Abstract
BACKGROUND: D-dimer (DD) is the most used fibrin-related marker and has been proposed, either alone or in combination with other variables, as prognostic factor in patients with sepsis. However, DD generation depends on both coagulation and fibrinolysis, meaning that it may give false negative results in conditions associated with marked fibrinolytic inhibition such as sepsis. In this study, we tested whether correction of DD for thrombin and plasmin generation could improve its prognostic significance in septic patients. MATERIAL AND METHODS: We performed a nested study in 269 septic patients from the ALBIOS trial. DD, prothrombin fragment 1+2 (F1+2) and plasmin-antiplasmin complex (PAP) were assayed at day 1. Corrected DD (DD(corr)) was calculated by the formula DD×PAP/F1+2, such that the lower the DD(corr) the greater the imbalance in favour of fibrin formation over fibrin lysis, and vice-versa. Primary outcome was 90-day mortality. RESULTS: DD(corr) showed a J-shaped relationship with mortality, which was highest in the first DD(corr) tertile (low fibrinolysis), intermediate in the 3(rd) (high fibrinolysis), and lowest in the 2(nd) (balanced fibrinolysis), suggesting an increased risk whenever the coagulation-fibrinolysis balance is tilted (p<0.0001). Neither DD, nor PAP or F1+2 showed a comparable association with mortality. DD(corr) was an independent prognostic factor in multivariable Cox models and significantly improved risk stratification (cNRI≥0.28). Finally, by combining DD(corr) and SOFA tertiles, we developed a score with high discriminatory power. DISCUSSION: DD(corr) is a good marker of the in vivo coagulation-fibrinolysis balance and displays a prognostic value in sepsis much higher than DD.