Systematic Review of Controlled Trials on Erythropoietin to Support Evidence-Based Guidelines

ABSTRACT: To support evidence-based clinical guidelines on erythropoietin use for anemia in oncology, we conducted systematic reviews of controlled trials on four patient groups. These were patients with treatment-related anemia; patients with disease-related anemia; patients transplanted with allogeneic hematopoietic stem cells; and those transplanted with autologous hematopoietic stem cells. Two reviewers followed a prospective protocol to select studies, abstract relevant outcomes, evaluate study quality, and conduct meta-analysis where data sufficed. For treatment-related anemia, meta-analysis of available evidence (22 trials; N = 1,927) demonstrated reduced odds of transfusion after erythropoietin, but higher-quality trials reported smaller odds reductions. In several trials, erythropoietin improved quality of life for groups with mean baseline hemoglobin In 1997, the Agency forHealthCare Policy and Research, nowknown as the Agency for Healthcare Research and Quality (AHRQ), launched aninitiative to promote evidence-based practice in everyday care by establishing12 Evidence-based Practice Centers. AHRQ establishes contracts with theEvidence-based Practice Centers to develop evidence reports and technologyassessments on clinical topics that are common, expensive, and/or significantfor the Medicare and Medicaid populations. Through this program, AHRQ partnerswith private and public organizations to improve the quality, effectiveness, andappropriateness of clinical care by facilitating the translation of researchevidence into clinical practice.

Approach to Systematic Reviews

Program Overview

In 1998, the American Society of Hematology (ASH) and theAmerican Society of Clinical Oncology (ASCO) jointly nominated the topic"Uses of Erythropoietin in Oncology" to AHRQ for an Evidence-basedPractice Center systematic review and evidence report. A panel appointed jointlyby ASH and ASCO intended to use the resulting evidence report to supportdevelopment of a clinical guideline for dissemination to their members. AHRQcontracted with the Blue Cross and Blue Shield Association (BCBSA) TechnologyEvaluation Center Evidence-based Practice Center to conduct the systematicreview and evidence report on erythropoietin.

Evidence-based Practice Centers develop evidence reports andtechnology assessments based on rigorous, comprehensive syntheses and analysesof relevant scientific literature, emphasizing explicit and detaileddocumentation of methods, rationale, and assumptions. These scientific synthesesmay include meta-analyses and cost analyses. Each Evidence-based Practice Centercollaborates with other medical and research organizations so that a broad rangeof experts is included in the development process. More detailed information onthe Evidence-based Practice Centers program, the topic nomination process, andthe list of centers is available at http://www.ahrq.gov/clinic/epcix/htm.[1]Executive summaries and full copies of completed reports (with bibliographiesand evidence tables) are available at the same URL for viewing or complimentarydownloading. Complimentary single, printed copies also may be obtained from theAHRQ Publications Clearinghouse (1-800-358-9295).

Systematic Review Methods

Protocols for systematic review are prospectively designed todefine study objectives and key questions; search strategy; patient populationsof interest; study selection criteria and methods to determine studyeligibility; outcomes of interest; data elements to be abstracted andabstraction methods; and methodsto assess study quality. Usually, two independent reviewers completeeach step of the protocol. Reviewers individually evaluate studies againstselection criteria, abstract data separately, and compare their results aftereach step. Disagreements are generally resolved by consensusbut may require resolution by a third reviewer.

A technical advisory group provides ongoing guidance on allphases of each Evidence-based Practice Centers’ review. Six technical advisorsparticipated in the evidence report on use of erythropoietin in oncologypatients. ASCO and ASH each appointed two of the six advisors (including theguideline panel co-chairs) and the Technology Evaluation Center Evidence-basedPractice Center staff recruited the remaining two.

Evidence-based Practice Centers reviews begin with acomprehensive literature search that attempts to identify all publications ofrelevant controlled trials. The search strategy for the review on erythropoietinis described briefly in the Executive Summary posted on the AHRQ web site[1] andmore completely in the full evidence report.[2] The Medline, Cancerlit, andEmbase databases, last searched in December 1998, yielded 2,915 references. Weidentified 28 additional reports by supplementary searches (eg, CurrentContents, bibliographies from manufacturers) through October 30, 1999, for atotal yield of 2,943 references.

Next, studies are selected for data abstraction usingcriteria specified in the protocol. The primary study selection criteria for theerythropoietin review required that studies be designed as controlled trialscomparing the outcomes of managing anemia with and without erythropoietin in apatient population relevant to one of four clinical settings. These were (1)anemia due primarily to cancer therapy; (2) anemia due primarily to amalignancy; (3) high-dose myeloablative therapy followed by an allogeneictransplant of hematopoietic stem cells from peripheral blood or bone marrow; and(4) high-dose myeloablative therapy followed by an autologous transplant ofhematopoietic stem cells. We defined the setting as anemia primarily due tocancer therapy if trials limited enrollment to patients undergoing concurrentchemotherapy or radiation therapy with conventional nonmyeloablative doses. Wedefined the setting as anemia primarily due to malignancy if some enrolledpatients did not receive concurrent chemotherapy or radiation therapy while onstudy.[2] Trials were excluded if there were < 10 similarly treated evaluablepatients in each arm.

