Matt Harmon

Chapter four 89 from this study can guide future studies on the pathophysiology of the hypo- thermic response. There are also several limitations to this study. First of all, pathway analysis based on transcriptional profiles does not clarify whether the upregulation or downregulation of pathways indicate a lack, or overabundance of a specific protein substrate. Future studies on the metabolite and protein products of the herein inferred canonical signaling pathways are warranted. Second, we did not standardize the timing and method of temperature measurements. However, core temperature measurements are common practice in the ICU setting and we also controlled for potentially incorrect temperature measurements. Furthermore, in this analysis we chose only to compare hypothermic patients to those with fever. As a result, there are large temperature differences between these two groups potentially minimizing the effect of any measurement inaccuracies. Also, blood sampling was performed in the first 24 hours of ICU admission. The sampling did not necessarily coincide with the hypothermic or febrile temperature measurement. However, we do not think that sampling needs to be simultaneous with temperature measurements in order to characterize this group of patients. A single hypothermic measurement in the first 24 hours of ICU admission is significantly associated with adverse outcome 1,7 , and changes associated with hypothermia likely persist beyond the hypothermic measurement. Finally, in the matched analysis, age was significantly higher in the hypothermic group compared to the febrile group. Though age (and gender) was added as a covariate in our transcriptome analysis, the former may confound results, at least in part, since prior studies have shown that age is an independent risk factor for hypothermia. 48 Finally, due to the inherent nature of observational studies, cause- effect relationships cannot be established. Conclusions In conclusion, hypothermic patients were characterized by largely similar, but also significant changes in leukocyte transcriptomes. Genes were associated with distinct cellular biological pathways, including tryptophan metabolism and serotonin-signaling. Cold-inducible mRNA binding protein (CIRBP) expression was particularly elevated in sepsis patients with hypothermia. These signaling pathways provide a possible link to the modulation of body temperature and immunosuppression. Future functional studies on the canonical signaling pathways and specific genes presented in this paper are warranted.