A Plasma Marker Panel Diagnostic for Dismal Transplant-Related Complications

Background: Non-infectious transplant-related complications (TRCs) following allogeneic hematopoietic cell transplantation (allo-HCT) are associated with unfavorable prognosis, but it's difficult to predict the severity at the beginning of clinical symptoms. Inflammatory endothelial injury and hemostatic abnormality were found to play a critical role in the development of severe TRCs. This study aimed to explore plasma markers that could serve as independent predictors for dismal TRCs.

Patients and methods: Patients who received allo-HCT between 2000 and 2013 and had available blood samples within 3 days after the initial symptoms of TRCs were included in the analysis. Patients who had undergone any prior transplantation within 6 months were excluded. A total of 185 patients were evaluated and the median age of patients was 46 years (range:16-68). TRCs within 100 days were categorized into three groups; complications with endothelial damage (ED), acute GVHD, and the other complications (OC). ED was defined as sinusoidal obstruction syndrome (SOS), transplant-associated microangiopathy (TAM), capillary leak syndrome (CLS) and idiopathic pulmonary syndrome (IPS). OC included pre-engraftment immune reaction, engraftment syndrome, hemophagocytic syndrome and organ dysfunction. If a patient experienced multiple TRC episodes, we analyzed the episode that developed first. Plasma angiopoietin-2 (ANG2), VEGF, thrombomodulin (TM), D dimer, plasmin-α2 plasmin inhibitor complex (PIC) and C-reactive protein (CRP) were measured. Multiple characteristics including plasma markers were evaluated for their association with overall survival (OS) and non-relapse mortality (NRM).

Results: Median follow-up duration of surviving patients after transplant was 4.5 years. Sixty-five percent of the patients received allo-HCT for acute myeloid leukemia or myelodysplastic syndrome and the disease risk was high (not in remission) in 40% of the patients. The source of graft was bone marrow from unrelated donor in 38%, cord blood in 30% of the patients. Eighty-six percent of the patients received myeloablative conditioning regimen.

Accumulative incidence of total non-infectious TRCs was 78.9% at day 100. ED was observed in 56 patients SOS (n=17), TAM (n=26), CLS (n=10) and IPS (n=3). These include 16 patients who also experienced grade III to IV GVHD in their post-transplant course. One-year NRM in patients with ED (45.8%) was significantly high compared to 13 patients with grade III-IV GVHD in the absence of ED (16.2%), 52 patients with grade I-II GVHD (10.3%), 26 patients with OC (4.8%), and 18 control patients who never experienced any TRCs or relapse within 100 days after transplant (0%).

All candidate biomarkers except for VEGF and PIC at the onset of ED were significantly higher compared to the peak levels of the control patients by Kruskal-Wallis rank sum test. ANG2 and D dimer were significant for SOS, TAM and CLS, whereas CRP was significant for SOS, and TM was significant for SOS and TAM. Any biomarkers were not significantly elevated at the onset of GVHD or OC. Receiver operating characteristic curves for ANG2, D dimer, CRP and TM distinguished ED from the others with AUC of 0.92, 0.83, 0.76 and 0.74, respectively. ANG2 at the concentration of 4.8ng/ml provided a positive predictive value (PPV) of 77% and a negative predictive value (NPV) of 91% for ED. PPV and NPV reached 91% and 89% with a specificity of 97% by combining ANG2 positivity with either high D dimer (>4.3μg/ml) or high CRP (>4.3mg/dl). Additional biomarkers provided no further increased diagnostic accuracy.

With adjusted analysis in 134 patients who developed non-infectious TRCs, the combined marker was independent risk factor for NRM (HR: 4.2; 95%CI, 2.2-7.9; p <0.001) and OS (HR: 3.1; 95%CI, 1.9-5.2; p<0.001). In patients with ED, the 3-year OS was significantly shorter in the high-ANG2 group than in the low-ANG2 group (28% versus 79%; p=0.022). In patients with GVHD in the absence of ED, the biomarker panel was unable to predict survival.

Conclusions: Combination of ANG2, D dimer and CRP had significant powers for confirming the diagnosis of dismal TRCs at the onset of clinical symptoms. The high diagnostic accuracy would make them useful for real-time clinical judgment and early intervention.