This review concluded, in patients with acute myeloid leukaemia, prophylactic colony-stimulating factor treatment reduced the time to neutrophil recovery, hospitalisation duration and antibiotic treatment duration, but not documented infection incidence; administration during chemotherapy did not affect complete remission or survival rates. Given significant differences between included trials and their unclear quality, the reliability of the authors' conclusions is unclear.

To evaluate the impact of colony-stimulating factors during or after chemotherapy in patients with acute myeloid leukaemia.

MEDLINE, Scopus and the Cochrane Database of Systematic Reviews were searched up to August 2009 for articles in any language. Search terms were reported. Bibliographies of included articles and relevant published reports were handsearched. Researchers in the field were contacted for further references.

Randomised controlled trials (RCTs) that compared primary granulocyte-colony stimulating factor or granulocyte-macrophage colony-stimulating factor prophylaxis with placebo or untreated control group in previously untreated adults with acute myeloid leukaemia were eligible for inclusion. Eligible patients were receiving induction or consolidation chemotherapy. RCTs that compared randomly assigned colony-stimulating factor treatment concurrent with induction chemotherapy versus placebo or no treatment in adult acute myeloid leukaemia patients were also eligible for inclusion. Trials where prophylactic antibiotics were administered were included if they were allowed equally in both trial arms. Economic analyses or studies administering granulocyte-colony stimulating factor for established neutropenia or febrile neutropenia were excluded.

Included prophylactic trials evaluated granulocyte-macrophage colony-stimulating factor or granulocyte-colony stimulating factor (Lenograstim or Filgrastim) in varying doses administered during induction, consolidation, or combined induction and consolidation phases. In trials of granulocyte-colony stimulating factor or granulocyte-macrophage colony-stimulating factor concurrent with chemotherapy, treatment started from two days before to half a day after chemotherapy began. The number of cycles with colony-stimulating factors ranged from one to 10. In over half of the trials, colony-stimulating factors were continued after chemotherapy.

Outcomes reported were time to neutrophil recovery, incidence of documented infections, length of hospitalisation, duration of intravenous antibiotic treatment, remission rates, event-free survival, disease-free survival and overall survival.

The authors did not state how many reviewers performed the study selection.

The authors did not appear to have formally assessed trial quality.

Time to event data were extracted and used to calculate hazard ratios (HRs) or number of events. Dichotomous data were calculated as odds ratios (ORs). Peto's method was used to calculate the corresponding point and variance estimators. For continuous data, the mean difference and sampling variance were calculated. Where the mean was not reported, the median was used.

Two reviewers independently extracted the data for review.

For continuous data, studies were combined using the mean difference (MD) with 95% confidence intervals (CIs). For dichotomous data, odds ratios with 95% confidence intervals were calculated. Fixed-effect models were used. Statistical heterogeneity was assessed using the Q statistic. Summary statistics were calculated separately for trials of prophylactic colony-stimulating factor use and concurrent colony-stimulating factor use. Where a trial was believed to contribute significantly to study heterogeneity, this was excluded.

Publication bias was assessed using Egger's funnel plot test.

Twenty-five RCTs were included in the review (n=7,510 patients). Eleven trials were of prophylactic colony-stimulating factor use (n=2,992), eleven trials were of concurrent colony-stimulating factors use with chemotherapy (n=3,441), and three trials evaluated both prophylactic and concurrent colony-stimulating factors use (n=1,077).

Prophylactic use of colony-stimulating factors: Patients treated with granulocyte or granulocyte-macrophage colony-stimulating factors had significantly shorter time to neutrophil recovery (mean difference -4.13 days, 95% CI -4.23 to -4.04; 11 RCTs; n=3,782), significantly less incidence of febrile neutropenia (OR 0.7, 95% CIs not reported; four RCTs; n=944), significantly less time in hospital (mean difference -2.06 days; 95% CI -2.36 to -1.76; six RCTs; n=2,277), and significantly shorter duration of antibiotic treatment (mean difference -0.88 days; 95% CI -1.24 to -0.52; five RCTs; n=1,803). There was evidence of significant statistical heterogeneity for time to neutrophil recovery and duration of antibiotic treatment (p<0.001), but not for duration of hospital stay. There was no significant difference between colony-stimulating factor-treated and control groups in incidence of documented infections, complete remission rates, event-free survival, disease-free survival, and overall survival. There was significant statistical heterogeneity for the outcome of complete remission (p<0.001).

Concurrent use of colony-stimulating factors with chemotherapy: There was no difference between colony-stimulating factor-treated and control groups in complete remission, disease-free survival and event-free survival (combined analysis), or overall survival. There was no evidence of statistical heterogeneity for any of these outcomes.

Prophylactic colony-stimulating factors administration reduced the time to neutrophil recovery, duration of hospitalisation and duration of intravenous antibiotic treatment, but not the incidence of documented infections. Colony-stimulating factor use did not affect complete remission or survival rates.

The review addressed a clear question with well-defined inclusion criteria for study design, intervention and participants. Three relevant databases were searched. Appropriate steps were taken to minimise the risk of language bias. Some attempts appear to have been made to identify unpublished material. Although the reviewers reported that they assessed publication bias, no results appear to have been reported, so publication bias could not be ruled out. Suitable steps were taken in the data extraction process to minimise reviewer error and bias, but it was unclear whether similar steps were taken at the study selection stage.

The quality of included trials did not appear to have been assessed, so the reliability of the findings was unclear. In light of the significant levels of statistical heterogeneity for some outcomes, the appropriateness of the synthesis was unclear. Potential sources of heterogeneity did not appear to have been assessed.

Given the unclear quality of the included trials and the presence of significant statistical heterogeneity, the reliability of the authors' conclusions is unclear

Practice: The authors did not state any implications for practice.

Research: The authors stated that retrospective analyses of past clinical trials should be conducted to identify which molecularly defined subgroups of patients may respond to priming strategies. Further research was needed into the use of colony-stimulating factors during treatment consolidation. Economic analyses of colony-stimulating factor use in patients with acute myeloid leukaemia were needed.

Deutsche Krebshilfe; H. W. & J. Hector Stiftung; Deutsche-Jose-Carreras Leukamie-Stiftung e.V..

This is a critical abstract of a systematic review that meets the criteria for inclusion on DARE. Each critical abstract contains a brief summary of the review methods, results and conclusions followed by a detailed critical assessment on the reliability of the review and the conclusions drawn.