In the available trials, erythropoietin treatment (withtransfusion used as necessary) was always compared with red blood cell (RBC)transfusion alone; no trials compared erythropoietin to any other alternative.All randomized controlled trials relevant to any of the four clinical settingswere included. Studies that used nonrandomized concurrent or historical controlswere included if the reviewers could determine that similar patients wereincluded in the treatment and control groups. Nonrandomized trials wereconsidered to be of lesser quality than randomized controlled trials. Outcomesof interest included

• Magnitudeof change in hemoglobin (Hgb) levels.

• Percentageof patients who met criteria for a hematologic response as defined in each study’sprotocol.

• Percentageof patients who were transfused.

• Numberof RBC units transfused per patient normalized to a 4-week period.

• Qualityof life.

• Symptomsof anemia (besides fatigue or other components measured by quality-of-lifequestionnaires).

• Adverseeffects of treatment.

The systematic review addressed the following key questionsseparately for each clinical setting:

1. Effects ofErythropoietin Treatment—What were therelative effects on outcomes of managing anemia with erythropoietin comparedwith transfusion alone? In settings other than stem-cell transplants, what werethe relative effects of erythropoietin treatment when different Hgb thresholdswere used to initiate erythropoietin treatment?

2. Variations inErythropoietin Regimens—In the includedstudies, did variations in the erythropoietin treatment regimen (such as dose,frequency, duration, route) affect the outcomes of treatment? Were thesevariations likely to confound interpretation of the evidence on the relativeeffects of erythropoietin treatment according to the alternative Hgb thresholdsfor initiating treatment?

3. Identification of Patient Response—Were there populationsor subgroups of patients more or less likely to benefit from erythropoietintreatment? Were there laboratory measurements that either predicted or permittedearly identification of patients whose anemia responded to erythropoietin?

4. Data on Adverse Effects—Whatwere the incidence and severity of adverse effects associated with the use oferythropoietin and how did these compare with the adverse effects oftransfusion?

Data abstraction for adverse events also was limited tocontrolled trials so that effects of erythropoietin could be distinguished fromeffects of disease progression or concurrent therapies for the underlyingmalignancies.

To supplement the systematic review, we conducted aliterature-based meta-analysis of the effect of erythropoietin on the odds oftransfusion for patients with anemia or at risk of anemia due primarily tocancer therapy. A random effects model was used to calculate the combined oddsratio of transfusion for the 12 randomized controlled trials that reportednumbers or percentages of patients transfused, with or without erythropoietinadministered subcutaneously, for treatment-related anemia. The odds ratioexpresses the relative likelihood that erythropoietin-treated patients will betransfused compared with the likelihood for controls. Published data wereinsufficient for literature-based meta-analysis of other outcomes, or of odds oftransfusion in other clinical settings.

Sensitivity analysis compared results of higher-qualitytrials to those of lesser-quality trials. A trial was classified as higherquality when it was randomized and double-blinded and met our criteriaconcerning limits on the number of subjects excluded from the analysis ofresults. We required that < 10% of subjects within each study arm wereexcluded from the analysis and that the ratio of exclusions from each arm wasless than 2:1; or, alternatively, that results were reported as anintention-to-treat analysis.

External Review

AHRQ requires that Evidence-based Practice Center reportsundergo extensive review by external experts and representatives of stakeholderorganizations. Early in each project, these individuals review and provide inputto modify the study protocol. Later, they review and comment on the report’sinitial draft. However, each Evidence-based Practice Center has ultimateresponsibility for the final draft of its reports, subject to AHRQ review.

For the erythropoietin report,[2] the BCBSA TechnologyEvaluation Center Medical Advisory Panel, which includes nationally recognizedexperts in technology assessment and hematology/oncology, reviewed a preliminaryanalysis of the evidence base. Additionally, 20 external reviewers critiqued thestudy protocol and draft report, and revisions were made based on theircomments. Eight reviewers were invited by Technology Evaluation Center based ontheir expertise in medical oncology, hematology, transfusion medicine,quality-of-life and systematic review methodology. One reviewer directed anotherAHRQ Evidence-based Practice Center and is a medical oncologist. Ten reviewerswere appointed by professional organizations other than ASCO or ASH and bypatient advocacy groups. These reviewers included clinical and researchspecialists involved in the treatment of cancer and/or management ofcancer-related anemia and patient advocacy representatives. One externalreviewer was from the technical staff of Ortho Biotech, Inc. Lists of theTechnology Evaluation Center Medical Advisory Panel members, external reviewers,and technical advisors are included in the evidence report’s appendices.[2]

Erythropoietin for Anemia Primarily Due to Cancer Therapy

Erythropoietin vs Transfusion

The evidence review is based on data abstraction and analysisof 22 controlled trials with a total enrollment of 1,927 patients.[3-24] Alltrials compared the outcomes of using erythropoietin to manage anemia inpatients undergoing therapy for a malignancy with the outcomes of RBCtransfusion alone. As shown in Table 1, 18 trials with 1,698 enrolled patients(88%) were randomized,[3-20] and 7 of these (853 patients; 44%) were placebo-controlled anddouble-blind.[3-9] For all 22 trials, the number of patients reported asevaluable is 1,838, which is 95% of all enrolled patients. We classified the 22trials into three categories defined by the study patients’ mean Hgb atenrollment: Hgb ≤10 g/dL[3-7,9,11,15,16,23]; Hgb >10 but< 12 g/dL[8,10,12,17,21,22,24]; and Hgb ³12 g/dL[13,14,18-20]. No trials directly compared the outcomes of initiatingerythropoietin treatment at different Hgb thresholds.

