DELIBERATE hypothermia is used in a variety of therapeutic settings. Clinical applications of hypothermia include cerebral protection for out-of-hospital cardiac arrest and traumatic brain injury. 1,2Hypothermia is also widely used intraoperatively, primarily for cerebral protection during neurosurgical procedures. 3The rationale for hypothermia is to protect ischemic cells from injury by decreasing their metabolic demands, and, secondarily, to inhibit inflammatory mediator production. 4Whereas this is laudable in areas of focal ischemia, the global implications of hypothermia are significant. Myriad processes are adversely affected by hypothermia, including increased rates of wound infection, 5coagulopathy with increased blood loss, 6adverse cardiac events, 7and even prolonged hospital length of stay. 8In this issue of the Journal, Torossian et al. examine the effects of hypothermia in a rodent model of abdominal sepsis. 9Sepsis was induced with peritoneal contamination and infection with human stool bacteria. The primary outcome measure was survival, and in this clinically relevant model, hypothermia substantially increased mortality. Pretreatment of the rats with granulocyte colony-stimulating factor (G-CSF) completely reversed the hypothermia-induced mortality effect, actually improving survival beyond that seen with normothermia.
Whereas the benefits of decreasing cellular oxygen demands in ischemia are easily understandable, sorting out the implications of hypothermia on complex processes such as wound infection, inflammation, and hemostasis is much more difficult. These processes rely on the coordinated interaction of multiple proteins, whose conformations may be altered by relatively subtle changes in temperature. It is the nature of this interaction that determines the balance between adequate host defense on the one hand, and overwhelming inflammation and multiple organ damage on the other. For example, significant coagulopathy results from even minor degrees of hypothermia. 10The consequences of systemic hypothermia are profound. In patients with trauma, hypothermia is part of the grim prognostic triad of hypothermia, coagulopathy, and metabolic acidosis, and is associated with mortality, independent of fluid administration. 11Therefore, hypothermia in the setting of nonneurologic trauma is clearly harmful, despite that massive trauma represents a clinical entity of global tissue ischemia—further evidence that hypothermia provides benefit primarily in areas of focal ischemia.
Sepsis is the systemic inflammatory response to infection. Like coagulation, the systemic response relies on the complex interaction of multiple proteins. Fever is a characteristic sign of infection, but sepsis may also present as hypothermia, as emphasized in the 2001 International Sepsis Definitions Conference (Washington, D.C., December 8–9, 2001). 12The prognosis of patients with sepsis presenting with hypothermia, like those presenting with leukopenia, is thought to be worse than those presenting with leukocytosis and fever, respectively. Fever, clearly, is an important adaptive response that should be preserved. For example, it has long been clinical practice to treat fever in patients with infection. This practice is now being questioned. Active external cooling of healthy volunteers with fever does not reduce core temperature, but it increases metabolic rate and activates the autonomic nervous system. 13Treatment of fever from viral syndromes with nonsteroidal antiinflammatory drugs may increase viral shedding and duration of symptoms, 14and a randomized, placebo-controlled trial of ibuprofen in patients with severe sepsis did not demonstrate efficacy. 15These and other studies confirm the importance of maintaining host defense in the setting of infection. It is also naïve to treat the inflammatory aspects of the systemic inflammatory response syndrome with high doses of nonspecific antiinflammatory agents, which may diminish the inflammatory response, but, ultimately, cause increased mortality. 16
What are the mechanisms for an abnormal immune response in the setting of hypothermia? Both humoral and cellular immunity are adversely affected with lower temperature. In bacterial infections, neutrophil chemotaxis is an essential component of host defense. Hypothermia inhibits both neutrophil chemotaxis and killing via the respiratory burst, and delays induction of pro-inflammatory cytokine production by macrophages. 17Torossian et al. showed that circulating levels of the cytokine interleukin-6 and the chemokine macrophage inflammatory protein-2 were both increased with hypothermia and ameliorated by either G-CSF administration or normothermia. Although the mortality data are robust, it is premature to speculate on mechanisms. It is tempting to conclude that the harmful effects of hypothermia were reversed via G-CSF's stimulation of neutrophil function, but close examination of the data suggests caution. For example, although neutrophil counts were increased by both hypothermia and G-CSF, functional phagocytic activity was not different in any of the groups. Cytokine and chemokine levels were decreased by normothermia and G-CSF, but little is known about what levels of these proteins are appropriate. It may be that G-CSF's salutary effects were on other leukocytes or on other aspects of neutrophil function. Cytokine concentrations are difficult to interpret in any study; recent data suggest that cytokine levels must be interpreted in the context of time and space, with high levels of pro-inflammatory cytokines advantageous early in infection and harmful later. 18This construct would suggest that it might indeed be harmful to dampen the cytokine response early in infection, when host defense is critical.
How, then, should we proceed in patients in whom induced hypothermia is contemplated and who are at risk for systemic infection? Hypothermia alone clearly results in an increased rate of wound infection, coagulopathy, and other postoperative complications 5; therefore, it should be used sparingly. More promising therapies using focal cooling, especially for the brain, are on the horizon. 19Is it possible to ameliorate the adverse effects of hypothermia with administration of G-CSF or other immune stimulants to these patients? The data to date do not support this practice. However, enough provocative animal and human data now exist to justify a large, randomized trial.