- Email: [email protected]
Effectiveness of erythropoietin in the treatment of patients with malignancies: methods and preliminary results of a Cochrane review
Best Practice & Research Clinical Haematology Vol. 18, No. 3, pp. 449–454, 2005 doi:10.1016/j.beha.2005.01.022 available online at http://www.sciencedirect.com
8 Effectiveness of erythropoietin in the treatment of patients with malignancies: methods and preliminary results of a Cochrane review Julia F. Bohlius*
MD, MSc PH
Research Fellow
Simon Langensiepen Andreas Engert
MD
MD
Cochrane Hematological Malignancies Group, Department I of Internal Medicine, University of Cologne, Kerpener-Street 9, 50924 Cologne, Germany
Guido Schwarzer MSc, PhD Institute of Medical Biometry and Medical Informatics, Freiburg, Germany Charles L. Bennett MD, PhD VA Chicago Healthcare System, The Feinberg School of Medicine, Northwestern University, Chicago, USA Cancer and cancer therapy-associated anemia may have an impact on tumor response and overall survival. Additionally, anemia represents an important economic factor. Therefore, therapeutic alternatives such as erythropoietin (EPO) and red blood cell transfusions have to be evaluated systematically. The effectiveness of recombinant human EPO to prevent or alleviate anemia in patients with malignant disease was determined. Randomized controlled trials comparing prophylaxis or treatment of anemia with EPO plus red blood cell transfusion (RBCT) or RBCT only in patients with malignant disease undergoing antineoplastic therapy were included. The endpoints needed for RBCT were hematological response (hemoglobin increase of 2 g/dL or hematocrit increase of 6%), tumor response, and overall survival. Medical databases (Cochrane Library, MEDLINE, EMBASE) and conference proceedings were searched (1985–2001). Full-text and abstract publications were included as well as unpublished data. Data extraction and quality
* Corresponding author. Tel.: C49 221 710 770 22; Fax: C49 221 710 770 24. E-mail address: [email protected] (J.F. Bohlius). 1521-6926/$ - see front matter Q 2005 Published by Elsevier Ltd.
450 J. F. Bohlius et al
assessment were done in duplicate. All authors were contacted to obtain missing data. Out of 33 eligible studies, 27 trials with 3,287 randomized patients were included. Key words: cancer therapy; recombinant human erythropoietin; systematic review.
Anemia commonly accompanies malignancies and may result from changes in the production of or response to erythropoietin (EPO) caused by the cancer and/or its treatment. Various clinical trials have shown that EPO increases red blood cell counts and hemoglobin concentrations in a large number of patients with anemia caused by cancer or cancer therapy. A comprehensive evidence-based guideline was published by the American Society of Clinical Oncology and the American Society of Hematology.1 They incorporate the evidence that was published in the systematic review of the Agency for Health Care Research and Quality.2,3 Although this report provides evidence that EPO significantly increases hemoglobin levels in anemic cancer patients and reduces blood transfusion requirements, the critical question whether EPO influences tumor response and prolongs survival time has not been answered yet. Also, uncertainty remained with respect to the optimal hemoglobin level to initiate erythropoietin therapy. A Cochrane Review meta-analysis is being conducted of the trials on EPO used in the treatment or prevention of anemia. The methods and first results from this meta-analysis are presented here. The objectives of the study included the impact of EPO on hematological parameters measured by hematological response and red blood cell transfusions (RBCT). Other assessments are made on the effect EPO had on tumor response, overall survival, and quality of life, as well as the incidence of adverse events.