The systematic review found adequate and consistent evidencethat erythropoietin increased Hgb levels and percentage of patientsdemonstrating hematologic response, when compared with controls managed bytransfusion alone (Table 2). This was true for pediatric patients as well asadults.

Most trials (17 of 22) reported the number or percentage ofpatients transfused (Tables 1 and 3), although many reported that differencesbetween arms were not statistically significant, and a few did not test for thestatistical significance of differences they reported. Just above half thetrials (12 of 22) reported the number of RBC units transfused per patient (Tables1 and 3). No trials reported effects of erythropoietin use on symptomsof anemia other than fatigue (not shown).

For all randomized studies that gave erythropoietinsubcutaneously,[3-7,10-12,14,16,19,20] meta-analysis showed that erythropoietinreduced the odds of transfusion by a factor of 0.38 compared with controls notgiven erythropoietin (Table 4; Figure1). This indicated that the likelihood oftransfusion for erythropoietin-treated patients was 38% of the likelihood forcontrols. The overall number needed to treat (NNT) calculated for this group ofstudies was 4.4 (95% confidence interval [CI] = 3.6-6.1), which suggested fouror five patients must be treated with erythropoietin to spare one patient fromtransfusion.

Sensitivity analysis found a smaller magnitude of riskreduction for higher-quality studies,[3-7] which were double-blinded (Figure2; Table 4). For higher-quality studies, the calculated NNT was 5.2 (95% CI = 3.8-8.4),and for lower-quality studies,[10-12,14,16,19,20] the calculated NNT was 2.6(95% CI = 2.1-3.8). Thus, higher-quality studies predicted one patient would avoid transfusionfor every five to six patients treated with erythropoietin, while thelesser-quality studies predicted one for every two to three patients treated.There was evidence that in unblinded studies physicians may have been moreaggressive in transfusing patients in the control arm, thus overestimating theobserved effect of erythropoietin.[14]

QOL Outcomes

The strongest evidence for an effect of erythropoietin onquality-of-life outcomes (Table 5) was a randomized, double-blinded,placebo-controlled trial available only as an abstract but later published infull.[7] In a patient population with mean baseline Hgb level ≤10 g/dL, investigators reported statistically significant differences in scorechanges that favored the erythropoietin-treated arm for three questions thatused visual analog scales (number of patients evaluable = 335) and for theFunctional Assessment of Cancer Treatment-Anemia (FACT-An) (number of patientsevaluable = 290). The study also reported statistically significant positivecorrelations between changes in Hgb levels and changes in quality-of-lifescores.

However, key methodologic features of administering thequality-of-life instruments[2] were not described. Additionally, the minimum changes inquality-of-life scores considered clinically significant were not definedprospectively or in the discussion of results. Furthermore, this trial did notactually compare the quality-of-life effects of initiating erythropoietintreatment at alternative thresholds of baseline Hgb, nor did the analysisstratify patients by different Hgb levels at entry. Eight other publishedstudies,[3,4,6,13,14, 19,20,23] which included a total of 516 evaluablepatients, did not provide consistent evidence that erythropoietin improvedquality-of-life outcomes. As shown in Table6, two of these reported onlywithin-arm comparisons of initial scores to final scores, but did not comparethe experimental and control arms to each other.[3,14]

Relative Effects at DifferentHgb Thresholds

The most robust evidence that erythropoietin improvestransfusion outcomes for patients undergoing therapy for malignancy comparedwith transfusions alone came from trials in patient groups with baseline Hgb ≤10 g/dL.[3-7,9,11,15,16,23] Transfusion outcomes did not appear to be superiorin trials where erythropoietin treatment was initiated in groups of patients whohave mean Hgb > 10 g/dL compared with trials where mean Hgb was ≤10 g/dL (Table 3). Among trials in adult patients with a baseline Hgb≤10 g/dL, the range of differences between erythropoietin and control arms for the percentage of patientstransfused was 9% to 45%.[3-7,11,15] For a baseline level of Hgb > 10 but< 12 g/dL, the range was 7% to 47%[8,10,12,17,21,22,24]; and 7% to 39% for abaseline level of Hgb ³12 g/dL.[13,14,18-20] However, these ranges are wide, and it is uncertainwhether the three groups of studies compared patient populations that weresimilar except for baseline Hgb.

The available evidence was inadequate to determine whetheroutcomes of erythropoietin treatment were superior when treatment was initiatedin groups who have mean Hgb > 10 g/dL, compared with groups where mean Hgb is£ 10 g/dL.Randomized controlled trials, double-blinded and adequately powered, arenecessary to compare the outcomes of erythropoietin treatment initiated atvarious Hgb thresholds. Inferences from indirect comparison of the results ofthe available trials cannot resolve this question.

