Induced Hypothermia during Emergency Department Thoracotomy: An Animal Model

Induced Hypothermia during Emergency DepartmentThoracotomy: An Animal ModelPeter Rhee, MD, MPH, Eric Talon, MD, Sandra Eifert, MD, Dan Anderson, BS, Kathy Stanton, MS, Elena Koustova, PhD,Geoff Ling, MD, PhD, David Burris, MD, Christoph Kaufmann, MD, MPH, Paul Mongan, MD, Norman M. Rich, MD,Michael Taylor, PhD, and Leon Sun, MD, PhD

Background: Induced hypothermia is used clinically to pre-vent ischemic injury during elective procedures. We present ananimal model of asanguinous hypothermic (10°C) circulatoryarrest, induced through a left anterior lateral thoracotomy afterexsanguinating uncontrolled hemorrhage.

Methods: Through a left anterior thoracotomy, 26 swine(45–70 kg) sustained a laceration of the descending thoracicaorta, producing exsanguinating uncontrolled hemorrhage. Af-ter 5 minutes of severe hypotension (systolic BP<20 mm Hg), a22 French Foley catheter was directed cephalad through theenlarged aortic wound. A solution (containing 42.5 mmol/L K1

and precooled to 1°C) was infused to arrest/preserve the heartand brain. A second 24 French Foley catheter was then directedcaudally through the same wound. The right atrium was openedto drain the venous system. The animal was cooled with acardiopulmonary bypass pump (>5L/min) through the Foleycatheters. Once 10°C was reached, a cannula was placed to theaortic root and the aortic laceration repaired. The animal wasmaintained at 10°C for a total of 90 minutes. Before the rewarm-ing process, the circulation was rinsed with a solution containingnormal levels of electrolytes followed by infusion of whole blood.Rewarming was performed by maintaining a 10-degree gradienton the heat exchanger. The first 16 animals were used in non-survival experiments to develop the technique and to recorddural temperatures and electroencephalogram tracings. The last10 animals were used to determine long-term survival and neu-rologic outcome. Group I: seven animals were kept at< 10°C

with flows less than 2L/min. Group II: three animals underwent20, 30, and 40 minutes of no flow once they were cooled to 10°C.After 6 weeks of survival and neurologic examinations, thebrains were fixed for histologic evaluations.

Results:The average time to cool the head to 18°C and 10°Cwas 6 minutes and 12 minutes, respectively. The hematocrit fellbelow 2% by the end of the cooling period. A total of 7 of the 10animals from the long-term study survived. Group I: five ofseven animals survived. Four of the survivors had no apprecia-ble neurologic deficits, were fully functional at 6 weeks, and hadno evidence of histologic injury. One of the five survivors in thisgroup had moderate neurologic disability. Of the two animalsthat died, one died from air embolism from the i.v. line. Thesecond death was in an animal for which maximal cooling to2.7°C was attempted. Group II: The first two animals that had“no flow” for 20 and 30 minutes were fully functional and hadnormal neurologic examinations. However, the second animalwas found to have brain injury on histologic examination. Thelast animal in this group died of accidental extubation duringrecovery.

Conclusion: Induction of hypothermic arrest through thechest after exsanguination is possible. The further developmentof this technique may provide an extended state of “suspendedanimation” to allow for repairs of hemorrhaging injuries intrauma patients who require emergency department thoracot-omy.

The goal of this study was to determine whether profoundhypothermia could be induced in a swine model ofexsanguinating uncontrolled hemorrhage. This would

create a state of “suspended animation” to allow time toperform surgical repairs.1,2 In this study, we set out to de-velop a technique for inducing hypothermic arrest in a clin-ically relevant model by using currently available techniquesand resources. We also wanted to test the concept of infusingcold acellular fluid to arrest metabolism to create a state ofsuspended animation. Therefore, the aim was to induce hy-pothermia through an emergency department thoracotomy

(EDT) incision after uncontrolled exsanguinating arrest in aswine model.

Currently, patients with penetrating trauma who do notrespond to resuscitation may undergo EDT. The survivalrates after this procedure are generally low and depend onmany factors, including type and location of injury as well asthe physiologic status at the time of the procedure.3,4 Al-though those who undergo EDT often have injuries that arerepairable, the survival rates are not optimal and the proce-dure can be costly in terms of the resources used. Despitemany advances in medicine, little progress has been made toimprove outcome after this procedure over the past severaldecades. If these patients in extremis could be put into a stateof “suspended animation,” surgeons may be able to repair theinjuries in a bloodless field.

The concept that cold can be protective at the cellular levelis not new. Cold slows biological activity and refrigerationhas made a major impact in modern life. Hypothermia canslow down the metabolism of living organisms by slowingactive ion transport, homeostasis, and enzyme activity. Be-

Submitted for publication September 24, 1999.Accepted for publication December 17, 1999.From the Department of Surgery (P.R., E.T., S.E., D.A., K.S., E.K., D.B.,

C.K., N.M.R., L.S.), Department of Anesthesia (G.L., P.M.), UniformedServices University of the Health Sciences, Bethesda, Maryland, and OrganRecovery Systems, Inc (M.T.), Charleston, South Carolina.

Presented at the 59th Annual Meeting of the American Association for theSurgery of Trauma, September 16–18, 1999, Boston, Massachusetts.

Address for reprints: Peter Rhee, MD, MPH, Department of Surgery,Uniformed Services University of the Health Sciences, 4301 Jones BridgeRoad, Bethesda, MD 20814; email: [email protected].

1079-6061/00/4803-0439The Journal of Trauma: Injury, Infection, and Critical CareCopyright © 2000 by Lippincott Williams & Wilkins, Inc.

Vol. 48, No. 3Printed in the U.S.A.

439

cause hypothermia suppresses metabolism, it minimizes thedemand for oxygen and extends tissue tolerance forischemia.5 Metabolism can also be decreased by controllingbiologic regulators that are used during hibernation. How-ever, the alteration of metabolism through pharmacologiccontrol is not yet available. These concepts have led to thedevelopment and use of techniques to induce hypothermicarrest in humans.6 Hypothermia is currently used in neuro-logic, pediatric, and cardiothoracic surgery.7–10 In humans,the limits of hypothermic arrest has been found to be approx-imately 1 hour at 15°C.11–13The question then becomes, Canthe advantages of hypothermic arrest be used in the setting oftrauma?

Recent research in this field has been led by Dr. PeterSafar, who demonstrated that hypothermia can be inducedwith complete recovery after normothermic controlled hem-orrhagic shock and profound hypothermic circulatory arrestof 60 minutes in dogs.14 The development of acellular fluidsto protect and preserve the whole body has further advancedthis field. Taylor et al. have demonstrated that 3 hours ofprofound hypothermic cardiac arrest can be achieved in dogswith normal neurologic preservation by using blood replace-ment with acellular fluids.15 Acellular fluids also addressproblems associated with hypothermia, such as increasedplasma viscosity and cellular sludging through the microvas-culature. The fluids developed by Taylor were “Hypother-mosol P”, which contains normal levels of electrolytes, and“Hypothermosol M”, which contains impermeants and 42.5mmol/L of potassium (Table 1). Experiments by Safar andTaylor established the circumstances and feasibility of hypo-thermic arrest, but they were performed under reasonablycontrolled conditions with closed chest and peripheral can-nulation. We wanted to determine whether asanguinous hy-

pothermic circulatory arrest could be induced through anEDT incision after exsanguination with the use of currentlyavailable equipment.