METHODS This is an international study and the title was registered within the Cochrane Collaboration. A protocol was set up, peer-reviewed, and published in the Cochrane Library.4 A comprehensive search of the literature was done, based on the highly sensitive search strategy for identifying reports of randomized controlled trials developed by Dickersin et al. 5 Medical databases from the Cochrane Library, MEDLINE, EMBASE, Lilacs, and others, as well as registers of ongoing studies, were scanned from the years 1985 to 2001. There were no language restrictions. Conference proceedings from the American Society of Clinical Oncology, American Society of Hematology, and the European Society of Medical Oncology were hand searched. In addition, we contacted experts in the field and pharmaceutical companies who manufacture EPO. In this analysis, all authors were contacted to obtain additional unreported information. The identification of relevant trials, as well as subsequent steps such as critical appraisal, data extraction, and computing, were done in duplicate by two independent reviewers (SL,JB). Study selection Inclusion criteria were met only by randomized controlled studies, with or without blinding, using EPO to treat or prevent anemia in patients with malignant disease. Placebo control, as opposed to no treatment, was not required for inclusion, but was considered in evaluating study quality. Trials in which patients were allocated by a quasi-random method; e.g. date of birth or day of month were excluded because we
Effectiveness of erythropoietin in the treatment of patients with malignancies 451
considered this study design to be of poor quality leading to unreliable results. More than 10 similarly treated evaluable patients in each study arm or relevant stratum were required for inclusion. Ongoing studies and interim analyses were not included. Only participants diagnosed with malignant disease, using clinical and histological/cytological criteria were included regardless of type or stage of the disease or previous therapy. All study participants had to be anemic or at risk for anemia from chemotherapy and/or radiotherapy or the underlying malignant disease. Other reasons for anemia, such as hemolysis, iron deficiency, and occult bleeding, had to be excluded. The identified studies were stratified on patients’ baseline characteristics of hemoglobin below 10, 10–12 g/dL and above 12 g/dL. There were no age restrictions, but at the end of the study only adult patients were included. Epoetin alfa or epoetin beta had to be administered subcutaneously or intravenously at a dose of at least 300 U/kg body weight per week given for at least 4 weeks. Dose adaptation of EPO was allowed, depending on hematological response. Studies on new erythropoiesis stimulating substances like darbepoetin were not considered for this review. Concomitant supportive treatments such as G-CSF or iron supplementation had to be given equally in all study arms. Trials on high-dose myeloablative chemotherapy regimens followed by bone marrow or peripheral blood stem cell transplantation were excluded, as were trials using EPO for short-term preoperative treatment to correct anemia or to support collection of autologous blood prior to cancer surgery. Published, full-text articles and abstracts, as well as unpublished data, are included in the meta-analysis. Quality assessment All trials were analyzed and subjected to a quality assessment that determined whether the trials were randomized and double blind, whether there was concealment of allocation, an intent-to-treat analysis, and documentation of dropouts and withdrawals. In addition we assessed, whether the patient characteristics were similar at baseline in the study groups compared. Studies were excluded from the analysis if they were not truly randomized or inadequately concealed allocation. A data questionnaire was sent out to all the study offices requesting hemoglobin levels, interventions, baseline characteristics, RBCT, mortality dropouts, tumor response, and overall survival information. Sensitivity analyses were performed for baseline hemoglobin, different malignancies, therapies, iron supplementation, granulocyte colony stimulating factor, methodological aspects such as concealment of allocation, masking, completeness of data, and intent-to-treat analyses, the type of publication, and the duration of study. To date, sensitivity analyses have not been completed concerning different dosages of EPO, dosage increases during treatment, and different methods of administration. Data analysis A fixed effect model was assumed in all meta-analyses. For binary data, the relative risk was used as a measure of treatment effect and the Mantel-Haenszel Method was used for pooling. For continuous data, weighted mean differences were calculated. Time to event data such as overall survival were calculated as hazard ratios based on individual patient data. If these data were not available the hazard ratio was calculated from published reports, using methods described in Parmar et al 6 The p-value of the homogeneity test was only used to describe the extent of heterogeneity inherent
452 J. F. Bohlius et al
in a meta-analysis. For primary outcome measures potential causes of heterogeneity were explored by performing sensitivity and subgroup analyses. In meta-analyses with at least four trials, a funnel plot was generated and a linear regression test7 was performed to examine the likely presence of bias in meta-analysis. A p-value less than 0.1 was considered significant for the linear regression test. Analyses were performed using Review Manager, Version 4.2.
