Hypothermia Used in Medical Applications for Brain and Spinal Cord Injury Patients
- 1.3k Downloads
Despite more than 80 years of animal experiments and clinical practice, efficacy of hypothermia in improving treatment outcomes in patients suffering from cell and tissue damage caused by ischemia is still ongoing. This review will first describe the history of utilizing cooling in medical treatment, followed by chemical and biochemical mechanisms of cooling that can lead to neuroprotection often observed in animal studies and some clinical studies. The next sections will be focused on current cooling approaches/devices, as well as cooling parameters recommended by researchers and clinicians. Animal and clinical studies of implementing hypothermia to spinal cord and brain tissue injury patients are presented next. This section will review the latest outcomes of hypothermia in treating patients suffering from traumatic brain injury (TBI), spinal cord injury (SCI), stroke, cardiopulmonary surgery, and cardiac arrest, followed by a summary of available evidence regarding both demonstrated neuroprotection and potential risks of hypothermia. Contributions from bioengineers to the field of hypothermia in medical treatment will be discussed in the last section of this review. Overall, an accumulating body of clinical evidence along with several decades of animal research and mathematical simulations has documented that the efficacy of hypothermia is dependent on achieving a reduced temperature in the target tissue before or soon after the injury-precipitating event. Mild hypothermia with temperature reduction of several degrees Celsius is as effective as modest or deep hypothermia in providing therapeutic benefit without introducing collateral/systemic complications. It is widely demonstrated that the rewarming rate must be controlled to be lower than 0.5 °C/h to avoid mismatch between local blood perfusion and metabolism. In the past several decades, many different cooling methods and devices have been designed, tested, and used in medical treatments with mixed results. Accurately designing treatment protocols to achieve specific cooling outcomes requires collaboration among engineers, researchers, and clinicians. Although this problem is quite challenging, it presents a major opportunity for bioengineers to create methods and devices that quickly and safely produce hypothermia in targeted tissue regions without interfering with routine medical treatment.
1 History of Cooling as a Treatment Tool in Medicine
Unlike heating, which has been used since ancient times as an infection prevention method for treating wounds, the medical benefits of cooling were not well understood and implemented until the recent centuries. It has been documented that cooling was used to minimize bleeding and to fight fevers. During Napoleon’s war against Russia in the nineteenth century, it was an observation by physicians that the soldiers left in snow had an improved survival rate than those given a warm blanket and/or placed close to a fire. They also observed that fast thawing frostbite extremities directly over open fires expedited tissue destruction. Those lessons learned in the early wars were implemented in the medical practice of the twentieth century, which included development of various devices that actively removed heat from the body with a great level of temperature control. One pioneer in this field is Dr. Temple Fay, who hypothesized that tumor cell multiplication and progression could be inhibited by cooling, since it was rare to see tumors developing in the body extremities. He later tested this hypothesis in cancer patients. He also implemented cooling for pain relief. In the 1940s, Dr. Fay published the first case study of using hypothermia for cerebral injury patients (Fay 1945). Along a similar timeline, industrial development of refrigeration systems began in the 1930s, which assisted medical professionals with better implementation and control of cooling methods to patients than previous passive techniques, such as whole-body immersion into cold water or ice baths.
Human tissue requires a constant supply of oxygen-rich blood. Among all tissues, heart and brain tissues in the head and spinal regions are most valuable to ischemia. Unlike liver or muscle tissues, a shortage of oxygen delivery to the brain would immediately impair neuronal functions since the brain does not have a storage for glucose. It is well documented that brain tissue cannot recover after only 5 min of normothermic (37 °C) brain ischemia, but this duration increases to more than 45 min when the brain tissue temperature is lowered to 10 °C (Wass et al. 1995). Therefore, induced hypothermia has been studied as a means to protect the brain from ischemia since the 1940s (Kabat 1940; Fay 1945; Miller 1949; Bigelow et al. 1950; Lazorthes and Campan 1958; Sedzimir 1959; Vandam and Burnap 1959a, b; Drake and Jory 1962). Another application is in open-heart and open-neck surgeries when surgeons need to repair malformation in those regions. Surgeons could work efficiently if they operate on a bloodless field and minimize the risk of hemorrhage. In the 1950s and 1960s, cardiopulmonary bypass in conjunction with induced hypothermia was usually employed to provide cardiac surgeons sufficient time for many neuro-procedures while avoiding permanent brain/heart/spinal cord damage. Another field was in cerebral resuscitation, where cerebral hypothermia was introduced to cardiac arrest patients in order to maximize the ability of the brain tissue to survive anoxic no-flow states. In the 1970s, numerous animal experiments and clinical studies showed the benefits of hypothermia in prolonging tolerance of brain tissue to ischemia. Hypothermia contributed significantly to reduced mortality associated with previously inoperable cardiac and cerebral pathologies. However, most of the neuroprotective effects of lowering body temperature have been more evident in animal experiments than in patients.
Early studies commonly used deep states of hypothermia with temperatures ≤30 °C. For example, Dr. Fay’s initial experiment lowered the body temperature of the patient to 27 °C. Deep hypothermia introduced detrimental systemic complications, such as pneumonia, ventricular fibrillation, and acidosis, which are especially fatal to elderly and fragile patients. In the 1970s and 1980s, research in hypothermia to enhance brain tissue recovery after ischemic attack became dormant. This may be due to advancements in medicine to identify alternative and less risky treatment methods. In addition, thermal management difficulties made it hard to demonstrate benefits of deep hypothermia on neuroprotection. In the 1980s, researchers began to reinvestigate the idea of mild hypothermia since it can be relatively easy to control cooling extent and implement cooling methods. Extensive experimental studies on canine, swine, and rodent models have been conducted to evaluate efficacies of different cooling methods, temperature reductions, hypothermia durations, cooling initiation relative to the ischemic event, and rewarming rates on tissue recovery from ischemia. The neuroprotective mechanisms of hyperthermia have become increasingly better understood in the past three decades due to those animal experiments. Unfortunately, supporting data in human subjects are still limited, especially for randomized multicenter clinical trials (Clifton et al. 2001, 2011). In clinical studies, hypothermia therapy seems more successful in open-heart and open-neck surgery than in traumatic head injury. This is not unexpected, because cooling initiation after traumatic head injury is usually implemented several hours following the event, which may be too late to maximize the benefits conferred by hypothermia. However, these challenges can create opportunity for future investigations to develop more effective and well-controlled approaches for treating tissue ischemia. It is anticipated that hypothermia will be proven as an effective method to limit and eliminate injury and death associated with ischemic injury and benefit a vast patient population.
2 Biological and Chemical Reactions to Injury and Neuroprotective Mechanisms Associated with Cooling
Tissue ischemia can be the result of ischemic stroke, cardiovascular and respiratory disorders, and external physical trauma. If the initial damage is limited, the tissue may be able to recover. If the injury is extensive, secondary tissue damage occurs, which includes intracranial hypertension (brain swelling), hemorrhage, hypoxia, and edema. It is well known that brain oxygen stores become exhausted within 15 s and brain energy stores become exhausted within 5 min after global ischemia (Wass et al. 1995). Energy loss in tissue results in depolarization of cell membranes. A series of biochemical reactions and cascades initiated by the trauma will then follow. Those cascade events evolve gradually and may last several days after the initial trauma. The increase in extracellular K+, energy depletion, disruption of the blood-brain barrier, free radical and glutamate release, excitotoxicity, and inflammation are typical consequences of those cascade events. The loss of selective neurons following global or local brain ischemia may lead to permanent neurologic deficit and even death. Thus, secondary injury can be more fatal to patients than the initial injury, and it is not uncommon to hear news reports on deaths occurring several hours or several days after the initial trauma.
Originally, researchers suggested that cooling may decrease blood perfusion in the injured region, similar to cold-induced halting of bleeding. Brain oxygen consumption can decrease by approximately 5–7% for every degree Celsius decrease in tissue temperature. Thus, a reduction in the energy expenditures of cerebral metabolic rates of glucose and oxygen was observed in deep hypothermia implementations. Later, people started to question this mechanism due to the observed benefits associated with mild hypothermia of lowering tissue temperature by only 1 or 2 °C. In the past several centuries, many researchers began to suggest that cooling may play an important role in deterring deleterious biochemical actions in secondary tissue injury triggered by the initial trauma. It has been demonstrated that hypothermia may modify a wide range of cell necrosis mechanisms (Marion et al. 1996; Maier et al. 2002; Yenari et al. 2000). Cooling retards progression of the ischemic cascade and pathological neuroexcitation. In brain tissue hypothermia, cooling may attenuate the opening of the blood-brain barrier, reduce glutamate release, inhibit inflammation, and slow down free radical generation and release. It may also impair glutamate-mediated calcium influx or directly inhibit calcium-mediated effects on calcium/calmodulin kinase. As a result, hypothermia preserves high-energy phosphates that may facilitate the maintenance of membrane integrity during ischemia, limit edema formation, lower intracranial pressure, and interrupt necrosis and apoptosis (Xu et al. 2002). All of the mechanisms lead to prolonged cell survival and can improve outcomes after reperfusion and rewarming.
