compared with placebo.

The analyses presented here indicate that a dose-responserelationship does exist for darbepoetin alfa with respect to hemoglobin,including the time taken to observe a response. Low doses of darbepoetin alfaappear to result in notable improvements for both hemoglobin and transfusion endpoints compared with placebo. In these analyses, after 12 weeks ofdarbepoetin alfa therapy, the low-dose group (doses ranging from 0.5 to1.0 µg/kg/wk) had a mean increase in hemoglobin greater than 1 g/dLfrom baseline, a hematopoietic response rate of approximately 50%, and ahemoglobin response rate of approximately 40%.

In the medium-dose group (average weekly dose of 1.5 to2.25 µg/kg), the impact on the hemoglobin variables evaluated included anearly 2.0-g/dL mean increase in hemoglobin from baseline, hematopoieticresponse rates of approximately 65%, hemoglobin responses of approximately 55%,and median time to response of 10 weeks. In the high-dose group (average weeklydose of 2.5 to 8.0 µg/kg), the impact on the hemoglobin variables evaluatedincluded a greater than 2-g/dL increase in hemoglobin from baseline,hematopoietic response rates in excess of 75%, hemoglobin responses ofapproximately 70%, and a median time to response of less than 8 weeks.

In our analysis, anemic patients with cancer who were treatedwith darbepoetin alfa had a decreased likelihood of RBC transfusions, relativeto patients who received placebo. Additionally, in all groups, fewer patientsreceiving darbepoetin alfa required RBC transfusions compared with patientsreceiving placebo, regardless of whether analyzed only from week 5 to the end ofthe treatment phase (18% to 31% vs 50%, respectively) or for the entiretreatment period (28% to 38% vs 58%, respectively).

It is interesting to note that no apparent dose response wasobserved for RBC transfusions, in contrast to the dose response observed forchange in hemoglobin. The change in hemoglobin for the low-dose group wassubstantially higher than that for placebo, although not as great inthe low-dose group as those forthe other two dose groups. This indicates that the magnitude of hemoglobinincrease attained with relatively low doses appears to be sufficient to reduce transfusions, even whenbaseline hemoglobin levels are greater than 10 g/dL at the initiation oftreatment.

The characterization of the dose range in these analyses isimportant. For example, based on these data, if a patient has an inadequateinitial response to darbepoetin alfa, an increase in dose with darbepoetin alfashould be considered. However, unlike anemic patients with chronic renal failurewhere almost all patients respond to erythropoietic therapy, it appears that aminority of anemic patients with cancer are unresponsive to therapy, at leastwith respect to the commonly estimated measures of efficacy. The impact ofproinflammatory cytokines and other possible suppressors of erythropoiesis andthe use of iron supplementation may be important both to help identify thoseunlikely to respond robustly, as well as to provide adjunctive treatmentoptions. If a hyporesponsive/nonresponsive population can be identified eitherprospectively or early into treatment, studies designed to look at treatmentbenefit using alternative measures (eg, stabilization of hemoglobinconcentration) should also be considered.

Various patient characteristics have been historicallyassociated with potential risks of anemia in patients with cancer, including ageand sex. In addition, the severity and frequency of anemia have been observed tobe dependent upon both chemotherapy and tumor types.[5] In particular,platinum-containing compounds have been observed to induce a persistentdeficiency in erythropoietin during treatment.[28] Abels and colleagues[11]noted a slight difference in transfusion rates between patients receivingnonplatinum-containing chemotherapy (41%) and those receivingplatinum-containing chemotherapy (53%). When both hemoglobin response and RBCtransfusion rates were analyzed individually for these factors, patients treatedwith darbepoetin alfa still exhibited increased rates of hemoglobin response anddecreased rates of RBC transfusions, regardless of any of these characteristics.Additionally, no major difference in treatment effect size was apparent,indicating that none of the particular subgroups studied need to be considereddifferentially with regard to changes in starting doses.

