Use of granulocyte colony-stimulating factor (G-CSF) and outcome in patients with non-chemotherapy agranulocytosis

The use of granulocyte colony-stimulating factor (G-CSF) in the treatment of non-chemotherapy drug- induced agranulocytosis is controversial. We aimed at assessing the effect of G-CSF on the duration of agranulocytosis.

To assess the effect of G-CSF on the duration of agranulocytosis, a Cox proportional hazard model with an estimated propensity score covariate adjusting for several prognostic factors was used.

One hundred and forty-five episodes of agranulocytosis were prospectively collected from January 1994 to December 2000 in Barcelona (Spain). No differences were found in the case-fatality rate between treated (9 of 101, 8.9%) and not treated (5 of 44, 11.4%) patients. The median time to reach a neutrophil count > or =1.0 x 10(9)/L was 5 days (95%CI 5-6) in patients treated with G-CSF compared to 7 days (95%CI 6-8) in those not treated, with a hazard ratio of 1.58 (95% CI 1.1-2.3).

G-CSF shortens time to recovery in patients with agranulocytosis. However, as an effect on case-fatality has not been recorded, and data on cost-effectiveness are lacking, it would be wise to restrict its use to high-risk patients.