3 Cooling Parameters Affecting the Protective Outcomes of Hypothermia
Numerous controlled animal experiments have demonstrated the benefit of initiation of cooling before the brain injury to confer significant neuroprotective outcomes. The general consensus is to initiate brain hypothermia as early as possible following an ischemia-precipitating event, although some studies have shown neuroprotective effects even when the treatment was delayed by up to 6 h (Colbourne and Corbett 1995). Brain injury mechanisms typically progress rapidly within 3–6 h after the initial injury. Hypothermia can reduce the initial inflammatory response after head trauma and minimize or prevent secondary brain injury. It is well accepted that for animal models, there is a 1–2 h treatment window after which the benefits of hypothermia are strongly diminished. However, it is difficult to correlate animal data directly to humans. Unfortunately, it is still a challenge to shorten the time between the injury and induction of hypothermia. Usually, the process of transporting the patient to the hospital is unproductive for this objective. A portable sensor attached to the patient that alerts the medical center of the impending case can minimize delay of treatment. Further, simple cooling approaches that can be implemented by the emergency medical service (EMS) personnel in the ambulance have been suggested to achieve this goal.
Precooling is an option only for circumstances under which the onset of tissue injury is foreseen. It is a clinical option for open-heart and open-neck surgical procedures. Experimental data from a rat model demonstrated the greatest benefits of neuroprotection when hypothermia was induced during global ischemia (Dietrich et al. 1993). Additionally, initiation of hypothermia before the injury completely prevented secondary brain damage in gerbils (Welsh et al. 1990).
Because tissue temperature has profound effects on local metabolism and cellular activities, it is expected that different depths of hypothermia have varying effects on the clinical outcome of minimizing secondary tissue injury. Hypothermia can be categorized as deep or severe (less than <28 °C), moderate (28–32 °C), and mild (32–35 °C) (Kirkpatrick et al. 1999). The traditional view that colder is better (Michenfelder 1988) has been questioned in the past decades (Tisherman et al. 1999). As brain temperature decreases, great systemic toxicity is observed, and the adverse effects outweigh the neuroprotection mechanisms associated with therapeutic hypothermia (Chambers 1999). Deep hypothermia is associated with arrhythmias (Mouritzen and Anderson 1966), cardiac complications (Frank et al. 1993), coagulopathies (Rohrer and Natale 1992), and pulmonary infections (Bohn et al. 1986). In addition, severe or moderate cooling usually requires sedation and mechanical ventilation. Therefore, deep cooling approaches are limited to facilities that have intensive care units and situations that medically justify the collateral complications. On the other hand, there are numerous studies in recent years suggesting that the preferred cooling temperature for neuroprotection is between 32 and 35 °C (Colbourne and Corbett 1995; Gunn and Gunn 1998). Many studies have shown improved neurological outcomes using mild hypothermia, compared with either the deep cooling group or the normothermia group. Additional studies are needed to determine which targeted temperature is most associated with the beneficial effects of hypothermia in patients suffering from specific injury.
Secondary effects of brain injury, including edema and elevated pressure, are known to persist several days after focal cerebral ischemia (Dirnagl et al. 1999). Therefore, prolonging brain hypothermia therapy may benefit those patients. Experimental studies have tested whether a long cooling duration is safe for patients. A duration of 24 h was proposed initially, and it was later extended to 48 h by Bernard and Rosalion (2008), Clifton et al. (1993), McIntyre et al. (2003), and Shiozaki et al. (1993). A review by Peterson et al. (2008) documents profound neuroprotective benefits observed in patients who underwent more than 48 h of hypothermia. Although prolonged cooling can maximize clinical neurological benefits, it may also increase systemic complications associated with cooling, which may be more severe in patients with compromised health. The optimal duration of treatment remains unknown for clinical applications, especially when other confounding factors are involved. Closely monitoring patients during hypothermia therapy should be an important planning consideration.
The importance of controlling the rewarming rate in the brain tissue from a state of hypothermia has been widely documented. Rapid rewarming may result in a dangerous rebound of intracranial pressure elevation and cerebral perfusion pressure reduction; the importance of gradual rewarming has been emphasized in multiple clinical studies (Shiozaki et al. 1993). Previous theoretical analysis (Diao et al. 2003) of simulated passive (and uncontrolled) rewarming by removal of a head-cooling device yields an average calculated rate of approximately 3 °C/h. Animal experiments conducted on rats (Diao and Zhu 2006) suggest that a fast rewarming rate might result in a mismatch between local blood perfusion and metabolism. As suggested in a study (Thoresen and Wyatt 1997), the rewarming rate in tissue should be conducted slowly at 0.25–0.5 °C/h. In some clinical trials, rewarming from hypothermia was conducted with a feedback control system over 18 h. It is critical to develop a thermoregulating system that allows not only fast cooling but also an adjustable and controlled rewarming rate.
4 Currently Used Cooling Approaches/Devices
The ability to cool deeply embedded interior tissues, such as the brain and spinal cord, is limited by considerations of the second law of thermodynamics. Unlike the heating process, which can be dissipative, cooling requires that a negative temperature gradient be applied across the interstitial transport medium. If conduction is used to drive the heat transfer process, the magnitude of cooling that can be produced is directly proportional to the gradient established across the medium and the thermal conductivity of the medium. Alternatively, under normothermic conditions, high blood perfusion rates in tissue regions can be used as the primary mechanism to remove heat from the body core and redistribute it to the surface area. This mechanism is dependent on the level of blood perfusion through the cutaneous circulation in conjunction with conduction through the skin and cooling on the surface. Unfortunately, a high blood perfusion rate in the brain or spinal cord acts to continuously rewarm the tissue, thus inhibiting cooling penetration from a cold front and limiting temperature reduction in those regions. A different option for cooling is by convection, which involves introducing a chilled perfusate solution directly into the circulatory system. However, this approach is highly invasive and has its inherent risks. Both approaches have been adopted for inducing tissue hypothermia, but heat removal by conduction through the overlying tissue has limited efficacy due to the poor thermal conductivity of the body tissue. Both approaches have been implemented in reducing tissue temperature either via systemic (whole-body) or selective (local) cooling. In general, the faster the achieved cooling, the more invasive the involved procedures.
Most of the clinical studies to date have examined systemic (whole-body) hypothermia. Either the whole-body mass or the blood in circulation is cooled using this approach. The major methodological drawback of this approach is the inherently slow cooling rate (∼0.5 °C/h) due to the large volume of the human body to be cooled and the increase in thermal resistance due to arteriovenous shunt vasoconstriction (Marion et al. 1997; Schwab et al. 1998a, b; Krieger et al. 2001). Among all methods, skin surface cooling is the most studied in hypothermia. Various techniques have been used to cool the skin, including convective air circulation, cooling blankets or jackets, water mattresses, ice packs, and alcohol or ice-water bathing. Surface cooling is an easy approach and does not require sophisticated equipment. The side effects include shivering and vasoconstriction. Full-body immersion into cold water makes it difficult to attach monitoring sensors to patients (Olsen et al. 2003). Also, vasoconstriction induced by skin surface cooling greatly increases the thermal resistance of the skin, further hindering heat transfer from the body core to the surface.
Direct cooling of the blood in circulation is an attractive technique because it bypasses the thermal resistance of superficial tissue layers. Intravascular catheters have been proposed as a method to induce a fast cooling rate (Georgiadia et al. 2001; Dixon et al. 2002; Doufas et al. 2002; Inderbitzen et al. 2002; Dae et al. 2003; Mack et al. 2003; Holzer et al. 2005). This method uses a covered cooling catheter that is inserted into the femoral vein and advanced to the inferior vena cava via abdominal X-ray guidance. Coolant is pumped into the catheter to achieve fast heat removal from the venous blood and with a cooling capacity of up to 150 W. Since its introduction to clinical trials, intravascular catheters have been tested in both brain and spinal cord injury patients, and their safety and effectiveness have been well documented. The drawback of this approach is the invasive nature of the surgical procedures involved, which can vary depending on the skill of the surgeon. Therefore, its clinical use has been limited to special hospitals, and it is not a technology available to personnel with limited training and resources (Georgiadia et al. 2001).
