Acute graft-versus-host disease (aGVHD), if not detected and treated early, is a common cause of morbidity and mortality. Drug hypersensitivity reactions (DHRs), the most frequent clinical and histopathological mimickers of early aGVHD, are often still distinguished from aGVHD by the presence of eosinophils within the inflammatory infiltrate on skin biopsy. Distinguishing these entities is important because the delay of appropriate treatment of aGVHD may lead to advanced stages of the disease process with a poor prognosis. To determine whether the existence or amount of eosinophilic infiltrate could be used to differentiate these entities, we employed a quantitative method of analyzing eosinophils in skin biopsies of rashes from patients with aGVHD and DHR. Eosinophils were counted in 50 high-power fields (HPFs) in skin biopsies of patients with clinical grade >or=2 aGVHD (+aGVHD), with clinical grade <2 aGVHD (-aGVHD), and those with clinical DHR (+DHR). The average number of eosinophils per 10 HPFs (ave. eos/10 HPFs) increased throughout each group. The ave. eos/10 HPFs in +DHR was significantly different from both aGVHD groups (P < 0.001). The specificity to completely rule out aGVHD did not reach 100% until 16.0 ave. eos/10 HPFs was observed. There is a significant difference between the numbers of eosinophils found in differentiating DHR from aGVHD, but a very high number (>16.0 ave. eos/10 HPFs) is necessary to rule out aGVHD completely. Therefore, a quantitative analysis of eosinophils in all biopsies to rule out aGVHD would be of limited value and should only be considered in those biopsies with significant eosinophilia.