Stage 0+ lower gastrointestinal (LGI) acute graft-versus-host disease (aGVHD) is associated with high response to non-systemic treatment, low incidence of moderate to severe chronic graft-versus-host disease (cGVHD) and low non-relapse mortality (NRM

Abstract Introduction: LGI aGVHD remains a major cause of morbidity and mortality after allogeneic hematopoietic cell transplantation (alloHCT). Stage 0+ LGI aGVHD, defined as low-volume diarrhea (1–499 mL/day or 1–2 episodes/day), is recognized in GVHD assessments but excluded from conventional grading systems. We hypothesize that patients (pts) with stage 0+ LGI aGVHD exhibit distinct outcomes compared to those without aGVHD (no aGVHD), those with aGVHD but without LGI involvement (no LGI aGVHD) and those with higher LGI aGVHD stages (stages 1-4). Methods: We evaluated 376 adults (median age 60, 57% male) who underwent alloHCT between 1/2019 and 12/2023. Pts were stratified by their day-100 GVHD assessment as follows: no aGVHD (n=123), no LGI aGVHD (n=157), stage 0+ (n=26) and stages 1-4 (n=70). Outcomes were assessed using the cumulative incidence function and Kaplan-Meier method from the day-100 landmark. Results: Most pts underwent first alloHCT (84%) for the treatment of acute leukemia (72%) using unrelated donors (73%: 37% matched, 36% mismatched) after reduced-intensity conditioning (64%). GVHD prophylaxis with post-transplant cyclophosphamide was used in 41% of pts. The median follow-up was 35.1 months (mo) (95% CI 32.4-37.8). Among patients with no LGI aGVHD, 36% had upper GI (UGI), 39% had skin, 24% had UGI + skin and 1% had liver involvement. Among patients with stage 0+, 4% had LGI involvement alone, 42% had concomitant UGI, 8% had concomitant skin, 42% had concomitant UGI + skin, and 4% had concomitant UGI + skin + liver. Among patients with stages 1-4, 7% had LGI involvement alone, 49% had concomitant UGI, 3% had concomitant skin, 40% had concomitant UGI + skin, and 1% had concomitant UGI + skin + liver. Overall day 28 GVHD treatment response was 89% in no LGI GVHD, 92% in stage 0+ and 78% in stages 1–4. LGI GVHD treatment response was 88% in stage 0+ vs 77% in stages 1-4. Systemic corticosteroids were used in 30% of no LGI GVHD, 31% of stage 0+ and 80% of stages 1-4 cases, while topical treatment (budesonide) without systemic steroids was used in 44%, 69% and 18% of pts, respectively. The cumulative incidence of moderate to severe cGVHD at 12 mo was 17% (95% CI 11-25) in no aGVHD, 11% (95% CI 6.7-17) in no LGI aGVHD, 8% (95% CI 1.3-23) in stage 0+ and 16% (95% CI 8-26) in stages 1-4. The cumulative incidence of relapse at 12 mo was 18% (95% CI 11-26), 17% (95% CI 11-23), 16% (95% CI 4.9-33) and 12% (95% CI 5.7-22), respectively (p=0.4). Notably, relapse accounted for 5/6 of deaths in stage 0+ pts. 12-mo NRM was 1.9% (95% CI 0.37-6.2) in no aGVHD, 3.5% (95% CI 1.3-7.5) in no LGI GVHD, 4% (95% CI 0.27-17) in stage 0+ and 20% (95% CI 11-31) in stages 1-4. 12-mo progression-free survival (PFS) was 80% (95% CI 73-88) in no aGVHD, 80% (95% CI 73-87) in no LGI aGVHD, 80% (95% CI 66-97) in stage 0+ and 68% (95% CI 57-80) in stages 1-4 (p=0.14). 12-mo overall survival (OS) was 83% (95% CI 77-90), 83% (95% CI 78-99), 85% (95% CI 72-100) and 73% (95% CI 63-84), respectively (p=0.6). Conclusions: In this cohort of adults with stage 0+ LGI aGVHD (n=26), we observed favorable clinical outcomes including low incidence of moderate to severe cGVHD and low NRM. Most (88%) stage 0+ pts had concomitant UGI GVHD involvement. Compared to stages 1–4 LGI aGVHD, stage 0+ pts were primarily treated with topical steroids alone, most often budesonide, achieving high overall and LGI GVHD response rates. Relapse was the leading cause of death in stage 0+, however this trend was not statistically significant, likely reflecting our limited sample size. These findings suggest that stage 0+ LGI aGVHD may represent an early, clinically meaningful disease state that warrants formal recognition. Incorporating this category into GVHD grading systems may improve risk stratification and support early interventions. Further studies are needed to validate these observations in larger cohorts and elucidate the biological mechanisms underlying relapse in this subgroup.