Transfusion-dependency is associated with poor prognosis in patients with MDS although the causal link for such association is disputed. This study tests thee hypotheses on the association between transfusion burden and prognosis in the MDS: (1) the cumulative transfusion burden is a confounder merely reflecting the time elapsed from diagnosis; (2) it is a surrogate for higher transfusion intensity, which would reflect a more severe disease; and (3) it is the total amount of transfused RBC units that influences on prognosis. We studied 191 transfusion-dependent patients with MDS or chronic myelomonocytic leukemia. Transfusion intensity was calculated at the time of each transfusion as the yearly-equivalent number of RBC units. The main outcome was acute leukemia-free survival from first transfusion. Median transfusion burden was 30 (range: 4–330) RBC units and 112 patients received ≥25 units after a median of 9 months from first transfusion. In nested Cox models, having received ≥25 RBC units had a significant effect on survival (P < 0.001) that was not abrogated by including follow-up ≥9 months as a time-dependent covariate. Including transfusion intensity in the model had a significant effect on leukemia-free survival (P < 0.001) and cancelled the prognostic value of having received ≥25 RBC units. In conclusion, transfusion intensity, instead of the cumulative transfusion burden, is the transfusion-related variable really influencing on the prognosis of patients with transfusion-dependent MDS. Am. J. Hematol. 86:245–250, 2011. © 2011 Wiley-Liss, Inc.