Prognostic factors for acute graft-versus-host disease after donor lymphocyte infusions

Acute graft-versus-host disease (AGVHD) complicates donor lymphocyte infusion (DLI) treatment in 40% to 60% of cases.1,2 Although factors predictive of GVHD after allogeneic stem cell transplantation (SCT) could also be valid after DLI, patients receiving DLI are a select group, having survived the GVHD after transplantation and relapsed. We therefore investigated the role of possible factors in the development of AGVHD after DLI.

Sixty-three consecutive patients with BCR-ABL–positive chronic myeloid leukemia (CML) who relapsed after allogeneic SCT (31 from their respective HLA-identical siblings and 32 from volunteer-matched unrelated donors) were started on treatment with DLI between August 1990 and April 1999. At the time of DLI, 7 patients were in molecular relapse, 15 were in cytogenetic relapse, and 26 were in hematologic relapse (19 in chronic phase, 7 in accelerated phase). Twenty-seven patients received a single infusion of donor lymphocytes (bulk-dose regimen [BDR]), and 36 patients received lymphocytes according to our escalating-dose regimen (EDR).3 Molecular remission was defined as the absence of detectable BCR-ABL transcripts by reverse transcriptase–polymerase chain reaction (RT-PCR) analysis of peripheral blood on 2 consecutive occasions.4 AGVHD was graded according to the Glucksberg criteria.5 Grades II to IV were deemed of clinical significance.

Forty-five patients (71%) treated with DLI achieved molecular remission. Fifteen patients (24%) developed grade II to IV AGVHD, 48 patients (76%) showed no or minimal AGVHD. Results of the univariate analyses are presented in Table 1. No association was found between AGVHD after DLI and AGVHD after transplantation. The vast majority of patients who developed AGVHD following DLI treatment did not have a history of AGVHD after their original transplantation. Recipient-donor sex mismatch, patient-donor CMV seropositivity, and increasing patient age (> 35 years) were found to be significantly associated with AGVHD in the DLI setting. In accord with previous observations,1,2 patients who received T-cell–replete allografts suffered less AGVHD after DLI compared to those who received a T-cell–depleted stem cell preparation. We did not observe any association between achievement of complete remission and development of AGVHD. This finding contrasts with the results reported by other groups, whereby 89% of the responders suffered AGVHD.6 Such a difference may be attributed to the high proportion of patients receiving DLI according to an EDR in our study. In fact, the influence of method of administration was highly significant. In a multivariate logistic analysis, any positive patient-donor CMV serostatus (relative risk [RR] = 19.5,P = .013), BDR (RR = 9.4, P = .008), and T-cell depletion (TCD) at transplantation (RR = 21.9,P = .011) were found to be independent predictors of GVHD after DLI. Although there is more AGVHD associated with patient-donor sex mismatch, in the mismatch-only group there is considerably more AGVHD associated with the use of BDR than with EDR (9 of 16 patients vs 2 of 11 patients, P = .048); this explains why sex mismatch is not significant in the multivariate analysis.

Our findings demonstrate that most of the factors known to be associated with AGVHD after allogeneic SCT are not predictive of AGVHD following DLI. Our data reinforce the notion that DLI can be safe if the effective cell dose is administered late after transplantation3,7; this has important practical implications for the use of DLI in the context of reduced-intensity conditioning SCT.