GvHD results in death in the majority of steroid-resistant patients. This report assesses the safety and efficacy of two regional intra-arterial steroid (IAS) treatment protocols in the largest published cohort of patients with resistant/dependent hepatic and/or gastrointestinal GvHD, as well as identification of predictors of response to IAS and survival. One hundred and twenty patients with hepatic, gastrointestinal GvHD or both were given IAS. Gastrointestinal initial response (IR) and complete response (CR) were documented in 67.9% and 47.6%, respectively, whereas hepatic IR/CR in 54.9% and 33.3%, respectively. The predictors of gastrointestinal CR were lower peak GvHD and steroid-dependent (SD) GvHD. The predictors for hepatic CR were male patient, reduced intensity conditioning and SD GvHD. Twenty-six of the 120 patients (21.6%) are currently alive (median follow-up for the survivors 91.5 months). The 12 months' overall survival is 30% with no treatment-associated deaths. Predictors of 12 months' survival were as follows: first transplant, age < 20 years, non-TBI regimen and GvHD CR. Shorter time to gastrointestinal IR but not time to hepatic IR was associated with improved 12 months' survival. IAS appears to be safe and effective. Gastrointestinal treatment is more effective than hepatic treatment. In our study, we conclude our current recommendations for IAS treatment.