Selective Brain Cooling using Intracarotid Cold Saline Infusion in
Humans: Theoretical Model and Practical Implications
？1？21,2Angelos-Aristeidis Konstas, Matthew A. Neimark, Andrew F. Laine, and John Pile-
？These authors contributed equally to the study and are joint first authors
1Department of Radiology, Columbia University, New York, USA 2Department of Biomedical Engineering, Columbia University, New York, USA
Correspondence and reprint requests to:
Department of Radiology,
177 Fort Washington Ave,
MHB 8SK, New York,
Running Head: Selective Brain Cooling
A three-dimensional mathematical model was developed to examine the transient and steady state temperature distribution in the human brain during selective brain cooling (SBC) by unilateral intracarotid freezing-cold saline infusion. To determine the combined effect of hemodilution and hypothermia from the cold saline infusion (as CBF is reduced by hypothermia and increased by hemodilution), data was pooled together from studies investigating the effect of hypothermia and hemodilution on CBF and an analytical expression was derived describing the combined effect of the two factors. The Pennes bioheat equation was used to propagate the evolution of brain temperature using the thermal properties of the different cranial layers and the effect of cold saline infusion on CBF. Both a healthy brain and a brain with stroke (ischemic core and penumbra) were modeled. The core and the penumbra were simulated by reducing the CBF and metabolic rate. Simulations using different saline flow rates were performed. Their results suggested that a flow rate of 30 ml/min is sufficient to induce moderate hypothermia within 10 minutes in the ipsilateral hemisphere. The brain with stroke cooled to lower temperatures than the healthy brain, mainly because the stroke itself limited the total intracarotid blood flow. Gray matter cooled twice as fast as white matter. The continuously falling hematocrit was the main time-limiting factor, limiting the SBC to a maximum of 3 hours. The study demonstrated that SBC using intracarotid saline infusion is feasible in humans and may be the fastest method of hypothermia induction.
Therapeutic hypothermia, Selective brain cooling, Intracarotid cold saline infusion, Ischemic stroke, Spatial and temporal brain temperature distributions.
The CNS is very vulnerable to both focal and global ischemia resulting from, respectively, acute ischemic stroke (102, 104) and cardiac arrest (8, 34, 49), where there is a cessation of blood flow to the whole brain. Traumatic brain injury (7, 15, 57) and perinatal asphyxia are two other clinical settings of global ischemia that result in severe CNS impairment (26, 97). In the United States alone, there are an estimated 731,000 strokes and 4 million stroke survivors annually, making stroke a major cause of long-term disability (10). Despite the enormous amount of research, there is a disappointingly small number of effective treatments for neuroprotection after brain injury. At present, thrombolysis with recombinant tissue-plasminogen activator (t-PA) is the only FDA approved therapy for the clinical treatment of acute thromboembolic stroke (103). New neuroprotective treatments are desperately needed.
A significant part of the cerebrovascular damage occurs several hours after the primary insult. Therefore early intervention is critical in the management of clinical problems such as ischemic stroke, cardiac arrest, traumatic brain injury and perinatal asphyxia (43, 50, 76), as it allows for a period of opportunity for therapeutic intervention. Attenuation of the secondary injury is the focus of early post-injury treatment (96, 97). Therapeutic hypothermia has been repeatedly shown to have a remarkable effect in global and focal ischemia. Several studies have demonstrated the effectiveness of post-ischemic
ohypothermia (29-35C) in animal models of global (93) or focal (52, 112) transient
cerebral ischemia. Moreover, intra-ischemic hypothermia resulted in dramatic neuroprotection by reducing infarction volume by 50-90% in all the published studies (13, 27, 40, 51, 68, 77), suggesting a role of hypothermia in preventing neurological injury
when applied prior to or simultaneously with the ischemic insult. The reports of the role of hypothermia in humans are encouraging. Two large randomized controlled trials have conclusively established the neuroprotective effects and safety of hypothermia after cardiac arrest (6, 34). Studies of hypothermia in acute ischemic stroke patients have also reported good preliminary results (39, 83).
In most clinical studies, hypothermia is induced by surface cooling through the use of cooling blankets, alcohol applied to exposed skin, or ice bags to groin, axilla and neck. Although whole-body surface cooling is the simplest and most cost-effective option for inducing hypothermia (24), it has two major drawbacks. First, it takes several hours to reach the target body core temperature. All studies report a 3 to 7 hour time period for
ocooling down to 32-34 C (39, 46, 82, 83, 92). Ideally, the target temperature should be reached as soon as possible in order not to miss the narrow therapeutic window of opportunity after the insult. Endovascular systemic cooling may be able to accelerate the rate of cooling and improve the efficacy of hypothermia (25). The second drawback of whole-body cooling methods (both non-invasive and invasive) is the high incidence of adverse effects such as impaired immune function, decreased cardiac output, pneumonia and cardiac arrhythmias/bradycardias (24, 48). Selective brain cooling without reducing the body core temperature can address both drawbacks: No adverse systemic effects should occur and the rate of cooling can be 10 to 30 times faster than that of whole-body surface cooling (89).
