Abstract

Using data submitted to the EBMT registry, we analyzed outcome on 344 patients (pts) who had received donor lymphocyte infusions (DLI) for relapse after allogeneic hematopoietic stem cell transplantation (HSCT) for chronic myeloid leukemia (CML) in 31 centers. 113/344 pts (33%) developed acute graft-versus-host disease (aGVHD) a median of 50 days post DLI (max grade: I=42, II=30, III=31, IV=6)(60% grade II–IV). Organs involved (%): skin (88), liver (42), gut (30). Median age was 38 (4–59), 58% pts were male, 62 transplants were HLA-identical sibling and 51 unrelated. 74 were T-cell depleted, 92 transplanted in CP1, 21 beyond CP1. Relapse was molecular in 19 pts, cytogenetic in 31, hematological in 49, accelerated or blastic in 12. Median initial cell dose was 107CD3+ cells/kg (0.01–32), median number of DLI was 1 (1–10). aGvHD was treated with prednisone in 92% of pts, CSA in 52 %, ATG and monoclonal antibodies in 2% and other in 19%. aGVHD resolved in 53% of the pts within a median of 63 d (7–546).

82/344 pts (24%) had chronic GVHD (cGVHD)(30 limited, 50 extensive, 2 not specified), of those 46 (56%) following aGVHD post DLI. Organs involved (%): skin (75), liver (35), lungs (13), mouth (43), eyes (22) and gut (5). Median age was 35 (6–58), 51% were male, stem cell source was PB in 15% and marrow in 85%, 43 underwent HLA-identical sibling HSCT and 39 unrelated donor HSCT. Forty-three were T-cell depleted, 66 transplanted in CP1, 16 beyond CP1. Relapse was molecular in 21 pts, cytogenetic in 29, hematological in 22, accelerated or blastic in 7. Median initial cell dose was 107 CD3+ cells/kg (0.05–40), median number of DLI was 1 (1–7). 61 pts are alive with a median follow-up of 50 mth. Treatment was with steroids in 83% of pts, CSA in 58 %, MMF in 20%, thalidomide in 15%, photopheresis in 15%, PUVA in 10% and other in 17%. cGVHD resolved in 39% of the pts within a median of 354 d (44–1588).

The estimated 5-y OS post-DLI was significantly lower in pts who developed aGVHD post-DLI, 61 ± 10% vs 74 ± 7% in the one that did not, p=0.007 and also a tendency to have a lower 5-y EFS, 58 ± 10% vs 65 ± 7%, p=0.19. Median duration of response to DLI in aGVHD pts was 4 y. aGVHD post-DLI did not influence the relapse rate (5 ± 5% vs 6 ± 5% in the absence of aGVHD). 5-y DLI related mortality was significantly higher in aGVHD pts, 31 ± 8% vs 4 ± 4%, p<0.00001. On the other hand, pts that developed cGVHD post-DLI had a tendency to have a better 5-y OS and EFS, 74 ± 11% and 71 ± 11% respectively vs 69 ± 6% and 62 ± 7% in those that did not, p=0.32 and 0.09. This was related to a tendency to lower incidence of relapse, 2 ± 3% in pts with cGVHD vs 9 ± 6% without, p=0.2. DLI related mortality was not different, 11 ± 8% vs 10 ± 5%, p=0.77.

aGVHD post-DLI for CML relapse is mainly of advanced stage and negatively influence OS and EFS with a higher DLI related mortality. cGVHD post-DLI is mainly extensive, but pts with cGHVD tend to have better outcome with better 5-y OS, EFS and less relapse than those without, although this was not statistically significant.

Author notes

Corresponding author