Tumor-bearing (TB) and control rats were maintained for four to seven days on total parenteral nutrition (TPN) in which the lipid component (11.1% of total calories) was Intralipid or fish oil. Although no TB rats maintained on standard Intralipid-based TPN died prematurely in these experiments, one-third of the TB rats maintained on fish oil-based TPN died of bleeding complications by Day 4 of TPN infusion. The bleeding appeared to be due to the particle size in the fish oil emulsion, because a subsequent study employing a better emulsified TPN with smaller lipid particles exhibited fewer bleeding problems. Tumor growth, skeletal muscle atrophy, and gut hypoplasia were not reduced in the TB group maintained on fish oil-based TPN. However, immunosuppression observed in the Intralipid-based TPN group, as measured by mitogen response of cultured splenocytes, was significantly improved in the rats maintained on fish oil-based TPN. Levels of eicosapentaenoic acid were reduced in all rats receiving TPN. Although the hypothesized antitumor and anticachexia effects of fish oil administration were not realized, significant improvement in immunosuppression was noted. These results suggest that intravenously administered fish oil may have potential positive effects for the host.