Article Tools
ORIGINAL REPORTS
Palliative and Supportive Care
Article Tools
Pooled Analysis of Individual Patient-Level Data From All Randomized, Double-Blind, Placebo-Controlled Trials of Darbepoetin Alfa in the Treatment of Patients With Chemotherapy-Induced Anemia
Abstract
Although numerous clinical trials have demonstrated the efficacy and tolerability of erythropoiesis-stimulating agents (ESAs) in patients with chemotherapy-induced anemia (CIA), results of some recent trials and one meta-analysis have suggested that ESAs may negatively impact survival and/or disease control in patients with cancer.
To assess the benefits and risks of ESAs in CIA, we conducted a pooled analysis of individual patient-level data from all randomized, double-blind, placebo-controlled trials in 2,122 patients with CIA receiving darbepoetin alfa (DA; n = 1,200) or placebo (n = 912).
DA did not increase mortality (hazard ratio = 0.97; 95% CI, 0.85 to 1.1) and had no effect on progression-free survival (hazard ratio = 0.93; 95% CI, 0.84 to 1.04) and disease progression (hazard ratio = 0.92; 95% CI, 0.82 to 1.03), but, as expected, increased the risk for thromboembolic events (hazard ratio = 1.57; 95% CI, 1.10 to 2.26). Overall and progression-free survival were not affected by baseline hemoglobin and seemed better in patients who achieved hemoglobin more than 12 or more than 13 g/dL. Transfusions and rates of hemoglobin increase (> 1 g/dL in 14 days; > 2 g/dL in 28 days) owing to transfusions were associated with an increased risk for death and disease progression in both treatment groups; in the absence of transfusions, rates of hemoglobin increase did not appear to increase the risk for adverse outcomes. Compared with placebo, DA significantly reduced the risk of receiving one or more transfusion.
The clinical effectiveness of erythropoiesis-stimulating agents (ESAs) for improving hemoglobin concentrations and reducing the need for RBC transfusions in patients with chemotherapy-induced anemia (CIA) has been repeatedly demonstrated.1–6
Data from numerous placebo-controlled clinical trials4–12 and meta-analyses13–16 have demonstrated that ESAs are generally well tolerated in patients with CIA. However, results of some trials17–26 and one meta-analysis27 have suggested that ESAs negatively impact survival and/or disease control in patients with cancer. Of the eight clinical trials noted in the product labels for the ESAs marketed in the United States, four trials were in patients receiving chemotherapy,19,22,24,25 two trialswere in the nonapproved setting of radiotherapy18,21 (all six studies targeted hemoglobin levels greater than those currently recommended), and two trials were in the nonapproved setting of patients with advanced cancer not receiving chemotherapy or radiotherapy.17,20 In 2008, the product labels for ESAs were updated with a boxed warning highlighting potential safety risks.28,29
The association between ESA use and an increased risk of cardiovascular/thromboembolic events has been quantified in multiple placebo-controlled studies, has remained stable over time, and is documented in ESA product labels. A published meta-analysis confirmed the increased risk of thromboembolic events in patients with cancer treated with an ESA versus no ESA (relative risk [RR] = 1.67; 95% CI, 1.35 to 2.06; 35 trials, 6,769 patients).13 Despite this increased risk, however, ESAs did not increase mortality in this patient population (hazard ratio [HR] = 1.08; 95% CI, 0.99 to 1.18; 42 trials, 8,167 patients).13
To reassess the benefits and risks of ESAs in the currently approved indication of CIA, we conducted the first pooled analysis of individual patient data from all randomized, placebo-controlled clinical trials of darbepoetin alfa (DA). Analyzing individual patient data provides the ability to standardize the definition of end points to reduce heterogeneity across trials, analyze data with time-to-event methodology, and explore baseline patient characteristics that might predict outcome. Analysis of large numbers of patients also allows detection of smaller differences between treatment groups that may not otherwise be observed among individual studies. We evaluated the effect of DA versus placebo on safety outcomes and transfusions as well as the effect of baseline hemoglobin, hemoglobin events (hemoglobin achieved and hemoglobin rate of increase), and transfusions on safety outcomes.
All six studies included in this analysis were completed, randomized, blinded, placebo-controlled studies in which DA was administered subcutaneously to patients with nonmyeloid tumors receiving concomitant chemotherapy (Table 1). With the exception of the small-cell lung cancer trial,12 all patients received one or more cycles of chemotherapy, had baseline hemoglobin ≤ 11 g/dL, and were scheduled to receive future chemotherapy (with or without radiotherapy). In the small-cell lung cancer trial,12 patients had baseline hemoglobin ≥ 9 g/dL and ≤ 13 g/dL and received their first dose of chemotherapy and DA simultaneously. DA was withheld if hemoglobin increased more than 14 g/dL for women or more than 15 g/dL for men; however, in the trials published by Pirker et al (2007)12 and Taylor et al (2005),7 DA was withheld if hemoglobin increased more than 14 g/dL and more than 13 g/dL, respectively, irrespective of sex.
