The Yin and Yang of Hypothermia in Trauma

Abstract

The relationship between hypothermia and outcome from severe trauma remains quite unclear, as hypothermia seems to be a double-edged sword. Clinical experience suggests that hypothermia is detrimental to patients, whereas multiple laboratory studies have demonstrated benefit. Translating the laboratory findings into positive results in clinical trials has proven to be challenging and frustrating. In this issue of the journal, Finkelstein and Alam have presented a comprehensive review of our current understanding of this fickle relationship between hypothermia and outcome in patients with trauma. The key points they make should be emphasized and put in perspective. What does all this mean and where do we go from here? Before proceeding, it is worth clarifying a few definitions. First is timing of hypothermia. Peter Safar has called treatment before a traumatic or ischemic insult ‘‘protection,’’ treatment during the insult ‘‘preservation,’’ and treatment after the insult ‘‘resuscitation.’’ For instance, use of hypothermia for protection during cardiac surgery is very different from hypothermia for resuscitation from cardiac arrest. Next is the depth of hypothermia. The definitions used by resuscitation researchers are somewhat arbitrary but different from those applied to the patient with exposure hypothermia. Finkelstein and Alam define mild hypothermia as 33 C to 36C; moderate 28 C to 32 C; deep 16 C to 27 C (we have used this term down to 11 C); profound 6 C to 15 C; and ultraprofound <5 C. The duration of hypothermia and the rapidity of cooling and rewarming are additional factors to consider. The insult to be discussed also should be clearly defined. ‘‘Cardiac arrest’’ refers to cessation of cardiac function. ‘‘Circulatory arrest’’ refers to cessation of all circulatory flow. This distinction is important, as ‘‘cardiac arrest’’ may not be important if cardiopulmonary bypass (CPB) is used, even with low flows at low temperatures; but when the pump stops (circulatory arrest), organ ischemia ensues. The concept of ‘‘metabolic arrest,’’ that is, inducing a state of zero metabolic demands, has not yet been achieved experimentally. Even at levels of profound hypothermic levels, cellular metabolic demands are not 0. ‘‘Induction of a state of hypometabolism’’ might be a better term. The mechanisms of action of hypothermia during and after traumatic or ischemic insults are multiple and poorly understood. Because small degrees of temperature difference can readily lead to outcome differences, it is doubtful that the main effect of hypothermia is decreasing cellular oxygen demands to better meet limited supply. Mild levels of hypothermia can affect many detrimental pathways, some systemic and some specific to individual organs. These multiple effects may have a greater impact on the outcome from life-threatening conditions, such as hemorrhagic shock or traumatic brain injury (TBI), than agents that target only 1 pathway. Patients with trauma are at significant risk of developing hypothermia for a number of reasons. These include exposure and infusion of cold fluids, complicated by an inability to produce heat because of shock, alcohol and drug intoxication, sedation, and anesthesia. In addition, hypothermia may just represent the end result of energy depletion from shock. The degree and risk of hypothermia do correlate with injury severity. Multiple retrospective clinical studies have demonstrated an association between the development of hypothermia in patients with trauma and increased mortality. The cause of this association has not been determined. Patients may develop shivering, catecholamine responses, and coagulopathy which may contribute to the detrimental effect of hypothermia. For trauma patients with active hemorrhage, hypothermia has 2 potential uses based on laboratory studies. The first use could be induction of mild, controlled hypothermia to delay or prevent cardiac arrest. In clinically relevant large animal models, hypothermia can improve survival. Reconciling these findings with the clinical experience is not straightforward. In contrast to patients with severe trauma, laboratory models often lack significant tissue trauma, use fresh whole blood (the animal’s shed blood) for resuscitation to prevent coagulopathy, and prevent shivering and catecholamine responses with anesthesia. Under controlled conditions with patients, shivering and catecholamine responses could similarly be mitigated and coagulopathy