Skin biopsies are commonly performed after allogeneic bone marrow transplantation (BMT) to help establish the origin of a new skin rash in a transplant recipient. Histologic criteria and a grading system for acute graft-versus-host reaction of the skin are well established. Histologic diagnosis, however, can be difficult and is based on interpretation of subtle changes that show significant overlap with features seen in other entities that can be responsible for a skin rash in the posttransplantation period such as drug reactions, viral exanthems, and the effects of chemotherapy. We retrospectively reviewed 179 skin biopsies from 137 patients who had undergone allogeneic BMT. We compared 98 skin biopsies from 71 patients with acute graft-versus-host disease (GvHD) with 81 biopsies from 66 patients who underwent biopsy to exclude GvHD but did not go on to develop the disease on clinical grounds. Two observers reviewed each slide without knowledge of the clinical situation and graded 16 histologic parameters. No single parameter (e.g., dyskeratotic keratinocytes, basal vacuolization, satellitosis, necrotic cells in appendages) achieved statistical significance on univariate analysis. A search for factors to separate GvHD biopsies from non-GvHD biopsies using logistic regression failed to reveal a single best predictor or a combination of predictors. We conclude that skin biopsies after allogeneic BMT are of limited use in predicting the progression of a skin rash to clinical grade II or higher GvHD.