It is possible that adequately powered comparative trialsmight show fewer transfusions when erythropoietin treatment starts at higherHgb levels than when it starts near 10 g/dL. However, examining the evidence wereviewed suggests two reasons why this is unlikely. First, patients withentry-level Hgb below the mean may have accounted for most of the transfusionsamong erythropoietin-treated patients in trials where baseline Hgb was ≤10 g/dL. Thus, the greatest yield for reducing the number of patients transfusedin this population might come from initiating erythropoietin before theHgb level falls substantially below 10 g/dL, rather than by initiatingerythropoietin treatment at a level substantially above 10 g/dL. Second, in alltrials, patients who were unresponsive to erythropoietin may have accounted fora substantial proportion of patients transfused. Initiating erythropoietintreatment at a higher Hgb level is not expected to reduce transfusions in thissubgroup of patients.

Effects of Different Methods for Administering Erythropoietin

The meta-analysis examined whether the characteristics oferythropoietin administration (dosing regimen, treatment duration, dose range)had an effect on the estimate of the summary odds ratio for transfusion. Onlyerythropoietin dose appeared to have an independent effect on transfusionoutcomes, but this was potentially confounded by study quality. However, theresults of two randomized controlled trials that directly compared lower andhigher doses of erythropoietin (450 vs 900 U/kg/wk) did not demonstrate that thehigher dose was superior in preventing transfusions.[12,14]

Effects of Patient Characteristics

• Age—Erythropoietinwas effective in preventing transfusion in pediatric patients.[16,23] No studiesreported outcomes stratified specifically for geriatric patients, but adults upto age 90 were included in some trials.[3,4,7]

• Malignant Disease—Therewas evidence that erythropoietin produced hematologic responses and probablyreduced transfusions in patients with nonmyeloid hematologic malignancies[7,15]to a similar degree as in patients with tumors of solid organs or tissues.

• Radiotherapy—Althougherythropoietin increased Hgb levels for patients managed with radiotherapyalone, mean Hgb levels of control patients did not decrease from baselinevalues.[13,17,22] The radiotherapy regimens utilized apparently did notcontribute to or exacerbate pre-existing mild anemia.

• Platinum Regimens—Theevidence demonstrated benefit from erythropoietin for patients givenchemotherapy regimens that included cisplatin or carboplatin, as well asregimens that did not include either of the platinum drugs.[2](See the full evidence report[2] for citations.)

• Predictors of Response—The22 trials included in this evidence base reported no significant predictors ofresponse to erythropoietin therapy.[2] In particular, neither baseline serumerythropoietin nor the ratio of observed to predicted serum erythropoietinlevels (O/P ratio) predicted response in any analysis.

Adverse Effects AssociatedWith Erythropoietin Use

Limited evidence on adverse events was available from thestudies included in this review, but the frequencies of those reported did notappear to differ markedly between erythropoietin-treated patients andcontrols.[2] The only statistically significant difference was a greaterfrequency of fatigue reported by patients in the control arms.

Anemia Due Primarily to Malignant Disease

The literature search identified six controlled trials (Table7), all randomized, with a total enrollment of 693 patients that met inclusioncriteria for this systematic review.[25-30] Three trials were placebo-controlledand double-blind (n = 332; 48%).[25-27] Of the 693 patients enrolled, 648(93.5%) were reported as evaluable. Patients in this evidence base had diagnosesknown to have a high occurrence of anemia of malignancy (multiple myeloma,non-Hodgkin’s lymphoma, chronic lymphocytic leukemia, and myelodysplasticsyndromes). With the exception of one trial in patients with myelodysplasticsyndromes,[25] the preponderance of patients in these trials received concurrenttherapy for their malignancy.

Erythropoietin vs Transfusion

There was consistent evidence that erythropoietin increasedHgb levels and percent of patients demonstrating hematologic response in patients with anemia of malignancy(Table 7).The evidence on transfusion outcomes was sparse, but suggested a favorableeffect of erythropoietin treatment. The only report on measurements of qualityof life was an abstract that did not provide sufficient detail forinterpretation of the results.[27] All patients included in these studies hadbaseline hemoglobin £10 g/dL. The evidence did not address alternative thresholds for initiatingerythropoietin treatment in patients with anemia of malignancy.

Effects of Different Methods forAdministering Erythropoietin

The studies suggested that starting doses in the 200 to 450U/kg/wk range were adequate to achieve hematologic response. However, the onlystudy of patients with myelodysplastic syndrome used a much higher dose, 1,050U/kg/wk, yet obtained a smaller increase in response rate.[25] The distinctmechanism of anemia in this clonal disorder probably contributed to the reducedresponse rate.

Effects of Patient Characteristics

• Malignant Disease—Astatistically significant hematologic response in the erythropoietin arm wasreported for all hematologic malignancies included in this review. However, thelimited evidence available suggested that hematologic response rates were lowerfor patients with myelodysplastic syndrome.

• Age—Allstudies were of adults. There were no studies of pediatric patients or thatseparately reported on geriatric patients.

• Prior Transfusion—Erythropoietinincreased hematologic responses or Hgb levels for patients with either multiplemyeloma or non-Hodgkin’s lymphoma, irrespective of history of priortransfusion.[28-30] A single study of myelodysplastic syndrome patients reportedthat erythropoietin increased hematologic responses for patients without previous history of transfusion, but notfor those previously transfused.[25] History of prior transfusion, however, maybe associated with other characteristics, such as duration and progression ofdisease (which may have affected erythropoiesis in myelodysplastic syndromepatients).

