Intravenous ribavirin was provided non-selectively for investigational open-label use among persons with suspected hantavirus pulmonary syndrome (HPS) in the United States between 4 June 1993 and 1 September 1994. Therapy was initiated prior to laboratory confirmation of hantavirus infection because most deaths from HPS occur within 48 h of hospitalization. Thirty patients with confirmed HPS, 105 patients without HPS and 5 patients without adequate diagnostic testing for HPS were enrolled. This observational study arguably provides the most complete information available on ribavirin-associated adverse effects. Although ribavirin was generally well tolerated, 71% of recipients became anaemic and 19% underwent transfusion. An apparent excess of hyperamylasaemia/pancreatitis was either therapy-associated or due to enrollment bias. The 30 enrolled HPS patients had a case-fatality rate of 47% (14/30). It is not possible to assess efficacy with this study design. However, comparison of survival curves for the 30 enrolled HPS patients and 34 patients who developed HPS during the same time period but were not enrolled did not suggest an appreciable drug effect. A randomized, placebo-controlled trial that enrolls patients during the prodrome phase would be necessary to assess the efficacy and further define the safety of intravenous ribavirin for HPS.