The Oncologist, Vol. 14, No. suppl_1, 6-15, September 2009; doi:10.1634/theoncologist.2009-S1-6 © 2009 AlphaMed Press
Update on Erythropoiesis-Stimulating Agents and Clinical Trials in OncologyaMultidisciplinary Oncology Institute, Genolier, Switzerland; bJohns Hopkins University, School of Medicine, Baltimore, Maryland, USA Key Words. Anemia • Erythropoiesis-stimulating agents • Guidelines Correspondence: Matti S. Aapro, M.D., IMO Clinique de Genolier, 1 Route du Muids, 1272 Genolier, Switzerland. Telephone: 41223669106; Fax: 41223669131; e-mail: maapro{at}genolier.net Received February 23, 2009; accepted for publication June 24, 2009.
Disclosures: Matti Aapro: Consultant/advisory role: ESMO, SIOG, ESO, EuroCancer, EORTC, IUCC; Honoraria: Amgen, Johnson & Johnson, Roche, Sanofi-Aventis, Bayer Schering, Merck-Serono, Merck, Helsinn, Novartis, Pfizer, Pierre Fabre; Research funding/contracted research: Amgen, Roche, Sanofi-Aventis, Merck-Serono, Merck, Helsinn, Cephalon, Pierre Fabre; Jerry L. Spivak: Consultant/advisory role: Amgen, Ortho Biotech, Roche.
Anemia commonly occurs among cancer patients receiving chemotherapy. In these patients, erythropoiesis-stimulating agents (ESAs) are effective in managing anemia but there is an increased risk for thrombovascular events. In more recent randomized clinical trials, there have been differing results regarding the impact of ESAs on overall survival and mortality. The balance between studies that show higher ESA-associated mortality and those that don't show ESA-associated mortality is examined in this review. This review discusses where we stand today on anemia management in cancer patients. Preliminary results from a recent independent patient data meta-analysis for on-study deaths and overall survival in patients receiving chemotherapy (the only oncology population for which ESA treatment is currently indicated) showed no statistically significant difference between the ESA and control groups (on-study deaths hazard ratio [HR], 1.10; 95% confidence interval [CI], 0.98–1.24; overall survival HR, 1.04; 95% CI, 0.97–1.11, compared with controls). Possible factors that could influence study results are discussed in this review. There are no convincing data to support ESA-induced tumor stimulation in patients. ESAs decrease RBC transfusion needs and sustain targeted hemoglobin levels, and this ESA response does not significantly impact overall survival or mortality when ESAs are used within guidelines and labeling. However, based on the currently available data and meta-analysis, the use of ESAs has to be carefully balanced against any possible risk for higher mortality.
Anemia in patients with cancer is multifactorial and frequent amongst cancer patients receiving chemotherapy [1]. Use of erythropoiesis-stimulating agents (ESAs) decreases the proportion of chemotherapy patients requiring RBC transfusions [2–4]. However, ESAs increase the risk for thrombovascular events in chemotherapy-induced anemia (CIA), chronic renal failure, and surgery. Data from >12,000 patients in controlled studies are available to provide information on the benefit–risk ratio. Outside the approved indication for the use of ESAs (i.e., ESAs are indicated only for the treatment of anemia resulting from concomitant myelosuppressive chemotherapy, ESAs are not indicated for patients receiving myelosuppressive therapy when the anticipated outcome is cure, and ESAs must be discontinued following the completion of a chemotherapy course), there is a higher risk for mortality; however, this is not the case if used within indication. The most rigorous study to date (20010145-small cell lung cancer [SCLC]), which was conducted to the most exacting standards, did not demonstrate greater mortality or tumor progression [5]. Furthermore, a relationship between epoetin receptor expression and tumor proliferation is unclear [6–9]. The need to assess the risks and benefits of ESAs has led to the convening of numerous regulatory meetings, including the Oncology Drug Advisory Committee (ODAC) of the U.S. Food and Drug Administration (FDA) 2004 and 2007 Advisory Board meetings, the 2007 American Society of Hematology/American Society of Clinical Oncology and European Organization for Research and Treatment of Cancer (EORTC) guideline update, the 2007 October Committee for Medicinal Products for Human Use (CHMP) safety committee, the 2008 CHMP and European Medicines Agency (EMEA): ESA label changes and harmonization, the 2008 March ODAC Advisory Board meeting, the 2008 May 15 Scientific Advisory Groups/EMEA hearing, and the Summer 2008 regulatory decisions. Based on the updated data, the regulatory authorities have narrowed the indications for the use of ESAs to improve benefits and minimize risks.
