High-dose or dose-intensive cytotoxic chemotherapy often causes myelosuppression and severe neutropenia among cancer patients. Severe neutropenia accompanied by fever, named febrile neutropenia (FN), is the most serious manifestation of neutropenia usually requiring hospitalization and intravenous antibiotics. FN and neutropenia can lead to chemotherapy treatment delays or dose reductions, which potentially compromises the effectiveness of cancer treatment and prospects for a cure. Granulocyte-macrophage (GM) and granulocyte colony-stimulating factors (G-CSFs) are administered during chemotherapy in order to prevent or reduce the incidence or the duration of FN and neutropenia. To assess the effect of prophylactic colony-stimulating factors (CSFs) in reducing the incidence and duration of FN, and all-cause and infection-related mortality during chemotherapy in patients with breast cancer. We searched the Cochrane Central Register of Controlled Trials (CENTRAL), MEDLINE, EMBASE, HEALTHSTAR, International Health Technology Assessment, SOMED, AMED and BIOSIS up to 8 August 2011. We also searched three Chinese databases (VIP, CNKI, CBM), the metaRegister of Controlled Trials, ClinicalTrials.gov, the World Health Organization's International Clinical Trials Registry Platform (WHO ICTRP) and OpenGrey.eu up to August 2011. Randomized controlled trials (RCTs) comparing CSFs (any dose) with placebo or no treatment in patients with breast cancer at any stage, at risk of developing FN while undergoing any type of chemotherapy. We used pooled risk ratios (RR) with 95% confidence intervals (CIs) for binary outcomes. At least two review authors independently extracted data and assessed the risk of bias of the included studies. Trial authors were contacted for further details when information was unclear. We included eight RCTs involving 2156 participants with different stages of breast cancer and chemotherapy regimens. The trials were carried out between 1995 and 2008 and judged as being at least at moderate risk of bias. The strength of the evidence was weak for the majority of outcomes, which was mostly because of the small numbers of evaluable patients, varying definitions, as well as unclear measurements of the trials' outcomes and uncertain influences of supportive treatments on them. In most trials, the chemotherapy regimens had a risk of FN that was below the threshold at which current guidelines recommend routine primary prophylaxis with CSFs. Using CSFs significantly reduced the proportion of patients with FN (RR 0.27; 95% CI 0.11 to 0.70; number needed to treat for an additional beneficial outcome (NNTB) 12) but there was substantial heterogeneity which can be explained by possible differential effects of G-CSFs and GM-CSFs and different definitions of FN. A significant reduction in early mortality was observed in CSF-treated patients compared to placebo or no treatment (RR 0.32; 95% CI 0.13 to 0.77; NNTB 79). This finding was based on 23 fatal events in 2143 patients; wherein 19 of these 23 events occurred in one study and 17 events were attributed to progression of the disease by the study authors. For infection-related mortality, there were no significant differences between CSF and control groups (RR 0.14; 95% CI 0.02 to 1.29). In CSF-treated patients, the risk for hospitalization was significantly reduced (RR 0.14; 95% CI 0.06 to 0.30; NNTB 13), as well as the use of intravenous antibiotics (RR 0.35; 95% CI 0.22 to 0.55; NNTB 18). The risks of severe neutropenia, infection or not maintaining the scheduled dose of chemotherapy did not differ between CSF-treated and control groups. CSFs frequently led to bone pain (RR 5.88; 95% CI 2.54 to 13.60; number needed to treat for an additional harmful outcome (NNTH) 3) and injection-site reactions (RR 3.59; 95% CI 2.33 to 5.53; NNTH 3). In patients with breast cancer receiving chemotherapy, CSFs have shown evidence of benefit in the prevention of FN. There is evidence, though less reliable, of a decrease of all-cause mortality during chemotherapy and a reduced need for hospital care. No reliable evidence was found for a reduction of infection-related mortality, a higher dose intensity of chemotherapy with CSFs or diminished rates of severe neutropenia and infections. The majority of adverse events reported from CSF use were bone pain and injection-site reactions but no conclusions could be drawn regarding late-term side effects.
