Changes in TNFR1 Levels in the First Week Post-Myeloablative HSCT Correlate with Severity and Incidence of GVHD and 1y TRM.

Graft-vs-host disease (GVHD) remains the major cause of morbidity and treatment-related mortality (TRM) after allogeneic HSCT, but there are no independent laboratory tests to either predict or confirm the clinical diagnosis of GVHD. Pre-clinical studies have shown that tumor necrosis factor-α (TNF) plays an important role in the pathogenesis of GVHD and that TNF plasma levels rise often several weeks before clinical disease becomes apparent. We tested the hypothesis that rises in TNF (on day 7 following allogeneic HSCT) will predict the development of significant GVHD and treatment-related mortality (TRM). We measured soluble TNF receptor 1 (TNFR1) as a surrogate for TNF because TNF circulates as a ligand-receptor complex. We studied samples obtained under informed consent from 438 patients undergoing allogeneic HSCT following myeloablative conditioning at the University of Michigan between 2000 and 2005. The conditioning regimens were based on Busulfan (68%), BCNU (20%), or TBI (12%). The median age of the patients was 42y (range 0–65y). The distribution of donors by degree of HLA-match and type was: 6/6 HLA-matched related donors (n=247), 5/6 matched related donors (n=20), 6/6 matched unrelated donors (n=124), 5/6 matched unrelated donors (n=47). Hematologic malignancy was the indication for 95% of HSCT. The median day of onset of GVHD grade 2–4 for related donors was 30d and for unrelated donors was 20d.

Because of the variablilty in baseline TNFR1 levels, we expressed the day 7 value as a ratio to pre-transplant baseline. The mean day 7 TNFR1 ratio strongly correlated with severity of GVHD. The mean TNFR1 ratio was 1.91±0.09 for pts with GVHD 0–1 (n=269), 2.32±0.20 for pts with GVHD 2 (n=83), and 2.92±0.26 for pts with GVHD 3–4 (n=86), p<0.001.

When treated as a continuous variable, change in TNFR1 on day 7 strongly correlated with both the likelihood of GVHD 2–4 (p<0.001) and 1yr TRM (p<0.001). When patients were grouped according to a day 7 TNFR1 ratio above or below 2.5 (corresponding to the 75th percentile and approximately the mean ratio for patients with GVHD 2–4), patients with the high day 7 TNFR1 ratios were much more likely to experience GVHD 2–4 and die within the first year from TRM.

Patients with a day 7 TNFR1 ratio ≤2.5 were more likely to be alive at 1yr post-transplant (64% vs 50%, p=0.008).

We conclude that even within risk groups stratified by donor source the magnitude of rise in early post-transplant TNFR1 ratios helps to predict the subsequent severity of GVHD, TRM and survival. A single threshhold at this early timepoint identifies roughly a quarter of patients who are at approximately twice the risk of significant GVHD and death. These informative changes are detectable often two to three weeks in advance of clinical manifestations of GVHD and may therefore provide the basis for development of a predictive laboratory test.