Meta-analyses are helpful to detect a signal of a rare event, and although they are not a substitute for properly conducted prospective studies, they have allowed progress in various areas of oncology. They can help in assessing data and indicate areas for further study of ESA treatment of anemic cancer patients on chemotherapy.
Survival and Safety In the meta-analysis of 12 randomized, double-blinded, placebo-controlled studies of both on-label and off-label epoetin alfa, no specific signals for survival outside the off-label use studies were observed [3, 10–12]. In this meta-analysis, only one study (the Breast Cancer Erythropoietin Survival Trial [BEST]) produced an HR that significantly favored placebo. In these 12 studies of anemia, the HR for on-study mortality was 1.13 (95% confidence interval [CI], 0.95–1.34), but this included studies in patients not receiving chemotherapy (Study H87-032, H87-014, and H87-015) and in which treatment went beyond the correction of anemia (BEST, N93-004). In the initial independent analysis published by Bohlius et al. [3], overall, the HR in 42 trials and 8,167 patients was 1.08 (95% CI, 0.99–1.18) in disfavor of ESA treatment use. However, this included cancer patients with or without concurrent chemotherapy [3]. In a further analysis presented by Amgen to the U.S. FDA, divided by on-label and off-label use, the trend is toward favoring the control for the off-label use whereas the on-label use favors the ESA [11, 13]. The fixed effects and random effects models in this particular meta-analysis gave a Peto odds ratio of 1.03 (95% CI, 0.93–1.15) [11, 13]. This meta-analysis involved 59 controlled ESA studies, with 15,249 patients, and divided the patients into three groups: anemia of cancer, radiation therapy setting (RTx), and CIA [11, 13]. In the RTx setting, there was a negative signal related to the head and neck (H&N) setting in studies of mainly nonanemic patients using radiation therapy without chemotherapy [Erythropoietin in Head and Neck Cancer Study (ENHANCE) and Danish Head and Neck Cancer Group Study (DAHANCA)] [11, 13]. These concerns are now well known and have led radiation oncologists to avoid epoetin use in the setting of standard radiation therapy without chemotherapy.
In the CIA data, the most recent studies regarded by authorities as favoring the control are the Gynecologic Oncology Group (GOG)-0191 study in cervical cancer and the PREPARE study in neoadjuvant breast cancer [14, 15]. GOG-0191 and PREPARE explored ESA use in nonanemic patients and/or a high hemoglobin (Hb) target (>13 g/dl), and in our opinion cannot be viewed as valid arguments against the use of ESAs. The GOG-0191 study aimed at maintaining an Hb level Overall data from eight of the 59 controlled phase III studies in oncology raised concern about the use of ESAs in cancer patients (Fig. 1) [11, 13, 14–17]. The 59 studies are divided as follows: (a) 46 of these studies included cancer patients receiving chemotherapy, (b) nine of these studies included cancer patients with CIA, and (c) four studies included cancer patients receiving radiotherapy.
The GOG-0191 and PREPARE studies were reported after the May 2007 ODAC meeting and led to the additional ODAC meeting held in March 2008 [14–16]. These resulted in the following recommendations of the ODAC 2008 panel to the FDA: to allow ESAs to continue to be marketed for CIA but recommend ESAs to be used with restriction for curable cancer, breast and H&N cancer. The signals in the GOG-0191 and PREPARE studies are inconclusive and not confirmed by well-controlled studies in similar settings (Table 1) [14, 15, 18, 19]. The other large cervical cancer study, conducted by Blohmer et al. [18], involving 250 patients, did not give a negative signal. In fact, in the Blohmer et al. [18] study, the relapse rate was in favor of the ESA group, although the difference was not statistically significant [18]. In the breast cancer setting, Möbus et al. [19] studied the effect of ESAs in the adjuvant setting and this study gave no negative signal whatsoever. Regarding these studies, we need to try to understand what went wrong, what went right, and why. These studies are discussed in more detail further on in this article.
