We hypothesized that severely overweight patients were at an increased risk of transplant-related toxicity and that this might be reflected in increased mortality soon after transplant. All patients transplanted between January 1985 and January 1992 were included in the analysis; data were complete for 2238 (1662 adults, 576 children ( > 18 years). Since we were interested primarily in acute post-transplant toxicity, survival through day 150 was the primary end-point analyzed. Children and adults were analyzed separately since the decision to transplant or not to transplant may differ between the two populations. Log rank statistics and multivariable Cox regression analyses showed that adults at 95% to > 145% of ideal body weight (BW) experienced comparable non-relapse mortality and survival. In contrast, survival among patients at 85-95% or < 85% of ideal BM was significantly worse (P = 0.004 and P = 0.0001, respectively) than in the 95-145% weight category. Results in children were similar except that the differences were less prominent (comparisons of < 85%, 85-95% and > 145% categories vs. 95-145% category showed P values of 0.22, < 0.01 and 0.66, respectively). The observed weight effect was not explainable in terms of the composition of the lower weight categories in regards to nine other known or suspected risk factors. In addition to weight, patient age, diagnosis, type of transplant, year of transplant, graft-versus-host disease (GVHD) prophylaxis and GVHD grades II-IV were also significantly associated with outcome. Thus, the present analysis shows that underweight patients are at an increased risk of death in the early period after transplant.