|Home | About | Journals | Submit | Contact Us | Français|
Induced hypothermia after ischemic stroke is a promising neuroprotective therapy and is the most potent in pre-clinical models. Technological limitations and homeostatic mechanisms that maintain core body temperature, however, have limited the clinical application of hypothermia. Advances in intravascular cooling and successful trials of hypothermia after global cerebral ischemia, such as in cardiac arrest and neonatal asphyxia, have renewed interest in hypothermia for stroke.
Based on experimental data and early human experience, hypothermia has been rated one of the most active modes of neuroprotection (Popovic et al., 2000; O'Collins et al., 2006). Hypothermia improves neurological outcome in survivors of cardiac arrest (Bernard et al., 2002; The Hypothermia After Cardiac Arrest Study, 2002) and its use after cardiac arrest is recommended by the International Liaison Committee on Resuscitation (ILCOR) (Nolan et al., 2003). In infants with hypoxic-ischemic encephalopathy, hypothermia of 33.5°C for 72h was safe, reduced fatality, and improved neurodevelopmental outcome (Shankaran et al., 2005).
The COOL-AID study group completed two clinical trials of hypothermia in acute ischemic stroke. The first study used surface cooling (Krieger et al., 2001); the second used endovascular cooling (De Georgia et al., 2004). Both trials were small and showed the feasibility of hypothermia after ischemic stroke, but neither was powered to answer questions regarding safety and efficacy.
The NIH-funded study of Intravascular Cooling in the Treatment of Stroke–Longer tPA window (ICTuS-L) (Guluma et al., 2006) study is currently under way and aims to test the combination of intravenous t-PA and hypothermia (Hemmen et al., 2006).
The Cochrane Database shows no completed, prospective, randomized, controlled trial of hypothermia in stroke patients (Correia et al., 2001). The 2005 guidelines on Acute Stroke Treatment by the American Stroke Association find that hypothermia after stroke a promising field of future research (Adams et al., 2005).
As in other forms of ischemia, the neuroprotective capabilities of hypothermia have long been suspected after ischemic stroke. Victims of near-drowning, who suffer global cerebral hypoxia, can have a remarkable neurological recovery if hypothermia is present (Young et al., 1980). In animal models using permanent focal ischemia, hypothermia shows modest neuroprotective effects (Baker et al., 1991). After temporary ischemia, however, hypothermia leads other putative therapies in reducing infarct size and improving behavioral outcome (Ridenour et al., 1992; Zausinger et al., 2000).
The effect of hypothermia after global cerebral ischemia depends on ischemia duration, and time between ischemia and onset of hypothermia. In experimental models, hypothermia initiated 5min after transient global ischemia protected cortical and hippocampal neurons. Cooling initiated 30min after ischemia was not effective (Ikonomidou et al., 1989). Dietrich et al. (1993) showed that hypothermia demonstrated the greatest benefits when initiated during global ischemia, while hypothermia initiated after global ischemia was not effective.
Although the precise mode of neuroprotective action in hypothermia is not known, it most likely exhibits multiple and synergistic effects on brain metabolism (Krieger and Yenari, 2004). Hypothermia decreases the cerebral metabolic rates of glucose and oxygen, and slows ATP breakdown (Erecinska et al., 2003). In the temperature range of 22–37°C, brain oxygen consumption reduces by approximately 5% for every degree fall in body temperature (Hagerdal et al., 1975). Hypothermia reduces glutamate release (Nakashima and Todd, 1996), inflammation, and free radical generation (Yenari et al., 2002). It lowers metabolic rate, limits edema formation, and interrupts necrosis/apoptosis (Eberspacher et al., 2005; Wang et al., 2005). In addition, hypothermia reduces intracellular calcium rises after ischemia; the mechanism of this effect is not clear. Hypothermia may impair glutamate-mediated calcium influx or directly inhibit calcium-mediated effects on calcium/calmodulin kinase (Hu et al., 1995; Takata et al., 1997).