Rigorous evaluation has not revealed greater tumor progression in chemotherapy studies. As mentioned, other studies, such as the 20010145-SCLC study, which studied tumor progression with central review (and was conducted in a homogeneous patient population), did not show greater tumor progression with ESA use. Importantly, in the PREPARE study, specified pathologic complete response was an endpoint, and there was no difference in the response rate in tumors with ESA use compared with controls (no ESA use). Thus, a negative signal might not be related to tumor progression as initially suggested by the Leyland-Jones [20] study. A later publication by Leyland-Jones et al. [21] found that tumor response and time to disease progression were similar between the epoetin and placebo groups.
A second meta-analysis on epoetin beta including recent studies that had not been included in the initial meta-analysis was recently published [22]. The newer studies included MF4449 (Henke et al. [23]), MO16375 (Management of Anemia Under RadioChemotherapy), and BA16756 (Breast Cancer-Anemia and the Value of Erythropoietin [BRAVE]) [22–24]. This second meta-analysis of epoetin beta included a total of 12 randomized, controlled studies and 2,297 patients (epoetin beta, n = 1,244; control, n = 1,053) [22]. This analysis evaluated the impact of epoetin beta on overall survival, tumor progression, and thromboembolic events, and a prespecified subgroup analysis also assessed effects in patients with a baseline Hb level
There are numerous other meta-analyses available assessing the overall survival of patients receiving ESAs. Reassuringly, the estimated HR for overall survival in the chemotherapy studies is around 1.02 or 1.03 (Table 3) [3, 4, 16, 17, 25–28].
Breast Cancer Update The BEST study showed higher early (4 months) mortality with epoetin, with higher thrombosis and disease progression mortality but, overall, a similar duration of progression-free survival [21, 24]. The Kaplan–Meier estimate of long-term survival in the BEST study showed a convergence of the survival lines at about 19 months after randomization [21]. A possible reason for the impaired survival seen with the ESA group in the BEST study in the first year of study (particularly in the first 4 months after randomization) may not have been a result of tumor progression, but possibly thromboembolic events. Chang et al. [29] conducted a study in 354 breast cancer patients receiving chemotherapy, comparing once-weekly epoetin alfa with standard care. Contrary to the higher early mortality seen in the BEST study, Chang et al. [29] did not observe a shorter survival time in the ESA group of patients. In the Chang et al. [29] study, the epoetin arm showed a similar 2-year survival rate and no difference in thromboses. The BRAVE study, in 463 metastatic breast cancer patients receiving chemotherapy (anthracycline and/or taxane based), compared once-weekly epoetin beta with control (transfusions allowed as clinically indicated in both arms) [24]. It found no difference in overall or progression-free survival (HR, 1.07; 95% CI, 0.89–1.30; p = .448), although there were more thromboses in the epoetin arm (but the incidence of serious thromboembolic events was 4% in the epoetin arm and 3% in the control arm) [24]. There was no difference in overall survival (HR, 1.07; 95% CI, 0.87–1.33; p = .522). The data shown in Table 1 from the PREPARE and Möbus et al. [19] studies are the latest data as of November 2008. These studies showed that there was no statistically significant difference in overall survival and progression-free survival when the Hb level was targeted at 12.5–13 g/dl [16, 17, 19]. These findings are in agreement with the Del Mastro et al. [30] study, which stressed the importance of preventing anemia and controlling Hb levels with epoetin in dose-dense CIA patients. In the Möbus et al. [19] study, breast cancer patients received adjuvant treatment with either a standard regimen (four cycles of epirubicin and cyclophosphamide followed by four cycles of paclitaxel) or a dose-dense regimen consisting of three courses each of epirubicin (150 mg/m2), paclitaxel (225 mg/m2), and cyclophosphamide (2,500 mg/m2) at 2-week intervals (the ETC regimen) with growth factor support [19]. Patients in the dose-dense ETC arm then underwent a second randomization and received either epoetin alfa or transfusion support (Fig. 2). The target Hb level was 12.5–13 g/dl. In that study, 1,255 patients were evaluable and 658 patients were randomized in the dose-dense ETC arm (333 received epoetin alfa), with a median follow-up of 62 months [19]. The ESA arm maintained a higher Hb level and anemia occurred significantly more often in the ETC-alone arm than in the ETC–ESA arm (p < .0001) [19]. There were significantly more transfusions in the ETC-alone arm than in the ETC–ESA arm (28% versus 13%, respectively; p < .0001) [19]. There was no significant difference in the 5-year overall survival rate (83% for the observation arm versus 81% for the epoetin alfa arm; p = .89) and in the disease-free survival rate (71% in the observation arm versus 73% in the epoetin alfa arm; p = .86) between the ETC-alone arm and the ETC–ESA arm [19]. There was a higher incidence of thromboembolic events in the ETC–ESA arm (3.0% versus 1.7%).