Chilled saline has been suggested as a preferred resuscitation fluid for patients with neurological and neurosurgical injuries (Polderman et al. 2005; Kollmar et al. 2009). It is inexpensive, easy to store, and frequently utilized as hydration fluid in hospitalized patients. This method of inducing hypothermia was tested in a swine model (Vanden Hoek et al. 2004). For a high intravenous infusion rate of 120 mL/min, a core cooling rate was measured at up to 18 °C/h. In a clinical study (Virkkunen et al. 2004), Ringer’s solution was infused at a rate of 100 mL/min to patients after resuscitation from nontraumatic cardiac arrest, and no serious adverse hemodynamic complications occurred. This approach is found to be very effective in achieving fast cooling. Intravenous coolants administered in an on-demand, temperature-guided, and supervised setting seem to be a reasonable approach for core cooling that avoids potentially unsafe use of extended fluid volumes and infusion time periods. However, it is questionable whether the patients can tolerate such a large infusion rate of fluid, especially in some patients with compromised kidney functions.
In sports medicine, controlling temperature elevation in the body may be related to athletic performance (Enomoto et al. 1996; Grahn et al. 1998, 2005, 2008). One technology uses an applied negative pressure on glabrous skin of the hand to mechanically distend arteriovenous anastomoses, thereby greatly increasing the cutaneous blood perfusion. By diverting a significant fraction of the cardiac output to the skin, where there can be effective heat transfer with the environment, it is possible to use the circulatory system as a convective conduit for thermal energy between the body surface and the thermal core, including the brain. In initial studies with humans, it has been possible to achieve cooling rates of core temperature of 10 °C/h and higher, which would be adequate for hypothermia induction in cases of TBI. One concern raised by neurosurgeons is the adverse effect of decreased blood pressure, which can be dangerous to patients with compromised health. However, this approach may still have potential to decrease body temperature and induce hypothermia at targeted brain or spinal cord regions without interference with simultaneous treatments at those targeted locations.
Localized cooling has been utilized in the past decades for patients suffering from tissue injury. It is the brain or spinal tissue that requires temperature reduction, not the entire body mass. Therefore, local cooling has been utilized in the past when the targeted tissue region becomes accessible in order to implement the cooling devices. Following spinal cord injury, the injured region is often exposed to allow surgical decompression. Hypothermia to the injured spinal cord region can be easily introduced using epidural heat exchanger or infusion of cold saline solution to the subarachnoid space. It has been reported that incorporating a cooling device into a routine neurocritical care device for monitoring interstitial pressure and/or draining ventricular fluid allows direct cooling of the injured intraparenchymal brain region (Zhu and Rosengart 2008). Cooling the dura surface directly was also proposed when the brain hemisphere is exposed. Selective brain cooling by targeting temperature reduction in the brain tissue alone has been well studied in the past. Feasible approaches include head surface cooling relying on cooling penetration from the scalp to the brain tissue, neck collar or interstitial cooling device targeting arterial blood supplied to the brain tissue, intracarotid flushing, and cooling in the ventricular space. Cooling on the head and neck surface requires 1 or 2 h to reach steady state with a penetration depth limited to the superficial regions of the brain cortex (Wang et al. 2004; Eginton 2007; Poli et al. 2013). An interstitial cooling device placed directly on the surface of the common carotid artery was proposed and tested (Wang and Zhu 2007, Wang et al. 2008, Wei et al. 2008). The approach might be too invasive for stroke patients, but it may be suitable for open-neck surgery when the carotid artery has already been exposed. Similar to the intravascular cooling catheter, intracarotid flushing advances a cooling catheter to the carotid artery in order to induce rapid cooling (4 °C temperature reduction in cerebral cortex within 10 min) (Ding et al. 2004). An endovascular cooling catheter may sound very invasive, but the rationale of implementing it was based on frequent clinical use of endovascular recanalization during some brain injury situations or cardiopulmonary bypass surgeries requiring circulatory arrest (Bachet and Guilmet 2002; Choi et al. 2006). One clinical trial was conducted to infuse cold saline (4–17 °C) at a rate of 33 mL/min (Choi et al. 2010). Similar to intravenous flushing, administration of a large volume of fluid may cause systemic volume overload and hemodilution (Esposito et al. 2014). The measured temperature reduction at the jugular venous bulb was lower than 1 °C after 10 min, which is smaller than theoretical predictions. The authors argued that the jugular venous bulb temperature does not correlate well with the deep brain temperature. Nasopharyngeal cooling takes advantage of the dense blood vessels in the nasal cavity, and cooling in the cavity may penetrate into the nearby brain regions (Esposito et al. 2014). Ventilation of cold air into the nasal cavity may not be very effective due to small thermal capacity of air. On the other hand, inserting a cooling catheter with coolant (RhinoChill™) seems effective in reducing the temperature of that region. Results from clinical trials on implementing the nasopharyngeal cooling were not very encouraging due to very small temperature drops and concerns on increasing blood pressure in patients (Andrews et al. 2005; Harris et al. 2007; Abou-Chebl et al. 2011; Covaciu et al. 2011; Poli et al. 2013; Springborg et al. 2013). Another approach is to cool the cerebrospinal fluid (CSF) in order to reduce the temperature of the brain. One way of implementation is to insert a cooling catheter into the lateral ventricles of the brain while maintaining normal body temperature. This method has been tested in sheep, and it reduced the brain temperature to 34.5 °C after 3 h (Moomiaie et al. 2012). A clinical study was performed to test ventricular infusion on O2 supply to the brain region, and a brain temperature drop was observed. A mathematical simulation was performed to explore the possibility of cooling the CSF on the back of the torso, resulting in cooling of the brain (Smith and Zhu 2010b).
5 Applications with Evaluation of Clinical Outcomes of Tissue Hypothermia
5.1 Traumatic Brain Injury
Traumatic brain injury (TBI) is a nondegenerative, noncongenital insult to the brain tissue due to externally inflicted trauma (McIntyre et al. 2003). The major consequences of head injury include skull fractures, intracranial hemorrhages, elevated intracranial pressure, and cerebral contusion. Unlike stroke, which is often associated with the elderly, TBI affects a predominantly young population, as TBIs occur mostly as a result of motor vehicle accidents. Neuroprotective outcomes in clinical trials that treated head injury using hypothermia are not consistent. There is marked heterogeneity among available clinical studies (Shaefi et al. 2016).
The lack of conclusive results of improved neurologic functions in TBI patients has deterred clinicians from implementing cooling as a standard treatment. In 1994, a multicenter, randomized, and prospective phase III study of systemic hypothermia in severe brain injury that involved seven medical centers in the United States (Clifton et al. 2001) was initiated based on the promise of decades of somewhat sporadic experiments and anecdotal positive clinical outcomes for hypothermia used to treat TBI in addition to a previously extensive phase II clinical trial (Clifton et al. 1993). This study enrolled 392 patients, ages 16–65, who suffered a coma from brain injury in the absence of other major traumas. A control group received standard management with a nominal core temperature of 37 °C. The treatment group received the same management protocol at a target temperature of 33 °C. There was a large variation in the cooling induction time due to many individual circumstances. The cooling process began at 4.1 ± 1.2 h post-injury via surface cooling technology and completed by 8.3 ± 3.0 h. Hypothermia was maintained for 48 h and was followed by slow rewarming distributed over 18 h. Some of the patients were already hypothermic (≤35 °C) owing to independent circumstances on admission to the hospital and were retained in a treatment subgroup. Overall, very minimal difference was observed between the treatment and control groups (Clifton et al. 2001), with the exception of significantly better outcomes from the subgroup of individuals already hypothermic at admission (Clifton et al. 2002). Although the results of this study were discouraging, they directly led to an enhanced appreciation for the need of achieving a hypothermic state quickly following a TBI in order for the therapy to be effective (Marion et al. 1997; Markgraf et al. 2001). Thus, the stage was set for current developments in the application of hypothermia for TBI. Unfortunately, a recent large clinical trial by the Clifton group in 2011 found no significant difference in outcomes between hypothermia and normothermia TBI patients even though targeted temperature reduction was achieved within 5 h, hypothermia was maintained for a long duration of 48 h, and rewarming was very gradual at 0.25 °C/h (Clifton et al. 2011). This clinical trial only suggested a possible benefit of hypothermia to patients with surgically resectable lesions (Clifton et al. 2011). Another randomized large clinical trial on severe TBI children in 2008 showed that hypothermia does not improve outcomes of those children and may also contribute to higher mortality in the participants of the cooling group (Hutchison et al. 2008). The results are disappointing since the trial enrolled a large patient group of 225 children and focused on early initiation of cooling (6.3 ± 2.3 h after injury) and slow rewarming (0.5 °C/h). Similarly, two recent clinical trials in Japan and Europe also failed to identify neurological benefits of hypothermia in treating TBI patients (Andrews et al. 2015; Maekawa et al. 2015).