RESULTS Overall, 1592 studies were screened and most (1511) were excluded immediately (Figure 1). The group was narrowed to 81 studies, which were subjected to more Potentially relevant RCTs identified and screened for retrieval (n=1,592)
RCTs excluded with reasons (n=1,511)
RCTs retrieved for more detailed evaluation (n=81) RCTsexcluded:did not fulfill the inclusion criteria (n=23)
Potentially appropriate RCTs to be included in the meta-analysis (n=58) RCTs excluded from metaanalysis: ongoing studies (n=24)
RCTs included in the meta-analysis (n=34)
RCTs withdrawn, by outcome, with reasons (n=7)
RCTs with usable information, by outcome (n=27) Figure 1. Improving the quality of reports of meta-analyses of randomized controlled trials: The QUORUM statement flow diagram.9
Effectiveness of erythropoietin in the treatment of patients with malignancies 453
detailed evaluation. Of the 81 studies, 23 did not fulfill the inclusion criteria and 24 are still ongoing and have to be included in the update of this review. Thirty-four were deemed to be eligible; however, of these, another seven studies had to be excluded. Two study reports did not report evaluable data, and we were not able to obtain additional data. Two randomized controlled trials turned out to be of poor quality. Two additional trials were excluded because one turned out to be a duplicate, and another was an interim analysis where a final analysis was never done. Overall, 27 trials with 3287 randomized patients were included. In the majority of studies (16, nZ2615), patients had a baseline mean hemoglobin !10 g/dL. In six studies, 348 patients had hemoglobin levels between 10 and 12 g/dL; in five studies, 324 patients had hemoglobin levels O12 g/dL. All of these five studies included patients with solid tumors. Overall, there were 13 studies of 969 patients with mostly solid tumors. There were six studies of patients with hematological malignancies amounting to 1059 patients, all of who had a mean hemoglobin at baseline of !10 g/dL. Two studies of 153 patients with myelodysplastic syndromes were included; their baseline hemoglobin was below 10 g/dL and the patients were given no therapy. There were six studies with 1106 patients with mixed diagnoses. Chemotherapy was given in 21 studies; most of the chemotherapy regimens contained platinum. One study administered radiotherapy alone (nZ50) in solid tumors. There was no anticancer therapy administered in three studies (nZ274). All trials compared EPO treatment initiated at study entry (plus transfusion if necessary) to transfusion of RBCs when the patient’s Hb level fell below a defined threshold or at the discretion of the treating physician. None of the trials compared directly the outcomes of initiating EPO treatment at alternative Hb thresholds. Two different EPO preparations were used. In 15 studies Epoetin alfa, and in eight studies Epoetin beta was given. In the majority of the trials EPO was given in a dosage of 450 U/kg body weight per week at 3 times per week. Several studies compared different EPO dosages, e.g. 450 versus 900 U/kg body weight per week or 300 versus 600 U/kg body weight per week. Overall, the majority of studies were fulltext publications (22), but some studies were included in abstract form (5). For 19 of the 27 included trials covering 86% of the included patients, additional unpublished data were provided by the authors or pharmaceutical companies. First analysis of the data suggests a significantly improved hematological response for the EPO group with respect to decreased need for RBCT and hemoglobin/hematocrit increase. There was some evidence for improved tumor control and overall survival in the EPO group; however, these data have to undergo further statistical analysis.8 More recently published randomized controlled trials contradict the effects observed in our meta-analysis.10,11 In these studies, survival for patients receiving EPO was worse compared to standard care.
CONCLUSIONS In summary, extensive literature search has identified 27 studies with 3287 patients were subjected to a meta-analysis. Additional unreported results for 19 of 27 trials, covering 86% of the patients included in this meta-analysis were kindly submitted by the investigators. Overall, we herewith collected the most comprehensive data on randomized controlled trials on EPO in clinical oncology to date. Data analysis is expected to be finished in spring 2004. An update of the data including studies for the years 2002 and following is planned.
