Despite frequent anemia and multiple transfusions in patients undergoing chemotherapy and allogeneic hematopoietic stem cell transplantation (allo-HSCT) for acute myeloid leukemia , recommendations for use of erythropoiesis-stimulating agents (ESAs) in these populations are still missing. The primary objective was the effect of ESA administration on patient's quality of life (QoL). Secondary objectives were hemoglobin (Hb) recovery, red blood cell (RBC) transfusions, overall survival, and event-free survival.
Adult patients with Hb ≤ 11 g/dL after consolidation chemotherapy for acute myeloid leukemia (group 1), or after allo-HSCT for any hematological diseases (group 2), were prospectively included. ESA was administered subcutaneously once per week during a maximum period of 6 months and was stopped when Hb level reached 12 g/dL. A paired-matched analysis using a historical control group was performed for secondary endpoints. Fifty-two patients were included in group 1, and 55 patients were in group 2.
For the global population, a significant improvement of QoL was noticed with ESA use; 83% (group 1) and 71% (group 2) of patients achieved an Hb level ≥ 12 g/dL without transfusion requirement. The pair-matched analysis showed a reduction of 4 RBC units per patient in group 1 (P = .0002) and 3 RBC units per patient in group 2 (P = .04). No significant difference in terms of thromboembolic events, overall survival, and event-free survival was observed between ESA and control groups. A RBC transfusion median savings of €1712 per patient was estimated in each group.
Anemia, defined as a deficiency in red blood cells (RBCs), is the most common hematological abnormality among cancer patients1, 2; it is associated with poor prognosis and outcomes,3, 4 with a detrimental impact on patient condition and quality of life (QoL).5, 6 The usual method used to treat anemia is RBC transfusions, which have only a short-lived therapeutic effect and are associated with 1) short-term life-threatening risks, such as exposure to infectious agents7 and the more common risk of transfusion-related acute lung injury,8 and 2) long-term life-threatening risks, especially iron overload.9, 10 Erythropoiesis-stimulating agents (ESAs) are currently the only therapeutic alternative to transfusions. The use of ESA remains very common in patients treated by chemotherapy for solid tumors11–13 and lymphoid malignancies.14, 15 Recently, the American Society of Hematology (ASH) and the American Society of Clinical Oncology (ASCO) published an updated guideline on the use of ESA in adult patients with cancer16 that recommends a strict benefit/risk evaluation by clinicians for its use compared with RBC transfusions. Many clinical trials and meta-analyses have demonstrated that ESAs are generally well-tolerated in patients with chemotherapy-induced anemia.11, 17–24 At the same time, results from other clinical trials have shown a negative impact on survival and/or disease control.25–29 There are also safety concerns with ESAs, including the established increased risk of venous thromboembolic events.30, 31 In patients receiving chemotherapy for acute myeloid leukemia (AML) or undergoing allogeneic hematopoietic stem cell transplantation (allo-HSCT), the use of ESAs has not been clearly assessed and is still controversial, which explains the need for prospective evaluation of ESAs in these 2 populations.
In our current prospective study, as primary objective, we evaluated the effect of ESA administration in anemic patients after either consolidation therapy for AML in complete remission (CR) or after allo-HSCT for hematological diseases, on patient QoL. Secondary objectives evaluated the impact of ESA administration on hemoglobin (Hb) recovery, the number of RBC transfusion units, and overall and event-free survival (OS and EFS). All secondary endpoints were achieved using a paired matched analysis. Total direct costs regarding transfusions and ESA administration were calculated.
MATERIALS AND METHODS
This prospective study included adult patients with Hb level ≤ 11 g/dL after consolidation chemotherapy (first, second, or third) for AML in CR (group 1); or after allo-HSCT for any hematological disease (group 2); or with no ESA therapy received within 3 weeks before study initiation (control). Patients with pregnancy, lactation, a history of thromboembolic events, uncontrolled hypertension, and/or concomitant cardiac, neurological, digestive, or pulmonary disease that would compromise the study outcomes were excluded from this study. Written informed consent was obtained for all participants.
