To evaluate the effect of 1-desamino-8-D-arginine vasopressin (desmopressin) on blood loss in surgery, we conducted a randomized, double-blind trial of the drug in 35 patients with normal hemostatic function who were having spinal fusion with Harrington rod instrumentation. Seventeen patients were designated to receive 10μg/m2 of desmopressin, and 18, to receive a placebo. Preoperative testing showed that desmopressin increased factor VIII coagulant activity, von Willebrand antigen concentrations, glass bead platelet retention, and prothrombin consumption and decreased the partial thromboplastin and bleeding times (p ≤ 0.0003). During surgery, desmopressin reduced blood loss by 32.5% (547 mL; 95% confidence interval [CI], 19 to 1075; p = 0.015) and reduced the need for concentrated erythrocyte transfusions by 25.6% (0.86 units; 95% CI, 0.08 to 1.65; p = 0.022). After surgery, desmopressin reduced the duration of treatment with analgesic agents by 13.1% (34.0 hours; 95% CI, — 5.2 to 72.7; p = 0.105), presumably by decreasing bleeding in the surgical wound. When adjusted for the origin of the scoliosis by two-way analysis of variance, this effect was even more evident (p = 0.014). Multiple regression analysis showed that the best three predictors of blood loss in surgery and transfusion requirements were the bleeding time, glass bead platelet retention, and the use of desmopressin.