PATIENTS AND METHODS

RESULTS

At the time of relapse, seven patients were classified as having a hematologic relapse, three as having a morphologic relapse, and four as having a cytogenetic relapse (Table 1). Three patients experienced progression of disease necessitating additional therapy before completion of filgrastim. All three of these patients had either AML or myelodysplastic syndrome, with hematologic relapses within 100 days of transplant.

Eleven patients completed their assigned therapy. There was no significant toxicity noted with filgrastim administration. Maximal WBC counts were all noted to be less than 50,000/mL. Six of the 11 patients who completed filgrastim therapy achieved a complete hematologic, morphologic, and cytogenetic response. Four of the six complete responses were observed on the day-28 evaluation after completion of filgrastim. The other two patients both had nearly complete responses at day 28 of filgrastim treatment and complete responses by day 100 after filgrastim treatment. On an intent-to-treat basis, the overall complete remission rate was 43% for all 14 patients. When response was evaluated based on relapse classification, three of four patients with a cytogenetic relapse, two of three patients with a morphologic relapse, and one of four patients with a hematologic relapse achieved a complete response. Four of six responders had CML; three of these had a cytogenetic relapse. Of the other two responders, one had AML with a morphologic relapse, and one had CLL with hematologic relapse. Two of the responding patients had not received any immunosuppressive agents in excess of 3 months before the initiation of filgrastim.

At the time of evaluation for response, all assessable patients’ bone marrows had cellularities in excess of 10%. Most patients were observed to have an increase or stabilization marrow cellularity. Subsequent monitoring failed to demonstrate any evidence of marrow aplasia.

Nine of 13 assessable patients developed either acute or chronic GVHD. Two patients had chronic GVHD before filgrastim administration. Six patients subsequently developed chronic GVHD within 1 to 6 months after filgrastim administration. Three patients developed limited chronic GVHD, and three developed extensive chronic GVHD. Chronic GVHD was observed in five of the six responders. Extensive chronic GVHD was generally limited to the skin with sclerodermatous changes. One patient had extensive chronic GVHD involving the mouth and tongue. This latter patient received a transplant from a five-of-six-antigen HLA-matched unrelated donor. The development of chronic GVHD was noted after the determination of response in all five patients.

At the time of this analysis, three of the six responders had relapsed after an initial response to filgrastim. The duration of their responses was greater than 12 months in all three patients. Two of these relapses were chloromas, one in the breasts and one in the left flank. The chloromas were not present at the time of their original relapse. The third relapse was in the patient with CLL, who had an isolated recurrence in an axillary lymph node. All three of these patients’ bone marrows were normal with all cells of donor origin. At a median follow-up of 14.5 months, the estimated event-free survival was 43% (Fig 1). The estimated overall survival was 73%; the median survival has not yet been reached.

DISCUSSION

The current options for patients who relapse after an alloSCT include a second transplant, removal of immunosuppression, interferon in selected cases of CML, or a DLI.^7,8 DLI has been successfully used in both children and adults.⁵ However, DLI has severe limitations, including a decreased activity in patients with acute leukemias or bulky disease (eg, non-Hodgkin’s lymphoma), a delayed response (weeks to months), and potentially severe toxicity in the form of GVHD, which is observed in more than 50% of patients.^5,9,10

Reports of the use of filgrastim for relapse have been anecdotal with the exception of the report by Giralt et al.^6,7,11 Filgrastim for relapse offers potential advantages over DLI. First, it is well tolerated; the primary acute side effects being those typically associated with filgrastim administration, including mild arthralgias and low-grade fevers. It may, however, be associated with the development of GVHD, both acute and chronic, which was observed in approximately 54% of patients in this analysis. Second, the response is relatively rapid; by a week after the completion of therapy (ie, 28 days), responses could be reasonably assessed. This may provide adequate time to make arrangements to secure donor leukocytes in the event that the donor is not readily available. Most importantly, marrow aplasia, which is observed in a significant number of patients receiving DLI, was not observed with filgrastim.^5,8

Similar to results with DLI, filgrastim administration had limited efficacy in patients with overt hematologic relapse of their acute leukemias. In this situation, cytoreduction of the leukemia may be necessary, either with conventional chemotherapy or a nonmyeloablative preparative regimen, which seems to have success in patients with sensitive disease.¹⁰

The primary question is why would this therapy work at all? There are several potential explanations. The first is that filgrastim had little or no effect, and the observed responses were a result of the removal of immunosuppression, permitting a graft-versus-leukemia effect.⁸ This may not provide an adequate explanation, however, because two of the observed responses occurred in patients receiving no immunosuppression at the time of their relapse. This clinical situation is not significantly different than for DLI where immunosuppression is also routinely withdrawn.^4,11 A second possible explanation is that the filgrastim provides a competitive advantage for normal hematopoiesis, potentially increasing the numbers of effector cells.¹² Another related explanation is that the increased hematopoiesis after filgrastim dilutes the leukemic cells present in the marrow, and the observed results were actually a false-positive. This seems highly unlikely given the high sensitivity of FISH and the continued maintenance of response long after the acute effects of filgrastim had subsided. A last possible explanation for the efficacy of filgrastim is a direct effect on the immune system. The administration of filgrastim is known to cause a shift of T cells from Type 1 to Type 2 cytokine pattern and a reduction of cytotoxic T-cell activity.¹³ These effects would seem contradictory to an enhanced antileukemic response; however, it has been demonstrated that cytotoxic T-cell activity is preserved through a perforin-dependent pathway after filgrastim exposure.¹⁴ Still, the overall effects of filgrastim on the immune system are not well known.

Regardless of the mechanism of action, these data suggest that the administration of filgrastim, either alone or in conjunction with discontinuation of immunosuppression if present, can result in rapid, complete cytogenetic remissions for patients who relapse after alloSCT. Further studies are required to substantiate these results and to provide further understanding of this treatment. However, given the ease of administration, minimal related complications, and the relatively rapid assessment of response, it seems reasonable to consider filgrastim as an initial treatment option for relapse after alloSCT. It should be considered as an adjunct and not a replacement to other treatments, particularly DLI. A practical strategy could include the use of filgrastim as initial management for relapse, permitting time to make arrangements for donor leukocyte collection. In the event there was no response to filgrastim, DLI could then be used with little predicted loss in efficacy. A more interesting alternative would be a direct comparison of filgrastim administration with DLI. The administration of filgrastim seems to be a practical and effective option for cytogenetic or molecular leukemic relapses after alloSCT.