METHODS

The institutional Laboratory Animal Review Board for thecare and use of animals approved this study. All research wasconducted in compliance with the Animal Welfare Act andother Federal statutes and regulations relating to animals andexperiments involving animals. Adherence to principlesstated in theGuide for the Care and Use of LaboratoryAnimals, NIH publication 86–23, 1985 edition was main-tained. All surgery was performed in designated veterinarysurgical suites with sterilized instruments and procedures.

A total of 26 female swine, each weighing 45 to 75 kg(Tom Morris Farms, Reisterstown, Md), were used. The first16 animals were used to develop the technique in nonsurvivalexperiments. Five of these animals were used to determinethe dural temperatures during this procedure and to observethe electroencephalogram (EEG) patterns. The dural temper-ature probes were inserted by incising the skin over the skulland placing a 1-cm burr hole placed over the right frontallobe. A thermistor was placed between the dura and bone.The burr hole was then filled with bone wax. EEG leads wereplaced around the skull for continuous measurements duringthe procedure.

Survival ExperimentsGroup I: Hypothermic Cardiac Arrest with Low Flow for90 MinutesSeven animals were used in this experiment. The mean arte-rial blood pressure as related to experimental events is shownin Figure 1. The animals were fasted overnight and sedatedwith an intramuscular injection of ketamine (10 mg/kg). Afterthe induction of general anesthesia with isoflurane and roomair, each animal was prepared with Betadine and draped withsterile sheets. A 9 French cordis catheter was placed in theleft external jugular vein through a cutdown. A continuousmixed venous, right ejection fraction pulmonary artery cath-eter (Baxter, Irvine, Calif) was inserted and floated into thepulmonary artery. A left anterior lateral thoracotomy wasperformed. A 20-gauge angiocatheter was inserted into the

TABLE 1. Components of hyperthermosol M and hyperthermosol P solutions

Components Hyperthermosol M(Maintenance solution)

Hyperthermosol P(Purge solution)

IonicSodium (mmol/L) 100 141.2Potassium (mmol/L) 42.5 3.0Calcium (mmol/L) 0.05 1.5Magnesium (mmol/L) 5.0 1.0Chloride (mmol/L) 17.1 111.0Sulfates (mmol/L) — 1.0

pH buffersH2PO4 (mmol/L) 10.0 1.2HCO3 (mmol/L) 5.0 25.0HEPES (mmol/L) 25.0 25.0

ImpermeantsLactobionates

(mmol/L)100.0 —

Sucrose (mmol/L) 20.0 —Glucose (mmol/L) 5.0 5.0Mannitol (mmol/L) 20.0 —

ColloidDextran-40 (%)

MetabolitesAdenosine (mmol/L) 2.0 1.0Glutathione (mmol/L) 3.0 3.0

Osmolality 350 305pH 7.6 7.6

FIG 1. Mean blood pressure schematic of the procedure.

The Journal of Trauma: Injury, Infection, and Critical Care March 2000

440

left subclavian artery for continuous blood pressure monitor-ing. A 28 French chest tube was placed through the 6thinterspace. Heparin (10,000 units i.v.), 1 g of cefazolin, and5 mg of dexamethasone were administered intravenously.The pericardium was opened. The distal thoracic aorta wasexposed superior to the azygous vein. A 5-0 surgical steelsuture was placed as a pull out wire to create a 1 cmlonglaceration along the circumference of the aorta. This injuryresulted in a hemorrhage of approximately 400 mL in the firstminute, at which time the mean arterial pressure fell to lessthan 30 mm Hg. Within 2 minutes after the aortic injury, theblood pressure was less than 20 mm Hg. Hemorrhaged bloodwas collected through the chest tube into a separate containerfor later use during the re-warming phase. The total bloodcollected during the first 6 minutes ranged 800–1,000 mL.

After the systolic blood pressure was less than 20 mm Hgfor 5 minutes, hypothermic arrest was induced by opening thethoracic tear widely (. 60% circumference) with Metzen-baum scissors. A 22 French Silastic urinary Foley catheterwith the tip cut off was directed cephalad through the openingand the balloon inflated. Flow was then initiated by usingprecooled (1°C) Hypothermosol M solution (Bio Life Tech-nologies, Rockville, Md; Table 1) at a rate of 1 L/min byusing a cardiopulmonary bypass pump (CBP) and water heatexchanger. Because this solution contains 42.5 mmol/L K1,instantaneous cardiac arrest ensued. The right atrial append-age was opened and a single stage venous cannula wasinserted and held in place by the assistant. We found that theinsertion of the venous cannula was relatively easy and thecannula did not have to be secured during this time. Excessblood and fluid in the chest was sucked into the CBP reser-voir. A second 24 French Foley catheter was placed into theaorta pointing caudally, and the balloon was inflated (Fig.2A). The flow was then increased to a maximal rate of 5L/min to allow rapid cooling. Isoflurane was discontinuedduring the cooling phase. A single 2-liter bag of ice wasplaced over the abdomen at the initiation of hypothermia andremoved during the rewarming phase.

The catheters and CBP with an oxygenator and 3-literreservoir were primed with 3.5 liters of Hypothermosol Msolution (1°C). The Foley catheters were attached to the0.5-inch arterial line tubing from the CPB pump with an Yconnector. The first Foley catheter was best placed by push-ing down on the distal aorta below the laceration with thecatheter. This approach collapsed the distal aorta under thecatheter and allowed the Foley to be easily inserted directlyinto the aorta. The second catheter was inserted by followingthe first catheter to the opening and directing it distally. Bothcatheters were inserted without visualization of the aorta.This method resulted in the fastest placement of the cathetersbecause blood in the thoracic cavity hindered visualization ofthe aorta. This two-catheter technique allowed for differentialcooling of the body. Different-sized Foley catheters wereused, because the upper portion of the swine was smaller thanthe lower; thus, cooling occurred faster in the upper half thanthe lower. After the esophageal temperature reached 18°C,the cranial catheter could be partially clamped as necessary topreferentially cool the lower body. After the rectal tempera-

ture had reached 18°C, the flow was reduced and the reser-voir drained. An additional 3 liters of Hypothermosol Msolution was placed into the reservoir. At this point, thehematocrit was less than 2%. The venous effluent changedcolor from dark red to a bright red color (same as that of theblood in the arterial line) at approximately 30°C. The animalwas then cooled to an esophageal and rectal temperature of10°C. At this point, the flow was reduced to 1 to 2 L/min anda 7-mm aortic cannula placed into the root of the aorta (Fig.2B). The aorta around the Foley catheters was then dissectedand vascular clamps placed. The Foley catheters were re-moved and the aortic laceration was repaired with running4-0 Surgilon. After a total of 90 minutes at 10°C, the fluid inthe CPB reservoir was drained and the circuit flushed with 7liters of Hypothermosol P (1°C) solution (Table 1) at 2-literincrements. When Hypothermosol P solution containing 3.0mmol/L K1 was infused into the animal, spontaneous ven-tricular contractions occurred at a rate of 10 per minute. Thereservoir was then drained and 2 liters of whole blood (1°C)containing citrate and glucose (Tom Morris Farms, MD) wasplaced in the reservoir, and the rewarming process was ini-tiated. A rewarming rate of 0.5°C/min was established bymaintaining a maximal differential of 10°C between the rectaltemperature and the heat exchanger. At 18°C, blood wasdrained from the reservoir and exchanged with an additional

FIG 2. (A) Photograph of Silastic Foley catheter in aorta, venous cannula in rightatrium. (B) Descending aorta repaired, 7-mm aortic cannula in root of aorta, andvenous cannula in right atrium.