|
| Study | Total* (N = 2,112) | Phase | Duration of Treatment (weeks) | Long-Term Follow-Up | Median Follow-Up Time (months) | Placebo | DA | ||
|---|---|---|---|---|---|---|---|---|---|
| Dose | No. of Patients | Dose | No. of Patients | ||||||
| Solid tumors: schedule 110 | 249 | II | 12 | No | 2.8 | Q3W | 51 | 4.5 μg/kg Q3W | 32 |
| 6.75 μg/kg Q3W | 17 | ||||||||
| 9 μg/kg Q3W | 46 | ||||||||
| 12 μg/kg Q3W | 28 | ||||||||
| 13.5 μg/kg Q3W | 35 | ||||||||
| 15 μg/kg Q3W | 40 | ||||||||
| Solid tumors: schedule 230 | 156 | II | 12 | No | 2.8 | Q4W | 31 | 9 μg/kg Q4W | 31 |
| 12 μg/kg Q4W | 31 | ||||||||
| 15 μg/kg Q4W | 33 | ||||||||
| 18 μg/kg Q4W | 30 | ||||||||
| Lymphoproliferative malignancies31 | 66 | II | 12 | No | 3.5 | QW | 11 | 1 μg/kg QW | 11 |
| 2.25 mcg/kg QW | 22 | ||||||||
| 4.5 μg/kg QW | 22 | ||||||||
| Lung cancer; treated with platinum chemotherapy5 | 314 | III | 12 | Yes | 7.5 | QW | 158 | 2.25 μg/kg QW | 156 |
| Lymphoproliferative malignancies(22,23) | 344 | III | 12 | Yes | 29.4 | QW | 170 | 2.25 μg/kg QW | 174 |
| Nonmyeloid malignancies7 | 386 | III | 15 | No | 4.2 | Q3W | 193 | 300 μg Q3W | 193 |
| Small-cell lung cancer; receiving platinum chemotherapy12 | 597 | III | 18 | Yes | 8.6 | QW for 4 weeks then Q3W | 298 | 300 μg QW for 4 weeks then 300 μg Q3W | 299 |
Abbreviations: DA, darbepoetin alfa; QW, every week; Q3W, every 3 weeks; Q4W, every 4 weeks.
*No. of patients randomly assigned who received at least one dose of investigational product.
Safety end points included overall survival, progression-free survival, and disease progression during treatment and in long-term follow-up and deaths and incidence of adverse events of interest (events identified resulting from known or hypothesized safety concerns) during treatment.
Deaths were identified from1 reasons given for discontinuation of study drug or study,2 a reported fatal adverse event; or3 the long-term follow-up case report form. Except for the small-cell lung cancer trial,12 collection of disease progression data was not prespecified; therefore, caution should be used when interpreting the progression-based end points. Disease progression was identified if1 the reason for discontinuing study drug/study was reported as disease progression,2 end-of-study disease status was progressive disease, or3 disease progression was noted on the long-term follow-up case report form. If disease progression was not noted, patients were censored at end of study or last long-term follow-up assessment. Time-to-disease progression was based on the earliest date that disease progression was reported; progression-free survival time was based on date of progression or date of death, whichever was earlier.
Adverse event data were coded using MedDRA, version 9. A prespecified list of preferred terms was used to identify adverse events of interest (cardiovascular/thromboembolic events, seizures, and hypertension). Adverse events were selected without regard to severity of event and reported relationship to investigational product.
The incidence of one or more transfusion was measured from week 5 to the end of treatment phase.
Safety and efficacy end points were analyzed using all randomly assigned patients who received one or more dose of investigational product. Patients were analyzed by randomized treatment group (patients randomly assigned into different DA dose groups within a study were treated as a single DA group); studies that allowed patients randomly assigned to placebo to crossover to DA were truncated to the double-blind treatment for analysis purposes. Data were analyzed using SAS version 9 (SAS Institute, Cary, NC).
Baseline demographics and clinical characteristics were summarized by number and percentage for categoric variables and mean (standard deviation) for continuous variables. Kaplan-Meier survival curves were created for overall survival, progression-free survival, and disease progression; all included long-term follow-up data.
The effect of investigational product on time-to-event end points (overall survival, progression-free survival, disease progression, adverse events, and transfusions) was characterized using Cox proportional hazards models stratified by study protocol. The first occurrence of an event was used in these analyses. The impact of baseline hemoglobin on the treatment effect of DA was characterized by estimating HRs for baseline hemoglobin subgroups (< 9 g/dL, ≥ 9 to 10 g/dL, ≥ 10 to 11 g/dL, ≥ 11 to 12 g/dL, and ≥ 12 g/dL).
Analyses were done to examine whether achieving hemoglobin more than 12 g/dL or more than 13 g/dL, an increase in hemoglobin more than 1 g/dL in 14 days or more than 2 g/dL in 28 days, or receipt of one or more transfusions were associated with an increased risk of death, disease progression, or thromboembolic events. The impact of each factor was assessed individually as a time-dependent covariate in a Cox proportional hazards model; HRs (with 95% CIs) were calculated. Analyses were performed for patients receiving DA and placebo separately. Analyses examining the impact of transfusions on hemoglobin were also performed.
The meta-analyses were examined for heterogeneity using Cochrane's Q and the I2 statistic. If evidence of heterogeneity existed, it was noted in the Results, and further analyses were conducted that assumed the effects being estimated in the different studies are not identical. The Cox models were examined for proportional hazards and any significant departures noted. Baseline imbalances were adjusted for multivariate models, and results are presented if estimated treatment effects differed meaningfully from the univariate models.
The difference in the incidence of adverse events (DA minus placebo) was calculated with 95% CIs.
The incidence of transfusions (with 95% CI) between week 5 and the end of treatment period was calculated by baseline hemoglobin category for each treatment group.