In the meantime, several clinical studies (Shiozaki et al. 1993; Marion et al. 1997; Metz et al. 1996) tested hypothermia therapy in patients with severe head injury and achieved improved outcomes compared to normothermia groups. Hypothermia benefits were evident in patients with a coma score of 5–6 at months 3 and 6 (Marion et al. 1997). It is reported that more than 62% of patients in the hypothermia group achieved good outcomes versus 39% in the control group. Jiang et al. (2000) evaluated the long-term benefits of hypothermia in patients with severe traumatic head injury. It was found that mortality decreased by 40% and the rate of favorable outcome increased by 70% in the hypothermia group 1 year after the head injury. One of the common benefits identified in the above studies was a marked reduction of the intracranial pressure in the hypothermia group after days of treatment. A recent study (Puccio et al. 2009) explored the use of cooling to counteract fever typically following TBI. Fever may exacerbate secondary injury severity after the initial injury by inducing excitotoxicity, free radical production, blood-brain barrier breakdown, cytoskeletal proteolysis, and inhibition of protein kinases. It has been found that higher fever was associated with higher risk of death in stroke and TBI patients (Saxena et al. 2015). In the Puccio study (Puccio et al. 2009), hypothermia reduced the fever burden in those patients, which may have potentially attenuated secondary injury. It was recommended that cooling should be maintained at least until after the intracranial pressure returns to the normal range (Shiozaki et al. 1993; Jiang et al. 2000). Some researchers (Safar and Kochanek 2001) attributed the failure to demonstrate the benefits by hypothermia to the delay of cooling initiation, which is often more than 4 h after the injury in clinical treatments (Clifton et al. 2001). A clinical trial by Gal et al. (2002) showed that more than 87% of patients achieved good neurological outcome in the hypothermia group, suggesting certain methodological discrepancies in the Clifton study. A clinical trial of hypothermia on traumatic head injury (Qiu et al. 2005) showed improved extradural pressure and reduced mortality from 51% in the control group (43 patients) to 26% in the hypothermia group (also 43 patients). Another clinical study by the same group (Liu et al. 2006) documented improved neurological outcomes in TBI patients 2 years after the hypothermia treatment, including complications that were manageable. A review by Peterson et al. (2008) of hypothermia-related neuroprotection on TBI patients identifies a cooling duration longer than 48 h as a favorable factor; however, it cautions the risk of pneumonia. The results also suggest favorable effects of hypothermia on the patients in trials having long-term follow-up periods of 1 or 2 years. There was one recent clinical trial conducted in China (Zhao et al. 2011) that attributed improved neurological outcomes in TBI patients by cooling to a reduction in blood glucose and lactate levels.
A retrospective and cross-sectional meta-analysis by Li and Yang (2014) concluded that moderate hypothermia benefitted TBI patients by resulting in better neurological outcomes, and reduced death rate is shown, although the data were not statistically significant. Overall, there is a lack of well-controlled, large clinical trials that provide convincing evidence of the benefits and safety of brain hypothermia as a treatment for traumatic head injury patients. A recent review (Saxena et al. 2015) suggested conducting clinical trials using mild hypothermia when the brain temperature is reduced by 1–2 °C in order to minimize systemic complications of hypothermia with targeted temperatures as low as 33 °C. The 4th edition of the Guidelines for the Management of Severe Traumatic Brain Injury, issued by the Brain Trauma Foundation in 2017, cautions on implementing mild hypothermia for patients with TBI while emphasizing adequate pharmacologic intervention as the 1st priority (Carney et al. 2017).
5.2 Brain Ischemic Injury from Stroke
The Guidelines on Acute Stroke Treatment by the American Stroke Association identify hypothermia after stroke as a promising field of research but state that well-established clinical trials are needed before endorsement. A review by Van Der Worp et al. (2007) concludes that hypothermia improved outcomes of animals suffering from ischemic stroke by more than 30%. This result is especially true in well-controlled animal stroke models (Barone et al. 1997). It is relatively easier to have reproducible models of transient global brain ischemia via occlusion of the distal middle cerebral and ipsilateral common carotid arteries than that of focal brain ischemia, for which it is more difficult to define experimental conditions. Computerized images of cerebral infarct volume and edema are commonly used to evaluate ischemic damage to the brain tissue. If the animals are kept alive after the injury, cognitive evaluation is conducted to assess the long-term protective outcome of cooling. The efficacy of hypothermia is better with a lower target temperature (Kurasako et al. 2003) and/or when it is initiated before or during the onset of ischemia. A study by Kurasako et al. (2003) on a spontaneously hypertensive rat model demonstrated a strong, linear, temperature-dependent reduction in infarct volume and edema progression in transient focal ischemia. Neuroprotection by hypothermia is more prominent in lower-temperature (28 °C and 30 °C) groups. More improvements are found in temporary models rather than in permanent ischemic models (Marion et al. 1997; Morikawa et al. 1992).
There are limited clinical trials to test the efficacy and safety of various hypothermia approaches for acute ischemic stroke patients. Overall, it has been demonstrated that more robust, randomized clinical trials with large patient enrollments are needed to provide conclusive evidence for the benefits of hypothermia in stroke patients. Clinical studies by Schwab et al. (1998a, b, 2001) showed that more than half of the patients survived severe stroke after up to 72 h of hypothermia at 33 °C. Most of the 38% mortality rate of the stroke patients occurred during the rewarming process, which suggests an important role of gradual rewarming after hypothermia. Although the clinical studies were not conducted using a control group, the observed 38% mortality rate is much lower than a typical death rate of 70% in severe stroke patients (Hacke et al. 1996; Berrouschot et al. 1998). A small clinical trial by Kammersgaard et al. (2000) using mild hypothermia (35 °C) in 17 patients also showed marked improvement of mortality rate of 12% compared to 23% in controls. Unlike the Schwab studies that were associated with many serious systemic complications, including thrombocytopenia, bradycardia, and pneumonia, the only reported side effect in the Kammersgaard trial was shivering, which was treated by pethidine. Kasner et al. (2002) conducted a clinical trial with mild hypothermia induced by acetaminophen administration in an attempt to achieve brain cooling rapidly. The results were inconclusive because acetaminophen produced only a very mild state of hypothermia of <36.5 °C, bringing into question its clinical efficacy. A randomized trial of hypothermia in infants with hypoxic-ischemic encephalopathy showed reduced risk of death or disability (Shankaran et al. 2005). However, neuroprotection from hypothermia was not obvious in additional two clinical studies (Krieger et al. 2001; de Georgia et al. 2004), even though the results favored hypothermia.
Endovascular cooling of blood has been investigated by various research groups as an effective cooling method to induce a fast cooling rate varying from 0.9 to 6.9 °C/h. Rather than cooling the entire body, this approach targets blood in circulation via a cooling catheter inserted into the femoral vein and advanced to the inferior vena cava. Several clinical trials in the early 2000s suggested safety and effectiveness of the fast endovascular approach on brain ischemic patients (de Georgia et al. 2004; Hemmen et al. 2010). One side effect noticed was shivering, but it can be managed by combining other medications and treatment. The Hemmen study enrolled 59 ischemic stroke patients and randomly assigned participants to a control group or a hypothermia group that used endovascular cooling to lower the body temperature to 33 °C. Based on the 90-day follow-up outcome, no statistically significant difference in mortality was reported (Hemmen et al. 2010). A smaller clinical trial published in 2006 with only 25 patients focused on evaluating the effect of hypothermia on patients with malignant supratentorial cerebral ischemia (Els et al. 2006). Again, an endovascular cooling catheter was utilized to quickly reduce body temperature to 35 °C. The 6-month follow-up outcome in the hypothermia group was better than that in the normothermia group and had similar side effects or complications (Els et al. 2006). The most recent clinical trial implemented mild hypothermia for 48 h using endovascular cooling in acute ischemic stroke patients. Based on the data measured in a larger participant group of 75 patients, it showed less cerebral edema, lower incidence of hemorrhagic transformation, and better outcomes in the cooling group than that of the normothermia group (Hong et al. 2014). The neurological outcome was measured using a modified Rankin Scale (mRS) score 3 months after the treatment, and more than half of the patients in the cooling group scored as “good,” while only 22% of the patients in the control group scored as “good.”