Patients in both groups were matched and compared to a historical population followed and treated at the same hospital with the only difference of not receiving ESA. The matching analysis takes into account sex, age, and disease status for the 2 groups; associated with cytogenetics, type, and number of chemotherapy for group 1; and diagnosis, conditioning, HSC source, related or unrelated donor, human leukocyte antigen matching, number of previous transplants, and graft-versus-host disease for group 2.
Study Design and ESA Administration
This study was undertaken between April 2006 and December 2009 at the Hematology Department of Edouard Herriot Hospital, Lyon, France. The design and study conduct complied with good clinical practice, in accordance with the Declaration of Helsinki and were approved by the French ethics committee. ESA was administered subcutaneously (sc) once per week during a maximum period of 6 months: for group 1, Aranesp (darbepoetin alfa injection) at 150 μg; for group 2, NeoRecormon (epoetin beta) at 30,000 IU. Hemoglobin level was monitored every week; injections were stopped once the Hb level reached 12 g/dL without any transfusion support. If after 4 injections, no improvement (increase of Hb < 1 g/dL or no decrease of RBC transfusion requirements) in Hb level was observed, doses were doubled, and if no improvement was observed after 8 injections, patient was taken off-study for ESA inefficiency. Transfusions were allowed as clinically indicated when the Hb level was ≤ 7 g/dL during hospitalization or Hb ≤ 8 g/dL outside hospitalization.
Quality of Life
The QoL was measured by patient completion of the Functional Assessment of Cancer Therapy-Anemia (Total FACT-An), which consists of the FACT-General (FACT-G) questionnaire, FACT-Fatigue (FACT-F), and FACT-Anemia (FACT-An) subscales.32 The QoL was measured at baseline, then patients were seen and evaluated at 1, 2, 3, and 6 months after study entry.
Hb Response Evaluation and Safety Assessment
Considering the Hb level, the CR (ESA CR) was defined as Hb level ≥ 12 g/dL; the Hb partial response (ESA PR) was defined as a ≥ 2 g/dL Hb increase, compared with the baseline value without any transfusion requirement. Adverse events and serious adverse events were recorded continuously throughout the study period, using Common Toxicity Criteria for Adverse Events, version 3, of the US National Cancer Institute.
Major resources used and direct medical costs that were related to transfusions and ESA administration were identified for each patient. Number of ESA injections (inside and outside hospitalization) and number of RBC transfusions were recorded. The ESA and RBC transfusion unit price were obtained from the accounting system of the Edouard Herriot Hospital; prices for ESA injections given outside hospitalization were calculated according to the national social insurance price. Total costs of ESA and RBC transfusions in the studied groups were compared to RBC transfusion costs in the matched populations.
The main aim of this study was to determine the patient's QoL at inclusion, and month 3 and month 6 after inclusion. The scores were calculated only if responses for more than 80% of the items were obtained. The evolution of the scores was analyzed with the nonparametric paired test of Friedman, and the comparison between each score was analyzed with a Wilcoxon Mann-Whitney test.
Secondary analyses were carried out using t tests to compare Hb levels and Wilcoxon Mann-Whitney test for transfusions number between studied groups and the matched population. OS and EFS were estimated by the Kaplan-Meier method with log-rank test for univariate analysis. The multivariate analysis studied the following variables: for group 1, age, sex, cytogenetics, type of chemotherapy (intensive vs standard), disease status, and baseline Hb levels; for group 2, age, sex, cytogenetics, type of disease, interval diagnosis-transplantation, disease status, cell source, sex-matching, kind of donor, human leukocyte antigen matching, ABO compatibility, conditioning, and Hb baseline levels. Statistical analysis was performed with R statistical software (version 2.9.2).
Between April 2006 and December 2009, 55 patients were included in group 1 and 61 patients were included in group 2. Patient characteristics for each group are summarized in Table 1. Patients with a follow-up of less than 2 months and who received fewer than 3 ESA injections were not considered eligible for evaluation. For the global population, there were 111 evaluable patients (52 [95%] in group 1 and 59 [97%] in group 2). Patients in group 1 were matched with 63 control patients, whereas patients in group 2 were matched with 65 control patients.