This integrated analysis included data from six randomized, double-blind, placebo-controlled trials in CIA conducted in 2,122 patients receiving DA (n = 1,200) or placebo (n = 912;Table 1). One trial of DA versus placebo in lymphoproliferative malignancies had worse overall survival in the DA group (estimated relative risk = 1.36; 95% CI, 1.02-1.82).23
Most patients were white, mean age was similar in both groups, and the proportion of women was higher in the DA group. A higher proportion of patients in the placebo group received platinum-containing chemotherapy (Table 2).
|
| Characteristic | Darbepoetin Alfa (n = 1,200) | Placebo (n = 912) | Total (N = 2,112) | |||
|---|---|---|---|---|---|---|
| No. | % | No. | % | No. | % | |
| Sex | ||||||
| Men | 596 | 49.7 | 493 | 54.1 | 1,089 | 51.6 |
| Women | 604 | 50.3 | 419 | 45.9 | 1,023 | 48.4 |
| Race | ||||||
| White | 1,129 | 94.1 | 866 | 95.0 | 1,995 | 94.5 |
| Black | 30 | 2.5 | 22 | 2.4 | 52 | 2.5 |
| Other* | 41 | 3.4 | 24 | 2.6 | 65 | 3.1 |
| Age, years | ||||||
| Mean | 61.9 | 62.0 | 61.9 | |||
| SD | 11.6 | 10.7 | 11.3 | |||
| < 65 | 659 | 54.9 | 522 | 57.2 | 1,181 | 55.9 |
| ≥ 65 | 541 | 45.1 | 390 | 42.7 | 931 | 44.1 |
| Platinum chemotherapy | 656 | 54.7 | 571 | 62.6 | 127 | 58.1 |
| Tumor types | ||||||
| Lung | 532 | 44 | 505 | 55 | 1,037 | 49 |
| Hematologic | 255 | 21 | 217 | 24 | 472 | 22 |
| Breast | 144 | 12 | 62 | 7 | 206 | 10 |
| Gastrointestinal | 106 | 9 | 56 | 6 | 162 | 8 |
| Gynecologic | 81 | 7 | 31 | 3 | 112 | 5 |
| Genitourinary | 41 | 3 | 10 | 1 | 51 | 2 |
| Other† | 41 | 3 | 31 | 3 | 72 | 3 |
| Disease stage | ||||||
| Stage II or lower/limited | 128 | 10.7 | 99 | 10.9 | 227 | 10.7 |
| Stage III or higher/extensive | 1,037 | 86.4 | 782 | 85.7 | 1,819 | 86.1 |
| Other | 23 | 1.9 | 17 | 1.9 | 40 | 1.9 |
| Unknown | 12 | 1.0 | 14 | 1.5 | 26 | 1.2 |
| ECOG performance status | ||||||
| 0 | 282 | 23.5 | 179 | 19.6 | 461 | 21.8 |
| 1 | 648 | 54.0 | 495 | 54.3 | 1,143 | 54.1 |
| 2 | 162 | 13.5 | 150 | 16.4 | 312 | 14.8 |
| 3 | 10 | 0.8 | 7 | 0.8 | 17 | 0.8 |
| Unknown | 98 | 8.2 | 81 | 8.9 | 179 | 8.5 |
| Baseline hemoglobin, g/dL | 1,169 | 876 | 2,045 | |||
| Mean | 10.5 | 10.5 | 10.5 | |||
| SD | 1.4 | 1.4 | 1.4 | |||
| Baseline endogenous EPO, mU/mL | 1,137 | 859 | 1,996 | |||
| Mean | 76.3 | 84.1 | 79.7 | |||
| SD | 134.5 | 209.5 | 170.9 | |||
Abbreviations: SD, standard deviation; ECOG, Eastern Cooperative Oncology Group; EPO, erythropoietin.
*Includes Asian, Hispanic, Japanese, Native American, Pacific Islander, and other.
†Includes bone sarcoma, head and neck, melanoma, oral, soft tissue sarcoma, unknown primary, other.
The most frequent primary tumor types were lung and hematologic cancers, reflecting the three large, phase III studies10,12,22 conducted in these patient populations (Table 2). These tumor types accounted for 65% of patients in the DA group and 79% in the placebo group. Most patients had later-stage disease, defined as stage III or higher/extensive. Eastern Cooperative Oncology Group performance status, baseline hemoglobin concentration, and baseline erythropoietin were generally well balanced between the two groups (Table 2).
For overall survival, progression-free survival, and disease progression (all including long-term follow-up), the Kaplan-Meier curves for DA and placebo nearly overlap, with HR of 0.97 (95% CI, 0.85 to 1.1) for overall survival, 0.93 (95% CI, 0.84 to 1.04) for progression-free survival, and 0.92 (95% CI, 0.82 to 1.03) for disease progression (Fig 1).
Fig 1. Kaplan-Meier curves of (A) overall survival, (B) progression-free survival, and (C) disease progression, all including long-term follow-up. Kaplan-Meier curves combined all patients without stratification by study protocol. HR, hazard ratio; DA, darbepoetin alfa.
