G-CSF and Leukemia Treatment: The Potential for Bad, the Observation of Good


Synopsis: Filgrastim (rhuG-CSF) has proven to be a useful adjunct to chemotherapy for a variety of solid tumors because it accelerates marrow recovery after intensive treatments. There has been some controversy over its use in hematological malignancies, particularly myeloid leukemias, because of the theoretical concern that malignant cells would be stimulated to grow. In this report, those concerns are laid to rest. In a multicenter trial involving 521 newly diagnosed acute myelogenous leukemia patients, treatment with filgrastim after both induction and consolidation chemotherapy resulted in shorter periods of neutropenia, less infection, and shorter hospitalization without any evidence for adverse effect on remission or survival. Although the theoretical concerns still exist, it appears that judicious use of this particular factor is both safe and of clinical benefit in the treatment of acute leukemia.

Source: Heil G, et al. Blood 1997:90;4710-4718.

The use of marrow stimulating factors has been a boon to chemotherapists, particularly those treating solid tumors. The experience has indicated that the enhanced myelopoiesis allows high doses of chemotherapy to be delivered with shortened periods of neutropenia, fewer antibiotics, and less need for hospitalization.1 However, there has been concern about using these agents for primary marrow disorders, particularly acute myelogenous leukemia. The concerns are theoretical and based upon the demonstration of G-CSF receptors on myeloid progenitor cells and the enhanced growth of leukemic blasts in culture.2 Fortunately, we now have a large, randomized trial to debunk these concerns.

This was a multi-center effort of the International Acute Myeloid Leukemia Study Group. A total of 521 consecutive patients with AML were randomized to receive Filgrastim (5 mcg/kg/d subcutaneously) or placebo after induction and consolidation until the absolute neutrophil count was greater than 1000 on each of three successive days. The remission rate was 68%, and the median, disease-free survival was 10 months for the entire group. There was no difference in these parameters between the filgrastim and placebo groups. However, patients receiving filgrastim experienced neutrophil recovery five days earlier after induction than those receiving placebo. Also, there were filgrastim-associated reductions in the duration of fever (7 vs 8.5 days), parenteral antibiotic use (15 vs 18.5 days), and hospitalization (20 vs 25 days). Similar reductions were seen after the second induction course (if needed) and the consolidation courses (the second of which was with high-dose cytosine arabinoside and daunorubicin). Furthermore, significantly fewer patients in the filgrastim group required antifungal therapy after induction.


This clinical trial took courage to conduct as an investigator and to participate in as a study subject. There was the distinct possibility that the proposed intervention could have had a stimulatory effect on those leukemic cells with appropriate receptors and that G-CSF-treated patients might have fared more poorly than the placebo control. Indeed, there have been a number of studies in which this or a similar growth factor was used in leukemia trials. However, the majority of these were confined to especially high-risk patients, such as the elderly, or patients with relapsed or refractory leukemia, and the results have been inconclusive. For example, in one trial with 102 patients using recombinant human GM-CSF, the remission rate was significantly lower in those receiving growth factor.3 In contrast, in 173 patients receiving recombinant human G-CSF reported by Dombret et al,4 the remission rate was higher in those receiving growth factor. In the other trials, however, remission rates were no different in the treatment and control groups.

The current study was remarkable because it included all adult patients with de novo leukemia. It was a multicenter, prospective, randomized, double-blinded trial and included growth factor treatment after both induction and consolidation courses. Despite the theoretical concerns about stimulating leukemic cell proliferation resulting in more aggressive tumor growth, the accumulated prior experience and the current trial support the conclusion that it is safe to treat leukemic patients with G-CSF. Remissions occurred with equal frequency in both treatment groups, and median survival was also the same.

In addition to being safe, this trial also established efficacy of filgrastim for parameters relevant to leukemia management. Filgrastim-treated patients had shorter periods of neutropenia, fewer infections, fewer serious infections, less parenteral antibiotics, and shorter in-hospital stays. The series was large enough that these parameters reached a level of statistical significance. Nevertheless, the impact was not huge, and, despite the fewer and less severe infections, therapy-induced neutropenia was prominent in both groups, and there were infectious deaths in both groups. Thus, although a number of parameters were influenced, and these are clearly important, it is a bit disappointing that overall survival was not improved. Perhaps this will come when we learn how much factor to use, when to use it, and whether or not a combination of factors, perhaps including a megakaryocyte stimulating factor, will enhance remission and allow the safe administration of higher doses of chemotherapy in the absence of marrow or stem cell transplant.


1. Crawford J, et al. N Engl J Med 1991;325:164.

2. Lowenberg B, Touw IP. Blood 1993;81:281.

3. Zittoun R, et al. J Clin Oncol 1996;14:2150.

4. Dombret H, et al. N Engl J Med 1995;332:1678.

Regarding the use of rHuG-CSF (filgrastim) in patients with acute myelogenous leukemia, which of the following statements is true?

a. Myeloid progenitor cells and leukemic blasts have been shown to have receptors specific for this agent, raising the concern that treatment might stimulate the leukemic clone.

b. Treatment in a large series of patients did not adversely affect remission rate or survival.

c. Treatment in a large series of patients showed efficacy in that there were fewer infections and shorter hospital stays in the treated group.

d. All of the above.

e. None of the above.