• Predictors of Response—Thisgroup of studies did not provide sufficient evidence to draw conclusions onpredictors of response. Only the serum concentration of endogenouserythropoietin at baseline[25,30] and the ratio of observed to expectedconcentrations of serum erythropoietin[28,30] were reported as significantpredictors of response in at least two trials.

Adverse Effects

There was a statistically significant increase inhypertension (10% vs 1%; P = .011) and a nonsignificant increase inthromboembolic events (3% vs 0%; P = .55) among those treated witherythropoietin. The reported frequency of adverse events other than hypertensionand thromboembolic events did not appear to differ between erythropoietin-treatedpatients and controls.

Anemia Resulting From Marrow Ablation and Allogeneic Stem-Cell Rescue

Erythropoietin vs Transfusion

The evidence concerning the use of erythropoietin afterhigh-dose chemotherapy and allogeneic stem-cell transplantation was derived fromseven controlled studies (total enrollment: 493) of patients with malignancies that are representative of those undergoing marrow-derived allogeneicstem-cell transplantation in clinical practice.[31-37] Of the seven controlledtrials, all but two[31-35] were randomized (total enrollment in randomizedstudies: 400); nonrandomized trials compared erythropoietin-treated patientswith historical controls.[36,37] The largest study enrolled and evaluated 215patients[31]; all other studies enrolled fewer than 100 patients. Outcomesreported from these seven trials are summarized in Table8, with the studieslisted in order of increasing erythropoietin dose.

These studies compared the outcomes of transfusion of redblood cells initiated at a predefined threshold with the outcomes oferythropoietin treatment supplemented with transfusion of red blood cells whennecessary. One study exclusively enrolled pediatric patients.[37] The enrolledpatients had a variety of hematologic tumors. All of the studies used marrow asthe stem-cell source, and all studies administered erythropoietin intravenously.

In four of five trials reporting this outcome, erythropoietinresulted in a statistically significant decrease in the time to RBC engraftment(see Table 8), as indicated by achievement of a predetermined Hgb levelindependent of transfusion support.[31-33,36] The range of reduction reportedwas 1 to 2 weeks. Reticulocyte measures, which tend to predict RBC engraftment,also suggested more rapid engraftment with erythropoietinadministration.[31,33-37]

Outcomes for day of last transfusion were related to andcorrelated with RBC engraftment by Hgb level results, with statisticallysignificant results favoring the erythropoietin-treated study arm.[31,32,36]

Erythropoietin administration is unlikely to spare anyonefrom transfusion, as recipients of allogeneic stem-cell transplantation areuniformly anemic following the procedure and response to erythropoietin, whetherendogenous or exogenous, is not immediate. The evidence suggested, however thaterythropoietin treatment may have decreased the number of RBC unitstransfused.[32,34,35,37]

Limited evidence suggested that erythropoietin treatment hadno significant effect on length of hospital stay.[32,33] This is not surprising,given the number of complications from allogeneic stem-cell transplantation thatare unrelated to anemia.

Effects of Different Methods for Administering Erythropoietin

Transfusion outcomes appeared to be associated with theduration of follow-up for reporting and statistical comparison: Shorterfollow-up was more often associated with a significant beneficialeffect,[34,35,37] whereas longer follow-up may have been complicated bytransfusions for graft-vs-host disease and resulted in nonsignificant outcomesfor erythropoietin.[31,33] For both RBC engraftment and RBC transfusionoutcomes, results obtained with erythropoietin dose extremes (525 or 3,500U/kg/wk) [35, 37] did not appear to differ from those obtained with the moderatedoses (700-1,050 U/kg/wk) used in most of the studies.

Effects of Patient Characteristics

• Age—Althoughonly one small study (nonrandomized, historical controls) specifically examinedthe use of erythropoietin in a pediatric population,[37] results were consistentwith those obtained in all other studies, which enrolled primarily adultpopulations. Additionally, significant results were obtained in this study usinga dose/kg/wk that was half or less than the doses used in studies of adultpatients.

Adverse Effects

There did not appear to be significant adverse eventsassociated with erythropoietin treatment in patients receiving allogeneicstem-cell transplants (reporting was sparse however). The available evidenceshowed no depression of platelet engraftment with erythropoietin treatment.

Anemia Resulting From Marrow Ablation Prior to Autologous Stem-Cell Rescue

The literature search and review for studies oferythropoietin use after autologous transplantation identified six controlledtrials (total enrollment: 321).[31,37-41] Three of the six trials[31,38,39] wererandomized (total enrollment: 169); nonrandomized trials compared erythropoietin-treatedpatients with historical controls.[37,40,41] Studies ranged in size from 20[38]to 114[31] enrolled patients. All of the studies used marrow as the exclusivesource of stem cells except for one[39] in which patients with Hodgkin’slymphoma were also given peripheral blood stem cells. Nevertheless, it appearsthat results from these studies can be generalized to patients transplanted withperipheral blood stem cells, the current standard of care. Study outcomes aresummarized in Table 9.

Erythropoietin vs Transfusion

The evidence did not support a beneficial effect oferythropoietin administration on RBC engraftment, RBC transfusion, or length ofhospital stay outcomes. It is particularly noteworthy that two studies,[31,37]that used the same erythropoietin protocol for both allogeneic and autologousstem-cell transplant patients reported several outcomes significantly improvedonly for allogeneic stem-cell transplant patients.

