Parenteral nutrition is advocated for sick foals and horses, although there is no report which has critically evaluated its benefits in this species. Therefore, the hypothesis that parenteral nutrition (PN) is beneficial for horses with severe gastrointestinal disease was investigated in a retrospective study. Records from 79 treatment courses in horses with gastrointestinal disease were reviewed for the years 1992–2000. The fatality rate (48.1%) was attributed to the severity of the primary disease. Complications due to PN were recorded and hyperglycaemia was the most common complication associated with PN. The number of laparatomies, proportion of horses that received lipid emulsion and the daily cost were higher in nonsurvivors than in survivors. The length of PN course was longer in survivors, although glucose peak was observed later in nonsurvivors after initiation of PN. The content of vitamins in PN solution and plasma protein at the beginning of PN were higher for survivors, while the volume of plasma administered was significantly larger in nonsurvivors. In horses with inflammatory diseases, the rate of glucose infusion and the amount of vitamins administered were higher than in horses with nonstrangulated obstruction. The proportion of horses that received lipid emulsion was higher among those with nonstrangulated and strangulated obstruction than in the group with inflammatory conditions. Because of the diverse group of diseases and the variability in the clinical signs, evaluating the effect of PN on disease outcome was not possible. A prospective study to compare horses with similar clinical conditions treated and not treated with PN is needed to evaluate fully the benefits of PN, and to establish guidelines for patients election for PN.