Pre-Transplant Beta2-Microglobulin Is an Independent Biomarker of Acute Graft-Versus-Host-Disease after Allogeneic Hematopoietic Cell Transplant in a Single Institutional Study

Introduction: Refinement of strategies capable to stratify the risk of acute Graft-versus-Host Disease (aGVHD) before HCT could facilitate critical clinical decisions in in the course of HCT, and improve studies about preventive and therapeutic interventions. The HCT comorbidity index (HCT-CI) is a relatively simple index that, accordign to recent data, is also capable to stratify the risk of aGVHD. More recently, it has been shown that the addition of laboratory biomarkers such as ferritin, albumin, C-reactive protein and platelet count, measured in the pre-transplant period, can further refine the accuracy of this index. β2-microglobulin (β2-m) is a component of the MHC class I complex that lies beside the α3 chain of this molecule and lacks a transmembrane region. The best characterized function of β2-m is the stabilization of the tertiary structure of the MHC class I. However, the protein also seems to be involved in iron metabolism and proliferation of cancer cells. β2-m is widely used as a prognostic factor in multiple myeloma. Interestingly, it also represents a biomarker of frailty in the elderly (Annweiler et al, 2011). In our Institution, β2-m is systematically measured before HCT in all patients since 2010. Based on the role of this protein in immune regulation, and on its association with frailty and other important comorbidities, we hypothesized that pre-transplant β2-m levels could be a biomarker of aGVHD.

Methods: This is a retrospective single-center observational diagnostic accuracy study. All consecutive patients (pts) submitted to allogeneic HCT in our Institution (n=103) between January 2010 and February 2014 were included in the analysis. Included pts underwent one single HCT using all types of conditioning, and unrelated or related HLA-compatible donors. Baseline levels of β2-m were systematically collected approximately 30 days before HSCT (along with several other laboratory tests evaluated in the pre-HCT visit), in the outpatient setting. β2-m was measured in the same clinical laboratory and by the same methodology (immunonephelometry) during the study period. Clinical and laboratory data were obtained from the medical records. Univariate and multivariate Cox regression were applied for searching independent variables influencing the risk for aGVHD. The Kaplan-Meier method and log-rank test were used for the calculation of overall survival (OS). aGVHD and TRM incidence were calculated considering early death and relapse as a competitive event (for acute GvHD), and relapse (for TRM), using the Gray’s test to compare the curves.

Results: β2-m levels were available for 97 pts with a median age of 44 years (16-68). High and low-dose conditioning were used in 59 and 38 pts respectively. Overall survival at 12 and 60 months were 60% (95%CI: 50-70%) and 44% (95%CI: 30-58%), respectively. TRM at 12 months was 20% (95%CI: 12-28%). Twenty-six (34%) pts presented aGVHD at a median of 62 days (20-140), of which 16 presented grade II aGVHD and 10 pts grade III-IV aGVHD. The cumulative incidence of aGVHD was 29% (95%CI: 19-29%) in 100 days. Higher levels of β2-m were observed in pts who developed aGVHD when compared to patients without this complication (p = 0.008). β2-m was also higher in older pts, in patients with more advanced disease, and submitted to low dose conditioning. In the multivariate Cox regression model for aGVHD, β2-m remained significant in different models incorporating variables such as age, conditioning dose, early x advanced disease, HCT-CI, diagnosis of lymphoproliferative diseases and ferritin, and adjusted for age and for the presence of lymphoproliferative disease (table 1). No association was observed between β2-m with TRM and OS.

Conclusion: the strong association between pre-transplant β2-m and the occurrence of aGVHD suggests that this protein could represent an additional biomarker of aGVHD. Larger multicenter studies are required to validate our results, and define whether the incorporation of β2-m levels into aGVHD risk stratification models could add prognostic value, and to understand the biological mechanisms underlying this association.

Table 1: Multivariate analysis of the risk of aGVHD

Variable	HR	p-value	CI 95%
Age-Adjusted
β2-m	3.72	0.03	1.12-12.3
Lymphoproliferative disease-adjusted
β2-m	4.44	0.03	1.12-17.5
No adjustment
β2-m	4.97	0.01	1.40-17.6
HCT-CI ≥2	2.79	0.04	1.05-7.42