METHODS Seventy-two consecutive patients undergoing orthotopic liver transplantation at the Department of Surgery I, University of Vienna Medical School (OLT nos. 1 to 72), were evaluated. Their mean age was 47 years (range: 18-63 years). The indications for liver transplantation are listed in Table 1. All transplant procedures were performed without using a bypass technique. The intraoperative management and surgical procedure have been described elsewhere [7]. Patients were categorized in two groups, each of which was divided in two subgroups. Group I consisted of 18 patients transplanted before the introduction of preoperative plasma exchange. These were retrospectively allocated to two subgroups on basis of their preoperative prothrombin times (PT): A (n = 9): preoperative PT less than 40%; B (n = 9): preoperative PT greater than 40%. The two subgroups of group 2, which contained 54 patients, were compared on a prospective basis: C (n = 32): preoperative PT above 40%; D (n = 22): PT on admission below 40%, preoperative plasma exchange. Comparison of the two subgroups was based on the following parameters: (1) pre-exchange PT (subgroup D); (2) preoperative PT (= PT post-plasma exchange in subgroup D; (3) intraoperative infusion volumes (balanced electrolyte solutions and human albumin to maintain an intravascular colloid osmotic pressure greater than 16 mm Hg); (4) transfusion volumes (whole blood stored for no more than 72 h or packed red cells and fresh plasma, as available; and (5) intraoperative sodium bicarbonate requirements to maintain an arterial pH greater than 7.20. RESULTS (Table 2) . Prothrombin time (PT): Group 1: Patients in subgroup A had a mean preoperative PT of 34% (range: 15%-40%). This was significantly lower than in subgroup B (74%; 52%-100%; P less than 0.001). Group 2: The pre-exchange mean PT in subgroup D was 27% (12%-39%) vs. 68% in subgroup C (45%-104%), the difference being highly significant (P less than 0.0001). In patients in subgroup D a mean plasma volume of 3638 ml was exchanged by plasmapheresis. This resulted in a significant increase in PT to 55% (Table 3). As a result, the preoperative post-exchange PT in subgroup D was slightly but significantly (P less than 0.005) less than in subgroup C. Transfusion volumes: Group 1: Patients in subgroup A needed significantly more blood units than those in subgroup B (55.3 units [19-110] vs. 18.7 [3-33]).(ABSTRACT TRUNCATED AT 400 WORDS)