Conventional intravenous fat emulsion, made of soybean oil, provide essential fatty acids (EFA) in excess of requirements and may be responsible for abnormal fatty acid profiles and increased peroxidation. This double blind, randomized study in children aims to compare a new 20% olive oil-based emulsion with reduced content of EFAs (20%) to a classical 20% soybean oil-based emulsion containing 60% EFAs. The study included 18 children, aged 4±0.7 yrs, on long-term TPN for short bowel syndrome (n=8), intractable diarrhea of infancy (n=7) and other (n=3). Children were randomized to receive for 2 months, 24% of nonprotein calories (1.80 g/kg.day) either as olive oil (O, n=9) or soybean oil-derived (S, n=9) emulsion. Daily nitrogen intakes were 300±20 mg/kg. All children underwent a preinclusion period (30 days) during which they received the same amount of lipid as MCT/LCT based emulsion. Assessment were performed at day -30, 0, 30 and 60 by ANOVA* on the deltas (D30-D0; D60 - D0) and included: liver function tests, plasma levels of phospholipids (PL), triglycerides (TG), cholesterol (CT), HDL, lipoprotein (apoA1, apoB), FA pattern of erythrocyte membrane and plasma phospholipids, and peroxidation index with the methods of Thiobarbitine Acid Reactive Substances (TBARS). None of the values differed between both group at the start of the study period (D0). Results: There was no difference for TG, apoA1, apoB, HDL-C while CT was significantly higher with S at D60 (p<0,047). On FA pattern of erythrocyte membrane values, C20:4 n-6 did not change between D0, D30 and D60 in both group as well as the ratio C20:3/C20:4 n-6. The main differences are as follows: Table LV-TBARS and ratios LDL-TBARS/LDL (CT+PL+TG) and LV-TBARS/LV(CT+PL+TG) were significantly lower in the O group. Clinical and biological tolerance of both emulsions was good and comparable. C18:1n-9 and C18:2n-6 increased in 0 and S group respectively, reflect the respective composition of the emulsions. In conclusion, the olive oil based emulsion with more balanced EFA composition is well tolerated, maintain a normal EFAs status and the conversion of linoleic acid to long chain n-6 fatty acids, and may be more suitable for prevention of lipid peroxidation.