Objective: To study the effect of a low (0.5g) or a standard (1g) tranexamic acid (TA) dose compared to placebo on clinical and biological endpoints in women experiencing postpartum hemorrhage (PPH) Design: TRACES trial is a double-blind, randomized, placebo-controlled, dose-ranging study Setting: 8 women hospitals in France. Population: Women experiencing PPH > 800 mL during caesarean section. Method: After informed consent, patients were randomized to receive either TA 0.5g (n=57), TA 1g (n=58), or a placebo (n=60). Data were collected at 8 time-points. Main outcome measures: Efficacy: additional blood loss after study drug, maternal morbidity, safety, biology: D-dimers, plasmin-antiplasmin complexes (PAP), simultaneous-generation-thrombin-plasmin-potential. Results: Compared to 1g dose, 0.5g TA was less effective to reduce additional blood loss (300 mL [95% confidence interval (95%CI) 68 to 630] vs 134 mL [95%CI50 to 419] (p=0.042)). Compared to placebo, 1g TA, but not 0.5g, inhibited hyperfibrinolysis as shown by plasmin generation potential, % increase in D-dimers from injection to 120 minutes (93% [95%CI 68 to 118] vs 58% [ 95%CI 32 to 84] (p=0.06) vs 38% [95%CI 13 to 63] (p=0.003) and % increase in PAP from injection to 30 minutes (56% [95%CI 25 to 87] vs 13% [ 95%CI 18 to 43] (p=0,051) vs -2% [95%CI -32 to 28] (p=0.009)). Conclusions: In this study, fibrinolysis inhibition was more sustained after the administration of 1g TA compared to 0.5g TA or a placebo. Further pharmacokinetic-pharmacodynamic modelling will be needed to determine the optimal TA dose to be administered in PPH. NCT 02797119