On 2017 Aug 28, Christian J. Wiedermann commented:

Flawed conclusion on sepsis and disseminated intravascular coagulation

In this systematic review of antithrombin (AT) in critically ill patients, no statistically significant effect of AT concentrate on mortality was found in any of the studied patient groups including the group of severe sepsis and disseminated intravascular coagulation (DIC) in 12 randomized controlled trials (RCT) with a total of 2,858 participants.

The KyberSept trial (Warren BL, 2001), a large-scale multicenter RCT directly assessing the effects of AT concentrate on mortality in patients with severe sepsis and septic shock, contributed 2,314 patients to the analysis in sepsis and DIC (weight, 81.4%); however, not all had DIC. In KyberSept patients, DIC was investigated only post hoc. Among the 563 participants on whom there were sufficient data to identify a subgroup of those with DIC, only 40.7% (229 of 563) had DIC at baseline (Kienast J, 2006). Consequently, in the subgroup analysis for patients with sepsis and DIC, the KyberSept trial is at high risk of bias, not at low risk.

This implies that in the meta-analysis by Allingstrup et al. of sepsis and DIC, at least 334 of the total of 2,858 participants definitely did not have DIC, thus, heavily invalidating its conclusions on survival of sepsis and DIC patients after treatment with AT concentrate.

On 2017 Aug 28, Christian J. Wiedermann commented:

Flawed conclusion on sepsis and disseminated intravascular coagulation

In this systematic review of antithrombin (AT) in critically ill patients, no statistically significant effect of AT concentrate on mortality was found in any of the studied patient groups including the group of severe sepsis and disseminated intravascular coagulation (DIC) in 12 randomized controlled trials (RCT) with a total of 2,858 participants.

The KyberSept trial (Warren BL, 2001), a large-scale multicenter RCT directly assessing the effects of AT concentrate on mortality in patients with severe sepsis and septic shock, contributed 2,314 patients to the analysis in sepsis and DIC (weight, 81.4%); however, not all had DIC. In KyberSept patients, DIC was investigated only post hoc. Among the 563 participants on whom there were sufficient data to identify a subgroup of those with DIC, only 40.7% (229 of 563) had DIC at baseline (Kienast J, 2006). Consequently, in the subgroup analysis for patients with sepsis and DIC, the KyberSept trial is at high risk of bias, not at low risk.

This implies that in the meta-analysis by Allingstrup et al. of sepsis and DIC, at least 334 of the total of 2,858 participants definitely did not have DIC, thus, heavily invalidating its conclusions on survival of sepsis and DIC patients after treatment with AT concentrate.