Objectives:

Anemia is relatively common in patients with either hematologic or solid tissue malignancies. When cancer treatment or the disease itself decreases production of or impairs response to endogenous erythropoietin, epoetin treatment may correct the resulting anemia. This systematic review compares outcomes of managing anemia with epoetin (and red blood cell [RBC] transfusion used as necessary) with using RBC transfusion alone. Four groups of patients with malignancy are included: (1) patients with anemia or at risk of anemia resulting primarily from cancer therapy; (2) patients with anemia resulting primarily from their malignant disease and who may also be receiving cancer therapy; and patients who are anemic as a result of bone marrow ablation prior to (3) allogeneic or (4) autologous stem-cell transplantation.

Search Strategy:

The MEDLINE, CancerLit, and EMBASE databases were searched from 1985 through 1998 and Current Contents on Diskette and Medscape Oncology through October 1999 for the terms: erythropoietin (Medical Subject Heading [MeSH^®]); epoetin alfa (MeSH^®); erythropoietin (tw); epoetin (tw); Epogen (tw); Procrit (tw); Eprex (tw); Marogen (tw); Recormon (tw); epo (tw); Anemia/drug therapy (MeSH^®; included all subheadings); Anemia/therapy (MeSH^®; included all subheadings); Anemia/diet therapy (MeSH^®; included all subheadings). The search was then limited to "neoplasms" or "myelodysplastic syndromes" and studies on human subjects. The yield was 2,943 references.

Selection Criteria:

This systematic review is limited to controlled trials comparing the outcomes of managing anemia with and without the use of epoetin in one of the four patient populations of interest. Uncontrolled trials were excluded.

Data Collection and Analysis:

We used a prospectively designed protocol conducted by two independent reviewers, with disagreements resolved by consensus. The meta-analysis used a random effects model to combine data on odds of transfusion in patients with anemia due primarily to cancer therapy.

Main Results:

For patients with anemia resulting primarily from cancer therapy, epoetin reduces the odds of transfusion. The overall number needed to treat (NNT) is 4.4 (95 percent confidence interval [CI], 3.6 to 6.1), which suggests four to five patients must be treated to spare one patient from transfusion. Sensitivity analysis found a smaller magnitude of risk reduction for double-blinded compared with unblinded studies. A large, double-blinded randomized trial, not yet published, found improvement in quality-of-life scores with epoetin. Assessment of the study methodology and clinical significance of the findings awaits publication of the full report. The most robust evidence that epoetin improves outcomes is from trials in patient groups with baseline hemoglobin (Hb) at or below 10 g/dL. The evidence is not adequate to determine whether outcomes are superior when epoetin treatment is initiated at higher hemoglobin thresholds.

As many as one-half of all patients did not achieve a hematologic response to epoetin. Thus, nonresponding patients may account for much of the transfusion use in the epoetin arms of these trials. To achieve the most efficient use of epoetin, more systematic evidence is needed on patient characteristics that predict responsiveness and on early indicators of response.

Anemia primarily a result of malignancy included patients with multiple myeloma, non-Hodgkin's lymphoma, chronic lymphocytic leukemia, and myelodysplastic syndromes. Epoetin increases Hb levels and achieves statistically significant hematologic response rates in these patients. The evidence on transfusion outcomes is sparse but suggests a favorable effect of epoetin. Hematologic response rates appear to be lower for patients with myelodysplastic syndrome; higher doses of epoetin may be necessary to achieve response.

For patients undergoing allogeneic stem-cell transplantation, epoetin decreased time to RBC engraftment by 1 to 2 weeks and may decrease the number of RBC units transfused. No reduction in length of hospitalization was reported. The evidence does not support a beneficial effect of epoetin for patients undergoing autologous stem-cell transplantation.

Conclusions:

For patients undergoing cancer therapy, evidence demonstrates that epoetin reduces transfusion if treatment is initiated when declining Hb levels near 10 g/dL. Randomized controlled trials, adequately powered, are needed to determine whether initiating treatment at higher baseline Hb levels yields additional benefits in reducing transfusion use or improving quality of life.

This review identified common deficiencies in the design and reporting of trials on epoetin. Some methodologic deficiencies may result in overestimation of the effects of epoetin and inadequacy of reporting may limit the ability to interpret and generalize results. Future trials should maintain a higher standard of methodologic quality and completeness of reporting.