The 2008 update of the Möbus et al. [19] study also supports these findings, and there was no statistically significant difference in the overall survival (p = .9388) and recurrence-free survival (p = .9709) durations in the ESA arm compared with the non-ESA arm [16, 17]. In that study, the investigators found that the dose-dense adjuvant ETC regimen resulted in significantly longer overall and disease-free survival times than in the standard arm, but it was associated with relevant hematological toxicity [19]. The epoetin alfa group had a significantly lower number of RBC tranfusions, greater median Hb value, and less anemia. At the same time, the addition of epoetin alfa to the ETC dose-dense adjuvant treatment had no influence on overall and disease-free survival [19]. A phase III prospective randomized trial compared adjuvant chemotherapy (docetaxel, doxorubicin, and cyclophosphamide or 5-fluorouracil, epirubicin, and cyclophosphamide) plus darbepoetin (ARA+) with adjuvant chemotherapy without darbepoetin (ARA–) in node-positive breast cancer patients and was recently reported in preliminary form [31]. In 1,234 patients, in total, 279 serious adverse events were reported: 150 in the ARA+ group versus 129 in the ARA– group [31]. In the ARA+ group of patients there was a modestly higher rate of thrombosis and cardiovascular events (ARA+, 24; ARA–, 13) [31]. There was significantly more anemia in ARA– patients and their Hb levels decreased during therapy. In ARA+ patients, Hb levels were stable over the whole treatment period and overstimulation was rare. In that study, darbepoetin was initiated at an Hb level <13 g/dl (which was a later amendment to an Hb level <12g/dl) and was stopped at an Hb level of 14 g/dl [31]. The investigators found that the short-term toxicity and safety results were reassuring in this study [31]. Thus, contrary to the results of the BEST study, subsequent studies have not demonstrated shorter survival in breast cancer patients treated with recombinant erythropoietin.
Cervical Cancer Update The design of the Blohmer et al. [14] study is shown in Figure 3. In that study, cervical cancer patients were randomized after surgery to either the epoetin arm (chemotherapy followed by radiotherapy with epoetin) or the control arm (chemotherapy followed by radiotherapy with transfusion support) [18]. The relapse rates, at a median of 105 weeks, were not statistically significantly different between the epoetin and control arms (17% versus 25%; p = .074), neither was the 2-year relapse-free survival rate (70% versus 81%; p = .058) [18]. The mean Hb level achieved in the epoetin arm was >12 g/dl, whereas that of the control arm was <12 g/dl. The investigators concluded that the addition of epoetin alfa to sequential chemoradiotherapy treatment resulted in a significantly lower incidence of grade 2 anemia (p = .0001) and significantly lower transfusion requirements (p = .0004) [18]. The mean Hb level was maintained at >12 g/dl throughout sequential chemoradiotherapy treatment in the epoetin alfa group, but not in the control group [18]. The addition of epoetin alfa treatment to sequential adjuvant chemoradiotherapy treatment resulted in a trend toward a lower rate of disease recurrences [18].
Therefore, contrary to initial impressions, recombinant erythropoietin therapy does not result in disease progression or shorter survival in anemic cervical cancer patients.
Update in Lymphomas The German Hodgkin's Study Group HD15-EPO Trial
Clearly, in Hodgkin's lymphoma, anemia is a risk factor, and data from various clinical trials have suggested that treatment with ESAs produces a better outcome in anemic patients. The German Hodgkin's Study Group investigated the use of ESAs in patients with advanced-stage Hodgkin's lymphoma in their HD15-EPO trial [34]. One of the endpoints of that study included a comparison of the safety of the epoetin alfa arm versus the placebo arm. The Hb target level was 12–13 g/dl. This is the biggest prospective randomized, placebo-controlled trial with epoetin alfa available in Hodgkin's lymphoma, with 1,379 patients randomized and 1,329 patients eligible for the epoetin alfa assessment [34]. The major endpoint was fatigue or quality of life. Secondary endpoints included overall survival, freedom from treatment failure (FFTF), progression-free survival, thromboembolism, adverse drug reactions, Hb cycle, and transfusion requirements. The final analysis of the study indicated that there was no statistically significant difference in terms of fatigue, response rate, FFTF, overall survival, tumor-related mortality, or side effects between those advanced stage Hodgkin's lymphoma patients receiving epoetin alfa and those receiving placebo [34]. There was no difference in the risk for thrombosis between the two groups. The number of RBC units given was significantly lower in the epoetin alfa group (the median number of transfusions for the epoetin alfa group was half that for the placebo group) [34]. The authors concluded that epoetin alfa was safe and efficacious in advanced stage Hodgkin's lymphoma. Furthermore, anemia and RBC transfusions are a relevant concern in this disease.
