Prophylactic drug treatment is one of several strategies to reduce postoperative blood loss and potentially limit homologous blood use in open heart surgery. A computerized MEDLINE search supplemented with manual bibliography reviews was performed for randomized clinical trials published in peer-reviewed English-language journals from January 1980 to June 1993. A metaanalysis was conducted of trials evaluating desmopressin (group DD, n = 13), epsilon-aminocaproic acid or tranexamic acid (group EA, n = 4), and aprotinin (group AP, n = 16). Eligible studies used placebo controls and administered the drug in a prophylactic manner. The primary study end point was postoperative chest tube loss (mL, mean +/- standard deviation). There was a significant reduction in postoperative chest tube loss detected for each of the active treatments versus the placebo (DD versus controls: percent reduction 0.11, p = 0.0021; EA versus controls: percent reduction 0.30, p < 0.0001; and AP versus controls: percent reduction 0.36, p < 0.0001). Therapy with EA or AP was associated with a greater reduction in chest tube loss than DD (EA versus DD, p = 0.0033, and AP versus DD, p < 0.0001). Secondary study end points were transfusion requirements, chest reexploration, and perioperative mortality. The volume of postoperative red cell transfusion (mean +/- standard deviation) was reduced with EA (p < 0.0001) or AP treatment (p < 0.0001) compared with a placebo or DD, whereas the proportion of patients given transfusions was limited only in the AP-treated patients (odds ratio 0.23; 95% confidence interval, 0.16 to 0.33; p < 0.0001).(ABSTRACT TRUNCATED AT 250 WORDS)