Hypothermia has significant effects on the production of hydroxyl radicals, and suppresses nitric oxide and peroxynitrite formation. Additionally, granulocyte and adhesion molecule ICAM-1 upregulation in response to ischemia is attenuated (Inamasu et al., 2001) Wang et al. (2002) showed that inflammatory responses are suppressed as late as 1 week after 2h of hypothermia.
In summary, hypothermia may reduce damage from excitotoxins, inflammation, free radicals, and necrosis. This could lead to longer neuronal survival and improved outcome after restoration of blood flow through revascularization methods such as thrombolysis. It may also reduce edema after ischemia and lower the risk of post-ischemia hemorrhage (Lyden et al., 2006).
Animal models of transient and permanent cerebral ischemia have used temperatures of 24–33°C (Krieger and Yenari, 2004). While infarct size was reduced in most models, animals with hypothermia of 28–34°C experienced greater recovery than those with temperatures below 28°C (Maier et al., 1998). This may be caused by a significant reduction of regional cerebral blood flow (CBF) at very low temperatures (Ehrlich et al., 2002).
Hypothermia below 28°C has not been tested in stroke patients. Medical complications such as hypokalemia, arrhythmia, infections, hypothyroidism, and heart failure occur below 28°C (Lyden et al., 2006). Hypothermia to 33°C was used after cardiac arrest in the Bernard et al. (2002) and HACA (The Hypothermia After Cardiac Arrest Study, 2002) studies, and for stroke patients in the COOL-AID pilot studies (Krieger et al., 2001; De Georgia et al., 2004). The ICTUS and ICTUS-L studies use hypothermia to 33°C and avoid endotracheal intubation by using a anti-shivering protocol that causes limited sedation only (Lyden, 2005). Hypothermia of 34–36°C was induced via surface cooling methods and acetaminophen therapy in clinical studies, and has proven safety and feasibility in stroke patients (Kammersgaard et al., 2000; Kasner et al., 2002). A larger study is currently underway to test the efficacy of mild hypothermia via early high-dose acetaminophen therapy in ischemic stroke (van Breda et al., 2005).
Surface cooling methods include convective air blankets, water mattresses, alcohol bathing, cooling jackets, and ice packing. These methods have been used for many years in the treatment of fever. Both cardiac arrest and the neonatal asphyxia studies used surface cooling. Patients with these conditions are usually comatose, endotracheally intubated, and ventilated. Therefore, using suffiecient sedation for anti-shivering therapy or neuromuscular blockade is possible. In the earlier studies of hypothermia in stroke, all patients were intubated (Krieger et al., 2001; Schwab et al., 2001). In general clinical practice, only a minority of stroke patients require endotracheal intubation and ventilation. A successful application of hypothermia in the majority of moderate or severely affected stroke patients would be more practical if intubation could be avoided (Lyden, 2005).
Advantages of surface cooling are that it does not require advanced equipment or expertise in catheter placement and averts the risk associated with central venous catheter placement. Cooling through external methods, however, is slower and for temperatures below 35°C necessitates the use of sedatives and paralytics to prevent discomfort and shivering in most cases. Using paralysis and sedation requires endotracheal intubation and ventilation, which in turn, is associated with significant risk of ventilator-associated pneumonia and other complications. Patients under paralytics render neurological assessments and detection of neuro-worsening during the cooling period virtually impossible. Cooling of the skin leads to vasoconstriction and reduces the heat exchange in cooled patients, which makes temperature control very difficult. This may lead to target temperature overshoot and lack of control during rewarming, which may be associated with reactive brain edema (Krieger et al., 2001).
New surface cooling devices, such as energy transferring skin pads (Mayer et al., 2004), may demonstrate a reduction in the time to target temperature and allow better temperature control, but this has not been confirmed in controlled clinical trials (Zweifler, 2003).
