Incidence, Clinical Risk Factors, and Biomarkers of SOS/VOD Following Allogeneic HSCT in Adults: A Real-Life Study By the Spanish Group of Hematopoietic Stem Cell Transplantation and Cell Therapy (GETH-TC)

Introduction

Sinusoidal obstruction syndrome (SOS) or hepatic veno-occlusive disease (VOD) is a serious complication following hematopoietic stem cell transplantation (HSCT) with an overall incidence ranging from 5% to 15% and a highly variable mortality rate. Recently, the European Society for Blood and Marrow Transplantation (EBMT) updated the diagnostic criteria, severity assessments, and risk factor evaluations for SOS/VOD (Mohty M et al. 2023). The EASIX score, a biomarker of endothelial dysfunction, has been associated to the incidence of SOS/VOD, overall survival (OS), and non-relapse mortality (NRM) following Allo-HSCT (Jiang S, et al. 2021), potentially aiding in early intervention. This study aims to analyze the incidence and impact of clinical risk factors and the EASIX score on Allo-HSCT outcomes in a large real-world, registry-based cohort.

Patients and methods

This multicenter, retrospective observational registry study included all consecutive Allo-HSCT procedures performed between Jan-2019 and Dec-2020 across nine centers members of GETH-TC. All patients (pts) had a minimum follow-up of one year. SOS/VOD definition and clinical risk factors analysis were assessed according to the EBMT-2023 guidelines. The EASIX score was calculated at three time-points after Allo-HSCT (EASIX-d0, d+7, d+14). Different cut-off points for EASIX were evaluated for their predictive ability using Cox regression and Harrell's C test. Statistical analysis were conducted using the Stata software package.

Results

A total of 796 adult pts (aged >18 yo) were included, with the following characteristics: median age of 53.7 years (range 18-74), male gender 59%. Most pts were undergoing their first Allo-HSCT (80%) and used peripheral blood as stem cell source (84%), with AML/MDS being the most frequent underlying diagnosis (67%). Donor types included: matched related donor (35%), haploidentical (33%) and matched unrelated donor (23%). Conditioning intensity: RIC (58%); MAC: (42%). SOS/VOD developed in 81 pts (10.2%), at a median time (IQR) of 15 (9-26) days. Probable SOS/VOD occurred in 41 pts (50.6%), clinical in 36 (44.4%), and proven in 4 (4.9%). The cumulative incidence of SOS/VOD at 30 and 90 days post-HSCT was 8.56% (95% CI, 6.75 -10.64) and 10.20 (95% CI, 8.22-12.43), respectively. Defibrotide was used in 45 pts (55.6%).

At a median follow-up of 35 months, OS and NRM of the entire cohort were 61.89% (95% CI, 58.40-65.19) and 24.10% (95% CI, 21.17-27.14), respectively. OS and NRM for patients developing and not developing SOS/VOD were 50.54% (95% CI, 38.26-61.59) vs. 65.8% (95% CI, 62.17-69.16), p=.085 and 33.42% (95% CI, 23.42-43.71) vs. 18.56% (95% CI, 15.80-21.49), p=.003, respectively

The D+90 cumulative incidence of SOS/VOD-related mortality was 1.52% (0.83-2.57) for the whole series and 15% (8.23-23.67) for pts developing SOS/VOD.

Univariate analysis identified significant clinical risk factors for SOS/VOD diagnosis, including advanced disease (≥2nd CR or relapse) (p=.026), non-HLA-matched donor (p=.016), ≥2nd HSCT (p<.0001), use of parenteral nutrition (p<.0001), elevated serum bilirubin (> 1.5 mg/dL/>26 μmol/L) (p=.034), prior use of gemtuzumab ozogamicin or inotuzumab ozogamicin (p=0.037), elevated levels of tacrolimus/sirolimus (>10 ng/mL) (p=.003) or INR (>1.5) (<.0001) before SOS/VOD diagnosis, baseline (d0) EASIX ≥2 (p= .032) and EASIX ≥ 6 on d+7 (p <.001) and on d+14 (p <.001). Patients with SOS/VOD showed an increased risk of NRM (sHR 1.95, p=.003)

Multivariate analysis confirmed the prognostic value (sHR) for SOS/VOD incidence of d0 EASIX ≥2 (1.60), d+7 EASIX ≥6 (2.62), d+14 EASIX ≥6 (2.77) adjusted by >1st HSCT, parenteral nutrition, elevated serum bilirubin, prior gemtuzumab use, elevated levels of tacrolimus/sirolimus, and elevated INR.

Conclusions

This large, registry-based multicenter retrospective study observed a SOS/VOD incidence similar to other series, with a significant impact on NRM. The prognostic factors analysis confirmed not only several high-risk factors included in the EBMT classification, such as second or subsequent HSCT and the use of parenteral nutrition, but also other less well stablished risk factors, including elevated tacrolimus/sirolimus or INR levels. EASIX score at d0, d+7 and d+14, was validated as predictor of SOS/VOD onset. These findings warrant confirmation in a prospective study.