Sung, L., Nathan, P.C., Alibhai, S.M.H., Tomlinson, G.A., & Beyene, J. (2007). Meta-analysis: Effect of prophylactic hematopoietic colony-stimulating factors on mortality and outcomes of infection. Annals of Internal Medicine, 147(6), 400–411.

To evaluate prophylactic colony-stimulating factors (CSFs) given concurrently with or after initiation of chemotherapy prior to the development of neutropenia compared with placebo or no therapy in patients with cancer undergoing chemotherapy or hematopoietic stem cell transplantation (HSCT)

The standard Quality of Reporting of Meta-Analyses (QUOROM) guidelines were used to guide the search.

DATABASES USED: Electronic searches of Ovid MEDLINE from 1966–April 24, 2007; EMBASE from 1980–April 26, 2007; and  the Cochrane Central Register of Controlled Trials Register (CENTRAL) through the second quarter of 2006 were performed. The pharmaceutical manufacturers of granulocyte CSFs (G-CSFs) and granulocyte macrophage CSFs (GM-CSFs) also were contacted.

INCLUSION CRITERIA: Patients randomly were assigned to CSFs or to placebo or no therapy. CSFs were given concurrently with or after initiation of chemotherapy or conditioning for stem cell transplantation but before neutropenia developed. Chemotherapy or conditioning regimens or other supportive care was not planned to systematically differ between study groups.

FINAL NUMBER STUDIES INCLUDED = 148 RCTs

TOTAL PATIENTS INCLUDED IN THE REVIEW: 16,839 participants or cycles; 8,474 randomly were assigned to CSF and 8,365 to placebo or no treatment.

KEY SAMPLE CHARACTERISTICS: The RCTs included adult or pediatric patients with cancer undergoing chemotherapy or HSCT. The results were analyzed at the study level, not at the patient level.

Compared with the control, prophylactic CSFs did not significantly affect

• Overall all-cause mortality (7.6 % rate in the CSF group and 8.0% in the control group)
• Risk for infection-related mortality (3.1% rate in the CSF group and 3.8% in the control group).

Compared with the control, prophylactic CSFs significantly reduced

• Documented infections by 15%
• Microbiologically documented infections by 14%
• Clinically documented infections by 25%
• Episodes of febrile neutropenia by 29%
• Duration of febrile neutropenia by a mean difference of 1.38 days
• Duration of fever by a mean difference of 0.45 days
• Time to absolute neutrophil count (ANC) of 500 cells/mcL or more by a mean difference of 3.79 days
• Time to ANC of 1,000 cells/mcL or more by a mean difference of 5.03 days
• Duration of parenteral antibiotic therapy by a mean difference of 1.81 days
• Duration of hospitalization by a mean difference of 2.41 days.

The median rate of febrile neutropenia in the placebo groups was 44.2% versus 25.3% in the CSF groups.

The use of G-CSFs had a greater effect than the use of GM-CSFs on reducing documented infections and febrile neutropenia, but all-cause mortality and infection-related mortality did not differ.