The composite endpoint graft-versus-host disease (GVHD)-free, relapse-free survival (GRFS) has recently been introduced as a tool to assess the success of allogeneic hematopoietic stem cell transplantation (HSCT) and has been incorporated into recent randomized trials of GVHD prophylaxis by the Blood and Marrow Transplant Clinical Trials Network. As developed, GRFS incorporates "chronic GVHD requiring systemic immunosuppression" as a measure of clinically significant chronic GVHD (cGVHD). However, the decision to start patients on immunosuppressive therapy for cGVHD is subjective and physician-dependent. We elected to assess a modification of the GRFS (m-GRFS) that uses a potentially more objective measure of cGVHD, specifically the development of National Institutes of Health grade moderate or severe cGVHD. A total of 613 patients who underwent a first allogeneic HSCT after an HLA-identical sibling (matched related donor [MRD]; n = 212), an 8/8 matched unrelated donor (MUD; n = 251) or T cell-replete haploidentical donor (HID) transplant with post-transplantation cyclophosphamide (n = 150) were included in this analysis. In the HID group, 86 patients (54%) received peripheral blood stem cells as the graft source. The median duration of follow-up was 50.2 months. The unadjusted Kaplan-Meier estimates for 1- and 2-year m-GRFS were 36% (95% confidence interval [CI], 32% to 40%) and 28% (95% CI, 25% to 32%), respectively. The 2-year m-GRFS was 30% (95% CI, 24% to 36%) for MRD graft recipients, 24% (95% CI, 19% to 30%) for MUD graft recipients, and 33% (95% CI, 26% to 41%) for HID graft recipients. A multivariate Cox model for m-GRFS identified donor type, Disease Risk Index (DRI) risk, donor-recipient sex mismatch, and year of transplantation as significant predictors of m-GRFS. Patients who received a MUD graft had worse m-GRFS compared to MRD graft recipients (hazard ratio [HR], 1.39; P = .003), whereas HID graft recipients had a similar m-GRFS as MRD graft recipients (HR, 1.10; P = .43). HID was associated with better m-GRFS compared with MUD (HR, .79; P = .046). These data show that m-GRFS is significantly affected by several modifiable factors, including donor type, donor-recipient sex match, and DRI. Adjusting donor choice and earlier referral of patients for evaluation of transplantation to improve the DRI can potentially overcome the negative impact of these factors.