Different methods for selective brain cooling exist (31). Non-invasive methods most commonly used are ice packs, cooling caps, and helmets. Invasive methods used are anterograde cerebral perfusion of extracorporeally cooled blood and intracarotid infusion
of cold saline. Invasive methods require more skill to perform and may involve more risks, but have two advantages over non-invasive methods. First, the rate of cooling is slow for non-invasive methods. During cardiac arrest, significant brain cooling was achieved after 1 hour of ice application to the scalp (20). Animal studies of anterograde cerebral perfusion (47, 64, 67, 84, 85) and intra-arterial saline infusion (21) suggested that a much faster rate of cooling could be achieved in humans. Second, theoretical analyses of surface cooling suggested that it is only effective in reducing the temperature in the superficial cerebral regions and not deep brain structures (19, 20, 71, 99). Moreover, intracerebral temperature monitoring in humans demonstrated a very limited effect of cooling helmets (17, 59). Although information is still incomplete regarding invasive methods of cooling, cold flow through the dense vascular network in brain should be able to cool both gray and white matter.
Although experience with this cooling method is limited, its clinical potential is realistic. Local saline infusion was found to be neuroprotective in dogs (106), baboons (3) and rats (21). In humans, feasibility of the method has been reported in cardiac arrest and neurosurgical patients (2, 55, 86, 107). Clinically, intra-arterial thrombolytic therapy by means of a microcatheter has been used successfully to open acutely occluded cerebral vessels (14). Saline is known to be safe and is already used on a regular basis for intra-arterial infusion during human endovascular procedures (21). The expertise interventional neuroradiologists have in the acute stroke setting further supports the concept that intracarotid saline infusion is technically feasible in clinical practice. Moreover, intracarotid saline infusion could potentially be used as an adjunct therapy in
endovascular procedures such as carotid stenting or cerebral aneurysm embolization, where thromboembolic events are of concern.
The volume of physiological saline infused is a critical issue that may affect feasibility and safety in clinical settings (21, 31). The relationship between internal carotid artery (ICA) blood flow volume and cerebral blood flow (CBF) is linear in humans (90). The ratio of regional CBF (i.e. ICA blood flow) to the saline infusion rate required to keep the ipsilateral brain temperature at the target hypothermic temperature is fixed. Animal
omodels and human studies reported a 5 to 15% reduction in CBF per 1 C drop in brain
temperature (Table 1). This has important implications for the induction and maintenance of intracerebral hypothermia, because the decrease in ICA blood flow rate will allow for the saline infusion rate to gradually decrease, hence the total volume of infused saline will be smaller than anticipated. On the other hand, saline infusion results in hemodilution that increases CBF in humans by 1 to 3% for each 1% drop in hematocrit (Table 2). The CBF increase is due to the decreased oxygen carrying capacity of blood rather than the resulting decrease in blood viscosity, since carbon monoxide could reproduce the effects of hemodilution on CBF (74). No animal or human studies have investigated the overall effect of cold saline infusion on CBF. The combined effect of hematocrit and temperature decrease on CBF will determine whether CBF will increase or decrease during intracarotid cold saline infusion, and whether the saline infusion rate will need to be adjusted accordingly.
In the first part of this study, data is pooled together from studies investigating the effect of hypothermia and hemodilution on CBF in order to derive an analytical expression of the combined effect of these two parameters on CBF during intracarotid cold saline
infusion. In the second part of this paper, a three-dimensional theoretical model is developed to examine the transient and steady state hemispheric temperature response to selective brain cooling with different cold saline flow rates in the ICA. The effect of regionally reduced blood perfusion rate in the brain tissue during focal ischemia is also examined. The time to reach the target hypothermic temperature is estimated using different cold saline infusion rates in healthy and ischemic adult brains.
THEORY AND METHODS
Effect of temperature and hematocrit on CBF
Michenfelder and Milde (62) demonstrated that at brain temperatures between 37 and 27 oC the metabolic rate decreased by a factor of 3 (Q= 3). Thus metabolic heat production 10
of the brain was estimated by Xu et al. (111) as follows:
10 (1) q！q?30
owhere q is the baseline metabolic rate at 37 C. Animal studies suggest that during o
circulatory arrest the reduction of metabolic rate nearly parallels the decrease in CBF (35). Moreover, there is direct evidence of coupling of CBF to cerebral oxygen metabolism in swine (22, 101) and dog (62) studies. However, there is evidence of uncoupling between
ooCBF and cerebral oxygen metabolism at deep (20-26 C) and profound (< 20 C)
hypothermia, suggesting that autoregulation of CBF in accordance with metabolic needs is not preserved in these low temperatures (22, 62).