Cochrane Systematic Reviews on ESAs
Once fully published, the IPD Cochrane meta-analysis will provide the most comprehensive and informative meta-analysis to date [35]. This IPD meta-analysis in cancer patients receiving chemotherapy, the cancer indication for which ESAs are approved, was based on 38 studies with 10,441 patients. Importantly, none of these studies used ESAs according to current label guidance (i.e., using the lowest ESA dose needed to avoid RBC transfusion, and using ESAs only for treatment of anemia resulting from concomitant myelosuppressive chemotherapy, etc.) [35, 36]. The question as to why a few of the more recent studies suggest that ESAs have a negative impact on overall survival in cancer patients remains open. A higher risk for thrombovascular events with ESAs is probably the most likely explanation. There are no convincing data to support epoetin alfa–induced tumor stimulation in patients. The complete IPD meta-analysis with nearly 14,000 patients might help to clarify this open question. In addition to off-label use, other possible reasons for the poorer results in some of the studies after 2002 may include: (A) changing to a more dose-dense ESA schedule in some clinical trials (i.e., the shifting of the ESA dose from three times per week to once weekly) and (B) moving the Hb target level from a correction of anemia to "beyond correcting" anemia. These two possible issues (ESA dosing schedule and Hb target level) are currently being addressed in oncology clinical trials. The 2007/2008 position of the EORTC Working Party regarding ESAs has not changed since their first publication in 2004: Clinical trials have established that ESAs decrease transfusion needs and sustain Hb levels, without the need for intermittent transfusions; ESAs increase quality of life in symptomatic patients; and ESAs should be used within guidelines. The EORTC will wait for the updated data from the patient-level Cochrane meta-analysis before deciding if anything should be changed in these guidelines. This EORTC decision is expected in September 2009. In conclusion, ESA therapy is effective in producing a hematopoietic response, and ESA response has been shown in a wide variety of tumor types, including solid and nonmyeloid hematologic malignancies. ESAs have a favorable risk–benefit ratio when used for labeled indications. This is further supported by the Paladini et al. [37] meta-analysis (17 studies with 3,788 patients), which assessed the safety of ESAs when used according to label indications, in patients with CIA (instead of cancer-induced anemia) with an Hb level <11g/dl. In that meta-analysis, there was no difference in the mortality rate associated with the use of ESAs (relative risk, 0.95; 95% CI, 0.88–1.03; p = .22), and the authors concluded that, when ESAs are used as indicated on the label, they are not associated with a higher mortality risk [37]. However, caution does need to be exercised with the use of ESAs, taking into account the patient's comorbidities and carefully balancing the benefits against any possible risk for higher mortality.
Conception/Design: Matti Aapro, Jerry L. Spivak Collection/assembly of data: Matti Aapro, Jerry L. Spivak Data analysis: Matti Aapro, Jerry L. Spivak Manuscript writing: Matti Aapro, Jerry L. Spivak Final approval of manuscript: Matti Aapro The authors take full responsibility for the content of this article and thank Rob Stepney, medical writer, and Julia O'Regan, Bingham Mayne and Smith, Edinburgh, supported by an educational grant from Ortho Biotech, a division of Janssen-Cilag Europe, for their assistance in preparing a first draft of the manuscript based on an oral presentation at a meeting held on November 20, 2008 in Sitges, Spain, organized by a Scientific Committee of Matti Aapro, Mario Dicato, Pere Gascón, Francesco Locatelli, Jerry Spivak, and Jay Wish.
The authors acknowledge and thank Andreas Engert (Cochrane Haematological Malignancies Group, Cologne, Germany) for his comments on this manuscript.
This article has been cited by other articles:
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||