Table 1. Baseline and Clinical Characteristics of Evaluable Patients in Groups 1, 2, and Their Controls
Controls for Group 1
Controls for Group 2
Abbreviations: allo-HSCT, allogeneic hematopoietic stem cell transplantation; ALL, acute lymphoblastic leukemia; AML, acute myeloid leukemia; CLL, chronic lymphocytic leukemia; CR, complete response; EPO, erythropoietin; HLA, human leukocyte antigen; MM, multiple myeloma; NA, not applicable; ND, not determined; NK, not known; PBSC, peripheral blood stem cell; PR, partial response.
Sex (male/female), N (%)
30 (58)/22 (42)
35 (56)/28 (44)
42 (71)/17 (29)
47 (72)/18 (28)
Age, y (median)
Weight, kg (median)
Interval diagnosis–inclusion, mo (median)
Interval diagnosis–allo-HSCT, mo (median)
Cytogenetics, N (%)
Initial diagnosis, N (%)
18 (30)/15 (25)/5 (9)
20 (31)/15 (23)/7 (11)
5 (9)/16 (27)
7 (11)/16 (24)
Disease status, N (%)
CR 1/CR 2
50 (96)/2 (4)
60 (95)/3 (5)
32 (54)/13 (22)
37 (57)/9 (14)
9 (15)/5 (9)
10 (15)/9 (14)
Type of chemotherapy, N (%)
25 (48)/27 (52)
29 (46)/34 (54)
Sequential chemotherapy, cycles no. 1/2/3, N (%)
44 (85)/7 (13)/1 (2)
52 (82)/11 (18)/0
Conditioning, N (%)
39 (66)/20 (34)
48 (74)/17 (26)
Stem cell source, N (%)
19 (32)/40 (68)
19 (29)/46 (71)
Sex matching, N (%)
26 (44)/16 (27)
30 (46)/11 (17)
9 (15)/8 (14)
16 (25)/8 (12)
HLA matching, N (%)
33 (56)/16 (27)
54 (83)/8 (12)
ABO compatibility, N (%)
10 (17)/12 (20)
5 (8)/11 (17)
Hemoglobin, g/dL (median)
Hematocrit, % (median)
Platelets, ×109/L (median)
EPO, IU/L (median)
Ferritin, ng/mL (median)
Quality of Life
For the global population (group 1 and 2), a significant improvement of QoL during the 6 months follow-up according to FACT-An subscale (P = .001) and to the FACT-F subscale (P = .003) (Fig. 1) were observed. According to Total FACT-An and to FACT-G, the QoL remained stable during the 6-month follow-up period. However, a statistically significant increase (P = .04) was observed in the physical well-being.
The median number of ESA injections per patient was 13 (range, 3-24) in group 1 and 8 injections (range, 2-28) in group 2. There were 85 ESA CR (43 [83%] in group 1 and 42 [71%] in group 2) and 3 (6%) ESA PR (only in group 1). Among patients who reached the 6-month follow-up in the 2 groups, 85% had a normal Hb level. Twelve patients (11%) were withdrawn (4 in group 1 and 8 in group 2) due to ESA inefficacy. In group 1, the median time to achieve an ESA CR was 34 days (range, 17-67 days) after first consolidation chemotherapy and 41 days (range, 12-67 days) after second consolidation (P = .35) with a median number of ESA injections of 5 (range, 2-9) and 6 (range, 2-9), respectively. In group 2, the median time to achieve ESA CR was 39 days (range, 14-180 days) after a median number of 6 (range, 2-9) injections. When compared between ESA patients and controls for the 2 groups, mean Hb level was not statistically different at baseline (11.3 g/dL ± 19 vs 10.8 g/L ± 19, respectively; P = .19), whereas ESA patients had significantly higher Hb level at hospital discharge (10.4 g/dL ± 1.6 vs 9.3 g/dL ± 7, respectively; P = .0001) and at the 6-month visit (11.2 g/dL ± 2.0 vs 10.3 g/dL ± 16, respectively; P = .012).
