Critically ill hospital patients were fed enteral formulas containing different fat substrates. Seven patients received formula X, which contained 28 g of structured triglycerides and menhaden oil to provide 7.6 g of medium-chain fatty acids, 2.5 g linoleic acid, 1.3 g eicosapentaenoic acid, and 0.4 g docosahexaenoic acid per 1000 mL of formula. Six patients received formula Y consisting of 36.8 g of medium-chain triglycerides and corn and soy oils providing 14.3 g medium-chain fatty acids and 11.7 g linoleic acid per 1000 mL. Feeding of formula X increased plasma total phospholipid levels of eicosapentaenoic acid on days 7 and 14 and docosahexaenoic acid levels on day 14. Plasma levels of linoleic acid were reduced in formula-X-fed in comparison to formula-Y-fed patients, whereas arachidonic acid was maintained in both groups during feeding. As a result of these changes, the patients receiving formula X had decreased ratios of arachidonic acid:eicosapentaenoic acid in plasma. Formula Y feeding did not alter eicosapentaenoic acid and docosahexaenoic acid levels in the plasma. In the erythrocyte, formula X resulted in a threefold increase in eicosapentaenoic acid from mean baseline levels of 0.4 +/- 0.4% to a mean value of 1.2 +/- 0.9% at day 7. The formula X feeding decreased linoleic acid levels on days 7 and 14, whereas levels of arachidonic acid and docosahexaenoic acid remained constant. Formula Y feeding did not affect any of the parameters measured for erythrocytes. The ability to alter plasma and erythrocyte levels of n-3 fatty acids and plasma arachidonic acid:eicosapentaenoic acid ratios may have important implications for patients.at risk for sepsis.