Since 1989, 35 bone marrow transplantations (BMT) from unrelated bone marrow donors (UBMD) were performed in our Paediatric BMT Unit for haematological or non haematological diseases. For 3 patients (pts), it was the second transplant procedure: 1 autograft, 2 BMT from HLA mismatch (MM) father with graft failure. In absence of HLA identical (Id) UBMD, searches were extended to HLA MM donors; they were selected on the basis of serology for HLA Class I, molecular biology for HLA Class II alleles and Mixed Lymphocyte Culture (MLC) below 20% in GVHD or Rejection sense. Finally, 19/35 BMT were practiced in HLA MM situation. For 3 pts, HLA MM was unknown at time of BMT. Only 3 pts received T-cell depleted marrow. For others, Anti Lymphocyte Globulin (ALG) was added to conditioning regimen with the aim to facilitate engraftment, to obtain a moderate T-cell depletion and to decrease the incidence of GVHD. Rabbit Thymoglobulin Merieux: 20 mg/kg total dose from J-7 to J-1 (4 doses) was used in pre graft in 15 cases, in post graft with increased doses for 2 pts, in pre and post in 2 cases. The final decision to include or not ALG was based on the underlying disease (Aplastic Anaemia), donor/recipient HLA MM pairs whatever the MM degree and high MLC percentage of RRS in rejection sense. HLA MM patients were transplanted without addition of ALG when CML was negative in rejection sense meanly for pts with high risk disease in order to obtain a possible GVL effect. Among the 19 pts who received ALG, 11 were HLA MM (from 1 to 4 MM) and 8 were HLA Id. to their donor; 9/11 had MLC positive results in rejection sense, the 2 others with AA received ALG despite MLC negative. For the 8 pts transplanted in HLA Id. situation, 3 had an MLC positive in rejection sense and 5 received ALG despite MLC negative (protocols). In ALG group, engraftment is non evaluable in 2 cases because of early death. One other pt failed to engraft and died. 16/19 engrafted (84%). 10/19 pts died: 2 VOD, 1 GVH IV, 1 PYO, 1 in aplasia with infection, 2 LMNH, 1 Lyell, 1 acral dermatitis, 1 IP CMV. Among the 9 pts who survive, 8 are in CR with functional graft (42%): 3 HLA MM and 5 HLA id. Considering the group of pts who not received ALG (n=16), 8 were in HLA MM situation, 8 HLA identical. 11/16 engrafted (68%): 2 non evaluable because of early death, 2 graft failure, 1 blastic evolution. 5/16 died: 1 GVH III, 2 blastic, 1 ulcer, 1 infection after rejection. Finally, 11/16 pts are alive but 7/11 CR with functional graft (44%): 4 HLA MM, 4 HLA Id. In this study, pre graft ALG improves engraftment. Post graft ALG increases the risk of LMNH or CMV. The incidence of GVHD≥II is low (5/35 despite HLA-MM). Survival in complete remission is equal in the two groups.