Peripheral intravenous nutrition with hyperosmolar solutions usually results in a high rate of venous complications. The aim of this multicenter double blind randomised study in 98 patients has been to measure: (a) tolerance by the peripheral veins being perfused with a protein-glucose-lipid nutritive mixture of 960 mOsm/l (group A, n = 33), (b) the protective effect of the additive to the nutritive mixture of either heparin: 1000 IU/1 (group B, n = 32) or heparin with hydrocortisone (5 mg/l) group C, n = 33). Tolerance by the veins was evaluated on a single vein site during a 48 h perfusion with 21 per day. The following complications: oedema, erythema, induration, thrombophlebitis led to the interruption of the perfusion. The rate of interruptions of perfusions for total venous complications and for thrombophlebitis has been respectively: at 24 h, in group A: 39 and 15%, in group B: 6 and 3%, in group C: 12 and 0%; at 48 h: in group A: 82 and 42%, in group B: 53 and 18%, in group C: 36 and 6%. Venous complication rates for 24 and 48 h were significantly lower in groups B and C (p<0.05) than in group A and there was no inter group difference between groups B and C. These results suggest that peripheral venous nutrition infusing 14.1 g of nitrogen and 8.5 MJ per day can be performed with a hyperosmolar solution of 960 mOsm/1, if that heparin be added to the nutritive mixture and the perfusion site be changed daily. Under these conditions the observed venous complications rate is equal to or less than 6% of cases.