Hepatic veno-occlusive disease (VOD) is the most common life threatening complication of preparative-regimen-related toxicity for bone marrow transplantation (BMT). The frequency of VOD varies greatly, from 1-2% in centers performing pediatric BMT for thalassemia to over 50% in some centers doing BMT for hematologic malignancy. The term liver toxicity syndrome is a clinicopathologic definition which encompasses the range of histopathology within the hepatic venules and surrounding sinusoids and hepatocytes. These histologic abnormalities are statistically associated with a clinical syndrome of jaundice, ascites, and painful hepatomegaly developing early post-transplant. Newer modalities which may aid accuracy are transvenous liver biopsy along with determination of the gradient between the wedged and free hepatic venous pressures, and measurement of blood coagulatory components, particularly protein C levels. Analyses of clinical risk factors for VOD are confounded by lack of a clear hierarchy of risk when comparing heterogeneous patient populations, the methods of patient selection and choice of controls, and whether analysis is univariate or multivariate. Prospective multivariate analyses indicate that the risk of developing liver toxicity is independently correlated with intensity of conditioning therapy, pre-transplant viral hepatitis, use of antimicrobial therapy with acyclovir, amphotericin, or vancomycin (reflecting fever), and mismatched or unrelated allogeneic marrow grafts. These analyses plus morphologic and biochemical data support the hypothesis that VOD is caused by cytoreductive injury to hepatocytes and endothelium in zone three of the liver acinus, and in turn strongly influenced by factors which induce the release of tumor necrosis factor-alpha (TNF-alpha) leading to enhancement or activation of coagulation with obstruction of hepatic sinusoids and venules. Pharmacokinetic measurements of busulfan as a conditioning agent demonstrate a correlation between high steady-state busulfan levels and liver toxicity and suggest that safer and/or more efficacious plasma busulfan concentrations can be obtained by making individual dose adjustments and by changing the schedule of administration. Conservative therapy of severe VOD, including the use of peritoneal-pleural shunts for relief of ascites, is unsatisfactory. Results from prophylactic studies aimed at preventing VOD by heparin or prostaglandin E1 indicate considerable differences with toxicity and efficacy. Use of the TNF-alpha blocker, pentoxifylline, has also shown promise in lessening VOD. A statistical model which predicts patients likely to have an unfavorable outcome from VOD has been used to select premorbid patients for promising new therapeutic modalities, such as recombinant tissue plasminogen activator.