5.3 Cardiac Arrest
Clinical trials using hypothermia treatment after cardiac arrest yield strong evidence to support beneficial effects of cooling to patients. A recent review by Schenone et al. (2016) analyzed 149 papers and later evaluated 24 papers related to clinical trials that implemented systemic hypothermia in cardiac arrest patients. Among them, 11 studies with randomized controlled trials and observational cohort studies compared results between hypothermia and normothermia groups. Most of the reports demonstrated increased odds of survival with good neurological outcomes when hypothermia was implemented to out-of-hospital cardiac arrest patients.
Two early randomized clinical trials published in 2002 enrolled a total of 352 adult comatose survivors after out-of-hospital cardiac arrest (OHCA) (Bernard et al. 2002; Holzer et al. 2002). A targeted core body temperature of approximately 33 °C was achieved in the hypothermia group, and the normothermia group had body temperatures higher than 37 °C, with or without fever control. Results at the 6-month follow-up showed that the percentage of patients with favorable neurological outcomes increased from approximately 30% in the normothermia group to approximately 50% in the hypothermia group. Further, the 6-month mortality rate was much lower in the cooling group than in the control group. In Lauren et al. (2005), they performed a randomized clinical study on a small patient group (61 cardiac arrest patients) to evaluate effects of high-volume hemofiltration with or without hypothermia on decreasing mortality rate. Compared to the control group, high-volume hemofiltration improved the overall prognosis after resuscitation from OHCA. The effect of hypothermia alone was not evaluated directly, but the results suggest that hypothermia if applied early might lower the mortality rate within several weeks after the treatment (Lauren et al. 2005). In a retrospective study, the effects of external cooling implemented to cardiac arrest patients (55 hypothermia vs. 54 normothermia) were evaluated (Oddo et al. 2006). They found that the patients in the cooling group had a much lower mortality rate with better neurological outcomes than that of the normothermia group. Therapeutic hypothermia was found particularly beneficial in patients with a short duration of cardiac arrest. Similarly, in another retrospective study (Belliard et al. 2007), external cooling using ice packs resulted in significantly higher survival rates than that of the normothermia group. They also commented that the cooling method was easy to implement without observing significant complications. In another retrospective study of 75 patients, targeted reduction of body temperature to 33 °C was achieved using either external or endovascular cooling (Ferreira et al. 2009). They found that hypothermia significantly increased the survival rate from 42% in the control group to 67% in the cooling group, and the percentage of patients with favorable neurological outcomes jumped from 19 to 51% in the cooling group. Interestingly, although there was no statistically significant difference in the survival rates of both cooling methods, the percentage of patients with favorable neurological outcomes in the endovascular subgroup was higher than that of the external cooling subgroup (62 vs. 40%). The authors observed a much shorter duration to reach targeted body temperature reduction using endovascular cooling. Effects of cooling were also evaluated in another retrospective study published in 2011, by comparing patients treated with a conservative approach with patients treated with active cooling (Petrovic et al. 2011). Hypothermia was found to decrease patient mortality rate from 81% in the normothermia group to 44% in the cooling group of cardiac arrest patients. Despite the above-reported positive effect of hypothermia on patient outcome and survival, a randomized large-scale clinical trial published recently by Nielsen et al. (2013) raised many questions on the benefits of hypothermia on cardiac patients. This study included more than 900 out-of-hospital cardiac arrest patients, and it showed no significant difference between patients in the cooling group and in the normothermia group when evaluating the survival rate and neurological outcomes at the 180-day follow-up. Overall, reports on the beneficial effects of hypothermia on the outcomes of cardiac arrest patients were frequently observed in the past 15 years, unlike outcomes when treating TBI or stroke patients.
Previous studies have also examined whether targeted temperature reduction matters. Unfortunately, most of the studies have not shown a significant difference when implementing a specific temperature over another during hypothermia. A recent review article suggested that in some of the clinical trials involving cardiac arrest patients, a delay in cooling initiation or a delay in reaching the targeted temperature increased risk of death and/or poor neurological outcomes (Fukuda 2016). Related to initiation of cooling, some studies have been performed to evaluate the efficacy of initiating hypothermia before the patient arrives at the hospital or before return of spontaneous circulation. In small clinical trials, the results from prehospital cooling and cooling before return of spontaneous circulation are encouraging due to data demonstrating improved survival rates and neurological outcomes. However, results from large randomized clinical trials are currently in progress and have not yet provided conclusive evidence of the benefits of cooling. Some studies suggested that fever control alone may be sufficient in resulting in improved neurological outcomes (Nielsen et al. 2013).
5.4 Brain Ischemic Injury During Surgery
In TBI, ischemic stroke, or cardiac arrest, it is difficult to implement hypothermia before the injury occurs. Previous well-controlled animal experiments have shown the importance of initiating cooling before the injury to confer the benefits of hypothermia in neuroprotection (Welsh et al. 1990; Dietrich et al. 1993). Cardiopulmonary surgery would make precooling possible. It is well known that brain injury is a common postsurgical observation after cardiopulmonary surgery (Miller 1993; Roach et al. 1996). It is believed that neurological dysfunction results from regional or global brain ischemia as a consequence of hypoperfusion due to circulatory arrest and formation of emboli (Nussmeier 2002).
Early animal studies (Gillinov et al. 1993; Zeiner et al. 1996a, b) have been performed to assess the improved neurological outcomes associated with hypothermia. It has been found that profound brain hypothermia significantly improved the neuro-deficit score of 48.3 in the control group to 19.2 in the cooling group (0 = normal and 100 = brain dead). Improved neurological outcome was also shown in a dog model of cardiac arrest (Sterz et al. 1991). In a study by Moyer et al. (1992), spontaneous cerebral hypothermia of 33 °C decreased focal infarction volume in a rat brain by 75%. A recent animal study performed on dogs also showed a 70% survival rate following surgery of mitral valve plasty after more than 90 min of cardiopulmonary bypass using profound hypothermia (20 °C).
Numerous experimental studies have shown that brain hypothermia enhances the brain’s tolerance to ischemia during cardiopulmonary surgery and thereby improves neurological outcomes. Hypothermic circulatory arrest temporarily suspends blood flow under very cold body temperatures. When the tissue is cooled to allow blood circulation to stop, cellular activity becomes low without introducing significant damage to the patient. This also allows sufficient time for the surgeon to repair under a bloodless field in open-heart and open-neck surgery. It is very rare to see published results of randomized clinical trials that compare patient outcomes in a hypothermia group with that of a normothermia group, since therapeutic hypothermia has been implemented as a standard therapy in the international resuscitation guidelines since the 1970s. Most of the current research is focused on evaluating the effects of the degree of hypothermia and various reperfusion techniques. One recent survey suggested that most clinical practices utilize deep hypothermia with the body temperature reduced to 20 °C or lower (Gutsche et al. 2014), allowing at least 30 min of time for cardiac repairment. Mild brain hypothermia was reported to minimize cerebral impairment in 1500 patients undergoing pulmonary endarterectomy (Jamieson et al. 2003). The protective effect of hypothermia was also evident in patients undergoing carotid endarterectomy (Kouchoukos et al. 1994), wherein there were no early postoperative strokes or reversible ischemic neurologic deficits following the surgery. An FDA-approved cooling system known as ChillerPad™ and ChillerStrip™ system (Seacoast Technologies, Inc., Portsmouth, NH) was employed in a clinical trial of aneurysm repair; hypothermia markedly reduced vasogenic edema and protected the blood-brain barrier (Wagner and Zuccarello 2005). One recent clinical study reported no significant difference in patient mortality between implementations of either profound hypothermia (∼15 °C) or deep hypothermia (∼20 °C). The authors suggested that deep hypothermia is a safe approach that is easy to implement without certain complications associated with profound hypothermia (Gong et al. 2016). Another clinical study focused on hypothermia at 31 °C in bypass surgery also demonstrated similar mortality rates to that of deep or profound hypothermia (Guo et al. 2014). Currently, the most recent studies in this field are evaluation of whether moderate hypothermia is equally effective as profound hypothermia in providing tissue protection. In addition, this research helps elucidate how to increase the bypass duration and allow surgeons a longer time to operate in complicated surgical procedures.
5.5 Spinal Cord Injury
The nervous system includes the brain, spinal cord, and a complex network of neurons. Like the brain, the spinal cord is covered by the meninges and contains both gray matter and white matter. Also similar to the brain tissue, the spinal cord is subjected to severe damage after injury, and hypothermia may provide neuroprotection after acute spinal cord injury.
Each year, more than 12,000 new cases of spinal cord injury are reported in the United States, and it affects all ages. Only 1% of those patients were discharged neurologically normal, and most patients suffered from either complete quadriplegia or paraplegia that required significant medical costs following their discharge from the hospital. It is a great challenge to develop new treatment options to preserve spinal cord tissue, thus facilitating recovery of some motor functions controlled by the spinal cord (Steeves et al. 2011).
