Pretransplantation Plasma ST2 Level as a Prognostic Biomarker of 1-Year Nonrelapse Mortality in Allogeneic Hematopoietic Cell Transplantation

Abstract

Soluble ST2 is established as a prognostic biomarker of nonrelapse mortality (NRM) when measured early after allogeneic hematopoietic cell transplantation (HCT). However, less is known about the prognostic value of ST2 measured before transplantation. We hypothesized that pretransplantation plasma ST2 level was associated with 1-year NRM and could add to our current prognostic assessment. Moreover, we aimed to investigate the associations between pretransplantation plasma ST2 levels and patient characteristics and other plasma biomarkers and to reproduce previous associations between post-transplantation plasma ST2 levels and outcomes of HCT. We conducted this cohort study of 374 adults who underwent allogeneic HCT at our center between July 2015 and December 2019 (median age, 59 years; 55% with a nonmyeloablative conditioning regimen). ST2 levels were measured by enzyme-linked immunosorbent assay in stored plasma samples obtained at a median of 23 days before HCT and also in samples obtained on days +7 and +14 post-HCT. A logistic regression model of 1-year NRM was fitted using an a priori defined set of covariates consisting of age, Hematopoietic Cell Transplantation-Specific Comorbidity Index (HCT-CI), and conditioning intensity (myeloablative versus nonmyeloablative), to which the pretransplantation ST2 level was added as a variable to assess its incremental prognostic value. Models also were fitted of 1-year all-cause mortality, relapse, and grade II-IV acute graft-versus-host disease (GVHD) for pretransplantation and post-transplantation ST2 levels. The median pretransplantation plasma ST2 level was 20.4 ng/mL (interquartile range, 15.2 to 27.2 ng/mL). Pretransplantation ST2 levels were higher in males compared with females (median, 22.2 ng/mL versus 18.1 ng/mL; P < .001) and were correlated with HCT-CI (Spearman ρ=.18; P < .001), body mass index (ρ=.10; P=.05), and plasma levels of C-reactive protein (ρ=.34; P < .001), creatinine (ρ=.17; P=.001), and albumin (ρ=-.17; P < .001). Pretransplantation ST2 levels added prognostic information about 1-year NRM to age, HCT-CI, and conditioning intensity (adjusted odds ratio [OR] of 1-year NRM per 10 ng/mL increase in ST2, 1.32; 95% confidence interval [CI], 1.05 to 1.65; P=.02). Although adding pretransplantation ST2 levels did not notably improve model discrimination (.674 to .675, ΔAUC=.001), it increased the diversity of the predicted risks (P=.02, likelihood ratio test). Pretransplantation ST2 levels also were prognostic of 1-year all-cause mortality (adjusted OR per 10-ng/mL increase, 1.23; 95% CI, 1.02 to 1.48; P=.03), but not of relapse (P=.47) or acute GvHD (P=.81). Plasma ST2 levels at day +7 were prognostic of 1-year NRM, all-cause mortality, relapse, and acute GVHD, whereas levels at day +14 were prognostic of 1-year NRM and all-cause mortality. Our results show that pretransplantation plasma ST2 levels added prognostic information about 1-year NRM to age, HCT-CI, and conditioning intensity, and suggest that ST2 has potential as a biomarker of pretransplantation vulnerability and should be considered in future developments of prediction models of NRM after allogeneic HCT.