The HR of on-study deaths suggested a slightly higher risk for death in the DA versus placebo group; however, the 95% CI spanned unity (HR = 1.11; 95% CI, 0.84 to 1.47;Table 3). The risk of having a cardiovascular/thromboembolic event was higher for DA compared with placebo (HR = 1.26; 95% CI, 1.00 to 1.59), primarily as a result of the higher incidence of thromboembolic events in the DA group (8% v5% in placebo; HR = 1.57; 95% CI, 1.10 to 2.26;Table 3). Myocardial infarction/coronary artery disease was also reported at a higher rate in the DA group; however, the percentage of patients was low and the CI for the HR was wide (Table 3). The incidence of cardiovascular events such as arrhythmia, cerebrovascular accident, and congestive heart failure was similar in both groups (Table 3).
|
| Event | Darbepoetin Alfa (n = 1,200) | Placebo (n = 912) | Hazard Ratio | 95% CI | ||
|---|---|---|---|---|---|---|
| No. | % | No. | % | |||
| Death | 110 | 9.2 | 92 | 10.1 | 1.11 | 0.84 to 1.47 |
| Adverse events of historical interest | ||||||
| Cardiovascular and thromboembolic events | 192 | 16.0 | 127 | 13.9 | 1.26 | 1.00 to 1.59 |
| Arrhythmia | 57 | 4.8 | 43 | 4.7 | 1.15 | 0.77 to 1.72 |
| Cerebrovascular accident | 20 | 1.7 | 18 | 2.0 | 1.05 | 0.55 to 1.98 |
| Congestive heart failure | 20 | 1.7 | 23 | 2.5 | 0.81 | 0.44 to 1.48 |
| Myocardial infarction or coronary artery disease | 21 | 1.8 | 13 | 1.4 | 1.34 | 0.66 to 2.71 |
| Embolism or thrombosis | 96 | 8.0 | 46 | 5.0 | 1.57 | 1.10 to 2.26 |
| Seizure | 7 | 0.6 | 11 | 1.2 | 0.62 | 0.24 to 1.60 |
| Hypertension | 47 | 3.9 | 31 | 3.4 | 1.33 | 0.84 to 2.10 |
DA did not increase mortality and had no effect on progression-free survival irrespective of baseline hemoglobin concentration—the 95% CIs spanned unity (Figs 2A and 2B). The expected increased risk of thromboembolic events associated with DA showed a tendency toward an association with baseline hemoglobin (Fig 2C).
Fig 2. (A) Risk of death, (B) disease progression, and (C) thromboembolic events by baseline hemoglobin concentration. The impact of baseline hemoglobin on the treatment effect of darbepoetin alfa was assessed by estimating hazard ratios (HRs) for baseline hemoglobin subgroups (< 9 g/dL, ≥ 9 to 10 g/dL, ≥ 10 to 11 g/dL, ≥ 11 to 12 g/dL, and ≥ 12 g/dL).
Not only DA but also transfusions could potentially influence the percentage of patients achieving hemoglobin concentrations more than 12 g/dL or more than 13 g/dL. Therefore, an exploratory investigation of the effect of transfusions on these hemoglobin levels was performed. The percentage of patients who achieved hemoglobin more than 12 g/dL or more than 13 g/dL was only slightly higher in the presence versus absence of transfusions: 37.8% versus 35.6% for DA; 15.8% versus 13.8% for placebo. Likewise, only small differences were observed for patients who achieved hemoglobin more than 13 g/dL: 62.4% versus 57.6% for DA; 40.8% versus 33.7% for placebo. Because transfusions had little impact on hemoglobin more than 12 g/dL or more than 13 g/dL, all hemoglobin values were used to evaluate the effect of hemoglobin achieved on adverse outcomes. Overall survival and progression-free survival seemed to be better in patients who achieved hemoglobin more than 12 g/dL or more than 13 g/dL as compared with those who did not (Fig 3A). There was no clear relationship between risk for embolism/thrombosis and achievement of hemoglobin more than 12 g/dL or more than 13 g/dL.
Fig 3. Risk of adverse outcomes resulting from (A) hemoglobin (Hb) ceiling of more than 12 g/dL or more than 13 g/dL; (B) Hb increase more than 1 g/dL in 14 days or (C) Hb increase more than 2 g/dL in 28 days in the presence and absence of transfusions; (D) transfusions. The impact of each factor was assessed individually as a time-dependent covariate in a Cox proportional hazards model. (*) n = 1, 200; (†) n = 912; (‡) n = 912. HR, hazard ratio; DA, darbepoetin alfa; PFS, progression-free survival.
We also investigated the effect of transfusions on rates of hemoglobin increase. In the absence of transfusions, the percentage of patients with a more than 1 g/dL in 14 days or more than 2 g/dL in 28 days increase in hemoglobin was 68.8% for DA and 52.3% for placebo or 39.1% for DA and 19.2% for placebo, respectively. When only hemoglobin values due to transfusions were included in the analysis, the percentage of patients with a more than 1 g/dL in 14 days or more than 2 g/dL in 28 days increase in hemoglobin was similar in both treatment groups (86.7% for DA, 87.5% for placebo; or 77.6% for DA, 75.3% for placebo, respectively). Therefore, we investigated the effect of hemoglobin rates of increase resulting from transfusions versus those not resulting from transfusions on adverse outcomes. A more than 1 g/dL in 14 days or more than 2 g/dL in 28 days increase in hemoglobin resulting from transfusions was associated with an increased risk of death and disease progression (Figs 3B and 3C). When transfusions were excluded from the analysis, these rates of hemoglobin increase were not associated with an increased risk for death or disease progression (Figs 3B and 3C). Again, there was no clear association between the risk for embolism/thrombosis and rate of hemoglobin increase in the presence or absence of transfusions (Figs 3B and 3C).
Transfusions were associated with a greater risk for death and disease progression in both treatment groups and with a greater risk for embolism/thrombosis in the DA group (Fig 3D).