Vol. 48, No. 3 Induced Hypothermia

441

2 liters of whole blood. At this stage, the cardiac activityincreased to a rate of 50 per minute. At 32°C, the blood thatwas previously collected through the chest tube during theinitial exsanguination was placed into the reservoir. When therectal temperature reached 32 to 35°C, calcium chloride,lidocaine, and sodium bicarbonate were used as needed forventricular fibrillation (65% of the time). Cardioversion withinternal paddles was performed as necessary. At this point,spontaneous motor activity returned and anesthesia was re-instituted with 1 to 2% isoflurane and pancuronium (0.1mg/kg). After rewarming, the animal was allowed to stabilizeand was separated from CPB. Separation from CPB requiredepinephrine in some cases. All animals were given one doseof dexamethasone (4 mg). The subclavian arterial line wasremoved. The chest was closed in layers with the chest tubein place. The last animal in this group was maximally cooledto a nadir rectal temperature of 2.7°C.

Postoperative anesthesia was maintained with isoflurane.Extubation was performed when normal hemodynamicscould be maintained without vasopressor support. Before

extubation, the chest tube and the pulmonary artery catheterwere removed and the neck incision was closed. Buprenor-phine was given before extubation for pain relief. The animalwas then placed in a warm incubator. Water was encouragedas tolerated and diet resumed at 24 hours. Postoperativeantibiotics were provided for up to 1 week for fevers orelevated white blood cell count. Postoperative pain medica-tion was administered for comfort with injection of buprenor-phine intramuscularly. Weekly neurologic testing was doneby using criteria as shown in Table 2. Blood was drawnweekly for chemical analysis and complete blood count.

Group II: Complete Circulatory Arrest ModelThree animals were used to determine whether completecirculatory arrest could be achieved with survival. All proce-dures were identical to group I, except that once 10°C hadbeen achieved, the Foley catheters in the thoracic aorta wereclamped and disconnected from the bypass circuit for 20, 30,and 40 minutes. This was done because complete cardiopul-monary arrest might be necessary in some situations to get

TABLE 2. Neurologic examination scoring sheet

Level of ConsciousnessAwake/normal 4Lethargic/arousable but not

to full normal state3

Obtunded/some reaction withstimulation

2

Coma 1Behavior

Normal 4Needs gentle prodding 3Needs vigorous prodding 2No interaction/spontaneous

activity1

Feeding Drinking

Normal 3 Normal 3Needs assistance 2 Needs assistance 2Won’t eat 1 Won’t drink 1

Cranial Nerves Pupils Corneal Gag Tongue

Left Right Left Right

Reactive 3 3 3 3 3 Midline 3Sluggish 2 2 2 2 2 Deviated L/R 2None 1 1 1 1 1 Limp 1

Motor Right Front Left Front Right Hind Left Hind

Normal 4 4 4 4Weak (mild) 3 3 3 3Weak (severe) 2 2 2 2Plegia (no mvmt) 1 1 1 1

Sensory Right Front Left Front Right Hind Left Hind Tail

Normal 4 4 4 4 4Impaired 3 3 3 3 3Only pain 2 2 2 2 2No sensation 1 1 1 1 1

Coordination Locomotion

Stands normally 4 Normal 3Stands but sways 3 Ataxia but more than 2 steps 2Stands but falls over 2 Cannot walk more than 2 steps 1Cannot stand 1

Total maximum score is 75

The Journal of Trauma: Injury, Infection, and Critical Care March 2000

442

control of the source of hemorrhage. If the protective effectcould be maintained without CPB for a short time, then itwould also simplify transport to a higher level facility.

Histologic Assessment of the BrainSix weeks after survival, the animals were anesthetized andunderwent transcardiac perfusion with 3.7% formaldehyde in0.1 mol/L phosphate-buffered saline (pH 7.4) after flushing withsaline. The brains were removed and placed in the same fixativeovernight at 4°C. Each perfused brain was cut into 4-mm thickcoronal sections and examined for gross lesions. Brain blockswere embedded in paraffin and 6-mm thick sections were cut.The following stains were performed: hematoxylin-eosin, im-munohistochemical localization of glial fibrillary acidic protein(GFAP), MAP-2, NF-68, NF-160, synaptophysin and Bax. An-tigen retrieval procedure was applied before immunostaining.The evaluation was done with 403 magnification of the follow-ing brain areas according to Stereotaxic Atlas of the Pig Brain:nucleus caudatus, putamen, frontal cortex, prepiriform cortex,and hippocampus.16

The following cell population and structures were evalu-ated: First, neurons, especially the presence of early (“eosin-ophilic neurons”) and late (“neuronal ghosts”) ischemicchanges. Early changes were described as loss of Nissl sub-stance and the development of cytoplasmic eosinophilia,shrinkage of the perikaryon, and relative preservation ofnuclei with indistinct nucleoli. The term “neuronal ghost”was used to refer to a state when cellular shrinkage becomesmore pronounced, neuronal nucleus disappears, leaving ahomogeneous, eosinophilic silhouette of the cell. The numberof neurons was determined for a uniform 1-mm2 area and for1-mm length of the central area of the hippocampal CA1subfield. The degeneration rates were calculated by using thefollowing formula: Degeneration rate5 (number of degen-erating neurons, both eosinophilic and ghosts/total number ofneurons)3 100.

Second, glial cells (GFAP). The increases in the gliotypicprotein GFAP were caused by astrocytic hypertrophy andhyperplasia, a well-known consequence of acute neuronaldamage. The response of astrocytes to the brain injury, asmonitored immunohistochemically, was reflected by substan-tial increases in GFAP immunoreactivity in affected regions.Rabbit anti-cow GFAP (DAKO, Z0334, 1:1,000) was used toshow the reactive gliosis, determined by an increase in GFAPimmunoreactivity. GFAP immunoreactivity was detectedwith secondary anti-rabbit IgG F(ab)2 fragment conjugated tohorseradish peroxidase (Sigma).

Third, neuronal axons (NF) and dendrites (MAP-2, synapto-physin). Monoclonal mouse anti-pig neurofilament 160 (SigmaN5264, 1: 500) was used for the localization of the neurofila-ment of molecular weight 160,000, and monoclonal mouse anti-pig neurofilament 68 (Sigma, N5139, 1:500) was used to local-ize the NF with molecular weight of 68,000. These were locatedbecause both NF are cytoplasmic and are two of the three majorcomponents of the neuronal cytoskeleton. Three protein subunitsNF 68, NF 160, and NF 200 maintain a neuron’s highly asym-metrical shape and provide mechanical strength and stability tothe soma and particularly to the axon. Monoclonal mouse anti–

MAP-2 (Sigma, M4403, 1:1,000) was used to identify microtu-bule-associated protein, located exclusively in neuronalperikarya and dendrites. Loss of MAP-2 precedes the significantcell loss after cerebral ischemia. Rabbit anti-cow GFAP(DAKO, Z0334, 1:1,000) was used to show the reactive gliosis,determined by an increase in GFAP immunoreactivity. GFAPimmunoreactivity was detected with secondary anti-rabbit IgGF(ab)2 fragment conjugated to horseradish peroxidase (Sigma).Anti-mouse IgG, Fab specific, horseradish peroxidase labeled(Sigma, A9917) was used for detection of MAP-2, Bax, NF 68,and NF 160. The signal was enhanced by using the tyramideamplification system (NEN Life Science Products). Vector VIP(Vector Laboratories) was used as substrate medium. Selectedsections were counterstained with hematoxylin.