Given the data in the literature relating CBF to temperature, a more precise formulation of the exponential relationship between these two quantities could be ascertained from the data of these studies. It was assumed that CBF was coupled to cerebral metabolic rate
owith brain temperatures as low as 25 C, thus the temperature-dependent CBF could be
;；：T，37?！??， (2) 0
Similarly, the relationship between brain metabolism and temperature could be expressed as:
;；：T，37q！q?， (3) 0
Data points from the studies listed in Table 1 were fit by the above function. The root mean square error (RMSE) between the data points and the function was minimized by varying ， and ： through the unconstrained nonlinear optimization (fminunc function in
Matlab). This resulted in a value of ， = 2.961 and ： = 0.08401. These values are similar
to those cited elsewhere (20, 62, 111). The RMSE was derived by subtracting the equation point values from the recorded data for corresponding temperatures, squaring the difference, summing the differences, and dividing the result by the degrees of freedom (number of data points subtracted by two because of the two data points that are needed to estimate the two parameters) yielding RMSE = 18.8%.
There is a nearly-linear relationship between mild and moderate hemodilution (HCT>30) and CBF (58). Data points from the studies listed in Table 2 were fit by the following linear function:
;；?！??1，~， (4) 0HCT
The mean square root error between the data points and the function was minimized by linear least squares fitting. This yielded a value of ~ = 2.245. In this case, the number of
degrees of freedom was only reduced by one to calculate RMSE=5.31%.
oA linear relationship similar to that of hematocrit and CBF for 37 C was assumed for all
other temperatures (i.e. utilizing the same value of ~). The combined effect of
temperature and hematocrit on CBF was expressed as:
T;；0.08401，37;；?！??2.9611，2.245， (5) cHCT0
where ω is the temperature and hematocrit corrected perfusion. c
Effect of infusion volume on hematocrit
When the bloodstream is infused with isotonic saline, part of it enters into the extravascular space. The remaining volume of saline hemodilutes the blood. An expression for the resulting change of hematocrit is given as follows:
pV，VRBCIV，！ (6) HCT;；VV？p，VIV0IV0IV
where V is the total red blood cell volume, V is the initial total intravascular volume RBCIV0
(including V), ΔV is the volume of added saline, and p is the fraction of extracellular RBCIV
water that is intravascular. An initial hematocrit of 42% and an initial V=5.0 L is IVO
assumed (implying V=2.1 L). RBC
In steady state, p is normally between 0.25 and 0.33 (1). However, since therapy will take place over relatively short periods of time relative to the time for intravascular volume to reach equilibrium, p should be somewhat higher in practice. To determine a working value of p, Eq. 5 was fit with the data in Table 3 by unconstrained nonlinear optimization
yielding p=0.4217. The number of degrees of freedom was reduced by one to calculate RMSE=0.0149.
Heat exchange between an insulated catheter and blood
From the femoral artery to the arch of the aorta there is a counterflow between cold saline in the catheter and the blood. From the aortic arch and in the carotid artery the coolant flows parallel to the blood. Heating of the cold saline flowing in the catheter is an important factor that determines the feasibility of ipsilateral intracerebral cooling. Hence, an experimental determination of the ability of catheters to deliver cold saline in the carotid artery bifurcation was carried out. A life-sized phantom model of the human arterial tree was used extending from the common femoral bifurcation to the ICAs (Flowtek, CA). Warm water was pumped continuously through the model. Inflow via the ascending aorta and outflow via the common iliac artery and left ICA closely reproduce the blood flow in the human arterial tree.
RA 100 cm 5F Torcon NB Advantage catheter (Cook, Bloomington, IN) was inserted into
a 90 cm 8F guiding catheter (Cordis, Miami, FL) with the distal tip of the 5F catheter extending 2 cm distally of the distal tip of the guiding catheter. The two catheters were glued together at the tip of the guiding catheter, resulting in an air-filled guiding catheter. Since air is an excellent thermal insulator, this created a thermally insulated catheter. To lower the inlet temperature of the saline to sub-zero temperatures, salt was added into the ice bath surrounding the saline bag. The insulated catheter was introduced to the model through the right femoral bifurcation and the distal tip was navigated to the left CCA bifurcation. Outlet temperature was monitored with a fluoroptic temperature probe (Luxtron, Santa Clara, CA) inserted in the 5F catheter and reaching within 1 cm from the