Baseline hemoglobin categorized as either < 10 or ³ 10 g/dL, which is the threshold between mild and moderate anemia, hasalso been identified as being a potential factor influencing the perceivedbenefit of erythropoietic treatment in the anemic cancer patient population.Despite physicians in many countries using hemoglobin concentrations ofapproximately 12 g/dL or above as a therapeutic target, few patients havetreatment initiated above 10 g/dL. In a recent review article involving ameta-analysis of several studies of rHuEPO use in patients with cancer receivingchemotherapy, Seidenfeld and colleagues concluded that evidence was insufficientto justify rHuEPO use in patients with baseline hemoglobin concentrations ³ 10 g/dL.[16]

Our subgroup analysis indicated that regardless of baselinehemoglobin category, patients had a greater likelihood of a hemoglobin responseand a decreased risk of transfusions when treated with darbepoetin alfa,compared with placebo (Figures 6 and 7). This is important because, despite thehigh baseline hemoglobin in this group, the benefits of antianemia treatmentincluded a substantial reduction in the risk of RBC transfusion. This suggeststhat there may be a potential benefit in treating all anemia, evenmild-to-moderate anemia, as the treatment effects are not limited to thehemoglobin-mediated benefits in quality of life commonly observed in anemicpatients with cancer receiving chemotherapy.[2] When the point estimates of thetransfusion rates were examined, we observed that fewer patients receivingdarbepoetin alfa required transfusions compared with placebo patients.

Furthermore, we observed that in patients receiving darbepoetinalfa, only 18% of patients with a baseline hemoglobin ³ 10 g/dL required atransfusion, compared with 42% of patients with a baseline hemoglobin< 10 g/dL. This supports a hypothesis that treatment withdarbepoetin alfa may provide transfusion-sparing effects, regardless of baselinehemoglobin, in contrast to the observations of Seidenfeld et al.[16]

In general, the safety profile in these studies was consistentwith patients with cancer who were undergoing chemotherapy, with the safetyprofiles being similar between patients receiving darbepoetin alfa and thosereceiving placebo. Treatment with darbepoetin alfa was not associated with anincreased risk of adverse events, especially with respect to hypertension andthrombotic events, both of which had previously been identified as a concern inpatients undergoing hemodialysis and receiving treatment with rHuEPO.Furthermore, treatment with darbepoetin alfa was not associated with thedevelopment of neutralizing antibodies.

In summary, this analysis indicates that darbepoetin alfatreatment in anemic patients with cancer undergoing chemotherapy increases thelikelihood of a patient exhibiting a hemoglobin response (ie, a ³ 2.0-g/dLincrease in hemoglobin from baseline) and decreases the likelihood of a patientrequiring a RBC transfusion, regardless of baseline hemoglobin, chemotherapytype, tumor type, age, and sex.


1. Miller CB, Jones RJ, Piantadosi S, et al: Decreasederythropoietin response in patients with the anemia of cancer. N Engl J Med322:1689-1692, 1990.

2. Ludwig H, Fritz E: Anemia in cancer patients. Semin Oncol25:2-6, 1998.

3. Means RT Jr: Pathogenesis of the anemia of chronic disease: Acytokine-mediated anemia. Stem Cells 13:32-37, 1995.

4. Moliterno AR, Spivak JL: Anemia of cancer. Hematol Oncol ClinNorth Am 10:345-363, 1996.

5. Groopman JE, Itri LM: Chemotherapy-induced anemia in adults:Incidence and treatment. J Natl Cancer Inst 91:1616-1634, 1999.

6. Walker RH: Special report: Transfusion risks. Am J ClinPathol 88:374-378, 1987.

7. Goodnough LT, Brecher ME, Kanter MH, et al: Transfusionmedicine. First of two parts—Blood transfusion. N Engl J Med 340:438-447,1999.

8. Glaspy J, Bukowski R, Steinberg D, et al: Impact of therapywith epoetin alfa on clinical outcomes in patients with nonmyeloid malignanciesduring cancer chemotherapy in community oncology practice. J Clin Oncol15:1218-1234, 1997.

9. Demetri GD, Kris M, Wade J, et al: Quality-of-life benefit inchemotherapy patients treated with epoetin alfa is independent of diseaseresponse or tumor type: Results from a prospective community oncology study. JClin Oncol 16:3412-3425, 1998.

10. Maraveyas A, Pettengell R: What is the role oferythropoietin in patients with solid tumours? Ann Oncol 9:239-241, 1998.

