Abstract • Objectives : To evaluate the use of plasma substitutes (albumin, gelatins, dextrans, starches) from 1989 to 1993 in a 3000-bed University hospital and to assess the impact of the 1989 consensus conference of the French speaking Society of Intensive Care Medicine on the choice of plasma substitutes for treatment of hypovolaemia. • Study design : Retrospective study of a case series. • Methods : Data on the use of albumin and artificial plasma substitutes were obtained from Blood Bank and Pharmacy. • Results : Between 1989 and 1993, the total amount of administered plasma substitutes decreased by 20 %, allowing a saving of 1.7 million FF. There was a 60 % decrease in the use of gelatins and dextrans. The starches, introduced in 1991, became in 1993 the most used plasma substitute, with 37 % of the total. The use of albumin showed only a 32 % decrease and was still in 1993 the main source of expenditure for plasma substitutes, with 3.8 million FF and 80 % of the total cost. The number of plasmapheresis increased by 15 %, whereas the use of albumin for plasmapheresis, almost constant in absolute value, increased from 31 % in 1989 to 45 % in 1993 of the total expenses. There were large differences between the services regarding the use of albumin. Intensive care and surgical units represented 70 % of the total (plasmapheresis excluded). In this group, the overall fall of use was 22 %, some units obtaining a 93 % decrease whereas other did not change in spite of constant admission numbers. Medical units decreased their albumin use by 53 %. • Conclusion : The decrease in the albumin use between 1989 and 1993 was not significant. The impact of the consensus conference, if any, was weak. A careful evaluation of the prescriptions would be necessary to improve the compliance of clinical practice with recommendations by the consensus conference.