Matt Harmon

108 Chapter five The effect of induced normothermia on the inflammatory host response. LPS resulted in an increase in C-reactive protein (CRP) (figure 2A; from 0.4 mg/ mL [0.3-0.9] at baseline to 7.7 mg/mL [6.85-10.0] at t=8 hours, p=0.03) and in leukocyte count (figure 2B; from 5.5 ± 1.5 x 10E9/L cells at baseline to 13.7 ± 3.2 x 10E9/L cells at t=8 hours, p <0.0001). Lymphocytes decreased after LPS infusion (figure 2C; from 1.6 ± 0.4 x 10E9/L cells at baseline to 0.3 ± 0.0 x 10E9/L cells at t=8 hours, p<0.0001). Induced normothermia was associated with lower CRP levels albeit not statistically significantly altered (p = 0.08). Induced normothermia did not alter leukocyte levels (p = 0.13). Lymphocyte counts were significantly lower in the normothermia compared to the fever group (p = 0.007), however the difference between groups was the largest at T=1 and likely not an effect of cooling. In the fever group, infusion of LPS resulted in an increase in IL-6 and IL-10 plasma levels with a peak at T=3 hours. TNFa levels peaked at T=1 hour after LPS infusion. Figure 4A-C shows the change in cytokine levels from baseline. Induced normothermia lowered plasma cytokine levels of iL-10 compared to febrile controls (figure 4F, p=0.04). but did not alter plasma cytokine levels of iL-6 and TNFa (figure 4D, E, p = 0.4 & p = 0.5 respectively). Supplemental table 2 and 3 show all results from the mixed model and AUC analyses. Discussion In this model of human endotoxemia, we found that induced normothermia (using external cooling and medication to prevent shivering) was well tolerated by awake volunteers. Cooling to normothermia successfully prevented the LPS induced febrile response. Induced normothermia also lowered heart rate but did not affect perfusion as measured by MAP and lactate levels. Induced normothermia lowered iL-10 levels compared to the fever group but did not lower CRP or plasma cytokine levels of iL-6 or TNFa. Induced normothermia did not have a profound effect on makers of inflammation in this study as only iL-10 levels were affected. Of note, the TNFa response is known to be very early. As cooling was initiated after 1 hour, the window to prevent a TNFa response possibly had already passed. However, it is surprising that iL-6 levels were not lower following induced normothermia. Experimental studies in animals have shown that cooling has a profound effect on proinflammatory cytokine levels. 15 However, these studies were mostly performed in animals cooled to hypothermia. Possibly, the temperature difference in our study was not large enough to exert a significant effect on iL-6. Additionally, the sample size of our study may have been too small to find an effect of cooling to