Matt Harmon

60 Chapter three poor outcome on the ICU 24 , the relationship between hypothermia and BMI and their combined role on outcome is unclear. Leptin, released from adipose tissue, has anti-inflammatory properties and may also mediate the hypothermic response, providing a possible link between the two. 25 Of interest, we also identified hypertension and chronic cardiovascular insufficiency as risk factors. Patients with cardiovascular disease may be hampered in raising or maintaining core temperature by a dysfunction in autonomic mechanisms such as increased heart rate and blood pressure and by shifting capillary blood flow from cutaneous to deep vascular beds. 26 Alternatively, the association between hypothermia and cardiovascular conditions may reflect the importance of an intact endothelial function in maintaining body temperature during sepsis. Interestingly, systemic fractalkine levels were significantly higher in hypothermic patients compared to nonhypothermic patients, and this difference was maintained after correcting for disease severity. Fractalkine is a chemokine that has been implicated as a mediator in a diverse spectrum of inflammatory conditions. 27 In critically ill patients with sepsis, increased levels of fractalkine are associated with adverse outcome. 27 Arterial and capillary endothelial cells have been identified as an important source of fractalkine during endotoxemia. 28 Also, levels of angiopoetin-2 and the ratio of angiopoietin-2/angiopoietin-1, which indicate impaired vascular integrity, were increased in the hypothermic patients compared to the nonhypothermic patients, albeit transiently. Taken together with the presence of mainly cardiovascular risk factors in patients with hypothermia, our data may suggest that the endothelium is somehow implicated in hypothermia through an as yet unknown mechanism. Although the current data cannot establish a causal link between fractalkine and hypothermia and the association of hypothermia and fractalkine in a population with significantly increased disease severity warrants further validation, the increased levels of fractalkine in hypothermia are intriguing and a detailed study on the role of the endothelium, in particular fractalkine, is warranted. There are several shortcomings to this study. First, the timing and method of temperature measurement was not standardized. Although this could have led to an increased variability in this study, we believe this effect will be limited due to the fact that core temperature measurements is standard practice in our ICUs. Second, blood sampling did not exactly coincide with timing of the temperature measurement. Although this might have diluted results, a single hypothermic temperature in 24 hours significantly increases mortality and we feel that the blood sampling does not necessarily need to be on exactly the same time as the hypothermic measurement to characterize this group. Third, results from this study are not applicable to all patients with sepsis as we excluded patients with