Differences in Erythropoietin Administration

Since the available evidence did not show a clear benefit forerythropoietin treatment, there was no evidence to favor a particular dose,dosing regimen, or treatment duration. Although it is possible that treatmentduration was too short in all included studies to significantly improveoutcomes, reticulocyte measures (an early indicator of RBC engraftment) did notindicate a probable response.[31,37,39]

Patient Characteristics

Erythropoietin did not show a beneficial effect for theentire population of patients treated in these studies. Results among thesubpopulations were consistent with overall results, and no subpopulation thatderived benefit from erythropoietin treatment could be identified.

The lack of response to erythropoietin in patients givenmarrow stem cells suggests that patients given peripheral blood stem cells alsowould be unlikely to respond. Preparations of peripheral blood stem cellsmobilized with growth factors contain progenitor cells from the erythroid (andother) lineage(s). These progenitors are farther along the maturation pathway tofunctional end-stage cells, and may be less dependent on erythropoietin than areunstimulated stem cells harvested from the marrow. The time to recovery of redcell counts and correction of anemia thus appears less likely to be shortened byerythropoietin therapy after infusion of peripheral blood stem cells than afterinfusion of marrow stem cells.

Adverse Effects

There did not appear to be significant adverse eventsassociated with erythropoietin treatment in patients receiving autologousstem-cell transplants (reporting was sparse, however). The available evidenceshowed no depression of platelet engraftment with erythropoietin treatment.

Future Research

The most robust evidence that erythropoietin treatmentimproves outcomes for oncology patients (and its most common oncologic use inthe United States) is for those patients concurrently undergoing cancer therapy.Consistent evidence demonstrates that erythropoietin reduces transfusionrequirements if treatment is initiated when declining Hgb levels approach 10 g/dL.More limited evidence suggests that erythropoietin also improves quality of lifefor mildly anemic patients.

Inferences from indirect comparisons of results fromavailable trials, however, were unable to resolve the question of an optimal Hgbthreshold level for initiating erythropoietin treatment. Inferences from largeuncontrolled studies[42-44] that enrolled patients with baseline Hgb levels ≤10.5 g/dL also cannot resolve this uncertainty, although they may providesuggestive evidence for a target at which Hgb levels should be maintained inthose who are treated. Randomized controlled trials, adequately powered, areneeded to directly compare treatment initiated at higher baseline Hgb levels (eg,£ 12 g/dL) with treatment delayed until Hgbapproaches 10 g/dL. Such direct comparisons can best determine whether earlierinitiation of erythropoietin therapy yields greater benefits than delayedtreatment, by further reducing transfusion use or improving quality of life.

The systematic review also identified common deficiencies inthe design and reporting of trials on erythropoietin. In addition to thepreponderance of unblinded studies, deficiencies common to this literatureincluded:

• Inadequatestatistical power.

• Failureto report on concealment of allocation.

• Failureto consistently report on a common set of clinically relevant outcomes.

• Failureto consistently test and report on statistical significance.

• Failureto account for patients lost to follow-up or excluded from analysis.

• Failureto use intent-to-treat analyses.

Some methodologic deficiencies may result in overestimationof the effects of erythropoietin, and inadequacy of reporting may limit theability to interpret and generalize results. Future trials should maintain ahigher standard of methodologic quality and completeness of reporting.

Published trials that reported on quality of life did notfollow recognized principles to minimize biases. Consequently, factors otherthan erythropoietin treatment may have affected outcomes. Future trials shouldmeasure effects of erythropoietin on quality of life more rigorously usingvalidated instruments, and by incorporating specific design features related toadministration of questionnaires and analysis and interpretation of results.

In nearly all trials, a substantial percentage of patientsdid not achieve a hematologic response to erythropoietin. Additionally,nonresponding patients may account for much of the transfusion use in theerythropoietin arms of these trials. To achieve the most efficient use oferythropoietin, additional systematic evidence is needed on baselinecharacteristics that predict responsiveness and on early indicators of response.

The reviewed evidence shows that initial doses oferythropoietin in the range of 300 to 450 U/kg/wk administered subcutaneouslyare adequate to increase Hgb and reduce the percentage of patients transfused.However, the optimal initial dose within this range has not been determined.Furthermore, within this dose range the team could not discern any difference inresponse rates between trials that used increasing dose regimens and those thatused decreasing dose regimens. To achieve the most efficient use oferythropoietin, comparative trials are needed to establish an optimal initialdose and to determine the optimal dosing regimen.

The team found evidence that patients with myelodysplasticsyndromes respond to erythropoietin, although response rates are much lower thanin other malignancies, and higher doses of erythropoietin appear to benecessary. To achieve the most efficient use of erythropoietin, additionalstudies are needed to determine which patients with myelodysplastic syndromesare most likely to respond. Studies are also needed to establish an optimal doseand dosing regimen.

This work was developed under contract with the Agency forHealthcare Research and Quality, Rockville, MD (AHRQ contract number290-97-0015). The authors of this article are responsible for its contents,including any clinical or treatment recommendations. No statement in thisarticle should be construed as an official position of the Agency for HealthcareResearch and Quality or of the U.S. Department of Health and Human Services.