454 J. F. Bohlius et al
REFERENCES *1. Rizzo JD, Lichtin AE, Woolf SH et al. Use of epoetin in patients with cancer: evidence-based clinical practice guidelines of the American Society of Clinical Oncology and the American Society of Hematology. J Clin Oncol 2002; 20: 4083–4107. *2. Aronson N, Seidenfeld J, Piper M, Flamm RC, Hasselblad, V. Use of Erythropoietin in Oncology. Evidence Report/Technology Assessment No. 30. Prepared by the Blue Cross and Blue Shield Association Evidence-based Practice Center under Contract No. 290-97-0015. AHRQ Publ No. 01-E009. Rockville (MD): Agency for Healthcare Research and Quality; June 2001. http://www.ahrq.gov/.2001. Agency for Healthcare Research and Quality, US Department of Health and Human Services, 2101 East Jefferson Street, Rockville, MD 20852. *3. Seidenfeld J, Piper M, Flamm C et al. Epoetin treatment of anemia associated with cancer therapy: a systematic review and meta-analysis of controlled clinical trials. J Natl Cancer Inst 2001; 93: 1204–1214. 4. Langensiepen S, Bohlius J & Seidenfeld J. Erythropoietin for Patients with Malignant Disease (Protocol for a cochrane Review). Chichester, UK: Wiley; 2003. *5. Dickersin K, Scherer R & Lefebvre C:. Identifying relevant studies for systematic reviews. BMJ 1994; 309: 1286–1291. *6. Parmar MK, Torri V& Stewart L:. Extracting summary statistics to perform meta-analyses of the published literature for survival endpoints. Stat Med 1998; 17: 2815–2834. *7. Egger M, Davey SG, Schneider M & Minder C:. Bias in meta-analysis detected by a simple, graphical test. BMJ 1997; 315: 629–634. *8. Bohlius J, Langensiepen S, Schwarzer G, Seidenfeld J, Piper M, Bennett CL et al. Erythropoietin for patients with malignant diseases(Cochrane Review). The Cochrane Library 2005; (1). *9. Moher D, Cook DJ, Eastwood S, Olkin I, Rennie D et al. Improving the quality of reports of meta-analyses of randomised controlled trials: the QUOROM statement. Quality of Reporting of Meta-analyses. Lancet 1999; 354: 1896–1900. *10. Henke M, Laszig R, Ruebe C, Schaefer U, Haase KD, Schilcher B et al. Erythropoietin to treat head and neck cancer patients with anaemia undergoing radiotherapy: randomised, double-blind, placebocontrolled trial. Lancet 2003; 362: 1255–1260. 11. Leyland-Jones B. Breast cancer trial with erythropoietin terminated unexpectedly. Lancet Oncology 2003; 4: 459–460.
8 Effectiveness of erythropoietin in the treatment of patients with malignancies: methods and preliminary results of a Cochrane review Julia F. Bohlius*
MD, MSc PH
Research Fellow
Simon Langensiepen Andreas Engert
MD
MD
Cochrane Hematological Malignancies Group, Department I of Internal Medicine, University of Cologne, Kerpener-Street 9, 50924 Cologne, Germany
Guido Schwarzer MSc, PhD Institute of Medical Biometry and Medical Informatics, Freiburg, Germany Charles L. Bennett MD, PhD VA Chicago Healthcare System, The Feinberg School of Medicine, Northwestern University, Chicago, USA Cancer and cancer therapy-associated anemia may have an impact on tumor response and overall survival. Additionally, anemia represents an important economic factor. Therefore, therapeutic alternatives such as erythropoietin (EPO) and red blood cell transfusions have to be evaluated systematically. The effectiveness of recombinant human EPO to prevent or alleviate anemia in patients with malignant disease was determined. Randomized controlled trials comparing prophylaxis or treatment of anemia with EPO plus red blood cell transfusion (RBCT) or RBCT only in patients with malignant disease undergoing antineoplastic therapy were included. The endpoints needed for RBCT were hematological response (hemoglobin increase of 2 g/dL or hematocrit increase of 6%), tumor response, and overall survival. Medical databases (Cochrane Library, MEDLINE, EMBASE) and conference proceedings were searched (1985–2001). Full-text and abstract publications were included as well as unpublished data. Data extraction and quality
* Corresponding author. Tel.: C49 221 710 770 22; Fax: C49 221 710 770 24. E-mail address: [email protected] (J.F. Bohlius). 1521-6926/$ - see front matter Q 2005 Published by Elsevier Ltd.
450 J. F. Bohlius et al
assessment were done in duplicate. All authors were contacted to obtain missing data. Out of 33 eligible studies, 27 trials with 3,287 randomized patients were included. Key words: cancer therapy; recombinant human erythropoietin; systematic review.