Among evaluable patients, those under ESA exhibited a significantly lower RBC transfusion rate compared to controls. In group 1, there were 319 RBC units used in the ESA group, whereas 487 RBC units were recorded in the control group, which represents a median reduction of 3.9 RBC units per patient (P = .0002). In group 2, there were 227 RBC units used in the ESA group, whereas 355 RBC units were recorded in the control group, which represents a median reduction of 3.2 RBC units per patient (P = .04). In terms of platelet transfusions, we noted a significantly lower number of transfusions in ESA patients only in group 1 compared to controls (322 units vs 396 units, P = .029), which represents a median reduction of 1.7 platelet units per patient. No significant difference was found between those receiving ESAs and controls in group 2.
The ESA injections were well-tolerated in both groups. Adverse events possibly related to ESA were 4 in group 1 (2 thrombotic events, 1 arthralgia, and 1 cutaneous allergy) and 5 in group 2 (all events were thrombovascular). One ESA-related serious adverse event was reported in group 2: hospitalization (11 days) due to a deep vein thrombosis of lower limbs. Furthermore, incidence of thromboembolic events in the ESA population and in the matched population were statistically similar (group 1: 3.6% vs 3.2%, P = .88; group 2: 9.8% vs 3.3%, P = .34).
The economic study showed that the median costs of ESA injections in group 1 were €3904 per patient (range, €2732-€6058) and €1862 per patient (range, €955-€3104) in group 2. These costs included ESA injections administered inside and outside hospitalization; median ESA inside hospital costs were €1596 per patient (range, €976-€2328) and 5€518 per patient (range, €339-€907) in group 1 and 2, respectively; median ESA outside hospitalization costs were €2219 per patient (range, €1281-€3500) and €307 per patient (range, €0-€1832) in group 1 and 2, respectively.
When we compared the RBC transfusion-related costs between patients who received ESAs and their controls, there was a significantly lower cost in the ESA groups. For group 1: €2568 per patient (range, € 1712-€4280) for ESA patients vs €4280 per patient (range, €2568-€5992) for controls (P = .007); for group 2: €856 per patient (range, €761-€3424) for ESA patients vs €2568 per patient (range, €856-€5041) for controls (P = .04). Therefore, a RBC transfusion-related median reduction of €1712 per patient was estimated in each group.
The multivariate analysis studying different confounding factors on the cumulative incidence of ESA CR showed a significant positive impact of younger age (P = .001) and intensive chemotherapy (P = .03) in group 1, and the negative impact of female recipient (P = .036) and major ABO incompatibility (P = .018) for group 2.
Overall and Event-Free Survival
After a median follow-up of 22.5 months (range, 6-49 months) for patients in group 1 and 25.9 months (range, 6-107 months) for their controls, we did not find any significant difference either in terms of OS (P = .77) or in terms of EFS (P = .57) (Fig. 2A,B). Similarly, after a median follow-up of 12.6 months (range, 2-45 months) for patients in group 2 and 22 months (range, 1-145 months) for their controls, there was also no significant difference either in terms of OS (P = .69) or in terms of EFS (P = .79) (Fig. 2C,D).