References:

1. http://www.ahrq.gov/clinic/epicix/htm, accessed on June 1,2001.

2. Aronson N, Seidenfeld J, Piper MA, et al: Uses ofErythropoietin in Oncology. Evidence Report/Technology Assessment No. 30.(Prepared by the Blue Cross and Blue Shield Association Evidence-based PracticeCenter under Contract No. 290-97-0015.) AHRQ Publication No. 01-E009. Rockville,MD: Agency for Healthcare Research and Quality, 2001.

3. Case DC Jr, Bukowski RM, Carey RW, et al: Recombinanthuman erythropoietin therapy for anemic cancer patients on combinationchemotherapy. J Natl Cancer Inst 85:801-806, 1993.

4. Henry DH, Brooks BJ Jr, Case DC Jr, et al: Recombinanthuman erythropoietin therapy for anemic cancer patients receiving cisplatinchemotherapy. Cancer J Sci Am 1:252-260, 1995.

5. 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.

6. Kurz C, Marth C, Windbichler G, et al: Erythropoietintreatment under polychemotherapy in patients with gynecologic malignancies: Aprospective, randomized, double-blind placebo-controlled multicenter study. GynecolOncol 65:461-466, 1997.

7. Littlewood TJ, Bajetta E, Nortier JWR, et al for theEpoetin Alfa Study Group: Effects of epoetin alfa on hematologic parameters andquality of life in cancer patients receiving nonplatinum chemotherapy: Resultsof a randomized, double-blind, placebo-controlled trial. J Clin Oncol19:2865-2874, 2001.

8. Wurnig C, Windhager R, Schwameis E, et al: Prevention ofchemotherapy-induced anemia by the use of erythropoietin in patients withprimary malignant bone tumors (a double-blind, randomized, phase III study). Transfusion36:155-159, 1996.

9. Porter JC, Leahey A, Polise K, et al: Recombinant humanerythropoietin reduces the need for erythrocyte and platelet transfusions inpediatric patients with sarcoma: A randomized, double-blind, placebo-controlledtrial. J Pediatr 129:656-660, 1996.

10. Quirt I, Coutere F, Pichette R, et al: The role ofrecombinant human erythropoietin (EPO) in reducing red blood cell transfusionsand maintaining quality of life (QOL) in patients with lymphoma and solid tumorsreceiving cytotoxic chemotherapy: Results of a randomized, double-blind,placebo-controlled clinical trial [abstract]. Blood 88(10 suppl 1):347a,1996.

11. 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.

12. ten Bokkel Huinink WW, de Swart CAM, van Toorn DW, et al:Controlled multicentre study of the influence of subcutaneous recombinant humanerythropoietin on anaemia and transfusion dependency in patients with ovariancarcinoma treated with platinum-based chemotherapy. Med Oncol 15:174-182,1998.

13. Sweeney PJ, Nicolae D, Ignacio L, et al: Effect ofsubcutaneous recombinant human erythropoietin in cancer patients receivingradiotherapy: Final report of a randomized, open-labeled, phase II trial. BrJ Cancer 77:1996-2002, 1998.

14. Thatcher N, De Campos ES, Bell DR, et al: Epoetin alphaprevents anaemia and reduces transfusion requirements in patients undergoingprimarily platinum-based chemotherapy for small cell lung cancer. Br J Cancer80:396-402, 1999.

15. Silvestris F, Romito A, Fanelli P, et al: Long-termtherapy with recombinant human erythropoietin (rHu-EPO) in progressing multiplemyeloma. Ann Hematol 70:313-318, 1995.

16. Varan A, Buyukpamukcu M, Kutluk T, et al: Recombinanthuman erythropoietin treatment for chemotherapy-related anemia in children. Pediatrics103:E161-E164, 1999.

17. Henke M, Guttenberger R, Barke A, et al: Erythropoietinfor patients undergoing radiotherapy: A pilot study. Radiother Oncol50:185-190, 1999.

18. Gamucci T, Thorel MF, Frasca AM, et al: Erythropoietinfor the prevention of anaemia in neoplastic patients treated with cisplatin. EurJ Cancer 29A(suppl 2):S13-S14, 1993.

19. Del Mastro L, Venturini M, Lionetto R, et al: Randomizedphase III trial evaluating the role of erythropoietin in the prevention ofchemotherapy-induced anemia. J Clin Oncol 15:2715-2721, 1997.

20. Welch RS, James RD, Wilkinson PM, et al: Recombinanthuman erythropoietin and platinum-based chemotherapy in advanced ovarian cancer.Cancer J Sci Am 1:261-266, 1995.

21. Markman M, Reichman B, Hakes T, et al: The use ofrecombinant human erythropoietin to prevent carboplatin-induced anemia. GynecolOncol 49:172-176, 1993.

22. Lavey RS, Dempsey WH: Erythropoietin increases hemoglobinin cancer patients during radiation therapy. Int J Radiat Oncol Biol Phys27:1147-1152, 1993.

23. Leon P, Jimenez M, Barona P, et al: Recombinant humanerythropoietin for the treatment of anemia in children with solid malignanttumors. Med Pediatr Oncol 30:110-116, 1998.

