Grading acute graft-versus-host disease (GVHD) is usually based on quantification of rash, serum bilirubin and diarrhea. Standard criteria have been developed and used for > 20 years by most transplant centers. However, neither the standard GVHD grading system nor any of several revisions has been validated in the context of GVHD prophylaxis with cyclosporine. The 1994 Consensus Conference on Acute GVHD Grading held in Keystone in January 1994 provided an opportunity to: (1) review data regarding these standard criteria; (2) determine if there are sufficient data to revise these criteria; and (3) develop recommendations for reporting results of GVHD prevention trials. Data were provided for 8249 patients from 12 large transplant centers and 2 transplant registries. Standard GVHD grading criteria were found to distinguish different mortality risks and treatment response rates. Analysis of new data suggested that persistent nausea with histologic evidence of GVHD but no diarrhea be included as stage 1 gastrointestinal GVHD. Additional studies were recommended to evaluate heterogeneity of outcome within GVHD grades prior to making further revisions. To improve comparability between publications, reports of GVHD prevention trials should include an accurate description of the grading system used and should report actuarial rates of grades II-IV and III-IV GVHD corrected for graft failure and potential interventions for early relapse. Additional information should include indications for therapy of GVHD and response.