Association Between Uric Acid Levels and Severe Acute Graft-Versus-Host Disease after Allogeneic Stem Cell Transplantation

Introduction and objectives: There is increasing knowledge on the relevance of danger signals activating the innate immune system in the pathogenesis of graft-versus-host disease (GvHD). Uric acid (UA) is a known danger signal that is released from injured cells during conditioning for allogeneic stem cell transplantation (alloSCT). UA has been identified to enhance dendritic cells maturation and to amplify T-cell responses. Previous reports analyzing the relationship between UA and acute GvHD (aGvHD) have shown controversial results (Ann Hematol 2015; 94: 139; Biol Blood Marrow Transplant 2014; 20: 730).The objective of this study is to investigate the association between UA levels pre-alloSCT and aGvHD.

Patients and methods: We retrospectively analyzed data from 548consecutive patients who received an alloSCT from January 2007 to December 2014 on four institutions. The cumulative incidence of aGvHD over the first 100 days post-alloSCT was estimated in the framework of competing risks by taking relapse or death without relapse as competing events. The association between pre-alloSCT UA levels (measured between day -7 and day 0), dichotomized at the median value, and the incidence of aGvHD was investigated by calculating the subhazard rate (SHR) according to Fine & Gray (J Am Stat Assoc 1999; 94: 496).

Results: Patients (median age 46y, range 16-67; 66% male) were transplanted for acute leukemia (51%), lymphoproliferative syndrome (22%), myelodysplastic syndrome (12%), multiple myeloma (8%), and other causes (7%). Disease status at the time of alloSCT was 49% complete remission, 17% partial response and 31% active disease. Fifty-one percent patients received alloSCT from related donors, 41% from unrelated donors, and 8% from alternative donors (cord blood units or haploidentical donors). The source of stem cells was peripheral blood in 84% patients. Fludarabine-based reduced intensity conditioning (RIC) regimens were used in 54% of transplants and myeloablative (MAC) in 46%. Most patients received immunosuppression with calcineurin inhibitors in combination with methotrexate (MAC) or mycophenolate mofetil (RIC). Sixteen percent also received ATG. Most patients (82%) received allopurinol 300 mg/day through conditioning regimen until day 0 of transplant. The median serum UA level prior to stem cells infusion was 3.7 mg/dL (range 0.7-8.8). Serum levels of UA (mg/dl) were higher in males than in females (4 vs. 3, p=0.0001), in patients submitted to RIC as compared with MAC alloSCT (4 vs 3.3, p=0.0001), in those not receiving ATG for GvHD prophylaxis (3.8 vs. 2.9, p=0.0002) and in the minority of patients not treated with allopurinol (4.7 vs. 3.3, p=0.0001). By day +100 post-alloSCT, 222 patients had been diagnosed of aGvHD II-IV, 79 had relapsed and 72 had died from causes unrelated to relapse or GvHD. The cumulative incidence of aGvHD II-IV was 40% (95% CI: 35%-44%) whereas the incidence of competing events (death or relapse) amounted to 11% (95% CI: 8%-13%). Patients with levels of UA above the median had an increased incidence of aGvHD II-IV (SHR: 1.4, 95% CI: 1.1-1.8; p=0.015) at the univariate analysis. However, such relationship vanished after multivariate adjustment for other factors associated with the peri-alloSCT UA levels and/or the risk of aGvHD (e.g. age, sex, type of donor, type of alloSCT, GvHD prophylaxis). Additional exploratory and subgroup analyses did not find any relationship between UA levels neither severe aGvHD (grades III/IV), early aGvHD (before day +60) nor type of conditioning regimen (RIC vs. MAC).

Conclusions: There is a weak association between higher peri-alloSCT levels of uric acid and increased incidence of aGvHD. Nevertheless, our results suggest that such associations reflect confounding effects rather than causal factors and, therefore, they do not support interventions aimed at preventing aGvHD by reducing uric acid levels.