Introduction

During severe sepsis, activation of the inflammatory cascade leads to cell damage and organ failure. In recent years, the importance of the cross-talk between coagulation and inflammation in severe sepsis has been well defined. This has led to the hypothesis that inhibitors of coagulation might have a dual effect, that is, interruption of the cascades of both coagulation and inflammation. Recombinant human activated protein C (rhAPC, drotrecogin alfa (activated), ** the rhAPC infusion. The subsequent ongoing bleeding from the puncture site ultimately required a red blood cell (RBC) transfusion. Blood transfusion requirements were similar in the rhAPC group and the control group (0.44 ± 0.53 versus 0.23 ± 0.35 RBC units per day, respectively, p = 0.27).

Discussion

In the present clinical study, we studied the dynamics of the anticoagulant, pro-fibrinolytic and anti-inflammatory action of rhAPC when used in severe sepsis, by comparing rhAPC treated patients with case controls. We demonstrated that sepsis-induced thrombin generation was reset by rhAPC, as reflected by a decrease in TAT and F1+2 levels within 8 h of infusion. We did not find any influence of rhAPC on parameters of fibrinolysis and inflammation. Although the delay between meeting the inclusion criteria and t = 0 was longer in the control group, we do not think that this difference has influenced our results. Indeed, shifting the control group curves in Figs 1, 2, 3, 4, 5 to the right for 12 h did not change the results of the comparison between the two treatment groups.

The inhibition of thrombin generation by rhAPC might be the main mechanism by which mortality reduction in patients with severe sepsis was achieved in the PROWESS study [1]. Mortality in severe sepsis is usually due to multiple organ failure, which is believed to be caused by microvascular thrombosis, impairing the blood supply to various organs [10, 11]. Under physiological circumstances, thrombin generation is regulated by the protein C system in order to prevent microvascular thrombosis. During sepsis, however, the expression of thrombomodulin and EPCR on the endothelial cell surface is downregulated, leading to inadequate activation of protein C and thus to inadequate inhibition of thrombin generation.

Our findings confirm the results of Dhainaut et al., who demonstrated that treatment with rhAPC attenuates thrombin generation, as reflected by a significant inhibition of TAT and F1+2 [12]. In our study, the inhibition was even more pronounced: treatment with rhAPC prevented the increase in thrombin generation that occurred in the control group. Interestingly, TAT and F1+2 levels did not change from 8 h until 7 days after starting the treatment, even after stop** the rhAPC infusion. These results are in contrast with those of Dhainaut et al., who found an increase in levels of TAT and F1+2 on day 5. There are several possible explanations for this difference. Firstly, we did not take measurements on days 5 and 6 and might have missed a transient rise in thrombin generation. Secondly, the rhAPC group in the PROWESS study might have been more severely ill at inclusion, as the mean APACHE II score was higher than in our rhAPC group (24.6 ± 7.6 versus 21 ± 6). It is conceivable that in more severely ill patients, normalization of thrombin generation takes more time. Thirdly, the time from inclusion to drug infusion was 17.5 ± 12.8 h in the PROWESS study, as compared with 12.3 ± 13.2 h in our study. It is also conceivable that the shorter delay to treatment might have influenced the speed of recovery. If rhAPC is indeed able to reset thrombin generation within 8 h in less severely ill patients when treated within 12 h of admission, one could argue that, under these circumstances, a shorter duration of rhAPC infusion might be sufficient to achieve the same extent of inhibition of thrombin generation. This could have important consequences for the recommended duration of treatment. However, based on the results of the present study, one cannot conclude that limitation of the duration of rhAPC treatment would yield the same results. Additional studies are needed to determine under which circumstances the duration of rhAPC infusion can be limited without influencing efficacy.

At baseline, 92% of our septic patients were protein C deficient with a mean protein C level of 45.8%. This finding is consistent with the results of earlier studies. Boldt et al. found a baseline protein C level of 47.8% in septic patients [13] and in the PROWESS study, Bernard et al. found median baseline protein C levels of 47 and 50% in the rhAPC group and the control group, respectively [1]. The depletion of protein C during sepsis is caused by a combination of degradation of protein C by neutrophil elastase and inadequate biosynthesis in the liver [11, 14]. In our study, the protein C levels returned to normal in the course of 2 days in both treatment groups, whereas in the study by Dhainaut et al., normalization of protein C levels took 3.5 days in the rhAPC group and 5 days in the control group [12]. The increased time needed for the normalization of protein C levels might reflect the greater severity of illness of patients in this study.

In the present study, we did not find a convincing effect of the administration of rhAPC on fibrinolysis. The levels of D-dimers remained unchanged over time in both the rhAPC group and the control group. This is in contrast with the findings of Bernard et al., who found a significant decrease in D-dimer levels in the rhAPC group as compared with the control group [1]. The fact that we did not find such an effect may be due to the small number of patients and the great interpatient variability in D-dimer levels. PAP levels showed a tendency to increase in the rhAPC group, but this increase was too small to reach statistical significance. In agreement with our findings, Dhainaut et al. did not find an effect of rhAPC on PAI-1, a marker of fibrinolysis, when they used the method of repeated measurements [12]. They concluded that their results do not provide a strong basis for a pro-fibrinolytic effect of rhAPC, and our results support this conclusion.

In the present study, we did not find an effect of rhAPC on cytokine levels. Levels of IL-6 and IL-10 gradually declined to normal in the course of 2 days and the level of TNF-alpha remained unchanged over time in both treatment groups. In the PROWESS study, the decrease in IL-6 levels was significantly greater in the rhAPC group as compared with the control group [1]. However, in the post-hoc analysis of the PROWESS data by Dhainaut et al., there were no significant differences in IL-6 levels between the rhAPC group and the control group [12]. Moreover, Dhainaut et al. did not find any difference in levels of TNF-alpha and IL-10 between the two treatment groups. Our findings confirm these results. Dhainaut et al. conclude that their results do not provide a strong basis for a systemic anti-inflammatory effect of rhAPC in vivo at the therapeutic dose used. Our results support this conclusion. Indeed, the anti-inflammatory effect of rhAPC has only been demonstrated in vitro to date [15], using rhAPC concentrations 100- to 1000-fold the concentration achieved in therapeutic circumstances [16, 17].

In the present study, no difference in outcome was found between the rhAPC group and the control group, which is probably due to the small number of patients. The numbers of serious adverse events did not differ between groups.

Conclusion

This study demonstrates that rhAPC resets sepsis-induced thrombin generation within the first 8 h of infusion, without influencing parameters of fibrinolysis and inflammation.

Key messages

  • Recombinant human activated protein C resets thrombin generation within the first 8 h of infusion.

  • The administration of recombinant activated protein C does not influence parameters of fibrinolysis and inflammation.