Reductions in the risk of receiving one or more transfusions from week 5 to the end of treatment period were observed for DA relative to placebo, overall and in each baseline hemoglobin category (Appendix Fig A1, online only). As expected, the incidence of transfusions in patients who initiated DA treatment at baselinehemoglobin 10 to 11 g/dL (19%) was less than half the incidence observed in patients with baseline hemoglobin less than 9 g/dL (41%; Appendix Fig A1, online only).
Results of some clinical studies in patients with cancer have raised concerns regarding a potential relationship between ESAs and increased risk of mortality and/or disease control. These negative safety signals have been discussed with the United States Food and Drug Administration at three Oncologic Drugs Advisory Committee meetings (May 2004, May 2007, and March 2008) and have been incorporated into the product labels in a boxed warning. Also, in light of these safety concerns, the National Comprehensive Cancer Network and the American Society of Clinical Oncology/American Society ofHematology updated their guidelines for the treatment of anemia in patients with cancer.32,33 It should also be noted that a recently published study-level meta-analysis27 and a recent abstract on a patient-level meta-analysis33a have reported a negative impact of ESA use on survival in cancer patients.
In this analysis of pooled individual data from patients with CIA, we did not detect increased mortality or increased disease progression with administration of DA. Likewise, a number of study-level meta-analyses of patients with cancer did not show a clear effect of ESAs on mortality or disease progression in those patients who received chemotherapy.13–16 Notably, those meta-analyses included trials conducted both within and outside the currently approved label.
Of the studies of ESA use in patients with cancer that resulted in negative safety outcomes, three studies initiated ESA therapy at relatively high baseline hemoglobin levels,19,21,25 and one study noted poorer outcomes in patients with baseline hemoglobin more than 11 g/dL.18 These observations have lead to speculation that high baseline hemoglobin may contribute to adverse outcomes. In the present analysis and in two large, study-level meta-analyses,13,14 however, there was no effect of baseline hemoglobin on mortality and/or disease progression in patients with cancer treated with ESAs.
It has also been suggested that high hemoglobin targets lead to negative safety outcomes: six of the trials with safety concerns targeted hemoglobin levels greater than the current upper safety limit of 12 g/dL.18,19,21,22,24,25 In the present study, there was no increased risk of death or disease progression among patients who achieved hemoglobin more than 12 or more than 13 g/dL or among patients not receiving transfusions who had rates of hemoglobin increase more than 1 g/dL in 14 days or more than 2 g/dL in 28 days. Patients' ability to achieve higher hemoglobin levels or increases in hemoglobin more than 1 g/dL in 14 days or more than 2 g/dL in 28 days may be reflective of a better biologic subgroup; therefore, this analysis may be confounded with responsiveness to treatment. Given these findings, we speculate that the negative safety signals observed in trials with high target hemoglobins may be attributed, in part, to patients who do not respond to ESAs. Future clinical trials to investigate whether failure to achieve the target hemoglobin level contributes to greater mortality would be informative.
Of note, transfusions and rates of hemoglobin increase due to transfusions were associated with increased risk of mortality and disease progression among patients treated with either DA or placebo. These negative outcomes may be confounded by the underlying health of the patients, since a requirement for transfusions may indicate poorer health. Thus, causality cannot be determined in this kind of analysis. Interestingly, an increased risk for death in patients receiving preoperative blood transfusions has been observed in a recent meta-analysis.34 Because the relationship between adverse outcomes and transfusions in patients with an increase in hemoglobin is a clinically relevant issue, further research in this area is warranted.
Research has demonstrated that patients with cancer are at a higher risk for thrombotic events relative to individuals without cancer.35,36 In our analysis, DA increased the absolute risk for thromboembolic events from 5% to 8% (57% relative risk), which is slightly lower than the 67% to 68% increased relative risk observed in two large study-level meta-analyses of ESAs in patients with cancer (both meta-analyses noted that the potential thrombophilic properties of ESAs are dependent on the underlying risk for thromboembolic complications).13,14 In concordance with the meta-analysis conducted by Seidenfeld et al (2006),14 the expected increased risk for thromboembolic events in our analysis did not seem to be strongly associated with baseline hemoglobin, and there was no clear effect of achieved hemoglobin levels. Likewise, there seemed to be no increased risk resulting from rates of hemoglobin increase more than 1 g/dL in 14 days or more than 2 g/dL in 28 days. However, to reduce the risk for thromboembolic events, health care providers should use the lowest dose of ESAs that will gradually increase hemoglobin to a level that will avoid the need for transfusions, as stated in the prescribing information.28,29
The increased risk of cardiovascular/thromboembolic events in oncology patients receiving ESAs might explain the increases in mortality associated with ESAs observed in several studies with nonanemic patients and/or those that targeted high hemoglobin concentrations.19,20,24 However, although increases in thromboembolic events have been observed almost uniformly across clinical trials and meta-analyses, adverse effects on survival have not. Thus other unknown factors, in addition to the known risk of thromboembolic events, may be contributing to the observed adverse outcomes in those studies.
