Five RCTs (n=167) were included in the review.
The studies were considered to be of reasonable quality (mean Jadad score 3.4, range: 1 to 5), with 4 studies scoring 3 or more.
No statistically significant differences between G-CSF and control were found for clinical resolution of infection, healing of wounds or the rates of wound healing.
Patients receiving G-CSF had significantly fewer lower extremity amputations (RR 0.41, 95% CI: 0.17, 0.95, P=0.38; NNT 8.60) than the control and significantly fewer invasive interventions (RR 0.38, 95% CI: 0.20, 0.69, P=0.002; NNT 4.5). A trend towards reduced hospital stay with G-CSF compared with control was also found (WMD -0.36 days, 95% CI: -1.39, 0.67, P=0.49). No statistically significant heterogeneity for any of the above meta-analyses was shown. Patients receiving G-CSF had significantly higher leukocyte counts during therapy than controls (WMD 25.24E9/L , 95% CI: 9.57, 40.92, P=0.002). Statistically significant heterogeneity was found. None of the studies reported any significant adverse effects with G-CSF.
There was no evidence suggesting publication bias from the funnel plot, Begg and Mazumdar's test (adjusted Kendall's score P=1; continuity corrected P=1) or Egger's test (P=0.79).