[Background] To achieve allogeneic hematopoietic stem cell transplantation (allo-HSCT) more safety, it is important to minimalize the risk of complications. Acute graft-versus-host-disease (GVHD) is a serious complication after allo-HSCT. Previous reports have indicated that young recipients, elderly donors, unrelated donors and acute leukemias are risk factors for the development of acute GVHD. While, there is a lack of reports on the risk factors involved in non-relapse mortality (NRM) in patients who develop acute GVHD. Patients who develop steroid-resistant acute GVHD have poor prognosis. The causes of death in patients with acute GVHD are diverse, but the main causes are multiorgan failure stemming from the target organ damage due to acute GVHD or severe infections caused by immunodeficiency due to long-term, high-dose steroid administration. Assessment of NRM risk at the onset of acute GVHD is important in determining early indications for second-line therapy for acute GVHD such as JAK inhibitors.

[Aim] Identify risk factors for NRM in patients who develop acute GVHD using clinical information available at the onset of acute GVHD.

[Methods] We performed a retrospective analysis of 173 patients with hematologic diseases who underwent initial allo-HSCT between 2012 and 2023 at our institution. Diagnoses consisted of AML (n=85), ALL (n=33), MDS (n=28), malignant lymphoma (n= 8), aplastic anemia (n=7), and others (n=12). The median age at allo-HSCT was 46 years (range 16-68). Graft sources were BM (n=57), PB (n=83), and CB (n=33). Thirty-four patients (19.6%) had a high (3 or more) HCT-CI score on allo-HSCT. Seventy-two patients (41.6%) received all-HSCT with high or very high disease risk as defined by disease risk index (DRI). We defined onset under steroids (OUS) as the onset of acute GVHD while receiving at least 0.2mg/kg prednisolone equivalent of steroids for at least 1 week prior to the onset of acute GVHD. The Gray's test was used to estimate the cumulated incidence of acute GVHD and NRM. Univariate analysis using Fisher's exact test was conducted to select candidate factors. Fine and Gray subdistribution hazard model was used to identify risk factors for NRM among patients with acute GVHD.

[Results] The cumulative incidence of acute GVHD was 42.3% (grade I 22.1%, grade II-IV 26.1%, grade III-IV 15.0%). In the patients with acute GVHD, the 1-year NRM was 13.0% and the 2-year NRM was 16.3%. OUS was present in 2 of 31 patients (6.4%) in grade I, 3 of 21 patients (14.2%) in grade II, 3 of 13 patients (23.0%) in grade III and 5 of 10 patients (50.0%) in grade IV acute GVHD. The reasons for OUS included engraftment syndrome (n=7), prophylaxis for serum sickness of ATG (n=5), and treatment for Sweet syndrome (n=1). The incidence of steroid-resistant acute GVHD was comparable between patients with OUS and patients without OUS who received systemic corticosteroids (50.0% vs 31.2%, p=0.45). In univariate analyses, Older age (over 50, p=0.04), grade III-IV of acute GVHD (p=0.002) and OUS (p=0.002) had significantly higher rates of NRM. In a multivariate analysis of patients with acute GVHD with age, donor, HLA compatibility, HCT-CI, conditioning therapy, severe GVHD (grade III-IV), and OUS as factors, severe GVHD (HR 6.81 [95%CI 1.96-23.58]) and OUS (HR 8.79 [95%CI 1.56-49.6]) were each independent risk factors for NRM. Six of 13 (46.1%) patients with OUS died without relapse, and the cause of death was bacterial infection (n=2), acute GVHD (n=2), idiopathic pneumonia syndrome (n=1) and cardiac failure (n=1). While, 5 of 63 (7.9%) patients without OUS died without relapse, and the causes of death were acute GVHD (n=3), veno-occlusive disease (n=1), and cardiac failure (n=1), with no death due to infection.

[Conclusion] The development of acute GVHD under steroid administration is a risk for NRM. Early second-line therapy and enhanced infection prophylaxis for patients with GVHD that develops under steroid administration will be considered.

Disclosures

No relevant conflicts of interest to declare.

This content is only available as a PDF.
Sign in via your Institution