Anemia commonly accompanies malignancies and may result from changes in the production of or response to erythropoietin (EPO) caused by the cancer and/or its treatment. Various clinical trials have shown that EPO increases red blood cell counts and hemoglobin concentrations in a large number of patients with anemia caused by cancer or cancer therapy. A comprehensive evidence-based guideline was published by the American Society of Clinical Oncology and the American Society of Hematology.1 They incorporate the evidence that was published in the systematic review of the Agency for Health Care Research and Quality.2,3 Although this report provides evidence that EPO significantly increases hemoglobin levels in anemic cancer patients and reduces blood transfusion requirements, the critical question whether EPO influences tumor response and prolongs survival time has not been answered yet. Also, uncertainty remained with respect to the optimal hemoglobin level to initiate erythropoietin therapy. A Cochrane Review meta-analysis is being conducted of the trials on EPO used in the treatment or prevention of anemia. The methods and first results from this meta-analysis are presented here. The objectives of the study included the impact of EPO on hematological parameters measured by hematological response and red blood cell transfusions (RBCT). Other assessments are made on the effect EPO had on tumor response, overall survival, and quality of life, as well as the incidence of adverse events.
METHODS This is an international study and the title was registered within the Cochrane Collaboration. A protocol was set up, peer-reviewed, and published in the Cochrane Library.4 A comprehensive search of the literature was done, based on the highly sensitive search strategy for identifying reports of randomized controlled trials developed by Dickersin et al. 5 Medical databases from the Cochrane Library, MEDLINE, EMBASE, Lilacs, and others, as well as registers of ongoing studies, were scanned from the years 1985 to 2001. There were no language restrictions. Conference proceedings from the American Society of Clinical Oncology, American Society of Hematology, and the European Society of Medical Oncology were hand searched. In addition, we contacted experts in the field and pharmaceutical companies who manufacture EPO. In this analysis, all authors were contacted to obtain additional unreported information. The identification of relevant trials, as well as subsequent steps such as critical appraisal, data extraction, and computing, were done in duplicate by two independent reviewers (SL,JB). Study selection Inclusion criteria were met only by randomized controlled studies, with or without blinding, using EPO to treat or prevent anemia in patients with malignant disease. Placebo control, as opposed to no treatment, was not required for inclusion, but was considered in evaluating study quality. Trials in which patients were allocated by a quasi-random method; e.g. date of birth or day of month were excluded because we
Effectiveness of erythropoietin in the treatment of patients with malignancies 451
considered this study design to be of poor quality leading to unreliable results. More than 10 similarly treated evaluable patients in each study arm or relevant stratum were required for inclusion. Ongoing studies and interim analyses were not included. Only participants diagnosed with malignant disease, using clinical and histological/cytological criteria were included regardless of type or stage of the disease or previous therapy. All study participants had to be anemic or at risk for anemia from chemotherapy and/or radiotherapy or the underlying malignant disease. Other reasons for anemia, such as hemolysis, iron deficiency, and occult bleeding, had to be excluded. The identified studies were stratified on patients’ baseline characteristics of hemoglobin below 10, 10–12 g/dL and above 12 g/dL. There were no age restrictions, but at the end of the study only adult patients were included. Epoetin alfa or epoetin beta had to be administered subcutaneously or intravenously at a dose of at least 300 U/kg body weight per week given for at least 4 weeks. Dose adaptation of EPO was allowed, depending on hematological response. Studies on new erythropoiesis stimulating substances like darbepoetin were not considered for this review. Concomitant supportive treatments such as G-CSF or iron supplementation had to be given equally in all study arms. Trials on high-dose myeloablative chemotherapy regimens followed by bone marrow or peripheral blood stem cell transplantation were excluded, as were trials using EPO for short-term preoperative treatment to correct anemia or to support collection of autologous blood prior to cancer surgery. Published, full-text articles and abstracts, as well as unpublished data, are included in the meta-analysis. Quality assessment All trials were analyzed and subjected to a quality assessment that determined whether the trials were randomized and double blind, whether there was concealment of allocation, an intent-to-treat analysis, and documentation of dropouts and withdrawals. In addition we assessed, whether the patient characteristics were similar at baseline in the study groups compared. Studies were excluded from the analysis if they were not truly randomized or inadequately concealed allocation. A data questionnaire was sent out to all the study offices requesting hemoglobin levels, interventions, baseline characteristics, RBCT, mortality dropouts, tumor response, and overall