There have been many animal studies conducted to evaluate the efficacy of spinal cord cooling under well-controlled injury settings. Animal models including rats and pigs have been proven valuable not only in isolating various factors in spinal cord injury but also understanding various molecular pathways involved in the injury (Strauch et al. 2004; Yoshitake et al. 2004; Ha and Kim 2008; Morino et al. 2008; Purdy et al. 2013; Grulova et al. 2013). Animal models have also been used to test various local and systemic cooling approaches or devices. Typically, spinal cord injury is induced by contusion with weight impact, compression of the spinal cord, distraction via stretching the spinal cord, dislocation by lateral displacement of the vertebra, or transection that severs the cord (Cheriyan et al. 2014). In a rat model, hypothermia of 32–33 °C was introduced to the spinal cord 30 min after the injury for a duration of 4 h, and the results were encouraging by showing improvement in locomotor deficits and reducing the area of damage (Yu et al. 2000). In one study using TUNEL (terminal deoxynucleotidyl transferase (TdT) dUTP nick-end labeling) stain as an index of cell damage on rats with spinal cord contusion injuries, the authors observed marked decrease in TUNEL in the hypothermia group of 32 °C in the evaluation at both the 72-h and 7-day follow-ups as compared to that of the control and sham groups (Shibuya et al. 2004). In another contusion model, modest hypothermia of 4 h also resulted in significant increases in white matter and gray matter volumes and preservation of neurons when compared with normothermic controls. The rats in the cooling group also had a faster rate of recovery in open-field locomotor ability and improved forelimb strength (Lo Jr et al. 2009). Hypothermia was shown to significantly improve the behavioral and histologic outcomes in rats that underwent compression injury (Batchelor et al. 2010; Maybhate et al. 2012). In one recent study, beneficial effects of hypothermia on survival of neural stem cells in a transplantation for recovery from SCI were demonstrated when compared with control groups (Wang and Pearse 2015). Overall the benefits of deep or moderate hypothermia on spinal cord injury have been well documented by many animal studies in controlled experimental settings.
In clinical settings, cooling initiation to spinal cord injury patients may not be introduced as early as that of well-controlled animal experiments. This is largely due to the transportation of patients to trauma centers and time required to obtain a patient’s consent to include hypothermia as an experimental treatment option. Therefore, results from clinical studies vary on neuroprotection and function recovery due to hypothermia. Early studies in the 1970s typically used an ice pack or cold saline infusion to the injured spinal cord for several hours. Unfortunately, those studies are often associated with small sample sizes and lack of control groups (Selker 1971; Tator 1972; Negrin Jr 1973; Meacham and McPherson 1973; Bricolo et al. 1976). The most recent clinical study using local cooling via an extradural saddle device on 20 SCI patients suggested an improved ASIA (American Spinal Injury Association) grade (Hansebout and Hansebout 2014). Although there was no control group of patients in normothermic conditions, the authors state that the outcomes due to cooling were better than that of traditional treatments (Hansebout and Hansebout 2014).
In recent years, systemic hypothermia for spinal cord injury has attracted a lot of attention due to development of an endovascular catheter that induces fast cooling and can quickly reach targeted temperature reduction. Unlike local cooling when laminectomy may be needed before cooling can be applied, systemic hypothermia can be implemented immediately after injury by cooling the blood using ice-cold saline or endovascular cooling catheters. Several clinical studies on implementing systemic cooling to SCI patients are encouraging. Some of them were carried out to evaluate safety of implementing the device within a small group of patients (Cappuccino et al. 2010; Tripathy and Whitehead 2011). Initial results by Levi et al. (2009, 2010) show improved patients’ AIS (ASIA Impairment Scale) grades with normal complications, and their follow-up study demonstrated favorable outcomes in those patients. A recent clinical trial by the same group enrolled more than 35 acute SCI patients (Dididze et al. 2013). At the 10-month follow-up evaluation, more than 43% of the patients improved at least one ISNCSCI (International Standards for Neurological Classification of Spinal Cord Injury) grade despite the relatively late cooling initiation of approximately 6 h after injury (Dididze et al. 2013). Currently there are no randomized large multicenter clinical trials to test systemic hypothermia on spinal cord injury patients. Although robust experimental and limited clinical studies have demonstrated that hypothermia is beneficial for acute SCI patients, further studies are critical in providing convincing evidence of the efficacy of hypothermia on the recovery of a devastating injury with very poor outcomes.
6 Engineering Contributions in Brain Hypothermia
6.1 Development of Cooling Devices
The major contribution of engineering in therapeutic brain hypothermia is the design of reliable and safe devices that achieve target cooling rates and temperatures. New technologies are typically evaluated for thermal efficacy and safety in small and large animal models to determine their suitability for testing in humans. During the design, patient safety is a prime consideration, and potential issues include leakage of coolant, induced freezing damage to tissue, physical trauma, etc. Sterilizability, size minimization consistent with thermal performance requirements, and physical packaging are typical design considerations during the development of the device. Another important thing often neglected by engineers is how the cooling device may interfere with normal medical treatments.
In general, there are two categories of device designs that are used clinically for therapeutic hypothermia: skin surface and endovascular cooling. More than six companies (Medivance, MTRE, Seabrook, Cincinnati Sub-Zero, Birchbrook, and Garnar) manufacture surface cooling devices, such as cooling blankets, garments, helmets, and pads. This category of devices is based on noninvasive methods and has been most broadly applied in clinical trials to date. An alternative approach is presented by AVAcore™ in which an applied negative pressure to glabrous skin in conjunction with cooling is used to enhance blood perfusion through arteriovenous anastomoses, thereby increasing convective energy transport with the body core. This technique has the potential for providing much faster cooling rates. There are three companies (Radiant Medical, Innercool Therapies, and Alsius) that have developed endovascular catheters for cooling human blood as it flows past the device that is inserted into a large vessel. Endovascular cooling is superior in providing a fast cooling rate. However, its major drawbacks are increased risk of blood clotting, invasive surgical insertion procedures, and complications associated with whole-body cooling. Other approaches, such as intracarotid cooling and interstitial cooling in the neck, are in their infant stages and have yet to be tested in large animal models and clinical studies.
6.2 Mathematical Simulations of Cooling and Rewarming Processes
In the past three decades, clinicians have realized the importance of closely monitoring brain temperature during hypothermia. Even a small reduction in brain temperature can improve neurological outcomes of patients suffering from brain injury. Temperature variation within normothermic brain tissue is usually very small. However, especially during therapeutic hypothermia, temperature gradients can develop not only between the brain and core but also within regions of the brain (Busto et al. 1987; Mellergard et al. 1990; Stone et al. 1995; Wass and Lanier 1996; Schwab et al. 1997, 1998a, b). Information on internal temperature gradients is difficult to obtain clinically by measurement with temperature probes owing to the risk of inducing additional tissue damage. Current noninvasive temperature measurements such as magnetic resonance imaging (±2 °C) lack the desired resolution to monitor small temperature variations in the brain region. Therefore, analytical methods for understanding the transient and spatial temperature distribution in brain tissue during hypothermia therapy are clinically valuable.
Quantitative thermal modeling aids in identification of an optimal treatment protocol, and appropriate constitutive input data creates a potential for designing personalized therapeutic regimens. Theoretical modeling provides clinicians with powerful tools to improve the ability to deliver safe and effective therapy. Most importantly, thermal modeling permits the identification and evaluation of critical monitoring sites to assess the cooling extent and to assure patient safety. Most of the current theoretical thermal models in brain hypothermia implement the Pennes bioheat equation (Pennes 1948) that simplifies the thermal contribution of local blood vessels as a simple heat source term added to the traditional heat conduction equation (Zhu and Diao 2001; Diao et al. 2003; Dennis et al. 2003; Ley and Bayazitoglu 2003, 2004; Janssen et al. 2005; Wang and Zhu 2007; Neimark et al. 2008; Zhu and Rosengart 2008; Zhu et al. 2009). With the advancement in computational resources, researchers have the capability to simulate the 3D temperature field of the head and neck and to consider multiple large blood vessels individually in the model. When point-to-point temperature nonuniformities are important, a model that incorporates perfusion through the vasculature is necessary to accurately predict the tissue temperature field. However, complex vascular geometry may prompt a researcher to model only selected large blood vessels individually and neglect the others. There were several attempts in the past to model large blood vessels, including the carotid arteries and jugular veins in the neck, during brain hypothermia (Zhu 2000; Bommadevera and Zhu 2002; Eginton 2007; Wang and Zhu 2007). In those studies, temperature decay along the large arteries was used to evaluate whether various techniques could induce sufficient cooling to the arterial blood that supplies the brain. Owing to the complicated vascular structure, it is difficult to model effects of individual blood vessels in the brain region. Van Leeuwen et al. (Van Leeuwen et al. 2000) predicted temperature contours based on detailed vasculature in the brain and found that they agree very well with those predicted by the Pennes model. This may be mainly due to the large rate and diffuse pattern of blood perfusion in the brain tissue, for which the Pennes bioheat equation is a preferred approach. Advanced computational methods also allow researchers to model the head with a more realistic geometry than a simple hemispherical structure. The head geometry is usually based on magnetic resonance images (MRIs), which can be imported into grid-generating algorithms and then interfaced with numerical software for temperature simulation (Van Leeuwen et al. 2000; Dennis et al. 2003; Ley and Bayazitoglu 2003). However, the results obtained from the more complicated head models to date are very similar to those predicted by a simple, layered head geometry.
