An in vitro skin explant assay as a predictive assay for graft-versus-hostdisease in a cohort of pediatric transplants

Severe acute graft-versus-host disease (GvHD) remains a serious complicatio n of allogeneic stem cell transplantation. An in vitro skin explant assay w as used to predict the occurence and severity of acute GvHD in a cohort of 30 pediatric patients undergoing human leucocyte antigen (HLA)-matched sibl ing transplants (20 patients) and matched or one antigen mismatched unrelat ed donor transplants (10 patients). In the cohort of HLA-matched sibling tr ansplants. the result appeared to reflect the degree of GVHD prophylaxis. T he skin explant assay correlated with GVHD outcome in 12 of 20 children, bu t this did not reach statistical significance (chi-square 0.95, d.f. =1, p= 0.32). These results support previous observations. In this present cohort, patients were treated either with cyclosporin A (CsA) monotherapy(n = 7)or with CsA plus additional methotrexate (MTX) (n=13). We have previously dem onstrated that the skin explant assay was not as predictive in patients rec eiving CsA plus additional MTX compared to cohorts treated with CsA alone. In the group of patients treated with CsA alone, four of five patients (80% ) predicted to develop GvHD developed acute GvHD of grade II or above: of t wo patients predicted to develop only grade 0-I GvHD, one patient developed no GvHD and the other grade II GvHD. In the CsA plus MTX group, nine patie nts were predicted to develop GvHD. Five of nine (55%) developed acute GVHD of grade II or above. while three of four with grade 0 or I skin explant a ssay results developed only grade 0-I GVHD. In a cohort of 10 patients who received unrelated donor transplants, the skin explant assay correlated wit h GvHD outcome in all 10 patients (Fisher's exact test p=0.008). Hence. the skin explant assay is a pretransplant in vitro GVHD predictive test that p redicts the occurence and severity of acute GvHD in pediatric unrelated don or transplants and to varying degrees, depending on the GvHD prophylaxis pr otocols. in HLA-matched sibling cohorts.