survival information. Sensitivity analyses were performed for baseline hemoglobin, different malignancies, therapies, iron supplementation, granulocyte colony stimulating factor, methodological aspects such as concealment of allocation, masking, completeness of data, and intent-to-treat analyses, the type of publication, and the duration of study. To date, sensitivity analyses have not been completed concerning different dosages of EPO, dosage increases during treatment, and different methods of administration. Data analysis A fixed effect model was assumed in all meta-analyses. For binary data, the relative risk was used as a measure of treatment effect and the Mantel-Haenszel Method was used for pooling. For continuous data, weighted mean differences were calculated. Time to event data such as overall survival were calculated as hazard ratios based on individual patient data. If these data were not available the hazard ratio was calculated from published reports, using methods described in Parmar et al 6 The p-value of the homogeneity test was only used to describe the extent of heterogeneity inherent
452 J. F. Bohlius et al
in a meta-analysis. For primary outcome measures potential causes of heterogeneity were explored by performing sensitivity and subgroup analyses. In meta-analyses with at least four trials, a funnel plot was generated and a linear regression test7 was performed to examine the likely presence of bias in meta-analysis. A p-value less than 0.1 was considered significant for the linear regression test. Analyses were performed using Review Manager, Version 4.2.
RESULTS Overall, 1592 studies were screened and most (1511) were excluded immediately (Figure 1). The group was narrowed to 81 studies, which were subjected to more Potentially relevant RCTs identified and screened for retrieval (n=1,592)
RCTs excluded with reasons (n=1,511)
RCTs retrieved for more detailed evaluation (n=81) RCTsexcluded:did not fulfill the inclusion criteria (n=23)
Potentially appropriate RCTs to be included in the meta-analysis (n=58) RCTs excluded from metaanalysis: ongoing studies (n=24)
RCTs included in the meta-analysis (n=34)
RCTs withdrawn, by outcome, with reasons (n=7)
RCTs with usable information, by outcome (n=27) Figure 1. Improving the quality of reports of meta-analyses of randomized controlled trials: The QUORUM statement flow diagram.9
Effectiveness of erythropoietin in the treatment of patients with malignancies 453
detailed evaluation. Of the 81 studies, 23 did not fulfill the inclusion criteria and 24 are still ongoing and have to be included in the update of this review. Thirty-four were deemed to be eligible; however, of these, another seven studies had to be excluded. Two study reports did not report evaluable data, and we were not able to obtain additional data. Two randomized controlled trials turned out to be of poor quality. Two additional trials were excluded because one turned out to be a duplicate, and another was an interim analysis where a final analysis was never done. Overall, 27 trials with 3287 randomized patients were included. In the majority of studies (16, nZ2615), patients had a baseline mean hemoglobin !10 g/dL. In six studies, 348 patients had hemoglobin levels between 10 and 12 g/dL; in five studies, 324 patients had hemoglobin levels O12 g/dL. All of these five studies included patients with solid tumors. Overall, there were 13 studies of 969 patients with mostly solid tumors. There were six studies of patients with hematological malignancies amounting to 1059 patients, all of who had a mean hemoglobin at baseline of !10 g/dL. Two studies of 153 patients with myelodysplastic syndromes were included; their baseline hemoglobin was below 10 g/dL and the patients were given no therapy. There were six studies with 1106 patients with mixed diagnoses. Chemotherapy was given in 21 studies; most of the chemotherapy regimens contained platinum. One study administered radiotherapy alone (nZ50) in solid tumors. There was no anticancer therapy administered in three studies (nZ274). All trials compared EPO treatment initiated at study entry (plus transfusion if necessary) to transfusion of RBCs when the patient’s Hb level fell below a defined threshold or at the discretion of the treating physician. None of the trials compared directly the outcomes of initiating EPO treatment at alternative Hb thresholds. Two different EPO preparations were used. In 15 studies Epoetin alfa, and in eight studies Epoetin beta was given. In the majority of the trials EPO was given in a dosage of 450 U/kg body weight per week at 3 times per week. Several studies compared different EPO dosages, e.g. 450 versus 900 U/kg body weight per week or 300 versus 600 U/kg body weight per week. Overall, the majority of studies were fulltext publications (22), but some studies were included in abstract form (5). For 19 of the 27 included trials covering 86% of the included patients, additional unpublished data were provided by the authors or pharmaceutical companies. First analysis of the data suggests a significantly improved hematological response for the EPO group with respect to decreased need for RBCT and hemoglobin/hematocrit increase. There was some evidence for improved tumor control and overall survival in the EPO group; however, these data have to undergo further statistical analysis.8 More recently published randomized controlled trials contradict the effects observed in our meta-analysis.10,11 In these studies, survival for patients receiving EPO was worse compared to standard care.