RESULTS

The first 16 animals were used to develop the model. Theremaining 10 animals were used for the survival studies.Mean values are reported with standard deviation.

Physiologic DataDural temperatures and rectal temperatures lagged the esoph-ageal temperatures by 1.5 to 3°C and 2 to 5°C, respectively(Fig. 3). These differences were dependent on the flow ratesand were greatest during rapid cooling. The EEG recordingsbecame isoelectric when the systolic blood pressure fell be-low 55 mm Hg or if the esophageal temperature was lowerthan 30°C. The EEG activity typically started to return within2 hours after the animal was warmed to 37°C.

The blood K1 level after the infusion of Hypothermosol Msolution increased to 10.96 1.3 mmol/L and fell to 6.66 1.1mmol/L after rinsing with Hypothermosol P solution. At thetime of separation from the cardiopulmonary bypass pump,the K1 level fell spontaneously to 4.56 0.8 mmol/L. Thechanges in the physiologic variables, blood gases and chem-istry during the experiments are shown in Table 3. The spunhematocrit during the 10°C arrest period was less than 2%.Average pump time was 203 minutes6 36 minutes.

SurvivalGroup I: Hypothermic Cardiac Arrest with Low Flow for90 MinutesFive of the seven animals survived. Four of the five survivorswere fully functional and had normal behavior. The neuro-

FIG 3. Temperature schematic during the procedure.

Vol. 48, No. 3 Induced Hypothermia

443

logic examination scores are shown in Table 4. The animalsgained weight appropriately during the 6-week period. Ani-mal 6 survived with subtle but notable neurologic deficitsmanifested by deviation of the tongue to the left and weak-ness of the left hind limb. Otherwise, this animal was func-tionally normal. Animal 4 in this series died of air embolismthrough the intravenous tubing during the immediate postop-erative period. Animal 7 in this series was cooled as much aspossible to an esophageal temperature of 2.7°C. This animalhad gross evidence of neurologic injury with lethargy, weak-ness, and deviation of the tongue immediately upon recovery.This animal died within 24 hours. Upon autopsy, fluid wasfound in the thoracic and abdominal cavity. The postopera-tive chemistries and blood count values of the survivors areshown in Table 5. The hematocrit returned to the baselinelevels, whereas the blood chemistries demonstrated a mildelevation in the total bilirubin, aspartate transaminase andg-glutamyltranspeptidase that persisted during the 6-weekperiod.

Group II: Complete No Flow Hypothermic AcellularArrest Model OutcomeTwo animals that underwent no flow hypothermic cardiopul-monary arrest for 20 minutes (animal 1) and 30 minutes(animal 2) survived without sequelae. Animal 2 survivedwithout clinically notable neurologic deficits, but the initialneurologic examination score was 60 on the first day aftersurgery. One week after surgery, the neurologic examinationscore was 75 and the animal seemed normal. Animal 3 in thisseries underwent 40 minutes of no flow arrest and died

because of accidental extubation during recovery 1 hour afterit was weaned off CBP (Table 4).

Histologic AssessmentsOnly five of the seven brains (three from group I and twofrom group II) were available for examination because oftechnical errors resulting in inadequate perfusion. Two ani-mals in group I and the animal in group II that had no flowfor 20 minutes had normal neurologic examination scores andhad no abnormal histopathologic changes observed in anybrain areas. The hippocampal pyramidal cells were well pre-served, and no disruptions of radial striated zone were de-tected. In addition, there was no appreciable loss of neuronsin CA1, cortex, and caudate-putamen and no ischemic neu-rons were found in nucleus caudatus and putamen.

Animal 6 in group I had residual neurologic deficits uponclose examination, with weakness of the left hind limb anddeviation of the tongue to the left. Macroscopical and micro-scopical examination of its brain revealed gross ischemia andnumerous hemorrhagic necrotic areas in cortex, hippocam-pus, and thalamus. The degeneration rates were found to bebetween 59.4% and 95.5%, depending on the area analyzed.Animal 2 in group II had normal neurologic exam scores afterthe first week but demonstrated high degeneration rates:45.7% and 51.1% in cortex, 61% in CA1.

Profound gliosis (increase in intensity of immunostainingand number of astrocytes positive for GFAP) was detected inboth of the brains with the ischemic changes described above.Hyperplastic astrocytes were irregular in outline and swollen.Astrocytes in the brains of the animals with normal degen-

TABLE 3. Physiologic, blood gas, and electrolyte valuesa

Parameter Baseline End of Cooling Beginning ofWarming Off CPB

CO (L/min) 4.4 6 1.7 NA NA 5.7 6 2.3MAP (mm Hg) 96.0 6 20.8 75.7 6 14.2pH 7.6 6 0.0 7.3 6 0.1 7.2 6 0.1 7.4 6 0.1BD (mmol/L) 5.4 6 2.1 25.6 6 3.3 26.5 6 3.5 22.5 6 4.7Lactate (mmol/L) 1.4 6 0.6 8.8 6 3.7 7.9 6 1.8 12.9 6 2.8Potassium (mmol/L) 3.8 6 0.3 10.9 6 1.3 6.6 6 1.1 4.5 6 0.8Magnesium (mmol/L) 0.35 6 0.2 0.85 6 0.1 0.36 6 0.04 0.4 6 0.1Calcium (mmol/L) 1.1 6 0.2 0.5 6 0.0 0.5 6 0.0 0.8 6 0.2Hematocrit (%) 30.4 6 2.7 1.7 6 0.3 16.9 6 2.2 23.3 6 3.3

a Values are means 6 standard deviation. CO, cardiac output; MAP, mean arterial pressure; BD, base deficit.

TABLE 4. Neurologic score and histopathology results

Animal Day 1 Week 1 Week 2 Week 3 Week 4 Week 5 Week 6 Histopathology Results

Group IAnimal 1 Survived fully intact neurologically but scores not available NAAnimal 2 75 75 75 75 75 75 75 NormalAnimal 3 75 75 75 75 75 75 75 NormalAnimal 4 Arrested 1 hour after surgery, suspected air embolism from i.v. line NAAnimal 5 75 75 75 75 75 75 75 NormalAnimal 6 75 75 72 73 73 73 73 Ischemic changes notedAnimal 7 Died 24 hours postoperatively NA

Group IIAnimal 8 75 75 75 75 75 75 75 NormalAnimal 9 60 75 75 75 75 75 75 Ischemic changes notedAnimal 10 Died, accidentally extubated

The Journal of Trauma: Injury, Infection, and Critical Care March 2000

444

eration rates were evenly distributed through the cerebralcortex with no signs of hypertrophy and swelling. The num-bers of reactive astrocytes in animals with low degenerationrates (histologically clean brains) were significantly differentthan in the group of animals with high degeneration rates(Fig. 4).

Stainings for MAP-2, NF-68, NF-160, and synaptophysinconfirmed the findings and showed similar trends. In allanimals with low degeneration rates, the immunoreactiveexpression of Bax was restricted to cell cytoplasm. Further-more, preferential localization to nuclei described for apopto-tic and necrotic neurons was not detected.17 In animals withhigh degeneration rates, signs of ischemic injury werepresent. This was seen in the animal with apparent neurologicdeficit (animal 6, group I) and in one animal with abnormalneurologic examination scores (animal 2 in group II). Re-

maining animals had no evidence of ischemic injury based onhistologic data.