Despite the recent ASH and ASCO recommendations for the use of ESAs in treatment of different oncology and hematology diseases,16 there is a lack of guidelines for its use in patients with myeloid malignancies and in patients undergoing allo-HSCT. In order to avoid any conflict of interest toward the use of a specific type of ESA, in our study, we decided to use the 2 available ESA molecules at that time: darbepoetin alfa and epoetin beta. There are no available powerful studies concerning the use of ESAs in myeloid malignancies and in patients undergoing allo-HSCT, and the studies that were present reported heterogeneous populations and small number of patients. In terms of Hb response, we observed ESA CR in 83% of patients (Hb level ≥ 12 g/dL) in group 1 and ESA CR in 71% of patients in group 2 with only 11% in total of ESA inefficiency in the 2 groups. In group 1, the median time to achieve an ESA CR was 34 days after first consolidation chemotherapy and 41 days after second consolidation. In group 2, the median time to achieve ESA CR was 39 days. When compared between ESA patients and controls, ESA patients had significantly higher Hb level at hospital discharge. Our results confirm previous results from studies evaluating the ESA response on Hb level in patients with solid tumors or with lymphoproliferative disorders and in patients undergoing chemotherapy or allogeneic HSCT showing a clear significant improvement in Hb level.33–36
Transfusion avoidance is generally the rationale for ESA administration in patients with cancer. The use of ESAs when Hb level is between 10 and 12 g/dL has been considered on the basis of the hypothesis that it may improve QoL in patients with cancer. Several clinical trials37–39 and one meta-analysis40 have included QoL as an endpoint for patients with chemotherapy-induced anemia randomly assigned to treatment with ESAs or placebo; these studies show that the use of ESAs has led to a statistically significant increase in QoL. Nevertheless, patients with AML or who were undergoing allo-HSCT were underreported in these studies. In our study, there was an improvement of QoL at 1, 2, 3, and 6 months follow-up with a significant positive effect at 6 months according to the FACT-An subscale (P = .001) and to the FACT-F subscale (P = .003). This improvement can appear with the natural history of the disease either for AML patients or those undergoing allo-HSCT, but anemia frequently persists during this follow-up with an important impact on patient performance status if no ESAs were given. According to the recent ASH/ASCO guidelines on the use of ESAs,16 assessment of QoL remains challenging.
Concerns regarding a potential relationship between ESAs and increased risk of mortality rate have been raised in many clinical trials, with controversial results.25–29 A study by Aapro et al evaluated the impact of treatment with epoetin-beta on survival in patients with cancer by pooling results of 12 randomized controlled studies.41 They did not find any significant effect on mortality and disease progression after use of ESAs compared with the control group. On the other hand, Bohlius et al evaluated the survival in different patient populations from 53 randomized studies,30 and they concluded that use of ESAs significantly increases the on-study mortality rate in the global population, whereas a nonsignificant mortality rate increase was observed in patients receiving chemotherapy and ESAs. In this meta-analysis, the authors noted evidence for interaction between ESA administration, Hb levels at baseline, target Hb level and mortality. In addition, when results were limited to patients with hematological malignancies, the mortality rate was not significantly affected by ESAs use; the authors also pointed out that one study27 (in breast cancers) has contributed 20% weight to the overall meta-analysis, and its exclusion reduced the overall hazard ratio. In our study, we did not find any significant difference between ESA patients in group 1 and their controls, either in terms of OS or in terms of EFS. Similarly, for patients in group 2 and their controls, there was no significant difference either in terms of OS or in terms of EFS.
When analyzing the number of RBC transfusions between ESA patients and controls, we noticed a significant benefit of ESA use leading to a median reduction of 3.9 RBC units per patient in group 1 (35% reduction) and 3.2 RBC units per patient in group 2 (36% reduction). Similarly, a meta-analysis by Lyman et al showed approximately 50% reductions of RBC transfusions when ESAs were introduced.42 In 2006, the Agency for Healthcare Research and Quality published a comparative effectiveness review of ESAs.43 They found that the majority of trials reported fewer transfusions among patients randomly assigned to ESA treatment compared with patients assigned to controls.
Many studies have warned about the thromboembolic events related to the use of ESAs,31, 40, 44 whereas other trials showed similar events between ESA patients and controls.45 In our study, thrombovascular events were not statistically different between groups 1 and 2 and their controls.
Finally, we have shown significantly fewer transfusions in the 2 groups, a reduction associated with an economic benefit of €1712 per patient. This economic benefit has to be added to the important QoL improvement and decrease of RBC transfusions with their side effects.
In conclusion, this prospective study shows a clinical and economic benefit of ESA administration, an achievement of a normal Hb level, and a significant reduction of number of RBC transfusions as well as toxicity related to RBC transfusions. ESA administration could improve a patient's QoL with no harmful effect on survival.