24. Dusenbery KE, McGuire WA, Holt PJ, et al: Erythropoietinincreases hemoglobin during radiation therapy for cervical cancer. Int JRadiat Oncol Biol Phys 29:1079-84, 1994.

25. Italian Cooperative Study Group for rHuEpo inMyelodysplastic Syndromes: A randomized double-blind placebo-controlled studywith subcutaneous recombinant human erythropoietin in patients with low-riskmyelodysplastic syndromes. Br J Haematol 103:1070-1074, 1998.

26. Garton JP, Gertz MA, Witzig TE, et al: Epoetin alfa forthe treatment of the anemia of multiple myeloma. A prospective, randomized,placebo-controlled, double-blind trial. Arch Intern Med 155:2069-2074,1995.

27. Rose E, Rai K, Revicki D, et al and the EPO in Anemia ofCLL Study Group: Clinical and health status assessments in anemic chroniclymphocytic leukemia (CLL) patients treated with epoetin alfa (EPO) [abstract]. Blood84(10 Suppl 1):526a, 1994.

28. Osterborg A, Boogaerts MA, Cimino R, et al: Recombinanthuman erythropoietin in transfusion-dependent anemic patients with multiplemyeloma and non-Hodgkin’s lymphoma—A randomized multicenter study. TheEuropean Study Group of Erythropoietin (Epoetin Beta) Treatment in MultipleMyeloma and Non-Hodgkin’s Lymphoma. Blood 87:2675-2682, 1996.

29. Dammacco F, Silvestris F, Castoldi GL, et al: Theeffectiveness and tolerability of epoetin alfa in patients with multiple myelomarefractory to chemotherapy. Int J Clin Lab Res 28:127-134, 1998.

30. Cazzola M, Messinger D, Battistel V, et al: Recombinanthuman erythropoietin in the anemia associated with multiple myeloma ornon-Hodgkin’s lymphoma: Dose finding and identification of predictors ofresponse. Blood 86:4446-4453, 1995.

31. Link H, Boogaerts MA, Fauser AA, et al: A controlledtrial of recombinant human erythropoietin after bone marrow transplantation. Blood84:3327-3335, 1994.

32. Klaesson S, Ringden O, Ljungman P, et al: Reduced bloodtransfusions requirements after allogeneic bone marrow transplantation: Resultsof a randomised, double-blind study with high-dose erythropoietin. BoneMarrow Transplant 13:397-402, 1994.

33. Biggs JC, Atkinson KA, Booker V, et al: Prospectiverandomised double-blind trial of the in vivo use of recombinant humanerythropoietin in bone marrow transplantation from HLA-identical sibling donors.The Australian Bone Marrow Transplant Study Group. Bone Marrow Transplant15:129-134, 1995.

34. Steegmann JL, Lopez J, Otero MJ, et al: Erythropoietintreatment in allogeneic BMT accelerates erythroid reconstitution: Results of aprospective controlled randomized trial. Bone Marrow Transplant10:541-546, 1992.

35. Vannucchi AM, Bosi A, Linari S, et al: High doses ofrecombinant human erythropoietin fail to accelerate platelet reconstitution inallogeneic bone marrow transplantation. Results of a pilot study. Haematologica82:53-56, 1997.

36. Link H, Brune T, Hubner G, et al: Effect of recombinanthuman erythropoietin after allogenic bone marrow transplantation. Ann Hematol67:169-173, 1993.

37. Locatelli F, Zecca M, Pedrazzoli P, et al: Use ofrecombinant human erythropoietin after bone marrow transplantation in pediatricpatients with acute leukemia: Effect on erythroid repopulation in autologous vsallogeneic transplants. Bone Marrow Transplant 13:403-410, 1994.

38. Vannucchi AM, Bosi A, Ieri A, et al: Combination therapywith G-CSF and erythropoietin after autologous bone marrow transplantation forlymphoid malignancies: A randomized trial. Bone Marrow Transplant17:527-531, 1996.

39. Chao NJ, Schriber JR, Long GD, et al: A randomized studyof erythropoietin and granulocyte colony-stimulating factor (G-CSF) versusplacebo and G-CSF for patients with Hodgkin’s and non-Hodgkin’s lymphomaundergoing autologous bone marrow transplantation. Blood 83:2823-2828,1994.

40. Ayash LJ, Elias A, Hunt M, et al: Recombinant humanerythropoietin for the treatment of the anaemia associated with autologous bonemarrow transplantation. Br J Haematol 87:153-161, 1994.

41. Pene R, Appelbaum FR, Fisher L, et al: Use of granulocyte-macrophagecolony-stimulating factor and erythropoietin in combination after autologousmarrow transplantation. Bone Marrow Transplant 11:219-222, 1993.

42. Glaspy J, Bukowski R, Steinberg D, et al: Impact oftherapy with epoetin alfa on clinical outcomes in patients with nonmyeloidmalignancies during cancer chemotherapy in community oncology practice. ProcritStudy Group. J Clin Oncol 15:1218-1234, 1997.

44. Demetri GD, Kris M, Wade J, et al: Quality-of-lifebenefit in chemotherapy patients treated with epoetin alfa is independent ofdisease response or tumor type: Results from a prospective community oncologystudy. Procrit Study Group. J Clin Oncol 16:3412-3425, 1998.

45. Gabrilove JL, 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.