DISCUSSION

The main goal of this study was to determine whether pro-found acellular hypothermic arrest could be induced throughthe chest after exsanguinating hemorrhage in a swine model.We purposely chose techniques and equipment that are avail-able in most trauma centers. This model has shown that it ispossible to induce acellular profound hypothermic arrest afterexsanguinating hemorrhage and then resuscitate the animalwithout detectable neurologic injury.

The recovery seemed to be complete in four of the fiveanimals that underwent 90 minutes of low flow acellular deephypothermia (10°C). No neurologic deficits were identifiedon clinical or histologic examinations. One animal in thisgroup survived with residual neurologic deficits. All animalsthat survived were extubated within 1 to 6 hours of closingthe chest. They were able to take water within 12 hours andusually fed themselves within 24 hours. Within 36 hours, theywere fully ambulatory with return of preoperative behavior.There was a mild elevation in the total bilirubin, aspartatetransaminase, andg-glutamyltranspeptidase during the6-week postoperative period. A similar rise in these enzymeshas also been reported in the canine model.18 One animal hadpersistent fever and leukocytosis that resolved with adminis-tration of antibiotics for 1 week.

The last animal in group I was cooled as much as possibleto a nadir temperature of 2.7°C and was kept at this temper-ature for 90 minutes. This was attempted because Tishermanet al. showed that cooling to 10°C resulted in better neuro-logic outcome than cooling to 15°C in dogs.19 Thus, we

TABLE 5. Postoperative blood count and serum chemistry valuesa

Parameter Week 1 Week 2 Week 3 Week 4 Week 5 Week 6

Hematocrit (%) 25.2 6 8.8 26.5 6 5.5 29.9 6 4.6 30.2 6 3.7 30.2 6 2.5 34.0 6 3.4White count (k/mL) 7.4 6 3.8 14.0 6 6.5 11.1 6 2.0 12.2 6 2.5 13.2 6 2.3 11.7 6 2.2Platelets (k/mL) 331.4 6 197.2 534.2 6 115.6 517.8 6 142.2 470.1 6 107.9 392.0 6 177.4 368.7 6 74.0

Sodium (142–149 mmol/L) 141.9 6 4.0 138.6 6 3.6 137.7 6 3.3 137.0 6 5.3 136.7 6 6.8 140.7 6 2.1Potassium (2.1–7.1

mmol/L)4.4 6 0.7 4.0 6 0.4 3.8 6 0.2 4.1 6 0.3 4.3 6 0.4 4.2 6 0.6

Glucose (85–160 mg/dL) 113.7 6 100.5 80.0 6 28.8 84.5 6 21.8 74.6 6 7.7 91.0 6 17.6 87.3 6 19.9Urea nitrogen (6–30 mg/

dL)7.3 6 1.1 6.4 6 1.9 7.5 6 1.8 7.3 6 1.0 9.9 6 2.5 9.8 6 3.3

Creatinine (0.5–2.1 mg/dL) 1.2 6 0.1 1.1 6 0.4 1.2 6 0.3 1.5 6 0.1 1.7 6 0.2 1.4 6 0.2Calcium (5–11.4 mg/dL) 10.6 6 1.7 9.5 6 2.6 9.7 6 0.7 9.2 6 0.3 9.3 6 0.4 9.7 6 1.0Total protein (6–8 g/dL) 5.5 6 0.7* 5.9 6 0.4* 6.1 6 0.6 6.2 6 0.4 6.2 6 0.3 6.2 6 0.5Total bilirubin (0.0–0.3

mg/dL)0.4 6 0.1 0.3 6 0.1 0.4 6 0.1 0.4 6 0.1 0.5 6 0.3 0.4 6 0.2

Alk Phos (92–294 U/L) 60.9 6 14.3* 78.4 6 41.4* 121.3 6 18.7 129 6 34.9 131.7 6 30.1 109.5 6 35.2LDH (575–3294 U/L) 6913.8 6 2521.5 2929.4 6 1201.2 2089.5 6 247.6 1671.7 6 249.2 1597.4 6 170.0 1565.8 6 262.7CK (50–3531 U/L) 482.3 6 252.6 334.3 6 171.9 547.2 6 144.7 708.9 6 362.8 690.7 6 268.9 684.0 6 395.3AST (16–65 U/L) 43.0 6 12.0 27.9 6 9.5 28.0 6 6.2 28.3 6 5.8 33.0 6 4.8 29.7 6 10.1ALT (9–43 U/L) 90.3 6 28.7 44.6 6 12.7 50.2 6 12.6 55.4 6 10.6 59.6 6 7.3 56.0 6 18.4GGT (16–30 U/L) 34.3 6 9.8 28.3 6 12.1 37.2 6 8.7 40.7 6 9.0 43.3 6 9.6 39.3 6 12.7Amylase (271–1198 U/L) 425.3 6 59.2 427.9 6 62.6 561.2 6 105.4 608.0 6 93.1 625.7 6 31.0 478.3 6 225.3

Mean chemistry values of survivors 6 standard deviation. , above normal range; *, below normal range. LDH, lactate dehydrogenase; CK,creatinine kinase; AST, alanine aminotransferase; ALT, aspartate transaminase; GGT, g-gutamyltranspeptidase.

FIG 4. Astrogliosis in animals with high (A) and low (B) degeneration rates.FC,frontal cortex; PC, prepiriform cortex; NC, nucleus caudatus;P, putamen. Theasterisks denotep < 0.05, by using Student’st test.

Vol. 48, No. 3 Induced Hypothermia

445

wanted to determine whether further decrease in temperature(, 10°C) offered better protection to the brain. However,upon resuscitation, the animal demonstrated neurologic in-jury within 3 hours of extubation. It has been previouslyshown that cooling below 5°C results in neurologic injury,pulmonary dysfunction, and systemic edema in dogs.5,20 Be-cause we performed only one animal experiment with max-imal cooling, it is not known whether this injury was causedby the nadir temperature or other technical problems.

The three animals in group II underwent complete circu-latory arrest for various time periods (20, 30, and 40 min-utes). This experiment represented a clinical scenario inwhich total cessation of flow might be helpful during surgicalrepairs. The second animal was not as responsive on postop-erative day 1 and 2, which was reflected in the neurologicscoring. This animal recovered completely by the third post-operative day with no clinical evidence of neurologic injury.However, the histologic examinations showed injury to thebrain. This emphasized the need for both the clinical andhistologic correlation. The third animal in this set of experi-ments that was left in arrest for 40 minutes recovered for 20hours, but demonstrated obvious neurologic deficits immedi-ately postoperatively. The animal was found dead 24 hoursafter surgery.

Of note, when the animals were reconnected to initiate lowflow, the venous blood/fluid was consistently dark with theoxygen content lower than arterial blood. Therefore, somemetabolism occurred during the no flow state, even at theselow temperatures. The esophageal temperature spontaneouslyincreased 1 to 1.5°C during the no flow state. In general, thevenous return usually became bright red and equal in color tothe arterial flow at temperatures around 30°C.

Because hypothermia protects against ischemia, it is key torapidly induce hypothermia for the purpose of protecting thebrain. This protection allows additional time for surgicalrepair. We chose to induce hypothermia through the chestbecause EDT is currently the standard approach in an exsan-guinating patient not responsive to resuscitation. Induction ofhypothermia through the femoral approach has been usedsuccessfully in a canine model, but because the chest isopened during the EDT and the descending thoracic aorta isreadily available, it was believed that this makes the thoracicapproach more clinically appealing. It can be expected thathalf of exsanguinating patients would require exploration ofthe chest cavity.