Most of the mathematical simulations are aimed at providing detailed temperature contours in a targeted tissue region, as well as predicting how long it takes to establish a steady state. These help determine a cooling rate at specific locations. Those simulation results are essential in evaluating the thermal feasibility of these cooling methods. Early studies simulated a temperature field inside the brain hemisphere with cooling on the scalp by a helmet (Van Leeuwen et al. 2000; Zhu and Diao 2001; Diao et al. 2003). It has been shown that cooling penetration from the scalp to the brain tissue is only limited to less than 8 mm unless the brain region is ischemic. Therefore, cooling in the brain white matter is not feasible. Later mathematical simulations were carried out based on image scan-generated realistic head models due to advancement in computational resources in the 2000s (Dennis et al. 2003; Janssen et al. 2005; Ley and Bayazitoglu 2003, 2004). Interstitial cooling on the common carotid artery was simulated to demonstrate the importance of the length of the cooling device to achieve desirable temperature reduction in the brain (Wang and Zhu 2007; Wang et al. 2008). It has been shown that cooling is limited when the blood flow rate of the common carotid artery is high. The neck collar has been proposed to cool the brain tissue during the summer time. Although it may provide cold sensation to the neck tissue, simulation of the temperature field in the neck and head region suggests that less than 0.3 °C of blood temperature reduction would be achieved using a cold neck collar (Eginton 2007). Cooling on the torso of patients suffering spinal cord injury was proposed as a noninvasive approach to cool the spinal cord and CSF. Since CSF in the spinal cord is connected with that in the head, simulations were performed to predict temperature reduction in the spinal cord tissue and possible hypothermia in the brain tissue, from the pulsation of the cooled CSF (Smith and Zhu 2010a, b). Theoretical models were developed before the implementation of endovascular infusion of cold saline into the common carotid artery. All the computational models simulate infusion of cold saline with an infusion rate varying from 7.5 to 50 mL/min (Konstas et al. 2007; Neimark et al. 2007a, b, 2008, 2013; Slotboom et al. 2004, Slotboom 2007). It has demonstrated a very fast temperature drop of the brain to 34 °C within 10 min. Those simulation predictions can be viewed as the upper limit of temperature reduction since they fail to address thermal resistance of the catheter as well as loss of cooling to the surrounding regions. Similarly, simulations were performed to determine the effects of intravenous infusion of saline in chilled or ice slurry on fever control in brain injury patients. The proposed approach would result in a maximal cooling rate of 0.45 °C/h when infusing an ice slurry saline at a maximal rate of 450 mL/h (Rosengart et al. 2009, Zhu et al. 2009).
Most importantly, mathematical simulation has also been used to extrapolate experimental measurements of cooling extent to a wide variety of situations. One example is during the early stages of designing and implementing an intraparenchymal cooling device (Zhu and Rosengart 2008). The cooling device consists of a 2.5-cm long, dual-lumen stainless steel shaft with cold water circulating from a slurry ice bath to the device. The diameter of the device is approximately 2.7 mm. Experiments to measure temperature on the cooling device surface and a brain location 6.4 mm from the probe surface were performed on a primate animal model with the cooling device inserted into the parenchymal region. Once validated with the animal experiment, the theoretical model was used to predict the performance of the cooling device in a human brain. The results were also extrapolated to show cooling penetration in ischemic brain regions, where the local blood perfusion rate is much lower than that of a healthy brain. An approximate relationship was established as the cooling penetration depth is inversely proportional to the square root of the local blood perfusion rate of the brain tissue. The relationship suggests that the cooling penetration depth doubles when the local blood perfusion rate decreases to 25% of its normal level.
One of the major challenges in mathematic simulation of cooling and rewarming processes is assessment of local blood perfusion variations in response to cooling. In most mathematical simulations, blood perfusion is usually considered as a constant during cooling or is considered to be decreasing as a function of the local tissue temperature following the Q10 law (Bering 1961; Hoffman et al. 1982). The predicted temperature distribution in the brain is in agreement with the observed temperature field during steady state. However, there is a large discrepancy between theoretical and experimental data on how long it takes to establish a steady state during cooling and rewarming.
In an experimental study performed on a rat model during head surface cooling, thermocouples were inserted into the brain to monitor the temperature reduction and recovery (Diao and Zhu 2006). The measured characteristic time to establish a steady state varied from 5 min to longer than 40 min, whereas the mathematical simulation predicted a much smaller characteristic time (less than 5 min). This is due to the small size of the rat head and constant blood flow rate during the simulation. These results imply that the change of the blood perfusion rate in tissue contributes significantly to the characteristic transient time constant. A similar conclusion can be drawn from other animal models (Zhu and Rosengart 2008) and alternative cooling approaches (Wang et al. 2008).
The Q10 law represents a linear relationship between 1/T, where T is tissue temperature, and log CMRO2 (the cerebral metabolic rate of oxygen consumption). This law states that the metabolic rate decreases by a factor of Q10 with each 10 °C reduction in temperature. Q10 is a parameter, which has been reported (Hoffman et al. 1982) to vary between 2 and 4.4 based on correlations with experimental measurements. Based on a Q10 value of 2, it has been calculated that hypothermia decreases cerebral metabolic rate by an average value of 7% for the first 1 °C reduction in temperature, whereas metabolic rate is reduced to one half of the normal value when the temperature reduction is by 10 °C. During normal conditions, cerebral blood flow (CBF) may follow the same pattern as that of cerebral metabolism due to their direct coupling. Several temperature simulations have incorporated the Q10 law (Dennis et al. 2003; Diao et al. 2003; Janssen et al. 2005; Zhu and Rosengart 2008). Because the local blood perfusion keeps decreasing with the temperature during cooling, the temperature field from the cold surface can penetrate more readily into the deep brain region, which, in turn, would further trigger perfusion reduction. If the temperature dependence of perfusion progresses during cooling, the result would be an extended time to reach a steady state.
During cerebral ischemia or head injury, not only the Q10 value may change but also CBF may be decoupled from metabolism. A number of studies have examined the variation of CBF during systemic hypothermia. Using the radioactive microsphere technique, Busija and Leffler (Busija and Leffler 1987) measured CBF in anesthetized newborn pigs. They concluded that systemic hypothermia reduced CBF secondary to the depression of cerebral metabolic rate. Verhaegen et al. (Verhaegen et al. 1993) measured the cortical blood flow in anesthetized rats using a laser Doppler flowmeter (LDF) and found that CBF was reduced during moderate hypothermia. Okubo et al. (Okubo et al. 2001) examined the effect of systemic cooling on cerebral metabolism and regional CBF variation in newborn piglets. They measured the regional CBF with colored microspheres and demonstrated that a reduction of cerebral cortex temperature resulted in a decrease in the blood flow in all brain regions. Unlike many experimental studies on CBF response during systemic cooling, there have been only few studies on the effect of selective brain cooling on CBF, and the various results are inconsistent. Laptook et al. (2001) examined the differences of CBF in newborn swine during selective brain cooling versus whole-body cooling and illustrated that the global CBF was reduced during both whole-body cooling and selective brain cooling. Ibayashi et al. (2000) demonstrated that the regional CBF decreased when selective brain cooling was implemented on rats. However, a previous study (Kuluz et al. 1993) using the LDF technique showed that the cortical CBF in normal, lightly anesthetized rats increased during selective brain cooling. In a study, the blood flow rate of the common carotid artery was measured continuously using an LDF applied in an in vivo setting to study the transient behavior of temperature and blood flow responses during selective brain cooling and rewarming using a cooling helmet (Diao and Zhu 2006). Very similar transient profiles of brain temperature and blood flow rate of the common carotid artery, as characterized by their characteristic time constants, were observed in rats (Diao and Zhu 2006). Nonetheless, more rigorous experimental verification is needed to establish the correlation between the CBF and blood flow rate of the common carotid artery. The accuracy of mathematical simulations can continue to be questionable unless the simulations can be verified with experimental data that simultaneously monitors the local blood perfusion rate during cooling and rewarming.