The present analysis reaffirmed the benefit of ESAs in reducing the risk of receiving a transfusion. Overall, patients treated with DA had a 54% lower risk of transfusions than patients treated with placebo. As expected, the absolute risk was dependent on baseline hemoglobin: HR was 0.50 at ≥ 9 to 10 g/dL and 0.32 at ≥ 12 g/dL. Similarly, a meta-analysis of studies that compared early intervention (generally, initiation of therapy < 11 g/dL) versus late intervention (generally, initiation of therapy when hemoglobin level decreases to < 10 g/dL) with ESA use demonstrated an approximate 50% reduction in the risk of transfusion favoring the early intervention approach.37 Also, a meta-analysis of 42 trials with 6,510 patients showed more of a reduction in the RR for transfusion in ESA-treated patients when treatment was initiated at hemoglobin 10 to 12 g/dL versus less than 10 g/dL (RR = 0.46, 95% CI, 0.40 to 0.53; v RR = 0.70, 95% CI, 0.65 to 0.70).13 Avoidance of transfusion is an important clinical benefit provided by ESAs; use of transfusions to relieve the symptoms of anemia carries inherent risks, which have been well documented,38–40 and transfusions are also disruptive for patients, caregivers, and physicians.
In summary, many clinical trials that evaluated ESAs for the treatment of CIA did not negatively impact survival or increase adverse cancer outcomes and provided significant health benefits by reducing transfusion requirements. Further support for these observations is provided in the present analysis. We acknowledge that our results are not applicable to the general population and do not negate the negative safety signals observed in some trials of ESAs in patients with cancer, most of which were conducted outside the approved indication; however, the lack of a clear effect of ESAs on the risk of death or disease progression in this analysis indicates that further research is needed to answer this question.
Supported (analysis) by Amgen Inc.
Presented in part at the 14th European Cancer Conference, September 23-27, 2007, Barcelona, Spain; the American Society of Hematology 49th Annual Meeting and Exposition, December 8-11, 2007, Atlanta, GA; and the Hematology Oncology Pharmacy Association/International Society of Oncology Pharmacy Practitioners Joint Annual Conference, June 18-21, 2008, Anaheim, CA.
Authors' disclosures of potential conflicts of interest and author contributions are found at the end of this article.
Clinical Trials repository link available on JCO.org.
Clinical trial information can be found for the following: AMG 980291, AMG 20010145, AMG 990114, AMG 20030232, AMG 98297.
Although all authors completed the disclosure declaration, the following author(s) indicated a financial or other interest that is relevant to the subject matter under consideration in this article. Certain relationships marked with a “U” are those for which no compensation was received; those relationships marked with a “C” were compensated. For a detailed description of the disclosure categories, or for more information about ASCO's conflict of interest policy, please refer to the Author Disclosure Declaration and the Disclosures of Potential Conflicts of Interest section in Information for Contributors.
Employment or Leadership Position: Alex Fleishman, Amgen Inc (C); Ken Bridges, Amgen Inc (C)Consultant or Advisory Role: Heinz Ludwig, Amgen (C), Roche (C), Ortho Biotech (C); Jeffrey Crawford, Amgen (C); David H. Henry, Amgen (C), Ortho Biotech (C), Watson Pharmaceutical (C); John A. Glaspy, Amgen (C)Stock Ownership: Alex Fleishman, Amgen Inc; Ken Bridges, Amgen IncHonoraria: Heinz Ludwig, Amgen, Roche, Ortho Biotech; Jeffrey Crawford, Amgen; Anders Österborg, AmgenResearch Funding: Jeffrey Crawford, Amgen; Anders Österborg, Amgen; Johan Vansteenkiste, Amgen; David H. Henry, Amgen, Ortho Biotech, Watson Pharmaceutical; John A. Glaspy, Amgen, Johnson & JohnsonExpert Testimony: None Other Remuneration: None
Conception and design: Heinz Ludwig, Jeffrey Crawford, Anders Österborg, Johan Vansteenkiste, David H. Henry, Alex Fleishman, Ken Bridges, John A. Glaspy
Provision of study materials or patients: Heinz Ludwig
Collection and assembly of data: Alex Fleishman
Data analysis and interpretation: Heinz Ludwig, Jeffrey Crawford, Anders Österborg, Johan Vansteenkiste, David H. Henry, Alex Fleishman, Ken Bridges, John A. Glaspy
Manuscript writing: Heinz Ludwig, Jeffrey Crawford, Anders Österborg, Johan Vansteenkiste, David H. Henry, Alex Fleishman, Ken Bridges, John A. Glaspy
Final approval of manuscript: Heinz Ludwig, Jeffrey Crawford, Anders Österborg, Johan Vansteenkiste, David H. Henry, Alex Fleishman, Ken Bridges, John A. Glaspy
| 1. | GD Demetri, M Kris, J Wade , etal: Quality-of-life benefit in chemotherapy patients treated with epoetin alfa is independent of disease response or tumor type: Results from a prospective community oncology study—Procrit Study Group J Clin Oncol 16: 3412– 3425,1998 Link, Google Scholar |
| 2. | J Glaspy: The impact of epoetin alfa on quality of life during cancer chemotherapy: A fresh look at an old problem Semin Hematol 34: 20– 26,1997 Medline, Google Scholar |
