Kuderer, N.M., Dale, D.C., Crawford, J., & Lyman, G.H. (2007). Impact of primary prophylaxis with granulocyte colony-stimulating factor on febrile neutropenia and mortality in adult cancer patients receiving chemotherapy: A systematic review. Journal of Clinical Oncology, 25, 3158–3167.

To evaluate primary granulocyte colony-stimulating factor (G-CSF) prophylaxis versus a placebo or untreated control group

DATABASES USED: Electronic databases through December 2006: MEDLINE, EMBASE, CANCERLIT, Cochrane Database of Systematic Reviews, Cochrane Central Register of Controlled Trials, Database of Abstracts of Reviews of Effect, and Conference Proceedings (American Society of Clinical Oncology and American Society of Hematology). In addition, references from included articles and relevant published reports were hand searched, and references from leaders in the field were solicited. The literature search had no language restrictions.

INCLUSION CRITERIA:

• Eligible studies were those of adult patients with cancer receiving conventional-dose chemotherapy for solid tumors or malignant lymphoma and randomly assigned to primary G-CSF prophylaxis versus a placebo or untreated control group.
• Primary G-CSF prophylaxis refers to G-CSF administration in the first cycle of chemotherapy before the onset of neutropenia, whereas secondary G-CSF prophylaxis is defined as G-CSF prophylaxis started in the chemotherapy cycle after the first episode of febrile neutropenia (FN). Patients may have received prophylactic antibiotics as long as they were permitted equally in both study arms.
• G-CSF must have been administered continuously until neutrophil recovery. Previous studies and guidelines concluded that myeloid growth factors are less effective if administered on the same day as chemotherapy, delayed more than four days following chemotherapy, or delayed until the onset of neutropenia. Therefore, studies were eligible only if the initiation of G-CSF was one to three days after the completion of myelosuppressive chemotherapy in each cycle.
• Studies in which control patients received secondary G-CSF prophylaxis after the first cycle with the same myeloid growth factor were permitted.

EXCLUSION CRITERIA:

• Studies were excluded if they used granulocyte-macrophage colony stimulating factor, were studies of children or patients with leukemia or multiple myeloma, included bone marrow or peripheral blood stem cell transplantation, or represented an economic analysis.
• Studies also were excluded if G-CSF was administered for established neutropenia or FN. Studies of patients receiving dose-dense or dose-escalation chemotherapy were excluded, as were studies that allowed differing drugs, doses, or schedules of chemotherapy or G-CSF in the different study arms.

FINAL NUMBER STUDIES INCLUDED = 17 RCTs of primary prophylactic G-CSF

TOTAL PATIENTS INCLUDED IN REVIEW = 3,493 patients

KEY SAMPLE CHARACTERISTICS:

• Ten studies (59%) used filgrastim, six (35%) used lenograstim, and one (6%) used pegfilgrastim.
• The cancer types consisted of solid tumors in 11 trials (65%) and aggressive non-Hodgkin lymphoma in six RCTs (35%).
• Four RCTs (24%) were limited to elderly patients with lymphoma.
• Eight studies used placebo control.
• At least three trials permitted secondary prophylaxis with G-CSF in control patients, with only two reports explicitly prohibiting this use of G-CSF. Eleven trials failed to specify whether secondary prophylaxis with G-CSF was allowed in controls.
• Prophylactic antibiotics were used in three trials, prohibited in five trials, and not specified in eight trials.

The occurrence of FN was reported as an outcome in 15 trials with 3,182 patients.

• G-CSF reduced the risk of FN by 46%.
• FN occurred one or more times in 39.5% of controls and 22.4% of G-CSF patients, resulting in a weighted summary RR of 0.54 (95% CI, 0.43 to 0.67; P < .0001).
• Filgrastim (RR = 0.61; 95% CI, 0.53 to 0.72) and lenograstim (RR = 0.62; 95% CI, 0.44 to 0.88) had similar efficacy.
• The single pegfilgrastim study demonstrated significantly greater efficacy (RR = 0.08; 95% CI, 0.03 to 0.18) compared with filgrastim and lenograstim (P < .0001).

Infection-related mortality was reported as an outcome in 12 trials including 1,454 control patients and 1,463 patients receiving G-CSF.

• Overall, infection-related mortality was observed in 2.8% of controls and 1.5% of G-CSF patients, for a weighted summary RR of 0.55 (95% CI, 0.34 to 0.90; P = .018).
• Reductions in infection-related mortality with G-CSF were observed among studies of filgrastim (RR = 0.53; 95% CI, 0.30 to 0.92; P = .024). Statistical analysis of the subset of patients who received lenograstim or pegfilgrastim was not possible because the sample size was too low to detect differences.

Early mortality was reported in 13 trials with 3,122 patients.

• Overall, early mortality was observed in 5.7% of control patients and 3.4% of G-CSF patients, resulting in a weighted summary RR of 0.60 (95% CI, 0.43 to 0.83; P = .002).
• Reductions in early mortality with G-CSF were observed among studies of filgrastim (RR = 0.60; 95% CI, 0.41 to 0.89; P = .010) and pegfilgrastim (RR = 0.36; 95% CI, 0.13 to 0.99; P = .047) but not lenograstim (RR = 0.84; 95% CI, 0.38 to 1.83; P = .657).

Relative dose intensity (RDI) was reported as an outcome in 10 trials. The average RDI among control patients in these studies ranged from 71.0%–95.0%, with a mean RDI of 86.7% (median RDI: 88.5%).

• Among G-CSF-treated patients, the average RDI ranged from 91.0%–99.0%, with a mean RDI of 95.1% (median RDI: 95.5%). RDI differences between study arms ranged from 2.8%–20.0%, with average differences of 8.4% (P = .001). None of the 10 G-CSF treatment arms reported an average RDI of less than 90%, whereas 6 of 10 control groups reported a mean RDI of less than 90%, with four control arms averaging an RDI of 85% or less.

Bone or musculoskeletal pain during the course of chemotherapy was reported as an outcome in 14 trials including 3,029 patients.

• Bone pain was reported in 10.4% of controls and 19.6% of G-CSF-treated patients, for a weighted summary RR of 4.023 (95% CI, 2.156 to 7.52; P < .0001).