RESUMO
Aims: There have been inconsistent reports regarding the effect of antithrombin on sepsis; furthermore, there are limited reports on how dosage affects therapeutic efficacy. Thus, we aimed to perform a systematic review and meta-analysis of the use of antithrombin for sepsis and a meta-regression analysis of antithrombin dosage. Methods: We included randomized controlled trials (RCTs) and observational studies of adult patients with sepsis who received antithrombin. Outcomes included all-cause mortality and serious bleeding complications. Statistical analyses and data synthesis were performed using a random-effects model; further, meta-regression and funnel plots were used to explore heterogeneity and biases. Results: Seven RCTs and six observational studies were included. Most patients in the RCTs and observational studies had severe sepsis and septic-disseminated intravascular coagulation (DIC), respectively. A meta-analysis using RCTs showed no significant differences in mortality between the antithrombin and control groups. However, the meta-analysis of observational studies indicated a trend of decreasing mortality rates with antithrombin administration (odds ratio [OR], 0.79; 95% confidence interval [CI], 0.68-0.92; p = 0.002). Bleeding complications were significantly higher in the antithrombin group than in the control group in both study types (OR, 1.90; 95% CI, 1.52-2.37; p < 0.01). The meta-regression analysis showed no correlation between antithrombin dosage and mortality. Conclusion: A meta-analysis of RCTs confirmed no survival benefit of antithrombin, whereas that of observational studies, which mostly focused on septic DIC, showed a significant beneficial effect on improving outcomes. Indications of antithrombin should be considered based on its beneficial and harmful effects.