| 3. | JL Gabrilove, CS Cleeland, RB Livingston , etal: Clinical evaluation of once-weekly dosing of epoetin alfa in chemotherapy patients: Improvements in hemoglobin and quality of life are similar to three-times-weekly dosing J Clin Oncol 19: 2875– 2882,2001 Link, Google Scholar |
| 4. | TJ Littlewood, E Bajetta, JW Nortier , etal: Effects of epoetin alfa on hematologic parameters and quality of life in cancer patients receiving nonplatinum chemotherapy: Results of a randomized, double-blind, placebo-controlled trial J Clin Oncol 19: 2865– 2874,2001 Link, Google Scholar |
| 5. | J Vansteenkiste, R Pirker, B Massuti , etal: Double-blind, placebo-controlled, randomized phase III trial of darbepoetin alfa in lung cancer patients receiving chemotherapy J Natl Cancer Inst 94: 1211– 1220,2002 Crossref, Medline, Google Scholar |
| 6. | A Osterborg, Y Brandberg, V Molostova , etal: Randomized, double-blind, placebo-controlled trial of recombinant human erythropoietin, epoetin beta, in hematologic malignancies J Clin Oncol 20: 2486– 2494,2002 Link, Google Scholar |
| 7. | K Taylor, P Ganly, V Charu , etal: Randomized, double-blind, placebo-controlled study of darbepoetin alfa every 3 weeks for the treatment of chemotherapy-induced anemia Blood 106,2005 abstr 3556 Medline, Google Scholar |
| 8. | T Grote, AL Yeilding, R Castillo , etal: Efficacy and safety analysis of epoetin alfa in patients with small-cell lung cancer: A randomized, double-blind, placebo-controlled trial J Clin Oncol 23: 9377– 9386,2005 Link, Google Scholar |
| 9. | TE Witzig, PT Silberstein, CL Loprinzi , etal: Phase III, randomized, double-blind study of epoetin alfa compared with placebo in anemic patients receiving chemotherapy J Clin Oncol 23: 2606– 2617,2005 Link, Google Scholar |
| 10. | D Kotasek, G Steger, W Faught , etal: Darbepoetin alfa administered every 3 weeks alleviates anaemia in patients with solid tumours receiving chemotherapy: Results of a double-blind, placebo-controlled, randomised study Eur J Cancer 39: 2026– 2034,2003 Crossref, Medline, Google Scholar |
| 11. | V Moebus, H Lueck, C Thomssen , etal: The impact of epoetin-alpha on anemia, red blood cell (RBC) transfusions, and survival in breast cancer patients (pts) treated with dose-dense sequential chemotherapy: Mature results of an AGO phase III study (ETC trial) J Clin Oncol 25: 20s,2007 suppl abstr 569 Google Scholar |
| 12. | R Pirker, R Ramlau, W Schuette , etal: A phase 3 randomized, double blind, placebo-controlled study of patients with previously untreated extensive-stage small cell lung cancer (SCLC) treated with platinum plus etoposide chemotherapy with or without darbepoetin alfa J Thorac Oncol 2,2007 abstr PD6-3-6 Google Scholar |
| 13. | J Bohlius, J Wilson, J Seidenfeld , etal: Recombinant human erythropoietins and cancer patients: Updated meta-analysis of 57 studies including 9353 patients J Natl Cancer Inst 98: 708– 714,2006 Crossref, Medline, Google Scholar |
| 14. | J Seidenfeld, M Pipe, JB , etal: Comparative effectiveness of epoetin and darbepoetin for managing anemia in patients undergoing cancer treatment: Comparative Effectiveness Review no. 3 Prepared by Blue Cross and Blue Shield Association Evaluation Center Evidence-based Practice Center under contract no. 290-02-0026 2006 Rockville, MD Agency for Healthcare Research and Quality www.effectivehealthcare.arhq.gov/reports/final.cfm Google Scholar |
| 15. | SD Ross, IE Allen, DH Henry , etal: Clinical benefits and risks associated with epoetin and darbepoetin in patients with chemotherapy-induced anemia: A systematic review of the literature Clin Ther 28: 801– 831,2006 Crossref, Medline, Google Scholar |
| 16. | M Aapro, B Coiffier, J Dunst , etal: Effect of treatment with epoetin beta on short-term tumour progression and survival in anaemic patients with cancer: A meta-analysis Br J Cancer 95: 1467– 1473,2006 Crossref, Medline, Google Scholar |
| 17. | RE Smith Jr, MS Aapro, H Ludwig , etal: Darbepoetin alpha for the treatment of anemia in patients with active cancer not receiving chemotherapy or radiotherapy: Results of a phase III, multicenter, randomized, double-blind, placebo-controlled study J Clin Oncol 26: 1040– 1050,2008 Link, Google Scholar |
| 18. | M Henke, R Laszig, C Rube , etal: Erythropoietin to treat head and neck cancer patients with anaemia undergoing radiotherapy: Randomised, double-blind, placebo-controlled trial Lancet 362: 1255– 1260,2003 Crossref, Medline, Google Scholar |
| 19. | B Leyland-Jones, V Semiglazov, M Pawlicki , etal: Maintaining normal hemoglobin levels with epoetin alfa in mainly nonanemic patients with metastatic breast cancer receiving first-line chemotherapy: A survival study J Clin Oncol 23: 5960– 5972,2005 Link, Google Scholar |
| 20. | JR Wright, YC Ung, JA Julian , etal: Randomized, double-blind, placebo-controlled trial of erythropoietin in non-small-cell lung cancer with disease-related anemia J Clin Oncol 25: 1027– 1032,2007 Link, Google Scholar |