7 Concluding Remarks
Hypothermia’s efficacy in improving treatment outcomes in patients suffering from cell and tissue damage caused by ischemia is still ongoing despite more than 80 years of animal experiments and clinical practice. This method can be applied to a wide variety of conditions, including spinal cord injury, traumatic brain injury, stroke, cardiopulmonary surgery, and cardiac arrest. Overall, an accumulating body of clinical evidence along with several decades of animal research and mathematical simulations has documented that the efficacy of hypothermia is dependent on achieving a reduced temperature in the target tissue before or soon following the injury-precipitating event. Mild hypothermia with a several degrees Celsius temperature reduction is as effective as modest or deep hypothermia in resulting in therapeutic benefits without introducing collateral/systemic complications. It is widely documented that rewarming rate must be controlled to be lower than 0.5 °C/h in order to avoid mismatch between local blood perfusion and metabolism. In the past several decades, many different cooling methods and devices have been designed, tested, and used in medical treatments with mixed results. Accurately and predictably designing treatment protocols to achieve specific cooling outcomes requires collaboration among engineers, researchers, and clinicians. Although the problem is quite challenging, it presents a major opportunity for bioengineers to design new methods and devices that can quickly and safely produce hypothermia in targeted tissue regions without interfering with routine medical treatment.
- Bering E (1961) Effect of body temperature change on cerebral oxygen consumption of the intact monkey. Am J Phys 200:417–419Google Scholar
- Busija DW, Leffler CW (1987) Hypothermia reduces cerebral metabolic rate and cerebral blood flow in newborn pigs. Am J Phys 253:H869–H873Google Scholar
- Eginton ML (2007) Evaluation of the effectiveness of a commercial cooling collar in reducing body temperature during heat stress: theoretical modeling of body temperature distribution. Master’s thesis. University of Maryland, Baltimore CountyGoogle Scholar
- Fay T (1945) Observations on generalized refrigeration in cases of severe cerebral trauma. Assoc Res Nerv Ment Dis Proc 4:611–619Google Scholar
- Hutchison JS, Ward RE, Lacroix J, Hébert PC, Barnes MA, Bohn DJ, Dirks PB, Doucette S, Fergusson D, Gottesman R, Joffe AR, Kirpalani HM, Meyer PG, Morris KP, Moher D, Singh RN, Skippen PW, for the Hypothermia Pediatric Head Injury Trial Investigators and the Canadian Critical Care Trials Group (2008) Hypothermia therapy after traumatic brain injury in children. N Engl J Med 358:2447–2456CrossRefPubMedPubMedCentralGoogle Scholar
- Kabat H (1940) The greater resistance of very young animals to arrest of the brain circulation. Am J Phys 122:588–599Google Scholar
- Kuluz JW, Prado R, Chang J, Ginsberg MD, Schleien CL, Busto R (1993) Selective brain cooling increases cortical cerebral blood flow in rats. Am J Phys 265:H824–H827Google Scholar
- Kurasako T, Zhan L, Pulsinelli WA, Nowak TS (2003) Transient cooling during early reperfusion attenuates delayed edema and infarct progression in the spontaneously hypertensive rat. Distribution and time course of regional brain temperature change in a model of postischemic hypothermic protection. J Cereb Blood Flow Metab 27:1919–1930CrossRefGoogle Scholar
- Ley O, Bayazitoglu Y (2004) Temperature distribution in a realistic human head during selective and whole body cooling and during circulatory arrest. Proc. IMECE, IMECE2004–60733, Nov. 13–19. Anaheim, CaliforniaGoogle Scholar
- Maekawa T, Yamashita S, Nagao S, Hayashi N, Ohashi Y, Brain-Hypothermia Study Group (2015) Prolonged mild therapeutic hypothermia versus fever control with tight hemodynamic monitoring and slow rewarming in patients with severe traumatic brain injury: a randomized controlled trial. J Neurotrauma 32:422–429CrossRefPubMedPubMedCentralGoogle Scholar
- Michenfelder J (1988) Protecting the brain. In: Michenfelder J (ed) Anesthesia and the brain. Clinical, functional, metabolic and vascular correlates, 1st edn. Churchill Livingstone, USA, pp 181–193Google Scholar
- Mouritzen CV, Anderson MN (1966) Mechanisms of ventricular fibrillation during hypothermia: relative change in myocardial refractory period and conduction velocity. J Thorac Cardiovasc Surg 51:579–584Google Scholar
- Neimark MA, Konstas AA, Lee L, Laine AF, Pile-Spellman J, Choi J (2013) Brain temperature changes during selective cooling with endovascular intracarotid Intraarterial cold saline infusion: simulation using human data fitted with an integrated mathematical model. J Neurointerv Sur 5:165–171CrossRefGoogle Scholar
- Nielsen N, Wetterslev J, Cronberg T, Erlinge D, Gasche Y, Hassager C, Horn J, Hovdenes J, Kjaergaard J, Kuiper M, Pellis T, Stammet P, Wanscher M, Wise MP, Aneman A, Al-Subaie N, Boesgaard S, Bro-Jeppesen J, Brunetti I, Bugge JF, Hingston CD, Juffermans NP, Koopmans M, Kober L, Langorgen J, Lilja G, Moller JE, Rundgren M, Rylander C, Smid O, Werer C, Winkel P, Friberg H, Investigators TTMT (2013) Targeted temperature management at 33 degrees C versus 36 degrees C after cardiac arrest. N Engl J Med 369:2197–2206CrossRefPubMedGoogle Scholar
- Qiu W, Liu W, Shen H, Wang W, Zhang Z et al (2005) Therapeutic effect of mild hypothermia on severe traumatic head injury. Chin J Traumatol 8:27–32Google Scholar
- Roach GW, Kanchuger M, Mangano CM, Newman M, Nussmeier N et al (1996) Adverse cerebral outcomes after coronary bypass surgery. Multicenter study of Perioperative Ischemia Research Group and the Ischemia Research and Education Foundation investigators. N Engl J Med 335:1857–1863CrossRefGoogle Scholar
- Safar P, Kochanek PM (2001) Comments on “lack of effect of induction of hypothermia after acute brain injury”. N Engl J Med 345:66Google Scholar
- Schwab S, Schwarz S, Aschoff A, Keller E, Hacke W (1998a) Moderate hypothermia and brain temperature in patients with severe middle cerebral artery infarction. Acta Neurochir 71(suppl):131–134Google Scholar
- Selker RG (1971) Icewater irrigation of the spinal cord. Surg Forum 22:411–413Google Scholar
- Shaefi S, Mittel AM, Hyam JA, Boone MD, Chn CC, Kasper EM (2016) Hypothermia for severe traumatic brain injury in adults: recent lessons from randomized controlled trials. Surg Neurol Int 103(1–11):7Google Scholar
- Steeves JD, Kramer K, Fawcett JW, Cragg J, Lammertse DP, Blight AR, Marion RJ, Ditunno JF Jr, Coleman WP, Geisler FH, Guest J, Jones L, Burns S, Schubert M, van Hedel HJ, Curt AEMSCI Study Group (2011) Extent of spontaneous motor recovery after traumatic cervical sensorimotor complete spinal cord injury. Spinal Cord 49:257–265CrossRefGoogle Scholar
- Yu CG, Jimenez O, Marcillo AE, Weider B, Bangerter K, Dietrich WD, Castro S, Yezierski RP, (2000) Beneficial effects of modest systemic hypothermia on locomotor function and histopathological damage following contusion-induced spinal cord injury in rats. J Neurosurg 93:85–93Google Scholar
- Zeiner A, Muellner M, Frossard M, Janata K, Behringer W et al (1996a) Mild therapeutic hypothermia to improve neurological outcome after cardiac arrest---a pilot study. Circulation 94:I9–I10Google Scholar
- Zeiner A, Mullner M, Sterz F, Frossard M, Laggner A (1996b) Mild resuscitative therapeutic hypothermia after cardiac arrest and its influence on hemodynamic parameters. Crit Care Med 24:A115Google Scholar
- Zhu L, Schappeler T, Cordero-Tumangday C, Rosengart AJ (2009) Thermal interactions between blood and tissue: development of a theoretical approach in predicting body temperature during blood cooling/rewarming. Adv Numer Heat Transfer 3:197–219Google Scholar