The highest survival rate from EDT is when the source ofhemorrhage is in the chest and is easily treatable. In thesecases, it may be more logical to try primary repair without theinduction of hypothermic arrest. There may be situationswhere induction of hypothermia would be more appropriatethrough the groin or in combination. If rapid high flows canbe achieved selectively by the development of specially de-signed catheters it would help to advance this field of re-search. Woods et al. have shown that a 500-mL 4°C salineflush by using a catheter with an occluding balloon designedto flush the heart and brain can provide protection afterprolonged exsanguination and 30 minutes of completearrest.21 This technique may also provide additional time to

access the chest. However, obtaining femoral vessel access ina hypovolemic patient may be problematic.

Ordinary Silastic urinary Foley catheters were used toinduce hypothermia because our intention was to use tech-niques and materials that are available in most trauma cen-ters. CPB roller pumps with a membrane oxygenator and heatexchanger are also available in most Level I trauma centers.Because there are rapid infusers and warmers primed andready for use in most hospitals, it was envisioned that, withmodification, a refrigerator unit with a reservoir could bemanufactured with pumps that could stand ready for use toinduce hypothermia.

The use of the precooled Hypothermosol M solution wasanother novel approach tested in these experiments. Thesolution contains 42.5 mmol/L of K1, which helps to arrestmetabolism and passive ionic exchange in cells when meta-bolic pumps are switched off during hypothermia. However,in our model, the serum K1 rose only to a maximal level of10.9 mmol/L. The Hypothermosol M solution was specifi-cally designed as an aqueous blood substitute for total bodyhypothermic perfusion. It embodies many of the principlesnow identified as contributory and important for success inorgan preservation.5,22–24 In contrast to the experiment byTaylor et al., we used the Hypothermosol M solution initiallyto arrest metabolism, whereas they induced hypothermia withHypothermosol P.15 The use of Hypothermosol P solutionalone had resulted in motor and sensory deficits.

The use of acellular solutions may offer an additionaladvantage in that precious blood is not lost while the surgicalrepair of the injury is being performed. The resources ofbanked blood are, thus, not being used until the source ofbleeding has been controlled. The use of manufactured acel-lular fluid is less costly than banked blood. We purchased theHypothermosol M solution for these experiments at a cost of$60.00 per liter. The cost of packed red blood cells can bemore than $250.00 for one 250-mL unit.

The limits of induced hypothermia have been tested overdecades.25–27 Recently, this field has made marked progressdue principally to the work done by Safar, Taylor, and theircolleagues. However, the models used in all these experi-ments were canine. Because swine cardiopulmonary and im-mune physiology more closely resembles that of human, webelieved that testing hypothermic arrest in this animal modelwould add valuable information to this field. The canine heartis relatively robust; in contrast the swine heart is very sensi-tive to arrhythmias. The swine is higher on the evolutionaryscale and its anatomy is closer to human.28

The model presented in this study does have several draw-backs. The exsanguinating hemorrhage was acute, and it maybe more clinically representative if the exsanguination periodwas longer. We plan to perform a set of experiments with aslower bleed from an abdominal vascular source. In thismodel, subtle neurologic deficits such as memory were notassessed. Further refinements in this model would includeperforming preoperative training and postoperative testing toassess learning and memory.

The induction of hypothermic arrest should not be com-pared with spontaneous hypothermia in trauma patients. It

The Journal of Trauma: Injury, Infection, and Critical Care March 2000

446

has been shown that spontaneous hypothermia after trauma isassociated with worse outcome29,30 and that active rewarm-ing may be beneficial.31 If the perfusion to tissues falls belowsome threshold, metabolism cannot keep up with the heatloss. Therefore, ischemia from hemorrhagic shock can con-tribute to hypothermia. Rewarming in these situations maydecrease the volume of blood and fluid needed, as coagu-lopathy becomes an important consideration during resusci-tation. Thus, the concept of rewarming spontaneously coldpatients is different than the induction of deep hypothermicarrest in normothermic patients to help establish surgicalcontrol of hemorrhage.

Surgeons facing the exsanguinating patient with potentiallyrepairable injuries often find themselves needing more time.Induced hypothermic arrest can provide this time. The ques-tion is whether hypothermia can be applied in the traumasetting. Here, we demonstrate a technique that can be appliedin combination with EDT in the setting of exsanguinatingarrest. We tested this in a swine model of uncontrolled ex-sanguination and found it to be possible. The technique ofinducing hypothermic arrest through the chest could be ofbenefit in a clinical setting.

REFERENCES

1. Bellamy R, Safar P, Tisherman S, et al. Suspendedanimation for delayed resuscitation.Crit Care Med.1996;24:S24–S47.

2. Simon D, Taylor MJ, Elrifai AM, et al. Hypothermic bloodsubstitution enables resuscitation after hemorrhagic shockand 2 hours of cardiac arrest.ASAIO J.1995;41:M297–M300.

3. Branney SW, Moore EE, Feldhaus KM, Wolfe RE. Criticalanalysis of two decades of experience with post injuryemergency department thoracotomy in a regional traumacenter.J Trauma.1998;45:87–95.

4. Rhee P, Acosta J, Bridgeman A, et al. Survival afteremergency department thoracotomy: review of publisheddata from the past 25 years.J Am Coll Surg.In press.

5. Taylor MJ, Elrifai AM, Bailes JE. Hypothermia in relationto the acceptable limits of ischemia for bloodless surgery.Adv Low Temp Biol.1996;3:1–64.

6. Hickey PR. Deep hypothermic circulatory arrest: current statusand future directions.Mt Sinai J Med.1985;52:541–547.

7. Baumgartner WA, Silverberg GD, Ream AK, et al.Reappraisal of cardiopulmonary bypass with deephypothermia and circulatory arrest for complex neurosurgicaloperations.Surgery.1983;94:242–249.

8. Wells FC, Coghill S, Caplan HL, et al. Duration ofcirculatory arrest does influence the psychologicaldevelopment of children after cardiac operation in early life.J Thorac Cardiovasc Surg.1983;886:823–831.

9. Haneda K, Sands MP, Thomas R, et al. Prolongation of thesafe interval of hypothermic circulatory arrest: 90 minutes.J Cardiovasc Surg.1983;24:15–21.

10. Spetzler RF, Hadley MN, Rigamonti D, et al. Aneurysms ofthe basilar artery treated with circulatory arrest,hypothermia, and barbiturate cerebral protection.J Neurosurg.1988;68:868–879.

11. Michenfelder JD. The hypothermic brain. In: MichenfelderJD, ed.Anesthesia and the Brain. Baltimore: Williams &Wilkins; 1987:23–34.

12. Hickey PR. Deep hypothermic circulatory arrest: current statusand future direction.Mt Sinai J Med.1985;52:541–547.

13. Livesay JJ, Cooley DA, Reul GJ, et al. Resection of aorticarch aneurysms: a comparison of hypothermia techniques in60 patients.Ann Thorac Surg.1983;36:19–28.

14. Capone A, Safar P, Radovsky A, et al. Complete recoveryafter normothermic hemorrhagic shock and profoundhypothermic circulatory arrest of 60 minutes.J Trauma.1996 40;388–395.