Endovascular cooling has become feasible via newly developed catheters with antithrombotic coverings. These can be inserted into the central venous system and cause cooling via indwelling heat transfer devices (Baumgardner et al., 1999). The new devices exchange heat via transduction through internal circulation within the catheter; no cooled fluid enters the patient (Mack et al., 2003). Thrombotic complications are reduced with the use of the antithrombotic-eluded catheter.
Also, endovascular cooling allows rapid heat exchange and faster cooling towards target temperature. Shivering is in part driven through skin receptors; endovascular cooling allows concurrent skin warming, which reduces shivering to make heat exchange more efficient and temperature control tighter. This control avoids target temperature overshoot and incidental rewarming, which can lead to increase in brain edema and intracranial pressure.
In animal studies, hypothermia during ischemia abolished the ischemic damage completely, and hypothermia within 3h significantly reduced infarct size (Chen et al., 1991; Zhang et al., 1993). Some reports found that delaying hypothermia for more than 3h after ischemia showed no significant neuroprotection (Dietrich et al., 1993), whereas Colbourne and Corbett (1995) showed that cooling over 24h has neuroprotective effects, even when the treatment was delayed by 6h. A recent study in fetal sheep confirmed the neuroprotective effect of delayed hypothermia. In this model of 72h of brain cooling, hypothermia was initiated with a 2- or 6-h delay and compared to sham-operated animals. Both the 6- and 2-h treatment-delay groups showed significant neuroprotective effects (Roelfsema et al., 2004). This long therapeutic window is very promising, as patients often present to the emergency department with a delay of a few hours after stroke onset.
The optimum duration of cooling is not known. Yanamoto et al. (1999) compared cooling for 22 and 3h after temporary middle cerebral artery occlusion (tMCAo). Animals in the 22-h group had better neuroprotection. In the gerbil two-vessel occlusion (2-VO) model, intra-ischemic hypothermia completely prevented hippocampal cell damage when continued for 4–6h; cooling for 2h showed less and 0.5–1h no protective effect (Dietrich et al., 1993). In addition, shorter periods of cooling may lead to only transient neuroprotection. Many experimental models sacrifice the animal within a few days and potentially miss late neuronal cell death after hypothermia. The studies with the best long-term outcome used hypothermia for 24h (Colbourne et al., 1999).
Treatment duration and time window may be interdependent, as 24h of cooling prevented neuronal damage in two models when delayed by up to 3 or 6h (Colbourne et al., 2000). Two hours of cooling was only effective when initiated within 2h, implying a shorter time window for brief hypothermia (Maier et al., 2001).
The thermoregulatory system tightly controls core body temperature near 37°C. Defenses against cooling include shivering and cutaneus vasoconstriction. Cutaneus vasoconstriction reduces heat conduction through the skin; shivering produces energy and heat through repetitive muscle contractions (Jessen, 1980). Shivering is the only heat-producing mechanism in humans and is the main factor in combating hypothermia.
In order to effectively cool patients and to avoid discomfort, shivering must be avoided. Shivering can be reduced with centrally active substances that alter hypothalamic thermoregulatory centers; via skin warming, which reduces the trigger of shivering through skin temperature sensors; and through muscle relaxation. Most anesthetics and narcotics change thermoregulatory control, and in general, the effect on thermoregulatory mechanisms is proportionate to their anesthetic properties. Most drugs that effectively control shivering, therefore, lead to either sedation or respiratory depression (Doufas et al., 2003).
Meperidine is the most important and clinically relevant drug that inhibits thermoregulatory responses (Burks et al., 1980; Alfonsi et al., 1998). The anti-shivering action of meperidine decreases the shivering threshold twice as fast as the vasoconstriction threshold and can be achieved at blood levels that do not lead to severe respiratory depression or sedation. A low dose of meperidine (25mg IV), which does not affect alertness or respiratory function, decreases the shivering threshold by 2°C (Kurz et al., 1993).