| 21. | J Overgaard, CM Hoff, H Sand Hansen , etal: Randomized study of the importance of novel erythropoiesis stimulating protein (Aranesp®) for the effect of radiotherapy in patients with primary squamous cell carcinoma of the head and neck (HNSCC): The Danish Head and Neck Cancer Group (DAHANCA 10) randomized trial Presented at 14th European Cancer Conference September 23-27, 2007 Barcelona, Spain abstr 6LB Google Scholar |
| 22. | M Hedenus, M Adriansson, J San Miguel , etal: Efficacy and safety of darbepoetin alfa in anaemic patients with lymphoproliferative malignancies: A randomized, double-blind, placebo-controlled study Br J Haematol 122: 394– 403,2003 Crossref, Medline, Google Scholar |
| 23. | Safety of Erythropoiesis-Stimulating Agents (ESAs) in Oncology: Background Information for Oncologic Drugs Advisory Committee 2003 Amgen: Update to Hedenus et al http://www.fda.gov/ohrms/dockets/ac/07/briefing/2007-4301b2-01-01-Amgen.pdf Google Scholar |
| 24. | G Thomas, S Ali, FJ Hoebers , etal: Phase III trial to evaluate the efficacy of maintaining hemoglobin levels above 12.0 g/dL with erythropoietin vs above 10.0 g/dL without erythropoietin in anemic patients receiving concurrent radiation and cisplatin for cervical cancer Gynecol Oncol 108: 317– 325,2008 Crossref, Medline, Google Scholar |
| 25. | Amgen announces interim results of Aranesp “PREPARE” study in breast cancer patients [Amgen press release]: November 30, 2007 Amgen Http://wwwext.amgen.com/media/media_pr_detail.jsp?year=2007&releaseID=1083091 Google Scholar |
| 26. | D Azria, A Zouhair, A Serre , etal: [Anemia in head and neck cancers] Bull Cancer 92: 445– 451,2005 Medline, Google Scholar |
| 27. | CL Bennett, SM Silver, B Djulbegovic , etal: Venous thromboembolism and mortality associated with recombinant erythropoietin and darbepoetin administration for the treatment of cancer-associated anemia JAMA 299: 914– 924,2008 Crossref, Medline, Google Scholar |
| 28. | Aranesp® (Darbepoetin alfa) Package Insert 2008 Amgen Inc Thousand Oaks, CA Amgen Google Scholar |
| 29. | L.P. PROCRIT® (Epoetin alfa) package insert 2008 Ortho Biotech Products Raritan, NJ Ortho Biotech Google Scholar |
| 30. | D Kotasek, M Albertsson, J Mackey , etal: Once per cycle dosing of darbepoetin alfa is feasible in anemic cancer patients receiving chemotherapy Ann Oncol 13,2002 abstr 626PD Medline, Google Scholar |
| 31. | M Hedenus, S Hansen, K Taylor , etal: Randomized, dose-finding study of darbepoetin alfa in anaemic patients with lymphoproliferative malignancies Br J Haematol 119: 79– 86,2002 Crossref, Medline, Google Scholar |
| 32. | JD Rizzo, MR Somerfield, KL Hagerty , etal: Use of epoetin and darbepoetin in patients with cancer: 2007 American Society of Clinical Oncology/American Society of Hematology clinical practice guideline update J Clin Oncol 26: 132– 149,2008 Link, Google Scholar |
| 33. | GM Rodgers: NCCN Practice Guidelines in Oncology: Cancer and treatment-related anemia, National Comprehensive Cancer Network Http://www.nccn.org/professionals/physician_gls/PDF/anemia.pdf Google Scholar |
| 33a. | J Bohlius, C Brillant, M Clarke , etal: Recombinant human erythropoiesis stimulating agents in cancer patients: Individual patient data meta-analysis on behalf of the EPO IPD Meta-Analysis Collaborative Group Blood 112,2008 abstr 6 Medline, Google Scholar |
| 34. | A Amato, M Pescatori: Perioperative blood transfusions for the recurrence of colorectal cancer Cochrane Database Syst Rev 1: CD005033,2006 Medline, Google Scholar |
| 35. | JA Heit, MD Silverstein, DN Mohr , etal: Risk factors for deep vein thrombosis and pulmonary embolism: A population-based case-control study Arch Intern Med 160: 809– 815,2000 Crossref, Medline, Google Scholar |
| 36. | JW Blom, CJ Doggen, S Osanto , etal: Malignancies, prothrombotic mutations, and the risk of venous thrombosis JAMA 293: 715– 722,2005 Crossref, Medline, Google Scholar |
| 37. | GH Lyman, J Glaspy: Are there clinical benefits with early erythropoietic intervention for chemotherapy-induced anemia? A systematic review Cancer 106: 223– 233,2006 Crossref, Medline, Google Scholar |
| 38. | LT Goodnough: Risks of blood transfusion Crit Care Med 31: S678– S686,2003 Crossref, Medline, Google Scholar |
| 39. | C Madjdpour, DR Spahn: Allogeneic red blood cell transfusions: Efficacy, risks, alternatives and indications Br J Anaesth 95: 33– 42,2005 Crossref, Medline, Google Scholar |
| 40. | BD Spiess: Red cell transfusions and guidelines: A work in progress Hematol Oncol Clin North Am 21: 185– 200,2007 Crossref, Medline, Google Scholar |
Acknowledgment
We thank Dianne Tomita, MPH, for her contribution to the analysis and Kathryn Boorer, PhD, for assistance with writing this manuscript.
Fig A1. Incidence and risk of transfusions by hemoglobin (Hb) concentration at initiation of treatment. Data include transfusions occurring between week 5 and the end of treatment period. Baseline hemoglobin values were missing for 60 patients: incidence of transfusions, 33% (95% CI, 17% to 54%) for darbepoetin alfa versus 45% (95% CI, 28% to 84%) for placebo; hazard ratio (HR), 0.59 (95% CI, 0.25 to 1.43).