15. Taylor M J, Bailes JE, Elrifai AM, et al. A new solution forlife without blood: asanguinous low-flow perfusion of awhole-body perfusate during 3 hours of cardiac arrest andprofound hypothermia.Circulation. 1995;91:431–444.

16. Felix B, Leger ME, Albe-Fessard D. Stereotaxic atlas of thepig brain.Brain Res Bull.1999;49:1–138.

17. Li Y, Chopp M, Powers C. Granule cell apoptosis and proteinexpression in hippocampal dentate gyrus after forebrainischemia in the rat.J Neurol Sci.1997;150:93–102.

18. Leavitt M, Bailes JE, Elrifai AM, et al. Blood parametersfollowing extracorporeal circulation of a blood substituteduring profound hypothermia in dogs.Proc Am AcadCardiovasc Perf.1992;13:49–53.

19. Tisherman S, Safar P, Radovsky A, et al. Profoundhypothermia (,10°C) compared to deep hypothermia (15°C)improves neurologic outcome in dogs after two hourscirculatory arrest induced to enable resuscitative surgery.J Trauma.1991;31:1051–1062.

20. Haneda k, Thomas R, Sands MP, et al. Whole bodyprotection during three hours of total circulatory arrest: anexperimental study.Cryobiology.1986;23:483–494.

21. Woods RJ, Safar, P, Takasu A, et al. Hypothermic aorticarch flush for preservation of brain and heart duringprolonged exsanguination cardiac arrest in dogs [abstract].J Trauma.1998;45:1116.

22. Belzer FO, Southard JM, Van Gulik TM, et al. Principles ofsolid-organ preservation by cold storage.Transplantation.1988;45:673–676.

23. Southard JH, Van Gulik TM, Ametani MS, et al. Importantcomponents of the UW solution.Transplantation.1990;49:251–257.

24. Southard JH, Belzer FO. New concepts in organpreservation.Clin Transplant.1993;7:134–137.

25. Connolly JE, Roy A, Guernsey J, Stemmer EA. Bloodlesssurgery by means of profound hypothermia and circulatoryarrest: effect of brain and heart.Ann Surg.1965;162:724–737.

26. Swan H, Virtue R, Blount SG, et al. Hypothermia insurgery: analysis of 100 clinical cases.Ann Surg.1955;142:382–400.

27. Edmunds LH, Folkman J, Snodress AB, Brown RB.Prevention of brain damage during profound hypothermiaand circulatory arrest.Ann Surg.1963;157:637–649.

28. Schaper W, Gorge G, Winkler B, Schaper J. The collateralcirculation of the heart.Prog Cardiovasc Dis.1988;31:57–77.

29. Luna GK, Maier RV, Palvin EG, et al. Incidence and effectof hypothermia in seriously injured patients.J Trauma.1987;27:1014–1018.

30. Jurkovich GJ, Greiser WB, Luterman A, Curreri PW.Hypothermia in trauma victims: an ominous predictor ofsurvival.J Trauma.1987;27:1019–1024.

31. Gentilello LM, Jurkovich GJ, Stark MS, et al. Ishypothermia in the victim of major trauma protective orharmful? A randomized prospective study.Ann Surg.1997;226:439–449.

DISCUSSION

Dr. Larry Gentilello (Seattle, Washington): I would like tothank the Association for extending me the privilege of dis-cussing this paper, and I would also like to congratulate theauthors on their work. It was a well-constructed scientific

Vol. 48, No. 3 Induced Hypothermia

447

study that was nicely presented and accompanied by a well-written manuscript.

Currently, recommendations are to prevent hypothermiaand to aggressively treat it once it has occurred. These rec-ommendations are based on historically controlled, prospec-tively controlled, and even randomized clinical trials thatdemonstrate that in humans hypothermia significantly re-duces the likelihood of successful resuscitation, and is anindependent risk factor for mortality.

On the other hand, the protective efforts of controlled deephypothermia during conditions of complete oxygen depriva-tion are obvious and well known. The length of time thatcirculatory arrest is safe is far longer than can be explained bysimple metabolic suppression of oxygen demand, and likelyinvolves synergistic suppression of a number of deleteriouschemical and metabolic pathways that mediate secondaryorgan damage.

It is natural and realistic to believe that some day traumasurgeons will develop techniques for reaping the potentialbenefits of hypothermia to treat severely injured patients.And while futuristic, this is an extremely important topic andworthy of support.

However, while the theoretical concepts behind the studyare sound, to date successful schemes for reaping the poten-tial benefits of hypothermia in trauma patients have beenextremely elusive. Even the use of mild hypothermia forcerebral protection following traumatic brain injury which,after a long series of failures in the 1980s was revived in the1990s, showed more harm than benefit in a recent NIH-sponsored randomized prospective multi-center nationalacute brain injury study, due to complications associated witheven slight decreases in core temperature.

Subjecting patients who are already candidates for emer-gency department thoracotomy to deep hypothermic cardiacarrest may likewise introduce serious problems related to theintrinsically deleterious properties of hypothermia itself.

In this study, which in fairness was constructed as a fea-sibility trial not an outcome trial, seven of nine animalssurvived 5 minutes of severe hypotension. One had obviousfunctional deficits. Another had approximately 50% degen-eration of the cortex and 60% degeneration of the hippocam-pus, with no apparent functional abnormalities, which high-lights the limitations of neurologic functional outcome testingin pigs. Thus, there were only five out of nine so-callednormal survivors of a period of only 5 minutes of severehypotension.

None of these comments should be taken as disparagingthe importance of the work. Breakthroughs in reanimationresearch are likely to occur in the future, and this study wasconducted to test a new technique for accomplishing deephypothermic cardiac arrest in the emergency department set-ting through a standard thoracotomy incision, and the authorswere successful in demonstrating the feasibility of theirmethod. If reanimatology develops a role in trauma surgery,their technique may be a major contribution.

I have several questions for the authors. These animalswere hypotensive for only 5 minutes, and while the bloodpressure was low, it was stable for those 5 minutes, and then

you initiated hypothermia. What recommendation would yougive to a clinician who is trying to decide that circulatoryarrest is superior to fluid resuscitation, or is the only means ofsalvaging such a patient, and, therefore, they should take themajor step of stopping a beating heart?

My second question is, would your experimental designhave been more realistic if hemorrhage was truly uncon-trolled? Unfortunately, most studies on hypothermic cardiacarrest use controlled phlebotomy; bleed then stop the bleed-ing, then introduce hypothermia. This is less of a problemthan uncontrollable bleeding. Do you think it would still beeasy to use this technique in a patient who is having massivebleeding in their chest?

My third question is that after rewarming from deep hy-pothermia, there is usually an increased antithrombin andfibrinolytic effect, as well as a relatively persistent thrombo-cytopenia, resulting in impaired coagulation. That is not aproblem in a Wigger’s type of model where bleeding iscontrolled, or when repairing cerebral aneurysms and you canuse meticulous surgical technique. My concern is about theeffect of coagulopathy on patients with major liver injuriesthat can only be partially repaired, pelvic fractures, and braininjuries.

Did you obtain coagulation data on these pigs, or plateletcounts and studies of their function after resuscitation?

My next question is, did you obtain base deficits or lactatelevels prior to inducing hypothermia? The goal of inducinghypothermia is to suppress utilization of energy during hyp-oxia so that the body can survive longer with a given amountof ATP. Elective cardiac or neurosurgical patients are not inshock prior to induction of hypothermia. Would the results bethe same if hypotension was prolonged and the animals wereprofoundly acidotic with already depleted ATP scores, whichis the clinical situation we most often face?