Mokhtarani et al. (2001) showed a synergistic effect of buspirone and meperidine. Buspirone reduces the shivering threshold to 35.0±0.8°C, high-dose meperidine reduces the threshold to 33.4±0.3°C, and the combination of buspirone and low-dose meperidine reduces the shivering threshold to 33.4±0.7°C (Mokhtarani et al., 2001). The combination did not cause sedation or respiratory depression. The mechanism of this synergistic effect is unknown. In addition to these pharmacological therapies, warming the skin (especially hands, feet, and face) during endovascular hypothermia treatment is highly affective in reducing shivering and patient discomfort (Cheng et al., 1995).
To be able to effectively cool patients who are awake, we combine a low-dose meperidine with buspirone and skin warming, which allows cooling to 33°C without significant sedation or respiratory depression in patients after ischemic stroke (Lyden et al., 2005).
In recent years, other pharmacological therapies have shown promise in the treatment of shivering after accidental or induced hypothermia (Mahmood and Zweifler, 2007). Clonidine and dexmedetomidine, two alpha-2 agonists, and Tramadol, a serotonin reuptake inhibitor, reduce shivering without significant respiratory depression or sedation when used in conjunction with meperidine (Delaunay et al., 1993; Doufas et al., 2003; Trekova et al., 2004).
Hypothermia is more effective in models of transient ischemia (Ridenour et al., 1992). Combining hypothermia with reperfusion therapy may be the most promising strategy in humans (Lyden et al., 2006).
However, the activity of t-PA may be reduced in hypothermia. In vitro analysis shows that cooling to 30–33°C decreases t-PA activity by 2–4% (Yenari et al., 1995). Wolberg et al. (2004) reported that t-PA activity declined to 50% when the clot temperature decreased from 40°C to 30°C. While thrombolysis is reduced by low temperature, platelet function deceases when temperature is lowered from 37°C to 33°C (Wolberg et al., 2004), and in trauma patients, platelet function and coagulation activity are decreased at temperatures below 34°C (Watts et al., 1998).
The experience regarding safety and efficacy of thrombolytic use and hypothermia is limited in stroke patients. The data available so far are insufficient to address the effect of hypothermia on t-PA activity and the resulting safety. The ICTuS-L study will provide further feasibility and early safety data regarding the use of thrombolysis in stroke patients treated with hypothermia, using state-of-the-art endovascular heat exchangers and intravenous t-PA (Lyden, 2005).
Whether hypothermia affects hemorrhagic complications after stroke is unknown. One could hypothesize opposite effects. On the one hand, cooling may reduce hemorrhage after stroke by reducing the activities of matrix metalloproteinases after ischemia (Sumii and Lo, 2002). On the other hand, cooling might reduce the activity of the rt-PA inhibitors, such as plasminogen activator-1, sufficiently that the lytic effect of rt-PA is enhanced, resulting in more hemorrhages. Again, trials such as ICTsS-L will begin to generate data on this point.
For an increase in treatment effect, it seems reasonable to combine neuroprotective strategies (Rogalewski et al., 2006). Over the last decades, many neuroprotective trials were undertaken and failed to prove clinical benefit in patients after stroke (Cheng et al., 2004). It is possible that the tested compounds failed to show a significant benefit when used alone. Finding synergistically active neuroprotective therapies could produce more potent effect after ischemia. Hypothermia is an ideal addition to neuroprotective compounds, and in animal models, the combination of hypothermia with tirilazad and magnesium was effective in reducing infarct size (Zausinger et al., 2003). The combination of hypothermia, caffeine, and alcohol (caffeinol) is currently being tested (Martin-Schild et al., 2008).
Neuroprotective strategies after ischemic stroke remain a promising field for therapeutic interventions, and hypothermia is the strongest available neuroprotective therapy in pre-clinical studies. Advances in hypothermia delivery using endovascular heat exchangers and novel anti-shivering protocols make routine clinical use of hypothermia after ischemic stroke possible (Hemmen and Lyden, 2007). Studies are underway to investigate the effect of hypothermia on clinical outcome. Future developments will also include combining hypothermia with thrombolysis and other neuroprotective strategies.