Eosinophilia indicates the evolution to acute graft-versus-host disease

We have read with great interest the paper from Daneshpouy et al,1 regarding the activated eosinophils in the upper gastrointestinal tract of patients with acute graft-versus-host disease (GVHD). The authors found that eosinophils were only present when there were histologic signs of GVHD. In addition, the eosinophil density correlated with GVHD severity. From those data, the authors drew the conclusion that tissue eosinophils might thus be a marker of acute inflammatory flare-ups in GVHD. Furthermore, the authors propose that systematic duodenal biopsy at the onset of digestive symptoms should allow early GVHD detection and that pathologic signs of GVHD, together with eosinophil density, might help to modulate immunosuppression therapy.

The data of our prospective study on the predictive value of bone marrow eosinophils in the evolution to acute GVHD confirm those interesting findings. We have performed prospectively bone marrow smears and histologic evaluation at day +30 and day +100 in 237 patients receiving transplants in our department from July 1997 to September 2000.2 Ninety-four patients were diagnosed with acute myelogenous leukemia (AML) or myelodysplasia syndrome (MDS), 68 with chronic myelogenous leukemia (CML), 46 with acute lymphocytic leukemia (ALL), 16 with non-Hodgkin lymphoma (NHL) and Hodgkin lymphoma, and 13 with nonmalignant hematologic disorders. One hundred twenty-nine patients (54%) underwent unrelated hematopoietic stem cell transplantation (HSCT), and 108 (46%) patients underwent related HSCT. Twenty patients died before the first bone marrow evaluation at day +30. A significant eosinophilia of more than 7%, determined at 400 bone marrow cells, was found in 32 patients (15% of patients). Twenty-eight (87.5%) of those patients developed acute GVHD grades II to IV (Table 1). Four patients died due to severe GVHD, and 24 patients developed chronic limited GVHD. We concluded that bone marrow eosinophilia after HSCT, probably mediated by endogenous interleukin-2, predicts severe acute GVHD (predictive positive value 87%). But the functional significance is not known and should be determined.

Few studies have explored the role of eosinophilia in the evolution of acute GVHD. Eosinophilia in HSCT following busulfan-and-cyclophosphamide (Bu+Cy) or total-body-irradiation-and-cyclophosphamide (TBI+Cy) preparative regimens has been regarded by some authors as a valuable indicator of evolution to chronic GVHD.3 But the significance of eosinophilia in acute GVHD is presently unknown. It has been demonstrated4 that blood eosinophils are activated in acute GVHD, and a significant CD25 antigen expression on the eosinophils has been demonstrated. The role of activated eosinophils in the development of cellular tissue destruction in acute GVHD is still a matter of speculation.

Daneshpouy et al's paper1 raises a critical issue of the predictive marker of severe-gut GVHD and consequently the early immunosuppressive modulation.