Impact of Graft-Versus-Host Disease on Relapse and Nonrelapse Mortality Following Posttransplant Cyclophosphamide-Based Transplantation

Abstract

Following conventional graft-versus-host disease (GVHD) prophylaxis, the development of acute and/or chronic GVHD is associated with lower relapse rates. However, the effects of GVHD on relapse and non-relapse mortality following post-transplant cyclophosphamide (PTCy)-based GVHD prophylaxis have not been well studied. To this end, we analyzed the impact of acute and chronic GVHD following PTCy-based haploidentical donor transplantation (HIDT). The analysis included 335 consecutive HIDT recipients transplanted at a single institution between 2005 and 2021. Landmark analysis (LA) and time-dependent multivariable analysis (MVA) were utilized to study the impact of GVHD development on transplant outcome. Landmarks were defined as Day +100 for acute GVHD and one-year for chronic GVHD. Recipient characteristics included a median age of 50 (19-80) years, most commonly transplanted for acute leukemia[/MDS [242]. PBSC was the graft source in 81%, and regimen intensity was myeloablative in 49%. Median follow-up was 65 (23-207) months. In landmark analysis, development of grade 3 to 4 acute GVHD (versus 0-1) was associated with inferior 3-year overall survival (OS 47% versus 64%, P = .041), due to higher NRM (25% versus 10%, P = .013). In contrast, development of grade 2 acute GVHD had no significant effect on NRM or survival. When restricted to acute leukemia/MDS patients, development of grade II acute GVHD was associated with improved OS (79% versus 58%, P = .027) and a trend towards lower relapse (24% versus 36%, P = .08). Development of moderate-to-severe chronic GVHD resulted in significantly higher NRM (15% versus 4%, P = .010), but had no impact on relapse, DFS or OS. In Cox multivariate analysis (MVA), grade 3 to 4 acute GVHD and moderate-to-severe chronic GVHD were both associated with significantly higher NRM (HR 3.38, P < .001 and HR3.35, P < .001, respectively). In addition, grade 3 to 4 acute GVHD predicted worse OS (HR 1.80, P = .007) and DFS (HR 1.55, P = .041). In contrast, relapse was not impacted by acute or chronic GVHD in MVA. Grade 2 acute GVHD was not associated with transplant outcome in MVA. In summary, both grade 3 to 4 acute and moderate-to-severe chronic GVHD were associated with higher NRM after PTCy-based HIDT, without an effect on relapse risk. Methods of early identification of such patients in order to augment GVHD prophylaxis are clearly needed.