Finally, studies on deep hypothermic cardiac arrest dem-onstrate persistent low cardiac output and an uncoupling ofoxygen delivery, blood flow, and metabolic rate during re-warming. Even after rewarming, deep hypothermia impairssubsequent cerebral blood flow, and studies have confirmedthat there is considerable jugular venous desaturation in pa-tients following hypothermic cardiac arrest and rewarming.

These animals were not brain injured. Do you have anopinion on whether or not you would experience the sameneurologic outcomes in the pressure-passive, brain-injuredpatients with depressed cardiac output and persistent cerebralvasoconstriction?

I would again like to congratulate the authors for con-tributing a new technique for instituting hypothermic car-diac arrest under conditions most likely to be encounteredin the emergency department. They bring up numerouscomplex but important issues and questions to be pursuedin order to establish the scientific justification for inducinghypothermia in critically injured patients. And I would liketo, again, thank the Association for the privilege of dis-cussing this paper.

Dr. Jack Bergstein (Morgantown, West Virginia): Dr.Rhee, I very much enjoyed your model and your presentation,and perhaps this comes a little closer to something that we

The Journal of Trauma: Injury, Infection, and Critical Care March 2000

448

can do in the emergency department than the models thathave come out of the Safar group.

It was not clear to me, though, when you inserted yourbypass cannulas. It appeared from your slide that the cannulaswere already in when you induced hemorrhage and hypother-mic flush. What I would ask you is how long does it take toput in those cannulas? I would expect that clinically thosewould go in after the patient is flushed.

And furthermore, how long would you expect to set up andprime the pump system? I am glad to see that you arestudying states of no flow, because I expect it is going to takeat least half an hour to get things running, particularly if youconsider the time to get the pump tech. into the trauma roomand things primed and going.

Dr. Matthew J. Wall, Jr (Houston, Texas): I would like tocongratulate you on a very thought-provoking study.

At what dural temperature did the EEGs go flat? When weresect arch-aneurysms in adults, we typically go to nasopha-ryngeal temperatures of 14 to 16° and try not to go too muchlower.

Second, I noticed you cannulated the right atrium to drainthe venous side. Have you given any consideration to sump-ing the left side of the heart? Often LV distention is a limitingfactor in resuscitating these patients while you are trying torewarm.

Third, a lot of us have abandoned circulatory arrest forarch-aneurysm resection and are now doing low-flow retro-grade cerebral perfusion. Would that be a consideration inthese patients?

And then finally, many of the circulatory arrest patients,after their aneurysmectomy, have grossly normal neurologicexams. However, if you do neuropsychological testing, youoften find significant deficits. I recognize you cannot see thatin the animals, but the presence of a normal neural exam doesnot always mean that they are unaffected. Thank you.

Dr. Peter Rhee(closing): I would like to first start with aclarification of the model. We created an aortic laceration bypulling on previously placed aortic wires so it bled freely inthe chest causing uncontrolled hemorrhage. Then we cut theaorta open at the site of the laceration and placed primedcatheters into the aorta through this site.

This procedure in the beginning was difficult, but we foundthat putting the Foley catheter down against the distal aortacollapsed it so it slid into the opening easily. You do not haveto visualize the opening to do it this way. The actual tech-nique takes approximately 10 to 15 seconds to place bothcatheters in the aorta.

As Dr. Gentilello and others have shown, spontaneoushypothermia rewarming is good and there are level I warmersin every hospital. We envision a refrigerator-like portableunit with a roller pump in every hospital in order to cool incertain situations.

There are situations when we know that we cannot controlthe bleeding and we wish that we could just cool the patientor do something else to give us additional time.

I envision that the first population what we would try thison are those who are going to die from uncontrolled exsan-guination who have very low survival rates. This is why we

chose to do this with the emergency department thoracotomyincision.

In the clinical setting, I would open the chest in penetratingtrauma and if there were something that I could do to controlthe hemorrhage, I would do so and not induce hypothermia.But if the injury is below the diaphragm, we know the successrate for those patients is very low. So this type of patient maybe the one I would go ahead and arrest metabolism with theinduction of hypothermia. Dr. Gentilello, I hope that answersyour first question.

This was an uncontrolled hemorrhage model as we pulleda wire through the aorta and let the animal bleed out freely,resulting in hypotension. In a clinical setting, if a person haspenetrating thoracic trauma, no palpable pressure, and theydo not respond to resuscitation, this is an indication foremergency department thoracotomy and it is in this situationthat you may consider this procedure.

As far as checking the extent of coagulopathy during therewarming phase, we did not have the sophistication to do allthe blood banking type of things. We were only able to getnumbers on physiology and chemistries. We bought freshwhole blood from a farm and used it at the end of theprocedure. We also used some of the animals’ own blood atthe end of the procedure to bring them back during resusci-tation.

In the earlier phase, we saw that there were some coagu-lation problems, but we started to use heparin at the begin-ning of the experiment and once you reversed the heparin,this eliminated the bleeding problems.

Lactate and base deficits were measured but were notmarkedly abnormal due to the short time period before thestart of the procedure. If we had a hemorrhagic shock modelthat was prolonged such as up to 30 minutes with a slowerbleed; the lactate level may build up. But we wanted todevelop an exsanguinating hemorrhage model similar to thetype that might require a emergency department thoracotomy.

In the next set of animals, I think should be where weinduce an injury in the abdomen, for example, the liver, andlet it bleed out slowly (over 30 minutes) so the lactate andbase deficit does build up. We do have preoperative values inthe manuscript and they were relatively normal before westarted the procedure because the exsanguination was quick.

Regarding Dr. Walls’ comments on neuropsychologicaltesting, we also thought this was an important aspect toconsider. This was why we reported the neurologic functionscores in addition to the sophisticated microscopic stainingssuch as the immunohistochemistries in various aspects of thebrain. This demonstrated that one animal with normal neu-rologic scores had brain injury. We currently are planning toperform memory testing on the next set of experiments in anattempt to incorporate some neuropsychological testing. TheEEGs in this model showed that they became isoelectricwhen the blood pressure fell below 50 mm Hg and remainedthis way until the animal was rewarmed. The LV distentionduring rewarming occurred occasionally, and this was treatedby venting the left atrium.

I would also like to make a point here that this concept iscompletely different than spontaneous hypothermia, which is

Vol. 48, No. 3 Induced Hypothermia

449

a result of ongoing ischemia. This causes the metabolism tobe below the minimal basal rate, and this is usually why weget cold spontaneously. These people with temperatures lessthan 32°C following trauma have a mortality rate that ap-proaches 100%. It has been demonstrated in these patientsthat resuscitation with active warming is probably better thandoing passive rewarming. This is different than inducinghypothermia to prevent ischemic tissue necrosis in a nonhy-pothermic patient. We are taking animals that are not severelyshocked but have been exsanguinated and driving their tem-perature straight down so that their metabolism is minimal.

When we warm these animals, we find that there is anoxygen debt that is dependent on flow. The lactate levelsare elevated and it takes a while (1 hour after rewarming)before the animal can recover and be separated from theCPB.

I believe that I have addressed most of the questionsregarding when I put the cannulas in. We do this after theinitial uncontrolled hemorrhage has taken place, not beforethe procedure.

I would like to thank the members of the AAST for theprivilege of the floor.

The Journal of Trauma: Injury, Infection, and Critical Care March 2000

450