Matt Harmon

122 Chapter six beats per minute (BPM) at baseline to 92.6 ± 10.9 BPM at t=4 hours). LPS caused a leucocytosis (5.5 ± 1.5 × 10 9 /L cells at baseline to 13.7 ± 3.2 × 10 9 /L cells at t=8 hours) and an increase in C-reactive protein (CRP) (0.4 [0.3 -0.9] mg/mL at baseline to 7.7 [6.9 – 10.0] at t=8 hours). The impact of fever control on LPS-induced consumption coagulopathy LPS resulted in a thrombocytopenia in the fever group (from 166 ± 29 × 10 9 /L cells at baseline to 97 ± 50 × 10 9 /L cells at t=3 hours, p < 0.0001). Platelet counts were significantly higher in the normothermia group compared to the fever group (194 ± 20 × 10 9 /L cells at t=3 hours, figure 1A; p = 0.003). LPS resulted in PT prolongation in the fever group (from 11.4 sec ± 0.4 at baseline to 12.2 sec ± 0.6 at t=6 hours, p = 0.03). PT was not altered in the normothermia group (figure 1B; p = 0.99). LPS decreased activated partial prothrombin time (aPTT) (from 25.7 ± 3.3 sec at baseline to 20.8 ± 1.0 sec at t=3 hours, 0.006), which was prevented in the normothermia group compared to the fever group (figure 1C; p = 0.005). LPS did not alter fibrinogen levels, nor were there differences between the fever group and the normothermia group (figure 1D; p = 0.78). LPS increased vWf antigen levels (from 87% ± 39 at baseline to 341% ± 122 at t=6 hours, p < 0.0001), which was prevented in the normothermia group (figure 1E; p = 0.03). LPS-infusion resulted in an increase in D-dimer at T=3 hours (from 0.2 mg/L [0.2-0.3] at baseline to 1.3 mg/L [0.8-1.9] at t=3 hours, p < 0.0001). This increase persisted until the end of the study period. D-dimer values did not differ between the groups (figure 1F; p = 0.06). Calculated ISTH DIC-scores were significantly lower in the normothermia group compared to the fever group (figure 2A, p = 0.04). Figure 2B shows the distribution of DIC scores between groups. In total, half of the subjects in the fever group reached a DIC-score above 4 at some time point during the study period whereas none of the subjects in the normothermia group reached a DIC- score above 4. Supplemental table 1 shows all the results of the mixed model used to assess correlations between the intervention and the measurements of interest. The impact of LPS and fever control on ROTEM values There was a large variation in ROTEM values while largely remaining within reference ranges. Figure 3 shows selected parameters from the ROTEM analyses. Cooling to normothermia did not alter most ROTEM parameters between groups (figure 3, supplemental table 2)., INTEM CT was higher in the normothermia group compared to the fever group (Figure 3D, p= 0.007). Of note, differences in INTEM CT levels between groups were already observable at T=1 hour, before