Biomarkers Predicting Acute GvHD and Transplant Outcomes in 120 Consecutive Allogeneic HSCT Recipients

Background: Allogeneic haematopoietic stem cell transplantation (allo-HSCT) is the only curative treatment for many haematological diseases. Despite advances in supportive therapies, acute Graft-versus-Host Disease (aGvHD) remains the leading cause of early morbidity and mortality. There are shortcomings in the prediction of aGVHD, indicating the urgent need for non-invasive and reliable laboratory tests to allow a precision-medicine tailored prophylactic approach.

Aims: We conducted a prospective observational study to ascertain the potential usefulness of the levels of eleven biomarkers, measured pre- and at +7 days post-alloHSCT, in predicting the development and severity of aGvHD in allo-HSCT patients. These time-points were chosen as early risk stratification may provide a window for additional prophylactic measures.

Methods: We collected data from 120 consecutive patients (41 female and 79 male; median age 52, range 19-77 years) who underwent allo-HSCT at our institute between 2014 and 2015. Most patients were affected by myeloid malignancies (AML=52%, MDS=11%, MPN=5%) while 32% were of lymphoid origin (ALL, MM and lymphomas). Revised disease-risk index (DRI, Armand et al.) was low or intermediate for 41% of pts, high for 44% and very high for 15% of them. Most patients (n=101) received peripheral blood stem cells (PBSC). Stem cell donors were unrelated (n=39, HLA matching 9/10 in 14 and 10/10 in 25), family haploidentical (n=56), HLA-identical sibling (n=21), or cord blood (n=4). Post-transplant GvHD prophylaxis was PT-Cy-bases in 65 patients, ATG-bases in 28 patients, both agents in 18 cases whilst 9 patients received neither. Additionally, sirolimus and MMF were used as additional GvHD prophylaxis according to institutional guidelines. The following biomarkers were measured 7 days pre- and post-transplant: interleukin-6 (IL6), Ceruloplasmin(CER), Cholinesterase (CHE), Albumin, Immunoglobulin A, Gammaglutamyl-transferase (GGT), White Blood Cells, Neutrophils, Haemoglobin,Platelets and Glycaemia. ROC curves were used to define optimal cut-offs of the biochemical variables to predict 100-day severe aGvHD and 100-days TRM by logistic regression. Then cumulative incidences curves (CI) for aGvHD and TRM and overall-survival (OS) curves were computed comparing patients under and over the identified cut offs for the variables with the best reported sensitivity and specificity. Multivariate Cox proportional hazard regression was finally applied.

Results: The 100-days CI of grade III-IV aGvHD was 13.5%, with a 100-days CI of TRM of 12%. The 1-year OS was 64%. Death was reported in 49 of the total 120 patients: 5/49 deaths were attributable to aGvHD. At baseline, IL6 threshold value of 2.5 pg/ml was significantly associated with the prediction of 100-days severe aGVHD (sens 65%, spec 61%) and TRM (sens 75%, spec 61%). Patients with IL-6 concentration equal or superior to 2.5 pg/ml had higher 100-days CI of severe aGvHD (16% Vs 6%, p 0,03) and TRM (12% Vs 3%, p 0,06). Interestingly, higher IL6 concentrations were associated to worse 1-year OS (42% Vs 82%, p<0,001) and were able to better stratify OS also in patients with the same DRI class (OS in low-intermediate risk: 58% vs 85%; high-risk 39% vs 78%; very high-risk 21% vs 49%; p 0,01). At 7 days after HSCT, GGT values were raised in patients who later developed grade III-IV GvHD (cut-off 129 U/l by ROC analysis, sens 69%, spec 67%). Indeed, patients with GGT>= 129 U/l levels had higher 100-days CI of severe aGVHD (15% vs 7%, p 0,16) and TRM (13% vs 4%, p 0,01), in addition to worse OS (53% vs 71% at 1y, p 0,03). By multivariate analysis (adjusting for age, DRI, Sorror comorbidity index, type of donor, source of stem cells, and backbone of GvHD prophylaxis) pre-transplant IL6 concentrations were significantly associated to severe aGvHD (HR 3, p 0.04), TRM (HR 3, p 0.06), and OS (HR 2.6, p 0,004).

Conclusion: In our series, baseline IL6 levels are predictors of aGvHD, especially with regards to grade III/IV aGvHD and translate in a higher risk of TRM and worse OS. This biomarker could be incorporated in a treatment algorithm to increase primary GvHD prophylaxis in patients at risk of severe aGVHD, potentially improving the final outcome of allo-HSCT.