11. Abels RI, Larholt KM, Krantz KD, et al: Recombinant humanerythropoietin (r-HuEPO) for the treatment of the anemia of cancer, in MurphyMJJ (ed): Blood Cell Growth Factors: Their Present and Future Use in Hematologyand Oncology; Proceedings of the Beijing Symposium, pp 121-141. Dayton, Ohio,AlphaMed, 1991.

12. Cascinu S, Fedeli A, Del Ferro E, et al: Recombinant humanerythropoietin treatment in cisplatin-associated anemia: A randomized,double-blind trial with placebo. J Clin Oncol 12:1058-1062, 1994.

13. De Campos E, Radford J, Steward W, et al: Clinical and invitro effects of recombinant human erythropoietin in patients receivingintensive chemotherapy for small-cell lung cancer. J Clin Oncol 13:1623-1631,1995.

14. Ludwig H, Sundal E, Pecherstorfer M, et al: Recombinanthuman erythropoietin for the correction of cancer associated anemia with andwithout concomitant cytotoxic chemotherapy. Cancer 76:2319-2329, 1995.

15. Oberhoff C, Neri B, Amadori D, et al: Recombinant humanerythropoietin in the treatment of chemotherapy-induced anemia and prevention oftransfusion requirement associated with solid tumors: A randomized, controlledstudy. Ann Oncol 9:255-260, 1998.

16. Seidenfeld J, Piper M, Flamm C, et al: Epoetin treatment ofanemia associated with cancer therapy: A systematic review and meta-analysis ofcontrolled clinical trials. J Natl Cancer Inst 93:1204-1214, 2001.

17. Vansteenkiste J, Pirker R, Massuti B, et al: Double-blind,placebo-controlled, randomized phase III trial of darbepoetin alfa in lungcancer patients receiving chemotherapy. J Natl Cancer Inst 94:1211-1220, 2002.

18. Hedenus M, Hansen S, Taylor K, et al: Randomised,dose-finding study of darbepoetin alfa in anaemic patients withlymphoproliferative malignancies. Br J Haematol 119:79-86, 2002.

19. Glaspy JA, Jadeja JS, Justice G, et al: Darbepoetin alfagiven every 1 or 2 weeks alleviates anaemia associated with cancer chemotherapy.Br J Cancer 87:268-276, 2002.

20. Glaspy J, Jadeja JS, Justice G, et al: Randomized,active-controlled, phase 1/2, dose-escalation study of NESP administered weeklyand every 2 weeks in patients with solid tumors (abstract 1546). Proc Am SocClin Oncol 20:387a, 2001.

21. Hedenus M, Hansen S, Dewey C, et al: A randomized, blinded,placebo-controlled, phase II, dose-finding study of novel erythropoiesisstimulating protein (NESP) in patients with lymphoproliferative malignancies(abstract 1569). Proc Am Soc Clin Oncol 20:393a, 2001.

22. Pirker R, Vansteenkiste J, Gateley J, et al: A phase 3,double-blind, placebo-controlled, randomized study of novel erythropoiesisstimulating protein (NESP) in patients undergoing platinum treatment for lungcancer (abstract 1572). Proc Am Soc Clin Oncol 20:394a, 2001.

23. Egrie JC, Grant JR, Gillies DK, et al: The role ofcarbohydrate on the biological activity of erythropoietin. Glycoconj J 10:263,1993.

24. Egrie JC, Browne JK: Development and characterisation ofnovel erythropoiesis stimulating protein (NESP). Br J Cancer 84(suppl 1):3-10,2001.

25. Kalbfleisch JD, Prentice RL: The Statistical Analysis ofFailure Time Data. New York, Wiley, 1980.

26. Gabrilove J, Cleeland CS, Livingston RB, et al: Clinicalevaluation of once-weekly dosing of epoetin alfa in chemotherapy patients:Improvements in hemoglobin and quality of life are similar to three-times-weeklydosing. J Clin Oncol 19:2875-2882, 2001.

27. Littlewood T, Bajetta E, Nortier J, et al: Effects ofepoetin alfa on hematologic parameters and quality of life in cancer patientsreceiving nonplatinum chemotherapy: Results of a randomized, double-blind,placebo-controlled trial. J Clin Oncol 19:2865-2874, 2001.

28. Wood PA, Hrushesky WJ: Cisplatin-associated anemia: Anerythropoietin deficiency syndrome. J Clin Invest 95:1650-1659, 1995.