ABSTRACT: Allogeneic hematopoietic stem cell transplantation (allo-HSCT) is still considered a definitive curative modality for refractory or relapsed non-Hodgkin's lymphoma (NHL). However, transplant-related morbidity and mortality remain a considerable challenge. The graft-versus-host disease (GVHD)-free with relapse-free survival (GRFS) rate and GRFS-related prognostic factors have not been fully examined for NHL alone. We evaluated 104 consecutive patients with refractory or relapsed aggressive NHL receiving allo-HSCT at a single institution. With a median follow-up of 31.5 months, the estimated 3-year overall survival (OS), disease-free survival (DFS), the cumulative incidence rates of relapse, and non-relapse mortality were 45.9%, 45.9%, 36.0%, and 17.0%, respectively. The patients with overall grades III-IV acute GVHD had markedly inferior OS and DFS (p?=?0.040 for OS and p?=?0.028 for DFS). However, patients with more than mild stage chronic GVHD showed superior OS and DFS (p?=?0.004 and p?=?0.008, respectively). The 1- and 3-year GRFS rates were 44.5% and 36.9%, respectively. The negative bone marrow involvement at diagnosis, chemosensitive disease status, and fewer exposure lines of chemotherapy before transplantation significantly increased the GRFS incidence. However, no transplant-associated factors were related to GRFS incidence. Furthermore, applying dynamic GRFS method which excepted patients whose chronic GVHD was fully resolved within short-period, survival rate significantly increased over time (36.9% vs. 41.9%, p?=?0.045 for conventional GRFS vs. dynamic GRFS at 3 years after transplantation). In conclusion, these results suggest that GRFS is also a useful endpoint to assess transplant outcomes, and the dynamic GRFS calculation, including rapidly manageable chronic GVHD, is a more practical method for patients with refractory or relapsed heterogenous subtypes of NHL.