CONCLUSIONS In summary, extensive literature search has identified 27 studies with 3287 patients were subjected to a meta-analysis. Additional unreported results for 19 of 27 trials, covering 86% of the patients included in this meta-analysis were kindly submitted by the investigators. Overall, we herewith collected the most comprehensive data on randomized controlled trials on EPO in clinical oncology to date. Data analysis is expected to be finished in spring 2004. An update of the data including studies for the years 2002 and following is planned.
454 J. F. Bohlius et al
REFERENCES *1. Rizzo JD, Lichtin AE, Woolf SH et al. Use of epoetin in patients with cancer: evidence-based clinical practice guidelines of the American Society of Clinical Oncology and the American Society of Hematology. J Clin Oncol 2002; 20: 4083–4107. *2. Aronson N, Seidenfeld J, Piper M, Flamm RC, Hasselblad, V. Use of Erythropoietin in Oncology. Evidence Report/Technology Assessment No. 30. Prepared by the Blue Cross and Blue Shield Association Evidence-based Practice Center under Contract No. 290-97-0015. AHRQ Publ No. 01-E009. Rockville (MD): Agency for Healthcare Research and Quality; June 2001. http://www.ahrq.gov/.2001. Agency for Healthcare Research and Quality, US Department of Health and Human Services, 2101 East Jefferson Street, Rockville, MD 20852. *3. Seidenfeld J, Piper M, Flamm C et al. Epoetin treatment of anemia associated with cancer therapy: a systematic review and meta-analysis of controlled clinical trials. J Natl Cancer Inst 2001; 93: 1204–1214. 4. Langensiepen S, Bohlius J & Seidenfeld J. Erythropoietin for Patients with Malignant Disease (Protocol for a cochrane Review). Chichester, UK: Wiley; 2003. *5. Dickersin K, Scherer R & Lefebvre C:. Identifying relevant studies for systematic reviews. BMJ 1994; 309: 1286–1291. *6. Parmar MK, Torri V& Stewart L:. Extracting summary statistics to perform meta-analyses of the published literature for survival endpoints. Stat Med 1998; 17: 2815–2834. *7. Egger M, Davey SG, Schneider M & Minder C:. Bias in meta-analysis detected by a simple, graphical test. BMJ 1997; 315: 629–634. *8. Bohlius J, Langensiepen S, Schwarzer G, Seidenfeld J, Piper M, Bennett CL et al. Erythropoietin for patients with malignant diseases(Cochrane Review). The Cochrane Library 2005; (1). *9. Moher D, Cook DJ, Eastwood S, Olkin I, Rennie D et al. Improving the quality of reports of meta-analyses of randomised controlled trials: the QUOROM statement. Quality of Reporting of Meta-analyses. Lancet 1999; 354: 1896–1900. *10. Henke M, Laszig R, Ruebe C, Schaefer U, Haase KD, Schilcher B et al. Erythropoietin to treat head and neck cancer patients with anaemia undergoing radiotherapy: randomised, double-blind, placebocontrolled trial. Lancet 2003; 362: 1255–1260. 11. Leyland-Jones B. Breast cancer trial with erythropoietin terminated unexpectedly. Lancet Oncology 2003; 4: 459–460.