The review concluded that parenteral iron significantly reduced the risk of transfusions and increased the chance of haematopoietic response when compared with ESAs alone. The authors' conclusion reflected the results obtained but some caution is warranted due to the uncertain trial quality and possibility of publication bias.

To determine the efficacy and safety of the addition of iron to erythropoiesis-stimulating agents (ESAs) in anaemic cancer patients.

MEDLINE and EMBASE were searched up to January 2011 for published studies. There were no language restrictions. Search terms were reported. Citations from selected articles were checked.

Randomised phase II/III trials that compared parental iron, no iron and oral iron in adult cancer patients with chemotherapy-related anaemia treated with ESAs (epoetins or darbepoetins) were eligible for inclusion. Only patients with active solid tumours or lymphomas were included. Patients with multiple myeloma were allowed. The primary outcome was haematopoietic response (defined as the proportion of patients in each group achieving erythropoietic response). Other endpoints of interest were time to haematopoietic response, transfusion requirements and safety.

Intravenous iron and cumulative dose ranged from 725mg to 3,000mg. ESAs used included epoetin alfa (40,000 U weekly), epoetin beta (30,000 U weekly) and darbepoetin alfa (150μg weekly to 500μg every three weeks); dose adjustments were permitted in some trials. Eligibility criteria for transferring saturation and serum ferritin and haemoglobin varied across trials. Types of parenteral or oral iron varied. Most trials included patients with solid tumours and lymphoma or leukaemia.

Three reviewers independently selected trials for inclusion in the review.

The authors did not state that they formally assessed the quality of the included trials.

Three reviewers independently extracted data on risk ratios (RRs) with associated confidence intervals (CIs) or calculated these using percentage of reported events/number of patients in each trial. Data analyses were performed on an intention-to-treat basis.

Pooled estimates were calculated using a fixed-effect model unless significant heterogeneity was found in which cases the DerSimonian and Laird random-effects model was used. Cochran's Q and Ι² were used to investigate statistical heterogeneity. Three different comparisons were performed: parental iron versus no iron; oral iron versus no iron; and parental iron versus oral iron. Subgroup analyses according to different ESA were performed. Publication bias was assessed with a visual inspection of funnel plots.

Eight RCTs were included in the meta-analyses (1,606 participants): seven phase III trials and one phase II trial. Twenty-two arms were available for comparison.

Haematopoietic response: The risk of obtaining a haematopoietic response was significantly increased with parenteral iron compared with no iron (RR 1.29, 95% CI 1.13 to 1.48; Ι²=63%; seven RCTs). No between-group difference was found for oral iron versus no iron (three RCTs) or for parenteral iron versus oral iron (four RCTs).

When the type of ESA was considered, a greater increase in haematopoietic response was found for epoetin (RR 1.85, 95% CI 1.43 to 2.39; Ι²=32%) than darbepoetin (RR 1.18, 95% CI 1.10 to 1.27; Ι²=6%) with parenteral iron. No significant differences were found with oral iron.

Transfusion rates: The risk of transfusion was significantly reduced with parenteral iron compared with no iron (RR 0.77, 95% CI 0.62 to 0.97, Ι²=0%; seven RCTs).

When type of ESA was considered, a significant reduction was found when parenteral iron was associated with darbepoetin but not with epoetins.

No statistically significant difference of risk of transfusion was found between oral iron and no iron (three RCTs) or between parental iron and oral iron (three RCTs).

Parenteral iron reduced the risk of transfusions by 23% and increased the chance of haematopoietic response by 29% when compared with ESAs alone. Oral iron did not increase haematopoietic response or reduce transfusion rate.

The review question was supported with clear inclusion criteria. Two relevant databases were searched without language restrictions but no attempt was made to locate unpublished trials and it was possible that relevant data were missed. Steps were taken to minimise the likelihood of error and bias during study selection and data extraction.

The synthesis appeared appropriate and some attempt was made to investigate heterogeneity. The authors did not state that they formally assessed the validity of the included trials and this limited interpretation of the results.

The authors' conclusion reflected the results obtained but some caution is warranted due to the uncertain trial quality and possibility of publication bias.

Practice: The authors indicated that intravenous iron was useful when added to ESAs and was likely to improve the proportion of non-responders, quality of life and overall cost. The choice between iron formulations would depend on ESA used, need for quick haematopoietic response, patient tolerance, underlying disease and expected adverse events.

Research: The authors did not state any implications for research.

Not reported.